Secondary Logo

Journal Logo

Original articles

Child mortality and HIV infection in Africa

a review

Newell, Marie-Louisea; Brahmbhatt, Heenab; Ghys, Peter Dc

Author Information
  • Free



In the year 2003 alone, an estimated 590 000–810 000 children were newly infected with HIV [1], nearly always through mother-to-child transmission (MTCT), which can occur before, during and after delivery [2,3]. The epidemic among children mirrors that among women of childbearing age. Currently, close to 90% of paediatric infections occur in sub-Saharan Africa, where the prevalence of HIV infection among women of childbearing age reaches 35% or more in some parts of southern Africa [1]. As most HIV-infected children die before their fifth birthday, HIV is already contributing to increased childhood mortality rates overall [4–6]. Of the 2.5–3.5 million deaths resulting from AIDS per year worldwide, approximately one-sixth occur in children below the age of 15 years [1].

Mother-to-child transmission of HIV infection

MTCT is rare during early pregnancy, relatively frequent in late pregnancy and during delivery, whereas breastfeeding contributes substantially to the overall risk [7,8]. In the absence of specific interventions, the estimated rate of MTCT ranges from 15 to 45%, with differences in MTCT rates between populations mostly associated with the additional risk through breastfeeding [5]. The postnatal transmission from mother-to-child through breastfeeding remains important when peripartum antiretroviral prophylaxis reduces the risk of vertical transmission late in pregnancy and during delivery [8].

Mothers with advanced clinical HIV disease or severe immunodepression are more than twice as likely to transmit than mothers who are clinically well and immunologically normal [9], probably because of the associated increased RNA viral load in plasma, which is known to be strongly associated with the risk of MTCT. RNA viral load in plasma is also high just after infection is acquired, and thus infants born to women with primary HIV infection are also at high risk of infection [9]. The avoidance of contact with infectious maternal secretions through elective caesarean section delivery (before labour and before membrane rupture) approximately halves the risk of transmission, even in women with low viral loads, with high CD4 cell counts, or on antiretroviral therapy [10,11]. Breastfeeding approximately doubles the risk of vertical transmission, and in populations where breastfeeding continues into the second year, the additional risk of transmission through breastfeeding is approximately 15–20% in absolute terms [7,12].

Prevention of mother-to-child transmission of HIV

Current approaches to interventions to reduce the MTCT of HIV infection target the peripartum period, but their application in populations where breastfeeding is the norm poses a considerable challenge [4,5,7]. Interventions shown to be effective include the reduction of maternal viral load through antiretroviral therapy in pregnancy and during delivery, the avoidance of exposure to contaminated maternal secretions through elective caesarean section delivery, and the avoidance of breastfeeding. Washing of the birth canal with an antiseptic to reduce exposure to contaminated secretions has been evaluated, and was found to have an effect only for women with a duration of ruptured membranes of more than 4 h or with a high concentration of the antiseptic agent [13,14]. Nutritional interventions during pregnancy or postpartum in the form of multivitamins or vitamin A alone are not associated with a reduction in the overall risk of MTCT in African populations [15,16].

However, although the risk of the peripartum transmission of HIV can be successfully reduced by antiretroviral therapy, neither elective caesarean section delivery nor refraining from breastfeeding is a safe or affordable option in many parts of the world where the prevalence of HIV is substantial. Whereas antiretroviral therapy to the pregnant woman also reduces the risk of transmission during labour and delivery, the postpartum acquisition of infection through breastfeeding remains an important mode of the acquisition of infection for infants in these settings. Research currently focuses on making breastfeeding by HIV-infected mothers safer by providing antiretroviral prophylaxis to the breastfeeding mother or to her breastfed infant, and results from ongoing trials will become available in 2 or 3 years time.

The results of various randomized controlled trials in African populations attest to the substantial reduction in the number of infected children born to HIV-infected women achievable with the application of these interventions. However, with the more widespread introduction of the prevention of MTCT programmes, systems need to be developed that will allow the evaluation of changes in HIV-related child morbidity and mortality at a population level.

