Trauma history and risk of irritable bowel syndrome in women veterans

Abstract

Background

Over 1.8 million women in the U.S. are veterans of the armed services. They are at increased risk of occupational traumas, including military sexual trauma. We evaluated the association between major traumas and irritable bowel syndrome among women veterans accessing VA healthcare.

Methods

We administered questionnaires to assess trauma history as well as IBS, PTSD and depression symptoms to 337 women veterans seen for primary care at VA Women’s Clinic between 2006 and 2007. Logistic regression was used to evaluate the association between individual traumas and IBS risk after adjustment for age, ethnicity, PTSD and depression.

Results

IBS prevalence was 33.5%. The most frequently reported trauma was sexual assault (38.9%). Seventeen of eighteen traumas were associated with increased IBS risk after adjusting for age, ethnicity, PTSD and depression, with six statistically significant (range of adjusted odds ratios (OR) between 1.85 [95% CI, 1.08–3.16] and 2.6 [95% CI, 1.28–3.67]). Depression and PTSD were significantly more common in IBS cases than controls, but neither substantially explained the association between trauma and increased IBS risk.

Conclusions

Women veterans report high frequency of physical and sexual traumas. Lifetime history of a broad range of traumas is independently associated with an elevated IBS risk.

Background

Irritable bowel syndrome (IBS) is a prevalent and costly functional gastrointestinal disorder. An IBS diagnosis is based upon presence of chronic or recurrent bowel abnormalities with abdominal pain relieved by defecation and not explainable by other causes. IBS is the most frequent cause for gastroenterologist visits in the U.S.(1) responsible for 3.7 million physician visits annually(2) and direct medical costs in excess of 1.6 billion dollars.(3) Considerable indirect costs, in excess of $19 billion dollars annually(3), also result from substantially higher total healthcare utilization and costs(4–6) and reduced work productivity.(5–7) Furthermore, a markedly decreased health-related quality of life is consistently reported in patients with IBS (4–6, 8–10), with an estimated 75% of cases still symptomatic after ten years.(11)

Although IBS is known to be clinically under-recognized among both genders in the general population(12), IBS prevalence in clinical settings is 2–3 times greater in women.(13) This female preponderance has been attributed to a combination of sex-based physiological differences and gender-based psychosocial and environmental differences, including likelihood of experiencing sexual trauma.(14) Several biological factors have been associated with IBS risk including GI infection(15–17), food intolerance/allergies(18, 19), and genetic susceptibility.(20–23) Psychological disorders including depression(24–27), anxiety(24, 27–29), and post-traumatic stress disorder (PTSD)(30–33) are more common in IBS cases.

The over 1.8 million women veterans in the U.S.(34) constitute a large occupational cohort at risk of IBS. Women also comprise approximately 14% of all active duty military personnel and 20% of all new military recruits. Currently, around 11% of all active-duty service members serving in Iraq and Afghanistan (OIF/OEF) conflicts are women, with women veterans of OIF/OEF now the single largest living group of women veterans (~187,000).(34–36) Women veterans are known to have increased risk of occupational trauma including in association with deployment and combat.(37) They also have high rates of lifetime sexual traumas including rape during military service.(38–40)

We previously reported a high survey-based prevalence of IBS as well as of psychological distress, including depression and post-traumatic stress disorder (PTSD), in a sample of women veterans receiving primary care at a Department of Veterans Affairs (VA) medical center.(32) The primary aims of our current research were: 1) to determine survey-based prevalence of a broad range of individual traumas in a larger sample of women veterans, and 2) to evaluate the association between individual traumas and IBS risk after adjustment for PTSD and depression. An associated exploratory aim was to perform a medical record review to assess clinical recognition of IBS in women veterans who access VA healthcare.

Methods

Sample and Questionnaires

We recruited consecutive women veterans aged 18–70 years old at the time of a scheduled primary medical care appointment in the Women’s Clinic at the Michael E. DeBakey VA Medical Center in Houston, Texas. All eligible women veterans were approached about study participation by a female researcher who was not part of the clinical staff. Study participants completed a self-administered survey prior to their clinical visit and received five-dollar remuneration. Our research protocol was approved by the Baylor College of Medicine Institutional Review Board.

