Risk of HIV Acquisition During Pregnancy and Postpartum: A Call for Action

Human immunodeficiency virus, pregnancy, postpartum, sexual transmission

(See the Major Article by Thomson et al, on pages 16–25.)

Prospective studies exploring the relationship between pregnancy and HIV acquisition in women have had inconsistent results, with some studies documenting an increased risk [1–3], while others finding no increase in risk — or even lower risk — after controlling for sexual behavior and other confounding factors such as sexually transmitted infections [4–7]. A meta-analysis of 19 cohort studies estimated a pooled HIV incidence during pregnancy and postpartum of 3.9 per 100 person-years (4.7 per 100 person-years during pregnancy and 2.9 per 100 person-years postpartum) [8], similar to that defined as “substantial risk” in nonpregnant individuals, such as female sex workers, by the World Health Organization (WHO) [9]. The pooled meta-analysis reported that HIV incidence was not significantly higher among pregnant or postpartum women than among nonpregnant/nonpostpartum women, but only 5 studies had appropriate data and were not able to control for a variety of confounding factors, limiting the power to detect associations.

The study reported by Thomson and colleagues in this issue of the Journal of Infectious Diseases was unique in its ability to control for many factors that other studies of HIV acquisition risk in pregnancy were not able to [10]. The authors have taken advantage of detailed databases from 2 clinical trials (Partners in Prevention Herpes Simplex Virus/HIV Transmission Study and Partners PrEP study) which enrolled serodiscordant couples in 7 high HIV prevalence African countries with median follow-up of 2 years; the studies included detailed standardized collection of monthly reports of sexual behavior, and data on male circumcision, sexually transmitted infections, contraceptive use, partner viral load, and antiretroviral therapy (ART).

An elegant analysis was conducted among 2751 African HIV-seronegative women enrolled in the trials, including 5069 person-years of follow-up and 615 incident pregnancies, to evaluate the risk of HIV acquisition during pregnancy and postpartum, comparing the per coital act HIV acquisition risk in early and late pregnancy (first 13 weeks compared to 14 weeks to gestation end) and through 6 months’ postpartum to nonpregnant/nonpostpartum intervals; visits were censored once the male partner reported initiation of ART, thus the study provides data on HIV acquisition risk in the absence of partner treatment. In addition to HIV antibody testing, the authors had archived quarterly plasma specimens; all seroconversions were tested for HIV RNA and, hence, the authors were able to more definitively determine the timing of acquisition of infection and assign infection events to the stage the woman was in near the time of HIV acquisition (early/late pregnant, postpartum, nonpregnant). Only genetically linked HIV acquisitions were included in the study, allowing linkage to the sexual behavior surveys of the partners. The authors report that HIV acquisition risk per condomless coital act was significantly increased during pregnancy through 6 months’ postpartum (adjusted relative risk [aRR] 2.76; 95% confidence interval [CI], 1.6–4.8) compared to nonpregnant/nonpostpartum time and was highest during late pregnancy (aRR 2.82; 95% CI, 1.3–6.2) and postpartum (aRR 3.97; 95% CI, 1.5–10.5). Additional analyses extending evaluation of the postpartum period through 52 weeks postdelivery found that the increased postpartum risk was concentrated in the first 6 months after delivery. The authors adjusted for time-varying measures of partner viral load and female use of active preexposure prophylaxis (PrEP) based on study arm, as well as age and additional demographic and clinical variables.

There are several important implications of this study. From the pathophysiologic viewpoint, because the study was able to control for major behavioral, demographic, and clinical confounders, these data strongly point toward a biologic association of pregnancy with HIV-acquisition risk. Pregnancy is associated with a complex interaction between sexual hormones and the immune system that modulates the maternal immune response to enable tolerance of the paternally derived fetal alloantigens, while maintaining reactivity against potential pathogens. With advancing pregnancy and increasing estradiol and progesterone levels, there is a decrease in natural killer cells and in the robustness of cytotoxic T-cell responses (T-helper cell type 1 immunity, Th1), with a shift toward the humoral T-cell response (T-helper cell type 2 immunity, Th2), while aspects of innate immunity as alpha-defensin levels and monocyte, dendritic cells, and polymorphonuclear cell activity are enhanced [11]. Pregnancy hormone-associated changes in the female genital tract, including changes in vaginal epithelial thickness, the vaginal and gut microbiome, and an increase in CCR5 coreceptor expression, may create a favorable milieu for HIV acquisition in the female genital tract during pregnancy [11–14]. The finding that the increased risk of HIV acquisition in the postpartum period was confined to the first 6 months postpartum is also consistent with this increased risk being related to biologic factors, as it takes weeks to months following delivery for the immunologic changes of pregnancy to reverse [15, 16].

