Interpreting antiretroviral exposure and drug–drug interactions in pregnancy: A clinically anchored framework for HIV pharmacology

In pregnancy, antiretroviral drug–drug interactions must be interpreted against dynamic changes in gastrointestinal physiology, plasma‐protein binding, renal elimination, hepatic enzyme activity and membrane transporter expression. Interaction studies quantify geometric mean exposure ratios relative to nonpregnant adults, but these measures describe relative pharmacokinetic change rather than whether effective exposure is preserved. For HIV treatment, validated exposure–response targets remain reasonable starting anchors when pregnancy‐specific targets are unavailable. No evidence indicates that gestational immune adaptation alone requires a different intrinsic antiviral target. Pregnancy alters target attainment by altering drug disposition and variability, and it can modify the pharmacokinetics of both the victim antiretroviral and the perpetrator drug. For long‐acting injectable PrEP, protective exposure, dosing interval and pharmacokinetic tail require separate consideration from treatment targets. This review evaluates candidate evidentiary anchors, including protein‐adjusted in vitro inhibitory concentrations, exposure–response relationships, therapeutic drug‐monitoring targets and model‐informed probability of target attainment. Evidence from integrase inhibitors, protease inhibitors, non‐nucleoside reverse transcriptase inhibitors, tenofovir formulations and long‐acting agents demonstrates that similar proportional reductions in exposure can have different clinical implications depending on therapeutic margin, resistance barrier, interaction mechanism, formulation and pharmacologically relevant matrix. A newly proposed six‐domain framework integrates clinical context, active analyte, exposure metric, evidentiary anchor, distributional target attainment and clinical outcomes. Physiologically based and population pharmacokinetic models provide evidence when pregnancy‐specific data are unavailable, but their use should reflect model verification and uncertainty. The central objective is therefore not to define a universal pregnancy threshold but to determine whether pharmacologically effective exposure is maintained in the setting of gestational physiology and c oncomitant medicines.