Frailty Trajectories in Men Aged 71 Years or Older Living in Western Australia: Results From the Health in Men Study

OBJECTIVES: To identify trajectories of frailty as older men age and to assess the associations between sociodemographic, lifestyle, and health factors with those trajectories. DESIGN: Community-based cohort study. SETTING AND PARTICIPANTS: Data from 1319 men aged 71 years or older, residing in Western Australia and participating in the Health in Men Study. METHODS: Frailty was assessed on at least 2 occasions between 2001 and 2019 using the five-point FRAIL scale that measures fatigue, resistance, ambulation, illnesses, and weight loss. Latent frailty trajectories over age were identified using a group-based trajectory model with an extension to allow for participant attrition because of death. Multinomial logistic regression was used to assess differences in sociodemographic, lifestyle, and health factors between the frailty trajectory groups. Sensitivity analyses were performed to assess the impact of including/excluding participants who dropped out. RESULTS: During follow-up, 999 (66.9%) participants died. Four distinct frailty trajectories were identified: prefrail to frail (11.7%), steady decline (39.2%), slower decline (37.2%), and sustained nonfrailty (11.9%). Smokers, those with less than 150 minutes of physical activity per week, or those having a diagnosis of cardiovascular disease or diabetes by age 70, were all more likely to be in the prefrail to frail trajectory than the sustained nonfrailty group. Sensitivity analyses showed similar results when those who were lost to follow-up were included in the analysis; however, when limiting the data set to only those who continued to participate throughout the study period, the most severe frailty trajectory was no longer detected. CONCLUSIONS AND IMPLICATIONS: We identified heterogeneity in frailty progression as men age. Frailty interventions may need to focus on lifestyle modifications and addressing or preventing underlying medical conditions prior to the age of 70. Adjustment for nonrandom attrition of participants is required to determine the long-term impact of frailty.