Stress 1

01-Linden.
qxd 9/13/2004 9:07 AM Page 1
Stress
Definitions and Pathways to Disease
Stress is essentially reflected by the rate of all wear and tear

caused by life.
—Hans Selye, 1956
Scientific and Popular Definitions
It makes little sense to write about stress management (SM) unless there
is clarity about the phenomenon that is to be managed. For this reason,
a broad but by no means exhaustive review of the term stress and its
importance for health is provided first. In this chapter, the meaning of the
terms stress and management is explored, and research is described that
reveals how they can be connected.
“Stress” has become so ubiquitous and so much a part of everyday
language that, at first glance, there appears to be no need for a definition.
Selye (1976), a pioneer of stress research, points out that “stress is a
scientific concept which has suffered from the mixed blessing of being too
well known and too little understood.” Consistent with Selye’s view, it is
argued here that when concepts from basic science become popularized,
there is potential for oversimplification or alteration of the term that may
ultimately belie its origins and add to confusion.
When seeking definitions, the general populace does not read
scientific journals. People are much more likely to refer to other “gold
standards” of definition like Webster’s Dictionary. Ideally, definitions
contained therein are in full accord with scientists’ definitions but just
1
01-Linden.qxd 9/11/2004 11:57 AM Page 2
2 STRESS MANAGEMENT
appear in a simpler or broader language. What, then, can one learn from
consulting a dictionary?
Webster’s dictionary (Webster’s Illustrated Encyclopedic Dictionary,
1990) gives six definitions, ranging from a generic definition to more specific
ones depending on areas of application. The first, most generic, one is:
“Importance, significance, or emphasis placed on something.” The second,
third, and fourth deal with stress as a feature of spoken language and sound:
“The degree of force with which a sound or syllable is spoken”; “The relative
emphasis given a syllable or word in verse in accordance with a metric
pattern”; and finally, “an accent” in music. The next definition relates stress
to physics: “An applied force or system of forces that tend to strain or
deform a body, measured by the force acting per unit area.” Finally, a
definition is given that is more psychological in nature: “A mentally or emo-
tionally disruptive or disquieting influence, or alternatively, a state of tension
or distress caused by such an influence.” One can easily see differences in
these many definitions such that only the definitions used for physics and
psychology contain elements of an action and a result, a challenge and a
response. The novelty of the term stress also provided a considerable
challenge for translation into other languages; for example, there are no
equivalent terms in French or German, and in the end it was largely decided
to use the word stress in the same way across many different languages.
Given that Selye (1976, p. 51) saw stress as the result of a process, he further
felt a need to label the beginning of the process in a manner distinct from the
outcome and coined the term stressor to refer to a causative agent, a trigger
for this process.
A two-step sequence of “stress” is also reflected in definitions found
in psychological textbooks. Girdano, Everly, and Dusek (1993, p. 7) state,
“Stress is the body reacting. It is psychophysiological (mind-body) arousal
that can fatigue body systems to the point of malfunction and disease.”
Hence, popular and scientific definitions see “stress” as a process in which
external and internal stimuli, forces, or systems interact, where triggers
activate a response system that may lead to exhaustion and vulnerability
(Wheaton, 1996).
My definition of stress, as applied to stress management, is this:
Stress is a mediational process in which stressors (or demands) trigger

an attempt at adaptation or resolution that results in individual distress
if the organism is unsuccessful in satisfying the demand. Stress respond-
ing occurs at physiological, behavioral, and cognitive levels. Stress is
more than just acute subjective or physiological activation and has its
potentially most deleterious health effects when it becomes chronic.
Stress: Definitions and Pathways to Disease 3
Before delving further into the history and basic research on the stress
concept, it should be clarified that, consistent with my definition of stress,
the emphasis is going to be on chronic stress and its health consequences
rather than on a singular, traumatic kind of stress exposure. Exposure to a
traumatic event, like witnessing or being subjected to violence, is a profound
event with potentially grave and long-lasting psychological sequelae; in
their most severe form, these sequelae qualify for posttraumatic stress dis-
order, which can be quite debilitating. Little is known about the long-term
physical health consequences of traumatic stress (with the exception of
early life exposure to trauma, discussed below) and the treatment tech-
niques embraced by stress management (Ong, Linden, & Young, 2004)
are not treatments of choice for posttraumatic stress disorder (Taylor,
Lerner, Sherman, Sage, & McDowell, 2003).
Also of importance is the recent introduction of the term “acute stress
disorder” (American Psychiatric Association [APA], 1994), which was meant
to describe initial reaction to trauma that in turn predicts posttraumatic
stress disorder. There is considerable debate whether science and clinical
practice are well served by having two disorders that are so closely inter-
linked, and that are really distinct only in the time period required for their
manifestation (2 days to 4 weeks relative to at least 1 month post trauma;
Harvey & Bryant, 2002). Notwithstanding this debate, neither acute stress
disorder nor posttraumatic stress disorder will receive much attention here.
How Can Stress Be Measured?
Having a clear definition of a construct is a useful and necessary precursor

for its measurement. In the case of stress, which was defined as a multistep
process, the answer to the question posed in the title of this section is
anything but simple. The great majority of what has been written about
the measurement of stress is really the measurement of the stress response,
that is, the result of the stress process. Because stress responding can occur
at behavioral, cognitive, and physiological levels, measures of stress
responses at each level would ideally correlate highly with each other,
meaning that they would be synchronous. Unfortunately they are often
desynchronous, and some observations made in my laboratory and a brief
anecdote may serve as illustrations of the relative desynchrony among
various stress measures.
When exposing individuals to controlled laboratory stressors, we
routinely request participants to provide a rating of the stressfulness of the
4 STRESS MANAGEMENT
experience, primarily to serve as a validation check. This method has been

very effective in showing that a mild stressor like exposure to white noise
receives a mean rating of 3 on a 10-point scale whereas an arithmetic chal-
lenge with interjected harassing feedback is likely to rate on average of 7
out of 10, thus validating the anticipated differences of the severity of the
stressors. However, irrespective of the type of stressor used, and provided
that a large sample was tested, some participants rated the exact same
stressor 1 out of 10 whereas somebody else rated it 10 out of 10. What
accounts for such whopping differences in perception? It is posited that
these often greatly varying ratings of the same stimulus reflect a blend of
(a) stable, natural response tendencies toward repressing or sensitizing to
individually relevant affective information; (b) the subjective, idiosyncratic
meaning of the stressor; (c) possible differences in the ability to perceive
simultaneous physiological activation; and (d) possible mood priming via
pleasant or unpleasant daily events that preceded participation in the
research study.
A researcher who sees such great variability in the judgment of the
exact same stressor develops great doubt about the comparability of sub-
jective stress ratings across individuals, and develops a cautiousness that
outsiders may not readily share. Along these lines, I had been approached
by a TV reporter who was aggregating information from “experts” and
the lay public about the presumably growing level of stress in the Canadian
populace. The TV production team wanted to measure the absolute level
of “stress” by conducting a representative telephone survey and using self-
reported stress levels as the index of “real population stress.” When I told
the reporter that in my opinion this method of assessment was not an ade-
quate way of measuring “population stress,” he was quite surprised and
asked for an explanation. The analogy I used was that measuring “stress”
was a lot like measuring “winter.” We all (especially we Canadians) know
what winter is, but it is also clear that there is no single defining charac-
teristic of “winter.” That notwithstanding, people know and even agree on
a number of features that jointly characterize winter, including below-
freezing temperatures, reduced daylight, snowfall (or increased rain in
some climate zones), a time epoch in the calendar ranging from December
to March, and so forth. My position with this reporter was that the more
of these features we measure, the more we capture the global phenomenon
“winter,” and that measuring “stress” is very much the same.
An attempt at applying this reporter’s crude definition to a real-world
problem may further serve to strengthen the point. If one accepted that self-
reported stress by one individual was a fully satisfactory definition of stress,
then workers’ compensation boards and various insurance companies would

likely go bankrupt as a consequence of the resulting number of “stress dis-
ability” claims they would have to pay out on. In a variety of jurisdictions,
workers’ compensation boards already have to deal with these issues and
have categorically decided that subjective self-report is clearly not sufficient
for a stress-related disability.
In principle, reliance on self-report of stress would make sense if there
were a close correlation (with high sensitivity and specificity) of self-
reported stress with the biological markers that are known to play a criti-
cal role in the process of activation, failure of recovery, and exhaustion.
Unfortunately, the literature indicates that biological changes and self-
reported stress, even under relatively transparent circumstances, are at best
moderately correlated, as research on acute physiological reactivity in the
laboratory shows. Even in well-controlled laboratory environments with
reduced stimulus complexity, physiological and parallel mood changes
rarely show correlations exceeding r = .3. Self-reported distress rarely
explains more than 10% of the variance in physiological change (Linden,
1987). While disappointing, this is not really surprising because (a) there are
few direct pathways between central nervous system activity and conscious
awareness, (b) researchers have observed marked individual differences
in ability and willingness to sense and report physiological changes
(Pennebaker, 1982), and (c) people rely heavily on contextual clues for
inferring physiological changes from mental representation of environmen-
tal events (“This is an important test and I know that I am ill-prepared, so
whatever I feel must be ‘stress’ and my fast-beating heart confirms this”).
At the level of sensation, there is inherently limited awareness of biological
markers, ranging from complete inability to sense, for example, lipid changes
or platelet aggregation in the blood, to a rather modest awareness of blood
pressure or heart rate changes, to reasonably accurate knowledge of breath-
ing rates or rising blood alcohol levels (although even in the latter cases false
feedback studies show that only large changes are accurately perceived).
Understanding the relationships between context use and accurate physio-
logical sensing requires delving into basic psychological research that dif-
ferentiates sensation from perception.
The problems that are endemic to people’s relative inability to accurately
sense stress-related biological changes, and to the unavoidable influence of
context variables on self-reports of stress, are particularly worrisome when
important decisions with long-term impact have to be made and objective
indices are hard to come by (see the discussion above on the workers’ com-
pensation systems that process claims for stress-related disability).
6 STRESS MANAGEMENT
Is there an answer to the question, “Can stress be measured?” In an

