Table 2.
Parameter estimates of the final population pharmacokinetic model, including NONMEM estimates with uncertainty, shrinkage and bootstrap median and confidence intervals (n = 2000)
| Parameter | Parameter description | Estimate | (RSE%) | Bootstrap | |
|---|---|---|---|---|---|
| [SHR%] | Median | 95% CI | |||
| CL (L/h)a | Clearance | 11.1 | (8.2) | 10.9 | 8.5–14.0 |
| Vc (L)b | Central volume of distribution | 5.13 | (28) | 4.84 | 1.5–17.6 |
| Vp (L)c | Peripheral volume of distribution | 18.2 | (12) | 16.8 | 8.0–24.8 |
| Q (L/h)d | Inter-compartmental clearance | 14.5 | (19) | 12.7 | 3.1–24.7 |
| θcov (eCLcr ≤ 120) | eCLcr covariate effect on CL | 6.67 | (17) | 6.61 | 4.0–9.0 |
| fCL, Vc | Scaling factor for IIV CL and Vc | 1.31 | (39) | 1.30 | 0.2–2.5 |
| IIV CL (%CV) | Inter-individual variability in CL | 49 | (15) [0.8] | 50 | 36–71 |
| IIV Vc (%CV) | Inter-individual variability in Vc | 64b | |||
| Proportional residual error (%CV) | Proportional residual error | 33.3 | (5.9) [7.8] | 32.9 | 29–37 |
RSE relative standard error, SHR shrinkage on standard deviation scale, CI confidence interval, θ population parameter, eCLcr estimated creatinine clearance (mL/min), IIV inter-individual variability, CV coefficient of variation, BW body weight at intensive care unit admission, CLi individual CL, Vci individual Vc, Vpi individual Vp, Qi individual Q, η1 the deviation from the typical parameter in the individual
aCLi = (CL × (BW/92)0.75 × (1 + θcov (eCLcr ≤ 120) × (eCLcr − 94)/1000)) × eη1,
If eCLcr > 120 mL/min: CLi = (CL × (BW/92)0.75 × (1 + θcov (eCLcr ≤ 120) × (120 − 94)/1000)) × eη1
bVci = (Vc × (BW/92)) × eη1 × fCL, Vc, the CV for Vc was derived as CV for CL times the estimated fCL,Vc
cVpi = Vp × (BW/92)
dQi = Q × (BW/92)0.75