CDI indicates Clostridium difficile infection.
Increased odds of recurrent Clostridium difficile infection (CDI) with gastric acid suppressants shown by the random-effects model.
No evidence of publication bias was noted. M-H indicates Mantel-Haenszel test; OR, odds ratio.
Plot demonstrates increased risk of recurrent Clostridium difficile infection with gastric acid suppressants by the random-effects model (odds ratio, 1.38; 95% CI, 1.08-1.76).
eTable 1. Search Strategy
eTable 2. Newcastle-Ottawa Scoring for All Included Studies
eTable 3. Covariables Adjusted in Each Study: Subgroup Analysis
eFigure 1. Forest Plots by Definition of Recurrence
eFigure 2. Forest Plots by Type of Acid Suppression Used
eFigure 3. Forest Plots by Type of Study
eFigure 4. Forest Plot of Studies in the Inpatient Setting Only
eFigure 5. Forest Plots by Type of Diagnostic Assay Used for C difficile Infection
Customize your JAMA Network experience by selecting one or more topics from the list below.
Tariq R, Singh S, Gupta A, Pardi DS, Khanna S. Association of Gastric Acid Suppression With Recurrent Clostridium difficile Infection: A Systematic Review and Meta-analysis. JAMA Intern Med. 2017;177(6):784–791. doi:10.1001/jamainternmed.2017.0212
Does concomitant use of gastric acid suppressant medications (proton pump inhibitors and histamine H2 receptor blockers) increase the risk of recurrent Clostridium difficile infection?
In this meta-analysis of 16 studies comprising 7703 patients, the use of gastric acid suppressants was associated with a significantly increased risk of recurrent C difficile infection. Subgroup analyses of studies with potential confounders also confirmed an increased risk of recurrent C difficile infection with use of these medications.
It may be reasonable to stop gastric acid suppressants in patients with C difficile infection; limiting unnecessary use of these medications may help to decrease both recurrent C difficile infection and health care costs.
Gastric acid suppression has been associated with an increased risk of primary Clostridium difficile infection (CDI), but the risk of recurrent CDI in patients taking gastric acid suppressant medications is unclear.
To perform a systematic review and meta-analysis to evaluate the association between gastric acid suppressants and recurrent CDI.
MEDLINE, EMBASE, the Cochrane Central Register, the Cochrane Database, and Web of Science were searched from January 1, 1995, to September 30, 2015, for studies assessing the association between gastric acid suppressant exposure and recurrent CDI. Search terms included Clostridium difficile, pseudomembranous colitis, proton pump inhibitor, and histamine H2 blocker.
Case-control studies, cohort studies, and clinical trials that included patients with CDI who did or did not receive gastric acid suppressant therapy and who were evaluated for recurrent CDI were included, with no restriction on study setting (inpatient or outpatient).
Data Extraction and Synthesis
The Newcastle-Ottawa scale was used to assess the methodologic quality of included studies. In this scale, case-control and cohort studies were scored on selection, comparability, and ascertainment of the outcome of interest. Data were independently abstracted to a predetermined collection form by 2 investigators. Summary odds ratio estimates with 95% CIs were calculated using the random-effects model and software to calculate the pooled effect size of studies reporting multivariate analyses.
Main Outcomes and Measures
Risk of recurrent infection in patients with CDI and its association with use of gastric acid suppressant medication.
Sixteen observational studies were included, together reporting 7703 patients with CDI; among these, 1525 patients (19.8%) developed recurrent CDI. The rate of recurrent CDI in patients with gastric acid suppression was 22.1% (892 of 4038 patients) compared with 17.3% (633 of 3665) in patients without gastric acid suppression, which indicated an increased risk by meta-analysis (odds ratio [OR], 1.52; 95% CI, 1.20-1.94; P < .001). There was significant heterogeneity among the studies, with an I2 value of 64%. Subgroup analyses of studies adjusting for age and potential confounders confirmed an increased risk of recurrent CDI with use of gastric acid suppressants (OR, 1.38; 95% CI, 1.08-1.76; P = .02).
