Measuring skeletal muscle area (SMA) at the third lumbar vertebra level (L3) using computed tomography (CT) is increasingly popular for diagnosing low muscle mass. The aim was to describe the effect of the CT L3 cut-off choice on the prevalence of low muscle mass in medical and surgical patients. Two hundred inpatients, who underwent an abdominal CT scan for any reason, were included. Skeletal muscle area (SMA) was measured according to Hounsfield units on a single CT scan at the L3 level. First, we calculated sex-specific cut-offs, adjusted for height or BMI and set at mean or mean-2 SD in our population. Second, we applied published cut-offs, which differed in statistical calculation and adjustment for body stature and age. Statistical calculation of the cut-off led to a prevalence of approximately 50 vs. 1% when cut-offs were set at mean vs. mean-2 SD in our population. Prevalence varied between 5 and 86% when published cut-offs were applied (p < 0.001). The adjustment of the cut-off for the same body stature variable led to similar prevalence distribution patterns across age and BMI classes. The cut-off choice highly influenced prevalence of low muscle mass and prevalence distribution across age and BMI classes.