Title: Apnea and bradycardia elicited by facial airstream stimulation in healthy infants in the first year of life
Full Citation
Permanent Link: http://ufdc.ufl.edu/UF00098625/00001
 Material Information
Title: Apnea and bradycardia elicited by facial airstream stimulation in healthy infants in the first year of life implications for detection of infants at risk for sudden infant death syndrome
Physical Description: x, 145 leaves : ill. ; 28 cm.
Language: English
Creator: Hurwitz, Barry Elliot, 1956-
Publication Date: 1984
Copyright Date: 1984
Subject: Sleep apnea syndromes   ( lcsh )
Bradycardia   ( lcsh )
Sudden infant death syndrome   ( lcsh )
Psychology thesis Ph. D
Dissertations, Academic -- Psychology -- UF
Genre: bibliography   ( marcgt )
non-fiction   ( marcgt )
Thesis: Thesis (Ph. D.)--University of Florida, 1984.
Bibliography: Bibliography: leaves 125-142.
General Note: Typescript.
General Note: Vita.
Statement of Responsibility: by Barry Elliot Hurwitz.
 Record Information
Bibliographic ID: UF00098625
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: alephbibnum - 000479215
oclc - 11795296
notis - ACP5939


This item has the following downloads:

PDF ( 6 MBs ) ( PDF )

Full Text







Dedicated with love to my wife, Debbie

and my son, Joseph


I thank Dr. W. Keith Berg for his guidance, astute

judgement and his infectious positive approach to life.

His active support of our past and present collaborative

efforts has meant a great deal to me. I look forward to a

long and fruitful future association. As well, I thank you

for so willingly being there whenever I needed you and for

patiently allowing me to interrupt your stories of the joys

of fatherhood in order to tell you mine.

I would also like to express my appreciation to my

doctoral committee, Drs. Merle E. Meyer, Peter Lang, Phil

Posner and Neil Rowland, for their assistance throughout my

graduate training. Also, I thank Drs. Don Stehouwer and Paul

Davenport for their helpful comments on this project.

In addition, I thank all those infants who fell asleep

for me and most notably I thank all those little kids who

held their breath for me and even those that didn't. Also,

I am grateful to the parents of the little kids for

generously contributing their time to this effort.

Special thanks are extended to my assistants, Alonso,

Judy, Kristen, Karen, Julio and Yolanda, for their help in

data collection at various times throughout the past two

years. I thank my lab mates Terry Blumenthal, Margaret

Davies and Allen Klempert for their enjoyable association

and patient tolerance.

I thank my parents and the rest of my family for their

tenacious support and everlasting expression of faith and

encouragement. I stand tall because of them. Finally, to

my wife and true friend Debbie, thank you for being there

when I wasn't and also when I was, and more importantly

thank you for my son, Joseph.


ACKNOWLEDGEMENTS. . . . . . . . . . . iii

LIST OF TABLES. . . . . . . . . ... ... .vii

LIST OF FIGURES . . . . . . . ... . viii

ABSTRACT. . . . . . . . . ... . . .ix


1 INTRODUCTION. . . . . . . . . . 1

Epidemiology. . . . . . . . . . 3
Pathology . . . . . . . . ... . 6
Cardiorespiratory Physiology. . . . . . 9
Control of Cardiac Function . . . . .. .10
Control of Respiratory Function . . . .. .13
Respiratory Patterns. . . . . . .. .13
Frequency of apnea. . . . . . .. .15
Control of ventilation. . . . . ... 20
Sleep and Breathing . . . . . . .. 24
Cardiorespiratory Responses To Airstream
Stimulation . . . . . . . . 30

2 METHODS . . . . . . . . ... . 38
Subjects. . . . . . . . .. . . 38
Apparatus . . . . . . . . . . 43
Experimental Manipulations and Design . . .46
Procedure . . . . . . . . . . 47
Data Quantification . . . . . . .. 49
Notes . . . . .. . . . . . . 52

3 RESULTS . . . . . . . . . . 53
Sleep State Incidence and Pattern . . . .. .53
Large Movement Incidence. . . . . . ... 55
Respiration Response to Airstream Stimulation 59
Percent of Elicited Apnea . . . . .. 59
Effect of airstream stimulation site. ... .59
Effect of sleep state . . . . . .. .62
Effect of preceding large movement. ..... 62
Latency to Apnea Onset. . . . . . .. .66
Effect of airstream stimulation site. ... .66
Effect of preceding large movement. ... .66

Duration of Apnea . . . . . . .. 68
Effect of airstream stimulation site. .. .. 68
Effect of preceding large movement. .. .... 71
Summary of the Respiration Response to
Airstream Stimulation . . . . . .. 73
HR Response Topography . . . . . . 74
Prestimulus Levels . . . . . . 75
Post-stimulus Changes . . . . . .. 75
HR response to the airstream stimulus:
face vs. abdomen. . . . . . ... 78
HR response to the facial airstream
stimulus: apnea vs. no apnea. . . .. 81
HR Response to the Facial Airstream Stimulus
from Apnea Onset . . . . . . .. 86
Notes . . . . . . . . . . . 92

4 DISCUSSION . . . . . . . . .. 94
Implication for Dive Reflex as the Mechanism
for Inducing Apnea and Bradycardia. . . 97
Topography of HR Response in the Absence
of Apnea. . . . . . . . . . 104
Detection of Infants At-Risk for SIDS . .. 111
Note. . . . . . . . . .. . 115




C INFORMED CONSENT . . . . . . . .. 119

AND AS STATES . . . . . . . . .122

LARGE MOVEMENT. . . . . . . . . 125

STIMULUS: FACE VS. ABDOMEN. . . . . ... 127

REFERENCES . . . . . . . . . . .. 130

BIOGRAPHICAL SKETCH . . . . . . . . . 143


1 Description fo the Newborn Subject Pool . . .. 40

2 Description of the 2-4 Month-Old Subject Pool . . 41

3 Description of the 8-12 Month-Old Subject Pool. . 42

4 Percent of Trials in Quiet Sleep, Acitive Sleep and
Transitional Sleep States . . . . . 54

5 Percent of Trials with Large Movement Elicited by
Face and Abdomen Airstream Stimulation. . . .. 58

6 Percent of Trials with Apnea Elicited by Airstream
Stimulation of the Face and Abdomen When
Sleep State Is Ignored and When Infants Were
in Qjiet Sleep. . . . . . . . .. . 60

7 Percent of Trials with Apnea Elicited by Airstream
Stimulation of the Face and Abdomen When
Preceded by Large Movement and When Large
Movement Was Absent . . . . . . ... 63

8 Latency to Apnea Onset in Seconds as Elicited by
Airstream Stimulation of the Face and Abdomen . 67

9 Latency to Apnea in Seconds Elicited by Airstream
Stimulation of the Face When There Was Preceding
Large Movement and When Large Movement Was Absent 69

10 Duration of Apnea in Seconds Elicited by Airstream
Stimulation of the Face and Abdomen . . . .. 70

11 Duration of Apnea in Seconds Elicited by Airstream
Stimulation of the Face When There Was Preceding
Large Movement and When Large Movement Was Absent 72

12 Mean Prestimulus HR Levels in bpm in the Presence
and Absence of Apnea and Large Movement for Facial
and Abdominal Airstream Stimulation . . . .. 76


1 Mean Sleep State Rating for the First and Last Ten
Trials. . . . . . . . . ... . . . 56

2 Mean HR Response to Facial Airstream Stimulation When
No Apnea or Large Movement Were Elicited. . . .. 79

3 Mean HR Response to Abdominal Airstream Stimulation
When No Apnea or Large Movement Were Elicited ... . 80

4 Mean HR Response to Facial Airstream Stimulation When
Apnea Was Elicited Without Preceding Large Movement . 82

5 Mean HR Response to Facial Airstream Stimulation When
Apnea Was Elicited Without Preceding Large Movement
Adjusted from Apnea Onset . . . . . . . 88

6 Mean HR Response to Facial Airstream Stimulation When
Apnea Occurred Following Large Movement Adjusted from
Apnea Onset . . . . . . . .. . . . 89


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




August, 1984

Chairman: W. Keith Berg
Major Department: Psychology

Cardiorespiratory responses to airstream stimulation

(10-seconds, 1.06 psi) were observed in 52 newborn (NB),

2-4 and 8-12 month-old healthy infants. Two areas of

stimulation were compared: the facial region, which

included the forehead, eyelids and nose, and the abdominal

region rostral to the umbilicus, which served as a control

stimulation site. The infants were tested while asleep,

during which EKG, heart rate and respiration were recorded.