Conclusions and Relevance
Meta-analyses of observational studies suggest that patients who receive gastric acid suppressants may be at increased risk for recurrent CDI. These data should be interpreted with caution because they may be confounded owing to the observational design of the individual studies. It may be reasonable to re-evaluate the need for these medications in patients with CDI.
Clostridium difficile infection (CDI) is the most common cause of hospital-acquired diarrhea and has recently shown increasing incidence, severity, and morbidity and mortality rates.1-3 A substantial proportion of CDI cases (up to 40%) are community acquired in patients who were previously thought to be at low risk, which suggests the presence of new risk factors.4-6 Novel risk factors for CDI development include the use of gastric acid suppressants, presence of systemic comorbid conditions, C difficile carriage in water and food sources, and close contact with patients with CDI in the community. These factors show the importance of environment-to-person and direct person-to-person transmission.3,5,7
Gastric acid suppressant medications, such as proton pump inhibitors (PPIs) and H2 receptor blockers (H2Bs), are commonly prescribed or obtained as over-the-counter products for gastroesophageal reflux disease, peptic ulcer disease, or functional dyspepsia, but they are also sometimes prescribed for unnecessary indications, which leads to overuse. Studies report an increased risk of primary CDI with gastric acid suppression,8 but these data are balanced by studies that do not demonstrate such an association after controlling for important confounders, such as age and comorbid conditions.9 For instance, in 1 cohort study, PPIs and H2Bs were associated with a 2-fold increase in the risk of CDI on univariate analysis, but after adjusting for age, length of stay, comorbid conditions, history of CDI, and antibiotic exposure, this association was no longer seen.10 Meta-analyses and systematic reviews have demonstrated that patients who are exposed to PPIs are at an increased risk for primary CDI.11 Despite the variability in the literature, the US Food and Drug Administration has issued a warning that PPIs are associated with an increased risk of CDI.12
Recurrent CDI after a primary infection is a major problem, with the risk being as high as 50% to 60% after 3 or more infections. Risk factors for recurrent CDI include older age, concomitant antibiotic use, and comorbid conditions.13,14 Up to 50% of patients with CDI are using concomitant gastric acid suppressants.15 Patients with primary CDI infection are rarely re-evaluated after the CDI episode to assess the necessity of these and other medications. Data on the association between gastric acid suppression and recurrent CDI are conflicting, and there is unexplained heterogeneity among the risk estimates.10,16 In a large, retrospective cohort study, PPI use was associated with a 1.5-fold increased risk of recurrent CDI.17 In other studies, including a population-based cohort study, the use of gastric acid suppressants was not associated with recurrent CDI.8,18 Subsequent systematic reviews and meta-analyses have shown a 1.5- to 1.7-fold increase in recurrent CDI with gastric acid suppression, but these analyses are limited by the exclusion of key studies.19,20 In addition, studies that have controlled for confounders may be better able to identify a true association. We performed a comprehensive, systematic review and meta-analysis to study the association between the use of gastric acid suppressants and the risk of recurrent CDI.
All procedures used in this meta-analysis were consistent with the Meta-analysis of Observational Studies in Epidemiology criteria21 for observational studies and the Preferred Reporting Items for Systematic Reviews and Meta-analyses guidelines for randomized clinical trials.22
The studies considered in this meta-analysis were case-control studies, cohort studies, and clinical trials that included a population of patients with CDI who did or did not receive gastric acid suppressant therapy (either PPIs or H2Bs) and evaluated the occurrence of recurrent CDI, with no restrictions on study setting (inpatient or outpatient). We excluded studies that did not evaluate recurrent CDI as an outcome. Studies were also excluded from meta-analyses if there were insufficient data to determine an estimate of an odds ratio (OR) and 95% CI. Studies with published full text were included, and those only in abstract form were excluded.