absolute sense, the answer has to be emphatically “no.” Given that stress
is not a fixed state but a process with multiple phases and with interacting
cognitive, behavioral, and physiological processes, we cannot readily index
stress and should not even attempt to draw inferences from it from any
single index, whether subjective or physiological. We can, however, aggregate
information from self-reported distress, observe behavior, and determine
physiological activation that is known to be relevant to the stress arousal
and exhaustion process. Grossi and his collaborators (Grossi, Perski,
Evengard, Blomkvist, & Orth-Gomer, 2003), for example, have compared
people with self-reported high and low burnout and found a reasonably
high level of parallel self-reported stress and physiological marker activity
in neuroendocrine and immune systems. Self-reported stress levels were
also sufficient to predict significantly greater mortality risk over 5 years in
a cohort of 6,920; this effect held true even after controlling for sociode-
mographic and known cardiovascular risk factors (Rasul, Stansfield, Hart,
Gillis, & Smith, 2004). No class of measurable phenomena that are corre-
lates of stress (and certainly not any single index within each class) can be
accepted as absolutely reflecting “stress”; any inference needs to involve
understanding of the context and needs to establish the concordance of
various stress markers. One can take, for example, cortisol and its precur-
sor ACTH, which are widely considered to be good markers of the stress
response. In addition to reflecting varying stress levels, they are also influ-
enced by naturally occurring diurnal patterns, individual differences, and
random fluctuations that prevent absolute inferencing of cortisol activity
to mean “stress.”
Nevertheless, I do believe that it is meaningful to study subjective
reports of stress in the same individual over time (as is done in diary stud-
ies) and then relate them to parallel occurring objective events, given that
the individual difference variables in stress reporting are presumably stable
over time. This claim needs to be tempered, however, with the fact that
reactive situations (like stress and pain reports in claimants for a disabil-
ity) threaten the trustworthiness of subjective stress reporting.
A History of Models for Stress and Health
The decision to start with a chapter on the history of stress should not be
taken to mean that an exhaustive review and discussion will follow; the
intent here is to focus on those features of previous theorizing that have most
prominently contributed to shaping this book and its objectives. Readers

who want more in-depth reviews of theories and proposed biological path-
ways from stress to disease can seek out a large number of books and review
articles. As such, there are many excellent undergraduate textbooks in health
psychology that provide broad overviews, and for greater depth of facts and
discussion, I recommend Lovallo’s (1997) excellent discourse, Stress and
Health, as well as McEwen’s (1998), Ray’s (2004), and Kelly, Hertzman,
and Daniels’s (1997) review articles.
These caveats notwithstanding, some background on major theories
and empirical findings needs to be presented early in this text so that
sound, empirically based psychophysiological rationales for stress man-
agement interventions can be offered, and so that my criticisms of extant
thinking and writing on stress management can be solidly grounded.
What may appear to be modern approaches to understanding stress
have roots in ancient views of health and disease that can be traced to
beliefs and practices of Oriental (around 2600 B.C.) and Greek physicians
(around 500 B.C.) who advocated moderation, avoidance of excess, and
concepts of balance and harmony. The current practices of acupuncture,
meditation, yoga, biofeedback, self-hypnosis, and Autogenic Training can
be traced back to these ancient views of health as a state of good balance.
Cannon (1928) can be credited with describing a view of physiologi-
cal balance that at once reconnects medicine with historical views of a
healthy balance and also represents a sound approach to physiology that
is actually measurable and quantifiable. His work underscored that the
autonomous nervous system response to challenges needs to be understood
as a dynamic interplay of sympathetic and parasympathetic activation
in the autonomous nervous system. These two regulatory forces of the
nervous system have opposing actions and both need to be strong and
responsive to achieve or maintain health.
Selye’s (1956) general adaptation syndrome can be seen as an elabo-
ration of Cannon’s work in that he showed conditions and pathways for
nervous system activity to become unbalanced. The general adaptation
syndrome describes stress as a potential 3-step sequence of events in which
a challenge (like the appearance of an aggressor—step 1) precedes the
body’s activation of its innate coping abilities to deal with the challenge:
fight, flee, or otherwise adapt (step 2). Frequently, the whole process of
challenge and response ends right there because the challenge has been
effectively met, and the constructive arousal that accompanied the stress
resistance and that allowed active responding is no longer needed and can
return to a physiological resting state. Consistent with Cannon’s work, it
8 STRESS MANAGEMENT
can be seen that after initial sympathetic activation, the body’s natural
inhibitory systems “kick in” in the form of counterregulatory, de-arousing,
parasympathetic activation. However, not all challenges are of time-
limited nature and/or allow quick, decisive responding, and the body con-
tinues to resist, becoming by necessity exhausted at some point (step 3).
This physiological exhaustion is considered to carry disease potential
because the body is now weak and unable to resist. Interestingly, the term
exhaustion found in Selye’s work has been carried forward into other
researchers’ work and, for example, a Dutch research group has coined the
term vital exhaustion, which they have shown as critically preceding
myocardial infarction (Van Diest & Appels, 2002). Vital exhaustion
describes a psychophysiological state of mental numbing that is related to
perceived low self-efficacy (i.e., an awareness of one’s own low level of
effectiveness), inability to cope, and a subjective sense of low energy and
fatigue, thus vividly describing a blend of biological and emotional features
that aptly represents Selye’s notion of exhaustion.
Selye’s original work posited a whole-body response such that exter-
nal challenges were held to lead to the same cascade of physiological
responses. The typically occurring physiological changes in response to
a challenge are well established and described in numerous textbooks.
Figure 1.1 describes the paths and the sequence of activities in Selye’s
activation-exhaustion model (1976).
This coarse model of Selye’s activation/exhaustion model leads to a
series of questions that need answering if the full process is to be under-
stood: (a) Which stimuli (or stimulus properties) activate the process?
(b) What is the physiological chain of actions that mark an activation
process? (c) Who, under what circumstances, adapts and who, under
what circumstances, becomes exhausted? Satisfactory answers to these
core questions ought to provide the stress management researcher with the
critical knowledge needed to develop a sound rationale for interventions.
Exhaustion
Stimulus Activation or
Adaptation
Figure 1.1 The Selye Model of Stress

Stress-triggering qualities of stimuli will not be discussed first

(a detailed review is found at the beginning of Chapter 2). Instead, the phys-
iological activation process is described first because it is central to the orig-
inality of Selye’s work. Selye believed in response universality; that is,
all stimuli above perception threshold are held to trigger similar physiolog-
ical stress responses across different species and across situations. This pre-
sumably universal cascade of events (adapted from Sarafino, 2002) is as
follows:
1. Environmental stimuli that are judged to be of subjective importance

trigger cortical activation that sends chemical messengers to the hypo-
thalamus, where
2. they stimulate the production of corticotrophic releasing factor (CRF)

and other chemical messengers that, in turn, activate two distinct tracks
of bodily reactions.
3. In the first track (also commonly referred to as the sympathetic adrenal

medullary axis), these messengers feed information forward to the
pituitary gland, which
4. changes the chemical structure of the messengers and releases adreno-

corticotropic hormone (ACTH) into the bloodstream.
5. When ACTH reaches the adrenal glands, it initiates the production of

cortisol, which, in turn, increases metabolic rate. Cortisol inhibits the
function of phagocytes and lymphocytes in the immune system (i.e., it
serves as a messenger for needed adaptations of the immune system).
6. On the second track (commonly referred to as the hypothalamic-

pituitary axis), chemical messengers leave the hypothalamus and trigger
electrochemical changes that advance as signals down the brain stem and
the spinal cord toward the adrenal glands.
7. At the level of the adrenal gland, this activation leads to release of

epinephrine, which supplies extra glucose to musculature and brain.
Epinephrine also increases suppressor T-cells and decreases helper
T-cells, thus revealing a second connective pathway of stress reactivity to
immune function.
8. The adrenal glands also release norepinephrine, which then speeds up

heart rate and increases cardiac output and blood pressure.
9. Ultimately these activities and their results are fed back to the hypothal-
amus, which serves as the “master controller” of this whole process.
10 STRESS MANAGEMENT
This cascade of physiological responses to challenge involves intricate

