Attachment and Psychosomatic Medicine: Developmental Contributions to Stress

and Disease
ROBERT G. MAUNDER, MD, FRCP(C), AND JONATHAN J. HUNTER, MD, FRCP(C)
Objective: The object of this study was to evaluate the evidence linking attachment insecurity to illness. Attachment
theory describes lifelong patterns of response to threat that are learned in the interaction between an infant and his
or her primary caregiver. Despite its biopsychosocial domain, attachment theory has only recently been applied to
psychosomatic medicine. Method: MEDLINE and PsychInfo databases were searched from 1966 to 2000 for English
language papers with key words “attachment” and “object relations.” Papers and their cited references were
reviewed if they were directly related to physical illness, symptoms, or physiology. A hypothetical causal model
was developed. Results: Direct and indirect evidence from survey studies supports an association between
attachment insecurity and disease. Animal studies and human experiments suggest that attachment contributes to
individual differences in physiological stress response. There is also less robust support for insecure attachment
leading to symptom reporting and to more frequent health risk behaviors, especially substance use and treatment
nonadherence. Evidence supports the prediction from attachment theory that the benefits of social support derive
more from attachment relationships than nonattachment relationships. Conclusions: Although the available data
are suggestive rather than conclusive, the data can be organized into a model that describe attachment insecurity
leading to disease risk through three mechanisms. These are increased susceptibility to stress, increased use of
external regulators of affect, and altered help-seeking behavior. This model warrants further prospective
investigation. Key words: Attachment, disease model, stress, development, biopsychosocial.

result of predicted and measurable individual differ-

ACTH ⫽ adrenocorticotropin; CSF ⫽ cerebrospinal
fluid; HbA1c ⫽ glycated hemoglobin; HPA ⫽ hypotha-
lamic-pituitary-adrenal; NE ⫽ norepinephrine; SOC ⫽
sense of coherence; UC ⫽ ulcerative colitis.
ences. The attachment model explains how repeated
crucial interactions between infant and caregiver re-
sult in lifelong patterns of stress-response, receptivity
to social support, and vulnerability to illness.

There is convincing evidence that social support (1),
depression, (2, 3) and stress (4) have important effects
on the progress of some diseases, although the rela-
tionship between these psychosocial factors and dis-
ease processes is complex and often inconsistent be-
tween studies (1, 2, 4). Furthermore, there is a growing
body of evidence describing a high prevalence of past
psychological trauma, especially sexual abuse, in per-
sons with a variety of physical conditions including
gastrointestinal disorders (5), fibromyalgia, (6, 7) and
pain syndromes (8, 9). One of the central questions for
any model that ascribes disease risk to relatively com-
mon events such as trauma, loss, isolation, or to the
ubiquitous experience of stress is how the model ac-
counts for individual differences in psychosocial
susceptibility.
This paper explores the possibility that attachment
theory could be extended to a biopsychosocial model
of development and health that explains inconsisten-
cies in the stress and social support literature as the
From the Department of Psychiatry, University of Toronto, Mount
Sinai Hospital, Toronto, Ontario, Canada.
Address reprint requests to: Dr. R. Maunder, Dept. of Psychiatry,
Mount Sinai Hospital, 600 University Ave, Toronto, Ontario, Can-
ada, M5G 1X5. Email: rmaunder@mtsinai.on.ca
Received for publication May 8, 2000; revisions received Novem-
ber 21, 2000.
ATTACHMENT THEORY
When Bowlby described attachment theory (10 –12),
he provided a biological basis for understanding close,
protective relationships. A central innovation of at-
tachment theory was the recognition that a child’s
desire for proximity to his or her mother is a biological
drive which has been selected in evolution, rather than
a behavior which is learned to satisfy other biological
drives such as hunger. If maintaining proximity is a
fundamental need, attachment behavior can be under-
stood as a set of strategies that have been learned to
achieve optimal proximity. An infant, defenseless on
its own, maintains proximity to her or his mother
through a complex system of communications and be-
haviors, which increase its chances of survival. At-
tachment behavior, such as smiling, vocalizing, crying,
and approaching, is the normal, adaptive response of a
mammal to threat. Organized patterns of attachment
behavior emerge at a relatively fixed time for each
species, in humans the second half of the first year
(13).
Before the emergence of organized attachment be-
havior (ie, before learning) infant signaling of needs is
innate and responsive to immediate stimuli. In this
pre-attachment period appropriate proximity relies on
parental anticipation of infant needs and sensitivity to
infant signals. In the transition to organized attach-
ment behavior, proximity is increasingly dependent

556 Psychosomatic Medicine 63:556 –567 (2001)

0033-3174/01/6304-0556
Copyright © 2001 by the American Psychosomatic Society
ATTACHMENT AND DISEASE
on patterns of approach and withdrawal that are rein-
forced and amplified by the parent. This operant con-
ditioning results in the laying down of procedural
memory in the infant. Procedural memory (learning
“how to”) is encoded learning which occurs develop-
mentally before the ability for conscious recall. Since
the neurological systems required for the laying down
of declarative memories will only be fully developed
by age 2 to 3, the behavioral patterns learned at this
critical stage are not available for conscious recall (14).
Attachment behaviors occur in patterns that show
substantial differences between infant-parent pairs but
occur in typical clusters. Infants can be reliably clas-
sified in an attachment typology based on Ainsworth’s
standardized Strange Situation in which the infant is
presented with stressful situations that include the
presence of a stranger, separation of the infant from his
or her primary caretaker, and reunion (13). Avoidant
infants cry minimally during separation, treat the
stranger similarly to the parent, and avoid contact on
reunion. Secure infants show distress at separation
and actively seek contact at the time of reunion, which
is soothing. Angry-ambivalent infants similarly show
separation distress and proximity-seeking on reunion,
but combine proximity-seeking with angry resistance
(13). The fourth category, disorganized attachment,
was created later to include subjects who present no
such coherent pattern.
The origin of these attachment phenotypes is
thought to reside in a complicated “goodness of fit”
between parent, child, and environment. Although
most attention has been paid to the parental contribu-
tions to attachment style it is also recognized that
parental attitudes are shaped by the temperamental
needs and abilities of the infant and by environmental
stresses and supports (15). Despite this complexity,
some generalizations are useful. In general, parenting
that is unresponsive relative to the needs of a child or
intolerant of a child’s distress predisposes to avoidant
attachment. The child learns that seeking proximity
through crying and clinging is futile. On the other
hand, independence is reinforced and valued. As a
result attachment behaviors are relatively deactivated.
Maintaining a degree of protective proximity depends
on avoiding communications of dependence (16).
Angry-ambivalent attachment typically follows
from interrupted or inconsistent parental care (17).
Confidence in oneself to respond appropriately to
threats does not develop adequately. The value of turn-
ing to the attachment figure for security is intermit-
tently reinforced, but the need to be vigilant for the
presence and loss of the other is more strongly rein-
forced. Proximity, once obtained, is often not soothing,
resulting in persistent anxiety.
Psychosomatic Medicine 63:556 –567 (2001)
Developmental Psychobiology of Attachment
The psychobiological basis of attachment has been
studied in animals, where it is possible to observe
extremes of attachment system failure. The first such
set of observations is of primates raised with various
degrees of maternal deprivation (18). The second is a
series of experiments with rats that has elegantly elu-
cidated the “hidden regulators” of infant physiology
that occur in the mother-infant interaction (19).
Brief separation of mammalian infants from their
mothers results in protest behavior, such as high-in-
tensity vocalization, agitated searching, and excessive
self-grooming (19). Persistent separation results in re-
duced responsiveness, decreased heart rate, and low-
ered body temperature often referred to as despair (18,
19). Monkeys that are kept in isolation through critical
periods of infancy demonstrate gross behavioral and
biological consequences. The behavioral conse-
quences of isolation rearing represent a disorganiza-
tion of innate social behaviors. What these animals are
missing, according to attachment theory, are the re-
peated experiences of parental amplification and rein-
forcement of innate behavioral patterns that result in
social norms becoming encoded in procedural mem-
ory (14). The biological consequences of isolation rear-
ing include abnormalities of neurotransmitter and hor-
mone function (18). Further consequences include
reductions in cerebellar (20) and motor cortical (21)
dendritic branching. These deficits illustrate the de-
gree to which normal brain development and normal
endocrine and neurotransmitter regulation are depen-
dent on the postnatal influence of crucial social
interactions.
The large conceptual gap between observations of
maternally deprived primates and the normal experi-
ence of parents and infants is partially bridged by the
description of a continuum of rehabilitation in de-
prived primates. In particular, Kraemer (18) has de-
scribed how the consequences of isolation rearing can
be partially compensated for by substitute mothers,
ranging from inanimate surrogates to peers to same-
species foster mothers. In a stepwise progression, the
degree of residual impaired function in isolation-
reared monkeys depends on the extent to which ma-
ternal functions are replaced by a surrogate and by the
age of the monkey at the time of replacement. Inani-
mate surrogates lead to increased attachment behavior
if the surrogates bear qualities such as feeling soft and
furry, warmth, mobility, and rocking motion. Inani-
mate surrogates will not suffice to correct the greatly
decreased CSF NE levels of the isolates, but peer sur-
rogates are partially effective. Peer surrogates rehabil-
itate behavior to a greater degree, such that isolated
557

