151751-Najiba-Chargi

264 CHAPTER 14 PHARMACOKINETIC RESULTS SMM and weight were significantly correlated with a Pearson correlation coefficient of 0.6, as shown in figure 2. Based on goodness-of-fit plots the data were well described by the used PK model. For 5 patients no SMA at the L3 slice was available, therefore SMA at the C3 slice was used. To examine whether other relevant covariates had influence on the PK model of cisplatin, and thereby could influence the effect of SMM, albumin and GFR were first tested as covariates on the baseline model. No effect of albumin on the PK model was found. Addition of the GFR as a covariate on CL bound resulted in a drop in OFV of 28 and 33 points, for the GFR calculated based on creatinine and cystatin C, respectively (nonhierarchical models). Also, a drop in IIV and a relevant difference in effect size were encountered. Therefore, GFR was also evaluated in combination with SMM. The PK model was extended with SMM, weight, and FFM as covariates on the PK parameters CL free , V free , Q, V p , CL bound , and V bound of cisplatin. Using estimated exponents, compared to fixed exponents based on the theory of allometric scaling, led to a substantially better description of the data, as indicated by the OFV. The exponent was unidentifiable for V free and Q. Addition of GFR, next to SMM, had no effect on the PKmodel (additional dOFV -5 and -4, for creatinine and cystatin C, respectively), which could be ex - plained by a relationship between weight and GFR (weight is even used to calculate creatinine clearance in the Cockcroft-Gault formula).Therefore, SMM, weight, and FFM were added as potential covariates on CL free , CL bound , and V bound of cisplatin while estimating the exponents. The OFV was significantly decreased by addition of SMM (dOFV -64, p <0.005), weight (dOFV -77, p <0.005), and FFM (dOFV -70, p <0.005). Since cisplatin is dosed based on BSA, BSA was also tested as covariate in the final model, which resulted in a significant drop in OFV (dOFV -86, p <0.005). The parameter estimates and estimated exponents for the model with SMM are shown in Supplementary materials S2. Figure 2. Correlation between weight and skeletal muscle mass. Males are displayed by black circles and females by grey triangles. R= 0.63, p<0.005