We used the validated Bowel Disorder Questionnaire (BDQ)(41, 42) to measure type, frequency and severity of gastrointestinal symptoms experienced during the previous year. We applied symptom-based diagnostic criteria adapted from gold-standard Rome II clinical guidelines to define if a woman veteran had IBS.(32)

We used the Mississippi Scale for Combat-Related PTSD (M-PTSD) to define presence of PTSD using the recommended diagnostic cutoff of ≥107, a level at which the M-PTSD demonstrated a 93% sensitivity in comparison to the gold-standard clinical diagnosis.(43) Depression was assessed using the validated Beck Depression Inventory, second edition (BDI-II).(44) We categorized total instrument sum score using recommended cut-points of 0–13 (no-minimal depression), 14–19 (mild depression), 20–28 (moderate depression) and 29–63 (severe depression).

We determined whether a woman veteran ever experienced a broad range of major life traumas using the validated 18-item Trauma History Questionnaire (THQ)(45). Our interim analyses suggested a very high prevalence of sexual traumas among our women veteran study participants. These results in conjunction with greatly increased national media and public awareness about the scope and burden of military sexual trauma(46) led to our addition of a second trauma measure, the Trauma Questionnaire (TQ)(47), to specifically capture timing of abusive traumas (i.e., sexual assault/rape, sexual harassment, and domestic violence) in relation to military service. The TQ was given to the final 25% of our study participants.

Electronic Medical Record Review

We evaluated VA electronic medical record of all BDQ-identified IBS cases for presence of alternate explanations for their symptoms. We also evaluated medical records for all participants for presence of a clinical IBS diagnosis and/or at least 2 cardinal symptoms consistent with a possible IBS diagnosis during the one-year time period prior to survey administration. Finally, we reviewed the records of a random sample of women veterans who refused to participate to see if they differed in likelihood of having IBS diagnosis or symptoms compared to study participants.

Statistical Analyses

We compared sociodemographic characteristics and IBS risk factors between women veterans with IBS (cases) and those without IBS (controls) using the χ² test for categorical variables and Student’s T-test for continuous variables. We evaluated how specific traumas influence IBS risk using logistic regression. We performed three sets of multivariate analyses. In the first minimal multivariate analysis, we adjusted for two well-established confounders, age and ethnicity. In the second intermediate analysis, we also adjusted for PTSD to see if any observed association between a specific trauma and IBS risk was attributable to its presence. In the third or full multivariate model, we additionally adjusted for presence of depression. We evaluated potential effect modification by creating all first-level interaction terms and assessing their significance using the Wald test. Interaction terms were included in reported models if significant at p<0.15. All logistic regression results are reported as odds ratios with associated 95% confidence intervals (CI).

Trauma timing in relation to military service

We calculated lifetime prevalence of abusive traumas in the final 83 women veterans who had been given the TQ. Participants were classified as having never experienced the trauma, experienced the trauma during military service, or experienced the trauma only outside of military service. To evaluate utility of the THQ and TQ to identify forcible sexual assault or rape, we calculated concordance between their most directly comparable questions using the kappa statistic.

Validation analyses of key study measures and clinical recognition of IBS

To assess our use of the BDQ to identify cases, we first calculated the false-positive rate as the proportion of BDQ-identified IBS cases with exclusionary medical conditions or symptoms in their medical records. We also calculated negative and positive predictive values (NPV and PPV respectively) or probability that the BDQ-derived IBS case-status accurately predicts IBS case-status in the medical record. Using the validated BDQ and adapted Rome II clinical guidelines as the ‘gold standard’, we calculated clinical under-recognition of IBS as the proportion of BDQ-identified IBS cases without a confirmed, suspected or differential diagnosis of IBS in their medical record. Finally, to assess potential selection bias, we used the χ² test to compare the proportion of participants with either a clinical IBS diagnosis or ≥ 2 symptoms suggestive of IBS reported in the medical record in both our entire cohort and in a random sample of non-participants.

All analyses were conducted using SPSS version 16.0 (SPSS Inc., Chicago, IL).

Results

Demographic and clinical features

We recruited 337 consecutive women veterans prior to their primary care visit at the Women’s Clinic at the Michael E. DeBakey VA Medical Center in Houston, Texas (11/2005–2/2006, 3/2007–6/2007). Over 95% of eligible women veterans approached about the study participated. The mean age of participants was 48.5 years, and most were African-American (48.1%) or non-Hispanic White (39.7%). (Table 1)

Table 1.

