Each year an estimated 529 000 maternal deaths occur globally, the majority in sub-Saharan Africa . Human immunodeficiency virus (HIV) is an indirect cause of maternal mortality, and although its contribution has not been accurately quantified , it has been reported to increase both pregnancy-related deaths  and deaths in women of child-bearing age in populations with a high HIV prevalence [3,4].
Women remain at risk beyond the conventional period of maternal mortality, that is 42 days after delivery [1,5]. In a large cohort of Zimbabwean women, mortality at 24 months after delivery was significantly higher in HIV-infected compared to HIV-uninfected women, with tuberculosis being the most common cause of death . Findings from the same study also showed that mortality amongst HIV-infected women with CD4 cell counts below 200 cells/μl was 54 times higher than the HIV-negative reference group, and 5.4 times higher compared to those with CD4 cell counts between 400 and 600 cells/μl .
In addition to increased mortality, HIV-infected women are more likely to experience higher morbidity within 2 years of delivery compared to HIV-uninfected women [6,8–10], including treatable conditions such as tuberculosis, pneumonia, and diarrhea. Although it has been suggested that breastfeeding by HIV-infected women may be detrimental to their health , these results have not been confirmed by other African studies [12–14].
There are a lack of data on the natural history of HIV disease after delivery in African women, with appropriate comparisons to background morbidity and mortality in HIV-uninfected women. In this study we describe morbidity and mortality in a large cohort of HIV-infected and uninfected African women, followed at regular intervals for a 2-year period postnatally, at a time when antiretroviral treatment (ART) was not available in the study area. Further, we quantify mortality stratified by maternal CD4 cell count, and investigate the impact of breastfeeding on maternal mortality.
Data were collected as part of the Vertical Transmission Study (VTS), a nonrandomized intervention cohort study, in KwaZulu-Natal, South Africa [15–17]. Pregnant women were recruited from nine, mainly rural, primary healthcare clinics, from 2001 to 2004. As previously reported, initially all HIV-infected women and a systematic subsample of uninfected women were enrolled, but for the last year of enrolment (2003–2004) all women were offered enrolment .
Delivery data were documented either from the woman directly or from the clinic registers. Prior to delivery, women were visited at home, by lay counsellors; socio-demographic data were collected, and women counselled on infant feeding options. At the time of the study highly active antiretroviral therapy (HAART) to delay disease progression was not available through the public health services and only minimally through private sources, but all HIV-infected women and their infants were given single-dose nevirapine to reduce the risk of mother-to-child transmission (MTCT) . All women received standard antenatal care mostly provided by professional midwives, and, as previously reported, approximately 90% of women had a vaginal delivery , and only 12% of women delivered at home .
After delivery, women were assessed by study nurses, monthly for the first 9 months and 3-monthly from 10 to 24 months. At each clinic visit, a medical history was taken, using a standardized questionnaire, and women were examined, documenting any current illness (using a predefined set of codes), weighed and had their mid-upper arm circumference (MUAC) measured. Women who chose to breastfeed were well supported by lay counsellors who conducted home visits at least fortnightly for the first 6 months after delivery; those who chose to formula-feed were supported by clinic staff and an infant feeding specialist . Infant feeding data were collected by field monitors who visited each mother weekly at her home, and documented all feeds and fluids given to the infant for every day of the preceding week .
CD4 cell counts and viral loads were determined in all HIV-infected women at their first antenatal booking visit, and in all women at around 26 weeks after delivery. No other laboratory tests were routinely carried out on women postnatally.
Each woman gave written informed consent; the study was approved by the Biomedical Research Ethics Committee of the University of KwaZulu-Natal, South Africa.
