Study setting and study population
The Centre for the AIDS Programme of Research in South Africa (CAPRISA) 002 study is a prospective cohort study which aimed to advance the understanding of HIV-1 subtype C acquisition, pathogenesis and clinical disease progression. The study was initiated in August 2004 and HIV negative women at high risk of HIV acquisition were enrolled at an urban clinical research site in eThekwini, Durban and a rural site in Vulindlela, KwaZulu-Natal, South Africa. The study recruitment procedures and eligibility criteria have been described previously [16]. Briefly, women 18 years and older self-identifying as sex workers or having had at least three partners in the 3 months prior to recruitment, and testing HIV negative were eligible for study participation [17]. Pregnant women were not eligible for enrolment into the HIV negative phase of the study due to the short follow-up period and nature of employment [16]. In the HIV negative study phase, women were followed up monthly for a maximum of 24 months or until diagnosis of HIV infection, at which time they were offered enrolment into the acute HIV infection phase of the study and followed-up during established HIV infection until they were ART requiring [16]. Acute HIV infection is defined as 0–3 months post-diagnosis of HIV infection, early HIV infection as ≥ 3–12 months post-diagnosis of HIV infection and established HIV infection as > 12 months post-HIV diagnosis, as per study protocol. In addition, CAPRISA undertakes observational studies in preparation for clinical trials in HIV prevention including microbicide and pre-exposure prophylaxis trials [18, 19]. Women acquiring HIV in any of these studies were also offered enrolment and long-term follow-up in the CAPRISA 002 study (Fig. 1). HIV infection as endpoint was based on a positive HIV-1 antibody test with a previously documented negative HIV-1 antibody test within 5 months; or the presence of HIV-1 RNA in the absence of HIV antibodies [16, 17]. Estimated time of HIV infection was defined as the midpoint between the last negative HIV ELISA and the first positive HIV ELISA or 14 days prior to a positive HIV-1 RNA test if the HIV ELISA is negative on the same day [16].
Once initiated on ART, participants have been in follow-up for up to 6 years [median 3.7 years, Interquartile range (IQR) 1.8–5.0] and remain in follow-up [20]. A subset of women were initiated on treatment in the CAPRISA 009 study, a randomised controlled trial which assessed ART outcomes in women exposed to tenofovir gel, when eligible for ART [21]. These women were transitioned back into the treatment phase of the CAPRISA 002 study once their participation in the CAPRISA 009 study ended.
Clinical evaluation
All participants provided written informed consent upon entry into the CAPRISA 002 study. Participants underwent clinical evaluation and had peripheral blood and urine samples collected at each scheduled study visit. Pregnancy status was determined by on-site urine pregnancy testing (Quickview One-Step hCG Combo Test, Quidel Corporation, San Diego, CA, USA) with minimum detection limit of 25mIU/mL or other standard urine pregnancy test, as available. Pregnancy testing was performed by trained nurses at entry into each study phase and on clinical suspicion of pregnancy at any study visit [16]. Estimated dates of conception, gestational age and pregnancy delivery were calculated from the date of last menstrual period. As standard of care all participants were offered contraception counselling and provision during the study. Participants who became pregnant were referred for antenatal care through the preferred public or private sector primary health care provider.
Based on the contemporaneous South African national treatment guidelines, ART was initiated at a CD4 count eligibility threshold applicable at the time or the presence of an AIDS-defining illness or pregnancy [20]. Pregnant women, if not eligible for combination ART at the time of their pregnancy, were offered a PMTCT single or dual drug regimen as standard of care. Prior to 2010 in South Africa, the ART initiation criterion was CD4 count threshold < 200 cells/µl or World Health Organisation (WHO) stage 4 and single dose nevirapine prophylaxis for women with a CD4 count above this threshold [6]. In 2010, WHO Option A was introduced, lowering the ART initiation threshold to CD4 count of < 350 cells/µl and adding zidovudine PMTCT prophylaxis from 14 weeks of gestation [6]. In 2013, Option B was implemented which included Option A but added pregnant women to initiate ART irrespective of CD4 count, to continue with ART throughout pregnancy and delivery until one-week post cessation of breastfeeding [6]. In 2015, Option B + was introduced which recommended the initiation of ART at any CD4 count and more importantly to continue on lifelong ART [6]. Pregnant women in the CAPRISA 009 study (2011 – 2014) were initiated on ART irrespective of CD4 count. Combination ART was provided to all pregnant women at the research clinic, while PMTCT prophylaxis regimens were accessed through the antenatal care facilities. In this analysis viral suppression is defined as HIV viral load < 400 copies/ml.
Laboratory evaluation
HIV infection was diagnosed using rapid tests (Determine: Abbott Laboratories, Tokyo, Japan and Capillus; Trinity Biotech, Jamestown, NY, USA) with confirmatory HIV enzyme immunoassay (BEP 2000; Dade Behring, Marburg, Germany) used for discordant antibody tests [16]. All participants had full haematological assessments including CD4 count (FACSCalibur Flow Cytometer or TruCOUNT, BD biosciences, San Jose, CA, USA) and HIV viral load (Roche Cobas Amplicor v.1.5, Taqman version 1.0 or version 2.0, Roche Diagnostics, Basel, Switzerland) monitoring was done at three to six-monthly intervals. STI screening, including testing for Neisseria gonorrhoeae, Chlamydia trachomatis, Trichomonas vaginalis, Mycoplasma genitalium, Herpes Simplex Virus (HSV) type 2 and bacterial vaginosis, was performed at the same frequency as previously described [22]. Testing for syphilis was provided by the primary health care provider.
Data management
All participant clinical, on-site testing and laboratory results were documented in study source notes and or case report form (CRF) by the study nurse or study clinician. A Pregnancy Outcome CRF was completed following face to face interview with a participant once a pregnancy outcome was reached. Pregnancy and laboratory data were captured onto standardised CRFs identifiable only by a participant identification number to maintain participant confidentiality. CRFs were faxed using the DataFax system (Clinical DataFax Sytems Inc., Ontario, Canada), with all data verified by data encoders for quality checks and stored in a secure study specific database.
Statistical analysis
Baseline and follow-up characteristics were summarized as either medians with interquartile range or means with standard deviation for continuous variables and proportions for categorical variables. Pregnancy incidence rates were calculated for the HIV negative study phase, acute, early and established HIV infection and post ART initiation separately. In women with more than one pregnancy in the same study phase, only the first pregnancy was included in the incidence calculation and Kaplan–Meier curve. However, if the same woman had pregnancies in different phases, all pregnancies were included in the survival analyses. For pregnancy outcomes, all pregnancies were counted including repeat pregnancies in each study phase.
Prior to HIV seroconversion, pregnancy-free survival time was calculated from enrollment to time of conception, last visit date for those who were terminated early without seroconverting, last date prior to estimated date of HIV infection (for those who seroconverted) or at 24 months post enrollment for those who completed the HIV negative phase without seroconverting. Whereas, after HIV acquisition, pregnancy-free survival time was calculated from estimated date of HIV infection until either time of conception or last visit date prior to ART initiation. Moreover, in the ART phase it was calculated from ART initiation date to time of conception, last visit date prior to or at 5-year post ART initiation (administrative censoring period). We used Poisson approximations to calculate 95% confidence interval (CI) for pregnancy incidence rates. Kaplan–Meier cumulative probability of falling pregnant was calculated separately for each phase. CD4 count and HIV viral load measurements close to the time of the delivery were compared over time using linear regression to assess for linear trend.
Ethics approval and consent to participate
The CAPRISA 002 study was reviewed and approved by the University of KwaZulu-Natal Biomedical Research Ethics Committee (E013/04). All methods were carried out in accordance with relevant guidelines and regulations. All study participants provided written informed consent at study entry.