The results revealed that in both NB and 2-4 month

infants, facial cooling with an airpuff elicited apnea with

greater frequency and duration than did the control

airpuff. Apnea occurred on about 27.2% of facial trials

compared to 7.8% ot control trials. The duration of apnea

was twice as long in the 2-4 month infants as in the NBs.

By 8-12 months the facial airpuff still elicited apnea

more reliably than did the abdominal airpuff but with

reduced frequency of 8.9% and 1.9% respectively. The

duration of apnea, however, was maintained at the longer

duration seen with the 2-4 month infants. Bradycardia on

trials when apnea was elicited differed in magnitude

between age groups such that the 2-4 and 8-12 month infants

decelerated to about -15 bpm; whereas no bradycardic

response to apnea was displayed by the NBs.

It has been demonstrated that in healthy sleeping NB

and 2-4 month infants an airstream directed to the face can

elicit apnea. However, by 8-12 months facial cooling does

not have as great an effect on inhibition of respiration.

Although apnea can be elicited in the NBs by facial

airpuff, this apnea was brief and resulted in no change in

cardiac response. Therefore, at 2-4 months, which is the

age of greatest vulnerability for Sudden Infant Death

Syndrome (SIDS), the airpuff was likely to produce more

frequent, prolonged apnea with concomitant bradycardia.

It has been suggested that SIDS infants produce prolonged

apnea because they have a higher threshold for hypoxia.

Therefore, by eliciting apnea with facial cooling, it

may be possible to identify a predisposition to SIDS by

analysis of the magnitude of cardiorespiratory change in

infants with a history of prcionged apnea.



"and this woman's child died in the night,
because she overlaid it"

This quote, which is from an account in the Book of

Kings 3:16-28 circa tenth century B.C., in which the word

"overlaid" appeared to imply suffocation of an infant

during the night, has been cited as perhaps the first

recorded instance of Sudden Infant Death Syndrome (SIDS)

(Peterson, 1980). Despite it's probable antiquity, SIDS

was largely ignored by the scientific community until, as

infant mortality rates declined, it became apparent that

these deaths comprised a substantial proportion of all

infant deaths within the first year of life. Currently the

incidence of SIDS throughout the world varies from 0.6 to

3.0 deaths per 1000 live births, accounting for about

7-9,000 deaths each year in the United States alone

(Peterson, 1980; Valdes-Dapena, 1980). About one-third of

all deaths in infants between the ages of one week and one

year are diagnosed SIDS and it is the most frequent cause

of death in this age group (Moore, 1981).

Prior to 1963, deaths that occurred suddenly,

unexpectedly, and inexplicably with no recognizable lethal

pathology at autopsy were generally certified to one or

another cause-of-death category in keeping with the

attending coroner's personal inclination (Peterson, 1980).

This practice circumvented the embarrassment of a factual

"don't know" and at the same time complied with legal

expectations. Following the 1963 conference on causes of

SIDS, awareness and acceptance of sudden unexpected death

as a legitimate diagnostic entity increased. Finally the

designation, Sudden Infant Death Syndrome, emerged from the

Second International SIDS Conference in 1969, wherein an

occurrence of SIDS was defined as the sudden death of any

infant which was unexpected by history, and in which a

thorough post-mortem examination fails to demonstrate an

adequate cause for death (Bergman, Beckwith & Ray, 1970).

The following text will review certain epidemiological,

pathological and physiological findings related to SIDS as

background for the proposed study. For a comprehensive

review of SIDS, the reader is referred to these recently

published reviews (Brooks, 1982; Kelly & Shannon, 1982;

Shannon & Kelly, 1984; Valdes-Dapena, 1980). In addition,

as a basis for comparison, developmental findings in normal

healthy infants will be included where relevant. Based on

this information, an experimental methodology, which could

be used as a screening test for detection of infants at

risk for SIDS, will be proposed to examine the development

of the normal infant's cardiorespiratory response in the

first year of life.


Epidemiologic evidence has been gathered by the

scientific community in an effort to identify a population

of infants who might be at increased risk for succumbing to

SIDS. In the absence of pathological evidence which could

directly account for SIDS, the epidemiological findings

also have been instrumental in generating hypotheses about

the etiology of SIDS. It is important that any hypothesis

about the cause of SIDS account for these epidemiological

factors. It has been found that SIDS occurs while the

infant is asleep in the early morning hours, between

midnight and six a.m., with no cry or stridor (Peterson,

1980). The incidence of infant mortality due to SIDS peaks

between two to four months of age. Less than ten percent

of all SIDS deaths occur when infants are younger than two

weeks and older than six months of age (Brooks, 1982). It

should be noted, however, that the peak age for infant

mortality of known causes also occurs at two to four months

of postnatal age (Kelly & Shannon, 1982).

About 60% of SIDS victims are male and incidence for

nonwhites is twice that for whites; SIDS is more

common in low birth weight infants (i.e., 11/1000 live

births subsequent siblings of SIDS victims (i.e., 21/1000

live births) and in the surviving twin of SIDS victims

(i.e., 42/1000 live births) (Peterson, 1980). These data

suggest that either genetic factors have promoted the

demise of these infants or an environmental influence has

affected their intrauterine or extrauterine development. It

should be noted, however, that no difference was found

between the concordance of SIDS episodes in monozygotic and

dizygotic twin pairs; in addition, the incidence of SIDS in

full first cousins of SIDS victims did not differ from the

incidence in the general population (Peterson, 1980).

Thus, overall the data do not tend to support a genetic

component to SIDS.

There are several epidemiological perinatal-factors

which indicate an environmental influence on SIDS. SIDS

infants are born with higher incidence to young, unmarried

mothers of lower socioeconomic levels who have had poor

prenatal care, have short interpregnancy intervals, have

previous fetal loss and who are smokers or have narcotic

dependency (Valdes-Dapena, 1980). Other environmental

factors which may play a role in SIDS are altitude of

residence and air temperature. Specifically, SIDS incidence

increases as the altitude of residence increases and

the postnatal age of the infant at death decreased with

altitude from a mean of 17 weeks at 300 meters to nine

weeks at 1200 meters (Getts & Hill, 1982). SIDS exhibits a

seasonal variation being more apt to occur during the

colder winter months. This has been shown in countries

north of the equator peaking between November and March and

countries south of the equator peaking between May and

September (Moore, 1981; Peterson, 1980; Taylor, 1982).

Moreover, a spectral analysis of the SIDS incidence in

Los Angeles County from 1974-79 confirmed this seasonal

peak of SIDS in addition to revealing the coincidence of

decreased air temperature with increases in frequency of

SIDS mortality within this period (Hoppenbrouwers &

Hodgman, 1982).

On the basis of epidemiologic studies, several groups

of infants are believed to be at enhanced risk for SIDS and

have been most extensively tested by physiologic studies.

They are 1) preterm (i.e., low birth weight) infants with

histories of apnea (i.e., cessation of breathing) and

bradycardia; 2) infants who have had a near-miss or aborted

episode of SIDS, including apnea, cyanosis or pallor and

unresponsiveness that responds to intervention of

stimulation or resuscitation; and 3) siblings of SIDS

infants. There has been some controversy regarding the use

of terms like "near-miss" or "aborted" SIDS because they

imply with some certainty that the infant almost died when

in fact there is no way of knowing whether or not the

infant would have spontaneously resumed breathing without

assistance (Brooks, 1982). Therefore, the term "Apnea of

Infancy" or AI, as suggested by Brooks (1982), will be

adopted and used throughout to describe "an unexplained

cessation of breathing for 20 seconds or longer, or a

shorter respiratory pause associated with bradycardia,

cyanosis or pallor." Clearly this term describes a

clinical syndrome of respiratory abnormality similar to the

sleep apnea of adults or the apnea of prematurity.


Report of excessive frequency of prolonged apneic

episodes during sleep, in infants who subsequent to their

examination died of SIDS, promoted study of the hypothesis

that SIDS victims were experiencing recurrent periods of

apnea-associated hypoxia prior to their death

(Steinschneider, 1972). Among the plethora of hypotheses

about the cause of SIDS, most investigatory work in the

past decade has centered on the apnea hypothesis (Valdes-

Dapena, 1980). A number of post-mortem studies were

published shortly after Steinschneider's report which

supported the hypothesis that the SIDS is the terminal

event in infants who were suffering from a prolonged

hypoxic process, which possibly resulted from anomalies in

ventilatory control (e.g., Naeye, 1973; 1976). Based on

this hypothesis, one would anticipate anomalies in the anatomic

sites of ventilatory control (brain stem and carotid body),

as well as the organs which are usually affected by chronic

hypoxia (the brain, pulmonary vasculature, liver, adrenal

glands and periadrenal brown fat).

