Abstract
OBJECTIVE: Previous studies have shown that relative handgrip strength, handgrip measure divided by body mass index (BMI), affects the future onset of diabetes and prediabetes. However, fat free mass (FFM) has been suggested to adjust for this effect better than BMI. In this study, we examined applicability of models that adjusted handgrip-diabetes relationship with either BMI or FFM. RESULTS: Of 1940 participants (56.2% male, average (SD) age, 57.2 [11.2] years), 267 (13.8%) had diabetes (DM) and 912 (47.0%) had prediabetes (pre-DM). The average handgrip measure for men was 40.0 kg (tertile measures, 37.4 kg and 42.5 kg) and for women 24.2 kg (tertile measures, 22.6 kg and 25.7 kg). Among both sexes, the percentage of people unaffected by DM or pre-DM was highest in the strong handgrip group and lowest in the weak handgrip group. Analysis using binary logistic models showed that an increase in handgrip measure was associated with a decrease in the chance of having either pre-DM or DM. This effect was detected by both BMI models and FFM models, even after adjustment for medical and lifestyle factors. Either or both should be used depending on the research aims, setting and methods.