Abstract 16049: Trends in Cause of Long-Term Death After Percutaneous Coronary Intervention
Background: The impact of changing demographics upon causes of long-term death after percutaneous coronary intervention (PCI) remains incompletely defined.
Methods: We evaluated secular trends in cause-specific long-term mortality after index PCI performed at a single center between 1991 and 2008. Deaths were ascertained by prospective surveillance. Final date of follow-up was 12/31/12. Cause of death was determined via telephone interviews and review of records, autopsy reports and death certificates. Competing-risks analysis of causes of long-term death was performed across three eras of PCI (1991-96, 1997-02, 2003-08). A separate adjusted analysis was performed by strata-weighting for age, gender, cardiac and non-cardiac variables and single vs multivessel disease across eras.
Results: 19,077 patients survived PCI-hospitalization, of whom 6,988 subsequently died (37%, 4.48 per 100 person-yrs). Cause of death was determined in 6,857 (98.1%). Across three time-periods there was a 33% decline in cardiac deaths at 5 yrs after PCI (incidence 9.8%, 7.4%, 6.6%) but a 57% increase in non-cardiac deaths (7.1%, 8.5%, 11.2%). A population study of local residents undergoing PCI exhibited near identical temporal changes. Similar trends were observed in every age group, in single and multivessel disease and whether PCI was performed for stable or acute coronary syndromes. Strata-weighted comparisons across three eras indicated a 50% decline in the incidence of cardiac deaths at 5yrs (8.2%, 5.5%, 4.1%) whereas the incidence of non-cardiac deaths remained relatively stable (7.5%, 7.1%, 8.1%).
Conclusions: This single-center study found a marked temporal switch from predominantly cardiac to predominantly non-cardiac causes of long-term death after PCI. Generalizability of the findings is supported by similar trends occurring in a local resident population, as well as in multiple subgroups undergoing PCI. Weighted adjustment suggests the temporal decline in cardiac mortality is independent of changes in measured clinical characteristics, whereas the rise in non-cardiac mortality may be related to an increased prevalence of non-cardiac disease at baseline. These findings have implications for care of patients with established coronary disease.
- © 2013 by American Heart Association, Inc.