Infant feeding information was categorized into five groups: ‘exclusive breastfeeding (EBF)’ when the mother reported only breastfeeding with no other types of feeding during a specific time period; ‘mixed feeding (breastfeeding 75% or more)’ when a mother gave breast milk for 75% or more of the time in addition to other forms of feeding; ‘mixed feeding (breastfeeding <75%)’ when a mother gave breast milk for less than 75% of the time in addition to other forms of feeding; ‘exclusive formula feeding’ (EFF) when a mother only fed using commercial formula milk; and ‘other’ when a mother used none of the above feeding methods – this category included giving formula milk with solids and water. The ‘EFF’ and ‘other’ groups were also classified as nonbreastfeeders. The mixed feeding group was divided into breastfeeding at least 75% or breastfeeding below 75% to capture the intensity of breastfeeding; previous analyses showed that most women exclusively breastfed, and over half of those who interrupted exclusive breastfeeding in the first 6 months of life returned to giving only breast milk within a median of 2 days .
Women enrolled in the VTS were excluded from this analysis if they had an unknown HIV status, had a stillbirth or spontaneous abortion, withdrew or moved in the immediate postnatal period, those whose infant died in week 1 and women with unknown feeding data in the first 2 days (Fig. 1).
Maternal morbidity was coded at each clinic visit by the study nurse using a system which grouped clinical diagnoses by systems (e.g. gastro-intestinal, respiratory). For these analyses, serious morbidity was defined as any of the following events: chronic diarrhoea more than 14 days, bloody diarrhoea, liver abscess, oral candidiasis, convulsions, meningitis, septic arthritis, chronic lung disease, pneumonia, pulmonary tuberculosis (TB), abscess (not of the breast), cellulitis, chicken pox, shingles, and chronic pyrexia.
The basic characteristics of HIV-infected and uninfected women were compared using chi-squared tests for comparison of proportions and Mann–Whitney rank-sum tests for comparison of population medians. Univariable and multivariable Cox proportional hazard regression assessed the factors associated with experiencing a serious morbidity event after delivery. Statistical analyses were carried out using STATA software version 10 (Stata Corporation, College Station, Texas, USA).
Maternal and infant characteristics
Information was analysed for 2624 women and 2670 children (46 sets of twins); approximately half of the women, 1326 (50.5%), were HIV-infected (Table 1). Approximately 50% of children of both HIV-infected and uninfected mothers were male and median birth weight was similar in infants of infected and uninfected mothers: 3000 g (range 1100–4700 g) and 3100 g (range 800–5500 g), respectively.
Maternal immunological status
Antenatally, this information was available for HIV-infected women only, with only 11% (139/1252) women having a CD4 cell count less than 200 cells/μl (Table 2). Median postnatal CD4 cell counts were significantly higher in HIV-uninfected than HIV-infected women (Mann–Whitney statistic 27.52, P < 0.001), with only 2.6% of 881 HIV-uninfected women having a postnatal CD4 cell count below 500 cells/μl (Table 2).
Among women and children who survived the first 3 months after delivery, 40.7% (1037/2545) of women exclusively breastfed for the first 3 months; this proportion did not differ by mother's HIV infection status. More HIV-infected than uninfected women (8.1 vs. 1.5%) practised ‘other’ feeding methods (the category which includes no breast milk, and nonexclusive formula feeding) during the first 3 months after delivery (Table 3).
Categorizing type of infant feeding into either intensive (exclusive breastfeeding or mixed feeding with breastfeeding 75% or more of the time) and less intensive (all other feeding categories, including mixed feeding with breastfeeding less than 75% of the time and exclusive formula feeding), the overall proportion of women intensively breastfeeding in the first 3 months after delivery was 72.6% (1844/2539). Among HIV-uninfected and HIV-infected women, 77.2 and 67.6% intensively breastfed for the first 3 months, respectively (P < 0.001) and similarly between 4 and 6 months significantly more HIV-uninfected than HIV-infected women intensively breastfed (69.9 vs. 58.3%, P < 0.001). Overall, during the first 6 months, significantly more HIV-uninfected women than HIV-infected women intensively breastfed (73.8 vs. 63.4%, respectively) (Table 3).
