Citation
Sleep apnea activity and its concomitants in a subclinical population

Material Information

Title:
Sleep apnea activity and its concomitants in a subclinical population
Creator:
Berry, David Thomas Reed, 1958-
Publication Date:
Language:
English
Physical Description:
viii, 132 leaves : ill. ; 29 cm.

Subjects

Subjects / Keywords:
Apnea ( jstor )
Chemical desaturation ( jstor )
Hypoxia ( jstor )
Index numbers ( jstor )
Memory ( jstor )
Oxygen ( jstor )
Rapid eye movement sleep ( jstor )
Sleep ( jstor )
Sleep apnea syndromes ( jstor )
Snoring ( jstor )
Apnea ( mesh )
Clinical and Health Psychology thesis Ph.D ( mesh )
Dissertations, Academic -- Clinical and Health Psychology -- UF ( mesh )
Respiration ( mesh )
Sleep Apnea Syndromes ( mesh )
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph.D.)--University of Florida, 1985.
Bibliography:
Bibliography: leaves 123-131.
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by David Thomas Reed Berry.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright [name of dissertation author]. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
022857682 ( ALEPH )
16877989 ( OCLC )
ACU4129 ( NOTIS )

Downloads

This item has the following downloads:

EOGSOKC06_8RP790.xml

AA00011824_00001.pdf

AA00011824_00001_0136.txt

sleepapneaactivi00berr_0066.txt

sleepapneaactivi00berr_0053.txt

AA00011824_00001_0070.txt

AA00011824_00001_0084.txt

AA00011824_00001_0048.txt

AA00011824_00001_0112.txt

sleepapneaactivi00berr_0099.txt

sleepapneaactivi00berr_0063.txt

sleepapneaactivi00berr_0001.txt

sleepapneaactivi00berr_0056.txt

sleepapneaactivi00berr_0048.txt

AA00011824_00001_0049.txt

AA00011824_00001_0120.txt

sleepapneaactivi00berr_0026.txt

sleepapneaactivi00berr_0082.txt

sleepapneaactivi00berr_0079.txt

AA00011824_00001_0010.txt

sleepapneaactivi00berr_0138.txt

AA00011824_00001_0043.txt

AA00011824_00001_0121.txt

AA00011824_00001_0005.txt

sleepapneaactivi00berr_0054.txt

AA00011824_00001_0004.txt

AA00011824_00001_0111.txt

AA00011824_00001_0079.txt

sleepapneaactivi00berr_0036.txt

sleepapneaactivi00berr_0007.txt

AA00011824_00001_0017.txt

sleepapneaactivi00berr_0055.txt

sleepapneaactivi00berr_0059.txt

sleepapneaactivi00berr_0084.txt

AA00011824_00001_0137.txt

sleepapneaactivi00berr_0019.txt

sleepapneaactivi00berr_0028.txt

AA00011824_00001_0114.txt

AA00011824_00001_0067.txt

AA00011824_00001_0098.txt

sleepapneaactivi00berr_0072.txt

AA00011824_00001_0059.txt

AA00011824_00001_0009.txt

AA00011824_00001_0092.txt

sleepapneaactivi00berr_0049.txt

AA00011824_00001_0051.txt

sleepapneaactivi00berr_0087.txt

AA00011824_00001_0023.txt

AA00011824_00001_0138.txt

AA00011824_00001_0041.txt

AA00011824_00001_0125.txt

AA00011824_00001_0045.txt

sleepapneaactivi00berr_0094.txt

AA00011824_00001_0034.txt

sleepapneaactivi00berr_0034.txt

AA00011824_00001_0061.txt

AA00011824_00001_0087.txt

sleepapneaactivi00berr_0021.txt

sleepapneaactivi00berr_0083.txt

sleepapneaactivi00berr_0009.txt

sleepapneaactivi00berr_0110.txt

AA00011824_00001_0093.txt

sleepapneaactivi00berr_0085.txt

sleepapneaactivi00berr_0016.txt

AA00011824_00001_0116.txt

sleepapneaactivi00berr_0033.txt

sleepapneaactivi00berr_0043.txt

AA00011824_00001_0026.txt

EOGSOKC06_8RP790_xml.txt

sleepapneaactivi00berr_0112.txt

AA00011824_00001_0069.txt

sleepapneaactivi00berr_0111.txt

sleepapneaactivi00berr_0080.txt

sleepapneaactivi00berr_0042.txt

sleepapneaactivi00berr_0032.txt

AA00011824_00001_0107.txt

AA00011824_00001_0108.txt

AA00011824_00001_0027.txt

AA00011824_00001_0071.txt

AA00011824_00001_0060.txt

AA00011824_00001_0097.txt

AA00011824_00001_0105.txt

AA00011824_00001_0143.txt

AA00011824_00001_0083.txt

AA00011824_00001_0011.txt

AA00011824_00001_0025.txt

sleepapneaactivi00berr_0051.txt

AA00011824_00001_0100.txt

AA00011824_00001_0014.txt

sleepapneaactivi00berr_0046.txt

AA00011824_00001_0052.txt

AA00011824_00001_0082.txt

sleepapneaactivi00berr_0014.txt

sleepapneaactivi00berr_0100.txt

AA00011824_00001_0110.txt

sleepapneaactivi00berr_0020.txt

AA00011824_00001_0024.txt

sleepapneaactivi00berr_0005.txt

sleepapneaactivi00berr_0040.txt

sleepapneaactivi00berr_0064.txt

AA00011824_00001_0018.txt

AA00011824_00001_0038.txt

sleepapneaactivi00berr_0025.txt

AA00011824_00001_0007.txt

sleepapneaactivi00berr_0124.txt

sleepapneaactivi00berr_0052.txt

sleepapneaactivi00berr_0050.txt

AA00011824_00001_0019.txt

AA00011824_00001_0008.txt

AA00011824_00001_0054.txt

AA00011824_00001_0037.txt

sleepapneaactivi00berr_0075.txt

AA00011824_00001_0141.txt

AA00011824_00001_0057.txt

AA00011824_00001_0089.txt

sleepapneaactivi00berr_0044.txt

sleepapneaactivi00berr_0002.txt

AA00011824_00001_0086.txt

sleepapneaactivi00berr_0114.txt

sleepapneaactivi00berr_0096.txt

sleepapneaactivi00berr_0132.txt

AA00011824_00001_0095.txt

AA00011824_00001_0036.txt

AA00011824_00001_0124.txt

AA00011824_00001_0134.txt

sleepapneaactivi00berr_0088.txt

AA00011824_00001_0016.txt

AA00011824_00001_0094.txt

AA00011824_00001_0020.txt

sleepapneaactivi00berr_0134.txt

sleepapneaactivi00berr_0120.txt

AA00011824_00001_0123.txt

sleepapneaactivi00berr_0141.txt

AA00011824_00001_0090.txt

sleepapneaactivi00berr_0135.txt

sleepapneaactivi00berr_0058.txt

sleepapneaactivi00berr_0041.txt

sleepapneaactivi00berr_0125.txt

sleepapneaactivi00berr_0115.txt

AA00011824_00001_0142.txt

sleepapneaactivi00berr_0011.txt

sleepapneaactivi00berr_0035.txt

sleepapneaactivi00berr_0081.txt

AA00011824_00001_0129.txt

AA00011824_00001_0073.txt

AA00011824_00001_0139.txt

AA00011824_00001_0103.txt

sleepapneaactivi00berr_0131.txt

sleepapneaactivi00berr_0133.txt

AA00011824_00001_0029.txt

sleepapneaactivi00berr_0095.txt

AA00011824_00001_0033.txt

AA00011824_00001_0030.txt

sleepapneaactivi00berr_0126.txt

AA00011824_00001_0099.txt

AA00011824_00001_0074.txt

AA00011824_00001_0122.txt

AA00011824_00001_0047.txt

AA00011824_00001_0113.txt

AA00011824_00001_0096.txt

AA00011824_00001_0126.txt

AA00011824_00001_0101.txt

AA00011824_00001_0066.txt

sleepapneaactivi00berr_0071.txt

AA00011824_00001_0003.txt

AA00011824_00001_0040.txt

sleepapneaactivi00berr_0105.txt

sleepapneaactivi00berr_0139.txt

sleepapneaactivi00berr_0004.txt

sleepapneaactivi00berr_0128.txt

AA00011824_00001_0056.txt

sleepapneaactivi00berr_0091.txt

sleepapneaactivi00berr_0074.txt

AA00011824_00001_0035.txt

AA00011824_00001_0013.txt

sleepapneaactivi00berr_0106.txt

AA00011824_00001_0021.txt

AA00011824_00001_0128.txt

AA00011824_00001_0118.txt

sleepapneaactivi00berr_0065.txt

sleepapneaactivi00berr_0060.txt

sleepapneaactivi00berr_0102.txt

AA00011824_00001_0002.txt

sleepapneaactivi00berr_0057.txt

sleepapneaactivi00berr_0018.txt

sleepapneaactivi00berr_0037.txt

sleepapneaactivi00berr_0086.txt

sleepapneaactivi00berr_0101.txt

AA00011824_00001_0081.txt

sleepapneaactivi00berr_0093.txt

sleepapneaactivi00berr_0070.txt

sleepapneaactivi00berr_0116.txt

AA00011824_00001_0109.txt

sleepapneaactivi00berr_0122.txt

AA00011824_00001_0088.txt

AA00011824_00001_0106.txt

sleepapneaactivi00berr_0109.txt

AA00011824_00001_pdf.txt

AA00011824_00001_0078.txt

AA00011824_00001_0053.txt

sleepapneaactivi00berr_0047.txt

sleepapneaactivi00berr_0031.txt

sleepapneaactivi00berr_0098.txt

sleepapneaactivi00berr_0103.txt

AA00011824_00001_0091.txt

sleepapneaactivi00berr_0000.txt

AA00011824_00001_0032.txt

AA00011824_00001_0132.txt

AA00011824_00001_0055.txt

AA00011824_00001_0133.txt

AA00011824_00001_0063.txt

AA00011824_00001_0050.txt

sleepapneaactivi00berr_0118.txt

sleepapneaactivi00berr_0003.txt

sleepapneaactivi00berr_0140.txt

sleepapneaactivi00berr_0029.txt

sleepapneaactivi00berr_0092.txt

AA00011824_00001_0001.txt

sleepapneaactivi00berr_0108.txt

AA00011824_00001_0130.txt

sleepapneaactivi00berr_0107.txt

sleepapneaactivi00berr_0022.txt

sleepapneaactivi00berr_0030.txt

AA00011824_00001_0042.txt

AA00011824_00001_0064.txt

sleepapneaactivi00berr_0077.txt

AA00011824_00001_0028.txt

sleepapneaactivi00berr_0129.txt

AA00011824_00001_0075.txt

sleepapneaactivi00berr_0010.txt

sleepapneaactivi00berr_0062.txt

AA00011824_00001_0072.txt

sleepapneaactivi00berr_0113.txt

sleepapneaactivi00berr_0097.txt

AA00011824_00001_0119.txt

AA00011824_00001_0104.txt

sleepapneaactivi00berr_0090.txt

sleepapneaactivi00berr_0073.txt

sleepapneaactivi00berr_0069.txt

sleepapneaactivi00berr_0012.txt

sleepapneaactivi00berr_0061.txt

AA00011824_00001_0140.txt

sleepapneaactivi00berr_0045.txt

AA00011824_00001_0046.txt

AA00011824_00001_0012.txt

sleepapneaactivi00berr_0006.txt

AA00011824_00001_0062.txt

AA00011824_00001_0022.txt

AA00011824_00001_0080.txt

sleepapneaactivi00berr_0008.txt

sleepapneaactivi00berr_0127.txt

sleepapneaactivi00berr_0027.txt

AA00011824_00001_0135.txt

AA00011824_00001_0039.txt

sleepapneaactivi00berr_0089.txt

sleepapneaactivi00berr_0024.txt

AA00011824_00001_0115.txt

sleepapneaactivi00berr_0142.txt

sleepapneaactivi00berr_0104.txt

AA00011824_00001_0006.txt

AA00011824_00001_0015.txt

AA00011824_00001_0085.txt

sleepapneaactivi00berr_0123.txt

sleepapneaactivi00berr_0136.txt

AA00011824_00001_0117.txt

sleepapneaactivi00berr_0067.txt

sleepapneaactivi00berr_0119.txt

AA00011824_00001_0044.txt

sleepapneaactivi00berr_0130.txt

AA00011824_00001_0076.txt

sleepapneaactivi00berr_0039.txt

sleepapneaactivi00berr_0068.txt

sleepapneaactivi00berr_0117.txt

AA00011824_00001_0031.txt

AA00011824_00001_0131.txt

AA00011824_00001_0127.txt

sleepapneaactivi00berr_0121.txt

sleepapneaactivi00berr_0078.txt

AA00011824_00001_0077.txt

sleepapneaactivi00berr_0137.txt

AA00011824_00001_0065.txt

AA00011824_00001_0102.txt

sleepapneaactivi00berr_0017.txt

AA00011824_00001_0068.txt

sleepapneaactivi00berr_0023.txt

sleepapneaactivi00berr_0076.txt

sleepapneaactivi00berr_0038.txt

AA00011824_00001_0058.txt

sleepapneaactivi00berr_0013.txt

sleepapneaactivi00berr_0015.txt


Full Text


















SLEEP APNEA ACTIVITY AND ITS CONCOMITANTS
IN A SUBCLINICAL POPULATION




By

DAVID THOMAS REED BERRY





















A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN
PARTIAL FULFILLMENT OF THE REQUIREMENTS
FOR THE DEGREE OF DOCTOR OF PHILOSOPHY


UNIVERSITY OF FLCOIDA

1985
















ACKEOWLEDGCI I:TS

The author would like to thank his chairman, Dr. Wilse

B. V.bt, for his moral and material support, as well as

guidance through the completion of the project. Also due

thanks are the other members of the supervisory committee,

Dr. A. J. Block, Dr. Russell M. Eauer, Dr. F. D. McGlynn,

ann Dr. Rudy Vuchinich, for their support, advice, and

contributions. Several people provided technical assistance

and are due a vote of thanks: Dan Switzer, Alex Kasterton,

Karla Smith, Alton Howard, Albert Briggs, and Edna Pearson.

The author wishes to acknowledge the moral and

financial support of his parents, Mr. and Mrs. John G.

Berry, without which the project would never have been

undertaken. Also of considerable help were the

encouragement and advice of C. i. Carswell, D. Haymaker, anc

S. Erody, all of whom deserve thanks. Cynthia Zimmerman

provided expert typing assistance for the manuscript.















TALE OF CONTENTS


Page

ACKNOWLEDG ENTS ....... .............. ..................... ii

LIST OF TABLES ........................................... v

ABSTRACT ...............................................vii

CHAPTER

C!E IITRODUCTION ..................................... 1

Definitions and Terminology......................2
History and Characteristics of Sleep Apnea
Syndrome ......................................
Incidence of Sleep Apnea Activity.................5
Temporal Characteristics of Apneas...............13
Variables Associated With Sleep Apnea
Syndrome and Activity.........................17
Deficits Associated With Sleep Apnea
Syndrome.....................................25
Cardiopulmonary Deficits.....................26
Arousal Deficits (Hypersomnolence)..........28
Oxygen-Deseturation and Cognitive Deficits..30
Measurement of Deficits Found in Sleep Apnea
Syndrome.......................................34
Measurement of Cardiopulmonary and Health
Deficits .................................. 34
Measurement of Arousal Deficits
(Hypersomnolence)........................35
Measurement of Hypoxia and Its Sequelae.....38
Statement of the Problem ........................46

TWO METHOD...................................... ... 48

Subjects ........................................48
Apparatus.......................................49
Measures........................................50
Procedure .. ................... ............... 54
Statistical Procedures..........................55

THREE RESULTS................. .... ................... 56

Demographics and Incidence of Nocturnal
Respiratory Disorder..........................56


iii









Nocturnal Respiratory and Health Variables......63
Nocturnal Respiratory and Sleep/Wake Variables..65
EEG Data.................................... 65
Daytime Sleepiness Data.....................69
Subjective Sleep Assessment.................70
Nocturnal Respiratory and Neuropsychological
Variables..................................... 74

FOUR DISCUSSION.....................................86

Demographics and Incidence of Ilocturnal
Respiratory Disorder..........................86
nocturnal Respiratory and Health Variables......90
l!octurnal Respiratory and Sleep/Wake Variables..92
EEG Data ....................................92
Daytime Sleepiness, Sleep Questionnaire,
and Sleep Log Data........................94
Nocturnal Respiratory and Neuropsychological
Variables.....................................96

FIVE SUMIIARY AND CONCLUSIONS........................103

APPENDIX

A SCORING PROTOCOL FOR RESPIRATORY VARIABLES..... 106

B RAk DATA TABLES................................109

REFERErCES...............................................123

EIOGRAPHICAL SKETCH.....................................132















LIST OF TABLES

Table Page

2-1 Interrater agreement for respiratory events
of 46 snoring males..............................51

3-1 Demographic variables from 46 snoring males.....56

3-2 Incidence of low and high levels of apnea/
hypopnea by age in 46 snoring males.............58

3-3 Significant (p<.05) Pearson correlations
between nocturnal respirEtory variables
and demographic variables in 46 snoring
males............................................ .

3-4 Means and standard deviations for selected
demographic and nocturnal respiratory
variables in 46 snoring males grouped by
level oi apnea/hypopnea.........................61

3-5 Means and standard deviations for health
related variables in 46 snoring males grouped
by level of apnea/hypopnea......................65

3-6 Means and standard deviations of selected
sleep variables from 43 snoring males............66

5-7 Meens end standard deviations of selected
sleep variables from 43 snoring males
groupec by level of apnea/hypopnea..............68

3-8 Leans and standard deviations for daytime
sleepiness variables in 46 snoring males
grouped by level of apnea/hypopnea..............69

3-9 Significant (p<.05) Pearson correlations
between nocturnal respiratory and sleep
questionnaire variables in 46 snoring males.....71

3-10 Means and standard deviations of sleep
questionnaire variables from 46 snoring
males groupec by level of apnea/hypopnea........71









3-11 Significant (p<.05) Pearson correlations
between nocturnal respiratory and sleep log
variables in 46 snoring males.....................75

3-12 Means and standard deviations of sleep log
variables in 46 snoring males grouped by
level of apnea/hypopnea......................... 73

3-13 Multivariate regression of demographic and
nocturnal respiratory variables on cognitive
scores in 46 snoring males......................76

3-14 Significant (p<.05) partial correlations,
controlling for age, between nocturnal
respiratory and cognitive variables in
46 snoring males................................ 79

3-15 Means and standard deviations of cognitive
scores in 46 snoring males grouped by
level of apnea/hypopnea.........................80

3-16 Multivariate regression of sleep and
nocturnal respiratory scores on cognitive
variables in 43 snoring males...................85

B-1 Demographic and medical variables from 46
snoring males...................................109

B-2 Daytime sleepiness variables from 46 snoring
males........................................... 111

B-3 Sleep questionnaire variables from 46 snoring
males...........................................113

8-4 Seven day sleep log means from 46 snoring
males ..........................................115

B-5 Neuropsychological variables from 46 snoring
males........................................... 117

B-6 Respiratory variables from 46 snoring males.....119

B-7 Sleep variables from 46 snoring males...........121















Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

SLEEP APNEA ACTIVITY AND ITS CONCOMITANTS
IN A SUBCLINICAL POPULATION

By

DAVID THOMAS REED BERRY

August 1985

Chairman: Wilse B. Webb
Major Department: Clinical Psychology

Sleep Apnea Syndrome (SAS) is a nocturnal respiratory

disorder with serious consequences. Patients with SAS

experience apneas (pauses in breathing of 10 or more

seconds) and hypopneas (declines in the amplitude of

breathing accompanied by oxygen desaturation) while

asleep. Recently, it has become apparent that apneas and

hypopneas may occur in otherwise apparently normal

subjects. A sample selected to display a wide range of

sleep disordered breathing (heavy snoring males in good

health and over 30 years of age) received detailed

measurement of their nocturnal respiration, sleep,

cognitive skills, and health in an attempt to clarify the

impact of apneas and hypopneas in a subclinical population.

Forty-six heavy snoring males in good health with a

mean age of 50 years and a mean weight of 190 pounds

comprised the sample. During an experimental evening they

vii









received testing and filled out various questionnaires

before sleeping overnight with their sleep and breathing

continuously recorded. Sixty-two per cent of these subjects

experienced at least one event, while 13% had more than 5

events per hour of sleep. Most events occurred in light

slow wave or REM sleep and were linked to obesity. Subjects

with high levels of apnea/hypopnea (more than 5 per hour)

experienced significantly exacerbated oxygen desaturation

relative to the remaining subjects. Oxygen desaturation was

linked to higher blood pressure readings. A typical "first

night" effect on the sleep of the subjects in the lab was

noted, resulting in a lighter sleep than usual and obscuring

possible relationships between sleep and respiratory events,

which were not observed. Deteriorating overnight

respiratory indices were associated with increased

sleepiness and napping. High levels of apnea/hypopnea were

associated with declining scores on tests measuring non-

verbal intelligence, verbal and non-verbal memory,

expressive verbal fluency, and cognitive flexibility. It is

speculated that hypoxia induced changes in cerebral

cholinergic synthetic pathways underlie these changes.

It is concluded that a significant subgroup of heavy

snoring males who experience multiple apnea/hypopneas are at

increased risk of significant oxygen desaturation, daytime

sleepiness, and cognitive changes. This group may fall on a

continuum with sleep apnea syndrome patients.


viii















CHAPTER ONE
INTRODUCTION

Sleep apneas are respiratory pauses during sleep.

According to the Association of Sleep Disorders Centers

Nosology of Sleep Disorders (ASDC, 1979), a sleep apnea

syndrome (SAS) is a potentially lethal condition charac-

terized by multiple apneas, excessive daytime sleepiness,

alterations of consciousness, and cardiopulmonary

complications. The apnea episodes are thought to be the

causal factor behind the other deficits associated with

SAS (Guilleminault et al., 1978).

Polysomnographic recordings in various populations

have revealed the occurrence of sleep apneas in the

general population, and particularly in aging popula-

tions. Estimates of the prevalence of at least a few

sleep apneas in older subjects have ranged as high as 75%

(Carskadon and Dement, 1981b). Carskadon et al. (1980)

have suggested that sleep apneas in older subjects may be

implicated in insomnia, cardiac ailments, and even

senility.

However, certain reports have documented the presence

of multiple sleep apnea episodes in subjects without

apparent pathology (Crr et al., 1979). This evidence

obscures the role of sleep apneas in the other pathologies









noted in SAS patients. The link between apneas and patho-

logy may be more complex than was originally thought,

given its apparently benign occurrence in otherwise

asymptomatic subjects.

The present study will examine the correlates of

sleep apnea episodes in a population selected to display a

wide range of such events. The relationship between sleep

apnea indices and sleep characteristics, age, weight and a

variety of deficit measurements will be analyzed.

A review of the current data available on various

aspects of sleep apneas and sleep apnea syndromes will be

presented as background for the study. The diagnosis,

history, incidence, and characteristics of sleep apneas

and sleep apnea syndrome will be reviewed, followed by a

discussion of variables associated with sleep apneas.

Lastly, the deficits commonly found in sleep apnea syn-

drome will be detailed, along with measurements thought to

be sensitive to these deficits.

Definitions and Terminology

Sleep apneas have been defined as cessations of air-

flow at the mouth and nose which last for 10 or more

seconds (Guilleminault et al., 1976). In a sleep apnea

syndrome (SAS), multiple sleep apneas are accompanied by

snoring, excessive daytime sleepiness, cardiopulmonary

complications, and altered states of consciousness. The

ease of quantification of sleep apneas has led to their

use as an important tool in diagnosing SAS. Guilleminault









et al. (1976) first suggested that the occurrence of 30 or

more apneas in 7 or more hours of sleep was diagnostic of

SAS, while Guilleminault et al. (1978) proposed the use of

an apnea index (AI; number of apneas/number of hours

sleep) in excess of 5 as a cut-off score to diagnose the

syndrome. Later, Carskadon et al. (1980) combined apneas

and hypopneas (a reduction in airflow of 50% or more at

the mouth and nose) to form an apnea + hypopnea index

(AHI) with 5 utilized as a diagnostic level of sleep apnea

syndrome. As can be appreciated from the preceding dis-

cussion, the diagnostic criteria for SAS have undergone a

rapid evolution, with remarkably little validation work

published.

History and Characteristics of Sleep Apnea Syndrome

Sleep Apnea Syndrome falls within the larger category

of Sleep Disordered Breathing (Block, 1980). Respiratory

disorders within this category involve changes in

breathing patterns during sleep. Syndromes described

within this category include Pickwickian Syndrome; Sleep

Apnea Syndrome; Chronic Obstructive Lung Disease; and

syndromes of Primary and Secondary Alveolar

Hypoventilation (Block, 1980; Guilleminault et al.,

1976).

The first disorder described in this group was the

Pickwickian Syndrome. These patients were initially

identified by a chronic daytime hypersomnolence, obesity,

Cheyne-Stokes breathing, hypercapnia and hypoxemia. This









syndrome typically results in complications such as right

sided congestive heart failure, pulmonary hypertension,

and peripheral edema (Block, 1980). An exacerbation of

disordered breathing (apnea) as well as hypercapnia and

hypoxemia (often severe) occurs with sleep onset. Treat-

ment consists of oxygen therapy, respiratory stimulants,

and maintenance of wakefulness.

The discovery of a sleep induced exacerbation of the

symptoms accompanying Pickwickian Syndrome spurred a

systematic investigation of sleep respiratory events

(Gastaut et al., 1965). The new emphasis on noctural

respiration led to a series of investigations which docu-

mented sleep induced respiratory dysrhythmias in non-obese

subjects (Lugarasi et al., 1966). A new group of sleep

disordered breathers designated as Sleep Apnea Syndrome

emerged, with several features which distinguished them

from Pickwickians. Among these differences were weight

(i.e., SAS patients were not necessarily obese), daytime

respiratory control (i.e., SAS patients were typically not

hypercapnic during waking as were Pickwickians), sex

(i.e., while AO; of Pickwickians were female, very few

premenapausal females have been identified with SAS), and

age (i.e, SAS patients are typically younger than

Pickwickians)(Block, 1980).

Two large samples of SAS patients have been reported

in the literature, allowing characterization of typical

SAS patients. Guilleminault et al. (1976) reported on a









series of 50 patients diagnosed with SAS over a period of

six years. Age ranged from 28-62 with a mean age of

45.5. The sample consisted of 48 males and 2 females.

Thirty-nine were referred on the basis of excessive day-

time sleepiness, while 7 were referred for loud snoring

and abnormal movements reported by their spouses. Sixty

percent of the patients were more than 15% above ideal

weight. Other symptoms included hypnagogic halluci-

nations, automatic behaviors, intellectual deterioration,

personality changes, impotence, morning headaches, and

hypertension. Weitzman, Kahn and Pollak (1980) report

data on a group of 10 SAS patients seen in their clinic.

These subjects displayed the following characteristics:

age between 38 and 47, obesity, male sex, nicotine depen-

dence, hypertension, complaints of excessive daytime

sleepiness, and the presence of serious cardiac arrhyth-

mias. (This group was selected from a larger sample of 38

patients.) Thus, pathological symptomatology associated

with a diagnosed SAS include excessive daytime somnolence,

cardiopulmonary complications and cognitive/intellectual

changes; a "typical" SAS patient is a middle aged male

who snores and is obese.

Incidence of Sleep Apnea Activity

The incidence of sleep apnea activity and sleep apnea

syndrome have been frequently confuse. While sleep

apneas have been an important diagnostic tool for identi-

fying sleep apnea syndrome, they have not been









definitively demonstrated to be the necessary and suffi-

cient cause of the syndrome. Thus, for the purposes of

the following review, the focus will be on occurrence and

level of sleep apnea activity in both sleep apnea syndrome

subjects and other populations. The full range of sleep

apnea activity will be reflected by noting the presence of

sleep apnea activity, and an indication of the quantita-

tive level of the activity will be presented in the apnea

index (AI) or the apnea + hypopnea index (AHI). Recall

that an AI of 5 has been frequently utilized as a cut-off

score for identifying SAS patients.

A review of the literature suggests that at least

three sampling strategies have been used for studying the

incidence of sleep epnea activity. Subjects have been

selected on the basis of one of the following three

criteria: 1) absence of sleep or respiratory complaints,

2) presence of sleep or respiratory complaints, or 3)

sampling without regard to these variables.

Because diagnosis of SAS is often dependent on number

of sleep apneas (or level of SAA), most researchers report

number of apneas observed, although some report only on

number of subjects who exceed one of the criteria for

SAS. A review of studies with samples of 20 or greater

will be presented with careful note of the sampling

strategy utilized, as well as the dependent measure of

apnea.









One sampling strategy which has been pursued is to

exclude subjects who complain of sleep or respiratory

difficulties. Screening for sleep complaints varies

widely. Guilleminault et al. (1978) studied a sample of

20 normalss" (presumably noncomplaining subjects). These

subjects ranged in age from 40-60 years. Males showed a

mean number of apneas of 7 (range 1-25), while females

showed a mean number of apneas of 2 (range 0-5). Thus,

none of these noncomplaining subjects presented a clinical

syndrome of sleep apnea, although sleep apneas did occur

apparently without pathology. Block et al. (1979) studied

49 subjects (M = 30, F = 19). Males in the sample had a

mean age of 38, while females had a mean age of 29. Any

potential subjects who complained of breathing difficul-

ties or sleep disturbances were excluded from the study.

Twelve males (40%) had episodes of sleep apnea while only

3 females (15%) suffered as well. For those subjects,

males had a mean of 4.2 episodes, while females had a mean

of 3. Within this sample, the number of apneas was not

significantly correlated with increasing age, although

occurrence of oxygen desaturation (a concomittant of

disordered breathing) was positively correlated with

age. Additionally, although the mean number of sleep

apneas was not significantly different between sexes (M =

4.2, F = 3), the more sensitive measure of apneas with

desaturation was (M = 3, F = 0). Webb (1974) studied 2C

males with a mean age of 44. Subjects were excluded from









the study if they reported any "serious" sleep com-

plaint. Two of the subjects reported consuming hypnotics

for sleep induction, indicating that the criterion used in

this study was less stringent than that of Block et al.

(1979) report. Nine of the subjects (45%) suffered at

least one episode of sleep apnea with a mean number of

episodes of 2.5. All of the subjects with episodes of

sleep apnea were older than the mean age of the sample,

leading Webb to conclude that apnea is related to age.

Bixler et al. (1982) constructed an age stratified patient

representative sample of 100 subjects (M = 41, F = 59).

In perhaps the most rigorous screening for sleep com-

plaints in this group of studies, Bixler et al. eliminated

any subject with a sleep complaint or chronic medication

usage. For the entire sample, 6 males (14.6%) and 5

females (10.25) had at least one episode of sleep apnea.

For subjects under 30, 3.3%; for those between 30 and 50,

12.8e; and those 50-74, 19.4; suffered episodes of sleep

apnea. For the subsample over 60, 25% had sleep apnea

activity with 8.5 episodes for males and 10.5 episodes for

females. In this sample, there was a nonsignificant trend

toward more apnea activity in males and a significant

positive correlation between age and sleep apnea.

block et al. (1980) report data on 20 post menapausal

women with a mean age of 59. Eight women (40%) had epi-

sodes of sleep apnea, with a mean number of episodes of









5. There was a significant correlation between apnea and

age.

For noncomplaining subjects then, males suffer a

higher incidence of apnea activity in virtually every

study which compares sex effects. For males, incidences

range widely from a low of 14.6% (Bixler et al., 1980) to

a high of 45% (Webb, 1974); while females range from a

low of 10.2% (Bixler et al., 1982) to a high of 40% (Block

et al., 1980). These differences are probably due at

least in part to age differences in male samples, as well

as age differences in females (only as they reflect mena-

pausal status). Reported number of apneas range from 2.5

(Webb, 1974) to 8.5 (Bixler et al., 1982) in males and

from 2 (Guilleminault et al., 1978) to 10.5 (Bixler et

al., 1982) in females. Many studies report a positive

correlation between number of apneas and increasing age.

A second sampling technique which has been utilized

has been studying subjects with sleep and breathing com-

plaints. Kales et al. (1982) studied 200 subjects with

primary complaints of insomnia. There were 82 males and

118 females in the sample with a mean age of 42.3. Sleep

apnea activity was noted in 10.5% of this group (M =

13.4%, F = 8.5%) with a mean number of 11.2 episodes for

males and 6.6 for females. Subjects with sleep apnea

activity were significantly older (46.8) than those with-

out (40.6). Ancoli-Israel et al. (1981) recruited 24

subjects whose answers to questionnaires provoked









suspicion of insomnia or nighttime breathing and muscular

events. Eleven males (7 = 72.5) and 13 females (R = 68.5)

were studied. All of these subjects, save one male and

one female, suffered at least one episode of sleep

apnea. For those subjects with at least one episode of

sleep apnea, males had a mean of 53 episodes of apnea,

while females had a mean of 32 episodes. In contrast to

the Kales et al. study, 6 males had a high level of sleep

apnea (AI>5), while 3 females met this criterion. A rela-

tionship between age and apnea was not reported in this

sample, possibly because of restricted range. These two

studies of sleep dist'r'bed subjects provide conflicting

data. That is, the insomniac alone sample did not demon-

strate an apnea incidence which was substantially

different from a noncomplaining sample (10.5%), while the

sample recruited for sleep induction, respiratory, and

muscular differences showed very discrepant apnea inci-

dences (nearly 100%). One potential explanation involves

age, as the Ancoli-Israel sample is substantially older

than Kales', although the aging samples have not reported

this level of apnea before. An alternative explanation is

that sampling insomniacs does not generate a high number

of apneas, whereas aging sleep and respiratory disturbed

subjects show a very high incidence of apnea. It would

seem then that the relevant "complaint" would involve

those reported by Ancoli-Israel et al. which are in

addition to those noted by Kales et al. These are (as








nearly as can be judged) questions surrounding

respiration.

The final category of incidence studies involves

those which do not consider sleep complaints in their

sampling. As in the other studies, these vary in their

generalizability. Carskadon and Dement (1981b) recruited

40 aging subjects, including 18 males (R = 72.7) and 22

females (R = 74). Their exclusion criteria are somewhat

problematic, involving excluding subjects who "spon-

taneously complain of sleep problems." Thus, it is

unclear what this is a representative sample of, specifi-

cally whether the distribution of those not spontaneously

complaining of sleep disturbances may be different from

those who actually do not suffer from sleep distur-

bances. These considerations aside, Carskadon and Dement

report that 16 males (88%) and 16 females (72%) demon-

strated at least one sleep apnea episode, while 8 males

(44%) and 7 females (31.8%) have high levels of sleep

apnea (AI>5). Males with apnea showed a mean number of

51, while females showed a mean number of 34. Sex

differences appear to have fallen out at this age, while

the restricted range of this extreme age sample does not

lend itself to a correlational analysis of age versus

apnea. In another study of limited generalizability,

Kreiss et al. (1982) randomly selected inpatients on a

Veterans Administration medical ward. These patients, all

presumably male, numbered 26. Further data reported were









sketchy, but 7 (27%) met the sleep apnea syndrome

criterion of >30 episodes of sleep apnea. The fact that

these subjects were patients on a medical ward obviously

restricts generalizability of these findings. Finally, in

the only truly random sampling reported, Ancoli-Israel et

al. (1982) present initial data on randomly selected sub-

jects. Forty subjects (sex not reported) with a mean age

of 71.5 were studied. Thirteen (32.5%) met the sleep

apnea syndrome criterion of 30 or more apneas. The sleep

apnea subgroup was older, 74.2, than the full sample.

The data collected from relatively random sampling

are primarily based on aging populations (>60). The data

from these studies suggest that between 32.5% (Ancoli-

Israel et al., 1981) and 37% (Carskadon and Dement, 1981b)

of aging subjects suffer high levels of sleep apnea

(AI>5). An even larger percentage, perhaps greater than

75%, suffer at least some episodes of sleep apnea. A

major deficiency in these randomly sampled studies is a

lack of data on younger subjects.

Thus data collected under different sampling proce-

dures yield variable incidence rates of sleep apnea and

AI>S. Subjects without sleep or respiratory complaints

have a very low incidence (perhaps 0) of AI>5, between 14%

and 455 of males surveyed had at least some apnea, while

females ranged between 10% and 40% experiencing some

apnea. Sampling subjects with sleep or respiratory com-

plaints suggested that subjects suffering respiratory








difficulties had very high incidences of apnea (almost

100% for both males and females) with increase of inci-

dence of AI>5 (55% for males, 23% for females). Sampling

without regard to sleep or respiratory difficulties is

largely confined to aging populations. Incidences of

apnea in these aging populations would seem to be rather

high (88% males, 72% females) while incidence of AI>5 seem

to be in the range of 32%-37%.

It is clear that different populations (as defined by

sampling strategies) suffer different incidences of sleep

apneas and AI>5. The clinician dealing with sleep dis-

orders would be wise to consider the acturial base rates

of apnea and AI>5 provided by these data. It is apparent

that sex, age, and respiratory status are significantly

related to the incidence rates of SAA and SAS. These

factors must be considered in any study of apnea and asso-

ciated deficits.

Temporal Characteristics of Apneas

The temporal characteristics of sleep apneas repre-

sent a potentially important variable. Above it was shown

that a wide variability in number of sleep apneas occurs,

and the arbitrary nature of the division between SAA and

SAS subjects was noted. Therefore, it is of interest to

determine whether the duration of sleep apneas provides a

more reliable distinction between the two groups. Another

variable of interest involves the placement of apneas

within sleep, as Guilleminault et al. (1976) noted that









normals might experience apneas during REM without

pathological consequence. Therefore, data available on

these two issues will be reviewed, including only studies

with a sample size great than 10 (with one exception).

Reports on normal subjects below 30 years of age are

somewhat rare, although some do exist. Bixler et al.

(1982) included a group aged 18-29 years (M = 13, F =

17). One subject in this group had sleep apneas (12) with

a mean duration of 15 seconds. Bixler et al. noted that

most apneas occurred in Stages 1 and REN sleep. Part of

the sample of Block et al. (1979), the females, had a mean

age of 29. Three of these subjects showed apneas with a

mean duration of 14 seconds. Again, a disproportionate

number of apneas occurred in Stages 1 and REM. Thus in

normal subjects under 30, apneas are somewhat rare, with a

mean duration of approximately 15 seconds when they do

occur. Most apneas occur in light and REM sleep.

More data are available on normal subjects between 30

and 60. Guilleminault et al. (1978) do not report on

duration of apneas in their sample aged 40-60, but they

note that apneas occurred only in Stages 1 and REM.

Bixler et al. (1982) reported that, in their 50-44 year

old group, 5 subjects experienced sleep apnea with a mean

duration of 13.1 seconds. A disproportionate number ol

apneas occurred in Stages 1 and REM. Block et al. (1979)

studied 30 males with a mean age of 38, with 12

experiencing sleep apneas. These apneas had a mean









duration of 20 seconds and occurred most frequently

(30/60) in Stages 1 and REM. Eight of 20 aging females,

in the report by Block et al. (1980), had sleep apneas

with a mean duration of 19 seconds. Sixty-two percent of

apnea episodes occurred in Stages 1, 2 or REM. Thus in

middle aged normal subjects, incidence of sleep apnea

appears to rise, with a mean duration of between 13 and 20

seconds, perhaps slightly longer than in younger groups.

Again, apneas occurred mostly in light (Stages 1 and 2)

and REM sleep.

Duration data on aging normals are limited.

Carskadon and Dement's (1981b) study of aging normals (M =

72, F = 74) does not report duration of apnea events.

Similar, Ancoli-Israel et al. (1982) do not provide this

information on their aged normals (R = 71.5). Bixler et

al. (1982) noted a high incidence of apneas in their sub-

jects over 60, but failed to report a mean duration for

this group. Given the sketchy details reported for this

group, it is difficult to draw any conclusions about

changes in temporal duration or placement of apneas in

aging normals.

To summarize the temporal data in normals, apneas

appear somewhat rarely in normals below 30, with a mean

duration of 15 seconds. Incidence of apneas increase

between 30 and 60, with a wider variation in mean dura-

tions reported (13-20 seconds). Apneas occur most

frequently in light and REM sleep in normals of all ages.









Temporal data on subjects diagnosed with sleep apnea

syndrome (>30 or AI>5) primarily include patients between

the ages of 30 and 60. The reason for the lack of older

SAS subjects is unclear, particularly in light of data

reported above, demonstrating very high percentages of SAS

in certain groups of aging subjects.

Garay et al. (1981) studied 13 sleep apnea syndrome

subjects. They divided these 12 males and 1 female into

two groups, based on their daytime resting CO2 levels.

Subjects who were eucapnic during the day had a mean apnea

duration of 16 seconds, while those who were hypercapnic

had a mean duration of 17 seconds. No data on distribu-

tion relative to sleep stages were reported.

Guilleminault et al. (1978) studied 50 predominantly male

(R age 48) sleep apnea syndrome patients. Apneas in this

population averaged 22 seconds in duration. Apneas were

longest in REM, and few occurred in Stages 3 or 4 sleep.

Fujita et al. (1981) studied 12 male SAS patients with a

mean age of 43. Apnea episodes averaged 23 seconds in

duration. Weitzman et al. (1980) studied 10 SAS males

whose age ranged from 38-57. The mean duration of apneas

in this group was 30 seconds. All apneas occurred in

light or REM sleep. Orr et al. (1979) studied 4 hyper-

somnolent males diagnosed with SAS. A control group of 4

nonhypersomnolent subjects was matched on number of ob-

structions per minute. Mean age was 57 for asymptomatic

and 42 for symptomatic subjects. Asymptomatic subjects








had apneas with a mean duration of 25.9 seconds, while

symptomatic subjects had a mean duration of 20.8 seconds.

Several characteristics emerge from the group of

reports on SAS subjects. There is wide variability in

duration of apneas, with reported means ranging from 16-30

seconds. The data available on placement of apneas in

sleep suggest that in SAS subjects most apneas occur in

light and REM sleep.

Comparison of data on normal and SAS subjects reveals

substantial overlap between mean duration of apneas in SAS

subjects (13-20 seconds) and SAS subjects (16-30

seconds). Particularly interesting was the Orr et al.

(1979) study which demonstrated a lower mean duration of

apneas for hypersomnolent subjects than for nonhypersomno-

lent subjects. Data on temporal placement of apneas show

most events occurring in light or REM sleep for both

normal and SAS subjects. Taken together, these data

suggest that substantial overlap exists on temporal

variables between normal and SAS subjects indicating a

limited utility for diagnostic purposes.

Variables Associated With Sleep Apnea
Syndrome and Activity

It was noted above that several variables have been

suggested as associated with increased incidences of sleep

apneas and SAS. These variables represent potential

"markers" of subpopulations which might experience higher

incidences of sleep apnea activity. These primary









associated variables include sex, age, snoring and

obesity. The data available on these variables and their

relationships to sleep apnea activity will be reviewed,

followed by a brief discussion of their relevance to

experimental methodology.

Regarding sex variables, as previously noted,

Guilleminault et al. (1978) found 96% of their SAS

patients to be males. In the Block et al. (1979) study of

nonsleep disturbed subjects, 40% of the males in the

sample had at least one episode of apnea, while only 15"

of the females so suffered. Bixler et al. (1982) noted a

trend toward more sleep apnea activity in males. Block et

al. (1980) studied 20 post menapausal females and found

apnea activity in 40% of them, leading them to suggest

that some factor associated with menapause (e.g., pro-

gesterone) might somehow "protect" premenapausal females

from apneas, and that this factor along with its benign

influence was lost with menapause. This explanation pre-

dicted the low level of apnea activity in younger females,

as well as the heightened apnea activity of aging

females. While the data of Carsakdon and Dement (1981b)

show aged females to be close to aged males in percentage

exhibiting sleep apnea activity, a report by Smallwood et

al. (1983) found no evidence of sleep apnea activity in

their 6 elderly (post menapausal) females, while elderly

males showed significant levels of apnea activity. The

limitations of the sample size in the Smallwood et al.








report might explain this discrepancy as representing a

sampling bias. Taken together, this evidence strongly

supports the notion that in young and middle aged popula-

tions, sleep apnea is a male phenomenon. Additionally,

evidence points toward a closing of the gap between

females and age matched males as the former group passes

menapause.

A second factor which has been implicated in sleep

apnea activity is age. Webb (1974) found that his sub-

jects exhibiting sleep apnea activity were significantly

older than his full sample. This was also the case for

the subjects of Kales et al. (1982) and Ancoli-Israel et

al. (1982). Smallwood et al. (1983) found that sleep

apnea was an age dependent phenomenon with male subjects

over 50 exhibiting significantly more apneas than those

under 30. Other evidence includes that of Bixler et al.

(1982) who found a positive correlation between age and

apnea in their full sample, as did Block et al. (1980) in

their post menapausal females. It seems clear from these

data then that increasing age is a factor predisposing for

sleep apnea activity.

Snoring has been closely associated with sleep apnea

(Lugaresi et al., 1982). This may be due to a causal

relationship or a single common pathway. Lugaresi et al.

propose the latter explanation. Specifically, they

suggest that both snoring and SAS are due to a sleep

induced stenosis of the upper airway. With this premise









in mind, they conducted a survey of 1000 subjects on the

incidence of snoring. Chronic snorers included 31% of the

males and 19% of the females. In an elaboration of this

study, Lugaresi et al. (1980) questioned 5713 indivi-

duals. In this sample, 24% of the males and 13% of the

females were chronic snorers. Snoring increased with age,

and by 60 years of age 60% of the males and 40% of the

females snored. Of note was the observation that hyper-

tension occurred more frequently in snorers than non-

snorers, suggesting that the former was more at risk for

cardiovascular complications.

Virtually every study of SAS patients notes the pre-

sence of snoring in all these subjects (Sullivan and Issa,

1980; Block, 1980; Coverdale et al., 1980). This is

true for both predominantly obstructive, and predominately

central SAS subjects (Guilleminault et al., 1978). How-

ever, the presence of heavy snoring in asymptomatic

(nonhypersomnolent) apneas is also well documented (Orr et

al., 1979; Fisher et al., 1978; Berry and Block, 1983),

as well as in otherwise normal subjects. These data may

be taken as evidence for snoring as a necessary, but not

sufficient, condition for the occurrence of sleep apnea.

The role of body weight in SAS is less clear.

Recall that Pickwickian patients are invariably obese, as

well as hypercapnic. Guilleminault et al. (1978) stress

that SAS may occur in nonobese as well as obese sub-

jects. However, Block (1980) states that he has found SAS









subjects to be uniformly obese. A review of the

literature is in order to clarify this point. First

normal, then obese, then SAS groups will be reviewed.

Bixler et al. (1982) found that their subgroup of

subjects with sleep apnea activity weighed significantly

more than the remainder of the sample. Kales et al.

(1982) found that within their sample of 200 insomniacs

and 100 normals, a significant positive correlation

emerged between a weight:height ratio and apnea acti-

vity. Block et al. (1979) found a positive correlation

between apneas, oxygen desaturation, and weight within

their normal males, while females did not demonstrate this

relationship. Using an older female sample, Block et al.

(1980) found a positive correlation between weight and

oxygen desaturation, but not apnea. Thus in normal sub-

jects, there appears to be a positive relationship between

weight and apnea activity, although this may be restricted

to males.

A research approach which bears on this question

involves studying obese subjects alone. Two studies on

successive referrals for gastric bypass surgery were

noted. Sicklesteel et al. (1981) found that of 19 suc-

cessive referrals, all 14 males exhibited sleep apneas,

while only 5 of the females showed evidence of oxygen

desaturation alone. Harmon et al. (1981) studied gastric

bypass patients and found 5 of 6 males to suffer apnea, as








opposed to none of the females. This evidence provides

further evidence for a role of increasing weight in apnea.

The data on symptomatic SAS subjects are more com-

plex. Several studies (Sullivan and Issa, 1980; beitzman

et al., 1980; Garay et al., 1981; Zwillich et al., 1982)

report that all their SAS subjects were obese, although

definitions of obesity either vary widely or remain un-

specified. In contrast, Guilleminault et al. (1978)

reported that only 60% of their SAS sample was greater

than 15% above ideal weight. Coverdale et al. (1980)

indicate that only two of their subjects proved to be

obese, according to their definition (>125% of ideal

weight). These reports suggest that most, although not

all, SAS subjects are obese.

Several investigators have sought to compare weight

in symptomatic (somnolent) and asymptomatic subjects with

enough sleep apnea episodes to be diagnosed as SAS. Orr

et al. (1979) compared 4 symptomatic SAS subjects with

matched asymptomatic subjects. In this sample, symptoma-

tic SAS subjects were heavier than asymptomatic sub-

jects. Examination of the data of Berry and Block (1963)

indicates that their symptomatic subjects were heavier

than their asymptomatic subjects (118 kg vs. 91 kg).

Standing in contrast to these data is a study reported by

Fisher et al. (1978), which included 19 subjects referred

for SAS evaluation. Based on number of apneas per night,

Fisher et al. divided their subjects into 3 groups. Group








I had >100 apneas per night, Group II had >30 but <100

apneas per night, while Group III had less than 20 apneas

per night. Thus Groups I and II were diagnosable as SAS

(>30), while Group III was not. Obesity was present in

Groups I and II, but in only 3 subjects in Group III.

However, these differences were not statistically signifi-

cant.

Taken together, this evidence suggests a positive

relationship between weight and apnea in the general popu-

lation. However, SAS patients are not invariably obese.

Further, weight loss fails to result in a decrease in

apneas in many SAS patients. Thus the relationship

between obesity and full blown SAS is less than clear.

The preceding pages have investigated the currently

understood role of several variables (sex, age, snoring

and weight) in sleep apnea activity. Presently available

data suggest that in young and middle aged subjects, apnea

is a predominantly male phenomenon, although post mena-

pausal females may close the gap. Increasing age is re-

lated to sleep apnea activity. Snoring, which occurs in a

significant percentage of the general population, is

thought to be a necessary but not sufficient condition for

SAS. Weight was shown to have a positive relationship to

sleep apnea activity in the general population, although

its presence or role in SAS is equivocable.

These conclusions have important implications for

experimental designs directed at understanding sleep apnea








activity and its correlates in a subclinical population.

It is clear from the earlier review of apnea incidence

that sampling from the general population would result in

a relatively low population of individuals with sleep

apnea, an inefficient approach from the perspective of the

limited resources available for research. Thus selection

variables must be considered with one eye on efficiency

and the other on protecting the generalizability of

results. Selection criteria should also aim toward eluci-

dating the relative roles and possible interactions of

these variables in sleep apnea. Therefore, selection

variables should have as straight-forward a relationship

as possible with sleep apnea, seeking to answer more ques-

tions than they raise.

With these caveats in mind, the variables available

as selection criterion will be reviewed and conclusions

drawn about their relative merits. The first variable

considered, sex, has been relatively well investigated.

It is clear that sleep apnea activity is primarily a male

circumstance. Therefore, the selection of female subjects

is contraindicated from the standpoint of efficiency. Age

has been demonstrated to exhibit a positive relationship

with sleep apnea, indicating selection of older sub-

jects. Snoring is another variable with a close relation-

ship to sleep apnea. The apparently common mechanism of

snoring and sleep apnea (stenosis of the upper airway)

indicates that possibly very few subjects without snoring








suffer sleep apneas. A significant proportion of snorers

might be expected to suffer apneas, while few nonsnorers

might have apnea. The last variable, weight, while having

a positive correlation with apnea activity in the general

population, is of uncertain pertinence to SAS. Addi-

tionally, changes in weight may occur without changes in

apnea activity. Therefore, this variable has a less than

explicit connection to apnea activity and violates the

second principle adopted to consider these variables.

This review of available data suggests that sampling

aging, snoring, males would generate a high proportion of

sleep apnea activity. External validity for this idea is

drawn from its similarity to the "typical" SAS patient

described earlier. Extant data suggest that it may also

prove useful to utilize weight measures as a covariate, in

an attempt to control its apparent influence on apnea

activity.

Deficits Associated With Sleep Apnea Syndrome

Sleep Apnea Syndrome is thought to result in several

deficits. These include cardiopulmonary complications,

hypersomnolence, hypoxia and intellectual changes. Some

researchers have suggested that these deficits may be

present, in an attenuated form, in subclinical apnea

(Ancoli-Israel et al., 1981; Carskadon and Dement,

1981b). A discussion of the evidence for the presence of

these deficits will be presented, followed by proposals

for appropriate measures of these variables.









Cardiopulmonary Deficits

Cardiopulmonary complications have been reported as

concomittents of SAS (Guilleminault et al., 1978).

Coccagna et al. (1972) reported high incidences of hyper-

tension and congestive heart failure in their SAS sub-

jects. Schroeder et al. (1978) studied 22 SAS patients

with a mean age of 47. Six subjects had waking systemic

hypertension, and several had waking cardiac abnorma-

lities. During sleep studies, 20 patients developed

significant rises in systemic arterial pressure which

cycled with episodes of apnea, while 21 patients developed

pulmonary arterial hypertension. Tracheostomy reversed

all these abnormalities, strongly implicating the apneas

in their etiology. Tilkian et al. (1978) studied 25 male

SAS subjects with a mean age of 44. Twenty-four of the 25

showed sleeping sinus arrhythmias, while 9 developed more

serious symptoms such as asystole. Seventeen of these

patients received tracheostomy, which abolished all of

these abnormalities. Experimental occlusion of the

tracheostomy during sleep in 6 of these patients resulted

in a return of the cardiopulmonary symptoms. Zwillich et

al. (1982) studied six consecutive male SAS patients. In

all patients, bradycardia accompanied any apnea with

significant desaturation. Additionally, a significant

correlation was noted between degree of desaturation and

severity of bradycardia. Bradycardia was abolished or

attentuated during administration of 02 enriched air.









Zwillich et al. proposed that bradycardia during apnea

results from increased vagal tone mediated by carotid body

chemoreceptors. Fujita et al. (1981) reported on 12 male

SAS patients. All reportedly suffered cardiac arrhythmias

during sleep. The body of data suggests that SAS patients

frequently suffer from cardiac abnormalities such as

hypertension, arrhythmias, and bradycardia.

A second body of data exists which indicates that SAS

patients do not invariably have cardiopulmonary compli-

cations. Coverdale et al. (1980) studied 14 patients, of

whom 8 were diagnosed with SAS. Only 2 of these patients

had evidence of corpulmongle, while 2 more had systemic

hypertension. None of these patients demonstrated severe

brady or tachycardia during sleep. In Orr et al. (1979),

8 patients, all of whom were technically diagnosable as

having SAS, all symptomatic (somnolent) patients had evi-

dence of right sided heart failure as well as frequent

arrhythmias during apneas. Asymptomatic subjects had none

of these symptoms. Kreiss et al. (1982) studied 26

Veterans Administration ward patients and found 7 to have

a SAS. The SAS patients had significantly more signs of

congestive heart failure, but not hypertension or

angina. Ancoli-Israel et al. (1981) sampled 24 elderly

subjects with complaints of sleeping respiratory ois-

orders. Nine of these subjects were diagnosable as SAS,

but this group did not have significantly more heart








disease or hypertension. Thus a group of SAS patients

without evidence of heart disease exists.

Taken together, these data indicate that cardiac

complications frequently, but not inevitably, accompany a

sleep apnea syndrome. That a large number of apneas may

be present without a significant increase in cardio-

pulmonary complications is demonstrated by the Orr et al.

and Ancoli-Israel et al. studies. As the relationship

between these complications and a full blown SAS is

unclear, the possibility of cardiopulmonary sequelae from

a subclinical level of apneas is even more uncertain.

However, an assessment of cardiopulmonary status is

indicated by the frequent association of cardiopulmonary

complications with sleep apnea syndrome.

Arousal Deficits (Hypersomnolence)

Excessive Daytime Sleepiness (EDS), or hypersomno-

lence, has been described as the single most common result

of SAS (Dement et al., 1978). Indeed, patients who meet

the criterion of number of apnea episodes and exhibit EDL

are described as symptomatic, while those with number of

apneas alone are labeled asymptomatic (Orr et al.,

1979). As an aside, this distinction illustrates again a

central issue of SAS. If patients exhibit the number of

apneas qualifying them for SAS, but no other pathology,

are they suffering from a pathological process? Dement et

al. (1978) believe that they are. They suggest that most,

if not all, patients with a clinically elevated number of









apneas actually suffer LDS, but that it is masked by two

factors. Dement et al. noted that SAS patients may deny

somnolence while literally falling asleep before the

clinician. They speculate that this results from a re-

sponse bias against acknowledging illness or a change in

subjective frame of reference about just what alertness

is. Thus EDS may be denied by patients for reasons which

may be outside the clinician's control or knowledge.

Alternatively, Block et al. (1979) propose that asympto-

matic SAS patients are suffering a subclinical level of

fallout from their frequent apneas. As years pass, in-

creasing weight or cumulative effects of desaturation

eventually lead to a full blown hypersomnolence--Sleep

Apnea Syndrome. A final possibility is that nonhyper-

somnolent subjects with frequent apneas exhibit a com-

pletely benign process totally unrelated to that found in

symptomatic SAS patients. It would seem that delineation

of the level of EDS in subjects with subclinical levels of

apnea may help clarify this issue. This notion will be

elaborated below.

The somnolence deficit which is noted in full blown

SAS patients is commonly reported. Dement et al. (1978)

describe impairments in continuous performance of

virtually any activity. Sleep Apnea Syndrome patients

reportedly fall asleep at outdoor stadiums, in front of

classes, and while treating patients. Excessive Daytime

Somnolence is described in every SAS patient reported in








experimental protocols by Weitzman et al. (1980), Fujita

et al. (1981), Zwillich et al. (1982), Garay et al. (1981)

and Sullivan and Issa (1980). At the same time, subjects

wno are asymptomatic for EDS but with the necessary number

of apneas to be diagnosed as SAS are described by Orr et

al. (1979) and Smirne et al. (1980) as well as others.

However, it should be noted that the latter studies relied

on global reports of somnolence, rather than quantified

data. Recalling the criticism of Dement et al. (1978) of

this technique as vulnerable to subjective bias, these

asymptomatic patients may have unreported or undetected

somnolence.

Excessive Daytime Somnolence appears to be frequently

concomittant with Sleep Apnea Syndrome. Thus its possible

occurrence in subclinical populations seems worthy of

investigation.

Oxygen Desaturation and Cognitive Deficits

One deficit which appears in SAS patients is oxygen

desaturation which accompanies apneas. When apnea begins,

arterial oxygen saturation begins to fall, and may con-

tinue falling throughout the event. An early technique

utilized in studying oxygen saturation levels involved

periodic blood samples drawn from arterial sources.

Birchfield et al. (1958) studied 11 normal males with a

mean age of 23. Daytime oxygen saturation in these sub-

jects averaged 96.5%. These values fell during sleep to

95.3%. Later, Orr et al. (1979) studied 8 males who had








been matched on number of apnea events. Asymptomatic

(nonhypersomnolent) subjects fell from a baseline 02

saturation of 80 mm (Hg) to a mean maximum desaturation of

54 mm (Hg) during sleep. Symptomatic subjects fell from a

baseline of 54 mm (Hg) to a mean maximum desaturation of

35 mm (Hg) while asleep. Orr et al. suggested that the

more severe desaturation which accompanied apneas in

symptomatic subjects underlaid their somnolence and other

complications. However, these conclusions are vulnerable

to certain methodological criticisms, as Block (1980)

points out. Specifically, periodic sampling may well fail

to detect the multiple brief desaturations associated with

SAS, rendering this technique potentially insensitive to

important events.

In an attempt to provide more representative data on

desaturation, Block et al. (1979) utilized an ear oximeter

in studying saturation in normal subjects. An ear oxi-

meter provided a continuous, accurate readout of moment to

moment arterial oxygen levels. Thirty males (mean age 38)

and 19 females (mean age 27) were studied. Of the 30

males, seventeen suffered at least one episode of desatu-

ration, falling from a mean saturation of 95% to a maximum

desaturation of 84% during sleep. In contrast, no epi-

sodes of desaturation were noted in the sample of premena-

pausal females (baseline 02 saturation: 96%). The oxygen

desaturations of the males were always found in asso-

ciation with breathing abnormalities or snoring. A stay








of oxygen saturation in 20 post menapausal females was

reported by Block et al. (1980). Eleven of these women

exhibited desaturation with several subjects desaturating

to less than 85%. Dolly and Block (1982) studied a group

of 17 males and 3 females with a mean age of 49 years.

These normal subjects dropped from a mean baseline satura-

tion of 96.3% to a mean maximum desaturation of 88.3%.

This evidence indicates that desaturations of 10-127 are

common in normal males and post menapausal females while

premenapausal females rarely, if ever, desaturate.

Oxygen saturation levels in SAS patients are not

commonly reported. Using an ear oximeter, Garay et al.

(1981) studied 11 SAS patients who were divided into 6

daytime eucapnics and 7 daytime hypercapnics. Eucapnics

fell from a baseline saturation of 95% to a mean maximum

desaturation of 75.5%, while hypercapnics fell from a

baseline saturation of 89% to a maximum desaturation of

62.8% (calculated from tabled data). These data were

collected during daytime "nap" studies, however, and hence

is somewhat questionable. Berry and Block (1983) recorded

9 heavy snoring males who suffered many apneac events.

