Does the frailty phenotype at the age of 66 to 71 predict death? A 14-year survival analysis of the Lc65+ study.

Background: Frailty is a health characteristic resulting from the loss of physiological reserve of multiple organs, leading to exposure to adverse outcomes, and is possibly reversible in its earliest stages. It is identified by a specific phenotype that contributes to the practice of geriatric medicine, where it is considered a potential target for preventive action. This phenotype has recently attracted interest in other medical specialties for risk assessment before stressful interventions in older adults. Whereas frailty is unusual in sexagenarians, pre-frailty is common. This longitudinal study aimed to evaluate the significance of fulfilling at least one criterion of the frailty phenotype in the late sixties as a predictor of short- and long-term mortality in males and females.

Methods: Data came from the first sample of the Lc65+ cohort, representative of the community-dwelling Lausanne population born between 1934 and 1939 (n = 1315). After baseline assessment of the five criteria of Fried's frailty phenotype (shrinking, exhaustion, muscular weakness, motor slowness and low physical activity) in 2005 (age 66-71 years), deaths were recorded over 14 years. We separated individuals into non-frail (fulfilling 0 criterion) and (pre-)frail (1+ criteria). The relationship between the phenotype and mortality was investigated graphically using Kaplan-Meier survival curves and quantified in Cox models. Multivariable analyses incrementally controlled age, socioeconomic and health characteristics. The prediction of fully adjusted models was evaluated using the Harrell's C index.

Results: Overall, 401 persons (30.5%) were (pre-)frail at baseline. A quarter of the 1315 participants died over 14 years (n = 336, 25.6%). The mortality rate was significantly higher in males in the (pre-)frail subgroup only. Survival curves showed a significant effect of (pre-)frailty on the risk of dying for both sexes. The effect of (pre-)frailty on mortality was stronger during the first 4 years of the follow-up. In males, it was significant both in short (0-4 years) and longer (>4-14 years) terms. In females, it was significant in the short term only. In all models, the estimated effect was stronger in males. The fully adjusted model was fairly predictive of death in the short term both in males (Harrell's C 0.79) and females (0.75).

Conclusions: The significantly higher mortality of individuals presenting 1+ frailty criteria supports the appropriateness of a systematic assessment of the frailty phenotype at the age of 66-71 years. In both females and males, early identification of pre-frailty has the potential to limit or reverse the development of frailty and extend lifespan through adequate individual management.