Metabolic syndrome and mortality in the elderly: A time-dependent association

Abstract

Aims: To evaluate the association between metabolic syndrome (MetS) components and mortality over time. Methods: 3086 residents aged ≥49 years were followed in the Blue Mountains Eye Study, Australia. MetS components as defined by the International Diabetes Federation criteria were measured at baseline (1992-1994), 5-year (1997-1999) and 10-year (2002-2004). Using Cox proportional hazards and competing risks models with MetS as a time-dependent covariate, we estimated the effects of MetS on all-cause and cause-specific mortality. Time-dependent receiver-operating-characteristic curve, integrated-discrimination-improvement and net-reclassification-improvement tests assessed predicting abilities of individual and combined MetS components. Results: Effect of MetS on mortality increased with time: all cause: 2-year: adjusted hazard ratio 0.96 [95% confidence interval 0.69-1.34]; 5-year: 1.06 [0.84-1.32]; 10-year: 1.23 [1.01-1.51]; and CHD: 2-year: 0.46 [0.20-1.03]; 5-year: 0.70 [0.41-1.21]; 10-year: 1.62 [1.02-2.59]. Conversely, MetS was associated with an increased risk of cancer death at 2-year only: 1.62 [1.01-2.62]; but not 5-year: 1.30 [0.94-1.81] or 10-year: 0.90 [0.57-1.44]. The discrimination analyses showed that different MetS components were associated with different causes of death. Conclusions: The long-term effect of MetS on all-cause and CHD mortality in an older population was detected using time-dependent models while simulating the real scenarios of MetS changes over time. © 2012 Elsevier Ireland Ltd.