Sixty (2%) women died postnatally at a median time after delivery of 8.5 months (range 1 day–22.8 months). The overall postnatal mortality rate was 30.72 per 1000 years of follow-up. Among HIV-uninfected women the mortality rate was 8.6 deaths per 1000 person-years, among HIV-infected women it was 55.68 deaths per 1000 person-years. Mothers who died postnatally were significantly more likely to be HIV-infected (χ2 = 32.08, P < 0.001); older at delivery (Mann–Whitney = -3.3, P = 0.001), with median age of 26 vs. 24 years among those who survived; less heavy 6 weeks after delivery (Mann–Whitney = 4.3, P ≤ 0.001), with median weight among those who died of 55.4 vs. 62.5 kg among those who survived; less well educated (no education compared to some primary, secondary or tertiary education) (χ2 = 12.05, P = 0.002); have a poorer water source (borehole, tank, well or river/stream compared to pipe water) (χ2 = 14.81, P = 0.001) and were more likely to live in a peri-urban than urban setting (χ2 = 6.81, P = 0.009), than women who survived.
In multivariable regression analysis being HIV-infected was associated with a 2.8 times increase (95% CI 1.27–6.08) in the likelihood of mortality in the first 24 months after delivery in comparison to HIV-uninfected women (P = 0.020). The model also adjusted for child's birth weight [adjusted hazard ratio (AHR) for birth weight less than 2500 g: 1.73, 95% CI 0.95–3.16, P = 0.072, compared to greater than 2500 g], maternal antenatal HIV status, maternal age at delivery (AHR 1.04 per 1 year increase, 95%CI 0.99–1.09, P = 0.095), maternal CD4 cell count, maternal weight 6 weeks after delivery (AHR 0.94 per 1 kg increase, 95% CI 0.92–0.97, P < 0.001), mothers' highest education level (AHR for those with no education 1.92, 95% CI 0.80–4.57, P = 0.142, compared to those with secondary or higher education), water source (AHR for stream water: 2.17, 95% CI 1.21–3.90, P = 0.009, compared to those with piped water) and clinic setting (AHR for peri-urban 3.40, 95% CI 1.12–10.33, P = 0.03, compared to urban).
The prevalence of serious morbidity at any time in the first 2 years after delivery in those who died postnatally was 30.6% (22/72) compared to 9.2% (234/2552) in those who survived (χ2 = 36.4, P < 0.001). Among women who died the most prevalent serious morbidities, although not necessarily cause of death, were bloody diarrhoea and pulmonary TB, accounting for 36.4% (8/22) and 31.8% (7/22) of events, respectively. A total of 62 women were infected with TB, 55 (89%) were co-infected with HIV of whom seven died postnatally. The unadjusted odds of dying postnatally was over four times greater among women who had at least one serious morbidity event after delivery (OR 4.4, P < 0.001); this association remained after adjusting for maternal antenatal HIV status.
Eighty-eight percent of the 72 women who died postnatally were HIV-infected; the median antenatal CD4 cell count in these women was 234 cells/μl (range 21–878 cells/μl) compared to 460 cells/μl (range 15–1835 cells/μl) in the HIV-infected women who survived (Mann–Whitney 5.9, P < 0.001). Time to death in women with an antenatal CD4 cell count below 200 cells/μl was similar to that in women with an antenatal CD4 cell count at least 200 cells/μl (P = 0.479) but the number of women in each category was small (24 and 48 women, respectively). The overall postnatal mortality rate in women with an antenatal CD4 cell count at least 200 cells/μl was 19.35 per 1000 person-years of follow-up compared to 171.43 per 1000 person-years of follow-up in women with an antenatal CD4 cell count below 200 cells/μl. Compared to HIV-uninfected women, mortality amongst HIV-infected women with higher antentatal CD4 cell counts of at least 200 cells/μl was significantly higher (8.6 per 1000 person-years, 95% CI 7.07–10.13 in HIV-uninfected vs. 19.6 per 1000 person-years, 95% CI 17.27–21.93 in HIV-infected women, P < 0.0001).