Progression of vertically acquired HIV infection

In a largely untreated population living in Europe or the United States, approximately 20% of children vertically infected with HIV rapidly progress to AIDS or death in the first year of life [17,18]. The progression of disease is slower thereafter, and by 6 years of age an estimated 40% of infected children will have progressed to AIDS or death [17,19,20]. With the widespread introduction of antiretroviral therapy at an early stage, progression to serious disease or death is now rare in Europe and the USA [18].

Disease progression is likely to be more rapid in settings with a higher background prevalence of other infections and nutritional problems, but there are few data to substantiate this. It is likely that the pattern of mortality by age in infected children is similar across settings, although the level may be substantially different. In one randomized trial to evaluate the effectiveness of peripartum antiretroviral therapy in west Africa, mortality among infected children was as high as 60% by 18 months [21]. In that trial, maternal CD4 cell counts of less than 200 cells/mm3 at the time of delivery [relative risk (RR) 3.25], maternal death (RR 9.65), and diagnosis of HIV infection in the infant before day 12 (RR 18.1) or between days 13 and 45 (RR 7.63) were significantly and independently associated with the risk of dying in HIV-infected children.

There is some evidence from studies in developed countries that advanced maternal disease not only increases the risk of infection for the child, but is also associated with more rapid progression in the infected child. Other virological characteristics may also influence the survival of HIV-infected children, although in a study in Uganda [22] the survival of infected children with sub-type A virus was similar to that of children infected with sub-type D. In that study, 38% of infected children died before 18 months.

In the randomized controlled trial of breast versus formula infant feeding in Nairobi, estimated 2-year mortality rates among all infants (infected and uninfected) were similar in the two arms: 20% versus 24% in the formula and breast group, respectively [23], similar to the overall risk at 18 months seen in the west African trial mentioned above. Infection with HIV was associated with a ninefold increased mortality risk. Although infants in the breastfeeding arm tended to have better nutritional status, especially during the first 6 months of life, than infants who were formula fed, the incidence of diarrhoea was similar in both groups (155 versus 149 per 100 person-years, respectively). The incidence of pneumonia was 62 per 100 person-years in both breastfed and formula-fed infants, and there was no significant difference between the groups in other morbidity.

In a study of infected children identified during their first year of life and followed up for more than 2 years thereafter [24], the HIV-RNA viral load and CD4 cell count at baseline were significantly associated with disease progression. A one log10 increase in viral level more than doubled the hazard of child mortality [24].

Mortality in HIV-infected breastfeeding women

Although the primary aim of the randomized trial in Nairobi was to estimate the risk of transmission associated with breastfeeding, information collected was also used to inform the debate about the possible adverse effects of infant feeding mode on the health of infected mothers. A higher mortality rate was reported in women randomly assigned to breastfeeding than in those in the formula-feeding arm [25]. Approximately 6% of women (24/397) died in the 2 years after delivery: 18 women allocated to breastfeeding and six among women randomly assigned to the formula-feeding group. However, an assessment of mortality was not the main aim of the trial, and concerns have been expressed over the methodology used and potential bias in the assessments [26]. In another study in Durban, South Africa [27], women made an informed choice regarding infant feeding, and those who chose to breastfeed were advised to do so exclusively. In a detailed analysis over a period of follow-up of on average 11 months in both breastfeeders and non-breastfeeders, there was no evidence of an increase in either mortality or morbidity in the mothers who were exclusively breastfeeding [27]. Two out of 410 exclusively breastfeeding mothers died, as did three of 156 never-breastfeeding women. There was no evidence of an association between clinical problems in the mother and the duration of breastfeeding.

In an ongoing analysis of data on more than 4200 HIV-infected breastfeeding mothers, pooled in a large international initiative, mortality rates over a 12–18-month period after delivery in HIV-infected women were being assessed. Preliminary information from data from one of the Petra trial sites in Tanzania [28] included in the meta-analysis, suggests a cumulative mortality risk of 7% at 24 months, associated with advanced disease as indicated by a CD4 cell count at delivery below 200 cells/mm3 and an RNA viral load greater than 100 000 copies/ml. In the larger dataset, overall mortality was approximately 2.8% at 12 months and 3.2% person-years of follow-up at 18 months, with maternal CD4 cell counts less than 200 cells/mm3 being strongly associated with the risk [29].