interplays and feedback loops of biochemical and electrophysiological
processes that regulate autonomic nervous, endocrine, and immune system
activity. Within the autonomic nervous system, activation of the sympa-
thetic branch (via release of ACTH) prepares organs for the fight-or-flight
response by dilating pupils and bronchi, increasing the rate and force of
the heart’s pumping action, constriction of blood vessels, secretion of epi-
nephrine, and decreasing peristalsis. An important physiological control
function is assigned to the parasympathetic branch of the autonomous ner-
vous system that opposes these actions and is functionally designed to
facilitate recovery. One can readily see that the teaching and learning of
techniques to maximize parasympathetic system flexibility is of critical
importance to physiological stress management.
This activation process, that is, the fight-or-flight response, does
not carry within its definition any connotation of inherent maladap-
tiveness; if anything, the opposite is true as long as fight or flight is
really necessary for survival. In the evolution of species, fight-or-flight
responses are highly useful tools for survival because they maximize
the availability of muscular energy, sensory acuity, and protection of
tissues from injury. However, a critical and widely accepted feature of
Selye’s view of the activation response is that our biological systems
cannot sustain this activation for a long period of time; all living crea-
tures become exhausted if the fight-or-flight response does not lead to
resolution of the challenge. The changed flow of blood during activa-
tion relative to rest can be used to show how initial adaptiveness of the
activation process can turn into a long-term health threat. During rest,
most of the body’s blood volume circulates between the heart and the
viscera so that needed nutrients and oxygen are available for organ
function; that is, about 60% of blood flow during rest goes to the
kidneys, skin, digestive system, and bones. During fight-or-flight
responses, however, the bulk of the total blood volume is made avail-
able to the muscles and the brain, with only a fraction still being avail-
able to support visceral functions (in that case, only about 20% of the
blood volume goes to skin, digestive system, kidneys, and bones;
Astrand & Rodahl, 1970). Just as whole living beings cannot stay alive
without nutrients or oxygen for an extended period of time, neither
can individual organs. Recovery from the stress-induced blood redis-
tribution to muscles and brain needs to occur relatively quickly if
the organ tissues are not to suffer damage from lack of nutrients and
oxygen. This relatively simplistic description of how activation can

turn to exhaustion is quite suitable for explaining the stress-exhaustion
process to patients who present with stress-related psychophysiological
disorders.
Selye’s idea of response universality has been challenged by subsequent
models and research findings discussed in more detail below. What was ini-
tially a simple stimulus-response model was expanded to include genetic
and early learning differences in responsivity as well as acute response mod-
ulation through behavioral and cognitive activity.
In order to understand how macro-level effects at the community
level can affect cellular activity and vice versa, Brody has described a
17-layer system in which levels of influence for stress and health are orga-
nized on a continuum from mini to macro activity, from atom to mole-
cule, to cell and tissue, to organs, systems, whole person, and finally
community and society (Brody, 1973). He presumes that each level of
activity influences the ones directly above and below so that ultimately
a connection can be shown between molecular change and societal
change. The direction of influence can be ascending and descending on
this continuum.
A particularly influential model of stress and health is Levi’s (1972)
interactional model (Figure 1.2), which extends Selye’s work by arguing
that the magnitude of a stress response can be better predicted by under-
standing the stressor in the context of a person’s predisposition, thus open-
ing the door for a better understanding of individual differences in the
stress response magnitude.
STIMULUS RESPONSE
Psychological stress
Stressor Stress
Physiological stress
ENVIRONMENT PERSON
Figure 1.2 The Stress Diathesis Model

While useful for driving and organizing further research, Selye’s and Levi’s
models are now considered simplistic by more cognitively oriented researchers.
Before continuing to discuss more expansive models, it may be of use to orga-
nize existing models into categories. For the purpose of this book, the catego-
rization structure proposed by Feuerstein, Labbe, and Kuczmierczyk (1986,
p. 122) was adopted here; it differentiates the following four types of models:
(1) response-based models, (2) stimulus-based models, (3) interactional mod-
els, and (4) information processing models.
Selye’s model is considered a response-based model because it pre-
sumes that all stimuli trigger the same cascade of physiological responses
and that the nature of the response (activation and resolution vs. activa-
tion and exhaustion) ultimately dictates the health outcome. Levi’s model
is more of an interaction model because it opens up the possibility that the
response to a stimulus is not universal but affected by the type of stressor
and the context, which, if understood, allows a prediction of the magni-
tude and likelihood of a subsequent stress response. The problem with
simple response- and stimulus-based models is that they do not account for
the great variation that people actually show under equally demanding
conditions. Interactional and information processing models allow for the
existence of feedback systems and presume a more cyclical nature of stress
processes. In one such model (Cox & McKay, 1978), a critical element in
determining the magnitude of a potential stress response is the cognitive
appraisal of a demand situation in that stress is presumed to result when
the perceived capability to cope does not meet the perceived demand. This
“imbalance” is by definition “stress.” Perceived ability to cope is in part
determined by actual capability, and perceived demand is in part deter-
mined by actual demand. Discrepancies between actual and perceived
demand and actual and perceived coping capability are further moderated
by prior learning and personality differences. Cox and McKay further
posit that individuals judge the adaptivity of their coping attempts and that
information is fed back to the individual, who may then reevaluate the
stressfulness of the situation, or the levels of (perceived and actual) demand
and capability.
Information-processing models are fairly similar in that they focus
on attention, appraisal, and memory processes. Hamilton (1980) sees per-
ceived stress levels as the aggregated result of a stressor recognition and
evaluation process that integrates the importance and meaning of the stres-
sor, the amount of attention that the system allocates to it, and memories
about past effective and noneffective coping experiences. According
to Hamilton, stressors themselves can have three challenging qualities:
(1) anticipation of physical pain or danger, (2) threat of social isolation or

rejection, and (3) stimulus complexity involving either concurrent response
demands or novelty and complexity. More than one of these qualities may
be present at the same time, and it is reasonable to predict that stressors
are ultimately leading to greater stress responses if two or even three of
these qualities are present simultaneously.
When trying to use models of stress to build an empirically grounded
rationale for stress management, one needs to realize how application of
various stress theories to different species also creates unique challenges
and opportunities for studying stress-related disease etiology. Applying
basic research findings from the animal model to human applications is
particularly challenging for understanding attempts at prevention and
intervention. This can be illustrated by relating stress models to plants,
animals, and humans and by highlighting similarities and differences. Bernard
(1961) makes the case for a reductionist approach to understanding
survival of simple organisms like plants. These organisms are dependent
on their environment to supply nutrients, water, light, and so forth, and if
these “supplies” become exhausted, the organism dies. Yet even this reduc-
tionist perspective allows adaptation in that plants may evolve to thrive in
more or less light, or varying degrees of water supply; ultimately, however,
they are still supply-dependant. Animals, on the other hand, have mobility
and can actively seek out supplies, and this search can lead to extreme
efforts such as the annual migration of geese over thousands of miles, even
from one hemisphere to the other, in order to obtain stable food supplies.
This greater mobility translates into much greater activation potential of
the organism, but that, in and of itself, also carries the potential for quicker
physical exhaustion as can be seen, again, in the example of migratory
birds who may not survive a lengthy migratory flight to their winter (or
summer) location. In humans, there is the same principal dependence on
environmental features as applies to all organisms (e.g., oxygen and nutri-
ents), there also is mobility (fight or flight) as is seen in animals, and, in
addition, humans possess cognition and mental representation that can
serve as stress triggers (which will be discussed later in this book). Cognition
is also a potential source of stress resolution. Without denying the possi-
bility of thought processes and a degree of consciousness in animals, find-
ings from tests of animal models offer little generalizability to human
cognitions. Stress triggers for humans can range from very physical (like
lack of air supply during an asthma attack, which is of course a factual
challenge to survival of the organism), to very psychological and human-
specific triggers, like a sudden fear response to the teacher’s announcement
of a pending test for which a student did not study, or a husband’s

sudden realization that he forgot to get a birthday gift for his wife. Hence,
the gamut of potential stress triggers for humans is almost infinite, cover-
ing physical, objective challenges inherent in the environment at one
end of the spectrum, as well as symbolic and learned challenges at the
other end.
The most recent models of stress and health have challenged Cannon’s
(1935) relatively rigid notion that a homeostatic state is needed and desir-
able for survival, and have attempted to classify the types of challenges that
do not merely trigger acute responses but also possess potential to become
chronic stress triggers. Theorizing about ideally functioning physiological
systems had initially focused on the idea that there are ideal values (reflect-
ing homeostasis in the sense of Cannon) and elevated, that is, maladaptive,
values (as is the case with blood pressure). However, this model implied a
rather static definition of homeostasis and has been replaced by the recog-
nition that living systems must adapt, change, and accommodate to chang-
ing circumstances to guarantee survival of a species. A system’s ability to
achieve long-term stability and health through ongoing adjustment change
has been labeled allostasis (Sterling & Eyer, 1988). Allostasis presumes
that physiological systems strive to remain within a healthy range of func-
tion that allows optimal responsivity to external challenges while main-
taining their own control functions. The allostasis concept is not meant to
be in contradiction to the homeostasis concept; rather it modifies the def-
inition of homeostasis as a desired variability within a healthy range
instead of a fixed static level. A good example of the danger of “excessive
stability” in the physiological domain is the observed lack of variability in
the intervals between heartbeats that precede sudden cardiac death
(Kamarck & Jennings, 1991). Similarly, the immune system works at its
best when responding to immune challenges, and it is argued that repeated
challenges can strengthen the system’s ability for future adaptive respond-
ing; this principle is, of course, the well-known rationale for vaccinations.
The same can be stated for the value of physical exercise, which, when
applied in moderation, serves as a stress buffer because it enhances the
body’s ability to adjust, to respond quickly to challenges. At a psycholog-
ical level, parents, for example, must adjust their expectations about rea-
sonable rights and responsibilities for their growing children; as the
children mature, curfews and allowances may need to be adjusted upward,
and increasing levels of responsibility can be expected in return. Failing to
make such adjustments is almost guaranteed to create family strain and
prevent desired maturation.
The extent of challenges or demands on a system is, then, referred to