Sociodemographic characteristics of 337 women veterans receiving primary medical care at a VA Women’s Clinic stratified according to presence of irritable bowel syndrome (IBS)¹.

	IBS+	IBS−	IBS+ vs. IBS− 2
	N=113 (33.5%)	N=224 (66.5%)	p-value
Sociodemographic Characteristic
Age in years Mean (SD)	46.9 (10.0)	49.3 (12.9)	0.06
Age-group in years n (%)			0.02*
18–30	5 (4.4%)	21 (9.4%)
31–45	48 (42.5%)	61 (27.2%)
46–60	51 (45.1%)	112 (50.0%)
61+	9 (8.0%)	30 (13.4%)
Ethnicity n (%)			0.02*+
African-American	55 (48.7%)	107 (47.8%)
White	52 (46.0%)	82 (36.6%)
Hispanic/Other	6 (5.3%)	34 (15.2%)
Missing	-	1 (0.4%)
Education n (%)			0.07
High school graduate or less	9 (8.0%)	38 (17.0%)
Some college	70 (61.9%)	122 (54.5%)
College graduate	33 (29.2%)	61 (27.2%)
Missing	1 (0.9%)	3 (1.3%)
Marital Status n (%)			0.61
Currently Married	35 (31.0%)	58 (25.9%)
Previously Married	59 (52.2%)	122 (54.5%)
Never Married	19 (16.8%)	43 (19.2%)
Missing	-	1 (0.4%)

IBS, irritable bowel syndrome; SD, standard deviation

^*Statistically significant at p<0.05.

¹Presence or absence of IBS determined from responses to the validated Bowel Disorder Questionnaire (BDQ) consistent with a diagnosis of IBS using adapted Rome II criteria. (see also Methods section)

²Missing not included in calculation of test statistic as less than 1.5% of either IBS+ cases or IBS− controls had missing values.

⁺Although overall ethnic distributions significantly different, % African-Americans not significantly different in IBS+ and IBS− (48.7% vs. 47.8% respectively).

Approximately 33.5% (n=113) were BDQ-defined as IBS cases, with IBS prevalence among Whites modestly higher than among African-Americans (38.8% vs. 34.0% respectively, N.S.). IBS cases were slightly younger than IBS-free controls (mean age 46.9 vs. 49.3 years, P=0.06). Although the proportion of African–Americans was similar among IBS cases and controls (48.7% vs. 47.8% respectively, N.S.), cases were significantly more likely to be White (46.0% vs. 36.6%) and much less likely to be Hispanic (5.3% vs. 15.2) than controls. IBS cases were significantly more likely to have PTSD (22.1% vs. 10.7%, P=0.006) or depression (44.2% vs. 29.5%, P=0.01) than IBS-free controls. (Table 1) As (98%) of IBS cases had diarrhea as a contributing or primary clinical feature (data not shown), all subsequent multivariate analyses apply to the entire sample without adjustment for clinical features.

Trauma history as predictor of IBS

The most frequently reported traumas on the THQ were “ever being forced to have sex against your will” (46.6%) and “being fondled under force or threat” (41.8%). (Table 2) Both were more common in IBS cases than controls.

Table 2.

Prevalence of selected traumas¹, post-traumatic stress disorder (PTSD)² and depression³ among 337 women veterans seen for primary care at a VA Women’s Clinicaccording to presence of irritable bowel syndrome (IBS)⁴.