In general, studies that have monitored the survival of HIV-infected women who recently gave birth have found that these women have much lower mortality rates than HIV-infected women in the population as a whole. For example, in the studies summarized by Porter in the first paper of this volume, we noted that the mortality of HIV-infected women is of the order of 10% per year, whereas those who recently gave birth have mortality rates of approximately 5% per year for the first 2 years after giving birth. This would appear to be a result of a selection effect, whereby healthier HIV-infected women can become pregnant more easily than their counterparts with more advanced disease.

Impact of vertical transmission and parental survival on child mortality

The census and household survey data that are generally used to produce estimates of child mortality do not allow the assessment of the effect of maternal HIV infection on child mortality to be reliably quantified. Such surveys have traditionally not included HIV testing; moreover, independence between the mortality experience of mothers and children, an assumption underlying the conventional demographic estimation techniques, can no longer be assumed. However, data are now becoming available from population-based longitudinal studies in which adult seroprevalence has been assessed at regular intervals, together with household surveys relating to deaths.

In an analysis of pooled data from three such longitudinal community based studies from Uganda, Tanzania and Malawi, in which births were classified by maternal HIV status, the excess risk of child mortality was estimated by the maternal HIV infection status. Information on the child's HIV infection was not generally available. In Masaka, southern Uganda, demographic surveillance covered 15 villages, with a maximum population of 17 000 people. Annual censuses were held since 1989, and serological surveys of adults over 13 years of age follow each census. In Kisesa, northern Tanzania, six villages and a semi-urban roadside settlement covering up to 27 000 people are included. Censuses were conducted at 4-monthly intervals between 1994 and 1996, 6-monthly or annually thereafter. Serological surveys were conducted in 1994, 1996–1997 and 1999–2000. The third study is located in the Karonga district in rural northern Malawi. Seroprevalence was estimated from filter paper blood samples collected between 1981 and 1989, and 197 HIV-infected people with 396 HIV-uninfected controls were contacted again between 1996 and 2000 to form a retrospective cohort. All the spouses and children of these cohort members are also included in the study. A census was administered to the total 2273 family members still living in the district in 1999, with serotesting for HIV [30–32].

Allowing for geographical and temporal differences in background levels of mortality, the child's sex and the mother's age, children born to HIV-infected mothers were estimated to be three times more likely to die than children born to uninfected mothers, with the effect lasting throughout the childhood years (Fig. 1). The year before and the year after a mother's death was a high-risk period for the child, and the estimated excess risk of child mortality associated with maternal death was nearly threefold in the year preceding the mother's death and nearly fivefold in the year after a mother's death; these risks were independent of maternal HIV status. The effect of maternal HIV infection was restricted to children under 5 years of age, and the effect of maternal survival to children less than 3 years of age. At later ages, the numbers followed are too small to yield statistically significant results. The results may suggest that having a dead or dying HIV-negative mother is associated with similar adverse consequences as having a healthy HIV-infected mother.

Fig. 1.:
Age-specific child mortality rates, by HIV status of mother at birth. – – ○ – – HIV negative, Masaka; – – □ – – HIV negative, Kisesa; – – ▵ – – HIV negative, Karonga; — ● — HIV positive, Masaka; — ▪ — HIV positive, Kisesa; — ▴ — HIV positive, Karonga. Data from Nakiying et al. [30], Ng'weshemi et al. [31], and Crampin et al. [32].

Although this paper clearly shows increased mortality of children born to HIV-infected mothers compared with those of uninfected mothers, because the HIV infection status of the child was not determined in the studies, it was not possible to estimate directly the mortality caused by HIV infection in the child itself, net of any effect of the health and HIV status of the mother on survival of her children.

