DXA‐Measured Total and Regional Fat‐to‐Lean Mass Ratio and Mortality Risk in Chinese Older Adults: A 20‐Year Prospective Study

BACKGROUND: The association of fat-to-lean mass ratio (FLR) with mortality remains unclear in older adults. This study aimed to investigate the sex-specific associations of total and regional FLR with mortality in older Chinese adults. METHODS: This analysis included 2000 men (mean age: 72.4 ± 5.0 years) and 2000 women (mean age: 72.6 ± 5.4 years) aged ≥ 65 years from the Hong Kong OS cohort. Total and regional (trunk, abdominal, arms, legs) FLRs were measured using DXA at baseline. Mortality was ascertained via Hong Kong death registry. Multivariable Cox and Fine-Gray competing risk models were used to assess the associations of FLR with all-cause, cardiovascular (CVD) and cancer mortality, with results presented as hazard ratios (HRs) and 95% confidence intervals (CIs). C-index was used to quantify the discriminative ability of FLR measures and BMI in estimating mortality risk. RESULTS: Over a median follow-up of 18.2 years (IQR: 11.6-20.7), 2446 deaths (CVD: 511, cancer: 644) occurred. The 1-, 10- and 20-year survival rates were 0.990, 0.753 and 0.354 for men, and 0997, 0.858 and 0.500 for women. In men, FLR was inversely associated with all-cause mortality when analysed continuously for whole-body (HR per SD increase: 0.94, 95% CI: 0.89-1.00), trunk (HR: 0.92, 0.87-0.97), abdominal (HR: 0.92, 0.86-0.97) and arm FLR (HR: 0.94, 0.89-1.00), with all p < 0.05. In categorical analyses, men in the lowest tertile (T1) of whole-body, trunk, abdominal and arm FLR had significantly higher mortality risk compared to those in the middle tertile (HRs > 1, all p < 0.05). In women, trunk (HR per SD increase: 0.94, 0.88-1.00, p = 0.041) and abdominal FLRs (HR: 0.91, 0.85-0.97) were inversely associated with all-cause mortality, with T1 of abdominal FLR showing increased risk (HR: 1.18, 1.01-1.37) compared to T2. Restricted cubic splines revealed L-shaped associations for most FLR measures (p-nonlinearity < 0.05), except leg FLR in women (P-nonlinearity = 0.060). For cause-specific mortality, no significant associations were observed for CVD mortality in either sex, only leg FLR was inversely associated with cancer mortality in men (HR per SD increase: 0.89, 0.80-0.99; HR for T1 vs. T2: 1.29, 1.00-1.66, both p < 0.05), but not in women. All FLR measures showed higher discriminative ability than BMI for mortality prediction (all p < 0.01). CONCLUSIONS: In Chinese older adults, lower FLR was associated with higher all-cause mortality in an L-shaped relationship for both sexes, but showed no significant association with CVD mortality. Lower leg FLR was associated with increased cancer mortality only in men. FLR measures were more predictive than BMI for mortality risk stratification.