Background:  Efavirenz (EFV) is commonly used in children over 3 years worldwide, but there is only limited pharmacokinetic (PK) information available in African children.

Methods:  In this study, 41 HIV-infected Ugandan children aged 3 to 12 years on generic EFV plus 3TC+ABC were enrolled in a cross-over PK study of twice to once daily 3TC+ABC 36 weeks after ART initiation in the ARROW trial. Once-daily EFV doses following WHO weight-bands were 200/250/300*/350* mg for children weighing 10 to 15, 15 to 20, 20 to 25, and 25 to 30 kg respectively, using EFV capsules or *halved 600 mg tablets. Intensive plasma PK sampling (t = 0, 1, 2, 4, 6, 8, and 12 hr post observed ingestion) was performed on twice-daily ART at steady state (PK1) and repeated 4 weeks later (PK2, including a further 24 hr sample). EFV daily area under the curve (AUC_0-24) and clearance (CL/kg) were estimated using WinNonlin. Predictors of log₁₀AUC and CL were assessed using multivariate mixed models, fitting random effects for each child.

Results:  During this study, 39 and 37 children had evaluable EFV profiles at PK1 and PK2, respectively. Participant demographic characteristics were as follows: 16/39(41%) children were boys; 18 were aged 3 to 6 years; and 21 were 7 to 12 years. There were 5, 16, 15, and 3 children in the 10 to 15, 15 to 20, 20 to 25, and 25 to 30 kg weight-bands, respectively. The geometric mean (%CV) AUC_0-24 was 50.4 (91.7%) and 54.0 (80.8%) h.mg/L at PK1 and PK2 respectively, with no significant variation across weight-bands (P =0.51). Between- and within- child %CV were 81% and 28% respectively. During the study, 3 sub-populations were identified from normal mixture modeling: 40% children with geometric mean AUC_0-24 27.2 h.mg/L, 32% with 49.9 h.mg/L, and 28% with 137 h.mg/L. Six children at PK1 and 7 at PK2 had sub-therapeutic C_8h and/or C_12h (<1.0 mg/L); 7/39(18%) at either visit. At PK2, 14/37(38%) children had C_24h <1.0 mg/L (median (IQR) [range] 1.1 (0.7 to 2.5) [0.3 to 18.4] mg/L). Nine children at PK1 and 10 at PK2 had C_8h and/or C_12h and/or C_24h >4.0 mg/L; 11/39(28%) at either visit. Overall mean (SD) clearance was 6.8 (3.9) and 6.2 (3.7) L/h at PK1 and PK2 respectively (P =0.04). CL increased by 0.50L/h for every year older (P =0.05), but did not depend on weight (P =0.30), weight-for-age (P =0.56) or height-for-age (P =0.82).

Conclusions:  African children aged 3 to 12 years, on daily EFV using WHO weight-bands, had lower and highly variable EFV PK parameters compared to data from adults. There were no differences across weight-bands, suggesting no major effect of some using half tablets. Increased EFV doses for children should be investigated, but risk increasing the proportion of children with toxic levels further.