Association of Childhood Abuse With Incident Inflammatory Bowel Disease Authors: Adjoa Anyane-Yeboa, Kevin Casey, Andrea L. Roberts, Emily Lopes, Kristin Burke, Ashwin Ananthakrishnan, James Richter, Yvette C. Cozier, Karenstan C. Koenen, Andrew T. Chan, Hamed Khalili Published: 2024 ABSTRACT INTRODUCTION:A link between inflammatory bowel disease (IBD), stressful life events, and psychological factors has previously been reported. Our objective was to examine the relationship between emotional, physical, and sexual abuse of childhood and risk of IBD using a large cohort of female health professionals.METHODS:We included participants in the Nurses' Health Study II who completed the Physical and Emotional Abuse Subscale of the Childhood Trauma Questionnaire and the Sexual Maltreatment Scale of the Parent-Child Conflict Tactics Scale in 2001. Diagnosis of IBD was determined by self-report and confirmed independently by 2 physicians through review of medical records. We used Cox proportional hazard modeling to estimate the risk of Crohn's disease (CD) and ulcerative colitis (UC) while adjusting for covariates.RESULTS:Among 68,167 women followed from 1989 until 2017, there were 146 incident cases of CD and 215 incident cases of UC. Compared with women with no history of abuse, the adjusted hazard ratios of CD were 1.16 (95% confidence interval [CI] 0.67–2.02) for mild, 1.58 (95% CI 0.92–2.69) for moderate, and 1.95 (95% CI 1.22–3.10) for severe abuse (Ptrend = 0.002). We did not observe an association between childhood abuse and risk of UC.DISCUSSION:Women who reported early life severe abuse had an increased risk of CD. These data add to the growing body of evidence on the critical role of early life stressors in development of CD. INTRODUCTION Crohn's disease (CD) and ulcerative colitis (UC), collectively known as inflammatory bowel disease (IBD), are chronic inflammatory diseases of the gastrointestinal tract with rising worldwide prevalence (1). The etiology of IBD is unknown but thought to be the result of a complex interaction between host genetic factors, gut microbiota, and environmental factors. Among environmental factors, chronic psychological stress has been shown to modulate gut inflammation (2–4). Specifically, in animal models of colitis, chronic stress has been linked to proinflammatory gut microbiota, activation of immune response, and worsening colitis phenotype (2–4). In parallel, observational studies have shown a link between anxiety and depression, short-term perceived stress, and psychosocial stressors such as serious illness and sudden death in the family and risk of developing IBD or experiencing flares among those with established disease (5–11). Childhood abuse is a chronic stressor that has been shown to be related to numerous metabolic and immune-mediated diseases (12–14). Childhood abuse has been associated with hospital admissions for autoimmune diseases such as rheumatoid arthritis, type 1 diabetes mellitus, autoimmune thrombocytopenic purpura, idiopathic pulmonary fibrosis, and systemic lupus (12,14). In addition, childhood abuse has been associated with elevated inflammatory markers such as c-reactive protein and interleukin-6 (IL-6) (15,16). Given these studies and emerging data on the importance of early life stressors in shaping host immune function and the gut microbiome (17–20), we sought to examine the relationship between childhood abuse and risk of IBD. Our hypothesis is that a history of childhood physical, emotional, or sexual abuse is associated with an increased risk of both incident CD and UC. MATERIALS AND METHODS Study population Our study population included participants from the Nurses' Health Study II, a prospective cohort study consisting of 116,430 female patients, who were nurses in the United States at the time of enrollment. Participant ages ranged from 25 to 44 years at baseline in 1989, with follow-up questionnaires sent every other