Relevant risk of carboplatin underdosing in cancer patients with normal renal function using estimated GFR: lessons from a stage I seminoma cohort.

Optimal dosing of carboplatin (C) based on measured glomerular filtration rate (GFR) appears crucial in the adjuvant treatment of patients with stage I seminoma. Contrary to the pivotal phase III trial GFR is often estimated by formulae. Our analysis demonstrates that the use of formulae leads to underdosing of C in a relevant number of patients and should not be regarded as standard of care.Background Seminoma stage I is the most frequent testis cancer and single-dose carboplatin (AUC7) is an effective and widely used adjuvant treatment. Underdosing of carboplatin by 10% has been shown to almost double the rate of relapse and hence correct dosing based on accurate GFR measurement is crucial. The gold standard of GFR measurement with a radiolabelled isotope is expensive and not readily available. In many institutions, it is replaced by GFR estimation with the Cockcroft–Gault formula, which might lead to significant carboplatin underdosing and potentially inferior clinical outcome. Methods Retrospective analysis of all patients with stage I seminoma treated with adjuvant carboplatin between 1999 and 2012. All patients had serum creatinine measured and underwent GFR measurement with a radioisotope (51Cr EDTA or 99mTc DTPA), which was compared with seven standard GFR estimation formulae (Cockcroft–Gault, CKD-EPI, Jelliffe, Martin, Mayo, MDRD, Wright) and a flat dosing strategy. Bias, precision, rates of under- and overdosing of GFR estimates were compared with measured GFR. Bland–Altman plots were done. Results A total of 426 consecutive Caucasian male patients were included: median age 39 years (range 19–60 years), median measured GFR 118 ml/min (51–209), median administered carboplatin dose 1000 mg (532–1638). In comparison to isotopic GFR measurement, a relevant proportion of patients would have received ≤90% of carboplatin dose through the use of GFR estimation formulae: 4% using Mayo, 9% Martin, 18% Cockcroft–Gault, 24% Wright, 63% Jelliffe, 49% MDRD and 41% using CKD-EPI. The flat dosing strategy, Wright and Cockcroft–Gault formulae, showed the smallest bias with mean percentage error of +1.9, +0.4 and +2.1, respectively. Conclusions Using Cockcroft–Gault or any other formula for GFR estimation leads to underdosing of adjuvant carboplatin in a relevant number of patients with Seminoma stage I and should not be regarded as standard of care. [ABSTRACT FROM AUTHOR]