We identified a total of 2377 articles in the initial search for observational studies, and we reviewed 60 abstracts and 18 full articles. We included 8 of these articles in our analysis. We identified 8513 randomized controlled trials, and we reviewed 914 abstracts and 35 full articles. We included 23 of these articles and 2 bibliographies of relevant articles in the study. In summary, we included five case–control studies,6,7,14,22,23
three cohort studies,2,10,24
and 23 randomized controlled trials25–47
in the final analysis ().
Figure 1: Selection of observational (case–control and cohort) studies (A) and randomized controlled trials (B) evaluating the risk of pneumonia in association with use of acid-suppressive drugs. *Includes nested case–control studies. H2RA = histamine (more ...)
and summarize the general characteristics of the 31 studies that were included in the analysis.2,6,7,10,14,22–47
The mean quality scores were 8.4 for the observational studies (maximum score 9) and 3.1 for the randomized controlled trials (maximum score 5).
Characteristics of case–control and cohort studies included in the final analysis of acid-suppressive drugs* and risk of pneumonia
Characteristics of randomized controlled trials (RCTs) included in the final analysis
Description of studies
The selected studies were published between 1985 and 2009. Five articles reported population-based studies,2,6,7,14,23
and 26 articles, including the 23 randomized controlled trials, reported hospital-based studies.10,22,24–47
Of the observational studies, five evaluated the association between use of acid-suppressive drugs and risk of community-acquired pneumonia,2,6,7,14,23
and three evaluated the association between use of these drugs and risk of hospital-acquired pneumonia.10,22,24
Main pooled analyses and heterogeneity
Meta-analyses for observational studies with the two types of acid-suppressive drug showed significant positive associations between use of proton pump inhibitors and risk of pneumonia (adjusted OR 1.27, 95% CI 1.11–1.46, I2 90.5%) and between use of histamine2 receptor antagonists and risk of pneumonia (adjusted OR 1.22, 95% CI 1.09–1.36, I2 0.0%) ().
Figure 2: Meta-analyses of observational studies evaluating the risk of pneumonia among patients receiving acid-suppressive drugs, based on random-effects model. Adjusted odds ratios (ORs) greater than 1 indicate increased risk of pneumonia. CI = confidence interval, (more ...)
Meta-analysis of the randomized controlled trials examining risk of hospital-acquired pneumonia in association with use of histamine2 receptor antagonists confirmed the findings of the observational studies (relative risk 1.22, 95% CI 1.01–1.48, I2 30.6%) ().
Figure 3: Meta-analysis of randomized controlled trials evaluating the risk of hospital-acquired pneumonia among patients using histamine2 receptor antagonists, based on random-effects model. Relative risk (RR) values greater than 1 indicate increased risk of pneumonia. (more ...)
In subgroup analyses by type of pneumonia, we observed a significant positive association between use of proton pump inhibitors and community-acquired pneumonia (adjusted OR 1.34, 95% CI 1.14–1.57, I2 93.6%) and between use of histamine2 receptor antagonists and hospital-acquired pneumonia (adjusted OR 1.24, 95% CI 1.05–1.47, I2 0.0%) ().
Subgroup analyses for use of acid-suppressive agents and risk of pneumonia using random-effects model for observational studies
Subgroup analyses by dose indicated a dose–response relationship. A higher dose of proton pump inhibitors was more strongly associated with pneumonia (adjusted OR 1.52, 95% CI 1.31–1.76, I2 27.5%) than the usual dose (adjusted OR 1.37, 95% CI 1.08–1.74, I2 86.5).
Subgroup analyses by duration of exposure showed that the strength of the association between use of proton pump inhibitors and risk of pneumonia decreased with longer duration of therapy before the index date (date of diagnosis of pneumonia). There were significant positive associations between risk of pneumonia and use of proton pump inhibitors within 7 days before the index date (adjusted OR 3.95, 95% CI 2.86–5.45, I2 0.0%), within 30 days before the index date (adjusted OR 1.61, 95% CI 1.46–1.78, I2 30.6%) and from 30 to 180 days before the index date (adjusted OR 1.36, 95% CI 1.05–1.78, I2 84.3%). The risk of pneumonia was greater with the use of histamine2 receptor antagonists within 7 days before the index date (adjusted OR 5.21, 95% CI 4.00–6.80, I2 not available). The risk also appeared greater with the use of these drugs within 30 days before the index date (adjusted OR 1.49, 95% CI 0.82–2.72, I2 80.4%) and from 30 to 180 days (adjusted OR 1.21, 95% CI 0.94–1.56, I2 27.6%), but these associations were not statistically significant.
Subgroup analyses of the 23 randomized controlled trials by comparators showed a significant positive association between use of histamine2 receptor antagonists and risk of pneumonia in studies that employed sucralfate as a control (relative risk 1.33, 95% CI 1.04–1.69, I2 24.7%). Placebo-controlled studies also indicated an overall increase in the risk of pneumonia with these drugs, but the result was not statistically significant (relative risk 1.09, 95% CI 0.80–1.48, I2 37.9%).
We conducted subgroup meta-analyses of the observational studies and randomized controlled trials by methodologic quality. Among the observational studies, we observed a significant positive association for both high-quality studies (adjusted OR 1.29, 95% CI 1.17–1.42, I2 0.0%) and low-quality studies (adjusted OR 1.15, 95% CI 1.00–1.32, I2 82.1%). Among the randomized controlled trials, the risk of pneumonia appeared greater in low-quality studies (relative risk 1.35, 95% CI 1.10–1.67, I2 12.5%), whereas there was no effect among the high-quality studies (relative risk 0.96, 95% CI 0.65–1.43, I2 47.0%).