Improving Hip Fracture Prediction by Using Sarcopenia‐Specific Cut‐Offs of T‐Scores for Osteosarcopenia: A Prospective Cohort Study
TL;DR
Osteosarcopenia defined with these specific cut‐offs was associated with an increased risk of hip fractures independent of clinical risk factors in both men and women and suggested that including possible sarcopenia and adopting sex‐ and sarcopenia‐specific cut‐offs to define osteosarcopenia could improve hip fracture prediction.
Abstract
ABSTRACT Background Older adults with osteosarcopenia, defined as the coexistence of osteopenia/osteoporosis and sarcopenia, have been associated with an increased risk of fractures compared with those with normal bone mineral density (BMD) and without sarcopenia. However, the conventional definition of osteosarcopenia has led to mixed results on its predictive accuracy for hip fractures. We aimed to define osteosarcopenia by identifying sex‐specific cut‐offs for low BMD according to different sarcopenia statuses and compare its predictive performance with osteosarcopenia defined conventionally and osteoporosis alone. Methods A cohort of 4000 community‐dwelling older adults (2000 females, mean age 72.5 ± 5.2 years) was recruited. Body composition and BMD of hip, lumbar spine and femoral neck were measured using dual energy X‐ray absorptiometry at baseline. According to the Asian Working Group for Sarcopenia 2019 consensus, all participants were classified as non‐sarcopenia, possible sarcopenia and sarcopenia. Incidence of hip fractures was documented during the follow‐up period from 2001 to 2013. The classification and regression tree (CART) analysis was performed to identify the optimal cut‐off values of T‐scores for participants with possible sarcopenia and those with sarcopenia, respectively. Cox proportional hazards regression models were used to estimate the associations between osteosarcopenia and hip fractures. The discrimination and predictive ability of osteosarcopenia were evaluated by Uno's concordance index (C‐index), time‐dependent net reclassification improvement (NRI) and integrated discrimination improvement (IDI). Results During an average of 9.2 years of follow‐up, 63 (3.2%) men and 69 (3.5%) women had at least one hip fracture. Sex‐ and sarcopenia‐specific cut‐offs of T‐scores were identified (men with possible sarcopenia: T‐score < −2.2; men with sarcopenia: T‐score < −2.0; women with possible sarcopenia: T‐score < −2.1; women with sarcopenia: T‐score ≤ −2.5). Osteosarcopenia defined with these specific cut‐offs was associated with an increased risk of hip fractures independent of clinical risk factors in both men and women (HR = 5.232, 95% CI = 3.172–8.631, and HR = 3.574, 95% CI = 2.01–6.355, respectively). It had the highest Uno's C‐index (men: 0.774 and women: 0.750, respectively) and outperformed osteosarcopenia using conventional definition and osteoporosis alone in predicting hip fractures by using NRI and IDI approaches. Conclusions Our findings suggested that including possible sarcopenia and adopting sex‐ and sarcopenia‐specific cut‐offs to define osteosarcopenia could improve hip fracture prediction. Further studies are warranted to validate the cut‐offs of T‐score identified in our cohort.