We estimated the glomerular filtration rate in 33 patients from our diabetic clinic using three methods: the creatinine clearance measured from a timed urine sample and a serum creatinine; the creatinine clearance calculated from the serum creatinine according to the formula of Cockcroft and Gault; and, the plasma clearance of ethylenediaminetetra-acetic acid (EDTA) labeled with 51Cr ([51Cr]EDTA). We repeated the measurements in seven subjects. The measured creatinine clearance was not reproducible. The other two methods were correlated, but not according to the formula y = x. The calculated creatinine clearance significantly underestimated the [51Cr]EDTA clearance particularly at higher [51Cr]EDTA clearance rates. [51Cr]EDTA clearance has been shown by others to parallel, but underestimate, inulin clearance, the optimal method of estimating glomerular filtration rate but difficult to perform in routine practice. Accurately measuring renal function in routine clinical practice is difficult, and this must be borne in mind when making clinical decisions based on current measurements.