Evaluation of creatinine-based formulas in dosing adjustment of cancer drugs other than carboplatin

BACKGROUND: Glomerular filtration rate (GFR) is often used to determine initial dosing of renally excreted cancer drugs. GFR can be calculated using the Cockcroft-Gault (CG) or the modified diet in renal diseases (MDRD) study formulas, both of which are based on serum creatinine levels. The MDRD formula is more accurate in noncancer patients, does not require patient weight, and is reported automatically by all laboratories in British Columbia, Canada. We previously showed that the CG and MDRD formulas have similar accuracy for carboplatin dosing in patients with gynecological malignancies. We now examine dosing of all renally excreted cancer drugs in the general cancer population. Since this setting does not include routine measurement of GFR, we report the concordance of estimates of GFR derived from the CG and MDRD formulas. METHODS: Patient data were collected retrospectively at the BC Cancer Agency. The primary outcome was the proportion of patients who would have received a different initial dose due to difference in the GFR. Each patient's dose was determined from dose adjustment tables stated in specific treatment protocols. The secondary outcome was concordance of the GFR derived from CG and MDRD, using the method of Bland and Altman. A difference of >30% was assumed to be clinically significant because this difference would usually lead to dose adjustment based on reclassification of renal function. RESULTS: A total of 313 patients were evaluated, with 40% male. The median age was 56 years, weight 67.5 kg, height 166 cm, and serum creatinine 74 micromol/L (0.84 mg/dL). The median GFR derived from the CG and MDRD formulas were 86.8 mL/min (mean 91 mL/min, SD +/- 30 mL/min) and 87.6 mL/min (mean 88 mL/min, SD +/- 26 mL/min), respectively. A total of 8.6% (27/313) of patients would have received a different dose due to difference in the GFR; of these, 67% (18/27) would have received a higher dose. A difference of >30% in GFR was found in 17.9% (56/313) of patients. CONCLUSIONS: There is good concordance of the GFR derived from the CG and MDRD formulas for most cancer patients, with less than 10% of patients expected to receive a different initial dose of chemotherapy. The MDRD formula may be a reasonable alternative to the CG formula for dosing of cancer drugs which are renally excreted or nephrotoxic.