year. Information on child abuse were ascertained on a separate questionnaire in 2001 measuring exposure to early life stressors, and nearly, 75% of participants who received it responded (N = 91,279), response rate—68,376. We excluded participants with a history of IBD at baseline in 1989 and those who did not respond to the supplementary questionnaire. Outcome ascertainment Starting in 1991 and in every subsequent questionnaire, participants in Nurses' Health Study II were queried about diagnosis of CD and UC. Those responding positively were mailed a supplementary questionnaire asking for additional information about disease characteristics and treatment. Participants were also asked for permission to obtain their medical records. Two gastroenterologists blinded to exposure data independently reviewed medical records to confirm diagnosis of CD or UC, according to previously established criteria (21). Ascertainment of childhood trauma Questions on child abuse were ascertained by a 2001 supplementary questionnaire measuring exposure to early life stressors. Physical and emotional abuse during childhood was evaluated using the Physical and Emotional Abuse Subscale of the Childhood Trauma Questionnaire (CTQ) (22). This subscale consists of 5 questions related to participant's experiences up to and including the age of 11 years: “(i) people in my family hit me so hard that it left me with bruises and marks; (ii) the punishments I received seemed cruel; (iii) I was punished with a belt, a board, a cord, or some other hard object; (iv) someone in my family yelled and screamed at me; and (v) people in my family said hurtful or insulting things to me.” The mean score for each participant was derived according to responses (0: never, 1: rarely, 2: sometimes, 3: often or 4: very often true), and quartiles were created using these mean scores, with categories labeled as: “None” (score: 0 to ≤0.4), “Mild” (score range 0.4 to ≤0.8), “Moderate” (score range: 0.8 to ≤1.4), and “Severe” (score: >1.4). The subscale has previously been shown to have high internal consistency (α = 0.94) (23). Childhood sexual abuse information (before the age of 18 years) was obtained through the Sexual Maltreatment Scale of the Parent-Child Conflict Tactics Scale (24) (PCTS). This subscale is scored 0–8 and based on participant's responses of “0: never,” “1: once,” or “2: more than once” for 4 survey items, possible categories included as follows: “No Abuse” (0 points), “Mild Abuse” (1–2 points), “Moderate Abuse” (3–4 points), or “Severe Abuse” (5+ points), in accordance with the existing literature (14). Our primary exposure here was overall childhood trauma (before the age of 18 years) defined according to participants' combined responses to the CTQ subscale and the PCTS. Categories included “None” (No abuse reported in either the CTQ or PCTS), “Mild” (Mild abuse reported in either the CTQ or PCTS, but not both), “Moderate” (Moderate abuse in either the CTQ or PCTS and no abuse in the other, or mild abuse in both), and “Severe” (Any response of severe, or moderate on both scales, or a combination of moderate and mild). Assessment of other covariates Data on weight, smoking, and use of nonsteroidal anti-inflammatory drugs (NSAIDs) and oral contraceptives or menopausal hormone therapy were assessed at each biennial questionnaire starting at baseline in 1989. Information on weight at age the age of 18 years and height was assessed at baseline in 1989. Diet, physical activity, and alcohol intake were assessed alongside the survey every 4 years starting at baseline in 1989 using validated questionnaires (25,26). Dietary pattern of participants were characterized using Alternative Healthy Eating Index as previously described (27). Clinical diagnosis of depression was assessed biennially from 2003 to 2011, while use of antidepressant use was assessed biennially from 1993 to 2011. Low Mental Health Inventory score-5 (subscale score of <60) was ascertained from general questionnaires in 1993, 1997, and 2001. We also considered parental education status at infancy obtained at baseline in 1989 as an indicator of socioeconomic status. Statistical analyses Person time was calculated from the return of the baseline questionnaire in 1989 until the end of study period on June 1, 2017, date of death, date of diagnosis of IBD, or last returned questionnaire, whichever came first. We used Cox proportional hazards regression to calculate hazard ratios (HR) and 95% confidence interval (CI) for incident CD and UC according to quartiles of childhood abuse. The first model was initially adjusted for demographics including age, race (White, Black/African American, Asian, and other or multiple races) and parental education (less than high school, one or both parents graduated high school, one or both parents graduated college) (model 1). The second model was additionally adjusted for lifestyle factors including body mass index (BMI) (continuous in kg/m2), diet (continuous), alcohol intake (0, 0.1–14.9, ≥ 15 g/d), physical activity (continuous), and smoking (never, past, current) (model 2), and the third was adjusted for ever (yes or no) use of medications including NSAIDs, oral contraceptives, menopausal hormone therapy, and appendectomy (model 3). All covariates, with the exception of race, ethnicity, parental education at infancy, and BMI at the age of 18 years, were modeled as time-varying. Missing values were obtained from previous questionnaires where possible and were otherwise set to the reference group in categorical variables and the median in continuous variables, to preserve the distribution of the data. Specifically, the frequency of missing values were BMI (0.5%, n = 369), diet (8.9%, n = 6,061), alcohol intake (8.9%, n = 6,058), physical activity (0.04%, n = 28), smoking (0.1%, n = 82), oral contraceptive use (0.2%, n = 115), and menopausal hormone therapy use (0.004%, n = 3). Finally, to explore whether the association between childhood abuse and risk of IBD was mediated by depression, we additionally adjusted for history of depression (model 4). For all models, trend was evaluated by modeling the response categories as continuous variables. Statistical analysis was performed using SAS 9.4 (SAS Institute, Cary, NC). Statistical significance was defined as a P value less than 0.05. RESULTS Among 68,167 women, we confirmed 146 cases of CD and 215 cases of UC over 1,926,571 person-years of follow-up. The average follow-up for participants was 25.0 years (SD = 3.3 years). Of the 68,167 women, 18,133 (26.6%) reported no childhood abuse, 14,745 (21.6%) reported mild childhood abuse, 12,851 (18.9%) noted moderate childhood abuse, and 22,438 (32.9%) reported severe (Table 1). The population was predominantly White at 94.9%. Compared with those without a history of abuse, a higher percentage of those with moderate-to-severe childhood abuse were smokers and had a history of depression (Table 1). Table 1.Baseline characteristics of NHS2 women by childhood traumaChildhood traumaOverall (n = 68,167)None (n = 18,133)Mild (n = 14,745)Moderate (n = 12,851)Severe (n = 22,438)Age35.1 (4.7)35.1 (4.8)35.0 (4.7)34.9 (4.6)35.2 (4.5)BMI, kg/m223.9 (4.9)23.5 (4.6)23.7 (4.6)23.9 (4.8)24.5 (5.4)Race (%) White94.996.495.694.893.3 Black/African American1.10.61.01.31.5 Native American1.21.01.21.21.3 Asian/Asian American2.81.92.22.73.9 Hispanic1.60.91.31.62.3Smoking status Never, %66.371.869.066.260.2 Former, %21.718.420.021.725.4 Current, %12.09.811.012.114.4Physical activity, MET hr/wk22.7 (27.3)22.3 (26.1)22 (26.4)22.7 (27.2)23.6 (29)Diet score, AHEI43.7 (9.8)43.2 (9.7)43.5 (9.7)43.6 (9.8)44.4 (10)Alcohol intake g/d 0 g per d, %47.346.847.345.948.6 0.1 g–14.9 g per d, %49.049.949.250.547.4 15.0 g + per d, %3.63.23.43.64.0 Menopause status, %2.21.92.02.02.8Menopausal hormone therapy use Never, %97.998.298.198.197.3 Former, %0.10.10.10.10.1 