Life stress


Material Information

Life stress impact on genital herpes recurrences
Physical Description:
vii, 112 leaves : ill. ; 29 cm.
Hoon, Emily Franck
Publication Date:


Subjects / Keywords:
Herpes Genitalis   ( mesh )
Stress, Psychological   ( mesh )
Life Change Events   ( mesh )
Clinical and Health Psychology thesis Ph.D   ( mesh )
Dissertations, Academic -- Clinical and Health Psychology -- UF   ( mesh )
bibliography   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph.D.)--University of Florida, 1986.
Bibliography: leaves 99-110.
Statement of Responsibility:
by Emily Franck Hoon.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 000560494
oclc - 17865016
notis - ACY6046
System ID:

Full Text








This project involved the time, talent, effort, and cooperation

of many individuals. Because I appreciate their contributions and

want to recognize them and thank them, I address the following:

Ken Rand--patient, enthusiastic, humorous, encouraging,

knowledgable, accessible--friend: contributed freely--remembered with

warmth forever!

Jim Johnson--guidance, sense of organization, perspective,

calming presence, wisdom--a role model deserving emulation--my sincere


Randy Carter and his assistant statisticians who helped make

sense of voluminous computer printouts.

And to the rest of the best committee a Ph.D. candidate ever

had: Eileen Fennell, Rudy Vuchinich, and Nancy Norvell.

Wendy and Robin Hoon--young people grown independent, before my

eyes, out of necessity while coping with the vagaries of their Mom's

graduate school schedyle--I love both of you, and I thank you for your

help and cooperation.

Ernest Franck, my dad, who has always encouraged me to follow my

curiosity and to accept challenges.

The late Emily Franck, my mom, who helped me to appreciate the

value of social relationships.

Pete Hoon, my ex-husband, with whom I shared my early research

experiences and with whom I continue to enjoy a satisfying

collaborative relationship as a professional and coparent.

Rob Martin, my present husband, who helped me in uncountable ways

to tie up loose ends and who provides inspiration for getting the most

out of work time and play time.

Subjects--courageously sharing personal information to advance

knowledge regarding herpes.

And the following cheerful helpers: research assistants,

laboratory personnel, and Clinical Research Center staff.


ACKNOWLEDGMENTS..... ................................................ i



ONE LITERATURE REVIEW ....................................... 1

Stress and Illness .................. ......... ..... .... .. 1
Moderator Variables........................................4
Social Support .................... ................ ... 4
Locus of Control .....................................7
Arousal Seeking ................................ ..... .9
Methodological Issues....................................12
Retrospective Design.................................12
Measures of Illness.................................. 14
Confounding of Events and Outcome Measures...........15
Value and Impact of Events......................... .16
Minor Events.......................................17
Recurrent Genital Herpes as a Model.......................18
Clinical Nature of Herpes Simplex Virus..............18
Recurrences............................. ......... 19
Present Lack of Cure........................... 21
Laboratory Documentation...................... 21
Endogenous Pathogen ............................... 22
Stress and Herpes ....................................... 23
Anecdotal Reports ............................ .23
Survey Research......................................25
Retrospective Research..............................26
Prospective Research ............................... 28
Summary.......................................... 31
Theoretical Framework.....................................33
Perception of Stressful Events....................... 33
Proposed Physiological Mechanisms....................36
Present Investigation...................................44
Hypotheses .................... .. ...........45

TWO METHODS.....................................................48

Research Design ..........................................48
Sample Selection............ .............. .......... 49

Measures...... ................................... ...... 50
Stress Measures......................................50
Life Experiences Survey (LES)...................50
Hassles Scale (HS).......................... .....51
Moderator Variables................................. 52
Locus of Control (LC)...........................52
Arousal Seeking (AS).......................... 53
Social Support Questionnaire (SSQ)..............54
Physiological Measures..............................55
Screening procedure: HSV antibody..............55
Dependent variable: Virus isolation and
Procedure ................................................. 56
Initial Contact...... ................................56
Second Contact ........................................57
Continuing Contact.................................58
Final Contact....................................... 58

THREE RESULTS................................................. 59

Sample Characteristics....................................59
Undesirable Life Events.............................60
Presumed Moderator Variables......................... 64
Definition of Recurrence.................................65
Stress/Recurrence Relationship..........................67
Influence of Moderator Variables..........................68
The Nature of the Life Stress/Recurrence Relationship.....75

FOUR DISCUSSION............................................... 78


A STRUCTURED INTERVIEW ............................. ....91

RECURRENCE VARIABLES...................................94

DEFINITIONS.............................................. 98

REFERENCES............................................... ........ 99

BIOGRAPHICAL SKETCH............................................ 111

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




May, 1986

Chairman: James H. Johnson
Major Department: Clinical Psychology

The present study of the effects of major and minor life events

on recurrences of genital herpes addressed the need for prospective

research with objective indices of illness (i.e., viral culture) and

provided empirical information regarding a stress/recurrence link.

For 6 months, 122 HSV seropositive subjects monitored stress and

had all recurrent lesions cultured. Subjects were administered the

Life Experiences Survey (LES) upon enrollment regarding major events

experienced over the past 6 months. This information was analyzed to

determine the extent of a prospective relationship to subsequent

recurrences. Thereafter the LES was administered monthly with the

Hassles Scale (HS) which concerns minor stressors. To identify

personal characteristics which might be related to frequent

recurrences under conditions of stress, the following presumed

moderator variables were examined: social support, locus of control

orientation, and arousal seeking tendency.

In accordance with the first hypothesis, correlational procedures

provided evidence for a positive association between the frequency of

major negative life events and the number of HSV recurrences during

the study. This relationship was significant for concurrent LES

scores, but not when initial LES scores were considered in a

prospective manner to subsequent recurrences.

A logistic regression procedure using all subjects provided

evidence for a moderating influence of social support satisfaction,

the nature of which was contrary to expectation. A significant

positive stress/recurrence relationship was found only at high levels

of satisfaction. As hypothesized, information from recurring subjects

(n=49) submitted to a regression analysis revealed a significant

positive stress/recurrence relationship only for externally oriented

individuals. Neither arousal seeking nor hassles were significantly

related to recurrences in any analysis.

Within-subject analyses were performed to examine the nature of

the concurrent stress/recurrence relationship. Stress was not found

to be significantly elevated prior to recurrences. This suggests

either a shorter latency exists between stress and recurrence onset

than could be detected; using monthly evaluations or a reciprocal

relationship exists in which stress may cause or be caused by



Stress and Illness

The notion of a link between psychosocial factors and illness is

not new. Physicians have prescribed environmental and behavioral

interventions for years based upon their clinical experience and

intuition. What is new is a flurry of activity designed to document

and clarify this relationship.

With regard to life stress as a psychosocial factor in the

etiology of illness, Holmes and Rahe (1967) can be credited with

initiating a new wave of research with the development of their

measure, the Schedule of Recent Experience (SRE). This was the first

attempt to quantify, using a simple, 43-item questionnaire format, the

stressful life events believed to heighten vulnerability to illness.

The SRE and various modifications of this measure have been

widely used to provide empirical evidence for the relationship between

life stress (defined ip terms of major life changes) and illness.

Consistent but modest correlations have been confirmed with a variety

of illnesses and illness-related measures in the past two decades.

Elevated life stress has been shown to be associated with higher

reported illness rates among many groups (Holmes & Masuda, 1974; Marx,

Garrity, & Bowers, 1975; Rahe, Mahan, & Arthur, 1970) the seriousness

of illness (Wyler, Masuda, & Holmes, 1971), athletic injuries

(Bramwell, Masuda, Wagner, & Holmes, 1975), work-related accidents

(Levenson, Hirschfeld, Hirschfeld, & Dzubay, 1983), and the

complications of pregnancy and childbirth (Gorsuch & Key, 1974; Newton

& Hunt, 1984; Nuckolls, Cassel, & Kaplan, 1972; Rizzardo et al.,

1982). Illness-related behaviors such as the seeking of medical

attention and taking disability days have also been associated with

higher scores on life change indices (Gortmaker, Eckenroade, & Gore,

1982; Liao, 1977; Miller, Ingham, & Davidson, 1976; Murphy & Brown,

1980; Tessler, Mechanic, & Dimond, 1976; Thurlow, 1971).

Similarly, the onset of particular diseases has been associated

with elevated life change stress scores. These include myocardial

infarction (Connolly, 1976; Theorell & Rahe, 1971), sudden cardiac

death (Rahe & Lind, 1971), childhood cancer (Jacobs & Charles, 1980),

gastrointestinal disorders (Craig & Brown, 1984), appendicitis (Creed,

1981), tuberculosis (Hawkins, Davies, & Holmes, 1957), mononucleosis

(Kasl, Evans, & Niederman, 1979), hypertension (Lal, Ahuja, &

Madhukar, 1982), menstrual abnormalities (Harris, 1984; cited in

Creed, 1985; Siegel, Johnson, & Sarason, 1979), chronic yeast

infections (Williams & Deffenbacher, 1983), streptococcal infections

(Meyer & Haggerty, 1962), and experimentally induced cold infections

(Totman, Kiff, Reed, & Craig, 1980). What is striking in these

examples is the wide range of health-related problems which bear a

significant relationship to measures of life stress. It is as though

excess stress increases one's vulnerability to illness in general, not

just a few specific disorders. Indeed, this is consistent with the

views of Holmes and Masuda (1974) that life stress results in illness

through increased susceptibility to a range of health-related


In addition to studying the relationship between life stress and

illness onset, other studies have focused on the relationship between

life stress and fluctuations of health status in individuals with

chronic illness. Children with a mixed group of chronic conditions

(e.g., diabetes, asthma, blindness) who attended a 3-week summer camp

were observed by Bedell, Giordani, Armour, Tavormina, and Boll

(1977). A relationship between self-reported life change and

frequency of illness-related symptomatic episodes was found.

Similarly, asthmatic individuals monitored prospectively were found to

require more medication under conditions of increased life stress

(de Araujo, Van Arsdel, Holmes, & Dudley, 1973). Diabetic control has

also been investigated. Bradley (1979) found life stress to be

related to increases in blood sugar, the need for clinic visits, and

necessary prescription changes, particularly for insulin-dependent

diabetics. Similarly, insulin-dependent diabetics were found to be

more vulnerable to life stress than age- and sex-matched noninsulin

dependent diabetics and matched nondiabetic controls (Linn, Linn,

Skyler, & Jensen, 1983;). Insulin-dependent diabetics perceived more

stress and showed poorer metabolic control despite similar compliance

with the medical regimen. They also demonstrated decreased immune

responsivity (Linn et al., 1983). Jacobson, Rand, and Hauser (1985)

studied the impact of life events on the long term complications of a

group of insulin-dependent diabetics. A small but significant

relationship was found between negative life events and glycemic

control for all subjects. In the subgroup of diabetics with

proliferative retinopathy of recent onset, a stronger relationship was

found between lack of glycemic control and the occurrence of negative

life events (Jacobson, Rand, & Hauser, 1985).

Moderator Variables

Despite the consistent finding of statistically significant

correlations, the correlation coefficients are of low magnitude,

typically less than .30, which suggests that life events measures

typically can only account for 9% or so of the variance in illness

measures (Rabkin & Struening, 1976). To improve upon this situation,

the life events measures have been administered in conjunction with

other measures believed to assess moderators of life stress.

Constitutional predisposition (Kobasa, Maddi, & Courington, 1981),

social support (Nuckolls et al., 1972; Schaefer, Coyne, & Lazarus,

1981), locus of control (Toves, Schill, & Ramanaiah, 1981), hardiness

(Kobasa, 1979), physical fitness (Roth & Holmes, 1985), and a tendency

to engage in arousing situations (Johnson, Sarason, & Siegel, 1979)

have been shown to moderate the effects of life stress.

Social Support

A frequently researched moderator variable has been social

support. This refers to the degree to which individuals have access

to social resources. Resources depend upon the relationships people

have to spouse, family, friends, neighbors, community groups, and

social institutions (Johnson & Sarason, 1984). Social support has

been measured in a variety of ways, from simply confirming the

availability of a confidante (Brown, Bhrolchain, & Harris, 1975) to

administering a 48-item inventory with four subscales to evaluate

esteem support, instrumental support, social companionship, and

informational support (Cohen & Hoberman, 1983).

In a recent comprehensive review of over 40 research articles,

Cohen and Wills (1985) found evidence consistent with both an overall

beneficial effect of support on well-being and a buffering

(interactive) role in relation to life stress. In the former, health

benefits are perceived to derive directly from a supportive network.

In the latter, social support is perceived to play a moderating role,

to be related to well-being only for persons experiencing the adverse

effects of stressful events. That is, in the face of major life

stress, an adequate social support system may serve to protect the

stressed individual from a pathologic outcome. Authors point out that

social support may simultaneously exert both a direct and a buffering

effect (Cohen & Wills, 1985).

Social support was found to exert buffering or positive main

effects on most of the psychological distress variables reviewed, such

as depression, loneliness, and anxiety. In fact, in one investigation

(Lin, Ensel, Simeone, & Kuo, 1979), the contribution of social support

to predicting psychiatric symptoms was more important than stressful

life events. However, findings regarding physical symptomatology have

been more variable. A clear, consistent link between social support

and decreased mortality has been demonstrated (Berkman & Syme, 1979),

but studies utilizing other health outcomes are less clear. They are

mixed when the health outcome is self-reported symptomatology. When

more stringent health measures have been used such as clinical

diagnosis and alterations in physiological functioning, a beneficial

role for social support has not always been consistently demonstrated

(Cohen & Wills, 1985).

A few examples will be given to illustrate the diversity of

outcomes regarding social support. Evidence consistent with a stress-

buffering hypothesis comes from an early study of pregnant women by

Nuckolls et al. (1972). Those experiencing high levels of stress and

low levels of social support experienced almost three times as many

pregnancy and birth complications as their similarly highly stressed

peers who had high levels of social support. Similarly, adult

asthmatics in a high life stress-low social support group required

significantly more medication to control symptoms than their peers

with either less stress or more social support (de Araujo et al.,

1973). These examples demonstrate a protective, positive role for

social support with respect to certain physical outcomes.

