Volume 14, Number 12, 2004
© Mary Ann Liebert, Inc.
Stress and Thyroid Autoimmunity
Tetsuya Mizokami,1Audrey Wu Li,2Samer El-Kaissi,1and Jack R. Wall1,3
While many studies have shown a connection between stress and autoimmune disease, most of the evidence
for stress contributing to the onset and course of autoimmune disease is circumstantial and the mechanisms by
which stress affects autoimmune disease are not fully understood. The best circumstantial evidence for an ef-
fect of stress on autoimmune thyroid disease is the well-known relationship between the onset of Graves’ hy-
perthyroidism and major stress but even this is debated. However, most of the recent case-control studies have
supported stress as a factor that affects the onset and clinical course of Graves’ disease. On the other hand,
there have been few reports concerning the possible relationship between stress and Hashimoto’s thyroiditis.
Because the onset and course of Hashimoto’s thyroiditis is generally insidious, the effect of stress on Hashimoto’s
thyroiditis might be overlooked. Numerous human and animal studies have demonstrated that psychological
and physiologic stressors induce various immunologic changes. Stress affects the immune system either di-
rectly or indirectly through the nervous and endocrine systems. These immune modulations may contribute to
the development of autoimmunity as well as the susceptibility to autoimmune disease in genetically predis-
posed individuals. Stress can be one of the environmental factors for thyroid autoimmunity.
action of the body in which normal homeostasis, or the
steady-state internal milieu, is disturbed or threatened” (2,3).
The stress system serves to counteract the effect of stressors
to maintain homeostasis. However, stress affects multiple
sites and induces various changes in all organ systems (4).
Activation of the stress system leads to a “general adapta-
tion state,” or the “stress syndrome,” which can be defined
as a cluster of time-limited psychological and physiologic
changes such as anorexia, body weight loss, and depression
(5). In addition, a hyperactive or hypoactive response to
stress may itself produce or contribute to various clinical dis-
orders (5). Numerous studies have indicated a connection
between stress and autoimmunity and that stress may trig-
ger or worsen autoimmune disease (6–9).
LTHOUGH THERE ARE MANY definitions of stress (1), stress
can be viewed as, “the complex psychophysiological re-
Stress and Graves’ Disease
The relationship between stressful life events and the on-
set of Graves’ disease was initially documented by Parry in
1825, and was subsequently noted by Graves, Basedow, and
others (10). There are many reports on the association be-
tween stress and the onset of Graves’ disease. However, most
of the early reports are anecdotal, and had considerable
drawbacks, such as inappropriate epidemiologic methods,
small sample size, improper controls, poor differential diag-
nosis within thyrotoxicosis, and imprecise definition of
stressful life events (11,12). Recently, many epidemiologi-
cally improved studies have demonstrated that patients with
Graves’ disease had more stressful life events than control
subjects prior to the onset or diagnosis of Graves’ hyperthy-
roidism and that stress had an unfavorable effect on the
prognosis of Graves’ disease.
Studies on the onset of Graves’ disease
The report of Winsa et al. (13) in 1991 appears to be the
first large population-based case-control study demonstrat-
ing a relationship between stress and the onset of Graves’
disease. By using a self-rated questionnaire in this study, 208
Swedish patients with newly diagnosed Graves’ disease had
more negative life events in the 12 months preceding the di-
agnosis and higher negative life event scores than 372
matched controls (odds ratio 6.3 for the category with the
highest negative score). Patients and controls had a similar
number of positive events. Sonino et al. (14) in Italy reported
that patients with Graves’ disease had significantly more
1Department of Clinical and Biomedical Sciences: Barwon Health, The Geelong Hospital, Geelong, Victoria, Australia.
2Thyroid Research Laboratory, Dalhousie University, Halifax, Nova Scotia, Canada.
3Department of Medicine, University of Sydney, Nepean Hospital, Penrith, N.S.W., Australia.
positive and negative life events than controls (patients 1.51
total events, controls 0.54; p ? 0.001). They interviewed 70
patients with Graves’ disease and a control group of 70
healthy subjects by investigating the occurrence of stressful
life events in the year before the first sign of disease onset
(Paykel’s interview for recent life events).
Kung (15) in Hong Kong examined life events and daily
stress in 95 patients with Graves’ disease prior to the onset
of the disease by using a self-rated questionnaire in the 12
months preceding the diagnosis. Graves’ disease patients ex-
perienced more negative life events (patients 1.24 events,
controls 0.63; p ? 0.003) and perceived them with higher rat-
ings. Similarly, patients with Graves’ disease reported more
daily hassles (patients, 14.58 hassles; controls, 8.68; p ? 0.001)
and had higher hassle scores. The number of positive and
neutral events was not significantly different between the
two groups. Radosavljevi´ c et al. (16) conducted a case-con-
trol study that included 100 newly diagnosed Yugoslavian
patients with Graves’ disease and 100 matched controls who
responded to Paykel’s interview. The patients had signifi-
cantly more stressful life events in the 12 months preceding
the diagnosis (patients 3.18 events, controls 1.85; p ? 0.0001).
The stressful events related to work, relationships, separa-
tion, and financial difficulties were more prevalent among
patients than controls.
Yoshiuchi et al. (17) investigated 228 Japanese patients
(182 women and 46 men) with newly diagnosed Graves’ dis-
ease and matched controls by using a self-rated question-
naire and multivariate analysis. They found that stressful life
events in the 12 months preceding the diagnosis were sig-
nificantly associated with the risk of developing Graves’ dis-
ease in women (relative risk, 7.7; highest stress score com-
pared to lowest score) but not in men. There was no
significant difference in the daily hassle score between pa-
tients and controls in either women or men, which is not con-
sistent with Kung’s result (15). Matos-Santos et al. (18) in
Portugal evaluated stressful life events for the 12 months pre-
ceding the onset of symptoms of thyroid disease among 31
patients with Graves’ disease, 31 with toxic nodular goiter
(nonautoimmune hyperthyroidism) and 31 healthy control
subjects using the same interview. Patients with Graves’ dis-
ease not only had a significantly greater number of stressful
life events (Graves’ disease, 174 total events; toxic goiter, 79;
control, 50; p ? 0.016) but also a higher number (Graves’ dis-
ease, 141 total events; toxic goiter, 65; control, 27; p ? 0.001)
and greater impact of negative stressful life events compared
to the other two groups. The differences between toxic nodu-
lar goiter group and the control group were not significant.
