Pharmacokinetic-Pharmacogenetic-and-Pharmacodynamic Adherence Relationships in Cohort South African HIV Infected Children on Lopinavir-and Nevirapine-Based Regimens

Abstract

Background: Antiretroviral therapy (ART), notably lopinavir and nevirapine substantially reduces Human immune-deficiency virus (HIV) associated morbidity and mortality in HIVinfected children. Low concentrations of nevirapine and lopinavir have been linked to inferior virological outcomes; it is recommended that lopinavir and nevirapine concentrations are maintained above 1 mg/L and 3 mg/L, respectively, in order to maintain viral suppression. Adherence to both lopinavir and nevirapine ART, respectively has long known to be a crucial contributor to HIV treatment success. Lopinavir and nevirapine pharmacokinetics demonstrate considerable inter-individual variability, which may affect treatment outcomes. At least part of this variability may be explained by host genetic factors. Associations between human genetic variants and exposure to lopinavir and nevirapine are incompletely understood, and have not been studied in a South African paediatric population. Data in this thesis were from a clinical trial conducted at Rahima Moosa Mother and Child Hospital in Johannesburg to assess whether NVP can be re-used (Post-randomization Phase) among 323 children exposed to NVP for PMTCT if they are first suppressed on ritonavir-boosted lopinavir based regimen (Pre-randomization Phase). This thesis assessed the relationship between serial clinic visits lopinavir (Pre-and-Post-randomization) and nevirapine (Postrandomization) concentrations and/or percentage adherence(Pre-and-Post-randomization) and virological outcomes in children. Moreover, population pharmacokinetics models were used to characterise lopinavir and nevirapine parameters. From the final models parameters were derived and were used to assess the relationship between lopinavir and nevirapine pharmacokinetics and genetic polymorphism relevant to both drugs Methods: Cox proportional hazard regression modelling for multiple failure events was used to estimate the crude and adjusted hazard effect of lopinavir (Pre-and Post-randomization) and nevirapine(Post-randomization) concentrations and/or percent adherence(Pre-and Post-randomization) of viral load>400 copies/mL (Pre-randomization) and >50 copies/mL (Post-randomization), respectively. The population means and variances of lopinavir and nevirapine pharmacokinetic parameters at steady state were estimated using non-linear mixed-effects regression. The final models of lopinavir and nevirapine were used to derive individual clearances (CL/F), minimum concentrations (Cmin) and area under the concentration time curves (AUC). The associations between model-derived pharmacokinetic parameters and genotypes in selected genes relevant to lopinavir or nevirapine were explored. Results: In 237 children pre-randomization with viral loads and lopinavir concentrations, the crude and adjusted Cox models revealed significant associations between virologic failure (viral load>400 copies/mL) and both lopinavir plasma concentrations (<1/mg/L) and pretreatment height-for-age z-scores but not percent adherence. In 99 children postrandomization, lopinavir concentrations >1 mg/L reduced the risk of viremia (viral load >50 copies/mL) with about 40%, compared to children with LPV <1 mg/L. No association was found with percent adherence in this group. In 95 children on nevirapine post-randomization, nevirapine concentrations were not significantly associated with increased hazard of viremia (viral load >50 copies/mL). Similarly, there was no significant association with percent adherence in this group. Lopinavir and nevirapine pharmacokinetics were both separately best described with a one compartment models with absorption lag time and transit compartment absorption models, respectively. There was an age driven effect on lopinavir and nevirapine relative bioavailability, respectively. After adjusting for multiple testing, there was no significant association between lopinavir CL/F, Cmin and AUC and genetic polymorphisms in the ABCB1, CYP3A4, CYP3A5 and SLCO1B1. CYP2B6 516G→T and CYP2B6 983T→C were associated with NVP CL/F. CYP2B6 983T→C was associated with NVP Cmin and AUC. Additionally, polymorphisms in the ABCB1 and CYP3A5 were independently associated with NVP CL/F, Cmin and AUC. Conclusions: Lopinavir concentrations <1mg/L were associated with the increased hazard of viremia (viral load >400 copies/mL or >50 copies/mL). The results suggest that lopinavir plasma concentration monitoring at a routine clinic visit may be a useful tool in identifying sub-therapeutic antiretroviral concentrations in children, and this could be used as a guide to therapeutic drug monitoring in children. There was no statistically significant association between polymorphisms in the ABCB1, CYP3A4, CYP3A5 and SLCO1B1 and lopinavir pharmacokinetics. Polymorphisms in the ABCB1, CYP2B6 CYP3A4 and CYP3A5 predicted nevirapine pharmacokinetics.