monkeys can become indistinguishable from peers
when in familiar social situations. To achieve near
normal levels of CSF NE requires same-species adult
female surrogates. Importantly, none of these surro-
gates is able to provide isolates with the ability to
perceive social relationships, to respond well to social
cues or to behave normally when competing for re-
sources. Thus primate isolation rearing provides
strong evidence of both biological and behavioral ef-
fects of early social deprivation and of the partial plas-
ticity of these effects to rehabilitation by improved
social relationships.
Hofer’s rat model brings us closer to the regulatory
processes that occur between mother and infant under
more nearly normal conditions. The infant rat depends
on its mother as a regulator of a number of functions
critical to survival, including nutrition and regulation
of temperature and activity (19). This body of work
reveals two factors that are important to understanding
the association of attachment to health. First, the
mother is understood to be supplying necessary regu-
latory functions for the infant. Understanding that the
mother serves as an external regulator of the infant’s
physiology allows us to see that the loss of mother at
this critical stage directly constitutes a loss of regula-
tion. More than just a stress, maternal loss constitutes
an actual undoing of the system. Second, the nature of
these regulatory functions is highly specific and not
intuitively obvious. For example, infant activity is in-
creased by warmth (provided by close proximity to
mother) but decreased by tactile contact with mother.
Furthermore the degree of contact with and stimula-
tion by mother is regulated by infant activity, such that
the mother-infant pair taken together comprises a bal-
anced feedback loop. Hofer’s experiments demonstrate
the mechanisms by which proximity-maintaining
communications occur developmentally before the
feelings that are recognized, in older humans, as “se-
curity.” Attachment is first and foremost a biologically
based behavioral system.
The development of autonomous function requires
a transition from the interdependent mother-infant
regulatory system to relatively independent internal
self-regulation of physiological functions. This transi-
tion depends on the unfolding of genetically pro-
grammed neurological development under the con-
stant modification of social interactions (22). Thus, the
same processes of mutual engagement and interaction
that initially serve to maintain proximity as an end in
itself are the crucial prerequisites of both emotional
attachment security and physiological self-regulation.
These intertwined developmental processes ulti-
mately result in an organism that can independently
maintain homeostasis and mount a physiological re-
558
R. G. MAUNDER AND J. J. HUNTER
sponse to external threat and in a set of expectations
related to threat.
Expectations learned after the development of de-
clarative memory form a cognitive schema that defines
what sorts of events are threatening, how one should
react to threats, and what one can expect of others
under these circumstances. On the other hand, earlier
learning encoded in procedural memory does not form
a cognitive “expectation” but a nonverbal working de-
scription of what counts as a threat and what is done
when one is detected. Together, this set of expecta-
tions relating to threat, proximity, and separation are
called the internal working model.
Adult Attachment Patterns and the Internal
Working Model
Any relationship in which proximity to the other
affects felt security is an attachment relationship (23).
The primary example of adult attachment is between
committed sexual partners, but many other relation-
ships may include a degree of attachment, including
the relationship of patient and clinician.
In Bowlby’s theory, the internal working model pro-
vides the continuity of attachment style over time.
Models of self and other vary from very positive ex-
pectations (resiliency of self, responsiveness of other)
to very negative expectations (fragile or incompetent
self, unavailable or unreliable other). Since expecta-
tions of self and other are relatively independent (24),
an individual’s attachment status can be described by
their position on each of these dimensions, which has
led to a four-category, two-dimension model of attach-
ment (25).
In the four-category model a person who has posi-
tive expectations of both self and other is secure. Se-
cure attachment is associated with an internalized
sense of being worthy of care, of being effective in
eliciting care when required, and a sense of personal
efficacy in dealing with most stressors independently
(23). Secure individuals are described as adaptable,
capable, trusting, and understanding (26). In this de-
scription, we are choosing to emphasize aspects of
attachment that relate to security, threat, and stress
and to de-emphasize feelings related to loveability and
comfort. This is consistent with Bowlby’s psychobio-
logical model.
Insecure attachment can be categorized as preoccu-
pied, dismissing, and fearful. A person who expects to
cope with stress inadequately but whose expectations
of others are more positive is preoccupied. Preoccu-
pied attachment, the adult correlate of infant angry-
ambivalent attachment, is associated with excessive
care-seeking, separation protest, and fear of loss (23).
Psychosomatic Medicine 63:556 –567 (2001)
ATTACHMENT AND DISEASE
Although the preoccupied individual seeks care, the
soothing that results from contact is partial and tran-
sient. Preoccupied individuals are described as anx-
ious, dependent, emotional, impulsive, and approval-
seeking (26).
People who distrust the effectiveness of social sup-
ports but have a positive view of themselves empha-
size independence and are classified as dismissing. A
self-sufficient and undemanding attitude is often
highly valued, but the associated distrust and avoid-
ance of intimacy communicates the underlying inse-
curity. Situations that demand relinquishing control
and depending on others, such as hospitalization for
acute illness, may result in crisis. Dismissing attach-
ment is characterized by coldness to others and com-
petitiveness (25).
Finally, a person whose expectations of both self
and other are negative is fearful. Fearful attachment is
the least-studied category, having been introduced by
Bartholomew in 1991 (25). Fearful individuals are de-
scribed as cautious, doubting, self-conscious, shy, and
suspicious (26).
Dismissing and fearful attachment taken together
are the adult correlate of infant avoidant attachment.
Most studies do not distinguish between dismissing
and fearful attachment, often labeling the composite
category “avoidant.” To maintain consistency, the
term avoidant will be used to describe dismissing and
avoidant styles (except where studies specifically dis-
tinguish dismissing and fearful styles), whereas “pre-
occupied” will be used to designate the attachment
type that is variously referred to as preoccupied, anx-
ious, ambivalent, and resistant.
Studies of attachment types in different age groups
have demonstrated that the prevalence of avoidant
attachment is fairly constant in all age cohorts (about
20% to 25%) whereas the prevalence of preoccupied
attachment is less in older cohorts (from about 20% at
college age to 5% to 8% in middle age) (26, 27). Re-
cently, a 31-year longitudinal study in women sup-
ported the view that these changes represent trends in
individuals over time, rather than cohort effects, by
showing a modest decline in preoccupied attachment,
a similar increase in secure attachment, and stable
avoidance attachment scores over adult life (26).
The importance of the internal working model for
understanding attachment and health is that it clarifies
the dilemma of whether attachment insecurity should
be understood as an emotional state or a personality
trait. An attachment type is the result of an internal
working model that guides affects and behavior when
a threat is perceived. Presumably the neurological sub-
strate for this internal working model has developed
through the effects of highly salient conditioning in
Psychosomatic Medicine 63:556 –567 (2001)
the first attachment. Attachment type is then under-
stood as a disposition toward certain perceptions of
others, certain perceptions of self, and certain pre-
ferred strategies that will be triggered by the presence
of a perceived threat. The emergence of attachment
behavior is largely context dependent, a state phenom-
enon, but the internal working model that provides
consistency to the pattern of behavior that emerges is a
trait.
METHODS
MEDLINE and PsychInfo databases were searched from 1966 to
2000 for English language papers with key words “attachment” and
“object relations.” Peer-reviewed papers and their cited references
were included for review if they were directly related to physical
illness, symptoms, or physiology, and not included if they ad-
dressed only psychopathology. This literature was first reviewed for
evidence for and against an overall association of insecure attach-
ment and disease. Possible causal mechanisms linking insecure
attachment to disease were then reviewed. A causal model was
developed and the evidence for and against hypotheses contained in
the model was reviewed.
ATTACHMENT INSECURITY AND DISEASE
To assess whether there is an overall association of insecure
attachment and disease there are four relevant aspects of attachment
literature. The evidence for an association of attachment and disease
in childhood, an association in adulthood, an association of child-
hood precursors of insecure attachment and adult disease, and the
association of secure attachment and later health will be reviewed.
Is Childhood Attachment Insecurity Associated
With Childhood Disease?
The clinical observation that parents of ill infants experience
anxiety that might interfere with reliable and consistent parenting
has led to the study of attachment security in several diagnostic
groups. Remarkably, no excess of attachment insecurity has been
found in premature infants (28, 29), in preschoolers with develop-
mental delays (29), or in infants with cleft lip and palate, (30)
compared with either healthy controls or community norms.
On the other hand, toddlers with certain chronic illnesses have
been found to have a higher prevalence of insecure attachment.
Children with recurrent otitis media over six months exhibit more
attachment insecurity than children without recurrent otitis media
(31). Twenty-six children aged 12 to 22 months with failure to thrive
exhibited less secure attachment and more anxious, disorganized
attachment in the Strange Situation, regardless of organic or nonor-
ganic etiology of failure to thrive (32). Cystic fibrosis is associated
with a higher prevalence of disorganized insecurity than found in
healthy controls (33). Epilepsy is associated with a higher preva-
lence of insecure attachment than cerebral palsy or normal controls
(34). A higher prevalence of insecure attachment has also been
reported in asthmatic preschool children (35).
In congenital heart disease there is a significant excess of
avoidant insecurity in patients (37%) compared with healthy con-
trols (9%) (36). Of further interest, functional severity of heart dis-
ease was not different between secure and insecure infants at diag-
nosis but insecure infants were significantly less likely to show
improvement at follow-up at age 1 year.
559

