Do single or clustered body composition parameters predict mortality? A multicenter Italian study

Studies on age-related changes in body composition often included ethnically different subjects, or few men or women with limited age range.

We retrospectively studied 1860 (1304 female, 556 male) free-living subjects, 20-to-98 years-old, including 324 with type 2 diabetes (T2DM). We measured absolute and total body mass(TBM)-adjusted fat mass (FM), lean mass (LM), appendicular lean mass (ALM), bone mineral content (BMC), TBM, skeletal muscle index (SMI) by DXA. From a subset of 20-50 years-old non-obese subjects, we defined sex-specific cutoffs for high-FM (mean +2SD), and low-LM, -ALM, -BMC, -SMI (mean -2SD), to investigate their impact on mortality. Women had higher FM, and lower absolute and TBM-adjusted LM, ALM, BMC. and SMI than men. Absolute and TBM-adjusted FM positively correlated with age, whereas all the other parameters decreased with age, with negative correlations being steeper in men. The regressions of TBM-adjusted values differed between subjects with or without T2DM. T2DM negatively interacted with the correlations of FM and FM/TBM with age, and positively with that between LM/TBM and age. TBM-adjusted parameters shifting from the reference distribution influenced mortality, with the respective HR being: 1.65 (95%CI:1.31-2.07; p < 0.0001) for FM/TBM; 1.58 (95%CI:1.25-2.00; p < 0.0001) for LM/TBM; 1.53 (95%CI:1.23-1.90; p = 0.00012) for BMC/TBM; 1.05 (95%CI:0.83-1.34; p = 0.68) for ALM/TBM; 1.77 (95%CI:1.37-2.28; p < 0.0001) for SMI; 1.66 (95%CI:1.30-2.11; p < 0.0001) for low-ALM + low-BMC; 1.47 (95%CI:1.12-1.93; p = 0.0056) for high-FM + low-ALM; 1.51 (95%CI:1.18-1.94; p = 0.0009) for high-FM + low-BMC.

Age-related body composition changes differ between sexes. Several individual or clustered parameters shifting from the distribution of younger subjects affect mortality.