Survival of infected and uninfected children of HIV-infected mothers

In a prospective community-based study in Rakai District, Uganda, children born to HIV-infected women were followed up until at least 24 months to assess their HIV infection status and mortality. Antenatal HIV prevalence was 17%, and the rate of MTCT was 21% overall. Two-year mortality in infected children of infected mothers was 547 per thousand, in uninfected children of infected mothers 166 per thousand, and in children born to uninfected mothers 128 per thousand. These estimates are thus in line with previously reported data from west Africa and Kenya. The risk of mortality in all children born to HIV-infected mothers was significantly increased, and was associated with maternal RNA viral load, which ties in with the observations above of increased child mortality risk in the year before and after a maternal death [33].

In a different study in the same setting [34], the possible association of polygyny on child survival was assessed. In a cohort of over 4000 pregnant women and their infants with 5300 person-years of observation, it was estimated that polygyny increased the hazard of child mortality overall (RR 1.36, P < 0.001) and in mothers who were HIV infected (RR 2.17, P < 0.001). There was no effect on child mortality of infants born to HIV-uninfected mothers. This would suggest that in polygynous households, where not all wives may be HIV infected, resources are diverted away from children of infected mothers.

In the Gambia, children born to 101 HIV-1-infected women, 243 HIV-2-infected and 468 uninfected women were followed up to 18 months of age [35]. Overall, 15 children born to HIV-1-infected women died, 14 in the group of HIV-2-infected women and 25 in the uninfected group. The estimated mortality among children born to HIV-1-infected women was 150 per thousand, substantially higher than mortality among children of HIV-2-infected women (60 per thousand) and among children of uninfected women (70 per thousand). These differences reflect the lower rate of MTCT for HIV-2, which is also less pathogenic [35]. Six of the 17 HIV-1-infected children died, compared with none of the eight HIV-2-infected children, but this difference was not statistically significant because of the small sample size (P = 0.13). A high RNA viral load in either the infant or mother was associated with increased mortality risk in the infant.

Child survival by timing of acquisition of HIV infection

The timing of MTCT is not precisely quantified, but it is generally accepted that in non-breastfeeding populations approximately two-thirds of transmission occurs in the peripartum period, whereas in breastfeeding populations this is reduced to a little less than half, and a similar proportion occurs over 18–24 months of breastfeeding [3]. Different routes may be associated with the intensity and amount of virus. Whether the route and timing of HIV-1 transmission from mothers to their infants influence morbidity and mortality is not quantified, but it is likely that in-utero transmission across the placenta, or intra-partum by exposure to contaminated maternal secretions in the birth canal, to an immunologically immature fetus has a poor outcome in terms of survival of the infant [33]. On the other hand, among the approximately 30–40% of infected children in breastfeeding populations who acquired their infection postnatally, it could be expected that these children who acquired infection when they were immunologically more mature may be less likely to progress rapidly to serious disease or death. Also, an infant who is infected later has already survived the neonatal period, a time of high mortality risk for both HIV and other causes, and this would be an additional reason to expect a lower mortality rate.

In a small study in Kigali, Rwanda [36], early infection with HIV was associated with a subsequent risk of death or AIDS. Among 54 HIV-infected children, 39 were estimated to have been infected early, before the first 3 months of life, which would be intra-uterine, intra-partum and early postpartum acquisition, whereas the remaining 15 were infected later, through breastfeeding. The risk of death was five times higher in those with early infection compared with those infected later, but the confidence interval was wide, and the analysis did not adequately allow for the later time of becoming infected and the duration of follow-up while infected. Recently, in a study in Tanzania [37], multivitamins given to the mother during pregnancy and lactation were associated with a slight but not significant reduction in breastfeeding transmission and a significant reduction in the risk of mortality (by 24 months) among children with evidence of acquisition of infection through breastfeeding after 6 weeks or age. This was taken to imply a possible benefit of multivitamins on breastfeeding-induced pathogenesis (vertical transmission and mortality), although similar methodological problems as mentioned above hinder the interpretation.