Is Adult Attachment Insecurity Associated With
Adult Disease?
Adult attachment pattern can be reliably classified using the
Adult Attachment Interview, while dimensions of attachment secu-
rity are measured with a number of self-report instruments (37).
While adult attachment has been extensively studied in relationship
to psychopathology (38) there has been much less investigation of
the relationship between adult attachment and physical illness. In
idiopathic spasmodic torticollis, avoidant attachment, measured by
the Attachment Q-sort, is over-represented compared with normal
controls (39). In UC, avoidant attachment (compulsive self-reliance)
measured by the Reciprocal Attachment Questionnaire is more prev-
alent in a subgroup of patients who lack an antibody marker than in
a clinically similar subgroup of UC patients with the marker. This
finding is taken as tentative confirmation of a hypothesis of distinct
etiological paths in these two groups (40).
SOC is a construct that describes a psychological predisposition
to life stress response (41). SOC describes a person’s confidence in
their ability to understand and manage stimuli from their internal
and external environments and, thereby, cope effectively with stres-
sors. SOC is very similar in description to the secure end of the
expectation of self dimension of the two-dimension, four-category
model of attachment. An association of low SOC and coronary heart
disease risk has been found in white-collar men, but not in blue-
collar men (42).
Are Childhood Risk Factors for Attachment
Insecurity Associated With Adult Disease?
The best prospective test of the influence of insecure attachment
on adult disease would be to determine attachment security in
children and follow them until adulthood. Such studies are not yet
available in humans, despite the fact that some cohorts of children
observed in the Strange Situation are now old enough that their
adult health status could be investigated. There are animal studies
and less direct human studies, however, which address this
question.
In animal studies, premature maternal separation is an experi-
mental analogue of the naturally occurring parental behaviors that
contribute to attachment insecurity. In a strain of rats that is genet-
ically susceptible to stress-induced gastrointestinal ulceration, early
weaning results in significantly more stress-induced ulceration after
the pups have grown to maturity (43). Inconsistent results have been
reported in cancer-prone rats. In one study separation of rat pups
from their mothers for 20 hours per day for 18 days was associated
with longer survival time after inoculation of Erlich carcinoma (44).
In another study of the same animal model, rats that were com-
pletely weaned from their dams prematurely showed longer survival
after implantation of Erlich carcinoma if they were paired with a
littermate after weaning (45).
In humans, one useful strategy to examine this relationship is to
look at childhood events that are often precursors of insecure attach-
ment. Evidence supports the clinical observation that experiences of
trauma, neglect, parental loss, and separation increase a child’s risk
of insecure attachment (46). While attachment security is the result
of a complex interaction between a child and his or her caretakers
and environment, and not all traumatized children become inse-
curely attached, trauma can be understood as the most provocative
test of the resiliency of the infant-caretaker system. Study of the
potential precursors of attachment insecurity is important because
events like trauma and parental death can be reported retrospec-
tively with more reliability than reports of the subjective quality of
parent-child attachment. Studying the precursors of insecurity al-
560
R. G. MAUNDER AND J. J. HUNTER
lows causal inferences to be made about attachment in the absence
of prospective studies. A very large study of primary practice med-
ical patients provides strong evidence for an association between
precursors of attachment insecurity and adult diseases. The risk
factors assessed were psychological, physical, or sexual abuse, vio-
lence against mother, and living with family members who were
substance abusers, mentally ill, suicidal, or imprisoned. The disease
outcomes studied were ischemic heart disease, cancer, chronic lung
disease, skeletal fractures, and liver disease. In 9,508 patients, a
graded relationship was found between the number of retrospec-
tively reported childhood family risk factors and adult disease (47).
Is Secure Attachment Associated With Adult
Health?
If insecure attachment is associated with disease it should also be
true that secure attachment is associated with health. Support for
this association can be inferred from the 50-year prospective study
of a cohort of men recruited at university age (48). Adult attachment
was not explicitly measured at the onset of this study because the
concept had not been described in 1942. However, childhood
strengths were measured in several ways that taken together suggest
that secure attachment may be a predictor of physical health. Spe-
cifically, the cohesiveness of the home before age 20 and whether or
not the boy’s relationships with his mother and father were condu-
cive to trust, autonomy, and initiative were predictors of the man’s
health at age 50 (49). At age 60, early adult closeness to siblings
became the more significant predictor. Maturity of defenses at age 20
to 47 also predicts physical health at 60 (48).
MODEL OF INSECURE ATTACHMENT
CONTRIBUTING TO DISEASE
Because there is suggestive evidence from a range of studies
favoring an overall association of attachment insecurity and illness,
we wish to examine the possibility that insecure attachment con-
tributes to disease processes. In reviewing the literature, we have
developed a model describing paths by which insecure attachment
could affect the course of disease. Figure 1 describes potential psy-
chosomatic mechanisms that could account for a correlation of in-
secure attachment and disease. The model describes how insecure
attachment may be associated with disturbances of stress regulation
(Path 1), use of external regulators of affect (Path 2), and nonuse of
protective behaviors (Path 3). These possible paths organize the
literature review that follows. The theoretical reasoning behind each
path is followed by an empirical review.
Path 1: Insecure Attachment May Affect Stress
Regulation
Of the three proposed paths by which insecure attachment could
affect disease, the role of stress is of particular interest from a
physiological point of view. Attachment and stress are related de-
velopmentally because, on the one hand, the stress response has
been selected in evolution to be triggered by environmental threat
while, on the other hand, the function of the attachment system is to
increase security in the face of environmental threat. We propose
three ways by which attachment may determine individual differ-
ences in the stress response.
Attachment insecurity may increase perceived stress. The inter-
nal working model describes the evaluative processes by which an
individual deals with threat by negotiating proximity to trusted
others. Included in this evaluation is a determination of what con-
Psychosomatic Medicine 63:556 –567 (2001)
ATTACHMENT AND DISEASE
Fig 1. Model of hypothesized mechanisms by which attachment security could contribute to disease.
stitutes a threat, which may involve both qualitative experiential
features of the environmental situation and a threshold for triggering
attachment behavior. Preoccupied attachment involves a self-per-
ception of vulnerability, which may lead to a lower threshold for
activating attachment behavior. In one model of hypochondriasis
and somatization, preoccupied persons have acquired a sense of
personal vulnerability and vigilance so intense that interoceptive
sensation (normal perception of physiological operations) is per-
ceived as a potential threat (50). Avoidant attachment, on the other
hand, includes an attitude of interpersonal distrust that may result
in situations that require intimacy or interdependence (including a
situation of apparent “social support”) being perceived as
threatening.
Mikulincer and Florian (51) have studied the relationship of
adult attachment to perceived threat and distress under two condi-
tions of real-life stress. In 92 male military recruits undergoing four
months of combat training appraisal of the training in terms of threat
was more pronounced in ambivalent (preoccupied) persons than in
avoidant or secure persons. In a retrospective study of students who
had experienced missile attacks in the Gulf War, attachment type
had a significant effect on the intensity of perceived distress. All
insecure subjects reported higher levels of hostility and somatiza-
tion than secure subjects, while ambivalently attached (preoccu-
pied) subjects also reported more anxiety and depression distress
(52).
Attachment insecurity may affect the intensity or duration of the
physiological stress response. The stress response is highly regu-
lated at multiple physiological levels. It is reasonable to assume that
this complexity allows that central factors (such as insecure attach-
ment) may modulate the system partially, allowing not only for
individual differences in the initiation or noninitiation of the stress
cascade but also for modification of its intensity or duration.
In rats, an increase in maternal care elicited by separating an
infant mouse from the mother (“handling”) for a few minutes each
day in the first two weeks of life results in lifelong reduction in the
HPA response to stressors. If the separation is longer, mothers ignore
the pups on their return and the pups suffer a lifelong increase in the
reactivity of ACTH and cortisol to stress (53). Furthermore, a strong
correlation has been found between the amount of early maternal
licking and grooming and the amount of messenger ribonucleic acid
for glucocorticoid receptors that is expressed in the hippocampal
regions, which are critical for negative feedback on the HPA axis
(54).
There is also human experimental evidence to support the rela-
Psychosomatic Medicine 63:556 –567 (2001)
tionship of attachment insecurity to the intensity of the cardiovas-
cular and cortisol response to acute stress in children and in adults.
Sroufe and Waters (55) measured changes in heart rate in children
during the Strange Situation. They reported that all children show
heart-rate increases during separation, which remain elevated until
reunion with the parent. At reunion secure infants exhibited a return
to baseline heart rate in less than a minute. Both resistant (preoccu-
pied) and avoidant children exhibited elevations of heart rate much
longer into the reunion sequence in spite of clear differences in
behavior (with the avoidant children displaying very little distress).
Subsequent investigations have revealed a more complicated pic-
ture. Adrenocortical stress reactivity during the Strange Situation
has been reported to occur in ambivalently attached infants but not
in secure or avoidant infants (56). Salivary cortisol may be elevated
only in those insecurely attached infants who also display behav-
ioral inhibition (57).
In university students, a history of childhood bereavement and a
self-report of disturbed family environment each exert a main effect
on the cardiovascular response to a laboratory stressor. In the same
protocol, family disturbance also predicted an enhanced cortisol
response to the stress of viewing a movie depicting the death of a
boy’s mother (58). Similarly, women with a history of childhood
sexual abuse exhibit increase ACTH, cortisol, and heart rate re-
sponses to a standardized stress compared with women without
such a history (59). While these studies allow inferences to be drawn
about attachment insecurity, one study of stress response has di-
rectly measured adult attachment. In idiopathic spasmodic torticol-
lis (but not in healthy controls), avoidant attachment measured by
the Attachment Q-sort is associated with elevated salivary cortisol
response to the stress of discussing attachment relationships (39).
[Attachment pattern may determine the success of social support
in buffering stress.] Social support has been widely studied as a
mediator of illness and is considered to be beneficial to a range of
health outcomes (60), despite controversy about the details of the
association and its mechanism (61). Attachment theory may be re-
quired for a coherent reading of the complexities of the relationship
of social support and health. Attachment theory, for example, pre-
dicts that attachment relationships will provide “felt security” in the
face of threat or loss whereas other (nonattachment) relationships
will not. Given the close developmental relationship between at-
tachment style and stress response, it can be hypothesized that
relationships that provide “felt security” are also more effective in
buffering stress. West (62) has used this prediction to explain why
close, intimate, and confiding relationships buffered women against
561

