Author + information
- Selcuk Adabaga,b,
- Tien N. Voa,b,
- Brent C. Taylora,b,
- Lisa Langsetmoa,b,
- John T. Schousboea,b,
- Allyson Katsa,b,
- Peggy Cawthona,b,
- Katie L. Stonea,b,
- Marcia L. Stefanicka,b,
- James Shikanya,b and
- Thuy-Tien L. Dama,b
Background: Frailty may be an important factor to consider in clinical decisions involving elderly patients. Although frailty is associated with cardiovascular disease (CVD) mortality, older patients also have a high competing risk of non-CVD mortality. We hypothesized that failure to account for the competing risk of non-CVD death among frail, older men would result in overestimation of their absolute probability and adjusted risk of CVD death.
Methods: Among 3,135 community-dwelling older men participating in the prospective MrOS sleep study, frailty status was assessed at baseline using the validated Cardiovascular Health Study (CHS) index and categorized as frail, intermediate, or robust. Men were contacted every 4 months to ascertain vital status; deaths were centrally adjudicated using death certificates and medical records and categorized as due to CVD or non-CVD cause. Absolute probability of CVD mortality during follow-up was estimated by traditional Kaplan-Meier (KM) method and the cumulative incidence function with non-CVD mortality as a competing risk. Risk of CVD mortality associated with frailty status was calculated using traditional Cox proportional hazards models and Fine-Gray subdistribution hazards models with non-CVD mortality as a competing risk.
Results: Mean age was 76.4 years and 475 (15%) of the men were categorized as frail. During an average follow-up of 9.2 years, 1,410 men died (419 CVD deaths, 808 non-CVD deaths and 183 uncertain cause). Cumulative absolute probability of CVD mortality was 13% (95% CI 10%-17%) at 5 years and 32% (95% CI 27%-37%) at 10 years using traditional KM method vs. 12% (95% CI 9%-15%) at 5 years and 24% (95% CI 20%-28%) at 10 years using the competing risk approach. Compared to robust men, the adjusted risk of CVD mortality among frail men was 2.72-fold higher (95% CI 1.99-3.70) using the traditional Cox model vs. 2.01-fold higher (95% CI 1.45-2.78) using the using the competing risk model.
Conclusions: Frailty is associated with a higher risk of CVD death among older men. However, ignoring the competing risk of non-CVD death among men with frailty overestimates their long-term probability and adjusted risk of CVD death.
Poster Hall, Hall C
Saturday, March 18, 2017, 9:45 a.m.-10:30 a.m.
Session Title: Updates on Risk Factors for Cardiovascular Disease
Abstract Category: 32. Prevention: Clinical
Presentation Number: 1187-056
- 2017 American College of Cardiology Foundation