Four subjects were somnolent, while 5 were not. Symptoma-

tic subjects fell from a baseline of 95% to a mean maximum

desaturation of 44%, while asymptomatic subjects desetu-

rated from a baseline of 94% to a mean maximum desatura-

tion of 85%.








Data on desaturation are interesting from several

perspectives. First of all, Sleep Apnea Syndrome subjects

seem to desaturate more heavily than normals, with Berry

and Block's symptomatic group desaturating up to 50% and

the subjects of Garay et al. desaturating about 22%,

compared with normal desaturation levels of 10-12%.

Secondly, amount of desaturation distinguished symptomatic

(hypersomnolent) from asymptomatic subjects in the data of

both Orr et al. and Berry and Block. Additionally, two

points relevant to methodology emerge. Measures of oxygen

desaturation seem crucial to understanding sleep apnea and

its effects, and ear oximetry appears to provide a proven,

valid measure of this variable.

Weitzman (1979) has suggested that the pathological

process in SAS is the hypoxia of desaturation in apneac

episodes. While cognitive and intellectual changes in SAS

have been only anecdotally reported (Guilleminault et al.,

1978), hypoxia does seem a plausible potential cause of

these changes. Thus oxygen desaturation appears to be a

reliable consequence of sleep apnea syndrome, while cogni-

tive/intellectual changes are putative sequelae.

In summary, a review of deficits found in sleep apnea

syndrome subjects suggests that cardiopulmonary complica-

tions, excessive daytime sleepiness, oxygen desaturation

and cognitive/intellectual changes are frequently asso-

ciated with this syndrome. An assessment of these

variables in a subclinical population seems indicated by








these data. Below a discussion of measurement issues and

proposal assessment of these variables appears.

Measurement of Deficits Found in Sleep Apnea Syndrome

Above it was noted that cardiopulmonary complica-

tions, excessive daytime sleepiness, nocturnal oxygen

desaturation and cognitive/intellectual changes are common

deficits of sleep apnea syndrome. A discussion of

measurement issues and proposed measurement devices for

these variables will be presented below.

Measurement of Cardiopulmonary and Health Deficits

Cardiopulmonary deficits are thought to be frequent

sequelae of sleep apnea syndrome. Thus an assessment of

blood pressure as well as self reports of hypertension and

heart trouble seems indicated in subclinical apnea sub-

jects. Additionally, a general survey of health status,

such as that provided by the Cornell Medical Incex (CiI;

Broadman et al., 1949) would also screen for other health

deficits. The CMI provides an overall score indicating

number of symptoms endorsed, as well as subscales on

several symptom categories. Separate examination of sub-

scales on respiratory, cardiopulmonary, and neurological

subscales seems appropriate. Assessment of these

variables provides a broad health screening, as well as a

detailed analysis of symptoms found in sleep apnea syn-

drome.








Measurement of Arousal Deficits (Hypersomnolence)

Excessive daytime sleepiness is a common sequel of

sleep apnea syndrome. Measurements of the sleep/wake

cycle of subclinical apnea patients thus seems indi-

cated. A sleep questionnaire assessing trait sleep habits

seems in order, as well as 1 week sleep logs following the

experimental night. The sleep questionnaire and sleep

logs have been used in ongoing research in W.B. Webb's

studies of aging and sleep and seem an adequate assessment

of sleep patterns of subclinical apnea subjects. Addi-

tionally, electroencephalographic recordings of sleep

during the experimental night seem in order using a stan-

dard recording and scoring system (Agnaw and Webb,

1972). The assessment of daytime sleepiness is relatively

new, only recently being the subject of investigation.

William Dement and associates began to develop methods of

quantifying somnolence in the early 1970s. The first

attempt produced a subjective rating scale, the Stanford

Sleepiness Scale (SSS; Hoddes et al., 1973). The SSS was

a 7 point Likert scale with endpoints anchored on 1) feel

active and vital; and 7) almost in reverie, sleep onset

soon; with 5 points with statements reflecting various

degress of somnolence in between. Patients rated their

introspective level of sleepiness on this scale.

Unfortunately, the SSS proved to be vulnerable to exactly

the same difficulties as a global rating scale used on SAS








patients, i.e., response bias and altered frame of

reference (Dement et al., 1978).

It was clear then that an objective measure of day-

time somnolence was needed. With this goal in mind,

Carskadon and Dement (1977) introduced the Multiple Sleep

Latency Test (MSLT). This technique involved measuring

the objective sleep latency during multiple nap attempts

throughout the day. In a typical paradigm, a subject

would be wired to an EEG and attempt to fall asleep at 2

hour intervals through the day. The latency to sleep in

each nap was measured by determining the latency to the

first epoch of Stage 1 sleep. In order to prevent signi-

ficant amounts of sleep accumulating during testing, sub-

jects were awakened after an epoch of sleep was

observed. If no sleep was noted, subjects were discon-

nected after 20 minutes and allowed to resume other acti-

vities. Carskadon and Dement believe that this measures

underlying physiological sleep tendency. This underlying

tendency is thought to be modulated by alerting stimuli in

the environment (Carskadon and Dement, 1982a). The vali-

dity of the test was established on normal subjects, who

exhibited a biphasic curve of sleep tendency throughout a

24 hour period. Extension of sleep over typical lengths

in normals (who are thought to be moderately chronically

sleep deprived) saw a significant rise in average sleep

latencies on the MSLT (Carskadon and Dement, 1979b), while

chronic sleep restriction (5 hrs sleep per night) caused








significant declines in average sleep latencies (Carskadon

and Dement, 1981a). Finally, total sleep deprivation led

to a drastic decline in MSLT values, with averages

plummeting below one minute (Carskadon and Dement,

1979a). Mean MSLT values showed recovery to baseline

values after a full night's sleep for partially (5 hrs)

deprived subjects, and recovery after two nights of ad lib

sleep for totally sleep deprived subjects (Carskadon and

Dement, 1979; 1981a). This body of evidence suggests

that the MSLT is a reliable and valid index of sleepiness

in normal subjects.

With its internal validity established, the MSLT was

applied to somnolent populations. Dement et al. (1978) as

well as Richardson et al. (1978) demonstrated that SAS and

narcoleptic patients have significantly shorter latencies

to sleep on the MSLT than normal subjects do. Hartse et

al. (1979) found SAS subjects to have an average latency

to Stage 1 of 2.5 minutes, while narcoleptics had a mean

of 3.2 minutes. Both these latencies were significantly

lower than those of a miscellaneous control group whose

latencies were about 11.2 minutes. Hartse et al. (1980)

found latencies to Stage 1 sleep as follows for various

patient groups: Normals 12.2 m, Insomniacs 16.2 m, SAS

2.5 m, and Narcoleptics 3.2 m. The SAS and narcoleptic

patients had significantly shorter latencies to Stage 1

sleep than insomniacs or normals. Roth et al. (1980)

compared 10 SAS patients with 10 age matched controls.








The normals averaged 20.4 m to Stage 1 sleep while the SAS

patients averaged 1.9 m. Zorrick et al. (1982) found mean

MSLT latencies to Stage 1 sleep of 3.1 m for SAS subjects

and 2.9 m for narcoleptics, significantly shorter than for

patients with psychiatric disorders.

The chief difficulty remaining in judging the MSLT

"reliable and valid" is external validation demonstrating

that exceeding a certain mean latency is correlated with

other deficits of performance. In some ways, low laten-

cies on the MSLT may suffer the same limitations as sleep

deprivation protocols; that is, demonstrating serious

deficits is surprisingly difficult. At the same time, few

would deny the subjective unpleasantness of sleep depriva-

tion, which may be applicable to low MSLT scores. Thus

the MSLT appears indicated in assessing excessive daytime

sleepiness while the Stanford Sleepiness Scale (SSS) pro-

vides an index of the subjective experience of the sub-

ject. Application of both these measures to a subclinical

apnea population would seem to adequately assess daytime

sleepiness in these subjects.

Measurement of Hypoxia and Its Sequelae

Noctural desaturation is frequently noted in SAS

patients. Therefore, it seems in order to examine the

known sequelae of hypoxia of other etiologies in an

attempt to determine variables or behaviors likely to be

sensitive to the hypoxia found in SAS and SAA. Research

on hypoxia varies systematically along two dimensions,









human vs. animal subjects and acute vs. chronic hypoxia.

First, the data available on animals will be reviewed,

followed by that extant on human subjects. Finally,

behavioral variables sensitive to changes in these

variables will be summarized.

The effects of chronic hypoxia on animals have not

been exhaustively studied (Davis, 1975). One project

which did examine this variable involved exposure of rats

to up to 36 hours of 10% oxygen. Davis (1975) observed

the behavior of these animals and concluded that they

appeared normal. Although hyperventilation was noted, no

permanent effects of the hypoxia were found. Hanbauer et

al. (1981) examined the effects of 10% 02 on rats exposed

for up to 4 weeks. They found that relatively short term

(2 days) hypoxia resulted in an increase in dopamine (DA)

content in the carotid body. An increase in norepine-

phrine (NE) content was noted after 1 week of exposure.

As the carotid bodies are implicated in the control of

respiration, the authors concluded that different

mechanisms are operative in the adaptation to short and

long term hypoxia. Unfortunately, little other data have

been reported on the chronic effects of hypoxia on more

complex functioning. This is particularly problematic as

it are these functions which are thought to be affected in

humans exposed to hypoxia.

Certain aspects of acute hypoxia in rats have been

carefully examined. Gibson et al. (1981) reviewed the









evidence which has been collected on neurotransmitter

alterations in hypoxia. In brief, mild to moderate

hypoxia, cerebral levels of ATP are normal, implying that

energy supplies of neurons are not disturbed. In con-

trast, the turnover of certain neurotransmitters is

altered. Gibson et al. note that oxygen is an important

substrate for the synthesis of several neurotrans-

mitters. When rats are exposed to acute hypoxia, syn-

thesis of dopamine and serotonin is decreased, while

absolute levels of tryosine, tryptophan, catecholamines,

and serotonin remain constant. Continuing hypoxia leads

to an adaptation resulting in normal turnover of these

neurotransmitters. Gibson and Blass (1976) examined the

effect of acute hypoxia on central acetylcholine (ACh).

As in dopamine (DA) and serotonin (5-HT), acute hypoxia

results in a decrease in synthesis in ACh, but not in the

absolute level of ACh. Thus there is a substantial body

of evidence indicating changes in central neurotransmitter

levels in hypoxia. Generalizing from the animal data to

human subjects, Gibson and Duffy (1981) suggest that these

changes underlie the behavioral deficits of hypoxia in

humans (which will be reviewed below).

Other variables have been examined in acute hypoxia

in animals. Annau (1972) found that rats exposed to brief

hypoxia exhibited decreased appetite, thirst, and self

stimulation rates. Annau observed that brief hypoxia

depressed all behavioral variables. Gellhorn (1951) found








that brief hypoxia of 4-7.5% 02 abolished cortical re-

sponses to auditory stimulation. Sara (1974) trained rats

in active one way avoidance tasks. Immediately after

training, subjects were exposed to 3.5-4.0% oxygen. Rats

tested at 1 and 3 hours post hypoxia avoided correctly,

while those tested at 6 and 24 hours did not, leading Sara

to suggest that hypoxia resulted in a memory retrieval

deficit. Freides and Allweise (1978) also trained rats in

one way avoidance and exposed them to 2.0% hypoxia imme-

diately following training. These animals avoided at 12

and 200 minutes post hypoxia, but not at 90 minutes.

Friedes and Allweise suggested that a sensitive period for

hypoxia effects on training existed for 15 minutes after

learning. Tauber and Allweise (1975) followed the same

training/hypoxia paradigm and found avoidance impaired at

2.5 but not 4.5 hours. They suggest that hypoxia inter-

feres with a medium term memory. While some of these data

are conflicting, it appears that acute hypoxia depresses

certain behavioral variables and interferes with aspects

of memory processes.

While caution is in order when generalizing from

animals to humans, several seemingly relevant points

emerge from the review. Behaviorally, simple activities

may be spared, although some depression of appetitive

activities may occur, while more complex behavior depen-

dent on memory processes may be altered. Of particular

interest are the demonstrated alterations in









neurotransmitter turnover seen in hypoxia. It has been

suggested that these subtle alterations may underlie the

changes in complex behavior seen in hypoxic humans.

Data on human performance after hypoxia have neces-

sarily been reliant on fortuitous or weak manipulations.

Although this may limit the usefulness of these data, it

is at the very least highly suggestive of variables sensi-

tive to hypoxia. Van Liere and Stickney (1963) note that

acute hypoxia leads to confusion, headache, drowsiness,

weakness and incoordination. After effects include head-

ache, nausea and emotional liability. Richardson et al.

(1959) found that acute, fatal hypoxia from surgery

results in damage to cortical neurons, particularly in

layers III and IV. Susceptible subcortical neurons in-

clude those of the corpus striatum and cerebellum. Plum

et al. (1962) describe a syndrome of delayed postanoxic

encephalopathy. After the acute coma resolves in 4-5

days, a brief period of normal functioning returns.

Increasing irritability and confusion then emerge, with

loss of coordination and memory, and diminished attention

span. Typically, the lesion spares grey matter while

affecting white matter. Devereaux and Partnow (1975)

describe a patient who recovered from a delayed acute

encephalopathy. Although IQ was relatively spared, severe

dysarthria emerged. Relevant points which emerge from

studies of acute hypoxia in humans suggest that sublethal









levels of hypoxia affect coordination, memory, attention

span, and may result in headache, drowsiness, and nausea.

The effects of milder, chronic hypoxia are summarized

by Gibson et al. (1981). At the oxygen pressure equiva-

lent of 5,000 feet, impaired dark adaptation is noted. At

10,000 feet decreased concentration, hyperventilation, and

short term memory deficits are noted. By 15,000 feet,

euphoria and loss of both coordination and critical judge-

ment are observed. Christenson et al. (1977) found signi-

ficantly less visual signals detected at 17% 02 than at

21,%, suggesting reduced vigilance. McFarland (1937)

studied men in ascents to high altitudes and concluded

that subjects with slower rates of ascent performed better

on unspecified psychological tests. West (1984) reports

data collected on an expedition to Mt. Everest. Results

of a neuropsychological battery indicated deficits in

finger tapping, verbal fluency, verbal learning, short

term memory, and expressive language. Thus the general

affects of chronic mild hypoxia are thought to be deficits

in dark adaptation, short term memory, critical judgement,

motor coordination, verbal fluency, and vigilance.

One group of chronically hypoxia patients, those with

chronic obstructive lung disease (COLD), has been rela-

tively well studied with sensitive neuropsychological

instruments and is deserving of a more painstaking

review. Krop et al. (1973) studied COLD patients before

and after they received oxygen supplements. The WAIS,









Wechsler Memory Scale (WKS), Bender-Gestalt, Bender

Gestalt Interference Procedure, Facial Recognition and

Finger Tapping tests were administered to COLD and control

subjects. Subjects receiving supplementation improved

significantly more than controls on WAIS Full Scale IQ,

Performance IQ, Wechsler Memory Quotient, both Bender

procedures and Finger Tapping. Grant et al. (1982)

studied another group of COLD patients. They administered

the Halstead-Reitan Battery, Aphasia Screening Test,

Trailmaking Test, Russell modification of the Wechsler

Memory Scale, Tactile memory test, and the WAIS. In COLD

subjects, deficits were found relative to controls on all

variables except Reitan Rhythm, Aphasia Screening and WNS

Logical stories. Grant et al. concluded that COLD

patients exhibited deficits on a global impairment rating,

attention, abstracting ability, complex perceptual motor

tasks, simple sensory and motor tasks, and memory tasks.

Low but significant negative correlations were reported

between several measures of oxygen saturation and both

Halstead Impairment Index and Global Impairment Rating.

Grant et al. speculated that floor effects resulted in low

variability in PA02 levels among COLD patients and kept

these correlations low. They further suggested that the

neuropsychological deficits observed in these patients

resulted from the oxygen want found in these subjects.

Thus data available on humans indicate that dark

vision, memory, motor coordination, critical judgement,








verbal fluency, and signal detection are affected by acute

hypoxia. Animal data suggest that sensitivity of memory

processes to hypoxia, while data from chronic hypoxia

humans indicate that attention, vigilance, abstracting

ability, complex perceptual motor skills, as well as motor

and sensory abilities are also impaired by hypoxia. While

none of these findings are unequivocally directly

applicable to the brief, acute, desaturation occurring

during apnea, they are, at the very least, indicative of

abilities sensitive to hypoxia in humans.

A neuropsychological battery sensitive to these

abilities includes Wechsler Adult Intelligence Scale

(WAIS; Wechsler, 1955); Wechsler Memory Scale (WMS;

Wechsler, 1945) including the Russell delayed memory

aspects (Russell, 1975); Rey-Osterreith Complex Figure

(Osterreith, 1944); Hooper Visual Organization Test

(Hooper, 1958); Verbal Fluency Test (Halstead, 1947);

Wisconsin Card Sort (Berg, 1948); and the Finger Tapping

Test (Halstead, 1947). Administration of this test

battery provides a broad screening of neuropsychological

functioning, with a particular emphasis on those abilities

known to be sensitive to hypoxia. Application of this

battery to subclinical apnea subjects provides a sensitive

indicator of possible hypoxic brain dysfunction.

In summary, alterations of sleep/wake cycles appear

to be present in sleep apnea syndrome patients. Thus

assessment of these variables in a subclinical population









is indicated. An adequate description of the sleep/wake

cycle would seem to include sleep questionnaire, sleep

log, overnight EEG, multiple sleep latency test, as well

as several Stanford Sleepiness Scale ratings.

Statement of the Problem

Sleep Apnea Syndrome is a clinical entity with

seemingly serious consequences. Although the nocturnal

apneas found in this illness are thought to be the cause

of the associated deficits, the identification of

seemingly asymptomatic subjects with a number of apneas

comparable to those found in SAS throws doubt onto this

formulation. Despite this evidence, the recent discovery

tnat apneas occur in significant numbers in the general

population has led to speculation that this subclinical

level of apnea may be the cause of various pathological

processes. At present, there is little evidence to

support or refute this notion.

The present report evaluates this issue through a

detailed study of a group of subjects thought to be at

risk for sleep apneas. Aging, heavy snoring males were

recruited, and a comprehensive analysis of their respira-

tory, health, neuropsychological, and sleep/wake cycle

status was carried out. Thus potential relationships

between respiratory disturbances and variables known to be

disrupted in sleep apnea syndrome patients could be

evaluated. This approach also allowed exploration of the





47



possible risk factors associated with subclinical sleep

apneas.















CHAPTER TWO
METHOD

Subjects

Heavy snoring males were recruited via newspaper

advertisements, notices, and phone calls to a list of

aging subjects reported by Webb (1982). Participants were

required to be male, self reported heavy snorers, over 30

years of age, and in general good health (i.e., self

described healthy and normal and not under the active care

of a physician for illness). Subjects with a history of

head trauma or alcoholism were excluded. Potential

subjects who met these criteria were offered a chance to

participate in a "snoring study" in which they would

complete testing during an evening, sleep overnight in a

lab with several physiological parameters recorded, and

complete one week's worth of "sleep logs," for a payment

of fifty dollars. Volunteers were scheduled on a first

come, first served basis. A total of 60 subjects were

studied, with 9 cropped because of technical problems with

their overnight recordings and 5 excluded because of

health problems. Thus 46 subjects represented the final

sample. These 46 subjects had a mean age of 49.9 (sd

13.1) and represented a wide cross section of the

population, ranging from the unemployed to university








professors. All subjects completed and signed an informed

consent agreement.

Apparatus

During the evening testing, all subjects received a

blood pressure check with a standard hospital cuff. The

measurement was made from a seated position. Overnight

recordings were made in a quiet, darkened room with the

subject sleeping on a standard hospital bed. All noc-

turnal physiological recordings were routed through a

Grass Polygraph (model 7D). Electrodes for electro-

encephalographic recordings were mounted in three pairs

using sites F2/F8, PZ/T6, RE/LE from the international

10/20 system of placement. These electrodes were affixed

to the scalp with collodion soaked gauze pads dried with

an air hose. Respiratory measures included measurement of

chest and abdomen wall movement derived from impedence

measures monitored from surface electrodes (3M Ag/AgC1

#2246) placed on the lower chest and just above the

navel. The signal from these electrodes was passed

through an impedence convertor and onto the polygraph.

Oral and nasal airflow was recorded by thumistors (TCT-1R

Transducer-Grass) clipped to the respective orifices.

These signals were also routed through the polygraph.

Electrocardiograms were monitored from modified bipolar

chest leads (MCL2). Finally, blood oxygen saturation was

continuously monitored from a probe attached to the ear

lobe, and analyzed by a Biox ear oximeter (Ilodel IIA BTA









Co). All physiological variables were recorded on the

polygraph's chart paper which was run at a speed of 10

mm/s.

Measures

Two classes of independent variables are identified,

with the first being thumistor variables (those derived

primarily from the nose and mouth thumistors). These

included number of apneas and hypopneas, mean duration of

apneas and hypopneas, mean low oxygen saturation in apneas

and hypopneas, mean oxygen saturation change in apneas and

hypopneas, and total seconds spent in apneas and hypo-

pneas. Thumistor variables were separately scored by two

trained raters, whose agreement appears in Table 2-1. An

overall agreement of 85% was achieved. Disagreements were

resolved by consensus scoring. The second class of

independent variables included oxygen saturation variables

(derived from a minute by minute rating of highest and

lowest oxygen saturation from the overnight record).

These variables included mean highest saturation, mean

lowest saturation, and number of desaturations of 4 ana

10% or more. In selecting these scoring criteria, several

considerations were relevant. Initially, a system with

reasonable interrater reliability was necessary. While

published criteria for scoring apneas were not problem-

atic, reliable scoring of hypopneas was not achieved until

a desaturation criterion of 10% was introduced. In

selecting measures of oxygen saturation, a mean high and










Table 2-1. Interrater agreement for respiratory events of
46 snoring males.


Consensus Initial Disagreements
Scoring Agreements #1 #2

(178) -
91 90 4 6
78 76 1 1
61 52 3 8
57 56 9 2
32 29 3 3
28 27 0 2
25 25 0 0
19 16 4 0
17 17 0 0
15 16 1 0
13 13 0 0
7 7 0 0
S44 0 0
4 4 0 0
3 2 0 1
3 3 0 0
3 0 7 2
3 3 0 0
3 3 0 0
1 1 0 0
1 1 0 0
1 1 0 0
1 1 0 1
1 0 0 1
1 1 0 0
1 1 0 0
1 1 0 0


474(652) 450


Note


Scored by
Consensus
Consensus
Consensus
Consensus
Consensus
Consensus


consensus
added 1
added 2
added 9
added 1
added 3
added 1


Consensus added 3

Consensus added 1




Consensus added 1

Consensus added 3






Consensus added 1


37 56


Note: 28 of
hypopnea.


the 46 subjects experienced at least 1 apnea or









low score seemed obvious. In order to assess the quan-

titative "impact" of numerous desaturations, a sum of 4%

and 10% desaturation was tabulated. These selected

parameters for thumistor and saturation variables allowed

reliable scoring and a description of several aspects of

nocturnal respiration.

Dependent measures fell into three broad classes:

measures of sleep/wake status, measures of health status,

and measures of neuropsychological status. Measures of

the sleep/wake cycle included one week's sleep logs

(derived variables included number of reports of daytime

sleepiness, hours asleep, estimated sleep latency, number

of wakenings after sleep onset, and minutes of waking

after sleep onset), a sleep questionnaire (derived

variables including hours of sleep usually obtained,

number of naps, hours napping, quality of sleep,

restfulness on awakening, depth of sleep, and amount of

daytime sleepiness), the mean of the several Stanford

Sleepiness Scale ratings, the two multiple sleep latency

tests (derived variables included mean sleep latency and

total minutes asleep in both naps) and the overnight LEG

variables derived from the scoring system of Agnew and

Webb (1972). Electroencephalographic variables included

time in bed; pure sleep time; sleep efficiency index;

number stage 0 periods; time 1 stage 0; sleep latency;

time % stages 1, 2, 3, 4 and REM; latency first hLM









period; mean REM period; mean REM cycle length; $ slow

wave sleep; and number of stage changes.

Measures of health status included height, weight,

diastolic and systolic blood pressure, as well as the

Cornell Medical Index. From the CMI, reports of hyper-

tension and heart trouble were noted, as well as the total

number of symptoms endorsed. Additionally, the scores for

the scales tapping respiratory, cardiac, and neurological

symptoms were tabulated separately.

Measures of neuropsychological status included the

WAIS Verbal and Performance IQ's, the WMS Memory Quotient,

delayed recall of the logical stories and the visual re-

production subtests of the WMS, delayed recall of the Rey

complex figure, number of words generated in the verbal

fluency test, number of correct sorts of the Wisconsin

Card Sort, mean number of taps for the right and left

handed finger tapping tests, and lastly the number of

correct trials on the Hooper test.

These neuropsychological tests were chosen with two

aims. First, an attempt was made to sample from each of

several loosely bounded areas of cognitive skill. These

included intelligence (both verbal and non-verbal),

memory, both immediate and delayed (verbal and non-

verbal), visuo-perceptual/organizational skills, language,

and frontal self-regulatory skills. A second emphasis was

on a relatively detailed analysis of areas previously

shown to be sensitive to hypoxia. These sensitive areas









included memory, visuo-organizational skills, verbal

fluency, and motor coordination. The chosen neuropsycho-

logical battery reflected these two aims. Thus the WAIS

tapped intellectual functioning; the Wechsler Memory

Scale with delayed recall of logical stories and visual

reproductions as well as digit span and Rey Figure heavily

sampled verbal and non-verbal memory; the Hooper tapped

visuo-organizational skills; Verbal fluency sampled lan-

guage skills; the Wisconsin card sort tapped frontal

self-regulatory skill; and Finger tapping sampled motor

functions. The chosen battery provided a broad screening

as well as a detailed assessment of neuropsychological

functions thought to be impaired by hypoxia.

Procedure

Subjects followed the schedule appearing below for

the experimental night:

1e00 Sign informed consent, blood pressure,

height/weight, finger tapping, verbal fluency,

subjective sleepiness (SSS-made every 1/2 hour

till bedtime)

1830 Wiring for EEG

1900 Begin MSLT I (20 m)

1930 WAIS

2000 WMS, Hooper

2030 Graphesthesia, Luria motor programs

2100 Begin MSLT II (20 m)

2130 Rey figure, Wisconsin card sort









2200 Sleep Questionnaire, Cornell Medical Index,

remaining testing

2230 Wiring for thumistor, EKG, respiratory effort,

ear oximeter

2300 Lights out, begin recording

A night technician remained with the subjects throughout

the night. Subjects usually awakened between 5 and 6 am,

and later completed the one week's sleep logs.

Statistical Procedures

Several statistical procedures were utilized in

evaluating the present data. One of the primary

approaches was correlational, considered appropriate

because of the exploratory nature of the study. Bearing

in mind the possible alpha inflation of large corre-

lational matrices, other approaches were also utilized to

add confidence. Thus the sample was stratified by level

of apnea/hypopnea, and oneway ANOVA procedures were used

to evaluate between group differences. Additionally, non-

parametric procedures were utilized where appropriate.

Positive correlational results were considered against

possible replication with other measures in the study, and

in some cases multivariate procedures were addea for

further confidence. It is conceded that correlational

approaches may carry the burden of possible spurious

results, but it is believed that the various additional

procedures outline above lend confidence to the findings.















CHAPTER THREE
RESULTS

Demographics and Incidence of
Nocturnal Respiratory Disorder

Subjects were selected to fulfill several sampling

requirements, including male sex, presence of heavy

snoring, and self report of general good health. A total

of 60 subjects were studied. Five subjects were later

excluded when questioning revealed the presence of chronic

alcoholism (2) and serious head trauma (3). Additionally

9 overnight records were rejected because of missing or

unusable data. The characteristics of the final 46

subjects are detailed in Table 3-1. These snoring males

had a mean age of approximately 50 years, a mean weight of

approximately 190 pounds, and a mean educational level of

15 years of schooling.




Table 3-1. Demographic variables from 46 snoring males.



K SD Range

Age 49.9 13.1 30-75
height (Ibs) 189.3 34.9 125-250
Weight:Height Ratio (lbs:inches) 2.7 0.5 1.9-3.6
Eaucation (years) 14.7 2.5 9-20









The respiratory records were annotated by two hour

intervals following bedtime. From a total of 652 events

scored, 150 (23%) occurred in the first two hours

following bedtime, 185 (23%) occurred in the second two

hours, 232 (36%) occurred in the third two hours, while 83

(12%) occurred in the final two hours. It should be noted

that wakeup times varied considerably, making interpre-

tation of the attenuated number of events in the final two

hours somewhat problematic.

From a total of 578 events in which reliable sleep

staging was achieved, 459 (78%) occurred in light slow-

wave sleep (stages 1-2), while 118 (20%) began in REi

sleep. Only 6 events (01%) occurred in slow wave sleep.

These precentages must be considered in light of the

relative distribution of sleep stages in these subjects:

stages 1-2 (64%), stage REM (12%), stages 3-4 (10%), stage

0 (14%). This distribution suggests that the sleep time

spent in stages, 1, 2, or REM must have provided the

overwhelming preponderance of events in these stages.

Sleep characteristics will be discussed in greater detail

below.

As noted below, 28 of the 46 subjects experienced at

least one episode of apnea or hypopnea. Table 3-2

presents the data on occurrence of at least one

apnea/hypopnea, as well as the occurrence of a high level

of apnea/hypopnea. Almost 2/3 of the sample had at least

one apnea or hypopnea, while 13% had high levels of









Table 3-2. Incidence of low and high levels of
apnea/hypopnea by age in 46 snoring males.



Low Apnea/ High Apnea/
Age Group N Hypopnea Hypopnea

30-39 14 .71 0
40-49 7 .57 .29
50-59 11 .72 .27
60+ 14 .43 .07
Overall 46 .62 .13



apnea/hypopnea. Note that subjects with high levels of

apnea/hypopnea are included in the overall total of

subjects with at least one event for the following

analyses. The classification of "high" levels of

apnea/hypopnea utilizes the criterion of an apnea +

hypopnea index greater than 5, a frequently used clinical

cut-off score. The table also presents the frequency of

respiratory distress broken down by age. No simple age

related trend is apparent, and a X2 analysis of the two

factors of age and respiratory distress was not signi-

ficant for those with at least one event (X2=3.2; p=NS),

or for those with high levels of apnea/hypopnea (X2=5.9;

p=KS).

A further investigation of possible age by level of

apnea activity involved division of the sample into those

above ano below 60 years of age crossed with those above

and below 5 events/hour. The X2 value for this comparison

was also nonsignificant (X2=0.6; p=NS). Subjects were









also divided into those above ano below the mean weight:

height ratio (2.7) and crossed with those with and without

at least one event or high levels of apnea/hypopnea. Both

these non-parametric tests were also non-significant

(X2=.01; p=lS; X2=1.5; p=HS).

lablt 5-3 utilizes Pearson correlations to explore

possible relationships between demographic variables (age,

weight, weight:height ratio) and nocturnal respiratory

indices (apnea index, apnea + hypopnea index, seconds in

events, mean hiFh saturation, mean low saturation, number

of desaturations >_4, and number of desaturations >10Q).

Again, age displays little relationship with respiratory

indices, but substantial correlations between weight and

various respiratory indices are apparent.




Table 3-3. Significant (p<.05) Pearson correlations between
nocturnal respiratory variables ana demographic
variables in 46 snoring males.



Respiratory Weight:Height
Variable Ape Weieht Ratio

Number of Apneas
Lumber of Apneas+Hypopneas .265
Apnea Index
Apnea+hypopnea Index .260 .506*
Mean High Saturation -.323* -.363*
Mean Low Saturation -.397* -.460*
Mean Saturation Change .319* .373*
Number of Desaturations >4 .267 .368* .450*
lumber of Desaturations >10


*p<.01









In summary, non-parametric analyses failed to reveal

significant relationships between age or weight and

overnight respiration. However, Pearson correlations

suggested that increasing weight is associated with

deterioration of several noctural respiratory indices.

Because increasing apnea/hypopnea levels have been

suggested to carry increasing risks for various deficits,

the sample was stratified by level of apnea/hypopnea. The

first group included those subjects without events (No

events), those with at least one event but less than 5

events per hour (Low apnea/hypopnea), and those with 5 or

more events per hour (High apnea/hypopnea). It should be

noted that this division of subjects is slightly different

from the earlier non-parametric analyses, because in this

grouping low apnea/hypopnea and high apnea/hypopnea are

mutually exclusive, while in the earlier analyses the

group with at least one apnea/hypopnea included those with

more than 5 events per hour.

Table 3-4 utilizes oneway ANOVA procedures to compare

these groups on demographic and nocturnal respiratory

variables. There were no significant between group

differences on any of the demographic factors: age,

weight, weight:height ratio, and education [F(2,43)=2.1,

0.6,0.8,0.7; p=lIS]. Logically, both apnea and hypopnea

indices were significantly different between groups

[F(2,43)=9.6, p<.CO04; 20.5, p<.0000]. Followup

comparisons (Scheffe) indicated that the high






61
















0 (U CU a N n m

v00 i- 0 -- .
O <0 Z NX
CO Ov


C.>
CO 0 CL ) o OaN




0. ..





C O*
o >> < Qr co O











10.
200.


O "
o 0 m IN- CU .~



aC O O ..
OU 0 JO m C

CCO C0




C UC** O**


SH 0)
SE -- 3 m o T- NOC~ro O^^






C D
3J O N
Co ( I C J O-(1J'30 0.U






0 0 0L00 0 c\ 0 r
-> Q CO N 00


C 0 0

CU La a
Dc t co It
Ei > a fu cu O -H l



CU 0 Cd


v 0 L 0 I0. 0 0 0 HU
0 c >Q. H -(




9, Vm "0 u 0 cu 0.'

00. CU C w C0 OU Z o o3
Ili>10 C: CUQu )m L. (U CUE-P
,u Q) a co w w m 0 M 4J-

0. a X 4 C 00 00c
*u-H 00D0. UP)
a)o 0 z Qa o o 7H










w Q -1.0'*-H-aOO Q -41 4-j
00 c0 > a 0 OO -H-H
*H 'a CL 0 O CCQ
*4 -L +*C to (0
c a'o W0W c U C ru)C C c I

C 0E-H-H=E(ECCQ C, WXCU E E
< 00 ) 400 0 C: oo o t o oo C, r









apnea/hypopnea group was significantly different from the

other two groups on these two indices. While mean

duration of apneas and hypopneas was not significantly

different between the high and low apnea groups

[F(1,25)=3.5, 0.1; p=NS], total seconds in respiratory

events was [F(2,43)=50.4, p<.0000], with followup

comparisons showing that the high apnea/hypopnea group was

significantly different from tne other two groups.

Turning to the oxygen saturation data, mean high

saturation failed to reach significance [F(2,43)=0.3,

p=NS]. However, mean low saturation, number of

desaturations >4% and >10% were all significant

[F(2,43)=11.6, p<.0001; 7.0, p<.C02; 21.4, p<.0000], and

followup comparisons showed that the group with high

levels of apnea/hypopnea was significantly different from

the other two groups.

Thus, division of the sample into levels of

apnea/hypopnea resulted in groups which were not

significantly different on demographic variables including

age, weight, weight:height ratio, and education. However,

the group with high levels of apnea/hypopnea had

significantly more nocturnal events, time in events, and

oxygen desaturation.

In summary, the present sample of 46 snoring males

appears representative of those predisposed to experience

sleep disordered breathing (middle-aged, snoring males).

Sixty-two per cent of these subjects had at least one









apnea or hypopnea, while 135 had high levels of

apnea/hypopnea. Most events occurred in light slow-wave

or REM sleep. No effect of age on level of sleep

disordered breathing was observed, although increasing

weight was correlated with nocturnal respiratory events.

When the sample was stratified by level of apnea/hypopnea

activity, subjects with high levels of events had

significantly more events and desaturations than the

remaining subjects, although no significant differences

were observed on age, weight or education.

Nocturnal Respiratory and Health Variables

Health variables including blood pressure readings

and several scores derived from the Cornell Medical Index

(overall number of symptoms endorsed, as well as number of

symptoms endorsed from the respiratory, cardiopulmonary

and neurological subscales) were evaluated initially

through a matrix of Pearson correlations which included

the respiratory variables (apnea index, apnea + hypopnea

index, seconds in events, as well as mean high and low

saturation and number of desaturations >4% and >10%).

From a total of 42 correlations, three were significant at

p<.05 or less: Systolic blood pressure and mean low

saturation (r=-.261); Systolic blood pressure and number

of desaturations >4% (r=.373); Overall number of symptoms

endorsed on the CMI and mean high saturation (r=-.254).

As a further exploration of possible relationships between

blood pressure and apnea/hypopnea activity, subjects were









asked whether they had ever been diagnosed with hyper-

tension. Thirty-three per cent of those with no events

had diagnosed hypertension, as did 33% of those with low

levels of apnea/hypopnea, while only 15% of those with

high levels of apnea/hypopnea had diagnosed hyperten-

sion. These data do not support an increasing level of

hypertension in subjects with high levels of

apnea/hypopnea.

The subjects were grouped into the earlier delineated

levels of apnea/hypopnea. Table 3-5 presents the means

and standard deviations of health variables from the three

groups. No dramatic trends are apparent in this table,

and oneway ANOVA procedures confirmed this observation

with no significant differences for: Diastolic BP,

F(2,43)=.01, p=NS; Systolic BP, F(2,43)=.1, p=NS; CMI

overall, F(2,42)=.2, p=liS: CMI respiratory, F(2,42)=.5,

p=NS; CMI cardiopulmonary, F(2,42)=.5, p=NS; CMI

neurological, F(2,42)=2.2, p=NS.

In summary, while correlational procedures suggested

relationships between several oxygen saturation measures

and blood pressure as well as overall number of symptoms

endorsed on the CMI, oneway ANOVAs failed to reveal

significant between group differences when subjects were

stratified by level of nocturnal distress.









Table 3-5. Means and standard deviations for health related
variables in 46 snoring males grouped by level
of apnea/hypopnea.



No Apnea/ Low Apnea/ High Apnea/
Hypopnea Hypopnea Hypopnea

Systolic BPa 129.1(14.2) 128.7(12.6) 127.0(7.5)
Diagtolic BP 83.2(13.4) 83.9(11.0) 81.6(7.6)
CMIU Overall Score 14.7 (7.4) 16.1(10.5) 13.8(7.4)
CMI Respiratory Score 1.4 (1.2) 1.1 (1.8) 2.0(1.4)
CMI Cardiac Score 1.6 (1.3) 1.2 (1.2) 1.0(1.2)
CMI Neurological Score 1.2 (0.9) 1.6 (1.5) 0.8(0.8)

SBP = blood pressure.
b CMI = Cornell Medical Index.


Nocturnal Respiratory and Sleep/Wake Variables

EEG Data

Electroencephographic variables were recorded

overnight for all subjects. Records were scored by a

trained technician who utilized the system of Arnew and

Webb (1972) to determine sleep stages. In three records,

EEG tracings were judged inadequate for differential sleep

staging. However, a judgement of sleep vs. wake was made

for these subjects in order to form a basis for the

respiratory indices. These three EEG records were

excluded from the following analyses, leaving 43 records

as a base for analysis.

Fifteen variables (time in bed, pure sleep time,

sleep efficiency index, number of stage 0 periods, time E

stage 0, sleep latency, time % stages 1, 2, 3, 4, and RLM,

latency 1st REM period, mean REM period length, mean REM









cycle length, 0 slow wave sleep), reflecting multiple

aspects of the sleep of these subjects, were derived from

the sleep stage scoring. Table 3-6 presents the means ana

standard deviations for these subjects. Although subjects

spent over six hours in bed, the table indicates that not

quite 5 hours were spent asleep. The remaining variables

suggest that an increase in sleep latency, and time awake,

is responsible for the diminished sleep time, while

lighter sleep is also increased.



Table 3-i. e!eans and standard deviations of selected sleep
variables from 43 snoring males.



Standard
Mean Deviation

Time in Bed 377.3 45.1
Sleep Latency 29.2 32.9
Pure Sleep Time 288.7 75.8
Sleep Eificiency Index .76 .16
Number Stage 0 Periods 8.6 5.6
Time ? Stage 0 14.3 12.7
Time % Stage 1 3.3 2.8
Time 4 Stage 2 60.8 10.3
Time % Stage 5 3.4 2.4
Time % Stage 4 7.9 6.4
Time % Stage REN 11.5 7.2
Latency 1st HEI Period 124.7 92.2
Mean REIi Period Length 14.9 9.4
RL! Cycle Length 74.9 57.7
% Slow Wave Sleep 12.9 8.3


Note: see text for explanation of N.


Partial correlations, controlling for age, were

calculated between these sleep and the respiratory

variables. Out of a total of 105 correlations, only one









was significant at p<.05 or less: Number of desaturations

>4% vs REM latency (r=-.369).

Table 3-7 presents the sleep variable means for the

subjects grouped by level of apnea/hypopnea. The group

with high apnea/hypopnea appears to spend more time awake,

less time asleep, and achieve less slow wave sleep than

the others. However, none of these differences proved

significant. In fact, only one out of 15 oneway ANOVAs

achieved significance: Time in bed, F(2,40)=0.4, p=NS;

Sleep latency, F(2,40)=1.7, p-iS; Pure sleep time,

F(2,40)=0.3, p=IS; Sleep efficiency index, F(2,40)=0.2,

p=nlS; Number of stage C periods, F(2,40)=0.5, p=NS; Time

5 stage 0, F(2,40)=1.3, p=NS; Time % stage 1,

F(2,40)=1.3, p=NS; Time % stage 2, F(2,40)=0.04, p=NS;

Time % stage 3, F(2,40)=3.64, p<.04; Time % stage 4,

F(2,40)=1.32, p=NS; Time % stage REM, F(2,40)=1.5,

p=NS: Latency 1st REM, F(2,40)=0.8, p=NS; Mean REM

period length, F(2,40)=0.3, p=NS: Mean REM cycle length,

F(2,40)=0.9, p=NS; slow wave sleep, F(2,40)=2.4,

p=NS. Scheffe followup procedures for time % stage 3

revealed no significant between group comparisons.

To summarize, EEG sleep measures indicated lighter

sleep then normal in this sample, with increases in

awakenings and light slow wave sleep observed. Neither

between group comparisons nor correlational procedures

revealed significant interactions between sleep and

nocturnal respiratory variables.






68














0 O ^^-LnOOrnri^-

C7 1m1;N2rNtavNC



o 0c f













to . .
*r41 cu^





H nUN O O =3 CNc CO -cz
0. . .

0(U n N n C- N K C- %D Z C -' C

ID (U. c 0 X
co 0 1-J N m nL-
C O3 Q 0 * ., .


- 4q
u0 0








0. C
C. -
.cu Q

4-1 a)
au ra 'Z
o 'oQo H-t-- c )i o









cu w-


C.;F=r0 N 0t
10 N1 a
'0 00 O4- 0






000 ti Oo
)o a c0 4 -

0W > *.
a0 C 004 P 0+

0a- d 0 J2 0



0W"ICCO 0nCiQ)0 Z (U


) a C0 0C *0
i 00000000000C
ac C) ma0aaa )w











*P(U -102 0 0E) 0
1; .c, E EE F ao



n E aUCLWZ) E- E-4 E L, oZ C: H
*o~~~~~~z t: u a i im iiau u









Daytime Sleepiness Data

Included among the measures of daytime sleepiness

were the mean of the several Stanford Sleepiness Scale

ratings, and two variables calculated from the two evening

naps: mean sleep latency and total minutes asleep during

the naps. These variables were correlated with the

respiratory variables, resulting in a Pearson correlation

matrix. From these 21 correlations, two emerged signi-

ficant at p<.05 or less, both involving the mean Stanford

Sleepiness Scale ratings. Mean SSS ratings were corre-

lated with apnea index (r=.297), and apnea + hypopnea

index (r=.283).

Table 3-8 presents the means and standard deviations

for the subjects grouped by level of apnea/hypopnea. One-

way ANOVA procedures were non-significant for all these

variables [SSS: F(2,38)=2.4, p=NS; Mean sleep latency:

F(2,35)=0.4, p=NS; Total minutes asleep: F(2,35)=0.3,

p=NS].



Table 3-8. Means and standard deviations for daytime
sleepiness variables in 46 snoring males grouped
by level of apnea/hypopnea.


ho Apnea/ Low Apnea/ High Apnea/
Hypopnea Hypopnea Hypopnea
Stanford Sleepiness
Scale 3.2 (.7) 3.6 (.8) 4.0 (.7)
Mean Nap Latency 14.5 (5.7) 15.5 (5.3) 13.4(3.4)
Total Miinutes of
Sleep in Naps 11.3(11.9) 9.6(10.0) 13.5(6.6)









In summary, correlational procedures suggested

relationships between nocturnal respiratory variables and

subjective sleepiness, while no reliable between group

difference were noted on the nap or SSS variables.

Subjective Sleep Assessment

Two assessments of subjective sleep parameters were

completed, including the sleep questionnaire and sleep

log. Several variables were selected from the sleep

questionnaire and included: usual hours of sleep,

judgement of whether usual sleep was adequate, number of

naps per week, hours napping per week, number of nightly

wakenings, minutes of nightly wakenings, usual sleep

latency, depth of sleep, and daytime sleepiness. These

variables were correlated with the respiratory variables,

for a Pearson matrix of 63 correlations. Table 3-9

presents the significant correlations which emerged from

this matrix. Several respiratory variables were

associated with napping variables, while mean low

saturation and number of desaturations >4% were associated

with reported minutes of nightly wakenings.

Table 3-10 presents the means and standard deviations

from the subjects grouped by level of apnea/hypopnea.

Oneway ANOVA procedures revealed no significant between

group differences on the selected sleep questionnaire

variables [Usual hours sleep, F(2,43)=G.6, p=NS; Adequate

sleep, F(2,43)=0.3, p=NS; number of naps, F(2,43)=2.9,

p=NS; Hours napping, F(2,43)=0.8, p=NS; Number of











Table 3-9. Significant (p<.05) Pearson correlations between
nocturnal respiratory and sleep questionnaire
variables in 46 snoring males.



Minutes
Number hours Nightly
of Naps Napping Awakening

Apnea Index
Apnea+Hypopnea Index .249
Seconds in Events .331
Pean High Saturation
Mean Low Saturation .307
Lumber of Desaturations >4% .283 -.291
number of Deseturations 10% .403 .376










Table 3-10. Means and standard deviations of sleep
questionnaire variables derived from 46 snoring
males grouped by level of apnea/hypopnea.



No Apnea/ Low Apnea/ High Apnea/
Hypopnea hypopnea Hypopnea
eean h. Sleep 7.4 (.9) 7.0(1.6) 7.5(1.0)
Adecuate Amount? 1.9 (.4) 1.8 (.4) 2.0(0)
Number of laps/Week 1.8(1.5) 2.3(2.6) 4.3(2.2)
hours Napping/Week 2.6(1.8) 3.5(2.7) 4.0(2.7)
Dumber of Nightly
lakenings 1.1 (.9) 1.5(1.3) 1.6 (.4)
Minutes Nightly
Awakenings 1.9(1.1) 1.5(1.0) 1.5 (.8)
Sleep Latency 2.0 (.9) 2.0 (.6) 2.0 (.6)
Depth of Sleep 2.4 (.8) 2.1 (.9) 3.0(C)
Daytime Sleepiness 1.6 (.6) 1.9 (.4) 1.8 (.4)









awakenings, F(2,43)=0.9, p=LS: Ilinutes of awakening,

F(2,43)=0.8, p=NS; Sleep latency, F(2,43)=0.01, p=NS;

Depth of sleep, F(2,43)=2.6, p=NS; Daytime sleepiness,

F(2,43)=1.8, p=NS].

Means were calculated from the 7 day sleep logs, and

the following variables were selected for analysis:

reports of daytime sleepiness, number of naps, minutes

napping, total bedtime, sleep latency, number of

awakenings, minutes of awakenings. These variables were

correlated with the respiratory indices to form a Pearson

correlation matrix totalling 49 correlations. Table 3-11

presents the 6 significant (p<.05) correlations which

emerged. Again, number of desaturations were associated

with reports of napping, while mean high saturation was

associated with total bedtime, sleep latency, and minutes

awake, and mean low saturation was associated with minutes

awake as well.

Table 3-12 presents the means and standard deviations

for the subjects divided into levels of apnea/hypopnea.

Oneway ANCVA procedures failed to reveal significant

between group differences [Daytime sleepiness,

F(2,35)=0.6, p=NS; Number of naps, F(2,35)=0.3, p=NS;

Minutes of napping, F(2,35)=1.3, p=NS; Total bedtime,

F(2,35)=0.7, p=NS; Sleep latency, F(2.35)=0.2, p=NS;

Number of awakenings, F(2,35)=0.2, p=NS; Minutes of

wakenings, F(2,35)=.0001, p=NS].












Table 3-11.


Significant (p<.05) Pearson correlations between
nocturnal respiratory and sleep log variables in
46 snoring males.


Number Minutes
of Total Sleep Nightly
Naps Bedtime Latency Wakenings

Apnea Index
Apnea+Hypopnea Index
Seconds in Events
Mean High Saturation .237 .250 .359
Mean Low Saturation .265
lumber of
Desaturations >4% .240
Number of
Desaturations >10% .240









Table 3-12. Means and standard deviations of sleep log
variables derived from 46 snoring males grouped
by level of apnea/hypopnea.



No Apnea/ Low Apnea/ High Apnea/
hypopnea hypopnea Hypopnea
Daytime Sleepiness 1.9 (.9) 1.9(1.1) 2.4 (.3)
Number of Naps 0.2 (.2) .3 (.4) .3 (.3)
Minutes Napping 0.0(0) .2 (.5) 0 (0)
Total Bed Time 6.5(3.1) 5.8(3.4) 7.5 (.7)
Sleep Latency 10.5(8.1) 11.1(9.8) 8.4(4.7)
Number of Wakenings 1.3(0.8) 1.3 (.9) 1.3 (.2)
Minutes making 2.0(1.4) 1.7(1.1) 1.7 (.3)









In summary, no between group differences were noted

for either sleep questionnaire or log variables when

subjects were grouped by level of apnea/hypopnea.

Correlational procedures suggested relationships between

several respiratory variables and reports of napping and

wakening after sleep onset on the sleep questionnaire.

These findings were cross validated in the correlational

analysis of sleep log data, in which reports of napping

and wakening after sleep onset were related to respiratory

variables, as were total bedtime and sleep latency.

Nocturnal Respiratory and Neuropsychological Variables

Scores derived from the cognitive/neuropsychological

battery included WAIS Performance IQ, WAIS Verbal IQ,

Wechsler Memory Quotient, delayed recall of logical

stories, delayed recall of visual reproductions, delayed

recall of Rey complex figure, digit span, Hooper score,

Wisconsin card sort score, finger tapping right and left,

and Verbal fluency.

As a preliminary analysis, these scores were compared

with selected respiratory and demographic variables in a

multivariate regression procedure. In this type of

analysis, a set of predictor scores is used to

simultaneously predict a set of criterion variables while

controlling for intercorrelations amongst the measures.

An overall test of significance, such as the Hottelings

T2, establishes the probability of obtaining the observed

relationships from chance alone. Upon obtaining a








significant overall p value, univariate regressions are

used to individually predict criterion variables with the

multiple predictors. Contribution of individual

predictors may also be examined in these univariate

regressions.

The results from the multivariate regression of the

above noted cognitive scores and selected nocturnal

respiratory variables (Apnea index, Apnea + hypopnea

index, lean high saturation, Number of desaturations 14%,

Age, Weight:Height ratio) are presented in Table 3-13. An

overall Hotellings T2 score was significant (p<.003),

allowing examination of the univariate regressions.

Criterion variables which were significantly predicted by

the respiratory and demographic variables included WAIS

PIQ, VMS NQ, delayed recall of Rey figure and visual

reproductions, Hooper, and Wisconsin card sort. These

relationships reflected negative relationship between

overnight respiratory indices and neuropsychological

scores. Respiratory indices formed significant components

of the predictor combinations for all these variables save

Hooper score, suggesting relationships between nocturnal

respiratory indices and non-verbal intelligence, verbal

and visual memory, and ability to shift sets while problem

solving.

Because age is correlated with nocturnal respiratory

parameters in past reports, and also with neuropsycho-

logical scores in other samples, the various overnight









Table 3-13. Multivariate regression of demographic and
nocturnal respiratory variables on cognitive
scores in 46 snoring males.


Predictor Variables
Age
Weight:Height Ratio
Apnea Index
Apnea+Hypopnea Index
Mean High Saturation
Number of Desaturations >4%


Criterion Variables
WAIS Verbal IQ
WAIS Performance IQ
Wechsler Memory Scale Memory
Quotient
Delayed Recall Logical Stories
Delayed Recall Visual
Reproduction
Delayed Recall Hey Figure
Digit Span
Hooper Test
Finger Tapping Left Hand
Finger Tapping Right Hand
Verbal Fluency
Wisconsin Card Sort


Multivariate test of significance

Hotellings T2 = 4.841 (p<.003)


Univariate Regression with
Variable
WAIS Verbal IQ
WAIS Perlormance IQ
Wechsler Memory Scale
Memory Quotient
Delayed Recall
Logical Stories
Delayed Recall
Visual Reproductions
Delayed Recall Rey Figure
Digit Span
Hooper Test
Finger Tapping Right Hand
Finger Tapping Left Hand
Verbal Fluency
Wisconsin Card Sort


(6,3e) d.f.
Multiple R
.3646
.5808

.5345

.4713

.6450
.6461
.2432
.6786
.3907
.4270
.4441
.5143


Breakdown of significantly predicted criterion regressions


Criterion
WAIS Performance
10


Predictor
Age
Weight:Height Ratio
Apnea Index
Apnea+Hypopnea Index
Mean High Saturation
Number of
Desaturations >4?'


Eeta
.1240
.0170
-.2636
-.1553
.0920


Sig.
.382
.914
.059
.344
.533


-.3978 .024


F.
.971
3.22

2.53

1.80

4.51
4.54
.39
5.40
1.14
1.41
1.55
2.27


Sig.
.458
.012

.037

.123

.002
.001
.875
.000
.358
.235
.187
.050









Table 3-13-continued.


Criterion
Wechsler Memory
Scale Eiemory
Quotient





Delayed Recall
Visual
Reproductions





Delayed Recall
Rey FiFure






Hooper Test







iisconsin Card
Sort


Predictor
Age
Weight:Height Ratio
Apnea Index
Apnea+Hypopnea Index
Mean High Saturation
Number of
Desaturations >44%

Age
Weight:Height Ratio
Apnea Index
Apnea+Hypopnea Index
Mean High Saturation
Number of
Desaturations >45,

Age
Weight:Height Ratio
Apnea Index
Apnea+Hypopnea Index
Mean High Saturation
Number of
Desaturations >45%

Age
l\eight:Height Ratio
Apnea Index
Apnea+Hypopnea Index
Mean High Saturation
Number of
Desaturations >,4%

Age
Weight:Height Ratio
Apnea Index
Apnea+Hypopnea Index
Nean High Saturation
Number of
Desaturations >45


Leta
.0272
.0996
-.3315
-.3421
.2384

-.50,7

-.3254
-.1214
-.4037
-.2026
.1368

-.2888

-.3550
.1795
-.2206
-.0572
.3191

-.3228

-.5658
.1202
-.1428
.1478
.1072

-.2042

-.3409
.0258
-.3205
.1029
.2383

-.0801


Lig.
.852
.546
.024
.049
.124

.007

.018
.416
.003
.191
.325

.078

.010
.231
.091
.709
.025

.049

.000
.403
.251
.318
.421

.190

.027
.887
.031
.550
.130

.657









scores were entered into a partial correlation matrix,

controlling for age, with the cognitive scores. Table

3-14 presents the resulting correlations significant at

p<.05 or less. The table incicates that every cognitive

test but digit span and Hooper correlated with at least

one overnight respiratory index. Of additional impact is

the observation that in every case of a significant

correlation, deteriorating respiratory indices are

associated with worsening neuropsychological scores.

Respiratory indices with particularly heavy loadings on

neuropsychological scores included those quantifying

apneas, (Apnea index, Apnea + hypopnea index) as well as

those more directly representing deepening hypoxia (Mean

low saturation and Number of desaturations >4%).

Subjects were grouped by level of nocturnal

respiratory disorder (no apneas/hypopneas, subclinical

level of apnea/hypopnea, and clinically significant level

of apnea/hypopnea). Means and standard deviations for the

neuropsychological scores from these groups are presented

in Table 3-15. The table suggests a general deterioration

of cognitive scores as level of nocturnal respiratory

distress increases. Oneway AIIOVA procedures were applied

to the neuropsychological scores from each test. Signifi-

cant between group differences emerged for WAIS PIQ

[F(2,43)=6.9, p<.002] delayed recall of logical stores

[F[2,43)=4.0, p<.03] delayed recall of visual reproduc-

tions [F(2,45)=3.4, p<.04], delayed recall of Hey complex














-OO 00 0










co
0 0
0 r- ?N1



*
I f I I I I I




E C E
C6 CO CUD N .f

*-" U-C- 0 0 L 0N 3


0 C

OCU
S4-) *
C4E o c
0 N 0 0 N o ,

0 I0 I .


0 <;
O C.
1 0 > ZC
4-<

CL











41 r r- o
O -P ( C X *










r-O co 0 9 z
-P C
V C











0U o 0

Q0 0 c O w
*Uc- t C







.H0 CO to
4OCO C40 F- C C0

e 33 a c
C "O 2i C >, a CUUflC
c] CU a 4-1 0 C a) w i


rC o) -r ai 4 W ., M aU
a:ECUE W0 0CU C: CrC

>CO 0 0. O C H- C





E cC 0 0 0 00 C. O *C -C .> *
ar) H 0 H .1 a H > a











Table 3-15. teans and standard deviations of cognitive
variables in 46 snoring males groupec by level
of apnea/hypopnea.


Variable

WAIS Full IQ
WAIS Verbal IQ
WAIS Performance IQ
Weschler Memory
Scale memory
Quotient
Delayed Recall
Logical Stories
Delayed Recall
Visual
productions
Digit Span
Delayed Recall
Rey Figure
Hooper Test
.isconsin Card Sort
Finger Tapping
Right Hand
Finger Tapping
Left Hard
Verbal Fluency


No Apnea/ Low Apnea/ High Apnea/
Hypopnea Hypopnea Hypopnea


124.9(11.2)
123.7(12)
122.8(11.5)


118.3(13.9)
118.5(14.5)
114.5(14.4)


111.5 (9.6)
114.3(10.4)
100.1(12.5)a


124.7(14.3) 117.1(19.0) 107.8(18.7)

9.9 (2.4) 8.6 (3.2) 5.8 (4.0)a


10.5 (5.7) 10.4 (3.4) (.1 (4.7)
6.8 (0.8) 6.8 (1.2) t.6 (.8)


25.6 (5.4)
25.2 (2.6)
4.1 (1.2)


23.2 (6.0)
25.5 (4.8)
3.4 (1.5)


15.1 (8.4)a,b
22.8 (4.1)
2.8(1.6)


62.6 (8.5) 57.5(10.4) 56.1 (8.3)

59.1 (9.0) 54.0 (9.7) 54.8(15.1)
14.6 (4.5) 13.1 (4.2) 9.3 (2.6)e


a Significantly different from no sleep apnea/hypopnea.
b Significantly different from low sleep apnea/hypopnea.








figure [F[2,45)=6.5, p<.OO3], and Verbal fluency,

[F[2,43)=3.6, p<.04]. Followup comparisons (Scheffe)

showed the group with high apnea/hypopnea to be

significantly impaired relative to those with no sleep

apnea/hypopnea on WAIS PIQ, delayed recall of logical

stories and Rey complex figure as well as verbal

fluency. Additionally, the high apnea/hypopnea group was

impaired relative to those with subclinical apnea/hypopnea

on aelayea recall of the Rey complex figure. These

findings are bolstered by the lack of significant between

group differences on age, weight, or education (see Table

3-15), all of which are potential alternative explanations

of the between group differences.

As a check, an analysis was run to evaluate the

relative contribution of nocturnal respiratory variables

and sleep variables reflecting fragmentation in predicting

cognitive/neuropsychological scores. A multivariate

regression of these scores was carried out. Demographic

and respiratory scores included age, apnea index, apnea +

hypopnea index, and number of desaturations >4%. A subset

of the earlier noted sleep variables was chosen to reflect

measures thought to be most responsive to sleep

fragmentation. These included sleep efficiency, number of

stage O periods, time % of stage 0, time % of stage 2, and

number of stage changes. These respiratory and sleep

parameters were used to predict the 5 cognitive scores

which were earlier shown to be predicted by nocturnal








respiratory scores: WAIS PIQ, Wechsler Memory Quotient,

delayed recall of visual reproductions and Hey figure, and

Wisconsin card sort. Table 3-16 presents the data from

this analysis. The Hotellings T2 was significant,

(p<.001) with delayed recall of visual reproductions and

Rey figure significantly predicted by the predictor

scores. Inspection of the individual regressions for

these two measures discloses that age, apnea index, and

number of desaturations >4% all formed significant

components of the prediction of delayed recall of visual

reproductions, while sleep efficiency was the sole

significant predictor of the Key figure. These data seem

supportive of an independent contribution of both

nocturnal respiratory variables and sleep variables at a

limited level, to the prediction of cognitive/neuro-

psychological scores.

