Background and aims In older adults, body mass index (BMI) may underestimate clinically relevant adiposity because aging is accompanied by loss of lean mass and preferential accumulation of visceral and ectopic fat. We examined whether DXA-derived visceral adipose tissue (VAT) identifies hidden liver-metabolic risk beyond BMI in non-obese older adults. Methods We performed a cross-sectional analysis of the first available assessment from a hospital-based geriatric database. Participants with BMI <30 kg/m2 and non-missing baseline DXA-derived VAT were included. VAT quartiles were used as the main exposure. Primary outcomes were non-invasive steatosis scores, namely fatty liver index (FLI) and hepatic steatosis index (HSI). Supportive secondary outcomes were alanine aminotransferase (ALT) and gamma-glutamyl transferase (GGT), which were interpreted as sensitive but non-specific liver-related biomarkers. Multivariable robust linear regression models with HC3 standard errors were used, adjusting for age, sex, diabetes, Mini Nutritional Assessment score, and number of pathologies. Additional analyses examined waist circumference as a proxy for VAT, Lipschitz age-adapted BMI categories, and combined muscle-mass/VAT phenotypes. Model-specific complete-case analysis was used. Results The analytic cohort comprised 891 non-obese older adults (mean age 81.5 ± 7.2 years), including 619 women (69.5%) and 272 men (30.5%). Across increasing VAT quartiles, BMI, waist circumference, fat mass, android fat percentage, diabetes prevalence, HSI, and FLI increased. In adjusted analyses, the highest VAT quartile versus the lowest remained associated with higher FLI (beta +33.90, 95% CI 28.72 to 39.08), higher HSI (beta +6.72, 95% CI 5.75 to 7.68), higher ALT (+17.5%, 95% CI 3.6%–33.3%), and higher GGT (+33.9%, 95% CI 11.6%–60.5%). Waist circumference correlated with VAT (Spearman rho = 0.500, p < 0.001) and identified high VAT with an AUC of 0.734; the Youden-optimal cut-point was 86.0 cm. VAT discriminated FLI ≥ 60 with an AUC of 0.846 and a Youden-optimal cut-point of 1249.3 g. Participants with low muscle mass and high VAT had higher adjusted FLI, HSI, ALT, and GGT than those with preserved muscle and low VAT. Conclusions DXA-derived regional adiposity identifies hidden liver-metabolic vulnerability beyond BMI in non-obese older adults. Body-composition phenotyping may refine nutritional and metabolic risk assessment in later life, particularly when apparently reassuring BMI values obscure adverse fat distribution. When DXA is unavailable, waist circumference may provide a feasible first-line proxy for identifying older adults with high visceral adiposity. Extrapolation to non-elderly adults should be tested in external cohorts.
DXA-derived regional adiposity identifies hidden liver-metabolic risk beyond BMI in non-obese older adults: A cross-sectional study
Perna, Simone;Mazzola, Giuseppe;Rondanelli, Mariangela
2026-01-01
Abstract
Background and aims In older adults, body mass index (BMI) may underestimate clinically relevant adiposity because aging is accompanied by loss of lean mass and preferential accumulation of visceral and ectopic fat. We examined whether DXA-derived visceral adipose tissue (VAT) identifies hidden liver-metabolic risk beyond BMI in non-obese older adults. Methods We performed a cross-sectional analysis of the first available assessment from a hospital-based geriatric database. Participants with BMI <30 kg/m2 and non-missing baseline DXA-derived VAT were included. VAT quartiles were used as the main exposure. Primary outcomes were non-invasive steatosis scores, namely fatty liver index (FLI) and hepatic steatosis index (HSI). Supportive secondary outcomes were alanine aminotransferase (ALT) and gamma-glutamyl transferase (GGT), which were interpreted as sensitive but non-specific liver-related biomarkers. Multivariable robust linear regression models with HC3 standard errors were used, adjusting for age, sex, diabetes, Mini Nutritional Assessment score, and number of pathologies. Additional analyses examined waist circumference as a proxy for VAT, Lipschitz age-adapted BMI categories, and combined muscle-mass/VAT phenotypes. Model-specific complete-case analysis was used. Results The analytic cohort comprised 891 non-obese older adults (mean age 81.5 ± 7.2 years), including 619 women (69.5%) and 272 men (30.5%). Across increasing VAT quartiles, BMI, waist circumference, fat mass, android fat percentage, diabetes prevalence, HSI, and FLI increased. In adjusted analyses, the highest VAT quartile versus the lowest remained associated with higher FLI (beta +33.90, 95% CI 28.72 to 39.08), higher HSI (beta +6.72, 95% CI 5.75 to 7.68), higher ALT (+17.5%, 95% CI 3.6%–33.3%), and higher GGT (+33.9%, 95% CI 11.6%–60.5%). Waist circumference correlated with VAT (Spearman rho = 0.500, p < 0.001) and identified high VAT with an AUC of 0.734; the Youden-optimal cut-point was 86.0 cm. VAT discriminated FLI ≥ 60 with an AUC of 0.846 and a Youden-optimal cut-point of 1249.3 g. Participants with low muscle mass and high VAT had higher adjusted FLI, HSI, ALT, and GGT than those with preserved muscle and low VAT. Conclusions DXA-derived regional adiposity identifies hidden liver-metabolic vulnerability beyond BMI in non-obese older adults. Body-composition phenotyping may refine nutritional and metabolic risk assessment in later life, particularly when apparently reassuring BMI values obscure adverse fat distribution. When DXA is unavailable, waist circumference may provide a feasible first-line proxy for identifying older adults with high visceral adiposity. Extrapolation to non-elderly adults should be tested in external cohorts.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