Graves’ disease had a higher impact of positive stressful life
events than toxic nodular goiter.
The above case-control studies all demonstrate that pa-
tients with Graves’ disease experienced more stressful life
events than control subjects. The study designs are different
(Table 1) and the results about the effects of stress do not
completely coincide between the reports. Only Yoshiuchi et
al. (17) found a gender difference in the effect of stressful life
events on the onset of Graves’ disease, and speculated that
this might be related to differences in immune response to
stress, race, or iodine intake. The subjects in their studies
were Japanese, whose iodine intake is generally high. How-
ever, the sample size of the male subjects was too small (46
men) to make any conclusions. Matos-Santos et al. (18) com-
pared Graves’ disease and toxic nodular goiter to correct for
the effect of the “thyrotoxocosis.” However, the degree of
hyperthyroidism were not documented in each group. Be-
cause patients with Graves’ hyperthyroidism tend to be more
thyrotoxic than those with toxic nodular goiter, the effects of
thyrotoxicosis on the result cannot be ignored (11).
Studies on the clinical course of Graves’ disease
In contrast to studies addressing the relationship between
stress and the onset of Graves’ disease, there have been fewer
studies about the relationship between stress and the clini-
cal course of Graves’ disease after its onset. Stewart et al. (19)
treated 293 patients with Graves’ disease using 131I. Eighty-
one patients with Graves’ disease who had a stressful event
antedating the onset of overt clinical symptoms became hy-
pothyroid earlier than nonstressed patients (50% rate of cu-
mulative hypothyroidism at 12 months for the stressed
group versus at 36 months for the nonstressed group; p ?
0.01). The authors speculated that an enhanced autoimmune
reaction against the thyroid acted synergistically with the ra-
diation to cause hypothyroidism earlier. However, the as-
sessment of stress was less scientifically precise. Benvenga
(20) retrospectively reviewed the clinical records of patients
with Graves’ disease. Most of the relapsers had taken bro-
mazepam only in the acute phase of thyrotoxicosis (17/23,
or 74%) whereas most of the nonrelapsers had taken bro-
mazepam for a longer period (14/21, or 71%). The author
suggested that stress management was effective in improv-
ing the prognosis of Graves’ hyperthyroidism. The possible
effect of stress on the outcome of radioactive iodine treat-
ment and the effect of tranquilizers for stress management
on the clinical course of Graves’ disease, remain to be clari-
fied and need further study.
On the other hand, there are case-control studies that sug-
gest that stress has a negative impact on the prognosis of
Graves’ disease. Yoshiuchi et al. (21) investigated the asso-
ciation between the short-term outcome of 230 patients (182
women and 48 men) with newly diagnosed Graves’ disease,
assessed 12 months after the beginning of antithyroid drug
therapy, and stressful life events, by questionnaire and mul-
tivariate analysis. Daily hassles at 6 months after beginning
therapy were significantly and independently associated
with the hyperthyroid state 12 months after beginning ther-
apy of Graves’ disease only in women (relative risk, 3.9; high
daily hassles scores compared to lower daily hassles scores).
However, the male sample size of this study may have been
too small to show a significant effect. Fukao et al. (22) stud-
ied the effects of emotional stress and patients’ personality
traits on the prognosis of hyperthyroidism in 69 antithyroid
drug-treated euthyroid patients with Graves’ hyperthy-
roidism using questionnaires. Stress scores correlated sig-
nificantly with higher serum levels of thyrotropin (TSH) re-
ceptor antibodies and larger thyroid volumes after the
cessation of antithyroid drugs. Daily hassles and four per-
sonality traits namely, hypochondriasis, depression, para-
noia and psychasthenia (mental fatigue), were more com-
mon in relapsers (n ? 41) than in patients who remained in
remission (n ? 28).
In contrast to studies relating to the onset of Graves’ dis-
ease, the above case-control studies on the clinical course of
Graves’ disease are prospective. Although the number of
MIZOKAMI ET AL.1048
STRESS AND AUTOIMMUNITY1049
TABLE 1. CASE–CONTROL STUDIES ON THE RELATIONSHIP BETWEEN STRESS AND GRAVES’ DISEASE
Period of stress evaluation/
Subjects (Number, Male, Female)
Thyroid status at the time of study
Gray 1985 (31)
Thyrotoxicosis (50; M 39, F11)
6 months prior to the first symptoms
Non-toxic goiters (50; M 45, F5)
Winsa 1991 (13)
Graves’ (208; M 37, F 171)
1 year prior to diagnosis
Less than 2 weeks after diagnosis
Sonino 1993 (14)
Graves’ (70; M 12, F 58)
1 year prior to the first signs
Euthyroid after treatment
Kung 1995 (15)
Graves’ (95; M 15, F 80)
1 year prior to diagnosis
Radosavljevi´ c 1996 (16)
Graves’ (100; M 7, F 93)
1 year prior to diagnosis
Yoshiuchi 1998 (17)
Graves’ (228; M 46, F 182)
1 year prior to diagnosis
Within 1 month after diagnosis
Yoshiuchi 1998 (21)
Graves’ (230; M 48, F 182)
1 year after the treatment
Martin-du Pan 1998 (33)
Graves’ (98; M 12, F 86)
Hashimoto (95; M 4, F 91)
Benign nodule (97; F 97)
Chiovato 1998 (27)
Panic disorder (87; M 17, F 70)
No Graves’ disease
Matos-Santos 2001 (18)
Graves’ (31; M 9, F 22)
1 year prior to the first symptoms
Toxic nodular goiter (31)
Euthyroid after treatment
Fukao 2003 (22)
Graves’ (69; M 4, F 65)
1 year after the cessation of anti-thyroid
Control (32; M 1, F 31)
Euthyroid after 2–5 years’ treatment
studies concerning the clinical course of disease is much
smaller than those relating to disease onset, the studies sug-
gest that major life events, some personality traits, and daily
problems, may negatively impact on the prognosis of an-
tithyroid drug-treated Graves’ hyperthyroidism.