In summary, various statistical procedures indicate

relationships between nocturnal respiratory measures and

non-verbal intelligence, verbal and non-verbal memory,

expressive verbal fluency, and ability to shift cognitive

set while problem solving. Division of subjects by level

of nocturnal respiratory distress indicated that subjects

with high amounts of apnea/hypopnea were impaired relative

to controls on measures of non-verbal intelligence, verbal

and non-verbal memory, and expressive verbal fluency, as

well as being impaired relative to subjects with

subclinical levels of apnea/hypopnea on measures of











Multivariate regression of sleep and nocturnal
respiratory on cognitive variables in 43
snoring males.


Predictor Variables

Age
Apnea Index
Apnea+Hypopnea Index
Number of Desaturations >4%
Sleep Efficiency
Number Stage 0 Periods
Time % Stage 0
Time C Stage 2
Lumber of Stage Changes


Criterion Variables

WAIS Performance IQ
Wechsler Memory Scale Memory
Quotient
Delayed Recall Visual
Reproduction
Delayed Recall Rey Figure
Verbal Fluency
Wisconsin Card Sort


Sultivariate test of significance


Hotellings T2 = 4.063

Univariate Regression
Variable
WAIS Performance IQ
Vechsler Memory Scale
Memory Quotient
Delayed Recall Visual
Reproductions
Delayed Recall Rey
Figure
Verbal Fluency
Wisconsin Card Sort


p<.001

with (9,33) d.f.
Multiple R
.610

.566

.655

.668
.578
.453


Breakdown of significantly predicted criterion regressions
Criterion Predictor Beta Sig.
Delayed Recall Age -.3596 .020
Visual Apnea Index -.3218 .021
Reproduction Apnea+Hypopnea Index .2529 .218
Number of
Desaturations >4% -.3778 .041
Sleep Efficiency .2303 .408
Number of Stage 0 Periods .3260 .209
Time % Stage 0 -.1136 .731
Time % Stage 2 .0005 .998
Number of Stage Changes -.2235 .391


Table 3-16.


F.
2.17

1.73

2.91

2.96
.85
.95


Si .
.051

.121

.012

.011
.096
.499









Table -1I-continuec.


Criterion
Delayed recall
Key Figure


Predictor
Age
Apnea Index
Apnea+Hypopnea Index
Lumber of
Desaturations >4%
Sleep Efficiency
lumber of Stage 0 Periods
Time S Stage 0
Time % Stage 2
Number of Stage Changes


Beta
-.2623
-.1519
-.3286

-.0795
.6832
.2487
.4559
-.0465
-. 054


.074
.259
.085

.b57
.016

.172
.802
.057








non-verbal memory. Addition of variables reflecting sleep

fragmentation to the analysis indicated that both

nocturnal respiratory and sleep variables make significant

and independent contributions to prediction of

cognitive/neuropsychological scores.














CHAPTER FOUR
DISCUSSION

Demographics and Incidence of
Nocturnal Respiratory Disorder

The present sample of 46 snoring males was selected

as a group likely to display a wide range of sleep

disordered breathing. These subjects, with a mean age of

50 years and a mean weight of 190 pounds, reflect a mostly

middle-aged to older male population which is moderately

obese. The demographic characteristics of this sample are

similar to those of the "typical" sleep apnea syndrome

patient (Guilleminault et al., 1978), although the

requirement of general good health insured that subjects

with a classic SAS were excluded. This sample may be

considered representative of subjects predisposed to sleep

disordered breathing, although at no more than a

subclinical level. Therefore it seems valid to consider

various risk factors as they relate to nocturnal

respiration.

In the present sample, 62% experienced at least one

apnea or hypopnea, while 13% suffered higher levels of

apnea/hypopnea. Only two other reports have described

samples selected for snoring or respiratory complaints.

Ancoli-Israel et al. (1981) examined 24 subjects with

complaints of respiratory difficulties, daytime








sleepiness, and muscular events (nocturnal myoclonus).

Further, these subjects had a mean age of almost 70 years,

indicating significant sampling and age differences from

the present study. Ninety-three per cent of these

subjects had at least one event, while 25% had high

numbers of apnea/hypopnea. These rates are much higher

than those found in the present sample. This is probably

attributable to the sampling differences noted above.

However, a recent report by Miles and Simmons (1984)

described apneas in a series of 190 patients referred for

complaints of heavy snoring. In this sample, 73% had at

least one event, while 9% had severe levels of apnea.

These data are quite comparable to the prevalences

observed in the present sample, 62% for at least one event

and 13i for high levels of apnea/hypopnea, and lend some

confidence to the present finding.

Thus the overall frequency of sleep disordered

breathing in middle-aged subjects with heavy snoring

appears to be between 62 and 75% for at least one event,

and between 9 and 135 for more severe levels of apnea.

The relatively close concordance of the two reports on

heavy snoring subjects is reassuring.

Factors which were found to be related to the

occurrence of sleep apneas in the previous review included

male sex, complaints of snoring, increasing age, and

obesity. The first two factors could not be examined in

the present sample because of their use as selection









criteria. A wide range of age was achieved in this

sample, but neither non-parametric nor correlational

procedures revealed significant relationships between

increasing age and sleep disordered breathing. Although

several previous reports have noted relationships between

increasing age and sleep disordered breathing, typically

this relationship is moderate for apnea (Block et al.,

1979; Carskadon et al., 1980; Bixler et al., 1982).

Additionally, no other correlational data are available in

subjects selected for sleep and breathing complaints. It

may be speculated that the effect of snoring masks the

effect of age in this group. Alternatively, some quirk of

sampling may have generated this effect. Correlational

procedures, however, revealed substantial relationships

between weight and sleep disordered breathing. This

finding is consistent with several past reports. To

summarize, the present data support relationships between

increasing weight and nocturnal events, while

relationships between age and sleep disordered breathing

are not found in this sample of heavy snoring males,

perhaps because the effect of age is masked by the effect

of snoring.

The temporal characteristics of the apneic events

assessed by two hour interval suggested an increase in

events between 3 and 5 am, although differences in sleep

onset time and wakeup time render any interpretation of








these data somewhat tenuous. This finding needs

replication before any conclusions may be drawn.

Ninety-eight per cent of apneic events occurred in

light slow wave and REM sleep, an observation which is not

surprising given that subjects spent nearly 80% of their

sleep in these stages. The preponderance of events in

these stages of sleep is consistent with the bulk of past

reports in this area (e.g, Bixler et al., 1982; Block et

al., 1979; Krieger et al., 1985).

When subjects were divided into groups by level of

apnea/hypopnea, significant differences emerged on oxygen

desaturation variables. Specifically, comparisons

indicated that the group with a high level of

apnea/hypopnea events experienced deeper desaturations

than the remaining two groups. This difference occurred

despite a lack of differences on age and weight among the

three groups, and appears to implicate the respiratory

events in the genesis of the desaturations. This is

consistent with an earlier report by McGinty et al. (1982)

who demonstrated that subjects with significant levels of

sleep disordered breathing were impaired on oxygen

desaturation variables, relative to subjects without

respiratory events. Subjects with heavy snoring and

multiple nocturnal apneic respiratory events appear to be

at risk for nocturnal oxygen desaturation.

In summary, the present sample of 46 snoring males

appears representative of those predisposed to experience









sleep disordered breathing (middle-aged, snoring males).

A prevalence of 62k was observed for occurrence of at

least one event, while 13% had high numbers of

apnea/hypopnea, data which are consistent with an earlier

report drawn from heavy snoring males. Most events

occurred in light slow wave or REM sleep. No effect of

age on level of sleep disordered breathing was observed,

although increasing weight was linked to nocturnal

respiratory events. When the sample was divided by level

of apnea/hypopnea, subjects with high numbers of

apnea/hypopnea suffered significantly exacerbated oxygen

desaturation, although no significant differences were

noted for weight or age among the groups.

ijocturnal Respiratory and Health Variables

The finding of only very limited relationships

between nocturnal respiratory events and health variables

must be tempered by an acknowledgement of possible bias

against such findings in this sample. Specifically, the

sample was selected to be normal and healthy, a criterion

which undoubtably restricted the range of possible health

pathology, and hence probably attenuated the correlational

findings. Additionally, the self-report health checklist,

the Cornell Medical Index, cannot be regarded as a

powerful measure of health pathology. In a very real

sense then, the odds were against finding reliable

relationships between nocturnal respiratory and health

variables.








In light of these caveats, it is of interest that two

measures of desaturation, mean low saturation, and number

of desaturations >04 were associated with increasing blood

pressure readings. While a report by Guilleminault et al.

(1978) noted an increasing prevalence of hypertension in

subjects diagnosed with sleep apnea syndrome, a

relationship between desaturation and hypertension has not

been previously reported. This finding was observed in

spite of the fact that subjects prescribed hypertensives

were not medication free for the pressure reading.

Additionally, a check of self reports of diagnosed

hypertension showed six subjects in the no event group as

having diagnosed hypertension, as opposed to six in the

low apnea/hypopnea group, and one in the high

apnea/hypopnea group, data which did not link

apnea/hypopnea, per se, with hypertension in these

subjects. The isolated finding of a relationship between

mean high saturation and overall number of symptoms

endorsed on the CMI is more puzzling and does not follow

any logical analysis.

In summary, the present study's assessment of health

variables must be regarded as somewhat limited. In spite

of this, an association between two measures of desatura-

tion and a measure of blood pressure was noted. It is

speculated that blood pressure may be more responsive to

nocturnal oxygen saturation tnan to the apneic events

alone.








Nocturnal Respiratory and Sleep/Wake Variables

EEG Data

The overnight sleep in the laboratory for these

snoring subjects was noteworthy for a general shift

towards lighter sleep than usual. A longer sleep latency,

more awakenings, more light sleep, and less slow wave and

REM sleep than normal are apparent. These findings

parallel the report by Block et al. (1979) who found

shifts to lighter sleep and more awakenings than in a

sample of normal subjects. These changes were said to

result from a first night in a strange bed and the

discomfort of the multiple clips and sensors used to

monitor respiration. Recording montages were identical

for the present subjects, and these two factors

undoubtably influenced the sleep of the present subjects

as well.

An intriguing observation was a lack of covariation

between respiratory and EEG sleep variables derived from

the recording night. This finding is counterintuitive, as

one would expect arousals from the respiratory events to

influence the sleep characteristics of subjects with

events. However, correlational procedures failed to

support this notion. It may be that a ceiling effect was

exerted on sleep measures sensitive to disturbance, one

that resulted from the first night effect compounded by

the discomfort of the recording procedure. The already

elevated disturbance sensitive sleep variables might not




Full Text
128
Hooper, H. (1958), The Hooper Visual Organization Test;
Hanue 1. Los Angeles; Western Psychological Services.
Kales, A., Bixler, E., Soldatos, C., Vela-Bueno, A.,
Caldwell, A., & Cadieux, R. (1982). Insomnia: The
role of sleep apnea and nocturnal myoclonus.
Psychosomatics, 25 187-195*
Kreiss, P., Kripke, D., & Ancoli-Israel, S. (1982).
Prevalence of sleep apnea among representative
inpatients on a veterans administration medical ward.
Abstract published for 1982 APSS meeting, San Antonio,
Texas.
Krieger, J., Turlot, J., Margin, P., & Kurtz, D. (1585).
Breathing during sleep in normal young and elderly
subjects: Hypopneas, apneas, and correlated factors.
Sleep, 6^(2), 106-120.
Krop, H., Block, A., & Cohen, E. (1975). neuropsycho
logical efiects of continuous oxygen therapy in chronic
obstructive pulmonary disease. Chest, 64, 517-522.
Krop, II., Block, A., Cohen, E., Croucher, H., & Schuster,
J. (1977). Neuropsychological effects of continuous
oxygen therapy in the aged. Chest, 72, 757-745*
Lugaresi, E., Cirignotta, F., Coccagna, G., & Baruzzi, A.
(1982). Snoring and the obstructive sleep apnea
syndrome. EEG, Suppl No. 55, 421-450.
Lugaresi, E., Cirignotta, F., Coccagna, G., & Piaa, C.
(1980). Some epidemiological data on snoring and
cardiocirculatory disturbances. Sleep, _3(3/4), 221-224.
Lugaresi, E., Coccagna, G., Berti, G., & Mantovani, M.
(1968). La "maledizione di Ondine" il disturbo del
respiro del sonno neil ipoventalizione alveolari
primaria. List Neurologie, 20, 27-37.
McFarland, R. (1937). Effects of partial acclimitization
on psychological tests at high altitudes. Journal of
Comparative Psychology, 23, 227-230.
McGinty, D., Littner, M., Beahm, E., Ruiz-Primo, E., Young,
L., & Sowers, J. (1982). Sleep related breathing
disorders in older men: A search for underlying
mechanisms. Neurobiology of Aging, 337-350.
leehl, P. (1977). Psychodiagnosis. New York: MAI. Norton
and Co.


62
apnea/hypopnea group was significantly different; from the
other two groups on these two indices. While mean
duration of apneas and hypopneas was not significantly
different between the high and low apnea groups
[F(1,25)=3*3, 0.1; p=NS], total seconds in respiratory
events was [F(2,43)=50.4, pC.0000], with followup
comparisons showing that the high apnea/hypopnea group was
significantly different from tne other two groups.
Turning to the oxygen saturation data, mean high
saturation failed to reach significance LF(2,43)=0.3,
p*NS]. However, mean low saturation, number of
desaturations and >10% were all significant
[F(2,43)=11.6, p<.0001; 7-0, p<.C02; 21.4, pC.0000], and
followup comparisons showed that the group with high
levels of apnea/hypopnea was significantly different from
the other two groups.
Thus, division of the sample into levels of
apnea/hypopnea resulted in groups which were not
significantly different on demographic variables including
age, weight, weight:height ratio, and education. However,
the group with high levels of apnea/hypopnea had
significantly more nocturnal events, time in events, and
oxygen desaturation.
In summary, the present sample of 46 snoring males
appears representative of those predisposed to experience
sleep disordered breathing (middle-aged, snoring males).
Sixty-two per cent of these subjects had at least one


76
Table 3-13. Multivariate regression of demographic and
nocturnal respiratory variables on cognitive
scores in 46 snoring males*
Criterion Variables
VAIS Verbal IQ
VAIS Performance IQ
Wechsler Memory Scale Memory
Quotient
Delayed Recall Logical Stories
Delayed Recall Visual
Reproduction
Delayed Recall Rey Figure
Digit Span
Hooper Test
Finger Tapping Left Hand
Finger Tapping Right Hand
Verbal Fluency
Wisconsin Card Sort
Multivariate test of significance
Hotellings T^ = 4.841 (p<*003)
Univariate Regression with
(6,3e) d.i.
Variable
Multiple R
F.
Sig.
VAIS Verbal IQ
.3646
.971
.458
WAIS Performance IQ
.5808
3.22
.012
Wechsler Memory Scale
Memory Quotient
5345
2.53
.037
Delayed Recall
Logical Stories
.4713
1.80
.123
Delayed Recall
Visual Reproductions
.6450
4.51
.002
Delayed Recall Rey Figure
. 6461
4.54
.001
Digit Span
.2432
.39
.875
Hooper Test
.6786
5.40
.000
Finger Tapping Right Hand
.3907
1.14
.358
Finger Tapping Left Hand
.4270
1.41
.235
Verbal Fluency
.4441
1.55
.187
Wisconsin Card Sort
.5143
2.27
.050
Breakdown of significantly
predicted criterion regressions
Criterion Predictor
£eta
Sig.
WAIS Performance Age
.1240
.382
10 Weight
¡Height Ratio
.0170
.914
Apnea
Index
-.2636
.059
Apnea+Hypopnea Index
-.1553
.34 4
Mean High Saturation
.0920
.533
Number
of
Desaturations >A%
-.3970
.024
Predictor Variables
AgG
Weight¡Height Ratio
Apnea Index
Apnea+Hypopnea Index
Mean High Saturation
Number of Desaturations >A%


40
evidence which has been collected on neurotransmitter
alterations in hypoxia. In brief, mild to moderate
hypoxia, cerebral levels of ATP are normal, implying that
energy supplies of neurons are not disturbed. In con
trast, the turnover of certain neurotransmitters is
altered. Gibson et al. note that oxygen is an important
substrate for the synthesis of several neurotrans
mitters. V/hen rats are exposed to acute hypoxia, syn
thesis of dopamine and serotonin is decreased, while
absolute levels of tryosine, tryptophan, catecholamines,
and serotonin remain constant. Continuing hypoxia leads
to an adaptation resulting in normal turnover of these
neurotransmitters. Gibson and Blass (1976) examined the
affect of acute hypoxia on central acetylcholine (ACh).
As in dopamine (DA) and serotonin (5-HT), acute hypoxia
results in a decrease in synthesis in ACh, but not in the
absolute level of ACh. Thus there is a substantial body
of evidence indicating changes in central neurotransmitter
levels in hypoxia. Generalizing from the animal data to
human subjects, Gibson and Duffy (1981) suggest that these
changes underlie the behavioral deficits of hypoxia in
humans (which will be reviewed below).
Other variables have been examined in acute hypoxia
in animals. Annau (1972) found that rats exposed to brief
hypoxia exhibited decreased appetite, thirst, and self
stimulation rates. Annau observed that brief hypoxia
depressed all behavioral variables. Gellhcrn (1951) found


104
fluency, and cognitive flexibility were observed. It was
speculated that hypoxic alterations in the cholinergic
synthetic pathways underlayed these changes. Apparently,
a significant subgroup of heavy snoring males fall on the
same continuum as subjects with sleep apnea syndrome,
occupying an intermediate position between normals and
those with SAS. From these and other data, it is
concluded that
1) between 62 and 73% of snoring males experience at
least one nocturnal respiratory event, while
between 9 and 13% appear to experience 5 or more
events per hour;
2) most events occur in light slow wave or REM
sleep;
3) increasing obesity is linked to nocturnal events
in snoring males;
4) nocturnal events are linked to increasing blood
pressure in snoring males;
5) nocturnal events are linked to increasing sleepi
ness in snoring males;
6) nocturnal events are linked to deteriorating
neuropsychological scores in snoring males;
7) a subset of heavy snoring men, those with more
than 5 events per hour, suffer increased oxygen
desaturation and decreased cognitive abilities in
the areas of non-verbal intelligence, verbal and


41
that brief hypoxia of 4-7*5% O2 abolished cortical re
sponses to auditory stimulation. Sara (1974) trained rats
in active one way avoidance tasks. Immediately after
training, subjects were exposed to 3.5-4.0# oxygen. Rats
tested at 1 and 3 hours post hypoxia avoided correctly,
while those tested at 6 and 24 hours did not, leading Sara
to suggest that hypoxia resulted in a memory retrieval
deficit. Freides and Allweise (1978) also trained rats in
one way avoidance and exposed them to 2,0% hypoxia imme
diately following training. These animals avoided at 12
and 200 minutes post hypoxia, but not at 90 minutes.
Friedes and Allweise suggested that a sensitive period for
hypoxia effects on training existed for 15 minutes after
learning. Tauber and Allweise (1975) followed the same
training/hypoxia paradigm and found avoidance impaired at
2.5 but not 4.5 hours. They suggest that hypoxia inter
feres with a medium term memory. While some of these data
are conflicting, it appears that acute hypoxia depresses
certain behavioral variables and interferes with aspects
of memory processes.
While caution is in order when generalizing from
animals to humans, several seemingly relevant points
emerge from the review. Behaviorally, simple activities
may be spared, although some depression of appetitive
activities may occur, while more complex behavior depen
dent on memory processes may be altered. Of particular
interest are the demonstrated alterations in


112
labie B-2-continued.
Mean Stanford
Mean Sleep
Minutes of
Subject
Sleepiness
Latency
Sleep
a
Hating
j- Naps
- Naps
42
1.5
06.0
30
43
2.5
-
-
44
3.8
05.5
35
43
3.4
-
-
46
2.1
-
-


2
noted in SAS patients. The link between apneas and patho
logy may be more complex than was originally thought,
given its apparently benign occurrence in otherwise
asymptomatic subjects.
The present study will examine the correlates of
sleep apnea episodes in a population selected to display a
wide range of such events. The relationship between sleep
apnea indices and sleep characteristics, age, weight and a
variety of deficit measurements will be analyzed.
A review of the current data available on various
aspects of sleep apneas and sleep apnea syndromes will be
presented as background for the study. The diagnosis,
history, incidence, and characteristics of sleep apneas
and sleep apnea syndrome will be reviewed, followed by a
discussion of variables associated with sleep apneas.
Lastly, the deficits commonly found in sleep apnea syn
drome will be detailed, along with measurements thought to
be sensitive to these deficits.
Definitions and Terminology
Sleep apneas have been defined as cessations of air
flow at the mouth and nose which last for 10 or more
seconds (Guilleminault et al., 1976). In a sleep apnea
syndrome (SAS), multiple sleep apneas are accompanied by
snoring, excessive daytime sleepiness, cardiopulmonary
complications, and altered states of consciousness. The
ease of quantification of sleep apneas has led to their
use as an important tool in diagnosing SAS. Guilleminault


35
Measurement of Arousal Deficits (Hypersomnolence)
Excessive daytime sleepiness is a common sequela of
sleep apnea syndrome. Measurements of the sleep/wake
cycle of subclinical apnea patients thus seems indi
cated. A sleep questionnaire assessing trait sleep habits
seems in order, as well as 1 week sleep logs following the
experimental night. The sleep questionnaire and sleep
logs have been used in ongoing research in W.E. Webb's
studies of aging and sleep and seem an adequate assessment
of sleep patterns of subclinical apnea subjects. Addi
tionally, electroencephalographic recordings of sleep
during the experimental night seem in order using a stan
dard recording and scoring system (Agnaw and Webb,
1972). The assessment of daytime sleepiness is relatively
new, only recently being the subject of investigation.
William Dement and associates began to develop methods of
quantifying somnolence in the early 1970s. The first
attempt produced a subjective rating scale, the Stanford
Sleepiness Scale (SSS; Hoddes et al., 1973)* The SSS was
a 7 point Likert scale with endpoints anchored on 1) feel
active and vital; and 7) almost in reverie, sleep onset
soon; with 5 points with statements reflecting various
degress of somnolence in between. Patients rated their
introspective level of sleepiness on this scale.
Unfortunately, the SSS proved to be vulnerable to exactly
the same difficulties as a global rating scale used on SAS


Table b-3
¡bleep questionnaire variables from 46 snoring males.
Subject
a
Mean
Hours
Sleep
Enough
?
Humber
of
Maps
Hours
Happing
Humber
of
Wakings
Minutes
Awake
Sleep
Latency
Depth
of
Sleep
Daytime
Sleepiness
1
7
2
4
4
2
0
1
3
2
2
o
2
6
5
1
0
1
*
>
2
3
7
1
"Z
J
3
0
2
2
3
2
4
8
2
0
0
1
2
2
3
2
5
6
2
0
0
0
0
3
3
2
6
7
2
1
2
0
0
2
3
2
7
8
2
0
0
1
2
3
1
2
8
1
3
2
3
3
2
1
3
2
9
8
2
7
4
3
2
2
2
2
10
9
2
2
2
2
2
3
2
2
11
7
2
1
1
1
1
2
2
2
12
'7
i
2
0
0
4
1
2
1
1
13
1
2
2
3
2
1
2
3
1
14
9
2
1
2
0
0
2
1
2
15
9
2
0
0
0
2
2
3
1
16
9
2
2
2
2
1
3
2
2
17
7
1
2
3
1
1
2
3
2
18
5
1
2
2
1
3
3
1
2
19
8
2
6
3
2
4
4
1
2
20
1
3
8
10
3
3
2
3
2
21
7
2
0
0

1
1
1
1
22
6
1
3
3
3
3
2
3
2
23
8
2
6
0
0
3
2
2
1
24
8
2
0
0
1
2
2
2
2


LIST OF TABLES
Table Page
2-1 Interrater agreement for respiratory events
of 46 snoring males 51
3-1 Demographic variables from 46 snoring males 56
3-2 Incidence of low and high levels of apnea/
hypopnea by age in 46 snoring males 58
3-3 Significant (p<.05) Pearson correlations
between nocturnal respiratory variables
and demographic variables in 46 snoring
males 59
3-4 Means and standard deviations for selected
demographic and nocturnal respiratory
variables in 46 snoring males grouped by
level ol apnea/hypopnee 61
3-5 Means and standard deviations for health
related variables in 46 snoring males grouped
by level of apnea/hypopnea 65
3-6 Means and standard deviations of selected
sleep variables from 43 snoring males 66
3-7 Means and standard deviations of selected
sleep variables from 43 snoring males
grouped by level of apnea/hypopnea 68
3-8 Leans and standard deviations for daytime
sleepiness variables in 46 snoring males
grouped by level of apnea/hypopnea 69
3-9 Significant (p<.05) Pearson correlations
between nocturnal respiratory and sleep
questionnaire variables in 46 snoring males 71
3-10 Means and standard deviations of sleep
questionnaire variables from 46 snoring
males groupec by level of apnea/hypopnea 71
v


Nocturnal Respiratory and Health Variables 63
Nocturnal Respiratory and Sleep/V.'ake Variables..65
ELG Data 65
Daytime Sleepiness Data 69
Subjective Sleep Assessment 70
Nocturnal Respiratory and Neuropsychological
Variables 74
FOUR DISCUSSION 86
Demographics and Incidence of Nocturnal
Respiratory Disorder 86
Nocturnal Respiratory and Health Variables 90
Nocturnal Respiratory and Sleep/Wake Variables,.92
EEG Data 92
Daytime Sleepiness, Sleep Questionnaire,
and Sleep Log Data 94
Nocturnal Respiratory and Neuropsychological
Variables 96
FIVE SUMMARY AND CONCLUSIONS 103
APPENDIX
A SCORING PROTOCOL FOR RESPIRATORY VARIABLES 106
B RAW DATA TABLES 109
REFERENCES 123
BIOGRAPHICAL SKETCH 132


CHAPTER FIVE
SUMMARY AND CONCLUSIONS
In this study, a detailed assessment of the nocturnal
respiration, health status, cognitive/neuropsychological
skills, and sleep/wake cycle of 46 snoring male subjects
was carried out. Sixty-two per cent of these subjects
experienced at least one event (apnea or hypopnea) while
13% exhibited more than 5 events per hour. Most events
occurred in light slow wave or REM sleep. Obesity was
linked to increasing nocturnal events. Subjects with more
than 5 events per hour suffered significantly exacerbated
oxygen desaturation relative to the remaining subjects.
Overnight oxygen desaturation was linked to increasing
blood pressure readings. As a whole, these subjects
experienced a longer sleep latency, increased wakening
after sleep onset, and more light sleep in the lab, the
classic "first night" effect. It was thought that this
first night effect obscured relationships between noc
turnal respiratory and sleep parameters, which were not
observed. However, overnight respiratory disturbances
were linked to increased sleepiness and napping. Addi
tionally, relationships between nocturnal events and
deteriorating scores on tests tapping non-verbal intelli
gence, verbal and non-verbal memory, expressive verbal
103


71
Table 3-9* Significant (p<.05) Pearson correlations between
nocturnal respiratory and sleep questionnaire
variables in 46 snoring males.
Minutes
Number
hours
Nightly
of Naps
Napping
Awakening
Apnea Index
Apnea+Hypopnea Index
.249
Seconds in Events
.331
Mean High Saturation
Mean Low Saturation
.307
Lumber of Desaturations >4%
.283
-.291
Lumber of Desaturations _>10%
.403
.378
Table 3-10. Means and standard deviations of sleep
questionnaire variables derived from 46 snoring
males grouped by level of apnea/hypopnea.
No Apnea/
Hypopnea
Low Apnea/
Hypopnea
High Apnea
Hypopnea
Mean h. Sleep
7.4 (.9)
7.0(1.6)
7.5(1.0)
Adequate Amount?
1.9 (.4)
1.6 (.4)
2.0(0)
Number of Naps/Y/eek
1.6(1.5)
2.3(2.6)
4.3(2.2)
Hours Napping/Y;eek
Number of Nightly
2.6(1.8)
3.5(2.7)
4.0(2.7)
Wakenings
Minutes Nightly
1.1 (.9)
1.5(1.3)
1.6 (.4)
V/akenings
1.9(1.1)
1.5(1.0)
1.5 (.8)
Sleep Latency
2.0 (.9)
2.C (.6)
2.0 (.6)
Depth of Sleep
2.4 (.8)
2.1 (.9)
3.0(C)
Daytime Sleepiness
1.6 (.6)
1.9 (.4)
1.8 (.4)


9
5. There was a significant correlation between apnea and
age.
For noncomplaining subjects then, males suffer a
higher incidence of apnea activity in virtually every
study which compares sex effects. For males, incidences
range widely from a low of 14.6% (Bixler et al., I960) to
a high of 45% (Webb, 1974); while females range from a
low of 10.2% (Bixler et al., 1982) to a high of 40% (Block
et al., 1980). These differences are probably due at
least in part to age differences in male samples, as well
as age differences in females (only as they reflect rnena-
pausal status). Reported number of apneas range from 2.5
(Webb, 1974) to 8.5 (Bixler et al., 1982) in males and
from 2 (Guilleminault et al., 1978) to 10.5 (Bixler et
al., 1982) in females. Many studies report a positive
correlation between number of apneas and increasing age.
A second sampling technique which has been utilized
has been studying subjects with sleep and breathing com
plaints. Kales et al. (1982) studied 200 subjects with
primary complaints of insomnia. There were 82 males and
118 females in the sample with a mean age of 42.5. Sleep
apnea activity was noted in 10.5% of this group (M =
13*4%, F = 8.5%) with a mean number of 11.2 episodes for
males and 6.6 for females. Subjects with sleep apnea
activity were significantly older (46.8) than those with
out (40.6). Ancoli-Israel et al. (1981) recruited 24
subjects whose answers to questionnaires provoked


129
Miles, L., & Simmons, F. (1964). Evaluation of 190
patients with loud and disruptive snoring. Sleep
Research, 13, 154.
Miller, W. (1962). Cardiac arrhythmias and conduction
disturbances in the sleep apnea syndrome. American
Journal of Medicine, 75* 517-521 .
Milner, E. (1970). Memory and the medial temporal regions
of the brain. In K. Pribram & D. Broadbent (eds),
Biology of Memory. New York: Academic Press.
Mitler, M., Gujaverty, K., Sampson, M., & Bowman, C.
(1982). Multiple nap approaches to evaluating the
sleepy patient. Sleep, , 119-127.
Orr, Vi., Martin, R., Imes, N., Rogers, R., & Stahl, M.
(1979). Hypersomnolent and nonhypersomnolent patients
with upper airway obstruction during sleep. Chest,
75(4), 418-422.
Csterrieth, P. (1944). Le test de copie d'une figure
complexe. Archives de Psychologie, 50. 206-556.
Plum, F., Posner, J., & Hain, R. (1962). Delayed
neurological deterioration after anoxia. Archives of
Internal Medicine, 110, 18-25*
Reynolds, C., Coble, P., Kupfer, D., & Holzer, B. (1982).
Application of tne multiple sleep latency tesx in
disorders of excessive somnolence.
Electroencephalography and Clinical Neurophysiology, 55.
445-452.
Richarcson, G., Carskadon, M., Flagg, Vi., van den Hoed, J.,
Dement, W., & Mitler, H. (1978). Excessive daytime
sleepiness in man: Multiple sleep latency measurements
in narcoleptic and control subjects.
Electroencephalograpny ano Clinical Neurophysiology, 45.
021-627.
Richardson, J., Chambers, R., & Heywood, P. (1959)*
Encephalopathies of anoxia and hypoglycemia. Archives
of Neurology, J_, 178-190.
Roth, T., Hartse, K., Zorick, F., & Conway, V. (1980).
Multiple naps in the evaluation of daytime sleepiness in
patients with upper airway sleep apnea. Sleep, (3/4)f
425-459.
Russell, E. (1975). A multiple scoring method for
assessment of complex memory functions. Journal of
Consulting and Clinical Psychology, 45. 800-809.


APPENDIX A
SCORING PROTOCOL FOR RESPIRATORY VARIAELES *
Data from the nose and mouth thermisters, as well as
the oximeter and strain gauges* were used in scoring
respiratory events.
An apnea was scored when a cessation of airflow at the
mouth and nose lasting for 10 or more seconds was
observed. Up to 3 mm of "noise" were tolerated in an apnea,
as long as no rhythmic pattern which might represent shallow
breathing was observed. An event began at the end of the
last exhalation preceding the apnea. Duration, lowest
saturation, and per cent saturation change were noted.
A hypopnea was scored when a reduction in amplitude of
breathing at the mouth and nose of 50# or mere was
accompanied by a 10# or larger desaturation. The event
began at the end of the last full exhalation. Duration,
lowest saturation, and per cent saturation change were
noted.
Each record was separately scored by two raters. Later
each record was jointly rescored and disagreements were
resolved by consensus.
A separate scoring of oxygen saturation was made by
rating each minute of the record for highest and lowest
saturation. Additionally, a separate tabulation of
106


59
also divided into those above and below the mean weight:
height ratio (2.7) and crossed with those with ana without
t least one event or high levels o apnea/bypopnea. Both
these non-parametrie tests were also non-significant
(X2=.01; p=I\S; X2=1.5; p-NS).
labia 3-3 utilizes Pearson correlations to explore
possible relationships between demographic variables (age,
weight, weight:height ratio) and nocturnal respiratory
indices (apnea index, apnea + hypopnea index, seconds in
events, mean high saturation, mean low saturation, number
of desaturations _>4^, and number of desaturations >10%).
Again, age displays little relationship with respiratory
indices, but substantial correlations between weight and
various respiratory indices are apparent.
Table 3-3. Significant (p<.05) Pearson correlations between
nocturnal respiratory variables and demographic
variables in 46 snoring males.
Respiratory
Variable
Age
Weight
WeightrHeight
Ratio
Number of Apneas
Number of Apneas+Hypopneas
Apnea Index
.265
Apnea+hypopnea Index
.260
.306*
Mean High Saturation
-.323*
-.363*
Mean Low Saturation
-.397*
-.460*
Mean Saturation Change
.319*
.373*
Number of Desaturations
Humber of Desaturations
>4 .287
2.10
.368*
.430*
*p<.01


90
sleep disordered breathing (middle-aged, snoring males).
A prevalence of 62& was observed for occurrence of at
least one event, while 135 had high numbers of
apnea/hypopnea, data which are consistent with an earlier
report drawn from heavy snoring males. Most events
occurred in light slow wave or REM sleep. No effect of
age on level of sleep disordered breathing was observed,
although increasing weight was linked to nocturnal
respiratory events. When the sample was divided by level
of apnea/hypopnea, subjects with high numbers of
apnea/hypopnea suffered significantly exacerbated oxygen
desaturation, although no significant differences were
noted for weight or age among the groups.
Nocturnal Respiratory and Health Variables
The finding of only very limited relationships
between nocturnal respiratory events and health variables
must be tempered by an acknowledgement of possible bias
against such findings in this sample. Specifically, the
sample was selected to be normal and healthy, a criterion
which undoubtably restricted the range of possible health
pathology, and hence probably attenuated the correlational
findings. Additionally, the self-report health checklist,
the Cornell Medical Index, cannot be regarded as a
powerful measure of health pathology. In a very real
sense then, the odds were against finding reliable
relationships between nocturnal respiratory and health
variables.


51
Table 2-1. Interrater agreement for respiratory events of
46 snoring males.
Consensus
Initial
Disagreements
Scoring
Agreements
#1
n
Note
(178)
_
.
Scored by
consensus
91
90
4
6
Consensus
added 1
78
76
1
1
Consensus
added 2
61
52
3
8
Consensus
added 9
57
56
9
2
Consensus
added 1
32
29
3
3
Consensus
added 3
28
27
0
2
Consensus
added 1
25
25
0
0
19
16
4
0
Consensus
added 3
17
17
0
0
15
16
1
0
Consensus
added 1
13
13
0
0
7
7
0
0
4
t
0
0
4
4
0
0
3
2
0
1
Consensus
added 1
3
3
0
0
3
0
7
2
Consensus
added 3
3
3
0
0
3
3
0
0
1
1
0
0
1
1
0
0
1
1
0
0
1
1
0
1
1
0
0
1
Consensus
added 1
1
1
0
0
1
1
0
0
1
1
0
0
474(652)
450
37
36
Note: 26 of the 46 subjects experienced at least 1 apnea or
hypopnea.


CHAPTER ONE
INTRODUCTION
Sleep apneas are respiratory pauses during sleep.
According to the Association of Sleep Disorders Centers
Nosology of Sleep Disorders (ASDC, 1979) a sleep apnea
syndrome (SAS) is a potentially lethal condition charac
terized by multiple apneas, excessive daytime sleepiness,
alterations of consciousness, and cardiopulmonary
complications. The apnea episodes are thought to be the
causal factor behind the other deficits associate^ with
SAS (Guilleminault et al., 1978).
Polysomnographic recordings in various populations
have revealed the occurrence of sleep apneas in the
general population, and particularly in aging popula
tions. Estimates of the prevalence of at least a few
sleep apneas in older subjects have ranged as high as 75£
(Carskadon and Dement, 1981b). Carskadon et al. (1SSC)
have suggested that sleep apneas in older subjects may be
implicated in insomnia, cardiac ailments, and even
senility.
However, certain reports have documented the presence
of multiple sleep apnea episodes in subjects without
apparent pathology (Orr et al., 1979)* This evidence
obscures the role of sleep apneas in the other pathologies
1


84
Pable ;;-16-con tinuec.
Criterion
Predictor
Beta
Sig
Delayed Recall
Age
-.2623
.074
Key Figure
Apnea Index
-.1519
259
Apnea+Hypopnea Inoex
Number of
-.3286
.085
Desaturations >4%
-0795
b57
Sleep Efficiency
.6832
.018
Number oi Stage 0 Periods
2487
354
Time % Stage 0
455S
.172
Time % Stage 2
-.0405
.602
Number oi Stage Changes
-.5054
.057


43
levels of hypoxia affect coordination, memory, attention
span, and may result in headache, drowsiness, and nausea.
The effects of milder, chronic hypoxia are summarized
by Gibson et al. (1981) At the oxygen pressure equiva
lent of 5*000 feet, impaired dark adaptation is noted. At
10,000 feet decreased concentration, hyperventilation, and
short term memory deficits are noted. By 15*000 feet,
euphoria and loss of both coordination and critical judge
ment are observed. Christenson et al. (1977) found signi
ficantly less visual signals detected at 17% O2 than at
21%, suggesting reduced vigilance. McFarland (1937)
studied men in ascents to high altitudes and concluded
that subjects with slower rates of ascent performed better
on unspecified psychological tests. West (1984) reports
data collected on an expedition to Mt. Everest. Results
of a neuropsychological battery indicated deficits in
finger tapping, verbal fluency, verbal learning, short
term memory, and expressive language. Thus the general
affects of chronic mild hypoxia are thought to be deficits
in dark adaptation, short term memory, critical judgement,
motor coordination, verbal fluency, and vigilance.
One group of chronically hypoxia patients, those with
chronic obstructive lung aisease (COLD), has been rela
tively well studied with sensitive neuropsychological
instruments and is aeserving of a more painstaking
review. Krop et al. (1973) studied COLD patients before
and after they received oxygen supplements. The WAIS,


15
duration of 20 seconds and occurred most frequently
(30/60) in Stages 1 and REM. Eight of 20 aging females,
in the report by Block et al. (1980), had sleep apneas
with a mean duration of 19 seconds. Sixty-two percent of
apnea episodes occurred in Stages 1, 2 or REM. Thus in
middle aged normal subjects, incidence of sleep apnea
appears to rise, with a mean duration of between 13 and 20
seconds, perhaps slightly longer than in younger groups.
Again, apneas occurred mostly in light (Stages 1 and 2)
and REM sleep.
Duration data on aging normals are limited.
Carskadon and Dement's (1981b) study of aging normals (M =
72, F = 74) does not report duration of apnea events.
Similary, Ancoli-Israel et al. (1982) do not provide this
information on their aged normals (x = 71.5). Bixler et
al. (1982) noted a high incidence of apneas in their sub
jects over 60, but failed to report a mean duration for
this group. Given the sketchy details reported for this
group, it is difficult to draw any conclusions about
changes in temporal duration or placement of apneas in
aging normals.
To summarize the temporal data in normals, apneas
appear somewhat rarely in normals below 30, with a mean
duration of 15 seconds. Incidence of apneas increase
between 30 and 60, with a wider variation in mean dura
tions reported (13-20 seconds). Apneas occur most
frequently in light and REM sleep in normals of all ages.


130
Sara, S. (1974). Delayed development of amnestic behavior
after hypoxia. Physiology and behavior, 13* 693-696.
Schmidt, H. (1982). Pupilometric assessment of disoraers
of arousal. Sleep, 5., 157-164.
Schroeder, J., Motta, J., & Guilleminault, C. (1978).
Hemodynamic studies in sleep apnea. In C. Guilleminault
& W. Dement (eds.), Sleep Apnea Syndromes. New York:
Alan R. Liss Inc.
Sicklesteel, J., Zorick, F., Wittig, R., Conway, W., & Roth,
T. (1981). Daytime somnolence and insufficient
sleep: A case series. Sleep Research, 10, 233-
Smallwood, R., Vitiello, M., Giblin, E., & Prinz, P.
(1983). Sleep apnea: Relationship to age, sex, anc
Alzheimers Dementia. Sleep, 60), 16-22.
Smirne, S., Francheschi, M., Comi, G., & Leder, F.
(1980). Sleep apneas in Alzheimers disease. Sleep
Research, 9., 224.
Sullivan, C., & Issa, F. (1980). Pathophysiological
mechanisms in obstructive sleep apnea syndrome. Sleep,
M3/4), 235-246.
Tauber, B., & Allweise, C. (1975). Effects of acute
hypoxia on memory. Israeli Journal of Medical Science,
11, 71.
Tilkian, A., Motta, J., & Guilleminault, C. (1978).
Cardiac arrhythmias in sleep apnea. In C. Guilleminault
& W. Dement (eds), Sleep Aonea Syndromes. New York:
Alan R. Liss Inc.
Van Liere, E., & Stickney, J. (1963). Hypoxia. Chicago:
University of Chicago Press.
Webb, P. (1974). Breathing during sleep. Journal of
Applied Physiology, 17(6), 899-903.
Webb, W.B. (1975). SleepThe Gentle Tyrant. Englewood
Cliffs, N.J.: Prentice-Hall Inc.
Webb, W.B. (1982). Measurement and characteristics of
sleep in older subjects. Neurobiology of Aging, 3_, 311-
319.
Wechsler, D. (1945). A standardized memory scale for
clinical use. Journal of Psychology, 19, 87-95*


33
Data on desaturation are interesting from several
perspectives. First of all, Sleep Apnea Syndrome subjects
seem to desaturate more heavily than normals, with Berry
and Block's symptomatic group desaturating up to 50% and
the subjects of Garay et al. desaturating about 22%,
compared with normal desaturation levels of 10-12%.
Secondly, amount of desaturation distinguished symptomatic
(hypersomnolent) from asymptomatic subjects in the data of
both Orr et al. and Berry and Block. Additionally, two
points relevant to methodology emerge. Measures of oxygen
desaturation seem crucial to understanding sleep apnea and
its effects, and ear oximetry appears to provide a proven,
valid measure of this variable.
Weitzman (1979) has suggested that the pathological
process in SAS is the hypoxia of desaturation in apneac
episodes. While cognitive and intellectual changes in SAS
have been only anecdotally reported (Guilleminault et al.,
1976), hypoxia does seem a plausible potential cause of
these changes. Thus oxygen desaturation appears to be a
reliable consequence of sleep apnea syndrome, while cogni
tive/intellectual changes are putative sequelae.
In summary, a review of deficits found in sleep apnea
syndrome subjects suggests that cardiopulmonary complica
tions, excessive daytime sleepiness, oxygen cesaturation
and cognitive/intellectual changes are frequently asso
ciated with this syndrome. An assessment of these
variables in a subclinical population seems indicated by


Table B-5
Ueuropsycholgica1 variables
V/AiS lw
Sub j.
II
Verba 1
Ferior-
mance
Full
WI'ISa
HQ
UI L>KVKC
Digit
Span
1
118
089
11b
157
11
11
7
2
07 5
07 b
074
Ob 7
02
00
4
3
1 2.'j
11b
121
12b
11
13
8
4
135
094
127
120
11
11
8
5
120
10b
115
135
10
11
8
b
122
110
118
124
12
14
7
7
108
10b
108
105
Ob
09
b
8
101
117
108
10b
Ob
11
7
9
120
140
130
129
Ob
08
8
10
128
122
12?
137
15
14
7
11
134
154
13b
124
11
12
8
12
1U2
112
10b
101
Ob
14
7
13
102
090
100
095
05
04
b
14
140
151
140
143
09
15
8
15
104
103
104
107
09
11
5
1b
115
125
118
089
05
10
5
17
113
111
113
094
05
14
b
18
122
110
118
112
11
11
7
19
095
121
107
100
Ob
00
b
20
151
159
1 5b
145
11
13
7
21
120
113
118
118
08
10
7
22
124
125
125
13o
10
10
8
23
119
118
120
142
10
14
8
24
127
119
125
120
10
15
7
25
122
120
122
120
11
11
7
2u
120
100
110
105
11
1u
r
(
46 snoring males
Hooper
L'l Finger Tapping
Hight Lelt Verbal
WOSb Hand Hand Fluency
17
24
4
44
40
03
Cb
12
1
25
2'j
04
25
30
5
52
47
08
19
22
1
54
52
1U
5b
25
5
55
55
15
24
28
4
59
58
13
1b
2
5
53
58
12
14
23
5
5b
53
07
24
28
5
55
53
20
28
28
4
b1
55
13
25
28
4
bO
53
1b
29
27
5
04
51
10
1b
24
3
5b
32
09
20
28
5
55
32
15
30
29
1
59
b1
13
24
29
5
53
30
12
24
29
5
57
59
12
27
28
2
7b
bo
17
22
22
2
b1
b2
03
29
2/
4
72
r *
/ c
19
25
25
2
bb
o4
09
28
25
1
5b
32
13
31
27
5
57
35
20
2b
27
5
b4
bO
18
1b
24
4
54
35
12
2b
25
s
bO
49
14


37
significant declines in average sleep latencies (Carskadon
and Dement, 1981a). Finally, total sleep deprivation led
to a drastic decline in MSLT values, with averages
plummeting below one minute (Carskadon and Dement,
1979a). Mean MSLT values showed recovery to baseline
values after a full night's sleep for partially (5 hrs)
deprived subjects, and recovery after two nights of ad lib
sleep for totally sleep deprived subjects (Carskadon and
Dement, 1979; 1981a). This body of evidence suggests
that the MSLT is a reliable and valid index of sleepiness
in normal subjects.
With its internal validity established, the MSLT was
applied to somnolent populations. Dement et al. (1978) as
well as Richardson et al. (1978) demonstrated that SAS and
narcoleptic patients have significantly shorter latencies
to sleep on the MSLT than normal subjects do. Hartse et
al. (1979) found SAS subjects to have an average latency
to Stage 1 of 2.5 minutes, while narcoleptics had a mean
of 3.2 minutes. Both these latencies were significantly
lower than those of a miscellaneous control group whose
latencies were about 11.2 minutes. Hartse et al. (1980)
found latencies to Stage 1 sleep as follows for various
patient groups: Normals 12.2 m, Insomniacs 16.2 m, SAS
2.5 m, and Narcoleptics 3.2 m. The SAS and narcoleptic
patients had significantly shorter latencies to Stage 1
sleep than insomniacs or normals. Roth et al. (1980)
compared 10 SAS patients with 10 age matched controls.


26
Cardiopulmonary Deficits
Cardiopulmonary complications have been reported as
concomittants of SAS (Guilleminault et al., 1978).
Coccagna et al. (1972) reported high incidences of hyper
tension and congestive heart failure in their SAS sub
jects. Schroeder et al. (1978) studied 22 SAS patients
with a mean age of 47. Six subjects had waking systemic
hypertension, and several had waking cardiac abnorma
lities. During sleep studies, 20 patients developed
significant rises in systemic arterial pressure which
cycled with episodes of apnea, while 21 patients developed
pulmonary arterial hypertension. Tracheostomy reversed
all these abnormalities, strongly implicating the apneas
in their etiology. Tilkian et al. (1978) studied 25 male
SAS subjects with a mean age of 44* Twenty-four of the 25
showed sleeping sinus arrhythmias, while 9 developed more
serious symptoms such as asystole. Seventeen of these
patients received tracheostomy, which abolished all of
these abnormalities. Experimental occlusion of the
tracheostomy during sleep in 6 of these patients resulted
in a return of the cardiopulmonary symptoms. Zwillich et
al. (1982) studied six consecutive male SAS patients. In
all patients, bradycardia accompanied any apnea with
significant desaturation. Additionally, a significant
correlation was noted between degree of desaturation and
severity of bradycardia. Bradycardia was abolished or
attentuated during administration of O2 enriched air.


53
period; mean REM period; mean REM cycle length; % slow
wave sleep; and number of stage changes.
Measures of health status included height, weight,
diastolic and systolic blood pressure, as well as the
Cornell Medical Index. From the CMI, reports of hyper
tension and heart trouble were noted, as well as the total
number of symptoms endorsed. Additionally, the scores for
the scales tapping respiratory, cardiac, and neurological
symptoms were tabulated separately.
Measures of neuropsychological status included the
WAIS Verbal and Performance IQs, the WMS Memory Quotient,
delayed recall of the logical stories and the visual re
production subtests of the WMS, delayed recall of the Rey
complex figure, number of words generated in the verbal
fluency test, number of correct sorts of the Wisconsin
Card Sort, mean number of taps for the right and left
handed finger tapping tests, and lastly the number of
correct trials on the Hooper test.
These neuropsychological tests were chosen with two
aims. First, an attempt was made to sample from each of
several loosely bounded areas of cognitive skill. These
included intelligence (both verbal and non-verbal),
memory, both immediate and delayed (verbal and non
verbal), visuo-perceptual/organizational skills, language,
and frontal self-regulatory skills. A second emphasis was
on a relatively detailed analysis of areas previously
shown to be sensitive to hypoxia. These sensitive areas


Table 3-14.
Significant (p<.05) partial correlations, controlling lor age, between
nocturnal respiratory and cognitive variables in 46 snoring males.
Apnea
Apnea/
ilypopnea
Index
Index
WAIS Verbal IQ
WAIS Performance IQ
-.311
-.414*
Wechsler Memory Scale
Memory Quotient
-.290
Delayed Recall Logical
Stories
-.343
Delayed Recall Visual
Reproductions
-.416
Digit Span
Delayed Recall Rey Figure
-.284*
-.305
Hooper Test
Wisconsin Card Sort
-338
Finger Tapping Right Hand
-.259
Finger Tapping Left Hand
-.258
Verbal Fluency
-.364*
Mean
Humber
of
Fvent
Saturations
L'esatura
tions
Time
High
Low
>A%
>_10>>
.274
-.321
.421*
.521*
-.461*
.302
.251
-.311
-.260
.260
-.276
.437*
.349*
.489*
-.337
.369*
.252
.332
. 265
.274
.329
.365*
-.329
vO
*p<.01.


85
non-verbal memory. Addition of variables reflecting sleep
fragmentation to the analysis indicated that both
nocturnal respiratory and sleep variables make significant
and independent contributions to prediction of
cognitive/neuropsychological scores.


APPENDIX E
RAW DATA TABLES


18
associated variables include sex, age, snoring and
obesity. The data available on these variables and their
relationships to sleep apnea activity will be reviewed,
followed by a brief discussion of their relevance to
experimental methodology.
Regarding sex variables, as previously noted,
Guillerainault et al. (1978) found 96% of their SAS
patients to be males. In the Block et al. (1979) study of
nonsleep disturbed subjects, 40% of the males in the
sample had at least one episode of apnea, while only 15%
of the females so suffered. Bixler et al. (1982) noted a
trend toward more sleep apnea activity in males. Block et
al. (1980) studied 20 post menapausal females and found
apnea activity in 40% of them, leading them to suggest
that some factor associated with menapause (e.g., pro
gesterone) might somehow "protect" premenapausal females
from apneas, and that this factor along with its benign
influence was lost with menapause. This explanation pre
dicted the low level of apnea activity in younger females,
as well as the heightened apnea activity of aging
females. While the data of Carsakdon and Dement (1981b)
show aged females to be close to aged males in percentage
exhibiting sleep apnea activity, a report by Smallwood et
al. (1983) found no evidence of sleep apnea activity in
their 6 elderly (post menapausal) females, while elderly
males showed significant levels of apnea activity. The
limitations of the sample size in the Smallwood et al.


xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EOGSOKC06_8RP790 INGEST_TIME 2012-09-24T13:31:42Z PACKAGE AA00011824_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES


Table 3-4. Means and standard deviations for selected demographic'ana nocturnal
respiratory variables in 46 snoring
apnea/hypopnea.
males
grouped
by level
of
No Apnea/
Low
Apnea/
High Apnea/
Nypopnea
Hypopnea
Hypcpnea
Age
52.3(12.7)
46.1
(13-5)
56.5
(9.3)
Weight (lbs)
185.6(34.8)
188.6
(37.6)
203-3
(24.3)
Weight:Height Ratio
2.7 (0.4)
2.7
(0.5)
2.9
(0.4)
Education (years)
15.2 (2.9)
14.6
(2.0)
13.8
(2.6) .
Number of Apneas
0
5.0
(7.7)
24.5
(34.1)a,fc
Number of Apneas+Hypopneas
0
7.3
(9.2)
81.7
(52.6)
Apnea Index
0
0.9
(1.1)
4.5
(5.y)*.b
Apnea+Hypopnea Index
0
1.4
(1.5)
19.5
(19 4)0 f3
Seconds in Apneas/Hypopneas
0
129.4(167.1)
2325 *5(1469.5)aD
Meen Apnea Duration
-
13.4
(4.0)
17.8
(1.3)
Mean Hypcpnea Duration
-
29.6
(21.2)
33.2
(14.5)
Mean High Saturation
95.5 (1.7)
95.3
(1.3)
94.9
(1.2) .
Mean Low Saturation
94.3 (2.1)
93.1
(1.6)
89.0
(4.6)0>l)
Number of Desaturations _>4/
29.5(54.1)
56.2
(36.6)
143.2
(1O5.9)0,b
Number of Desaturations >J\0%
0.2 (0.7)
3.5
(6.3)
51.0
(48.7)ab
Significantly different from no sleep apnea/hypopnea.
Significantly different from low sleep apnea/hypopnea.


