Cumulative Childhood Stress and Autoimmune Diseases in Adults
SHANTA R. DUBE, PHD, MPH, DELISA FAIRWEATHER, PHD, WILLIAM S. PEARSON, PHD, MHA, VINCENT J. FELITTI, MD,
ROBERT F. ANDA, MD, MS, AND JANET B. CROFT, PHD
Objective: To examine whether childhood traumatic stress increased the risk of developing autoimmune diseases as an adult.
Methods: Retrospective cohort study of 15,357 adult health maintenance organization members enrolled in the Adverse Childhood
Experiences (ACEs) Study from 1995 to 1997 in San Diego, California, and eligible for follow-up through 2005. ACEs included
childhood physical, emotional, or sexual abuse; witnessing domestic violence; growing up with household substance abuse, mental
illness, parental divorce, and/or an incarcerated household member. The total number of ACEs (ACE Score range ? 0–8) was used
as a measure of cumulative childhood stress. The outcome was hospitalizations for any of 21 selected autoimmune diseases and 4
immunopathology groupings: T- helper 1 (Th1) (e.g., idiopathic myocarditis); T-helper 2 (Th2) (e.g., myasthenia gravis); Th2
rheumatic (e.g., rheumatoid arthritis); and mixed Th1/Th2 (e.g., autoimmune hemolytic anemia). Results: Sixty-four percent
reported at least one ACE. The event rate (per 10,000 person-years) for a first hospitalization with any autoimmune disease was
31.4 in women and 34.4 in men. First hospitalizations for any autoimmune disease increased with increasing number of ACEs (p ?
.05). Compared with persons with no ACEs, persons with ?2 ACEs were at a 70% increased risk for hospitalizations with Th1,
80% increased risk for Th2, and 100% increased risk for rheumatic diseases (p ? .05). Conclusions: Childhood traumatic stress
increased the likelihood of hospitalization with a diagnosed autoimmune disease decades into adulthood. These findings are
consistent with recent biological studies on the impact of early life stress on subsequent inflammatory responses. Key words:
childhood abuse, traumatic stress, autoimmune diseases, stress, inflammatory response.
ACE ? adverse childhood experience; AD ? autoimmune disease;
Th1 ? T-helper 1; Th2 ? T-helper 2; CRP ? C-reactive protein;
CRH ? corticoid releasing hormone.
proximately 3% to 8% of the population in the United States
(1). The prevalence of AD (14.7 to 23.5 million people) is
reported to be steadily increasing (1,2). ADs occur when the
immune response damages tissues in the body; ADs are clas-
sified according to the organ, tissue, or system targeted by the
immune response. Disease may present as early as the second
decade of life, but typically peaks in the third to sixth decade
(3). Infectious agents and environmental factors are most
commonly cited as etiologic factors (4,5). Because many ADs
begin at a relatively young age and are chronic and treatments
are palliative (2), they represent a significant personal and
economic burden to individuals and their families.
utoimmune diseases (ADs) are a heterogeneous group of
70 to 80 different inflammatory disorders affecting ap-
AD immune response has been classified as T-helper 1
(Th1), T-helper 17 (Th17), or T-helper 2 (Th2) depending on
the release of interferon (IFN)-?, interleukin (IL)-17, or IL-4
from CD4?T cells. However, both Th1 and Th2 ADs involve
cell-mediated and antibody-mediated responses (5,6). ADs
can be grouped according to the predominant immune mech-
anisms as follows: Th1, mixed Th1/Th2, and Th2; AD clas-
sically thought of as “rheumatoid” are classified as a subgroup
of Th2 diseases. The risk for ADs may increase with age due
to an increasing Th2 response and increasing numbers of
Conservative estimates indicate that approximately 80% of
the individuals with AD are women (2) because the basic
immune response differs between men and women. For ex-
ample, women respond to infection, immunization, or trauma
with higher antibody production whereas inflammation is usu-
ally more severe in men (6,8–11). Sex differences in AD are
also likely to be linked to sex-specific differences in glucocor-
ticoid responses to stress because glucocorticoids decrease
cell-mediated Th1-type immunity in response to acute stress
(7). Estrogen transcriptionally upregulates glucocorticoid lev-
els in females whereas testosterone decreases glucocorticoid
levels in males (7,12–13).
The long-term health effects of childhood traumatic stress
are well documented. For example, childhood abuse, neglect,
and related forms of household dysfunction increase the risk
of substance abuse, mental illness, sexually transmitted dis-
eases, suicide attempts, and other health outcomes, such as
ischemic heart disease (14–35). To date, there have been few
studies that examine the contribution of childhood traumatic
stress to the risk of developing an AD. A recent study by
Danese and colleagues examined the association between
childhood trauma and C-reactive protein (CRP), a biomarker
of inflammation that may play a role in AD (36,37). After
controlling for current stress, they reported that childhood
maltreatment was associated with elevated CRP levels in
adults 20 years later, suggesting that childhood maltreatment
independently increases inflammation later in life (36).
From National Center for Chronic Disease Prevention and Health Promo-
tion (S.R.D., W.S.P. R.F.A., J.B.C.), Centers for Disease Control and
Prevention, Division of Adult and Community Health, Atlanta, Georgia;
Department of Environmental Health Sciences (D.F.), Bloomberg School of
Public Health and Department of Pathology, School of Medicine, Johns
Hopkins University, Baltimore, Maryland; and the Department of Preventive
Medicine (V.J.F.), Southern California Permanente Medical Group (Kaiser
Permanente), San Diego, California.
