HIV May Not Increase Risk of Stillbirth Among Women in Sub-Saharan Africa
In Sub-Saharan Africa, a pregnant woman's HIV status is not directly linked to the likelihood that her fetus will be stillborn, data from a randomized trial suggest.1 Instead, inflammation of the amniotic sac during pregnancy, hemorrhage during labor and delivery outside a hospital or clinic are the factors most strongly associated with stillbirth.
The study was a secondary analysis of data from a randomized, placebo-controlled trial that examined whether antibiotics would help prevent mother-to-child HIV transmission associated with chorioamnionitis, a bacterial infection of the amniotic sac. Women were eligible for the trial, which was conducted in 2001–2003 at four clinics in Malawi, Tanzania and Zambia, if they were between 20 and 24 weeks' gestation, had a documented HIV infection, had not recently received antibiotics and had no serious conditions that might complicate pregnancy outcome. Although the goal of the trial was to prevent perinatal HIV transmission, a small proportion of HIV-negative women were recruited at all but one site to help reduce the stigma associated with the trial.
During enrollment, women's demographic characteristics and obstetric and maternal health histories were collected via questionnaire, health care providers tested the women for STIs and other gynecological conditions, and a seven-day course of antibiotics (or placebo) was initiated. Check-ups were performed at 28 and 36 weeks' gestation. When the women went into labor, they received another course of antibiotics or placebo and were examined for signs of chorioamnionitis and other complications. In addition, HIV-positive participants received a dose of an antiretroviral drug prior to labor.
Among the 2,434 participants who gave birth to singleton infants, 80 experienced a fetal death, yielding a rate of 33 stillbirths per 1,000 deliveries. Nearly half (44%) of stillborn infants died before labor, the remainder (56%) during labor. Chorioamnionitis was present in one in three instances of stillbirth.
In univariate analyses, women who were formally employed had an elevated risk of stillbirth (odds ratio, 2.3), as did women who had had a previous stillbirth (2.3). Other factors associated with stillbirth included birth outside a health facility (3.7), hemorrhage during labor (14.4) and development of chorioamnionitis (20.9). Predictably, the risk of stillbirth increased as gestational age at birth decreased: Infants born before 34 weeks' gestation were nearly 23 times as likely to be stillborn as those delivered at 37 weeks or later.
HIV infection was not associated with increased risk of stillbirth in either univariate or multivariate analyses. In a separate univariate analysis of HIV-positive participants, women whose CD4 cell counts were 200 or fewer per microliter (indicating severe immune system impairment) had an increased likelihood of stillbirth; however, this association was not significant in the multivariate analysis.
The lack of an association between HIV and stillbirth contrasts with findings from past studies. The authors suggest that it is not HIV infection itself, but rather HIV-related immunosuppression, that "might account for the adverse pregnancy outcomes associated with the disease." If so, then the relatively healthy status of HIV-positive women in the current study may explain the similarity in outcomes between infected and uninfected women.
The researchers point out that several "simple and effective interventions to improve obstetric care in resource-limited settings" could reduce stillbirth. These measures include fetal monitoring and routine administration of magnesium sulfate. In addition, the fact that disproportionate numbers of deliveries occurring outside a health facility resulted in stillbirth indicates that "increasing the number of trained birth attendants and encouraging institutional delivery could also reduce high stillbirth rates."—H. Ball
1. Chi BH et al., Predictors of stillbirth in Sub-Saharan Africa, Obstetrics & Gynecology, 2007, 110(5):989– 997.