Selecting an equation to estimate glomerular filtration rate for use in renal dosage adjustment of drugs in electronic patient record systems.

STUDY OBJECTIVE: To identify a variant of the Cockcroft-Gault equation whose estimate would agree with the Modification of Diet in Renal Disease (MDRD) estimate of glomerular filtration rate (GFR) since the MDRD equation may not be programmable in some electronic patient record systems.

DESIGN: Prospective case series.

SETTING: A 625-bed, adults-only, private, tertiary care teaching hospital.

PATIENTS: Two hundred eight consecutive hospitalized patients with MDRD-estimated GFRs less than 90 ml/minute/1.73 m 2 . Seventeen patients were black (includes native African immigrants).

INTERVENTION: Chemistry assays were performed, and patients' records were reviewed for age, ethnic background, height, and actual weight. Ideal weight, corrected weight, body mass index, body surface area (BSA), and GFR by the original MDRD equation were calculated for each patient.

MEASUREMENTS AND MAIN RESULTS: Cockcroft-Gault estimates of renal clearance were calculated by using actual weight, ideal weight, and corrected weight both with and without correction for BSA. These estimates, as well as an estimate of GFR by the abbreviated MDRD equation, were compared with the original MDRD estimate. The results obtained with the abbreviated MDRD equation and with the Cockcroft-Gault equation that used corrected weight and BSA adjustment had the best agreement; both were within +/- 30% of the original MDRD equation in 80% of the 208 patients' results. All other Cockcroft-Gault variants tested were less accurate.

CONCLUSION: Electronic patient record databases may not contain a nominal variable database field for black or non-black status, which is required by the MDRD equations. The Cockcroft-Gault equation that used corrected weight and BSA adjustment performed about as well as the abbreviated MDRD equation and can be programmed into electronic patient records. Given the small number of black patients included in this study, further study in this patient population is recommended before applying this Cockcroft-Gault variant to this subgroup.