Address correspondence and reprint requests to Shanta R. Dube, Centers
for Disease Control and Prevention, National Center for Chronic Disease
Prevention and Health Promotion, Division of Adult and Community Health,
4770 Buford Highway, N.E., MS K-50, Atlanta, GA 30341-3717. E-mail:
Received for publication May 16, 2008; revision received July 28, 2008.
The findings and conclusions in this article are those of the authors and do
not necessarily represent the views of the Centers for Disease Control and
The Adverse Childhood Experiences Study was supported under coopera-
tive agreement #TS-44 to 10/11 from the Centers for Disease Control and
Prevention through the Association of Prevention Teaching and Research and
a grant from the Garfield Memorial Fund. This study was also partially
supported through a sole source contract (#200-2005-M-13275) with the
Kaiser Foundation Research Institute. Dr. Fairweather is supported by Grant
R01 HL087033 from the National Heart, Lung and Blood Institute.
243Psychosomatic Medicine 71:243–250 (2009)
Copyright © 2009 by the American Psychosomatic Society
Published Ahead of Print on February 1, 2009 as 10.1097/PSY.0b013e3181907888
Using longitudinal data from the Adverse Childhood Ex-
periences (ACEs) Study, we created an ACE Score, using an
integer count of eight interrelated and co-occurring (24,34,35)
exposures of childhood adversity, to measure cumulative
childhood traumatic stress. The relationship between the ACE
Score and the risk of 21 different ADs as an adult was
examined. We hypothesized that as the cumulative exposure
to childhood stress and trauma increased the risk of develop-
ing AD in adulthood would also increase.
The ACE Study has been described elsewhere (14–35). More than 50,000
adult members of the Kaiser Foundation Health Plan in San Diego, California
are evaluated annually at Kaiser Permanente’s San Diego Health Appraisal
Clinic (HAC). All Health Plan members who visit the HAC complete a
standardized evaluation that included assessment of health history, health-
related behaviors, a clinical review of systems, and psychosocial evaluations
(14). Health plan members who completed the standardized HAC evaluation
during 1995 to 1997 were eligible to participate in the ACE Study. They were
mailed an ACE Study questionnaire with items about childhood exposure to
abuse, neglect, domestic violence, and other related forms of serious house-
hold dysfunction (14). The survey was administered in two waves. Seventy
percent (n ? 9508) of adult members who were surveyed in Wave I (August
through November 1995 and January through March 1996) and 65% (n ?
8667) of adults who were surveyed in Wave 2 (June through October 1997)
responded. Because 754 plan members inadvertently were mailed the survey
twice, they were removed, resulting in a sample size of 17,421 at baseline. Of
the 17,421 participants at baseline, 84 persons had incomplete information on
race and educational attainment, leaving an analytic sample of 17,337 per-
sons. The study was approved by the Institutional Review Boards of Kaiser
Permanente and the Office for Protection from Research Risks at the National
Institutes of Health.
The present analyses are based on follow-up medical record data available
through December 31, 2005. Of the 17,337 participants included in prior
analysis of the baseline data, 724 (4.2%) were excluded from the prospective
phase of the study because the administrative enrollment database showed
that their health maintenance organization (HMO) membership lapsed before
their evaluation at the HAC or the member record number obtained from their
clinic appointment was not considered valid. It is possible that persons might
have disenrolled from the Kaiser Health Plan and then reenrolled back into the
health plan during the follow-up period. To account for this occurrence, we
calculated ratio of time disenrolled/total possible time enrolled. For persons
who disenrolled and reenrolled at least once (median/mean ? 1 time; range ?
1–9 times) during the follow-up period, we excluded 1248 (7.2%) persons
whose ratio of time disenrolled/total possible time enrolled was ?80%.
Although this “80% cut off” may be considered arbitrary, we considered such
persons to have inadequate continuity of follow-up to merit consideration for
inclusion in the prospective analysis. Thus, 15,365 of persons from the
baseline sample were included in this analysis. An additional eight observa-
tions were deleted because their hospital admission date occurred before their
baseline appointment date. Thus, the final sample size for analyses was
Questions used to define ACEs are listed in Table 1. All questions about
ACEs pertained to the respondent’s first 18 years of life (?18 years of age).
For questions adapted from the Conflict Tactics Scale (38), there were five
response categories: (“never,” “once or twice,” “sometimes,” “often,” or
“very often”). We defined the following three types of childhood abuse:
emotional abuse (2 questions) (38), physical abuse (2 questions) (38), or
contact sexual abuse (4 questions) (39). In addition, we defined five exposures
to household dysfunction during childhood which included household expo-
sure to substance abuse (defined by 2 questions) (40), mental illness (2
questions), violent treatment of female caretaker (4 questions) (38), criminal
behavior in the household (1 question), and parental separation or divorce
Case findings for AD were based on hospitalizations after the baseline
appointment date (prospective phase) with identified International Classifica-
tion of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes
for 21 ADs in hospital discharge data. Research recommendations and pre-
vious studies suggested that AD should be considered as a set of illnesses
when taking a public health approach to studying the diseases (1,2,41). For the
present study, AD included the following 21 illnesses identified by a recent
National Institutes of Health report (1): Addison’s disease, autoimmune
hemolytic anemia, autoimmune thrombocytopenia purpura, celiac disease,
dermatomyositis, Graves’ disease, Hashimoto’s thyroiditis, idiopathic myo-
carditis, idiopathic pulmonary fibrosis, insulin-dependent diabetes mellitus,
irritable bowel disease, multiple sclerosis, myasthenia gravis, pernicious
anemia, psoriasis, rheumatoid arthritis, scleroderma, Sjogren disease, sys-
temic lupus erythematosus, vitiligo, and Wegener’s granulomatosis. Persons
were considered at risk for hospitalization with ICD-9 coded AD as a
discharge diagnosis if they were currently enrolled in the HMO at the time of
the hospitalization and the hospitalization with ICD-9 coded AD occurred
between the baseline survey date and December 31, 2005.