In spite of the intuitive appeal of a positive effect of social

support, and in contrast to the above studies, Norbeck and Tilden

(1983) found no evidence for either a main effect or buffering role

with respect to pregnancy complications. They did, however, find a

positive main effect of emotional support on psychological

symptomatology (Norbeck & Tilden, 1983). New Zealand researchers

(Graves & Graves, 1985) failed to find the hypothesized buffering

effect of social support on self-reported symptoms of illness in three

distinct ethnic groups. In fact, of the correlations designated large

enough to be of clinical significance, all were in direct opposition

to the buffering hypothesis. It is suggested that the obligatory

reciprocity of dense social networks may be less helpful under

conditions requiring adaptation to life change (Graves & Graves,

1985). An interesting cyclical pattern of stress and social support

was demonstrated by Canadian investigators in a longitudinal study

(McFarlane, Norman, Streiner, & Roy, 1983). They found that help from

social support networks reduced exposure to stressful events.

However, increases in stressful events led to a reduction in perceived

help. No association was found between social support and physical

symptom reporting or physician visits (McFarlane et al., 1983).

Obviously, the relationship between social support, stress and

health is a complex one. Critics suggest the use of longitudinal

designs (Cohen & Wills, 1985; Thoits, 1982) and the measurement of

satisfaction (perceived quality) with available support as well as the

amount (quantity) of social support available (Sarason, Levine, &

Sarason, 1982) to further unravel the role of social support with

respect to major life events and health status.

Locus of Control

Another measurable variable that has shown promise as a moderator

of life stress is an individual's perception of the controllability of

situations. People vary in the degree to which they attribute the

responsibility for events to themselves or to fate. It is reasonable

to assume that individuals who perceive themselves as having little or

no control would feel more threatened in the face of undesirable life

events than those who feel capable of influencing outcomes. A more

intense adverse reaction could contribute to a more severe

physiological response to life stress.

The Rotter (1966) Locus of Control Scale measures the degree to

which individuals perceive themselves as having control over naturally

occurring life events. It has provided a tool for investigation of

perception of control, life stress, and psychological and physical

outcomes. Individuals are classified as internals on the scale if

they believe in personal responsibility and externals if they

attribute the control of events to chance. Johnson and Sarason (1978)

administered the Locus of Control Scale along with a major life events

scale and measures of depression and anxiety to college students.

They found negative life changes to be significantly related to both

trait anxiety and depression for externals, but not for internals. A

similar relationship was found for males in a study of life stress,

locus of control, and health status as measured by the Cornell Medical

Index (Toves et al., 1981). Negative life change was related to

health status only for external males. Females demonstrated a

relationship between negative events and illness regardless of

perception of control (Toves et al., 1981). In further support of the

notion of perceived control as a moderator of life stress, Kobasa

(1979) demonstrated that individuals with an internal locus of control

may be protected from adverse physical consequences of major life

change. Compared to business executives with high stress/high illness

scores, those with high stress/low illness indicated a more internal

locus of control. In this investigation, perception of internal

control emerged as one component of a constellation of factors named

"hardiness" (Kobasa, 1979).

The above studies are all retrospective in nature, and it could

be argued that individuals who have experienced recent illness would

see themselves as victimized and thus tend to respond to the locus of

control scale as externals because of their poor health status. To

strengthen the causal inference that hardiness (which includes an

internal locus of control orientation), Kobasa (1981) replicated her

findings in a prospective design suggesting that an internal locus of

control is related to resistance to illness under conditions of

stress. Manuck, Hinrichsen, and Ross (1975) followed college students

for 6 months after assessing locus of control orientation and life

stress. Although they found no difference in illness-related

treatment-seeking behavior between internals and externals under high-

stress conditions, more low-stressed externals sought treatment than

low-stressed internals (Manuck et al., 1975). Passer and Seese

(1983), however, failed to prospectively demonstrate a significant

moderating effect of perception of control with athletic injury as the

dependent variable. These findings suggest that locus of control

merits further investigation as a moderator of the effects of major

life stress.

Arousal Seeking

Individuals differ in the degree to which they seek out or

attempt to avoid arousing situations. Inventories have been devised

to evaluate this tendency enabling researchers to investigate this

characteristic as a moderator of life stress (Mehrabian, 1978;

Mehrabian & Russell, 1974; Zuckerman, 1971). It is hypothesized that

those with a low optimal level of arousal, who tend to avoid

stimulation (low arousal seekers), would be more likely to respond

negatively to undesirable life events than those who have a high

optimal level of arousal and thus seek out stimulation (high arousal

seekers). High arousal seekers are presumed to be better able to deal

with the increased arousal brought about by the experiencing of life

changes (Johnson & Sarason, 1984).

This prediction has been addressed by a few researchers. Smith,

Johnson, and Sarason (1978) found negative life change to be

significantly related to discomfort scores on the Psychological

Screening Inventory for only the college students low on a sensation

seeking measure. Extending these findings using a different measure

of arousal seeking and measures of anxiety, depression, and hostility,

Johnson et al. (1979) found negative change to be related to anxiety

and hostility only for low arousal seeking subjects. At least one

researcher (Cohen, 1982) has failed to replicate the above results.

These preliminary findings with regard to psychological outcomes

suggest that a moderating effect of arousal seeking tendency on life

stress with regard to certain physical outcomes may also be worth

investigation. Individuals low on the arousal-seeking dimension may

be much more likely to experience physiological reactivity in the face

of life stress than others, which could adversely affect general


In summary, as a means of investigating the variability in

individual susceptibility to health status changes in the face of

undesirable life stress, the investigation of the roles of moderator

variables appears to be warranted. The potential influence of social

support, locus of control, and arousal seeking tendency are of

particular interest based on findings to date.

With the exception of the Kobasa investigations (Kobasa, 1979;

Kobasa et al., 1981), moderator variables have been evaluated singly

regarding their potential to alleviate or exacerbate the negative

effects of life stress. More information is needed on how these

variables interact or combine in an additive fashion to influence

health status in the face of life stress. Multiple regression

techniques provide an analytic tool appropriate for clarifying the

relative importance of multiple potential moderator variables.

Increased documentation of significant associations between

stressful life events and the onset or course of physical illness has

changed the degree of complexity of the empirical questions. Instead

of asking if stress is related to a particular illness, the questions

to answer now are for which individuals, under what conditions, and to

what extent are life events related to this disease? Answering more

complex questions requires more complex research designs (Johnson &

Sarason, 1984). In the next section, some of the methodological

shortcomings of present research will be discussed. Addressing these

issues will enable us to be more specific about our research questions

and to further clarify the nature of the stress/illness relationship.

Methodological Issues

Methodological issues limit interpretations and findings in this

area of research. These include retrospective bias, inadequate

indices of illness, failure to discriminate between desirable and

undesirable events, and failure to account for minor stressful events.

Retrospective Design

The commonly utilized retrospective design relies on recall of

events and illnesses. Findings based on memory may reflect

individuals' expectations regarding stress and illness, their need to

justify illness (Rabkin & Struening, 1976), or their tendency to

report negatives due to feelings of depression (Johnson & Sarason,

1984). Retrospective reporting is also subject to forgetting as

demonstrated by Monroe (1982a). He investigated memory of major life

events utilizing prospectively collected information as a baseline to

indirectly estimate the most recent retrospectively reported four-

month period. Underreporting of events was estimated in this way to

be as high as 61%, with desirable events relatively more susceptible

to distortion with passing time than undesirable events (Monroe,

1982a). Funch and Marshall (1984) examined the rate of fall-off in

event reporting over a retrospective 30-month period. Using estimates

based on 12- and 18-month recall periods, they concluded that fall-off

is most rapid in the first 12 months, approximately 5% per month. It

then tends to level off. Fall-off was related to type of event in

their research, the more salient events showing the least fall-off.

Respondent variables such as income, education, and marriage were

related to the tendency to remember disruptive events. That is,

subjects with fewer resources were most likely to underreport

stressful events (Funch & Marshall, 1984). In a review of several

articles which address the issue of event underreporting, Paykel

(1983) found consistency in estimates of 4-5% loss per month in

retrospectively self-reported life events, with a lower rate (1-3%)

for event reporting by interview. To avoid retrospective bias due to

forgetting, frequent short-time intervals for reporting have been

recommended (Cleary, 1980; Monroe, 1982a), with interview techniques

used whenever possible (Paykel, 1983).

Prospective studies avoid this kind of distortion. Jenkins,

Hurst, and Rose (1979) reviewed studies on cardiac and cancer patients

finding prospective studies which contradicted or diminished the

significance of earlier findings of retrospective investigations.

Unfortunately, good prospective studies of the onset of illness are

very cumbersome and costly to implement. Many people must be

investigated who may never get the disease or diseases in question

during the course of study. For example, of 1400 cadets studied for

four years, only 194 became infected with infectious mononucleosis

during the investigation and, of these, only 48 developed clinical

symptomatology (Kasl, Evans, James, & Niederman, 1979). In another

example, 4,486 widowers were followed for 9 years to conclude that

there is a 40% increase in mortality within the first 6 months of

bereavement (Parkes, Benjamin, & Fitzgerald, 1969). Prospective

investigations of the effects of life stress on the clinical course

and outcome of disease are more easily and economically

accomplished. These involve prospectively following already

identified patients with respect to life stress along with

fluctuations in their health status regarding the illness in

question. The data from all subjects remain important for analysis,

not just those who happen to die or become infected with a particular


Measures of Illness

Another weakness of research in this area has been the failure to

employ strict measures of illness (Johnson & Sarason, 1984). Self-

reports of symptoms, the seeking of medical treatment, and medication

needs may not reflect the existence or severity of a physical

disorder. Additionally, high life stress may increase the tendency to

seek help for minor illness (Mechanic, 1975) or lower an individual's

threshold of complaint and tendency to adopt the sick role (Minter &

Kimball, 1978). Harney and Brigham (1984) experimentally demonstrated

that an individual's tendency to tolerate discomfort (cold pressor,

loud noise) is related to the amount of recent life change

experienced. This suggests that high life change individuals may be

more likely to seek aid than others for identical symptoms (Harney &

Brigham, 1984). One way to avoid these confounding factors is to

employ more rigid measures of illness. Rather than reliance on

illness behavior, medical records, or self-report, more objective

indices are suggested such as x-ray and other clinical and laboratory

data which provide biological evidence of the presence or severity of

disease (Minter & Kimball, 1978).

Confounding of Events and Outcome Measures

A concern of many in this area of research in the possible

confounding of independent and dependent variables in life stress

measures (Johnson & Sarason, 1984; Monroe, 1982b). Many items which

are considered to be stressful events may also be consequences of

illness (e.g., sexual problems, major personal injury, changes in

eating or sleeping habits). An argument can be made that such items

should be removed or analyzed separately from the domain of major life

events to avoid artificially inflating the stress scores of

individuals with illness (Dohrenwend & Dohrenwend, 1974). However,

such experiences may themselves represent significant sources of

stress. To ignore them in a consideration of the total stress of an

individual may be to neglect important information and result in a

loss of sensitivity in the life stress measure. In an investigation

of etiological factors associated with a particular disease, a

division of items into three subgroups for analysis is recommended:

(a) events that may be confounded with psychiatric condition, (b)

events consisting of physical illness and injury, and (c) events which

are independent of either an individual's health or psychiatric

condition (Dohrenwend & Dohrenwend, 1974; Monroe, 1982b). When the

major purpose of the investigation is to predict illness from stress

levels, it is reasonable to include all items as reflective of the

total stress experienced (Johnson & Sarason, 1984). This would be

particularly appropriate when the illness dependent variable is

specific and objective (e.g., laboratory documentation of the presence

of a disease organism). Confounding problems are more likely when

general, subjective measures of illness serve as the dependent

variable (e.g., reported symptoms, treatment-seeking behavior).

Value and Impact of Events

Other methodological considerations include the value (negative

or positive) and impact (weighting) of events. Until recently,

investigators have failed to discriminate between desirable and

undesirable events. In those investigations in which a distinction

has been made, negative change has proven to be more predictive of

personal maladjustment (Mueller, Edwards, & Yarvis, 1977; Vinokur &

Selzer, 1975) and physical illness (Hotaling, Atwell, & Linsky, 1978;

Johnson & Sarason, 1984; Sheehan, O'Donnell, Fitzgerald, Harvey, &

Ward, 1979).

Different methods have been explored for weighting the impact of

stressful events to obtain a more sensitive measure of individual or

consensus-derived perception of the stressfulness of particular life

events. However, raw event totals and weighted event totals are

typically highly correlated with each other. Zimmerman (1983) found

an average correlation of .94 in a review of 19 studies. Since the

differential weighting procedures are functionally equivalent to

unweighted procedures in group research, it is parsimonious to employ

the simplest method: a simple count of events (Johnson & Sarason,

1984; Monroe, 1982b; Skinner & Lei, 1980; Zimmerman, 1983). This does

not deny the clinical relevance of subjective weighting procedures for

understanding particular individuals. It simply acknowledges that

such refinements are unnecessary and redundant for group research.

Minor Events

The final methodological issue relates to the consideration of

minor as well as major life events. Because major life events per se

do not represent all stressful events, the Hassles Scale (HS) has been

developed (Kanner, Coyne, Schaefer, & Lazarus, 1981). In contrast to

major life events such as divorce, job loss, and pregnancy, the HS

deals with minor stresses, annoying practical problems such as making

mistakes, being bored, and having to wait. Already there is some

suggestion that hassle scores may be more strongly related to a

general health measure than life events scores (DeLongis, Coyne,

Dakof, Folkman, & Lazarus, 1982).

In order to account for a greater proportion of the variance in

the relationship between stress and illness, the present study was

designed to address some of the existing methodological limitations.

To avoid retrospective bias, a prospective research design with

monthly concurrent assessments of stressors was planned. The Life

Experiences Survey (LES) was chosen to measure major life stress

because it distinguishes between negative (undesirable) and positive

(desirable) events (Sarason, Johnson, & Siegel, 1978). The Hassles

Scale (HS) was utilized to evaluate minor stressful events (Kanner et

al., 1981). Additionally, three hypothesized moderator variables were

measured to evaluate their effectiveness in combination in

ameliorating the effects of stress. And, finally, a population was

chosen for which an objective index of illness could be obtained.

Recurrent Genital Herpes as a Model

The choice to study individuals with recurrent genital herpes

virus in this investigation was based on the nature of the illness,

the need for information concerning the precipitants of reactivations

of the virus, the ease of obtaining an objective measure of the

disease, and the availability of a motivated population of subjects.

Given these considerations, an investigation of herpes simplex virus

would seem to be a fruitful means of studying the relationship between

life stress and fluctuating health status in chronic illness.

Clinical Nature of Herpes Simplex Virus

Herpes simplex virus (HSV) types 1 and 2 are two of several

herpes viruses which affect humans and share a unique capacity to

produce lifelong latent infection despite host development of antibody

and cellular immunity. Serious complications can occur including

ocular infection, encephalitis, and neonatal infection. Fortunately,

these are rare and can usually be prevented. There also appears to be

an association between genital herpes and cervical cancer, but so far,

direct evidence of a causal link is lacking (Rand, 1982).

