Clinical validation of SARC-F by proxy as a practical tool to evaluate sarcopenia in dependent older adults.

Introduction: Sarcopenia is a prevalent disorder in older adults with significant adverse outcomes and regular screening is recommended for those at risk. The SARC-F questionnaire is the most commonly recommended screening tool for sarcopenia. However, as a self-reported tool, it cannot be applied to dependent individuals with communication problems. We hypothesized that implementation of the proxy-reported SARC-F (SARC-F by proxy) would be non-inferior in screening sarcopenia when compared with the standard SARC-F. Thus, we aimed to investigate the clinical validity of the SARC-F by proxy in identifying sarcopenia in older adults and to compare its performance with the standard SARC-F. Additionally, we aimed to determine the ideal cut-off of SARC-F by proxy in screening sarcopenia.

Materials And Methods: This is a validation study including older adults aged ≥60 years without communication problems and their close proxies. The participants were recruited from a geriatric outpatient clinic of a tertiary health center and a nursing home. Standard SARC-F was transformed to SARC-F by proxy and administered to the proxies of older adults, and standard SARC-F was administered to the patients simultaneously in different rooms. We defined sarcopenia as probable and confirmed by the EWGSOP2 consensus report. We performed receiver operating characteristics (ROC) and sensitivity/specificity analyses of SARC-F by proxy for diagnosing sarcopenia and compared its performance with standard SARC-F by the DeLong test.

Results: We included 172 older adults (median age: 72; 44.8% female) and 107 proxies in close contact (median age: 55, 63.2% female). The prevalence of probable and confirmed sarcopenia was 18.9% and 12.9%, respectively. For both definitions, area under the curve (AUC) values of SARC-F by proxy and standard SARC-F were moderate and similar [probable sarcopenia: 0.619 and 0.624 (p = 0.9); confirmed sarcopenia 0.613 and 0.645 (p = 0.7), respectively]. The best balance between sensitivity and specificity was achieved with a SARC-F by proxy score of ≥2 for both sarcopenia definitions (sensitivity levels were 74.7% and 77.8%, and specificity levels were 50.0% and 49.6%, for probable and confirmed sarcopenia, respectively).

Discussion: SARC-F by proxy showed a similar, non-inferior performance compared to the standard SARC-F in the evaluation of sarcopenia. Our results suggest that it can be used instead of standard SARC-F to screen sarcopenia in older patients with communication problems. Further validation studies in different populations are warranted to support our findings.