Abstract
BACKGROUND: Musculoskeletal conditions and physical frailty have overlapping constructs. We aimed to quantify individual contributions of musculoskeletal factors to frailty. METHODS: Participants included 347 men and 360 women aged ≥60 yr (median ages; 70.8 (66.1-78.6) and 71.0 (65.2-77.5), respectively) from the Geelong Osteoporosis Study. Frailty was defined as ≥3, pre-frail 1-2, and robust 0, of the following; unintentional weight loss, weakness, low physical activity, exhaustion, and slowness. Measures were made of femoral neck BMD, appendicular lean mass index (ALMI, kg/m(2)) and whole-body fat mass index (FMI, kg/m(2)) by DXA (Lunar), SOS, BUA and SI at the calcaneus (Lunar Achilles Insight) and handgrip strength by dynamometers. Binary and ordinal logistic regression models and AUROC curves were used to quantify the contribution of musculoskeletal parameters to frailty. Potential confounders included anthropometry, smoking, alcohol, prior fracture, FMI, SES and comorbidities. RESULTS: Overall, 54(15.6%) men and 62(17.2%) women were frail. In adjusted-binary logistic models, SI, ALMI and HGS were associated with frailty in men (OR = 0.73, 95%CI 0.53-1.01; OR=0.48, 0.34-0.68; and OR = 0.11, 0.06-0.22; respectively). Muscle measures (ALMI and HGS) contributed more to this association than did bone (SI) (AUROCs 0.77, 0.85 vs 0.71, respectively). In women, only HGS was associated with frailty in adjusted models (OR = 0.30 95%CI 0.20-0.45, AUROC = 0.83). In adjusted ordinal models, similar results were observed in men; for women, HGS and ALMI were associated with frailty (ordered OR = 0.30 95%CI 0.20-0.45; OR = 0.56, 0.40-0.80, respectively). CONCLUSION: Muscle deficits appeared to contribute more than bone deficits to frailty. This may have implications for identifying potential musculoskeletal targets for preventing or managing the progression of frailty.