Neither serious complications nor significant physical

consequences are responsible for the widespread fear and attention

elicited by the disease. There is evidence that the incidence of

genital herpes is rising rapidly. Consultations to clinicians for

genital herpes have increased almost nine-fold in the past 20 years

(Mertz & Corey, 1984). Its incurability, transmissability, rising

incidence, and sexual nature create the concern. This can only be

understood in light of information regarding the natural course of the


Herpes is almost always transmitted by intimate contact between

infected and noninfected skin surfaces, particularly the mucous

membranes of the oral cavity and genitalia. Once inside, herpes takes

over the protein-producing apparatus of the host cell and reproduces

itself until it is brought under control by the immune system. During

an initial (primary) infection, this process produces multiple,

painful blisters or ulcerative lesions accompanied by fever, swollen

lymph nodes, and malaise. The lesions last an average of 19 days

(Mertz & Corey, 1984).

Within hours of initial exposure, the virus retreats via nerve

fibers to a safe hiding place in the trigeminal ganglia (oral herpes)

or the sacral ganglia (genital herpes) which is located near the

spinal cord. It is this capacity of the virus to reside latently

without destroying the ganglion and to periodically reactivate to

cause recurrent lesions which frustrates patients, doctors, and



Almost all patients experience recurrences, from one in a

lifetime to two or three a month. The risk of recurrence when the

infectious agent is HSV-2 is higher than with HSV-1 regardless of the

site of the infection. Otherwise, the clinical manifestation of HSV-1

and HSV-2 is identical (Mertz & Corey, 1984). A diverse group of

factors are associated with reactivation of HSV including fever

(Keddie, Reeves, & Epstein, 1941), local irritation or tissue injury

(Blyth & Hill, 1984; Wickett, 1982), lack of sleep, overexertion, and

emotional stress (Bierman, 1982). Although there is widespread

acceptance of a stress/recurrence relationship, the evidence for it is

weak. The notion is based on anecdotal reports and a handful of

retrospective studies. The further clarification of the nature of

this relationship is important for proper disease management.

The physical consequences of recurrences are minimal. The

lesions cause discomfort but are rarely severe. They may last up to 2

weeks, but are sometimes fleeting, lasting a day or less. The patient

can usually carry out normal daily activities during recurrent

episodes. Between recurrences patients are completely well. However,

patients are highly infectious during a recurrent episode and may

transmit the disease even when asymptomatic (Mertz & Corey, 1984).

Patients find the periodic imposed abstinence from sexual activity to

be disruptive and to result in decreased sexual spontaneity. Female

patients have concerns about their reproductive capability. Single

individuals must deal with disclosure to potential partners, causing

embarrassment, shame, and fear of rejection (Luby & Klinge, 1985).

Unfortunately, because of its sexual nature, biased reporting

regarding genital herpes in the media has contributed to popular panic

about the disease, stigmatizing patients, and exacerbating the

emotional turmoil involved (Bierman, 1985).

Present Lack of Cure

At the present time, no cure is available for the disease. It

has recently been demonstrated that oral administration of daily doses

of acyclovir suppresses reactivation of the virus in patients with

frequently recurring genital herpes (Douglas et al., 1984; Straus et

al., 1984). However, when the prophylactic therapy ceases,

recurrences resume at the prior rate. And, unfortunately, the expense

of the drug and as yet unknown long term side effects prohibit its use

by many. Until a cure can be found, management of the disease must

consist of counseling patients to prevent transmission by practicing

sexual abstinence during recurrences and preventing recurrences.

Prevention of frequent recurrences with acyclovir is an option for

some. With better knowledge of the precipitants of frequent

recurrences of genital herpes, behavioral counseling for drug-free

prevention for all patients will improve. Thus, there is a compelling

need to learn more about the psychological and behavioral factors

associated with recurrences.

Laboratory Documentation

It is possible to obtain objective evidence of the presence of

herpes simplex virus by laboratory methods. The virus may be

cultivated in tissue culture, viral particles may be detected by

electron microscopy, or immunologic methods may be used to detect

viral antigen or cytopathologic change. The isolation of the virus in

tissue culture has been shown to be the most sensitive measure. The

virus has been successfully isolated in 82% of ulcerative lesions in

primary genital herpes episodes and 42% of ulcerative lesions in

recurrent episodes. The decrease in percent viral confirmation in

recurrent episodes as opposed to primary can be attributed to a

shorter duration of viral shedding in the former (Corey & Holmes,

1983). The fact that objective evidence of the presence of the

disease is possible also contributed to the choice to study this

particular group of individuals in a stress/illness paradigm.

Endogenous Pathogen

A final factor regarding recurrent genital herpes is that

extraneous exposure to an unknown dosage of an exogenous pathogen was

not an issue. Stress would be most important as a contributing factor

in the etiology of a disease when the host/microorganism relationship

is in delicate balance (Plaut & Friedman, 1981). Uncontrolled,

extraneous dosages of a pathogen may be so high that disease is

inevitable, or too low to cause disease under any conditions.

Likewise, the more virulent the microorganism, the lower the dose

necessary to introduce infection. Recurrent herpes lesions depend

upon an endogenous source of infection which remains stable within an

individual unless reinfection occurs from an outside source.

Genital herpes, therefore, provides a good model for the study of

stress/illness relationships. It is a chronic, recurring disease

which causes emotional turmoil for patients and their sexual partners

despite minimal physical consequences. Between recurrent episodes,

patients are unaffected by disease presence. Patients are thus

motivated to avoid recurrences and willing to participate in research

to learn more about precipitants of herpes reactivations. Viral

isolation is possible by laboratory techniques providing an objective

measure of the presence of the disease, and extraneous exposure to

pathogens of unknown dose and virulence is not a concern. Patients,

physicians, and individuals who have never had genital herpes will all

benefit from improved understanding of the disease.

Stress and Herpes

In the following section, investigations which have attempted to

clarify the relationship between stress and herpes will be

presented. Herpes simplex virus-type 2 most commonly affects the

genital area, but HSV-1 may also do so. Because they are clinically

indistinguishable (Rand, 1982), research including either HSV-1 or

HSV-2 will be reviewed.

Anecdotal Reports

Schneck (1947) reported the case of a male who could predict

HSV-1 recurrences which occurred within 24 hours of emotional

stress. These episodes were most likely to occur when appropriate

outlets for hostile feelings were unavailable. According to the

account, improved management of hostility resulted in decreased


Blank and Brody (1950) provided psychoanalysis for 10 patients

with recurrent oral herpes. Patients were seen weekly for 2 to 50

hourly sessions. Nine of the ten patients were described as passive,

anxious to please, dependent, immature, and hysterical. In two cases,

cancellation of an appointment by the therapist and the resultant

feeling of rejection is said to have precipitated a herpes

recurrence. From their experiences with this group of patients,

authors claim support for the use of psychotherapy in decreasing the

frequency of recurrences (Blank & Brody, 1950).

There have been reports of an association between hypnotic

suggestion and recurrent herpes episodes. Herlig and Hoff (cited in

Janicki, 1971) successfully induced oral herpes recurrences in three

female patients by reminding them of unpleasant emotional situations

and itching sensations while under hypnosis. Recurrent lesions

appeared in 1 or 2 days in all three subjects. Ullman (1947)

similarly induced an oral herpes episode in a male patient by giving

hypnotic suggestions that herpes blisters would form and that he

appeared debilitated. Within 24 hours the man had blisters on his

lower right lip which appeared to be oral herpes. Hypnosis has also

been reported to have a positive effect on the recurrence rate of

HSV-1 (McDowell, 1959).

Many supposed "cures" of herpes have been attributed to placebo

effects, or positive expectancy. One investigator reports favorable

reactions in 50% of herpes patients who received injections of sterile

water (Kern, 1979). Hamilton (1980) reports even more dramatic

placebo results. In a double-blind placebo-controlled trial of ether

as a treatment for herpes, 75% of the patients on placebo experienced

improvement of herpes symptomatology (Hamilton, 1980, p. 54).

Anecdotal reports of an association between hypnosis or

expectancy and episodes of recurrent herpes are suggestive, but

inconclusive. Failures to induce herpes blisters through suggestion

are unreported. However, the existence of a link between emotional or

cognitive factors and recurrent herpes as can be hypothesized from

these reports is worthy of more systematic investigation.

Survey Research

Results of epidemiological survey research also suggest a link

between emotional factors and recurrences. In a national survey of

3,148 individuals with genital herpes, 83% of the respondents

acknowledged stress as a factor in recurrences ("Help membership HSV

survey," 1981). The subjects attributed the following ill effects to

having contracted herpes: periodic depressions (84%), sense of

isolation (70%), conscious avoidance of intimacy (53%), cessation of

sexual activity (10%), diminished sex drive and/or impotence (35%),

suicidal feelings (25%), dissolution of a long term marriage or

relationship (18%), rejection by a potential sex partner (21%), and

loss of self esteem (40%). Herpes is cited as both a cause and effect

of stress by these participants who were recruited through the

newsletter of a support network for individuals with genital herpes.

Two-thirds of the sample reported experiencing more than five

recurrences in 1 year. Although the sample may not be representative

of all genital herpes patients, it does reflect the perceptions of a

substantial number of patients. Another, possibly more representative

sampling of 825 patients, 53% of whom believed they were cured of

genital herpes, yielded a similar finding. In this group, emotional

stress emerged as the most important self-reported mechanism for

triggering recurrences, cited by 86% of the sample (Bierman, 1982).

Retrospective Research

Taylor (1978) studied life events and herpes recurrences in a

sample of 60 female students, of whom 36 had genital herpes and 24

served as controls. All were administered the Life Experiences Survey

and a Genital Herpes Questionnaire. Women with genital herpes did not

report a higher frequency of stressful events over the prior year than

women without herpes. Within the herpes groups, however, women with a

high rate of recurrence (four or more in previous year) were compared

to those with a low recurrence rate (three or less). The high

recurrers reported having significantly more negative events than the

low recurrers. Due to the retrospective nature of the study, it is

impossible to know whether high levels of negative events were a cause

or effect of herpes reactivations. In the same study, 70% of the

women indicated that having genital herpes caused emotional stress.

Additionally, 92% of the herpes patients recalled experiencing stress

the week prior to a recent occurrence, while only 42% of normal

controls recalled stress at a comparable period. The potential

retrospective bias is apparent. In addition to patients' expectations

regarding stress and illness, they had a dramatic event (herpes

episode) to tie their recollections to while control subjects may not

have had a temporally contiguous event to aid in their recall of

stresses occurring at that time. These findings are supportive of a

stress/recurrence relationship, but are limited by retrospective bias,

the lack of objective documentation of the illness measure, and the

restricted nature of the sample (female students).

Watson (1983) extended Taylor's (1978) findings in an

investigation of 51 male and female volunteers who had at least one

recurrence of genital herpes within the 7 months prior to

participation. A significant positive relationship was found between

undesirable life events in the previous year as reported on the Life

Experiences Survey and genital herpes recurrences in the previous 6

months. Recurrences in the prior 6 months were also significantly

related to undesirable life events reported for the 12- to 6-month

period prior to the study, suggesting a long term effect of major life

events on herpes recurrences.

Watson (1983) is the only investigator to consider potential

moderators of life stress and herpes recurrences. He included

measures of perceived locus of control and social support in his

research design. Using the Rotter Internal/External Control Scale

(Rotter, 1966), Watson (1983) demonstrated an interaction between

locus of control and stress in which subjects who perceive internal

control experienced fewer recurrences under high stress conditions

than those with an external locus of control. This interaction effect

suggests a moderating influence of locus of control on the negative

effects of stress. Social support as measured by total helpfulness on

the Social Relationship Scale did not appear to have a similar

moderating effect on stress. However, a more direct effect was

found. Subjects reporting high levels of helpfulness had

significantly fewer recurrences than those reporting low helpfulness

(Watson, 1983). This investigation emphasizes the importance of

investigating potential moderating factors. It is limited by reliance

on retrospective reporting and the nonrepresentative nature of the

sample, primarily high recurrers recruited from a self-help support


Prospective Research

A series of prospective investigations have been undertaken to

examine the relationship of psychosocial variables and recurrences of

oral HSV (Friedman, Katcher, & Brightman, 1977; Katcher, Brightman,

Luborsky, & Ship, 1973; Luborsky, Mintz, Brightman, & Katcher,

1976). Independent measures included scores on the Life Change Index

reported retrospectively over the 2 years prior to the study, the

Cornell Medical Index, John Hopkins Symptoms Index, Clyde Mood Scale,

a Social Assets Scale, illness history, and blood analysis. Dependent

variables included illness records, infirmary visits, and, when

possible, documentation of HSV recurrences.

A year long investigation of 67 paid volunteers, 14 of whom had

significant antibodies to HSV, found previous history and blood serum

antibody presence most predictive of oral herpes recurrence, with the

unhappy factor of the mood scale also contributing significantly

(Katcher et al., 1973). When only the psychological variables were

considered, the unhappy factor accounted for 16% of the variance. The

relationship between predictors was different for herpes and systemic

illness, suggesting that recurrences do not reflect an illness

disposition. A significant negative relationship of social assets and

herpes episodes implies a possible moderating effect of social

assets. The 2-year retrospective life events measure was of minimal

value in all analyses and was disregarded in the subsequent

investigations (Friedman et al., 1977; Luborsky et al., 1976). While

life events were not an important variable in this study, this may be

explained on the basis of methodology. The number of subjects who

were seropositive for herpes was small. After attrition, only 38

subjects completed a full year of participation. The stepwise

multiple regression procedures were based on 19 episodes of herpes,

only 10 of which were documented by examination and viral culture.

The life events measure may have been insensitive due to memory

decline (Monroe, 1982a), the failure to consider negative events

separately (Dohrenwend & Dohrenwend, 1974; Johnson & Sarason, 1984),

or a lack of association between temporally remote events and

recurrences. However, the investigators anecdotally report an

increase of herpes recurrences during exam time for students (Katcher

et al., 1973).

A 3-month daily investigation of 20 herpes seropositive subjects

failed to demonstrate any systematic relationships between mood and

herpes reactivations (Luborsky et al., 1976). However, almost all

subjects reported stresses prior to episodes when responding

retrospectively. The authors suggest that by having moods reported in

the morning, the emotional impact of the day was missed. But it may

also be the case that moods are not necessarily equivalent to stress.

