BACKGROUND: Accurate assessment of nutrition status is essential in identifying subpopulations of critically ill patients who are malnourished and at higher mortality risk. The aim of this analytic observational study was to assess the performance of physical assessment and anthropometric measures commonly used in clinical research. METHODS: A prospective study was undertaken in 31 intensive care units (ICUs) with a focus on patients with short-term contraindications to enteral nutrition. Within 24 hours of admission to the ICU, the following measures were collected: the Subjective Global Assessment components measuring subcutaneous fat loss and muscle wasting, height, weight, mid-upper-arm circumference, and triceps skinfold thickness (TSF). Mid-arm muscle circumference and body mass index (BMI) were calculated. BMI was assessed as a continuous variable and categorized according to the World Health Organization (WHO) categories. The primary outcome was hospital discharge mortality. RESULTS: In total, 1363 patients were enrolled. BMI, analyzed according to WHO categories (P = .09), and TSF (P = .32) failed to demonstrate statistically significant predictive ability. TSF failed to demonstrate statistically significant clinical utility (area under the receiver operating characteristic curve, 0.52; 95% confidence interval, 0.48-0.56). All other individual measures demonstrated statistically significant predictive ability and statistically significant clinical utility. CONCLUSIONS: On the basis of the results of our ICU cohort, we recommend caution when using BMI categorized according to WHO definitions. We cannot recommend collection of TSF. More research is required to understand reliability, performance, and use before our results are able to be generalized to other ICU populations.