Similarly, the median postnatal CD4 cell count in 41 HIV-infected women who died postnatally and who had postnatal CD4 cell counts recorded was 314 cells/μl (range 41–1601 cells/μl), whereas in those who survived it was significantly higher, 810 cells/μl (range 1–3843 cells/μl, Mann–Whitney 6.5, P < 0.001). There was no significant difference between the intensity of infant feeding (breastfeeding for 75% or more of the time vs. breastfeeding for less than 75% of the time or formula feeding) during the first 3 months after delivery in women who died postnatally and those who survived, regardless of mother's HIV infection status (P = 0.112 and P = 0.530, respectively). Further, among mothers who died postnatally, there was no significant association between the intensity of feeding and maternal HIV infection status (P = 0.071).
Serious maternal morbidity was identified in 250 of 2624 (9.5%) women. These 250 women experienced 348 serious morbidity events at a median time since delivery of 5.3 months (range 5 days to 2 years). The most common first serious morbidity event was bloody diarrhoea (dysentery), experienced by 26.4% (66/250) of women, followed by pulmonary TB in 24.4% (61/250 women) and an abscess in 17.2% (43/250) of women. The remaining morbidities were experienced by less than 10% of women. Similar proportions of HIV-uninfected and infected women experienced each specific type of serious morbidity, with the exception of bloody diarrhoea and abscesses which were more common in HIV-uninfected women, and pulmonary TB and shingles which were more common in HIV-infected women (184/250 were HIV-infected women). Among 184 HIV-infected women who had a serious morbidity event within 2 years of delivery, 90% had a postnatal CD4 cell count at least 200 cells/μl.
In multivariable Cox regression analysis investigating the factors associated with time to first serious morbidity event after delivery, women who were HIV-infected, had lower levels of education, or a river/stream water source were more likely to have a shorter time to their first event (Table 4).
Univariable Cox regression showed a 1.5 times increase in the risk of a serious morbidity event for 1 unit increase in log10 viral load (P < 0.001) but too few women had this information available to include in the multivariable regression (Table 4).
As highlighted in all previous studies, HIV infection carries with it a very high burden of mortality. The crude mortality rate we report of 55.68 per 1000 person-years in HIV-infected women, compared to 8.6 per 1000 person-years in HIV-uninfected women, is similar to one other African study  but higher than in others [6,12,22]. One other large African study has simultaneously monitored mortality in HIV-infected and uninfected women  and reported a higher mortality rate for HIV-infected women. However, this Zimbabwean study found that relative to HIV-uninfected women, the HIV-infected women were 54.1 times more likely to die during the first 2 years postpartum, whereas in our study the mortality rate in HIV-infected women was only 6.5. Further, with the ability to allow for other factors known to increase mortality rate, the HIV-infected women in this study were only 2.8 times more likely to die during the same time period.
We report that HIV-infected women with CD4 cell counts below 200 cells/μl were at substantially increased risk of mortality. This observation is consistent with many other studies  and highlights the importance of using pregnancy as an important window of opportunity for HIV testing and immediate CD4 testing for women who are HIV-infected. Eligible women should receive HAART which will not only reduce the risk of transmission to their infants but will also have a significant impact on improving their health and presumably prevent some of the mortality associated with low CD4 cell counts which has been consistently observed. Furthermore, we show that even HIV-infected women with higher CD4 cell counts experienced serious morbidity events, and were also at significantly increased risk of dying compared to HIV-uninfected women, highlighting the urgency to implement the new World Health Organization guidelines recommending a higher CD4 threshold for lifelong ART for pregnant women .
Improving the survival of mothers has an added benefit of improving the survival of infants and children; African studies show increased mortality amongst children in the year before and after the death of their mother, regardless of maternal HIV status [24,25]. Infant outcomes have also been related to maternal morbidity, with infants being more likely to die if their mothers had advanced disease  or even when their mothers experienced a serious adverse event, irrespective of the infant's HIV status .