The main pathological findings have demonstrated that

SIDS victims differed from healthy infants who were victims

of accidental injury in that SIDS infants had hypertrophic

right ventricles in direct proportion to hypertrophy and

hyperplasia of pulmonary arteriolar smooth muscle,

increased volume of chromaffin cells in the adrenal

medulla, retention of periadrenal brown fat and presence of

hepatic extramedullary hematopoiesis (Naeye, 1973; Naeye,

1976; Naeye, Whalen, Ryser & Fisher, 1976). Moreover, upon

histologic analysis of SIDS cases, Naeye (1976) found

increases in brain stem astroglial density, particularly in

the midline area of the watershed zone of brain stem

circulation. Other recent findings of increased glomic

volume of the carotid body in those SIDS victims with

evidence of hypertrophy of the pulmonary vasculature have

been reported (Naeye, Fisher, Ryser & Whalen, 1976) and

reduction of myelinated fibers in the cervical vagus nerve

in some SIDS victims (Sachis, Armstrong, Becker & Bryan, 1981).

The finding of increased muscularity and hyperplasia

of the pulmonary arteries has been independently described

in SIDS victims (Shannon, 1980; Valdes-Dapena, Gillane,

Cassady, Catherman & Ross, 1980; Williams, Vawter &

Reid, 1979). These anatomic changes also have been induced

in animals exposed to repeated episodes of hypoxia

(Rabinovitch, Gamble, Nadas, Meittinen & Reid, 1979) and

are typical of children who die after prolonged exposure to

the hypoxic conditions of high altitude (Naeye, 1973). The

increased hypertrophy of the right ventricle may be a result

of the increased pulmonary arteriolar resistance, which forces

the heart to pump against a narrower opening. Brown fat

and hepatic hematopoiesis normally disappears within the

first year of life but its disappearance is delayed in SIDS

infants, which is a common finding also noted in autopsy of

hypoxic-children residing in high altitude (Naeye et al.,

1976). The astroglial proliferation in the brain stem has

affected areas such as the nucleus and tractus of

solitarius, nucleus ambiguous, raphe nucleus and dorsal

vagal nucleus, which are critical in the autonomic control

of the cardiovascular and respiratory systems. Similar and

more extensive CNS damage has been observed in post-mortem

examination of preterm infants with chronic hypoxia due to

respiratory distress syndrome (Brand & Bignami, 1969). In

addition, perinatal asphyxia for 10-15 minutes in monkeys

resulted in a pattern of brain damage which always

involved midline brain stem structures (Myers, 1972).

Takashima, Armstrong, Becker and Bryan (1978) have

postulated that periods of cerebral ischemia may accompany

the prolonged periods of apnea and hypoxemia in SIDS

infants, which in turn may result in the observed medullary

lesions and consequent astroglial proliferation. For this

to occur the medullary lesions would have to result from

discrete ischemia to brain stem otherwise more widespread

damage would be observed.

In sum, a case has been made for the etiology of SIDS,

whereby an infant acutely or chronically exposed to

hypoxic conditions may develop abnormal cardiorespiratory

control due to lesions in the critical brain stem areas.

The resultant prolonged periods of apnea or hypoventilation

could lead to further hypoxia and establish a positive

feedback loop of increasing pathologic tissue damage or at

least impaired function in regions of cardiorespiratory

control. Eventually, perhaps when challenged, the

cardiorespiratory regulatory mechanisms can no longer

compensate for the hypoxemic conditions and respiratory

failure results.

Cardiorespiratory Physiology

Since post-mortem studies can only provide clues to

the etiology of SIDS, physiologic studies have been

performed on living infants. What pre-mortem data on SIDS

victims are available were derived from studies of infants

who have died of SIDS subsequent to physiologic recording.

Other than this chance occurrence, there obviously can be

no systematic study of the physiologic responses of these

infants. However, investigators have studied infants

thought to be at risk for SIDS: infants with AI and

siblings of SIDS victims. They have typically been

contrasted with age-matched, healthy infants. It should

be noted that most studies of SIDS examine the target

physiological response longitudinally over the infant's

development. This is entirely appropriate because the risk

for SIDS seems to be a developmental phenomenon in which

the infant is at increasing risk from birth to 2-4 months

and then with maturation risk declines.

Control of Cardiac Function

Investigations of heart rate (HR) and HR variability

patterns have been performed to determine whether

manifestations of malfunctioning autonomic control

mechanisms could be detected in at-risk infants. During

maternal labor, fetal HR recordings of infants who

subsequently were SIDS victims revealed no differences from

control infants (Hoppenbrouwers, Zanini & Hodgman, 1979).

Siblings of SIDS infants, however, displayed greater fetal

HR variability and more frequent bradycardia than controls

but no difference in overall fetal HR levels when monitored

in-utero while their mothers slept (Hoppenbrouwers,

Hodgman, Harper & Sterman, 1981). The authors suggested

that the increased variability and bradycardic changes in

the siblings may reflect an increased level of intrauterine

hypoxic challenge in the siblings, whereas the effects of

hypoxia may have been concealed by the birthing process in

the SIDS infants. The results in the siblings, however,

should be viewed with caution since they are confounded by

maternal sleep disturbance, which was higher in mothers of

siblings of SIDS than control mothers.

The course of maturation of HR and HR variability for

healthy infants has been documented over the first six

months of life (Harper, Leake, Hodgman & Hoppenbrouwers,

1982). For HR levels, there is an increase in HR from

birth to one month after which rate declines steadily. In

contrast, HR was faster in both siblings of SIDS and in AI

infants during sleep, displaying a sharper rise and slower

decline in the first four months of life (Harper, Leake,

Hoppenbrouwers, Sterman, McGinty & Hodgman, 1978; Harper et

al., 1982; Leistner, Haddad, Epstein, Lai, Epstein, Eng &

Mellins, 1980). These differences were present only in

quiet sleep (QS) state (Harper et al., 1982); whereas

others observed these differences in both QS and active

sleep (AS) state (Leistner et al., 1980). For HR

variability in normals, sleep state had a more marked effect.

Awake variability was greater than asleep; during sleep, QS

was less variable than AS (Harper et al., 1982). Moreover,

during QS, HR variability declined rapidly up to one

month and then increased rapidly between one and three

months after which it became stable. In AS a steady

decline in variability was observed over the first six

months. In contrast, awake HR variability increased up to

two months and then rapidly declined between two and four

months after which it stabilized.

No difference has been observed between normals and

siblings of SIDS in the maturation of HR variability during

sleep up to six months of age, although risk infants

displayed reduced variability in the first two months when

awake (Harper et al., 1982). On the other hand, others

found less HR variability in AI infants than in healthy

controls during AS and QS between two and four months

(Leistner et al., 1980). Harper and colleagues (1978) have

attributed the increased HR levels in the at-risk infants

to a reduction in vagal tone. Leistner and colleagues

(1980) argue, instead, that the elevated HR levels are due

to enhanced sympathoadrenal drive. This is based on the

histologic evidence in SIDS victims of hypertrophy of the

chromaffin cell layer of the adrenal medulla, which may

indicate increased norepinephrine synthesis. Further

support of the latter hypothesis was derived from evidence

of a shorter Q-T electrocardiographic interval in AI

infants than controls over the first four months of life

regardless of state, which suggested sympathetic or adrenal

mediated enhanced-conduction (Haddad, Epstein, Epstein,

Mazza, Mellins & Krongrad, 1979). Although sympathoadrenal

mediation may be a possibility for the shortened Q-T

interval, ventricular repolarization may also be affected

by decreased vagal tone (Prystowsky, Jackman, Rinkenberger,

Heger & Zipes, 1981). Regardless of the autonomic

mediation, it still remains possible that the elevated HR

levels in the at-risk infants represent a compensatory

response to hypoxia; the increased cardiac output could act

to compensate for the oxygen deficiency.

Control of Respiratory Function

Respiratory patterns. The respiratory rates (RR) of

siblings of SIDS have been measured throughout the night

across the first six months of age (Hoppenbrouwers, Jenson,

Hodgman, Harper & Sterman, 1980). In both normals and

siblings, RR declined with increasing age. However,

regardless of sleep state, in each age group the RR rate of

the SIDS siblings was faster than the control infants, which

was interpreted as a compensatory response to hypoxia. Both

groups of infants displayed a developmental trend in RR

pattern throughout the night. In both AS and transitional

sleep (TS), normals and siblings displayed a U-shaped RR

pattern throughout the night at all ages. In QS this

U-shaped RR pattern did not appear in normals until three

months. Siblings, however, displayed this U-shaped pattern

prematurely at one month during QS. The authors posit that

this was due to a transient acceleration of maturation of

the nightly RR pattern in the siblings. Respiratory

variability decreased with increasing age in all infants

although siblings of SIDS displayed enhanced variability

in QS at three months.

