Are Proton Pump Inhibitors Associated With the Development of Community-Acquired Pneumonia?

A Meta-analysis

Christopher Giuliano; Sheila M Wilhelm; Pramodini B Kale-Pradhan


Expert Rev Clin Pharmacol. 2012;5(3):337-344. 

In This Article


We identified 817 studies initially, of which 724 studies were not relevant to the current analysis. A total of 93 abstracts were reviewed and 47 articles were excluded. Of the 46 remaining articles, 12 were included for full review. Three articles were excluded in the primary analysis owing to the inability to separate PPI and H-2 Receptor Antagonists (H2RA) data,[24] reporting of hospital-acquired pneumonia[25] and inappropriate methodology.[26] Two retrospective cohort analyses[27,28] and seven case–control studies[29–35] were included for the final analyses for the primary outcome. One study did not have sufficient information to be included in the primary outcome analysis but was included in a subgroup analysis.[24] The authors agreed on all included studies. Figure 1 illustrates the search strategy related to primary outcome.

Figure 1.

Systematic literature search for studies evaluating proton pump inhibitor use and association of community-acquired pneumonia. H2RA: Histamine 2 receptor antagonist; ICU: Intensive care unit; PPI: Proton pump inhibitor.

Nine studies with 120,863 pneumonia cases from 1987 to 2006 were included in the meta-analysis. PPI usage in pneumonia cases ranged from 4.3 to 47%. NOQAS scores ranged from 4 to 8 out of a maximum of 9. Six of the studies were population-based studies in various countries including the UK, The Netherlands, Canada and Denmark.[28,30,31,33–35] Two studies used the same UK patient database but evaluated different populations.[33,34] Other studies included a single-center hospital-based study, a stroke rehabilitation center database study, and an Australian Veteran Affairs study.[27,29,32] All studies made statistical adjustments for confounding variables for the development of pneumonia, although studies did not all evaluate the same confounding variables. Many of the studies did not evaluate gastroesophageal reflux disease, alcohol use and smoking as potential confounders. One study only controlled for dysphagia, tracheostomy and feeding tube.[32] Acid-suppressive agents used varied among the studies. Three studies evaluated only PPI use,[27,30,35] whereas other included studies evaluated both H2RA and PPI use.[28,29,31–34] Table 1 summarizes the study and patient characteristics of each study.

Statistical heterogeneity was present in this analysis (I2 = 97.8%); therefore, a random-effects model was used. Our primary analysis showed an increased risk of developing pneumonia with current use of a PPI (odds ratio [OR]: 1.39; 95% CI: 1.09–1.76). A preplanned sensitivity analysis showed no change in the association of PPI with the development of CAP. When excluding the study that had a NOQAS score <6, the risk remained significant (OR: 1.37; 95% CI: 1.05–1.78; I2 = 98.1%).[33] Also, after excluding studies that evaluated a specific patient population (elective surgery or stroke), risk remained significant (OR: 1.45; 95% CI: 1.13–1.89; I2 = 97.2%).[28,32] These results are displayed in Figure 2. Prespecified subgroup analysis included PPI dose and PPI duration. PPI use less than 30 days (OR: 1.65; 95% CI: 1.25–2.19), high PPI dose (OR: 1.50; 95% CI: 1.33–1.68) and low PPI dose (OR: 1.17; 95% CI: 1.11–1.24) were significantly associated with CAP. There was no association between CAP and PPI use for more than 180 days (OR: 1.10; 95% CI: 1.00–1.21). Heterogeneity decreased in the PPI dose subgroups, although the test for interaction was not significant (p = 0.15). Heterogeneity remained high in the short duration subgroup but decreased in the longer duration subgroup. Test for interaction was significant between the treatment duration subgroups (p < 0.005). These results are displayed in Figure 2.

Figure 2.

Proton pump inhibitor use and association of community-acquired pneumonia. CAP: Community-acquired pneumonia; PPI: Proton pump inhibitor.