All-Cause and Cause-Specific Mortality in Patients With Granulomatosis With Polyangiitis: A Population-Based Study

Objective To investigate all-cause and cause-specific mortality in patients with newly diagnosed granulomatosis with polyangiitis (GPA) between 2 calendar time periods, 1997-2004 and 2005-2012. Methods Using an administrative health database, we compared all patients with incident GPA with non-GPA controls matched for sex, age, and time of entry into the study. The study cohorts were divided into 2 subgroups based on the year of diagnosis ("early cohort [1997-2004] and "late cohort" [2005-2012]). The outcome was death (all-cause, cardiovascular disease [CVD]-related cancer-related, renal disease-related, and infection-related) during the follow-up period. Hazard ratios (HR) were estimated using Cox proportional hazards models, first adjusted for age, sex, and time of entry and then adjusted for selected covariates based on a purposeful selection algorithm. Results Three hundred seventy patients with GPA and 3,700 non-GPA controls were included in this study, contributing 1,624.8 and 1,8671.3 person-years of follow-up, respectively. Sixty-eight deaths occurred in the GPA cohort, and 310 deaths occurred in the non-GPA cohort. Overall, the age-, sex-, and entry time-adjusted all-cause mortality HR in the GPA cohort was 3.12 (95% confidence interval CI 2.35-4.14). There was excess mortality due to CVD-related causes, but not cancer, in the GPA cohort. Reports of death due to infection or renal disease was not permitted, because the numbers of death were insufficient (<6 deaths for each outcome). All-cause mortality significantly improved between the early cohort and late cohort time periods (HR 5.61 and 2.33, respectively; P for interaction = 0.017). Conclusion This population-based study showed increased all-cause and CVD-related mortality risks in patients with GPA. There was significant improvement in the all-cause mortality risk over time, but the risk remained increased compared with that in the general population.