Estimation of glomerular filtration rate in patients with abnormal body composition and relation with carboplatin toxicity

Abstract

Background: Carboplatin clearance is correlated with Glomerular Filtration Rate (GFR), which is usually evaluated with creatinine clearance using Cockcroft-Gault (CG) formula. Because serum creatinine (sCR) level is related to total muscle mass, we hypothesized that an abnormal body composition with a low lean body mass (LBM) percentage (LBM / weight) may result in GFR overestimation and inadequate carboplatin dose calculation by using CG formula. Serum cystatin C (sCY) is an alternative marker of GFR which is not affected by muscle mass. We aimed to correlate GFR creatinine (GFRCr) and cystatin C (GFRCy) to LBM percentage and the risk of toxicity. A high GFRCr/GFRCy ratio might estimate the miscalculation of GFR with sCR. Methods: sCR and sCY were prospectively measured in consecutive cancer patients ( pts) before chemotherapy. GFRCr and GFRCy were calculated with CG and GRUBB formula, respectively. The LBM was assessed using CT-scan. Febrile neutropenia and grade 3-4 thrombocytopenia were analysed in a subgroup of pts treated with carboplatin AUC 5 +/- paclitaxel. Results: In 134 cancer pts (males 51%), median age was 63.7 (20 - 89). In 128 pts without severe renal insufficiency, sCR was correlated with LBM (r = 0.30, Pearson correlation coefficient, p <0.005) but sCY was not (r = - 0.11, p = 0.24). GFRCr was not correlated with total LBM but significantly decreased in pts with the lowest LBM percentage. In this group of pts, the GFRCr was significantly higher than GFRCy, indicating an over estimation of GFR with sCR (p = 0.009, Wilcoxon test). In 24 pts treated with carboplatin AUC5 +/- paclitaxel, 6 pts had febrile neutropenia or grade 3-4 thrombocytopenia. Mean LBM percentage was lower in pts with limiting toxicities (49% vs 64%, p = 0.0002). Similarly mean GFRCr/GFRCy ratio was higher in pts with limiting toxicities (1.9 vs 1.1, p = 0.004). By ROC analysis, the GFRCr/GFRCy ratio value with the best sensitivity-specificity was 1.4. In the whole cohort, 43 pts (32%) have a GFRCr/GFRCy ratio > 1.4. Conclusions: Pts with a low LBM percentage (sarcopenic overweight pts) are at high risk of inadequate calculation of renal function, over dosage of carboplatin and increased acute limiting toxicity. High GFRCr/GFRCy ratio allows to identify these pts.