	IBS+	IBS−	IBS+ vs. IBS−
	N=113	N=224
	n (%)		p-value
Positive history of specific trauma1,5
Someone tried to take things from you by force/threat	44 (38.9%)	49 (21.9%)	0.001**
Someone tried and/or succeeded breaking into home while there	21 (18.6%)	27 (12.1%)	0.120
Serious accident anywhere	49 (43.4%)	81 (36.2%)	0.245
‘Man-made’ disaster like train crash, fire, etc. where felt you/loved ones were in physical danger	26 (23.0%)	29 (12.9%)	0.022*
Other situation where you were seriously injured	27 (23.9%)	34 (15.2%)	0.054^
Other situation feared might be killed/seriously injured	52 (46.0%)	72 (32.1%)	0.014*
Seen someone seriously injured/killed	46 (40.7%)	72 (32.1%)	0.148
Seen or had to handle dead bodies (not funerals)	54 (47.9%)	82 (36.6%)	0.058^
Close family/friend killed by drunk driver	19 (16.8%)	38 (17.0%)	0.943
Spouse/partner/child died	41 (36.3%)	62 (27.7%)	0.122
Serious or life-threatening illness	32 (28.3%)	40 (17.9%)	0.033*
Forced to have sex against will	63 (55.8%)	94 (42.0%)	0.019*
Fondled under force or threat	56 (49.6%)	85 (37.9%)	0.054^
Other situation with attempted force for unwanted sexual contact	41 (36.3%)	47 (21.0%)	0.003**
Attacked with weapon	36 (31.9%)	39 (17.4%)	0.003**
Attacked without weapon and seriously injured	29 (25.7%)	34 (15.2%)	0.023*
Any family member ever beaten/pushed hard enough to cause injury	38 (33.6%)	45 (20.1%)	0.007**
Experienced any other extraordinarily stressful event	37 (32.7%)	51 (22.8%)	0.057^
PTSD2,5
Present	25 (22.1%)	24 (10.7%)	0.006**
Depression3,5
Present	50 (44.2%)	66 (29.5%)	0.010*

^*Statistically significant at p<0.05.

^**Statistically significant at p<0.01.

^{^}Approach significance (0.05≤P≤0.06).

¹Presence of trauma determined by self-reported responses to the validated Trauma History Questionnaire (THQ).

²Presence of PTSD determined by self-reported responses to the validated Mississippi Scale for Combat-Related Trauma (M-PTSD) with score ≥107 consistent with PTSD diagnosis.

³Presence of depression of at least moderate level determined by self-reported responses to the validated Beck Depression Inventory-II (BDI-II)

⁴Presence or absence of IBS as determined by self-reported responses to the validated Bowel Disorder Questionnaire (BDQ) consistent with a diagnosis of IBS per adapted Rome II criteria. (see also Methods section)

⁵Less than 3% missing values

Collectively, nine traumas in univariate analysis were significantly associated with IBS (ORs=1.79–2.24). (Table 3) Another four traumas were associated with increased risk closely approaching significance (ORs=1.56–1.74, 0.05≤ P ≤0.06).

Table 3.

Logistic regression models evaluating lifetime history of individual traumas as potential risk factors for IBS¹ among 337 women veterans receiving primary medical care at a VA Women’s Clinic.

