The estimation of dietary protein intake in chronic renal failure.

Forty-five adult clinic patients with chronic renal failure each supplied a 4-day weighed dietary record, a 24-h urine collection, and a nocturnal spot urine sample. Total nitrogen (N) losses derived from the urines were corrected for proteinuria and non-urea nitrogen excretion. Individual estimates of N intake were compared by correlation and assessing the level of agreement. Daily urea N excretion derived from the spot sample correlated well with the 24-h collection P less than 0.001, but the degree of agreement was poor, mean difference being +1.62 g with 95% limits of +5.7 to -2.47 g. The correlation between the spot-sample-derived N loss and dietary N intake was poor, r = 0.42; P less than 0.05. For 12 patients taking a low-protein diet, N intake correlated well with 24-h urine derived N losses, P less than 0.001, mean difference being +0.59 g, 95% limits +2.39 to -1.21 g. The correlation and agreement was less satisfactory for the subjects who had not received dietary instruction, due largely to individual variation in day-to-day protein intake. Use of spot urine samples is too inaccurate for routine clinical practice. Single 24-h urine derived estimates of N intake are only of value for assessing patients previously prescribed a low-protein intake.