AbstractObjective:To evaluate the performance of mid-upper arm circumference (MUAC) to identify thinness in the late adolescence period (aged 15–19 years) in Ethiopia.Design:We conducted a school-based cross-sectional study. The receiver operating characteristics curve was used to examine the validity of MUAC compared with BMIZ-score to identify adolescents with thinness (BMIZ-score <−2sd).Settings:Fifteen high schools (grade 9–12) located in Addis Ababa, Ethiopia.Participants:A total of 851 adolescent (456 males and 395 females) were included in the study.Results:The prevalence of thinness and severe thinness among high-school adolescents in Addis Ababa was 9·5 % (95 % CI 7·7, 11·7 %). The overall AUC for MUAC against BMIZ-score <−2 SD was 0·91 (95 % CI 0·88, 0·93). The optimal MUAC cut-offs to identify thinness were 23·3 cm for males and 22·6 cm for females. These cut-off points give high sensitivity and specificity for both males (a sensitivity of 87·9 % and a specificity of 75·9 %) and females (a sensitivity of 100 % and a specificity 88·2 %).Conclusions:MUAC has a comparable level of accuracy with BMIZ-score to identify thinness in adolescents aged 15–19 years. Hence, MUAC could be used as an alternative tool for surveillance and screening of thinness among adolescents aged 15–19 years. The optimum cut-off proposed by this study may incorrectly include a large number of adolescents when used in a relatively well-nourished population. In this situation, it would be necessary to choose a cut-off with greater positive predictive value.