The researchers then followed a larger group of participants for

3 years (Friedman et al., 1977), during which 51 of 149 had at least

one HSV episode. Using herpes incidence as the dependent variable in

a multiple regression analysis, the social and psychological variables

were found to be of least predictive value (3-5% of explained

variance), biological factors (disease history and illnesses,

especially upper respiratory infection) being of greatest value

(80-90% of explained variance). Unfortunately, these investigators

did not include a stress measure. They replicated the earlier finding

that predictors were specific for herpes and unrelated to predictors

for upper respiratory infection (Friedman et al., 1977; Katcher et

al., 1973).

Daily samples of vaginal secretions from five women, three of

whom were positive for genital herpes were collected for a month by

investigators of asymptomatic viral shedding (Adam, Dressman, Kaufman,

& Melnick, 1980). Herpes simplex virus was isolated at least once

from each of the women with a history of herpes. Multiple positive

results by immunoperoxidase staining of cervical-vaginal smears were

found for all three. No statistical analyses were performed, but

graphical representation revealed temporal clusters of positive

results for each woman. These were anecdotally related to

menstruation, emotional stress, exam periods, and problems with

children. Unfortunately, the stress data were not collected in a

systematic manner making findings difficult to interpret (Adams et

al., 1980).

The data from a group of 58 psychiatric patients with primary

episodes of genital herpes who were followed prospectively for 28

weeks by Goldmeier and Johnson (1982) provide some support for a link

between psychosocial factors and recurrence. Subjects completed a 60-

item screening measure of psychiatric symptoms, the General Health

Questionnaire, on which they indicated which feelings and behavior

(symptoms) had been experienced over the past 4 weeks. They were then

requested to return to the clinic for viral cultures in the event of a

recurrence over the next 28 weeks. Subjects who did not return were

mailed a follow-up form requesting recurrence information. Of the 29

subjects reporting a recurrence, 7 returned for culture confirmation

of herpes. General Health Questionnaire Scores of recurrers were

compared to the nonrecurrers. The nonrecurrers had significantly

lower scores than the recurrers. Furthermore, the 29 subjects above

the cut-off score of 11 which indicates potential or overt psychiatric

problems had a significantly higher recurrence rate as determined by

actuarial recurrence-free curves. Investigators infer that anxiety or

obsessionality may increase production of adrenergic substances and

contribute to herpes reactivation. This hypothesis is consistent with

literature implicating the autonomic nervous system in decreases in

immune responsivity (Hall & Goldstein, 1981; Rogers, Dubey, & Reich,



To summarize, despite the widespread popular notion that stress

and recurrences of herpes are linked, consistent empirical support for

such a relationship is lacking. A few anecdotal reports suggest a

connection between psychological distress and recurrences (Blank &

Brody, 1950; Schneck, 1947) and hypnotic suggestion is reported to

influence recurrences (Janicki, 1971; Ullman, 1947). Strong placebo

effects have been apparent in controlled investigations (Hamilton,

1980; Kern, 1979). Survey research demonstrates that patients with

recurrent herpes believe there is a connection with emotional stress

(Bierman, 1982; "Help survey," 1981). In retrospective investigations

of patients with genital herpes, high recurrence rates have been

associated with high major life stress (Taylor, 1978) and, in one

study, negative life stress appears to have exerted a long term effect

on recurrence rate (Watson, 1983). Unfortunately, anecdotal reports,

survey research, and retrospective studies are all subject to


In one prospective investigation of oral herpes recurrences, a

negative mood factor contributed significantly to recurrence rate, but

a life events measure proved to be unimportant (Katcher et al.,

1973). Moods failed to hold up as important in subsequent prospective

studies in which life stress was unfortunately not measured (Freidman

et al., 1977; Luborsky et al., 1976). Anecdotally related stress

episodes were graphically associated with laboratory evidence of viral

shedding for three women (Adams et al., 1980). And finally, patients

demonstrating a strong potential for psychiatric problems on a

screening device subsequently reported more recurrences than those

falling below the critical cut-off score, suggesting a connection

between anxiety and recurrences (Goldmeier & Johnson, 1982). So far,

empirical support for a stress/herpes association is suggestive rather

than strong. A longitudinal prospective study with regular,

systematic collection of stress data and laboratory documentation of

recurrences would contribute valuable information.

In the following section, an attempt will be made to argue for

the significance of such a study in the broader scheme of

health/illness research. A reasonable proposal will be made of the

sequence of environmental, psychological, and physiological events

which could take place if stress is, indeed, a significant factor in

precipitating herpes recurrences.

Theoretical Framework

Perception of Stressful Events

Between an environmental event and a recurrence of genital herpes

many levels of complex interactions are possible. Figure 1

illustrates in simplified form the hypothesized process by which

environmental events could result in recurrences of genital herpes.

First, there must be a stimulus, the occurrence of a major or minor

life event. For it to be reported as a stressor or a hassle, it must

be perceived as such. Moderating influences may affect an

individual's perception of events in either negative or positive

ways. A person who craves excitement would probably view the event of

a flat tire enroute to work differently than a counterpart who thrives

on routines and predictability. A person who feels cared for and

supported by family and friends will probably be less threatened by

criticism on the job than a more solitary individual. The point is

that the same event could be perceived as a stressor by one

individual, but not another.





4- a


0C c


E 0.


4- I







- U

c/) 'r



(o D

0 I,

- a






I s




Once an event is perceived as a stressor, it may initiate a

variety of physiological responses as the individual struggles to

adapt to the new situation. Just as moderator variables may influence

a person's perception of stressors, they may influence biological

reactions as well. Internal adaptation responses may occur in the

nervous, endocrine, and immune systems. Awareness of internal

responses may, in turn, affect perception in a feedback loop in a

manner to exacerbate or diminish the experience of threat. The

resultant fluctuation may upset the delicate balance necessary for

herpes simplex virus to remain in its quiescent existence. The virus

could then reactivate and result in an episode of recurrent herpes.

In the following section, potential mechanisms by which perceived

stress could result in a herpes recurrence will be considered.

Proposed Physiological Mechanisms

The precise triggering mechanism which leads to reactivation of

herpes simplex virus is elusive. While residing in the ganglion, the

virus neither destroys the latently infected cells nor is eliminated

by host defense mechanisms or prolonged antiviral chemotherapy.

Something occurs at the molecular level to stimulate it to migrate to

the skin surface and begin viral replication (Bierman, 1983; Stevens,

1975). To explain this phenomenon, two major hypotheses have been

proposed. In one hypothesis, the dynamic state hypothesis, a small

number of cells constantly replicate virus. Specific immune defense

responses involving antibody, lymphocytes, and interferon continually

work to localize and prevent the clinical manifestations of recurrent

infection (Stevens, 1975). This theory, also called the "skin trigger

theory," proposes that recurrences result when local defenses are

temporarily suppressed, or when slight physiological changes (other

than immune mechanisms) alter the balance in favor of the virus (Hill

& Blyth, 1976).

The static state hypothesis or "ganglion trigger" theory proposes

that the virus remains inactive in the ganglia until nervous system

stimulation provokes it to migrate toward the skin surface. In this

theory, the viral genome remains in an unproductive, quiescent state

(Stevens, 1975). At present, the weight of evidence favors the static

state hypothesis. Investigators that once believed in the dynamic

hypothesis (Hill & Blyth, 1976) now consider the viral genome intact

but inactive in the ganglion. It is repressed and expresses few if

any viral antigens and, therefore, provokes no significant immune

response unless infection is reactivated (Blyth & Hill, 1984; Hill,

Blyth, Harbour, Berrie, & Tullo, 1983). The precise mechanism of

derepression allowing viral expression remains a mystery.

There is some indication that immune system status influences an

individual's vulnerability to recurrent herpes. This is particularly

true for cell-mediated; immune responses as opposed to humoral

immunity. Once infected with herpes simplex virus, individuals

develop neutralizing antibodies humorall immunity) within 1 to 4 weeks

of primary infection (Nahmias & Roizman, 1973c). Unfortunately, the

antibodies are not effective in preventing reinfection or recurrent

infection in either healthy or immunocompromised individuals (Nahmias

& Roizman, 1973a; Rand, 1982; Rand, Rasmussen, Pollard, Arvin, &

Merigan, 1977). The herpes viruses, being capable of cell to cell

spread, are well adapted to evade antibodies. It has been

demonstrated that the titers of neutralizing antibody only

occasionally fluctuate before, during, or after recurrences (Nahmias &

Roizman, 1973b).

Fluctuations in cell-mediated immune responsivity, on the other

hand, appear to be related to recurrent herpes infections. Evidence

for this relationship comes from a variety of sources. A defect in

cellular immunity is proposed to account for the severity of herpes

infections in the newborn and the immunologically impaired host

(Nahmias & Roizman, 1973c). A quicker, more intense in vitro

lymphocyte transformation response was related to faster healing in

primary infection in one investigation. Thirteen of the patients were

followed for a period which included one or more recurrence. The

herpes simplex virus stimulation index was declining when recurrences

took place. Because the stimulation index was declining even among

patients who did not experience recurrence, authors suggest this

decline may be a necessary but insufficient correlate of recurrence

(Corey, Reeves, & Holmes, 1978). Similar results were obtained in an

investigation comparing the lymphocyte responsivity of patients with

frequently recurring oral herpes to a control group of seropositive

blood bank donors not reporting frequent herpes recurrences (Kirchner,

Schwenteck, Northoff, & Schopf, 1978). Responsivity of the two groups

was similar during infection, but during the disease-free interval

responsivity of the patient group was significantly lower. The

severity of infection in immunocompromised patients has been related

specifically to decreased responsivity of lymphocytes as well (Meyers,

Flournoy, & Thomas, 1980; Rand et al., 1977).

Another parameter of cellular immunity, the specific cytotoxicity

of peripheral blood mononuclear leukocytes (PBML) to cells infected

with herpes was found to fall to low levels during the week prior to

recurrent HSV in a group of 14 healthy seropositive adults, showing

enhanced reactivity during acute recurrent infection (Rola-Plezczynski

& Lieu, 1984). T-cells and the percentage of natural cytotoxic

lymphocytes were also found to decrease prior to recurrent HSV

episodes and to increase during recurrences. Fluctuations in

interferon, another component of cellular immunity, have likewise been

implicated in recurrences (Corey & Holmes, 1983; Rand et al., 1977)

and administration of interferon has significantly reduced the risk of

reactivation of patients undergoing surgery on the trigeminal ganglion

(Pazin et al., 1979).

To summarize, it appears that humoral immunity does not play a

major part in the manifestation or recovery from recurrent HSV

infection. This does not rule out the possibility that antibody may

contribute in combination with complement or as a mediator of

lymphocyte cytotoxicity. On the other hand, a relationship to HSV

infection status has been demonstrated for a number of cell-mediated

immune parameters including quantity and responsivity of T-

lymphocytes, interferon release, and natural lymphocyte cytotoxicity

specific for HSV.

A deficiency in immunologic competence, particularly impairment

of cell-mediated responsivity, could be a link between psychosocial

stress and recurrent herpes. Alterations as a result of stress have

been demonstrated in parameters of cellular immunity such as decreases

in lymphocyte responsiveness and natural killer cell activity

(Bartrop, Luckhurst, Lazarus, Kiloh, & Penny, 1977; Locke, 1982;

Kiecolt-Glaser et al., 1984). Changes in cell-mediated immune

functions are in turn related to herpes recurrences. An individual's

response to stress, therefore, could cause immunosuppression which

would allow viral replication. Physiological evidence for such a

connection will be reviewed as well as evidence for a more direct path

of influence which does not necessarily involve immune status.

There is consistent evidence that immunosuppression has an effect

on recurrences of herpes simplex virus. Seropositive transplant

patients whose immune responses are pharmacologically suppressed to

prevent organ rejection experience herpes lesions that are

inordinately severe, extensive, and long lasting. This occurs

sometimes, but not always, with increased incidence over a presurgery

recurrence base-rate (Blyth & Hill, 1984). For example, the

recurrence rate 5 to 15 weeks after bone marrow transplant is 60-80%

(Meyers et al., 1980). In the first 3 months following cardiac

transplantation, 83% of seropositive patients experienced herpes

recurrences of prolonged duration averaging over 2 months in duration

(Rand et al., 1977). Fifty percent of the seropositive renal

transplant patients developed herpes lesions of 2 to 16 weeks duration

within a 4-month follow-up period (Pass et al., 1979).

One of the early investigations to indirectly implicate stress as

a factor in the reactivation of HSV was done on rabbits. Laibson and

Kibrick (1966) experimentally induced recurrent ocular herpes in

latently infected rabbits by intramuscular administration of

epinephrine, a hormone well known to be released in excess during

periods of stress. Investigators demonstrated significant viral

isolation and infection in experimentally inoculated animals relative

to control animals. One could postulate an autonomic nervous system

link between stress and recurrence based on this evidence.

Reactivation of herpes in animals and humans has also been

accomplished by direct neural stimulation. Low levels of electric

current delivered via electrodes implanted over the trigeminal

ganglion of latently infected rabbits consistently induce viral

shedding and herpes recurrences, depending on the interval between

stimuli (Green, Rosborough, & Dunkel, 1981). In humans, trigeminal

surgery appears to exert a direct triggering effect. Between 83-94%

of seropositive individuals who undergo trigeminal surgery experience

recurrence and/or viral shedding within a week. These high incidence

recurrences are unremarkable in duration and severity and are

restricted to the area supplied by the stimulated or sectioned portion

of the trigeminal sensory ganglion (Carton & Kilbourne, 1952; Pazin et

al., 1979). Excessive; stimulation to peripheral tissue which receives

sensory nerves from the infected ganglion induces reactivation in

animals (Hill et al., 1983). Such stimulation could be analogous to

reactivation by sun, wind, or physical injury in naturally occurring

herpes episodes. One could extend the concept of reactivation by

direct neural stimulation to include excessive neural firing as might

result from psychological as well as physical stimulation. Thus, just

as hypnosis or expectancy (placebo effects) may act directly to alter

recurrence rates via a postulated central nervous system mechanism,

anxiety and stress may also act directly to increase recurrence rates

by causing excessive neural firing or by lowering the threshold for

neural firing.