We conducted a comprehensive search of Ovid MEDLINE In-Process & Other Non-Indexed Citations, Ovid MEDLINE, Ovid EMBASE, Ovid Cochrane Central Register of Controlled Trials, Ovid Cochrane Database of Systematic Reviews, Web of Science, and Scopus from January 1, 1995, to September 30, 2015. The search strategy was independently designed and conducted by study investigators (R.T. and S.K.) and the Mayo Clinic library staff. The search was limited to studies published in English. Controlled vocabulary supplemented with key words was used to search for studies of PPI and H2B use and CDI. The main key words used in the search were Clostridium difficile, C diff, C difficile, Clostridium difficile infection, CDI, Clostridium difficile–associated diarrhea or CDAD, or pseudomembranous colitis; and proton pump inhibitor, proton pump inhibitors, PPI, PPIs, proton pump, histamine 2 blocker, H2 blocker or gastric acid suppression, and acid suppressive therapy; and outcomes, recurrence, reinfection, relapse, or recurrent infection. The detailed search strategy is shown in eTable 1 in the Supplement.
Two authors (R.T. and S.K.) independently reviewed the titles and abstracts of the identified studies, and those that did not answer the research question of interest were excluded. The full texts of the remaining articles were reviewed to determine whether the inclusion criteria were fulfilled. The reference lists of articles with information on the topic were reviewed for additional pertinent studies. We manually searched the abstracts from major gastroenterology and infectious diseases conferences from 2001 to 2015. A flow diagram of included studies is shown in Figure 1.
A modified Newcastle-Ottawa scale was used to assess the methodologic quality of case-control and cohort studies by 2 of us (R.T. and S.K.).23 In this scale, case-control studies were scored across 3 categories using the following factors: selection (4 questions), comparability (2 questions), and ascertainment of the outcome of interest (4 questions), and cohort studies were scored with selection (4 questions), comparability (2 questions), and ascertainment of the outcome of interest (5 questions). For each question, 1 point was given if the study met the criterion except for comparability of study groups, in which 2 points were awarded for question 2 (the study controlled for age, sex, or both and other confounding factors) (eTable 2 in the Supplement). Studies with a cumulative score of 7 or more were considered high quality. Any discrepancies were addressed by a joint re-evaluation of the original article.
Data were independently abstracted to a predetermined collection form by 2 of us (R.T. and S.K.). Data were collected for each study, including study setting and design, year of publication, location, primary outcome reported, type of gastric acid suppressant, number of patients in each group (exposed vs not exposed and recurrent vs no recurrent CDI). Conflicts in data abstraction were resolved by consensus.
Our primary analysis focused on assessing the risk of recurrent infection in patients with CDI and its association with gastric acid suppressant use. In the included studies, recurrent CDI was defined as recurrent infection within 30, 56, 60, or 90 days after symptom resolution.
We used the random-effects model described by DerSimonian and Laird24 to calculate meta-analytic ORs and 95% CIs. We assessed heterogeneity within groups with the I2 statistic, which estimates the proportion of total variation across studies that is due to heterogeneity in study patients, design, or interventions rather than chance; I2 values greater than 50% suggest substantial heterogeneity.25 We explored potential causes of heterogeneity with stratification by clinical and methodologic features of studies. This stratification included case definition (time interval of recurrence: within 60 days vs within 90 days), type of gastric acid suppressant (PPI and H2B reported together, PPI alone, or H2B alone), and study design. The presence of publication bias was assessed by visual inspection of funnel plots.25 All P values were 2-tailed. For all tests (except heterogeneity), a probability level <.05 was considered statistically significant. Calculations were performed and graphs were constructed using RevMan, version 5.3 (Review Manager; Cochrane Inc). MetaXL, version 5.1 (EpiGear International Pty Ltd) was used to calculate the pooled effect size of studies that reported ORs and 95% CIs on multivariate analysis.26
A priori–defined sensitivity analyses included studies that had controlled for potential confounders, case-control vs cohort studies, PPIs vs PPIs and H2Bs, and definition of recurrent CDI. Additional sensitivity analyses based on study setting and diagnostic testing for CDI were performed.
