A randomly selected population sample undergoing colonoscopy: prevalence of the irritable bowel syndrome and the impact of selection factors

ObjectiveTo analyse the epidemiology of irritable bowel syndrome (IBS) symptoms in a random sample of the general population and in a subsample consenting to a colonoscopy, and to what extent this introduces symptom selection bias.Materials and methodsOverall, 3347 randomly selected Swedish adults aged 18-70 years were mailed the validated Abdominal Symptom Questionnaire (ASQ). Responders (n=2293; 68.5%) were contacted by phone, and 745 consented to a colonoscopy. All nonresponders were contacted by phone; 265 were reached and asked seven key ASQ questions. Colonoscopy participants also completed the Rome II Modular Questionnaire.ResultsThe prevalence of IBS on the basis of the mailed ASQ (troublesome abdominal pain and bowel disturbance in the past 3 months) was 26.2% [95% confidence interval (CI): 24.4-28.0] among the ASQ responders and 36.6% (95% CI: 33.2-40.1) among the colonoscopy participants (P<0.001). Nonresponders had a lower prevalence of IBS (15.8%; 95% CI: 11.4-20.3, P<0.001) than ASQ responders. Colonoscopy participants were slightly older than noncolonoscoped participants completing the ASQ (P<0.001), but men and women were equally represented and no significant socioeconomic differences were identified. The prevalence of IBS was 14.8% (95% CI: 12.2-17.5) on the basis of the Rome II Modular Questionnaire in colonoscopy participants and 14.5% (95% CI: 11.9-17.2) when visible inflammatory disease was excluded. Of the colonoscopy participants, 31.9% (95% CI: 28.5-35.3) were symptom free.ConclusionIBS symptoms are common and rarely explained by visible inflammatory disease or cancer. There was a modest selection bias by IBS in participants accepting a screening colonoscopy, but still, one-third were symptom free. Thus, conclusions for the general population can be made from findings in the study cohort.