Background:  Premature birth is a risk factor for perinatal HIV transmission and other neonatal morbidity and mortality. Given that HIV prevalence is high in many resource- limited settings, and many HIV⁺ pregnant women may be on antiretrovirals (ARV) for prevention of mother-to-child transmission (PMTCT), assessing whether there is an association between use of ARV and premature birth or other adverse fetal outcomes is important. Data are conflicting on this issue. Our objective was to compare the rates of premature births and low birth weights among ARV-naïve HIV⁺ pregnant women exposed to 2 different ARV regimens given from 34 weeks’ gestation through 6 months post partum in the Kisumu Breastfeeding Study; a PMTCT trial initially using zidovudine/lamivudine/nevirapine (ZDV/3TC/NVP), but later revised:  those with CD4 <250 cells/ mm³ got ZDV/3TC/NVP and those with CD4 ≥250 cells/mm³ got ZDV/3TC and nelfinavir (NFV).   

Methods:  We determined gestational age using last menstrual period dates and encouraged all participants to deliver at the local hospital where infant weight could be obtained. We determined the duration of ARV exposure and the rates of preterm delivery (<37 weeks) and low birth weight (<2500 g) for different ARV regimens at delivery among participants who had a baseline CD4 ≥250 cells/mm³.

Results:  There were 403 women with CD4 ≥250 cells/mm³ who received ARV from 34 weeks’ gestation; 196(48%) received a NVP-based regimen, 207 (52%) received a NFV-based regimen. After 20 women withdrew before delivery, 383 participants remained to deliver 391 infants of which 384 (98%) were live births and 7 (2%) still births. The median duration of antenatal exposure to NVP (5.1 weeks, range 0.29 to 11.6) was significantly lower than NFV (6.0 weeks, range 0.57 to 12.2; p = 0.008). Of the live births, 65 (17%) were preterm and 51 (13.3%) were of low birth weight. After controlling for baseline maternal viral load and duration on ARV before delivery, the rate of preterm deliveries was not significantly different:  19% vs 14% for those on NVP and NFV, respectively (p = 0.08). Similarly, the rate of low birth weight was also not significantly different:  14% vs 12% for NVP and NFV, respectively (p = 0.7).

Conclusions:  We show that ZDV/3TC/NVP or ZDV/3TC/NFV taken by HIV⁺ pregnant women with CD4 ≥250 cells/mm³ in this study site after 34 weeks’ gestation are associated with similar risk of preterm delivery and low birth weight. Assessment of other safety features and tolerance should also be used to guide choice of NVP or NFV in PMTCT ARV regimens.