Measurement of glomerular filtration rate (GFR) following the injection of one of several suitable marker substances remains the best method to determine the severity of renal insufficiency as well as its rate of progression. However, the expense of these procedures continues to restrict their use. A second alternative is the determination of creatinine clearance (CCr) after oral administration of cimetidine. This drug blocks tubular secretion of creatinine almost completely, and CCr measured under these conditions is reported to be nearly identical to GFR in mild or severe renal failure. The optimal dose and timing of cimetidine for this purpose is still uncertain, but a single 1,200-mg dose 2 hours before beginning urine collection is probably suitable. A third alternative is simply the measurement of serum or plasma creatinine (PCr) concentration. However, it is well established that substantial reductions in renal function may occur before PCr becomes abnormal. GFR in adults with chronic renal failure can be approximately estimated from PCr, provided it is greater than 2 mg/dL, with the aid of additional demographic and biochemical variables. Gender, height, weight, age, and race should be taken into account. Further study is needed to derive the best formula for predicting GFR from PCr and other variables. Finally, CCr (without cimetidine) is still in use. This is unfortunate because it has been established that rather than improving on the estimation of GFR from PCr, CCr (determined from urinary creatinine measurements as well as PCr) is a less reliable guide to GFR than PCr alone. The CCr/GFR ratio is almost always greater than unity and increases with decreasing GFR to a maximum of approximately 1.7 at a GFR of approximately 20 mL/min. Furthermore, the variability of CCr is greater than that of PCr. Measurement of CCr (without cimetidine) is an anachronism and should be abandoned.