BACKGROUND: The dosing of drugs in patients with kidney dysfunction is often based on the estimates of kidney function. OBJECTIVE: To systematically compare the performance of the Modification of Diet in Renal Disease (MDRD) and Cockcroft-Gault (CG) equations for dosage adjustment. METHODS: We assessed agreement (concordance, kappa statistics [κ,κ _ω ]) between CG and MDRD using a Food and Drug Administration database to evaluate the effect of renal function on the pharmacokinetics of 36 approved drugs. Across the approved drugs, we compared the correlation between these 2 equations for renal drug clearance (Cl _ren ) and area under the concentration-time curve. For 26 approved drugs that require renal dose adjustment, we also compared dosing regimens and expected exposure using these equations. Sensitivity analyses were performed by adjusting the MDRD estimates for individualized body surface area and/or range of serum creatinine assay calibration errors. RESULTS: In the pharmacokinetic database with 973 subjects (age 18-95 years, weight 35-153 kg, female 33%), we found that the CG and the MDRD classification of renal function generally agreed (64.2%, κ = 0.54, κ _ω = 0.73). Among the subjects studied for drugs requiring renal dose adjustment, dosages in 12% were changed to a higher or lower dosing category by the MDRD compared to the CG equation. In particular, using MDRD in place of CG for dosage modification yielded higher dosing recommendations for subjects with a combination of age >80 years, weight 0.7 and ≤1.5 mg/dL; the coefficient of determination was also higher by CG than MDRD in trials that enrolled these or similar patients. CONCLUSIONS: For patients with advanced age, low weight, and modestly elevated serum creatinine, further work is needed before the MDRD equations can replace the CG equation for dose adjustment in the labeling.