It should be noted here that there appears to be a subtle

difference in how direct neural stimulation and immunosuppression

influence herpes recurrences. Direct neural stimulation results in

more immediate and higher rates of recurrence. These directly induced

recurrences are of average extent and duration. The recurrence rate

in immunosuppressed individuals is not necessarily higher, but the

recurrences are more extensive and long lasting. Of course, any

surgical treatment, whether immunosuppression is involved or not,

involves both physical and psychological trauma, providing a

confounded picture. Presumably, patients undergoing trigeminal or

transplant surgery experience grossly similar psychological

responses. A comparison of herpes simplex virus activation in these

two groups of patients suggests that direct neural stimulation is

associated with reliable reactivation of the virus, while

immunosuppression is more frequently associated with prolonged

severity and duration of recurrent herpes episodes.

One could conjecture more than one route by which stress could

influence the reactivity of HSV. Stress could directly stimulate the

latent virus via the central nervous system by causing increased

neural firing or a lower threshold for neural firing. Stress could

also be indirectly involved, by causing a deficiency in immune

surveillance. It is also possible that both are necessary to provoke

the clinical manifestation of herpes symptomatology, a trigger to

reactivate the virus and an immunocompromised environment in which it

may flourish. At the present time, the exact mechanisms that control

the latency and reactivation of herpes simplex virus in humans are

unknown. Immune response could play a vital role, but the

physiological state of the neuron could also be a key factor (Blyth &

Hill, 1984).

Because of its compatibility with existing evidence, the

supposition that both neural stimulation and immunosuppression are

necessary for herpes reactivation is intuitively appealing. However,

examinations of the validity of this and the other hypotheses present

serious difficulties. While it is known that some triggers are more

effective than others (e.g., trigeminal surgery more frequently

results in reactivation of HSV than sun exposure), the exact nature of

neural stimulation necessary to provoke a recurrence is unknown and,

therefore, difficult to measure or manipulate. Likewise, differences

in immunocompetency are difficult to define and even more difficult to

manipulate, particularly in humans, due to the ethical issues

involved. The preferred course at present is to continue research at

all levels of organization from the precise cellular/biochemical

measurements obtainable in the lab to the more global psychosocial

measurements obtainable through systematic naturalistic observations

of individuals over extended time periods. An integrated effort by

professionals with expertise in immunology, endocrinology, neurology,

pathology, and the behavioral sciences will be necessary to discern

meaningful patterns from divergent multiple measures. Ideally,

converging evidence from all of these disciplines will bring us closer

to the answers.

Present Investigation

The present study was undertaken to extend our knowledge of the

relationship between stress and health status. It was designed to

address methodological shortcomings of existing investigations.

Genital herpes was chosen as a disease model because of its recurrent

nature, the theoretical basis for a physiological mechanism linking

stress and recurrences, the compelling need for empirical information

regarding stress and recurrences, and the availability of an objective

laboratory diagnostic measure for disease confirmation. Prospective

in nature, the study design called for an evaluation of major and

minor life events upon enrollment and on a monthly basis throughout

the 6-month period of investigation. Documentation of herpes lesions

occurred within 24-36 hours of each recurrent episode. Thus, both

prospective and concurrent information regarding stress and

recurrences was collected. Additionally, three variables which have

shown promise in prior research as moderators of life stress were

assessed. The use of multiple regression techniques enabled the

clarification of the relative importance of social support, locus of

control, and arousal seeking tendency in influencing recurrence rates

in the face of stress. Thus the possibility of identifying subgroups

of individuals who were thought to be particularly vulnerable was

explored. In the present state of the art of psychobiological

research in which the single-cause, single-effect model is

acknowledged as simplistic, we must depend upon converging evidence

from the different levels of many disciplines to obtain a reasonable

picture of the many internal and external events and dynamic

interactions that ultimately result in a disease outcome (Schwartz,


This particular investigation involves the second and fifth

levels as illustrated in Figure 1: the perception of events as

stressful or not and the occurrence of viral reactivation as

manifested in recurrent herpes episodes. No attempt was made to

objectively monitor environmental events in the lives of subjects by

trained neutral observers. Nor have attempts been made to measure

neural, endocrine, or immune parameters. The independent variables in

this study consist of self-report measures of major and minor

stressors and social and personality variables which hypothetically

moderate the perception and influence of the stressor variables. The

dependent variable measured was a manifestation of genital herpes

symptomatology, the isolation of virus from a lesion. While this

particular investigation has not attempted to elucidate the complex

web of physiological events between stressor and symptomatology, it

has endeavored to shed light on the psychological conditions under

which a relationship between stress and illness exists.

Several specific hypotheses were investigated in the present

investigation. As stated earlier, the present investigation was

designed to explore the relationship between stressful events and

recurrences using both prospectively and concurrently collected

information. The first hypothesis, using a prospective approach,

concerns the strength of the relationship between an individual's

experience of major life events and subsequent experience of herpes

recurrences. Specifically, the first hypothesis states:

1. The occurrence of undesirable life events prior to entry into

the study will be positively related to herpes recurrences during the

6-month period of investigation.

The next two hypotheses concern the concurrent experience of both

major and minor life events and herpes recurrences. Specifically,

2. The occurrence of undesirable life events during the 6-month

period of investigation will be positively related to herpes

recurrences over the same period of time.

3. The occurrence of monthly hassles will be positively related

to herpes recurrences during the period of investigation.

Additionally, it was hypothesized that three moderator variables,

social support, locus of control, and arousal seeking tendency, would

interact with each other in such a way that an individual's

vulnerability to herpes recurrences under high stress conditions would

be predictable on the basis of scores on the moderator variables.

That is,

4. The following variables which have previously been shown to

moderate the effects of life stress will be related to HSV recurrences

in the following ways: (a) Low levels of social support, a perception

of events as externally controlled, and a tendency to avoid arousing


situations will be associated with more frequent HSV recurrences. (b)

The antithesis of the above will serve to protect individuals from

frequent HSV recurrences. That is, individuals with a high level of

social support, with an internal perception of control, who tend to be

arousal seeking will have fewer HSV recurrences. (c) The probable

frequency of recurrence for a given individual will be predictable on

the basis of scores on the moderator variables.


Research Design

The longitudinal design of this study allowed for the

investigation of prospective and concurrent information regarding the

occurrence of major and minor stressful events and episodes of

recurrent genital herpes. Data concerning past stressful events

collected from initial interviews and questionnaires provided

information to relate to ensuing recurrences of genital herpes

prospectively documented over the 6-month period of the study.

Concurrent information on major and minor stressful events was

collected monthly during a subject's 6 months of participation to

relate to documented recurrences for the same period of time. This

design provided the opportunity to avoid distortion due to

retrospective bias in an investigation of the relationship of major

and minor stressful events and herpes recurrences.

The assessment of; subjects' sense of perceived control, tendency

toward arousal seeking, and social support provided an opportunity to

evaluate the hypothesized stress-moderating effects of these

variables. In addition to studying the effects of the presumed

moderator variables separately, multivariate statistical techniques

made it possible to evaluate the manner in which these variables

related in an additive or interactive manner to possibly enhance or

diminish the resistance of individuals to recurrent episodes of

genital herpes in the face of life stress.

Sample Selection

Subjects were recruited over a 16-month period (March, 1983-July,

1984) through advertisements and referrals. Advertisements were

placed in both the university and local newspaper and on bulletin

boards on campus and in the J. Hillis Miller Health Center. Sources

of referral included the University of Florida Student Health Center,

the clinics of Shands Teaching Hospital, and participating

volunteers. Subjects were telephone-screened regarding their

suitability, availability, and willingness to make a 6-month

commitment to the sudy. The research was described as an effort to

understand more about the psychological factors involved in the

recurrence of genital herpes in order to improve the counseling

techniques of health care providers.

After obtaining informed consent, 10 cc of blood was drawn from

each participant. Those with a positive antibody titer to herpes were

enrolled in the study. Those subjects with a negative titer were not

enrolled unless they had a positive culture from a genital site to

herpes simplex virus.

Each of the 122 subjects who completed the full 6 months of

filling out questionnaires and having genital herpes recurrences

cultured received 100 dollars at the conclusion of the study for

his/her participation.


Stress Measures

The occurrence of major and minor events in the lives of subjects

was assessed by self-report questionnaires. Subjects completed the

questionnaires upon enrollment and monthly during the study.

Life Experiences Survey (LES). This questionnaire was used to

evaluate the experience of subjects with major life events. The LES

is comprised of 57 life events, 3 of which are fill-ins and the last

10 of which are specifically relevant to a college population

(Sarason, Johnson, & Siegel, 1978). Subjects indicate the frequency

of occurrence of each of the events over a specified time period.

They also rate the impact of each event on a scale from -3 (extremely

negative) to +3 (extremely positive), a zero rating designating a

neutral impact. Means and standard deviations are available for the

47- and 57-item versions based on 345 college students. Since many

subjects were college students, the longer inventory was used.

However, group statistical analyses were based upon the 47 items which

all subjects completed.

Desirable and undesirable events are assessed separately on the

LES, an advantageous feature for this investigation. Negative events

have been found to be more predictive of illness in general (Hotaling

et al., 1978; Sheehan et al., 1979) and genital herpes in particular

(Taylor, 1978; Watson, 1983). This is consistent with the literature

on adjustment (Mueller et al., 1977; Vinokur & Selzer, 1975).

Although impact (subject weighted) scores were available, the measure

of interest in this investigation was the frequency of undesirable

events. Several studies providing data on both simple frequency

scores and weighted scores have shown them to be equally predictive

where group data are concerned (Monroe, 1982b; Skinner & Lei, 1980;

Zimmerman, 1983). Developers of the scale have provided information

regarding significant positive correlations with several different

measures of anxiety, noncomformity, discomfort, depression, locus of

control, and a nonsignificant correlation with the Marlowe Crowne

Social Desirability Scale to demonstrate validity. Additionally, the

scale has been found to discriminate between normal controls and

counseling clients (Sarason et al., 1978). This scale has been used

in two retrospective investigations of subjects with recurrent

herpes. In one, the scale discriminated between subjects experiencing

high and low recurrences (Taylor, 1978). In the other, significant

positive correlations were found between the frequency and impact of

undesirable life events of the past year and the frequency and

duration of herpes recurrences for the past 6 months (Watson, 1983).

Hassles Scale (HS). The HS focuses on the minor events of the

past month, the irritating, frustrating, and distressing demands

characterizing daily transactions with the environment (Kanner et al.,

1981). The 117 items which include such things as care for pet,

having to wait, inconsiderate smokers, too many meetings, problems

with children, and preparing meals are checked if perceived as a

hassle and rated from 1 to 3 according to the severity of the

hassle. A frequency score is obtained by counting the number of

hassles checked. An impact score is obtained by summing the hassle

ratings. Norms are based on 100 middle class subjects ranging in age

from 45 to 64 who completed the scale for 9 consecutive months.

Test-retest reliability over that period is reported as .79 for

frequency scores, .48 for impact scores. Significant positive

correlations have been found with a negative affect scale, a life

events measure, a symptom check list (Kanner et al., 1981), and a

general health measure (DeLongis et al., 1982).

Moderator Variables

Upon enrollment in and termination from the research 6 months

later, subjects completed inventories to assess enduring

characteristics proposed to moderate the effects of stress on

illness. These included measures of social support, of perception of

internal or external controllability of events, and of arousal seeking


Locus of Control (LC). The LC is a 29-item forced-choice

inventory including six filler items which assesses the degree to

which individuals view environmental events as being under their

personal control (Rotter, 1966). Subjects scoring low on the LC

internalss) tend to perceive events as being controllable by their own

actions, while those scoring high (externals) tend to view events as

being influenced by factors other than themselves. It is a measure of

generalized expectancy, or belief, rather than preference. The

measure is widely used and extensive data are available regarding

internal consistency, test-retest reliability, and discriminant

validity. Correlations with social desirability and intelligence are

low. Means and standard deviations on the scale are provided for

numerous groups, including large groups of university students

(Rotter, 1966). A number of studies investigating psychological

adjustment and illness as related to life stress provide support for

the role of perceived control as a moderator of life stress. That is,

an external LC score is associated with instability and illness under

conditions of increased life stress (Johnson & Sarason, 1978; Manuck

et al., 1975; Toves et al., 1981).

Arousal Seeking (AS). Arousal seeking is defined as the tendency

of individuals to engage in or avoid situations which increase

stimulation to a personal optimal level. The AS scale was developed

specifically to assess individual differences in this trait (Mehrabian

& Russell, 1974). Subjects are instructed to indicate their degree of

agreement or disagreement (on a 9-point scale) to each of 40 items.

Means and standard deviations are reported for two large samples of

college students. Adequate internal consistency and reliability have

been demonstrated. The AS scale correlated positively with a measure

of extroversion and negatively with measures of anxiety.

(Correlations with anxiety are significant, but weak; it is not simply

an anxiety measure.) Low correlations have been found with social

desirability and succorance (Mehrabian & Russell, 1974). In an

investigation of the relationship between life stress and measures of

anxiety, depression, and hostility, a significant relationship was

found only for subjects low on the arousal seeking dimension.

Negative life change in subjects who were low in arousal seeking

resulted in greater anxiety and hostility. This suggests that the

dimension plays a moderating role on life stress (Johnson, Sarason, &

Siegel, 1979).

Social Support Questionnaire (SSQ). Social support may be

defined as the degree to which individuals have access to social

resources or relationships upon which they can rely, particularly in

times of need. A person's satisfaction with the quality of such

resources is an important dimension in addition to the number of

persons in a support system. The SSQ measures the size and

satisfaction rating of an individual's social network (Sarason,

Levine, Basham, & Sarason, 1981). It consists of 27 two-part items,

one part to indicate the number of support people available under

particular circumstances, the other to indicate the respondent's

satisfaction rating for available support. Norms exist for a sample

of 602 university students. Internal consistency and test-retest

reliability are high. Correlations with scales of maladjustment show

an inverse relationship, suggesting that people who have fewer and

less satisfying social supports are more likely to be anxious, more

labile emotionally, and more pessimistic than others. Subjects low on

the SSQ tend to have an external locus of control and be relatively

low in self-esteem.

There have been several studies which provide evidence for social

support as a moderator of life stress (de Araujo et al., 1973;

Hotaling et al., 1978; Nuckolls et al., 1972; Schaefer et al.,

1981). Those low in social support evidenced more maladjustment and

ill health under circumstances of high life stress. Therefore, high

levels of social support may play a stress-buffering role. In the

present investigation the SSQ was modified to include a pair of

questions regarding to whom a subject had disclosed he/she had herpes

and level of satisfaction with support received.

