Background:  Efavirenz (EFV) is a vital treatment option for HIV/TB co-infection, but pharmacokinetic variability and formulation issues have precluded dosing guidelines for young children. We sought to determine an EFV dose and study its efficacy and toxicity in HIV⁺ children <3 years from TB-endemic countries, and hypothesized that CYP2B6 516 geno (TT = slow and GG /GT = rapid metabolism) would correlate with EFV pharmacokinetics.

Methods:  IMPAACT P1070 is an ongoing, prospective, phase I/II open-label 24-week trial of EFV in children ages 3 to <36 months without TB or with unreported TB. Subjects initiated EFV and 2 NRTI using weight-band EFV doses of ~1600 mg x (weight in kg/70)^0.7 as opened capsules once-daily. Intensive pharmacokinetic (dried blood spot and plasma) samples and dried blood spot CYP2B6 geno were drawn at week 2. Dried blood spot pharmacokinetic analysis was performed within 2 weeks and doses adjusted if outside the target AUC (35 to 180 mg*h/mL.). HIV-1 RNA and toxicity labs were drawn every 4 to 8 weeks.

Results:  We enrolled 26 subjects without TB (17 males, median age 25 months [IQR 14 to 28]) with median follow-up of 24 weeks (IQR 14 to 24); 19 CYP2B6 GG/GT and 7 TT-initiated EFV at a median dose of 40.4 mg/kg (IQR 36.7 to 44.1). The median EFV AUC was significantly higher in subjects with TT geno (490.2, range 324 to 1031 mg*h/mL) than GG/GT (106.3, range 8 to 823 mg*h/mL) (p = 0.0005 by Kruskal-Wallis). Of 26 subjects, 13 met the week 2 AUC target (13 of 19 GG/GT, 0 of 7 TT), 5 were below target (5 of 19 GG/GT, 0 of 7 TT), and 8 were above (1 of 19 GG/GT, 7 of 7 TT). Possibly treatment-related toxicities ≥grade 3 occurred in 2 of 7 TT subjects (1 grade 4 neutropenia and 1 grade 3 hemoglobinemia), and in 1 of 19 GG/GT subject (grade 3 neutropenia) who had excessive EFV levels. All toxicities resolved after discontinuing EFV. Of the total, 10 subjects discontinued EFV before week 24—5 for excessive EFV levels, 1 for grade 4 neutropenia, and 4 for reasons unrelated to treatment. Baseline median RNA (n = 25) was 5.47 log₁₀ copies/mL (IQR 5.05 to 5.88) and 19 to 26 (73% ITT; 95%CI 52% to 88%) achieved either ≥1 log₁₀ drop from baseline RNA or <400 copies/mL (14 of 19) by week 8; all 16 subjects who completed week 24 had RNA <400 copies/mL.

Conclusions:  CYP2B6 geno strongly influences EFV pharmacokinetics and safety early in life. Aggressive dosing (~40 mg/kg) produces therapeutic EFV concentrations in most (68%) children <3 years with GG/GT geno, however, this leads to excessive exposure in those with TT geno. This suggests that optimal use of EFV in children <3years requires pretreatment genotyping, which will be performed for subsequently enrolled subjects.