OBJECTIVE The aim of this study was to determine the prevalence of muscle strength-based sarcopenia and to determine possible predictors. METHODS This is a cross-sectional population-based study in the community-dwelling Turkish elderly. Anthropometric measurements, namely body height, weight, triceps skin fold (TSF), mid upper arm circumference (MUAC), waist circumference (WC) and calf circumference (CC), were noted. The midarm muscle circumference (MAMC) was calculated by using MUAC and TSF measurement. Sarcopenia was assessed, adjusted for body mass index (BMI) and gender, according to muscle strength. Physical performance was determined by 4 m walking speed (WS; m/s). The receiver operating curve analysis was performed to determine cut-offs of CC, MAMC and 4 m WS. RESULTS A total of 879 elderly subjects, 50.1% of whom were female, were recruited. The mean handgrip strength (HGS) and s.d. was 24.2 (8.8) kg [17.9 (4.8) female, 30.6 (7.1) male]. The muscle function-dependent sarcopenia was 63.4% (female 73.5%, male 53.2%). The muscle mass-dependent sarcopenia for CC (<31 cm) and MAMC(<21.1 cm in males, <19.9 cm in females) was 6.7% and 7.3%, respectively. The prevalence of low 4 m WS (≤ 0.8 m/s) was 81.8% (91.3% in females and 72.3% in males, respectively). We compared MAMC, CC and 4 m WS and found that AUC for 4 m WS was the best predictor of sarcopenia. CONCLUSIONS An adequate muscle mass may not mean a reliable muscle function. Muscle function may describe sarcopenia better compared with muscle mass. The CC, MAMC and 4 m WS cut-offs may be used to assess sarcopenia in certain age groups.