as the allostatic load, which at some point may exceed a given system’s
ability to cope. McEwen (1998) describes four types of situations in which
the nature of the challenge and the resulting allostatic load may exceed
the body’s capacity to respond and therefore lead to damage. The first such
scenario is the frequent repetition of exposure to a stimulus with no time
for recovery before the next stressor is represented. Figure 1.3 displays this
“repeated hits” scenario.
LOAD
EVENTS OVER TIME
Figure 1.3 Repeated Hits Model
A real-world scenario that represents this type of situation would be

emergency room staff in a busy city hospital where recent personnel cuts
have led to serious understaffing that requires frequent overtime work.
Anybody watching an episode of E.R. on television will see this scenario
come to life! The potential for quick burnout and exhaustion in this
environment is readily apparent.
A second type of allostatic overload is also characterized by repeated
stressor appearance but differs in that some individuals fail to show
adaptation that others are capable of. An example would be adjustment
(or the lack thereof) to living close to a fire station; some people may develop
the capacity of sleeping through repeated siren noise in the night whereas
others simply fall farther and farther into sleep deprivation. This scenario is
displayed in Figure 1.4.
A third type of stressor situation with high “overload potential” is one

where badly needed recovery is delayed or fails to happen. This type of sit-
uation (displayed in Figure 1.5) arises, for example, through chronic work
stress or in response to interpersonal conflict, as diary studies of marital
interactions (DeLongis, Folkman, & Lazarus, 1988) and interpersonal
LOAD
Person A
Person B
EVENTS OVER TIME
Figure 1.4 Lack of Adaptation Model
Event
LOAD
Beginning End Recovery Phase

TIME
Figure 1.5 Lack of Recovery Model
laboratory stressor studies (Earle, Linden, & Weinberg, 1999; Linden,

Rutledge, & Con, 1998) have shown. Such scenarios are not particularly
rare, as Frankenhaeuser (1991) has shown in her psychophysiological eval-
uations of working mothers. These women showed understandable stress
responses to heavy demands at the workplace, and the return home

provided no reprieve either (as was shown by tracking stress hormone
release) because there still were housework and other family responsibili-
ties waiting for them.
A fourth situation is not displayed graphically because it involves
multiple, complex, interacting systems; here the inadequate response of
one allostatic system triggers a compensatory response in other systems.
An example of this nature provided by McEwen (1998) describes stress
hormone activity. If cortisol secretion in response to a challenge does not
occur, then inflammatory cytokines are released that make animals (and
presumably humans alike) susceptible to autoimmune disturbances.
The addition of the allostatic load concept to stress research represents
an expansion, refinement, and evolution of Selye’s concept of the adaptation
syndrome but does not contradict its basic premises. With the addition of
research on allostasis, there now is a richer, better-documented picture of
how to differentiate normal from pathological stress responding, and exten-
sive descriptions exist of various critical situations as well as a rich body of
studies on the physiological pathways and their interactions for understand-
ing the linkage of stress to disease outcomes. Interestingly, neither Selye’s
original work nor the expansion model arising from the allostasis concept
deals with how the stimuli themselves come into existence, whether or not
they deserve attention, and whether they can be eliminated or manipulated.
These stress models do not address the question of potential modifiability of
the stimuli themselves. The inherent message is that stressors exist and that
physiological systems (and the people who are governed by these systems)
must respond; that is, people are not seen as creators or shapers of their own
environment, they are simply considered to be reactors.
Understanding Stress Responsivity
Individual Differences in
Stress Responding and Their Origins
As any astute reader of scientific articles on stress reactivity will

quickly find out (especially when reviewing tables of results in stress reac-
tivity studies), people’s responses to the exact same stimulus vary greatly.
In my laboratory, we have obtained such evidence in numerous controlled
studies where participants were exposed to contrived stressors (like
physical exertion, mental arithmetic, simulated public speeches, or affect
provocation). It is normal and typical that under the exact same stimulus
condition one can see an average increase of 10 beats in heart rate and
maybe an average of 10-mmHg (millimeters of mercury) increase of sys-
tolic and diastolic blood pressure (SBP and DBP, respectively) in a healthy
sample. Yet one not-infrequently observed extreme in response variabil-
ity can be an actual drop in heart rate and blood pressure (of −5 points)
under acute stress (relative to resting baseline). At the other extreme, heart
rate increases of +50 beats/minute, SBP increases of +40 mmHg and DBP
increases of +30 mmHg are also often observed in the most responsive
research participants, who, incidentally, may not show any signs of car-
diovascular disease at the time. Understanding the reasons for this remark-
able variability in subjective and in cardiovascular stress responsivity may
hold important clues for changing reactivity, and it may provide informa-
tion for what to build into the rationales for stress management.
For this reason, the current section is devoted to a description of the wide
range of variables that are known to mediate and moderate stress reactivity.
The reader should note that (similar to the above section on stress-disease
pathways) this section can serve only as an illustration and not an exhaustive
review of all prior research on stress reactivity. For more detail, the reader can
consult Lovallo’s (1997) book, or a special issue of Psychosomatic Medicine
where the status quo of the cardiovascular reactivity concept was placed
under a magnifying glass (Kamarck & Lovallo, 2003; Linden, Gerin, &
Davidson, 2003; Schwarz et al., 2003; Treiber et al., 2003). In the following
pages, known classes (or types) of moderating and mediating factors are
described, and many examples are given of how they interact to buffer against
stress consequences or how they may act to worsen stress responses.
Genetic Predispositions
On the surface, it makes sense that SM researchers and practitioners are

knowledgeable about but not overly interested in genetic predispositions for
exaggerated stress responses because genetics are not open to modification;
however, the same argument also underlines the importance of modifying
those risk factors that are actually open to change. When one is known to be
at risk, the subjective importance of modifying what is open to change ought
to increase. Especially individuals with a genetic predisposition for stress-
related diseases ought to have a keen interest in changing modifiable risk fac-
tors. Ewart (1991), for example, has presented a persuasive pathway model
of how genetics and environment interact in the etiology of hypertension. He
presents evidence that hostility aggravates the frequency and intensity of
stress activation and that hostility is at least in part transmitted within

families. This observation is supported by data on intra-class coefficients for
siblings’ personalities; among monozygotic twins the heritability coefficient
for personality is estimated as .50; this decreases to .25 for dizygotic twins,
and drops to .05 for adopted siblings. In sum, it is well established that
family history of hypertension increases risk of the disease, and it is paired
with typically heightened reactivity to stress. Personality, which is also partly
genetic, has been shown to predict physical disease (Booth-Kewley &
Friedman, 1987), and some personality features like hostility are predictive
of greater stress responses.
Early Learning
In addition to genetic predispositions for differential stress reactivity,

there is growing evidence that exposure to traumatic events early in life can
serve a similar predispositional role that leaves some individuals more sus-
ceptible to long-term stress effects. Kendler and his colleagues (2000), for
example, have investigated life event stress exposure and its predictive
power for development of major depression and generalized anxiety dis-
order. They assessed life events that included dimensions of loss, humilia-
tion, entrapment, and danger in a sample of 7,322 male and female twins
and determined depression and anxiety prevalence. Not only did life events
indeed predict prevalence of affective disturbance, but distinguishable
event dimensions had specific consequences in that loss alone was not very
predictive of long-term affect, but loss paired with humiliation was partic-
ularly predisposing for depression whereas loss and danger experiences
jointly were more likely to lead to elevated anxiety.
The Kendler et al. study is one of many that provide compelling epi-
demiological evidence that adverse experience during childhood increases
the likelihood of alcohol and drug dependence, eating disorders, affective
disorders, posttraumatic stress disorder, and suicidal behavior (for a review
see Surtees et al., 2003). Although epidemiological work can provide only
“surface” descriptions of relationships, other researchers have extended this
work by studying biological pathways, using animal models that center
around maternal separation and abandonment stress (Caldji et al., 2001;
Meaney et al., 1996). These researchers have shown consistent magnifying
effects of early stress exposure on the nature of the hypothalamic-pituitary-
adrenal axis response to stress. Similar impact of early stress exposure
on cell-mediated immunity and subsequent survival has also been shown
in rhesus monkeys (Lewis, Gluck, Petitto, Hensley, & Ozer, 2000), MRI
technology has permitted showing how early trauma affects brain