Physiological Measures

Screening procedure: HSV antibody. Since enrollment in the

study was limited to those who were herpes simplex virus (HSV)

seropositive, subjects were screened in order to ensure that they had

had prior exposure to HSV. This work was performed in a virology

laboratory and was done by a microtiter complement fixation (CF)

method previously described (Rand, Kramer, & Johnson, 1982). No

attempt was made to discriminate between antibodies to herpes simplex

virus-1 (HSV-1) and herpes simplex virus-2 (HSV-2) because (a) sera

that is seropositive by CF is seropositive for both HSV-1 and HSV-2 by

an accepted discrimination technique (ELISA); and (b) although HSV-2

usually causes genital herpes, HSV-1 is also a frequent cause.

Dependent variable: Virus isolation and transport. All subjects

were instructed to call the Clinical Research Center (CRC) as soon as

they were aware of a recurrent genital herpes lesion. An appointment

was made for a culture to be obtained within 24-36 hours by trained

personnel in the CRC. The CRC personnel were available on a 24-hour

daily basis.

Cultures for HSV were obtained by swabbing the base of unroofed

vesicular or ulcerated lesions with a cotton swab. Swabs were

immersed in transport media [minimal essential medium (MEM)

supplemented with 2 mm glutamine, 5000 u/l penicillin, 50 ug/ml

streptomycin, 100 ug/ml gentamicin, 2 ug/ml amphotericin B and 1%

gelatin] and held at 4C until transported to the laboratory; previous

work in the laboratory (unpublished) and elsewhere has shown this

method does not result in any significant decrease in recovery of

HSV-2 (Bettoli, Brewer, Oxtoby, Azidi, & Guinan, 1982; Yaeger, Morris,

& Prober, 1977). Cultures were then either frozen at -70C or

inoculated immediately into primary cultures of human foreskin

fibroblasts (Rand et al., 1982). All isolates were identified by

typical cytopathologic effect.

Because subjects would not always be able to come to the CRC in

time, they were given Virocult swabs (Medical Wire and Equipment Co.,

Cleveland, OH) upon enrollment and instructed in the procedure for

obtaining self-cultures. Pilot work indicated that patients self-

obtained cultures were positive as frequently as staff-obtained

cultures. Herpes simplex virus has been shown to be stable for at

least 48 hours when refrigerated under these conditions (Medical Wire

and Equipment Co., product information).


Initial Contact

Subjects were seen singly at the CRC by the investigator or a

trained research assistant. Confidentiality was ensured, study

procedures were explained in detail, questions answered, and informed

consent obtained. Blood was drawn for screening and subjects were

instructed to return to the CRC to have herpes lesions cultured every

time they had a recurrence during the 6-month period of the study.

They were provided with a Virocult swab and trained to obtain a self-

culture in the event that it was impossible to get to the CRC within

24-36 hours. A culture could be safely kept in the refrigerator for 2

to 3 days until transport to the clinic was possible. A sample copy

of the LES was provided with a calendar on which to record major life

events. This eased the filling out of questionnaires each month and

helped to avoid overlap or duplication of major life events from month

to month. Subjects were then given a packet of initial questionnaires

and instructed in the procedure for completion. The packet contained

the LES (for the past 6-month and 1-year periods), the HS, LC, AS, and

SSQ questionnaires. An appointment was made for the second contact in

1 to 2 weeks time when the blood screening results would be available.

Second Contact

Subjects returned with the completed questionnaires of the

initial packet. Their questions were answered, and they were given a

second opportunity to sign the consent form to signify continuing

commitment to the investigation at a time when blood screening results

were known and they were thoroughly familiar with all procedures and

questionnaires. Subjects were not considered fully enrolled until

their second signature was obtained. Completed questionnaires were

reviewed for proper completion and a structured interview was

conducted to obtain demographic information and the subjects' history

of herpes. (Please see Appendix A for a copy of the structured


Continuing Contact

Continuing contact was by mail, phone, and clinic visits. In the

last week of each month, subjects received by mail a packet containing

preceded questionnaires (LES and HS) to be completed on the final day

of the month with respect to events of the month; self-addressed,

stamped envelopes for their return; a newsletter; and educational

materials regarding genital herpes to maintain motivation. Subjects

were requested to date the major events reported on the LES to avoid

overlap with preceding and following months. Phone contact was made

to subjects to clarify material on the questionnaires, or to prod for

their return as necessary. Subjects initiated phone contact at any

time to ask questions about the study, or to obtain information about

herpes. Clinic visits occurred at any time during the 6 months of

participation for the purpose of obtaining culture material from

recurrent lesions. Subjects received counseling for disease

management by the CRC staff or the investigators if appropriate during

clinic visits.

Final Contact

A debriefing interview was conducted with each subject as soon as

the questionnaires for the sixth month were completed. Payment was

made to subjects following this final interview.


Sample Characteristics

The original sample consisted of 153 individuals. Five were

excluded because they did not have HSV antibodies in the initial

screening and thus were not subject to recurrent HSV. Attrition

occurred for the following reasons: (a) moved, n=8; (b) too busy,

n=4; (c) lost interest, n=2; (d) received experimental vaccine, n=l;

(e) believed participation worsened condition, n=l; (f) visual

handicap, n=l; and (g) lost to follow-up, n=8. Considering only the

seropositive subjects, this represents a loss of 26, or 18%, of the

eligible subjects.

The final sample consisted of 122 subjects, 83 of whom were

female (68%) and 39 of whom were male (32%). Average age was 27.3

(SD=6.1) with a range of 18 to 55. Educational level was high, with

the average number of years of education being 14.3 (range=9-18).

Sixty-three (52%) of the participants were currently in college or

graduate school. The majority of the participants (94; 78%) were

single, although 83 (68%) reported having a regular sex partner.

Subjects reported having herpes an average of 3 years, 11 months

(range of 1 month to 18 years). For the year prior to the study,

46.7% reported four or more recurrences, 35.8% reported one to three

recurrences, and 17.5% reported none.

Undesirable Life Events

The measures of major and minor life events in this study each

produced two scores of potential relevance to the present

investigation, a simple frequency of negative events score and an

impact score reflecting subjects' individual weighting of these

events. Life events scores of importance included a 6-month recall of

major events occurring just prior to enrollment, and monthly reports

of life events which were summed to produce a 6-month concurrent total

of events experienced during the investigation. The HS was also

completed monthly and summed to provide a 6-month total hassles

index. (See Appendix C for subjects' scores on all variables.)

Consistent with previous investigations (Zimmerman, 1983),

Pearson product-moment correlations of frequency and impact scores

resulted in extremely high correlations. The correlation between the

6-month retrospective LES frequency and impact scores was .95 (n=120;

Z=.0001), between LES concurrent frequency and impact .95 (n=124,

p=.0001), and between concurrent HS frequency and impact .97 (n=124;

p=.0001). Although it might be argued that the weighted measures may

be more sensitive to an individual's perception of the stressfulness

of major and minor events, given the redundancy in information

provided by these indices, they were not included in further group

analyses. Simple frequency counts of negative events were used for

both the LES and HS measures as they were seen as providing the most

objective and straightforward indices of major and minor life changes,


Table 1 provides means and standard deviations obtained on stress

measures for the 6 months prior to the study (past) and for the 6

months during the study (concurrent).

Table 1
Means and Standard Deviations of Past and Concurrent Stress Measures

Time Period

Past 6 months Concurrent 6 months
Stress Measure n = 120 n = 117

Frequency of Major Negative 4.43 11.90
Life Events (3.11) (8.94)

Impact of Major Negative 9.57 21.02
Life Events (7.16) (17.24)

Hassles Frequency 121.48

Note. Means appear above standard deviations which are in
a Concurrent measures were computed by summing monthly scores over the
6-month period of the investigation.

Despite the fact .that subjects dated major events for monthly

administrations of the LES to avoid overlap from month to month, the

average frequency of major negative life events is higher for the

concurrent 6-month period than for the 6-month period prior to entry

into the study, as recalled at the time of enrollment. There may be

several possible reasons for this discrepancy. First, it is possible

that the actual number of events experienced over the two 6-month time

periods differed. Secondly, the discrepancy may be attributed to

memory loss for events that occurred over the more remote 6-month

period prior to entry into the study (Monroe, 1982a; Paykel, 1983).

Memory loss for the concurrent period of investigation would be less

severe as recall was required over a shorter time period of 1 month.

Repeated reporting of major negative events that tend to recur

could also have contributed to the higher concurrent scores. For

example, the eight most frequently reported negative life events in

the concurrent information include events that could occur more than

once in a 6-month period. Arranged in order from the most frequent

they include sexual difficulties, major change in sleeping habits,

changed work situation, major change in financial status, major change

in eating habits, more or less arguments with spouse, breaking up with

boyfriend or girlfriend, and a major change in social activities.

These items were among those most frequently endorsed on the 6-month

recall as well. Such events could occur more often than once every 6

months and be counted a maximum of six times per subject in the

concurrent information (which is based on a total of six monthly LES

scores) but would only have been counted once on the 6-month recall of

major events experienced prior to the study.

It is of interest to compare LES scores of the present sample

with those of other groups. Normative data are provided on the LES

impact ratings of 345 male and female students enrolled in

introductory psychology classes (Sarason et al., 1978). Although

impact scores were not used in analyses in the present study, they

were available and are used here for comparison purposes since

normative data have been provided for LES impact but not LES frequency

scores. The mean 6-month recall negative impact score of herpes

subjects was 8.57 (SD = 7.16) as contrasted to a mean score of 4.66

(SD = 4.36) for males or 5.64 (SD = 6.43) for females for a 12-month

recall administration of the LES. One would have expected the figures

to be reversed since the data on the normative sample are based on a

time period that is twice as long (12 as opposed to 6 months). The

difference is even more striking for concurrent information with an

average frequency of major negative life events of 21.02 (SD =

17.24). The latter, as already discussed, includes information on

frequently occurring and potentially repeated major events. Authors

of the original measure suggest that the LES score of the student

(normative) population may be low relative to subjects from the

general population (Sarason et al., 1978). It appears that the

present, more heterogeneous sample has perceived relatively more

stress than the normative group.

Scores on the HS are more directly comparable to normative data

as both are based on monthly administrations (Kanner et al., 1980).

The 100 normative subjects were older, ranging in age from 45 to 64 as

opposed to the age range in the present sample of 18 to 55. The

average monthly hassles frequency score of the normative group was

20.50 which is very similar to the average for the herpes sample of


Presumed Moderator Variables

Information regarding the herpes sample as a whole on arousal

seeking tendency, locus of control, and social support is provided

here for comparison purposes.

Subjects' scores on arousal seeking tendency were approximately

normally distributed with a mean of 35.02 (n = 120). This is slightly

higher than the average of 32.0 reported for the normative sample of

536 subjects (Mehrabian & Russell, 1974). The standard deviations of

the two samples are similar--28.58 for the herpes subjects and 29.0

for the normative group.

Locus of Control (LC) scores were also approximately normally

distributed with a mean score of 10.07 (SD = 3.87, n = 122). Average

scores in the literature range from 5.94 (SD = 3.36, n = 155) for

Peace Corps volunteers to 9.22 (SD = 3.88, n = 303) for college

students (Rotter, 1966). Although the standard deviations are

similar, the average score of the herpes sample is slightly higher,

indicating that, as a group, the present subjects tend to be more

externally oriented than others.

When the average Social Support Satisfaction (SSS) score is

compared to normative data, it initially appears that the herpes

subjects' SSS scores may be slightly lower. The average SSS score for

herpes subject is 141.87 (SD = 18.92, n = 119); for the normative

group of 602 undergraduates, the mean was 145.26 (SD not provided;

Sarason et al., 1981). However, the distribution of scores from the

herpes subjects ranged from 2 to 213 and was highly skewed with most

of the scores grouped together at the very satisfied end of the scale

(median = 156.5). Subjects who were dissatisfied with their perceived

social support were not well represented in the present sample which

limits the generalizability of findings with regard to social support.

As a group, the herpes subjects appear to be slightly higher on

arousal seeking tendency and social support satisfaction and more

externally oriented than the respective normative samples described in

the literature.

As expected, major and minor events were significantly related.

The Pearson product-moment correlation of concurrent LES and HS totals

was .68 (n=122; p=.0001). The HS scores were retained for analyses of

hypotheses specifically concerning minor events.

Definition of Recurrence

True herpes recurrent lesions do not always produce positive

culture results in the laboratory due to a relatively short duration

of viral shedding. Corey and Holmes (1983) report obtaining a 42%

culture positive rate from recurrent lesions. If the definition of

the dependent measure was restricted to culture-positive lesions, as

many as 58% of true episodes of herpes would be missed. If the

definition included all reported lesions, the value of laboratory

documentation in this investigation would be lost.

The decision was made to define a recurrence for the purposes of

the present study as any reported recurrence for an individual subject

for whom at least one culture-positive result was obtained during the

course of the study. This compromise was adopted to maximize clinical

sensitivity and scientific credibility. The different recurrence

definitions demonstrated a high degree of association with each

other. Spearman correlation coefficients of the number of recorded

recurrences and the number of positive cultures with the number of

recurrences given one positive were .66 and .71, both statistically

significant at the .0001 level. Furthermore, correlations of each of

the recurrence measures with the stress measures were very similar to

each other. (Refer to Appendix D for correlation coefficients of the

stress and recurrence measures.) Therefore, subsequent analyses

include all reported recurrences for subjects for whom at least one

positive was documented. Hereafter, unless otherwise stated, the term

"recurrences" will refer to those meeting the requirements of this

definition. During the study, 95 of the 122 total subjects reported

258 herpes lesions. Of these, 195 were cultured (76%), and 104 of

those cultured were positive (53%). This positive rate compares

favorably to that reported earlier (Corey & Holmes, 1983). There were

31 subjects whose lesions remained unconfirmed by laboratory


Using recurrences as defined above, the hypotheses of the present

investigation were tested. For ease of presentation, results

concerning the existence and extent of a direct relationship between

stress and recurrences will be introduced first. Linear correlation

coefficients were used to examine this association. Following that,

analyses pertaining to the conditions under which such a relationship

exists are presented. Included here are the results of multivariate

analyses designed to investigate the combined influence of the stress

measures and presumed moderator variables. Finally, findings of

within-subject analyses accomplished for the purpose of exploring the

nature of the obtained relationship will be offered.

Stress/Recurrence Relationship

The first three hypotheses were tested using Scearman

correlations of stress scores and the number of herpes recurrences

experienced during the period of investigation. As defined above, all

recurrences were counted for each subject who had at least one

documented by viral culture during the 6-month study. This

information was correlated with the frequency of undesirable events

reported in the 6 months prior to the time of investigation to test

the first hypothesis. This is the purest prospective investigation of

stress and herpes recurrences since all stress information was

collected prior to the monitoring of recurrences. The result, based

on 120 individuals, was nonsignificant (r=.040; p=.656).