Current, %2.01.71.81.82.5Oral contraceptive: ever use, %83.782.082.784.185.4Appendectomy, %13.312.212.512.715.1History of depression, %22.515.519.122.630.5Parent highest education level Both parents less than high school, %9.57.38.59.412.0 One parent high school degree, %69.668.369.069.770.9 Both parents high school degree, %20.924.422.520.917.1AHEI, Alternative Healthy Eating Index; BMI, body mass index; NHS2, Nurses' Health Study II.Values are means (SD) or medians (Q25, Q75) for continuous variables; percentages or ns or both for categorical variables.Values of polytomous variables may not sum to 100% due to rounding. Baseline characteristics of NHS2 women by childhood trauma AHEI, Alternative Healthy Eating Index; BMI, body mass index; NHS2, Nurses' Health Study II. Values are means (SD) or medians (Q25, Q75) for continuous variables; percentages or ns or both for categorical variables. Values of polytomous variables may not sum to 100% due to rounding. Compared with those without a history of childhood abuse, the age-adjusted HR of incident CD was 1.19 (95% CI 0.69–2.06) among those with a history of mild childhood abuse, 1.63 (95% CI 0.96–2.76) for moderate childhood abuse, and 2.13 (95% CI 1.35–3.36) for severe childhood abuse (Ptrend = 0.0003) (Table 2, model 1). Compared with those without a history of childhood abuse, the multivariate-adjusted HR of CD was 2.00 (95% CI 1.26–3.17) (Ptrend = 0.001) among those who reported severe childhood abuse after additional adjustment for individual lifestyle factors such as BMI, physical activity, smoking status, diet, and alcohol intake (Table 2, model 2). With additional adjustment for medication use (NSAIDs, oral contraceptives, menopausal hormone therapy use) and history of appendectomy (Table 2, model 3), the risk was largely unchanged and remained statistically significant (HR = 1.97, 95% CI 1.24–3.12) (Ptrend = 0.002). Last, adjustment for history of depression during adulthood did not materially alter our estimates (multivariate-adjusted HR = 1.95 (95% CI 1.22–3.10), Ptrend = 0.002) (Table 2, model 4). By contrast, we did not find an association between childhood abuse and risk of incident UC (model 4, Ptrend = 0.628) (Table 2). In our fully adjusted model (Table 2, model 4), compared with women who reported no abuse during childhood, the multivariate-adjusted HR of UC among those who reported severe abuse was 0.89 (95% CI 0.62–1.27). Table 2.Combined childhood trauma and risk of Crohn's disease and ulcerative colitisHazard ratio (95% CI)Cases/PYModelaModelbModelcModeldCrohn's disease None26/514,2061.00 (ref)1.00 (ref)1.00 (ref)1.00 (ref) Mild25/418,3571.19 (0.69–2.06)1.17 (0.67–2.03)1.17 (0.67–2.02)1.16 (0.67–2.02) Moderate29/363,7321.63 (0.96–2.76)1.60 (0.94–2.72)1.59 (0.93–2.70)1.58 (0.92–2.69) Severe66/630,2762.13 (1.35–3.36)2.00 (1.26–3.17)1.97 (1.24–3.12)1.95 (1.22–3.10) Ptrend0.00030.0010.0020.002Ulcerative colitis None61/514,1771.00 (ref)1.00 (ref)1.00 (ref)1.00 (ref) Mild42/418,3350.86 (0.58–1.27)0.85 (0.57–1.25)0.84 (0.57–1.25)0.83 (0.56–1.23) Moderate41/363,7210.96 (0.64–1.43)0.95 (0.64–1.41)0.93 (0.63–1.39)0.90 (0.60–1.34) Severe71/630,2690.98 (0.69–1.39)0.96 (0.67–1.35)0.95 (0.67–1.34)0.89 (0.62–1.27) Ptrend0.9430.9420.8930.628AHEI, Alternative Healthy Eating Index; CI, confidence interval; PY, person-years.aAdjusted for race, age, parents' education level.bFurther adjusted for body mass index, physical activity, smoking status, diet (AHEI), alcohol intake (g/d).cFurther adjusted for medications (nonsteroidal anti-inflammatory use, oral contraceptive use, menopausal hormone therapy use) and surgery (appendectomy).dFurther adjusted for history of depression. Combined childhood trauma and risk of Crohn's disease and ulcerative colitis AHEI, Alternative Healthy Eating Index; CI, confidence interval; PY, person-years. Adjusted for race, age, parents' education level. Further adjusted for body mass index, physical