In the randomized trial of breastfeeding versus formula feeding conducted in Nairobi, Kenya, between 1992 and 1998, early infection was defined as a positive HIV-DNA polymerase chain reaction (PCR) assay before 2 months of age (indicating a combination of intra-uterine, intra-partum and very early postpartum transmission), whereas late infection was defined as a PCR negative test at 2 months of age, followed by a positive test result thereafter. Of a total of 80 HIV-infected children, one third had evidence of late and two-thirds of early infection. Mortality was 12 per 100 person-years for those infected late but 71 per 100 person-years for those infected early [hazard ratio (HR) 0.2, P = 0.02]. Infants infected late had a lower peak viral load, and a lower viral setpoint than those infected early (B. Richardson, personal communication, 2002). A similar pattern was observed in a large cohort of infants born to HIV-infected women in Harare, Zimbabwe. These observations need urgent confirmation as they have implications for predicted trends in child mortality in settings with high HIV prevalence, where programmes to prevent MTCT in the peripartum period are now being introduced.

In a meta-analysis of data from seven trials in sub-Saharan Africa [38], the crude mortality rate in 3468 children born to HIV-infected mothers was assessed. Preliminary results suggest that overall mortality was 100 per 1000; higher in east and west Africa than in South Africa; higher for infants born to women with HIV-related immunosuppression and those whose mothers died in the year after delivery; and higher for infected than for uninfected children. Among the infected children, those who had evidence of being infected at, or soon after, birth had a higher mortality risk than those with evidence of acquisition of infection after 4 weeks of age, which was confirmed in a further analysis that allowed for age at acquisition of infection among the infected children [38].

Trends in under-five mortality caused by HIV in sub-Saharan Africa

Outside relatively small clinical trials or observational studies, summarized above, generally with a limited duration of follow-up, there is a lack of data on the mortality of HIV-infected children in the first five or more years of life at a population level. Because it is not possible to diagnose HIV infection among newborns of HIV-infected mothers using antibody tests, given the passive transfer of maternal antibodies and the cost of alternative diagnostic methods such as DNA PCR, estimates of the number of newly infected children per year have been derived from adult HIV antibody prevalence estimates, using various assumptions relating to the representativeness of prevalence studies, fertility, transmission rates, and mortality in women of childbearing ages, as previously described [39]. For each country, age-specific patterns of fertility from the United Nations Population Division were used as the basis for estimating the number of children born to HIV-positive mothers. Adjustments were made to these national fertility patterns to address the representativeness of HIV-infected young women and to correct for the effects of HIV/AIDS on fertility. First, for women aged 15–19 years who are HIV infected, fertility was assumed to be 50% higher than the national rate for that age because all HIV-infected women are sexually active and exposed to the risk of pregnancy. For all other age groups, the age-specific fertility rate was reduced by 20% to account for the reduction of fertility caused by HIV/AIDS. The fertility of HIV-negative women was adjusted downwards (15–19) or upwards (20–49) to compensate for the increase or reduction among HIV-infected women in order to leave the overall fertility rate in each age group unchanged.

Under-five deaths caused by HIV were generated from the number of newly infected children per year by applying a net HIV/AIDS survival function, based among other sources on the Rakai data that were discussed above. The survival time for children is described by a double Weibull distribution, with median survival of less than 3 years for children not receiving antiretroviral therapy [40]. According to that distribution, 60% of children infected via MTCT will die before the age of 5 years as a result of HIV/AIDS.

The results of this exercise show that in sub-Saharan Africa, HIV infection is an ever-increasing cause of child mortality (Fig. 2). By 2002, it represented close to 10% of all deaths among children under 5 years of age in the region. However, the effect of HIV on child mortality varies widely between countries: the effect is substantial in southern Africa, causing up to half of all child deaths in the worst affected countries, whereas it hardly contributes at all in several countries of west Africa where antenatal prevalence is less than 5%.

Fig. 2.:
Estimated under-five mortality caused by HIV/AIDS per 1000 births in sub-Saharan Africa, 1994–2002.

