Abstract 17951: The Effect of Proton Pump Inhibitors on Acute Coronary Syndrome Risk in Patients Receiving Clopidogrel after Coronary Drug Eluting Stent Placement in 10,677 Members of Kaiser Permanente Northern California

Abstract

The primary aim of this study was to quantify associations between the use of proton-pump inhibitors (PPI) and H2 receptor antagonists (H2RA) and the risk of acute coronary syndrome (ACS) after coronary drug-eluting stent (DES) placement. The secondary aims were to describe the mechanisms causing ACS and to explore the effects of a system-wide pharmacy warning discouraging PPI use after DES.

Methods: This was a retrospective cohort study using database searches for clinical (ICD-9 codes) and pharmacy data. The inclusion criteria were DES between 2003 and 2009 and a > 90 day supply of clopidogrel during the year after DES. The exposure variable was a > 90 day prescription for PPI versus H2RA versus neither drug. All outcomes were confirmed by chart review and included ACS occurring within 1 year after DES (primary) and gastrointestinal bleeding (GIB). The causes of ACS were described after review of angiographic data. Cox proportional hazards analysis was performed to quantify the association between PPI use and ACS after adjustment for 19 covariates. We also analyzed the rates of PPI prescription and the outcomes before and after 6/1/2009, when a warning was established which discouraged co-prescription of PPI with clopidogrel.

Results: The study included 10,677 DES recipients; 23% received PPI and 13% received H2RA. ACS (N=230) was more common in the PPI (3.2%) than the H2RA (2.9%) and unexposed groups (1.7%). Stent thrombosis caused only 17% of the ACS events and did not differ according to PPI status. PPI use was independently associated with ACS (HR 1.37, CI 1.03-1.84, P=0.03). In the 1,199 patients who received a DES after the pharmacy warning, we observed a reduction in PPI use (25% to 8%, chi-square P<0.001), a rise in H2RA use (12% to 27%, P<0.001), and non-significant reductions in both ACS (2.2% to 1.5%, chi-square P=0.098) and GIB (1.2% to 0.8%, P=0.16).

Conclusion: In this large observational study, the use of PPI after coronary DES was independently associated with an increased risk of ACS, though only a minority of the events were due to stent thrombosis. A system-wide pharmacy warning discouraging PPI use resulted in a marked reduction in PPI prescription and a non-significant reduction in ACS, a result which requires confirmation in a larger population.