Reports that suggest an effect of
stress on Graves’ disease
The incidence of Graves’ disease has been observed to in-
crease during major wars in early surveys (10,11). The inci-
dence of Graves’ disease dramatically increased by a factor
of 5 in eastern Serbia (former Yugoslavia) during the civil
war between 1994 and 1996 while the incidence of toxic ad-
enoma (Plummer’s disease) did not increase during the same
period (23). The authors speculated that civil war and eco-
nomic destruction with catastrophic inflation might trigger
an increase in the incidence of Graves’ disease. However, the
incidence of treated thyrotoxicosis has not changed during
the ongoing civil unrest in Northern Ireland (24). Chiovato
et al. (25) suggested that stressful events relating to one’s
personal life might be more important than “social stress.”
There are clinical case reports that suggest a possible re-
lationship between stress and Graves’ disease in various sit-
uations. The following are relatively recently published re-
ports. Unfortunately, the number of patients is too small to
generalize the relationship from these reports. Matusbayashi
et al. (26) reported two patients with Graves’ disease occur-
ring in the setting of previously established euthyroid panic
disorder. The interval between the onset of panic disorder
and that of hyperthyroidism was 4 and 5 years, respectively.
Although they suggested that a panic disorder might pre-
dispose to Graves’ disease, Chiovato et al. (27) could not
show an increased incidence of Graves’ disease in patients
with a panic disorder (vide infra). Misaki et al. (28) reported
three cases of hyperthyroid Graves’ disease that occurred af-
ter partial thyroidectomy for papillary carcinoma. They spec-
ulated that surgical stress might have altered immunologi-
cal homeostasis, converting preclinical Graves’ disease into
full-blown hyperthyroidism. Because the onset of Graves’
hyperthyroidism after tumor resection has only rarely been
reported, anesthesia and surgery are assumed not to be ma-
jor stressors for the development of Graves’ disease. A rela-
tionship between stress and the onset and clinical course of
Graves’ disease has also been reported in children. Morillo
and Gardner (29) reported four children with an age range
of 8 to 14 years in whom separating experiences appeared
to be related to the onset or relapse of Graves’ disease. They
suggested that in some children psychological events were
important triggers of Graves’ disease (30). The prevalence of
Graves’ disease in children is much less than that in adults,
and there are no case-control studies concerning the rela-
tionship between stress and Graves’ disease in children.
Some authors obtained contradictory findings, and con-
cluded that stress did not cause Graves’ disease (31). Limit-
ing our survey to studies utilizing established epidemiologic
methods, we review Gray and Hoffenberg’s report published
in 1985 and reports published thereafter. Gray and Hoffen-
berg (31) did not find any significant difference in the num-
ber and nature of stressful life events up to 6 months before
the onset of thyrotoxicosis between 50 patients with thyro-
toxicosis and 50 with nontoxic goiters. However, they in-
terviewed untreated “thyrotoxic” patients and did not dis-
tinguish between Graves’ disease and other causes thy-
rotoxocosis. Chiovato et al. (27) in Italy could not find past
or present Graves’ hyperthyroidism in 87 patients with panic
disorder encompassing a total of 478 patient-years of expo-
sure to recurrent endogenous stress unrelated to life events.
They speculated that failure to activate the hypothalamic-pi-
tuitary axis by endogenous stress resulting from panic dis-
order, as opposed to exogenous stress resulting from life
events, might be the reason why panic disorder does not pre-
cipitate Graves’ hyperthyroidism. Recently, Simon et al. (32)
found a significantly elevated rate of thyroid dysfunction by
combining their own data and those from 12 previous panic
disorder studies that examined both thyroid history and
blood hormone levels. They concluded that the lifetime
prevalence of thyroid dysfunction appeared to be elevated
in patients with panic disorder. However, each study had
relatively small sample sizes and it is not known whether
panic disorder precedes thyroid dysfunction or vice versa.
In addition, the cause of thyroid dysfunction was uncertain
was not certain in all studies. Martin-du Pan (33) in France
evaluated the role of major stress and pregnancy in trigger-
ing autoimmune thyroid disease in 98 patients with Graves’
disease and 97 patients with benign thyroid nodules. There
were no significant differences of stress factors between the
two groups, and generally the role of stress in triggering
Graves’ disease seemed weak and dubious compared to the
role of pregnancy and the postpartum period.
Criticisms of case-control studies
Almost all of the case-control studies addressing the effect
of stress on Graves’ disease can be criticized. These criticisms
apply not only to reports of a positive relationship between
stress and Graves’ disease but also to those showing a neg-
ative relationship. There are some general methodologic
problems and limitations in studies dealing with stress, es-
pecially retrospective studies based on the assessment of life
events preceding the onset of thyrotoxicosis or the diagno-
sis of Graves’ disease (11,12,34–36). Firstly, the main scien-
tific problem is the difficulty in defining “stress” and objec-
tively quantifying individual stressors. The impact of life
events vary between individuals, despite apparently similar
events. Second, the recall bias cannot be avoided in retro-
spective studies. Patients with Graves’ disease may be more
prone to recall stressful events than healthy controls. Third,
it is impossible to date the onset of Graves’ disease precisely,
which is sometimes insidious and thyrotoxic symptoms are
often absent in subclinical disease. Thus, stressful events may
occur after the onset of Graves’ disease. Some studies in-
vestigated life events in the 12 months before the diagnosis,
rather than before the first symptoms or signs. However,
some events could have occurred between the onset and di-
agnosis. Finally, thyrotoxicosis itself can cause psychologi-
cal disturbance and behavioral changes such as anxiety and
emotional liability, which may have an effect on life events
(37). Some stressful events associated with Graves’ disease
may be the consequence rather than the trigger for disease
In summary, the effects of stress on Graves’ disease seem
MIZOKAMI ET AL. 1050
likely from the clinical and epidemiological point of view.
However, the evidence for stressful life events contributing
to the development and prognosis of Graves’ disease are cir-
cumstantial, and all studies on stress have several method-
ological limitations. Moreover, the biologic mechanism by
which stress affects Graves’ disease remains uncertain (vide
Stress and Hashimoto’s Thyroiditis
In contrast to Graves’ disease, there have been few reports
concerning a possible association between stressful events
and Hashimoto’ thyroiditis. Martin-du Pan (33) evaluated
the triggering role of major stressors and pregnancy in the
occurrence of autoimmune thyroid disease in 95 patients
with Hashimoto’s thyroiditis and 97 patients with benign
thyroid nodules as controls. He concluded that stress and
pregnancy did not have any triggering role in Hashimoto’s
thyroiditis. Oretti et al. (38) investigated the possible associ-
ation between life events and the development of postpar-
tum thyroid dysfunction in 115 pregnant women with thy-
roid autoantibodies and 123 thyroid autoantibody negative
pregnant women. Women who developed postpartum thy-
roid dysfunction did not report an excess of life events (to-
tal, negative or neutral) in the preceding year and there was
no difference in the number of life events between antibody
positive and antibody negative women.
