Background: TB APPRISE (IMPAACT P1078), a Phase IV randomized, multi-country non-inferiority trial assessing the safety of 28 weeks of isoniazid preventive therapy (IPT) initiated during pregnancy (immediate IPT) versus deferring to week 12 postpartum (deferred IPT) in people living with HIV on antiretroviral therapy, showed higher than expected hepatotoxicity. We investigated the potential roles of antiretrovirals, isoniazid, pharmacogenetics and other factors.

Methods: Hepatotoxicity was defined as Grade≥3 liver enzyme elevations; or Grade≥2 enzyme elevations with elevated bilirubin or symptomatic hepatitis. We performed Poisson regression of all-cause hepatotoxicity on study arm, antiretroviral regimen, pharmacogenetics of isoniazid and efavirenz metabolism (NAT2, CYP2B6) and other participant characteristics. Adjusted models included study arm and covariates with p<0.25 in unadjusted models. Antiretroviral regimen and pharmacogenetics interactions with study arm were evaluated.

Results: All 945 pregnant participants with follow-up liver function measurements were on antiretrovirals (85% with efavirenz, 13% with nevirapine); 63 (6%) experienced hepatotoxicity events; 29 (6%) in immediate and 34 (7%) in deferred arm; only 5 events (8%) occurred in pregnancy; 49 (78%) occurred between delivery and 24 weeks postpartum. Higher risk of hepatotoxicity was observed with nevirapine use in the immediate arm, but there was no difference by study arm in participants on efavirenz. Slow efavirenz metabolizers had increased risk of hepatotoxicity.

Conclusions: It is critical to monitor for hepatotoxicity in early postpartum, where there is higher risk compared to antepartum. ARV regimen and pharmacogenetics should also be considered in making decisions on when to initiate IPT in pregnant and postpartum populations.