In a parallel study of AI infants over the first six

months, the data do not entirely replicate these findings

(Hodgman, Hoppenbrouwers, Geidel, Hadeed, Sterman, Harper &

McGinty, 1982). There was no difference in RR or

respiratory variability between AI infants and controls,

although the decrease in RR with increasing age was

replicated. The only other longitudinal study of AI

infants observed faster RR and reduced tidal volume in the

AI infants at only one and two months. At three months no

differences were observed and at four months AI infants' RR

was less than controls (Haddad, Leistner, Lai & Mellins,


In studies on preterm infants who were presumably

hypoxemic, faster HR and RR than term infants has been

observed throughout the first six postnatal months, with

maximal differences between 2.5 and 3.5 months, the peak

period of SIDS vulnerability (Katona, Frasz & Egbert,

1980). It was concluded that since preterm HR and RR

maturational patterns were not delayed by the mean

difference in post-conceptional age, that both conceptional

age and postnatal experience were factors in the

development. Therefore, relative to healthy controls all

three SIDS risk groups displayed typical patterns of faster

RR over the first six months of life, which could be

related to maturational differences and/or compensatory

responses to hypoxemia.

Frequency of apnea. There are four types of apnea:

central, periodic, obstructive and mixed. Central apnea is

defined as a cessation of respiration with no chest or air

movement; periodic apnea is also of central origin except

that the apnea appears in constellations of short pauses

usually greater than three seconds in duration and repeated

three or more times with interapnea intervals of 20 seconds

or less; obstructive apnea is a disruption in ventilation

consisting of a lack of air flow but continued respiratory

movements; mixed apnea is initially a central apnea but

towards the end of the apneic episode some chest movements

appear with no air flow indicating some obstruction (Kelly

& Shannon, 1982).

Apnea with duration less than nine seconds is a common

finding in healthy infants over the first six months of

life (Hoppenbrouwers, Hodgman, Arakawa, Harper & Sterman,

1980). In this study short apnea (2-5 seconds) was more

abundant than longer apnea (6-9 seconds) and the incidence

of short apnea increased with age while longer apnea

incidence decreased with age. It was speculated that this

increase with age could probably be accounted for by the

same mechanism which was responsible for the concomitant

increase in interbreath interval. The frequency of apnea

longer than ten seconds was common only in the first week

when incidence of all apnea types was greatest. Otherwise,

prolonged apneic episodes were infrequent. Additionally,

sleep state affected apnea frequency, with more apnea

observed in AS than QS. When apnea does occur it has been

found to be of greater duration in QS than AS (Mellins &

Haddad, 1983).

There are conflicting data on whether the incidence of

apnea is enhanced in at-risk infants. In the fourteen

years since Steinschneider (1972) reported his findings of

prolonged and excessive apnea in infants who subsequently

died of SIDS, a number of studies have also reported

increased incidence of apnea in infants at risk for SIDS.

In another study, Steinschneider (1977) observed longer and

more frequent central apnea as well as increased frequency

of periodic apnea in infants with AI. However, apnea

shorter than 20 seconds was not examined. An increase in

frequency and duration of periodic apnea in both siblings of

SIDS and AI infants has also been reported (Kelly &

Shannon, 1979; Kelly, Walker, Cahen & Shannon, 1980).

In contrast, Guilleminault and colleagues have observed

no significant differences between AI infants and controls in

incidence of periodic apnea or in central apnea greater

than ten seconds (Guilleminault, Ariagno, Korobkin, Nagel,

Baldwin, Coons & Owen, 1979; Guilleminault et al., 1981).

However, they did report an increase in short (i.e., three

to ten seconds) mixed and obstructive apnea frequency in AI

infants from three weeks to four months of age.

Unfortunately, they failed to report or analyze data on

central apneas less than ten seconds. They report the

means of their 1979 study for central apneas from three to

ten seconds in a more recent publication but failed again

to analyse them (Guilleminault, Souquet, Ariagno, Korobkin

& Simmons, 1984). Inspection of these means suggested that

these short central apneas were more frequent than the

combined mixed and obstructive frequencies of similar

duration. In addition, it appeared that the AI infants

had a greater incidence of short central apneas at three

and six weeks of age but not at twelve weeks.

Additional contradictions were evident in other

research. Greater incidence of short apneas in subsequent

siblings of SIDS than controls was observed only when in AS

at two weeks of age (Kelly, Twanmoh & Shannon, 1982). No

difference in these infants was observed at eight and

fourteen weeks of age. Some researchers report no

difference in the frequency or duration of respiratory

pauses in AI and control infants (Mellins & Haddad, 1983).

Conversely, other researchers have found a reduced

incidence of central apnea in siblings of SIDS than in

controls in AS and QS (Hoppenbrouwers, Hodgman, McGinty,

Harper & Sterman, 1980). In this study, the incidence of

mixed and obstructive apnea and central apnea greater than

ten seconds was too low to analyse. However, it was found

that from one week to six months of age, differences

between infant groups were greatest for short apneas of two

to five seconds in QS; whereas for longer apneas of six to

nine seconds differences were greatest in AS and TS. In

another study, using AI and normal infants from one week to

six months of age, this same laboratory reported no

differences in periodic, mixed, obstructive or central

apnea frequency (Hodgman et al., 1982). However, although

the trends were nonsignificant, AI infants did display

lower frequency of short and longer central apnea in this


In summary, both the Shannon and Guilleminault

research groups report excessive incidence of apnea in

infants at risk for SIDS, although they don't agree on

which type of apnea is significant. Shannon claims that

its periodic apnea that is excessive and Guilleminault

emphasizes mixed and obstructive apneas. Alternatively,

the Haddad group reports no difference in apnea frequency

and the Hoppenbrouwers group found frequency reductions in

central apneic episodes less than ten seconds in duration.

Guilleminault and colleagues have recently emphasized that

because of differences in study design, definition of AI,

subject age, time and duration of study, recorded

measurements, analysis of respiratory events, and setting

of session, it is difficult to compare between laboratory

studies (Guilleminault, Ariagno, Korobkin, Coons, Owen-

Boeddiker & Baldwin, 1981). The lack of standardization

may therefore be responsible for the differences or lack

thereof in apnea frequency between studies. It seems clear

from the review above that there has been little

interlaboratory replication and although there may be

differences one way or another in the frequency of a

certain type of apnea, these are restricted to the

particular circumstances of measurement. Hence, the role

of apnea in SIDS remains equivocal.

The finding of an increase or decrease in apnea

incidence has been posited to relate to defective

respiratory control interacting with a myriad of potential

factors such as sleep state and CNS maturation (Shannon &

Kelly, 1982). The lack of consistent replicable finding

suggests that the apnea is of varied origin and probably

constitutes only one of the many triggering stimuli in SIDS

infants. Based upon the reported rates of apnea in the

above studies, it can be calculated that the risk groups

spent approximately only one minute per hour in apnea

(McGinty & Sterman, 1980). At this level the frequency of

apnea can hardly be considered a significant factor in the

induction of hypoxic conditions in risk infants. Apnea may

only represent the final common pathway in a series of

events preceding death.

Control of ventilation. There is evidence that the

mechanisms which control breathing begin to mature and

respond to respiratory stimuli such as level of carbon

dioxide and metabolic acidosis during fetal development.

For a review of the ontogeny of fetal respiratory

development see Maloney and Bowes (1983). The fetus stops

initiating breathing when the partial pressure of arterial

oxygen (PaO2) is lowered to about 15 mmHg (Wilds, 1978).

The adult, however, increases ventilation when PaO2 is

lowered below 50 mmHg (Dejours, 1963). The adult's

hyperpnea or ventilatory increase is sustained until

normoxic levels are achieved. The newborns' response to a

5% decrement in inspired oxygen differs from that of both

the fetus and the adult (Rigatto, 1977). In the newborn,

hypoxia induces a biphasic compensatory response: an

initial ventilatory increase within the first two minutes

followed by a decrement to and below baseline ventilatory

levels. The newly-born preterm infant at 33 and 37 weeks

gestation responds the same as the newborn (Rigatto,

1982). Therefore, gestational age does not seem to be a

factor. However, postnatal maturation may play a role,

since the initial ventilatory increase in the preterm

becomes larger with postnatal age and the adult-like

sustained hyperpnea to hypoxia begins about one month after

birth (Rigatto, 1982). Full term infants, on the other

hand, begin sustained hypoxia-induced hyperpnea at about

two weeks of postnatal age (Rigatto, 1977). The

progressive increase in ventilation seen in the preterm

infants suggests that maturational changes occurred in

either respiratory pump mechanics or peripheral

chemoreception, or both.