BioBreeding/Worcester (BB/W) rats and nonobese dia-
betic (NOD) mice spontaneously develop autoimmune dia-
betes and thyroiditis with lymphocytic infiltration and au-
toantibodies. The onset of diabetes in BB/W rats and NOD
mice is clearly advanced by various stressors (39,40). A com-
bination of behavioral stressors, such as restraint and crowd-
ing, were found to lower the age of onset of diabetes in BB/W
rats and a greater percentage of BB/W rats became diabetic
after chronic stress, compared with unstressed controls
(39,41). These experimental animals have specific genetic ab-
normalities that affect their immune system and it may not
be appropriate to extrapolate the findings from animal ex-
periments to humans.
The onset and course of Hashimoto’s thyroiditis are often
insidious and patients do not become symptomatic until they
develop overt hypothyroidism or a goiter, making it diffi-
cult to evaluate the role of stress in the onset and course of
Stress and the Neuroendocrine-immune System
The orchestrated neuronal interplay in the brain underlies
the stress system (42). Activation of the stress system affects
the nervous, endocrine, and immune systems, and leads to
changes of the internal milieu that regulates homeostasis.
The immune system receives signals directly from the brain
and indirectly from the nervous and endocrine systems. The
principal components of the stress system are corticotropin-
releasing hormone (CRH) in the hypothalamus, the locus
ceruleus, and other noradrenergic neurons in the brain stem,
and their peripheral effectors. These peripheral effectors
are the pituitary-adrenal axis and the systemic and
adrenomedullary sympathetic nervous systems (sympa-
thoadrenal system). The hypothalamic-pituitary-adrenal
(HPA) axis and sympathoadrenal system have a close mu-
tual interaction. In addition to the HPA axis and sympa-
thoadrenal system, the stress system is also associated with
hypothalamic-pituitary-gonadal (HPG) and other neuroen-
docrine axes, which also affect the immune system (43–45).
Hypothalamic-pituitary-adrenal axis and glucocorticoids
CRH stimulates adrenocorticotropin (ACTH) secretion
from the anterior pituitary, which in turn stimulates the
adrenal cortex to secrete glucocorticoids. Glucocorticoids ap-
pear to play a major role in the stress-induced immune re-
action. Glucocorticoids modulate the immune response at
multiple levels, altering leukocyte migration, diminishing
antigen presentation and expression of major histocompati-
bility complex class II, suppressing lymphocyte proliferation
and differentiation and decreasing the production and ef-
fects of cytokines and other mediators (46). Although phar-
macologic doses of glucocorticoids are nonspecifically im-
munosuppressive at virtually every level of the immune
response, physiologic levels of glucocorticoids secreted by
the adrenal glands regulate immune function to maintain ho-
meostasis and are immunomodulatory rather than solely im-
munosuppressive (45). Glucocorticoids inhibit the differen-
tiation and function of T-helper 1 (TH1) lymphocytes and the
secretion of TH1 cytokines, but do not affect or potentiate the
secretion of TH2 cytokines (47). The effect of high physio-
logic levels of glucocorticoids, such as those induced by
stress, is therefore to shift the immune response from a TH1-
to a TH2-type pattern (47).
STRESS AND AUTOIMMUNITY 1051
neuronal, endocrine, and immune systems.
Stress and autoimmunity: The relationship between
Systemic/adrenomedullary sympathetic nervous systems
(sympathoadrenal system) and other nervous systems
The autonomic nervous system regulates the immune sys-
tem regionally through the innervation of lymphoid organs
by the sympathetic nervous system and through the pe-
ripheral nervous system. Locally released norepinephrine
from the sympathetic nerve terminals and circulating cate-
cholamines affect lymphocyte traffic and proliferation, and
modulate local lymphocyte functions (48). Catecholamines
inhibit the production of TH1 cytokines via ?2-adrenergic re-
ceptors, but do not directly affect the production of TH2 cy-
tokines, thereby favoring a TH2 shift (48). The nervous sys-
tem also regulates local immune responses through the
peptidergic nerves with the release of neuropeptides such as
CRH, substance P, and vasoactive intestinal polypeptide (49).
Hypothalamic-pituitary-gonadal axis and sex hormones
There is a bidirectional relationship between the HPA axis
and the HPG axis. CRH inhibits hypothalamic gonadotro-
pin-releasing hormone (GnRH) secretion directly and via
?-endorphin (44). The final effector of the HPA axis, gluco-
corticoids, inhibits hypothalamic GnRH and pituitary
luteinizing hormone secretion (43). Glucocorticoids suppress
the production of sex hormones (estrogens and androgens)
and the action of these hormones on target tissues (50). Es-
trogens enhance and androgens inhibit the HPA axis re-
sponsiveness to stress, therefore females have higher HPA
axis responsiveness than males (51).
Sex hormones play an important role in immune modu-
lation and contribute to the greater incidence of autoimmune
disease seen in females (52). Sex hormones may act directly
on the immune system, modulating aspects of antigen pres-
entation, lymphocyte activation, cytokine expression and
homing of immune cells. Estrogens enhance autoantibody
production, while androgens diminish the number of B cells
and depress autoantibody production (53).
Hypothalamic-pituitary-thyroid axis and thyroid hormones
Stress alters the hypothalamic-pituitary-thyroid (HPT)
axis function. Although the daily rhythm of TSH production
is preserved in stress, the secretion of pituitary TSH is sup-
pressed and TSH response to thyrotropin-releasing hormone
(TRH) is blunted. The conversion of the relatively inactive
thyroxine to the biologically active triiodothyronine in pe-
ripheral tissues is decreased during stress (54). On the other
hand, the HPT axis has a close bidirectional relationship with
the HPA axis and the sympathoadrenal system (55–57). Glu-
cocorticoids inhibit TRH-induced TSH secretion and cate-
cholamines may enhance selected responses to triiodothyro-
nine. Conversely, thyroid hormones enhance the actions of
glucocorticoids and ?-adrenergic effects (56,57). Once the
normal relationship between these endocrine axes is dis-
turbed, thyrotoxicosis can induce a vicious circle.
