This multisite study was conducted in Africa at a time when there were few opportunities to access ART or adequate care for opportunistic infections. Since late 2003, national programs have scaled up and now provide ART to large numbers of infected individuals, dramatically improving the survival and quality of life for those involved. These data describe the natural history of HIV disease in antenatal women and their infants before ART availability.
In this study, mortality among mothers and infants did not differ among sites.
Two percent of the HIV-infected women died during the follow-up period that began in their second trimester of pregnancy and continued for 12 months after delivery. There were no deaths among the 367 HIV-negative women and only 3% of these women had a SAE, illustrating the differences between HIV uninfected and infected women. Each of the countries where the study was conducted have some of the highest maternal mortality ratios in the world: Malawi 984 per 100,000 live births,10
Tanzania 578 per 100,000 live births,11
Zambia 729 per 100,000 live births.12
However, although the majority of SAEs occurred during labor and delivery for all women, 90% of deaths occurred postnatally and no deaths occurred during labor and delivery. These findings suggest that the high maternal mortality reported in these countries may be more directly influenced by the underlying health status, such as HIV-status and anemia.
The primary cause of death among these women was anemia. The prevalence of anemia is high among women in sub-Saharan Africa. National population-based surveys estimate 44% of women in Malawi and 48% of women in Tanzania are anemic.10,11
Anemia is a leading contributor to maternal morbidity and mortality in developing countries13
and is caused primarily by malaria, intestinal parasites and HIV itself. Our results support the role of reducing anemia in women to reduce maternal mortality, especially among HIV-infected women.
Infant mortality, with or without HIV infection, is high in these countries. In contrast to the maternal cohort in which only 25% of SAEs resulted in death, the majority of SAEs (82%) among the infants resulted in death. Mortality was high in this cohort of infants where 86% of the mothers were HIV-infected and 20% of the infants were HIV infected by the end of the 12 month follow-up period. Almost 15% of the infants in the cohort died during this follow-up period (146 per 1000 live births), higher than the all infant mortality in Tanzania (68 per 1000 live births),11
Zambia (95 per 1000 live births),12
and Malawi (76 per 1000 live births).10
Mortality among HIV-1-infected infants was more than 5 times as high as uninfected infants regardless of the timing of infection. By the end of the follow-up period, 42% of infected infants had died, compared with 7.2% of uninfected HIV-exposed infants and 4.8% of HIV unexposed infants. The mortality was the highest among infants infected early (at birth or 4–6 weeks) and confirms other studies of perinatal transmission.14,15
The mortality rate among infants infected early is more than 3-fold higher than the mortality rate of infants infected after 8 weeks. This high mortality underscores the importance of early diagnosis and immediate referral to ART. Increasing access to appropriate therapy is essential to prevent early mortality.16
Infant outcomes were strongly related to maternal outcomes. Infants in our study were significantly more likely to die if their mother experienced an SAE or died, regardless of the infant's HIV status. In univariate analysis, the mortality rate was more than 2-fold higher among infants born to mothers who experienced an SAE and >4-fold higher among infants born to mothers who died. Among the HIV-infected mothers, the relationship of the mother's health to the outcome of the infant can also be quantified by the maternal CD4 count and the HIV viral load at baseline. The lower the mother's CD4 count or the higher the viral load at enrollment, the more likely the infant was to die in the first year. An infant of a mother with a CD4 count <200 was 3 times more likely to die than a child of a mother with a CD4 count >500. All the unadjusted maternal factors in the analysis of infant mortality continued to be significant in the multivariate analysis including HIV infection, maternal morbidity and mortality, CD4 count, and RNA viral load. Clinical diagnosis or WHO staging will not accurately identify all those with a low CD4 count. Therefore, CD4 count should be included as part of routine prenatal HIV testing after post-test counseling of all infected women.
The limitations of this study include the underestimation of HIV-related morbidity and mortality among infants. The HIV-status was not determined at death and some of these infants may have been infected after their last negative test. Because of the testing sequence, an infant with a negative HIV-test at birth and at 6–8 weeks who subsequently seroconverted and died before the next follow-up visit would be classified as HIV-negative or unknown. Mothers or infants receiving care elsewhere for an SAE were not captured by this analysis. Additionally, data from 3 sites with diverse populations were pooled for this analysis and differences may exist in standard of care, protocols for making clinical diagnoses, care-seeking behavior, and access. However, these differences are not likely to be significant as our analysis demonstrated similar outcomes in all 3 populations. Also, the exclusion criteria for this study selected for asymptomatic women and excluded the symptomatic or chronically ill. This likely reduced the maternal morbidity and mortality in this study population relative to a population-based sample.
Because of the significant difference in mortality of infants born to HIV-infected compared with uninfected mothers, the counseling message to all HIV-positive women should state this increased risk of death to a newborn. This message will help women make informed reproductive decisions. Additional programs should provide early and routine HIV testing for all pregnant women, CD4 monitoring at the time of testing for infected mothers, early diagnosis, and appropriate early treatment for HIV-exposed infants, a family support program for infants of sick or deceased mothers, and appropriate and timely ART for infected mothers and infants, including ART regimens for the PMTCT.