the psychiatric sequelae of stressful life events in Brown’s Camber-
well study, when no such affect was found with other close affiliate
relationships, even to first degree relatives (63). Furthermore, attach-
ment theory predicts that the degree of buffering provided by an
attachment figure would be greater for secure individuals than in-
secure ones and greater in preoccupied insecurity than in avoidant
insecurity.
It is expected that attachment style will also influence the degree
to which social contact is sought at times of stress. Secure attach-
ment is thought to be associated with a greater willingness to seek
support and a related expectation that support will be available and
effective. Avoidant attachment is associated with a desire for inter-
personal distance and self-reliance (23), whereas preoccupied at-
tachment may be associated with less support-seeking because of a
fear of unavailability or rejection.
A longitudinal study of 60 recently widowed and 60 married
men and women supports the prediction that attachment relation-
ships are special, in that the emotional loneliness resulting from the
loss of an attachment partner cannot be compensated by nonattach-
ment social support (64). With respect to stress and health variables,
the ability of the proximity of an attachment figure to reduce phys-
iological stress responses has received support from both animal and
human studies. In guinea pigs the presence of a bonded partner is
more effective than the presence of another same species animal in
reducing cortisol response (65). In humans, partner support results
in lower cortisol reactivity to a laboratory stress in men, compared
with stranger support or no support (66). Family emotional involve-
ment has a stronger association to cardiovascular health behavior
than other forms of social support (67). Support from a spouse or
family member is more effective in long-term reduction of diastolic
blood pressure than other social support (68).
The hypothesis that different patterns of attachment are associ-
ated with different patterns of using social support has also been
studied. In students, both the perception of available support and
the extent to which support is sought at times of stress were greater
in secure subjects than in avoidant or ambivalent (preoccupied)
subjects (69). In 92 military recruits avoidant attachment was asso-
ciated with less support-seeking than secure attachment (51). Secure
women are more likely than avoidant women to seek emotional
support from partners in an anxiety-provoking laboratory experi-
ence (70). In 81 healthy young adults dismissing and fearful attach-
ment styles were associated with much less support seeking than
secure and preoccupied styles. Support-seeking in this study was
significantly mediated by the perception of available support (71).
Path 2: Insecure Attachment May Result in
Altered Use of External Regulators of Affect
The next proposed etiological pathway concerns regulation of
affect. Since insecure attachment results in deficits in internal affect
regulation (72, 73), it is expected that insecurity will be associated
with greater use of external regulators. A number of behavioral
strategies that are used to regulate dysphoric affect (to soothe, to
distract, or to excite) are also risk factors for disease. Included among
these are smoking tobacco, drinking alcohol, using other psychoac-
tive drugs, over-eating, under-eating, and engaging in risky sexual
activity. This path focuses on regulatory behavioral strategies, some-
what arbitrarily distinguishing these from the relationship factors
(social support) that have been described above.
Substance use. The relationship between adolescent coping, ad-
olescent-parent relationship, and substance use is fairly well estab-
lished, although the inferences that can be drawn regarding attach-
ment are generally indirect. Childhood abuse and trauma predict
adolescent smoking and alcohol abuse (74) and adult smoking, al-
562
R. G. MAUNDER AND J. J. HUNTER
cohol consumption, alcoholism, and drug abuse (47, 75). Direct
studies of attachment reveal that adults with avoidant attachment
drink alcohol to enhance positive affect (76). Adolescent substance
users value social conformity less than nonusers (74, 77, 78), which
is closely related to parental support (77) and strength of emotional
bonds to family and friends (74). Drug use in young adults is also
related to parent-child attachment (79).
Eating behavior. Patterns of food consumption are thought to be
related to developmental relationships, and food consumption has
physiological and affective regulatory effects. It is predicted, there-
fore, that insecure attachment will be a risk factor for disturbances of
eating behavior. Insecure attachment is associated with weight con-
cern and lower self-esteem in adolescents (80). Parental overprotec-
tion is associated with oral control and anorectic behavior in ado-
lescent girls (81). Eating behavior is also associated with smoking
and alcohol consumption (81). External regulation through food
intake has also been suggested as a mechanism responsible for
obesity. One study has provided confirmatory evidence (82), al-
though this has not yet been replicated with better-validated mea-
sures of attachment insecurity.
Sexual behavior. It is expected that attachment style will have a
strong influence on sexual behavior (83). Unfortunately, the litera-
ture currently available is quite modest. As with substance use,
social conformity and peer bonding have been associated with lower
sexual activity in teenagers (81). Unsafe sexual behavior has been
associated with impaired object relations (84).
Path 3: Insecure Attachment May Alter Use of
Protective Factors
In addition to an excess of health risk behaviors, insecure attach-
ment may also contribute to illness through the failure or nonuse of
protective factors. One of the most important protective factors,
social support, has been described above. Additionally insecure
attachment may affect treatment adherence.
Treatment adherence. Diabetes is an example of a disease in
which treatment adherence is crucial. Viederman and Hymowitz
(85) proposed a model of diabetic control in which “the effective-
ness of [the patient’s] engagement and participation in treatment” is
determined by early relationship experience. The experience of de-
veloping autonomy, mastery, and reliable warm support during the
toddler phase provides a positive body image and sense of self-
control that persists through life and allows for sensitivity to bodily
needs and for appropriate self-control at times of health crisis. In the
absence of such a sense of autonomy and flexibility, health crises
produce defensiveness, especially denial. The mobilization of mal-
adaptive defenses during a health crisis results in an inability to
autonomously monitor blood sugar and respond adaptively by ad-
justing insulin and diet, and the inability to benefit from the avail-
ability of supportive resources. Recently this proposition has been
empirically studied using HbA1c as a measure of diabetic control in
15 type 1 diabetics. HbA1c is an established biological index that is
elevated in poorly controlled diabetes. Subjects with avoidant at-
tachment had significantly higher levels of HbA1c than those with
secure or preoccupied attachment (86). While it is presumed that
this worse control is due to nonadherence, it is also possible that
increased glucocorticoids in avoidantly attached persons contribute
to insulin resistance.
Symptom reporting. Two studies directly support the link be-
tween attachment insecurity and symptom reporting. In students
who were making the transition from home to living at university, a
relationship of avoidant attachment and physical symptoms was
found, mediated by emotional self-control and emotion-focused
coping (87). In another study an excess of medically unexplained
Psychosomatic Medicine 63:556 –567 (2001)
ATTACHMENT AND DISEASE
symptoms was found in fearful and preoccupied Hepatitis C patients
compared with securely attached individuals with the same disease
(88).
There is also indirect support for this association. In university
students who report a low SOC there is a positive correlation be-
tween negative life events and physical symptoms (89). In addition,
low SOC was found to predict sickness-absences from work in
women, but not in men, over a four-year longitudinal study of
municipal workers (90). Further indirect evidence comes from stud-
ies of adults who were sexually abused in childhood who have more
pain complaints, at more sites and greater health care utilization
than nonabused controls (9).
DISCUSSION
This review finds supportive evidence for the hy-
pothesis that attachment insecurity contributes to
physical illness. A model is presented that describes
how attachment insecurity may contribute to three
mechanisms that increase disease risk: altered stress
physiology, increased use of external regulators of af-
fect, and altered use of health-protective behaviors.
Thus attachment style may be a predictor of stress-
vulnerability and disease risk for a number of diseases.
A rationale for differentiating between individuals
who are highly susceptible to stress, physiologically
and behaviorally, and those whose response to stress is
well-regulated is crucial for etiological research. Since
heterogeneity in stress susceptibility is likely to dilute
true stress relationships in population-based research,
this rationale should provide new research approaches
to the several diseases in which stress is thought to be
a risk factor, but for which research on this association
has resulted in inconsistent results.
The attachment paradigm provides a truly biopsy-
chosocial model of disease. The model describes how
repeated, critical interactions occur at key develop-
mental periods between a child and his or her care-
taker (each of whom come to the interaction with their
own set of more or less determined temperamental
needs and vulnerabilities). While the need for such
interactions and the basic “vocabulary” of interaction
is determined for a species over an evolutionary time-
scale, within this domain there is much flexibility
allowing an individual to develop patterns of behavior
and physiological response which provide the best fit
between the individual, his or her caretaker, and their
environment. The consequences of this developmental
setup are both physiological and behavioral, and help
to explain the wide range of expression of distress and
coping behavior that is found in humans.
This review has focused on the possibility that pat-
terns of attachment influence disease risk. It must be
recognized, however, that the relationship of attach-
ment to illness is very likely to be bidirectional. Illness
events are near-perfect triggers for the mobilization of
Psychosomatic Medicine 63:556 –567 (2001)
attachment behavior. Because illness events can rep-
resent loss, threat, isolation, and dependency, they are
precisely the types of stress that will initiate attach-
ment behaviors. The concept of the internal working
model helps to clarify the distinction between attach-
ment behaviors, which will often be elicited by illness
stresses, and attachment patterns, which are more en-
during. Nonetheless attitudes that indicate attachment
insecurity may well be easier to elicit when the attach-
ment system is mobilized during illness stress. Clearly
there is a need for prospective clinical studies and
controlled experimental studies to test the causal pro-
cesses that link correlated states.
We propose that the deleterious health effects of
attachment insecurity are not universal. Given the
ubiquity of stress and the high prevalence of insecure
attachment (35% to 40% of the general population
(27)), attachment insecurity cannot logically be a suf-
ficient condition for any moderately uncommon dis-
ease. Furthermore, it would not make evolutionary
sense for a highly conserved behavioral pattern (91) to
be a sufficient condition for major illness. It is more
persuasive that attachment insecurity contributes to
disease by adding risk as a “second hit” on a preexist-
ing vulnerability, such as would be provided by ex-
pression of a gene or by an infectious agent, which
provides the disease specificity.
The evidence reviewed here provides modest sup-
port for each of the hypotheses examined. Considered
together, the direct and indirect evidence that favors a
role for attachment in disease forms a coherent group
of literature that calls for careful evaluation. The
model that has been described and reviewed is com-
plex, but each of the several hypotheses that it gener-
ates seems to be open to be operationalized and tested.
The association of attachment insecurity (and its
precursors) and the physiological stress response has
been more thoroughly studied than the association of
attachment security and the behavioral risk factors that
we were able to identify. As a result there is more
robust evidence for the physiological impact of attach-
ment insecurity than for its behavioral impact cur-
rently, which may run counter to intuition. Investiga-
tion of the role of attachment insecurity in treatment
nonadherence and other health risk behaviors should
be a priority.
The model has significant implications. If attach-
ment insecurity contributes to disease course or onset,
then this may lead to treatment opportunities. One can
speculate that physiologic dysregulation which ac-
companies insecure attachment could lead, for exam-
ple, to testing of benzodiazepines, or modulators of
corticotrophin-releasing factors to reduce disease risk
in persons with a specific vulnerability to a particular
563

