PPI and Pneumonia

Most of us have been guilty of prescribing proton pump inhibitors (PPIs) to our hospitalized patients without prudently assessing their need and often disregarding the actual indications for their use. Perhaps the article titled “Acid-Suppressive Medication Use and the Risk for Hospital-Acquired Pneumonia,” published in this week’s JAMA, will motivate us to reconsider prescribing PPIs without giving due consideration to potential adverse events.
According to the study, an estimated 40-70% of medical inpatients are treated with either a PPI or H2-receptor antagonist during hospitalization; approximately 50% of the patients are newly started on these medications, and nearly 50% of these patients are discharged on a PPI. While several outpatient studies suggest an increased risk of community-acquired pneumonia in patients receiving acid-suppressive medications, no large-scale studies exist to determine the association between acid-suppressive medications and hospital-acquired pneumonia. Herzig, et al, conducted a large, prospective cohort study from January 2004 through December 2007, including patients over the age of 18 (median age 54) admitted to the hospital for at least 3 days. Excluded from this study were all patients who spent any time in the ICU. The final cohort consisted of 63,878 admissions, of which 52% received either a PPI (83%) or H2-receptor antagonist (23%) during hospitalization. The primary outcome of hospital-acquired pneumonia occurred in 2219 admissions (3.5%). In the group of patients receiving acid-suppressive medications, there was a higher unadjusted incidence of hospital-acquired pneumonia compared to the unexposed group: 4.9% vs 2.0%, OR 2.6 (95% CI, 2.3-2.8). Additionally, there was a significant association for both aspiration pneumonia and non-aspiration pneumonia, with the diagnoses based on ICD-9 codes used at the time of hospitalization. After adjusting for potential confounders, the adjusted OR for hospital-acquired pneumonia in the group receiving acid-suppressive therapy was 1.3 (95% CI, 1.1-1.4). Interestingly, after adjustment, the association was significant only for PPI’s and not H2-receptor antagonists. Overall, the use of acid-suppressive medications, in particular PPIs, was associated with a 30% increased odds of developing hospital-acquired pneumonia. Given the estimated mortality rate of 18% for hospital-acquired pneumonia, these results are somewhat alarming. Thus, this study clearly emphasizes the need to closely examine our patient’s medication regimen at the time of hospitalization and determine if a PPI is truly warranted.