INTRODUCTION: We conducted a systematic review and meta-analysis to compare the prevalence of small intestinal bacterial overgrowth (SIBO) in patients with irritable bowel syndrome (IBS) and controls.
METHODS: Electronic databases were searched up to December 2018 for studies reporting SIBO prevalence in patients with IBS. Prevalence rates, odds ratios (ORs), and 95% confidence intervals (CIs) of SIBO in patients with IBS and controls were calculated.
RESULTS: We included 25 studies with 3,192 patients with IBS and 3,320 controls. SIBO prevalence in patients with IBS was significantly increased compared with controls (OR = 3.7, 95% CI 2.3-6.0). In studies using only healthy controls, the OR for SIBO in patients with IBS was 4.9 (95% CI 2.8-8.6). With breath testing, SIBO prevalence in patients with IBS was 35.5% (95% CI 33.6-37.4) vs 29.7% (95% CI 27.6-31.8) in controls. Culture-based studies yielded a SIBO prevalence of 13.9% (95% CI 11.5-16.4) in patients with IBS and 5.0% (95% CI 3.9-6.2) in controls with a cutoff value of 105 colony-forming units per milliliter vs 33.5% (95% CI 30.1-36.9) in patients with IBS and 8.2% (95% CI 6.8-9.6) in controls with a cutoff value of 103 colony-forming unit per milliliter, respectively. SIBO prevalence diagnosed by lactulose breath test is much greater in both patients with IBS (3.6-fold) and controls (7.6-fold) compared with glucose breath test. Similar difference is seen when lactulose breath test is compared with culture methods. OR for SIBO in patients with IBS-diarrhea compared with IBS-constipation was 1.86 (95% CI 1.83-2.8). Methane-positive breath tests were significantly more prevalent in IBS-constipation compared with IBS-diarrhea (OR = 2.3, 95% CI 1.2-4.2). In patients with IBS, proton pump inhibitor was not associated with SIBO (OR = 0.8, 95% CI 0.5-1.5, P = 0.55).
DISCUSSION: This systematic review and meta-analysis suggests a link between IBS and SIBO. However, the overall quality of the evidence is low. This is mainly due to substantial "clinical heterogeneity" due to lack of uniform selection criteria for cases and controls and limited sensitivity and specificity of the available diagnostic tests.