Background: Recent evidence suggests a role for enteric bacteria in the development of symptoms in subjects with irritable bowel syndrome (IBS). The concept was initially based on the common finding of an abnormal lactulose breath test to suggest the presence of small intestinal bacterial overgrowth (SIBO). Despite successful antibiotic studies based on these findings, the premise of bacterial overgrowth was met with skepticism due to the perceived inaccuracies of breath testing in humans. In this study, we conduct a systematic review and meta-analysis of the studies using breath testing in IBS subjects with comparison to healthy controls.
Methods: A literature search was conducted to identify studies in which breath testing was conducted in subjects with IBS. Once abstract titles were identified, studies that examined breath testing in a case-control fashion were identified for paper review. After exclusion criteria were applied, the remaining papers were examined using meta-analysis, among which the prevalence of an abnormal breath test in IBS subjects was compared. Further analysis was conducted among studies with appropriate age- and sex-matched controls.
Results: Eleven studies met criteria for inclusion in the meta-analysis. Although heterogeneity was identified among studies on the basis of substrate used, lactulose was the most common substrate. In pooling the 11 studies, breath testing was more often abnormal among IBS subjects than health controls (OR=4.46, 95% CI=1.69-11.80). This was further evident in more appropriately designed age- and sex-matched studies (OR=9.64, 95% CI=4.26-21.82). The overall sensitivity and specificity for breath testing in separating IBS from healthy controls was 43.6 and 83.6%.
Conclusions: This study demonstrates that the breath test findings in IBS appear to be valid. While this meta-analysis does not suggest that the breath test findings imply SIBO, the abnormal fermentation timing and dynamics of the breath test findings support a role for abnormal intestinal bacterial distribution in IBS.