activity, smoking status, diet (AHEI), alcohol intake (g/d). Further adjusted for medications (nonsteroidal anti-inflammatory use, oral contraceptive use, menopausal hormone therapy use) and surgery (appendectomy). Further adjusted for history of depression. In exploratory analysis, we considered the possibility that our observed association may be mediated by changes in anthropometry. After additional adjustment for weight change (age 18 years to present), our finding was not materially changed (see Supplementary Table 1, http://links.lww.com/CTG/B215). Specifically, compared with participants without reported history of childhood abuse, those who experienced severe childhood abuse had multivariable-adjusted HRs of 1.94 (95% CI 1.22–3.10) (Ptrend = 0.002) for CD and 0.89 (95% CI 0.62–1.26) (Ptrend = 0.623) for UC. We also explored the independent association between the component of total trauma and risk of CD and UC. Specifically, we observed a positive association between childhood physical and emotional abuse and risk of CD (Ptrend = 0.004), but not UC (Ptrend = 0.726) (Table 3, model 4). Similarly, childhood sexual abuse was associated with increased risk of CD, albeit not statistically significant (Ptrend = 0.062) (Table 4, model 4), while we did not observe an association with UC (Ptrend = 0.832). Table 3.Childhood emotional and physical abuse and risk of Crohn's disease and ulcerative colitisHazard ratio (95% CI)Cases/PYModelaModelbModelcModeldCrohn's disease None35/672,1301.00 (ref)1.00 (ref)1.00 (ref)1.00 (ref) Mild28/416,4081.29 (0.78–2.12)1.28 (0.78–2.11)1.28 (0.78–2.10)1.27 (0.77–2.10) Moderate37/381,9511.87 (1.18–2.97)1.84 (1.16–2.94)1.83 (1.15–2.90)1.81 (1.14–2.89) Severe46/456,0821.98 (1.27–3.09)1.87 (1.20–2.93)1.85 (1.18–2.89)1.83 (1.16–2.87) Ptrend0.00080.0020.0030.004Ulcerative colitis None78/672,0881.00 (ref)1.00 (ref)1.00 (ref)1.00 (ref) Mild43/416,3970.90 (0.62–1.31)0.89 (0.61–1.30)0.89 (0.61–1.29)0.88 (0.60–1.27) Moderate41/381,9420.94 (0.64–1.37)0.93 (0.64–1.36)0.92 (0.63–1.35)0.89 (0.61–1.31) Severe53/456,0751.04 (0.73–1.48)1.02 (0.71–1.45)1.01 (0.70–1.43)0.95 (0.66–1.35) Ptrend0.8660.9610.9870.726AHEI, Alternative Healthy Eating Index; CI, confidence interval; PY, person-years.aAdjusted for race, age, parents' education level.bFurther adjusted for body mass index, physical activity, smoking status, diet (AHEI), alcohol intake (g/d).cFurther adjusted for medications (nonsteroidal anti-inflammatory use, oral contraceptive use, menopausal hormone therapy use) and surgery (appendectomy).dFurther adjusted for history of depression. Childhood emotional and physical abuse and risk of Crohn's disease and ulcerative colitis AHEI, Alternative Healthy Eating Index; CI, confidence interval; PY, person-years. Adjusted for race, age, parents' education level. Further adjusted for body mass index, physical activity, smoking status, diet (AHEI), alcohol intake (g/d). Further adjusted for medications (nonsteroidal anti-inflammatory use, oral contraceptive use, menopausal hormone therapy use) and surgery (appendectomy). Further adjusted for history of depression. Table 4.Childhood sexual abuse and risk of Crohn's disease and ulcerative colitisHazard ratio (95% CI)Cases/PYModelaModelbModelcModeldCrohn's disease None81/1,281,1401.00 (ref)1.00 (ref)1.00 (ref)1.00 (ref) Mild45/455,2411.62 (1.12–2.33)1.56 (1.08–2.25)1.53 (1.06–2.21)1.52 (1.05–2.20) Moderate14/125,2171.77 (1.00–3.13)1.66 (0.94–2.95)1.64 (0.92–2.91)1.61 (0.91–2.87) Severe6/64,9731.44 (0.62–3.31)1.27 (0.55–2.94)1.24 (0.54–2.88)1.22 (0.52–2.83) Ptrend0.0140.0400.0510.062Ulcerative colitis None145/1,281,0731.00 (ref)1.00 (ref)1.00 (ref)1.00 (ref) Mild45/455,2430.88 (0.63–1.23)0.87 (0.62–1.21)0.86 (0.62–1.21)0.85 (0.61–1.19) Moderate13/125,2150.92 (0.52–1.63)0.90 (0.51–1.59)0.89 (0.50–1.57)0.85 (0.48–1.50) Severe12/64,9711.63 (0.90–2.96)1.59 (0.88–2.90)1.60 (0.88–2.92)1.51 (0.83–2.75) Ptrend0.5480.6370.6550.832AHEI, Alternative