99
with greater physical deterioration (Feinberg et al.,
19^7), subjects selected for health do not display this
effect. A third embarrassment for the sleep fragmentation
hypothesis lies in the study of Orr et al. (1979) who
described one group of symptomatic sleep apnea syndrome
subjects and a group of asymptomatic snoring subjects that
were matched on number of nocturnal events. Presumably,
arousals resulting from events were matched in the two
groups, yet the symptomatic subjects had serious impair
ments relative to the asymptomatic groups.
Despite these difficulties, the present study found
that one of the five cognitive scores significantly pre-
cicted by the nocturnal respiratory indices was also
correlated with sleep efficiency on the experimental
night. However, given the earlier observed problems with
the experimental nights* sleep, the inconsistent past
evidence, and the limited (1) relationship observed, this
finding remains inconclusive. A more ambitious study,
examining sleep and respiration across several nights, is
necessary to clarify this issue.
An alternative, although admittedly speculative,
explanation of the cognitive changes lies in the cerebral
hypoxia accompanying respiratory events. Brain dysfunc
tion is known to occur in mild to moderate hypoxia. Y/hile
the brain possess various homeostatic mechanisms to com
pensate for chronic hypoxia (hence the adaptation of long
term residents of mountainous regions) it may be that the


100
multiple brief, acute desaturations experienced by the
subject with apnea/hypopnea fail to activate these protec
tive mechanisms effectively (Gibson et al., 1981).
Recently, a body of evidence has accumulated which indi
cates that turnover of certain neurotransmitters is drama
tically altered by acute hypoxia. The cholinergic synthe
tic pathway, which results in the formation of the neuro-
transmitter acetylcholine (ACh) appears to be particularly
sensitive to the effects of acute hypoxia (Gibson et al.,
1981)* That ACh may be the neurotransmitter most affected
by hypoxia of nocturnal events is provocative, in light of
the "Cholinergic hypothesis" (Bartus et al., 1985) of the
memory and cognitive impairment often found in aging sub
jects. This formulation notes that cholinergic blockade
in young subjects inhibits memory, while various aspects
of the cholinergic synthetic pathv/ay are altered in aged
subjects with memory and cognitive deficits. Alterations
in the cnolinergic synthetic pathway are thought to under
lie the cognitive changes in both populations. It may be
speculated that the cognitive dysfunction found in the
present subjects with multiple apnea/hypopnea is also a
product of a disruption of cholinergic synthesis.
To summarize the logic of the "cerebral hypoxia hypo
thesis," it is first noted that subjects with nocturnal
respiratory events experience multiple, acute hypoxic
episodes. This acute hypoxia appears to evade cerebral
protective mechanisms which normally respond to chronic


19
report might explain this discrepancy as representing a
sampling bias. Taken together, this evidence strongly
supports the notion that in young and middle aged popula
tions, sleep apnea is a male phenomenon. Additionally,
evidence points toward a closing of the gap between
females and age matched males as the former group passes
menapause.
A second factor which has been implicated in sleep
apnea activity is age. Webb (1974) found that his sub
jects exhibiting sleep apnea activity were significantly
older than his full sample. This was also the case for
the subjects of Kales et al. (1982) and Ancoli-Israel et
al. (1982). Smallwood et al. (1983) found that sleep
apnea was an age dependent phenomenon with male subjects
over 50 exhibiting significantly more apneas than those
under 50. Other evidence includes that of Bixler et al.
(1982) who found a positive correlation between age and
apnea in their full sample, as did Block et al. (1980) in
their post menapausal females. It seems clear from these
data then that increasing age is a factor predisposing for
sleep apnea activity.
Snoring has been closely associated with sleep apnea
(Lugaresi et al., 1982). This may be due to a causal
relationship or a single common pathway. Lugaresi et al.
propose the latter explanation. Specifically, they
suggest that both snoring and SAS are due to a sleep
induced stenosis of the upper airway. With this premise


63
apnea or hypopnea, while 13% had high levels of
apnea/hypopnea. Most events occurred in light slow-wave
or REM sleep. No effect of age on level of sleep
disordered breathing was observed, although increasing
weight was correlated with nocturnal respiratory events.
When the sample was stratified by level of apnea/hypopnea
activity, subjects with high levels of events had
significantly more events and desaturations than the
remaining subjects, although no significant differences
were observed on age, weight or education.
Nocturnal Respiratory and Health Variables
health variables including blood pressure readings
and several scores derived from the Cornell Medical Index
(overall number of symptoms endorsed, as well as number of
symptoms endorsed from the respiratory, cardiopulmonary
and neurological subscaies) were evaluated initially
through a matrix of Pearson correlations which included
the respiratory variables (apnea index, apnea + hypopnea
index, seconds in events, as well as mean high and low
saturation and number of desaturations >A% and >10%)*
From a total of 42 correlations, three were significant at
p<.05 or less: Systolic blood pressure and mean low
saturation (r=-.21); Systolic blood pressure and number
of desaturations >A% (r=.373); Overall number of symptoms
endorsed on the CMI and mean high saturation (r=-.254).
As a further exploration of possible relationships between
blood pressure and apnea/hypopnea activity, subjects were


52
low score seemed obvious. In order to assess the quan
titative "impact" of numerous desaturations, a sum of 4%
and 10% desaturation was tabulated. These selected
parameters for thumistor and saturation variables allowed
reliable scoring and a description of several aspects of
nocturnal respiration.
Dependent measures fell into three broad classes:
measures of sleep/wake status, measures of health status,
and measures of neuropsychological status. Measures of
the sleep/wake cycle included one week's sleep logs
(derived variables included number of reports of daytime
sleepiness, hours asleep, estimated sleep latency, number
of wakenings after sleep onset, and minutes of waking
after sleep onset), a sleep questionnaire (derived
variables including hours of sleep usually obtained,
number of naps, hours napping, quality of sleep,
restfulness on awakening, depth of sleep, and amount of
daytime sleepiness), the mean of the several Stanford
Sleepiness Scale ratings, the two multiple sleep latency
tests (derived variables included mean sleep latency and
total minutes asleep in both naps) and the overnight EEC
variables derived from the scoring system of Agnew and
Webb (1972). Electroencephalographic variables included
time in bed; pure sleep time; sleep efficiency index;
number stage 0 periods; time % stage 0; sleep latency;
time % stages 1, 2, 3, 4 and REM; latency first REM


89
these data somewhat tenuous- This finding needs
replication before any conclusions may be drawn.
Ninety-eight per cent of apneic events occurred in
light slow wave and REK sleep, an observation which is not
surprising given that subjects spent nearly 80£ of their
sleep in these stages. The preponderance of events in
these stages of sleep is consistent with the bulk of past
reports in this area (e.g, Bixler et al., 1982; Block et
al., 1979; Krieger et al., 1983).
When subjects were divided into groups by level of
apnea/hypopnea, significant differences emerged on oxygen
desaturation variables. Specifically, comparisons
indicated that the group with a high level of
apnea/hypopnea events experienced deeper desaturations
than the remaining two groups. This difference occurred
despite a lack of differences on age and weight among the
three groups, and appears to implicate the respiratory
events in the genesis of the desaturations. This is
consistent with an earlier report by McGinty et al. (1982)
who demonstrated that subjects with significant levels of
sleep disordered breathing were impaired on oxygen
desaturation variables, relative to subjects without
respiratory events. Subjects with heavy snoring and
multiple nocturnal apneic respiratory events appear to be
at risk for nocturnal oxygen desaturation.
In summary, the present sample of 46 snoring males
appears representative of those predisposed to experience


97
included Krop et al. (1973) and Krop et al. (1977) both
of which notea xhr.it continuous nocturnal oxygenation
treatment significantly improved neuropsychological scores
in normal aged and chronic obstructive lung disease
patients, respectively. The present results are also
consistent with a report by Grant et al. (1982)
demonstrating considerable neuropsychological impairment
in chronic obstructive lung disease patients, a group
knov/n to experience hypoxia which is significantly
exacerbated during sleep (Block, 1981). Additionally,
West (1984) found impairments in finger tapping speed,
shcrt-term memory, and expressive verbal fluency in a
group of climbers ascending F:t. Everest. Taken together,
the present results and past reports present a consistent
picture of relationships between both daytime and
nighttime hypoxia and neuropsychological impairments.
Although the selection requirement of general good
health ruled out the inclusion of subjects with a full
blown sleep apnea syndrome, six of the present subjects
demonstrated levels of respiratory disturbance sufficient
to meet current diagnostic criteria for SAS-(A+H)I>5
These subjects showed significant impairments, relative to
those with no sleep apnea/hypopnea, on measures tapping
non-verbal intelligcnce, delayed verbal and non-verbal
memory, and verbal fluency. These data are suggestive
that heavy snoring males with multiple apneas/hypopneas
may be at risk for cognitive sequelae. If the impairments


UNIVERSITY OF FLORIDA
illlllllllH
3 1262 08554 7908


75
significant overall p value, univariate regressions are
used to individually predict criterion variables with the
multiple predictors. Contribution of individual
predictors may also be examined in these univariate
regressions.
The results from the multivariate regression of the
above noted cognitive scores and selected nocturnal
respiratory variables (Apnea index, Apnea + hypopnea
index, Mean high saturation, Number of desaturations >4%,
Age, Veight:Height ratio) are presented in Table 5-13* An
overall Hotellings T2 score was significant (p<.003),
allowing examination of the univariate regressions.
Criterion variables which were significantly preaicted by
the respiratory and demographic variables included WAIS
PIQ, VMS MQ, delayed recall of Hey figure and visual
reproductions, Hooper, and Wisconsin card sort. These
relationships reflected negative relationship between
overnight respiratory indices and neuropsychological
scores. Respiratory indices formed significant components
of the predictor combinations for all these variables save
Hooper score, suggesting relationships between nocturnal
respiratory indices and non-verbal intelligence, verbal
and visual memory, and ability to shift sets while problem
solving.
Because age is correlated with nocturnal respiratory
parameters in past reports, and also with neuropsycho
logical scores in other samples, the various overnight


3
et al. (1976) first suggested that the occurrence of 30 or
more apneas in 7 or more hours of sleep was diagnostic of
SAS, while Guilleminault et al. (1978) proposed the use of
an apnea index (AI; number of apneas/number of hours
sleep) in excess of 5 as a cut-off score to diagnose the
syndrome. Later, Carskadon et al. (1980) combined apneas
and hypopneas (a reduction in airflow of 50% or more at
the mouth and nose) to form an apnea + hypopnea index
(AHI) with 5 utilized as a diagnostic level of sleep apnea
syndrome. As can be appreciated from the preceding dis
cussion, the diagnostic criteria for SAS have undergone a
rapid evolution, with remarkably little validation work
published.
History and Characteristics of Sleep Apnea Syndrome
Sleep Apnea Syndrome falls within the larger category
of Sleep Disordered Breathing (Block, 1980). Respiratory
disorders within this category involve changes in
breathing patterns during sleep. Syndromes described
within this category include Pickwickian Syndrome; Sleep
Apnea Syndrome; Chronic Obstructive Lung Disease; and
syndromes of Primary and Secondary Alveolar
Hypoventilation (Block, 1980; Guilleminault et al.,
1976).
The first disorder described in this group was the
Pickwickian Syndrome. These patients were initially
identified by a chronic daytime hypersomnolence, obesity,
Cheyne-Stokes breathing, hypercapnia and hypoxemia. This


17
had apneas with a mean duration of 25.9 seconds, while
symptomatic subjects had a mean duration of 20.8 seconds.
Several characteristics emerge from the group of
reports on SAS subjects. There is wide variability in
duration of apneas, with reported means ranging from 16-30
seconds. The data available on placement of apneas in
sleep suggest that in SAS subjects most apneas occur in
light and REM sleep.
Comparison of data on normal and SAS subjects reveals
substantial overlap between mean duration of apneas in SAS
subjects (13-20 seconds) and SAS subjects (16-30
seconds). Particularly interesting was the Orr et al.
(1979) study which demonstrated a lower mean duration of
apneas for hypersomnolent subjects than for nonhypersomno-
lent subjects. Data on temporal placement of apneas show
most events occurring in light or REM sleep for both
normal and SAS subjects. Taken together, these data
suggest that substantial overlap exists on temporal
t
variables between normal and SAS subjects indicating a
limited utility for diagnostic purposes.
Variables Associated With Sleep Apnea
Syndrome anc Activity
It was noted above that several variables have been
suggested as associated with increased incidences of sleep
apneas and SAS. These variables represent potential
"markers" of subpopulations which might experience higher
incidences of sleep apnea activity. These primary


92
Nocturnal Respiratory and Sleep/Wake Variables
EEG Data
The overnight sleep in the laboratory for these
snoring subjects was noteworthy for a general shift
towards lighter sleep than usual. A longer sleep latency,
more awakenings, more light sleep, and less slow wave and
REM sleep than normal are apparent. These findings
parallel the report by Block et al. (1979) who found
shifts to lighter sleep and more awakenings than in a
sample of normal subjects. These changes were said to
result from a first night in a strange bed and the
discomfort of the multiple clips and sensors used to
monitor respiration. Recording montages were identical
for the present subjects, and these two factors
undoubtably influenced the sleep of the present subjects
as well.
An intriguing observation was a lack of covariation
between respiratory and EEG sleep variables derived from
the recording night. This finding is counterintuitive, as
one would expect arousals from the respiratory events to
influence the sleep characteristics of subjects with
events. However, correlational procedures failed to
support this notion. It may be that a ceiling effect was
exerted on sleep measures sensitive to disturbance, one
that resulted from the first night effect compounded by
tne discomfort of the recording procedure. The already
elevated disturbance sensitive sleep variables might not


6
definitively demonstrated to be the necessary and suffi
cient cause of the syndrome. Thus, for the purposes of
the following review, the focus will be on occurrence and
level of sleep apnea activity in both sleep apnea syndrome
subjects and other populations. The full range of sleep
apnea activity will be reflected by noting the presence of
sleep apnea activity, and an indication of the quantita
tive level of the activity will be presented in the apnea
index (AI) or the apnea + hypopnea index (AHI). Recall
that an AI of 5 has been frequently utilized as a cut-off
score for identifying SAS patients.
A review of the literature suggests that at least
three sampling strategies have been used for studying the
incidence of sleep apnea activity. Subjects have been
selected on the basis of one of the following three
criteria: 1) absence of sleep or respiratory complaints,
2) presence of sleep or respiratory complaints, or 3)
sampling without regard to these variables.
because diagnosis of SAS is often dependent on number
of sleep apneas (or level of SaA), most researchers report
number of apneas observed, although some report only on
number of subjects who exceed one of the criteria for
SAS. A review of studies with samples of 20 or greater
will be presented with careful rote of the sampling
strategy utilized, as well as the dependent measure of
apnea.


44
Wechsler Memory Scale (WMS), Bender-Gestalt, Bender
Gestalt Interference Procedure, Facial Recognition and
Finger Tapping tests were administered to COLD and control
subjects. Subjects receiving supplementation improved
significantly more than controls on WAIS Full Scale IQ,
Performance IQ, Wechsler Memory Quotient, both Bender
procedures and Finger Tapping. Grant et al. (1982)
studied another group of COLD patients. They administered
the Halstead-Reitan Battery, Aphasia Screening Test,
Trailmaking Test, Russell modification of the Wechsler
Memory Scale, Tactile Memory test, and the WAIS. In COLD
subjects, deficits were found relative to controls on all
variables except Reitan Rhythm, Aphasia Screening and WMS
Logical stories. Grant et al. concluded that COLD
patients exhibited deficits on a global impairment rating,
attention, abstracting ability, complex perceptual motor
tasks, simple sensory and motor tasks, and memory tasks.
Low but significant negative correlations were reported
between several measures of oxygen saturation and both
Halstead Impairment Index ana Global Impairment Rating.
Grant et al. speculated that floor effects resulted in low
variability in PA2 levels among COLD patients and kept
these correlations low. They further suggested that the
neuropsychological deficits observed in these patients
resulted from the oxygen want found in these subjects.
Thus data available on humans indicate that dark
vision, memory, motor coordination, critical judgement,


74
In summary, no between group differences were noted
for either sleep questionnaire or log variables when
subjects were grouped by level of apnea/hypopnea.
Correlational procedures suggested relationships between
several respiratory variables and reports of napping and
wakening after sleep onset on the sleep questionnaire.
These findings were cross validated in the correlational
analysis of sleep log data, in which reports of napping
and wakening after sleep onset were related to respiratory
variables, as were total bedtime and sleep latency.
Nocturnal Respiratory and Neuropsychological Variables
Scores derived from the cognitive/neuropsychological
battery included WAIS Performance IQ, WAIS Verbal IQ,
YJechsler Memory Quotient, delayed recall of logical
stories, delayed recall of visual reproductions, delayed
recall of Rey complex figure, digit span, Hooper score,
Wisconsin card sort score, finger tapping right and left,
and Verbal fluency.
As a preliminary analysis, these scores were compared
with selected respiratory and demographic variables in a
multivariate regression procedure. In this type of
analysis, a set of predictor scores is used to
simultaneously predict a set of criterion variables while
controlling for intercorrelations amongst the measures.
An overall test of significance, such as the Hottelings
T2, establishes the probability of obtaining the observed
relationships from chance alone. Upon obtaining a


102
at risk for cognitive sequelae. These data may be taken
as limitec support for the anecdotal reports of cognitive
changes in SAS patients. Two explanations of possible
underlying mechanisms of the cognitive changes may be
considered. The first, the "sleep fragmentation hypo
thesis," suggests that the cognitive deficits result from
the multiple arousals experienced by the subject. This
notion is suspect as a plausible explanation because sleep
deprivation and sleep measures tapping sleep fragmentation
fail to intercorrelate consistently with cognitive scores
in other populations. Additionally, the extremely limited
relationship observed in the present study is less than
conclusive. It is proposed that cerebral hypoxia induces
alterations in synthesis of ACh, a neurotransmitter impli
cated in cognitive abilities in other populations, and
thus causes the cognitive changes. The cerebral hypoxia
theory remains speculative, although several testable
hypotheses may be generated from it. Further research
into factors affecting the sleeping brain may prove
heuristic for neuropsychologists.


16
Temporal data on subjects diagnosed with sleep apnea
syndrome (>50 or AI>5) primarily include patients between
the ages of 30 and 60. The reason for the lack of older
SAS subjects is unclear, particularly in light of data
reported above, demonstrating very high percentages of SAS
in certain groups of aging subjects.
Garay et al. (1981) studied 13 sleep apnea syndrome
subjects. They divided these 12 males and 1 female into
two groups, based on their daytime resting CO2 levels.
Subjects who were eucapnic during the day had a mean apnea
duration of 16 seconds, while those who were hypercapnic
had a mean duration of 17 seconds. No data on distribu
tion relative to sleep stages were reported.
Guilleminault et al. (1978) studied 50 predominantly male
(x age 48) sleep apnea syndrome patients. Apneas in this
population averaged 22 seconds in duration. Apneas were
longest in REM, and few occurred in Stages 3 or 4 sleep.
Fujita et al. (1981) studied 12 male SAS patients with a
mean age of 43* Apnea episodes averaged 23 seconds in
duration. Weitzman et al. (198C) studied 10 SAS males
whose age ranged from 38-57* The mean duration of apneas
in this group was 30 seconds. All apneas occurred in
light or REM sleep. Orr et al. (1979) studied 4 hyper-
somnolent males diagnosed with SAS. A control group of 4
nonhypersomnolent subjects was matched on number of ob
structions per minute. Mean age was 57 for asymptomatic
and 42 for symptomatic subjects. Asymptomatic subjects


10
suspicion o insomnia or nighttime breathing and muscular
events. Eleven males (x = 72.5) and 13 females (x = 68.5)
were studied. All of these subjects, save one male and
one female, suffered at least one episode of sleep
apnea. For those subjects with at least one episode of
sleep apnea, males had a mean of 53 episodes of apnea,
while females had a mean of 32 episodes. In contrast to
the Kales et al. study, 6 males had a high level of sleep
apnea (AI>5), while 3 females met this criterion. A rela
tionship between age and apnea was not reported in this
sample, possibly because of restricted range. These two
studies of sleep disturbed subjects provide conflicting
data. That is, the insomniac alone sample did not demon
strate an apnea incidence which was substantially
different from a noncomplaining sample (10.5%) while the
sample recruited for sleep induction, respiratory, and
muscular differences showed very discrepant apnea inci
dences (nearly 100£). One potential explanation involves
age, as the Ancoli-Israel sample is substantially older
than Kales', although the aging samples have not reported
this level of apnea before. An alternative explanation is
that sampling insomniacs does not generate a high number
of apneas, whereas aging sleep and respiratory disturbed
subjects show a very high incidence of apnea. It would
seem then that the relevant "complaint" would involve
those reported by Ancoli-Israel et al. which are in
addition to those noted by Kales et al. These are (as


30
experimental protocols by Weitzman et al. (1980), Fujita
et al. (1981), Zwillich et al. (1982), Garay et al. (1981)
and Sullivan and Issa (1980). At the same time, subjects
wno are asymptomatic for EDS but with the necessary number
of apneas to be diagnosed as SAS are described by Orr et
al. (1979) and Smirne et al. (1980) as well as others.
However, it should be noted that the latter studies relied
on global reports of somnolence, rather than quantified
data. Recalling the criticism of Dement et al. (1978) of
this technique as vulnerable to subjective bias, these
asymptomatic patients may have unreported or undetected
somnolence.
Excessive Daytime Somnolence appears to be frequently
concomittant with Sleep Apnea Syndrome. Thus its possible
occurrence in subclinical populations seems worthy of
investigation.
Oxygen Desaturation and Cognitive Deficits
One deficit which appears in SAS patients is oxygen
desaturation which accompanies apneas. V.'hen apnea begins,
arterial oxygen saturation begins to fall, and may con
tinue falling throughout the event. An early technique
utilized in studying oxygen saturation levels involved
periodic blood samples drawn from arterial sources.
Eirchfield et al. (1958) studied 11 normal males with a
mean age of 23* Daytime oxygen saturation in these sub
jects averaged 96.5$. These values fell during sleep to
95.3#. Later, Orr et al. (1979) studied 8 males who had


Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
SLEEP APNEA ACTIVITY AND ITS CONCOMITANTS
IN A SUBCLINICAL POPULATION
By
DAVID THOMAS REED BERRY
August 1985
Chairman: Wilse B. Webb
Major Department: Clinical Psychology
Sleep Apnea Syndrome (SAS) is a nocturnal respiratory
disorder with serious consequences. Patients with SAS
experience apneas (pauses in breathing of 10 or more
seconds) and hypopneas (declines in the amplitude of
breathing accompanied by oxygen desaturation) while
asleep. Recently, it has become apparent that apneas and
hypopneas may occur in otherwise apparently normal
subjects. A sample selected to display a wide range of
sleep disordered breathing (heavy snoring males in good
health and over 30 years of age) received detailed
measurement of their nocturnal respiration, sleep,
cognitive skills, and health in an attempt to clarify the
impact of apneas and hypopneas in a subclinical population.
Forty-six heavy snoring males in good health with a
mean age of 50 years and a mean weight of 190 pounds
comprised the sample. During an experimental evening they
vii


11
nearly as can be judged) questions surrounding
respiration.
The final category of incidence studies involves
those which do not consider sleep complaints in their
sampling. As in the other studies, these vary in their
generalizabil.ity. Carskadon and Dement (1981b) recruited
40 aging subjects, including 18 males (x = 72.7) and 22
females (x = 74). Their exclusion criteria are somewhat
problematic, involving excluding subjects who "spon
taneously complain of sleep problems." Thus, it is
unclear what this is a representative sample of, specifi
cally whether the distribution of those not spontaneously
complaining of sleep disturbances may be different from
those who actually do not suffer from sleep distur
bances. These considerations aside, Carskadon and Dement
report that 16 males (88%) and 16 females (72.%) demon
strated at least one sleep apnea episode, while 8 males
(44$) and 7 females (31.8%) have high levels of sleep
apnea (AI>5). Males with apnea showed a mean number of
51, while females showed a mean number of 34. Sex
differences appear to have fallen out at this age, while
the restricted range of this extreme age sample does not
lend itself to a correlational analysis of age versus
apnea. In another study of limited generalizability,
Kreiss et al. (1982) randomly selected inpatients on a
Veterans Administration medical ward. These patients, all
presumably male, numbered 26. Further data reported were


received testing and filled out various questionnaires
before sleeping overnight with their sleep and breathing
continuously recorded. Sixty-two per cent of these subjects
experienced at least one event, while 13% had more than 5
events per hour of sleep. Most events occurred in light
slow wave or REM sleep and were linked to obesity. Subjects
with high levels of apnea/hypopnea (more than 5 per hour)
experienced significantly exacerbated oxygen desaturation
relative to the remaining subjects. Oxygen desaturation was
linked to higher blood pressure readings. A typical "first
night*' effect on the sleep of the subjects in the lab was
noted, resulting in a lighter sleep than usual and obscuring
possible relationships between sleep and respiratory events,
which were not observed. Deteriorating overnight
respiratory indices were associated with increased
sleepiness and napping. High levels of apnea/hypopnea were
associated with declining scores on tests measuring non
verbal intelligence, verbal and non-verbal memory,
expressive verbal fluency, and cognitive flexibility. It is
speculated that hypoxia induced changes in cerebral
cholinergic synthetic pathways underlie these changes.
It is concluded that a significant subgroup of heavy
snoring males who experience multiple apnea/hypopneas are at
increased risk of significant oxygen desaturation, daytime
sleepiness, and cognitive changes. This group may fall on a
continuum with sleep apnea syndrome patients.
viii


4b
is indicated. An adequate description of the sleep/wske
cycle would seem to include sleep questionnaire, sleep
log, overnight EEG, multiple sleep latency test, as well
as several Stanford Sleepiness Scale ratings.
Statement of the Problem
Sleep Apnea Syndrome is a clinical entity with
seemingly serious consequences. Although the nocturnal
apneas found in this illness are thought to be the cause
of the associated deficits, the identification of
seemingly asymptomatic subjects with a number of apneas
comparable to those found in SAS throws doubt onto this
formulation. Despite this evidence, the recent discovery
that apneas occur in significant numbers in the general
population has led to speculation that this subclinical
level of apnea may be the cause of various pathological
processes. At present, there is little evidence to
support or refute this notion.
The present report evaluates this issue through a
detailed study of a group of subjects thought to be at
risk for sleep apneas. Aging, heavy snoring males were
recruited, and a comprehensive analysis of their respira
tory, health, neuropsychological, and sleep/wake cycle
status was carried out. Thus potential relationships
between respiratory disturbances and variables known to be
disrupted in sleep apnea syndrome patients could be
evaluated. This approach also allowed exploration of the


29
apneas actually suffer EDS, but that it is masked by two
factors. Dement et al. noted that SAS patients may deny
somnolence while literally falling asleep before the
clinician. They speculate that this results from a re
sponse bias against acknowledging illness or a change in
subjective frame of reference about just what alertness
is. Thus EDS may be denied by patients for reasons which
may be outside the clinician's control or knowledge.
Alternatively, Block et al. (1979) propose that asympto
matic SAS patients are suffering a subclinical level of
fallout from their frequent apneas. As years pass, in
creasing weight or cumulative effects of desaturation
eventually lead to a full blown hypersomnolenceSleep
Apnea Syndrome. A final possibility is that nonhyper-
somnolent subjects with frequent apneas exhibit a com
pletely benign process totally unrelated to that found in
symptomatic SAS patients. It would seem that delineation
of the level of EDS in subjects with subclinical levels of
apnea may help clarify this issue. This notion will be
elaborated below.
The somnolence deficit which is noted in full blown
SAS patients is commonly reported. Dement et al. (1978)
describe impairments in continuous performance of
virtually any activity. Sleep Apnea Syndrome patients
reportedly fall asleep at outdoor stadiums, in front of
classes, and while treating patients. Excessive Daytime
Somnolence is described in every SAS patient reported in


I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.
Wilse B. V/ebb, Chairman
Graduate Research Professor of
Psychology
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.
Assistant Professor of Clinical
Psychology
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.
T. 37Block
Processor of Medicine
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.
F. Dudley McGlyriri
Professor of Clinical
Psychology


49
professors. All subjects completed and signed an informed
consent agreement.
Apparatus
During the evening testing, all subjects received a
blood pressure check with a standard hospital cuff. The
measurement was made from a seated position. Overnight
recordings were made in a quiet, darkened room with the
subject sleeping on a standard hospital bed. All noc
turnal physiological recordings were routed through a
Grass Polygraph (model 7D). Electrodes for electro-
encephalographic recordings were mounted in three pairs
using sites F2/F8, PZ/T6, RE/LE from the international
10/20 system of placement. These electrodes were affixed
to the scalp with collodion soaked gauze pads dried with
an air hose. Respiratory measures included measurement of
chest and abdomen wall movement derived from impedence
measures monitored from surface electrodes (3F Ag/AgCl
#2246) placed on the lower chest and just above the
navel. The signal from these electrodes was passed
through an impedence convertor and onto the polygraph.
Oral and nasal airflow was recorded by thumistors (TCT-1R
Transducer-Grass) clipped to the respective orifices.
These signals were also routed through the polygraph.
Electrocardiograms were monitored from modified bipolar
chest leads (MCL2). Finally, blood oxygen saturation was
continuously monitored from a probe attached to the ear
lobe, and analyzed by a Biox ear oximeter (flodel IIA BTA


61
figure [F[2,43)=6.5> p<.003], and Verbal fluency,
[F[2,43)=3-6 p<.04]. Followup comparisons (Scheffe)
showed the group with high apnea/hypopnea to be
significantly impaired relative to those with no sleep
apnea/hypopnea on WAIS PIQ, delayed recall of logical
stories and Rey complex figure as well as verbal
fluency- Additionally, the high apnea/hypopnea group was
impaired relative to those with subclinical apnea/hypopnea
on aelayea recall of the Hey complex figure. These
findings are bolstered by the lack of significant between
group differences on age, weight, or education (see Table
3-15) all of which are potential alternative explanations
of the between group differences.
As a check, an analysis was run to evaluate the
relative contribution of nocturnal respiratory variables
and sleep variables reflecting fragmentation in predicting
cognitive/neuropsychological scores. A multivariate
regression of these scores was carried out. Demographic
and respiratory scores included age, apnea index, apnea +
hypopnea index, and number of desaturations >4%. A subset
of the earlier noted sleep variables was chosen to reflect
measures thought to be most responsive to sleep
fragmentation. These included sleep efficiency, number of
stage 0 periods, time % of stage 0, time % of stage 2, ano
number of stage changes. These respiratory and sleep
parameters were used to predict the 5 cognitive scores
which were earlier shown to be predicted by nocturnal


7e
scores were entered into a partial correlation matrix,
controlling for age, with the cognitive scores. Table
5-14 presents the resulting correlations significant at
p<.05 or less. The table incicates that every cognitive
test but digit span and Hooper correlated with at least
one overnight respiratory index. Of additional impact is
the observation that in every case of a significant
correlation, deteriorating respiratory indices are
associated with worsening neuropsychological scores.
Respiratory indices with particularly heavy loadings on
neuropsychological scores included those quantifying
apneas, (Apnea index, Apnea + hypopnea index) as well as
those more directly representing deepening hypoxia (Mean
low saturation and Humber of desaturations ^>4^).
Subjects were grouped by level of nocturnal
respiratory disorder (no apneas/hypopneas, subclinical
level of apnea/hypopnea, and clinically significant level
of apnea/hypopnea). Means and standard deviations for the
neuropsychological scores from these groups are presented
in Table 5-15. The table suggests a general deterioration
of cognitive scores as level of nocturnal respiratory
distress increases. Oneway AHOVA procedures were applied
to the neuropsychological scores from each test. Signifi
cant between group differences emerged for V/AIS PIQ
LF(2,45)=6.9 p<.002] delayed recall of logical stores
[F[2,45)=4.0, p<.05] delayed recall of visual reproduc
tions [F(2,45)=5.4, p<.04], delayed recall of Rey complex


sleepiness, and muscular events (nocturnal myoclonus).
Further, these subjects had a mean age of almost 70 years,
indicating significant sampling and age differences from
the present study. Ninety-three per cent of these
subjects had at least one event, while 25% had high
numbers of apnea/hypopnea. These rates are much higher
than those found in the present sample. This is probably
attributable to the sampling differences noted above.
However, a recent report by Miles and Simmons (1984)
described apneas in a series of 1S0 patients referred for
complaints of heavy snoring. In this sample, 73% had at
least one event, while 9% had severe levels of apnea.
These data are quite comparable to the prevalences
observed in the present sample, 62% for at least one event
and 13% for high levels of apnea/hypopnea, and lend some
confidence to the present finding.
Thus the overall frequency of sleep disordered
breathing in middle-aged subjects with heavy snoring
appears to be between 62 and 73% for at least one event,
and between 9 and 13% for more severe levels of apnea.
The relatively close concordance of the two reports on
heavy snoring subjects is reassuring.
Factors which were found to be related to the
occurrence of sleep apneas in the previous review included
male sex, complaints of snoring, increasing age, and
obesity. The first two factors could not be examined in
the present sample because of their use as selection


28
disease or hypertension. Thus a group of SAS patients
without evidence of heart disease exists.
Taken together, these data indicate that cardiac
complications frequently, but not inevitably, accompany a
sleep apnea syndrome. That a large number of apneas may
be present without a significant increase in cardio
pulmonary complications is demonstrated by the Orr et al.
and Ancoli-Israel et al. studies. As the relationship
between these complications and a full blown SAS is
unclear, the possibility of cardiopulmonary sequelae from
a subclinical level of apneas is even more uncertain.
However, an assessment of cardiopulmonary status is
indicated by the frequent association of cardiopulmonary
complications with sleep apnea syndrome.
Arousal Deficits (Hypersomnolence)
Excessive Daytime Sleepiness (EDS), or hypersomno
lence, has been described as the single most common result
of SAS (Dement et al., 1978). Indeed, patients who meet
the criterion of number of apnea episodes and exhibit EDS
are described as symptomatic, while those with number of
apneas alone are labeled asymptomatic (Orr et al.,
1979). As an aside, this distinction illustrates again a
central issue of SAS. If patients exhibit the number of
apneas qualifying them for SAS, but no other pathology,
are they suffering from a pathological process? Dement et
al. (1978) believe that they are. They suggest that most,
if not all, patients with a clinically elevated number of


116
A-CCOC^>aM^CO I CTiCT\^L^COCOt>-
OOOOOOOt- ioooooooo
v~ Al Al AJ Al Al AJ I Al fA A fA AJ Al
CNJ t v~t~ CNJ IrrrrrtMrtM
vO m ltn "3- co ^ v- i r-i>-c\j>-m^£jcom
r~ rf'i rr (\| r- t
OOr-OO--r-v IOOt-OOOt-O
Os Al fA C\] f A AJ Al AJ lAJAJs-rAAJIAAJfA
O
I t-N- i
I OO l
i c\j i
I C\i AJ I
I lAvO I
I OO I
I Al AJ I
vCNCU^O^-tMi'Av N CO CA O Al f A ^ LA M)
AJ Al Al Aj f A f A ¡A fA NA fA (A fA fA rA-v^-^^--^'^'-^-


126
Feinberg, I., Koresko, R.L., & Heller, N. (1967). EEG
sleep patterns as a function of normal and pathological
aging in man. Journal of Psychiatric Research, 107-
144.
Fisher, J., de la Pena, A,, Mayfield, D., & Flickinger, R,
(1978). Psychobiologic determinants of subgroups of
sleep apnea. Sleep Research, _7, 221.
Francheschi, M., Zamponi, P.f Crippa, D., & Smirna, S.
(1982). EDS: A 1 year study in an unselected inpatient
population. Sleep, (3), 239-247.
Freides, E., & Allweise, C. (1978). Transient hypoxic
amnesia. Behavioral Biology, 22, 178-189.
Fujita, S., Zorick, F., Koshoreck, G., Wittig, R., Conway,
W. & Roth, T. (1981). Treatment of upper airway sleep
apnea with UPP: Recent experience. Sleep Research, 10,
197.
Garay, S., Rapaport, D., Sorkin, E., Epstein, H., Feinberg,
I.f & Goldring, R. (1981). Regulation of ventilation
in the obstructive sleep apnea syndrome. American
Review of Respiratory Disease, 124, 451-457.
Gastaut, H., Tassinari, C., & Duron, B. (1965). Etude
polygraphique des manifestations episoiques diurnes et
nocturnes du syndome de Pickwick. Revue Neurolgie, 112,
573-579.
Gellhorn, E. (1951). Sensitivity of the auditory
projection area to anoxia. American Journal of
Physiology, 164. 748-751.
Gibson, G., & Blass, J. (1976). Impaired synthesis of
acetylcholine accompanying mild hypoxia and
hypoglycemia. Journal of Neurochemistry, 27. 37-46.
Gibson, G., & Duffy, J. (1981). Impaired synthesis of
acetylcholine by mild hypoxia or nitrous oxide. JournaI
of Neurochemistry, 36(1), 28-53.
Gibson, G., Pulsinelli, W., Blass, J., & Duffy, T.
(1981). Brain dysfunction in mild to moderate
hypoxia. American Journal of Medicine, 70, 1247-1254.
Grant, I., Heaton, R., McSweeny, J., Adams, K., & Timms,
R. (1982). Neuropsychological findings in hypoxemic
chronic obstructive pulmonary disease. Arcnives ci
Internal Medicine, 142, 1470-1479.


e
the study if they reported any "serious" sleep com
plaint. Two of the subjects reported consuming hypnotics
for sleep induction, indicating that the criterion used in
this study was less stringent than that of Block et si.
(1979) report. Nine of the subjects (45%) suffered at
least one episode o sleep apnea with a mean number of
episodes of 2.5. All of the subjects with episodes of
sleep apnea were older than the mean age of the sample,
leading Webb to conclude that apnea is related to age.
Bixler et al. (1982) constructed an age stratiiied patient
representative sample of 100 subjects (M =41, F = 59).
In perhaps the most rigorous screening for sleep com
plaints in this group of studies, Bixler et al. eliminated
any subject with a sleep complaint or chronic medication
usage. For the entire sample, 6 males (14.6%) and 5
females (10.2%) had at least one episode of sleep apnea.
For subjects under 30, 3*3%; for those between 30 and 50,
12.8%; and those 50-74, 19.4% suffered episodes of sleep
apnea. For the subsample over 60, 25% had sleep apnea
activity with 8.5 episodes for males and 10.5 episodes for
females. In this sample, there was a nonsignificant trend
toward more apnea activity in males and a significant
positive correlation between age and sleep apnea.
Block et al. (1980) report data on 20 post menapausal
women with a mean age of 59. Eight women (40%) had epi
sodes of sleep apnea, with a mean number of episodes of


CHAPTER FOUR
DISCUSSION
Demographics and Incidence o
Nocturnal Respiratory Disorder
The present sample of 46 snoring males was selected
as a group likely to display a wide range of sleep
disordered breathing. These subjects, with a mean age of
50 years and a mean weight of 190 pounds, reflect a mostly
middle-aged to older male population which is moderately
obese. The demographic characteristics of this sample are
similar to those of the "typical" sleep apnea syndrome
patient (Guilleminault et al., 1978), although the
requirement of general good health insured that subjects
with a classic SAS were excluded. This sample may be
considered representative of subjects predisposed to sleep
disordered breathing, although at no more than a
subclinical level. Therefore it seems valid to consider
various risk factors as they relate to nocturnal
respiration.
In the present sample, 62% experienced at least one
apnea or hypopnea, while 13% suffered higher levels of
apnea/hypopnea. Only two other reports have described
samples selected for snoring or respiratory complaints.
Ancoli-Israel et al. (1981) examined 24 subjects with
complaints of respiratory difficulties, daytime
86


12
sketchy, but 7 (27%) met the sleep apnea syndrome
criterion of >30 episodes of sleep apnea. The fact that
these subjects were patients on a medical ward obviously
restricts generalizability of these findings. Finally, in
the only truly random sampling reported, Ancoli-Israel et
al. (1982) present initial data on randomly selected sub
jects. Forty subjects (sex not reported) with a mean age
of 71.5 were studied. Thirteen (32.5$) met the sleep
apnea syndrome criterion of 30 or more apneas. The sleep
apnea subgroup was older, 74.2, than the full sample.
The data collected from relatively random sampling
are primarily based on aging populations (>60). The data
from these studies suggest that between 32.5$ (Ancoli-
Israel et al., 1981) and 37$ (Carskadon and Dement, 1981b)
of aging subjects suffer high levels of sleep apnea
(AI>5). An even larger percentage, perhaps greater than
75$ suffer at least some episodes of sleep apnea. A
major deficiency in these randomly sampled sxudies is a
lack of data on younger subjects.
Thus data collected under different sampling proce
dures yield variable incidence rates of sleep apnea and
AI>S. Subjects without sleep or respiratory complaints
have a very low incidence (perhaps 0) of AI>5, between 14$
and 45$ of males surveyed had at least some apnea, while
females ranged between 10$ and 40$ experiencing some
apnea. Sampling subjects with sleep or respiratory com
plaints suggested that subjects suffering respiratory


Tabic B-4
Seven day sleep log means from 46 snoring males
ubject
f?
Daytime
Sleepiness
Dumber
of
Daps
Sleep
Latency
Number
of
Wakenings
Minutes
Waking
Minutes
Napping
Total
Bedtime
1
2
0
3
1
2
0
7
2
2
1
5
0
1
0
9
3
2
0
15
4
2
0
6
4
2
0
7
2
3
0
8
5
3
0
12
3
2
0
6
6
3
0
16
2
1
0
6
7






8
3
1
9
2
2
0
7
9
2
1
21
2
4
1
7
10
3
0
19
2
3
0
9
11
3
0
6
2
2
0
8
12
-
-
-
-
-
-
_
13
3
0
9
1
1
0
7
14
2
0
17
1
1
0
9
15

-
-
-
-
-
-
16
3
0
12
2
3
0
9
17
-
-

-

-
-
18
2
0
35
2
3
0
n
/
19
2
1
30
2
5
0
7
20
2
0
7
2
2
0
7
21
3
0
14
1
2
0
9
22
2
0
14
2
4
0
7
23
3
0
9
0
1
0
8
24

-
-
*
-
-
-
25
1
1
5
1
o
2
5


82
respiratory scores: WAIS PIQ, Wechsler Memory Quotient,
delayed recall of visual reproductions and Hey figure, and
Wisconsin card sort. Table 3-16 presents the data from
this analysis. The Hotellings T2 was significant,
(p<.001) v;ith delayed recall of visual reproductions and
Rey figure significantly predicted by the predictor
scores. Inspection of the individual regressions for
these two measures discloses that age, apnea index, and
number of desaturations >A% all formed significant
components of the prediction of delayed recall of visual
reproductions, while sleep efficiency was the sole
significant predictor of the Hey figure. These data seem
supportive of an independent contribution of both
nocturnal respiratory variables and sleep variables at a
limited level, to the prediction of cognitive/neuro
psychological scores.
In summary, various statistical procedures indicate
relationships between nocturnal respiratory measures and
non-verbal intelligence, verbal and non-verbal memory,
expressive verbal fluency, and ability to shift cognitive
set while problem solving. Division of subjects by level
of nocturnal respiratory distress indicated that subjects
with high amounts of apnea/hypopnea were impaired relative
to controls on measures of non-verbal intelligence, verbal
and non-verbal memory, and expressive verbal fluency, as
well as being impaired relative to subjects with
subclinical levels of apnea/hypopnea on measures of


I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy,
Assistant Professor of Clinical
Psychology
This dissertation was submitted to the Graduate Faculty of
the College of Health Related Professions and to the
Graduate School and was accepted as partial fulfillment of
the requirements for the degree of Doctor of Philosophy.
August 1985
(?
Dean, College of Health Related
Professions
Dean, Graduate School


Table E-7
Sleep variables from 46 snoring
Number
ub j.
tt
TOT
bleep
La tency
bleep
Liliciency
bleep
Tine
Time '/
Sta^e 0
Ota£e 0
Periods
1_
1
}Ob
0
0.0
103
30.4
3
1.7
2
373
3
O.b
210
12.9
0
5.0
5
A
401
0
0.0
304
4.2
10
5.9
j
374
b
0.7
20 3
22. b
12.0
3.0
0
487
2
o.o
450
3.6
0
5.9
7
431
1d
o.o
371
10.1
14
5.3
0
30b
14
0.7
203
23.1
9
3.2
4bri
54
O.b
3b5
7.5
2
1.9
1U
33b
94
0.3
110
37.0
3
1.9
11
40/
30
o.b
341
0-3
7
2.1
12
402
154
0.5
199
9.9
6
1.4
13
300
50
0.7
2b7
23.5
7
3-7
14
33b
135
0.6
203
0.1
9
0.5
1b
302
34
O.b
335
3.7
7
0.3
1b
33b
30
0.6
219
26.0
10
2.0
17
374
25
O.b
30b
1.3
2
0.9
1d
343
11
0.9
320
3.6
7
2.7
1V
320
24
0.2
73
33.0
2
1.9
20
44b
67
0.7
321
13.7
7
4 .0
21
323
17
0.7
23b
15.1
9
1.1
22
348
9
O.b
200
17.4
7
3.9
2 >
2b 4
t>
0.9
2bb
3.3
4
1 .b
24
37 b
19
0.0
290
19.2
12
2.0
2b
32o
0
o.o
203
7.5
12
2.3
2b
35b
23
0.6
221
33.b
33
2.1
27
3b4
2b
O.b
232
20.5
1b
6.9
2d
327
19
0.0
2b5
4.3
4
2.9
20
344
b
0.0
27 3
19-2
11
2.7
50
340
20
o.o
295
7.0
6
3.1
31
-
-
-
-

-

ales
% Slow Number
lime
Jo OLcl/IG
Wave
UEft
Mean
MEM
HEM
2
3.
4
bleep
Chan/'es
La tency
HEM
6ycii
0
60.0
o
0
0
14
0
0
0
0
71.0
9.9
0.4
11.9
51
0
0
0
17.9
40.3
0.7
16.5
26.3
63
02
1b
90.7
10.9
46.3
4.9
11.5
21.2
32
75
20
233.0
16.2
55.5
2.9
15.0
19.5
47
72
13
bb. wj
10.2
69-9
2.1
1.7
4.3
64
163
1 3
61 .0
12.6
63.0
1.0
0.5
3.1
38
190
23.5
bb.o
25.4
47.3
6.1
11.0
19.3
36
3b
2b. 5
1 19.7
1.5
26.0
1.9
9.9
2b. 1
1b
221
4
0
21 .0
65.4
2.7
1.6
4.7
42
bb
27
145
0
07.0
0
0
0
16
0
0
0
2.9
60.5
2.6
6.9
12.3
30
247
5
49
17.0
65.5
4.5
4.5
9.8
32
02
19
167.0
17.0
59.2
4.3
15.5
20.6
23
212
29.3
10b. 0
2.6
b2.5
6.0
0.0
8.2
44
292
0
0
17.1
bb. 3
4.5
9.7
14.4
21
97
2b.5
120
10.1
53.b
2.1
19.9
22.8
30
71
15
75.7
0
41.7
2.5
0
5.4
11
0
0
0
1b.7
57.3
2.9
7.5
12.1
40
153
31
122
14.0
32.5
7.5
9.7
20.1
43
131
19.5
114
7.9
55.9
2.3
14.3
20.4
32
07
15.3
1/8
9.0
07.2
6.9
13.7
21.3
34
114
15.5
93
0.9
61.3
2.2
5.0
9.6
39
291
52
0
21.1
60.4
2.6
7.1
10.5
44
43
10.0
47.0
0
64.3
0
0
0
71
0
0
0
9.b
63
0
0
0
50
262
2b
0
13
6b. 4
4.7
0.7
13.9
3o
01
9
64
13.9
62.1
6
0
0
32
153
1b
3b. 3
15
63.4
4.7
6
3.1
30
65
1b
111
(V)


65
Table 3-5. Means and standard deviations for health related
variables in 46 snoring males grouped by level
of apnea/hypopnea.
No Apnea/
Low Apnea/
High Apnea/
Hypopnea
Hypopnea
Hypopnea
Systolic BPa
129-1(14.2)
128.7(12.6)
127.0(7.5)
Dia¡
stolic BP
85.2(13.4)
83.9(11.0)
81.6(7.6)
CMI*
3 Overall Score
14.7 (7.4)
16.1(10.5)
13.8(7.4)
CMI
Respiratory Score
1.4 (1.2)
1.1 (1.8)
2.0(1.4)
CMI
Cardiac Score
1.6 (1.3)
1.2 (1.2)
1.0(1.2)
CMI
Neurological Score
1.2 (0.9)
1.6 (1.5)
0.8(0.6)
"f BP = blood pressure.
b CMI = Cornell Medical Index.
Nocturnal Respiratory and Sleep/V/ake Variables
EEG Data
Electroencephographic variables were recorded
overnight for all subjects. Records were scored by a
trained technician who utilized the system of Agnev and
Webb (1972) to determine sleep stages. In three records,
EEG tracings were judged inadequate for differential sleep
staging. However, a judgement of sleep vs. wake was made
for these subjects in order to form a basis for the
respiratory indices. These three EEG records were
excluded from the following analyses, leaving 43 records
as a base for analysis.
Fifteen variables (time in bed, pure sleep time,
sleep efficiency index, number of stage 0 periods, time %
stage 0, sleep latency, time % stages 1, 2, 3, 4, and REM,
latency 1st REM period, mean REM period length, mean REM


131
Wechsler, D. (1955)* Wechsler Adult Intelligence Scale:
Manual. New York: Psychological Corporation.
Weitzman, E. (1979). The syndrome o hypersomnia and sleep
induced apnea. Chest, 75, 414-415*
Weitzman, E., Kahn, E., & Poliak, C. (1980). Quantitative
analysis o sleep and sleep states before and after
tracheostomy in patients with the hypersomnia sleep
apnea syndrome. Sleep, (3/4), 407-423*
Weitzman, E., Poliak, C., Borowieki, B., Burak, B.,
Shprintzen, R., & Rakoff, S. (1978). The hypersomnia
sleep apnea syndrome: Site and mechanism of upper
airway obstruction. In C. Guilleminault & W. Dement
(eds), Sleep Apnea Syndromes. New York: Alan R. Liss
Inc.
West, J. (1984). Human pnysiology at extreme altitudes.
Science, 223, 784-786.
Zorrick, T., Roehrs, T., Koshorek, G.,
Hartse, K., Wittig, R., & Roth, T.
sleepiness in various disorders of
174.
Sicklesteei, J.,
(1982). Patterns of
EDS. Sleep, , 165-
Zwillich, C., Devlin, T., White, D., Douglass, N., Weil, J.,
& Martin, R. (1982). Bradycardia during sleep apnea.
Journal of Clinical Investigations, 69, 1286.


95
disturbed nocturnal breathing. In the first case,
increasing respiratory disturbance causes sleep
deprivation, manifested in increased subjective sleepiness
(SSS ratings) and compensatory sleeping (increased
napping). The other side of this coin is a sleep pressure
v/hicn causes disturbed subjects to fall asleep quickly
(lowered sleep latency), and subjectively fail to report
nightly wakenings, while subjects with intact respiration
and hence less sleep pressure experience a longer sleep
latency and awareness of nightly awakenings. Admittedly
this is speculative, and two immediate problems with this
formulation are evident. First, the lack of correlation
between LEG sleep and respiration of the experimental
night is problematic, as well as the lack of relationships
between respiration and the evening "naps." However, it
has already been speculated that an exaggerated first
night effect obscured correlations between sleep and
respiration, and it is new speculated that the two naps,
at 7 and S pm, missed the afternoon trough of sleepiness
which might have better reflected sleepiness.
The interpretation of increasea sleepiness from
nocturnal respiratory disturbance is useful in explaining
the pattern of findings observed in this study, although
some problems with this formulation were noted.
To summarize the main aspects of the discussion of
sleep/vake variables, it is noted that these snoring
subjects achieved a night of sleep comparable to those


21
subjects to be uniformly obese* A review of the
literature is in order to clarify this point. First
normal, then obese, then SAS groups will be reviewed.
Bixler et al. (1982) found that their subgroup of
subjects with sleep apnea activity weighed significantly
more than the remainder of the sample. Kales et al.
(1982) found that within their sample of 200 insomniacs
and 100 normals, a significant positive correlation
emerged between a weight:height ratio and apnea acti
vity. Block et al. (1979) found a positive correlation
between apneas, oxygen desaturation, and weight within
their normal males, while females did not demonstrate this
relationship. Using an older female sample, Block et al.
(1980) found a positive correlation between weight and
oxygen desaturation, but not apnea. Thus in normal sub
jects, there appears to be a positive relationship between
weight and apnea activity, although this may be restricted
to males.
A research approach which bears on this question
involves studying obese subjects alone. Two studies on
successive referrals for gastric bypass surgery were
noted* Sicklesteel et al. (1981) found that of 19 suc
cessive referrals, all 14 males exhibited sleep apneas,
while only 5 of the females showed evidence of oxygen
desaturation alone. Harmon et al. (1981) studied gastric
bypass patients and found 5 of 6 males to suffer apnea, as


25
suffer sleep apneas. A significant proportion of snorers
might be expected to suffer apneas, while few nonsnorers
might have apnea. The last variable, weight, while having
a positive correlation with apnea activity in the general
population, is of uncertain pertinence to SAS. Addi
tionally, changes in weight may occur without changes in
apnea activity. Therefore, this variable has a less than
explicit connection to apnea activity and violates the
second principle adopted to consider these variables.
This review of available data suggests that sampling
aging, snoring, males would generate a high proportion of
sleep apnea activity. External validity for this idea is
drawn from its similarity to the "typical'' SAS patient
described earlier. Extant data suggest that it may also
prove useful to utilize weight measures as a covariate, in
an attempt to control its apparent influence on apnea
activity.
Deficits Associated With Sleep Apnea Syndrome
Sleep Apnea Syndrome is thought to result in several
deficits. These include cardiopulmonary complications,
hypersomnolence, hypoxia and intellectual changes. Some
researchers have suggested that these deficits may be
present, in an attenuated form, in subclinical apnea
(Ancoli-Israel et al., 1981; Carskadon and Dement,
1981b). A discussion of the evidence for the presence of
these deficits will be presented, followed by proposals
for appropriate measures of these variables.


22
opposed to none of the females. This evidence provides
further evidence for a role of increasing weight in apnea.
The data on symptomatic SAS subjects are more com
plex* Several studies (Sullivan and Issa, 1980; Iveitzman
et al., 198C; Garay et al., 1981; Zwillich et al., 1982)
report that all their SAS subjects were obese, although
definitions of obesity either vary widely or remain un
specified. In contrast, Guilleminault et al. (1978)
reported that only 60# of their SAS sample was greater
than 15/S above ideal weight. Coverdale et al. (1980)
indicate that only two of their subjects proved to be
obese, according to their definition (>125# of ideal
weight). These reports suggest that most, although not
all, SAS subjects are obese.
Several investigators have sought to compare weight
in symptomatic (somnolent) and asymptomatic subjects with
enough sleep apnea episodes to be diagnosed as SAS. Orr
et al. (1979) compared 4 symptomatic SAS subjects with
matched asymptomatic subjects. In this sample, symptoma
tic SAS subjects were heavier than asymptomatic sub
jects. Examination of the data of Berry and Block (1963)
indicates that their symptomatic subjects were heavier
than their asymptomatic subjects (118 kg vs. 91 kg).
Standing in contrast to these data is a study reported by
Fisher et al. (1978), which included 19 subjects referred
for SAS evaluation. Based on number of apneas per night,
Fisher et al. divided their subjects into 3 groups. Group


91
In light of these caveats, it is of interest that two
measures of desaturation, mean low saturation, and number
of desaturations >4% were associated with increasing blood
pressure readings. While a report by Guilleminault et al.
(1978) noted an increasing prevalence of hypertension in
subjects diagnosed with sleep apnea syndrome, a
relationship between desaturation and hypertension has not
been previously reported. This finding was observed in
spite of the fact That subjects prescribed hypertensives
were not medication free for the pressure reading.
Additionally, a check of self reports of diagnosed
hypertension showed six subjects in the no event group as
having diagnosed hypertension, as opposed to six in the
low apnea/hypopnea group, and one in the high
apnea/hypopnea group, data which did not link
apnea/hypopnea, per se, with hypertension in these
subjects. The isolated finding of a relationship between
mean high saturation and overall number of symptoms
endorsed on the CMI is more puzzling and does not follow
any logical analysis.
In summary, the present study's assessment of health
variables must be regarded as somewhat limited. In spite
of this, an association between two measures of desatura
tion and a measure of blood presure was noted. It is
speculated that blood presure may be more responsive to
nocturnal oxygen saturation than to the apneic events
alone.


66
cycle length, % slow wave sleep), reflecting multiple
aspects of the sleep of these subjects, were derived from
the sleep stage scoring. Table 3-6 presents the means ana
standard deviations for these subjects. Although subjects
spent over six hours in bed, the table indicates that not
quite 5 hours were spent asleep. The remaining variables
suggest that an increase in sleep latency, and time awake,
is responsible for the diminished sleep time, while
lighter sleep is also increased.
Table 3-6. Means and standard deviations of selected sleep
variables from 43 snoring males.
Time in Bed
Sleep Latency
Pure Sleep Time
Sleep Efficiency Index
Number Stage 0 Periods
Time % Stage 0
Time % Stage 1
Time % Stage 2
Time % Stage 3
Time % Stage 4
Time £ Stage REM
Latency 1st REM Period
Mean REM Period Length
REM Cycle Length
% Slow Wave Sleep
Standard
Mean Deviation
377.3
45.1
29.2
32.9
288.7
75.8
.76
.16
8.6
5.6
14.3
12.7
3.3
2.8
60.8
10.3
3*4
2.4
7.9
6.4
11.5
7.2
124.7
92.2
14.9
9.4
74.9
57.7
12.9
8.3
Note: see text for explanation of N.
Partial correlations, controlling for age, were
calculated between these sleep and the respiratory
variables. Out of a total of 105 correlations, only one


114
C\JC\]C\JC\C\jT-C\U\irO|CSjT-C\J C\]C\JtC\]C\JC\Jt-C\IOOJ
T- fO, k\ io NA fi"\ n rr\ 1^1 CM ro 00 r^i O to
C\| C\¡ rr\ CM C\l C\J OJ C\J CM rO A C\J i^r(M(MtOr rr Or
r- C\j CM CM O OJ CM rr\ jo* x- M ¡*^ s- CM CM lA t- CM CM O r-
^0CMC\JCM0x-C\JCMT-T-C\^-C\J^-T-r-0 O CM CM ^ r- h*, O r- ro> K\ CM CM ^rf'.OrOCO
O v-
COT-C\Jirur\t-tACOC\Jf^C\JC\J rCM^OCVO^v-Or
r-C\jC\JC\¡C\JC\JC\JC\jC\JC\JC\¡C\jC\|C\jC\iC\jCvJ\]C\Jv-OC\J
O C > 00 N- CQ O'. C ce C"- N- ¡>
CO NO CO
m^NCj OiOr-CM W'N ^ iT\O N CC C>0 r* C\J
CM CM CM CM CM rO NN n fA ro IN n ^ ^ ' N '


39
human vs. animal subjects and acute vs. chronic hypoxia.
First, the data available on animals will be reviewed,
followed by that extant on human subjects. Finally,
behavioral variables sensitive to changes in these
variables will be summarized.
The effects of chronic hypoxia on animals have not
been exhaustively studied (Davis, 1975). One project
which did examine this variable involved exposure of rats
to up to 36 hours of *10% oxygen. Davis (1975) observed
the behavior of these animals and concluded that they
appeared normal. Although hyperventilation was noted, no
permanent effects of the hypoxia were found. Hanbauer et
al. (1981) examined the effects of 10? O2 on rats exposed
for up to 4 weeks. They found that relatively short term
(2 days) hypoxia resulted in an increase in dopamine (DA)
content in the carotid body. An increase in norepine
phrine (WE) content was noted after 1 week of exposure.
As the carotid bodies are implicated in the control of
respiration, the authors concluded that different
mechanisms are operative in the adaptation to short and
long term hypoxia. Unfortunately, little other data have
been reported on the chronic effects of hypoxia on more
complex functioning. This is particularly problematic as
it are these functions which are thought to be affected in
humans exposed to hypoxia.
Certain aspects of acute hypoxia in rats have been
carefully examined. Gibson et al. (1981) reviewed the


105
non-verbal memory, expressive verbal fluency, anc
cognitive flexibility;
8) the subset of heevy snoring males with more than
5 events per hour appear to occupy an inter
mediate position between normals and those with
sleep apnea syndrome.


88
criteria. A wide range of age was achieved in this
sample, but neither non-parametric nor correlational
procedures revealed significant relationships between
increasing age and sleep disordered breathing. Although
several previous reports have noted relationships between
increasing age and sleep disordered breathing, typically
this relationship is moderate for apnea (Block et al.,
1979; Carskadon et al., 1980; Bixler et al., 1982).
Additionally, no other correlational data are available in
subjects selected for sleep and breathing complaints. It
may be speculated that the effect of snoring masks the
effect of age in this group. Alternatively, some quirk of
sampling may have generated this effect. Correlational
procedures, however, revealed substantial relationships
between weight and sleep disordered breathing. This
finding is consistent with several past reports. To
summarize, the present data support relationships between
increasing weight and nocturnal events, while
relationships between age and sleep disordered breathing
are not found in this sample of heavy snoring males,
perhaps because the effect of age is masked by the effect
of snoring.
The temporal characteristics of the apneic events
assessed by two hour interval suggested an increase in
events between '5 and 5 am, although differences in sleep
onset time and wakeup time render any interpretation of


2 V
10U
090
099
066
01
00
28
129
133
133
135
09
12
29
122
116
120
114
09
09
30
137
133
138
120
10
09
31
121
102
114
097
03
02
32
123
114
120
118
08
11
35
123
112
119
109
13
O'j
34
133
110
127
129
09
10
33
113
120
11(1
120
09
12
36
122
102
114
121
10
08
37
133
127
134
133
15
14
36
138
116
1 31
118
07
00
39
134
133
137
133
08
12
4
100
109
105
107
03
09
41
130
123
129
133
08
12
42
134
135
130
143
15
12
43
111
111
112
100
09
03
44
123
133
131
13b
09
13
43
136
142
131
143
13
14
46
120
114
119
137
13
11
WMSMQ = Wechsler Memory Scale Memory yuotient
Delayed Recall Logical Stories
Delayed Recall Visual Reproductions
^ Delayed Recoil Rey figure
Wisconsin Card Sort
6
OO
16
2
57
60
09
8
20
26
3
70
bO
19
8
25
26
4
57
5b
09
b
24
26
5
60
68
1b
7
1y
2U
0
53
44
11
u
20
27
4
70
77
13
5
3
24
1
49
42
Ot
7
30
26
4
80
72
11
7
21
27
4
71
70
1
7
26
25
5
54
50
11
7
34
25
5
66
6/
17
0
20
12
4
55
47
26
b
30
25
5
73
74
17
7
19
23
2
58
49
10
7
30
28
5
69
61
14
5
2b
28
5
53
50
18
7
13
17
3
50
49
17
7
28
26
4
74
65
10
7
2/
25
5
55
48
22
9
25
29
2
6b
60
13
3 U


93
have enough tree variance to respond to the respiratory
events of the night, Alternatively, the LEG scoring
system, which averages across 1 minute periods, might miss
the short arousals ana fragmentation resulting from
events. Whatever the reason, the lack of covariation of
sleep and respiratory measures is puzzling in light of the
report of Guilleminault et al. (1978) who described
arousals as a frequent concommitant of nocturnal
respiratory events, although Bixler et al. (1982) failed
to replicate this in a group of normal subjects.
Between group differences on the sleep variables were
also absent when subjects grouped by level of respiratory
distress were compared on sleep characteristics. Again it
may be speculated that the effect of an exacerbated first
night might obscure possible differences.
In summary, these snoring subjects appeared to
achieve a night of sleep comparable to tnose observed in
similar studies. Longer sleep latency, more awakenings,
and lighter sleep were observed. Significantly, these
changes characterize a first night effect (Agnew, Webb,
and Williams, 1966) of lighter sleep in a strange bed, and
may have been increased by the discomfort of the recording
procedures. Less than a chance number of correlations
were observed between nocturnal respiratory and sleep
parameters. This surprising result is attributed to an
exacerbated first night effect which probably functioned
to obscure possible respiratory disturbed sleep parameter


58
Table 3-2. Incidence of low and high levels of
apnea/hypopnea by age in 46 snoring males.
Age Group
U
Low Apnea/
Hypopnea
High Apnea/
Hypopnea
50-39
14
.71
0
40-49
7
.57
.29
50-59
11
.72
.27
60+
14
.43
.07
Overall
46
.62
.15
apnea/hypopnea. Note that subjects with high levels of
apnea/hypopnea are included in the overall total of
subjects with at least one event for the following
analyses. The classification of "high" levels of
apnea/hypopnea utilizes the criterion of an apnea +
hypopnea index greater than 5 a frequently used clinical
cut-off score. The table also presents the frequency of
respiratory distress broken down by age. No simple age
related trend is apparent, and a X2 analysis of the two
factors of age and respiratory distress vas not signi
ficant for those with at least one event (X2=3.2; p=NS),
or for those with high levels of apnea/hypopnea (X2=5*9;
p=NS).
A further investigation of possible age by level of
apnea activity involved division of the sample into those
above ana below 60 years of age crossed with those above
and below 5 events/hour. The X2 value for this comparison
was also nonsignificant (X2=0.6; p=NS).
Subjects were


TAELE OF CONTENTS
Page
ACKNOWLEDGMENTS ii
LIST OF TABLES v
ABSTRACT vii
CHAPTER
ChE INTRODUCTION 1
Definitions and Terminology 2
History and Cheracteristics of Sleep Apnea
Syndrome 3
Incidence of Sleep Apnea Activity 5
Temporal Characteristics of Apneas 13
Variables Associated With Sleep Apnea
Syndrome and Activity 17
Deficits Associated With Sleep Apnea
Syndrome 23
Cardiopulmonary Deficits 26
Arousal Deficits (Hypersomnolence) 28
Oxygen-Desaturation and Cognitive Deficits..30
Measurement of Deficits Found in Sleep Apnea
Syndrome 34
Measurement of Cardiopulmonary and Health
Deficits 34
Measurement of Arousal Deficits
(Hypersomnolence) 35
Measurement of Hypoxia and Its Sequelae 38
Statement of the Problem 46
TWO METHOD 48
Subjects 48
Apparatus 49
Measures 50
Procedure 54
Statistical Procedures.. 55
THREE RESULTS 56
Demographics and Incidence of Nocturnal
Respiratory Disorder 56
iii


32
of oxygen saturation in 20 post menapausal females was
reported by Block et al. (1980). Eleven of these women
exhibited desaturation with several subjects desaturating
to less than 85#. Dolly and Block (1982) studied a group
of 17 males and 3 females with a mean age of 49 years.
These normal subjects dropped from a mean baseline satura
tion of 96*3% to a mean maximum desaturation of 88.3$.
This evidence indicates that desaxurations of 10-12?' are
common in normal males and post menapausal females while
premenapausa1 females rarely, if ever, desaturate.
Oxygen saturation levels in SAS patients are not
commonly reported. Using an ear oximeter, Garay et al.
(1981) studied 11 SAS patients who were divided into 6
daytime eucapnics and 7 daytime hypercapnics. Eucapnics
fell from a baseline saturation of 95% to a mean maximum
desaturation of 73*5/, while hypercapnics fell from a
baseline saturation of Q9% to a maximum desaturation of
62.8# (calculated from tabled data). These date were
collected during daytime "nap" studies, however, and hence
is somewhat questionable. Berry and Elock (1983) recorded
9 heavy snoring males who suffered many apneac events.
Four subjects were somnolent, while 5 were not. Symptoma
tic subjects fell from a baseline of 95# to a mean maximum
desaturation of 44#, while asymptomatic subjects desatu-
rated from a baseline of 94# to a mean maximum desatura
tion of 83#.


