Survivors of acute respiratory distress syndrome (ARDS) experience severe muscle wasting. Upper arm anthropometrics can provide a quick, non-invasive estimate of muscle status, but its accuracy is unknown. This study examines the accuracy of upper arm percent muscle area (UAMA) with reference measures of lean mass from dual energy X-ray absorptiometry (DXA). Data are from 120 ARDS survivors participating in a multicenter national study. Receiver operating characteristic (ROC) curves, by patient sex, demonstrated that UAMA did no better than chance in discriminating low appendicular skeletal muscle mass identified using DXA findings (c-statistics, 6 months: 0.50-0.59, 12 months: 0.54-0.57). Modest correlations of UAMA with DXA measures (whole-body: r = 0.46-0.49, arm-specific: r = 0.50-0.51, p < 0.001) and Bland-Altman plots indicate poor precision. UAMA is not an appropriate screening measure for estimating muscle mass when compared to a DXA reference standard. Alternate screening measures should be evaluated in ARDS survivors.