Type of trauma2	Unadjusted Estimates	Age- and Ethnicity-Adjusted Estimates	Age-, Ethnicity-, and PTSD3- Adjusted Estimates	Age-, Ethnicity-, PTSD3-, and Depression4-Adjusted Estimates
	History of trauma+	History of trauma+	History of trauma+	History of trauma+
	Odds ratio (95% CI)	Odds ratio (95% CI)	Odds ratio (95% CI)	Odds ratio (95% CI)
Someone tried to take things from you by force/threat	2.24 (95% CI 1.36–3.68)**	2.38 (95% CI 1.42–3.98)*	2.18 (95% CI 1.29–3.69)*	2.16 (95% CI 1.28–3.67)*
Someone tried and/or succeeded breaking into home while there	1.63 (95% CI 0.88–3.05)	1.83 (95% CI 0.96–3.49)	1.71 (95% CI 0.89–3.31)	1.71 (95% CI 0.88–3.30)
Serious accident anywhere	1.32 (95% CI 0.83–2.10)	1.38 (95% CI 0.85–2.23)	1.27 (95% CI 0.78–2.08)	1.21 (95% CI 0.75–2.04)
‘Man-made’ disaster like train crash, fire, etc. where felt that you/loved ones in physical danger	1.97 (95% CI 1.10–3.56)*	2.19 (95% CI 1.19–4.02)*	2.05 (95% CI 1.10–3.80)*	2.06 (95% CI 1.11–3.82)*
Other situation where you were seriously injured	1.74 (95% CI 0.99–3.07)^	1.76 (95% CI 0.99–3.16)^	1.51 (95% CI 0.83–2.77)	1.48 (95% CI 0.88–2.72)
Other situation feared might be killed/seriously injured	1.79 (95% CI 1.12–2.87)*	1.83 (95% CI 1.13–2.98)*	1.55 (95% CI 0.93–2.58)	1.52 (95% CI 0.91–2.55)
Seen someone seriously injured/killed	1.42 (95% CI 0.88–2.28)	1.40 (95% CI 0.86–2.27)	1.27 (95% CI 0.77–2.08)	1.25 (95% CI 0.76–2.06)
Seen or had to handle dead bodies (not including funerals)	1.56 (95% CI 0.98–2.49)^	1.57 (95% CI 0.97–2.53)^	1.48 (95% CI 0.91–2.41)	1.50 (95% CI 0.92–2.44)
Close family/friend killed by drunk driver	0.98 (95% CI 0.53–1.79)	0.99 (95% CI 0.53–1.85)	0.92 (95% CI 0.49–1.74)	0.93 (95% CI 0.49–1.77)
Spouse/partner/child died	1.47 (95% CI 0.90–2.39)	1.78 (95% CI 1.05–3.03)*	1.65 (95% CI 0.96–2.83)^	1.65 (95% CI 0.96–2.82)^
Serious or life-threatening illness	1.79 (95% CI 1.04–3.05)*	2.01 (95% CI 1.14–3.54)*	2.00 (95% CI 1.12–3.55)*	1.96 (95% CI 1.09–3.51)*
Forced to have sex against will	1.74 (95% CI 1.09–2.77)*	1.72 (95% CI 1.06–2.79)*	1.58 (95% CI 0.97–2.60)^	1.55 (95% CI 0.94–2.57)^
Fondled under force or threat	1.57 (95% CI 0.99–2.50)^	1.48 (95% CI 0.92–2.40)	1.29 (95% CI 0.78–2.13)	1.26 (95% CI 0.75–2.10)
Other situation with attempted force for unwanted sexual contact	2.16 (95% CI 1.30–3.57)**	2.09 (95% CI 1.25–3.49)*	1.87 (95% CI 1.10–3.19)*	1.85 (95% CI 1.08–3.16)*
Attacked with weapon	2.20 (95% CI 1.30–3.72)**	2.43 (95% CI 1.40–4.19)*	2.19 (95% CI 1.25–3.82)*	2.15 (95% CI 1.23–3.78)*
Attacked without weapon and seriously injured	1.91 (95% CI 1.09–3.34)*	2.11 (95% CI 1.18–3.78)*	1.84 (95% CI 1.01–3.35)*	1.80 (95% CI 0.97–3.23)^
Family member beaten/pushed you hard enough to cause injury	2.01 (95% CI 1.20–3.35)*	2.23 (95% CI 1.30–3.80)*	1.95 (95% CI 1.12–3.38)*	1.91 (95% CI 1.08–3.34)*
Experienced any other extraordinarily stressful event	1.63 (95% CI 0.98–2.71)^	1.73 (95% CI 1.02–2.91)*	1.65 (95% CI 0.97–2.80)^	1.61 (95% CI 0.95–2.75)^

^*Statistically significant at p<0.05.

^**Statistically significant at p<0.01.

^{^}Approach significance (0.05≤P≤0.06).

CI, confidence interval; IBS, irritable bowel syndrome; PTSD, Post-traumatic stress disorder.

¹Presence of IBS determined according to responses to the Bowel Disorder Questionnaire (BDQ) consistent with a diagnosis of IBS per adapted Rome II criteria. (see also Methods section)

²Trauma history determined by responses to the Trauma History Questionnaire (THQ)

³Presence of PTSD determined by responses to the Mississippi Scale for Combat-Related Trauma (M-PTSD) with score ≥107 consistent with diagnosis of PTSD.

⁴Presence of depression of at least moderate severity determined by responses to the Beck Depression Inventory-II (BDI-II) with score ≥20.

In the first minimal multivariate model which adjusted for age and ethnicity, eleven traumas (61%) were associated with significantly increased risk (ORs=1.72 – 2.43), with another four (22%) closely approaching significance. (Table 3) In the second model which also adjusted for PTSD, the excess IBS risk associated with individual traumas was slightly attenuated in comparison to the minimal model, with seven traumas (39%) still associated with significant excess risk (ORs=1.84 – 2.19). Similar to results from the first or minimal model, all but one remaining trauma, although non-significant, were associated with excess risk (ORs=1.27–1.71). (Table 3) Across all trauma histories, PTSD was consistently associated with an independent 100% increase in IBS risk (ORs =2.02–2.70, data not shown). In the final multivariate model that also adjusted for depression, IBS risk was again attenuated only slightly, significant excess risk was still observed for six traumas, with another four closely approaching significance. Further, additional adjustment for depression only minimally attenuated the independent excess IBS risk associated with PTSD for all but three traumas (home break-in, seeing/handling dead bodies not related to funerals, and serious illness), where there was evidence of strong confounding by depression. (data not shown)

Only one trauma, “having someone close to you killed by a drunk driver”, was not associated with increased IBS risk in univariate or multivariate analysis.

