Abstract Body

The PROMISE trial found that antiretroviral therapy (ART) in pregnancy reduced mother-to-child transmission, but also increased the frequency of several adverse birth outcomes (ABOs) compared to antenatal zidovudine alone.

PROMISE randomized HIV-infected women ≥14 weeks gestation and not in labor to receive 1 of 3 antenatal regimens: ZDV only (Arm A), ZDV+3TC+LPV/r (Arm B), or TDF+FTC+LPV/r (Arm C). In the early versions of the protocol, women could be randomized only to Arms A and B unless they tested HBsAg+; in version 3.0, women were randomized with equal probability into all 3 arms. We studied the association between antiretroviral regimen and ABOs: delivery <37 weeks (PTD); infant birthweight <2500g (LBW); composite of PTD, LBW, stillbirth (SB), and spontaneous abortion (AB); delivery <34 weeks (VPTD); infant birthweight <1500 g (VLBW); and composite of VPTD, VLBW, SB, and AB. Gestational age at delivery was estimated primarily by Ballard score. We adjusted for baseline factors (maternal age, BMI, HIV viral load, CD4, alcohol use, country, gestational age at entry) and obstetrical complications across all multivariable models. If there were zero counts in specific strata, the variable as a whole was not included in the model. In primary analysis, we included data from all participants; in sensitivity analysis, we restricted data to only those enrolled in version 3.0.

3423 women who enrolled and delivered in PROMISE were allocated to Arm A (n=1507), Arm B (n=1497), or Arm C (n=419). When we considered outcomes with PTD and/or LBW, women on ZDV+3TC+LPV/r (Arm B) and TDF+FTC+LPV/r (Arm C) each had higher risk for ABO compared to ZDV alone (Arm A). When analysis was restricted to severe outcomes (i.e., VPTD, VLBW), the risk associated with Arm C remained elevated. In head-to-head comparisons between the two ART regimens, Arm C had a higher risk of severe ABO such as VPTD (AOR: 2.55, 95%CI:1.46-4.44) and VLBW (AOR: 3.06, 95%CI:1.23-7.59). Findings remained consistent with protocol version 3.0 data alone.

LPV/r-containing ART was associated with a significantly elevated risk of ABO after adjustment for multiple obstetrical and clinical factors. For severe outcomes, this risk was higher among women on TDF-FTC compared to ZDV-3TC. Further study is needed to determine whether this is an independent effect of TDF-FTC, a result of drug-drug interactions with LPV/r, or due to other factors.