morphology in humans (DeBellis, 2001), and an autopsy study of Japanese
children has confirmed adverse early trauma sequelae in terms of compro-
mised immune function (Fukunaga et al., 1992). Using health care use as an
index of the cost of stress, Biggs, Aziz, Tomenson, and Creed (2003) have
also demonstrated that childhood adversity was an independent predictor
of health care use in functional gastrointestinal disorders.
Personality
A connection of personality (i.e., enduring patterns of behavior) and

disease may come about via different, typically indirect pathways. One
such influence pathway sees certain personality types as translating into
response predispositions that ultimately affect physiological regulatory sys-
tems (Schwartz et al., 2003). Another potential pathway is via stable indi-
vidual differences in magnitude and frequency of stress responses (details
follow in a later section), and a third one is via the influence that person-
ality has on influencing other risk factors for disease.
Small but significant, simple linear relationships of personality predic-
tors of blood pressure exist and have been found for trait anger/hostility,
anxiety, depression, and defensiveness (Rutledge & Hogan, 2002). Rutledge
and Hogan conducted a meta-analysis of studies that measured personality
features and that also studied blood pressure change with follow-ups of
1 year or longer (averaging 8.4 years). Significant r values between .07
and .09 were observed for anger, depression, anxiety, and defensiveness,
with defensiveness being the least often studied but overall strongest
predictor.
Finally, there is evidence for linkages of personality factors to tonic phys-
iological indices, which themselves are disease predictors. Using ambulatory
blood pressure monitoring, Linden and his collaborators (Linden, Chambers,
Maurice, & Lenz, 1993) showed that low social support in women, and high
hostility and defensiveness in men, were associated with elevated levels of
blood pressure even after the statistical effects of traditional risk factor effects
had been partialled out. They also observed that high social support was neg-
atively correlated with hostility, thus indicating a link between a personality
factor and a stress-buffering feature. Further, in a randomized, controlled
clinical trial of psychotherapy for hypertension, those patients with the great-
est hostility reductions and those with improvements in use of constructive
anger expression behavior showed the greatest blood pressure reductions
during treatment (Linden, Lenz, & Con, 2001).
Miller and his collaborators (Miller, Cohen, Rabin, Skoner, & Doyle,
1999) assessed major dimensions of personality and tonic cardiovascular,
neuroendocrine, and immunological parameters in 276 healthy adults.
While neuroticism was generally unrelated to any physiological function,
low extraversion was associated with higher blood pressure, epinephrine
and norepinephrine, and natural killer cell activity. Low agreeableness
(which is conceptually similar to hostility) was positively related to higher
systolic and diastolic blood pressure and epinephrine. The magnitude of
personality and physiology intercorrelation was small, accounting for no
more than 7% of the variance. Interestingly, health practices that are pre-
sumed to represent one possible path for stress leading to disease did not
mediate the association between physiology and personality.
Miller et al.’s findings map well onto results of Denollet and his
collaborators (Denollet, Sys, & Brutsaert, 1995; Denollet et al., 1996),
who have shown that a novel personality construct (coined “Type D,” and
consisting of social introversion and emotional inhibition) is highly pre-
dictive of cardiac death in cardiac patient populations. When the Type D
construct was used to predict acute reactivity in the laboratory, the over-
all Type D construct did not predict cardiovascular reactivity per se
(Habra, Linden, Andersen, & Weinberg, 2003). However, the two sub-
factors of Type D were independently predictive of differential cardiac and
endocrine reactivity in a harassing laboratory paradigm; this was particularly
true in men (Habra et al., 2003).
While most of the attention in the personality-disease literature has
been given to indices of affective distress (i.e., anxiety, depression, anger/
hostility), there is also a growing literature referred to as the “positive psy-
chology” movement that attempts to identify psychological traits that buffer
and protect from stress consequences (Lutgendorf, Vitaliano, Tripp-Reimer,
Harvey, & Lubaroff, 1999). Some of that attention has been directed at the
construct of sense of coherence (Antonovsky, 1979) and cannot be written
off as simply being the opposite of negative affect.
Sense of coherence (SOC) is akin to possessing a meaning, a sense of
purpose, a positive spiritual strength. In a sample of older adults who were
about to relocate, SOC played a significant mediational role in buffering
against the stress of relocation that was indexed by natural killer (NK)
cell activity (Lutgendorf et al., 1999). This study compared healthy older
adults about to move with a matched control group that was not moving,
and found that poorest NK activity was seen in “movers” with a low sense
of coherence. Fournier, de Ridder, and Bensing (1999) have studied the
role of optimism in coping with multiple sclerosis; these researchers found
optimism incorporated three distinct subfactors, namely outcome

expectancies, efficacy expectancies, and unrealistic thinking. Unrealistic
thinking was clearly related to mobility restrictions and was considered
maladaptive; outcome and efficacy expectancies explained depression but
were unrelated to mobility. The presence of optimism and social support
were independent predictors of good physical outcomes in cardiac patients
during their rehabilitation phase (Shen, McCreary, & Myers, 2004).
In sum, there is a growing literature that supports personality factors
explaining small amounts of variance in predictor models of acute stress
reactivity and in the etiology of stress-related diseases. Both stress-increasing
and stress-buffering personality features have been identified. Although
such linear independent contributions of personality to stress-related
health indices have been established, current theoretical models place more
emphasis on interactive models such that personality may exacerbate
responses to acute stress and maintain chronic stress. Personality plays a
more potent role in disease development when it is seen as a response pre-
disposition that has its full effect when it is paired with the presence of
environmental triggers that activate its hyperarousal propensities.
Stressor Exposure and Stress

Reactivity as Predictors for Disease
The relationships between various stress response system markers,

types of stressors, and response mediators and moderators have been exten-
sively studied (for a review see Lovallo, 1997), and only summaries can be
provided here. The trend in this literature is to move away from simple,
direct cause-and-effect models and study multiple interacting systems and
their short- and long-term adaptations. There is little doubt that complex
models map much better onto the observed data on biological pathways
and that, as a by-product, it becomes ever more challenging for researchers
to advance the field because the best studies are the ones that broadly cap-
ture multiple response systems, predispositional factors, and consider both
short- and long-term psychological and physiological adaptation.
The cascade of physiological events in response to a stressor described
above includes activation of a multifaceted cardiac and hemodynamic
response cluster that, however, would not likely be of any health conse-
quence if recovery following activation was swift and complete (Linden,
Earle, Gerin, & Christenfeld, 1997; McEwan & Stellar, 1993). Hence it
was necessary for stress researchers to show how short-term activation can
lead to long-term changes that are deleterious in nature.
A full understanding of pathways for stress leading to disease requires

that researchers show how existing, adaptive self-regulatory systems change
(or get “corrupted”) by stress. A good example of a regulatory system that
is critically affected by chronic stress, and that shows compensatory and ulti-
mately harmful “adaptation,” is the baroreceptor control system for blood
pressure regulation. Baroreceptors are pressure sensors found in the walls of
blood vessels that serve a critical role in the feedback system of brain-heart
interactions. When pressure rises in response to a demand, the increased
blood pressure is detected by the baroreceptors, which inform higher corti-
cal centers that a fight-flight response has taken place. If the brain interprets
a stress response as no longer needed, it reduces cardiac activation and the
baroreceptors contribute by feeding information back to the brain that the
down-regulation process is, at some point, complete. In essence, this system
functions like a thermostat in a home, telling the “controller” that the heat
activation can be stopped because a heat comfort level (or threshold) has
been reached. Such systems have a set point that the regulatory systems are
trying to maintain. However, if the demand becomes chronic and no brain
signal for recovery is activated, then the baroreceptors react by actually
changing the set point or threshold so that now a higher baseline or tonic
level is considered the desired target. It is critically important to note that
such set points are relatively stable but not absolutely resistant to change and
resetting. Both animal (Dworkin, Filewich, Miller, & Craigmyle, 1979) and
human studies (Elbert, Pietrowsky, Kessler, Lutzenberger, & Birbaumer,
1985) have provided supportive evidence for a resetting phenomenon that
raises target levels and ultimately maintains higher blood pressure levels; that
is, an upshift in tonic blood pressure occurs when no signals of stress reso-
lution are fed back to cortical control centers.
Critical individual difference factors that mark higher risk for stress
reactivity are familial history of hypertension, greater exposure to acute
stressors, a propensity to show exaggerated acute responses to acute stres-
sors, and a lessened ability to recover quickly (Linden et al., 1997; Roy,
Kirschbaum, & Steptoe, 2001; Schwartz et al., 2003; Stewart & France,
2001; Treiber et al., 2003). Folkow (1982) has shown that individuals with
a positive family history of hypertension have higher vascular resistance that,
when activated, enhances cardiac responsivity to a stressor (Light, 1987). It
is important to note that both initial reactivity to a stressor and recovery
speed are individual difference factors that are relatively stable dispositions
over time (Burleson et al., 2003; Frankish & Linden, 1996; Rutledge,
Linden, & Paul, 2000), thus granting the candidate potential to play a
significant role in pathogenesis.
Frequency and magnitude of acute stress responding is exaggerated