We found no evidence of interaction between any main effects or of lack of model fit in any multivariate model.

Military sexual trauma (MST) and overlap with other abusive traumas

The final 83 study participants completed the TQ. There was high concordance (κ=0.75, P=0.001) between comparable TQ and THQ items for occurrence of forcible sexual assault expressed as ‘Has anyone ever used force or threat of force to have sex with you against your will?’ (TQ) and ‘Has anyone ever made you have intercourse, oral or anal sex against your will?’ (THQ) Specifically, 90% of women who reported experiencing sexual assault on the TQ also reported experiencing it on the THQ.

Approximately 60% of women who completed the TQ reported experiencing ≥1 sexual assault during their lifetime, with 21% reporting forcible sexual assault during military service. (Figure 1) Slightly more than one-third (34%) reported sexual abuse during childhood, with 25% subsequently reporting sexual assault or rape during military service. More women veterans reported domestic violence than sexual harassment or assault occurred during military service (49% vs. 38% vs. 21% respectively). Overall, almost two-thirds experienced ≥1 of these three abusive traumas (rape, sexual harassment and domestic violence) during military service, with 39% of abused women experiencing ≥2 traumas. (data not shown)

Fig 1.

History of Sexually Abusive Traumas reported on Trauma Questionnaire (TQ) by final 83 women veteran study participants^*.

^*Only 1 missing value (1%)

^{^}Sexual abuse in childhood by age 13 and perpetrated by someone at least 5 years older.

^{^^}TQ-defined sexual assault or sex against will by force or threat of force occuring at least once during military service.

^{^^^}Any non-childhood sexual assault that occurred at any time other than during military service.

^{^^^^}Given TQ question structure, can only determine if MSA or another non-MSA occured but not if both an MSA and a non-MSA occurred.

Validation of BDQ to identify IBS cases and evaluation of selection bias

Use of the BDQ to identify IBS cases was associated with a low false positive rate, as <4% had exclusionary conditions anywhere in their medical record. Use of the BDQ to identify presence of an IBS diagnosis in the medical record was associated with good NPV (0.73, [95% CI, 0.67–0.78]). However, the corresponding PPV was lower (0.65, [95% CI, 0.51–0.77]).

Overall, medical-record documented prevalence of either an IBS diagnosis or of ≥2 symptoms suggestive of IBS in the year prior to the index clinic visit were similar in study participants and in a random sample of 35 non-participants (14.3% vs. 16.8% for diagnosis and 17.1% vs. 17.1% for ≥2 symptoms, both N.S.).

Discussion

To our knowledge, the current study conducted in 337 women veterans receiving primary medical care at the VA is the first study to assess lifetime history of a broad range of multiple individual traumas as risk factors for symptom-confirmed IBS in any veteran or military population.

All but one of the broad range of traumas examined were associated with increased IBS risk in women veterans, most with modest to moderate 50–115% excess risk. The excess risk associated with individual traumas persisted with only minor attenuation in multivariate analyses even after adjusting for other well-established risk factors, specifically age, ethnicity, depression and PTSD. However, the number of traumas reaching or approaching significance was reduced given concomitant reduction in study power.

The very high prevalence of sexual traumas observed in this study is not completely surprising. A mandated VA-wide clinical screening for military sexual trauma demonstrated a 22% prevalence rate in women veterans.(38) To better understand how often a self-reported history of sexual assault may have in fact included forcible sexual assault during military service, the final 25% of participants completed a second trauma measure, the Trauma Questionnaire (TQ).(47) More than 40% of women reporting a history of non-childhood sexual assault on this measure reported experiencing forcible sexual assault during military service. However, the substantive reduction in sample size, particularly of IBS cases, precluded valid assessment of abusive traumas experienced specifically during military service as independent IBS risk factors.

