Dysmenorrhea in women with Crohn's disease: a case-control study.

Abdominal and pelvic symptoms in young women with Crohn’s disease (CD) are often attributed to inflammatory bowel disease (IBD) when, in fact, other entities may be responsible. Dysmenorrhea, or painful menstruation, is the most common disorder of reproductive-aged women and has symptoms which overlap with active CD [ 1 , 2 ]. Determining the prevalence of dysmenorrhea and its contribution to overall pain severity and to perceived CD activity may be helpful in tailoring therapy in women with CD and avoiding the over assignment of gastrointestinal (GI) symptoms to IBD. It is well established that painful disorders of the GI tract such as irritable bowel syndrome (IBS) are frequently co-morbid with primary dysmenorrhea, with some women reporting an increase in symptoms during menstruation [ 3 , 4 ]. The relationship between menstrual distress and inflammatory disorders of the GI tract such as CD is less well understood; however, prior studies suggest that menstrual symptoms may be perceived as increased CD activity due to their overlapping symptoms [ 5 ]. In addition, common pathophysiologic mechanisms may exist between the two disorders. Kane and colleagues suggested that the greater rates of diarrhea that women with IBD report during the premenstrual and menstrual phases of the cycle may be related to increased colonic contraction from intestinal prostaglandin production during these phases and/or increasing progesterone levels across the menstrual cycle [ 5 ]. It is important to note, however, that a direct correlation between plasma levels of ovarian hormones and perimenstrual symptoms has not been established [ 6 ]. To date, no study has determined the prevalence of dysmenorrhea or the impact of menstrual pain in women with CD using validated methods. Furthermore, no study has examined the relationship between dysmenorrhea and CD activity scores or between dysmenorrhea and HRQOL in women with CD. Closing this knowledge gap is imperative. The purpose of this study was to examine: 1) the impact of dysmenorrhea on overall pain severity and pain medication use in women with CD vs. healthy controls, 2) the relationship between dysmenorrhea and CD activity scores, and 3) the effect of dysmenorrhea on HRQOL in women with CD.

We recruited women aged 18 to 48 with regular menstrual cycles, defined as cycles ranging from 24 to 35 days and lasting 4 to 6 days for 3 consecutive months prior to study entry, and with CD as determined by endoscopy or radiographic imaging. Women aged 18 to 48 with regular menstrual cycles and without CD or other major co-morbidity were recruited as controls. Subjects (cases and controls) were excluded if they were pregnant or breastfeeding, using or recently (within 3 months) had used intradermal, injectable or intrauterine contraception, less than 3 months post-partum, within 3 years of menarche, or had been diagnosed with ulcerative colitis, indeterminate colitis, IBS or other functional bowel disorder (FBD), premenstrual syndrome, endometriosis, pelvic inflammatory disease, or other significant GI or gynecologic co-morbidity as determined by the principal investigator. Subjects were actively recruited in the primary care and gastroenterology clinics at the University of Wisconsin Hospital and Clinics from May 2010 to July 2011.. Passive recruitment was also conduced using fliers posted on bulletin boards at the University hospital. Eligible subjects were offered $30 for participation. Written informed consent was obtained. All data collection except for the Crohn’s Disease Activity Index (CDAI) score was conducted at a single visit. Data elements were collected by structured interview, self-completion of validated questionnaires, or standardized data abstraction from the medical record (see Table 1 ). Regarding the CDAI, cases were given a 7-day symptom diary to return by mail. This study was approved by the Institutional Review Board at the University of Wisconsin in April 2010. Dysmenorrhea was defined according to Andersch and Milsom [ 7 ] as severely painful menstruation during the past 3 months leading to missing work/school or significant life events. Global pain severity was evaluated using a 0 to 10 scale (0=no pain to 10=worst imaginable pain). Subjects were asked to rate their pain over the past 3 months. Menstrual symptoms were evaluated using Form C of the Moos Menstrual Distress Questionnaire (MDQ), a self-report inventory for measuring cyclical perimenstrual symptoms [ 8 ]. Form C consists of 46 items, each describing a specific symptom. Subjects are asked to rate their experience with each symptom on a 5-point scale (0=no experience of symptom; 4=symptom present and severe). Ratings of each symptom were made separately for the 3 phases of each subject's most recent cycle: menstrual (during menstruation), premenstrual (4 days before menstruation), and intermenstrual (the remainder of the cycle). The MDQ items are divided among 8 symptom scales as follows: pain, water retention, autonomic reactions, negative affect, impaired concentration, behavior change, arousal, and control. MDQ composite scores and pain subscores were calculated for each subject. CD activity was measured using the CDAI [ 9 ] and Harvey-Bradshaw Index (HBI) [ 10 ]. Quality of life was measured using the Short Form-36 (SF-36) [ 11 , 12 ], and the Inflammatory Bowel Disease Questionnaire (IBDQ) [ 13 ]. A total score and 2 standardized summary scores, the physical component summary (PCS) score and the mental health component summary (MCS) score, were calculated for each subject using the SF-36. Using the IBDQ, a total score ranging from 32 (very poor HRQOL) to 224 (perfect HRQOL) was calculated for each case along with 4 sub domain scores. A sample size of 118 (59 controls, 59 cases) was calculated based on a two-sided test with alpha = 0.05 and beta = 0.80. This is based on the published prevalence of dysmenorrhea in otherwise healthy women of 25% and a prevalence of dysmenorrhea in women with IBS of 50% which was extrapolated to be similar in women with IBD. Descriptive statistics, including means and standard deviations for continuous variables and frequencies for categorical variables, were calculated for cases and controls using standard methods. An overall odds ratio was calculated for having dysmenorrhea among cases and controls. Adjustment was then made for confounding factors, including young age (i.e. age < 20 years), nulliparity, depression/anxiety, heavy menses, and hormone-based contraceptive use. To compare epidemiological and clinical characteristics, parametric and non-parametric continuous variables were compared with Student’s t-test and the Mann–Whitney test respectively, and categorical variables were compared with the Chi-square test or Fisher’s exact test, as appropriate. Pearson’s r correlation coefficient was used for correlation between menstrual distress scores and CD activity scores. Linear regression was performed to determine the relationship between dysmenorrhea and the disease activity scores. All analyses were done using SPSS-PASW Statistics GradPack 17.0.

Fifty-four cases and 66 controls were recruited. The majority of cases had colonic disease only (64.8%). Thirty nine percent of cases had ileocolonic disease, 5.6% had small bowel disease only, and 3.7% had disease in other locations. Fistulas were present in 26% of cases: 16.7% had peri-anal fistulas and 9.3% had non-perianal fistulas. Five out of 54 cases (9.3%) had undergone IBD-related surgery: 1.9% had undergone partial colon resection, 5.5% were status post ileocolonic resection, and 1.9% had undergone perianal fistula drainage. Cases and controls did not differ significantly in demographic characteristics, level of educational achievement, and social habits ( Table 2 ). The prevalence of gynecologic surgeries did not differ significantly between cases and controls. nor did menstrual and pregnancy histories ( Table 3 ). We found significantly higher use of non-steroidal anti-inflammatory drugs (NSAIDS) by controls compared to cases (p<0.001). No significant difference in aspirin or hormone use was found between the 2 groups. The prevalence of dysmenorrhea was 40% in cases and 46% in controls. The odds ratio for dysmenorrhea in women with CD was 0.77 (95% CI: 0.37 to 1.60). We tested a multinomial logistic regression model for dysmenorrhea in cases with age, pregnancy, depression, heavy menses, and hormone use as covariates. The overall model was not significant nor were the likelihood ratio tests for the individual variables. Pain scores were significantly higher in subjects with dysmenorrhea compared to those without among both cases and controls (p<0.001, both). When evaluating only subjects with dysmenorrhea, there was no difference in pain scores between cases and controls ( Table 4 ). Pain medication use of any type for menstrual pain was not significantly different in the presence or absence of dysmenorrhea among controls. NSAID use, however, was significantly higher in controls with dysmenorrhea (p=0.005) in contrast to lower use among cases. Greater pain medication use for menstrual pain was found for cases with dysmenorrhea vs. cases without (p=0.0001); however, there was no difference in NSAID use between these groups. In women without dysmenorrhea, we found that more controls used pain relievers for menstrual pain (p=0.002). There was no significant difference in pain medication use of any type for menstrual pain between cases and controls with dysmenorrhea; however, controls with dysmenorrhea used more NSAIDS than did cases (p=0.001). Linear regression analysis was performed with dysmenorrhea and CD as predictors for pain severity. Dysmenorrhea was highly significant (p<0.0001) and CD trended towards significance (p=0.05). Multinomial logistic regression with an interaction term between dysmenorrhea and CD did not change the model. As expected, when comparing subjects (i.e. cases plus controls) with dysmenorrhea to subjects without dysmenorrhea, we found the MDQ composite score and the MDQ pain subscore to differ significantly between groups (p=0.03, both). However, when comparing cases with dysmenorrhea to controls with dysmenorrhea we found no significant differences in either score ( Table 5 ). Differences in MDQ composite scores and MDQ pain subscores were not significantly affected by CD location; however, the study was underpowered to find such differences. The presence of perianal fistulas also did not significantly