Healthy Eating Index; CI, confidence interval; PY, person-years.aAdjusted for race, age, parents' education level.bFurther adjusted for body mass index, physical activity, smoking status, diet (AHEI), alcohol intake (g/d).cFurther adjusted for medications (nonsteroidal anti-inflammatory use, oral contraceptive use, menopausal hormone therapy use) and surgery (appendectomy).dFurther adjusted for history of depression. Childhood sexual abuse and risk of Crohn's disease and ulcerative colitis AHEI, Alternative Healthy Eating Index; CI, confidence interval; PY, person-years. Adjusted for race, age, parents' education level. Further adjusted for body mass index, physical activity, smoking status, diet (AHEI), alcohol intake (g/d). Further adjusted for medications (nonsteroidal anti-inflammatory use, oral contraceptive use, menopausal hormone therapy use) and surgery (appendectomy). Further adjusted for history of depression. We considered the possibility of recall bias due to the cross-sectional nature of collecting data on childhood trauma. Therefore, we conducted a sensitivity analysis limiting our follow-up to after 2001 when women were initially asked to report childhood trauma (Table 5). Similar to our primary analyses, we observed that increased childhood trauma was associated with an increased risk of CD (model 4, Ptrend = 0.015), but not UC (model 4, Ptrend = 0.129). Table 5.Childhood combined trauma and risk of Crohn's disease and ulcerative colitis after assessment of CTQ (2001–2017)Hazard ratio (95% CI)Cases/PYModelaModelbModelcModeldCrohn's disease None13/301,9131.00 (ref)1.00 (ref)1.00 (ref)1.00 (ref) Mild17/245,6631.56 (0.76–3.22)1.55 (0.75–3.21)1.56 (0.76–3.23)1.55 (0.75–3.19) Moderate13/213,0801.39 (0.64–2.99)1.38 (0.64–2.98)1.39 (0.64–3.01)1.37 (0.63–2.96) Severe37/367,3422.34 (1.24–4.42)2.33 (1.23–4.42)2.35 (1.24–4.45)2.26 (1.18–4.32) Ptrend0.0090.0100.0100.015Ulcerative colitis None36/301,8921.00 (ref)1.00 (ref)1.00 (ref)1.00 (ref) Mild17/245,6530.60 (0.33–1.06)0.60 (0.33–1.06)0.59 (0.33–1.06)0.58 (0.33–1.04) Moderate21/213,0740.83 (0.48–1.42)0.83 (0.48–1.42)0.80 (0.47–1.38)0.78 (0.45–1.34) Severe29/367,3520.67 (0.41–1.10)0.69 (0.42–1.13)0.67 (0.41–1.10)0.63 (0.38–1.04) Ptrend0.2030.2350.1900.129AHEI, Alternative Healthy Eating Index; CI, confidence interval; PY, person-years.aAdjusted for race, age, parents' education level.bFurther adjusted for body mass index, physical activity, smoking status, diet (AHEI), alcohol intake (g/d).cFurther adjusted for medications (nonsteroidal anti-inflammatory use, oral contraceptive use, menopausal hormone therapy use) and surgery (appendectomy).dFurther adjusted for history of depression. Childhood combined trauma and risk of Crohn's disease and ulcerative colitis after assessment of CTQ (2001–2017) AHEI, Alternative Healthy Eating Index; CI, confidence interval; PY, person-years. Adjusted for race, age, parents' education level. Further adjusted for body mass index, physical activity, smoking status, diet (AHEI), alcohol intake (g/d). Further adjusted for medications (nonsteroidal anti-inflammatory use, oral contraceptive use, menopausal hormone therapy use) and surgery (appendectomy). Further adjusted for history of depression. DISCUSSION In this large prospective cohort of female nurses in the United States, we found that history of childhood abuse was associated with an increased risk of CD. The risk remained significant even after adjusting for lifestyle factors, medications, weight change, and depression. By contrast, we did not observe an association between childhood abuse and risk of UC. Chronic stress has been shown to be associated with an inflammatory response in many organ systems, including the gut. Specifically, stress is known to be associated with proinflammatory factors such as c-reactive protein, IL-6, tumor necrosis factor alpha, IL-1β, and nuclear factor kappa B2. Many of these pathways have been