Tamara van Donge

Dynamics of creatinine in ELBW neonates 133 7 Supporting Information 2). Comparing the equation for calculation of total body water suggested by Shaffer et al. (mean (min-max), 0.618 L (0.264-1.25 L) and our 0.7 L/kg (mean (min-max), 0.627 L (0.26-1.28 L)), resulted in a minor difference of 8.7 ml on average. 14 Table 2: Parameter estimates of final creatinine model together with effect size estimates. CV, coefficient of variation, MOD; mode of delivery (1 for C-section and 0 for vaginal delivery), IBU; ibuprofen treatment (1 for yes and 0 for no treatment). Median GA was set at 26.73 weeks. Parameter (unit) Estimates (RSE%) IIV [CV%] Population parameters Crea birth (mg/dl) 0.597 (2.31) 0.252 [25.6] Kin production (mg/day) 3.55 (1.44) - CL BL (L/day) 0.075 (3.95) 0.255 [25.9] Emax (day -1 ) 0.874 (0.36) - t 50 (days) 21.1 (8.72) 0.49 [52] Hill 1.31 (4.79) 0.481 [51] Covariate parameters Parameter-covariate relationship IBU effect on CL BL (no / yes) 2.73 (0.29) / 2.55 (0.61) CL !" ! = CL !" "#" × (1 + 2.73) / CL !" "#" × (1 + 2.55) GA effect on t 50 -2.35 (36.3) t50 ! = t50 "#" × GA ! GA $ %&.() × (1 − 0.24 × MOD) MOD effect on t 50 -0.24 (47) GA effect on Crea birth 1.6 (21.6) Crea !"#$% ! = Crea !"#$% "#" × GA & GA ' (.* Residual variability Proportional error 0.11 (1.5) [11] Inter-individual variability (IIV) could be estimated for the parameters Crea birth , CL BL , t 50 and Hill (Table 2). In the final model, typical creatinine concentration at birth (Crea birth ) was estimated to be 0.597 mg/dl with 25.6% IIV, and to increase in more mature ELBW cases (+6% per increasing week of GA, Table 2). For the typical patient, ibuprofen treatment (yes/no per day) accounted for a 5% decrease (for ibuprofen treatment) in CL BL (25.9% IIV). The t 50 estimate for a typical patient was 21.1 days and decreased in more mature patients (-8% per increasing week of GA; shorter t 50 in more mature patients). Additionally, MOD accounted for shorter t 50 when delivered by C-section (-24%, Table 2). The Hill factor was estimated at 1.31 with 51% IIV. RSEs of the typical population parameters and corresponding IIV values were less than 10%, and RSEs for the covariate-parameter relationships were not exceeding 47% (Table 2). No model misspecification was observed as the distribution of the observations was symmetrical