Thyroid hormones can directly cause immune alterations.
T-lymphocyte proliferative responses to mitogens are de-
creased (58) and primary humoral immune responses are de-
pressed in hypothyroid animals (59). However, the effect of
hyperthyroidism provoked by triiodothyronine or thyroxine
administration on humoral and cellular immunity is contro-
Human studies on stress and general
Human studies have demonstrated that psychological and
physiologic stress can cause various immunologic changes.
Biondi and Picardi (60) reviewed studies showing effects of
emotional stress on neuroendocrine function in healthy hu-
mans. They concluded that HPA axis and the sympathoad-
renal system were most tightly linked to an acute stress re-
sponse. The stress response was not caused by the nature of
the stressor per se, but by the ability of the individual to deal
with the stressor. The effects of various stressful life events
such as bereavement, marital discord, providing care for a
relative with a debilitating illness and the stress of academic
examinations on immunologic functions have been studied
(4,61). Herbert and Cohen (61) carried out a meta-analysis of
these reports on stress and immunity. They showed a rela-
tionship between stress and a decrease in functional immune
measures. Stress was associated with decreased proliferative
response of lymphocytes to mitogens, as well as natural killer
cell activity. In terms of cell numbers, stress was reliably as-
sociated with a higher number of circulating white blood
cells and lower numbers of circulating B-lymphocytes,
T-lymphocytes, and helper and suppressor/cytotoxic T-cells.
With respect to immunoglobulin levels, stress was con-
sistently associated with a decrease in total serum im-
munoglobulin M (IgM) levels. Cohen et al. (62) recently re-
viewed the literature on antibody responses to immunization
in humans. The literature supports a relationship between
psychological stress and suppression of the humoral im-
mune response to immunization. Lower secondary antibody
responses were found among patients with chronically high
levels of stress.
The above reports showed that stress was associated with
decreased immune function in healthy humans and that
stress could be a trigger for various immunologic distur-
bances. However, there are no clear biologic explanations as
to how the immunologic disturbances caused by stress af-
fect immune reactions or autoimmune disease. Stress may
cause different immunologic perturbations in immunologi-
cally and genetically defective humans compared to healthy
humans. The reports that demonstrated a relationship be-
tween stress and decreased immune function in healthy hu-
mans cannot be extrapolated to indicate a relationship be-
tween stress and autoimmunity.
Animal models of autoimmune disease
Experimental animal studies suggest that hyperactive or
hypoactive responses of the stress system to a variety of stim-
uli may be associated with susceptibility to autoimmune dis-
ease (47,63). The obese strain (OS) chicken spontaneously
develops autoimmune thyroiditis and is used as an animal
model of Hashimoto’s thyroiditis. The development of spon-
taneous autoimmune thyroiditis in the OS chicken is influ-
enced by several alterations in the immuno-endocrine com-
munication via the HPA axis and the effect of glucocorticoids
on immunocompetent cells (64). OS chickens have signifi-
cantly elevated serum levels of corticosteroid-binding glob-
ulin, which decreases the concentration of bioactive free
glucocorticoids, and cortisol treatment induced a significant
decrease in the frequency and severity of spontaneous au-
toimmune thyroiditis (64). Lewis (LEW/N) and Fischer
MIZOKAMI ET AL. 1052
(F344/N) rats are inbred, major histocompatibility locus-
compatible, rat strains that differ only at one minor histo-
compatibility locus (the Neu-1 locus) (63). Lewis rats are
extraordinarily susceptible to experimentally induced TH1-
mediated autoimmune conditions including uveitis, pol-
yarthritis and experimental allergic encephalomyelitis (2).
These rats display a blunted HPA axis response to a variety
of stressors (47,63). In contrast, Fischer rats, which have a
hyperresponsive HPA axis, are relatively resistant to the de-
velopment of these autoimmune conditions in response to
the same inflammatory stimuli (2,47,63).
The various animal models of spontaneous autoimmune
disease share abnormalities of the HPA axis or the immuno-
neuro-endocrine interaction (64). The association between a
blunted HPA axis and susceptibility to autoimmune disease
has been shown in these animal models, which suggest an
important role for the HPA axis and glucocorticoids in the
pathophysiology of autoimmune disease. Defective activa-
tion of the stress system may increase susceptibility to au-
toimmune disease. On the other hand, Hu et al. (65) pointed
out that there was no strong correlation between plasma glu-
cocorticoid levels and the severity of autoimmune disease.
They suggested that glucocorticoids only modulate the
severity of autoimmune diseases in susceptible individuals,
but do not play an essential role in their initiation.
How Does Stress Affect Graves’ Disease?
Susceptibility to Graves’ disease is determined by a mix-
ture of genetic and environmental factors. Of the genetic fac-
tors, the HLA and cytotoxic T-lymphocyte antigen-4 (CTLA-
4) have been established as major sites of susceptibility loci
(66). Environmental factors include infection, iodine, smok-
ing, and novel immunotherapeutic agents. Stress is consid-
ered to be one of the putative environmental causes of
Graves’ disease. Chiovato et al. (11) suggested that genetic
and environmental factors other than stress are relevant to
the development and course of the disease, because the dif-
ference of stressful life events, which are recorded more fre-
quently in Graves’ patients in many case-control studies, is
generally not great.
The association of stress with Graves’ disease is probably
caused by immunologic perturbations caused by the stress
system. Dysfunction of the stress system may affect the im-
mune response to the TSH receptor through the modulation
of hormones, neurotransmitters and cytokines in genetically
predisposed individuals. The HPA axis appears to play a
principal role in the relationship between stress and au-
toimmune disease, because glucocorticoids have versatile ef-
fects on immunological modulation and the various au-
toimmune animal models show some defects of the HPA
axis. Volpé (67) proposed that a defect of antigen-specific
suppressor T-lymphocytes is partially responsible for the ini-
tiation of Graves’ disease. Stress may cause a generalized
suppressor T-lymphocyte defect and TSH receptor antibod-
ies may be produced as a result of a specific defect in im-
munologic surveillance. The reaction to stress in a geneti-
cally susceptible individual may alter lymphocyte function
or augment a previous immune response and thereby initi-
ate clinical Graves’ disease (68).