implicated in the pathophysiology of IBD (28–32). Furthermore, it has been hypothesized that stress activates the sympathetic and parasympathetic nervous systems in the gastrointestinal tract and leads to increased permeability and altered immune function that could in turn result in chronic inflammation and colitis (33). In addition, stress has also been shown to reactivate quiescent colitis through increased colonic permeability, and activation of CD4+ cells and mast cells (34). Prior studies have shown relationships between physical and psychological stressors and both inflammation and IBD. For instance, IBD has been associated with an increased risk of depression and anxiety (35–37) and depression in turn has been associated with incident CD (38). Early childhood stress has also been associated with increased inflammation, impaired cellular immunity, and altered gut microbiota in adulthood (39,40). Furthermore, early childhood stressors such as altered family dynamics have also been associated with incident IBD (41), while stressful life events have been associated with flares of IBD (42,43). Nevertheless, our findings are consistent with other studies that have shown a relationship between childhood abuse, the immune system, and immune-mediated diseases. Childhood abuse has been found to increase the risk of incident lupus, an autoimmune disorder, in a cohort of women in the United States (14). More specific to this study, the relationship between child abuse and IBD has previously been studied in a Canadian cohort (36). The study found a greater than 2-fold increased risk of UC in participants with physical and sexual abuse and no increased risk of CD (44). These findings are in contrast with our results that show an association between childhood emotional and physical abuse and risk of CD, but not UC. However, the study by Fuller-Thomson et al (44). was limited by a cross-sectional design, where information on childhood abuse and diagnosis of IBD were ascertained at the same time, thus temporality could not be established. In addition, the diagnosis of IBD was based on self-report and was not confirmed by medical record review. Our study, which prospectively collected information on childhood emotional and physical abuse and confirmed all cases of CD and UC through medical record review, significantly expands on this prior work. Misclassification of CD and UC diagnosis as well as recall bias related to disease severity may in part explain the conflicting results between our study and those of Fuller-Thomson et al (44). However, in our study, it is unclear why there was an association between childhood abuse and risk of CD, but not UC. More work is needed to further understand the potential mechanism underlying our observed association. There are a few limitations in our study. First, childhood trauma history was self-reported and thus is subject to recall bias, as participants were asked about these childhood experiences in adulthood. Yet, when evaluated prospectively and limited to individuals who developed IBD after 2001, when childhood abuse was reported, the relationship between childhood abuse and incident CD persisted. Nevertheless, we cannot rule out the possibility of misclassification in assessing early childhood trauma. Second, the cohort that was used for the study is composed mainly of older White women with higher levels of parental education and socioeconomic status. This population may experience lower levels of chronic stress than minority individuals and those of lower socioeconomic status. In addition, women are more likely to experience symptoms of depression and anxiety and are more likely than men to seek care for these symptoms (45–48). Women are also more likely than more to report higher levels of psychological distress, missed days off work, disability, poor quality of life, and face more delays in diagnosis due to these symptoms (47,49–51). This may also