Because this result was surprising in light of two previous

retrospective investigations (Taylor, 1978; Watson, 1983), the

analysis was repeated for the subset of subjects who experienced

recurrences during the study, eliminating all subjects with zero

recurrences. This produced a subsample of recurring subjects, more

similar to the previous samples studied. The result was also

nonsignificant (r=-.040; n=53). These results indicate a total lack

of association between recent undesirable life events and subsequent

herpes recurrences, providing no support for Hypothesis 1.

In contrast, a significant positive relationship was obtained

between monthly totals of the frequency of reported undesirable life

events, summed over the 6-month period of the investigation, and

recurrences experienced over the same period of time. The correlation

obtained was .193 (p=.037; n=117). This significant positive

correlation indicates that increasing levels of reported negative life

events were associated with increasing recurrence rates when

concurrently monitored. Thus, some support was obtained for the

second hypothesis that stress and recurrences are concurrently


With regard to Hypothesis 3, concerning minor life events

experienced over the same period of time, no significant relationship

was demonstrated. The correlation of the total HS scores reported

monthly and recurrences was only .019 (n=117). It appears that minor

life events are not significantly associated with herpes recurrences,

at least when monthly hassles scores are accumulated over a 6-month

period and related to recurrences over the same time period.

Influence of Moderator Variables

The fourth hypothesis concerns finding the best possible

combination of variables for predicting recurrences and evaluating the

influence of presumed moderator variables. Because it was the

particular pattern or combination of these variables that was of

interest, multivariate methods of analysis were chosen. It would have

been possible to use an analysis of variance procedure with the data

collected in this investigation in order to examine the differences in

recurrence rates associated with high or low scores on the moderator

variables and high or low frequencies of stressful events. However,

multivariate regression analyses make better use of the continuous

nature of the data and can provide a more precise description of the

relationships among the variables (Linn, 1982). In linear models,

multivariate techniques allow for the assessment of the effect of each

variable while statistically holding other variables of interest

constant. They also allow for nonlinear models which can include

interaction terms and therefore make it possible to assess the

moderating effects of one variable on another (SAS User's Guide,


Because the findings of the concurrent data provided the only

evidence for a stress/recurrence relationship, analyses employing

these concurrent data are the focus of the remainder of this report.

One analysis included all subjects, the second included only subjects

who experienced at least one recurrence during the study. Details of

the methods and results of these analyses and the follow-up procedures

used to verify results are described below.

When all individuals were considered in the analysis, a highly

skewed distribution of recurrences resulted due to the large number of

individuals with zero recurrences. Therefore, the recurrence variable

was categorized into three groups according to the number of

recurrences: (a) nonrecurrers who experienced no documented

recurrences during the study, (b) recurrers who had one or two

recurrences, and (c) recurrers who had three or more recurrences.

Logistic regression analysis was chosen as the most appropriate

analysis for use with this ordered categorical data (Harrell, 1983).

A logistic regression can be used to estimate nonlinear relationships

and takes into account the ordered nature of the recurrence categories

(Afifi & Clark, 1984; Harrell, 1983).

A forward stepwise logistic regression procedure was performed to

identify which, if any, variables and/or interactions were related to

recurrence groups. Single independent variables available to enter

the model consisted of the total number of negative major life events

reported monthly and summed over the 6 months of the study (LESNT),

the total number of hassles reported monthly and summed over the 6

months of the study (HST), and the following three variables which

were measured at the beginning of the study: locus of control (LC),

satisfaction with social support (SSS), and arousal seeking tendency

(AS). Two-factor interactions, which included at least one stress

measure (HST x LC, HST x SSS, HST x AS, LESNT x LC, LESNT x SSS,

LESNT x AS, LESNT x HST), and three-factor interactions, which

included one stress measure (HST x LC x SSS, HST x LC x AS,


were all available to enter the model if important in predicting

recurrences. The level of significance required for a variable to

enter and remain in the model was .05. The only term selected as

significant in this manner was the LESNT x SSS interaction (T<.05).1

1 The logistic regression procedure was repeated with two
variations: (a) Subjects were eliminated whose cultures were never
documented by positive culture (n=31). In most cases, these cultures
were "documented" in the sense that a CRC nurse had judged the lesion
worthy of culture. Therefore, some were true recurrences for which an
attempt to isolate the virus failed. (b) Sex was included as a
variable in order to statistically control for a significant sex
difference on responses to the SSQ (t(117)=3.73; p=.0007). Removing
the "probable recurrers" from the group of nonrecurrers and
controlling for sex differences did not alter the logistic regression

A follow-up logistic regression analysis was performed with

LESNT, SSS, and their interaction as the independent variables. This

resulted in a prediction equation relating stress to recurrences. The

nature of the LESNT x SSS interaction, however, suggests that the

slope of recurrences on LESNT increases with increasing social support

satisfaction, meaning that at higher levels of SSS, there is a

positive relationship between stress and recurrences. Only those

extremely low on SSS demonstrate a negative relationship between

stress and recurrences. A change in slope occurred at approximately

100. For SSS scores greater than 100, the relationship between LESNT

and recurrences was positive, but negative for values less than 100.

In the present sample, SSS scores range from 69 to 168 with a mean of

141.2 and standard deviation of 19.8. The distribution is highly

skewed with more scores at the high end of the scale. Since less than

5% of the subjects' SSS scores fell below the critical score of 100

virtually no support was obtained for a beneficial moderating effect

of social support satisfaction on life stress. For the vast majority

of subjects, high life stress was associated with more frequent


Since previous investigations have focused on recurring subjects,

a regression analysis was performed using the subsample of the 52

recurrers for whom information on all independent variables was also

complete. The number of recurrences for an individual served as the

dependent variable. A square root transformation was applied to

stabilize the variance. A stepwise regression technique was utilized

in which variables which contribute significantly to the model enter,

remaining only if they contribute importantly in relation to all other

variables in the model (SAS User's Guide, 1985). The same independent

variables were available to enter the model as in the initial logistic

regression. The forward selection technique with a p level of .10

required for variables to enter and remain in the equation kept only

one term: the LESNT x LC interaction. The amount of variance

accounted for by this interaction was only 3%.

A follow-up regression analysis which included only LC, LESNT,

and their interaction was performed. This resulted in a prediction

equation which indicated that for individuals with a more external

locus of control (high values of LC), greater life stress resulted in

more recurrences. The opposite relationship held true for more

internally oriented individuals. The point at which the slope on

LESNT changed was approximately 3.43. So, for values of LC greater

than 3.43, LESNT was positively related to recurrences; for values of

LC less than 3.43 LESNT was negatively related to recurrences. In

this particular sample, LC scores ranged from 2 to 19 with a mean

score of 10.07 (SD = 3.87) and the distribution of LC scores was near

normal. As previously discussed, the herpes subjects tend to be more

externally oriented than normative groups (Rotter, 1966).

Follow-up regression analyses were performed separately for three

groups of individuals based upon their LC scores: (a) individuals

with LC scores between 2 and 8 (LC1), (b) those with LC scores between

9 and 13 (LC2), and (c) those with LC scores between 14 and 19

(LC3). Mean LESNT scores for these groups were 12.18, 13.48, and

13.75, respectively. The relationship between major life stress and

recurrences was different for these groups, as is illustrated in

Figure 2.

As can be seen by examination of this figure, the regression

lines for predicting the number of recurrences from an individual's

LESNT score are quite different in slope for individuals with the

lowest LC scores. Low LC scores are associated with belief in

personal (internal) control of events while high scores are associated

with belief in external control. Individuals with the lowest scores

(LC1) demonstrate a slightly negative nonsignificant slope, while

higher LC scores (LC2 and LC3) demonstrate a positive slope. In the

LC2 and LC3 groups, the estimated relationship was approximately the

same for both groups but only the slope for the LC2 group was

significant (Z=.02). This illustrates the finding that life stress

has an important association with the probability of recurrence.

Since both regression analyses confirmed a stress/recurrence

relationship, the question of causality was raised. Does life stress

result in an increase in recurrences or vice versa? While it is not

possible to actually address the issue of causality as variables were

not experimentally manipulated in the study, a closer look at the

temporal relationship and direction of the relationship between stress

and recurrences appeared warranted. The following section is devoted

to follow-up analyses designed to suggest the direction of the

relationship between LES and recurrences.






~ 2.2

I 2.0

0 1.9


u 1.7








0 5 10 15 20 25 30 35 40 45

Number of Major Negative Life Events
Monthly Totals Summed Over Six Month Study

Predicted recurrences as a function of major negative life
events for three ranges of Locus of Control (LC) scores.

O LC1 scores between 2 and 8
O LC2 scores between 9 and 13
* LC3 scores between 14 and 19

Figure 2.

The Nature of the Life Stress/Recurrence Relationship

The method of data collection in this investigation allowed for a

more refined, within-subject analysis of the stress/recurrence

relationship. Information regarding life events was obtained at

regular monthly intervals, with recurrences being recorded by the date

of onset. This information was utilized in an attempt to clarify the

directionality of the stress/recurrence relationship.

Specifically, the hypothesis was that stressful events would

precede recurrences and this would be reflected by an increased number

of stressful events in the month prior to that of a recurrence. If on

the other hand, recurrences resulted in increased stress, this would

be reflected in an increased number of stressful events in the month

following the recurrence. To investigate this hypothesis, stress

difference scores were computed and summed for subjects who reported

recurrences. The difference scores were derived by subtracting the

stress score for the month following the month of a recurrence from

the stress score reported the month prior to a recurrence. If the

mean of these individual difference scores was significantly different

from zero in a positive direction, it would indicate that stress is

more likely to precede; than follow a recurrence and strengthen the

nature of the causal inference to be made. A nonsignificant or

negative result would weaken the case that stress causes recurrences,

perhaps suggesting that recurrent episodes of herpes result in

increased stress or that some third factor is related to stress and

recurrences causing similar changes to both.

Since it is possible that stress scores during months of a

recurrence could be inflated due to factors related to the recurrence,

these months were avoided in the analysis. That is, scores used were

from the first clear month (month without a recurrence) prior to a

recurrence and the first clear month following. Thus, if a person had

continuous recurrences in months 2, 3, and 4, the difference score was

computed between months 1 and 5. Difference scores on the LES could

not be obtained for recurrences falling in the first month since a

prior monthly stress score was not available. Similarly, recurrences

falling in the last month lacked stress information from a clear month

following the recurrence. In the case of more than one difference

score per subject, the scores were averaged for that subject. In the

event of recurrences separated by only 1 month, the month between

could be considered both a month prior and a month following.

Therefore, the month prior to the first recurrence and the month

following the last recurrence were used to compute difference

scores. Multiple recurrences within 1 month were treated as one

recurrence for the month.

The application of this procedure to the frequency of major life

events reported on the; LES on a monthly basis resulted in an average

difference score across subjects of 0.073. A one-sample t-test for

this subsample of 55 subjects resulted in a t-value of 0.28 which is

not significant (p=.78). Because this was a within-subjects analysis

and weighted scores are more sensitive than simple frequency scores to

an individuals' perception of the stressfulness of events, this

analysis was repeated using impact scores. A similar nonsignificant

difference score was obtained. The average impact difference between

the month preceding and the month following a recurrence was 0.255

(t=0.48; U=.63). These results together indicate that the frequency

and impact rating of stressful events preceding a recurrence are not

significantly different from the frequency and impact ratings of

stressful events following a recurrence. This does not resolve the

issue of directionality.

Because some of the negative difference scores were large and

could have masked the fact that stress more often precedes

recurrences, another test was performed. A simple count of negative

and positive difference scores was obtained and tested for

significance with the McNemar test for changes (Siegel, 1956).

Twenty-one subjects reported more stress preceding a recurrence, 17

reported more stress following. The rest of the subjects had equal

prerecurrence and postrecurrence scores. McNemar's test with a

correction for continuity resulted in a Chi-square of 0.237 which is

not significant at the .05 level. Thus, neither of the within-subject

analyses provided support for the notion that negative major life

stress is a significant factor in precipitating herpes recurrences.


The major finding of this investigation was a significant

positive correlation between concurrent major life events and

documented recurrences of genital herpes. Subjects reporting more

stressful events over the 6-month period of study experienced more

recurrences. The predicted prospective relationship between stress

and recurrences was not confirmed by the data. Neither information

from the total sample nor from the subset of recurring subjects showed

any relationship between stressful events prior to the study and the

subsequent experience of herpes recurrences. Minor events, or

hassles, were not related to recurrences in any of the analyses.

The influence of several presumed moderators of life stress was

also investigated. Results differed depending on whether the total

sample or a subset of recurring subjects was considered. In the

former case, when information from all subjects was considered, an

interaction between social support and major negative life events was

found to be most predictive of recurrences. This interaction

indicated that high life stress was associated with more frequent

recurrences and that the effect of life stress became greater as

satisfaction with social support increased. This result was in direct

contradiction to prediction. In the latter case, when information

from only recurrers was analyzed, an interaction between major stress

and locus of control emerged as an important predictor of

recurrences. This finding was consistent with expectation. It

indicated that major stresses were associated with concurrent

recurrences for most subjects, excluding only those at the extreme

internal end of the locus of control distribution. Thus, an external

locus of control orientation and the experience of major stressors was

associated with higher recurrence rates. Arousal seeking tendency

failed to demonstrate any significant relationship to recurrence rate

in either analysis. In summary, with respect to those moderator

variables under study, only locus of control orientation interacted

with stress as predicted. Social support satisfaction had a

moderating influence directly opposite to that expected, and arousal

seeking tendency emerged as unimportant when other variables were


In the context of the present stress/illness literature, the

finding of a significant association between undesirable major life

events and an illness outcome measure is not new. It is important,

however, to have replicated this finding in an investigation designed

to avoid methodological flaws of earlier studies. In the present

study, stress elevations cannot be attributed to recall bias or a need

to justify illness (Rabkin & Streuning, 1976). Monthly reporting of

stressors was designed to avoid distortion due to forgetting (Funch &

Marshall, 1984; Monroe, 1982a; Paykel, 1983). Episodes of herpes,

objectively documented by viral culture, cannot be attributed to a

need to seek attention in the face of stress (Mechanic, 1975; Minter &

Kimball, 1978) or a lowered tolerance for discomfort due to stress

(Harney & Brigham, 1984). Therefore, the finding of a significant

association, albeit weak, is nonetheless important. The present study

employed stringent control measures and confirmed the findings of

previous studies. This enhances the credibility of earlier findings

and strengthens confidence in the existence of a stress/illness


With regard to genital herpes recurrences in particular, the

finding is even more important. As reviewed earlier, the belief in a

stress/recurrence link is based entirely on anecdotal reports (Adams

et al., 1980; Blank & Brody, 1950; Janicki, 1971; Ullman, 1947),

survey research (Bierman, 1982; "Help membership HSV survey," 1981),

and retrospective investigations (Taylor, 1978; Watson, 1983).

