About one in four pregnancy-related deaths in Sub-Saharan Africa are attributable to HIV, according to an analysis of demographic and surveillance data collected from 1989 to 2012.1 Seventeen percent of pregnant women in the study areas were infected with HIV, and the mortality rate was substantially higher among these women than among pregnant women without the virus (14 vs. 2 deaths per 1,000 person-years). Overall, about 88% of deaths among pregnant and postpartum women with HIV were attributable to the virus.
In 2011, the United Nations set a goal of reducing the number of HIV-related maternal deaths in low- and middle-income countries by 50% within five years. However, determining the proportion of maternal deaths that are attributable to HIV has been hindered by a lack of empirical data; recent mathematical models and systematic reviews have yielded global estimates ranging from 6% to 25%. To estimate the contribution of HIV to maternal mortality in eastern and southern Africa, where the prevalence of HIV is among the highest in the world, investigators analyzed data from six independently established cohort studies in five countries. Although their protocols differ, the six studies have all collected data on births, deaths, pregnancies and HIV (cases of which were identified both from biological testing and from “verbal autopsy” interviews with family, friends and caregivers). At least 15 years of HIV data were available from the studies in Kisesa, Tanzania (1994–2011); Manicaland, Zimbabwe (1994–2008); Masaka, Uganda (1989–2011); and Rakai, Uganda (1994–2009). About a decade of data were available from Karonga, Malawi (2003–2012), and from uMkhanyakude, South Africa (2003–2011).
The authors examined pregnancy-related mortality, defined as any death occurring during pregnancy or up to 42 days postpartum; because they counted all such deaths, they did not use the term “maternal mortality,” which excludes deaths from accidental or incidental causes. Women were assumed to be at risk for pregnancy and pregnancy- related mortality from ages 15 to 49. Mortality rates were calculated per 1,000 person-years. Women who tested positive for HIV were classified as having become infected halfway between their last negative and first positive tests; if a woman’s positive HIV status had been determined during a verbal autopsy, the researchers assumed she had become infected five years earlier.
Overall, the six studies provided data on about 138,100 women and 636,200 person-years of follow-up, during which 87,000 pregnancies and 235 pregnancy-related deaths were reported. Data on HIV status were available for 321,900 (51%) of the person-years, 49,700 (57%) of the pregnancies and 118 (50%) of the pregnancy-related deaths. The prevalence of HIV among women of childbearing age was 17% for the overall sample; at individual study sites, it ranged from 7% (Kisesa) to 35% (uMkhanyakude). The prevalence was lower among pregnant or postpartum women than among other women (11% vs. 18%).
Of the 118 pregnancy-related deaths among women whose HIV status was known, 60 occurred among women who had the virus. The mortality rate during pregnancy and the postpartum period was 14 per 1,000 person-years among women with HIV and 2 per 1,000 among uninfected women. The crude mortality rate among pregnant or postpartum women with HIV was eight times that among their uninfected counterparts (mortality rate ratio, 8.2); after adjustment for age, the rate ratio was slightly higher (9.0).
Overall, among pregnancy or postpartum women with HIV, about 12 deaths per 1,000 person-years, or about 88% of deaths, were attributable to the virus. Among all pregnant or postpartum women whose HIV status was known, the virus was responsible for 1.3 deaths per 1,000 person-years, or 45% of deaths. By extrapolating from these figures and using United Nations estimates of regional HIV prevalence, the researchers estimate that about 24% of deaths among pregnant or postpartum women in Sub-Saharan Africa can be attributed to HIV.
In general, the impact of HIV on mortality was substantially lower among women who were pregnant or postpartum than among those who were not. One reason, the authors note, is that fertility falls rapidly with duration of HIV infection; thus, pregnant women with HIV tend to have been infected more recently than other women with the virus, and are less likely than those women to die from HIV- related causes.
Limitations of the study, the researchers note, include the relatively small number of pregnancy-related deaths in the study areas, the reliance on verbal autopsies to identify some such deaths and the lack of information on many women’s HIV status. Given the high proportion of pregnancy-related deaths that are attributable to HIV, the authors recommend that HIV care, reproductive health services and safe motherhood programs be integrated, and that safe motherhood interventions strive to prevent not only obstetric causes of deaths but also non-obstetric ones. Ensuring that pregnant women with HIV have access to antiretroviral medication and related care, they add, “should cause HIV-related deaths in women of childbearing age to decline rapidly.”—P. Doskoch
1. Zaba B et al., Effect of HIV infection on pregnancy-related mortality in sub-Saharan Africa: secondary analyses of pooled community-based data from the network for Analysing Longitudinal Population-based HIV/AIDS data on Africa (ALPHA), Lancet, 2013, 381(9879):1763–1771.