Because various ADs have different immunopathology, we also examined
the relationship between the ACE Score and specific AD groupings (8). One
AD group with a cell-mediated, Th1-type immunopathology, which is more
prevalent in men, included the following: idiopathic myocarditis, idiopathic
pulmonary fibrosis, insulin-dependent diabetes mellitus, irritable bowel dis-
ease, and Wegener’s granulomatosis (7,8,42–44). Another AD group with an
antibody-mediated, Th2-type immunopathology that is more prevalent in
women included the following: autoimmune thrombocytopenia purpura, der-
matomyositis, Graves’ disease, Hashimoto’s thyroiditis, myasthenia gravis,
rheumatoid arthritis, scleroderma, Sjogren disease, and systemic lupus ery-
thematosus (42–44). A subgroup of the Th2 group included specifically
rheumatic diseases (dermatomyositis, rheumatoid arthritis, scleroderma,
Sjogren disease, and systemic lupus erythematosus) (43,44). The remaining
ADs have a mixed Th1/Th2 pathology without a clear difference in preva-
lence between men and women (43,44).
Previous reports from the ACE Study have shown that forms of childhood
abuse and household dysfunction tend to be interrelated and co-occur
(24,34,35). Therefore, to examine ACEs as cumulative stress, we created the
ACE Score (range ? 0–8), which is an integer count of the total number of
categories of ACEs reported by respondents. Four levels of ACE exposures
(0, 1, 2, or ?3) were used for examining hospitalizations with any of 21 ADs.
Because of the relatively limited number of ADs when examining specific
immunopathology types, we also used a three-level ACE Score (0, 1, or ?2).
Analyses were conducted with the summed score as three dichotomous
variables (yes/no) with zero ACE exposures as the referent for examination of
any of the 21 ADs and two dichotomous variables with zero ACE as the
referent for examination of specific immunopathology types. To test for a
trend (graded relationship) between the ACE Score and the risk of AD, the
ACE Score was also entered as an ordinal variable with adjustment for the
demographic covariates (sex, age, and race). The statistical validity and
test-retest reliability of the ACE Score has been published previously; Co-
hen’s ? was found to be 0.64, representing good agreement (45).
Cox proportional hazards regression models were used to estimate the
hazard ratio (HR) and 95% confidence interval (CI) for the likelihood of a first
hospitalization for AD during follow-up. Survival time in years was calcu-
lated from the date of the baseline survey until the first hospitalization for any
AD among AD cases or until December 31, 2005 for participants who
remained free of AD; no participants died during follow-up. Independent
variables in the models for any AD included age at baseline, race (White,
non-White), and the ACE Score, which all satisfied the proportional hazards
assumptions. For specific AD immunology types as the outcomes, the pro-
S. R. DUBE et al.
244Psychosomatic Medicine 71:243–250 (2009)
portional hazards assumption was satisfied for age, sex, and ACE Score;
however, race failed to satisfy the assumption for Th2 and Th2 rheumatic
diseases and was removed from the models. In addition, stratification by two
age categories (19 to 64 years and ?65 years) was conducted because ADs
tend to peak during the third and sixth decade of life (3).
Characteristics of the Study Population
The study population included 8293 women (54%) and 7064
men (46%). The mean ? standard deviation age was 56 ? 15
years; 23.9% were aged 19 to 44 years, 43.5% were aged 43 to
64 years, and 32.7% were aged ?65 years. Seventy-six percent
of participants were White, and 11% Hispanic, 4% Black, 7%
Asian, ?1% Native American, and 2% other. This was a highly
educated population such that 40% were college graduates; 36%
had some college education; 17% were high school graduates
(i.e., 12 years education); and only 7% had not completed high
The prevalence of each of the eight individual ACEs was as
follows: emotional abuse, 10%; physical abuse, 28%; sexual
abuse, 21%; household substance abuse, 27%; mental illness
in the home, 19%; witnessed domestic violence, 13%; crimi-
nal household member, 5%; and parental separation or di-
vorce, 23% (Table 1). Close to two thirds (64%) of the
respondents reported at least one ACE; 37% reported ?2
ACEs (Table 1).
Overall 372 (2.4%) of hospitalizations with a discharge
diagnosis of AD were identified during follow-up. Among the
hospital records with an AD diagnosis, 317 (85%) listed a
single AD as a discharge diagnosis whereas 15% listed more
than one AD (Table 2). The five most common ADs identified
were: insulin-dependent diabetes mellitus (23.1%), rheuma-
toid arthritis (18.8%), autoimmune thrombocytopenia purpura
TABLE 1. Definition and Baseline Prevalence of Categories of Adverse Childhood Experiences Reported by Adults Aged >19 Years by Sex:
Adverse Childhood Experiences Study 1995 to 1997
Adverse Childhood Experiences (ACEs)
(n ? 8293)
(n ? 7064)
(n ? 15,357)
Did a parent or other adult in the household . . .
1) Often or very often swear at you, insult you, or put you down?
2) Sometimes, often, or very often act in a way that made you
that you might be physically hurt?
Did a parent or other adult in the household . . .
1) Often or very often push, grab, slap, or throw something at
2) Ever hit you so hard that you had marks or were injured?
Did an adult or person at least 5 years older ever . . .
1) Touch or fondle you in a sexual way?
2) Have you touch their body in a sexual way?
3) Attempt oral, anal, or vaginal intercourse with you?
4) Actually have oral, anal, or vaginal intercourse with you?