Both endogenous glucocorticoids and catecholamines at
concentrations observed during periods of stress cause a se-
lective suppression of TH1 response and a shift toward TH2-
mediated immunity (47). This TH2 shift caused by the acti-
vation of the stress system may affect the onset or course of
Graves’ disease (5,44). Graves’ disease is generally consid-
ered to be a TH2-predominant disease. Patients with Graves’
disease have significantly higher serum levels of TH2 cyto-
kines. Graves’ disease is frequently associated with allergic
rhinitis and TH2-predominant conditions are frequently as-
sociated with allergic disease (69). Humanized anti-CD52
monoclonal antibody therapy for multiple sclerosis, which
causes the immune response to change from the TH1 phe-
notype, triggered Graves’ disease in one study (70). How-
ever, immune deviation toward TH2 induced a significant
inhibition of TSH receptor antibody production and a re-
duction in the prevalence of hyperthyroidism in the murine
Graves’ disease model (71). This suggests that not only TH2
but also TH1 immune responses are involved in the patho-
genesis of Graves’ disease. Both TH1 and TH2 responses may
participate and cooperate in inducing and sustaining Graves’
diseases, and it is difficult to explain the effect of stress on
Graves’ disease only by a TH2 shift.
Davies (72) hypothesized that stress-induced immuno-
suppression might be followed by immune system hyperac-
tivity, which could trigger autoimmune thyroid disease. It is
well known that postpartum thyroid dysfunction is medi-
ated by a rebound from the suppressed immune state dur-
ing pregnancy (73). Maternal immunity is suppressed dur-
ing pregnancy so as not to reject the fetus, but the immune
activity is intensified to above-normal levels after the deliv-
ery. Women with positive TSH receptor antibodies in early
pregnancy have a high risk of developing Graves’ disease
after delivery (73). Moreover, there are various types of post-
partum autoimmune conditions, such as autoimmune hy-
pohysitis, rheumatoid arthritis, and autoimmune hepatitis
(73). However, it is not certain whether the mechanisms of
immunosupression are similar in periods of stress and dur-
Finally, it is demonstrated that autoantibodies are typi-
cally present many years before the diagnosis of type 1 dia-
betes or systemic lupus erythematosus (SLE) (74,75). Al-
though there are no prospective studies on the development
of TSH receptor antibodies before the onset of Graves’ dis-
ease, Graves’ disease is unlikely to be an exception to that
subclinical finding. Stress may not be a direct immunologic
initiator of Graves’ disease, but may be rather one of the trig-
gers for the development to overt Graves’ hyperthyroidism
in patients with subclinical disease.
There are many epidemiologic and clinical reports that
demonstrate an association between stress and Graves’ dis-
ease, but a direct influence of stress on the onset and course
of Graves’ disease remains to be clarified. On the other hand,
there are very few reports on the relationship between stress
and Hashimoto’s thyroiditis. Stress affects the immune sys-
tem both directly and indirectly through the activation of the
neural and endocrine systems. Those immune modulations
caused by various hormones, especially glucocorticoids, neu-
rotransmitters, and cytokines, can contribute to the devel-
opment of autoimmunity as well as susceptibility or resis-
tance to autoimmune thyroid disease.
STRESS AND AUTOIMMUNITY 1053
1. Pacák K, Palkovits M 2001 Stressor specificity of central neu-
roendocrine responses: Implications for stress-related disor-
ders. Endocr Rev 22:502–548.
2. Wilder RL 1995 Neuroendocrine-immune system interac-
tions and autoimmunity. Ann Rev Immunol 13:307–338.
3. Chrousos GP 1998 Stressors, stress, and neuroendocrine in-
tegration of the adaptive response. The 1997 Hans Selye
memorial lecture. Ann NY Acad Sci 851:311–335.
4. Cohen S, Herbert TB 1996 Health psychology: Psychologi-
cal factors and physical disease from the perspective
of human psychoneuroimmunology. Ann Rev Psychol
5. Chrousos GP, Gold PW 1992 The concepts of stress and
stress system disorders. Overview of physical and behav-
ioral homeostasis. JAMA 267: 1244–1252.
6. Leclère J, Weryha G 1989 Stress and auto-immune endocrine
diseases. Horm Res 31:90–93.
7. Rogers MP, Fozdar M 1996 Psychoneuroimmunology of au-
toimmune disorders. Adv Neuroimmunol 6:169–177.
8. Walker JG, Littlejohn GO, McMurray NE, Cutolo M 1999
Stress system response and rheumatoid arthritis: A multi-
level approach. Rheumatology (Oxf) 38:1050–1057.
9. Jedryka-Goral A, Maslinski W, Ziolkowska M, Konarska M,
Zolnierczyk-Zreda D 2002 Stress—Where are we now? Does
immunity play an intrinsic role? Autoimmunity 35:421–426.
10. Rosch PJ 1993 Stressful life events and Graves’ disease.
11. Chiovato L, Pinchera A 1996 Stressful life events and Graves’
disease. Eur J Endocrinol 134:680–682.
12. Dayan CM 2001 Stressful life events and Graves’ disease re-
visited. Clin Endocrinol (Oxf) 55:13–14.
13. Winsa B, Adami H, Bergström R, Gamstedt A, Dahlberg PA,
Adamson U, Jansson R, Karlsson A 1991 Stressful life events
and Graves’ disease. Lancet 338:1475–1479.
14. Sonino N, Girelli ME, Boscaro M, Fallo F, Busnardo B, Fava
GA 1993 Life events in the pathogenesis of Graves’ disease.
A controlled study. Acta Endocrinol (Copenh) 128:293–296.
15. Kung AW 1995 Life events, daily stresses and coping in
patients with Graves’ disease. Clin Endocrinol (Oxf)
16. Radosavljevi´ c VR, Jankovi´ c SM, Marinkovi´ c JM 1996 Stress-
ful life events in the pathogenesis of Graves’ disease. Eur J
17. Yoshiuchi K, Kumano H, Nomura S, Yoshimura H, Ito K,
Kanaji Y, Ohashi Y, Kuboki T, Suematsu H 1998 Stressful
life events and smoking were associated with Graves’ dis-
ease in women, but not in men. Psychosom Med 60:182–185.
18. Matos-Santos A, Nobre EL, Costa JGE, Nogueira PJ, Macedo
A, Galvao-Teles A, de Castro JJ 2001 Relationship between
the number and impact of stressful life events and the on-
set of Graves’ disease and toxic nodular goitre. Clin En-
docrinol (Oxf) 55:15–19.