affect MDQ scores. We performed a one-way ANOVA to examine differences between cases and controls on the MDQ pain subscore and composite score with hormone, NSAIDS, and aspirin use as covariates. Cases reported significantly higher pain than controls on the pain subscore (p=0.04) but no significant difference was found between groups on the composite score. The mean CDAI score (SD) for all cases was 108.4 (80.6) and the mean HBI (SD) score was 3.4 (3.1). By either scoring system our cases, on average, were in remission. The differences in mean CDAI scores and HBI scores for cases with and without dysmenorrhea were not significant. Pearson correlation coefficients were calculated to evaluate for correlations between CD activity and the MDQ composite score and MDQ pain subscore. The Pearson r correlation between CDAI and MDQ composite score was weakly positive ( r =0 .42, p=0.004) as was the correlation between CDAI and MDQ pain subscore ( r =0 .44, p =0.003). The correlation between the HBI score and MDQ composite score trended towards significance ( r =0.27, p=0.051) and was weakly positive between the HBI score and MDQ pain subscore ( r =0.32, p=0.019). Total SF-36 scores were significantly lower in cases with dysmenorrhea compared to cases without (p=0.016). When comparing summary scores PCS scores, but not MCS scores, were also significantly lower (p=0.004) in cases with dysmenorrhea ( Table 6 ). The mean total IBDQ score for all cases was 168.9 (32.7). The mean total IBDQ scores for cases with and without dysmenorrhea were, 163.3 (33.3) and 173.1 (32.7), respectively. The difference in total scores was not significant; however, the difference in the systemic symptoms subscore was (p=0.035). No significant difference was found for the other IBDQ subscores. Univariate analysis with NSAID use and hormone use as covariates did not show any significant difference in SF-36 or IBDQ scores between cases with and without dysmenorrhea.

Dysmenorrhea is the most common gynecologic complaint and the leading cause of school and work absenteeism among women. The prevalence of dysmenorrhea ranges widely in the literature depending on the population studied and the measurement method used [ 7 , 14 – 17 ]. Conservatively dysmenorrhea affects 25% of menstruating adults. In addition, up to 90% of adolescents may be affected [ 1 ]. While lower abdominal cramps are the most common symptom of dysmenorrhea, women may also experience nausea and vomiting, loss of appetite, diarrhea, general aching as well as irritability, sleep disturbances, and depression [ 18 ]. The severity of dysmenorrhea varies depending on its interference with daily activities [ 19 ] and patients’ perception of pain [ 20 ]. Various pain scales and scores have been used to grade the severity of dysmenorrhea [ 21 – 23 ]. Dysmenorrhea is categorized as primary if there is no underlying pelvic pathology or secondary if pelvic pathology is present. Prostaglandin F 2α , a potent myometrial stimulant and vasoconstrictor in the uterus, is believed to mediate primary dysmenorrhea [ 24 ]. Leukotrienes [ 25 , 26 ] and vasopressin [ 27 , 28 ] may also contribute. Although the distinction between primary and secondary dysmenorrhea may be used to determine treatment approach, differences in pain experience or sensitivity have not been well-established between the two groups [ 29 ]. Given the prevalence of dysmenorrhea among reproductive-aged women and its shared symptoms with CD, symptom confusion between the two entities may occur. Thus, women with CD and dysmenorrhea may perceive their disease as being more active and undergo more testing and/or medication changes than their non-dysmenorrheic counterparts. Documenting the prevalence of dysmenorrhea in women with CD and determining the relationship between dysmenorrhea, CD and pain severity are therefore important. In this study we found dysmenorrhea was highly prevalent in women with CD, occurring in 40% of cases. We also found that it had an additive effect on global pain severity. Although the odds of dysmenorrhea was not significantly higher among cases compared to controls, prior studies suggest that menstruation causes GI symptoms, particularly diarrhea, more frequently in women with IBD compared with healthy controls and produces cyclic variation in symptoms in women with both UC and CD [ 30 , 31 ]. Therefore given the prevalence of dysmenorrhea and its impact on IBD symptoms it is a co-morbidity for which screening is needed. Failure to recognize dysmenorrhea may contribute to the higher frequency of bowel symptoms reported by women with IBD compared to men [ 32 ] and lead to the erroneous conclusion that abdominal pain is solely attributed to IBD We found no significant difference in pain medication use of any type for menstrual pain between controls with and without dysmenorrhea. However, among cases, pain medication use was significantly associated with dysmenorrhea. One possible explanation for this is that non-IBD women may be using pain medication for non-dysmenorrheic menstrual pain, whereas women with CD, who are habituated to a baseline level of pain, may only be using medications when menstrual pain is severe. With regards to NSAID use, we found the opposite trends. Controls with