illness (for example those at genetic risk) and/or those
of a specific attachment type.
An attachment perspective might modify behavioral
interventions for external regulators of affect that are
also disease risk factors. Increasing treatment adher-
ence in avoidantly attached individuals, for example,
might be facilitated by strategies that enhance per-
ceived personal control and respect interpersonal dis-
tance, whereas interventions to reduce excessive help-
seeking in individuals with preoccupied attachment
would take a very different approach.
Several caveats must be emphasized. One is that the
association of attachment insecurity with major de-
pression (92) has been given insufficient attention in
the literature reviewed here. This association is impor-
tant to the study of physical disease for a number of
reasons. First, depression is often associated with a
worse course of physical illness (2, 3). Second, depres-
sion may, in some cases, be a direct consequence of
severe attachment disruption. Third, there is an impor-
tant overlap in the physiological features of attach-
ment insecurity and depression. Depression, chronic
stress, and insecure attachment share an association to
HPA dysregulation (54, 58, 93) in which it is concep-
tually unclear which entity is primary. Thus, the bio-
psychosocial concepts of depression and insecure at-
tachment cannot be cleanly separated. At a minimum,
the inter-relationship needs to be addressed by mea-
suring and controlling for depressive symptoms in
studies of attachment insecurity and disease.
It is also worth noting the differences between at-
tachment and two other psychological constructs that
have been invoked to explain the relationship between
distressing experience and physical symptoms. SOC is
descriptively similar to the secure end of the self-
appraisal dimension of attachment in the four-category
model. Although this observation has not been directly
tested, it is possible that there is considerable overlap
between these constructs. It remains problematic that
SOC captures only half of the attachment construct.
The “appraisal of other” dimension appears not to
overlap with SOC. This is important because the ap-
praisal of other dimension is crucial to understanding
the avoidant stance which appears to be important in
treatment adherence (86), in symptom reporting (87),
and in the ability to benefit from the buffering poten-
tial of social supports. Neuroticism has also been pos-
tulated to mediate between the experience of child-
hood abuse and adult disease (94). Neuroticism, which
describes a perceived inability to process or cope with
intense affect, is not identical with insecure attach-
ment. The neuroticism construct does not distinguish
between attachment and nonattachment relationships
in their ability to provoke and buffer stress. The dif-
564
R. G. MAUNDER AND J. J. HUNTER
fering predictions of a neuroticism model and an at-
tachment model in this regard are empirically testable.
Finally, methodological concerns limit confidence
in the original research reviewed in this paper. At this
early stage of development of this literature most of the
studies that have been published are positive studies,
which introduces a potential bias. In addition, with
the exception of one large general practice cohort (47),
two prospective studies (36, 48), and the experiments
(18, 19, 58, 65), these studies report on relatively small
samples and are cross-sectional in design. Thus no
conclusions can be drawn about the validity of the
model.
In an effort to promote rigorous testing of a series of
hypotheses that are important to psychosomatic med-
icine, the methodology of this review has been inclu-
sive and synthetic. It is hoped that this exercise in
hypothesis generation will encourage the development
of a literature robust enough to allow for review that is
exclusive and analytic in the future. In particular, pro-
spective and longitudinal research is required to con-
trol for bidirectional influences between (context-de-
pendent) attachment behavior and illness events and
to determine if attachment types which are construed
to be traits are, in fact, present as predisposing factors
before disease onset.
REFERENCES
1. Broadhead WE, Kaplan BH, James SA, Wagner EH, Schoenbach
VJ, Grimson R, et al. The epidemiologic evidence for a relation-
ship between social support and health. Am J Epidemiol 1983;
117:521–37.
2. Wulsin LR, Vaillant GE, Wells VE. A systematic review of the
mortality of depression. Psychosom Med 1999;61:6 –17.
3. Watson M, Haviland JS, Greer S, Davidson J, Bliss JM. Influence
of psychological response on survival in breast cancer: a popu-
lation-based cohort study. Lancet 1999;354:1331– 6.
4. Chrousos GP, Gold PW. The concepts of stress and stress system
disorders. Overview of physical and behavioral homeostasis.
Jama 1992;267:1244 –52.
5. Drossman DA, Talley NJ, Leserman J, Olden KW, Barreiro MA.
Sexual and physical abuse and gastrointestinal illness. Review
and recommendations. Ann Intern Med 1995;123:782–94.
6. McBeth J, Macfarlane GJ, Benjamin S, Morris S, Silman AJ. The
association between tender points, psychological distress, and
adverse childhood experiences: a community-based study. Ar-
thritis Rheum 2000;42:1397– 404.
7. Walker EA, Keegan D, Gardner G, Sullivan M, Bernstein D,
Katon WJ. Psychosocial factors in fibromyalgia compared with
rheumatoid arthritis: II. Sexual, physical, and emotional abuse
and neglect. Psychosom Med 1997;59:572–7.
8. Golding JM. Sexual assault history and headache: five general
population studies. J Nerv Ment Dis 1999;187:624 –9.
9. Fillingim RB, Wilkinson CS, Powel T. Self-reported abuse his-
tory and pain complaints among young adults. Clin J Pain 1999;
15:85–91.
10. Bowlby J. Attachment and loss. Vol 1: Attachment. New York:
Basic Books; 1969.
Psychosomatic Medicine 63:556 –567 (2001)
ATTACHMENT AND DISEASE
11. Bowlby J. Attachment and loss. Vol 2: Separation: anxiety and
anger. New York: Basic Books; 1973.
12. Bowlby J. Attachment and loss. Vol 3: Loss: sadness and depres-
sion. New York: Basic Books; 1980.
13. Ainsworth M, Blehar M, Waters E, Wall S. Patterns of
attachment: a psychological study of the strange situation. Hill-
sdale (NJ): Erlbaum; 1978.
