Proton Pump Inhibitor Use and Small Bowel Intestinal Overgrowth (SIBO) - Another Allegation of Harm?

David A. Johnson, MD


August 18, 2010

Increased Incidence of Small Intestinal Bacterial Overgrowth During Proton Pump Inhibitor Therapy

Lombardo L, Foti M, Ruggia O, Chiecchio A
Clin Gastroenterol Hepatol. 2010;8:504-508

Study Summary

Concern about the potential link between potent acid suppression and enteric infections has been expressed. However, the literature regarding the effect of antisecretory therapy on the risk for enteric infections has been conflicting. The focus of these alleged associations to date has been on Salmonella, Campylobacter, Shigella, and Clostridia speciesand Escherichia coli. Lombardo and colleagues suggest a potential link between the use of proton pump inhibitors (PPIs) and small intestinal bacterial overgrowth (SIBO).

In this study from Italy, glucose hydrogen breath tests (GHBTs) were given to 450 consecutive patients -- 200 with gastroesophageal reflux disease (GERD) who were taking PPIs (median 36 months), 200 with the irritable bowel syndrome (IBS) who had not taken PPIs for at least 3 years, and 50 healthy controls. The GHBT was used as an indirect, surrogate test for SIBO, and all patients completed a gastrointestinal symptom score evaluation. The IBS patients were also classified as diarrhea (40%), constipation (40%), or mixed (20%) subtypes.

Overall, SIBO was suggested by a positive result on GHBT in 50% of patients receiving PPIs, 24.5% of patients with IBS, and 6% of controls. These differences were statistically significant: PPI vs IBS (odds ratio [OR], 3.14; 95% confidence interval [CI], 2.06-4.80), PPI vs control (OR, 16.0; 95% CI, 4.8-53.0), IBS vs control (OR, 6.12; 95% CI, 3.8-7.5). The prevalence of SIBO was similar in all age groups except one -- 41- to 60-year-olds, in which SIBO was more frequent in the PPI users (P < 0.01). Prevalence of SIBO was higher in patients who had 1 year of treatment with PPIs than in those who had <1 year (OR, 11; 95% CI, 5.5-21.8).

For those defined as patients with SIBO, rifaximin 400 mg 3 times daily for 14 days was given as an open-label treatment. Eradication was achieved in 87% of the PPI group and in 91% of the IBS group (P = NS). There was no apparent difference in the success of eradication and length of use of PPIs. Following eradication, absence was reported for bloating (90%), diarrhea (94%), and abdominal pain (92%).


At first glance, it might seem that SIBO could be added to the growing concerns regarding PPI use. Recognizably, the gastrointestinal tract is not sterile, and conceptually, alterations in the normal physiologic milieu might cause bacterial changes that have clinical consequences. The normal gut microflora is maintained primarily by 4 major mechanisms: gastric acid secretion, pancreatic secretion, visceral motility, and mucosal structural integrity. Helicobacter pylori status was established for only 68% of the patients. Clearly, H pylori infection can influence the acid secretion status of the patient and hence is another undefined potential source of bias for a sizeable number of patients studied. Acid secretion can be altered with the use of PPIs, but even the most potent PPIs, with twice-daily dosing, maintain gastric pH > 4 for 18-19 hours. Most bacteria are killed within 30 minutes at a pH of 4, so PPI use clearly results in an acidic exposure for a length of time that greatly exceeds this threshold. The investigators' suggestion that PPI use should be tailored to "tissue repair and physiologic respect for gastric pH" seems to ignore the aforementioned time elements for acid suppression with PPIs.

Another concern that should be addressed before we jump on the SIBO and PPI bandwagon is the lack of mention of a validated questionnaire for assessment of gastrointestinal symptoms. Furthermore, the retrospective evaluation of patients' symptoms for a 3-year period introduces a strong potential for recall bias. In fact, the patients recorded their more recent symptoms as being more severe -- hence, the suggestion by the investigators that more time of exposure to PPIs was associated with more severe symptom reporting.

Furthermore, although patients with GERD and those with IBS who had not taken PPIs for 3 years were separated, patients with GERD who might also have IBS were not evaluated. This could be a major undefined confounding variable.

Clearly, some response to the use of rifaximin for patients with suspected SIBO, as defined by the positive GHBT results, was evident. Although this breath test is used as indirect evidence of SIBO, some controversy has arisen with regard to its reliability for defining SIBO. It also seems peculiar that the patients with SIBO differed from patients with IBS only in 1 age range (41-60 years).

In summary, emerging evidence is clear on the role of bacterial flora changes in SIBO, particularly for patients with IBS. The data from the present study, however, are not convincing for extrapolation of PPI use for SIBO risk. Arguably, this needs further evaluation; however, at present, it is perhaps a question to consider for symptomatic patients but not a reason to stop using PPIs in patients who are receiving them for appropriate indications.