dysmenorrhea used significantly more NSAIDS than controls without dysmenorrhea, in contrast to cases who did not. This likely reflects the reluctance to use NSAIDS by patients with IBD given their association with increased disease activity [ 33 ] and hospitalization [ 34 ]. It also suggests that women with CD may be avoiding the preferred initial treatment of dysmenorrhea and using less effective medications for the treatment of menstrual pain [ 35 ]. Practitioners should consider on a case-by-case basis whether NSAIDs are an appropriate treatment for their dysmenorrheic IBD patients. A secondary aim of this study was to explore the impact of dysmenorrhea on measures of CD activity and on HRQOL. Given the overlap in symptoms between dysmenorrhea and active CD we hypothesized that cases with dysmenorrhea would have higher CD activity scores compared to cases without. However, we found that neither with the CDAI nor with the HBI was there a difference between groups. Thus at least for patients with CD in remission dysmenorrhea was not associated with higher CDAI or HBI scores. We did, however, find positive correlations between the MDQ composite score and pain subscore and the CDAI and HBI. This suggests that there may be a tendency for women with CD with higher levels of menstrual distress to have higher CD activity scores. Greater perceived CD activity may lead to more health-care seeking behavior by women with co-morbid dysmenorrhea. This, in turn, may lead to more rigorous evaluation for active disease and unwarranted changes to CD treatment regimens. To evaluate the impact of dysmenorrhea on HRQOL among women with CD, we used both generic and disease-specific measures. We found total SF-36 scores to be lower in cases with dysmenorrhea vs. controls with dysmenorrhea. PCS scores were also significantly lower and there was a trend towards lower MCS scores. These differences may be explained by the lower HRQOL that individuals with CD experience at baseline. Although there was no significant difference in total IBDQ scores, we found that cases with dysmenorrhea scored more poorly on the systemic symptoms subscore of the IBDQ compared to cases without dysmenorrhea. This subscore is based on questions pertaining to feeling tired and worn out, energy, general wellness, sleep disturbances, and weight problems. Interestingly, the bowel symptoms subscore of the IBDQ was not affected by dysmenorrhea despite its shared GI manifestations with active CD. Nevertheless, recognizing the impact of dysmenorrhea on the physical and systemic symptoms experienced by women with CD provides another target for improving QOL in this population. This study is potentially limited by several possible sources of bias. First, although we used a standard definition of dysmenorrhea, the perception of whether menstruation is painful is subjective, allowing for possible misclassification bias. Results from the MDQ, however, suggest that subjects did correctly self-identify as those with dysmenorrhea had greater menstrual distress scores compared to those without. Second, recall bias may have altered the assessment of menstrual pain severity and pain medication use. We used Form C of the MDQ to address this as it has been validated to retrospectively measure menstrual distress. Lastly, selection bias may have led to an overestimation of dysmenorrhea rates among non-IBD women as most of our controls were recruited from primary care clinics focused on women’s health. Thus the prevalence of dysmenorrhea among our controls could be greater than the true prevalence in the general population. If this were the case, the odds of dysmenorrhea may in fact be higher in women with CD which makes its recognition even more important. Given the paucity of data in IBD and menstruation, future directions for this work include evaluation of the effect of painful menstruation on perceived disease activity in ulcerative colitis. More precise correlation between menstrual symptoms an IBD activity is also needed. Studies with repeated use of Form T of the MDQ, a self-report instrument which prospectively records reactions or symptoms experienced by women on a specific day, a 30-day GI symptoms diary, and inflammatory markers may achieve this. In addition, intervention studies aimed at treating dysmenorrhea are needed to assess the improvement in HRQOL women with CD experience after their menstrual distress has been addressed.

Dysmenorrhea and CD are additive predictors for global pain severity. Dysmenorrhea in women with CD was associated with higher pain medication use for menstrual pain but not with higher NSAID use, suggesting that women with CD may be using less effective therapies to treat dysmenorrhea. Menstrual distress as measured by the MDQ positively correlates with CD activity scores suggesting that menstrual symptoms may impact perception of patients’ IBD activity. This may lead to more health-care seeking behavior and/or lower health status. Lower scores in some aspects of HRQOL were also found in subjects with CD and dysmenorrhea, suggesting that dysmenorrhea is a potential target for interventions to improve the overall health status of this population. Gastroenterologists should consider this entity when evaluating pain in women with CD.