CHAPTER TWO
METHOD
Subjects
Heavy snoring males were recruited via newspaper
advertisements, notices, and phone calls to a list o
aging subjects reported by Webb (1982), Participants were
required to be male, self reported heavy snorers, over 30
years o age, and in general good health (i.e., self
described healthy and normal and not under the active care
of a physician for illness). Subjects with a history of
head trauma or alcoholism were excluded. Potential
subjects who met these criteria were offered a chance to
participate in a "snoring study" in which they would
complete testing during an evening, sleep overnight in a
lab with several physiological parameters recorded, and
complete one week's worth of "sleep logs," for a payment
of fifty dollars. Volunteers were scheduled on a first
come, first served basis. A total of 60 subjects were
studied, with 9 dropped because of technical problems with
their overnight recordings and 5 excluded because of
health problems. Thus 46 subjects represented the final
sample. These 46 subjects had a mean age of 49*9 (sd
13-1) and represented a wide cross section of the
population, ranging from the unemployed to university
48


27
Zwillich et al. proposed that bradycardia during apnea
results from increased vagal tone mediated by carotid body
chemoreceptors. Fujita et al. (1981) reported on 12 male
SAS patients. All reportedly suffered cardiac arrhythmias
during sleep. The body of data suggests that SAS patients
frequently suffer from cardiac abnormalities such as
hypertension, arrhythmias, and bradycardia.
A second body of data exists which indicates that SAS
patients do not invariably have cardiopulmonary compli
cations. Coverdale et al. (1980) studied 14 patients, of
whom 8 were diagnosed with SAS. Only 2 of these patients
had evidence of corpulmonle, while 2 more had systemic
hypertension. None of these patients demonstrated severe
brady or tachycardia during sleep. In Orr et al. (1979),
8 patients, all of whom were technically diagnosable as
having SAS, all symptomatic (somnolent) patients had evi
dence of right sided heart failure as well as frequent
arrhythmias curing apneas. Asymptomatic subjects had none
of these symptoms. Kreiss et al. (1982) studied 26
Veterans Administration ward patients and iound 7 to have
a SAS. The SAS patients had significantly more signs of
congestive heart failure, but not hypertension or
angina. Ancoli-Israel et al. (1981) sampled 24 elderly
subjects with complaints of sleeping respiratory ais-
orders. Nine of these subjects were diagnosable as SAS,
but this group did not have significantly more heart


5
series of 50 patients diagnosed with SAS over a period of
six years. Age ranged from 28-62 with a mean age of
45.5* The sample consisted of 48 males and 2 females.
Thirty-nine were referred on the basis of excessive day
time sleepiness, while 7 were referred for loud snoring
and abnormal movements reported by their spouses. Sixty
percent of the patients were more than 15$ above ideal
weight. Other symptoms included hypnagogic halluci
nations, automatic behaviors, intellectual deterioration,
personality changes, impotence, morning headaches, and
hypertension. Veitzman, Kahn and Poliak (1980) report
data on a group of 10 SAS patients seen in their clinic.
These subjects displayed the following characteristics:
age between 38 and 47, obesity, male sex, nicotine depen
dence, hypertension, complaints of excessive daytime
sleepiness, and the presence of serious cardiac arrhyth
mias. (This group was selected from a larger sample of 38
patients.) Thus, pathological symptomatology associated
with a diagnosed SAS include excessive daytime somnolence,
cardiopulmonary complications and cognitive/intellectual
changes; a "typical" SAS patient is a middle aged male
who snores and is obese.
Incidence of Sleep Apnea Activity
The incidence of sleep apnea activity and sleep apnea
syndrome have been frequently confused. While sleep
apneas have been an important diagnostic tool for identi
fying sleep apnea syndrome, they have not been


13
difficulties had very high incidences of apnea (almost
100J2 for both males and females) with increase of inci
dence of AI>5 (55% for males, 23% for females). Sampling
without regard to sleep or respiratory difficulties is
largely confined to aging populations. Incidences of
apnea in these aging populations would seem to be rather
high (88% males, 72% females) while incidence of AI>5 seem
to be in the range of 32%-37%.
It is clear that different populations (as defined by
sampling strategies) suffer different incidences of sleep
apneas and AI>5. The clinician dealing with sleep dis
orders would be wise to consider the acturial base rates
of apnea and AI>5 provided by these data. It is apparent
that sex, age, and respiratory status are significantly
related to the incidence rates of SAA and SAS. These
factors must be considered in any study of apnea and asso
ciated deficits.
Temporal Characteristics of Apneas
The temporal characteristics of sleep apneas repre
sent a potentially important variable. Above it was shown
that a wide variability in number of sleep apneas occurs,
and the arbitrary nature of the division between SAA and
SAS subjects was noted. Therefore, it is of interest to
determine whether the duration of sleep apneas provides a
more reliable distinction between the two groups. Another
variable of interest involves the placement of apneas
within sleep, as Guilleminault et al. (1976) noted that


120
rO'rO^riO^OOOOOr OtOOO^OOr
N(\j r- v C\|
^oot-o o O'sTinv md ovcoA r- r--
CM ^-vD mo CO CO CM CT\ v t-cOf-T- 't lA
CM CM C\J
a 7- v- ^^r\jN^tri^vDNirvntn^inoir\'j)c\j
c^ cts xi o^ en o c'i ao cr. en o\ en ^ en en en en en en en en en
O A >- O A A O ^ X) CM CQ CO IAt- LHC0 NfOtO m
N'lTiLTiM^in' iT.VOsDNNO NC0 T
cncncncncncnencncncncncncn CTv cncncncncncncncn
OOCMs-^rOAOCMOaoOvO OfflOOOOOt^
a) Ti^r OCCO^ A KO
\S\ r~ \S\"~ V-
rA y- rA
CMOO^iTiO^O^OOOO 00 OOOOOOOOCNJ
OCMf ^O.^^MOOOO O O-OOOOOO
OOOOCn 0^-0^0000 C0 oooooooo
OOOOfMOrOCMOOOOOOOOOOOOO
A vC C0 Cn Q C\J A ^ AAD X- CQ O O v- C\J A -vi- LTVvG
CM CM CM CM CM ¡A -A A A rA A IA A '


£* ^ ^ ^ ^ V^J ^ \^> ,\J (V
C ui ^ ^ re
^ CT CT O' CT C? Ui er V CT A** en en efi en *o. c
e* O VC C C C \T Wj -* C ko ^vCU*^^
^ _i [\j u ^ i rvj 4* ^ i i ^ .__!
unroo^^cr rocrcroiv^j^crc^ecrro
CT -< CT cw cr -< cr -%: c cr -c ^ c -< -e c cr
ro-*-ii\j->i\;r\j-> -i-iiNjivro-^roivro-*
-i^cr.^c:^ ^uia^vc c ^c r
vj* rv er en Cz ru vx c C C re \r u. r\j [\j e* r\j e>j e*j ro ro f\: rv re fv ro i\j -> r\j re
* en o- vji C ^ crouiO'C^CvC^JC: crctvc^-
o Co f\: c o c n Cat c !T c tt u v£
f\J
^rccT'V)^iV^Cf\jf\-'C e?
CCC^O^UiCCa^CC
IV
c
ro rv v v
c->c^ococcoccccccccc
cr c v' c1 cr < v s c: *>; vc c: a cr c: -X t
crovicco^uicccc^cGNccc
ccccccccccccccccccc
cen-Ceri->ercc^
CCCCCCCCGCCCCCCCCCC
C^WCCCCC^fUCN'^fy-V;. V. e>- X
cccccoccccccccccccc
rv^f\j-k^^CJV-^CCUi re e> C -1
cccccocccoccccccccc
v ror po-^c
ou


ACKNOWLEDGMENTS
The author would like to thank his chairman, Dr. V/ilse
E. Webb, for his moral and material support, as well as
guidance through the completion of the project. Also due
thanks are the other members of the supervisory committee,
Dr. A. J. Elock, Dr. Russell M. Eauer, Dr. F. D. McGlynn,
and Dr. Rudy Vuchinich, for their support, advice, and
contributions. Several people provided technical assistance
and are due a vote of thanks: Dan Switzer, Alex Masterton,
Karla Smith, Alton Howard, Albert Briggs, and Edna Pearson.
The author wishes to acknowledge the moral anti
financial support of his parents, Mr. and Mrs. John G.
Berry, without which the project would never have been
undertaken. Also of considerable help were the
encouragement and advice of C. M. Carswell, D. Haymaker, and
S. Brody, all of whom deserve thanks. Cynthia Zimmerman
provided expert typing assistance for the manuscript.
ii


73
Table 3-11* Significant (p<.05) Pearson correlations between
nocturnal respiratory and sleep log variables in
46 snoring males.
Number
Minutes
of Total
Sleep
Nightly
Naps Bedtime
Latency
Wakenings
Apnea Index
Apnea+Hypopnea Index
Seconds in Events
Mean High Saturation
.237
.250
.359
Mean Low Saturation
Kumber of
.265
Desaturations >_4%
.240
Number of
Desaturations y]Q%
.240
Table 3-12. Means and
variables
by level
standard deviations of sleep log
derived from 46 snoring males grouped
of apnea/hypopnea.
No Apnea/
Low Apnea/
High Apnea/
Hypopnea
Hypopnea
Hypopnea
Daytime Sleepiness
1.9 (.9)
1.9(1.1)
2.4 (.3)
Number of Naps
0.2 (.2)
.3 (.4)
.3 (.3)
Minutes Mapping
0.0(0)
.2 (.5)
0 (0)
Total Bed Time
6.5(3.1)
5.8(3.4)
7.5 (.7)
Sleep Latency
10.5(8.1)
11.1(9.8)
8.4(4.7)
Number of Wakenings
1.3(0.8)
1.3 (.9)
1.3 (.2)
Minutes V, a king
2.0(1.4)
1.7(1.1)
1.7 (.3)


1C1
hypoxia. Acute hypoxia has recently been shown to signi
ficantly alter the synthesis of ACh, a neurotransrr.itter
implicated in memory and cognitive changes in young sub
jects experiencing cholinergic blockade, and older sub
jects experiencing age related disruption of cholinergic
pathways. The hypoxic alteration of ACh turnover is
proposed as the proximal cause of cognitive/memory
impairments found in subjects with nocturnal respiratory
events.
Although the cerebral hypoxia hypothesis is highly
speculative, it does generate predictions which might be
used to tost its efiicacy. For instance, subjects with
differing levels of hypoxia might be expected to exhibit
diiiering levels of ACh turnover, as well as cognitive
differences. Administration of nocturnal oxygen should
improve both cognitive and cholinergic variables. Addi
tionally, manipulation of cholinergic pathways might
improve cognition in subjects with nocturnal respiratory
events. However, until a demonstration of this sort is
carried out, this line of reasoning remains highly specu
lative.
In summary, the present findings indicate clear rela
tionships between nocturnal respiratory and cogni-
tive/neuropsychological variables. This result is con
sistent with past reports of neuropsychological
impairments in subjects experiencing hypoxia. Apparently,
heavy snorers with multiple apnea/hypopnea episodes may be


80
Table 3-15* Keans and standard deviations of cognitive
variables in 46 snoring males groupec by level
of apnea/hypopnea.
No Apnea/
Low Apnea/
High Apnea/
Variable
Hypopnea
Hypopnea
Hypopnea
VAIS Full IQ
124.9(11.2)
118.3(13.9)
111.5 (9.6)
WAIS Verbal IQ
123.7(12)
118.5(14.5)
114.3(10.4)
WAIS Performance IQ
122.8(11.5)
114.5(14.4)
100.1(12.5)a
Y/eschler Memory
Scale Memory
Quotient
124.7(14.5)
117.1(19.0)
107.8(18.7)
Delayed Recall
Logical Stories
9-9 (2.4)
6.6 (3.2)
5.8 (4.0)a
Delayed Recall
Visua1
Reproductions
10.5 (3.7)
10.4 (3.4)
f.1 (4.7)
Digit Span
6.6 (0.6)
6.8 (1.2)
6.6 (.6)
Delayed Recall
Rey Figure
25.6 (5.4)
23.2 (6.0)
15.1 (fc.4)s,b
Hooper Test
25.2 (2.6)
25.5 (4.6)
22.8 (4.1)
Wisconsin Card Sort
4.1 (1.2)
3.4 (1.5)
2.6(1.6)
Finger Tapping
Right Hand
62.6 (6.5)
57.5(10.4)
56.1 (8.3)
Finger Tapping
Left Hand
59.1 (9.0)
54.C (9.7)
54.8(13.1)
Verbal Fluency
14.6 (4.5)
13.1 (4.2)
So (2.6)
a
b
Significantly different from no sleep apnea/hypopnea.
Significantly different from low sleep apnea/hypopnea.


found in these subjects are taken as minimally comparable
with subjects experiencing the classic sleep apnea
syndrome, some support for the anecdotal reports of in
tellectual deterioration in SAS (Cuilleminault et el.,
1978) is derived. However, both these conclusions stand
in need of replication belore they may be confidently
asserted.
The mechanism underlying the link between nocturnal
respiratory and cognitive variables is a fascinating
enigma. To cate, two explanations have been proposed as
proximal causes of the changes observed in sleep apnea
syndrome. The first, or "sleep fragmentation" hypothesis
(Carskadon et al., 1980) notes the large number of
arousals which may accompany nocturnal respiratory events
and suggests that the resultant sleep fragmentation may
lead to daytime impairments, particularly in the
elderly. While intuitively appealing, there are several
difficulties with this explanation. First, the demonstra
tion of reliable cognitive deficits after sleep depriva
tion hrs proved surprisingly difficult (Webb, 1975).
Changes in cognitive test scores such as those used in the
present study have not been reliably observed following
sleep deprivation. A second line of evidence inconsistent
with the sleep fragmentation hypothesis is the observed
lack of correlation between sleep structure and cognitive
variables in normal subjects (Berry anc Webb, In press).
Although this covariation has been observed in subjects


SLEEP APNEA ACTIVITY AND ITS CONCOMITANTS
IN A SUBCLINICAL POPULATION
By
DAVID THOMAS REED BERRY
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN
PARTIAL FULFILLMENT OF THE REQUIREMENTS
FOR THE DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
1985

ACKNOWLEDGMENTS
The author would like to thank his chairman, Dr. Wilse
B. Webb, for his moral and material support, as well as
guidance through the completion of the project. Also due
thanks are the other members of the supervisory committee,
Dr. A. J. Block, Dr. Russell M. Eauer, Dr. F. D. McGlynn,
ann Dr. Rudy Vuchinich, for their support, advice, and
contributions. Several people provided technical assistance
and are due a vote of thanks: Dsn Switzer, Alex Kasterton,
Karla Smith, Alton Howard, Albert Briggs, and Edna Pearson.
The author wishes to acknowledge the moral and
financial support of his parents, Mr. and Mrs. John G.
Berry, without which the project would never have been
undertaken. Also of considerable help were the
encouragement and advice of C. ti. Carswell, D. Haymaker, and
S. Brody, all of whom deserve thanks. Cynthia Zimmerman
provioeri expert typing assistance for the manuscript.
ii

TAELE OF CONTENTS
Page
ACKNOWLEDGMENTS ii
LIST OF TABLES v
AESTRACT vii
CHAPTER
ONE INTRODUCTION 1
Definitions and Terminology 2
History and Characteristics of Sleep Apnea
Syndrome 3
Incidence of Sleep Apnea Activity 5
Temporal Characteristics of Apneas 13
Variables Associated With Sleep Apnea
Syndrome and Activity 17
Deficits Associated With Sleep Apnea
Syndrome 23
Cardiopulmonary Deficits 26
Arousal Deficits (Hypersomnolence) 28
Oxygen-Deseturation and Cognitive Deficits..30
Measurement of Deficits Found in Sleep Apnea
Syndrome 34
Measurement of Cardiopulmonary and Health
Deficits 34
Measurement of Arousal Deficits
(Hypersomnolence) 35
Measurement of Hypoxia and Its Sequelae 38
Statement of the Problem 46
TWO METHOD 48
Subjects 48
Apparatus 49
Measures 50
Procedure 54
Statistical Procedures 55
THREE RESULTS 56
Demographics and Incidence of Nocturnal
Respiratory Disorder
iii
56

Nocturnal Respiratory and Health Variables 63
Nocturnal Respiratory and Sleep/Y.’ake Variables..65
EEC Data 65
Daytime Sleepiness Data 69
Subjective Sleep Assessment 70
Nocturnal Respiratory and Neuropsychological
Variables 74
FOUR DISCUSSION 86
Demographics and Incidence of Nocturnal
Respiratory Disorder 86
Nocturnal Respiratory and Health Variables 90
Nocturnal Respiratory and Sleep/Víake Variables. .92
EEG Data 92
Daytime Sleepiness, Sleep Questionnaire,
and Sleep Log Data 94
Nocturnal Respiratory and Neuropsychological
Variables 96
FIVE SUMMARY AMD CONCLUSIONS 103
APPENDIX
A SCORING PROTOCOL FOR RESPIRATORY VARIABLES 106
B RAW DATA TABLES 109
REFERENCES 123
BIOGRAPHICAL SKETCH 132
i v

LIST ÃœF TABLES
Table Page
2-1 Interrater agreement for respiratory events
of 46 snoring males 51
3-1 Demographic variables from 46 snoring males 56
3-2 Incidence of low and high levels of apnea/
hypopnea by age in 46 snoring males 58
3-3 Significant (p<.05) Pearson correlations
between nocturnal respiratory variables
and demographic variables in 46 snoring
males
3-4 Keans and standard deviations for selected
demographic and nocturnal respiratory
variables in 46 snoring males grouped by
level ol apnea/hypopnea 61
3-5 Means and standard deviations for health
related variables in 46 snoring males grouped
by level of apnea/hypopnea 65
3-6 Means and standard deviations of selected
sleep variables from 43 snoring males 66
3-7 Means and standard deviations of selected
sleep variables from 43 snoring males
grouped by level of apnea/hypopnea 68
3-8 Leans and standard deviations for daytime
sleepiness variables in 46 snoring males
grouped by level of apnea/hypopnea 69
3-9 Significant (p<.05) Pearson correlations
between nocturnal respiratory and sleep
questionnaire variables in 46 snoring males 71
3-10 Means and standard deviations of sleep
questionnaire variables from 46 snoring
males grouped by level of apnea/hypopnea 71
v

3-11 Significant (p<.05) Pearson correlations
between nocturnal respiratory and sleep log
variables in 46 snoring males 73
3-12 Means and standard deviations of sleep log
variables in 46 snoring males grouped by
level of apnea/hypopnea 73
3-13 Multivariate regression of demographic and
nocturnal respiratory variables on cognitive
scores in 46 snoring males 76
3-14 Significant (p<.05) partial correlations,
controlling for age, between nocturnal
respiratory and cognitive variables in
46 snoring males 79
3-15 Means and standard deviations of cognitive
scores in 46 snoring males grouped by
level of apnea/hypopnea 80
3-16 Multivariate regression of sleep and
nocturnal respiratory scores on cognitive
variables in 43 snoring males 83
B-1 Demographic and medical variables from 46
snoring males 109
B-2 Daytime sleepiness variables from 46 snoring
males 111
B-3 Sleep questionnaire variables from 46 snoring
males 113
B-4 Seven day sleep log means from 46 snoring
males 115
B-5 Neuropsychological variables from 46 snoring
males 117
B-6 Respiratory variables from 46 snoring males....119
B-7 Sleep variables from 46 snoring males 121
vi

Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
SLEEP APNEA ACTIVITY AND ITS CONCOMITANTS
IN A SUBCLINICAL POPULATION
By
DAVID THOMAS REED BERRY
August 1985
Chairman: Wilse B. Webb
Major Department: Clinical Psychology
Sleep Apnea Syndrome (SAS) is a nocturnal respiratory
disorder with serious consequences. Patients with SAS
experience apneas (pauses in breathing of 10 or more
seconds) and hypopneas (declines in the amplitude of
breathing accompanied by oxygen desaturation) while
asleep. Recently, it has become apparent that apneas and
hypopneas may occur in otherwise apparently normal
subjects. A sample selected to display a wide range of
sleep disordered breathing (heavy snoring males in good
health and over 30 years of age) received detailed
measurement of their nocturnal respiration, sleep,
cognitive skills, and health in an attempt to clarify the
impact of apneas and hypopneas in a subclinical population.
Forty-six heavy snoring males in good health with a
mean age of 50 years and a mean weight of 190 pounds
comprised the sample. During an experimental evening they
vii

received testing and filled out various questionnaires
before sleeping overnight with their sleep and breathing
continuously recorded. Sixty-two per cent of these subjects
experienced at least one event, while 13# had more than 5
events per hour of sleep. Most events occurred in light
slow wave or REM sleep and were linked to obesity. Subjects
with high levels of apnea/hypopnea (more than 5 per hour)
experienced significantly exacerbated oxygen desaturation
relative to the remaining subjects. Oxygen desaturation was
linked to higher blood pressure readings. A typical "first
night" effect on the sleep of the subjects in the lab was
noted, resulting in a lighter sleep than usual and obscuring
possible relationships between sleep and respiratory events,
which were not observed. Deteriorating overnight
respiratory indices were associated with increased
sleepiness and napping. High levels of apnea/hypopnea were
associated with declining scores on tests measuring non¬
verbal intelligence, verbal and non-verbal memory,
expressive verbal fluency, and cognitive flexibility. It is
speculated that hypoxia induced changes in cerebral
cholinergic synthetic pathways underlie these changes.
It is concluded that a significant subgroup of heavy
snoring males who experience multiple apnea/hypopneas are at
increased risk of significant oxygen desaturation, daytime
sleepiness, and cognitive changes. This group may fall on a
continuum with sleep apnea syndrome patients.
viii

CHAPTER ONE
INTRODUCTION
Sleep apneas are respiratory pauses during sleep.
According to the Association of Sleep Disorders Centers
Nosology of Sleep Disorders (ASDC, 1979). a sleep apnea
syndrome (SAS) is a potentially lethal condition charac¬
terized by multiple apneas, excessive daytime sleepiness,
alterations of consciousness, and cardiopulmonary
complications. The apnea episodes are thought to be the
causal factor behind the other deficits associated with
SAS (Guilleminault et al., 1978).
Polysomnographic recordings in various populations
have revealed the occurrence of sleep apneas in the
general population, and particularly in aging popula¬
tions. Estimates of the prevalence of at least a few
sleep apneas in older subjects have ranged as high as 75/5
(Carskadon and Dement, 1981b). Carskadon et al. (1980)
have suggested that sleep apneas in older subjects may be
implicated in insomnia, cardiac ailments, and even
senility.
However, certain reports have documented the presence
of multiple sleep apnea episodes in subjects without
apparent pathology (Crr et al., 1979). This evidence
obscures the role of sleep apneas in the other pathologies
1

2
noted in SAS patients. The link between apneas and patho¬
logy may be more complex than was originally thought,
given its apparently benign occurrence in otherwise
asymptomatic subjects.
The present study will examine the correlates of
sleep apnea episodes in a population selected to display a
wide range of such events. The relationship between sleep
apnea indices and sleep characteristics, age, weight and a
variety of deficit measurements will be analyzed.
A review of the current data available on various
aspects of sleep apneas and sleep apnea syndromes will be
presented as background for the study. The diagnosis,
history, incidence, and characteristics of sleep apneas
and sleep apnea syndrome will be reviewed, followed by a
discussion of variables associated with sleep apneas.
Lastly, the deficits commonly found in sleep apnea syn¬
drome will be detailed, along viith measurements thought to
be sensitive to these deficits.
Definitions and Terminology
Sleep apneas have been defined as cessations of air¬
flow at the mouth and nose which last for 10 or more
seconds (Guilleminault et al., 1976). In a sleep apnea
syndrome (SAS), multiple sleep apneas are accompanied by
snoring, excessive daytime sleepiness, cardiopulmonary
complications, and altered states of consciousness. The
ease of quantification of sleep apneas has led to their
use as an important tool in diagnosing SAS. Guilleminault

3
et al. (1976) first suggested that the occurrence of 30 or
more apneas in 7 or more hours of sleep was diagnostic of
SAS, while Guilleminault et al. (1978) proposed the use of
an apnea index (AI; number of apneas/number of hours
sleep) in excess of 5 as a cut-off score to diagnose the
syndrome. Later, Carskadon et al. (1980) combined apneas
and hypopneas (a reduction in airflow of 50% or more at
the mouth and nose) to form an apnea + hypopnea index
(AHI) with 5 utilized as a diagnostic level of sleep apnea
syndrome. As can be appreciated from the preceding dis¬
cussion, the diagnostic criteria for SAS have undergone a
rapid evolution, with remarkably little validation work
published.
History and Characteristics of Sleep Apnea Syndrome
Sleep Apnea Syndrome falls within the larger category
of Sleep Disordered Breathing (Block, 1980). Respiratory
disorders within this category involve changes in
breathing patterns during sleep. Syndromes described
within this category include Pickwickian Syndrome; Sleep
Apnea Syndrome; Chronic Obstructive Lung Disease; and
syndromes of Primary and Secondary Alveolar
Hypoventilation (Block, 1980; Guilleminault et al.,
1976).
The first disorder described in this group was the
Pickwickian Syndrome. These patients were initially
identified by a chronic daytime hypersomnolence, obesity,
Cheyne-Stokes breathing, hypercapnia and hypoxemia. This

4
syndrome typically results in complications such as right
sided congestive heart failure, pulmonary hypertension,
and peripheral edema (Block, 1980). An exacerbation of
disordered breathing (apnea) as well as hypercapnia and
hypoxemia (often severe) occurs with sleep onset. Treat¬
ment consists of oxygen therapy, respiratory stimulants,
and maintenance of wakefulness.
The discovery of a sleep induced exacerbation of the
symptoms accompanying Pickwickian Syndrome spurred a
systematic investigation of sleep respiratory events
(Gastaut et al., 1965). The new emphasis on noctural
respiration led to a series of investigations which docu¬
mented sleep induced respiratory dysrhythmias in non-obese
subjects (Lugarasi et al., 1968). A new group of sleep
disordered breathers designated as Sleep Apnea Syndrome
emerged, with several features which distinguished them
from Pickwickians. Among these differences were weight
(i.e., SAS patients were not necessarily obese), daytime
respiratory control (i.e., SAS patients were typically not
hypercapnic during waking as were Pickwickians), sex
(i.e., while 40of Pickwickians were female, very few
premenapausal females have been identified with SAS), and
age (i.e, SAS patients are typically younger than
Pickwickians)(Block, 1980).
Two large samples of SAS patients have been reported
in the literature, allowing characterization of typical
SAS patients. Guilleminault et al. (1978) reported on a

5
series of 50 patients diagnosed with SAS over a period of
six years. Age ranged from 28-62 with a mean age of
45.5. The sample consisted of 48 males and 2 females.
Thirty-nine were referred on the basis of excessive day¬
time sleepiness, while 7 were referred for loud snoring
and abnormal movements reported by their spouses. Sixty
percent of the patients were more than 15& above ideal
weight. Other symptoms included hypnagogic halluci¬
nations, automatic behaviors, intellectual deterioration,
personality changes, impotence, morning headaches, and
hypertension. Weitzman, Kahn and Poliak (1980) report
data on a group of 10 SAS patients seen in their clinic.
These subjects displayed the following characteristics:
age between 38 and 47, obesity, male sex, nicotine depen¬
dence, hypertension, complaints of excessive daytime
sleepiness, and the presence of serious cardiac arrhyth¬
mias. (This group was selected from a larger sample of 38
patients.) Thus, pathological symptomatology associated
with a diagnosed SAS include excessive daytime somnolence,
cardiopulmonary complications and cognitive/intellectual
changes; a "typical" SAS patient is a middle aged male
who snores and is obese.
Incidence of Sleep Apnea Activity
The incidence of sleep apnea activity and sleep apnea
syndrome have been frequently confused. While sleep
apneas have been an important diagnostic tool for identi¬
fying sleep apnea syndrome, they have not been

6
definitively demonstrated to be the necessary and suffi¬
cient cause of the syndrome. Thus, for the purposes of
the following review, the focus will be on occurrence and
level of sleep apnea activity in both sleep apnea syndrome
subjects and other populations. The full range of sleep
apnea activity will be reflected by noting the presence of
sleep apnea activity, and an indication of the quantita¬
tive level of the activity will be presented in the apnea
index (AI) or the apnea + hypopnea index (AHI). Recall
that an AI of 5 has been frequently utilized as a cut-off
score for identifying SAS patients.
A review of the literature suggests that at least
three sampling strategies have been used for studying the
incidence of sleep apnea activity. Subjects have been
selected on the basis of one of the following three
criteria: 1) absence of sleep or respiratory complaints,
2) presence of sleep or respiratory complaints, or 3)
sampling without regard to these variables.
Because diagnosis of SAS is often dependent on number
of sleep apneas (or level of SAA), most researchers report
number of apneas observed, although some report only on
number of subjects who exceed one of the criteria for
SAS. A review of studies with samples of 20 or greater
will be presented with careful note of the sampling
strategy utilized, as well as the dependent measure of
apnea.

7
One sampling strategy which has been pursued is to
exclude subjects who complain of sleep or respiratory
difficulties. Screening for sleep complaints varies
widely. Guilleminault et al. (1978) studied a sample of
20 "normals" (presumably noncomplaining subjects). These
subjects ranged in age from 40-60 years. Males showed a
mean number of apneas of 7 (range 1-25), while females
showed a mean number of apneas of 2 (range 0-5). Thus,
none of these noncomplaining subjects presented a clinical
syndrome of sleep apnea, although sleep apneas did occur
apparently without pathology. Block et al. (1979) studied
49 subjects (M = 30, F = 19). Males in the sample had a
mean age of 38, while females had a mean age of 29. Any
potential subjects who complained of breathing difficul¬
ties or sleep disturbances were excluded from the study.
Twelve males (40$) had episodes of sleep apnea while only
3 females (15$) suffered as well. For those subjects,
males had a mean of 4.2 episodes, while females had a mean
of 3. Within this sample, the number of apneas was not
significantly correlated with increasing age, although
occurrence of oxygen desaturation (a concomittant of
disordered breathing) was positively correlated with
age. Additionally, although the mean number of sleep
apneas was not significantly different between sexes (M =
4.2, F = 3), the more sensitive measure of apneas with
desaturation was (M =3, F = 0). Webb (1974) studied 2C
males with a mean age of 44. Subjects were excluded from

8
the study if they reported any "serious" sleep com¬
plaint. Two of the subjects reported consuming hypnotics
for sleep induction, indicating that the criterion used in
this study was less stringent than that of Block et si.
(1979) report. Mine of the subjects (45$) suffered at
least one episode of sleep apnea with a mean number of
episodes of 2.5. All of the subjects with episodes of
sleep apnea were older than the mean age of the sample,
leading Webb to conclude that apnea is related to age.
Bixler et al. (1982) constructed an age stratified patient
representative sample of 100 subjects (M =41, F = 59).
In perhaps the most rigorous screening for sleep com¬
plaints in this group of studies, Bixler et al. eliminated
any subject with a sleep complaint or chronic medication
usage. For the entire sample, 6 males (14.6$) and 5
females (10.2ft) had at least one episode of sleep apnea.
For subjects under 30, 3.3ft; for those between 30 and 50,
12.6ft; and those 50-74, 19.4ft suffered episodes of sleep
apnea. For the subsample over 60, 25ft had sleep apnea
activity with 8.5 episodes for males and 10.5 episodes for
females. In this sample, there was a nonsignificant trend
toward more apnea activity in males and a significant
positive correlation between age and sleep apnea.
Block et al. (1980) report data on 20 post menapausal
women with a mean age of 59. Eight women (40ft) had epi¬
sodes of sleep apnea, with a mean number of episodes of

9
5. There was a significant correlation between apnea and
age.
For noncomplaining subjects then, males suffer a
higher incidence of apnea activity in virtually every
study which compares sex effects. For males, incidences
range widely from a low of 14.6% (Bixler et al., 1980) to
a high of 45% (Webb, 1974); while females range from a
low of 10.2% (Bixler et al., 1982) to a high of 40% (Block
et al., 1980). These differences are probably due at
least in part to age differences in male samples, as well
as age differences in females (only as they reflect mena-
pausal status). Reported number of apneas range from 2.5
(VJebb, 1974) to 8.5 (Bixler et al., 1982) in males and
from 2 (Guilleminault et al., 1978) to 10.5 (Bixler et
al., 1982) in females. Many studies report a positive
correlation between number of apneas and increasing age.
A second sampling technique which has been utilized
has been studying subjects with sleep and breathing com¬
plaints. Kales et al. (1982) studied 200 subjects with
primary complaints of insomnia. There were 82 males and
118 females in the sample with a mean age of 42.3. Sleep
apnea activity was noted in 10.5% of this group (M =
13-4%, F = 8.5%) with a mean number of 11.2 episodes for
males and 6.6 for females. Subjects with sleep apnea
activity were significantly older (46.8) than those with¬
out (40.6). Ancoli-Israel et al. (1981) recruited 24
subjects whose answers to questionnaires provoked

10
suspicion of insomnia or nighttime breathing and muscular
events. Eleven males (x = 72.5) and 13 females (x = 68.5)
were studied. All of these subjects, save one male and
one female, suffered at least one episode of sleep
apnea. For those subjects with at least one episode of
sleep apnea, males had a mean of 53 episodes of apnea,
while females had a mean of 32 episodes. In contrast to
the Kales et al. study, 6 males had a high level of sleep
apnea (AI>5), while 3 females met this criterion. A rela¬
tionship between age and apnea was not reported in this
sample, possibly because of restricted range. These two
sxudies of sleep disturbed subjects provide conflicting
data. That is, the insomniac alone sample did not demon¬
strate an apnea incidence which was substantially
different from a noncomplaining sample (10.5%), while the
sample recruited for sleep induction, respiratory, and
muscular differences showed very discrepant apnea inci¬
dences (nearly 100;*). One potential explanation involves
age, as the Ancoli-Israel sample is substantially older
than Kales', although the aging samples have not reported
this level of apnea before. An alternative explanation is
that sampling insomniacs does not generate a high number
of apneas, whereas aging sleep and respiratory disturbed
subjects show a very high incidence of apnea. It would
seem then that the relevant "complaint" would involve
those reported by Ancoli-Israel et al. which are in
addition to those noted by Kales et al. These are (as

11
nearly as can be Judged) questions surrounding
respiration.
The final category of incidence studies involves
those which do not consider sleep complaints in their
sampling. As in the other studies, these vary in their
generalizability. Carskadon and Dement (1981b) recruited
40 aging subjects, including 18 males (3c = 72.7) and 22
females (x = 74). Their exclusion criteria are somewhat
problematic, involving excluding subjects who "spon¬
taneously complain of sleep problems." Thus, it is
unclear what this is a representative sample of, specifi¬
cally whether the distribution of those not spontaneously
complaining of sleep disturbances may be different from
those who actually do not suffer from sleep distur¬
bances. These considerations aside, Carskadon and Dement
report that 16 males (88%) and 16 females (72%) demon¬
strated at least one sleep apnea episode, while 8 males
(44%) and 7 females (31.8%) have high levels of sleep
apnea (AI>5). Males with apnea showed a mean number of
51, while females showed a mean number of 34. Sex
differences appear to have fallen out at this age, while
the restricted range of this extreme age sample does not
lend itself to a correlational analysis of age versus
apnea. In another study of limited generalizability,
Kreiss et al. (1982) randomly selected inpatients on a
Veterans Administration medical ward. These patients, all
presumably male, numbered 26. Further data reported were

12
sketchy, but 7 (27%) met the sleep apnea syndrome
criterion of >30 episodes of sleep apnea. The fact that
these subjects were patients on a medical ward obviously
restricts generalizability of these findings. Finally, in
the only truly random sampling reported, Ancoli-Israel et
al. (1982) present initial data on randomly selected sub¬
jects. Forty subjects (sex not reported) with a mean age
of 71.5 were studied. Thirteen (52.5%) met the sleep
apnea syndrome criterion of 30 or more apneas. The sleep
apnea subgroup was older, 74.2, than the full sample.
The data collected from relatively random sampling
are primarily based on aging populations (>60). The data
from these studies suggest that between 32.5# (Ancoli-
Israel et al., 1981) and 37# (Carskadon and Dement, 1981b)
of aging subjects suffer high levels of sleep apnea
(AI>5). An even larger percentage, perhaps greater than
75#, suffer at least some episodes of sleep apnea. A
major deficiency in these randomly sampled studies is a
lack of data on younger subjects.
Thus data collected under different sampling proce¬
dures yield variable incidence rates of sleep apnea and
AI>S. Subjects without sleep or respiratory complaints
have a very low incidence (perhaps 0) of AI>5, between 14#
and 45# of males surveyed had at least some apnea, while
females ranged between 10# and 40# experiencing some
apnea. Sampling subjects with sleep or respiratory com¬
plaints suggested that subjects suffering respiratory

13
difficulties had very high incidences of apnea (almost
100?; for both males and females) with increase of inci¬
dence of AI>5 (55$ for males, 23$ for females). Sampling
without regard to sleep or respiratory difficulties is
largely confined to aging populations. Incidences of
apnea in these aging populations would seem to be rather
high (88$ males, 72$ females) while incidence of AI>5 seem
to be in the range of 32$-37$.
It is clear that different populations (as defined by
sampling strategies) suffer different incidences of sleep
apneas and AI>5. The clinician dealing with sleep dis¬
orders would be wise to consider the acturial base rates
of apnea and AI>5 provided by these data. It is apparent
that sex, age, and respiratory status are significantly
related to the incidence rates of SAA and SAS. These
factors must be considered in any study of apnea and asso¬
ciated deficits.
Temporal Characteristics of Apneas
The temporal characteristics of sleep apneas repre¬
sent a potentially important variable. Above it was shown
that a wide variability in number of sleep apneas occurs,
and the arbitrary nature of the division between SAA and
SAS subjects was noted. Therefore, it is of interest to
determine whether the duration of sleep apneas provides a
more reliable distinction between the two groups. Another
variable of interest involves the placement of apneas
within sleep, as Guilleminault et al. (1976) noted that

14
normals might experience apneas during REM without
pathological consequence. Therefore, data available on
these two issues will be reviewed, including only studies
with a sample size great than 10 (with one exception).
Reports on normal subjects below 30 years of age are
somewhat rare, although some do exist. Bixler et al.
(1982) included a group aged 18-29 years (M = 13, F =
17). One subject in this group had sleep apneas (12) with
a mean duration of 15 seconds. Bixler et al. noted that
most apneas occurred in Stages 1 and REM sleep. Part of
the sample of Block et al. (1979), the females, had a mean
age of 29. Three of these subjects showed apneas with a
mean duration of 14 seconds. Again, a disproportionate
number of apneas occurred in Stages 1 and REM. Thus in
normal subjects under 30, apneas are somewhat rare, with a
mean duration of approximately 15 seconds when they do
occur. Most apneas occur in light and REM sleep.
More data are available on normal subjects between 30
and 60. Guilleminault et al. (1978) do not report on
duration of apneas in their sample aged 40-60, but they
note that apneas occurred only in Stages 1 and REM.
Bixler et al. (1962) reported that, in their 30-44 year
old group, 5 subjects experienced sleep apnea with a mean
duration of 13.1 seconds. A disproportionate number of
apneas occurred in Stages 1 and REM. Block et al. (1979)
studied 30 males with a mean age of 38, with 12
experiencing sleep apneas. These apneas had a mean

15
duration of 20 seconds and occurred most frequently
(30/60) in Stages 1 and REM. Eight of 20 aging females,
in the report by Block et al. (1980), had sleep apneas
with a mean duration of 19 seconds. Sixty-two percent of
apnea episodes occurred in Stages 1, 2 or REM. Thus in
middle aged normal subjects, incidence of sleep apnea
appears to rise, with a mean duration of between 13 and 20
seconds, perhaps slightly longer than in younger groups.
Again, apneas occurred mostly in light (Stages 1 and 2)
and REM sleep.
Duration data on aging normals are limited.
Carskadon and Dement's (1981b) study of aging normals (H =
72, F = 74) does not report duration of apnea events.
Similary, Ancoli-Israel et al. (1982) do not provide this
information on their aged normals (x = 71.5). Bixler et
al. (1982) noted a high incidence of apneas in their sub¬
jects over 60, but failed to report a mean duration for
this group. Given the sketchy details reported for this
group, it is difficult to draw any conclusions about
changes in temporal duration or placement of apneas in
aging normals.
To summarize the temporal data in normals, apneas
appear somewhat rarely in normals below 30, with a mean
duration of 15 seconds. Incidence of apneas increase
between 30 and 60, with a wider variation in mean dura¬
tions reported (13-20 seconds). Apneas occur most
frequently in light and REM sleep in normals of all ages.

16
Temporal data on subjects diagnosed with sleep apnea
syndrome (>30 or AI>5) primarily include patients between
the ages of 30 and 60. The reason for the lack of older
SAS subjects is unclear, particularly in light of data
reported above, demonstrating very high percentages of SAS
in certain groups of aging subjects.
Garay et al. (1981) studied 13 sleep apnea syndrome
subjects. They divided these 12 males and 1 female into
two groups, based on their daytime resting CO2 levels.
Subjects who were eucapnic during the day had a mean apnea
duration of 16 seconds, while those who were hypercapnic
had a mean duration of 17 seconds. Ho data on distribu¬
tion relative to sleep stages were reported.
Guilleminault et al. (1978) studied 50 predominantly male
(x age 48) sleep apnea syndrome patients. Apneas in this
population averaged 22 seconds in duration. Apneas were
longest in REM, and few occurred in Stages 3 or 4 sleep.
Fujita et al. (1981) studied 12 male SAS patients with a
mean age of 43* Apnea episodes averaged 23 seconds in
duration. V/eitzman et al. (1980) studied 10 SAS males
whose age ranged from 38-57. The mean duration of apneas
in this group was 30 seconds. All apneas occurred in
light or REM sleep. Orr et al. (1979) studied 4 hyper-
somnolent males diagnosed with SAS. A control group of 4
nonhypersomnolent subjects was matched on number of ob¬
structions per minute. Mean age was 57 for asymptomatic
and 42 for symptomatic subjects. Asymptomatic subjects

17
had apneas with a mean duration of 25.9 seconds, while
symptomatic subjects had a mean duration of 20.8 seconds.
Several characteristics emerge from the group of
reports on SAS subjects. There is wide variability in
duration of apneas, with reported means ranging from 16-30
seconds. The data available on placement of apneas in
sleep suggest that in SAS subjects most apneas occur in
light and REM sleep.
Comparison of data on normal and SAS subjects reveals
substantial overlap between mean duration of apneas in SAS
subjects (13-20 seconds) and SAS subjects (16-30
seconds). Particularly interesting was the Orr et al.
(1979) study which demonstrated a lower mean duration of
apneas for hypersomnolent subjects than for nonhypersomno-
lent subjects. Data on temporal placement of apneas show
most events occurring in light or REM sleep for both
normal and SAS subjects. Taken together, these data
suggest that substantial overlap exists on temporal
f
variables between normal and SAS subjects indicating a
limited utility for diagnostic purposes.
Variables Associated With Sleep Apnea
Syndrome anc Activity
It was noted above that several variables have been
suggested as associated with increased incidences of sleep
apneas and SAS. These variables represent potential
"markers" of subpopulations which might experience higher
incidences of sleep apnea activity. These primary

18
associated variables include sex, age, snoring and
obesity. The data available on these variables and their
relationships to sleep apnea activity will be reviewed,
followed by a brief discussion of their relevance to
experimental methodology.
Regarding sex variables, as previously noted,
Guilleminault et al. (1978) found 96$ of their SAS
patients to be males. In the Block et al. (1979) study of
nonsleep disturbed subjects, 40$ of the males in the
sample had at least one episode of apnea, while only 15$
of the females so suffered. Bixler et al. (1982) noted a
trend toward more sleep apnea activity in males. Block et
al. (1980) studied 20 post menapausal females and found
apnea activity in 40$ of them, leading them to suggest
that some factor associated with menapause (e.g., pro¬
gesterone) might somehow "protect" premenapausal females
from apneas, and that this factor along with its benign
influence was lost with menapause. This explanation pre¬
dicted the low level of apnea activity in younger females,
as well as the heightened apnea activity of aging
females. While the data of Carsakdon and Dement (1981b)
show aged females to be close to aged males in percentage
exhibiting sleep apnea activity, a report by Smallwood et
al. (1983) found no evidence of sleep apnea activity in
their 6 elderly (post menapausal) females, while elderly
males showed significant levels of apnea activity. The
limitations of the sample size in the Smallwood et al.

19
report might explain this discrepancy as representing a
sampling bias. Taken together, this evidence strongly
supports the notion that in young and middle aged popula¬
tions, sleep apnea is a male phenomenon. Additionally,
evidence points toward a closing of the gap between
females and age matched males as the former group passes
menapause.
A second factor which has been implicated in sleep
apnea activity is age. Webb (1974) found that his sub¬
jects exhibiting sleep apnea activity were significantly
older than his full sample. This was also the case for
the subjects of Kales et al. (1982) and Ancoli-Israel et
al. (1982). Smallwood et al. (1983) found that sleep
apnea was an age dependent phenomenon with male subjects
over 50 exhibiting significantly more apneas than those
under 30. Other evidence includes that of Bixler et al.
(1982) who found a positive correlation between age and
apnea in their full sample, as did Block et al. (1980) in
their post menapausal females. It seems clear from these
data then that increasing age is a factor predisposing for
sleep apnea activity.
Snoring has been closely associated with sleep apnea
(Lugaresi et al., 1982). This may be due to a causal
relationship or a single common pathway. Lugaresi et al.
propose the latter explanation. Specifically, they
suggest that both snoring and SAS are due to a sleep
induced stenosis of the upper airway. With this premise

20
in mind, they conducted a survey of 1000 subjects on the
incidence of snoring. Chronic snorers included 31$ of the
males and 19$ of the females. In an elaboration of this
study, Lugaresi et al. (1980) questioned 5713 indivi¬
duals. In this sample, 24% of the males and 13% of the
females were chronic snorers. Snoring increased with age,
and by 60 years of age 60% of the males and 40% of the
females snored. Of note was the observation that hyper¬
tension occurred more frequently in snorers than non-
snorers, suggesting that the former was more at risk for
cardiovascular complications.
Virtually every study of SAS patients notes the pre¬
sence of snoring in all these subjects (Sullivan and Issa,
1980; Block, 1980; Coverdale et al., 1980). This is
true for both predominantly obstructive, and predominately
central SAS subjects (Guilleminault et al., 1978). How¬
ever, the presence of heavy snoring in asymptomatic
(nonhypersomnolent) apneas is also well documented (Orr et
al., 1979; Fisher et al., 1978; Berry and Block, 1983),
as well as in otherwise normal subjects. These data may
be taken as evidence for snoring as a necessary, but not
sufficient, condition for the occurrence of sleep apnea.
The role of body weight in SAS is less clear.
Recall that Pickwickian patients are invariably obese, as
well as hypercapnic. Guilleminault et al. (1978) stress
that SAS may occur in nonobese as well as obese sub¬
jects. However, Block (1980) states that he has found SAS

21
subjects to be uniformly obese. A review of the
literature is in order to clarify this point. First
normal, then obese, then SAS groups will be reviewed.
Bixler et al. (1982) found that their subgroup of
subjects with sleep apnea activity weighed significantly
more than the remainder of the sample. Kales et al.
(1982) found that within their sample of 200 insomniacs
and 100 normals, a significant positive correlation
emerged between a weight:height ratio and apnea acti¬
vity. Block et al. (1979) found a positive correlation
between apneas, oxygen desaturation, and weight within
their normal males, while females did not demonstrate this
relationship. Using an older female sample, Block et al.
(1980) found a positive correlation between weight and
oxygen desaturation, but not apnea. Thus in normal sub¬
jects, there appears to be a positive relationship between
weight and apnea activity, although this may be restricted
to males.
A research approach which bears on this question
involves studying obese subjects alone. Two studies on
successive referrals for gastric bypass surgery were
noted. Sicklesteel et al. (1981) found that of 19 suc¬
cessive referrals, all 14 males exhibited sleep apneas,
while only 5 of the females showed evidence of oxygen
desaturation alone. Harmon et al. (1981) studied gastric
bypass patients and found 5 of 6 males to suffer apnea, as

22
opposed to none of the females. This evidence provides
further evidence for a role of increasing weight in apnea.
The data on symptomatic SAS subjects are more com¬
plex. Several studies (Sullivan and Issa, 1980; Iveitzman
et al., 198C; Garay et al., 1981; Zwillich et al., 1982)
report that all their SAS subjects were obese, although
definitions of obesity either vary widely or remain un¬
specified. In contrast, Guilleminault et al. (1978)
reported that only 60$ of their SAS sample was greater
than 15$ above ideal weight. Coverdale et al. (1980)
indicate that only two of their subjects proved to be
obese, according to their definition (>125$ of ideal
weight). These reports suggest that most, although not
all, SAS subjects are obese.
Several investigators have sought to compare weight
in symptomatic (somnolent) and asymptomatic subjects with
enough sleep apnea episodes to be diagnosed as SAS. Orr
et al. (1979) compared 4 symptomatic SAS subjects with
matched asymptomatic subjects. In this sample, symptoma¬
tic SAS subjects were heavier than asymptomatic sub¬
jects. Examination of the data of Berry and Block (1983)
indicates that their symptomatic subjects were heavier
than their asymptomatic subjects (118 kg vs. 91 kg).
Standing in contrast to these data is a study reported by
Fisher et al. (1978), which included 19 subjects referred
for SAS evaluation. Based on number of apneas per night,
Fisher et al. divided their subjects into 3 groups. Group

23
I had >100 apneas per night, Group II had >30 but <100
apneas per night, while Group III had less than 20 apneas
per night. Thus Groups I and II were diagnosable as SAS
(>30), while Group III was not. Obesity was present in
Groups I and II, but in only 3 subjects in Group III.
However, these differences were not statistically signifi¬
cant.
Taken together, this evidence suggests a positive
relationship between weight and apnea in the general popu¬
lation. However, SAS patients are not invariably obese.
Further, weight loss fails to result in a decrease in
apneas in many SAS patients. Thus the relationship
between obesity and full blown SAS is less than clear.
The preceding pages have investigated the currently
understood role of several variables (sex, age, snoring
and weight) in sleep apnea activity. Presently available
data suggest that in young and middle aged subjects, apnea
is a predominantly male phenomenon, although post mena-
pausal females may close the gap. Increasing age is re¬
lated to sleep apnea activity. Snoring, which occurs in a
significant percentage of the general population, is
thought to be a necessary but not sufficient condition for
SAS. Weight was shown to have a positive relationship to
sleep apnea activity in the general population, although
its presence or role in SAS is equivocable.
These conclusions have important implications for
experimental designs directed at understanding sleep apnea

24
activity and its correlates in a subclinical population.
It is clear from the earlier review of apnea incidence
that sampling from the general population would result in
a relatively low population of individuals with sleep
apnea, an inefficent approach from the perspective of the
limited resources available for research. Thus selection
variables must be considered with one eye on efficiency
and the other on protecting the generalizability of
results. Selection criteria should also aim toward eluci¬
dating the relative roles and possible interactions of
these variables in sleep apnea. Therefore, selection
variables should have as straight-forward a relationship
as possible with sleep apnea, seeking to answer more ques¬
tions than they raise.
With these caveats in mind, the variables available
as selection criterion will be reviewed and conclusions
drawn about their relative merits. The first variable
considered, sex, has been relatively well investigated.
It is clear that sleep apnea activity is primarily a male
circumstance. Therefore, the selection of female subjects
is contraindicated from the standpoint of efficiency. Age
has been demonstrated to exhibit a positive relationship
with sleep apnea, indicating selection of older sub¬
jects. Snoring is another variable with a close relation¬
ship to sleep apnea. The apparently common mechanism of
snoring and sleep apnea (stenosis of the upper airway)
indicates that possibly very few subjects without snoring

25
suffer sleep apneas. A significant proportion of snorers
might be expected to suffer apneas, while few nonsnorers
might have apnea. The last variable, weight, while having
a positive correlation with apnea activity in the general
population, is of uncertain pertinence to SAS. Addi¬
tionally, changes in weight may occur without changes in
apnea activity. Therefore, this variable has a less than
explicit connection to apnea activity and violates the
second principle adopted to consider these variables.
This review of available data suggests that sampling
aging, snoring, males would generate a high proportion of
sleep apnea activity. External validity for this idea is
drawn from its similarity to the "typical" SAS patient
described earlier. Extant data suggest that it may also
prove useful to utilize weight measures as a covariate, in
an attempt to control its apparent influence on apnea
activity.
Deficits Associated With Sleep Apnea Syndrome
Sleep Apnea Syndrome is thought to result in several
deficits. These include cardiopulmonary complications,
hypersomnolence, hypoxia and intellectual changes. Some
researchers have suggested that these deficits may be
present, in an attenuated form, in subclinical apnea
(Ancoli-Israel et al., 1981; Carskadon and Dement,
1981b). A discussion of the evidence for the presence of
these deficits will be presented, followed by proposals
for appropriate measures of these variables.

26
Cardiopulmonary Deficits
Cardiopulmonary complications have been reported as
concomittants of SAS (Guilleminault et al., 1978).
Coccagna et al. (1972) reported high incidences of hyper¬
tension and congestive heart failure in their SAS sub¬
jects. Schroeder et al. (1978) studied 22 SAS patients
with a mean age of 47. Six subjects had waking systemic
hypertension, and several had waking cardiac abnorma¬
lities. During sleep studies, 20 patients developed
significant rises in systemic arterial pressure which
cycled with episodes of apnea, while 21 patients developed
pulmonary arterial hypertension. Tracheostomy reversed
all these abnormalities, strongly implicating the apneas
in their etiology. Tilkian et al. (1978) studied 25 male
SAS subjects with a mean age of 44. Twenty-four of the 25
showed sleeping sinus arrhythmias, while 9 developed more
serious symptoms such as asystole. Seventeen of these
patients received tracheostomy, which abolished all of
these abnormalities. Experimental occlusion of the
tracheostomy during sleep in 6 of these patients resulted
in a return of the cardiopulmonary symptoms. Zwillich et
al. (1982) studied six consecutive male SAS patients. In
all patients, bradycardia accompanied any apnea with
significant desaturation. Additionally, a significant
correlation was noted between degree of desaturation and
severity of bradycardia. Bradycardia was abolished or
attentuated during administration of 02 enriched air.

27
Zwillich et al. proposed that bradycardia during apnea
results from increased vagal tone mediated by carotid body
chemoreceptors. Fujita et al. (1981) reported on 12 male
SAS patients. All reportedly suffered cardiac arrhythmias
during sleep. The body of data suggests that SAS patients
frequently suffer from cardiac abnormalities such as
hypertension, arrhythmias, and bradycardia.
A second body of data exists which indicates that SAS
patients do not invariably have cardiopulmonary compli¬
cations. Coverdale et al. (1980) studied 14 patients, of
whom 8 were diagnosed with SAS. Only 2 of these patients
had evidence of corpulmonále, while 2 more had systemic
hypertension. None of these patients demonstrated severe
brady or tachycardia during sleep. In Orr et al. (1979),
8 patients, all of whom were technically diagnosable as
having SAS, all symptomatic (somnolent) patients had evi¬
dence of right sided heart failure as well as frequent
arrhythmias during apneas. Asymptomatic subjects had none
of these symptoms. Kreiss et al. (1982) studied 26
Veterans Administration ward patients and found 7 to have
a SAS. The SAS patients had significantly more signs of
congestive heart failure, but not hypertension or
angina. Ancoli-Israel et al. (1981) sampled 24 elderly
subjects with complaints of sleeping respiratory Dis¬
orders. Nine of these subjects were diagnosable as SAS,
but this group did not have significantly more heart

28
disease or hypertension. Thus a group of SAS patients
without evidence of heart disease exists.
Taken together, these data indicate that cardiac
complications frequently, but not inevitably, accompany a
sleep apnea syndrome. That a large number of apneas may
be present without a significant increase in cardio¬
pulmonary complications is demonstrated by the Orr et al.
and Ancoli-Israel et al. studies. As the relationship
between these complications and a full blown SAS is
unclear, the possibility of cardiopulmonary sequelae from
a subclinical level of apneas is even more uncertain.
However, an assessment of cardiopulmonary status is
indicated by the frequent association of cardiopulmonary
complications with sleep apnea syndrome.
Arousal Deficits (Hypersomnolence)
Excessive Daytime Sleepiness (EDS), or hypersomno¬
lence, has been described as the single most common result
of SAS (Dement et al., 1978). Indeed, patients who meet
the criterion of number of apnea episodes and exhibit EDS
are described as symptomatic, while those with number of
apneas alone are labeled asymptomatic (Orr et al.,
1979). As an aside, this distinction illustrates again a
central issue of SAS. If patients exhibit the number of
apneas qualifying them for SAS, but no other pathology,
are they suffering from a pathological process? Dement et
al. (1978) believe that they are. They suggest that most,
if not all, patients with a clinically elevated number of

29
apneas actually suffer EDS, but that it is masked by two
factors. Dement et al. noted that SAS patients may deny
somnolence while literally falling asleep before the
clinician. They speculate that this results from a re¬
sponse bias against acknowledging illness or a change in
subjective frame of reference about just what alertness
is. Thus EDS may be denied by patients for reasons which
may be outside the clinician's control or knowledge.
Alternatively, Block et al. (1979) propose that asympto¬
matic SAS patients are suffering a subclinical level of
fallout from their frequent apneas. As years pass, in¬
creasing weight or cumulative effects of desaturation
eventually lead to a full blown hypersomnolence—Sleep
Apnea Syndrome. A final possibility is that nonhyper-
somnolent subjects with frequent apneas exhibit a com¬
pletely benign process totally unrelated to that found in
symptomatic SAS patients. It would seem that delineation
of the level of EDS in subjects with subclinical levels of
apnea may help clarify this issue. This notion will be
elaborated below.
The somnolence deficit which is noted in full blown
SAS patients is commonly reported. Dement et al. (1978)
describe impairments in continuous performance of
virtually any activity. Sleep Apnea Syndrome patients
reportedly fall asleep at outdoor stadiums, in front of
classes, and while treating patients. Excessive Daytime
Somnolence is described in every SAS patient reported in

30
experimental protocols by Weitzman et al. (1980), Fujita
et al. (1981), Zwillich et al. (1982), Garay et al. (1981)
and Sullivan and Issa (1980). At the same time, subjects
wno are asymptomatic for EDS but with the necessary number
of apneas to be diagnosed as SAS are described by Orr et
al. (1979) and Smirne et al. (1980) as well as others.
However, it should be noted that the latter studies relied
on global reports oi somnolence, rather than quantified
data. Recalling the criticism of Dement et al. (1978) of
this technique as vulnerable to subjective bias, these
asymptomatic patients may have unreported or undetected
somnolence.
Excessive Daytime Somnolence appears to be frequently
concomittant with Sleep Apnea Syndrome. Thus its possible
occurrence in subclinical populations seems worthy of
investigation.
Oxygen Desaturation and Cognitive Deficits
One deficit which appears in SAS patients is oxygen
desaturation which accompanies apneas. When apnea begins,
arterial oxygen saturation begins to fall, and may con¬
tinue falling throughout the event. An early technique
utilized in studying oxygen saturation levels involved
periodic blood samples drawn from arterial sources.
Eirchfield et al. (1958) studied 11 normal males with a
mean age of 23* Daytime oxygen saturation in these sub¬
jects averagea 96.5?. These values fell during sleep to
95.3?. Later, Orr et al. (1979) studied 8 males who had

31
been matched on number of apnea events. Asymptomatic
(nonhypersomnolent) subjects fell from a baseline 02
saturation of 80 mm (Hg) to a mean maximum desaturation of
54 mm (Hg) during sleep. Symptomatic subjects fell from a
baseline of 54 mm (Hg) to a mean maximum desaturation of
35 mm (Hg) while asleep. Orr et al. suggested that the
more severe desaturation which accompanied apneas in
symptomatic subjects underlaid their somnolence and other
complications. However, these conclusions are vulnerable
to certain methodological criticisms, as Block (1980)
points out. Specifically, periodic sampling may well fail
to detect the multiple brief desaturations associated with
SAS, rendering this technique potentially insensitive to
important events.
In an attempt to provide more representative data on
desaturation, Block et al. (1979) utilized an ear oximeter
in studying saturation in normal subjects. An ear oxi¬
meter provided a continuous, accurate readout of moment to
moment arterial oxygen levels. Thirty males (mean age 38)
and 19 females (mean age 27) were studied. Of the 30
males, seventeen suffered at least one episode of desatu¬
ration, falling from a mean saturation of 95% to a maximum
desaturation of 84% during sleep. In contrast, no epi¬
sodes of desaturation were noted in the sample of premena-
pausal females (baseline 02 saturation: 96%). The oxygen
desaturations of the males were always found in asso¬
ciation with breathing abnormalities or snoring. A study

32
of oxygen saturation in 20 post menapausal females was
reported by Block et al. (1980). Eleven of these women
exhibited desaturation with several subjects desaturating
to less than 85/>. Dolly and Elock (1982) studied a group
of 17 males and 3 females with a mean age of 49 years.
These normal subjects dropped from a mean baseline satura¬
tion of 96.3% to a mean maximum desaturation of 88.3%.
This evidence indicates that desaturations of 10-12?' are
common in normal males and post menapausal females while
premenapausal females rarely, if ever, desaturate.
Oxygen saturation levels in SAS patients are not
commonly reported. Using an ear oximeter, Garay et al.
(1981) studied 11 SAS patients who were divided into 6
daytime eucapnics and 7 daytime hypercapnics. Eucapnics
fell from a baseline saturation of 95% to a mean maximum
desaturation of 73.5%, while hypercapnics fell from a
baseline saturation of 89% to a maximum desaturation of
62.8% (calculated from tabled data). These data were
collected during daytime "nap" studies, however, and hence
is somewhat questionable. Berry and Block (1983) recorded
9 heavy snoring males who suffered many apneac events.
Four subjects were somnolent, while 5 were not. Symptoma¬
tic subjects fell from a baseline of 95% to a mean maximum
desaturation of 44“', while asymptomatic subjects desatu-
rated from a baseline of 94% to a mean maximum desatura¬
tion of 85%.

