rss
Evid Based Med 18:193-194 doi:10.1136/eb-2012-101057
  • Aetiology
  • Systematic review with meta-analysis

Proton pump inhibitor use may be associated with an increased risk of Clostridium difficile infection

  1. Edwin van Wijngaarden2
  1. 1 Department of Pharmacotherapy, Pharmacotherapy Outcomes Research Center, University of Utah College of Pharmacy, Salt Lake City, Utah, USA
  2. 2 Public Health Sciences, University of Rochester School of Medicine and Dentistry, Rochester, New York, USA
  1. Correspondence to : Dr Vanessa Stevens
    Pharmacotherapy Outcomes Research Center, University of Utah College of Pharmacy, 421 Wakara Way, Suite 208, Salt Lake City, UT 84103, USA; vanessa.stevens{at}pharm.utah.edu

Commentary on: [CrossRef][Medline]Google Scholar.

Context

Proton pump inhibitors (PPI) are a safe and effective treatment of acid-related disorders such as gastroesophageal reflux disease and Barrett's oesophagus. However, they may be associated with an increased risk of Clostridium difficile infection (CDI).

Methods

Kwok and colleagues conducted a systematic review and meta-analysis of observational studies evaluating the association between PPI and the risk of primary or recurrent CDI. Relevant articles were identified through a comprehensive literature search including MEDLINE, EMBASE (from database inception to December 2011) and reference lists of articles retrieved from these bibliographical databases. Studies were eligible for inclusion in the meta-analysis if they were either case–control or cohort design, and either reported OR, risk ratios (RR) or provided sufficient data to allow manual calculation of these effect measures. Data were abstracted by one of three reviewers and verified by at least one other reviewer. Study authors were contacted where clarification was required. The primary outcome of interest in this analysis was CDI as confirmed by a laboratory, by the presence of billing codes in the administrative data, or by the prescription of oral vancomycin. Inverse variance methods were used to estimate pooled RRs using random effect models. Heterogeneity was assessed using I2 statistics, and subgroup analyses were undertaken according to pre-specified criteria. Numbers needed to harm (NNH) estimates were generated from pooled RRs.

Findings

Individuals who received PPI were 1.74 times (95% CI 1.47 to 2.85) more likely to develop incident CDI and 2.5-fold (95% CI 1.16 to 5.44) more likely to develop recurrent CDI as compared with patients who did not receive PPI. The NNH was estimated to be 67 in the absence of antibiotic exposure and 22 in patients with concomitant antibiotics. The meta-analysis showed a substantial degree of statistical heterogeneity, even when analyses were stratified by study design or other potential sources of heterogeneity (I2=67–97%, dependent upon subgroup).

Commentary

Clostridium difficile infection (CDI) has been the subject of renewed interest in the last decade, primarily owing to changing incidence rates attributed, in part, to the emergence of a novel epidemic strain with increased resistance to fluoroquinolones. Antibiotics are considered the most important risk factor for CDI. In addition, there is emerging evidence that proton pump inhibitors (PPI) constitute a risk factor for CDI. However, there is debate as to whether this body of research constitutes convincing evidence, owing to methodological limitations of the existing studies.1 ,2 Randomised controlled trials remain the gold standard for determining the beneficial effects of treatment alternatives, but are not appropriate for the evaluation of adverse events for obvious ethical reasons. As such, a synthesis of the existing observational research results can provide a valuable insight into the current state of the literature.

Overall, it appears that PPI are associated with increased risk of incident and recurrent CDI, based on a consistent direction of association observed in the majority of studies included in the meta-analysis. However, the validity of results from a meta-analysis depends heavily upon the quality of the studies included. The evaluation of the association between PPI and the risk of CDI was not the primary objective of the majority of studies. This may have resulted in the lack of detailed exposure information and variability in covariates controlled for, both of which may have contributed to the statistical heterogeneity. The authors did not describe the differences in interpretation of effect measures obtained in case control and cohort studies with respect to the direction of inquiry in study design. That is to say, for case–control studies, the interpretation should be that patients with CDI had increased odds of exposure to PPI. In contrast, the interpretation of cohort studies should be that patients with PPI exposure had an increased risk of developing CDI. Combining these two effect measures into one overall summary relative risk estimate is typically not desirable. The authors state that they considered none of the included studies to be at low risk of bias owing to the lack of detailed information on PPI exposure. However, they did not consider quality of exposure assessment methods in their assessment. At least one study categorised exposure according to frequency of administration, providing a rough dose-response evaluation,3 and another allowed PPI exposure to change over time during the study period to account for any changes in therapy.4 Both studies found an increased risk of CDI among PPI users. There are methodological limitations in this review, but it provides some support for an association between PPI use and CDI, especially among patients who are also receiving antibiotics.

Footnotes

  • Funding None.

  • Competing interests None.

References

Free Sample

This recent issue is free to all users to allow everyone the opportunity to see the full scope and typical content of EBM.
View free sample issue >>


Don't forget to sign up for content alerts so you keep up to date with all the articles as they are published.