51Cr-EDTA, inulin and creatinine clearances were measured simultaneously in 11 diabetic children and 12 healthy young adults. The clearances in all individuals were 115 +/- 24 ml/min per 1.73 m2SA for 51Cr-EDTA, 118 +/- 25 ml/min per 1.73 m2SA for inulin and 157 +/- 35 ml/min per 1.73 m2SA for creatinine. Values were higher in the diabetic children but the differences were not significant. 51Cr-EDTA clearance significantly underestimated that of inulin by a mean of -7.4 +/- 2.5 (SEM) ml/min per 1.73 m2SA (p less than 0.01) and creatinine clearance significantly overestimated that of inulin by 37.6 +/- 3.3 (SEM) ml/min per 1.73 m2SA (p less than 0.0001). Similarly, the mean ratio of 51Cr-EDTA to inulin clearance was 0.94 (95% CI 0.90-0.98) and that of creatinine to inulin clearance was 1.32 (95% CI 1.27-1.37); the differences between diabetics and controls were not significant. Correlation coefficients were 0.93 between 51Cr-EDTA and inulin clearances, and 0.95 between inulin and creatinine clearances. The pooled coefficient of variation between clearances within an individual was higher with inulin, 10.3 +/- 6.5% (SD), than 51Cr-EDTA, 7.3 +/- 5.1% (p less than 0.001, t test). These results show that 51Cr-EDTA clearance underestimates that of inulin to a similar extent in both diabetic children and healthy controls and creatinine clearance overestimates inulin clearance to a greater but similar extent in both groups. The methodological variation in 51Cr-EDTA measurement is less than with the other 2 methods. Therefore, we recommend the use of the renal clearance of 51Cr-EDTA for the measurement of GFR in diabetic children.