33
Data on desaturation are interesting from several
perspectives. First of all, Sleep Apnea Syndrome subjects
seem to desaturate more heavily than normals, with Berry
and Block's symptomatic group desaturating up to 50% and
the subjects of Garay et al. desaturating about 22%,
compared with normal desaturation levels of 10-12%.
Secondly, amount of desaturation distinguished symptomatic
(hypersomnolent) from asymptomatic subjects in the data of
both Orr et al. and Berry and Block. Additionally, two
points relevant to methodology emerge. Measures of oxygen
desaturation seem crucial to understanding sleep apnea and
its effects, and ear oximetry appears to provide a proven,
valid measure of this variable.
Weitzman (1979) has suggested that the pathological
process in SAS is the hypoxia of desaturation in apneac
episodes. While cognitive and intellectual changes in SAS
have been only anecdotally reported (Guilleminault et al.,
1978), hypoxia does seem a plausible potential cause of
these changes. Thus oxygen desaturation appears to be a
reliable consequence of sleep apnea syndrome, while cogni¬
tive/intellectual changes are putative sequelae.
In summary, a review of deficits found in sleep apnea
syndrome subjects suggests that cardiopulmonary complica¬
tions, excessive daytime sleepiness, oxygen desaturation
and cognitive/intellectual changes are frequently asso¬
ciated with this syndrome. An assessment of these
variables in a subclinical population seems indicated by

34
these data. Below a discussion of measurement issues and
proposal assessment of these variables appears.
Measurement of Deficits Found in Sleep Apnea Syndrome
Above it was noted that cardiopulmonary complica¬
tions, excessive daytime sleepiness, nocturnal oxygen
desaturation and cognitive/intellectual changes are common
deficits of sleep apnea syndrome. A discussion of
measurement issues and proposed measurement devices for
these variables will be presented below.
Measurement of Cardiopulmonary and Health Deficits
Cardiopulmonary deficits are thought to be frequent
sequelae of sleep apnea syndrome. Thus an assessment of
blood pressure as well as self reports of hypertension and
heart trouble seems indicated in subclinical apnea sub¬
jects. Additionally, a general survey of health status,
such as that provided by the Cornell Medical Index (CMI;
Broadman et al., 1949) would also screen for other health
deficits. The CMI provides an overall score indicating
number of symptoms endorsed, as well as subscales on
several symptom categories. Separate examination of sub¬
scales on respiratory, cardiopulmonary, and neurological
subscales seems appropriate. Assessment of these
variables provides a broad health screening, as well as a
detailed analysis of symptoms found in sleep apnea syn¬
drome.

35
Measurement of Arousal Deficits (Hypersomnolence)
Excessive daytime sleepiness is a common sequela of
sleep apnea syndrome. Measurements of the sleep/wake
cycle of subclinical apnea patients thus seems indi¬
cated. A sleep questionnaire assessing trait sleep habits
seems in order, as well as 1 week sleep logs following the
experimental night. The sleep questionnaire and sleep
logs have been used in ongoing research in W.B. Webb's
studies of aging and sleep and seem an adequate assessment
of sleep patterns of subclinical apnea subjects. Addi¬
tionally, electroencephalographic recordings of sleep
during the experimental night seem in order using a stan¬
dard recording and scoring system (Agnaw and Webb,
1972). The assessment of daytime sleepiness is relatively
new, only recently being the subject of investigation.
William Dement and associates began to develop methods of
quantifying somnolence in the early 1970s. The first
attempt produced a subjective rating scale, the Stanford
Sleepiness Scale (SSS; Hoddes et al., 1973). The SSS was
a 7 point Likert scale with endpoints anchored on 1) feel
active and vital; and 7) almost in reverie, sleep onset
soon; with 5 points with statements reflecting various
degress of somnolence in between. Patients rated their
introspective level of sleepiness on this scale.
Unfortunately, the SSS proved to be vulnerable to exactly
the same difficulties as a global rating scale used on SAS

36
patients, i.e., response bias and altered frame of
reference (Dement et al., 1976).
It was clear then that an objective measure of day¬
time somnolence was needed. With this goal in mind,
Carskadon and Dement (1977) introduced the Multiple Sleep
Latency Test (MSLT). This technique involved measuring
the objective sleep latency during multiple nap attempts
throughout the day. In a typical paradigm, a subject
would be wired to an EEG and attempt to fall asleep at 2
hour intervals through the day. The latency to sleep in
each nap was measured by determining the latency to the
first epoch of Stage 1 sleep. In order to prevent signi¬
ficant amounts of sleep accumulating during testing, sub¬
jects were awakened after an epoch of sleep was
observed. If no sleep was noted, subjects were discon¬
nected after 20 minutes and allowed to resume other acti¬
vities. Carskadon and Dement believe that this measures
underlying physiological sleep tendency. This underlying
tendency is thought to be modulated by alerting stimuli in
the environment (Carskadon and Dement, 1982a). The vali¬
dity of the test was established on normal subjects, who
exhibited a biphasic curve of sleep tendency throughout a
24 hour period. Extension of sleep over typical lengths
in normals (who are thought to be moderately chronically
sleep deprived) saw a significant rise in average sleep
latencies on the MSLT (Carskadon and Dement, 1979b), while
chronic sleep restriction (5 hrs sleep per night) caused

37
significant declines in average sleep latencies (Carskadon
and Dement, 1981a). Finally, total sleep deprivation led
to a drastic decline in MSLT values, with averages
plummeting below one minute (Carskadon and Dement,
1979a). Mean MSLT values showed recovery to baseline
values after a full night's sleep for partially (5 hrs)
deprived subjects, and recovery after two nights of ad lib
sleep for totally sleep deprived subjects (Carskadon and
Dement, 1979; 1981a). This body of evidence suggests
that the MSLT is a reliable and valid index of sleepiness
in normal subjects.
With its internal validity established, the MSLT was
applied to somnolent populations. Dement et al. (1978) as
well as Richardson et al. (1978) demonstrated that SAS and
narcoleptic patients have significantly shorter latencies
to sleep on the MSLT than normal subjects do. Hartse et
al. (1979) found SAS subjects to have an average latency
to Stage 1 of 2.5 minutes, while narcoleptics had a mean
of 3.2 minutes. Both these latencies were significantly
lower than those of a miscellaneous control group whose
latencies were about 11.2 minutes. Hartse et al. (1980)
found latencies to Stage 1 sleep as follows for various
patient groups: Normals 12.2 m, Insomniacs 16.2 m, SAS
2.5 m, and Narcoleptics 3.2 m. The SAS and narcoleptic
patients had significantly shorter latencies to Stage 1
sleep than insomniacs or normals. Roth et al. (1980)
compared 10 SAS patients with 10 age matched controls.

38
The normals averaged 20.4 m to Stage 1 sleep while the SAS
patients averaged 1.9 m. Zorrick et al. (1982) found mean
MSLT latencies to Stage 1 sleep of 3.1 m for SAS subjects
and 2.9 m for narcoleptics, significantly shorter than for
patients with psychiatric disorders.
The chief difficulty remaining in judging the MSLT
"reliable and valid" is external validation demonstrating
that exceeding a certain mean latency is correlated with
other deficits of performance. In some ways, low laten¬
cies on the MSLT may suffer the same limitations as sleep
deprivation protocols; that is, demonstrating serious
deficits is surprisingly difficult. At the same time, few
would deny the subjective unpleasantness of sleep depriva¬
tion, which may be applicable to low MSLT scores. Thus
the MSLT appears indicated in assessing excessive daytime
sleepiness while the Stanford Sleepiness Scale (SSS) pro¬
vides an index of the subjective experience of the sub¬
ject. Application of both these measures to a subclinical
apnea population would seem to adequately assess daytime
sleepiness in these subjects.
Measurement of Hypoxia and Its Sequelae
Noctural desaturation is frequently noted in SAS
patients. Therefore, it seems in order to examine the
known sequelae of hypoxia of other etiologies in an
attempt to determine variables or behaviors likely to be
sensitive to the hypoxia found in SAS and SAA. Research
on hypoxia varies systematically along two dimensions,

39
human vs. animal subjects and acute vs. chronic hypoxia.
First, the data available on animals will be reviewed,
followed by that extant on human subjects. Finally,
behavioral variables sensitive to changes in these
variables will be summarized.
The effects of chronic hypoxia on animals have not
been exhaustively studied (Davis, 1975). One project
which did examine this variable involved exposure of rats
to up to 36 hours of 10$ oxygen. Davis (1975) observed
the behavior of these animals and concluded that they
appeared normal. Although hyperventilation was noted, no
permanent effects of the hypoxia were found. Hanbauer et
al. (1981) examined the effects of 10$ on rats exposed
for up to 4 weeks. They found that relatively short term
(2 days) hypoxia resulted in an increase in dopamine (DA)
content in the carotid body. An increase in norepine¬
phrine (NE) content was noted after 1 week of exposure.
As the carotid bodies are implicated in the control of
respiration, the authors concluded that different
mechanisms are operative in the adaptation to short and
long term hypoxia. Unfortunately, little other data have
been reported on the chronic effects of hypoxia on more
complex functioning. This is particularly problematic as
it are these functions which are thought to be affected in
humans exposed to hypoxia.
Certain aspects of acute hypoxia in rats have been
carefully examined. Gibson et al. (1981) reviewed the

40
evidence which has been collected on neurotransmitter
alterations in hypoxia. In brief, mild to moderate
hypoxia, cerebral levels of ATP are normal, implying that
energy supplies of neurons are not disturbed. In con¬
trast, the turnover of certain neurotransmitters is
altered. Gibson et al. note that oxygen is an important
substrate for the synthesis of several neurotrans¬
mitters. When rats are exposed to acute hypoxia, syn¬
thesis of dopamine and serotonin is decreased, while
absolute levels of tryosine, tryptophan, catecholamines,
and serotonin remain constant. Continuing hypoxia leads
to an adaptation resulting in normal turnover of these
neurotransmitters. Gibson and Blass (1976) examined the
affect of acute hypoxia on central acetylcholine (ACh).
As in dopamine (DA) and serotonin (5-hT), acute hypoxia
results in a decrease in synthesis in ACh, but not in the
absolute level of ACh. Thus there is a substantial body
of evidence indicating changes in central neurotransmitter
levels in hypoxia. Generalizing from the animal data to
human subjects, Gibson and Duffy (1961) suggest that these
changes underlie the behavioral deficits of hypoxia in
humans (which will be reviewed below).
Other variables have been examined in acute hypoxia
in animals. Annau (1972) found that rats exposed to brief
hypoxia exhibited decreased appetite, thirst, and self
stimulation rates. Annau observed that brief hypoxia
depressed all behavioral variables. Gellhorn (1951) found

41
that brief hypoxia of 4-7.52 02 abolished cortical re¬
sponses to auditory stimulation. Sara (1974) trained rats
in active one way avoidance tasks. Immediately after
training, subjects were exposed to 3.5-4.0# oxygen. Rats
tested at 1 and 3 hours post hypoxia avoided correctly,
while those tested at 6 and 24 hours did not, leading Sara
to suggest that hypoxia resulted in a memory retrieval
deficit. Freides and Allweise (1978) also trained rats in
one way avoidance and exposed them to 2.02 hypoxia imme¬
diately following training. These animals avoided at 12
and 200 minutes post hypoxia, but not at 90 minutes.
Friedes and Allweise suggested that a sensitive period for
hypoxia effects on training existed for 15 minutes after
learning. Tauber and Allweise (1975) followed the same
training/hypoxia paradigm and found avoidance impaired at
2.5 but not 4.5 hours. They suggest that hypoxia inter¬
feres with a medium term memory. V/hile some of these data
are conflicting, it appears that acute hypoxia depresses
certain behavioral variables and interferes with aspects
of memory processes.
While caution is in order when generalizing from
animals to humans, several seemingly relevant points
emerge from the review. Behaviorally, simple activities
may be spared, although some depression of appetitive
activities may occur, while more complex behavior depen¬
dent on memory processes may be altered. Of particular
interest are the demonstrated alterations in

42
neurotransmitter turnover seen in hypoxia. It has been
suggested that these subtle alterations may underlie the
changes in complex behavior seen in hypoxic humans.
Data on human performance after hypoxia have neces¬
sarily been reliant on fortuitous or weak manipulations.
Although this may limit the usefulness of these data, it
is at the very least highly suggestive of variables sensi¬
tive to hypoxia. Van Liere and Stiokney (1963) note that
acute hypoxia leads to confusion, headache, drowsiness,
weakness and incoordination. After effects include head¬
ache, nausea and emotional lability. Richardson et al.
(1959) found that acute, fatal hypoxia from surgery
results in damage to cortical neurons, particularly in
layers III and IV. Susceptible subcortical neurons in¬
clude those of the corpus striatum and cerebellum. Plum
et al. (1962) describe a syndrome of aelayed postanoxic
encephalopathy. After the acute coma resolves in 4-5
days, a brief period of normal functioning returns.
Increasing irritability and confusion then emerge, with
loss of coordination and memory, and diminished attention
span. Typically, the lesion spares grey matter while
affecting white matter. Devereaux and Partnow (1975)
describe a patient who recovered from a delayed acute
encephalopathy. Although IQ was relatively spared, severe
dysarthria emerged. Relevant points which emerge from
studies of acute hypoxia in humans suggest that sublethal

43
levels of hypoxia affect coordination, memory, attention
span, and may result in headache, drowsiness, and nausea.
The effects of milder, chronic hypoxia are summarized
by Gibson et al. (1981). At the oxygen pressure equiva¬
lent of 5,000 feet, impaired dark adaptation is noted. At
10,000 feet decreased concentration, hyperventilation, and
short term memory deficits are noted. By 15,000 feet,
euphoria and loss of both coordination and critical judge¬
ment are observed. Christenson et al. (1977) found signi-
ficanxly less visual signals detected at 17% O2 than at
21%, suggesting reduced vigilance. McFarland (1937)
studied men in ascents to high altitudes and concluded
that subjects with slower rates of ascent performed better
on unspecified psychological tests. West (1984) reports
data collected on an expedition to Kt. Everest. Results
of a neuropsychological battery indicated deficits in
finger tapping, verbal fluency, verbal learning, short
term memory, and expressive language. Thus the general
affects of chronic mild hypoxia are thought to be deficits
in dark adaptation, short term memory, critical judgement,
motor coordination, verbal fluency, and vigilance.
One group of chronically hypoxia patients, those with
chronic obstructive lung disease (COLD), has been rela¬
tively well studied with sensitive neuropsychological
instruments and is deserving of a more painstaking
review. Krop et al. (1973) studied COLD patients before
and after they received oxygen supplements. The UAIS,

44
Wechsler Memory Scale (WMS), Bender-Gestalt, Bender
Gestalt Interference Procedure, Facial Recognition and
Finger Tapping tests were administered to COLD and control
subjects. Subjects receiving supplementation improved
significantly more than controls on WAIS Full Scale IQ,
Performance IQ, Wechsler Memory Quotient, both Bender
procedures and Finger Tapping. Grant et al. (1982)
studied another group of COLD patients. They administered
the Halstead-Reitan Battery, Aphasia Screening Test,
Trailmaking Test, Russell modification of the Wechsler
Memory Scale, Tactile Memory test, and the WAIS. In COLD
subjects, deficits were found relative to controls on all
variables except Reitan Rhythm, Aphasia Screening and WKS
Logical stories. Grant et al. concluded that COLD
patients exhibited deficits on a global impairment rating,
attention, abstracting ability, complex perceptual motor
tasks, simple sensory and motor tasks, and memory tasks.
Low but significant negative correlations were reported
between several measures of oxygen saturation and both
Halstead Impairment Index and Global Impairment Rating.
Grant et al. speculated that floor effects resulted in low
variability in P¿02 levels among COLD patients and kept
these correlations low. They further suggested that the
neuropsychological deficits observed in these patients
resulted from the oxygen want found in these subjects.
Thus data available on humans indicate that dark
vision, memory, motor coordination, critical judgement,

45
verbal fluency, and signal detection are affected by acute
hypoxia. Animal data suggest that sensitivity of memory
processes to hypoxia, while data from chronic hypoxia
humans indicate that attention, vigilance, abstracting
ability, complex perceptual' motor skills, as well as motor
and sensory abilities are also impaired by hypoxia. While
none of these findings are unequivocally directly
applicable to the brief, acute, desaturaticn occurring
during apnea, they are, at the very least, indicative of
abilities sensitive to hypoxia in humans.
A neuropsychological battery sensitive to these
abilities includes Wechsler Adult Intelligence Scale
(WAIS; Wechsler, 1955); Wechsler Memory Scale (WMS;
Wechsler, 1945) including the Russell delayed memory
aspects (Russell, 1975); Rey-Osterreith Complex Figure
(Osterreith, 1944); Hooper Visual Organization Test
(Hooper, 1958); Verbal Fluency Test (Halstead, 1947);
Wisconsin Card Sort (Berg, 1948); and the Finger Tapping
Test (Halstead, 1947). Administration of this test
battery provides a broad screening of neuropsychological
functioning, with a particular emphasis on those abilities
known to be sensitive to hypoxia. Application of this
battery to subclinioal apnea subjects provides a sensitive
indicator of possible hypoxic brain dysfunction.
In summary, alterations of sleep/wake cycles appear
to be present in sleep apnea syndrome patients. Thus
assessment of these variables in a subclinioal population

4b
is indicated. An adequate description of the sleep/wake
cycle would seem to include sleep questionnaire, sleep
log, overnight EEG, multiple sleep latency test, as well
as several Stanford Sleepiness Scale ratings.
Statement of the Problem
Sleep Apnea Syndrome is a clinical entity with
seemingly serious consequences. Although the nocturnal
apneas found in this illness are thought to be the cause
of the associated deficits, the identification of
seemingly asymptomatic subjects with a number of apneas
comparable to those found in SAS throws doubt onto this
formulation. Despite this evidence, the recent discovery
that apneas occur in significant numbers in the general
population has led to speculation that this subclinical
level of apnea may be the cause of various pathological
processes. At present, there is little evidence to
support or refute this notion.
The present report evaluates this issue through a
detailed study of a group of subjects thought to be at
risk for sleep apneas. Aging, heavy snoring males were
recruited, and a comprehensive analysis of their respira¬
tory, health, neuropsychological, and sleep/wake cycle
status was carried out. Thus potential relationships
between respiratory disturbances and variables known to be
disrupted in sleep apnea syndrome patients could be
evaluated. This approach also allowed exploration of the

47
possible risk factors associated with subclinical sleep
apneas.

CHAPTER TWO
METHOD
Subjects
Heavy snoring males were recruited via newspaper
advertisements, notices, and phone calls to a list of
aging subjects reported by Webb (1982). Participants were
required to be male, self reported heavy snorers, over 30
years of age, and in general good health (i.e., self
described healthy and normal and not under the active care
of a physician for illness). Subjects with a history of
head trauma or alcoholism were excluded. Potential
subjects who met these criteria were offered a chance to
participate in a "snoring study" in which they would
complete testing during an evening, sleep overnight in a
lab with several physiological parameters recorded, and
complete one week's worth of "sleep logs," for a payment
of fifty dollars. Volunteers were scheduled on a first
come, first served basis. A total of 60 subjects were
studied, with 9 dropped because of technical problems with
their overnight recordings and 5 excluded because of
health problems. Thus 46 subjects represented the final
sample. These 46 subjects had a mean age of 49.9 (sd
13.1) and represented a wide cross section of the
population, ranging from the unemployed to university
48

49
professors. All subjects completed and signed an informed
consent agreement.
Apparatus
During the evening testing, all subjects received a
blood pressure check with a standard hospital cuff. The
measurement was made from a seated position. Overnight
recordings were made in a quiet, darkened room with the
subject sleeping on a standard hospital bed. All noc¬
turnal physiological recordings were routed through a
Grass Polygraph (model 7D). Electrodes for electro-
encephalographic recordings were mounted in three pairs
using sites F2/F8, PZ/T6, RE/LE from the international
10/20 system of placement. These electrodes were affixed
to the scalp with collodion soaked gauze pads dried with
an air hose. Respiratory measures included measurement of
chest and abdomen wall movement derived from impedence
measures monitored from surface electrodes (3K Ag/AgCl
#2246) placed on the lower chest and just above the
navel. The signal from these electrodes was passed
through an impedence convertor and onto the polygraph.
Oral and nasal airflow was recorded by thumistors (TCT-1R
Transducer-Grass) clipped to the respective orifices.
These signals were also routed through the polygraph.
Electrocardiograms were monitored from modified bipolar
chest leads (MCLg). Finally, blood oxygen saturation was
continuously monitored from a probe attached to the ear
lobe, and analyzed by a Biox ear oximeter (tlodel IIA BTA

50
Co). All physiological variables were recorded on the
polygraph's chart paper which was run at a speed of 10
mm/s.
Measures
Two classes of independent variables are identified,
with the first being thuraistor variables (those derived
primarily from the nose and mouth thumistors). These
included number of apneas and hypopneas, mean duration of
apneas and hypopneas, mean low oxygen saturation in apneas
and hypopneas, mean oxygen saturation change in apneas and
hypopneas, and total seconds spent in apneas and hypo¬
pneas. Thumistor variables were separately scored by two
trained raters, whose agreement appears in Table 2-1. An
overall agreement of 85$ was achieved. Disagreements were
resolved by consensus scoring. The second class of
independent variables included oxygen saturation variables
(derived from a minute by minute rating of highest and
lowest oxygen saturation from the overnight record).
These variables included mean highest saturation, mean
lowest saturation, and number of desaturations of 4 ana
10$ or more. In selecting these scoring criteria, several
considerations were relevant. Initially, a system with
reasonable interrater reliability was necessary. While
published criteria for scoring apneas were not problem¬
atic, reliable scoring of hypopneas was not achieved until
a desaturation criterion of 10$ was introduced. In
selecting measures of oxygen saturation, a mean high and

51
Table 2-1. Interrater agreement for respiratory events of
46 snoring males.
Consensus
Initial
Disagreements
Scoring
Agreements
#1
H2
Note
(178)
_
.
.
Scored by
consensus
91
90
4
6
Consensus
added 1
78
76
1
1
Consensus
added 2
61
52
3
8
Consensus
added 9
57
56
9
2
Consensus
added 1
32
29
3
3
Consensus
added 3
28
27
0
2
Consensus
added 1
25
25
0
0
19
16
4
0
Consensus
added 3
17
17
0
0
15
16
1
0
Consensus
added 1
13
13
0
0
7
7
0
0
4
4
0
0
4
4
0
0
3
2
0
1
Consensus
added 1
3
3
0
0
3
0
7
2
Consensus
added 3
3
3
0
0
3
3
0
0
1
1
0
0
1
1
0
0
1
1
0
0
1
1
0
1
1
0
0
1
Consensus
added 1
1
1
0
0
1
1
0
0
1
1
0
0
474(652)
450
37
36
Note: 26 of
the 46
subjects
experienced at least 1 apnea or
hypopnea.

52
low score seemed obvious. In order to assess the quan¬
titative "impact" of numerous desaturations, a sum of A%
and 10% desaturation was tabulated. These selected
parameters for thumistor and saturation variables allowed
reliable scoring and a description of several aspects of
nocturnal respiration.
Dependent measures fell into three broad classes:
measures of sleep/wake status, measures of health status,
and measures of neuropsychological status. Measures of
the sleep/wake cycle included one week's sleep logs
(derived variables included number of reports of daytime
sleepiness, hours asleep, estimated sleep latency, number
of wakenings after sleep onset, and minutes of waking
after sleep onset), a sleep questionnaire (derived
variables including hours of sleep usually obtained,
number of naps, hours napping, quality of sleep,
restfulness on awakening, depth of sleep, and amount of
daytime sleepiness), the mean of the several Stanford
Sleepiness Scale ratings, the two multiple sleep latency
tests (derived variables included mean sleep latency and
total minutes asleep in both naps) and the overnight EEC
variables derived from the scoring system of Agnew and
Webb (1972). Electroencephalographic variables included
time in bed; pure sleep time; sleep efficiency index;
number stage 0 periods; time % stage 0; sleep latency;
time % stages 1, 2, 3, 4 and REM; latency first REM

53
period; mean REM period; mean REM cycle length; % slow
wave sleep; and number of stage changes.
Measures of health status included height, weight,
diastolic and systolic blood pressure, as well as the
Cornell Medical Index. From the CMI, reports of hyper¬
tension and heart trouble were noted, as well as the total
number of symptoms endorsed. Additionally, the scores for
the scales tapping respiratory, cardiac, and neurological
symptoms were tabulated separately.
Measures of neuropsychological status included the
WAIS Verbal and Performance IQ's, the V/MS Memory Quotient,
delayed recall of the logical stories and the visual re¬
production subtests of the WMS, delayed recall of the Rey
complex figure, number of words generated in the verbal
fluency test, number of correct sorts of the Wisconsin
Card Sort, mean number of taps for the right and left
handed finger tapping tests, and lastly the number of
correct trials on the Hooper test.
These neuropsychological tests were chosen with two
aims. First, an attempt was made to sample from each of
several loosely bounded areas of cognitive skill. These
included intelligence (both verbal and non-verbal),
memory, both immediate and delayed (verbal and non¬
verbal), visuo-perceptual/organizational skills, language,
and frontal self-regulatory skills. A second emphasis was
on a relatively detailed analysis of areas previously
shown to be sensitive to hypoxia. These sensitive areas

54
included memory, visuo-organizational skills, verbal
fluency, and motor coordination. The chosen neuropsycho¬
logical battery reflected these two aims. Thus the WAIS
tapped intellectual functioning; the Wechsler Memory
Scale with delayed recall of logical stories and visual
reproductions as well as digit span and Rey Figure heavily
sampled verbal and non-verbal memory; the Hooper tapped
visuo-organizational skills; Verbal fluency sampled lan¬
guage skills; the Wisconsin card sort tapped frontal
sell-regulatory skill; and Finger tapping sampled motor
functions. The chosen battery provided a broad screening
as well as a detailed assessment of neuropsychological
functions thought to be impaired by hypoxia.
Procedure
Subjects followed the schedule appearing below for
the experimental night:
1800 Sign informed consent, blood pressure,
height/weight, finger tapping, verbal fluency,
subjective sleepiness (SSS-made every V2 hour
till bedtime)
1830 Wiring for EEG
1900 Begin HSLT I (20 m)
1930 WAIS
2000 VMS, Hooper
2030 Graphesthesia, Luria motor programs
2100 Begin MSLT II (20 m)
2150 Rey figure, Wisconsin card sort

55
2200 Sleep Questionnaire, Cornell Medical Index,
remaining testing
2230 Wiring for thumistor, EKG, respiratory effort,
ear oximeter
2300 Lights out, begin recording
A night technician remained with the subjects throughout
the night. Subjects usually awakened between 5 and 6 am,
and later completed the one week's sleep logs.
Statistical Procedures
Several statistical procedures were utilized in
evaluating the present data. One of the primary
approaches was correlational, considered appropriate
because of the exploratory nature of the study. Bearing
in mind the possible alpha inflation of large corre¬
lational matrices, other approaches were also utilized to
add confidence. Thus the sample was stratified by level
of apnea/hypopnea, and oneway ANOVA procedures were used
to evaluate between group differences. Additionally, non-
parametric procedures were utilized where appropriate.
Positive correlational results were considered against
possible replication with other measures in the study, and
in some cases multivariate procedures were added for
further confidence. It is conceded that correlational
approaches may carry the burden of possible spurious
results, but if is believed that the various additional
procedures outlined above lend confidence to the findings.

CHAPTER THREE
RESULTS
Demographics and Incidence of
Nocturnal Respiratory Disorder
Subjects were selected to fulfill several sampling
requirements, including male sex, presence of heavy
snoring, and self report of general good health. A total
of 60 subjects were studied. Five subjects were later
excluded when questioning revealed the presence of chronic
alcoholism (2) and serious head trauma (3). Additionally
9 overnight records were rejected because of missing or
unusable data. The characteristics of the final 4b
subjects are detailed in Table 3-1. These snoring males
had a mean age of approximately 50 years, a mean weight of
approximately 190 pounds, and a mean educational level of
15 years of schooling.
Table 3-1. Demographic variables
from 46
snoring
males.
K
SD
Range
Age
49.9
13.1
30-75
height (lbs)
189.3
34.9
125-250
Veight:Height Ratio (lbs:inches)
2.7
0.5
1.9-3.6
Education (years)
14.7
2.5
9-20
56

57
The respiratory records were annotated by two hour
intervals following bedtime. From a total of 652 events
scored, 150 (23%) occurred in the first two hours
following bedtime, 185 (23$) occurred in the second two
hours, 232 (36?!) occurred in the third two hours, while 83
(12?!) occurred in the final two hours. It should be noted
that wakeup times varied considerably, making interpre¬
tation of the attenuated number of events in the final two
hours somewhat problemmatic.
From a total of 578 events in which reliable sleep
staging was achieved, 45S (78%) occurred in light slow-
wave sleep (stages 1-2), while 118 (20%) began in REM
sleep. Only 6 events (01?!) occurred in slow wave sleep.
These precentages must be considered in light of the
relative distribution of sleep stages in these subjects:
stages 1-2 (64%), stage REM (12%), stages 3-4 (10?!), stage
0 (14%). This distribution suggests that the sleep time
spent in stages, 1, 2, or REM must have provided the
overwhelming preponderance of events in these stages.
Sleep characteristics will be discussed in greater detail
below.
As noted below, 28 of the 46 subjects experienced at
least one episode of apnea or hypopnea. Table 3-2
presents the data on occurrence of at least one
apnea/hypopnea, as well as the occurrence of a high level
of apnea/hypopnea. Almost 2/3 of the sample had at least
one apnea or hypopnea, while 13?! had high levels of

58
Table 3-2. Incidence of low and high levels of
apnea/hypopnea by age in 46 snoring males.
Age Group
N
Low Apnea/
Hvpopnea
High Apnea/
Hvpopnea
50-39
14
.71
0
40-49
7
.57
.29
50-59
11
.72
.27
60+
14
.43
.07
Overall
46
.62
.13
apnea/hypopnea. Note that subjects with high levels of
apnea/hypopnea are included in the overall total of
subjects wixh at least one event for the following
analyses. The classification of "high" levels of
apnea/hypopnea utilizes the criterion of an apnea +
hypopnea index greater than 5, a frequently used clinical
cut-off score. The table also presents the frequency of
respiratory distress broken down by age. No simple age
related trend is apparent, and a X2 analysis of the two
factors of age and respiratory distress was not signi¬
ficant for those with at least one event (X¿=3.2; p=NS),
or for those with high levels of apnea/hypopnea (X2=5.9;
P-NS).
A further investigation of possible age by level of
apnea activity involved division of the sample into those
above ana below 60 years of age crossed with those above
and below 5 events/hour. The X2 value for this comparison
was also nonsignificant (X2=0.6; p=NS). Subjects were

59
also divided into those above and below the mean weight:
height ratio (2.7) and crossed with those with and without
át least one event or high levels of apnea/hypopnea. Both
these non-parametric tests were also non-significant
(X2=.01; p=KS; X2=1.5; p=NS).
1 able 3-3 utilizes Pearson correlations to explore
possible relationships between, demographic variables (age,
weight, weight:height ratio) and nocturnal respiratory
indices (apnea index, apnea + hypopnea index, seconds in
events, mean high saturation, mean low saturation, number
of desaturations >_A%, anc number of desaturations _>10&).
Again, age displays little relaxionship with respiratory
indices, but substantial correlations between weight and
various respiratory indices are apparent.
Table 3-3. Significant (p<.05) Pearson correlations between
nocturnal respiratory variables and demographic
variables in 46 snoring males.
Respiratory
Variable
Number of Apneas
Number of Apneas+Hypopneas
Apnea Index
Apnea+Nypopnea Index
Kean High Saturation
Mean Low Saturation
Mean Saturation Change
Number of Desaturations _>4
Humber of Desaturations >10
Weight:Height
Age Weight Ratio
.265
260
.306*
323*
-.363*
3S7*
-.460*
319*
.373*
368*
.43-0*
p<.01

60
In summary, non-parametric analyses failed to reveal
significant relationships between age or weight and
overnight respiration. However, Pearson correlations
suggested that increasing weight is associated with
deterioration of several noctural respiratory indices.
Because increasing apnea/hypopnea levels have been
suggested to carry increasing risks for various deficits,
the sample was stratified by level of apnea/hypopnea. The
first group included those subjects without events (Wo
events), those with at lease one event but less than 5
events per hour (Low apnea/hypopnea), and those with 5 or
more events per hour (High apnea/hypopnea). It should be
noted that this division of subjects is slightly different
from the earlier non-parametric analyses, because in this
grouping low apnea/hypopnea and high apnea/hypopnea are
mutually exclusive, while in the earlier analyses the
group with at least one apnea/hypopnea included those with
more than 5 events per hour.
Table 3-4 utilizes oneway ANOVA procedures to compare
these groups on demographic and nocturnal respiratory
variables. There were no significant between group
differences on any of the demographic factors: age,
weight, weight:height ratio, and education [F(2,43)=2.1,
0.6,0.8,0.7; p=HS]. Logically, both apnea and hypopnea
indices were significantly different between groups
[F(2,43)=9.6, p<.0004; 20.5, pC.0000]. Followup
comparisons (Scheffe) indicated that the high

Table 3-4. Means and standard deviations for selected demographic'ana nocturnal
respiratory variables in 46 snoring males grouped by level of
apnea/hypopnea.
No Apnea/
Low
Apnea/
High Apnea/
Hypopnea
Hypopnea
Hypopnea
Age
52.3(12.7)
46.1
(15-5)
56.5
(9.3)
Weight (lbs)
185.6(34.8)
188.5
(37.6)
203.3
(24.3)
Weight:Height Ratio
2.7 (0.4)
2.7
(0.5)
2.9
(0.4)
Education (years)
15.2 (2.9)
14.6
(2.0)
13.8
(2.6)
Number of Apneas
0
5.0
(7.7)
24.5
(34.1)a,b
Number of Apneas+Hypopneas
0
7.3
(9.2)
81 .7
(52.6)
Apnea Index
0
0.9
(1.1)
4.5
(5.9)a’b
Apnea+Hypopnea Index
0
1.4
(1.5)
19.5
(19.4)3.6
Seconds in Apneas/Hypopneas
0
129.4(167.1)
2325.5(1469.5)a,D
Hern Apnea Duration
-
13.4
(4.0)
17.8
(1.3)
Mean Hypopnea Duration
-
29.6
(21.2)
33-2
(14.5)
Mean High Saturation
95.5 (1.7)
95.3
(1.3)
94.9
(1.2)
Mean Low Saturation
94.3 (2.1)
93.1
(1.6)
89.0
( 4.6)a ’ b
Number of Desaturations >4£
29.5(54.1)
5e.2
(58.6)
143.2
(105.9)a,b
Number of Desaturations >10¡S
0.2 (0.7)
3.5
(6.3)
51 .0
(48.7)a,b
a
b
Significantly different from no sleep apnea/hypopnea.
Significantly different from low sleep apnea/hypopnea.

62
apnea/hypopnea group v/as significantly different from the
other two groups on these two indices. While mean
duration of apneas and hypopneas was not significantly
different between the high and low apnea groups
[F(1,25)=3.3. 0.1; p=NS], total seconds in respiratory
events was [F(2,43)=50.4, p<.0000], with followup
comparisons showing that the high apnea/hypopnea group was
significantly different from tne other two groups.
Turning to the oxygen saturation data, mean high
saturation failed to reach significance LF(2,43)=0.3>
P=NS]. However, mean low saturation, number of
desaturations >_4% and >10% were all significant
[F(2,43)=11.6, pC.0001; 7.0, p<.C02; 21.4, pC.OOOO], and
followup comparisons showed that the group with high
levels of apnea/hypopnea was significantly different from
the other two groups.
Thus, division of the sample into levels of
apnea/hypopnea resulted in groups which were not
significantly different on demographic variables including
age, weight, weight:height ratio, and education. However,
the group with high levels of apnea/hypopnea had
significantly more nocturnal events, time in events, and
oxygen desaturation.
In summary, the present sample of 46 snoring males
appears representative of those predisposed to experience
sleep disordered breathing (middle-aged, snoring males).
Sixty-two per cent of these subjects had at least one

63
apnea or hypopnea, while 13£ had high levels of
apnea/hypopnea. Most events occurred in light slow-wave
or REM sleep. No effect of age on level of sleep
disordered breathing was observed, although increasing
weight was correlated with nocturnal respiratory events.
When the sample was stratified by level of apnea/hypopnea
activity, subjects with high levels of events had
significantly more events and desaturations than the
remaining subjects, although no significant differences
were observed on age, weight or education.
Nocturnal Respiratory and Health Variables
Health variables including blood pressure readings
and several scores derived from the Cornell Medical Index
(overall number of symptoms endorsed, as well as number of
symptoms endorsed from the respiratory, cardiopulmonary
and neurological subscales) were evaluated initially
through a matrix of Pearson correlations which included
the respiratory variables (apnea index, apnea + hypopnea
index, seconds in events, as well as mean high and low
saturation and number of desaturations >J\% and _>105?).
From a total of 42 correlations, three were significant at
p<.05 or less: Systolic blood pressure and mean low
saturation (r=-.26l); Systolic blood pressure ana number
of desaturations _>4& (r=.373); Overall number of symptoms
endorsed on the CMI and mean high saturation (r=-.254).
As a further exploration of possible relationships between
blood pressure and apnea/hypopnea activity, subjects were

64
asked whether they had ever been diagnosed with hyper¬
tension. Thirty-xhree per cent of those with no events
had diagnosed hypertension, as did 33% of those with low
levels of apnea/hypopnea, while only 15% of those with
high levels of apnea/hypopnea had diagnosed hyperten¬
sion. These data do not support an increasing level of
hypertension in subjects with high levels of
apnea/hypopnea.
The subjects were grouped into the earlier delineated
levels of apnea/hypopnea. Table 3-5 presents the means
and standard deviations of health variables from the three
groups. No dramatic trends are apparent in this table,
and oneway ANOVA procedures confirmed this observation
with no significant differences for: Diastolic BP,
F(2,43)=.01, p=NS; Systolic BP, F(2,43)=.1, p=NS; CMI
overall, F(2,42)=.2, p=NS: CHI respiratory, F(2,42)=.5>
p=NS; CHI cardiopulmonary, F(2,42)=.5, p=NS; CMI
neurological, F(2,42)=2.2, p=NS.
In summary, while correlational procedures suggestea
relationships between several oxygen saturation measures
and blood pressure as well as overall number of symptoms
endorsed on the CMI, oneway ANOVAs failed to reveal
significant between group differences when subjects were
stratified by level of nocturnal distress.

65
Table 3-5. Means and standard deviations for health related
variables in 46 snoring males grouped by level
of apnea/hypopnea.
No Apnea/
Low Apnea/
High Apnea/
Hypopnea
Hypopnea
Hypopnea
Systolic BPa
129.1(14.2)
128.7(12.6)
127.0(7.5)
Dia¡
stolic BP
85.2(13.4)
83.9(11.0)
81.6(7.6)
CMI1
3 Overall Score
14.7 (7.4)
16.1(10.5)
13.8(7.4)
CMI
Respiratory Score
1.4 (1.2)
1.1 (1.8)
2.0(1.4)
CMI
Cardiac Score
1.6 (1.3)
1.2 (1.2)
1.0(1.2)
CMI
Neurological Score
1.2 (0.9)
1.6 (1.5)
c.e(o.e)
Tj BP = blood pressure.
b CMI = Cornell Medical Index.
Nocturnal Respiratory and Sleep/V/ske Variables
EEG Data
Electroencephographic variables were recorded
overnight for all subjects. Records were scored by a
trained technician who utilized the system of Agnew and
Webb (1972) to determine sleep stages. In three records,
EEG tracings were judged inadequate for differential sleep
staging. However, a judgement of sleep vs. wake was made
for these subjects in order to form a basis for the
respiratory indices. These three EEG records v/ere
excluded from the following analyses, leaving 43 records
as a base for analysis.
Fifteen variables (time in bed, pure sleep time,
sleep efficiency index, number of stage 0 periods, time %
stage 0, sleep latency, time % stages 1, 2, 3, 4, and REM,
latency 1st REM period, mean REM period length, mean REM

66
cycle length, % slow wave sleep), reflecting multiple
aspects of the sleep of these subjects, were derived from
the sleep stage scoring. Table 3-6 presents the means ana
standard deviations for these subjects. Although subjects
spent over six hours in bed, the table indicates that not
quite 5 hours were spent asleep. The remaining variables
suggest that an increase in sleep latency, and time awake,
is responsible for the diminished sleep time, while
lighter sleep is also increased.
Table 3-6. Means and standard deviations of selected sleep
variables from 43 snoring males.
Mean
Standard
Deviation
Time in Bed
377.3
45.1
Sleep Latency
29.2
32.9
Pure Sleep Time
288.7
75.8
Sleep Efficiency Index
.76
.16
Number Stage 0 Periods
8.6
5.6
Time % Stage 0
14.3
12.7
Time % Stage 1
3.3
2.8
Time ;3 Stage 2
60.8
10.3
Time £ Stage 3
3.4
2.4
Time 53 Stage 4
7.9
6.4
Time 53 Stage REM
11.5
7.2
Latency 1st REM Period
124.7
92.2
Mean REM Period Length
14.9
9.4
RLK Cycle Length
74.9
57.7
>3 Slow Wave Sleep
12.9
8.3
Mote: see text for explanation of N.
Partial correlations, controlling for age, were
calculated between these sleep and the respiratory
variables. Out of a total of 105 correlations, only one

67
was significant at p<.05 or less: Number of desaturations
2.4% vs REM latency (r=-.3&9).
Table 3-7 presents the sleep variable means for the
subjects grouped by level of apnea/hypopnea. The group
with high apnea/hypopnea appears to spend more time awake,
less time asleep, and achieve less slow v/ave sleep than
the others. However, none of these differences proved
significant. In fact, only one out of 15 oneway ANOVAs
achieved significance: Time in bed, F(2,40)=0.4, p=NS;
Sleep latency, F(2,40)=1.7> p=NS; Pure sleep time,
F(2,40)=0.3. P=NS; Sleep efficiency index, F(2,40)=0.2,
p=NS; Number of stage C periods, F'(2,40) =0.5, p=NS; Time
% stage 0, F(2,40)=1.3, p=NS; Time % stage 1,
F(2,40)=1.3, p=«S; Time % stage 2, F(2,40)=0.04, p=NS;
Time % stage 3, F(2,40)=3.64, p<.C4; Time % stage 4,
F(2,40)=1.32, p=NS; Time ¡5 stage REM, F(2,40)=1.5,
p=NS: Latency 1st REM, F(2,40)=0.S, p=NS; Kean REM
period length, F(2,40)=0.3> p=NS: Mean REM cycle length,
F(2,40)=0.9, p=NS; % slow wave sleep, F(2,40)=2.4,
p=NS. Scheffe followup procedures for time % stage 3
revealed no significant between group comparisons.
To summarize, EEC sleep measures indicatec lighter
sleep then normal in this sample, with increases in
awakenings and light slow wave sleep observed. Neither
between group comparisons nor correlational procedures
revealed significant interactions between sleep and
nocturnal respiratory variables.

Table 3-7. Means and standard deviations 01 selected sleep variables from 43
snoring males grouped by level oi' apnea/hypopnea.
Ho Apnea/
Low Apnea/
High Apnea/
Hypopnea
Hypopnea
Hypopnea
Time in Bed
376.1 (45.3)
382.2
(47.4)
361.2 (38.0)
Sleep Latency
20.2 (12.7)
38.5
(43-5)
20.4 (18.7)
Pure Sleep Time
292.7 (73.5)
290.9
(01.5)
265-4 (69.5)
Sleep Efficiency Index
.77 (.17)
.76
(.16)
.73 (.14)
Number Stage 0 Periods
9.5 (7.6)
7.7
(3-2)
9.2 (5.5)
Time % Stage 0
15.8 (12.9)
11 .6
(12.4)
20.9 (12.6)
Time % Stage 1
4.1 (3.7)
2.6
(1.8)
3.5 (2.0)
Time % Stage 2
61.3 (8.0)
60.3
(12.9)
61.3 (1.3)
Time % Stage 3
2.0 (1.9)
4.3
(2.5)
1.5 (2.3)
Time % Stage 4
8.5 (7.2)
0.5
(5.6)
3.6 (5.1)
Time % Stage REM
9.7 (6.0)
13.3
(7.8)
9.1 (7.6)
Latency 1st REM Period
136.5(106)
107.5
(71.5)
156.4(125.4)
Mean REM Period Length
13.9 (0.0)
16.1
(9.8)
13.5 (11.1)
REM Cycle Length
80.0 (56.1)
78.7
(61.7)
41.6 (42.7)
% Slow Wave Sleep
12.8 (8.0)
14.6
(7.4)
5.9 (0.1)
Note: see text for explanation of N.

69
Daytime Sleepiness Data
Includeo among the measures of daytime sleepiness
were the mean of the several Stanford Sleepiness Scale
ratings, and two variables calculated from the two evening
naps: mean sleep latency and total minutes asleep during
the naps. These variables were correlated with the
respiratory variables, resulting in a Pearson correlation
matrix. From these 21 correlations, two emerged signi¬
ficant at p<.05 or less, both involving the mean Stanford
Sleepiness Scale ratings. Mean SSS ratings were corre¬
lated with apnea index (r=.297), and apnea + hypopnea
index (r=.283).
Table 3-8 presents the means and standard deviations
for the subjects grouped by level of apnea/hypopnea. One¬
way ANOVA procedures were non-significant for all these
variables [SSS: F(2,38)=2.4, p=NS; Mean sleep latency:
F(2,35)=0.4, p=NS; Total minutes asleep: F(2,35)=0.3,
p=NS].
Table 3-8. Means and standard deviations for daytime
sleepiness variables in 46 snoring males grouped
by level of apnea/hypopnea.
Stanford Sleepiness
Scale
Mean Nap Latency
Total Minutes of
Sleep in Naps
No Apnea/
Hypopnea
3.2 (.7)
14.5 (5.7)
11.3(11.9)
Low Apnea/
Hypopnea
3.6 (.e)
15.5 (5-3)
9.6(10.8)
High Apnea/
Hypopnea
4.0 (.7)
13.4(3.4)
13.5(6.6)

70
In summary, correlational procedures suggested
relationships between nocturnal respiratory variables and
subjective sleepiness, while no reliable between group
difference were noted on the nap or SSS variables.
Subjective Sleep Assessment
Two assessments of subjective sleep parameters were
completed, including the sleep questionnaire and sleep
log. Several variables were selected from the sleep
questionnaire and included: usual hours of sleep,
judgement of whether usual sleep was adequate, number of
naps per week, hours napping per week, number of nightly
wakenings, minutes of nightly wakenings, usual sleep
latency, depth of sleep, and daytime sleepiness. These
variables were correlated with the respiratory variables,
for a Pearson matrix of 63 correlations. Table 3-9
presents the signficant correlations which emerged from
this matrix. Several respiratory variables were
associated with napping variables, while mean low
saturation and number of desaturations _>_4% were associated
with reported minutes of nightly wakenings.
Table 3-10 presents the means and standard deviations
from the subjects grouped by level of apnea/hypopnea.
Oneway AHOVA procedures revealed no significant betv'een
group differences on the selected sleep questionnaire
variables [Usual hours sleep, F(2,45)=0.6, p=NS; Adequate
sleep, F(2,43)=0.3> p=HS; Number of naps, F(2,43)=2.9,
p=NS; Hours napping, F(2,43)=0.8, p=NS; Number of

71
Table 3-9- Significant (p<.05) Pearson correlations between
nocturnal respiratory and sleep questionnaire
variables in 46 snoring males.
Minutes
Number
Hours
Nightly
of ¡Japs
Napping
Awakening
Apnea Index
Apnea+Hypopnea Index
.249
Seconds in Events
Mean High Saturation
Mean Low Saturation
.331
â–  307
Lumber of Desaturations _>4^
Lumber of Desaturations >10£
.283
.403
.378
-.291
Table 3-10. Means and standard deviations of sleep
questionnaire variables derived from 46 snoring
males grouped by level of apnea/hypopnea.
No Apnea/
Low Apnea/
High Apnea,
Hvpopnea
Hvpopnea
Hvpopnea
Mean h. Sleep
7-4 (.9)
7.0(1.6)
7.5(1.0)
Adequate Amount?
1.9 (-4)
1.8 (.4)
2.0(0)
Lumber of Naps/Week
1.6(1.5)
2.3(2.6)
4.3(2.2)
hours Napping/Week
LumLer of Nightly
2.6(1.8)
3.5(2.7)
4.0(2.7)
Wakenings
Minutes Nightly
1.1 (.9)
1.5(1.3)
1.6 (.4)
Wakenings
1.9(1.1)
1.5(1.0)
1.5 (.8)
Sleep Latency
2.0 (.9)
2.0 (.6)
2.0 (.6)
Depth of Sleep
2.4 (.8)
2.1 (.9)
3.0(C)
Daytime Sleepiness
1.6 (.6)
1.9 (.4)
1.8 (.4)

72
awakenings, F(2,43)=0.9, p=NS: Minutes of awakening,
F(2,43)=0.8, p=NS; Sleep latency, F(2,43)=0.01, p=KS;
Depth of sleep, F(2',43)=2.6, p=NS; Daytime sleepiness,
F(2,43)=1.8, p=NS].
Means were calculated from the 7 day sleep logs, and
the following variables were selected for analysis:
reports of daytime sleepiness, number of naps, minutes
napping, total bedtime, sleep latency, number of
awakenings, minutes of awakenings. These variables were
correlated with the respiratory indices to form a Pearson
correlation matrix totalling 49 correlations. Table 3-11
presents the 6 significant (p<.05) correlations which
emerged. Again, number of desaturations were associated
with reports of napping, while mean high saturation was
associated with total bedtime, sleep latency, and minutes
awake, and mean low saturation was associated with minutes
awake as well.
Table 3-12 presents the means and standard deviations
for the subjects divided into levels of apnea/hypopnea.
Oneway ANCVA procedures failed to reveal significant
between group differences [Daytime sleepiness,
F(2,35)=0.6, p=NS; Number of naps, F(2,35)=0.3, p=NS;
Minutes of napping, F(2,35)=1-3, p=NS; Total bedtime,
F(2,35)=0.7, p=NS; Sleep latency, F(2.35)=0.2, p=NS;
Number of awakenings, F(2,35)=0.2, p=K'S; Minutes of
wakenings, F(2,35)=.0001, p=NS],

75
Table 3-11. Significant (p<.05) Pearson correlations between
nocturnal respiratory and sleep log variables in
46 snoring males.
Number
Minutes
of
Total
Sleep
Nightly
Naps
Bedtime
Latency
Wakenings
Apnea Index
Apnea+Hypopnea Index
Seconds in Events
Mean High Saturation
.237
.250
• 359
Mean Low Saturation
Lumber of
.265
Desaturations >J[%
Number of
.240
Desaturations _>10^
.240
Table 3-12. Means and
variables
by level
standard deviations of sleep log
derived from 46 snoring males grouped
of apnea/hypopnea.
No Apnea/
Low Apnea/
High Apnea/
hvbopnea
Hvoopnea
Hvoopnea
Daytime Sleepiness
1.9 (.5)
1.9(1 .1)
2.4 (.3)
Number of Naps
0.2 (.2)
.3 (.4)
.3 (.3)
Minutes Mapping
0.0(0)
.2 (.5)
0 (0)
Total Bed Time
6.5(3.1)
5.8(3.4)
7.5 (.7)
Sleep Latency
10.5(8.1)
11.1(9.8)
8.4(4.7)
Number of Wakenings
1.3(0.8)
1.3 (.9)
1.3 (.2)
Minutes Waking
2.0(1.4)
1.7(1.1)
1.7 (.3)

74
In summary, no between group differences were noted
for either sleep questionnaire or log variables when
subjects were grouped by level of apnea/hypopnea.
Correlational procedures suggested relationships between
several respiratory variables and reports of napping and
wakening after sleep onset on the sleep questionnaire.
These findings were cross validated in the correlational
analysis of sleep log data, in which reports of napping
and wakening after sleep onset were related to respiratory
variables, as were total bedtime and sleep latency.
Nocturnal Respiratory and Neuropsychological Variables
Scores derived from the cognitive/neuropsychological
battery included WAIS Performance IQ, WAIS Verbal IQ,
Wechsler Memory Quotient, delayed recall of logical
stories, delayed recall of visual reproductions, delayed
recall of ftey complex figure, digit span, Hooper score,
Wisconsin card sort score, finger tapping right and left,
and Verbal fluency.
As a preliminary analysis, these scores were compared
with selected respiratory and demographic variables in a
multivariate regression procedure. In this type of
analysis, a set of predictor scores is used to
simultaneously predict a set of criterion variables while
controlling for intercorrelations amongst the measures.
An overall test of significance, such as the Hottelings
T¿, establishes the probability of obtaining the observed
relationships from chance alone. Upon obtaining a

75
significant overall p value, univariate regressions are
used to individually predict criterion variables with the
multiple predictors. Contribution of individual
predictors may also be examined in these univariate
regressions.
The results from the multivariate regression of the
above noted cognitive scores and selected nocturnal
respiratory variables (Apnea index, Apnea + hypopnea
index, Mean high saturation, Number of desaturations >4%,
Age, Veight:Height ratio) are presented in Table 3-13- An
overall Hotellings T2 score was significant (p<.003),
allowing examination of the univariate regressions.
Criterion variables which were significantly predicted by
the respiratory and aemographic variables included WAIS
PIO, WMS MQ, delayed recall of Rey figure and visual
reproductions, Hooper, and Wisconsin card sort. These
relationships reflected negative relationship between
overnight respiratory indices and neuropsychological
scores. Respiratory indices formed significant components
of the predictor combinations for all these variables save
Hooper score, suggesting relationships between nocturnal
respiratory indices and non-verbal intelligence, verbal
and visual memory, and ability to shift sets while problem
solving.
Because age is correlated with nocturnal respiratory
parameters in past reports, and also with neuropsycho¬
logical scores in other samples, the various overnight

76
Table 3-13. Multivariate regression of demographic and
nocturnal respiratory variables on cognitive
scores in 46 snoring males.
Predictor Variables
Age
Weight:Height Ratio
Apnea Index
Apnea+Hypopnea Index
Mean High Saturation
Number of Desaturations >4%
Criterion Variables
V/AIS Verbal IQ
V/AIS Performance IQ
Wechsler Memory Scale Memory
Quotient
Delayed Recall
Delayed Recall
Reproduction
Delayed Recall
Digit Span
Hooper Test
Finger Tapping
Finger Tapping
Verbal Fluency
Wisconsin Card
Logical Stories
Visual
Rey Figure
Left Hand
Right Hand
Sort
Multivariate test of significance
fiotellings T^ = 4-841 (p<.003)
Univariate Regression with
(6.38) d.f.
Variable
Multiple R
F.
Sig.
VAIS Verbal IQ
. 3646
.971
.458
WAIS Performance 10
.5808
3.22
.012
Wechsler Memory Scale
Memory Quotient
.5345
2.53
.037
Delayed Recall
Logical Stories
.4713
1.80
.123
Delayed Recall
Visual Reproductions
.6450
4.51
.002
Delayed Recall Rey Figure
. 6461
4.54
.001
Digit Span
.2432
.39
.875
Hooper Test
.6786
5.40
.000
Finger Tapping Right Hand
.3907
1.14
.358
Finger Tapping Left Hand
.4270
1.41
.235
Verbal Fluency
.4441
1.55
.187
Wisconsin Card Sort
.5143
2.27
.050
Breakdown of significant.lv
predicted criterion regressions
Criterion Predictor
Beta
Sig.
V/AIS Performance Age
.1240
.382
10 Weight
:Height Ratio
.0170
.914
Apnea
Index
-.2636
.059
Apnea+
Hypopnea Index
-.1553
.344
Mean High Saturation
.0920
.533
Number
of
Desa
â– durations >4?/
-.3978
.024

77
Table p-15-continued.
Criterion
Predictor
Beta
Sip.
Wechsler Memory
Age
.0272
.852
Scale Memory
Weight:Height Ratio
.0996
.546
Quotient
Apnea Index
-.3515
.024
Apnea+Hypopnea Index
-.3421
.049
Mean High Saturation
Number of
.2384
.124
Desaturations >A%
-.5037
.007
Delayed Recall
Age
-.3254
.018
Visual
Weight:Height Ratio
-.1214
.416
Reproductions
Apnea Index
-.4037
.003
Apnea+Hypopnea Index
-.2026
.191
Mean High Saturation
Number of
.1368
.325
Desaturations _>4%
-.2888
.078
Delayed Recall
Age
-.3550
.010
Rey Figure
Veight:Height Ratio
.1795
.231
Apnea Index
-.2206
.091
Apnea+Hypopnea Index
-.0572
• 709
Mean High Saturation
Number of
.3191
.025
Desaturations 2.4%
-.3226
.049
Hooper Test
Age
-.5658
.000
Weight:Height Ratio
.1202
.403
Apnea Index
-.1428
.251
Apnea+Hypopnea Index
.1476
.318
Mean High Saturation
Number of
.1072
.421
Desaturations 2.4%
-.2042
.190
V.'isconsin Card
Age
-.3409
.027
Sort
Weight:Height Ratio
.0236
.867
Apnea Index
-.3205
.031
Apnea+Hypopnea Index
.1029
.550
Mean High Saturation
Number of
.2383
.130
Desaturations 2.4%
-.0801
.657

76
scores were entered into a partial correlation matrix,
controlling for age, with the cognitive scores. Table
3-14 presents the resulting correlations significant at
p<.05 or less. The table incicates that every cognitive
test but digit span and Hooper correlated with at least
one overnight respiratory index. Of additional impact is
the observation that in every case of a significant
correlation, deteriorating respiratory indices are
associated with worsening neuropsychological scores.
Respiratory indices with particularly heavy loadings on
neuropsychological scores included those quantifying
apneas, (Apnea index, Apnea + hypopnea index) as well as
those more directly representing deepening hypoxia (Mean
low saturation and Humber of desaturations >_H%).
Subjects were grouped by level of nocturnal
respiratory disorder (no apneas/hypopneas, subclinical
level of apnea/hypopnea, and clinically significant level
of apnea/hypopnea). Means and standard deviations for the
neuropsychological scores from these groups are presented
in Table 5-15. The table suggests a general deterioration
of cognitive scores as level of nocturnal respiratory
distress increases. Oneway AHOVA procedures were applied
to the neuropsychological scores from each test. Signifi¬
cant between group differences emerged for V/AIS PIQ
LF(2,43)=6.9, p<.002] delayed recall of logical stores
[F[2,43)=4.0, p<.03] delayed recall of visual reproduc¬
tions [F(2,43)=3.4, p<.04], delayed recall oí Rey complex

Table 3-14.
Significant (p<.05) partial correlations, controlling for age, between
nocturnal respiratory and cognitive variables in 46 snoring males.
Apnea/
Mean
Humber
of
Apnea
llypopnea
Fvent
Saturations
Desatura
tions
Index
Index
Time
High
Low
24S
>10>,
WAIS Verbal IQ
.274
-.321
WAIS Perforinance IQ
Vechsler Memory Scale
-.311
-.414*
-.421*
.521*
-.461*
.302
Memory Quotient
Delayed Recall Logical
-.290
.251
-.311
Stories
-.345
-.260
Delayed Recall Visual
Reproductions
-.416
.260
-. 276
Digit Span
Delayed Recall Rev Figure
-.204*
-.305
-.437*
.349*
.489*
-.337
.369*
Hooper Test
V.'isconsin Card Sort
Fing;er Tapping Right Hand
-.336
-.259
.252
.332
Finger Tapping Left Hand
-.258
.265
.274
Verbal Fluency
-.364*
-. 329
.365*
-. 329
p<.01.

80
Table 3-15. Keans and standard deviations of cognitive
variables in 46 snoring males groupec by level
of apnea/hypopnea.
Variable
No Apnea/
Hypopnea
Low Apnea/
Hvpopnea
High Apnea/
Hvpopnea
WAIS Full IQ
124.9(11.2)
118.3(13.9)
111.5 (9-6)
VAIS Verbal IQ
123.7(12)
118.5(14.5)
114.3(10.4)
VAIS Performance IQ
122.8(11.5)
114.5(14.4)
100.1(12.5)a
V/eschler Memory
Scale Memory
Quotient
124.7(14.3)
117.1(19.0)
107.8(18.7)
Delayed Recall
Logical Stories
9-9 (2.4)
8.6 (3.2)
5.8 (4.0)a
Delayed Recall
Visual
Reproductions
10.5 (3.7)
10.4 (3.4)
f.1 (4.7)
Digit Span
6.8 (0.8)
6.8 (1.2)
6.6 (.8)
Delayed Recall
Rey Figure
25.6 (5.4)
23.2 (6.0)
15-1 (6.4)s,b
Hooper Test
25.2 (2.6)
25.5 (4.8)
22.8 (4.1)
Visconsin Card Sort
4.1 (1.2)
3.4 (1.5)
2.e(1.6)
Finger Tapping
Right Hand
62.6 (8.5)
57.5(10.4)
56.1 (8.3)
Finger Tapping
Left Hand
59.1 (9.0)
54.0 (9.7)
54.8(13.1)
Verbal Fluency
14.6 (4.5)
13.1 (4.2)
5.3 (2.6)®
a
b
Significantly different from no sleep apnea/hypopnea.
Significantly different from low sleep apnea/hypopnea.