19. Stewart T, Rochon J, Lenfestey R, Wise P 1985 Correlation
of stress with outcome of radioiodine therapy for Graves’
disease. J Nucl Med 26:592–599.
20. Benvenga S 1996 Benzodiazepine and remission of Graves’
disease. Thyroid 6:659–660.
21. Yoshiuchi K, Kumano H, Nomura S, Yoshimura H, Ito K,
Kanaji Y, Kuboki T, Suematsu H 1998 Psychosocial factors
influencing the short-term outcome of antithyroid drug
therapy in Graves’ disease. Psychosom Med 60:592–596.
22. Fukao A, Takamatsu J, Murakami Y, Sakane S, Miyauchi A,
Kuma K, Hayashi S, Hanafusa T 2003 The relationship of
psychological factors to the prognosis of hyperthyroidism in
antithyroid drug-treated patients with Graves’ disease. Clin
Endocrinol (Oxf) 58:550–555.
23. Paunkovic N, Paunkovic J, Pavlovic O, Paunovic Z 1998 The
significant increase in incidence of Graves’ disease in east-
ern Serbia during the civil war in the former Yugoslavia
(1992 to 1995). Thyroid 8:37–41.
24. Hadden DR, McDevitt EDG 1974 Environmental stress and
thyrotoxicosis. Absence of association. Lancet 2:577–578.
25. Chiovato L, Barbesino G, Pinchera A 2000 Graves’ disease.
In: DeGroot LJ, Jameson JL (eds) Endocrinology, 4th ed. WB
Saunders Company, Philadelphia, 2001, pp. 1422–1449.
26. Matsubayashi S, Tamai H, Matsumoto Y, Tamagawa K,
Mukuta T, Morita T, Kubo C 1996 Graves’disease after the
onset of panic disorder. Psychother Psychosom 65:277–280.
27. Chiovato L, Marinò M, Perugi G, Fiore E, Montanelli L, Lapi
P, Cavaliere R, Ciampi M, Patronelli A, Placidi G, Placidi
GF, Cassano GB, Pinchera A 1998 Chronic recurrent stress
due to panic disorder does not precipitate Graves’ disease.
J Endocrinol Invest 21:758–764.
28. Misaki T, Iwata M, Kasagi K, Iida Y, Akamizu T, Kosugi S,
Konishi J 2000 Hyperthyroid Graves’ disease after hemithy-
roidectomy for papillary carcinoma: Report of three cases.
Endocr J 47:191–195.
29. Morillo E, Gardner LI 1979 Bereavement as an antecedent
factor in thyrotoxicosis of childhood: Four case studies with
survey of possible metabolic pathways. Psychosom Med
30. Morillo E, Gardner LI 1980 Activation of latent Graves’ dis-
ease in children. Review of possible psychosomatic mecha-
nisms. Clin Pediatr (Phila) 19:160–163.
31. Gray J, Hoffenberg R 1985 Thyrotoxicosis and stress. Q J
32. Simon NM, Safren SA, Otto MW, Sharma SG, Lanka GD,
Pollack MH 2002 Longitudinal outcome with pharma-
cotherapy in a naturalistic study of panic disorder. J Affect
33. Martin-du Pan RC 1998 Triggering role of stress and preg-
nancy in the occurrence of 98 cases of Graves’ disease com-
pared to 95 cases of Hashimoto thyroiditis and 97 cases of
thyroid nodules. Ann Endocrinol (Paris) 59:107–112.
34. Harris T, Creed F, Brugha TS 1992 Stressful life events and
Graves’ disease. Br J Psychiatry 161:535–541.
35. Ben-Shlomo Y, Chaturvedi N 1992 Stress and Graves’ dis-
ease. Lancet 339:427.
36. Petticrew M 1992 Stress and Graves’ disease. Lancet 339:427.
37. Stern RA, Robinson B, Thorner AR, Arruda JE, Prohaska ML,
Prange AJ Jr 1996 A survey study of neuropsychiatric com-
plaints in patients with Graves’ disease. J Neuropsychiatry
Clin Neurosci 8:181–185.
38. Oretti RG, Harris B, Lazarus JH, Parkes AB, Crownshaw T
2003 Is there an association between life events, postnatal
depression and thyroid dysfunction in thyroid antibody pos-
itive women? Int J Soc Psychiatry 49:70–76.
39. Carter WR, Herrman J, Stokes K, Cox DJ 1987 Promotion
of diabetes onset by stress in the BB rat. Diabetologia
40. Durant S, Coulaud J, Amrani A, el Hasnaoui A, Dardenne
M, Homo-Delarche F 1993 Effects of various environmental
stress paradigms and adrenalectomy on the expression
of autoimmune type 1 diabetes in the non-obese diabetic
(NOD) mouse. J Autoimmunol 6:735–751.
41. Lehman CD, Rodin J, McEwen B, Brinton R 1991 Impact of
environmental stress on the expression of insulin-dependent
diabetes mellitus. Behav Neurosci 105:241–245.
MIZOKAMI ET AL.1054
42. Habib KE, Gold PW, Chrousos GP 2001 Neuroendocrinol-
ogy of stress. Endocrinol Metab Clin North Am 30:695–728.
43. Chrousos GP, Torpy DJ, Gold PW 1998 Interactions between
the hypothalamic-pituitary-adrenal axis and the female re-
productive system: Clinical implications. Ann Intern Med
44. Ferin M 1999 Stress and the reproductive cycle. J Clin En-
docrinol Metab 84:1768–1774.
45. Sternberg EM 2001 Neuroendocrine regulation of autoim-
mune/inflammatory disease. J Endocrinol 169:429–435.
46. Sapolsky RM, Romero LM, Munck AU 2000 How do glu-
cocorticoids influence stress responses? Integrating permis-
sive, suppressive, stimulatory, and preparative actions. En-
docr Rev 21:55–89.
47. Elenkov IJ, Chrousos GP 1999 Stress, cytokine patterns and
susceptibility to disease. Baillière’s Clin Endocrinol Metab
48. Elenkov IJ, Wilder RL, Chrousos GP, Vizi ES 2000 The sym-
pathetic nerve—An integrative interface between two su-
persystems: The brain and the immune system. Pharmacol
49. Green PG, Miao FJ, Strausbaugh H, Heller P, Janig W, Levine
JD 1998 Endocrine and vagal controls of sympathetically de-
pendent neurogenic inflammation. Ann NY Acad Sci
50. Rivier C, Rivest S 1991 Effect of stress on the activity of the
hypothalamic-pituitary-gonadal axis: Peripheral and central
mechanisms. Biol Reprod 45:523–532.
