Several diagnostic algorithms exist to detect sarcopenia in older adults. We compared the prevalence of sarcopenia according to the selected diagnostic algorithms.

This cross-sectional study compared the European Working Group of Sarcopenia in Older People (EWGSOP) 2010, updated EWGSOP 2019, the Foundation for National Institutes of Health (FNIH) and the International Working Group on Sarcopenia (IWGS) criteria in 778 outpatients of the Geriatric Clinic aged 60 to 89 years. Bioimpedance analysis (BIA) to estimate muscle mass, hand-held hydraulic dynamometer to measure muscle strength, the TUG test and gait speed to assess physical function were used.

The prevalence of sarcopenia varied from 0% to 6.43% depending on the algorithm. For the majority of associations between the different definitions of sarcopenia the agreement was null or fair (Cohen’s kappa between 0.2 and 0.4). Moderate agreement (Cohen’s kappa between 0.4 and 0.6) was found for only three relationships. Nevertheless, for these three relationships, McNemar’s test has given different results, indicating that even in the moderately agreeing algorithms, the shared diagnoses of sarcopenia concerned only part of subjects.

According to diagnostic algorithms the prevalence of sarcopenia is low in independent community-dwelling older adults. The agreement between the different definitions is poor.