81
figure [F[2,43)=6.5> p<.003], and Verbal fluency,
[F[2,43) =3-6, p<.04]. Followup comparisons (Scheffe)
showed the group with high apnea/hypopnea to be
significantly impaired relative to those with no sleep
apnea/hypopnea on V/AIS PIQ, delayed recall of logical
stories and Rey complex figure as well as verbal
fluency. Additionally, the high apnea/hypopnea group was
impaired relative to those with subclinical apnea/hypopnea
on aelayea recall of the Rey complex figure. These
findings are bolstered by the lack of significant between
group differences on age, weight, or education (see Table
3-15), all of which are potential alternative explanations
of the between group differences.
As a check, an analysis was run to evaluate the
relative contribution of nocturnal respiratory variables
and sleep variables reflecting fragmentation in predicting
cognitive/neuropsychological scores. A multivariate
regression of these scores was carried out. Demographic
and respiratory scores included age, apnea index, apnea +
hypopnea index, and number of desaturations >4%. A subset
of the earlier noted sleep variables was chosen to reflect
measures thought to be most responsive to sleep
fragmentation. These included sleep efficiency, number of
stage 0 periods, time % of stage 0, time % of stage 2, ana
number of stage changes. These respiratory and sleep
parameters were used to predict the 5 cognitive scores
which were earlier shown to be predicted by nocturnal

82
respiratory scores: V/AIS PIQ, V.'echsler Memory Quotient,
delayed recall of visual reproductions and Rey figure, and
Wisconsin card sort. Table 3-16 presents the data from
this analysis. The Hotellings T2 was significant,
(p<.001) with delayed recall of visual reproductions and
Rey figure significantly predicted by the predictor
scores. Inspection of the individual regressions for
these two measures discloses that age, apnea index, and
number of desaturations _>4£ all formed significant
components of the prediction of delayed recall of visual
reproductions, while sleep efficiency was the sole
significant predictor of the Rey figure. These data seem
supportive of an independent contribution of both
nocturnal respiratory variables and sleep variables at a
limited level, to the prediction of cognitive/neuro¬
psychological scores.
In summary, various statistical procedures indicate
relationships between nocturnal respiratory measures and
non-verbal intelligence, verbal and non-verbal memory,
expressive verbal fluency, and ability to shift cognitive
set while problem solving. Division of subjects by level
of nocturnal respiratory distress indicated that subjects
with high amounts of apnea/hypopnea were impaired relative
to controls on measures of non-verbal intelligence, verbal
and non-verbal memory, and expressive verbal fluency, as
well as being impaired relative to subjects with
subclinical levels of apnea/hypopnea on measures of

83
Table 3-16. Multivariate regression of sleep and nocturnal
respiratory on cognitive variables in 43
snoring males.
Predictor Variables
Criterion Variables
Age
Apr.ea Index
Apnea+Hypopnea Index
Mumber of Desaturations _>4/»
Sleep Efficiency
Kumber Stage 0 Periods
Time 5» Stage 0
Time % Stage 2
Dumber of Stage Changes
WAIS Performance IQ
Wechsler Memory Scale Memory
Quotient
Delayed Recall Visual
Reproduction
Delayed Recall Rey Figure
Verbal Fluency
Wisconsin Card Sort
Multivariate test of significance
Hotellings = 4.063 p<.001
Univariate Regression with (9.33) d.f.
Variable
Multiple R
F.
Sig.
WAIS Performance IQ
Wechsler Memory Scale
fSl5
2.17
.051
Memory Quotient
Delayed Recall Visual
Ul
cr
O'.
1.73
.121
Reproductions
Delayed Recall Rey
.655
2.91
.012
Figure
.668
2.96
.011
Verbal Fluency
.578
.85
.096
Wisconsin Card Sort
.453
.95
.495
Breakdown of significantly predicted criterion regressions
Criterion
Predictor
Beta
Sig.
Delayed Recall
Age
-.3596
.020
Visual
Apnea Index
-.3218
.021
Reproduction
Apnea+Hypopnea Index
Dumber of
.2329
.218
Desaturations >455
-.3778
.041
Sleep Efficiency
.2303
.408
Number of Stage 0 Periods
.3260
.209
Time % Stage 0
-.1136
.731
Time 55 Stage 2
.0005
• 998
Number of Stage Changes
-.2235
.391

Table ,v-1 O-continuec.
Criterion
Predicxor
Beta
Sip.
Delayed Recall
Age
-.2623
.074
Key Figure
Apnea Index
-.1315
.259
Apnea+Hypopnea Index
Number of
-.3266
.085
Desaturations >4%
-.0795
.657
Sleep Efficiency
.6632
.016
number of Stage 0 Periods
.2467
.334
Time Í Stage 0
.4559
.172
Time % Stage 2
-.0465
.602
Number of Stage Changes
-.5054
.057

85
non-verbal memory. Addition of variables reflecting sleep
fragmentation to the analysis indicated that both
nocturnal respiratory and sleep variables make significant
and independent contributions to prediction of
cognitive/neuropsychological scores.

CHAPTER FOUR
DISCUSSION’
Demographics and Incidence of
Nocturnal Respiratory Disorder
The present sample of 46 snoring males was selected
as a group likely to display a wide range of sleep
disordered breathing. These subjects, with a mean age of
50 years and a mean weight of 190 pounds, reflect a mostly
middle-aged to older male population which is moderately
obese. The demographic characteristics of this sample are
similar to those of the "typical" sleep apnea syndrome
patient (Guilleminault et al., 1978), although the
requirement of general good health insured that subjects
with a classic SAS viere excluded. This sample may be
considered representative of subjects predisposed to sleep
disordered breathing, although at no more than a
subclinical level. Therefore it seems valid to consider
various risk factors as they relate to nocturnal
respiration.
In the present sample, 62% experienced at least one
apnea or hypopnea, while 13% suffered higher levels of
apnea/hypopnea. Only two other reports have described
samples selected for snoring or respiratory complaints.
Ancoli-Israel et al. (1981) examined 24 subjects with
complaints of respiratory difficulties, daytime
86

87
sleepiness, and muscular events (nocturnal myoclonus).
Further, these subjects had a mean age of almost 70 years,
indicating significant sampling and age differences from
the present study. Ninety-three per cent of these
subjects had at least one event, while 25?» had high
numbers of apnea/hypopnea. These rates are much higher
than those found in the present sample. This is probably
attributable to the sampling differences noted above.
However, a recent report by Miles and Simmons (1984)
described apneas in a series of 190 patients referred for
complaints of heavy snoring. In this sample, 73?» had at
least one event, while 9% had severe levels of apnea.
These data are quite comparable to the prevalences
observed in the present sample, 625» for at least one event
and 13% for high levels of apnea/hypopnea, and lend some
confidence to the present finding.
Thus the overall frequency of sleep disordered
breathing in middle-aged subjects with heavy snoring
appears to be bexween 62 and 73% for at least one event,
and between 9 and 13% for more severe levels of apnea.
The relatively close concordance of the two reports on
heavy snoring subjects is reassuring.
Factors which were found to be related to the
occurrence of sleep apneas in the previous review included
male sex, complaints of snoring, increasing age, and
obesity. The first two factors could not be examined in
the present sample because of their use as selection

88
criteria. A wide range of age was achieved in this
sample, but neither non-parametric nor correlational
procedures revealed significant relationships between
increasing age and sleep disordered breathing. Although
several previous reports have noted relationships between
increasing age and sleep disordered breathing, typically
this relationship is moderate for apnea (Block et al.,
1979; Carskadon et al., 1980; Bixler et al., 1982).
Additionally, no other correlational data are available in
subjects selected for sleep and breathing complaints. It
may be speculated that the effect of snoring masks the
effect of age in this group. Alternatively, some quirk of
sampling may have generated this effect. Correlational
procedures, however, revealed substantial relationships
between weight and sleep disordered breathing. This
finding is consistent with several past reports. To
summarize, the present data support relationships between
increasing weight and nocturnal events, while
relationships between age and sleep disordered breathing
are not found in this sample of heavy snoring males,
perhaps because the effect of age is masked by the effect
of snoring.
The temporal characteristics of the apneic events
assessed by two hour interval suggested an increase in
events between 3 and 5 am, although differences in sleep
onset time and wakeup time render any interpretation of

89
these data somewhat tenuous. This finding needs
replication before any conclusions may be drawn.
Hinety-eight per cent of apneic events occurred in
light slow wave and REM sleep, an observation which is not
surprising given that subjects spent nearly BOSS of their
sleep in these stages. The preponderance of events in
these stages of sleep is consistent with the bulk of past
reports in this area (e.g, Bixler et al., 1982; Block et
al., 1979; Krieger et al., 1983).
Vihen subjects were divided into groups by level of
apnea/hypopnea, significant differences emerged on oxygen
desaturation variables. Specifically, comparisons
indicated that the group with a high level of
apnea/hypopnea events experienced deeper desaturations
than the remaining two groups. This difference occurred
despite a lack of differences on age and weight among the
three groups, and appears to implicate the respiratory
events in the genesis of the desaturations. This is
consistent with an earlier report by McGinty et al. (1982)
who demonstrated that subjects with significant levels of
sleep disordered breathing were impaired on oxygen
desaturation variables, relative to subjects without
respiratory events. Subjects with heavy snoring and
multiple nocturnal apneic respiratory events appear to be
at risk for nocturnal oxygen desaturation.
In summary, the present sample of 46 snoring males
appears representative of those predisposed to experience

90
sleep disordered breathing (middle-aged, snoring males).
A prevalence of 62% was observed for occurrence oi at
least one event, while 13/» had high numbers of
apnea/hypopnea, data which are consistent with an earlier
report drawn from heavy snoring males. Most events
occurred in light slow wave or REM sleep. Mo effect of
age on level of sleep disordered breathing was observed,
although increasing weight was linked to nocturnal
respiratory events. When the sample was divided by level
of apnea/hypopnea, subjects with high numbers of
apnea/hypopnea suffered significantly exacerbated oxygen
desaturation, although no significant differences were
noted for weight or age among the groups.
nocturnal Respiratory and Health Variables
The finding oi only very limited relationships
between nocturnal respiratory events and health variables
must be tempered by an acknowledgement of possible bias
against such findings in this sample. Specifically, the
sample was selected to be normal and healthy, a criterion
which undoubtably restricted the range of possible health
pathology, and hence probably attenuated the correlational
findings. Additionally, the self-report health checklist,
the Cornell Medical Index, cannot be regarded as a
powerful measure of health pathology. In a very real
sense then, the odds were against finding reliable
relationships between nocturnal respiratory and health
variables.

91
In light of these caveats, it is of interest that two
measures of desaturation, mean lowr saturation, and number
of desaturations _>45» were associated with increasing blood
pressure readings. While a report by Guilleminault et al.
(1978) noted an increasing prevalence of hypertension in
subjects diagnosed with sleep apnea syndrome, a
relationship between desaturation and hypertension has not
been previously reported. This finding was observed in
spite of the fact that subjects prescribed hypertensives
were not medication free for the pressure reading.
Additionally, a check of self reports of diagnosed
hypertension showed six subjects in the no event group as
having diagnosed hypertension, as opposed to six in the
low apnea/hypopnea group, and one in the high
apnea/hypopnea group, data which did not link
apnea/hypopnea, per se, with hypertension in these
subjects. The isolated finding of a relationship between
mean high saturation and overall number of symptoms
endorsed on the CMI is more puzzling and does not follow
any logical analysis.
In summary, the present study's assessment of health
variables must be regarded as somewhat limited. In spite
of this, an association between two measures of desatura¬
tion and a measure of blood presure was noted. It is
speculated that blood presure may be more responsive to
nocturnal oxygen saturation than to the apneic events
alone.

92
Nocturnal Respiratory and Sleep/VJake Variables
EEC Data
The overnight sleep in the laboratory for these
snoring subjects was noteworthy i'or a general shift
towards lighter sleep than usual. A longer sleep latency,
more awakenings, more light sleep, and less slow wave and
REM sleep than normal are apparent. These findings
parallel the report by Block et al. (1979) who found
shifts to lighter sleep and more awakenings than in a
sample of normal subjects. These changes were said to
result from a first night in a strange bed and the
discomfort of the multiple clips and sensors used to
monitor respiration. Recording montages were identical
for the present subjects, and these two factors
undoubtably influenced the sleep of the present subjects
as well.
An intriguing observation Vías a lack of covariation
between respiratory and EEG sleep variables derived from
the recording night. This finding is counterintuitive, as
one would expect arousals from the respiratory events to
influence the sleep characteristics of subjects with
events. However, correlational procedures failed to
support this notion. It may be that a ceiling effect was
exerted on sleep measures sensitive to disturbance, one
that resulted from the first night effect compounded by
the discomfort of the recording procedure. The already
elevated disturbance sensitive sleep variables might not

93
have enough free variance to respond to the respiratory
events of the night. Alternatively, the EEG scoring
system, which averages across 1 minute periods, might miss
the short arousals and fragmentation resulting from
events. Whatever the reason, the lack of covariation of
sleep and respiratory measures is puzzling in light of the
report of Guilleminault et al. (1978) who described
arousals as a frequent concommitant of nocturnal
respiratory events, although Bixler et al. (1982) failed
to replicate this in a group of normal subjects.
Between group differences on the sleep variables were
also absent when subjects grouped by level of respiratory
distress were compared on sleep characteristics. Again it
may be speculated that the effect of an exacerbated first
night might obscure possible differences.
In summary, these snoring subjects appeared to
achieve a night of sleep comparable to tnose observed in
similar studies. Longer sleep latency, more awakenings,
and lighter sleep were observed. Significantly, these
changes characterize a first night effect (Agnew, Webb,
and Williams, 1966) of lighter sleep in a strange bed, and
may have been increased by the discomfort of the recording
procedures. Less than a chance number of correlations
were observed between nocturnal respiratory and sleep
parameters. This surprising result is attributed to an
exacerbated first night effect which probably functioned
to obscure possible respiratory disturbed sleep parameter

94
correlations in this sample. Additionally, an EEG scoring
system which averages across one minute periods might
obscure brief sleep fragmentation from nocturnal events.
Although the lack of correlation is puzzling, Bixler et
al. (1962) reported similar findings. In their normal
sample, few apnea induced arousals were noted. Multiple
night recordings of all these parameters appear necessary
to clarify the possible interaction of respiration and
sleep fragmentation in a first night effect.
Daytime Sleepiness, Sleep Questionnaire, and
Sleep Log Data
The three assessments of subjective aspects of sleep
and sleepiness will be discussed together, as the pattern
of findings lend themselves to this organization. The
core finding here is a relationship between nocturnal
respiratory variables and sleepiness variables manifesting
itself in two forms. Daytime sleepiness, reflected in the
mean SSS ratings and self report of napping on both the
sleep questionnaire and logs, was positively related to
various indices of nocturnal events (apnea index, apnea +
hypopnea index, seconds in events, and number of
desaturations). A second aspect of this relationship
involved relationships between decreased saturation and
reports of decreased sleep latency and nightly wakenings
on the sleep logs and questionnaires.
These seemingly disparate findings may be interpreted
as the sequelae of sleep oisruption resulting from

95
disturbed nocturnal breathing. In the first case,
increasing respiratory disturbance causes sleep
deprivation, manifested in increased subjective sleepiness
(bSS ratings) and compensatory sleeping (increased
napping). The other side of this coin is a sleep pressure
which causes disturbed subjects to fall asleep quickly
(lowered sleep latency), and subjectively fail to report
nightly wakenings, while subjects with intact respiration
and hence less sleep pressure experience a longer sleep
latency and awareness of nightly awakenings. Admittedly
this is speculative, and two immediate problems with this
formulation are evident. First, the lack of correlation
between FEG sleep and respiration of the experimental
night is problematic, as well as the lack of relationships
between respiration and the evening "naps." However, it
has already been speculated that an exaggerated first
night effect obscured correlations between sleep and
respiration, and it is now speculated that the two naps,
at 7 and 9 pm, missed the afternoon trough of sleepiness
which might have better reflected sleepiness.
The interpretation of increased sleepiness from
nocturnal respiratory disturbance is useful in explaining
the pattern of findings observed in this study, although
some problems with this formulation were noted.
To summarize the main aspects of the discussion of
sleep/wake variables, it is noted that these snoring
subjects achieved a night of sleep comparable to those

reported in similar studies. A first night effect pro¬
bably caused increased sleep latency, wakenings, and
lighter sleep, a result which may have functioned to
obscure possible relationships between sleep and respira¬
tory variables, which were not observed. A pattern of
correlations between increased respiratory disturbance and
sleepiness was noted, perhaps because these variables
reflected subjective reports which were largely indepen¬
dent of the experimental nights' sleep (SSS ratings made
before sleep, sleep questionnaires reflecting trait sleep
variables, and sleep logs averaged over a 7 day period
following the experimental nights' sleep). Multiple night
recordings appear necessary to sort out a possible syner¬
gistic interaction between the first night effect and the
discomfort of the recording procedures.
Nocturnal i-espira'cory snc Neurophysiological Variables
The major finding of the respiratory/cognitive
analysis was the demonstration of relationships between
nocturnal respiratory parameters and cognitive/neuro¬
psychological scores tapping non-verbal intelligence,
verbal end non-verbal memory, expressive verbal fluency,
and cognitive flexibility. Most of these relationships
remained when age, weight, and educational differences
were ruled out as plausible alternatives.
These findings extend the boundaries set by previous
work which indicated a relationship between hypoxia and
neuropsychological scores.
These previous reports

97
included Krop et al. (1973) and Krop et al. (1977), both
of which noted That continuous nocturnal oxygenation
treatment significantly improved neuropsychological scores
in normal aged and chronic obstructive lung disease
patients, respectively. The present results are also
consistent with a report by Grant et al. (1982)
demonstrating considerable neuropsychological impairment
in chronic obstructive lung disease patients, a group
known to experience hypoxia which is significantly
exacerbated during sleep (Block, 1981). Additionally,
West (19GA) found impairments in finger tapping speed,
short-term, memory, end expressive verbal fluency in a
group of climbers ascending Mt. Everest. Taken together,
the present results and past reports present a consistent
picture of relationships between both daytime and
nighttime hypoxia and neuropsychological impairments.
Although the selection requirement of general good
health ruled out the inclusion of subjects v'ith a full
blown sleep apnea syndrome, six of the present subjects
demonstrated levels of respiratory disturbance sufficient
to meet current diagnostic criteria for SAS-(A+H)I>5.
These subjects showed significant impairments, relative to
those with no sleep apnea/hypopnea, on measures tapping
non-verbal intelligcnce, delayed verbal and non-verbal
memory, and verbal fluency. These data are suggestive
that heavy snoring males with multiple apneas/hypopness
may be at risk for cognitive sequelae. If the impairments

found in these subjects are taken as minimally comparable
with subjects experiencing the classic sleep apnea
syndrome, some support for the anecdotal reports of in¬
tellectual deterioration in SAS (Cuilleminault et el.,
1978) is derived. However, both these conclusions stand
in need of replication before they may be confidently
asserted.
The mechanism underlying the link between nocturnal
respiratory and cognitive variables is a fascinating
enigma. To cate, two explanations have been proposed as
proximal causes of the changes observed in sleep apnea
syndrome. The first, or "sleep fragmentation" hypothesis
(Carskadon et al., 1980) notes the large number of
arousals which may accompany nocturnal respiratory events
and suggests that the resultant sleep fragmentation may
lead to daytime impairments, particularly in the
elderly. While intuitively appealing, there are several
difficulties with this explanation. First, the demonstra¬
tion of reliable cognitive deficits after sleep depriva¬
tion has proved surprisingly difficult (Webb, 1975).
Changes in cognitive test scores such as those used in the
present study have not been reliably observed following
sleep deprivation. A second line of evidence inconsistent
with the sleep fragmentation hypothesis is the observed
lack of correlation between sleep structure and cognitive
variables in normal subjects (Berry anc Webb, In press).
Although this covariation has been observed in subjects

99
with greater physical deterioration (Feinberg et al.,
1967), subjects selected for health do not display this
effect. A third embarrassment for the sleep fragmentation
hypothesis lies in the study of Orr et al. (1979) who
described one group of symptomatic sleep apnea syndrome
subjects and a group of asymptomatic snoring subjects that
were matched on number of nocturnal events. Presumably,
arousals resulting from events were matched in the two
groups, yet the symptomatic subjects had serious impair¬
ments relative to the asymptomatic groups.
Despite these difficulties, the present study found
that one of the five cognitive scores significantly pre¬
dicted by the nocturnal respiratory indices was also
correlated with sleep efficiency on the experimental
night. However, given the earlier observed problems with
the experimental nights' sleep, the inconsistent past
evidence, and the limited (1) relationship observed, this
finding remains inconclusive. A more ambitious study,
examining sleep and respiration across several nights, is
necessary to clarify this issue.
An alternative, although admittedly speculative,
explanation of the cognitive changes lies in the cerebral
hypoxia accompanying respiratory events. Brain dysfunc¬
tion is known to occur in mild to moderate hypoxia. While
the brain possess various homeostatic mechanisms to com¬
pensate for chronic hypoxia (hence the adaptation of long
term residents of mountainous regions) it may be that the

10C
multiple brief, acute desaturations experienced by the
subject with apnea/hypopnea fail to activate these protec¬
tive mechanisms effectively (Gibson et al., 1981).
Recently, a body of evidence has accumulated which indi¬
cates that turnover of certain neurotransmitters is drama¬
tically altered by acute hypoxia. The cholinergic synthe¬
tic pathway, which results in the formation of the neuro-
transmitter acetylcholine (ACh) appears to be particularly
sensitive to the effects of acute hypoxia (Gibson et al.,
1981). That ACh may be the neurotransmitter most affected
by hypoxia of nocturnal events is provocative, in light of
the "Cholinergic hypothesis" (Bartus et al., 1985) of the
memory and cognitive impairment often found in aging sub¬
jects. This formulation notes that cholinergic blockade
in young subjects inhibits memory, while various aspects
of the cholinergic synthetic pathway are altered in aged
subjects with memory and cognitive deficits. Alterations
in the cholinergic synthetic pathway are thought to under¬
lie the cognitive changes in both populations. It may be
speculated that the cognitive dysfunction found in the
present subjects with multiple apnea/hypopnea is also a
product of a disruption of cholinergic synthesis.
To summarize the logic of the "cerebral hypoxia hypo¬
thesis," it is first noted that subjects with nocturnal
respiratory events experience multiple, acute hypoxic
episodes. This acute hypoxia appears to evade cerebral
protective mechanisms which normally respond to chronic

1 Cl
hypoxia. Acute hypoxia has recently been shown to signi¬
ficantly alter the synthesis of ACh, a neurotransrcitter
implicated in memory and cognitive changes in young sub¬
jects experiencing cholinergic blockade, anc older sub¬
jects experiencing age related disruption of cholinergic
pathways. The hypoxic alteration of ACh turnover is
proposed as the proximal cause of cognitive/memory
impairments found in subjects with nocturnal respiratory
events.
Although the cerebral hypoxia hypothesis is highly
speculative, it does generate predictions which might be
used to test its eflicacy. For instance, subjects with
differing levels of hypoxia might be expected to exhibit
difiering levels of ACh turnover, as well as cognitive
differences. Administration of nocturnal oxygen should
improve both cognitive and cholinergic variables. Addi¬
tionally, manipulation of cholinergic pathways might
improve cognition in subjects with nocturnal respiratory
events. However, until a demonstration of this sort is
carried out, this line of reasoning remains highly specu¬
lative .
In summary, the present findings indicate clear rela¬
tionships between nocturnal respiratory and cogni¬
tive/neuropsychological variables. This result is con¬
sistent with past reports of neuropsychological
impairments in subjects experiencing hypoxia. Apparently,
heavy snarers with multiple apnea/hypopnea episodes may be

102
at risk for cognitive sequelae. These data may be taken
as limited support for the anecdotal reports of cognitive
changes in SAS patients. Two explanations of possible
unde-rlying mechanisms of the cognitive changes may be
considered. The first, the "sleep fragmentation hypo¬
thesis," suggests that the cognitive deficits result from
the multiple arousals experienced by the subject. This
notion is suspect as a plausible explanation because sleep
deprivation and sleep measures tapping sleep fragmentation
fail to intercorrelate consistently with cognitive scores
in other populations. Additionally, the extremely limited
relationship observed in the present study is less than
conclusive. It is proposed that cerebral hypoxia induces
alterations in synthesis of ACh, a neurotransmitter impli¬
cated in cognitive abilities in other populations, and
thus causes the cognitive changes. The cerebral hypoxia
theory remains speculative, although several testable
hypotheses may be generated from it. Further research
into factors affecting the sleeping brain may prove
heuristic for neuropsychologists.

CHAPTER FIVE
SUMMARY AND CONCLUSIONS
In this study, a detailed assessment of the nocturnal
respiration, health status, cognitive/neuropsychological
skills, and sleep/wake cycle of 46 snoring male subjects
was carried out. Sixty-two per cent of these subjects
experienced at least one event (apnea or hypopnea) while
1exhibited more than 5 events per hour. Most events
occurred in light slow wave or REM sleep. Obesity was
linked to increasing nocturnal events. Subjects with more
than 5 events per hour suffered significantly exacerbated
oxygen desaturation relative to the remaining subjects.
Overnight oxygen desaturation was linked to increasing
blood pressure readings. As a whole, these subjects
experienced a longer sleep latency, increased wakening
after sleep onset, and more light sleep in the lab, the
classic "first night" effect. It vas thought that this
first night effect obscured relationships between noc¬
turnal respiratory and sleep parameters, which were not
observed. However, overnight respiratory disturbances
were linked to increased sleepiness and napping. Addi¬
tionally, relationships between nocturnal events and
deteriorating scores on tests tapping non-verbal intelli¬
gence, verbal and non-verbal memory, expressive verbal
103

104
fluency, and cognitive flexibility were observed. It was
speculated that hypoxic alterations in the cholinergic
synthetic pathways underlayed these changes. Apparently,
a significant subgroup of heavy snoring males fall on the
same continuum as subjects with sleep apnea syndrome,
occupying an intermediate position between normals and
those with SAS. From these and other data, it is
concluded that
1) between 62 and 73$ of snoring males experience at
least one nocturnal respiratory event, while
between 9 and 13$ appear to experience 5 or more
events per hour;
2) most events occur in light slow wave or REM
sleep;
3) increasing obesity is linked to nocturnal events
in snoring males;
4) nocturnal events are linked to increasing blood
pressure in snoring males;
5) nocturnal events are linked to increasing sleepi¬
ness in snoring males;
6) nocturnal events are linked to deteriorating
neuropsychological scores in snoring males;
7) a subset of heavy snoring men, those with more
than 5 events per hour, suffer increased oxygen
desaturation and decreased cognitive abilities in
the areas of non-verbal intelligence, verbal and

105
non-verbal memory, expressive verbal fluency, and
cognitive flexibility;
8) the subset of heavy snoring males with more than
5 events per hour appear to occupy an inter¬
mediate position between normals and those with
sleep apnea syndrome.

APPENDIX A
SCORING PROTOCOL FOR RESPIRATORY VARIABLES •
Data from the nose and mouth thermisters, as well as
the oximeter and strain gauges, were used in scoring
respiratory events.
An apnea was scored when a cessation of airflow at the
mouth and nose lasting for 10 or more seconds was
observed. Up to 3 mm of "noise" were tolerated in an apnea,
as long as no rhythmic pattern which might represent shallow
breathing was observed. An event began at the end of the
last exhalation preceding the apnea. Duration, lowest
saturation, and per cent saturation change were noted.
A hypopnea was scored when a reduction in amplitude of
breathing at the mouth and nose of 5C% or mere was
accompanied by a 10S6 or larger desaturation. The event
began at the end of the last full exhalation. Duration,
lowest saturation, and per cent saturation change were
noted.
Each record was separately scored by two raters. Later
each record was jointly rescored and disagreements were
resolved by consensus.
A separate scoring of oxygen saturation was made by
rating each minute of the record for highest and lowest
saturation. Additionally, a separate tabulation of
106

107
desaturaticns exceeding 4 and 10£ was kept. Decause of the
relative ease of this rating, only one rater scored
saturation records.

APPENDIX E
RAW DATA TABLES

109
Table B-1. Demographic and medical variables from 46 snoring males
Subj.
ti
Age
Educa¬
tion
Height
(in)
Weight
Weigh t:
Height
Ha tio
Blood
Diastolic
Pressure
bystolic
Overa 11
score
Uorneil Medical Index
Res- Cardio
pirat.ory pulmonary
heuro-
logica 1
1
58
12
60
250
5.67
150
090
2
bO
12
o5
225
5.57
140
100
022
01
b2
02
5
52
1b
67
148
2.10
110
065
02b
01
00
01
4
68
15
70
168
2.4
140
100
007
00
01
00
5
52
12
74
245
5.51
120
085
018
05
00
01
b
50
1b
b7
155
2.51
110
060
015
00
02
02
7
51
12
69
155
1.95
150
085
00/
00
00
01
8
' 48
09
07
155
2.51
120
090
005
01
02
00
9
bO
20
bb
157
2.57
150
100
014
00
02
05
10
5b
1b
bb
255
5.45
155
090
005
00
00
02
11
49
18
70
250
5.28
120
100
050
01
02
01
12
5b
15
74
250
5.10
125
090
001.
00
01
01
15
48
15
70
187
2.67
120
070
007
00
01
01
14
58
15
75
190
2.55
120
070
005
01
00
â–  -60
15
54
14
b7
125
1.99
114
07b
006
00
00
"01
1 o
50
1b
71
1bO
2.25
150
080
014
02
04
01
17
52
1b
72
167
2.51
120
085
005
01
00
02
1b
56
1b
69
180
2.60
120
090
050
02
01
01
19
57
12
72
190
2.71
120
090
015
01
02
01
20
50
16
71
250
5.52
155
100
042
06
02
Ob
21
47
14
71
198
2.79
125
090
007
00
02
02
22
66
14
b5
155
2.04
126
Ob 0
018
02
00
CO
25
55
16
72
145
2.01
120
074
015
01
01
01
24
51
1b
72
105
2.29
115
070
010
04
00
02
25
51
15
bO
160
2.55
122
064
025
00
01
04
2b
64
18
bb
215
5.1b
145
100
u10
01
01
02
27
75
12
bO
185
2.69
140
090
026
05
00
01

-b. .b. ja. .b. .C». Vj» O. V_1 Vj. V* O. Vjí V.» fVJ ro
c v.ii».V.-Jrc -» cxC'JrviJayiiy-iC'i:!
■e*- WOUVCCCCVflvrOi-kUC1.' rwc vj: v£ 'o.
vjiroO-=>--^0'(\JO'COrv-^-c*-C'C:i\:crC'ro
c -v! cr cw a ~< CW CCTW- crca'ji-a'vOrovcrv.'vjv^-avj-, ccx-^j^ct-^
-‘SC'^cvji —k vj-! vjvjiiv-fe-vjuicxrov-rcccfc'ji
-* VJ1 O VJ1 C -E* CrV3U1C^VCvCV-N
■‘OOMCOucvO'jíca'ví
c-*c-*ocqccocccc
?8uicco^uiccccjíc
c c c c
rv: v>j
CVJ1 C O
cccccccccc
NJ-a^OJVJiCvCVj-CUi
coccccccccccoc
CANCCCCC^WC^JN
c c c c c
-> O- V.-. <0. !V
c c c c
NAIVJ J
cccccoccccccccccroc
W M C W -1 C C V- - J W -i c C w -* C 'o.
ou
1bV 2.49 129 OHO U^4
200 2.94 124 070 010
210 1.ÃœJ) 129 009 010
205 2.08 120 082 015
195 2.00 120 080 011

tr
1
2
5
4
5
6
7
8
o
1C
11
12
13
14
15
16
17
16
19
20
21
22
23
24
25
26
27
26
29
30
31
32
33
34
35
36
37
38
39
40
41
111
Daytime sleepiness variables from 46 snoring
males.
Mean Stanford
Kean Sleep
Minutes of
Sleepiness
Latency
Sleep
Rating - Haps - Haps
-
09-0
22
-
16.0
05
-
11.5
18
4.5
-
-
3.5
06.5
26
3.5
-
-
4.7
20.C
01
2.7
13.5
13
3.8
09.5
17
4.3
20.0
00
3.4
20.0
00
3.3
16.0
09
4.5 20.0 00
4.1
14.0
14
3-6
20.0
00
2.8
08.5
25
2.5
11.5
16
-
15.C
11
3.6
20.0
00
3.7
20.0
00
2.4
14.5
05
4.5
14.5
12
4.4
20.0
00
3-3
04.0
34
3.3
-
-
4.8
19.0
03
3.5
20.0
00
4.1
15.0
10
2.6
17.5
07
4.3
14.0
13
3.1
11.0
16
5.1
11.5
16
3-1
-
-
3.8
20.0
00
3.4
20.0
00
3.4
13-5
14
2.2
20.0
00
2.9
06.5
25
3.9
19.0
08
4.0
08.0
27

112
Table B-2-continued.
Subject
a
Kean Stanford
Sleepiness
Ratine
42
43
44
45
46
1.5
2.5
3.8
3.4
2.1
Mean Sleep
Latency
- Naps
06.0
03.5
Minutes of
Slee
- Map
30
35
a w

Table b-3. bleep questionnaire variables from 4b snoring males.
Subject
r,
Mean
Hours
Sleep
Enough
?
Humber
ol'
Hups
Hours
Happing
Humber
of
Wakings
Minutes
Awake
Sleep
Latency
Depth
of
Sleep
Daytime
Sleepines
1
7
2
4
4
2
0
1
3
2
2
Q
2
6
5
1
0
1
X
2
3
7
1
z
J
3
0
2
2
3
2
4
8
2
0
0
1
2
2
3
2
5
6
2
0
0
0
0
3
3
2
6
7
2
1
2
0
0
2
3
2
7
8
2
0
0
1
2
3
1
2
8
1
3
2
3
5
2
1
3
2
9
8
2
7
4
3
2
2
2
2
10
9
2
2
2
2
2
3
2
2
11
7
2
1
1
1
1
2
2
2
12
y
2
0
0
4
1
2
1
1
13
7
2
2
3
2
1
2
3
1
14
9
2
1
2
0
0
2
1
2
15
9
2
0
0
0
2
2
3
1
16
9
2
2
2
2
1
3
2
2
17
7
1
2
3
1
1
2
3
2
18
5
1
2
2
1
3
3
1
2
19
8
2
6
3
2
4
4
1
2
20
1
3
8
10
3
3
2
3
2
21
7
2
0
0
0
1
1
1
1
22
6
1
3
3
3
3
2
3
2
23
8
2
0
0
0
3
2
2
1
24
8
2
0
0
1
2
2
2
2

114
C\JCMC\JC\JC\jT-C\JC\JK^C\J^-C\JC\JC\Jr-C\IC\JC\Jv-C\JOC\J
V- ?A ;A rA .A IA fA fA IA IA IA IA PA (A IA C\J IA c\J IA O IA
CM C\j fA C\J CM CM CM C\| CM IA IA CNJ r (M tM fOr r r O r
' 'íOC\ICMC\JOT*C\|CMrrf\Jr(MvrrOí\IC\lO'Ñf
O CM CM A rA *
- IA IA CM CM '¡J-IAO r O CO 1
CCv-CMiAlAr-LACOCMrACMCM ^WfAOCMO'i^-Or
r-CMCMCMCMCMCMCMCMCMCMCM CMCMCMCMCMCMCMv-OCM
- ^ O O- L" ^ CO C^~ CO [>- O', c cc C— C— I>- c— c
co iv o co
IA A- CO 01 o v CM IA -3- LA O t>- CO CT\ O v- CM A A MO
CM CM CM CM CM iA 'A fA fC\ IA IA ^ N"\ rA ^ ^ M' ^ ^ ^

Table B-4. Seven clay sleep log means from 46 snoring males.
lubject
a
Daytime
Sleepiness
Dumber
of
Daps
Sleep
Latency
Dumber
of
Wakenings
Minutes
Waking
Minutes
Mapping
Tota 1
Bedtime
1
2
0
3
1
2
0
7
p
2
1
5
0
1
0
9
3
2
0
15
4
2
0
6
4
2
0
7
2
3
0
e
5
3
0
12
3
2
0
6
6
3
0
16
2
1
0
6
8
3
1
Qj
2
2
0
7
9
2
1
21
2
4
1
7
10
3
0
19
2
3
0
9
11
3
0
6
2
2
0
8
13
3
0
9
1
1
0
7
14
2
0
17
1
1
0
9
15
-
-
-
-
-
~
-
16
3
0
12
2
3
0
9
17
-
-
-
-
-
-
-
18
2
0
55
2
3
0
7
19
2
1
30
2
5
0
7
20
2
0
7
2
2
0
7
21
3
0
14
1
2
0
9
22
2
0
14
2
4
0
7
23
3
0
9
0
1
0
8
24
-
-
-
-
-
-
-
25
1
1
5
1
O
2
5

^ ^ ^
CTNVJ1
i ro r\j i
l O O I
I CT\UJ I
i ro ro i
i ro-u i
i oo i
i —o —^ i
o
oj uj v>i \^i l>j vjj rvi ro r\j r\)
ro O VC CD -0 O VX V>! (\) -i CAC ¡X 'J .7
v_>j fv> v_si ro ^ ro ro i ro ro ro v?j ro \j¡ ro ro
C-^OOC-^OO I -*_s-kC0->00
vji cc crui
ro
ro -c
-^v^- ro-*ro-*-i-*-*-*i r\j->-*-^-*-^-*ro
ro -£* ro vjj vj,- v»vj ro i ro ro ro ro ro ro -*
CCOOCOCOI -iOCCOCOC
-<]Cecc-<3-':C3vcvc i cs'cvo-c-^ccro-o
9 U

Table B-5
. Neuropsychological variables iron) 46 snoring males.
V/A1S 1 w Kinder Tapping
Subj. Ferior- WMSa Digit Hooper Hight Left Verbal
//“
Verbal
manee
Full
l“]Q
UKLSb
LiKVItc
8pan
L'HHtb
"lest
vu;;e
Hand
Hand
Fluene y
1
110
009
11b
137
11
11
7
17
24
4
44
40
05
2
07 5
07 b
074
ob 7
02
00
4
Ob
12
1
23
2-j
04
3
123
11b
121
12b
11
13
8
25
30
5
52
47
00
4
135
094
127
120
11
11
0
19
22
1
54
52
10
5
120
10b
115
135
10
11
0
5b
25
5
55
55
15
b
122
110
118
124
12
14
7
24
20
4
59
50
15
7
100
10b
108
105
Ob
09
b
1b
2
5
53
58
12
8
101
117
108
10b
Ob
11
7
14
23
y
5b
53
07
9
120
140
130
129
Ob
08
8
24
20
5
53
55
20
1U
128
122
127
137
13
14
7
28
20
4
b1
55
15
11
134
134
13b
124
11
12
8
25
28
4
bO
55
1b
12
102
112
1Cb
101
Ob
14
7
29
27
5
b4
51
10
13
102
090
100
095
03
04
b
1b
24
3
5b
52
09
14
140
131
140
14;>
09
13
8
20
28
5
55
52
15
15
104
103
104
107
09
11
5
30
29
1
59
01
15
1b
113
123
118
089
05
10
5
24
29
5
53
50
12
17
113
111
113
094
05
14
b
24
29
5
57
59
12
18
122
110
118
112
11
11
7
27
20
2
7b
bo
17
19
095
121
107
100
Ob
00
b
22
22
2
b1
b2
05
20
131
139
13b
143
11
13
7
29
27
4
72
72
19
21
120
113
110
110
08
10
7
25
25
2
bb
b4
09
22
124
123
125
13b
10
10
8
28
25
1
5b
52
19
23
119
118
120
142
10
14
8
31
27
5
57
55
20
24
127
119
125
120
10
15
7
2b
27
5
b4
1)0
16
25
122
120
122
120
11
11
7
18
24
4
54
55
12
2u
120
100
110
105
11
10
7
2b
25
5
bH
49
14

27
100
090
099
08b
01
00
28
129
13b
133
13b
09
12
29
122
116
120
114
09
09
30
137
135
138
120
10
09
31
121
102
114
097
0 j
02
32
123
114
120
118
08
11
33
123
112
119
109
13
0-j
34
133
116
127
129
09
10
3b
11b
120
116
120
09
12
36
122
102
114
121
10
08
37
133
127
134
133
1b
14
36
138
11b
131
118
07
08
39
1 j/4
133
137
133
08
12
40
100
109
10b
107
03
09
41
130
123
129
133
08
12
42
134
13b
136
143
1b
12
4;>
111
111
112
100
09
0^
44
123
13b
131
136
09
13
4b
136
142
131
143
13
14
4b
120
114
119
137
13
11
® WMSHQ = Wechsler Memory Scale Memory quotient
Delayed Recall Logical Stories
c Delayed Recall Visual Reproductions
d Delayed Recoil Rey Figure
e Wisconsin Card Sort
b
00
1b
8
20
2b
6
2b
2b
b
24
2b
7
13
20
b
20
27
b
23
24
7
30
26
7
21
27
7
2b
2b
7
34
2b
8
20
12
6
30
2b
7
19
23
7
30
28
b
2b
28
7
13
17
7
26
2b
7
27
2b
9
2b
29
57
bO
09
70
bO
19
57
bb
09
bO
b8
1b
53
44
11
70
77
13
49
42
08
80
72
11
71
70
12
54
50
11
6b
07
17
5b
47
20
73
74
17
bb
49
10
69
b1
14
53
50
18
50
49
17
74
65
10
5b
48
22
6b
60
13
2
3
4
5
0
4
1
4
4
5
b
4
b
2
b
5
3
4
b
2
ne

Table R-6. Respiratory variables from 46 snoring males.
ibject
Apnea
Apnea+
Hypopnea
Seconds
in
Kean
Saturation
Hut
Des:
a
Index
Index
Lvents
High
Low
>4?;
1
O.C
58.4
4058
93-4
80.9
220
2
1.7
4.2
350
95.4
88.7
214
3
0.5
0.5
36
94.9
95.0
2b
4
2.5
2.5
144
96.9
94.0
74
5
0.4
0.9
44
95.6
92.0
123
6
3-4
3-7
500
95.8
91.7
90
7
0.2
0.2
10
95.6
91.7
5
8
0.0
0.0
0
95.6
94.9
0
9
3.6
4.8
522
98.4
95-4
145
10
0.5
0.5
10
95.0
94.2
2
11
0.0
0.0
0
91.6
90.0
22
12
0.3
0.9
109
95.1
92.1
148
13
13.0
13.0
1498
94.3
90.0
3
14
0.8
0.8
36
95.1
93.7
10
15
0.2
0.2
23
95.6
94.5
4
16
0.8
0.8
48
95.1
94.1
2
17
0.2
0.2
21
95.4
94.1
7
10
0.2
0.2
10
94.3
93.2
0
19
0.0
0.0
0
96.5
95.1
18
20
0.7
0.7
40
95.1
93.3
49
21
0.0
0.2
20
94.7
92.3
60
22
0.0
0.0
0
94.9
94.3
1
23
0.0
0.0
0
95-4
94.9
0
24
0.0
0.0
0
92.8
92.0
1
>10?.
73
21
OOOWOCCOCOCOWOO^CO>]WWC

25
0.0
0.2
80
26
0.0
0.0
0
27
0.0
13.0
3192
28
0.0
4.3
441
29
2.9
5.5
514
30
0.0
0.0
0
31
11.1
12.1
1105
32
0.0
14.6
3586
33
2.7
2.7
192
34
O.C
0.0
0
35
0.0
0.0
0
36
0.0
0.0
0
37
0.0
0.0
0
38
0.8
0.8
36
39
0.0
0.0
0
40
0.0
1 .8
168
41
0.0
0.0
0
42
0.0
0.0
0
43
0.0
0.0
0
44
0.0
0.0
0
45
0.0
0.0
0
46
0.0
0.2
7
94.6
93.0
24
1
95-5
91.6
218
0
95.7
69.0
261
74
93.6
91.5
56
21
96.8
94.4
60
19
95.1
93.3
85
1
94.3
92.0
89
13
95.0
87.7
226
127
94.1
91.3
99
0
96.4
95.6
4
0
96.8
95.5
10
0
97.2
96.6
4
0
97.8
97.0
5
0
96.8
95.8
11
1
97.3
95.6
84
0
98.1
95.6
74
8
95.5
94.2
16
0
95.8
95.0
9
0
91 .7
90.3
48
3
96.3
95.5
6
0
96.6
96.5
1
0
94.5
92.6
57
1
120

Table B-7
Sleep variables from 46 snoring males.
Number % Slow Number
iubj.
H
TbT
Sleep
La tency
Sleep
Lii iciency
Sleep
Time
Time 5»
Sta^e 0
Stu¿e 0
Periods
1
Time
KLi-1
/» Sta*:e
2 5
4
Wave
Sleep
Stafee
Chan/'es
Kbh
La tency
Mean
HUM
Util
Cycle
1
30b
9
O.b
103
50.4
b
1.7
0
60.0
0
0
0
14
0
0
0
2
373
3
O.b
210
12.9
b
b.o
0
71.0
9.9
0.4
11.9
51
0
0
0
5
401
0
0.9
304
4.2
10
b.9
17-9
40.5
0.7
16.5
26.3
6b
02
10
90.7
5
374
0
0.7
205
22.6
12.0
3.6
10.9
4b. b
4.9
11.5
21.2
52
75
20
255.0
6
4B7
2
0.9
4bb
3.6
0
b.9
1b.2
5b.b
2.9
15.0
19.5
47
72
15
06.3
7
451
1b
0.9
371
10.1
14
b. 3
16.2
69.9
2.1
1.7
4.3
64
163
1 3
61.0
0
39b
14
0.7
203
23.1
9
3.2
12.6
63.0
1 .0
0.5
3.1
30
190
25.5
00.0
9
4brt
b4
O.b
303
7-b
2
1.9
2b. 4
47.3
6.1
1 1.65
19.3
56
30
26. 5
1 I *».v
1Ü
33b
94
0.3
110
b7.b
5
1.9
1.5
26.0
1.9
9.9
20.1
1b
221
4
0
11
407
30
0.0
341
9.3
7
2.1
21.0
65.4
2.7
1.6
4.7
42
60
27
145
12
402
1 b4
O.b
199
9.9
6
1.4
0
07.0
0
0
0
16
0
0
0
599
bO
0.7
2b7
25.b
7
3-7
2.9
60.5
2.6
6.9
12.3
50
247
5
49
14
35b
13b
0.6
20b
0.1
9
O.b
17.0
bb.b
4.5
4.5
9.0
52
02
19
167.U
1b
3B2
34
0.0
33b
3.7
7
0.5
17.0
59.2
4.5
15.5
20.6
29
212
29.5
10o. 0
1b
3bb
b9
0.6
219
26.0
10
2.0
2.6
b2. b
6.0
0.0
0.2
44
292
0
0
1 t
374
2b
0.0
30b
1.5
2
0.9
17.1
6b. b
4.5
9.7
14.4
21
97
2b.5
120
10
343
11
0.9
320
5.6
7
2.7
10.1
53.6
2.1
19.9
22.0
30
71
15
75.7
1b
329
24
0.2
75
b5.0
2
1.9
0
41.7
2.5
0
5.4
11
0
0
0
20
44b
67
0.7
321
13.7
7
4.0
16.7
57.3
2.9
7.5
12.1
40
135
51
122
21
32b
17
0.7
23b
1b.1
9
1.1
14.0
52.5
7.5
9.7
20.1
4 5
151
19.5
114
22
340
9
O.b
200
17.4
7
3.9
7.9
55.9
2.5
14.5
20.4
52
07
15.5
170
25
204
6
0.9
2bb
5.5
4
1.0
9.0
67.2
6.9
13.7
21.3
34
114
15.5
65
24
37 b
19
O.b
290
19.2
12
2.b
0.9
61.3
2.2
5.0
9.6
39
291
52
0
2b
32b
0
0.9
2bb
7.5
12
2.3
21.1
60.4
2.6
7.1
10.5
44
45
10.0
4 7.0
2b
3bb
23
O.b
221
33.6
35
2.1
0
64.5
0
0
0
71
0
0
0
27
5b4
2b
O.b
252
20. b
1b
6.9
9-b
63
0
0
0
50
262
20
0
2b
32/
19
o.b
2bb
4.5
4
2.9
15
bb.4
4.7
0.7
15.9
30
01
9
64
29
344
b
o.b
273
19.2
11
2.?
1b. 9
62.1
0
0
0
32
153
10
50.5
50
340
20
0.9
29b
7.0
6
3.1
15
63.4
4.7
6
5.1
30
bb
16
111
51
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
121

>■ A Vji I f U •
â–º -e -* ro ie rc vjw re x
>e we r.cvj.i*-»uic^v
ccccc cocccoccc
I
cvvett'i a vr -: a x: -< \c r
e- v>. e*. v* o i e*.e»1 lev. ie re iee>. .•
C iv -< vr es -» 03 r." v e* -fe- c ^
e> cc er re ■&» cy-*u;i-c.s-*er.
erx^erc^
C oi v.r o>. j | C^W'X^CTCav
C er re e>- -* vj fu -»re v -» r.
c c-kc o*. v re iv c v a e>. cv c c
a c vp (v ^civcc^cce
-< -e* c v -£»- v -o c; vo & v>- c v* vr
-»tv cr ve ■< erroy-k-i-t-vv^vx
e e. w s v iv-ii. jj, j jee
e- a. a. e* re •*■ ve ve C -» O' re -< iv
-> -e cr v£ '.■ -< e>.v er cr v e> e> ■
w cvc wvr >r o'v.'- o* -&■ a e- -»i
ro _» en vo -k xcam£o re ie x
ZZ l

REFERENCES
Agnew, H.W., & Webb, W.E. (1972). Measurement oí' sleep
onset by EEC criteria. American Journal oí EEC
Technology, 12, 127-134.
Agnew, H.W., Webb, W.B., & Williams, R.L. (1966). The
first night effect: An LEG study of sleep.
Psychophysiology, _2, 263-260.
Ancoli-Israel, S., Kripke, D., Mason, V., & Kaplan, W.
(1982). A prevalence study of sleep disorders in
seniors: Six-months results. Abstract published for
1982 APSS meeting, San Antonio.
Ancoli-Israel, S., Kripke, D., Mason, W., & Messin, S.
(1981). Sleep apnea and nocturnal myoclonus in a senior
population. Sleep, 4_(4), 349-358.
Annau, Z. (1972). The comparative effects of hypoxic and
carbon monoxide hypoxia on behavior. In B. Weiss & U.
Laites (eds), Behavioral Toxicology. New York: Plenum
Press.
Appel, D., Schmidt-Nowara, V., Poliak, C., & Wietzman, E.
(1982). Effects of tracheostomy closure on sleep and
breathing in sleep apnea patients with long term
tracheostomy. Abstract published for APSS meeting 1982,
San Antonio, p. 131.
Association of Sleep Disorder Centers (1979). Diagnostic
classification of sleep and arousal disorders. Sleep,
2, 1-137.
Eartus, R., Dean, R., & Bear, b. (1983). The cholinergic
hypothesis of impaired memory in the aged. In D. Harman
& K. Gray (eds), Aging: Nutrition anc Aging. New
York: Raupa Press.
Berg, I. (1948). A simple objective technique for
measuring flexibility in thinking. Journal of General
Psychology, 39, 15-22.
Berry, D., t Webb, W.B. (In Press). Sleep and cognitive
variables in normal aging subjects. Journal of
Gerontology
125

124
Berry, R., & Block, A.J. (1983). Positive nasal
pressure: A cure for snoring. American Review of
Respiratory Disease, 127. 84. ' ~
Birohfield, R., Seiker, H., & Heyman, A. (1958).
Alterations in blood gasses in natural sleep and
narcolepsy. Neurology. _0, 107-112.
Bixler, E., Kales, A., Soldatos, C., Vela-Bueno, A., Jacoby,
J., & Scarone, S. (1982). Sleep apneac activity in a
normal population. Research Communications in Chemical
Pathology and Pharmacology, 36(l)i 141-152.
Block, A.J. (1980). Respiratory disorders during sleep
(part 1). Heart and Lung, _9(6), 1101-1124.
Block, A.J. (1981a). Respiratory disorders during sleep
(part 2). Heart and Lung, 10(1), 90-96.
Block, A.J. (1981b). Polysomnography: Some difficult
questions. Annals of Internal Medicine, 95_(5), 644-656.
Block, A., Boysen, P., Wynne, J., & Hunt, L. (1979). Sleep
apnea, hypopnea, and oxygen desaturation in normal
subjects. New England Journal of Medicine, 300(10),
513-517.
Block, A., Wynne, J., & Boysen, P. (1980). Sleep
disordered breathing and nocturnal oxygen desaturation
in post-menapausal women. American Journal of Medicine,
69, 75-79.
Broadman, K., Erdmann, A., Lorge, I., Wolf, H., & Broadbent,
T. (1949). The Cornell Medical Index. Journal of the
American Medical Association. 140, 530-534.
Broughten, R. (1982). Performance and evoked potential
measures of various states of daytime sleepiness.
Sleep, ¿(5), 135-146.
Carskadon, M., & Dement, W. (1977). Sleep tendency: An
objective measure of sleep loss. Sleep Research, 6,
200.
Carskadon, M., & Dement, W. (1979a). Sleep tendency during
extension of nocturnal sleep. Sleep Research. _8, 147.
Carskadon, M., & Dement, W. (1979b). Effects of total
sleep loss on sleep tendency. Perceptual and Motor
Skills, 48, 495-506.

125
Carskadon, M., & Demenx, Vi. (19&1e). Cumulative effects of
sleep restriction on daytime sleepiness.
Psychophysiology, 16, 107-113.
Carskadon, K., & Dement, W. (1981b). Kespiraticn during
sleep in the aged human. Journal of Gerontology, 36(4),
420-423.
Carskadon, i-i., & Dement, W. (1982a). The multiple sleep
latency test: What does it measure? Sleep, _5, 567-572.
Carskadon, M., & Dement, W. (1962b). Nocturnal
determinants of daytime sleepiness. Sleep, _5, 573-581.
Carskadon, M., van den hoed, J., & Dement, W. (1980).
Insomnia and sleep disturbances in the elderly. Journa1
of C-erlatric Psychiatry, 13, 135-151.
Christenson, C., Gliner, J., Horvath, S., & Wagner, J.
(1977). Effects of three kinds of hypoxia on vigilance
performance. Aviation Space and Environmental Medicine,
46(6), 491-497.
Ccccagna, G., Mantovani, M., Bergmani, T., Parchi, C., h
I.ugaresi, E. (1972). Tracheostomy in hypersomnia with
periodic breathing. Bulletin of Physiological Pathology
and Respiration, 8_, 1217-1227-
Coverdale, S., Head, D., Woolcock, A., & Schoeffel, R.
(1980). The importance of suspecting sleep apnea as a
common cause of excessive daytime sleepiness.
Australian and Mew Zealand Journal of Medicine, 1C, 284-
268.
Davis, J. (1975). Adaptation of brain monoamine synthesis
to hvpoxia in the rat. Journal of Applied Physiology,
39(2), 215-220.
Dement, Vi., & Carskadon, M. (1982). Current perspective in
daytime sleepiness: The issues. Sleep, _5_, 556-566.
Dement, V.'., Carskadon, M., & Richardson, G. (1978).
Excessive daytime sleepiness in the sleep apnea
syndrome. In C. Guilleminault & W. Dement (eas), Sleep
Apnea Syndromes. New York: Alan R. Liss Inc.
Devereaux, K., & Partnow, M. (1975). Delayed hypoxic
encephalopathy without cognitive dysfunction. Archives
of Neurology, 52, 704-705.
Dolly, F., & block, A. (1982). Effect of flurazepam on
sleep disordered breathing in asymptommatic subjects.
American Journal of Medicine, 75, 239-243.

126
Feinberg, I., Koresko, R.L., & Heller, N. (1967). EEG
sleep patterns as a function of normal and pathological
aging in man. Journal of Psychiatric Research, 5, 107-
144.
Fisher, J., de la Pena, A., Mayfield, D., & Flickinger, R.
(1978). Psychobiologic determinants of subgroups of
sleep apnea. Sleep Research. _7, 221.
Francheschi, M., Zamponi, P., Crippa, D., & Smirna, S.
(1982). EDS: A 1 year study in an unselected inpatient
population. Sleep, _5(3), 239-247.
Freides, B., & Allweise, C. (1978). Transient hypoxic
amnesia. Behavioral Biology, 22, 178-189.
Fujita, S., Zorick, F. , Koshoreck, G. , Wittig, R., Conway,
W., & Roth, T. (1981). Treatment of upper airway sleep
apnea with UPP: Recent experience. Sleep Research, 10,
197.
Garay, S., Rapaport, D., Sorkin, B., Epstein, H., Feinberg,
I., & Goldring, R. (1981). Regulation of ventilation
in the obstructive sleep apnea syndrome. American
Review of Respiratory Disease, 124, 451-457.
Gastaut, H., Tassinari, C., & Duron, B. (1965). Etude
polygraphique des manifestations episoiques diurnes et
nocturnes du syndome de Pickwick. Revue Meurolgie, 112,
573-579.
Gellhorn, E. (1951). Sensitivity of the auditory
projection area to anoxia. American Journal of
Physiology, 164, 748-751.
Gibson, G., & Blass, J. (1976). Impaired synthesis of
acetylcholine accompanying mild hypoxia and
hypoglycemia. Journal of Heurochemistry, 27, 37-46.
Gibson, G. , & Duffy, J. (1981). Impaired synthesis of
acetylcholine by mild hypoxia or nitrous oxide. Journai
of Heurochemistry, 36(1), 28-53.
Gibson, G., Pulsinelli, Vi., Blass, J., & Duffy, T.
(1981). Brain dysfunction in mild to moderate
hypoxia. American Journal of Medicine, 70, 1247-1254.
Grant, I., Heaton, R., HcSweeny, J., Adams, K., & Timms,
R. (1982). Neuropsychological findings in hypoxemic
chronic obstructive pulmonary disease. Archives of
Internal Medicine. 142, 1470-1479.

127
Guilleminault, C., van den Hoed, J., & Mitler, M. (1978).
Clinical overview oí the sleep apnea syndromes. In C.
Guilleminault & W. Dement (eds), Sleep Apnea
Syndromes. New York: Alan R. Liss Inc.
Guilleminault, C., Simmons, F., Motta, J., Cumminsky, J.,
Rosekind, M., Shroeder, J., & Dement, W. (1981).
Obstructive sleep apnea syndrome and tracheostomy.
Archives oi Internal Medicine, 141. 985-988.
Guilleminault, C., Tilkian, A., & Dement, W. (1976). The
Sleep Apnea Syndromes. Annual Review oi Medicine,27,
465-470.
Halstead, W. (1947). Brain and Intelligence. Chicago:
University of Chicago Press.
Hanbauer, I., Karoum, F., Hellstrom, S., & Lahiri, S.
(1981). Effects of hypoxia lasting up to one monrh on
the catecholamine content in rat carotid body.
Neuroscience, _6, 81-86.
Harmon, E., Wynne, J., Block, A., & Malloy-Fisher, L.
(1981). Sleep-disordered breathing and oxygen
desaturation in obese patients. Chest, 79, 256-260.
Hartse, K., Roth, T., & Zorrick, F. (1982). Daytime
sleepiness and daytime wakefulness: The effect of
instruction. Sleep, _5, 107-118.
Hartse, K., Zorick, F., Roth, T., Kaffeman, M., & Hoyles,
T. (1980). Daytime sleep tendency in normal,
insomniac, and somnolent populations. Sleep Research,
1, 204.
Hartse, K., Zorick, F., Sicklesteel, J., Piccione, P., &
Roth, T. (1979). Nap recordings in the diagnosis of
daytime somnolence. Sleep Research, _8, 190.
Hebb, D. (1961). Distinctive features of learning in the
higher animal. In D. Delafresneye (ed), Brain
Mechanisms and Learning. London: Oxford University
Press.
Hoddes, E., Dement, W., & Zarcone, V. (1972). The
development and use of the Stanford Sleepiness Scale.
Psychophysiology, _9, 15C-160.
Hoddes, E., Zarcone, V., Smythe, H., Phillips, R., & Dement,
Vi. (1973). Quantification of sleepiness: A new
approach. Psychophysiology. 10, 431-436.