The initial increase in ventilation has been

attributed to stimulation of the peripheral chemoreceptors

because of the rapidity of the ventilatory response to

hypoxia onset (Rigatto, 1982). The subsequent fall in

ventilation has been postulated to be due to adjustments in

pulmonary mechanics (e.g., lung compliance decrease),

central depression of respiratory drive, and immature

interaction between carotid sinus chemoreceptive afferent

feedback and the central respiratory mechanisms (Haddad,

Schaeffer & Bazzy, 1983). Hence, it was suggested that in

infants at-risk for SIDS hypoxic depression in an

immature CNS could induce an imbalance of excitatory and

inhibitory influences, which may supercede the

chemoreceptor effect on breathing. This could result in a

state of chronic hypoventilation or be manifested by a

noncompensatory decrement in ventilation to an acute

hypoxic challenge.

Current data suggest that, when compared with adults,

infants have a reduced PaO2, indicating that infants are

normally exposed to a hypoxic challenge (Brooks, Schluete,

Nabelet & Tooley, 1978). More systematic study of this

hypoxic stimulus revealed that periodic apnea could be

induced in preterm infants when subjecting them to mild

hypoxia (Rigatto, 1982). After prolonged exposure,

however, incidence of periodic apnea decreased and central

apnea increased during the decreased ventilation of the

secondary limb of the biphasic response. This suggested

that hypoxia may be an important triggering stimulus for

apnea when the respiratory system is depressed in immature

nervous systems. A similar finding was observed in AI

infants exposed to a 3% decrease in oxygen; most notably,

however, these at-risk infants displayed greater incidence

of periodic apnea and central apnea to this hypoxic

stimulus than did the control infants (Brady, Ariagno,

Watts, Goldman & Dumpit, 1978).

There have been numerous reports of periods of

prolonged apnea in infants at-risk for SIDS (Shannon, Kelly

& O'Connell, 1977; Steinschneider, 1977; Kaun, Blum,

Engleman & Waterschoot, 1982). Increased durations of

apnea in AI infants have been associated with corresponding

greater reductions in the transcutaneous partial pressure

of oxygen (tcPO2), which is a noninvasive measure of PaO2

(Kaun et al., 1982). The prolonged duration of apnea in

the AI infants could be related to a higher peripheral or

central chemoreceptor response threshold to hypoxia or

hypercapnia. That is, during apnea greater hypoxic and/or

hypercapnic levels are accumulated before respiration is

initiated resulting in prolonged apneic episodes.

Therefore, it would be expected that infants with this

elevated threshold would be sluggish to respond to a

hypoxic or hypercapnic challenge.

Support for this suggestion comes from studies of

ventilatory response to hypoxic and hypercapnic stimuli in

at-risk infants. Impaired ventilatory increases to a 5%

reduction in inspired oxygen (i.e., 15% oxygen) were

observed in AI infants (Hunt, McCulloch & Brouillette,

1981). Also, blunted ventilatory responses and a more

pronounced subsequent depression of ventilation were

observed in AI infants to hypoxic challenges of 15% oxygen

compared to controls (Haidmayer, Kurz, Kenner, Wurm & Pfeiffer,

1982). Concomitant with the depression in ventilation,

were more pronounced and rapid declines in tcPO2 values in

the AI infants. Moreover, other studies have found that AI

infants displayed less ventilatory response to hypercapnic

challenges of 4-5% carbon dioxide than controls (Haidmayer

et al., 1982; Hunt et al., 1981; Marotta, Fort, Mondestin,

Hiatt & Hegyi, 1984; Shannon et al., 1977). Therefore,

infants at risk for SIDS have reduced hypoxic and

hypercapnic ventilatory responses, suggesting reduced

sensitivity to chemoreceptive challenge.

In summary, the majority of evidence from

cardiorespiratory physiologic investigations has revealed

that infants at-risk for SIDS display elevated HR levels

and reduced HR variability; RR is also elevated and these

infants exhibit a transient acceleration in the

developmental pattern of RR while sleeping. There are

conflicting data on which type of apnea is present in these

infants and whether the frequency of apnea during sleep is

excessive or diminished. However, when apnea does occur,

the duration is more prolonged and is associated with

corresponding decrement in PaO2. When breathing gas

mixtures of decreased oxygen content or increased carbon

dioxide content, at-risk infants initially displayed a

sluggish ventilatory increase and with prolonged exposure,

subsequently exhibited more pronounced depression of


Sleep and Breathing

The finding that SIDS infants are typically found dead

in the early morning hours, when they were presumed to be

asleep, has led some to speculate that sleep precipitates

the events leading to SIDS (McGinty & Sterman, 1980). There

are rapid maturational changes in sleep and respiratory

physiology in the first six months of life. (See Berg and

Berg (1979) for a review of the ontogeny of sleep,

cardiorespiratory physiology and sensory function in

infancy.) During this time, the sleep states become well

defined and become more organized within the sleeping

periods. There is a consolidation of sleep and waking into

longer and longer uninterrupted periods. By four to six

months, these become coupled so that the longest of the

sleeping periods follows the longest of the waking periods

(Coons & Guilleminault, 1982). By three months stereotypic

elements of adult electroencephalographic (EEG) sleep

patterns can be used to discriminate between AS and QS and

stages within QS (Coons & Guileminault, 1982). For a

recent comprehensive review of the neurophysiology of sleep

the reader may refer to McGinty and Beahm (1984).

In QS breathing is very regular and inspiration is

prolonged. It has been suggested that during QS the

forebrain and the brain stem are uncoupled (Harper, 1983).

However, slow-rhythmic descending discharges from regions

of the forebrain have been documented, which may be pacing

the lower brain stem structures in QS (Harper & Sieck,

1980). Other forebrain areas have been implicated in the

gating of QS onset (Harper, 1983). The influence of the

forebrain in the ontogeny of QS may be evidenced by the

observation of the emergence of a dramatic postnatal

increase in the proportion of total sleep time spent in QS

associated with concomitant increases in postnatal

forebrain development (Parmalee, Stern & Harris, 1972).

During AS, breathing is extremely variable and RR is

increased; desynchronous descending influences of the

forebrain and other areas on brain stem respiratory and

motor neurons affect rhythm and rate of breathing and also

diminish muscle tone (Schulte, Albani, Schmizer & Bentele,

1982). Phasic bursts of rapid eye movement and twitches in

the limbs and peripheral musculature occur. The rib cage

loses stability and during inspiration paradoxical

deflation of the rib cage may be observed due to depression

of intercostal and abdominal muscle activity (Henderson-

Smart & Read, 1978). This might indicate that AS

represents the state of greatest vulnerability to apnea and

hypoxia. However, other observations indicate that AS may

be associated with improved breathing. For example, oxygen

level in the blood of infants is improved in AS compared to

QS (Brooks et al., 1978).

There is evidence accumulating that the state of

greatest vulnerability for apneic episode may be QS. When

kittens were chronically exposed to hypoxic conditions, 28%

of them failed to compensate in QS and exhibited

hypoventilation and prolonged apnea, in some cases

progressing to death; in AS the hypoxemic conditions in

these animals were transiently reversed with faster RR and

decreased incidence of apnea (McGinty & Sterman, 1980).

This evidence suggests that the conflicting reports of

apnea incidence in the at-risk infants may reflect

differential sampling of infants who may or may not be

capable of producing compensatory cardiorespiratory

responses. In support of this possibility, it may be

recalled that the compensatory pattern of elevated HR and

RR with reduced incidence of apnea (which was more

pronounced in QS) was observed in siblings of SIDS who, by

definition, had had a healthy neonatal course with no

documented prolonged apnea attack (Hoppenbrouwers et al.,

1980); whereas no compensatory pattern was observed in AI

infants who had a prior history of prolonged apnea (Hodgman

et al., 1982). This suggests that the siblings of SIDS may

differ from the AI infants in the ability to compensate

with increased cardiac output and ventilation to a hypoxic

challenge; whereas the AI infants having previously

demonstrated some pathophysiologic disorder by virtue of

their prolonged apneic attack, may no longer be capable of

producing compensatory cardiorespiratory responses.

Other evidence suggests some alteration in the

maturation of sleep physiology in at-risk infants.