explain the higher rates of reported abuse in our cohort. Specifically, we found that the prevalence of childhood abuse was about 75%, which is significantly higher than those reported from the Centers for Disease Control and Prevention approximately (14.3%) in the past year (52) and another study that showed 25% of children in the United States will face abuse (53). Therefore, it is unclear if the findings of our study would be generalizable to other racial and ethnic groups, those of lower socioeconomic status or men. Last, we cannot exclude the possibility that our findings may be subject to residual confounding. In our models, we adjusted for variables that are captured in our cohort and have previously been shown to be associated with CD or UC. Unfortunately, other potential confounders such as antibiotic use during early childhood were not available in our study. Using a large US-based cohort of professional women, we found a dose-response association between childhood trauma and incident CD. Our work adds to the growing body of evidence on the critical role of early life stressors in the development of CD and further emphasizes the critical role of psychosocial stressors on gastrointestinal health. Future studies should aim to understand the mechanisms underlying this possible relationship between childhood abuse and incident CD. CONFLICTS OF INTEREST Guarantor of the article: Hamed Khalili, MD, MPH. Specific author contributions: A.A.Y.: conception and design, drafting of article, review and revisions, final approval of article. K.C.: acquisition of data and analysis, drafting of article, review and revisions, final approval. A.L.R.: design, review and revisions, final approval. E.L.: review and revisions, final approval. K.B.: review and revisions, final approval. A.A.: review and revisions, final approval. J.R.: review and revisions final approval. Y.C.: review and revisions, final approval. K.K.: review and revisions, final approval. A.T.C.: review and revisions, final approval. H.K.: conception and design, drafting of article, review and revisions, supervision, final approval. Financial support: This work was supported by U01 CA176726 NHSII cohort infrastructure grant, Crohn's and Colitis Senior Research Award to H.K., and NCI P50CA244433 Diversity Supplement award to A.A.Y. Potential competing interests: H.K. has received consulting fees from Takedaand research funding from Takeda and Pfizer. A.A.Y. has received consulting fees from Janssen, Takeda Pharmaceuticals, and Exact Sciences and research funding from Pfizer Medical Grants Program. A.C. has served as a consultant for Pfizer Inc., and Boehringer Ingelheim for unrelated studies and grant support from Pfizer Inc., Zoe Ltd, and Freenome for unrelated studies. The remaining authors have no relevant conflicts. IRB approval statement: The study is approved by the Massachusetts General Hospital IRB. Protocol number: 2001P001128.Study HighlightsWHAT IS KNOWN✓Psychological factors and chronic stress affect IBD.WHAT IS NEW HERE✓Childhood abuse may be associated with the risk of Crohn's disease.✓The association was not mediated by depression.✓The findings may be independent of lifestyle factors. Study Highlights WHAT IS KNOWN ✓Psychological factors and chronic stress affect IBD. ✓Psychological factors and chronic stress affect IBD. Psychological factors and chronic stress affect IBD. WHAT IS NEW HERE ✓Childhood abuse may be associated with the risk of Crohn's disease.✓The association was not mediated by depression.✓The findings may be independent of lifestyle factors. ✓Childhood abuse may be associated with the risk of Crohn's disease. Childhood abuse may be associated with the risk of Crohn's disease. ✓The association was not mediated by depression. The association was not mediated by depression. ✓The findings may be independent of lifestyle factors. The findings may be independent of lifestyle factors. Supplementary Material