Abstract Body

Background: The 2013 WHO guidelines for prevention of mother-to-child transmission (PMTCT) recommend that countries offering antenatal zidovudine (Option A) transition to 3-drug antiretroviral treatment (ART) for all pregnant women, regardless of CD4 count (Option B or B+). In 2009, Botswana became one of the first African countries to transition from WHO Option A to Option B. Although midwives at antenatal clinics administered Option A, women were referred to ART clinics to start ART under Option B. We evaluated the impact of this transition on projected MTCT. Methodology: From May 2009 to October 2012, covering both the initial pilot in selected regions and the general roll-out of Option B, we abstracted all obstetric records at 6 of the busiest maternity wards in Botswana to determine the PMTCT interventions received in pregnancy and at the time of delivery. We used published estimates for MTCT rates by maternal CD4 strata and interventions received to assign an MTCT prevention score (MPS) to each delivery. Multivariable regression accounting for clustering was used to estimate the impact of Option B on the MPS while adjusting for maternal and facility characteristics. Results: Of 36597 Botswana citizens delivering, 29.7% were HIV infected, 69.1% were HIV-uninfected, and 1.2% had undocumented HIV status. Among 10867 HIV-infected women, 27.2% started ART prior to pregnancy and 72.8% were ART-naive. Of ART-naive women, 65.2% entered antenatal care at a clinic offering Option A and 34.8% at a clinic offering referral for Option B (an onsite ART clinic was available to provide Option B for 65.3% of women). Trends in antiretroviral use and MPS are shown in Figures 1 and 2. In adjusted analyses, registering at a clinic offering referral for Option B was associated with increased ART use (adjusted odds ratio [aOR] 2.68, 95% confidence interval [CI] 2.32 to 3.10), but also an increased likelihood of no antenatal antiretroviral use by the time of delivery (aOR 2.05, 95%CI 1.69 to 2.47). Overall, in the first 3 years of the program, Option B implementation was significantly associated with higher projected MTCT rate, MPS +0.14 (P<0.001). Conclusions: Implementation of Option B in Botswana was associated with an overall increased projected MTCT rate, due to increased numbers of women receiving no antenatal antiretrovirals. Programs starting Option B or B+ should ensure that ART is initiated at antenatal clinics and that zidovudine monotherapy remains an available option in the event of delayed ART initiation.