Influence of Dietary Habits on the Risk of Inflammatory Bowel Disease: A Case Control Study

Introduction: At the Los Angeles County University of Southern California Medical Center (LAC+USC) we service an inflammatory bowel disease (IBD) clinic for a low economic population. We evaluated the frequency of the types of IBD and identified risk factors associated with ulcerative colitis (UC) versus Crohns disease (CD). Methods: This is a retrospective, cross-sectional review of patients with the diagnosis of IBD including UC, CD or indeterminate colitis that were seen at the Roybal Comprehensive Health Clinic between 1/01/2004 through 12/31/2009. Ulcerative colitis was defined as a continuous inflammation of the colon involving the distal rectum and with architectural distortion on histology. Crohn's disease was defined by characteristic small and large bowel changes on imaging studies and/or segmental colitis in untreated patients on colonoscopy. Using an IBD database, ICD-9 diagnosis codes, electronic medical records, and the Clinical Outcomes Research Initiative (CORI) system, the type of colitis, extent of disease, and demographic information were obtained. Univariate and stepwise multivariate analysis were performed. Results: A total of 287 patients with IBD were identified. The frequency of UC in this population was 68% (196/287); (95% Confidence Interval (CI): 63%-74%); CD 24% (70/287); (95% CI; 19%-29%); and indeterminate colitis (IC) 7% (21/287); (95% CI: 4-10). In the 266 patients with UC or CD, 55% were male, 65% over 30 years of age, 69% Hispanic, 12% African American (AA), 9% Asian and 9% non-Hispanic American. Sixty-one percent were immigrants and 52% were from Latin America. We did not include in the analysis education, history of smoking and family history of IBD because of the high percentage of missing data 37, 30 and 47% respectively. Table 1 provides the association of each risk factor with UC vs CD. Stepwise logistic regression yielded two significant independent risk factors: AA ethnicity [adjusted OR: 0.34; 95% CI: 0.15, 0.77; p=0.01] and immigrant [adjusted OR: 2.16; 95% CI: 1.19, 3.94, p=0.01). The R-squared for AA was 7% and that for immigrant was 3%. These two factors explained 10% of the variation between UC and CD. Conclusion: In the LAC+USC IBD clinic which serves low socioeconomic patients, there is a 3 to 1 ratio of UC to CD. Immigrant and non-African American ethnicity are predictive of UC whereas non-immigrant and African American ethnicity are predictive of CD.