When we investigated the causes of mortality, co-infection with TB was a predictor of mortality which supports the findings of an earlier study from South Africa  which reported an increase in pregnancy-related deaths in women who were co-infected with TB, as well as supporting the findings of a Zimbabwean study  which found a substantial mortality effect of co-infection with TB in women who were followed for 2 years postpartum. This highlights the importance of instituting guidelines that promote TB screening of all HIV-infected pregnant or lactating women as a matter of urgency, particularly in areas of high TB prevalence as in our study population . We would also call for integrated facilities that allow sputum testing, X-rays and TB treatment to be offered at the same site as PMTCT and HIV treatment facilities.
Bloody diarrhoea, most commonly due to dysentery, was another common serious morbidity. Additionally, lower educational levels and lack of piped water were shown to increase the risk for both morbidity and mortality, and highlights the importance of nonmedical interventions such as improved water, sanitation and empowerment and education of women, in improving the health of the population. Finally, when we investigated the influence of infant feeding mode on maternal mortality and morbidity we find that our results corroborate the findings of most previous studies, that breastfeeding even in HIV-infected women is not associated with an increased risk of mortality. This confirmatory finding is reassuring in the light of recent WHO recommendations that HIV-infected women breastfeed for up to 12 months while they or their infants receive antiretroviral drug prophylaxis to prevent HIV transmission.
Wellcome Trust supported this study (Grant #063009/Z/00/2) and the Africa Centre (Grant #050534).
1. Ronsmans C, Graham WJ. Maternal mortality: who, when, where, and why. Lancet 2006; 368:1189–1200.
2. Khan M, Pillay T, Moodley JM, Connolly CA. Maternal mortality associated with tuberculosis-HIV-1 co-infection in Durban, South Africa. AIDS 2001; 15:1857–1863.
3. Fawcus SR, van Coeverden de Groot HA, Isaacs S. A 50-year audit of maternal mortality in the Peninsula Maternal and Neonatal Service, Cape Town (1953–2002). BJOG 2005; 112:1257–1263.
4. Hosegood V, Vanneste AM, Timaeus IM. Levels and causes of adult mortality in rural South Africa: the impact of AIDS. AIDS 2004; 18:663–671.
5. Hoj L, da Silva D, Hedegaard K, Sandstrom A, Aaby P. Maternal mortality: only 42 days? BJOG 2003; 110:995–1000.
6. Zvandasara P, Hargrove JW, Ntozini R, Chidawanyika H, Mutasa K, Iliff PJ, et al
. Mortality and morbidity among postpartum HIV-positive and HIV-negative women in Zimbabwe: risk factors, causes, and impact of single-dose postpartum vitamin A supplementation. J Acquir Immune Defic Syndr 2006; 43:107–116.
7. Hargrove JW, Humphrey JH. Mortality among HIV-positive postpartum women with high CD4 cell counts in Zimbabwe
. AIDS 24
8. Chilongozi D, Wang L, Brown L, Taha T, Valentine M, Emel L, et al
. Morbidity and mortality among a cohort of human immunodeficiency virus type 1-infected and uninfected pregnant women and their infants from Malawi, Zambia, and Tanzania. Pediatr Infect Dis J 2008; 27:808–814.
9. Collin SM, Chisenga MM, Kasonka L, Haworth A, Young C, Filteau S, Murray SF. Factors associated with postpartum physical and mental morbidity among women with known HIV status in Lusaka, Zambia. AIDS Care 2006; 18:812–820.
10. Walson JL, Brown ER, Otieno PA, Mbori-Ngacha DA, Wariua G, Obimbo EM, et al
. Morbidity among HIV-1-infected mothers in Kenya: prevalence and correlates of illness during 2-year postpartum follow-up. J Acquir Immune Defic Syndr 2007; 46:208–215.