128
Hooper, H. (1958). The Hooper Visual Organization Test:
Manus1. Los Angeles: Western Psychological Services.
Kales, A., Bixier, E., Soldatos, C., Vela-Bueno, A.,
Caldwell, A., & Cadieux, R. (1982). Insomnia: The
role of sleep apnea and nocturnal myoclonus.
Psychosomatics, 25, 187-195.
Kreiss, P., Kripke, D., & Ancoli-Israel, S. (1982).
Prevalence of sleep apnea among representative
inpatients on a veterans administration medical ward.
Abstract published for 1982 APSS meeting, San Antonio,
Texas.
Krieger, J., Turlot, J., Margin, P., & Kurtz, D. (1983).
Breathing during sleep in normal young and elderly
subjects: Hypopneas, apneas, and correlated factors.
Sleep, _6(2), 106-120.
Krop, H., Block, A., £ Cohen, E. (1973). Neuropsycho¬
logical effects of continuous oxygen therapy in chronic
obstructive pulmonary disease. Chest, 64, 317-322.
Krop, H., Block, A., Cohen, E., Croucher, R., & Schuster,
J. (1977). Neuropsychological effects of continuous
oxygen therapy in the aged. Chest, 72. 737-743.
Lugaresi, E. , Cirignotta, F., Coccagna, G., & Baruzzi, A.
(1982). Snoring and the obstructive sleep apnea
syndrome. ELG, Suppl No. 35, 421-430.
Lugaresi, E., Cirignotta, F., Coccagna, G., £ Piaña, C.
(1980). Some epidemiological data on snoring and
cardiocirculatory disturbances. Sleep, _3(3/4), 221-224.
Lugaresi, E., Coccagna, G., Berti, G., & Mantovani, H.
(1968). La "maledizione di Ondine" il disturbo del
respiro del sonno neil ipoventalizione aiveolari
primaria. Sist I-leurologie, 20, 27-37.
McFarland, R. (1937). Effects of partial acclimitization
on psychological tests at high altitudes. Journal of
Comparative Psychology, 23. 227-230.
McGinty, D., Littner, M., Beahm, E., Ruiz-Primo, E., Young,
L., & Sowers, J. (1982). Sleep related breathing
disorders in older men: A search for underlying
mechanisms. Neurobiology of Aging, _3, 337-350.
tleehl, P. (1977). Psychodiapnosis. New York: Y/.W. Norton
and Co.

Hiles, L., & Simmons, F. (1984). Evaluation of 190
patients with loud and disruptive snoring. Sleep
Research, 13. 154.
Killer, Vi. (1982). Cardiac arrhythmias and conduction
disturbances in the sleep apnea syndrome. American
Journal of Medicine, 75, 317-321.
Milner, B. (1970). Memory and the medial temporal regions
of the brain. In K. Pribram & D. Broadbent (eds),
Biology of Memory. New York: Academic Press.
Hitler, M., Gujaverty, K., Sampson, M., & Bowman, C.
(1982). Multiple nap approaches to evaluating the
sleepy patient. Sleep, _5, 119-127.
Orr, Vi., Martin, R. , Imes, N. , Rogers, R., & Stahl, M.
(1979). Hypersomnolent and nonhypersomnolent patients
with upper airway obstruction during sleep. Chest,
75(4), 418-422.
Csterrieuh, P. (1944). Le test de copie d'une figure
complexe. Archives de Psychoiogie, 30. 206-356.
Plum, F., Posner, J., & Hain, R. (1962). Delayed
neurological deterioration after anoxia. Archives of
Internal Medicine, 110, 18-23.
Reynolds, C., Coble, P., Kupfer, D., & Holzer, B. (1982).
Application of the multiple sleep latency test in
disorders of excessive somnolence.
Electroencephalography and Clinical Neurophysiology, 53,
443-452.
Richarason, G., Carskadon, M., Flagg, V., van den Hoed, J.,
Dement, W., & Hitler, M. (1978). Excessive daytime
sleepiness in man: Multiple sleep latency measurements
in narcoleptic and control subjects.
Electroencephalography and Clinical neurophysiology, 45,
021-627.
Richardson, J., Chambers, R., & Heywood, P. (1959).
Encephalopathies of anoxia and hypoglycemia. Archives
of Heurology, _1_, 178-190.
Roth, T., Hartse, K., Zorick, F., & Conway, V. (1980).
Multiple naps in the evaluation of daytime sleepiness in
patients with upper airway sleep apnea. Sleep, _5_(3/4),
425-439.
Russell, E. (1975). A multiple scoring method for
assessment of complex memory functions. Journal of
Consulting and Clinical Psychology, 43, 800-809.

130
Sara, S. (1974). Delayed development of amnestic behavior
after hypoxia. Physiology and Behavior, 13, 693-696.
Schmidt, H. (1982). Pupilometric assessment of disorders
of arousal. Sleep, _5_, 157-164.
Schroeder, J., Motta, J., & Guilleminault, C. (1978).
Hemodynamic studies in sleep apnea. In C. Guilleminault
& W. Dement (eds.), Sleep Apnea Syndromes. New York:
Alan R. Liss Inc.
Sicklesteel, J., Zorick, F. , Wittig, R., Conway, Vi., & Roth,
T. (1981). Daytime somnolence and insufficient
sleep: A case series. Sleep Research, 10, 233.
Smallwood, R., Vitiello, M., Giblin, E., & Prinz, P.
(1983). Sleep apnea: Relationship to age, sex, and
Alzheimers Dementia. Sleep, _6H), 16-22.
Smirne, S., Francheschi, M., Comi, G., & Leder, F.
(1980). Sleep apneas in Alzheimers disease. Sleep
Research, _9, 224.
Sullivan, C., & Issa, F. (1980). Pathophysiological
mechanisms in obstructive sleep apnea syndrome. Sleep,
1(3/4), 235-246.
Tauber, B., & Allweise, C. (1975). Effects of acute
hypoxia on memory. Israeli Journal of Medical Science,
21, 71.
Tilkian, A., Motta, J., & Guilleminault, C. (1978).
Cardiac arrhythmias in sleep apnea. In C. Guilleminault
& W. Dement (eds), Sleep Apnea Syndromes. New York:
Alan R. Liss Inc.
Van Liere, E., & Stickney, J. (1963). Hypoxia. Chicago:
University of Chicago Press.
Webb, P. (1974). Breathing during sleep. Journal of
Applied Physiology, 37(6), 899-903.
Webb, W.B. (1975). Sleep—The Gentle Tyrant. Englewood
Cliffs, N.J.: Prentice-Hall Inc.
Webb, W.B. (1982). Measurement and characteristics of
sleep in older subjects. Neurobiology of Aging, _3_, 311 —
319.
Wechsler, D. (1945). A standardized memory scale for
clinical use. Journal of Psychology, 19. 87-95.

Vechsler, D. (1955). Wechsler Adult Intelligence Scale:
tienual. New York: Psychological Corporation.
Weitzman, E. (1979). The syndrome of hypersomnia and sleep
induced apnea. Chest, 75. 414-415-
Weitzman, E., Kahn, E., & Poliak, C. (1980). Quantitative
analysis of sleep and sleep states before and after
tracheostomy in patients with the hypersomnia sleep
apnea syndrome. Sleep, ¿(3/4), 407-423.
Weitzman, E., Poliak, C., Borowieki, B., Burak, B.,
Shprintzen, R., & Rakoff, S. (1978). The hypersomnia
sleep apnea syndrome: Site and mechanism of upper
airway obstruction. In C. Guilleminault & Vi. Dement
(eds), Sleep Apnea Syndromes. New York: Alan R. Liss
Inc.
West, J. (1984). Human physiology at extreme altitudes.
Science, 223, 784-788.
Zorrick, T. , Roehrs, T., Koshorek, G., Sicklesteel, J.,
Hartse, K., Wittig, R., & Roth, T. (1982). patterns of
sleepiness in various disorders of EDS. Sleep, _5, 165-
174. '
Zwiilion, C., Devlin, T., White, D., Douglass, N., Weil, J.,
& Martin, R. (1982). Bradycardia during sleep apnea.
Journal of Clinical Investigations, 69, 1286.

BIOGRAPHICAL SKETCH
David T. R. Berry was born on April 3, 1958, in
California. He attended high school in Montgomery, Alabama,
received a bachelor's degree in psychology from Auburn
University, and a master's degree in experimental psychology
from the University of Alabama in Birmingham. After
completing an internship in clinical psychology in Chicago,
he plans to seek a faculty position in psychology.
132

I certify that X have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.
Wilse B. V/ebb, Chairman
Graduate Research Professor of
Psychology
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.
Assistant Professor of Clinical
Psychology
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.
A. áy.Block
Processor of Medicine
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.
Professor of Clinical
Psychology

I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.
Assistant Professor of Clinical
Psychology
This dissertation was submitted to the Graduate Faculty of
the College of Health Related Professions and to the
Graduate School and was accepted as partial fulfillment of
the requirements for the degree of Doctor of Philosophy.
August 1985
^X»J) \
Dean, College of Health Related
Professions
Dean, Graduate School




70
In summary, correlational procedures suggested
relationships between nocturnal respiratory variables and
subjective sleepiness, while no reliable between group
difference were noted on the nap or SSS variables.
Subjective Sleep Assessment
Two assessments of subjective sleep parameters were
completed, including the sleep questionnaire and sleep
log. Several variables were selected from the sleep
questionnaire and included: usual hours of sleep,
judgement of whether usual sleep was adequate, number of
naps per week, hours napping per week, number of nightly
wakenings, minutes of nightly wakenings, usual sleep
latency, depth of sleep, and daytime sleepiness. These
variables were correlated with the respiratory variables,
for a Pearson matrix of 63 correlations. Table 3-9
presents the signficant correlations which emerged from
this matrix. Several respiratory variables were
associated with napping variables, while mean low
saturation and number of desaturations >4% were associated
with reported minutes of nightly wakenings.
Table 3-10 presents the means and standard deviations
from the subjects grouped by level of apnea/hypopnea.
Oneway ANOVA procedures revealed no significant between
group differences on the selected sleep questionnaire
variables [Usual hours sleep, F(2,43)=G.6, p-NS; Adequate
sleep, F(2,43)=0.3> p=NS; Number of naps, F(2,43)-2.9
p=NS; Hours napping, F(2,43)=0.8, p=NS; Number of


45
verbal fluency, and signal detection are affected by acute
hypoxia. Animal data suggest that sensitivity of memory
processes to hypoxia, while data from chronic hypoxia
humans indicate that attention, vigilance, abstracting
ability, complex perceptual' motor skills, as well as motor
and sensory abilities are also impaired by hypoxia. While
none of these findings are unequivocally directly
applicable to the brief, acute, desaturaticn occurring
during apnea, they are, at the very least, indicative of
abilities sensitive to hypoxia in humans.
A neuropsychological battery sensitive to these
abilities includes Wechsler Adult Intelligence Scale
(WAIS; Wechsler, 1955); Wechsler Memory Scale (WMS;
Wechsler, 1945) including the Russell delayed memory
aspects (Russell, 1975); Rey-Osterreith Complex Figure
(Osterreith, 1944); Hooper Visual Organization Test
(Hooper, 1958); Verbal Fluency Test (Halstead, 1947);
Wisconsin Card Sort (Berg, 1948); and the Finger Tapping
Test (Halstead, 1947). Administration of this test
battery provides a broad screening of neuropsychological
functioning, with a particular emphasis on those abilities
known to be sensitive to hypoxia. Application of this
battery to subclinical apnea subjects provides a sensitive
indicator of possible hypoxic brain dysfunction.
In summary, alterations of sleep/wake cycles appear
to be present in sleep apnea syndrome patients. Thus
assessment of these variables in a subclinical population


Table 3-16* Multivariate regression of sleep and nocturnal
respiratory on cognitive variables in 43
snoring males.
Predictor Variables
Criterion Variables
Age
WAIS Performance IQ
Apnea Index
V/echsler Memory Scale
Memory
Apnea+Hypopnea Index
Quotient
Number of Desaturations >4%
Delayed Recall Visual
Sleep Efficiency
Reproduction
Number Stage 0 Periods
Delayed Recall Rey Figure
Time % Stage 0
Verbal Fluency
Time % Stage 2
Number of Stage Changes
Wisconsin Card Sort
Multivariate test of significance
Hotellings = 4.063 p<.001
Univariate Regression with
(9.33) d.f.
Variable
Multiple R F.
VAIS Performance IQ
V.echsler Memory Scale
.610 2.17
.051
Memory Quotient
Delayed Recall Visual
.566 1.73
.121
Reproductions
Delayed Recall Rey
.655 2.91
.012
Figure
.668 2.96
.011
Verbal Fluency
578 .85
.096
Wisconsin Card Sort
.453 .95
.499
Ereakdown of significantly
predicted criterion regressions
Criterion
Predictor
Eeta
Sig.
Delayed Recall
Age
-.3596
.020
Visual
Apnea Index
-.3218
.021
Reproduction
Apnea+Hypopnea Index
Number of
.2329
.218
Desaturations >4%
-.3778
.041
Sleep Efficiency
.2303
.408
Number of Stage 0 Periods
.3260
.209
Time % Stage 0
-.1136
.731
Time % Stage 2
.0005
998
Number of Stage Changes
-.2235
391


24
activity and its correlates in a subclinical population.
It is clear from the earlier review of apnea incidence
that sampling from the general population would result in
a relatively low population of individuals with sleep
apnea, an inefficent approach from the perspective of the
limited resources available for research. Thus selection
variables must be considered with one eye on efficiency
and the other on protecting the generalizability of
results. Selection criteria should also aim toward eluci
dating the relative roles and possible interactions of
these variables in sleep apnea. Therefore, selection
variables should have as straight-forward a relationship
as possible with sleep apnea, seeking to answer more ques
tions than they raise.
With these caveats in mind, the variables available
as selection criterion will be reviewed and conclusions
drawn about their relative merits. The first variable
considered, sex, has been relatively well investigated.
It is clear that sleep apnea activity is primarily a male
circumstance. Therefore, the selection of female subjects
is contraindicated from the standpoint of efficiency. Age
has been demonstrated to exhibit a positive relationship
with sleep apnea, indicating selection of older sub
jects. Snoring is another variable with a close relation
ship to sleep apnea. The apparently common mechanism of
snoring and sleep apnea (stenosis of the upper airway)
indicates that possibly very few subjects witnout snoring


124
Berry, R., & Block, A.J. (1983) Positive nasal
pressure: A cure for snoring. American Review of
Respiratory Disease, 127 84.
Birchfield, R., Seiker, H., & Heyman, A. (1958).
Alterations in blood gasses in natural sleep and
narcolepsy. Neurology, _C, 107-112.
Bixler, E., Kales, A., Soldatos, C., Vela-Bueno, A., Jacoby,
J., & Scarone, S. (1982). Sleep apneac activity in a
normal population. Research Communications in Chemical
Pathology and Pharmacology, 56(1), 141-152.
Block, A.J. (1980). Respiratory disorders during sleep
(part 1). Heart and Lung. 9.(6), 1101-1124.
Block, A.J. (1981a). Respiratory disorders during sleep
(part 2). Heart and Lung, 10(1), 90-96.
Block, A.J. (1981b). Polysomnography: Some difficult
questions. Annals of Internal Medicine, 95(5), 644-656.
Block, A., Boysen, P., Wynne, J., & Hunt, L. (1979). Sleep
apnea, hypopnea, and oxygen desaturation in normal
subjects. New England Journal of Medicine. 300(10),
513-517.
Block, A., Wynne, J., & Eoysen, P. (1980). Sleep
disordered breathing and nocturnal oxygen desaturation
in post-menapausal women. American Journal of Medicine,
69 75-79.
Eroadman, K., Erdmann, A., Lorge, I-, Wolf, H., & Broaabent,
T. (1949). The Cornell Medical Index. Journal of the
American Hectical Association, 140, 530-534.
Broughten, R. (1982), Performance and evoked potential
measures of various states of aaytime sleepiness.
Sleept (5), 135-146.
Carskadon, M., & Dement, W. (1977). Sleep tendency: An
objective measure of sleep loss. Sleep Research, 6,
200.
Carskadon, M., & Dement, W. (1979a). Sleep tendency during
extension of nocturnal sleep. Sleep Research, 8_, 147.
Carskadon, M., & Dement, W. (1979b). Effects of total
sleep loss on sleep tendency. Perceptual and Motor
Skills, 48, 495-506.


47
possible risk factors associated with subclinical sleep
apneas


Carskadon, M., & Dement, Vi. (1981a). Cumulative effects of
sleep restriction on daytime sleepiness.
Psychophysiology, 16, 107-113.
Carskadon, K., & Dement, W. (1981b). Respiration during
sleep in the aged human. Journal of Gerontology, 36(4),
420-423.
Carskadon, M.f & Dement, V. (1982a). The multiple sleep
latency test: What does it measure? Sleep, _5, 367-372.
Carskadon, M., & Dement, W. (1982b). Nocturnal
determinants of daytime sleepiness. Sleep, _5, 573-581 .
Carskadon, M., van den Hoed, J., & Dement, V. (1980).
Insomnia and sleep disturbances in the elderly. Journal
of Geriatric Psychiatry, 13, 135-151.
Christenson, C., Gliner, J., Horvath, S., & Vagner, J.
(1977). Effects o three kinds of hypoxia on vigilance
performance. Aviation Space and Environmental Medicine,
18(6), 491-497.
Coccagna, C., Hantovani, H. bergmani, T., Parchi, C., b
Lugaresi, E. (1972). Tracheostomy in hypersomnia with
periodic breathing. Bulletin of Physiological Pathology
and Respiration, _8, 1217-1227.
Coverctale, S., Read, D. Voolcock, A., & Schoeffel, R.
(1980). The importance of suspecting sleep apnea as a
common cause of excessive daytime sleepiness.
Australian and New Zealand Journal of Medicine, 10, 284-
20&:
Davis, J. (1975). Adaptation of brain monoamine synthesis
to hvpoxia in the rat. Journal of Applied Physiology,
39(2), 215-220.
Dement, V., & Carskadon, M. (1982). Current perspective in
daytime sleepiness: The issues. Sleep, 5., 556-566.
Dement, V/., Carskadon, M. & Richardson, G. (1978).
Excessive daytime sleepiness in the sleep apnea
syndrome. In C. Cuilleminault & V. Dement (eos), Sleep
Apnea Syndromes. New York: Alan R. Liss Inc.
Devereaux, M., & Partnow, M. (1975). Delayed hypoxic
encephalopathy without cognitive dysfunction. Archives
of Neurology, 32, 704-705.
Dolly, F., & Block, A. (1982). Effect of flurazepam on
sleep disordered breathing in asymptommatic subjects.
American Journal of Medicine, 73. 239-243.


55
2200 Sleep Questionnaire, Cornell Medical Index,
remaining testing
2230 Wiring for thumistor, EKG, respiratory effort,
ear oximeter
2300 Lights out, begin recording
A night technician remained with the subjects throughout
the night. Subjects usually awakened between 5 and 6 am,
and later completed the one week's sleep logs.
Statistical Procedures
Several statistical procedures were utilized in
evaluating the present data. One of the primary
approaches was correlational, considered appropriate
because of the exploratory nature of the study. Bearing
in mind the possible alpha inflation of large corre
lational matrices, other approaches were also utilized to
add confidence. Thus the sample was stratified by level
of apnea/hypopnea, and oneway ANOVA procedures were used
to evaluate between group differences. Additionally, non-
parametric procedures were utilized where appropriate.
Positive correlational results were considered against
possible replication with other measures in the study, and
in some cases multivariate procedures were added for
further confidence. It is conceded that correlational
approaches may carry the burden of possible spurious
results, but it is believed that the various additional
procedures outlined above lend confidence to the findings.


42
neurotransmitter turnover seen in hypoxia. It has been
suggested that these subtle alterations may underlie the
changes in complex behavior seen in hypoxic humans.
Data on human performance after hypoxia have neces
sarily been reliant on fortuitous or weak manipulations.
Although this may limit the usefulness of these data, it
is at the very least highly suggestive of variables sensi
tive to hypoxia. Van Liere and Stickney (1963) note that
acute hypoxia leads to confusion, headache, drowsiness,
weakness and incoordination. After effects include head
ache, nausea and emotional lability. Richardson ex al.
(1959) found that acute, fatal hypoxia from surgery
results in damage to cortical neurons, particularly in
layers III and IV. Susceptible subcortical neurons in
clude those of the corpus striatum and cerebellum. Plum
et al. (1962) describe a syndrome of delayed postanoxic
encephalopathy. After the acute coma resolves in 4-5
days, a brief period of normal functioning returns.
Increasing irritability and confusion then emerge, with
loss of coordination and memory, and diminished attention
span. Typically, the lesion spares grey matter while
affecting white matter. Devereaux and Partnow (1975)
describe a patient who recovered from a delayed acute
encephalopathy. Although IQ was relatively spared, severe
dysarthria emerged. Relevant points which emerge from
studies of acute hypoxia in humans suggest that sublethal


SLEEP APNEA ACTIVITY AND ITS CONCOMITANTS
IN A SUBCLINICAL POPULATION
By
DAVID THOMAS REED BERRY
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN
PARTIAL FULFILLMENT OF THE REQUIREMENTS
FOR THE DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
1985


D
*=93
O.^
3/2
3.1
5
2.0
33
4 9
2
0.9
5/y
11.1
8
1 b
3 4
378
33
0. V
2b 4
0
10
13.8
35
555
51
o.y
240
21.8
b
1.5
5<>
321
2 b
0.9
284
2.7
4
5.8
3/
424
28
0.8
55y
8.9
9
2.5
58
325
21
o.y
221
9.4
7
1 b
39
41b
5
0.9
505
b. 5
7
2.2
40
41
42
372
1
0.8
519
0.9
10
3.2
35o
1
0.9
354
3.7
11
11.5
43
442
28
0.8
552
8.8
6
3.6
44
425
5
0.9
575
9.b
1.
J
2.9
45
4 54
22
o.y
528
20.4
25
5.b
4b
5/0
54
0.8
305
2.8
b
0.8
I
1o.9
oU. 9
5.2
11.2
18.y
40
70
18. 3
100.7
0.5
85.3
0.7
3.1
4.2
27
102
1
0
10.0
99.7
1.2
5.3
5.7
37
67
1 1
106
1.5
09.1
1 .b
5*b
9.2
23
q298
2
5
10.b
56.1
5.1
28.4
50.3
29
147
15.5
1U9
9.9
75.1
1 .0
5.5
6.9
35
1o9
13
133
14.8
85.9
4.9
5.5
11.5
30
127
18
95
10.7
67.2
1.9
13.6
1 b o
37
177
14.7
98
25.5
59.5
2.4
7.9
10.9
40
91
26.7
68
12.4
81.7
->.2
7.5
11.1
12.4
81.7
10.8
<2
15.5
47.9
7.5
19.2
26.9
5b
113
20
123.5
10.b
84 6
2.4
6.9
12.5
40
95
17.3
63.7
5.4
80.2
5.4
7.0
13.1
70
393
7
17
16.7
03.1
5.9
11.2
17.2
29
123
28
1 3*5
r\j
ro


34
these data. Below a discussion of measurement issues and
proposal assessment of these variables appears.
Measurement of Deficits Found in Sleep Apnea Syndrome
Above it was noted that cardiopulmonary complica
tions, excessive daytime sleepiness, nocturnal oxygen
desaturation and cognitive/intellectual changes are common
deficits of sleep apnea syndrome. A discussion of
measurement issues and proposed measurement devices for
these variables will be presented below.
Measurement of Cardiopulmonary and Health Deficits
Cardiopulmonary deficits are thought to be frequent
sequelae of sleep apnea syndrome. Thus an assessment of
blood pressure as well as self reports of hypertension and
heart trouble seems indicated in subclinical apnea sub
jects. Additionally, a general survey of health status,
such as that provided by the Cornell Medical Incex (CMI;
Broadman et al., 1949) would also screen for other health
deficits. The CMI provides an overall score inaicating
number of symptoms endorsed, as well as subscales on
several symptom categories. Separate examination of sub
scales on respiratory, cardiopulmonary, and neurological
subscales seems appropriate. Assessment of these
variables provides a broad health screening, as well as a
detailed analysis of symptoms found in sleep apnea syn
drome.


CHAPTER THREE
RESULTS
Demorraphics and Incidence oi'
nocturnal Respiratory Disorder
Subjects were selected to fulfill several sampling
requirements, including male sex, presence of heavy
snoring, and self report of general good health. A total
of 60 subjects were studied. Five subjects were later
excluded when questioning revealed the presence of chronic
alcoholism (2) and serious head trauma (3). Additionally
9 overnight records were rejected because of missing or
unusable data. The characteristics of the final 4b
subjects are detailed in Table 3-1- These snoring males
had a mean age of approximately 30 years, a mean weight of
approximately 190 pounds, and a mean educational level of
15 years of schooling.
Table 3-1. Demographic variables
from 46
snoring
males.
M
SD
Range
Age
49.9
13.1
30-75
Weight (lbs)
189.3
34.9
125-250
Weight:Height Ratio (lbs:inches)
2.7
0.5
1.9-3.6
Education (years)
14.7
2.5
9-20
56


20
in mind, they conducted a survey of 1000 subjects on the
incidence of snoring. Chronic snorers included 31# of the
males and 19% of the females. In an elaboration of this
study, Lugaresi et al. (1980) questioned 5713 indivi
duals. In this sample, 24% of the males and 13% of the
females were chronic snorers. Snoring increased with age,
and by 60 years of age 60% of the males and 40% of the
females snored. Of note was the observation that hyper
tension occurred more frequently in snorers than non-
snorers, suggesting that the former was more at risk for
cardiovascular complications.
Virtually every study of SAS patients notes the pre
sence of snoring in all these subjects (Sullivan and Issa,
1980; Block, 1980; Coverdale et al., 1980). This is
true for both predominantly obstructive, and predominately
central SAS subjects (Guilleminault et al., 1978). How
ever, the presence of heavy snoring in asymptomatic
(nonhypersomnolent) apneas is also well documented (Orr et
al., 1979; Fisher et al., 1978; Berry and Block, 1983),
as well as in otherwise normal subjects. These data may
be taken as evidence for snoring as a necessary, but not
sufficient, condition for the occurrence of sleep apnea.
The role of body weight in SAS is less clear.
Recall that Pickwickian patients are invariably obese, as
well as hypercapnic. Guilleminault et al. (1978) stress
that SAS may occur in nonobese as well as obese sub
jects. However, Block (1980) states that he has found SAS


67
was significant at p<.G5 or less: Number of desaturations
>4% vs REM latency (r=-.369).
Table 3-7 presents the sleep variable means for the
subjects grouped by level of apnea/hypopnea. The group
with high apnea/hypopnea appears to spend more time awake,
less time asleep, and achieve less slow wave sleep than
the others. However, none of these differences proved
significant. In fact, only one out of 15 oneway ANOVAs
achieved significance: Time in bed, F(2,40)=0.4, p=NS;
Sleep latency, F(2,40)=1.7> p-NS; Pure sleep time,
F(2,40)=C.3, p=NS; Sleep efficiency index, F(2,40)=0.2,
p=NS; Number of stage 0 periods, F(2,40)=0.5 p=NS; Time
% stage 0, F(2,40)=1.3 p=NS; Time % stage 1,
F(2,40)=1.3, p=NS; Time % stage 2, F(2,40)=0.04, p=NS;
Time % stage 3 F(2,40)=3.64, p<.C4; Time % stage 4
F(2,40)=1.32, p=NS; Time % stage REM, F(2,40)=1.5,
p=NS: Latency 1st REM, F(2,40)=0.6, p=NS; Mean REM
period length, F(2,40)=0.3> p=NS: Mean REM cycle length,
F(2,40)=0.9, p=NS; % slow wave sleep, F(2,40)=2.4,
p=NS. Scheffe followup procedures for time % stage 3
revealed no significant between group comparisons.
To summarize, EEG sleep measures indicatec lighter
sleep then normal in this sample, with increases in
awakenings and light slow wave sleep observed. Neither
between group comparisons nor correlational procedures
revealed significant interactions between sleep and
nocturnal respiratory variables.


64
asked wnether they had ever been diagnosed with hyper
tension. Thirty-three per cent of those with no events
had diagnosed hypertension, as did 33% of those with low
levels of apnea/hypopnea, while only 15# of those with
high levels of apnea/hypopnea had diagnosed hyperten
sion. These data do not support an increasing level of
hypertension in subjects with high levels of
apnea/hypopnea.
The subjects were grouped into the earlier delineated
levels of apnea/hypopnea. Table 3-5 presents the means
and standard deviations of health variables from the three
groups. No dramatic trends are apparent in this table,
and oneway ANOVA procedures confirmed this observation
with no significant differences for: Diastolic BP,
F(2,43)-*01, p-NS; Systolic BP, F(2f43).1, p=NS; CMI
overall, F(2,42)=.2, p=NS: CMI respiratory, F(2,42)=.5*
p=I\!S; CMI cardiopulmonary, F(2,42)=.5, p=NS; CMI
neurological, F(2,42)=2.2, p=NS.
In summary, while correlational procedures suggested
relationships between several oxygen saturation measures
and blood pressure as well as overall number of symptoms
endorsed on the CMI, oneway ANOVAs failed to reveal
significant between group differences when subjects were
stratified by level of nocturnal distress.


111
Table B-2.
Daytime sleepiness
males.
variables from
46 snoring
Subject
Mean Stanford
Sleepiness
Mean Sleep
Latency
Minutes of
Sleep
if
Rating
- Naps
- Naps
1
_
09.0
22
2
-
16.0
05
5
-
11.5
18
4
4.5
-
-
5
3.5
06.5
26
6
5.5
-
-
7
4.7
20.0
01
8
2.7
13.5
13
o
3.8
09.5
17
1C
4.3
20.0
00
11
3.4
20.0
00
12
3.3
16.0
09
13)
-
-
-
14
4.5
20.0
00
15
4.1
14.0
14
16
3.6
20.0
00
17
2.8
08.5
25
18
2.5
11 .5
18
19
-
15.C
11
20
3.6
20.0
00
21
3.7
20.0
00
22
2.4
14.5
05
23
4-5
14.5
12
24
4.4
20.0
00
25
3.3
04.0
34
26
3.3
-
-
27
4.8
19.0
03
28
3.5
20.0
00
29
4.1
15.0
10
30
2.6
17.5
07
31
4.3
14.0
13
32
3.1
11.0
18
33
5.1
11.5
18
34
3-1
-
-
35
3.8
20.0
00
36
3.4
20.0
00
37
3.4
13.5
14
38
2.2
20.0
00
39
2.9
06.5
25
40
3.9
19.0
08
41
4.0
C8.0
27


96
reported in similar studies. A first night effect pro
bably caused increased sleep latency, wakenings, and
lighter sleep, a result which may have functioned to
obscure possible relationships between sleep and respira
tory variables, which were not observed. A pattern of
correlations between increased respiratory disturbance and
sleepiness was noted, perhaps because these variables
reflected subjective reports which were largely indepen
dent of the experimental nights sleep (SSS ratings made
before sleep, sleep questionnaires reflecting trait sleep
variables, and sleep logs averaged over a 7 day period
following the experimental nights' sleep). Multiple night
recordings appear necessary to sort out a possible syner
gistic interaction between the first night effect and the
discomfort of the recording procedures.
I.ccturnrl hespratory cnc h'c-crophysi oloriccl Variables
The major finding of the respiratory/cognitive
analysis was the demonstration of relationships between
nocturnal respiratory parameters and cognitive/neuro-
psychological scores tapping non-verbal intelligence,
verbal and non-verbal memory, expressive verbal fluency,
and cognitive flexibility. Most of these relationships
remained when age, weight, and educational differences
were ruled out as plausible alternatives.
These findings extend the boundaries set by previous
work which indicated a relationship between hypoxia and
neuropsychologica1 scores. These previous reports


BIOGRAPHICAL SKETCH
David T. R. Berry was born on April 3, 1938, in
California. He attended high school in Montgomery, Alabama,
received a bachelor's degree in psychology from Auburn
University, and a master's degree in experimental psychology
from the University of Alabama in Birmingham. After
completing an internship in clinical psychology in Chicago,
he plans to seek a faculty position in psychology.
132


38
The normals averaged 20.4 m to Stage 1 sleep while the SAS
patients averaged 1.9 m. Zorrick et al. (1982) found mean
MSLT latencies to Stage 1 sleep of 3.1 m for SAS subjects
and 2.9 m for narcoleptics, significantly shorter than for
patients with psychiatric disorders.
The chief difficulty remaining in Judging the MSLT
"reliable and valid is external validation demonstrating
that exceeding a certain mean latency is correlated with
other deficits of performance. In some ways, low laten
cies on the MSLT may suffer the same limitations as sleep
deprivation protocols; that is, demonstrating serious
deficits is surprisingly difficult. At the same time, few
would deny the subjective unpleasantness of sleep depriva
tion, which may be applicable to low MSLT scores. Thus
the MSLT appears indicated in assessing excessive daytime
sleepiness while the Stanford Sleepiness Scale ( sSS) pro
vides an index of the subjective experience of the sub
ject. Application of both these measures to a subclinical
apnea population would seem to adequately assess daytime
sleepiness in these subjects.
Measurement of Hypoxia and Its Sequelae
Moctural desaturation is frequently noted in SAS
patients. Therefore, it seems in order to examine the
known sequelae of hypoxia of other etiologies in an
attempt to determine variables or behaviors likely to be
sensitive to the hypoxia found in SAS and SAA. Research
on hypoxia varies systematically along two dimensions,


4
syndrome typically results in complications such as right
sided congestive heart failure, pulmonary hypertension,
and peripheral edema (Block, 1980). An exacerbation of
disordered breathing (apnea) as well as hypercapnia and
hypoxemia (often severe) occurs with sleep onset. Treat
ment consists of oxygen therapy, respiratory stimulants,
and maintenance of wakefulness.
The discovery of a sleep induced exacerbation of the
symptoms accompanying Pickwickian Syndrome spurred a
systematic investigation of sleep respiratory events
(Gastaut et al., 1965). The new emphasis on noctural
respiration led to a series of investigations which docu
mented sleep induced respiratory dysrhythmias in non-obese
subjects (Lugarasi et al., 1966). A new group of sleep
disordered breathers designated as Sleep Apnea Syndrome
emerged, with several features which distinguished them
from Pickwickians. Among these differences were weight
(i.e., SAS patients were not necessarily obese), daytime
respiratory control (i.e., SAS patients were typically not
hypercapnic during waking as were Pickwickians), sex
(i.e., while A0% of Pickwickians were female, very few
premenapausal females have been identified with SAS), and
age (i.e, SAS patients are typically younger than
Pickwickians)(Block, 1980).
Two large samples of SAS patients have been reported
in the literature, allowing characterization of typical
SAS patients. Guilleminault et al. (1976) reported on a


54
included memory, visuo-organizational skills, verbal
fluency, and motor coordination. The chosen neuropsycho
logical battery reflected these two aims. Thus the WAIS
tapped intellectual functioning; the Wechsler Memory
Scale with delayed recall of logical stories and visual
reproductions as well as digit span and Rey Figure heavily
sampled verbal and non-verbal memory; the Hooper tapped
visuo-organizational skills; Verbal fluency sampled lan
guage skills; the Wisconsin card sort tapped frontal
sell-regulatory skill; and Finger tapping sampled motor
functions. The chosen battery provided a broad screening
as well as a detailed assessment of neuropsychological
functions thought to be impaired by hypoxia.
Procedure
Subjects followed the schedule appearing below for
the experimental night:
1800 Sign informed consent, blood pressure,
height/weight, finger tapping, verbal fluency,
subjective sleepiness (SSS-made every V2 hour
till bedtime)
1830 Wiring for EEG
1900 Begin HSLT I (20 m)
1930 WAIS
2000 VMS, Hooper
2030 Graphesthesia, Luria motor programs
2100 Begin MSLT II (20 m)
2130 Rey figure, Wisconsin card sort


14
normals might experience apneas during REM without
pathological consequence. Therefore, data available on
these two issues will be reviewed, including only studies
with a sample size great than 10 (with one exception).
Reports on normal subjects below 30 years of age are
somewhat rare, although some do exist. Bixler et al.
(1982) included a group aged 18-29 years (M = 13, F =
17). One subject in this group had sleep apneas (12) with
a mean duration of 15 seconds. Bixler et al. noted that
most apneas occurred in Stages 1 and REM sleep. Part of
the sample of Block et al. (1979), the females, had a mean
age of 29. Three of these subjects showed apneas with a
mean duration of 14 seconds. Again, a disproportionate
number of apneas occurred in Stages 1 and REM. Thus in
normal subjects under 30, apneas are somewhat rare, with a
mean duration of approximately 15 seconds when they do
occur. Most apneas occur in light and REM sleep.
More data are available on normal subjects between 30
and 60. Guilleminault et al. (1978) do not report on
duration of apneas in their sample aged 40-60, but they
note that apneas occurred only in Stages 1 and REM.
Bixler et al. (1962) reported that, in their 30-44 year
old group, 5 subjects experienced sleep apnea with a mean
duration of 13.1 seconds. A disproportionate number of
apneas occurred in Stages 1 and REM. Block et al. (1979)
studied 30 males with a mean age of 38, with 12
experiencing sleep apneas. These apneas had a mean


72
awakenings, F(2,43)=0.9 p-NS: Minutes of awakening,
F(2,43)=0,e, p=NS; Sleep latency, F(2,43)=0.01, p=NS;
Depth of sleep, F(2',43)=2.6, p*NS; Daytime sleepiness,
F(2,43)=1.8, p=NS].
Means were calculated from the 7 day sleep logs, and
the following variables were selected for analysis:
reports of daytime sleepiness, number of naps, minutes
napping, total bedtime, sleep latency, number of
awakenings, minutes of awakenings. These variables were
correlated with the respiratory indices to form a Pearson
correlation matrix totalling 49 correlations. Table 3-11
presents the 6 significant (p<.05) correlations which
emerged. Again, number of desaturations were associated
with reports of napping, while mean high saturation was
associated with total bedtime, sleep latency, and minutes
awake, and mean low saturation was associated with minutes
awake as well.
Table 3-12 presents the means and standard deviations
for the subjects divided into levels of apnea/hypopnea.
Oneway ANCVA procedures failed to reveal significant
between group differences [Daytime sleepiness,
F(2,35)-0.6, p=NS; Number of naps, F(2,35)=0.3, p=NS;
Minutes of napping, F(2,35)-1*3> p=NS; Total bedtime,
F(2,35)=0.7, p=NS; Sleep latency, F(2.35)0.2, p=NS;
Number of awakenings, F(2,35)=0.2, p=NS; Minutes of
wakenings, F(2,35)=.0001, p=NS].


127
Guilleminault, C., van den Hoed, J., & Mitler, M. (1978).
Clinical overview of the sleep apnea syndromes. In C.
Guilleminault & V. Dement (eds), Sleep Apnea
Syndromes. New York: Alan R. Liss Inc.
Guilleminault, C., Simmons, F., Motta, J., Cumminsky, J.,
Rosekind, M., Shroeder, J., & Dement, W. (1981).
Obstructive sleep apnea syndrome and tracheostomy.
Archives of Internal Medicine, 141, 985-988.
Guilleminault, C., Tilkian, A., & Dement, W. (1976). The
Sleep Apnea Syndromes. Annual Review oi Medicine,27,
465-470.
Halstead, W. (1947). Drain and Intelligence. Chicago:
University of Chicago Press.
Hanbauer, I., Karoum, F., Hellstrom, S., & Lahiri, S.
(1981). Effects of hypoxia lasting up to one monrh on
the catecholamine content in rat carotid body.
Neuroscience, 6., 81-86.
Harmon, E., Wynne, J., Block, A., & Mailoy-Fisher, L.
(1981). Sleep-disordered breathing and oxygen
desaturation in obese patients. Chest, 79, 256-260.
Hartse, K., Roth, T., & Zorrick, F. (1982). Daytime
sleepiness and daytime wakefulness: The effect of
instruction. Sleep, , 107-118.
Hartse, K., Zorick, F., Roth, T*, Kaffeman, M., & Hoyles,
T. (1980). Daytime sleep tendency in normal,
insomniac, and somnolent populations. Sleep Research,
9., 204.
Hartse, K., Zorick, F., Sicklesteel, J., Piccione, P., &
Roth, T. (1979). Nap recordings in the diagnosis of
daytime somnolence. Sleep Research, _6, 190.
Hebb, D. (1961). Distinctive features of learning in the
higher animal. In D. Delafresneye (ed), Brain-
Mechanisms and Learning. London: Oxford University
Press.
Hoddes, E., Dement, W., & Zarcone, V. (1972). The
development and use of the Stanford Sleepiness Scale.
Psychophysiology, 15C-160.
Hoddes, E., Zarcone, V., Smythe, H., Phillips, R., & Dement,
Vf. (1973). Quantification of sleepiness: A new
approach. Psychophysiology, 10, 431-436.


31
been matched on number of apnea events. Asymptomatic
(nonhypersomnolent) subjects fell from a baseline 02
saturation of 80 mm (Hg) to a mean maximum desaturation of
54 mm (Hg) during sleep. Symptomatic subjects fell from a
baseline of 54 mm (Hg) to a mean maximum desaturation of
35 mm (Hg) while asleep. Orr et al. suggested that the
more severe desaturation which accompanied apneas in
symptomatic subjects underlaid their somnolence and other
complications. However, these conclusions are vulnerable
to certain methodological criticisms, as Elock (1980)
points out. Specifically, periodic sampling may well fail
to detect the multiple brief desaturations associated with
SAS, rendering this technique potentially insensitive to
important events.
In an attempt to provide more representative data on
desaturation, Block et al. (1979) utilized an ear oximeter
in studying saturation in normal subjects. An ear oxi
meter provided a continuous, accurate readout of moment to
moment arterial oxygen levels. Thirty males (mean age 38)
and 19 females (mean age 27) were studied. Of the 30
males, seventeen suffered at least one episode of desatu
ration, falling from a mean saturation of 95% to a maximum
desaturation of 84% during sleep. In contrast, no epi
sodes of desaturation were noted in the sample of premena-
pausal females (baseline 02 saturation: 96%). The oxygen
desaturations of the males were always found in asso
ciation with breathing abnormalities or snoring. A study


107
aesaturaticns exceeding 4 and
relative ease of this rating,
saturation records.
10% was kept. because of the
only one rater scored


3-11 Significant (p<.05) Pearson correlations
between nocturnal respiratory and sleep log
variables in 46 snoring males 73
3-12 Means and standard deviations of sleep log
variables in 46 snoring males grouped by
level of apnea/hypopnea 73
3-13 Multivariate regression of demographic and
nocturnal respiratory variables on cognitive
scores in 46 snoring males 76
3-14 Significant (p<*05) partial correlations,
controlling for age, between nocturnal
respiratory and cognitive variables in
46 snoring males 79
3-15 Means and standard deviations of cognitive
scores in 46 snoring males grouped by
level of apnea/hypopnea 80
3-16 Multivariate regression of sleep and
nocturnal respiratory scores on cognitive
variables in 43 snoring males 83
B-1 Demographic and medical variables from 46
snoring males 109
B-2 Daytime sleepiness variables from 46 snoring
males 111
B-3 Sleep questionnaire variables from 46 snoring
males 115
B-4 Seven day sleep log means from 46 snoring
males 115
B-5 Neuropsychological variables from 46 snoring
males 117
B-6 Respiratory variables from 46 snoring males...,119
B-7 Sleep variables from 46 snoring males 121
vi


7
One sampling strategy which has been pursued is to
exclude subjects who complain of sleep or respiratory
difficulties. Screening for sleep complaints varies
widely. Guilleminault et al. (1978) studied a sample of
20 "normals" (presumably noncomplaining subjects). These
subjects ranged in age from 40-60 years. Males showed a
mean number of apneas of 7 (range 1-25), while females
showed a mean number of apneas of 2 (range 0-5). Thus,
none of these noncomplaining subjects presented a clinical
syndrome of sleep apnea, although sleep apneas did occur
apparently without pathology. Block et al. (1979) studied
49 subjects (M = 30, F = 19). Males in the sample had a
mean age of 38, while females had a mean age of 29. Any
potential subjects who complained of breathing difficul
ties or sleep disturbances were excluded from the study.
Twelve males (40$) had episodes of sleep apnea while only
3 females (15$) suffered as well. For those subjects,
males had a mean of 4.2 episodes, while females had a mean
of 3. Within this sample, the number of apneas was not
significantly correlated with increasing age, although
occurrence of oxygen desaturation (a concomittant of
disordered breathing) was positively correlated with
age. Additionally, although the mean number of sleep
apneas was not significantly different between sexes (M =
4.2, F = 3), the more sensitive measure of apneas with
desaturation was (M =3, F = 0). Webb (1974) studied 2C
males with a mean age of 44. Subjects were excluded from


77
Table 2,-15-continued.
Criterion
Predictor
Beta
Sig.
Wechsler Memory
Age
.0272
.852
Scale Memory
Weight¡Height Ratio
.0996
.546
Quotient
Apnea Index
-.3315
.024
Apnea+Hypopnea Index
-.3421
.049
Mean High Saturation
Number of
.2384
.124
Desaturations >4%
-.5037
.007
Delayed Recall
Age
-.3254
.018
Visual
V/eight:Height Ratio
-.1214
.416
Reproductions
Apnea Index
-.4037
.005
Apnea+Hypopnea Index
-.2026
.191
Mean High Saturation
Number of
.1368
.325
Desaturations >4%
-.2888
.078
Delayed Recall
Age
-.3550
.010
Rey Figure
V/eight ¡Height Ratio
.1795
.251
Apnea Index
-.2206
.091
Apnea+Hypopnea Index
-.0572
.709
Mean High Saturation
Number of
.3191
.025
Desaturations >_4%
-.3228
.049
Hooper Test
Age
-.5658
.000
V.eight:Height Ratio
.1202
.405
Apnea Index
-.1428
.251
Apnea+Hypopnea Index
.1478
.318
Mean High Saturation
Number of
.1072
.421
Desaturations >4%
-.2042
.190
\vi scon sin Card
Age
-.3409
.027
Sort
V/eight ¡Height Ratio
.0258
.887
Apnea Index
-.3205
.031
Apnea+Hypopnea Index
.1029
.550
Mean High Saturation
Number of
.2383
.130
Desaturations _>4^
-.0801
.657


Table B-
1. Demographic ana med
ical vai
r i a b 1 e s
from 46
snoring
males.
Weight:
Corneli He
dical index
Sub j.
bduca-
Height
Height
Blood
Hressure
Overa 1 i
Kes-
Cardio
heuro-
a
Age
tion
(in)
Weight
Hu tio
biastolic
Sys to!ic
score
pira t.ory
pulmonu ry
logica 1
i
58
12
00
250
5.67
130
090
2
00
12
U5
225
5.57
140
100
022
01
\j
02
5
52
10
07
148
2.10
110
065
020
01
00
01
4
bb
15
70
106
2.4
140
100
007
00
01
OO
5
52
12
74
245
3.31
120
085
018
03
00
U1
0
50
10
07
155
2.51
110
060
015
00
02
02
7
51
12
09
155
1.95
130
085
00/
oo
00
01
b '
4
09
07
155
2.31
120
090
005
01
02
00
y
00
20
O0
157
2.57
130
100
014
00
02
05
10
50
10
08
255
3.45
155
090
005
00
00
02
11
49
18
70
250
5.28
120
100
030
01
02
01
12
5b
15
74
250
5.10
125
090
UOO
00
01
01
15
48
15
70
187
2.67
120
070
007
00
01
01
14
58
15
75
190
2.55
120
070
005
01
00
-Cu
15
54
14
07
125
1.99
114
076
000
00
oo
01
1o
50
10
71
100
2.25
150
080
014
02
04
01
17
32
1b
72
107
2.31
120
085
005
01
00
02
1b
5b
1b
b9
180
2.00
120
090
0;50
02
01
01
19
57
12
72
190
2.71
120
090
015
01
02
01
2U
50
16
71
250
3.52
155
100
042
06
02
Ol
21
47
14
71
198
2.79
125
090
007
00
02
02
22
66
14
05
155
2.04
120
ObO
018
02
00
CO
25
55
16
72
145
2.01
120
074
013
01
01
01
24
51
10
72
105
2.29
115
07U
010
04
oo
02
25
51
15
00
100
2.35
122
004
025
00
01
04
2b
04
10
00
215
1o
149
100
u10
U1
C1
02
27
75
12
00
185
2.09
14u
090
020
0.-)
00
01


36
patients, i.e., response bias and altered frame of
reference (Dement et al., 1978).
It was clear then that an objective measure of day
time somnolence was needed. With this goal in mind,
Carskadon and Dement (1977) introduced the Multiple Sleep
Latency Test (MSLT). This technique involved measuring
the objective sleep latency during multiple nap attempts
throughout the day. In a typical paradigm, a subject
would be wired to an EEG and attempt to fall asleep at 2
hour intervals through the day. The latency to sleep in
each nap was measured by determining the latency to the
first epoch of Stage 1 sleep. In order to prevent signi
ficant amounts of sleep accumulating during testing, sub
jects were awakened after an epoch of sleep was
observed. If no sleep was noted, subjects were discon
nected after 20 minutes and allowed to resume ether acti
vities. Carskadon and Dement believe that this measures
underlying physiological sleep tendency. This underlying
tendency is thought to be modulated by alerting stimuli in
the environment (Carskadon and Dement, 1982a). The vali
dity of the test was established on normal subjects, who
exhibited a biphasic curve of sleep tendency throughout a
24 hour period. Extension of sleep over typical lengths
in normals (who are thought to be moderately chronically
sleep deprived) saw a significant rise in average sleep
latencies on the MSLT (Carskadon and Dement, 1979b), while
chronic sleep restriction (5 hrs sleep per night) caused


Table 3-7. Means and standard deviations 01 selected sleep variables from 43
snoring males grouped by level oi' apnea/hypopnea.
Mo Apnea/
Low Apnea/
High Apnea/
Hypopnea
Hypopnea
Hypopnea
Time in Bed
376.1 (45.3)
382.2
(47.4)
361.2 (38.0)
Sleep Latency
20.2 (12.7)
38.3
(43.5)
20.4 (18.7)
Pure Sleep Time
292.7 (73.5)
290.9
(81.5)
265.4 (69.5)
Sleep Efficiency Index
.77 (.17)
.70
( .16)
.73 (.14)
Number Stage 0 Periods
9.5 (7.6)
7.7
(3.2)
9.2 (5-5)
Time % Stage 0
15.8 (12.9)
11 .6
(12.4)
20.9 02.6)
Time % Stage 1
4.1 (3.7)
2.6
(1.8)
3.5 (2.0)
Time % Stage 2
61.3 (8.0)
60.3
(12.9)
61.3 (1.3)
Time % Stage 3
2.8 (1.9)
4.3
(2.5)
1.5 (2.3)
Time % Stage 4
8.5 (7.2)
8.5
(5-6)
3.6 (5-1)
Time % Stage REM
9.7 (6.C)
13.3
(7.8)
9.1 (7.6)
Latency 1st REM Period
136.5(106)
107.5
(71.5)
156.4(125.4)
Mean REM Period Length
13.9 (0.8)
16.1
(9.8)
13.5 (11.1)
REM Cycle Length
80.0 (56.1)
78.7
(61.7)
41.6 (42.7)
% Slow Wave sleep
12.8 (8.0)
14.6
(7.4)
5.9 (0.1)
Note: see text for explanation of N.


23
I had >100 apneas per night, Group II had >30 but <100
apneas per night, while Group III had less than 20 apneas
per night. Thus Groups I and II were diagnosable as SAS
(>30), while Group III was not. Obesity was present in
Groups I and II, but in only 3 subjects in Group III.
However, these differences were not statistically signifi
cant.
Taken together, this evidence suggests a positive
relationship between weight and apnea in the general popu
lation. However, SAS patients are not invariably obese.
Further, weight loss fails to result in a decrease in
apneas in many SAS patients. Thus the relationship
between obesity and full blown SAS is less than clear.
The preceding pages have investigated the currently
understood role of several variables (sex, age, snoring
and weight) in sleep apnea activity. Presently available
data suggest that in young and middle aged subjects, apnea
is a predominantly male phenomenon, although post mena-
pausal females may close the gap. Increasing age is re
lated to sleep apnea activity. Snoring, which occurs in a
significant percentage of the general population, is
thought to be a necessary but not sufficient condition for
SAS. Weight was shown to have a positive relationship to
sleep apnea activity in the general population, although
its presence or role in SAS is equivocable.
These conclusions have important implications for
experimental designs directed at understanding sleep apnea


REFERENCES
Agnev, H.W., & Webb, W.E. (1972). Measurement of sleep
onset by EEG criteria. American Journal of EEG
Technology, 12, 127-134.
Agnew, H.V., Webb, W.B., & Williams, R.L. (1966). The
first night effect: An EEG study of sleep.
Psychophysiology, _2, 263-260.
Ancoli-Israel, S., Kripke, D., Mason, V., & Kaplan, W.
(1982). A prevalence study of sleep disorders in
seniors: Six-months results. Abstract published for
1982 APSE meeting, San Antonio.
Ancoli-Israel, S., Kripke, D., Mason, W., & Messin, S.
(1981). Sleep apnea and nocturnal myoclonus in a senior
population. Sleep, _4(4), 349-338.
Annau, Z. (1972). The comparative effects of hypoxic and
carbon monoxide hypoxia on behavior. In B. Weiss & U.
Laites (eds), Behavioral Toxicology. New York: Plenum
Press.
Appel, D., Schmidt-Nowara, V., Poliak, C., & Wietzman, E.
(1982). Effects of tracheostomy closure on sleep and
breathing in sleep apnea patients with long term
tracheostomy. Abstract published for APSS meeting 1982,
San Antonio, p. 131.
Association of Sleep Disorder Centers (1979). Diagnostic
classification of sleep and arousal disorders. Sleep,
2_, 1-137.
Bartus, R., Dean, R., & Bear, B. (1983). The cholinergic
hypothesis of impaired memory in the aged. In D. Harman
& K. Gray (eds), Aging: Nutrition anc Aging. New
York: Raupa Press.
Berg, I. (1948). A simple objective technique for
measuring flexibility in thinking. Journal of General
Psychology, 39 15-22.
Berry, D., l. Webb, W.B. (In Press). Sleep and cognitive
variables in normal aging subjects. Journal of
Gerontology
123


Table l-6
Respiratory variables from 46
Subject Apnea
ij_ Index
1
0.0
2
1.7
3
0.5
4
2.3
5
0.4
6
3*4
7
0.2
8
0.0
9
3*6
10
0.5
11
0.0
12
0.3
13
13.0
14
0.8
15
0.2
16
0.8
17
0.2
10
0.2
19
0.0
20
0.7
21
0.0
22
0.0
23
0.0
24
0.0
Apnea+
Hypopnea
Index
Seconds
in
Lvents
38.4
4058
4.2
350
0.3
36
2.5
144
0.9
44
3*7
500
0.2
10
0.0
0
4.8
522
0.5
10
0.0
0
0.9
109
13*0
1498
0.8
36
0.2
23
0.8
48
0.2
21
0.2
10
0.0
0
0.7
40
0.2
20
0.0
0
0.0
0
O.C
0
rin£ males
Rumber of
Mean
Saturation
Desatura
tions
Rifh
Low
>4%
>105-
93*4
80.9
220
73
93*4
88.7
214
21
94.9
93.0
26
0
96.9
94.0
74
2
95*6
92.0
123
2
95.8
91.7
90
7
95.6
91.7
5
0
93.6
94.9
0
0
98.4
95-4
145
9
95.0
94.2
2
0
91.6
90.0
22
0
95.1
92.1
148
3
94.3
90.0
3
0
95.1
93.7
10
0
95.6
94.5
4
0
95.1
94.1
2
0
95.4
94.1
7
0
94.3
93.2
0
0
96.5
95.1
18
0
95.1
93.3
49
0
94.7
92.3
60
2
94.9
94.3
1
0
95.4
94.9
0
0
92.8
92.0
1
0
\D


69
Daytime Sleepiness Data
Included among the measures of daytime sleepiness
were the mean of the several Stanford Sleepiness Scale
ratings, and two variables calculated from the two evening
naps: mean sleep latency and total minutes asleep during
the naps. These variables were correlated with the
respiratory variables, resulting in a Pearson correlation
matrix. From these 21 correlations, two emerged signi
ficant at p<.05 or less, both involving the mean Stanford
Sleepiness Scale ratings. Mean SSS ratings were corre
lated wixh apnea index (r=.297)* and apnea + hypopnea
index (r=.283).
Table 3-6 presents the means and standard deviations
for the subjects grouped by level of apnea/hypopnea. One
way ANOVA procedures were non-significant for all these
variables [SSS: F(2,38)=2.4, p=N5; Mean sleep latency:
F(2,35)-0.4, p=NS; Total minutes asleep: F(2,35)=0.3,
p=NS].
Table 3-8. Means and standard deviations for daytime
sleepiness variables in 46 snoring males grouped
by level of apnea/hypopnea.
No Apnea/
Low Apnea/
High Apnea/
Hypopnea
Hypopnea
Hypopnea
Stanford Sleepiness
Scale
3.2 (.7)
3.6 (.e)
4.0 (.7)
Mean Nap Latency
Total Minutes of
14.5 (5.7)
15.5 (5.3)
13.4(3.4)
Sleep in Naps
11.3(11.9)
9.6(10.6)
13.5(6.6)


57
The respiratory records were annotated by two hour
intervals following bedtime. From a total of 652 events
scored, 150 (23%) occurred in the first two hours
following bedtime, 185 (23%) occurred in the second two
hours, 232 (36%) occurred in the third two hours, while 83
(12%) occurred in the final two hours. It should be noted
that vakeup times varied considerably, making interpre
tation of the attenuated number of events in the final two
hours somewhat problemmatic.
From a total of 578 events in which reliable sleep
staging was achieved, 459 (78%) occurred in light slow-
wave sleep (stages 1-2), while 118 (20%) began in REM
sleep. Only 6 events (01%) occurred in slow wave sleep.
These precentages must be considered in light of the
relative distribution of sleep stages in these subjects:
stages 1-2 (64%), stage REM (12%), stages 3-4 (10%), stage
0 (14%). This distribution suggests that the sleep time
spent in stages, 1, 2, or REM must have provided the
overwhelming preponderance of events in these stages.
Sleep characteristics will be discussed in greater detail
below.
As noted below, 28 of the 46 subjects experienced at
least one episode of apnea or hypopnea. Table 3-2
presents the data on occurrence of at least one
apnea/hypopnea, as well as the occurrence of a high level
of apnea/hypopnea. Almost 2/3 of the sample had ax least
one apnea or hypopnea, while 13% had high levels of


94
correlations in this sample. Additionally, an EEC scoring
system which averages across one minute periods might
obscure brief sleep fragmentation from nocturnal events.
Although the lack of correlation is puzzling, Bixler et
al. (1982) reported similar findings. In their normal
sample, few apnea induced arousals were noted. Multiple
night recordings of all these parameters appear necessary
to clarify the possible interaction o respiration and
sleep fragmentation in a first night effect.
Daytime Sleepiness, Sleep Questionnaire, and
Sleep Lop Data
The three assessments of subjective aspects of sleep
and sleepiness will be discussed together, as the pattern
of findings lend themselves to this organization. The
core finding here is a relationship between nocturnal
respiratory variables and sleepiness variables manifesting
itself in two forms. Daytime sleepiness, reflected in the
mean SSS ratings and self report of napping on both the
sleep questionnaire and logs, was positively related to
various indices of nocturnal events (apnea index, apnea +
hypopnea index, seconds in events, and number of
desaturations). A second aspect of this relationship
involved relationships between decreased saturation and
reports of decreased sleep latency and nightly wakenings
on the sleep logs and questionnaires.
These seemingly disparate findings may be interpreted
as the sequelae of sleep oisruption resulting from


60
In summary, non-parametric analyses failed to reveal
significant relationships between age or weight and
overnight respiration. However, Pearson correlations
suggestea that increasing weight is associated with
deterioration of several noctural respiratory indices.
Because increasing apnea/hypopnea levels have been
suggested to carry increasing risks for various deficits,
the sample was stratified by level of apnea/hypopnea. The
first group included those subjects without events (Ho
events), those with at least one event but less than 5
events per hour (Low apnea/hypopnea), and those with 5 or
more events per hour (High apnea/hypopnea). It should be
noted that this division of subjects is slightly different
from the earlier non-parametric analyses, because in this
grouping low apnea/hypopnea and high apnea/hypopnea are
mutually exclusive, while in the earlier analyses the
group with at least one apnea/hypopnea included those with
more than 5 events per hour.
Table 3-4 utilizes oneway ANOVA procedures to compare
these groups on demographic and nocturnal respiratory
variables. There were no significant between group
differences on any of the demographic factors: age,
weight, weightiheight ratio, and education [F(2,43)-2.1,
0.6,0.8,0.7; p=US]. Logically, both apnea and hypopnea
indices were significantly different between groups
[F(2,43)=9-6, p<-C004; 20.5, pC.0000]. Followup
comparisons (Scheffe) indicated that the high


50
Co). All physiological variables were recorded on the
polygraph's chart paper which was run at a speed of 1C
ram/s.
Measures
Two classes of independent variables are identified,
with the first being thumistor variables (those derived
primarily from the nose and mouth thumistors). These
included number of apneas and hypopneas, mean duration of
apneas and hypopneas, mean low oxygen saturation in apneas
and hypopneas, mean oxygen saturation change in apneas and
hypopneas, and total seconds spent in apneas and hypo
pneas. Thumistor variables were separately scored by two
trained raters, whose agreement appears in Table 2-1. An
overall agreement of 85# was achieved. Disagreements were
resolved by consensus scoring. The second class of
independent variables included oxygen saturation variables
(derived from a minute by minute rating of highest and
lowest oxygen saturation from the overnight record).
These variables included mean highest saturation, mean
lowest saturation, and number of desaturations of 4 ana
10# or more. In selecting these scoring criteria, several
considerations were relevant. Initially, a system with
reasonable interrater reliability was necessary. While
published criteria for scoring apneas were not problem
atic, reliable scoring of hypopneas was not achieved until
a desaturation criterion of 10# was introduced. In
selecting measures of oxygen saturation, a mean high and