The validity of the urinary urea nitrogen (UUN) estimate of total urinary nitrogen (TUN) was tested in patients who required iv nutrition. UUN and TUN were determined in 120 urine collections from ten preoperative, 13 postoperative, and 11 stressed patients. The relationship between TUN and UUN was examined by linear regression, and analysis of covariance was used on log-transformed data to assess differences between the patient groups. Although there was a close relationship between UUN and TUN for the preoperative patients (r2 = .94, total range of differences = 3.85 g N), this was not as accurate in postoperative and stressed patients (r2 = .69 and .76, respectively, total range of differences = 16.8 and 10.7 g N, respectively). There was no significant difference between the slopes of the regression lines for the relationship between UUN and TUN for three groups (f = 1.1, df = 2114, p less than .3), but the intercepts of the regression lines differed between the preoperative and stressed patient groups (t = 3.47, v = 114, p less than .001). The relationship between TUN and UUN for the whole group was improved by the inclusion of the independent variables of both the patient's clinical state and the urinary creatinine excretion. Arm muscle circumference, which is an estimate of muscle mass, may replace creatinine excretion with little loss in prediction accuracy.