Depression and PTSD are among the top three diagnostic codes assigned to women veterans using the VA, with both shown to increase IBS risk in women veterans.(31, 32) PTSD was associated with an independent 100% increased IBS risk for 15 of 18 traumas we assessed. However, PTSD neither substantially explained nor modified the independent increased IBS risk conveyed by individual traumas. This modest to moderate increased IBS risk persisted for most traumas even after additional adjustment for depression.

A number of studies have evaluated individual abusive traumas like sexual abuse as IBS risk factors,(48–57) with most reporting increased IBS risk.(49–54, 56, 57) One group evaluated abusive traumas as IBS risk factors in a manner generally comparable to ours, i.e., a survey-based epidemiologic study conducted in a single general adult population not selected based on disease or clinical sub-specialty and with adjustment for concomitant psychological comorbidity. Their initial study was a mail survey to residents of a Penrith, Australia(48), with subsequent interview of a small sample of initial participants.(55) In contrast to our findings, they found almost no abusive traumas were associated with excess IBS risk after adjustment for general overall psychological morbidity level and neuroticism. Some of this discrepancy between our study findings is likely attributable to substantial differences in our underlying target populations, including our restriction to women, and specifically women veterans, who have a high lifetime trauma burden, and in measures of psychological comorbidity. Other relevant factors may include our use of direct recruitment with high participation rate compared to their use of a mailed survey and follow-up interviews in a sub-sample of participants with lower participation rates. The consistency of our findings suggestive of excess IBS risk with many abusive as well as non-abusive traumas supports the internal validity of our findings. Future comparably conducted research may help clarify whether abusive as well as non-abusive traumas are independently associated with excess IBS risk after adjusting for similar psychological comorbidities in other civilian and military populations.

Most studies that have evaluated the association between military service and risk of GI symptoms or disorders including IBS examined deployment in veterans of the first Gulf War. A 2010 Institute of Medicine (IOM) report evaluating this literature highlighted several key individual study findings including: 1) epidemiologic data demonstrating substantial deployment-related increases in IBS; 2) physiologic data demonstrating persistent low-grade colonic inflammation and increased visceral hypersentivity among deployed Gulf War veterans with IBS or IBS-type symptoms; and 3) epidemiologic data demonstrating greatly increased risk of IBS among veterans with a history of gastroenteritis, with effects particularly elevated among deployed veterans.(58) Unfortunately, we could not assess if there was a deployment-IBS association in our women veterans because we did not have access to necessary Department of Defense data to confirm deployment, including a history of combat or deployment-related gastroenteritis. However, our finding that multiple individual traumas, including many that may plausibly happen during deployment, independently increased risk of symptom-confirmed IBS in women veterans is consistent with the IOM findings. Our study offers several additions to the IOM-cited literature. First, we demonstrated that a trauma-IBS association exists for a broad range of traumas and persists even after adjustment for PTSD and depression. These two key psychological comorbidities in veterans that often co-occur with IBS were not controlled for in most of these Gulf War studies which often evaluated IBS as only one of multiple potential health outcomes. Second, because our study participants spanned a broad age-range, our results suggest a trauma-IBS association is unlikely to be limited to a single military cohort or era. Finally, our findings demonstrate a trauma-IBS association exists in women veterans, a population that was either absent or represented only a very small minority of the veterans evaluated in these earlier Gulf War studies.

Recent experimental and clinical studies have evaluated the association between early life trauma and stress in relation to IBS. In the neonatal maternal separation model (NMS) of IBS in rodents, several traumatic exposures during the early neonatal period increase visceral hyperalgesia,(59–62) a hallmark IBS symptom. For example, recent NMS studies demonstrated receipt of noxious stimuli like colorectal distention in this hyporesponsive periods leads to alterations in nociception, including increased immune and neurochemical responsivity at the level of the enteric nervous system.(61, 63) Others have shown this NMS-related visceral hyperalgesia persist in adulthood in response to subsequent noxious stimuli.(60, 62) Additional indirect evidence comes from epidemiologic studies demonstrating perinatal gastric suction(64) or living in a wartime environment during specific intervals in infancy(65) increases risk of functional gastrointestinal disorders including IBS in adulthood.

