Comparative analysis of sarcopenic obesity detection and associated mental–cognitive outcomes in community-dwelling older adults: impact of using body mass index versus waist circumference

BACKGROUND: Sarcopenic obesity (SO), defined as the coexistence of sarcopenia and obesity, is associated with adverse physical, psychological, and cognitive outcomes in older adults. In community settings lacking advanced body composition measurement tools, simplified screening indicators are often used. The Ishii score, based on age, handgrip strength, and calf circumference, offers high predictive accuracy for sarcopenia. Body mass index (BMI) reflects overall adiposity, whereas waist circumference (WC) represents central obesity. However, evidence directly comparing SO prevalence and related characteristics using BMI versus WC in Chinese community populations remains limited. OBJECTIVES: To compare SO prevalence, agreement, and associated factors when using BMI or WC as the obesity index in combination with the Ishii score, and to examine differences in sociodemographic, psychological, and cognitive characteristics across SO phenotypes. METHODS: A cross-sectional study was conducted among 880 community-dwelling older adults in Guangxi, China. Sarcopenia was defined using the Ishii score (≥ 105 for men, ≥ 120 for women), and obesity was defined by BMI (≥ 28 kg/m²) or WC (≥ 90 cm in men, ≥ 85 cm in women). Participants were categorized into four phenotypes: non-obesity/non-sarcopenia (NONS), isolated obesity (IO), isolated sarcopenia (IS), and sarcopenic obesity (SO). Cohen's Kappa assessed agreement between methods. Group differences were analyzed using chi-square(χ²) tests, t tests, one-way analysis of variance (ANOVA), and logistic regression. Depressive symptoms and cognitive impairment were assessed using the 9-item Patient Health Questionnaire (PHQ-9) and the Ascertain Dementia-8 (AD-8), respectively. RESULTS: SO prevalence was 3.2% (BMI method) versus 14.1% (WC method), with moderate agreement (Kappa = 0.572). Multinomial logistic regression revealed that WC-based SO was significantly associated with male sex, low monthly income, and lack of attention to nutritional information. BMI-based SO was primarily associated with older age and the presence of multimorbidity. For mental-cognitive outcomes, both SO phenotypes were associated with a higher risk of depressive symptoms and cognitive impairment. However, the strength of association with depressive symptoms was markedly greater for the BMI-defined SO group (OR = 7.823, 95% CI: 3.198-19.139) than for the WC-defined SO group (OR = 2.105, 95% CI: 1.097-4.042). CONCLUSIONS: The choice of obesity index substantially influences SO detection, with WC identifying a much larger and likely earlier-stage at-risk population. The BMI-based definition, while identifying fewer cases, captured a severe SO phenotype with exceptionally strong associations with depressive symptoms.