Population pharmacokinetic and pharmacodynamics analysis of lipid emulsion propofol in pediatric patients

Aim: This study aimed to characterize the pharmacokinetics and pharmacodynamics of lipid emulsion propofol administered by a target-controlled infusion (TCI) anesthesia in pediatric surgery.

Method: Forty patients (ASA PS 1,2) aged 2–12 years were given an intravenous bolus of 2propofol (Fresofol?Fresenius Kabi Korea Ltd., Korea) 3 mg/kg, followed by continuous infusion at the rate of 200 mg/kg/min for variable periods. Arterial concentrations of propofol were measured at preset intervals and bispectral index (BIS) values were recorded throughout the study period. Pharmacokinetic and pharmacodynamic characteristics were evaluated using a population analysis with nonlinear mixed effects modeling.

Results: Pharmacokinetics and pharmacodynamics of propofol in children were best described by a two compartment model and inhibitory sigmoid Emax with effect-compartment model, respectively. The pharmacokinetic and pharmacodynamic models were directly connected by effect compartment. The final pharmacokinetic model was constructed by using coefficient and exponential terms. Population parameter estimates, inter-individual variability, and median parameter values (2.5–97.5%) of the non-parametric bootstrap replicates of the final pharmacokinetic and pharmacodynamic models are shown in table 1 and 2, respectively. In pediatric pharmacokinetic model, lean body mass (JLBM), body fat (BF), and height (HT) were included as a significant covariate for λ1 and λ2. In pharmacodynamic model, age and remifentanil dose (TREMI) were significant covariates for Ce50. The blood-brain equilibration half-time was 1.43 min for BIS.

Conclusion: In children, propofol pharmacokinetics was influenced by lean body mass, body fat, and height. With increasing age or total dose of remifentanil, Ce50of propofol was decreased.

Table 1. Population pharmacokinetic parameter estimates, inter-individual variability, and median parameter values (2.5 – 97.5%) of the non-parametric bootstrap replicates of the final pharmacokinetic model of propofol (n=39)

 

Parameter

Estimate (RSE, %CV)

Median, 2.5 – 97.5%

C1

0.35 (21.23, 65.04)

0.38, 0.19 – 1.15

C2

0.016 (7.45, 28.65)

0.016, 0.012 – 0.020

??3

0.89 (7.30, –)

0.95, 0.53 – 2.91

??7

0.51 (23.93, –)

0.51, 0.18 – 0.74

??8

0.66 (38.33, –)

0.65, 0.24 – 1.04

??4

0.023 (4.80, –)

0.023, 0.020 – 0.026

??9

2.23 (22.96, –)

2.25, 8.64 × 10-12 – 4.18

s2

0.29 (4.43, –)

0.28, 0.23 – 0.35

Inter-individual and residual random variability were modeled using a log-normal model and a constant coefficient of variation (CV) model, respectively.Nonparametric bootstrap analysis was repeated 2000 times. s2: variance of residual random variability, RSE: relative standard error = SE estimate1 ´ 100 (%). JLBM: lean body mass; BF: body fat; HT: height.

Table 2. Population pharmacodynamic parameter estimates, inter-individual variability, and median parameter values (2.5 – 97.5%) of the non-parametric bootstrap replicates of the final pharmacokinetic model of propofol (n=39)

Parameter

Estimate (RSE, %CV)

Median, 2.5–97.5%

E0

80.7 (2.2924, –)

81.2, 77.2–84.645

Emax

29.1 (9.5189, 29.95)

28.4, 20–37.345

Ce50 (ug/ml) =??3−??6×(age)−??7×(TREMI )

??3

4.27 (13.4895, 29.85)

4.19, 3.03–5

??6

0.121 (57.1074, –)

0.12, 0.1–0.2

??7

2.61 (44.4444, –)

2.35, 0.01–5

??

2.14 (8.6449, –)

2.1, 1.46–3

ke0 (/min)

0.485 (0.0079, 93.381)

0.469, 0.337–0.6

s2

68.6 (1.5306, –)

11.6, 0–37.665

Inter-individual and residual random variability were modeled using a log-normal model and a constant coefficient of variation (CV) model, respectively. Nonparametric bootstrap analysis was repeated 2000 times. s2: variance of residual random variability, RSE: relative standard error = SE estimate–1 ´ 100 (%). E0: baseline effect; Emax: maximum effect; TREMI: remifentanil dose administered.