The described search strategy revealed 483 potentially relevant studies; abstracts were screened and relevant full-text articles were obtained (Figure 1). In all, 36 full-text articles were reviewed, 20 of which were excluded for various reasons. A total of 16 studies were included in this meta-analysis: 15 observational studies and 1 post hoc analysis of 2 clinical trials of treatment of CDI that evaluated use of gastric acid suppressants.16-18,27-39 The post hoc analysis of clinical trials was considered a prospective cohort study because the primary aim of the trial was to assess response to CDI treatments, and the risk of recurrent CDI from use of gastric acid suppressants was a secondary analysis.39 Together, the 16 studies reported 7703 patients with CDI; 1525 (19.8%) of these patients developed recurrent CDI.
The median Newcastle-Ottawa quality score for case-control studies was 6 of 11 possible points (range, 6-8) and for cohort studies was 10 of 12 possible points (range, 8-10). The Table depicts the methodologic quality of all included studies.
The characteristics of the 16 included studies are described in the Table. The studies were performed in North America, Europe, Korea, Japan, and Israel. The study time periods range from 1991 to 2013. Most observational studies assessed medication exposure through a review of medical records.
Of the 7703 patients with CDI, 4038 (52.4%) were using gastric acid suppressants. The rate of recurrent CDI in patients receiving these drugs was 22.1% (892 of 4038) compared with 17.3% (633 of 3665) in patients without use of gastric acid suppressants. Meta-analysis of all studies using the random-effects model demonstrated an increased risk of recurrent CDI in patients receiving gastric acid suppressants (OR, 1.52; 95% CI, 1.20-1.94; P < .001) (Figure 2). There was significant heterogeneity among the studies, with an I2 value of 64%. No publication bias was seen (Figure 3).
Given the significant heterogeneity in the meta-analysis of all the included studies, we performed subgroup analyses to better understand the heterogeneity. However, no single source of heterogeneity was identified; the I2 value remained increased in all subgroup analyses.
Of 16 studies, 9 studies had performed multivariable analysis after adjusting for potential confounders, including age and other comorbid conditions (eTable 3 in the Supplement). We calculated the pooled effect size of these studies by combining reported ORs and 95% CIs. Meta-analysis of these studies revealed an increased risk of CDI with use of gastric acid suppressants (OR, 1.38; 95% CI, 1.08-1.76; P = .02) (Figure 4).
On the basis of expert opinion, CDI is considered recurrent if symptoms recur within 56 days of the initial episode and are associated with a positive stool C difficile test.40 Since a possible source of heterogeneity is variability in the definition of recurrent CDI, we separated studies that defined recurrent CDI as being within 60 days or 90 days of the initial episode. One study did not mention the follow-up period and was excluded from this subgroup analysis. These analyses revealed an increased risk of recurrent CDI for patients using gastric acid suppressants in studies that defined recurrence within 90 days (OR, 1.53; 95% CI, 1.07-2.19; P = .02) (eFigure 1A in the Supplement) and in studies that defined recurrence within 60 days (OR, 1.54; 95% CI, 1.04-2.28; P = .03) (eFigure 1B in the Supplement).
Several studies included patients using only PPIs, and others included both PPIs and H2Bs. We separated the studies with PPIs only or with PPIs and/or H2Bs (reported together). Meta-analyses revealed an increased risk of CDI recurrence with PPIs (OR, 1.66; 95% CI, 1.18-2.34; P = .004) (eFigure 2A in the Supplement) but not in studies that mentioned use of both PPIs and/or H2Bs (OR, 1.37; 95% CI, 0.95-1.99; P = .09) (eFigure 2B in the Supplement). Only 1 study reported H2B use alone.