The second and third trimesters of pregnancy and the early postpartum period through 6 months appeared to the periods of highest risk for HIV acquisition [10]. While there has been significant attention to the need to improve retention of HIV-positive women in care, little attention has been paid toward the need for repeat testing in pregnancy and postpartum following an initial HIV-negative test in early pregnancy. Transmission of HIV from mother-to-child (MTCT) among women who acquire HIV during pregnancy or breastfeeding is double to triple that of women who acquire HIV prior to pregnancy [8]. Elimination of MTCT will not occur without elimination of new infections among women. In a study in South Africa, an estimated 3.3% of mothers who had at least 1 HIV-negative test antenatally in 2011–2012 seroconverted during pregnancy; they accounted for 26% of early MTCT in infants age 4–8 weeks [17]. In a mathematical model of MTCT, the proportion of MTCT attributable to maternal seroconversion during late pregnancy or postpartum in South Africa was estimated to be 34% in 2014 [18].

Programs for prevention of MTCT have almost exclusively been directed toward women who are HIV seropositive at their first antenatal visit. Without incorporating interventions focused on keeping HIV-seronegative pregnant and breastfeeding women negative, control of the pediatric HIV epidemic will be difficult. Although many HIV-endemic countries recommend repeat testing of seronegative pregnant women in the third trimester, this is often not done [19], and few programs include HIV testing during the postpartum breastfeeding period. Male partner involvement in antenatal care and HIV testing of male partners of pregnant women are low and, thus, many women lack knowledge of their partners’ HIV serostatus [20–22].

The most effective way to limit transmission from women who seroconvert during pregnancy or breastfeeding is to prevent them from becoming infected in the first place. The data from Thomson et al point toward the critical need to prioritize interventions targeted to HIV seronegative pregnant and breastfeeding women to prevent HIV acquisition. Given the complexity of the HIV epidemic, this will require a multifaceted approach including structural, behavioral, and biomedical interventions. A community-based combination prevention intervention targeted to HIV-seronegative pregnant women was recently studied in South Africa. The intervention included lay community workers who provided individualized HIV prevention counseling and performed every-3-months home and clinic-based individual and couples’ HIV testing, with referral of male partners for circumcision, sexually transmitted infections, or HIV treatment, as appropriate [23]. The antenatal and postnatal HIV incidence was 1.49 and 1.03 infections per 100 person-years, respectively, substantially lower than historical seroconversion rates or the 4.8 per 100 person-years during pregnancy/postpartum in the meta-analysis [8].

Tenofovir-based preexposure prophylaxis is highly effective in preventing HIV acquisition among adherent women [24]. Although adherence in placebo-controlled clinical trials has been variable, experience outside of the clinical trial setting has been promising. In a study of serodiscordant couples in Kenya and Uganda, 78% of women took ≥6 doses and 88% took ≥4 doses per week while on PrEP, providing an estimated 93% protection against HIV acquisition [25]. Similarly, a clinical trial in South Africa found 75% adherence to a daily PrEP regimen in women, with no seroconversions [26]. Even in the context of a placebo-controlled trial, high medication adherence was observed during the periconception periods in HIV-uninfected women in the Partners PrEP trial, with women who experienced pregnancy taking 97% of prescribed doses of study drug overall [27]. The Partners Demonstration Project in Kenya and Uganda demonstrated that integrated delivery of time-limited PrEP until sustained ART use and viral suppression in African HIV-1–serodiscordant couples was feasible, demonstrated high uptake and adherence, and resulted in near elimination of HIV-1 transmission, with an observed HIV incidence of <0.5% per year compared to an expected incidence of >5% per year [28]. Increasing data on safety of tenofovir in pregnancy and breastfeeding indicate no increase in adverse birth/infant outcomes and minimal penetration of tenofovir into breast milk [29–32].

The article by Thomson and colleagues provides definitive evidence that the risk of HIV acquisition increases during pregnancy and the early postpartum period. These data serve to emphasize that HIV-seronegative pregnant and postpartum women in HIV-endemic areas need to be considered key populations at high risk for HIV acquisition, requiring urgent attention to the development of interventions to detect HIV seroconversion and initiate ART to prevent transmission to their infants and sexual partners and, even more critically, to maintain their HIV-seronegative status.

Notes

Potential conflicts of interest.  Author certifies no potential conflicts of interest. The author has submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.