Effects of age and gender on body composition indices as predictors of mortality in middle-aged and old people

Abstract To determine whether body composition indices interact with age and gender as a predictor of all-cause mortality, 1200 participants at least 40 years of age were recruited in 2009 and 2010. A multi-frequency bioelectrical impedance analysis device was used to measure each participant’s body composition indices, including the fat mass index (FMI), fat free mass index (FFMI), skeletal muscle mass index (SMMI), and visceral fat area index (VFAI). A baseline questionnaire was used to collect demographic information about lifestyle habits, socioeconomic status, and medical conditions. All claimed records of death from 2009 to 2018 in the National Health Insurance Research Databank were identified. The all-cause mortality rate was 8.67% after a mean follow-up period of 5.86 ± 2.39 person-years. The Cox proportional hazard model analysis showed significantly negative associations between FFMI or SMMI with all-cause mortality in the total group and those aged $$\geq$$ ≥ 65 y/o. The FFMI and SMMI were negative predictors of mortality in both genders. The FMI and VFAI were positive predictors of mortality exclusively in females. In conclusion, the SMMI is a better predictor of mortality than the BMI, FMI, and FFMI, especially in older adults. A higher fat mass or visceral fat distribution may predict higher mortality in females.