Table 2. Population parameter estimates from the final lumefantrine pharmacokinetic model.

Parameter	Fixed effects		Random effects
	Population estimate (BS estimate)	%RSE (95% CI)	%CV for IIV (BS estimate)	%RSE (95% CI)
F	1 (fixed)	—	70.3 (70.2)	5.95 (65.3–75.3)
Box–Cox shape parameter on F	−0.343 (−0.342)	19.5 (−0.469 to 0.215)	—	—
ka (h−1)	0.0386 (0.0388)	2.72 (0.0368 to 0.0410)	—	—
CL/F (l/h)	1.35 (1.36)	29.7 (0.538 to 2.19)	—	—
VC/F (l)	11.2 (11.6)	30.3 (4.65 to 18.8)	144 (141)	10.8 (115–165)
Q/F (l/h)	0.344 (0.350)	29.8 (0.137 to 0.566)	—	—
VP/F (l)	59.0 (59.5)	29.7 (23.6 to 96.4)	—	—
Dose50 (mg/kg) on F	3.86 (4.07)	41.5 (1.25 to 8.04)	—	—
Pregnancy on ka	0.352 (0.355)	21.2 (0.212 to 0.510)	—	—
Parasitaemia on F	−0.643 (−0.635)	13.0 (−0.793 to 0.473)	—	—
σ	0.323 (0.323)	4.88 (0.293 to 0.357)	—	—

Coefficient of variation (%CV) for inter-individual variability (IIV) was calculated as $100\times \sqrt{{\mathrm{e}}^{\mathrm{e}\mathrm{s}\mathrm{t}\mathrm{i}\mathrm{m}\mathrm{a}\mathrm{t}\mathrm{e}}-1}$. The relative standard error (%RSE) was calculated as $100\times \frac{\mathrm{S}\mathrm{t}\mathrm{a}\mathrm{n}\mathrm{d}\mathrm{a}\mathrm{r}\mathrm{d}\mathrm{d}\mathrm{e}\mathrm{v}\mathrm{i}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}}{\mathrm{A}\mathrm{v}\mathrm{e}\mathrm{r}\mathrm{a}\mathrm{g}\mathrm{e}\mathrm{p}\mathrm{a}\mathrm{r}\mathrm{a}\mathrm{m}\mathrm{e}\mathrm{t}\mathrm{e}\mathrm{r}\mathrm{e}\mathrm{s}\mathrm{t}\mathrm{i}\mathrm{m}\mathrm{a}\mathrm{t}\mathrm{e}}$ from 1,000 iterations of a non-parametric bootstrap (BS) procedure. The 95% confidence interval (95% CI) is displayed as the 2.5th to 97.5th percentile of the BS estimate, and the BS estimate as the average value.

F: relative bioavailability; Box–Cox shape parameter: shape parameter on Box–Cox transformation; k_a: absorption rate constant; CL/F: elimination clearance; V_C/F: apparent central volume of distribution; Q/F: inter-compartmental clearance; V_P/F: apparent peripheral volume of distribution; dose₅₀: dose (mg/kg) needed for half of maximum dose-dependent saturation of F; pregnancy: categorical covariate effect of pregnancy on k_a; parasitaemia: continuous covariate effect of enrolment parasite density on F; and σ: additive residual error on log scale. All parameters were centred on a non-pregnant patient weighing 42 kg with an admission parasitaemia of 15,800 parasites/μl.

Clearance and volume parameters were centred on the median body weight (WT) and scaled allometrically ($\mathrm{C}\mathrm{L}\mathrm{a}\mathrm{n}\mathrm{d}Q=\mathrm{\theta }\left(n\right)\times {\frac{\mathrm{W}\mathrm{T}}{42}}^{\frac{3}{4}},V=\mathrm{\theta }\left(n\right)\times \frac{\mathrm{W}\mathrm{T}}{42}$); dose-dependent absorption was implemented as a saturation model ($F=\mathrm{\theta }\left(n\right)\times (1-\frac{\mathrm{D}\mathrm{o}\mathrm{s}\mathrm{a}\mathrm{g}\mathrm{e}}{{\mathrm{\theta }}_{{\mathrm{D}\mathrm{o}\mathrm{s}\mathrm{a}\mathrm{g}\mathrm{e}}_{50}}+\mathrm{D}\mathrm{o}\mathrm{s}\mathrm{a}\mathrm{g}\mathrm{e}})$); baseline parasitaemia was implemented as an exponential relationship centred on the median natural logarithm transformed value ($F=\mathrm{\theta }\left(n\right)\times \left({\mathrm{e}}^{{\mathrm{\theta }}_{\mathrm{p}\mathrm{a}\mathrm{r}\mathrm{a}\mathrm{s}\mathrm{i}\mathrm{t}\mathrm{a}\mathrm{e}\mathrm{m}\mathrm{i}\mathrm{a}}-\left(\mathrm{p}\mathrm{a}\mathrm{r}\mathrm{a}\mathrm{s}\mathrm{i}\mathrm{t}\mathrm{a}\mathrm{e}\mathrm{m}\mathrm{i}\mathrm{a}-4.2\right)}\right)$), and the categorical pregnancy effect was implemented as a proportional effect (k_a = θ(n) × (1 + θ_pregnancy)).