Substance abuse in household
1) Live with anyone who was a problem drinker or alcoholic?
2) Live with anyone who used street drugs?
Mental illness in household
1) Was a household member depressed or mentally ill?
2) Did a household member attempt suicide?
Mother treated violently
Was your mother or stepmother . . .
1) Sometimes, often, or very often pushed, grabbed, slapped, or
had something thrown at her?
2) Sometimes, often, or very often kicked, bitten, hit with a fist,
or hit with something hard?
3) Ever repeatedly hit over at least a few minutes?
4) Ever threatened with or hurt by a knife or gun?
Incarcerated household member
1) Did a household member go to prison?
Parental separation or divorce
1) Were your parents ever separated or divorced?
29.2 23.6 26.7
CHILDHOOD STRESS AND AUTOIMMUNE DISEASES
245 Psychosomatic Medicine 71:243–250 (2009)
(16.7%), idiopathic pulmonary fibrosis (9.1%), and systemic
lupus erythematosus (8.1%).
The unadjusted rates (per 10,000 person-years) for AD
hospitalizations increased with increasing age (8.8 for ages 19 to
44 years; 26.0 for ages 45 to 64 years, and 56.8 for ages ?65
years (p ? .05)). Because of the expected sex differences in
the prevalence of certain ADs (8,42–44), we examined the
age-specific relationship between sex and any AD and found
that, for persons aged 19 to 64 years, women were 50% more
likely than men to be hospitalized with an AD (HR ? 1.5;
95% CI ? 1.1–2.2); among persons ?65 years, women were
less likely than men to have a first hospitalization with AD
(HR ? 0.7; 95% CI ? 0.5–0.9). There were no race differ-
ences in hospitalizations for ADs.
ACE Score and Risk of Hospitalizations for Any AD
We examined the sex-specific relationship between ACEs
and the likelihood of hospitalization with any AD (Table 3).
For both men and women, the likelihood of a first hospital-
ization for any AD was higher among adults with 2 or ?3
ACEs compared with those with no ACE; however, the rela-
tionship was statistically significant only for women (p ? .05)
(Table 3). A test for linear trend was also performed, using the
ACE Score as an ordinal variable in the models. For every
increase in the ACE Score, the likelihood of a first hospital-
ization with any AD increased 20% (p ? .001) for women and
10% for men (p ? .05).
When stratified by age, the relationship between the ACE
Score and a hospitalization with any of the 21 ADs was
stronger among adults aged 19 to 64 years than it was for the
cohort aged ?65 years (Figure 1). A test for linear trend was
performed, using the ACE Score as an ordinal variable in the
models. The results indicated that, for every increase in the
ACE Score, the likelihood of a hospitalization with AD in-
creased 20% for persons aged 19 to 64 years (p ? .05)
whereas it increased only 10% for ages ?65 years (p ? .08).
TABLE 2. Autoimmune Disease Grouping by Immunopathology and Number of Cases Identified Through Hospital Discharge Records Between
Baseline and December 31, 2005 Among 15,357 Adults by Sex: Adverse Childhood Experiences Study 1995 to 2005
Autoimmune Disease (ICD-9 Code)Target Organ, Tissue, or Receptor
Number of Cases
Autoimmune diseases associated with cell-
mediated, Th1-type immunopathology and/
or increased incidence in males
Idiopathic myocarditis (422.91)
Idiopathic pulmonary fibrosis (515.0)
Insulin-dependent diabetes mellitus (250.01,
250.03, 250.13, 250.23, 250.41, 250.51,
Irritable bowel disease (556.0 to 556.9)
Wegner granulomatosis (446.4)
Autoimmune diseases with mixed Th1/ Th2
Pernicious anemia (281.0)
Addison disease (255.4)
Autoimmune hemolytic anemia (283.0)
Celiac disease (579.0)
Multiple sclerosis (340.0)
Autoimmune diseases associated with antibody-
mediated, Th2-type immunopathology and/
or increased incidence in females (includes
rheumatic diseases below)
Graves’ disease (242.0)
Autoimmune thrombocytopenia purpura
Myasthenia gravis (358.0)
Hashimoto’s thyroiditis (245.2)
Th2-type rheumatic diseases subgroup
Systemic lupus erythematosus (710.0)
Sjogren disease (710.2)
Rheumatoid arthritis (714.0)
Red blood cells
ICD-9 ? International Classification of Diseases, Ninth Revision; Th1 ? T-helper 1; Th2 ? T-helper 2.
Among the 372 hospital records with any diagnosis of an autoimmune disease, 15% include more than one autoimmune disease.
S. R. DUBE et al.
246 Psychosomatic Medicine 71:243–250 (2009)
Specific Autoimmune Groupings
Women were 40% less likely than men (p?.05) to have
hospitalizations for ADs associated with Th1-type immuno-
pathology (Table 4). However, women had a 50% greater
(p ? .05) likelihood than men for hospitalization with a
Th2-type AD (Table 4) and were also more than twice as
likely as men to have a hospitalization for a Th2 specific
rheumatic disease (HR ? 2.5; 95% CI ? 1.6–3.9) (Table 4).
Compared with persons with no ACEs, those with ?2 ACEs
were at a 70% increased likelihood for hospitalization with
Th1-type ADs (p ? .05), 80% increased likelihood for Th2
type ADs (p ? .05), and twice the likelihood for hospitaliza-
tion with Th2 rheumatic diseases (p ? .05), after controlling
for sex and age (Table 4). The test for linear trend revealed a
20% increased risk for Th1-type AD, 20% increased risk for
Th2-type AD, and 30% increased risk for rheumatic diseases
for every level of increase in the ACE Score. No significant
relationship between ACEs and hospitalizations with mixed
Th1/Th2-type AD was observed.