Prospective investigations have either not found a significant

association, possibly due to inadequate methodology (Katcher et al.,

1973) or failed to systematically monitor stress (Freidman et al.,

1977; Luborsky et al., 1976). The best empirical evidence for a

stress/recurrence association has been provided by Taylor (1978) and

Watson (1983) in retrospective studies based on self-reported,

undocumented episodes of recurrent herpes. Taylor (1978) found that

women who reported a high rate of recurrence also reported a higher

frequency of stressful events over the previous year. Watson's (1983)

replication of the finding in a larger sample of males and females

strengthens confidence in it. The correlation between major negative

life events and recurrences found in the present study (r=.19) is in

agreement with (significant and positive), but weaker than, those

found by Watson (1983; r=.32, r=.41). However, it is based on

evidence less subject to confounds than prior research.

The degree of confidence one may have in the findings of a

particular investigation can be related to the type of evidence it has

provided. One can conceptualize at least three levels of evidence

related to methodology in stress/illness research. Prospective

research in which stress is evaluated prior to the occurrence of

illness provides the purest and therefore the strongest evidence.

Concurrent research in which stress and illness are evaluated over the

same time period is somewhat weaker. The significant findings of the

present investigation are based on concurrent stress and recurrence

measures. Retrospective findings which form the bulk of evidence are

the weakest because the possibilities for distortion are highest.

Retrospective investigations are valued more for their heuristic

quality than for definitive findings.

The ability to demonstrate a prospective relationship to a large

extent depends upon the timing of event recording and occurrence of

illness. The impact that life events have may vary over different

periods of time (Sarason et al., 1982). Whether remote or recent

events affect a particular illness is also dependent upon the

characteristics of the illness. The timing of event-reporting in

life/stress research has been generally a matter of convenience or

convention. In an early investigation by Rahe and Holmes (cited in

Holmes & Masuda, 1974), physicians retrospectively reported stressful

events and health changes over a 10-year period. More recently, in

light of information regarding forgetting and distortion (Funch &

Marshall, 1984; Monroe, 1982a), shorter intervals have been used. A

decision regarding the optimal reporting interval to use in order to

observe a relationship could also be made considering physiological

information of the particular disease in question and related

stress/illness research findings.

With respect to genital herpes, there are reasons to look both at

long and very short latency periods. A long latency period is

suggested by Watson's (1983) research. He found a strong relationship

between stress during one 6-month period and recurrences in the

following 6-month period. The present study failed to replicate those

findings prospectively. Stresses in the 6-month period prior to the

present study were not related to documented recurrences during the 6-

month study. To investigate the possibility that stressful events

which preceded the study were related to recurrences in the first few

months as opposed to the entire 6-month period of the study, Spearman

correlations were done between LES scores for the 6-month period prior

to the study and recurrences in the first 2 months and the first 3

months of the study. These were -.13 (R=.23, n=85) and -.07 (j=.48,

n=85), respectively, indicating no significant relationship between

stress recalled for the 6 months prior to the study and recurrences in

the first 2 or 3 months of investigation. It is possible that the

earlier finding may have been influenced by retrospective bias or the

homogeneity of the sample (Watson, 1983) and resulted in a spurious

finding. More research is needed to resolve this issue.

An argument can be made on the basis of the present investigation

for a shorter latency. The significant findings were in the

concurrent information as opposed to the prospective information. The

attempt to use the monthly nature of data collection to examine the

essence of this stress/recurrence relationship more closely did not,

however, suggest elevated stress in the months prior to recurrences.

Although there could be many explanations for this observation, one

strong possibility is that the temporal relationship between stress

and herpes is of such brief duration that monthly measures are too

infrequent to capture the variability in stress that relates to

recurrences. If stress is causally related to recurrences and the

latency between the occurrence of a stressful event and a herpes

recurrence is a matter of 5 days or less, an event that occurs in the

beginning of the month could be related to a recurrence in the same

month. The elevation of stress that occurred prior to the recurrence

would not have been reflected in the score of the preceding month.

Similarly, if recurrences are causally related to stress and a

recurrence occurs in the beginning of the month, the resultant stress

could be reflected in LES scores in the same month and may not be

reflected in the scores of the following month. Thus, the analysis

based on monthly measures may not have been sensitive enough to detect

stress elevations that may be relevant regarding the directionality of

influence between stress and herpes.

The latency between the occurrence of stress and reactivation of

HSV may, in fact, be only a matter of days or hours. In support of

this, a latency of 1 to 2 days was reported by Herlig and Hoff (cited

in Janicki, 1971) and Ullman (1947) between emotional upset and

appearance of blisters. Direct neural stimulation such as occurs

during trigeminal surgery reliably results in recurrent herpes

episodes in 3-5 days in humans (Carton & Kilbourne, 1952; Pazin et

al., 1979). In an animal model, electrical stimulation of the nerve

produces herpes blisters within 3 days (Green et al., 1981).

The temporal relationship is an important consideration in the

design of life/stress studies. The present findings suggest the

possibility of a brief latency between stress and herpes. Therefore,

a recommendation for future research is to evaluate stress on a weekly

or even daily basis, while continuing to evaluate the possibility of

longer latencies. The data from more frequent monitoring can be

collapsed as desired to evaluate the impact of stress over a variety

of time intervals.

One of the aims of the current research was to enhance the

ability to predict illness outcome by evaluating minor stressors

(hassles) and presumed stress moderators as well as major life

events. There is a notion that the ill effects of major life stress

may be mediated by an increase in the number of hassles experienced

(Kanner et al., 1981). That is, the ill effects of major events

(e.g., moving) may be due to increased hassles (e.g., trouble getting

meals, being lonely, too many things to do). In the present research,

hassles did not demonstrate a direct relationship to recurrences and

the inclusion of hassles did not enhance the predictability of

recurrences in either regression analysis. In this sample at least,

the association of major negative life events and recurrences was

independent of an association with hassles.

With regard to the moderator variables, it was predicted that

they would interact with each other in such a way that a person's

scores on each of the measures would help to determine susceptibility

to recurrences under varying conditions of stress. This did not occur

in accordance with expectations for each measure. Of the three

presumed moderators chosen for study, only locus of control

orientation interacted with major stress as predicted. It was

hypothesized that individuals who do not feel in control of the events

in their lives would react more to stress and therefore be more likely

to have recurrences if stressed. There is support for this with

regard to adjustment outcomes (Johnson & Sarason, 1978), treatment

seeking (Manuck et al., 1975), and retrospectively reported herpes

recurrences (Watson, 1983). In general, a relationship is seen

between stress and illness only for externally oriented individuals.

The present study found support for this in the subsample of subjects

who had recurrences. No relationship was found between stress and

recurrences for individuals on the extreme low end of the scale

internalss). A significant positive relationship was found between

stress and recurrences for externals providing further support for

locus of control as a stress moderator.

Social support has been widely researched by investigators using

a wide variety of measures. In general, either an overall beneficial

effect or a buffering role in relation to life stress has been

attributed to social support (Cohen & Wills, 1985). These beneficial

effects have been more consistently demonstrated with psychological

outcomes than with illness measures however. Watson (1983) found a

direct effect on genital herpes recurrences such that those high on

social support (perceived helpfulness) had significantly fewer

recurrences. It was important to replicate these findings in a

prospective investigation because retrospective bias and recruitment

procedures in the earlier study limit the generalizability of the

findings. Forty of the 46 subjects were recruited through a herpes

self-help group which may have biased the findings regarding perceived

helpfulness (Watson, 1983).

The findings of the present study with regard to social support

are not in accord with those suggesting either a beneficial main

effect or those demonstrating a mitigating effect on stress. The

nature of the moderating role uncovered for social support revealed

instead a significant positive relationship between stress and herpes

for individuals high on social support satisfaction. That means

people who reported more satisfaction with social support also had

more recurrences when under stress. This unexpected finding was

closely examined. In the present investigation, social support scores

were highly skewed. Most subjects' scores suggested that subjects

were either very satisfied or very very satisfied with support

received. It is possible that individuals in this particular sample

were predominantly very active socially and happy with that. However,

subjects may have overextended themselves socially, thus negatively

affecting their resistance to recurrences. Unfortunately, the

restricted nature of the sample with regard to social support limits

the generalizability of these findings.

It was considered that subjects reporting high satisfaction with

social support may have been more sexually active which, in turn, led

to more frequent recurrences of genital herpes. Because information

was available on the frequency of sexual activity on a monthly basis,

this possibility was empirically examined. Frequency of sexual

activity was found to be related to social support satisfaction

(r=.21; _=.024) but the number of recurrences was negatively related

to frequency of sexual activity (r=-.25; _=.007). This suggests that

the number of recurrences may influence the frequency of sexual

activity rather than vice versa.

It is unfortunate that the sources of support are not

identifiable on the SSQ used in this investigation. It was possible

to discern network size and perceived quality of support, but not

possible to know if the perceived sources of support were family,

friends, professionals, peers, confidante, or a significant other.

With regard to genital herpes in particular, the support of a

significant other could be more important than others.

It was also not possible to differentiate between types of

support on the social support measure. Cohen and Wills (1985) provide

a useful typology of four support resources: esteem support,

informational support, social companionship, and instrumental

support. They suggest that there must be a reasonable match between

coping requirements and available support in order for stress

buffering to occur. Relating this to the present findings, it is

conceivable that respondents' reported perceived satisfaction on the

SSQ is more reflective of the type of support labeled social

companionship by Cohen and Wills (1985). Thus, despite the tendency

to be quite pleased with available support, it may not be the type of

support that assists one in dealing with increased stress. As

suggested by Graves and Graves (1985) whose measures of illness

symptoms also provided evidence contrary to a stress-buffering role

for social support, perhaps the obligatory nature of dense social

networks is maladaptive under conditions of life change.

Together with the wide range of findings regarding social support

in the literature (see reviews by Cohen & Wills, 1985; Thoits, 1982),

the present results further emphasize the complexity of the phenomenon

of social support, and the need for more specific information

concerning source and type of support.

Arousal seeking tendency was hypothesized to exert a stress

moderating role on the basis of studies dealing with adjustment

outcomes. Subjects low on arousal seeking tendency have been shown to

be more subject to psychological discomfort under high stress

conditions than high arousal seekers (Johnson et al., 1979; Smith et

al., 1978). The present investigation unsuccessfully attempted to

replicate and extend those findings with regard to a physical outcome

(genital herpes recurrences).

Although the logistic regression and stepwise regression

procedures resulted in interesting findings, only 3% of the variance

in recurrences was accounted for. Furthermore, because a different

moderator variable was chosen in each regression analysis depending

upon whether the full sample or only recurrers were utilized, it

suggests the effects may not be robust. Social support was important

only when predicting who will or will not have recurrences, locus of

control was important when predicting how many recurrences the

recurrers would have.

Thus, the present investigation has provided support for a

significant but weak association between stress and illness. This

raises questions of how to proceed in future investigations of the

nature of recurrences and their relationship to stress. In order to

enhance predictability as originally planned, further methodological

refinements will have to be accomplished. The use of more frequent

stress monitoring to evaluate a variety of impact intervals and a more

specific measure of social support have already been recommended.

On the basis of the suggestive results of the present study

concerning the effects of social support and locus of control,

continued exploration of individual differences appears warranted.

Other psychological factors might also be considered. While it was

inferred that the three characteristics chosen for study in the

present investigation would enhance one's ability to cope with stress

and thus mitigate the illness-producing effects of stressful life

events, coping skills such as problem-solving ability could be more

directly evaluated.

In addition to psychological factors, the role of biological

factors should be examined. Results of an earlier prospective study

of oral herpes suggest that biological factors are better predictors

of oral herpes recurrence than psychological factors including mood

and social assets. Disease history and illnesses accounted for 80-90%

of explained illness (Friedman et al., 1977). A measure of illness

history and predisposition could potentially improve predictability.

Because of the interest in the mechanism of recurrence, it might also

be useful to evaluate hormonal and immunological parameters to examine

fluctuations which relate to stress and recurrence onset. To maximize

the probability of observing recurrences, high recurrers could be

evaluated in a single-subject design with frequent stress and

physiological monitoring.

The present findings regarding major negative life events and

recurrent herpes provide limited empirical support for the long held

notion of a stress/recurrence link. Because this evidence is

consistent within the study and based on concurrent findings, it is

stronger than that of previous retrospective findings. However,

despite stringent control measures, 96-97% of the variance with regard

to genital herpes recurrences remains unexplained. It behooves

researchers and clinicians to collaborate with professionals in an

interdisciplinary manner to further untangle the complexities of



Initial Structured Interview

Demographic Information

1. Sex:
(1) F
(2) M

2. Age:

3. Marital Status:
(1) Single
(2) Living with significant other
(3) Married
(4) Separated
(5) Divorced
(6) Widowed

4. How many children do you have?
(1) 0
(2) 1
(3) 2
(4) 3
(5) 4
(6) 5 or more

5. What level of education have you completed?
(1) Grade school
(2) High school graduate
(3) Some college
(4) College graduate
(5) Some graduate work
(6) Master's degree
(7) Ph.D., M.D., or other advanced degree

6. Are you presently a student?
(1) Yes
(2) No

7. Have



8. What

9. When

you had any major illnesses over the past 6 months?

Do you have any major illnesses or chronic conditions?

Are you taking any medication?

prompted you to volunteer for this study?

did you first experience problems with genital herpes?

10. When was genital HSV first diagnosed?

By whom?


11. When was your most recent HSV recurrence?

12. How many HSV recurrences have you had in the last year?

None 1-3 4 or more





What do you think causes recurrences for you?

Do you at present have a regular sex partner? Yes No

Does your partner have herpes? Yes No

respect to the last 6 months, what has been your usual frequency

16. Genital/genital contact?

17. Oral/genital contact?

18. Masturbation?

19. With approximately how many different partners have you had
sexual contact over the past 6 months?

20. Have you had sexual contact with any partners who have genital
HSV over the past 6 months?


21. Has there been a change in your usual pattern of sexual activity
over the past 6 months?

If so, describe:

To what do you attribute this change?

22. Do you have any information that you want to add to what we've
already talked about?