Background This study aims to assess the construct validity of a body composition-defined definition of sarcopenic obesity based on low appendicular lean mass relative to fat mass (ALMI FMI) and high fat mass index (FMI) and to compare with an alternative definition using appendicular lean mass index (ALMI) and percent body fat (%BF). Methods This is a secondary analysis of two cohort studies: the National Health and Examination Survey (NHANES) and the Health, Aging, and Body Composition study (Health ABC). Sarcopenic obesity was defined as low ALMI FMI combined with high FMI and was compared with a widely used definition based on ALMI and %BF cut-points. Body composition Z-scores, self-reported disability, physical functioning, and incident disability were compared across body composition categories using linear and logistic regression and Cox proportional hazards models. Results Among 14, 850 participants from NHANES, patients with sarcopenic obesity defined by low ALMI FMI and high FMI (ALMI FMI-FMI) had above-average FMI Z-scores [mean (standard deviation): 1.00 (0.72)]. In contrast, those with sarcopenic obesity based on low ALMI and high %BF (ALMI-%BF) had below-average FMI Z-scores. A similar pattern was observed for 2846 participants from Health ABC. Participants with sarcopenic obesity based on ALMI FMI-FMI had a greater number of disabilities, worse physical function, and a greater risk of incident disability compared with those defined based on ALMI-%BF. Conclusions Body composition-defined measures of sarcopenic obesity defined as excess adiposity and lower-than-expected ALMI relative to FMI are associated with functional deficits and incident disability and overcome the limitations of using %BF in estimating obesity in this context.