We had the unique opportunity to examine the relationship
between ACEs and the likelihood of hospitalizations with AD
identified through hospital discharge records. As the number
of ACEs increased, the likelihood of hospitalizations with
Th1, Th2, Th2-rheumatic, and any of 21 types of ADs also
increased. Moreover, the relationship between the ACE Score
and AD hospitalizations was stronger among younger adults.
To our knowledge, this study is the first to demonstrate a
relationship between multiple types of childhood adversity on
hospitalizations for AD during adulthood.
Research on nervous, endocrine, and immune interactions
has revealed that these systems are anatomically and functionally
interconnected (11–13). Stressors, such as infections, toxins,
and/or psychological trauma, stimulate the hypothalamic-pi-
tuitary-adrenal axis to release corticoid-releasing hormone
(CRH), resulting in elevated systemic levels of corticoste-
roids, such as glucocorticoids. Acute stress initially may in-
crease inflammation through acute-phase mediators like IL-1,
IL-6, and CRP that are eventually downregulated by glucocor-
ticoids thereby maintaining homeostasis (46). Chronic stress
has the opposite effect and decreases glucocorticoid levels. A
recent epidemiologic study confirmed the link between child-
hood abuse and long-term changes in immune response (36);
in this longitudinal study, childhood abuse was associated
with elevated CRP levels, white blood cell counts, and other
markers of inflammation 20 years later (36).
Women and female rodents have higher systemic baseline
levels of glucocorticoids than their male counterparts due to the
transcriptional regulation of CRH by estrogen (11). In a sim-
ilar manner, estrogen increases IL-4 levels in females, result-
ing in a greater Th2-type immune response (7). Increased
TABLE 3.Association Between ACE Score and First Hospitalization of Any of 21 Autoimmune Diseases Identified Through Hospital Discharge
Records Between Baseline and December 31, 2005 Among 15,357 Adults by Sex: Adverse Childhood Experiences Study 1995 to 2005
Autoimmune Diseases (No.)a
(per 10,000 Person-Years)
ACE ? adverse childhood experience; HR ? hazard ratio; CI ? confidence interval.
aAutoimmune diseases: Addison’s disease, autoimmune hemolytic anemia, autoimmune thrombocytopenia purpura, celiac disease, dermatomyositis, Graves’
disease, Hashimoto’s thyroiditis, idiopathic myocarditis, idiopathic pulmonary fibrosis, insulin-dependent diabetes mellitus, irritable bowel disease, multiple
sclerosis, myasthenia gravis, pernicious anemia, psoriasis, rheumatoid arthritis, scleroderma, Sjogren disease, systemic lupus erythematosus, vitiligo, and
bHR and 95% CI obtained from sex-specific Cox proportional hazards regression model that included ACE score, age, and race.
association of adverse childhood events with the development of any auto-
immune disease, by age groups: Adverse Childhood Experiences Study 1995–
Adjusted hazard ratios and 95% confidence intervals for the
CHILDHOOD STRESS AND AUTOIMMUNE DISEASES
247Psychosomatic Medicine 71:243–250 (2009)
glucocorticoid levels in females further enhance IL-4 produc-
tion (11–13). In contrast, testosterone reduces glucocorticoid
and IL-4 levels in males, resulting in a predominantly IFN-?,
Th1-type immune response to infection or trauma (11). Re-
search in rodents (6,42,47,48) and the observed immunopa-
thology and sex prevalence of certain ADs (8,43,44) suggests
that sex-specific immune mechanisms may account for the
observed differences. In our study, hospitalizations for Th1-
associated ADs were more common in men in this cohort,
whereas hospitalizations for Th2-associated ADs, such as
rheumatoid disorders, were more common in women. This is
the first epidemiological study confirming this immunopatho-
Rheumatic diseases are a group of inflammatory disorders
in which autoantibodies and immune complex deposition pro-
duce tissue damage (44). One of the characteristics of rheu-
matic diseases is the production of rheumatoid factor (RF),
which is an autoantibody that binds other antibodies. RF is
usually produced following viral infections (44), suggesting
that infections may contribute to the development of AD (49).
Fairweather and Frisancho-Kiss have found that social stress
occurring before viral infection in rodents increased inflam-
matory heart disease in both sexes, but especially in females
(unpublished results). These findings and the present study
suggest that childhood stressful events may increase ADs
independently as well as amplify the effect of other environ-
mental factors, such as infections. Thus, a possible explana-
tion for the increased prevalence of ADs in females is that
females respond to similar stressful events differently than
males due to sex differences in their physiology and neurobi-
ology (i.e., greater Th2 and glucocorticoid levels that are
further amplified by stress) (13,50).
In addition, physiological and anatomical changes in the
brains of individuals who have experienced childhood abuse
have been documented. For example, Teicher et al. conducted
electroencephalograms to measure limbic irritability and
found the percentage of clinically significant brain-wave ab-
normalities to be higher among individuals who had a history
of early trauma versus those who did not experience early
trauma (51). Magnetic resonance imaging has revealed reduc-
tions in hippocampal volumes among severely sexually
abused women, and reductions in the intracranial and cerebral
volumes among maltreated children compared with non-
abused individuals (51–53).