and recovery from acute stress is slower in hostile individuals (Earle et al.,
1999; Suls & Wan, 1993) and in defensive individuals (Rutledge &
Linden, 2003); this finding is particularly true for men. A defensive per-
sonality style predicted blood pressure change over 3 years in a sample of
125 research participants, as did initial blood pressure hyperreactivity;
when both variables were entered into a mediational model, it could be
shown that individuals with high defensiveness and high initial reactivity
also showed the relatively greatest blood pressure change over time
(Rutledge & Linden, 2003). Similarly, the additive effects of three risk pre-
dictors (i.e., family history of hypertension, initial hyperreactivity to stress,
and acutely high stress levels) represented a much greater odds ratio for
hypertension development than either predictor alone, or than a combina-
tion of two predictors (Light et al., 1999).
The importance of personality style as a mediating factor in stress
effects can be seen in a prospective study of 166 young adults who were
studied over a 2-year period (Twisk, Snel, Kemper, & van Mechelen,
1999). The researchers tracked changes in daily hassles, life events, and
behavioral and biological risk factors and assessed how changes were
interlinked. The results showed that increases in daily hassles were predic-
tive of a worsening lipid profile, decreased physical activity, and increased
smoking behavior. All of these connections were particularly strong in par-
ticipants with a “rigid” personality style.
A particularly useful model of the synergistic results of acute stress
effects when superimposed on chronic strain is the study of caregivers
to elderly Alzheimer’s patients (Vitaliano, Zhang, & Scanlan, 2003).
Caregiving itself is considered a chronic stressor, and the additional effects
of vulnerabilities and resources on preclinical and clinical disease states can
be studied in this model. A graphical display of the factors involved in this
path model may facilitate the explanation of critical interrelationships.
While most concepts in this model are self-explanatory, the term metabolic
syndrome should be more clearly specified here. It refers to an intercorrelated
cluster of risk factors that include elevated glucose levels, lipid levels, insulin
activity, and obesity risk inherent in genetics and sedentary lifestyle. Although
the sheer number of arrows in this pathway model (see Figure 1.6) gives the
appearance of “conceptual clutter,” the complex interactions suggested in
Vitaliano et al.’s model are well justified in light of the evidence on risk
factor interactions.
A review of the intricate relationships and relative contributing
weights of these predictors for disease has been undertaken in a major
review (Vitaliano et al., 2003) where it is shown that the presumed chronic
stress of caregiving alone is not a sufficient predictor for disease. Chronic
stress, however, accentuates and sets the stage for many risk aggregations.
Caregivers of chronically ill patients, for example, are more likely to have
poor health habits that contribute to metabolic syndrome. One can also
expand or redraw this model by thinking of disease itself can as a “chronic
stress platform” that accentuates more disease (Vitaliano et al., 2003).
Given the typically advanced age of many caregivers, they themselves may
be ill, and caregiver samples can be subdivided into those with and with-
out history of heart disease or cancer. Caregivers who themselves were ill
had worse health habits than had those without heart disease, for example,
and they also reported fewer uplifting life events (Vitaliano et al., 2003).
Suggestive evidence for one discrete pathway linking acute and chronic
stress effects in caregivers comes from von Kaenel, Dimsdale, Patterson,
and Grant’s work (2003) on blood coagulation. Caregivers with high addi-
tional life stressors (assessed via structured interviews) showed poorer
hemostatic function than did caregivers with the same level of caregiving
demand and other medical risk factors but without the acute stressor
exposure.
Social
Vulnerability
Resources
Personal
Acute Distress
Resources
Metabolic Coronary
Syndrome Disease
Poor Health
Chronic Stress
Habits
Figure 1.6 A Mediational Model for the Role of the Metabolic Syndrome
Another interesting model of study for synergistic effects of high stress

exposure and other risk factors is that of job strain. Job strain (defined as
being in a high-demand and low-control job) was found to cluster with
negative affect, reduced levels of social support, and a preponderance of
negative sentiments towards coworkers (Williams et al., 1997). Similarly,
high job stress seen in firefighters has been shown to negatively affect their
overall stress responsivity and lead to greater alcohol consumption (Murphy,
Beaton, Pike, & Johnson, 1999), thus showing an indirect pathway for stres-
sor exposure to lead to negative health outcomes.
The study of stress reactivity has been of great interest because
controlled stress provocations are considered to provide a window into
real-life stress and its consequences (Linden et al., 2003). Although stress
reactivity research has flourished (especially in the cardiovascular arena),
there is growing but still imperfect evidence for its predictive power in
cardiovascular disease etiology (Schwartz et al., 2003). The predominant
models of reactivity research have been challenged as representing too
small a window because they pay little attention to recovery (Linden et al.,
1997) despite many promising studies that elucidate the critical role that
recovery plays in understanding how stress can lead to disease.
Epidemiological as well as acute provocation studies provide intri-
guing insights. In one of the largest studies on lifestyle changes to date,
researchers have also studied the health effects of vacation as a type of
planned recovery (Gump & Matthews, 2000). A sample of 12,338 patients
who had completed a lifestyle program was followed for 9 years and the
effects of holidays on mortality and morbidity were studied. The researchers
reported a roughly 30% reduction in cardiac mortality and morbidity for
those who took regular annual vacations.
Sleep is also worth studying in the recovery context given that is con-
sidered a primordial element in the body’s “program” to seek low activation
states and to facilitate recuperation from fatigue. Research on sleep is ham-
pered by the difficulty of measuring sleep quality because reliance on self-
reported data is problematic and acquisition of hard data by studying large
numbers of individuals in sleep laboratories is often prohibitively expensive.
The result is a general lack of trustworthy data. Fortunately, more reliable
findings have become available in the form of a prospective study with 185
healthy older adults who provided sleep-lab data and whose mortality rates
were then studied during a mean follow-up length of 12.8 years (Dew et al.,
2003). Defining sleep quality via use of EEG data, these researchers showed
a 2.13:1 mortality risk ratio for those with long sleep latencies (i.e., an index
of poor sleep quality) after controlling for coexisting medical burden.
Elevated risk (1.94:1) was similarly apparent for lower sleep efficiency. In the
same vein, Akerstedt et al. (2004), in a sample of 5,720 healthy employed
men and women, have shown that disturbed sleep was a stronger predictor
of fatigue than workload or lack of exercise.
Evidence Linking Stress and Stress-Related

Risk Factors to Specific Disease Pathways
The previous sections were partially designed to describe linkages of

