The association between frailty, the metabolic syndrome, and mortality over the lifespan

Abstract

Frailty and the metabolic syndrome are each associated with poor outcomes, but in very old people (90+ years) only frailty was associated with an increased mortality risk. We investigated the relationship between frailty, metabolic syndrome, and mortality risk, in younger (20-65 years) and older (65+ years) people. This is a secondary analysis of the US National Health and Nutrition Examination Survey (NHANES) datasets for 2003-2004 and 2005-2006, linked with mortality data up to 2011. The metabolic syndrome was defined using the International Diabetes Federation criteria. Frailty was operationalized using a 41-item frailty index (FI). Compared to the younger group (n = 6403), older adults (n = 2152) had both a higher FI (0.10 ± 0.00 vs. 0.22 ± 0.00, p < 0.001) and a greater prevalence of the metabolic syndrome (24.1 vs. 45.5%, p < 0.001). The metabolic syndrome and FI were correlated in younger people (r = 0.25, p < 0.001) but not in older people (r = 0.08, p < 0.1). In bivariate analyses, the FI predicted mortality risk in both age groups whereas the metabolic syndrome did so only in the younger group. In Cox models, adjusted for age, sex, race, education, and each other, the FI was associated with increased mortality risk at both ages (younger HR 1.05 (1.04-1.06); older HR 1.04 (1.03-1.04) whereas the metabolic syndrome did not contribute to mortality risk. The FI better predicted mortality than did the metabolic syndrome, regardless of age.

Keywords: Deficit accumulation; Frailty in older people; Frailty index; Metabolic syndrome; NHANES.

Figures

Fig. 1

Kaplan-Meier curves for 10-year survival probability stratified by frailty index level for the whole sample (a), younger (under 65 years, b), and older (65 years and older, c) groups. Log-rank analysis showed a significant difference between the curves for each age group (p < 0.001)

Fig. 2

Kaplan-Meier curves for 10-year survival probability stratified by metabolic syndrome status for the whole sample (a), younger (under 65 years, b), and older (65 years and older, c) groups. Log rank analysis showed a significant difference between the curves for the whole sample and the younger group (p < 0.001), but not the older group (p = 0.48)