TABLE 4. Associations of Sex and Baseline ACE Score to First Hospitalization With Selected Autoimmune Disease (Immunopathology) Types
Among 15,357 adults: Adverse Childhood Experiences Study 1995 to 2005
Autoimmune Diseases (n)
(per 10,000 Person-Years)
ACE ? adverse childhood experience; HR ? hazard ratio; CI ? confidence interval; Th1 ? T-helper 1; Th2 ? T-helper 2.
aTh1 ? idiopathic myocarditis, idiopathic pulmonary fibrosis, insulin-dependent diabetes mellitus, irritable bowel disease, and Wegener granulomatosis.
bTh2 ? autoimmune thrombocytopenia purpura, dermatomyositis, Graves’ disease, Hashimoto’s thyroiditis, myasthenia gravis, rheumatoid arthritis, sclero-
derma, Sjogren disease, and systemic lupus erythematosus.
cMixed Th1/Th2 ? pernicious anemia, vitiligo, Addison disease, autoimmune hemolytic anemia, psoriasis, celiac disease, multiple sclerosis.
dTh2 rheumatic subgroup of Th2 ? dermatomyositis, rheumatoid arthritis, scleroderma, Sjogren disease, and systemic lupus erythematosus.
eCox proportional hazard regression model included sex, age, race, and ACE score for analyses of Th1 and mixed Th1/Th2.
fRace excluded from the model because it did not meet the proportional hazards assumption for total Th2 and the Th2 rheumatic subgroup.
S. R. DUBE et al.
248 Psychosomatic Medicine 71:243–250 (2009)
Although the effects cannot be defined to any specific area
of the brain, it has been shown that the limbic system, which
is responsible for emotional response, is adversely affected.
Because ACEs rarely occur in isolation (14,15), the cumula-
tive effect of multiple ACEs shown in our study may have an
even more powerful negative effect on a young child’s devel-
oping brain via repeated activation of the stress response. This
repeated “dosing” of the developing central nervous system by
adrenal catecholamines and corticosteroids may contribute to
central nervous system- and endocrine-mediated differences
in immune function that result in an increased risk for AD.
There are several limitations to the present study. Inflam-
matory biomarkers, such as CRP or white blood cell counts,
were not compared with ACE Scores because biological sam-
ples were not available. Our data cannot provide certainty
about the temporal relationship between stress exposure
(ACEs) and AD, because of the lack of information about the
age at which ACEs occurred and also lack of information on
the age at onset of the AD. However, given the age at onset for
most ADs, it is likely that the ACEs antedated disease onset in
most cases. Also, the lack of information about the exact age
at which ACEs occurred potentially limits any specific infer-
ences that might be made about the developmental pathway of
ACEs on AD.
There are potential limitations with retrospective reporting
of childhood experiences. Respondents may have difficulty
recalling certain events. However, longitudinal follow-up of
adults whose childhood abuse was documented has shown that
their retrospective reports of childhood abuse are likely to
underestimate actual occurrence (54,55). Difficulty recalling
childhood events likely results in misclassification (classify-
ing persons truly exposed to ACEs as unexposed) that would
bias our results toward the null (56). This potential weakness
may result in underestimates of the true strength of the rela-
tionships between ACEs and hospitalization for AD.
ADs were identified through hospitalizations and not out-
patient data. Future studies may be strengthened through the
use of clinical data because most ADs are diagnosed through
outpatient visits. Finally, it is possible that other unmeasured
factors at the time of hospitalizations that were not included in
our analyses could have affected the strength of our estimates
(either upward or downward) of association between ACEs
and ADs. Despite this weakness, our data provide preliminary
evidence of the association between ACEs and ADs.
It is important to note that the prevalence estimates for
childhood exposures we report are nearly identical to those
reported in surveys of the general population. We found that
16% of the men and 25% of the women met the case definition
for contact sexual abuse; a national telephone survey of adults
in 1990 conducted by Finkelhor et al., using similar criteria,
estimated that 16% of men and 27% of women had been
sexually abused (57). As for physical abuse, 28% of the men
from our study had experienced this abuse as boys, which
closely parallels the percentage (31%) found in a population-
based study of Ontario men that used questions from the same
scales (58). The similarity of the estimates from the ACE
study to those of population-based studies suggests that our
findings are likely to be applicable in other settings.
This is the first study to find an association between early
childhood stressors and the development of AD decades later.
Our epidemiological findings, coupled with the documented
immunopathology of AD, provide preliminary evidence of the
relationship between early childhood stress with the human
physiological and immunological response, which may also
contribute to and expand on the theory of developmental
origins of adult disease and health (59). Because childhood
adverse events are common and ADs are chronic and often
debilitating, expansion of research in this area may further
elucidate the impact of stress on adult chronic disorders such
1. United States Department of Health and Human Services. Progress in
Autoimmune Diseases Research, Report to Congress, National Institutes
of Health, The Autoimmune Diseases Coordinating Committee, March
2005. Publication No. 05-5140.
2. Jacobsen DL, Gange SJ, Rose NR, Graham NMH. Epidemiology and
estimated population burden of selected autoimmune diseases in the
United States. Clin Immunol Immunopathol 1997;84:223–243.
3. Rosen, A. Autoimmunity. In: Warrell DA, Cox TM, Firth JD, eds. Oxford
Textbook of Medicine. 4thed. New York, NY: Oxford University Press,
4. Dooley MA, Hogan SL. Environmental epidemiology and risk factors for
autoimmune disease. Curr Opin Rheumatol 2003;15:99–103.
5. Fairweather D. Autoimmune disease: mechanisms. In: Encyclopedia of
Life Sciences. Chichester, UK: John Wiley & Sons, Ltd; January 2007:
http://www.els.net/ [DOI: 10.1002/9780470015902.a0020193].
6. Frisancho-Kiss S, Davis SE, Nyland JF, Frisancho JA, Cihakova D, Rose
NR, Fairweather D. Cutting Edge: Cross-regulation by TLR4 and T cell
Ig mucin-3 determines sex differences in inflammatory heart disease.
