LAUSR

Objective To investigate the association of normal BMI with central obesity (CO), high BMI with CO, high BMI without CO, and normal BMI without CO, with function and cognition in older adults. Methods Cross-sectional study involving 754 participants ≥ 65 years. Data collected include demographics, cognition, and physical measurements. Results Females had a higher prevalence of high BMI with CO and a lower prevalence of high BMI without CO than males (61.0% vs 44.6% and 4.6% vs 15.0%, respectively). Within gender, CO groups, regardless of BMI, had lower mini-mental state examination (MMSE), handgrip strength (HGS), and longer timed-up-and-go (TUG) scores. Overall, the high BMI without CO group had the highest MMSE scores, HGS, and shortest TUG. Amongst males, HGS was significantly lower in the normal BMI with CO group (B −3.28, 95% CI −6.32 to −0.23, P = 0.04). CO, regardless of normal/high BMI, had significantly longer TUG time (B 2.65, 95% CI 0.45 to 4.84, P = 0.02; B 1.07, 95% CI 0.25 to 1.88, P = 0.01, respectively) than normal BMI without CO group. CO was associated with lower MMSE scores in both genders but significant only in males with normal BMI and CO (B −1.60, 95% CI −3.15 to −0.06, P = 0.04). Conclusion CO may be a better predictor of obesity and adverse outcomes in older adults. High BMI without CO was associated with better outcomes especially in males but require further validation. Prospective longitudinal studies are needed to ascertain the impact of BMI and/or CO on function, cognition, mortality, and gender differences.