Using similar methods, Zaba et al. [41] estimated the contribution of HIV infection to child mortality in selected countries with different trends in child mortality before the advent of the HIV/AIDS epidemic and with different levels of HIV prevalence. This method does not require assumptions about fertility differences between infected and uninfected women, and makes more direct use of information about HIV prevalence in pregnant women. In their conclusions they stressed the importance of the level and trend of child mortality from causes unrelated to HIV in determining overall child mortality trends. Therefore, in a country such as Botswana, which had achieved fairly stable low child mortality rates before the epidemic (60 per thousand in the late 1980s), and where HIV prevalence in pregnant women is estimated to be over 35%, child mortality has doubled in the 1990s. By contrast, in Malawi, where child mortality was nearly 250 per thousand before the epidemic, but declining rapidly, overall child mortality levels are estimated to have continued to fall throughout the 1990s, even though HIV prevalence has reached nearly 15%. In Botswana, HIV-attributable mortality in 2000 was over 60%, in Malawi just over 10%. These results are confirmed to some extent by nationally representative household surveys such as Demographic and Health Surveys, although the authors question the reliability of estimates based on retrospective reports of birth histories that go back more than 5 years before the survey, because mothers who have been infected recently are more likely to be included in the survey sample than mothers who were living with HIV/AIDS during most of their reproductive years and therefore had a larger proportion of their children infected with HIV.

In conclusion, in the absence of vital registration and reliable cause of death information, evidence on the impact of HIV infection on child mortality is limited to data from either relatively small studies or to more extensive population-based data that have been used together with specific assumptions to arrive at estimated rates of child mortality as a result of HIV/AIDS. This review shows that child mortality and the association with HIV infection as estimated in various demographic models is in line with the limited empirical data available from cohort studies in some settings across Africa, although there is a need to improve the assumptions used in the models as new information becomes available.

The child mortality estimates from community-based cohorts demonstrate that children of HIV-infected mothers have higher mortality rates than children of uninfected mothers, and that child mortality is closely linked with maternal health status, but because the proportion of vertically infected children is unknown, the value of these studies is limited to some extent. Models that use HIV surveillance data together with a set of assumptions derived from available evidence provide a valuable tool to estimate HIV-related mortality in children, allowing their use for informing the required policy decisions. However, to improve the projections of the overall effect that the HIV epidemic will have on child mortality at the population level in countries with generalized epidemics requires reliable age-specific mortality rates in infected and uninfected children, particularly if MTCT rates change after MTCT prevention efforts and changes in breastfeeding practices. The findings of the study in Rakai, Uganda are thus of considerable interest, as they do provide such information. However, there is a need to develop methods that would provide the opportunity to assess trends in child mortality allowing for HIV infection in the child and the mother without resorting to cohort studies, which are very expensive and time-consuming.

In many ways, the data on all-cause child mortality are more plentiful and of higher quality than the corresponding data for adults. Even countries that do not have vital registration systems can rely on time series of child mortality estimates that go back to the 1970s, thanks to census-based indirect estimates using the Brass techniques based on ratios of children surviving to children ever born and more recent direct estimates based on birth histories collected in household surveys. On the other hand, we are not as well informed about HIV prevalence among children as among adults, necessitating the use of models to convert information on HIV prevalence in pregnant women to estimates of the proportion of children infected. If national household surveys begin to collect information on the HIV status of women in the near future it will be important to tabulate birth history information separately for infected and uninfected women.

The relative abundance of child mortality data that illustrate wide contrasts between African countries, even those in the same regional grouping, should caution against broad generalizations about the effects of HIV on child mortality on the continent as a whole. Both disparate trends in HIV prevalence and varying levels of non-HIV-associated child mortality will ensure very different impacts in different countries.