Siblings of SIDS victims display an accelerated maturation

of 12-15 Hz EEG activity in QS and 4-7 Hz EEG activity in

AS between one and two months of age (Sterman, Harper,

Hoppenbrouwers, McGinty & Hodgman, 1979). It was suggested

that an abnormal stimulus such as hypoxia may have caused

this transient acceleration in EEG development. It has

been speculated that an accelerated maturation of the

forebrain inhibitory influences on medullary respiratory

centers may be present in these infants (Hoppenbrouwers &

Hodgman, 1982). Other evidence indicates that, rather than

being accelerated, the sleep state development in at-risk

infants is delayed. The percent time in QS was less in AI

infants than in controls at three and four months of age,

suggesting a maturational delay or disturbance in the AI

infants (Haddad, Walsh, Leistner, Grodin & Mellins, 1981).

Analysis of the temporal sequencing of sleep-wake epochs in

normal infants and siblings of SIDS, revealed disturbed

patterns of state organization at two to three months of

age (Harper, Frostig, Taube, Hoppenbrouwers & Hodgman,

1983). Siblings were observed to have longer intervals

between periods of wakefulness, a shorter mean QS cycle

period and an increase in the frequency of state

transitions. The prolonged periods of uninterrupted sleep

may reflect an inability to arouse from sleep (Harper et

al., 1983). Hunt (1981) observed that AI infants also

showed a tendency to remain asleep in the face of hypoxic

challenge; whereas control infants displayed an increased

frequency of arousal from sleep when inspired oxygen was

progressively decreased to 15%. The authors posited that

the failure to arouse may be due to a defect in peripheral

chemoreceptor sensitivity. Others have hypothesized that

the chronic hypoxemia of SIDS, and the hypoventilation and

failure to arouse in at-risk infants may be due to tonic

overactive synthesis of endorphins, which have previously

been associated with the CNS suppression of respiration and

wakefulness (Kuich & Zimmerman, 1981). Some support for

this theory was observed in post-mortem examination of

met-enkephalin (an endogenous endorphin) levels in

cerebrospinal fluid, when it was found that SIDS victims

had greater concentrations than age-matched controls

(Rappaport, Kozakewich, Fenton, Mandell, Vawter & Yang,

1984). Regardless of the etiology, the failure to make

transition from sleep to waking may place these infants at

greater risk. The finding of increased incidence of apnea,

occurring antecedent to awakening in infants in the

first six months of life, suggests that arousal from sleep

may be an important mechanism in terminating apneic

episodes (Guilleminault et al., 1981).

Cardiorespiratory Responses to Airstream Stimulation

The evidence from pathological and physiologic

investigations indicate that infants at-risk for SIDS may

be chronically hypoxic, are sluggish to respond to hypoxic

and hypercapnic stimuli with a ventilatory increase and

are less likely to arouse from sleep during such a

challenge; also, while asleep they consistently exhibit

prolonged periods of apnea with concomitant bradycardia.

Hence one of the most striking deficits in these infants

is their higher tolerance of hypoxia and hypercapnea

before responding appropriately. This deficit is

particularly apparent from two to four months of age,

which is the period of most rapid CNS development and also

the age of greatest vulnerability for SIDS. Therefore it

may be possible to detect an infant at risk for SIDS by

eliciting apnea and thereby establishing a hypoxic or

hypercapnic challenge to which they would be expected to

respond sluggishly. The prolonged apneic episodes that are

spontaneously produced by these infants are probably

related to a more general CNS dysfunction of autonomic

depressor reflexes involving the complex interactions of

the medullary chemoreceptors with inputs from the

trigeminal, glossopharyngeal and vagus cranial nerves,

peripheral chemoreceptors, sleep control mechanisms and

cardiorespiratory regulating mechanisms.

Investigations in infants of the ontogeny of reflex

mechanisms which when stimulated can induce apnea are

limited to four areas of research: 1) laryngeal

chemoreceptor stimulation and gastroesophageal reflux;

2) ventilatory response to hypoxic and hypercapnic

exposure; 3) vagal-mediated reflexes in compensation to

inspiratory load; and 4) dive reflex and facial cooling

with an airstream.

In animals stimulation of the laryngeal chemoreceptors

with chemical irritants have been found to evoke prolonged

apnea (Downing & Lee, 1975). There has been a report of an

association of apnea and gastroesophageal reflux in

premature infants, which presumably induces reflex apnea by

acidic reflux stimulation of the laryngeal chemoreceptors

(Herbst, Minton & Book, 1979). However, others have found

no such relationship in premature and full-term infants, and

AI infants (Ariagno, Guilleminault, Baldwin & Owen-

Boeddiker, 1982; Walsh, Farrell, Keenan, Lucas & Kramer,

1981). Some have studied hypercapnic ventilatory

responsiveness and sustained hypoxic hyperpnea and have

observed that at some critical level of PaO2 and PaCO2

periodic apnea is induced and with further exposure central

apnea eventually occurs (e.g., Rigatto, 1982). However,

the role of this mechanism in initiating prolonged apnea

has not been examined. Inflation of infants lungs in order

to activate vagal-inhibition of respiration has revealed

that premature infants are more susceptible to respiratory

inhibition than full-term healthy infants at any age (e.g.,

Boychuk, Rigatto & Sheshia, 1977; Kirkpatrick, Olinsky,

Bryan & Bryan, 1976). Although this technique when studied

developmentally would provide information on the ontogeny

of vagal reflexes in compensation to an inspiratory load,

it is subject to confounding influences of changes in

pulmonary mechanics due to sleep state (Henderson-Smart,

1984). Nevertheless, this test is more likely to provide

information on inspiratory load compensation than

information on the development of mechanisms of prolonged

respiratory cessation. Although research on these first

three areas could be of importance to a better understanding

of apneic episodes in SIDS, apnea induction by the dive

reflex and facial cooling may have the greatest relevance.

The dive reflex is a depressor reflex which appears to

resemble the acute cardiorespiratory response observed in

infants during an apneic episode. Diving in animals like

seals elicits apnea, profound bradycardia and peripheral

vasoconstriction. These responses have the clear

physiological function of conserving oxygen by reducing

circulation to tissues of greater anaerobic capacity (e.g.,

skin, mesentary, muscle) and providing circulation to

tissues less tolerant of the effects of oxygen deprivation

(e.g., heart, brain). The ophthalmic division of the

trigeminal nerve, which in humans innervates the forehead,

eyelids and dorsum of the nose, was found to be primarily

responsible for completing the afferent pathway by which

dive reflexes were elicited in the duck (Anderson, 1963).

The dive reflex has also been observed in humans

(e.g., Elsner, Franklin, Van Citters & Kenney, 1966;

Kawakami, Natelson & DuBois, 1967). More recently,

elaboration of the relevant eliciting stimuli has revealed

that mere submersion of the face in cold water was

sufficient to elicit the bradycardia and peripheral

vasoconstriction typical of the dive reflex (Hurwitz &

Furedy, 1979). Furthermore, it was the combination of cold

facial stimuli and breath holding that produced greater

bradycardia and vasoconstriction than in either control

conditions of face immersion (while breathing through a

snorkel) or breath holding alone. Others have found that

the dive reflex was even more pronounced with colder water

temperatures (Furedy, Morrison, Heslegrave & Arabian, 1983;

Song, Lee, Chung & Hong, 1969). The dive reflex literature

illustrates the significance of cold facial stimulation in

effecting cardiovascular and respiratory responses.

In human infants apneic and bradycardic responses have

been elicited by nasopharyngeal suction and by blowing

oxygen over the infants' face (Brown, Ostheimer, Bell &

topography of the HR response from stimulus onset, which

would have provided more information on the temporal effect

of the stimulus and would have been a less biased measure.

Brief apneic episodes in response to the airstream stimulus

were observed but no difference in apnea duration existed

between the two infant groups. Unfortunately, no further

attempt was made to examine the apnea responses

quantitatively. Moreover, no systematic analysis was

performed to partial out the effects of sleep state,

postnatal age or presence or absence of apnea from their

possible effects on the cardiorespiratory responses


The proposed study, herein, was designed to establish

the normative response to facial airstream stimulation in

infants throughout the first year of life. This study has

a focus similar to that of McCubbin et al. (1977) but

utilized a more systematic design, including the

appropriate control conditions to compare the

cardiorespiratory responses. The study was designed to

assess in sleeping infants developmental changes in

cardiorespiratory responses to airstream stimulation in the

first year of life by 1) determining whether airstream

stimulation of the facial region supplied by the ophthalmic

branch of the trigeminal nerve (i.e., forehead, eyelids and

dorsum of the nose) would elicit apnea more frequently than

a control airstream stimulus directed to the abdomen (i.e.,

rostral to the umbilicus); 2) measuring the HR response on

a second-by-second basis and average a sufficiently large

number of homogeneous trials to determine the precise

topography of the cardiac response to airstream

stimulation; and 3) examining the effect of apnea

elicited by the airstream stimulus on the HR response


This study will be informative for the following

reasons. First, the examination of the cardiorespiratory

responses to repeated airstream stimulation will provide

information on the maturation of cardiorespiratory control

in the production of depressor reflexes in normal infants

cross-sectionally throughout the first year of life. It

will be informative to determine whether the cardio-

respiratory responses in infants, during the period of most

rapid CNS development (i.e., two to four months of age) and

vulnerability to SIDS, differ from responses during

earlier or later postnatal age. Second, it will provide

more information on the cardiorespiratory effects of

conditions of facial cooling similar to those which exist

presently in hospital delivery protocol manipulations such

as nasopharyngeal suction and oxygen blown over the face.