NMS in rodents also leads to chronic dysregulation in the limbic-pituatary-hypothalamic-adrenal (LPHA) axis,(66–68) with chronic LPHA dysregulation in humans associated with PTSD and depression(69), conditions often co-occurring with IBS. A recent structural MRI imaging study found increased hypothalamic gray matter and thinning of the anterior midcingulate limbic cortex in IBS cases compared to matched healthy controls.(70) Selective neural thinning that was observed only in IBS cases occurred in the supraspinal dorsolateral prefrontal cortex, which is extensively interconnected with both the limbic system and neocortex association areas and plays an important role in motor regulation and executive function. A recent functional MRI imaging study evaluated neural activation and pain perception in response to rectal distention in women, half of whom had IBS and half of whom had a history of abuse.(71) It demonstrated that both pain perception and a history of abuse were independently associated with increased activity in posterior and middle dorsal cingulate limbic cortex, with greatest effects observed in women who had both IBS and a history of abuse. Decreased activity in another limbic structure, the supragenual anterior cingulate cortex, was only observed in women with IBS who also had an abuse history. However, given the cross-sectional design and limited sample size of these studies, it is currently unknown to what extent these structural and functional differences may reflect increased susceptibility to versus are a result of or are reinforced by clinical presence of IBS.

Our results that a broad range of traumas are associated with increased IBS risk in women veterans are consistent with this emerging data suggesting potential biological linkage between trauma and IBS within the complex reciprocal interplay of the gut and the enteric nervous, central nervous and neuroendocrine systems. Our results also suggest the period in which traumas influence IBS risk may extend beyond early childhood into adulthood and also include a much broader range of major traumas than just abusive traumas. Additional indirect evidence supportive of potential role for trauma and stressors experienced in adulthood to increase IBS susceptibility comes from a recent experimental study in young healthy adult women that demonstrated visceral hypersensitivity and maladaptive intestinal epithelial responsitivity, both well-known to occur with IBS, were increased after receipt of noxious jejunal stimulation, with effects larger in healthy women placed under conditions of moderate compared to low background stress.(72)

Our study has multiple strengths including the prospective recruitment of a large and multiethnic sample of women veterans, restriction to a general care as opposed to specialty- or disease-specific clinical setting, high participation rate and use of validated measures. We also verified all IBS cases and controls in our study were veterans since up to half of women accessing VA healthcare are non-veterans (e.g., spouses/dependents or employees).(73) Our physician-performed medical record review substantiated our measure for the presence of IBS. It also provided data suggestive of substantial under-recognition of IBS among women veterans using the VA. Finally, to our knowledge, our study is the most comprehensive evaluation of individual traumas including multiple major lifetime traumas not previously examined for IBS in any population.

Our study also has several limitations. First, our study evaluated women veterans who have very high prevalence of IBS symptoms, trauma and psychological disorders; therefore, our findings may not generalize to women veterans who do not use the VA, to civilian women, or male veterans. The cross-sectional design precludes causal inferences as we cannot reliably establish whether IBS began before or after a trauma occurred. However, the consistent finding that a broad range of traumas, both abusive and non-abusive, experienced over a lifetime increased IBS risk suggests trauma is likely antecedent to IBS in at least some cases. Lastly, although recent community data suggests most IBS cases are symptomatic a decade later,(11) we were unable to evaluate differences in trauma-IBS association between those with long-standing versus only recent IBS.

Our study demonstrated lifetime history of a broad range of major life traumas beyond those experienced in early childhood or that are abusive in nature are associated with increased IBS risk in women veterans, even after adjusting for their most common psychological comorbidities, depression and PTSD. Our findings also have potentially important implications for healthcare delivery for women veterans. Specifically, they suggest women veterans who use the VA and who screen positive for PTSD, depression or military sexual trauma during system-mandated universal screenings for these conditions may also benefit from additional screening for IBS. Finally, our results suggestive of considerable underdiagnosis of IBS in women veterans, also suggest that obtaining a detailed trauma history may be useful in facilitating appropriate clinical recognition and diagnosis of IBS.

Abbreviations

BDI-II

Beck Depression Inventory, 2^nd edition

BDQ

Bowel Disorder Questionnaire

CI

confidence interval

IBS

irritable bowel syndrome

M-PTSD

Mississippi Scale Combat-Related PTSD

MST

military sexual trauma

PTSD

post-traumatic stress disorder

THQ

Trauma History Questionnaire

TQ

Trauma Questionnaire

VA

Department of Veteran Affairs