A subgroup analysis based on the study design—separating cohort and case-control studies—revealed an increased risk of recurrent CDI for patients using gastric acid suppressants in both the case-control studies (OR, 1.91; 95% CI, 1.20-3.03; P = .006) (eFigure 3A in the Supplement) and cohort studies (OR, 1.34; 95% CI, 1.01-1.78; P = .04) (eFigure 3B in the Supplement).
Of the 16 studies, 9 included patients only from the inpatient setting, and 7 included both inpatients and outpatients. Subgroup analysis of the studies with patients from the inpatient setting only also revealed an increased risk of recurrent CDI with use of gastric acid suppressants (OR, 1.44; 95% CI, 1.05-1.97; P = .02) (eFigure 4 in the Supplement).
Subgroup analysis on the basis of the diagnostic assay used for CDI—polymerase chain reaction (PCR) vs enzyme-linked immunosorbent assay (ELISA)—revealed an increased risk of recurrent CDI among studies that used ELISA to detect recurrent CDI (OR, 2.54; 95% CI, 1.76-3.67; P < .001) but not among the studies that used PCR to detect CDI (OR, 1.21; 95% CI, 0.82-1.80; P = .33) (eFigure 5 in the Supplement).
Gastric acid suppression has been implicated in the pathogenesis of both primary and recurrent CDI owing to loss of the protective effect of gastric acid and/or perturbations in the gut microbiota. Studies demonstrating the association between gastric acid suppressants and recurrent CDI have shown conflicting results. In this systematic review and meta-analysis of 16 studies, we found that gastric acid suppressants were associated with an increased risk of recurrent CDI compared with nonuse of these medications. Subgroup analyses including only studies with multivariate analysis adjusting for age and potential confounders also revealed an increased risk of CDI recurrence. There was significant heterogeneity among the studies, most likely owing to the diverse patient populations included and methodologic differences between the studies. Subgroup analyses separating case-control and cohort studies and the type of gastric acid suppressant used (PPIs and H2Bs vs PPIs) did not identify a cause of the heterogeneity.
The pathophysiology of recurrent CDI involves a complex interplay between host and microbial factors. Established risk estimators of recurrent CDI include advanced age, prior CDI episodes, systemic antibiotic re-exposure (with different classes of antibiotics being associated with different levels of risk of recurrence), and low anti-toxin–A immunoglobulin levels.41,42 Antibiotic therapy for CDI is inactive against C difficile spores and may also lead to increased disruption of the gut microbiota, which further contributes to recurrent CDI. It is postulated that C difficile spores are resistant to degradation in the acid environment of the stomach and that gastric acid suppression may not be associated with CDI,43 although ingested vegetative forms would be sensitive to stomach acid. Gastric acid suppressants alter the distal gut microbiome, which leads to a substantial decrease in bacterial diversity and may contribute to both primary and recurrent CDI.44,45 However, this association needs to be further explored in the setting of robust clinical data that allow controlling for confounders.
Subgroup analysis of studies that had performed multivariate analysis revealed an increased risk of recurrent CDI with gastric acid suppression after controlling for potential confounders. It is possible that use of gastric acid suppressants may be greater among the elderly, patients using concomitant antibiotics, and those with comorbid conditions. All of these are independent risk factors for recurrent CDI and, hence, potential confounders. Our meta-analysis showed that, even after controlling for important confounders, gastric acid suppressants emerge as an independent risk factor for recurrent CDI. This finding indicates that the association between PPIs and recurrent CDI might be a true association.
Subgroup analysis on the basis of the definition of recurrent CDI (time of recurrence from initial episode of 90 days vs 60 days) demonstrated that studies defining CDI recurrence either way found an increased risk of recurrent CDI with gastric acid suppression. Standard treatment guidelines based on expert opinion recommend that recurrent CDI be defined as occurring within 8 weeks of an initial episode and CDI episodes after 8 weeks be classified as reinfection. It may be postulated that, after treatment of the initial infection, gastric acid suppression may also have a role in reinfection.
