Background. The formula of Cockcroft and Gault, relating age, weight, sex and plasma creatinine is widely used to estimate glomerular function. Its ability to predict [Tc-99m]DTPA clearance, an accepted measure of GFR, is therefore of interest. Aim: To test prospectively the reliability of the formula in predicting the [Tc-99m]DTPA clearance in subjects with normal and sub-normal renal function. Methods: The 24 hour creatinine clearance (Ccr), the derived clearance using the formula (C(der)), and the [Tc-99m]DTPA clearance have been compared in 47 male and 47 female subjects, aged between 15 and 71 years. Obese and non-obese subjects (males > 90 kg, females > 80 kg body weight) were considered separately. Results. Across the range of [Tc-99m]DTPA clearance studied (14 to 125 Ml/minute), in non-obese subjects, both Ccr and C(der) were correlated similarly with [Tc-99m]DTPA clearance in males and females. In obese subjects as a group, C(der) was also closely correlated with [Tc-99m]DTPA clearance but the scatter of data was wider. Among non-obese subjects, correlations were high between either Ccr or C(der) and [Tc-99m]DTPA clearance values < 101 mL/minute but were not significant at [Tc-99m]DTPA clearance values > 100 mL/minute. For [Tc-99m]DTPA clearance values between 14 and 100 mL/minute, C(der) was a more direct and more precise predictor of [Tc-99m]DTPA clearance [DTPA clearance = 1.01(C(der)) + 2.34, with 95% confidence limits of +/- 18 mL/minute] than of Ccr [Ccr = 1.34(C(der)C) - 10.86, confidence limits +/- 28 mL/minute]. Conclusion: For very mild through to moderate renal failure, the formula of Cockcroft and Gault can evidently be used as a reasonable estimate of GFR.