11. Nduati R, Richardson BA, John G, Mbori-Ngacha D, Mwatha A, Ndinya-Achola J, et al
. Effect of breastfeeding on mortality among HIV-1 infected women: a randomised trial. Lancet 2001; 357:1651–1655.
12. Breastfeeding and HIV International Transmission Study Group. Mortality among HIV-infected women according to children's feeding modality
. J Acquir Immune Defic Syndr
13. Coutsoudis A, Coovadia H, Pillay K, Kuhn L. Are HIV-infected women who breastfeed at increased risk of mortality? AIDS 2001; 15:653–655.
14. Kuhn L, Kasonde P, Sinkala M, Kankasa C, Semrau K, Vwalika C, et al
. Prolonged breast-feeding and mortality up to two years postpartum among HIV-positive women in Zambia. AIDS 2005; 19:1677–1681.
15. Coovadia HM, Rollins NC, Bland RM, Little K, Coutsoudis A, Bennish ML, Newell ML. Mother-to-child transmission of HIV-1 infection during exclusive breastfeeding in the first 6 months of life: an intervention cohort study. Lancet 2007; 369:1107–1116.
16. Rollins NC, Coovadia HM, Bland RM, Coutsoudis A, Bennish ML, Patel D, Newell ML. Pregnancy outcomes in HIV-infected and uninfected women in rural and urban South Africa. J Acquir Immune Defic Syndr 2007; 44:321–328.
17. Bland R, Coovadia H, Coutsoudis A, Rollins N, Newell M. Cohort profile: mamanengant or the Africa Centre Vertical Transmission Study. Int J Epidemiol
18. KwaZulu Natal Provincial Department of Health. Protocol for the phased implementation of a comprehensive package of care for the prevention of mother to child transmission of HIV in KwaZulu Natal
. 4th ed. Pietermaritzburg, South Africa: KwaZulu Natal Provincial Department of Health; 2003.
19. Patel D, Bland R, Coovadia H, Rollins N, Coutsoudis A, Newell ML. Breastfeeding, HIV status and weights in South African children: a comparison of HIV-exposed and unexposed children. AIDS 2010; 24:437–445.
20. Bland RM, Little KE, Coovadia HM, Coutsoudis A, Rollins NC, Newell ML. Intervention to promote exclusive breast-feeding for the first 6 months of life in a high HIV prevalence area. AIDS 2008; 22:883–891.
21. Leroy V, Msellati P, Lepage P, Batungwanayo J, Hitimana DG, Taelman H, et al
. Four years of natural history of HIV-1 infection in african women: a prospective cohort study in Kigali (Rwanda), 1988–1993. J Acquir Immune Defic Syndr Hum Retrovirol 1995; 9:415–421.
22. Taha TE, Kumwenda NI, Hoover DR, Kafulafula G, Fiscus SA, Nkhoma C, et al
. The impact of breastfeeding on the health of HIV-positive mothers and their children in sub-Saharan Africa. Bull World Health Organ 2006; 84:546–554.
24. Newell ML, Coovadia H, Cortina-Borja M, Rollins N, Gaillard P, Dabis F. Mortality of infected and uninfected infants born to HIV-infected mothers in Africa: a pooled analysis. Lancet 2004; 364:1236–1243.
25. Zaba B, Whitworth J, Marston M, Nakiyingi J, Ruberantwari A, Urassa M, et al
. HIV and mortality of mothers and children: evidence from cohort studies in Uganda, Tanzania, and Malawi. Epidemiology 2005; 16:275–280.
26. Kuhn L, Kasonde P, Sinkala M, Kankasa C, Semrau K, Scott N, et al
. Does severity of HIV disease in HIV-infected mothers affect mortality and morbidity among their uninfected infants? Clin Infect Dis 2005; 41:1654–1661.
27. Houlihan C, Mutevedzi P, Lessells R, Cooke G, Tanser F, Newell M. The TB challenge in a rural South African HIV programme. BMC Infect Dis