J Immunol 2007;178:6710–6714.
7. Beeson PB. Age and sex associations of 40 autoimmune diseases.
Am J Med 1994;96:457–462.
8. Fairweather D, Frishancho-Kiss S, Rose NR. Sex differences in autoim-
mune diseases from a pathological perspective. Americal Journal of
Pathology 2008; 173:600–609.
9. Lang TJ. Estrogen as an immunomodulator. Clin Immunol 2004;113:
10. Kher A, Wang M, Tsai BM, Pitcher JM, Greenbaum ES, Nagy RD, Patel
KM, Wairiuko GM, Markel TA, Meldrum DR. Sex differences in the
myocardial inflammatory response to acute injury. Shock 2005;23:1–10.
11. Cutolo M, Capellina S, Sulli A, Serioli B, Secchi ME, Villaggio B, Straub
RH. Estrogens and autoimmune diseases. Ann NY Acad Sci 2006;1089:
12. Elenkov IJ, Chrousos GP. Stress hormones, proinflammatory and anti-
inflammatory cytokines and autoimmunity. Ann NY Acad Sci 2002;966:
13. Viau V. Functional cross-talk between the hypothalamic-pituitary-
gonadal and –adrenal axes. J Neuroendocrinol 2002;14:506–513.
14. Felitti VJ, Anda RF, Nordenberg D, Williamson DF, Spitz AM, Edwards
V, Koss MP, Marks JS. The relationship of adult health status to child-
hood abuse and household dysfunction. Am J Prev Med 1998;14:
15. Anda RF, Croft JB, Felitti VJ, Nordenberg D, Giles WH, Williamson DF,
Giovino GA. Adverse childhood experiences and smoking during ado-
lescence and adulthood. J Am Med Assoc 1999;82:1652–1658.
16. Dietz PM, Spitz AM, Anda RF, Williamson DF, McMahon PM, Santelli
JS, Nordenberg DF, Felitti VJ, Kendrick JS. Unintended pregnancy
among adult women exposed to abuse or household dysfunction during
their childhood. J Am Med Assoc 1999;82:1359–1364.
17. Hillis SD, Anda RF, Felitti VJ, Nordenberg D, Marchbanks PA. Adverse
childhood experiences and sexually transmitted diseases in men and
women: a retrospective study. Pediatrics 2000;106:E11.
18. Hillis SD, Anda RF, Felitti VJ, Marchbanks PA. Adverse childhood
CHILDHOOD STRESS AND AUTOIMMUNE DISEASES
249 Psychosomatic Medicine 71:243–250 (2009)
experiences and sexual risk behaviors in women: a retrospective cohort Download full-text
study. Fam Plann Perspect 2001;33:206–211.
19. Anda RF, Felitti VJ, Chapman DP, Croft JB, Williamson DF, Santelli J,
Dietz PM, Marks JS. Abused boys, battered mothers, and male involve-
ment in teen pregnancy. Pediatrics 2001;107:e19.
20. Dube SR, Anda RF, Felitti VJ, Chapman D, Williamson DF, Giles WH.
Childhood abuse, household dysfunction and the risk of attempted suicide
throughout the life span: findings from Adverse Childhood Experiences
Study. J Am Med Assoc 2001;286:3089–3096.
21. Dube SR, Anda RF, Felitti VJ, Edwards VJ, Croft JB. Adverse childhood
experiences and personal alcohol abuse as an adult. Addict Behav 2002;
22. Williamson DF, Thompson, TJ, Anda, RF, Dietz, WH, Felitti VJ. Body
weight, obesity, and self-reported abuse in childhood. Int J Obes Relat
Metab Disord 2002;26:1075–1082.
23. Anda RF, Whitfield CL, Felitti VJ, Chapman D, Edwards VJ, Dube SR,
Williamson DF. Adverse childhood experiences, alcoholic parents, and
later risk of alcoholism and depression. Psychiatr Serv 2002;53:
24. Dube SR, Felitti VJ, Dong M, Chapman DP, Giles WH, Anda RF.
Childhood abuse, neglect, and household dysfunction and the risk of
illicit drug use: the Adverse Childhood Experience Study. Pediatrics
25. Dube SR, Felitti VJ, Dong M, Giles WH, Anda RF. The impact of
adverse childhood experiences on health problems: evidence from four
birth cohorts dating back to 1900. Prev Med 2003;37:268–277.
26. Dong M, Anda RF, Dube SR, Felitti VJ, Giles WH. Adverse childhood
experiences and self-reported liver disease: new insights into a causal
pathway. Arch Intern Med 2003;163:1949–1956.
27. Hillis SD, Anda RF, Dube SR, Felitti VJ, Marchbanks PA, Marks JS. The
association between adolescent pregnancy, long-term psychosocial out-
comes, and fetal death. Pediatrics 2004;113:320–327.
28. Dong M, Giles WH, Felitti VJ, Dube, SR, Williams JE, Chapman DP,
Anda RF. Insights into causal pathways for ischemic heart disease: the
Adverse Childhood Experiences Study. Circulation 2004;110:
29. Chapman DP, Anda RF, Felitti VJ, Dube SR, Edwards VJ, Whitfield CL.
Epidemiology of adverse childhood experiences and depressive disorders
in a large health maintenance organization population. J Affect Disord
30. Dube SR, Anda RF, Whitfield, CL, Brown DW, Felitti VJ, Dong M,
Giles WH. Long-term consequences of childhood sexual abuse by gender
of victim. Am J Prev Med 2005;28:430–438.