single stress-related factors, in direct influence models, to disease etiology
and to health maintenance at large. This effort revealed limited evidence for
direct influence models; in its stead, most of the previous sections ended
with demonstrations of how individual difference factors in stress respond-
ing relate to one another and how their aggregation and interactions aug-
ment health risks. Although previous sections focused on descriptions of
generic pathways from stress to disease, the following section is designed to
provide more specific evidence for the usefulness of the complex pathway
models to describe specific disease outcomes and its unique predictor sets.
A 3-year prospective cohort study of 10,432 Australian women revealed
that perceived stress was an independent predictor of new diagnosis of
symptomatic coronary heart disease even if traditional risk factors had been
accounted for. The associated risk (odds ratio [OR]) was 2.4:1, which
exceeded the risk ratios observed for other risk factors. Note that an odds
ratio (OR) of 2.4:1 means that an individual who carries a certain risk is
2.4 times more likely to develop the disease than one without the risk factor
present. Perceived time pressure was not a significant predictor, whereas
lack of social support (OR 1.4:1), body mass index (OR 1.9:1), poor nutri-
tion (OR 2.1:1), diabetes (OR 1.9:1), and hypertension (OR 1.7:1) were.
Alcohol intake and a physically active lifestyle, on the other hand, served as
protective factors (ORs 0.4:1 and 0.6:1, respectively).
Also, Eaker, Pinsky, and Castelli (1992) reported outcomes from a
20-year follow-up of 749 healthy women from the Framingham study. Low
education, high perceived tension, and no vacations predicted MI or cardiac
death after controlling for age, smoking, blood pressure (BP), diabetes, gen-
der, cholesterol, and obesity. Although a study like Eaker et al.’s cannot tell
us what the pathway to disease was, there was strong experimental support
that at least part of the influence of stress on health is via the influence that
subjective stress has on unhealthy behaviors. Based on a survey of 12,000
individuals in 26 different worksites, Ng and Jeffrey (2003) reported that
self-reported high stress was consistently linked (for men and women alike)
with higher-fat diets, less frequent exercise, increased smoking, and greater
number of relapses from being an ex-smoker.
Support for the needed demonstration of pathways can also be derived
from controlled investigations of cardiac dysfunction under mental stress
showing reduced supply of blood to the heart (Jiang et al., 1996; Krantz
et al., 1999; Rozanski et al., 1988), reduced ability of blood to clot
(Grignani et al., 1992), dysfunction in blood vessel walls (Ghiadoni et al.,
2000), and reduced blood flow at sites with atherosclerotic plaque deposits
(Yeung et al., 1991). Recent evidence on the linkage between inflamma-
tory processes, stress, and cardiovascular disease has been diligently
reviewed by Black and Garbutt (2002) and suggests a complex but empir-
ically well-documented pathway from stress to cardiovascular disease via
causation and/or aggravation of inflammatory processes. These authors
argue that the inflammatory process is contained within the acute stress
response, and—up to a point—stress responding is described as adaptive
within the definitions of the fight-or-flight response. Hence, any psycho-
logical intervention that may reduce inflammation brought on by chronic
stress also has considerable potential for preventing cardiovascular disease.
An additional pathway for linking cardiovascular health outcomes
to psychological factors has been described in a series of programmatic stud-
ies that link specific cardiovascular changes to loneliness (Cacioppo et al.,
2002). Cacioppo and his collaborators showed greater age-related blood
pressure changes and poorer sleep in lonely than nonlonely older adults.
Evidence also links stress to immune dysfunction. Cohen et al. (1998)
inoculated 300 volunteers with the common cold virus and monitored
them for symptoms of illness. Participants who reported exposure to stres-
sors lasting more than 1 month developed colds at a two to three times
greater rate than those with low stress. Similarly, it has been demonstrated
that stress negatively affects duration and intensity of illnesses such as her-
pes, hepatitis B, meningitis C, and human immunodeficiency virus (Cohen,
Miller, & Rabin, 2001; Herbert & Cohen, 1993; Kemeny, Cohen, Zegans,
& Conant, 1989). Social stress appears to play a distinctive vulnerability
role in outbreaks of latent herpes viruses (Padgett, Sheridan, Berntson,
Candelora, & Glaser, 1998). Owen and Steptoe (2003) tested the rela-
tionship of acute mental stress, immune and cardiovascular function in
211 middle-aged adults. Independent of other risk factors, high heart rate
reactivity was associated with plasma interleukin 6 and tumor necrosis
alpha whereas heart rate variability (an index of the heart’s ability to
adjust to varying needs) was not associated with immune function. Given
that differences on absolute levels of heart rate had different effects than
variations in heart rate variability, Owen and Steptoe concluded that indi-
vidual differences in sympathetically driven cardiac stress responses are
associated with compromised immune function, hence giving support to
the rationale that stress management should target sympathetic arousal
reduction.
While research on cardiovascular adjustments to acute, contrived
stressors suggests that initial hyperreactivity is a useful predictor of long-
term hypertension development (Treiber et al., 2003), the lab reactivity
paradigm is not as useful for the study of acute stress and immune func-
tion. Segerstrom and Miller (2004) culled data from 300 studies and
concluded that exposure to brief, contrived stressors was typically associ-
ated with adaptive up-regulation of immune function, in particular for
natural killer cell activity. Only chronic stressors were associated with sup-
pression of both cellular and humoral measures of the immune system.
These findings suggest that time-limited stressor exposure may actually
serve to strengthen immune function and entail a vaccination-type effect.
Interestingly, it is not necessary to document quantifiable, salient envi-
ronmental stress triggers to show a link with disease vulnerability; mere
perception of high stress was associated with lowered antibody production
after vaccination (Burns, Drayson, Ring, & Carroll, 2002).
Despite limited evidence for the reliability of stress self-report due to
contextual factors and individual differences (for more detail see section titled
“‘Take-Home Messages’ That Are Pertinent to Stress Management,” below),
self-reported, perceived stress is consistently related to certain stimulus
environments and has been shown to associate with other risk factors.
Perceived stress (a) is higher in some occupations than others (low decision
control, high demand), which themselves carry differential risks for disease
(Williams et al., 1997); (b) is higher in low socioeconomic strata; (c) is asso-
ciated with greater consumption of tobacco, higher relapse in previous smok-
ers, worse diets, reduced physical activity, and poorer sleep (Cartwright et al.,
2003; Ng & Jeffrey, 2003); and (d) contributes to physical inactivity and
poorer diet thus worsening lipid profiles and contributing to the metabolic
syndrome.
Research on cancer progression has shown mixed evidence for the role
of stress but does elucidate a role for indirect linkages: Stressed cancer
patients showed decreases in healthy behaviors and increases in unhealthy
behaviors (like smoking), which in turn may affect cancer progression
(Lacks & Morin, 1992). Another interesting development in immune
research has been the discovery that not only does the brain regulate
immune function, but the immune system itself triggers brain activity by
alerting the brain to infection or injury by releasing a protein called pro-
inflammatory cytokine (Maier & Watkins, 1998). This protein triggers a
cascade of responses, including fever and listlessness, that in turn are held
to serve an adaptive function by reducing energy output.
Earlier in this book there was mention of Selye’s belief in the universality
of a fight-or-flight response to challenge. This belief had to give way to the
recognition that there are other possible types of responses. A particularly
intriguing suggestion of a different response type is that of energy-preserving
response (Maier & Watkins, 1998) that appeared to explain Kemeny and
Gruenewald’s (2000) results; these researchers have shown a cognitive equiv-
alent to a proinflammatory cytokine activation in that HIV-positive men with
rather optimistic outlooks developed AIDS symptoms less quickly than did
those with negative though realistic aspirations.
While by no means exhaustive or comprehensive, the above discus-
sion makes a strong case for the shared pathways of stress, genetics, and
behavioral risk factors for physiological dysregulation and diseases of the
cardiovascular, immune, and endocrine systems. One can readily see how
stress reduction would affect other behavioral and biological risk factors
and disease outcomes; these relationships can be made apparent by a
more visual display. In Table 1.1, evidence is summarized about which
risk factors predict which disease. Note that this table was selected for its
illustrative value, not because it reflects the latest evidence. At the time of
writing this book, the evidence has actually grown stronger.
Each x in the table signifies that, according to the Center for Disease
Control (USDHHS, 1986), scientific evidence for a consistent link of risk
to a specific disease outcome exists; it is striking how few cells in this table
remain empty. What this table cannot show is how risk factors themselves
are linked to one another and then cause synergistic effects on health. A
great deal of information regarding clustering of risk factors was described
above, and a particularly good demonstration of such clustering was
provided above in Figure 1.6, which in turn is backed by statistical, meta-
analytic review data (Vitaliano et al., 2003).
To summarize, a great deal of evidence links quantifiable stress (both
self-report and physiological indices) to disease markers (Bunker et al.,
2003). Although the term disease marker was consciously chosen to reflect
imperfect evidence of causality, the available knowledge makes a strong
case for a likely causal role of physiological exhaustion and lack of recov-
ery in disease development.
Table 1.1 Diseases Potentially Brought On by Risk Factors
Risk Factor Disease
Heart Disease Stroke Cancer Diabetes
Tobacco x x x x
Alcohol x
Cholesterol x x x x
Hypertension x x x
Diet x x x x
Obesity x x x x
Inactivity x x x
Stress x x x
Drug Use x x
Occupation x x x
SOURCE: Adapted from the U.S. Department of Health and Human Services, 1986.
“Take-Home Messages” That Are

Pertinent to Stress Management
Stress is well understood at the subjective, experiential level; it is simply

part of modern life (if one is allowed to use such a platitude). Yet it is dif-
ficult to define operationally and then to measure because it is a process
and not a state. This review of the literature revealed that the field has
steadily evolved by expanding and elaborating early models, and has
revealed much consistency in the observation of powerful and complex,
interconnected pathways for stress-disease linkages, showing mutual influ-
ences of the nervous, endocrine, and immune systems. Studies of short-term
as well as long-term exposure to stress, in animals and humans, confirm
the critical role of chronic stress in disease development even though
many of these pathways remain insufficiently understood (Kelly et al.,
1997; Lovallo, 1997). Both self-reported stress levels and physiological
markers possess usefulness in showing stress disease linkages, especially in

repeated measures designs.
Early models like Selye’s have in good part held up to scrutiny, and
more recent models have typically expanded on earlier thinking rather
than invalidating it. One notable exception to the claimed veracity of early
models is the criticism leveled against Selye’s belief that stress responses are
ubiquitous, “whole system” responses in which all physiological response
components operate in synchrony. One could wish that he had been right,
because that would make the measurement of stress a lot easier! However,
as knowledge of nervous system function has expanded, Selye’s belief in a
whole system response has required modification.
At least partly consistent with the general adaptation syndrome is the
later subdivision of the nervous system responses into activities along a
sympathetic-adrenal axis versus activation of a hypothalamic-pituitary-
adrenocortical axis (HPA), but the presumption that both are always oper-
ating in parallel and are of equal importance to disease development is no
longer held to be true (Dienstbier, 1989). The work of Frankenhaeuser
(1991), Haynes, Gannon, Oromoto, O’Brien, and Brandt (1991), and
Linden et al. (1997) make a strong case that physiological arousal of the
sympathetic axis is not likely disease-contributing unless it is paired with
substantial activation of the HPA axis and/or unless physiological recov-
ery is delayed. This also implies that researchers interested in studying
stress-disease linkages need to include measures of sympathetic activation
(e.g., electrodermal activity) as well as HPA activation (e.g., cortisol) and
use study protocols that are of sufficient length to permit adequate study
of poststressor recovery. Cortisol changes relative to resting baseline, for
example, may be demonstrable for as long as 1 hour poststressor even in
what are clearly contrived, relative minor stressor exposure paradigms
(Linden et al., 1998). Stress management should therefore target arousal
reduction skills to facilitate and accelerate recovery, as well as teach skills
to minimize initial reactivity that exceeds the biological/survival needs
inherent in a given challenge.
In sum, at this time we have a much clearer sense of how diverse stress-
related events and processes affect short- and long-term adjustments of the
cardiovascular system (Kop, 1999), the immune system (Cohen et al., 1998;
Sklar & Anisman, 1981), and the endocrine system (Dienstbier, 1989;
Frankenhaueser, 1991). On the whole, there is weak, at best mixed, evidence
that mere exposure to stressors is a sufficient trigger for disease, but there is
overwhelming evidence in support of interaction models, such that predis-
positions to hyperreactivity (e.g., via genetics, personality, prior exposure)
paired with acute challenges lead to exaggerated responses, slow recovery,

