Estimation of creatinine clearance in patients with metastatic ovarian cancer

The authors evaluated the predictive performance of four methods used to estimate creatinine clearance (Cl(cr)) in patients with metastatic ovarian cancer. Methods described by Cockcroft and Gault, Jelliffe, and two equations derived from cancer patients by Robinson and Tsubaki, were evaluated. Estimated Cl(cr) values obtained by each method using actual weight (ABW), ideal weight (IBW) and lower of ABW and IBW were compared with measured values determined by a 12- or 24-hour urine collection for 14 patients enrolled in a controlled clinical trial. The mean prediction errors (ME) and mean absolute errors (MAE) were calculated to evaluate the bias and precision, respectively, of each method. The relationship between predicted and measured Cl(cr) is poor (r = 0.38 to 0.54). Cockcroft and Gault using ABW (p = 0.21), Robinson using ABW (p = 0.44), and Jelliffe (p = 0.17) were equally unbiased predictors of measured Cl(cr). All other methods significantly underestimated measured Cl(cr). All methods appeared to be equally imprecise (p<0.05). The use of standard equations for estimating Cl(cr) in patients with ovarian cancer is predictive of the measured 24-hour value. The use of oncology specific equations does not improve the accuracy or precision of these estimates.