Third, since the airstream stimulus is not substantially

different from that which the infant might commonly


experience in every day life, this study may shed some

light on the eliciting conditions of apnea and bradycardia

in normal healthy infants. Fourth, this simple reflex test

of the integrity of trigeminal-brainstem-cardiorespiratory

function could have prognostic significance for screening

of cardiorespiratory risk for SIDS. However, baseline

data in a normative sample are required initially so that

the results of this study will provide information against

which data from high risk infants may be compared.

Maternal and infant hospital medical records of the

newborn group were surveyed to insure that all criteria were

met. For the older groups information on the infants'

delivery and neonatal course and the maternal obstetrical

history was obtained in a questionnaire during testing (see

Appendix A). Table 1, 2 and 3, respectively, shows the sex,

gestational age (GA), postnatal age (PNA) at the time of

testing, birth weight (BW), APGAR scores, race (R), maternal

age (MA), and maternal medication (MED) at delivery for

each infant in the three age groups included in this study.

Newborn subjects were recruited by obtaining permission

from their mother or father in the neonatology ward of the

J. Hillis Miller Health Center. These subjects received no

remuneration for their participation. The older subjects

were recruited through the mail. Names of infants born in

the appropriate age group were obtained through the Alachua

County Health Department. Parents were sent a general letter

describing the study and were requested to indicate whether they

were interested in participating in the study on an enclosed

stamped-postcard. Those parents responding affirmatively

were telephoned to schedule an appointment for testing. In

this conversation parents were informed about the purpose

and specifics of the study and an appointment was scheduled

at a time that would correspond to the infant's nap schedule.
These subjects received five dollars for their participation.


The factors of critical interest in this study were

the site of airstream stimulation and the age of the infant

when tested. Knowledge of these effects indicate whether

cardiorespiratory responses to stimulation of the facial

region supplied by the ophthalmic branch of the trigeminal

nerve may be unique and if so, whether it changes

throughout the first year of life. However, the behavioral

observations revealed that two other factors, sleep state

and presence of elicited-large movement, may also have

influenced the cardiorespiratory response. Before the

effects of these two latter factors on the cardio-

respiratory response could be examined, it was necessary

to assess the infants' sleep state pattern during testing

and to determine the extent to which large movement was

elicited by the airstream stimuli.

Sleep State Incidence and Pattern

Table 4 presents the percentage of trials in which

the sleep state of the infants were rated as QS, TS and

AS across the age groups. The most notable finding was

that the majority of data were collected while the infants

were in QS. However, the NB group tended to exhibit a

0 0

iE N
d H

r- 4-
- ""'r---


I : om

IJ I -4 I rL
at -O CD
D7 ji c

o.)-J uLL L- 'C' H- C -- LS

,'" -: I- 1-- I1'

Prestimulus Levels

Prior to the analysis of the HR response topography

the prestimulus levels of HR were evaluated to determine

whether the post-stimulus HR response could have been

influenced by initial prestimulus levels. This was of

particular concern since it was to be expected that

prestimulus levels would co-vary with infant age (Harper et

al., 1982). Table 12 depicts the mean HR values in beats

per minute (bpm) for the second preceding stimulus onset

across the age groups during trials with and without


The analysis of the prestimulus HR level in the trials

in which no apnea or LM were elicited used Age as the

between-subject factor and Site of stimulation as the

within-subject factor. Only the Age effect was

significant, F(2/47)=5.9, 2<.01. With no significant

difference between the two younger groups, the difference

between the age groups, therefore, can be accounted for by

diminished prestimulus levels in the 8-12 infants

Post-stimulus Changes

The interpretation of the post-stimulus HR response

topography is complicated by the apparent relationship

between prestimulus HR levels and the infants postnatal

age. Hence, the post-stimulus HR levels and perhaps

topography may be partially determined by the prestimulus

Mean Prestimulus
Apnea and Large

HR Levels in bpm in the Presence and Absence of
Movement (LM) for Facial (F) and Abdominal (A)
Airstream Stimulation Trials

Apnea Movement Condition NB 2-4 8-12 Mean

No No LM F 117.92 122.76 110.52 117.83

A 118.72 122.17 112.11 118.26

Mean 118.32 122.47 111.32

N 13 22 15

Yes LM F 123.56 130.03 106.65 124.90

N 11 10 2

Yes No LM F 121.21 121.45 117.16 120.31

8 13

HR levels. This statistical problem is known as the Law of

Initial Values as outlined by Wilder (1950). Following the

guidelines suggested by Richards (1980) for countering the

effects of prestimulus HR on postimulus HR, the assumption

of homogeneity of regression coefficients for the data

were tested over age for regression of each of the 25 post-

stimulus seconds on the first prestimulus second. Each of

these analyses revealed that the assumption of homogeneity

of regression coefficients was not violated. Therefore,

the effects of the experimental parameters on the HR

response topography could be compared across age. However,

to control for the effects due to the Law of Initial
Values, the analysis of covariance (ANACOVA) was used.

Since the interpretation of HR change was of primary

concern, difference scores were calculated by taking the

algebraic difference between the prestimulus value and the

post-stimulus values. The correlation between prestimulus

and difference scores have typically been found lower than

the correlation between prestimulus and post-stimulus

scores (Benjamin, 1967). Therefore, the use of difference

scores, herein, provided an additional statistical

advantage. Despite lower correlations, however, difference

scores still remain subject to the Law of Initial Values.

The HR difference scores, therefore, were all analysed with

ANACOVAs using the HR value in the second preceding stimulus

onset as the covariate, Age as the between-subject factor

and Seconds as the within-subject factor. It should be

noted that all significant effects reported below remained

significant when the conservative Greenhouse-Geisser

approximation of degrees of freedom was used (Greenhouse &

Geisser, 1959).

HR Response to the Airstream Stimulus: Face vs. abdomen

Figures 2 and 3 depict the HR change scores at each

age in response to facial and abdominal stimulation in the

absence of apnea and LM. For both stimulation sites these

figures show that the a short monophasic acceleration to

the airstream was produced by the NBs. However, the two

older groups displayed a triphasic response to the

airstream stimuli comprised of a brief deceleration

followed by an acceleration and a subsequent deceleration.

After stimulus offset HR levels in these infants gradually

recovered toward prestimulus levels.

More specifically, ANACOVA analyses (see Appendix F)

revealed that an initial deceleration was elicited in the

older infants within one second of stimulus onset but was

not elicited in the NBs. An acceleration followed

maximizing at the third post-stimulus second in all age

groups. This acceleration, which was greater when elicited

by the facial airpuff than by the abdominal airpuff,

differentiated all the age groups such that the 8-12


0 1

**--I -

r. 4 a


S 0 -- 0.0
r- 1i

I I I :. I : I
0~~~~~~~~L uj;6: '" 'JCO0J ^'1C

rck~~ ~ ~ *zr~iu 'cz cici

LO C) Iq-4



k! CO C

- 00
< L.. .



SCt n a: l-I-'-LU --,- CuJI i.


C l I I

ax' 'X Ir r, -':1 c~. ix' ej

infants displayed the fastest acceleratory recruitment

followed by the 2-4 infants and then the NBs. As in the

early deceleration, the NBs were differentiated from the

older infants by the ensuing deceleration from seconds

three to ten, which reached its nadir by the seventh post-

stimulus second in the older infants but merely returned to

pretrial levels in the NBs. Only the 2-4 infants displayed

an effect of the stimulation site during this interval with

greater deceleration to the facial airpuff than the

abdominal airpuff. Following stimulus offset all groups

displayed a cubic trend over seconds 11 to 25, which was

unaffected by the stimulation site, appearing to decrease

slightly 5 to 10 seconds after stimulus offset followed by

a gradual return to pretrials levels.

