Abstract: SA-PO623
Cause Specific Mortality Attributable to Proton Pump Inhibitor
Session Information
- Pharmacology
October 27, 2018 | Location: Exhibit Hall, San Diego Convention Center
Abstract Time: 10:00 AM - 12:00 PM
Category: Pharmacology (PharmacoKinetics, -Dynamics, -Genomics)
- 1700 Pharmacology (PharmacoKinetics, -Dynamics, -Genomics)
Authors
- Xie, Yan, VA Saint Louis Health Care System, Saint Louis, Missouri, United States
- Bowe, Benjamin Charles, Research and Development Service, Veterans Affairs St Louis Health Care System, St. Louis, Missouri, United States
- Li, Tingting, Washington University in St. Louis, Saint Louis, Missouri, United States
- Yan, Yan, Washington University Medical School, Saint Louis, Missouri, United States
- Xian, Hong, Saint Louis University College for Public Health & Social Justice, St. Louis, Missouri, United States
- Al-Aly, Ziyad, VA Saint Louis Health Care System, Saint Louis, Missouri, United States
Background
Proton pump inhibitors (PPI) are widely used; and their use is associated with increased risk of adverse events and death. However, a detailed analysis of the cause of the death distribution among users of PPI is not available. In this work, we aimed to characterize the cause of death among users of PPI.
Methods
A cohort of 325,307 new users of PPI or H2 blockers was built. Cause of death data from National Death Index was used to examine the association between PPI and cause specific morality. Multiple survival models including proportional hazards models, accelerated failure time models, and additive hazard models were used. we used an instrumental variable approach based on clinicians’ prescribing behavior and high-dimensional propensity score in domains including diagnoses, procedures, medications and laboratory results to account for measured and unmeasured confounders, respectively. For each outcome, model-covariate combination with smallest Brier score was selected and the survival possibility for the cohort was calculated.
Results
During 10 years of follow up, PPI use was significantly associated with all-cause mortality, and mortality from circulatory, and genitourinary causes. Compared to users of H2 blockers, event rate per 1000 persons attributable to PPI use was 5.44 (17.57, 33.43), 9.91 (3.94, 15.79) and 5.17 (2.83, 7.51) for all-cause, circulatory and genitourinary mortality, respectively. The excess death accounted for 6.94% (4.85, 9.03), 7.15% (2.90, 11.17) and 32.58% (19.34, 43.56) of all the all-cause, circulatory, and genitourinary mortality in PPI users. Further analyses within the circulatory and genitourinary causes showed PPI use was associated with 8.07 (3.29, 12.77) and 3.66 (1.95, 5.46) excess death per 1000 persons from coronary artery disease and chronic kidney disease, which accounted for 10.22% (4.27, 15.72) and 35.97% (21.07, 48.23) of the related deaths. Death from acute kidney injury was not associated with PPI use (excess death 0.61 (-0.27, 1.51) and population attributable fraction 29.32% (-15.42, 56.11)).
Conclusion
In this work, we used advanced statistical methodologies to map the cause of death distribution among users of PPI. Our findings suggest that PPI use is associated with increased mortality from circulatory and genitourinary causes, but not other causes of death. PPI should only be used when medically indicated.
Funding
- Veterans Affairs Support