Implication of grip strength assessment for the management of body weight in disability prevention in older adults.

Background: Although body mass index (BMI) and grip strength (GS) are both predictors of disability, their joint effect on predicting incident disability remains uncertain. We examined whether the effect of BMI on incident disability can be modulated by GS in community-dwelling older people in Japan.

Methods: A total of 1486 community-dwellers in Japan (731 men and 755 women, aged 65-82 years) who participated in the 2nd to 7th waves of the National Institute for Longevity Sciences-Longitudinal Study of Aging (NILS-LSA) project, were followed up to 20 years. Twelve subgroups were created according to the intersections of GS (sex-specific tertiles: low, intermediate, and high) and BMI (underweight 1, <18.5; underweight 2, 18.5-<21.5; normal, 21.5-<25.0; and overweight/obese, ≥25.0 kg/m ). Incident disability was defined as requiring care need level 1 or higher, as per the long-term care insurance system in Japan. The hazard ratios (HR) and corresponding 95% confidence intervals (CI) of incident disability (the 'high GS-normal BMI' subgroup was used as the reference) were calculated using a multivariable-adjusted Cox proportional hazards model. The model was adjusted for baseline characteristics on sex, participation wave, smoking, depressive symptoms, total physical activity, disease history, education level, residential status, and energy intake.

Results: Compared with that for participants in the 'high GS-normal BMI' group, the risk of incident disability was higher for participants in the 'low GS-normal BMI', 'low GS-overweight/obese', 'intermediate GS-underweight 1' and 'intermediate GS-overweight/obese' groups. The multivariable-adjusted HR (95% CI) were 1.72 (1.27-2.32, P value<0.001), 1.81 (1.27-2.58, P value = 0.001), 2.42 (1.35-4.32, P value = 0.003) and 1.53 (1.06-2.20, P value = 0.023), respectively. The results did not change substantially when participants with disability occurring within 1 and 2 years of follow-up were excluded, or when the competing events were death without incident disability or dementia.

Conclusions: The joint effect of BMI and GS was more pronounced in those of normal weight or overweight/obese status and low GS, along with underweight or overweight/obese status and intermediate GS in predicting disability. The lack of observed joint effect for those underweight with low GS was likely due to insufficient sample size. GS consideration is necessary for weight management in older adults, and interventions for individuals who are underweight or overweight/obese with adequate GS need not be prioritized for disability prevention.