14. Kandel ER. Biology and the future of psychoanalysis: a new
intellectual framework for psychiatry revisited. Am J Psychiatry
1999;156:505–24.
15. Belsky J, Rosenberger K, Crnic K. The origins of attachment
security: “Classical” and contextual determinants. In: Goldberg
S, Muir R, Kerr J, editors. Attachment theory: social develop-
mental, and clinical perspectives. Hillsdale (NJ): The Analytic
Press; 1995.
16. Main M. Avoidance in the service of attachment: A working
paper. In: Immelman K, Barlow G, Main M, Petrinovitch L,
editors. Behavioral Development: The Bielefeld Interdiscipli-
nary Project. New York: Cambridge University Press, 1981;
651–93.
17. Bowlby J. Making and breaking of affectional bonds: 1. aetiology
and psychopathology in the light of attachment theory. Brit
J Psychiatry 1977;130:201–10.
18. Kraemer GW. A psychobiological theory of attachment. Behav
Brain Sci 1992;15:493–511.
19. Hofer MA. Hidden regulators: implications for a new under-
standing of attachment, separation and loss. In: Goldberg S,
Muir R, Kerr J, editors. Attachment theory: social developmen-
tal, and clinical perspectives. Hillsdale (NJ): The Analytic Press,
1995.
20. Floeter MK, Greenough WT. Cerebellar plasticity: modification
of Purkinje cell structure by differential rearing in monkeys.
Science 1979;206:227–9.
21. Struble RG, Riesen AH. Changes in cortical dendritic branching
subsequent to partial social isolation in stumptailed monkeys.
Dev Psychobiol 1978;11:479 – 86.
22. Schore AN. Affect Regulation and the Origin of the Self. Hills-
dale, N.J.: Lawrence Erlbaum Associates 1994.
23. West ML, Sheldon-Kellor AE. Patterns of Relating: An Adult
Attachment Perspective. New York: Guilford Press 1994.
24. Brennan KA, Clark CL, Shaver PR. Self-report measurement of
adult attachment: An integrative overview. In: Simpson JA,
Rholes WS, editors. Attachment theory and close relationships.
New York: Guilford; 1998.
25. Bartholomew K, Horowitz LM. Attachment styles among young
adults: a test of a four-category model. J Pers Soc Psychol 1991;
61:226 – 44.
26. Klohnen EC, John OP. Working models of attachment: A theory-
based prototype approach. In: Simpson JA, Rholes WS, editors.
Attachment theory and close relationships. New York: Guilford;
1998.
27. Mickelson KD, Kessler RC, Shaver PR. Adult attachment in a
nationally representative sample. J Pers Soc Psychology 1997;
73:1092–106.
28. Frodi A, Thompson R. Infants’ affective responses in the Strange
Situation: effects of prematurity and quality of attachment.
Child Dev 1985;56:1280 –90.
29. Goldberg S. Risk factors in infant-mother attachment. Canadian
Journal of Psychology 1988;42:173– 88.
30. Speltz ML, Endriga MC, Fisher PA, Mason CA. Early predictors
of attachment in infants with cleft lip and palate. Child Dev
1997;68:12–25.
31. McCallum MS, McKim MK. Recurrent otitis media and attach-
Psychosomatic Medicine 63:556 –567 (2001)
ment security: a path model. Early Education & Development
1999;10:517–34.
32. Ward MJ, Kessler DB, Altman SC. Infant-mother attachment in
children with failure to thrive. Infant Mental Health Journal
1993;14:208 –20.
33. Simmons RJ, Goldberg S, Washington J, Fischer-Fay A. Infant-
mother attachment and nutrition in children with cystic fibro-
sis. J Dev behav Pediatr 1995;16:183– 6.
34. Marvin RS, Pianta RC. Mother’s reaction to their child’s
diagnosis: relations with security of attachment. J Clin Child
Psychol 1996;25:436 – 45.
35. Mrazek DA, Casey B, Anderson I. Insecure attachment in se-
verely asthmatic preschool children: is it a risk factor? J Am
Acad Child Adolesc Psychiatry 1987;26:516 –20.
36. Goldberg S, Simmons RJ, Newman J, Campbell K, Fowler RS.
Congenital heart disease, parental stress and infant-mother re-
lationships. J Pediatr 1991;119:661– 6.
37. Crowell JA, Fraley RC, Shaver PR. Measurement of individual
differences in adolescent and adult attachment. In: Cassidy J,
Shaver PR, editors. Handbook of attachment. New York:
Guilford; 1999. p. 434 – 65.
38. Dozier M, Stovall KC, Albus KE. Attachment and psychopathol-
ogy in adulthood. In: Cassidy J, Shaver PR, editors. Handbook of
Attachment: Theory, Research, and Clinical Applications. New
York: Guilford Press, 1999;497–519.
39. Scheidt CE, Waller E, Malchow H, Ehlert U, Becker-Stoll F,
Schulte-Moenting J, Lucking CH. Attachment representation
and cortisol response to the Adult Attachment Interview in
idiopathic spasmodic torticollis. Psychother Psychosom 2000;
69:155– 62.
40. Maunder RG, Lancee WJ, Greenberg GR, Hunter JJ, Fernandes
BJ. Insecure attachment in a subgroup with ulcerative colitis
defined by ANCA status. Dig Dis Sci 2000; 45:2127–32.
41. Antonovsky H, Sagy S. The development of a sense of coherence
and its impact on responses to stress situations. J Soc Psychol
1986;126:213–25.
42. Poppius E, Tenkanen L, Kalimo R, Heinsalmi P. The sense of
coherence, occupation and the risk of coronary heart disease in
the Helsinki Heart Study. Soc Sci Med 1999;49:109 –20.
43. Ackerman SH, Hofer MA, Weiner H. Age at maternal separation
and gastric erosion susceptibility in the rat. Psychosom Med
1975;37:180 – 4.
44. LaBarba RC, White JL. Maternal deprivation and the response to
Erlich carcinoma in BALB/c mice. Psychosom Med 1971;33:
458 – 61.
45. LaBarba RC, Lazar JM, White JL. The effects of maternal separa-
tion, isolation and sex on the response to Erlich carcinoma in
BALB/c mice. Psychosom Med 1972;34:557–9.
46. Lyons-Ruth K, Jacobvitz D. Attachment disorganization: Unre-
solved loss, relational violence, and lapses in behavioural and
attentional strategies. In: Cassidy J, Shaver PR, editors. Hand-
book of attachment: theory, research, and clinical applications.
New York: Guildford; 1999;520 –54.
47. Felitti VJ, Anda RF, Nordenberg D, Williamson DF, Spitz AM,
Edwards V, Koss MP, Marks JS. Relationship of childhood abuse
and household dysfunction to many of the leading causes of
deaths in adults: The adverse childhood experiences (ACE)
study. Am J Prev Med 1998;14:245–58.
48. Vaillant GE, Vaillant CO. Natural history of male psychological
health, XII: a 45-year study of predictors of successful aging at
age 65. Am J Psychiatry 1990;147(1):31–7.
49. Vaillant GE. Natural history of male psychological health, II:
some antecedents of healthy adult adjustment. Arch Gen Psy-
chiatry 1974;31:15–22.
565