and sometimes exhaustion. The coexistence of these features in the same
person contributes strongly to disease development. In support of this now
dominant line of thinking, there is rapidly growing knowledge of how phys-
iological systems interact, and there is a solid understanding of how stress
can affect disease, possibly in a causal manner.
Knowing the effects of stress on multiple, pivotal regulatory physio-
logical functions, it can no longer surprise that stress can affect many dif-
ferent disease processes. Even if not clearly demonstrated to be causal, it
can at least be shown how maintenance and exacerbation of health prob-
lems can occur under high chronic stress (Lovallo, 1997; Segerstrom &
Miller, 2004). This knowledge can and has been applied to diseases of the
immune system (e.g., cancer, AIDS, lupus, multiple sclerosis, common
colds), endocrine dysfunctions (e.g., diabetes), and cardiovascular health
(e.g., high blood pressure, myocardial infarction, sudden cardiac death).
As such, stress has been shown to be important to almost all causes of
mortality and chronic disease processes.
Stress plays a critical role in understanding the impact of early trauma
responses on disease susceptibility. The magnitude of acute stress
responses and speed of recovery are also likely predictive of disease devel-
opment, although the research base in this area is still weak. Outcomes are
positively affected by the presence of buffers (physical fitness, presence of
support, an optimistic outlook, a sense of meaning, enjoyable activities)
and negatively influenced by the presence of chronic strain in everyday life,
negative mood, defensiveness, anger/hostility, anxiety, and depression.
There is evidence of gender differences in that men and women may
not benefit equally from buffers like social support (i.e., women tend to
benefit more) or be equally detrimentally affected by the presence of hos-
tility and rumination (in this case, men are more likely affected). Similarly,
reactivity to particular stressors may be gender- and population-specific.
In aggregation, this overview makes a convincing case that stress
reduction (however brought about) has far-reaching beneficial conse-
quences for physiological adaptation and health maintenance, and preven-
tion of exhaustion. Stress reduction benefits achieved in one critical
physiological system (e.g., the cardiovascular system) are likely to show
generalized benefits for other systems (e.g., strengthening of immune func-
tion). Creation of stress buffers (i.e., forces or personality characteristics that
protect against stress consequences) can also contribute to benefits across
many physiological functions. This observation implies that research on
stress reduction outcome can reveal benefits in physiological functions that
were not even the primary target of intervention and measurement, and
that broad, multisystem measurement is needed to uncover the full benefit
of stress reduction efforts.
If stress management is taught to initially healthy individuals, then it is
logically necessary to follow them for many, many years if researchers do
not want to miss out on seeing initially nonexistent, then slowly accruing
but potentially powerful benefits. Failing to see large immediate physio-
logical benefits of stress management training applied to healthy individu-
als does not imply failure because healthy individuals will begin training at
healthy levels of physiological function; detecting change is therefore diffi-
cult due to floor effects.
Not only do researchers need to consider long-term, parallel, interact-
ing, and counterproductive effects in multiple regulatory systems, they
should also differentiate potential health benefits in physiological resting
functions from those reflected in differential acute reactivity to a challenge.
This is a standard approach in studying immune system functionality and
stress-health linkage (Glaser, Rabin, Chesney, Cohen, & Natelson, 1999)
and provides unique opportunities to study outcomes. Lastly, documented
immediate stress reduction does not necessarily lead to long-term benefits,
and the researcher has a mandate to demonstrate lasting effects.
Given that the experience of stress has objective and subjective com-
ponents, and that pertinent physiological systems are interconnected and
interdependent, researchers can and should consider a wide range of
potential measurement targets. For example, Miller and Cohen (2001)
conducted a meta-analysis of the effects of psychological treatments on
immune function (for results, see Chapter 3) and reported that the various
studies had evaluated no fewer than 15 different markers of immune health!
Similarly, adaptive cardiac functions can be subdivided into resting mea-
sures as well as reactivity indices and may be represented in more than a
dozen endpoints, including systolic and diastolic blood pressure, pulse
pressure, pulse-transit time, blood volume pulse, peripheral resistance,
baroreceptor activity, time- and frequency-modulated aspects of heart rate
variability, and so forth. In addition, subjective distress reports can be
obtained from research participants, and dozens of standardized scales are
available to achieve this; they may include simple one-item rating scales, a
variety of standardized stress coping tools, self-report of negative affect, and
peer or clinician ratings. Although the section on stress measurement was
critical of self-report, this was not meant to discourage all self-report, given
existing evidence of its predictive validity when studied within the same
person. Finally, there are many behavioral expressions and consequences
of high stress loads that can be tapped, ranging from acute changes in
facial expressions of affect to return-to-work statistics or hospital emer-
gency visits. Researchers in stress management are urged to broadly assess
as many pertinent outcomes as possible. While this expectation greatly
complicates the life of the researcher and drives up the cost of research, it
also enhances the opportunity to show extensive, generalized benefits of
“hard” outcomes that can convince consumers and policymakers of the
value of psychological intervention (Linden & Wen, 1990).

Stress 1

Uploaded by

Document Informationclick to expand document information

Copyright:

Available Formats

Stress 1

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Stress 1

Uploaded by

Copyright:

Available Formats

01-Linden.

qxd 9/13/2004 9:07 AM Page 1

Stress is essentially reflected by the rate of all wear and tear

Scientific and Popular Definitions

Stress is a mediational process in which stressors (or demands) trigger

Stress: Definitions and Pathways to Disease 3

How Can Stress Be Measured?

Having a clear definition of a construct is a useful and necessary precursor

experience, primarily to serve as a validation check. This method has been

Stress: Definitions and Pathways to Disease 5

then workers’ compensation boards and various insurance companies would

Is there an answer to the question, “Can stress be measured?” In an

A History of Models for Stress and Health

Stress: Definitions and Pathways to Disease 7

prominently contributed to shaping this book and its objectives. Readers

Figure 1.1 The Selye Model of Stress

Stress: Definitions and Pathways to Disease 9

Stress-triggering qualities of stimuli will not be discussed first

1. Environmental stimuli that are judged to be of subjective importance

2. they stimulate the production of corticotrophic releasing factor (CRF)

3. In the first track (also commonly referred to as the sympathetic adrenal

4. changes the chemical structure of the messengers and releases adreno-

5. When ACTH reaches the adrenal glands, it initiates the production of

6. On the second track (commonly referred to as the hypothalamic-

7. At the level of the adrenal gland, this activation leads to release of

8. The adrenal glands also release norepinephrine, which then speeds up

This cascade of physiological responses to challenge involves intricate

Stress: Definitions and Pathways to Disease 11

oxygen. This relatively simplistic description of how activation can

Figure 1.2 The Stress Diathesis Model

Stress: Definitions and Pathways to Disease 13

(1) anticipation of physical pain or danger, (2) threat of social isolation or

of a pending test for which a student did not study, or a husband’s

Stress: Definitions and Pathways to Disease 15

The extent of challenges or demands on a system is, then, referred to

EVENTS OVER TIME

Figure 1.3 Repeated Hits Model

A real-world scenario that represents this type of situation would be

A third type of stressor situation with high “overload potential” is one

EVENTS OVER TIME

Figure 1.4 Lack of Adaptation Model

Beginning End Recovery Phase

Figure 1.5 Lack of Recovery Model

laboratory stressor studies (Earle, Linden, & Weinberg, 1999; Linden,

Stress: Definitions and Pathways to Disease 17

responses to heavy demands at the workplace, and the return home

Understanding Stress Responsivity

As any astute reader of scientific articles on stress reactivity will

On the surface, it makes sense that SM researchers and practitioners are

Stress: Definitions and Pathways to Disease 19

stress activation and that hostility is at least in part transmitted within

In addition to genetic predispositions for differential stress reactivity,

technology has permitted showing how early trauma affects brain

A connection of personality (i.e., enduring patterns of behavior) and

Stress: Definitions and Pathways to Disease 21

optimism incorporated three distinct subfactors, namely outcome

Stressor Exposure and Stress

The relationships between various stress response system markers,

Stress: Definitions and Pathways to Disease 23