Objective: The primary aims of this trial were to evaluate the reproducibility of a portable handheld calorimeter (Medgem) in a clinical population, and to compare its measures with a calorimeter in typical use with these patients. Design: Cross-sectional clinical validation study. Setting: Outpatient Clinical Research Center. Subjects: A total of 24 stable home nutrition support patients. Interventions: In random order three measures of resting metabolic rate (RMR) were taken after a 4-h fast, 15 min rest and 2-h abstention from exercise. Two measures were taken with the same Medgem (MG) and one with the traditional calorimeter (Deltatrac). Reproducibility of MG measures and their comparability to a Deltatrac measure were assessed by Bland-Altman analysis, with 47250 kcal/day established a priori as a clinically unacceptable error. In addition, disagreement between the two types of measures was defined as greater than 10% difference. Results: The mean difference between two MG measures was À6.8 kcal/day, with limits of agreement between 233 and À247 kcal/day and clinically acceptable. The mean difference between the Deltatrac and mean of two MG measures was À162 kcal/day, with limits of agreement between 577 and À253 kcal/day and clinically unacceptable. In all, 80% of the repeated MG RMR measures agreed within 10%, and the mean MG reading agreed with the Deltatrac in 60% of cases. Conclusions: RMR obtained using the MG calorimeter has an acceptable degree of reproducibility, and is acceptable to patients. The MG measures, however, are frequently lower than traditional measures and require further validation prior to application to practice in this vulnerable patient group.