50. Stuart S, Noyes R. Attachment and interpersonal communica-
tion in somatization. Psychosomatics 1999;40:34 – 43.
51. Mikulincer M, Florian V. Appraisal of and coping with a real-
life stressful situation: the contribution of attachment styles.
Personality and Social Psychology Bulletin 1995;21:406 –14.
52. Mikulincer M, Florian V, Weller A. Attachment styles, coping
strategies, and posttraumatic psychological distress: the impact
of the Gulf War in Israel. J Pers Soc Psychol 1993;64:817–26.
53. Levine S, Haltmeyer GC, Kaas GG, Penenberg VH. Physiological
and behavioral effects of infant stimulation. Physiol Behav 1967;
2:55– 63.
54. Liu D, Diorio J, Tannenbaum B, Caldji C, Francis D, Freedman A,
Sharma S, Pearson D, Plotsky PM, Meaney MJ. Maternal care,
hippocampal glucocorticoid receptors, and the hypothalamic-
pituitary-adrenal responses to stress. Science 1997;277:
1659 – 62.
55. Sroufe LA, Waters E. Heart rate as a convergent measure in
clinical and developmental research. Merrill-Palmer Quarterly
1977;23:3–27.
56. Spangler G, Schieche M. Emotional and adrenocortical re-
sponses of infants to the strange situation: the differential func-
tion of emotional expression. Int J Behav Dev 1998;22:681–706.
57. Nachmias M, Gunnar MR, Mangelsdorf S, Parritz RH, Buss K.
Behavioral inhibition and stress reactivity: the moderating role
of attachment security. Child Dev 1996;67:508 –22.
58. Luecken LJ. Childhood attachment and loss experiences affect
adult cardiovascular and cortisol function. Psychosom Med
1998;60:765–72.
59. Heim C, Newport DJ, Heit S, Graham YP, Wilcox M, Bonsall R,
Miller AH, Nemeroff CB. Pituitary-adrenal and autonomic re-
sponses to stress in women after sexual and physical abuse in
childhood. Jama 2000;284:592–7.
60. House JS, Landis KR, Umberson D. Social relationships and
health. Science 1988;241:540 –5.
61. Uchino BN, Cacioppo JT, Kielcolt-Glaser JK. The relationship
between social support and physiological processes: a review
with emphasis on underlying mechanisms and implications for
health. Psychol Bull 1996;119:488 –531.
62. West M, Livesley WJ, Reiffer L, Sheldon A. The place of attach-
ment in the life events model of stress and illness. Can J Psy-
chiatry 1986;31:202–7.
63. Brown GW, Bhrolchain MN, Harris T. Social class and psychi-
atric disturbance among women in an urban population. Soci-
ology 1975;9:225–54.
64. Stroebe W, Stroebe M, Abakoumkin G. The role of loneliness
and social support in adjustment to loss: A test of attachment
versus stress theory. J Pers Soc Psychol 1996;70:1241–9.
65. Sachser N, Matthias D, Hirzel D. Social relationships and the
management of stress. Psychoneuroendocrinology 1998;23:
891–904.
66. Kirschbaum C, Klauer T, Filipp SH, Hellhammer DH. Sex-
specific effects of social support on cortisol and subjective re-
sponses to acute stress. Psychosom Med 1995;57:23–31.
67. Franks P, Campbell TL, Shields CG. Social relationships and
health: the relative roles of family functioning and social sup-
port. Soc Sci Med 1992;34:779 – 88.
68. Levine DM, Green LW, Deeds SG, Chwalow J, Russell RP, Finlay
J. Health education for hypertensive patients. Jama 1979;241:
1700 –3.
69. Florian V, Mikulincer M. Effects of adult attachment style on the
perception and search for social support. J Psychol 1995;129:
665–76.
70. Simpson JA, Rholes WS, Nelligan JS. Support seeking and sup-
port gicing within couples in an anxiety provoking situation: the
566
R. G. MAUNDER AND J. J. HUNTER
role of attachment styles. J Pers Soc Psychology 1992;62:
434 – 46.
71. Ognibene TC, Collins NL. Adult attachment styles, perceived
social support, and coping strategies. J Soc Pers Relations 1998;
15:323– 45.
72. Mikulincer M. Adult attachment style and affect regulation:
strategic variation in self-appraisals. J Pers Soc Psychol 1999;75:
420 –35.
73. Kobak R, Sceery A. Attachment in late adolescence: working
models, affect regulation, and representation of self and others.
Child Dev 1999;59:135– 46.
74. DeFronzo J, Pawlak R. Effects of social bonds and childhood
experience on alcohol abuse and smoking. J Soc Psychol 1993;
133:635– 42.
75. Springs FE, Friedrich WN. Health risk behaviors and medical
sequalae of childhood sexual abuse. Mayo Clin Proc 1992;67:
603– 4.
76. Magai C. Affect, imagery and attachment. In: Cassidy J, Shaver
PR, editors. Handbook of attachment: theory research and clin-
ical applications. New York: Guilford Press; 1999.
77. Wills TA, Cleary SD. How are social support effects mediated? A
test with parental support and adlescent substance use. J Pers
Soc Psychol 1996;71:937–52.
78. McBride CM, Curry SJ, Cheadle A, Anderman C, Wagner EH.
School-level application of a social bonding model to adolescent
risk-taking behaviour. J Sch Health 1995;65:63– 8.
79. Brook JS, Whiteman M, Finch S, Cohen P. Mutual attachment,
personality, and drug use: pathways from childhood to young
adulthood. Genetic Social and General Psychology Monographs
1998;124:492–510.
80. Sharpe TM, Killen JD, Bryson SW, Shisslak CM, Estes LS, Gray
N, Crago M, Taylor CB. Attachment style and weight concerns in
preadolescent and adolescent girls. Int J Eat Disord 1998;23:
39 – 44.
81. Lavik NJ, Clausen SE, Pedersen W. Eating behaviour, drug use,
psychopathology, and parental bonding in Norway. Acta Psy-
chiatr Scand 1991;84:387–90.
82. Raynes E, Auerbach C, Botyanski NC. Level of object represen-
tation and psychic structure deficit in obese persons. Psychol
Rep 1989;64:291– 4.
83. Feeney JA, Raphael B. Adult attachments and sexuality: impli-
cations for understanding risk behaviours for HIV infection.
Aust N Z J Psychiatry 1992;26:399 – 407.
84. Bassman LE. Object relations and self-reported AIDS self-care
behaviour. Psychol Rep 1991;68:915–23.
85. Viederman M, Hymowitz P. A developmental-psychodynamic
model for diabetic control. Gen Hosp Psychiatry 1988;10:34 – 40.
86. Ciechanowski P, Katon W, Hirsch I. Attachment style and ad-
herence in the diabetic patient [abstract]. Psychosom Med 1999;
61:110.
87. Kotler T, Buzwell S, Romeo Y, Bowland J. Avoidant attachment
as a risk factor for health. Br J Med Psychol 1994;67:237– 45.
88. Ciechanowski P, Dwight M, Katon W, Rivera-Ball D. Attachment
classification associated with unexplained medical symptoms
in patients with chronic hepatitis C. Proceedings of the Second
International Attachment and Psychopathology Conference;
1998 Toronto, Canada.
89. Jorgensen RS, Frankowski JL, Carey MP. Sense of coherence,
negative life events and appraisal of physical health among
university students. Personality and Individual Differences
1999;27:1079 – 89.
90. Kivimaki M, Feldt T, Vahtera J, Nurmi JE. Sense of coherence
and health: evidence from two cross-lagged longitudinal sam-
ples. Soc Sci Med 2000;50:583–97.
Psychosomatic Medicine 63:556 –567 (2001)
ATTACHMENT AND DISEASE