51. Da Silva JAP 1999 Sex hormones and glucocorticoids: In-
teractions with the immune system. Ann NY Acad Sci
52. Ahmed SA, Hissong BD, Verthelyi D, Donner K, Becker K,
Karpuzoglu-Sahin E 1999 Gender and risk of autoimmune
diseases: Possible role of estrogenic compounds. Environ
Health Perspect 107(Suppl 5):681–686.
53. Olsen NJ, Kovacs WJ 1996 Gonadal steroids and immunity.
Endocr Rev 17:369–384.
54. Cremaschi GA, Gorelik G, Klecha AJ, Lysionek AE, Genaro
AM 2000 Chronic stress influences the immune system
through the thyroid axis. Life Sci 67:3171–3179.
55. Silva JE 2000 Catecholamines and the sympathoadrenal sys-
tem in thyrotoxicosis. In: Braverman LE, Utiger RD (eds),
Werner and Ingbar’s The Thyroid: A Clinical and Funda-
mental Text, 8thed. Lippincott Williams & Wilkins, Phila-
delphia, pp. 642–651.
56. Dluby RG 2000 The adrenal cortex in thyrotoxicosis. In:
Braverman LE, Utiger RD (Eds), Werner and Ingbar’s The
Thyroid: 8th ed. Lippincott Williams & Wilkins, Philadel-
phia, pp. 637–641.
57. Kamilaris TC, DeBold CR, Johnson EO, Mamalaki E, List-
wak SJ, Calogero AE, Kalogeras KT, Gold PW, Orth DN 1991
Effects of short and long duration hypothyroidism and hy-
perthyroidism on the plasma adreno corticotropin and cor-
ticosterone responses to ovine corticotropin-releasing hor-
mone in rats. Endocrinology 128:2567–2576.
58. Silberman DM, Wald M, Genaro AM 2002 Effects of chronic
mild stress on lymphocyte proliferative response. Participa-
tion of serum thyroid hormones and corticosterone. Int Im-
59. Dorshkind K, Horseman ND 2000 The roles of prolactin,
growth hormone, insulin-like growth factor-I, and thyroid
hormones in lymphocyte development and function: In-
sights from genetic models of hormone and hormone re-
ceptor deficiency. Endocr Rev 21:292–312.
60. Biondi M, Picardi A 1999 Psychological stress and neuroen-
docrine function in humans: The last two decades of re-
search. Psychother Psychosom 68:114–150.
61. Herbert TB, Cohen S 1993 Stress and immunity in humans:
A meta-analytic review. Psychosom Med 55:364–379.
62. Cohen S, Miller GE, Rabin BS 2001 Psychological stress and
antibody response to immunization: A critical review of the
human literature. Psychosom Med 63:17–18.
63. Ligier S, Sternberg EM 1999 Neuroendocrine host factors
and inflammatory disease susceptibility. Environ Health
Perspect 107(Suppl 5):701–707.
64. Wick G, Hu Y, Schwarz S, Kroemer G 1993 Immunoendocrine
communication via the hypothalamo-pituitary-adrenal axis in
autoimmune diseases. Endocr Rev 14:539–563.
65. Hu Y, Dietrich H, Herold M, Heinrich PC, Wick G 1993 Dis-
turbed immuno-endocrine communication via the hypo-
thalamo-pituitary-adrenal axis in autoimmune disease. Int
Arch Allergy Immunol 102:232–241.
66. Tomer Y, Davies TF 2003 Searching for the autoimmune thy-
roid disease susceptibility genes: From gene mapping to
gene function. Endocr Rev 24:694–717.
67. Volpé R 1991 Graves’ disease: pathogenesis. In: Braverman
LE, Utiger RD (eds) Werner and Ingbar’s The Thyroid: A
Clinical and Fundamental Text, 6th ed. Lippincott Williams
& Wilkins, Philadelphia, pp. 648.
68. DeGroot LJ 1996 Graves’ disease and the manifestation of
thyrotoxicosis. In: DeGroot LJ, Larsen PR, Hennemann G
(eds) The Thyroid and its Diseases. Churchill Livingstone,
New York, pp. 371.
69. Amino N, Hidaka Y, Takano T, Izumi Y, Tatsumi K, Nakata Y
2003 Association of seasonal allergic rhinitis is high in Graves’
disease and low in painless thyroiditis. Thyroid 13:811–814.
70. Coles AJ, Wing M, Smith S, Coraddu F, Greer S, Taylor C,
Weetman A, Hale G, Chatterjee VK, Waldmann H, Comp-
ston A 1999 Pulsed monoclonal antibody treatment and
autoimmune thyroid disease in multiple sclerosis. Lancet
71. Nagayama Y, Mizuguchi H, Hayakawa T, Niwa M, McLach-
lan SM, Rapoport B 2003 Prevention of autoantibody-medi-
ated Graves’-like hyperthyroidism in mice with IL-4, a Th2
cytokine. J Immunol 170:3522–3527.
72. Davies TF 2000 Graves’ disease: pathogenesis. In: Braver-
man LE, Utiger RD (eds) Werner and Ingbar’s The Thyroid,
The Clinical and Fundamental Text, 8th ed. Lippincott
Williams & Wilkins, Philadelphia, pp. 518.
73. Amino N, Tada H, Hidaka Y, Izumi Y 2000 Postpartum au-
toimmune thyroid syndrome. Endocr J 47:645–655.
74. Arbuckle MR, McClain MT, Rubertone MV, Scofield RH,
Dennis GJ, James JA, Harley JB 2003 Development of au-
toantibodies before the clinical onset of systemic lupus ery-
thematosus. N Engl J Med 349:1526–1533.
75. Batstra MR, Aanstoot HJ, Herbrink P 2001 Prediction and
diagnosis of type 1 diabetes using beta-cell autoantibodies.
Clin Lab 47:497–507.
Address reprint requests to:
Jack R. Wall, M.D., Ph.D.
Department of Clinical and Biomedical Sciences: Barwon Health
The Geelong Hospital
PO Box 281
STRESS AND AUTOIMMUNITY 1055