Proton pump inhibitor discontinuation in long-term care

Abstract

Objectives: To determine factors associated with proton pump inhibitor (PPI) discontinuation in long-term care.

Design: Retrospective cohort analysis.

Setting: Veterans Affairs (VA) long-term care facilities.

Participants: Veterans admitted for nonhospice care in 2005 with a length of stay of 7 days or more who were prescribed a PPI within 7 days of admission (N = 10,371).

Measurements: Prescribed medications and comorbidities were determined from VA pharmacy and administrative databases and functional status from Minimum Data Set records. Associations between participant characteristics and PPI discontinuation were determined using Cox proportional hazard ratios (HRs), censoring at death, discharge, or 180 days after admission.

Results: Participants were predominantly male (97%) and had a median age of 73 (interquartile range 60-81). There were 2,749 (27%) PPI discontinuations; 43% of these occurred within 28 days of admission. Hospitalizations (HR = 1.22, 95% confidence interval (CI) = 1.01-1.46), preadmission PPI use (HR = 1.35, 95% CI = 1.16-1.56), and lowest functional status (HR = 1.22, 95% CI = 1.03-1.45) were associated with early PPI discontinuation in adjusted models. Participants with gastric acid-related disease (HR = 0.53, 95% CI 0.46-0.61), diabetes mellitus (HR = 0.82, 95% CI 0.72-0.94), and those who were prescribed six or more medications (6-7 medications, HR = 0.78, 95% CI = 0.66-0.92; 8-10 medications, HR = 0.64, 95% CI = 0.54-0.76; ≥ 11 medications 0.51, 95% CI = 0.42-0.62) were less likely to have early discontinuation. No PPI discontinuer had PPIs resumed during the study, and few (9%) had histamine-2 receptor antagonist substitutions.

Conclusion: Although there may be clinical uncertainty regarding PPI discontinuation, more than one-quarter of participants prescribed a PPI upon admission to long-term care had it discontinued within 180 days. Targeting individuals prescribed PPIs for medication appropriateness review may reduce prescribing of potentially nonindicated medications.