31. Dube SR, Miller JW, Brown DW, Giles WH, Felitti VJ, Dong M, Anda
RF. Adverse childhood experiences and the association with ever using
alcohol and initiating alcohol use during adolescence. J Adolesc Health
32. Anda RF, Felitti VJ, Walker J, Whitfield, CL, Bremner JD, Perry BD,
Dube SR, Giles WH. The enduring effects of abuse and related adverse
experiences in childhood: a convergence of evidence from neurobiology
and epidemiology. Eur Arch Psych Clin Neurosci 2006;256:174–186.
33. Edwards VJ, Anda RF, Gu D, Dube SR, Felitti VJ. Adverse childhood
experiences and smoking persistence in adults with smoking-related
symptoms and illness. Permanente J 2007;11:5–7.
34. Dube SR, Anda RF, Felitti VJ, Croft JB, Edwards VJ, Giles WH.
Growing up with parental alcohol abuse: exposure to childhood abuse,
neglect, and household dysfunction. Child Abuse Neglect 2001;25:
35. Dong M, Anda RF, Felitti, VJ, Dube SR, Williamson DF, Thompson TJ,
Loo CM, Giles WH. The interrelatedness of multiple forms of childhood
abuse, neglect, and household dysfunction. Child Abuse Neglect 2004;
36. Danese A, Pariante CM, Caspi A, Taylor A, and Poulton R. Childhood
maltreatment predicts adult inflammation in a life-course study. Proc Nat
Acad Sci 2007;104:1319–1324.
37. Szalai AJ. C-reactive protein (CRP) and autoimmune disease: facts and
conjectures. Clin Dev Immunol 2004;11:221–226.
38. Straus M, Gelles RJ. Physical violence in American families: risk factors
and adaptations to violence in 8,145 families. New Brunswick, NJ:
Transaction Press, 1990.
39. Wyatt GA. The sexual abuse of Afro-American and white American
women in childhood. Child Abuse Negl 1985;9:507–19.
40. Schoenborn CA. Exposure to alcoholism in the family: United States,
1988. Adv Data 1991;205:1–13.
41. Boscarino JA. Posttraumatic stress disorder and physical illness: results
from clinical and epidemiological studies. Ann NY Acad Sci 2004;1032:
42. Fairweather D, Rose NR. Women and autoimmune diseases. Emerg
Infect Dis 2004;10:2005–2011.
43. Rose NR, Mackay IR, eds. The Autoimmune Diseases. 4thed. St. Louis,
MO: Elsevier Academic Press; 2006.
44. Kumar V, Abbas AK, Fausto N, eds. Robbins and Cotran Pathologic
Basis of Disease. 7thed. Philadelphia, PA: Elsevier Saunders; 2005.
45. Dube SR, Williamson DF, Thompson T, Felitti VJ, Anda RF. Assessing
the reliability of retrospective reports of adverse childhood experiences
among adult HMO members attending a primary care clinic. Child Abuse
46. McEwen BS. The neurobiology of stress: from serendipity to clinical
relevance. Brain Res 2000;886:172–189.
47. Frisancho-Kiss S, Nyland JF, Davis SE, Frisancho JA, Barrett MA, Rose
NR, Fairweather D. Sex differences in coxsackievirus B3-induced
myocarditis: IL-12R?1 signaling and IFN-? increase inflammation in
males independent from STAT4. Brain Res 2006;1126:139–147.
48. Fairweather D, Frisancho-Kiss S, Yusung SA, Barrett MA, Gatewood
SJL, Davis SE, Njoku DB, Rose NR. IFN-? protects against chronic viral
myocarditis by reducing mast cell degranulation, fibrosis, and the profi-
brotic cytokines TGF-?1, IL-1?, and IL-4 in the heart. Am J Pathol
49. Fairweather D, Frisancho-Kiss S, Rose NR. Viruses as adjuvants for
autoimmunity: evidence from coxsackievirus-induced myocarditis. Rev
Med Virol 2005;15:17–27.
50. Tolin DF, Foa EB. Sex differences in trauma and posttraumatic stress
disorder: a quantitative review of 25 years of research. Psych Bull
51. Teicher MH, Ito Y, Glod CA, Andersen SL, Dumont N, Ackerman E.
Preliminary evidence for abnormal cortical development in physically
and sexually abused children using EEG coherence and MRI. Ann NY
Acad Sci 1997;821:160–175.
52. Perry BD, Pollard RA, Blakely TL, Baker WL, Vigilante D. Childhood
trauma, the neurobiology of adaptation and use-dependent development
of the brain: how states become traits. Infant Mental Health J 1995;16:
53. Van der Kolk BA, Fisler RE. Childhood abuse and neglect and loss of
self-regulation. Bull Menninger Clin 1994;58:145–168.
54. Femina DD, Yeager CA, Lewis DO. Child abuse: adolescent records vs.
adult recall. Child Abuse Negl 1990;14:227–231.
55. Williams LM. Recovered memories of abuse in women with documented
child sexual victimization histories. J Trauma Stress 1995;8:649–673.
56. Rothman KJ. Modern epidemiology. Boston: Little, Brown; 1986.
57. Finkelhor D, Hotaling G, Lewis IA, Smith C. Sexual abuse in a national
survey of adult men and women: prevalence, characteristics, and risk
factors. Child Abuse Negl 1990;14:19–28.
58. MacMillan HL, Fleming JE, Trocme N, Boyle MH, et al. Prevalence of
child physical and sexual abuse in the community: results from the
Ontario Health Supplement. JAMA 1997;278:131–5.
59. Lawlor DA. The developmental origins of health and disease: Where do
we go from here? Epidemiology 2008;19(2):206–208.
S. R. DUBE et al.
250 Psychosomatic Medicine 71:243–250 (2009)