HR Response to the Facial Airstream Stimulus: Apnea vs. no apnea

Figure 4 shows the response to facial stimulation when

apnea was present without preceding LM. By comparing these

responses with the responses in Figure 2 in which no apnea

was elicited by the facial airstream, the effect of apnea

elicitation on HR can be examined. Although the overall

topography of the response was similar with and without

apnea, when apnea was elicited, the older groups displayed

a more prolonged late deceleration which was of greater

magnitude than when no apnea was elicited. The NBs

response was unaffected by whether or not apnea was induced.

,.- '..

m I I


I- *- -A
I- Il H!

-= -4 C'

r OJ 0 -J -t LID I- s
i I I -I - -*

T -"" =' --Ei '-7" .SJ LLI O LU .. "

The ANACOVAs of the overall HR response used seconds 1

to 25 as levels of the within-subject Seconds factor, Age

as the between-subject factor and presence or absence of

Apnea as the second within-subject factor A significant

interaction between Age, Apnea and Seconds, F(48/576)=2.0,

r<.001, was found. The next analysis examined whether the

differences described by this interaction existed during

the period of acceleration across seconds 1 to 3. It did

not, since only a significant Seconds effect, F(2/48)=6.7,

2<.005, emerged. Therefore, up to this point in the

analysis the HR response was unaffected by the age of the

infant or by the presence or absence of apnea.

The age groups were differentiated by the HR response

during the late-deceleration from seconds 3 to 12, as

indicated by a significant interaction between Age, Apnea

and Seconds, F(18/216)=3.1, p<.001. The subanalysis

consisted of examining the NB group alone. Only a

significant linear trend across Seconds, F(1/6)=21.1, p<.005,

was found, indicating that there was a linear HR

decrement toward base level both in trials with apnea and

trials without apnea. Subsequent t-tests revealed that the

decrement in the NBs HR during this interval did not

proceed below prestimulus levels. When the two older groups

were contrasted a significant interaction between Age, Apnea

and Seconds, F(9/162)=4.7, p<.001, emerged. This was due

to changes in both magnitude.and topography of the

deceleration when apnea was elicited. When the two older

groups were compared using only the non-apnea responses,

only a significant quadratic trend over Seconds,

F(1/18)=7.5, p<.05, emerged, indicating that there was no

difference between groups in the magnitude of the

deceleration when no apnea was elicited. However, on

trials when apnea was elicited, the magnitude of

deceleration was increased to a greater extent in the

8-12 infants than in the 2-4 infants, as substantiated by

a significant interaction between Age and Seconds,

F(9/162)=6.3, p<.001. Moreover, the effect of apnea was

to prolong the deceleratory response differentially in

these older infants. As can be seen in Figure 4, the

deceleration in the 2-4 group reached its nadir and began

its recovery toward pretrial levels sooner than the 8-12

infants. This was confirmed by an orthogonal trend

analysis of each age group for apnea data only, with a

significant decreasing quadratic trend over Seconds,

F(1/12)=31.6, p<.001, for the 2-4 infants and a decreasing

linear trend over Seconds, F(1/6 )=39.1, p<.001, for the

8-12 infants.

The analysis of seconds 12 to 25, which represented

the period of HR recovery, yielded a significant main effect

of Apnea, F(1/23)=7.0, p<.05, and an interaction between Age

and a quadratic trend over Seconds, F(2/24)=5.2, p<.05.

Neither the effects of Apnea or Seconds nor their interaction

were significant when analysing only the NB group's HR

recovery, indicating that regardless of whether apnea or no

apnea was elicited the NB HR was stable throughout this

period maintaining a mean HR level of -0.65 bpm. When

comparing the younger groups only a significant Age effect,

F(1/17)=6.2, p<.05, was found, indicating that the 2-4

group operated at a lower level than the NB group, with

mean HR bpm levels of -0.65 and -5.01 for the NB and 2-4

groups respectively. When comparing the 2-4 and 8-12

groups, however, an interaction between Age, Apnea and

quadratic trend over Seconds, F(1/18)=4.5, r<.05, emerged.

No difference was found in the recovery trends of the

trials with no apnea. When apnea did occur, however, the

enhanced late-deceleration and the resultant phase

shift between these groups resulted in interacting

decreasing and increasing quadratic trends over Seconds,

F(1/18)=7.5, p<.05. Inspection of Figure 5 reveals that by

post-stimlulus second 12 the HR response of the 2-4 group

had partially recovered, while the 8-12 group had just

reached its nadir. Hence, HR levels decreased to a mean of

-7.25 bpm for the 2-4 infants while the HR levels increased

to a mean level of -8.58 bpm for the 8-12 infants.

In summary, the NBs HR response when apnea was

elicited was not distinguishable from the response observed
when no apnea was elicited. The main effect of apnea on

the HR response appeared in the older groups. Apnea

resulted in an accentuated deceleration during seconds 3 to

12 post-stimulus onset reaching a maximum at seconds 7 and

12 respectively for the 2-4 and 8-12 groups. Apnea also

had an effect on the HR recovery during seconds 12 to 25 in

the older groups, exhibiting less tendency to return to

pretrial levels.

HR Response to the Facial Airstream Stimulus from Apnea Onset

The previous HR analysis used HR values scored from

stimulus onset. When apnea was present, it could occur

anywhere in the trial within the 20-second post-stimulus

observation window. Therefore, the topography of the HR

response in the presence of apnea was difficult to

interpret. This was especially the case, since the latency

to apnea onset was significantly related to age, with older

infants displaying longer apnea latencies. To obtain a

more precise determination of the effect of apnea on the

response topography, HR was aligned relative to apnea onset

rather than stimulus onset. To maintain a constant

standard of reference, these HR values, as in the above HR

statistics, were represented relative to prestimulus

level. For example, the HR value at second zero

represented the HR difference from prestimulus level of the

interval just preceding apnea onset.

The mean HR values observed during apnea as elicited

by the facial airpuff (uncomplicated by LM) were plotted in

Figure 5. The analysis of the HR response from apnea onset

did not examine the effect of preceding LM on the HR

topography. As can be seen in Figure 6, the effect of LM

was to substantially increase HR (relative to prestimulus

level) prior to apnea onset. Otherwise the response

topography remained similar to the response when no LM

preceded apnea. As can be seen in Figure 5, the HR

response of the NBs was not affected by the production of

apnea; whereas, in both of the older groups apnea induced

rapidly-recruiting, large-magnitude decelerations, which

began to recover to prestimulus levels when apnea

terminated. Note that the arrowhead located below each

curve in both Figures 5 and 6 corresponds to the point

where apnea ended and breathing resumed.

ANACOVA of the no LM data (Figure 5) assessed the

pre-apnea HR levels. Seconds -5 to 0 were used as levels

of the Seconds factor, Age as the between-subject factor

and the unadjusted HR value in the second preceding apnea

onset (i.e., second 0) as the covariate. There were no

differences to emerge during this period as neither the

effects of Age or Seconds nor their interaction were


El -n) 0

mi l


0- -o
44 I i UI

S-- C

s^ S-1 a^!

n' c '- '.k J LU
-T I s *

-4 -- -

| i I |

1 i C I I I
0.3 s O W .*.sUs.. C' OJ -.'a C
Il~ il

-l- I I'T---I -q C _CL.


S- I

l lG -H-

LM =-4 k-H b
" ... ,,- -.,
I II! &

iLLI n

?.'. C'i ^- T-L 1 -- j

*) 1: I5-~ r ? ii

Inspection of the trends in Figure 5 suggested that

there was an initial deceleration upon apnea onset in the

older groups, which was absent in the NB infants. The

analysis of this trend used seconds 0 to 9 as the period of

deceleration after apnea onset. The ANACOVA yielded a

significant Age effect, F(2/24)=4.0, p<.05, a Seconds

effect, F(9/225)=7.7, p<.001 and a two-way interaction of

these factors, F(18/225)=1.9, p<.05. The subanalysis of

the most salient of these effects the Age-by-Seconds

interaction consisted first of assessing whether there

was a Seconds effect in the NB group alone; there was not,

F(9/63)=1.0, indicating that when apnea was elicited it had

no effect on the NBs HR response. An ANACOVA was then

performed on the two older groups. Only a significant

Seconds effect, F(9/162)=4.6, p<.001, was found.

Therefore, from the onset of the apneic period to the ninth

second of the trial, the two older groups exhibited HR

deceleration, which was of similar magnitude and which was

totally absent in the NB group. The 2-4 infants reached a

maximum deceleration of -14.46 bpm at seven seconds post-

apnea onset; whereas the 8-12 infants decelerated to -15.91

bpm by the ninth second after apnea began.

Analysis of seconds 9 to 25 was performed to assess

the period of recovery from the apneic episode. The NB

group displayed no significant change over these seconds,

University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs