Maternal mortality ratio and deaths beyond 42 days up to 5-years post pregnancy in women living with HIV on life-long antiretroviral therapy in a resource limited setting.

backgroundMaternal mortality ratio (MMR) remains poorly described in women living with HIV (WLWH), more so during this era of lifelong antiretroviral therapy (ART).

methodsIn a case-control cohort study design with HIV as the exposure variable, we determined MMRs from ≥ 20 weeks' gestational age up to 42-days postpartum (PP) in WLWH and their peers living without HIV (WLWoH). Furthermore, mortality was assessed 43-days PP up to 5-years following delivery of the index baby. Study midwives administered socio-demographic, HIV-related, household water and sanitation questionnaires in pregnancy. At follow up visits, documented live births, maternal deaths, signs and symptoms of illnesses at death or verbal autopsy, including the place of death as per the Medical Research Council of Zimbabwe' Serious Adverse Event Reporting requirements. In R-statistical package, using univariate and multivariate automated variable elimination (AVE) models, we assessed the risk factors associated with maternal deaths.

resultsSix-hundred (600) WLWoH, and 608 WLWH mainly on Tenofovir/Lamivudine/Efavirenz (89.3%), ART-naïve (10.4%) and 0.3% ART-defaulters were enrolled. Eighteen maternal deaths were recorded over the 5-years with 56% occurring at home. Three pregnancy-related fatalities were recorded, all 3 occurring in WLWH; one case each in pregnancy (sepsis), childbirth (obstetric haemorrhage) and within 42-days PP (sepsis); giving a cohort MMR of 3/1202; 249.6 (51.5-727.6)/100000 live births, being 3/602;498.3 (102.9-1449.4)/100000 live births in WLWH. No deaths were recorded in WLWoH within 6-weeks PP while one death was observed well beyond six weeks post-delivery. WLWH were more likely to die compared to WLWoH, odds ratio (OR) 17 (2.29-129.9), p=0.006. Compared with their living peers, mortality in WLWH was associated with a new diagnosis of HIV infection, hence not on ART, p=0.027, and non-disclosure of HIV sero-status, p=0.014. In AVE regression models, WLWH with plasma HIV-RNA load of more than 1000 copies per mL were more likely to die compared to their alive HIV-RNA suppressed peers, OR 5 (1.2-24.8), p=0.03.

conclusionWLWH were 17 times more likely to die when compared to WLWoH. Non-disclosure of HIV sero-status was associated with the increased odds of dying, probably due to poor adherence leading to suboptimal ART-exposure. In addition, a new HIV diagnosis, hence being ART-naïve was also a significant contributing factor. There is need to address barriers hindering HIV counselling and testing, treatment, adherence and continuation to suppress HIV-RNA load to reduce MMR in WLWH on lifelong ART.

trial registrationwww. CLINICALTRIALS: gov,trialregistrationnumber:NCT04087239 , Registered 12 September 2019.