91. Simpson JA. Attachment theory in modern evolutionary per-
spective. In: Cassidy J, editor. Hanbook of Attachment: Theory
Research and Clinical Applications. New York: Guilford, 1999;
115– 61.
92. West M, Rose MS, Verhoef MJ, Spreng S, Bobey M. Anxious
attachment and self-reported depressive symptomatology in
women. Can J Psychiatry 1998;43:294 –7.
93. Heim C, Nemeroff CB. The impact of early adverse experi-
ences on brain systems involved in the pathophysiology of
anxiety and affective disorders. Biol Psychiatry 2000;46:
1509 –22.
94. Talley NJ, Boyce PM, Jones M. Is the association between irri-
table bowel syndrome and abuse explained by neuroticism? A
population-based study. Gut 1999;42:47–53.

ANNOUNCEMENT

Minority Research Training in Psychiatry

The Program for Minority Research Training in Psychiatry (PMRTP) is funded by the National Institute of
Mental Health. Through it, the American Psychiatric Institute for Research and Education sponsors training
of minority medical students, Psychiatric residents, and fellows who are interested in research by providing
advice, placement assistance, tuition, stipends, travel and other expenses. The director of the Program is
James Thompson, M.D., M.P.H.; the project manager is Ernesto Guerra. For more information call the toll
free number for the PMRTP, 1-800-852-1390, or 202-682-6225, email eguerra@psych.org, or write to PMRTP
at the American Psychiatric Institute for Research and Education, 1400 K Street, NW, Washington, DC
20005.

Psychosomatic Medicine 63:556 –567 (2001) 567