Subgroup analysis was performed separating case-control and cohort studies (based on different risks of bias); the risk of recurrent CDI was increased with gastric acid suppression in both types of studies. Our subgroup analysis based on the type of gastric acid suppressant showed an increased risk of recurrent CDI with PPIs but not with H2Bs and/or PPIs. Proton pump inhibitors have a greater gastric acid suppressive effect, so they may have a larger effect on the microbiome than do H2Bs.44,45 It is also plausible that concomitant use of PPIs and antibiotics might have an additive effect on CDI recurrence rate, but data were not available on antibiotic use in individual studies. Most of the included studies did not mention the dosage and duration of gastric acid suppressive therapy, so we were unable to analyze the possible effects of these factors. Of all included studies, 3 mentioned that gastric acid suppressants were used continuously,17,27,28 but similar information was not included in the other studies, which limits our ability to analyze the effect of continuous vs intermittent use of these medications. Differences in drug dosage and duration may explain some of the heterogeneity observed in the analysis.
An analysis of studies that included only inpatients also revealed increased risk of recurrent CDI with gastric acid suppressants. A higher risk of recurrent CDI was seen among studies that used ELISA to detect recurrent CDI but not among those that used PCR. This difference could be due to an oversensitivity of the PCR test.
The strengths of our study include a comprehensive literature review with a large patient population compared with previous meta-analyses, which excluded several key studies. Our results were stable on all subgroup analyses.
Our study has several limitations. The individual studies included in the meta-analysis varied in several ways, including design, patient population, definition of recurrent CDI, tests used to diagnose initial and recurrent CDI episodes, North American pulsed-field gel electrophoresis type 1 status of the primary infection, re-exposure to antibiotics, and duration of use of gastric acid suppressants. These different aspects led to substantial heterogeneity. Because the individual studies controlled for different confounding factors, we were unable to perform analyses in which all confounding factors could be accounted or controlled for, including demographics (age and sex), the number of prior CDI episodes, continuous vs intermittent use of gastric acid suppressants, duration and dose of the medications, first recurrence vs multiple recurrences, adherence to infection control practices, and the possibility that positive stool assays represented colonization rather than active infection.
Gastric acid suppressants may be associated with an increased risk of recurrent CDI. Development of CDI might require discontinuation of these medications. Unnecessary use of gastric acid suppressants in both the inpatient and outpatient settings without proper indication should be limited. Decreasing unnecessary use of gastric acid suppressants may help to reduce the rates of primary and recurrent CDI as well as health care costs. These data should be interpreted with caution because they may be confounded due to the observational design of the individual studies. Large, better-designed prospective studies controlling for confounding variables are required to understand the association between gastric acid suppression and recurrent CDI. These studies should focus on evaluating the effect of concomitant use of antibiotics with PPIs, dosage and duration of gastric acid suppressants, and CDI recurrence. A randomized clinical trial evaluating the effect of continuing the use of gastric acid suppressants vs stopping them in patients with no serious indication might be helpful in assessing the true association between gastric acid suppressants and the risk of recurrent CDI.
Accepted for Publication: January 17, 2017.
Corresponding Author: Sahil Khanna, MBBS, MS, Division of Gastroenterology and Hepatology, Mayo Clinic, 200 First St SW, Rochester, MN 55905 (firstname.lastname@example.org).
Published Online: March 27, 2017. doi:10.1001/jamainternmed.2017.0212
Author Contributions: Dr Khanna had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Tariq, Singh, Gupta, Khanna.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Tariq, Gupta, Khanna.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Tariq, Singh, Gupta, Khanna.
Administrative, technical, or material support: Tariq, Khanna.
Study supervision: Pardi, Khanna.
Conflict of Interest Disclosures: None reported.
Meeting Presentation: The study was presented at Digestive Disease Week; May 24, 2016; San Diego, California.