1. UNAIDS/WHO. AIDS epidemic update. Geneva: UNAIDS; 2003.
2. Dabis F, Newell ML, Fransen L, Saba J, Lepage P, Leroy V, et al. Prevention of mother-to-child transmission of HIV in developing countries: recommendations for practice.Health Policy Planning 2000; 15:34–42.
3. Newell ML. Mechanisms and timing of mother-to-child transmission of HIV-1.AIDS 1998; 12:831–837.
4. Dabis F, Ekpini ER. HIV-1/AIDS and maternal and child health in Africa.Lancet 2002; 359:2097–104.
5. De Cock K, Fowler MG, Mercier E, de Vincenzi I, Saba J, Hoff E, et al. Prevention of mother-to-child HIV transmission in resource-poor countries–translating research into policy and practice.JAMA 2000; 283:1175–1182.
6. Walker N, Schwartländer B, Bryce J. Meeting international goals in child survival and HIV/AIDS.Lancet 2002; 360:284–289.
7. Fowler MG, Newell ML. Breast-feeding and HIV-1 transmission in resource-limited settings.J Acquired Immune Defic Syndr 2002; 30:230–239.
8. Leroy V, Karon JM, Alioum A, Ekpini ER, Meda N, Greenberg AE, et al. 24-Month efficacy of a maternal short-course zidovudine regimen to prevent mother-to-child transmission of HIV-1 in West Africa.AIDS 2002; 16:631–641.
9. Leroy V, Montcho C, Manigart O, Van de Perre P, Dabis F, Msellati P, et al.Maternal plasma viral load, zidovudine and mother-to-child transmission of HIV-1 in Africa: DITRAME ANRS 049a trial.AIDS 2001; 15:517–522.
10. European Collaborative Study. HIV-infected pregnant women and vertical transmission in Europe since 1986.AIDS 2001; 15:761–770.
11. Ioannidis JPA, Abrams EJ, Ammann A, Bulterys M, Goedert J, Gray L, et al. Perinatal transission of human immunodeficiency virus type 1 by pregnant women with RNA virus loads 1000 copies/ml.J Infect Dis 2001; 183:539–545.
12. Nduati R, John G, Ngacha DA, Richardson S, Overbaugh J, Mwatha A, et al. Effect of breastfeeding and formula feeding on transmission of HIV-1: a randomised clinical trial.JAMA 2000; 283:1167–1174.
13. Biggar RJ, Miotti PG, Taha TE, Mtimavalye L, Broadhead R, Justesen A, et al. Perinatal intervention trial in Africa: effect of a birth canal cleansing intervention to prevent HIV transmission.Lancet 1996; 347:1647–1650.
14. Gaillard P, Mwanyumba F, Verhofstede C, Claeys P, Chohan V, Goetghebeur E, et al. Vaginal lavage with chlorhexidine during labour to reduce mother-to-child HIV transmission: clinical trial in Mombasa, Kenya.AIDS 2001; 15:389–396.
15. Coley J, Msamanga GI, Smith-Fawzi MC, Kaaya S, Hertzmark E, Kapiga S, et al. The association betweem maternal HIV-1 infection and pregnancy outcomes in Dar es Salaam, Tanzania.Br J Obstet Gynaecol 2001; 108:1125–1133.
16. Coutsoudis A, Pillay K, Spooner E, Kuhn L, Coovadia H, for the South African Vitamin A Study Group. Randomized trial testing the effect of vitamin A supplementation on pregnancy outcomes and early mother-to-child HIV-1 transmission in Durban, South Africa.AIDS 1999; 13:1517–1524.
17. The French Pediatric HIV Infection Study Group, European Collaborative Study. Morbidity and mortality in European children vertically infected by HIV-1.J Acquired Immune Defic Syndr Hum Retrovirol 1997; 14:442–450.
18. European Collaborative Study. Fluctuations in symptoms in HIV-infected children: the first 10 years of life.Pediatrics 2001; 108:116–122.
19. Diaz C, Hanson C, Cooper ER, Read JS, Watson JE, Mendez H, et al. Disease progression in a cohort of infants with vertically acquired HIV infection observed from birth: the women and infants transmission study (WITS).J Acquired Immune Defic Syndr Hum Retrovirol 1998; 18:221–228.
20. European Collaborative Study. Natural history of vertically acquired human immunodeficiency virus-1 infection.Pediatrics 1994; 94:815–819.
21. Dabis F, Elenga N, Meda N, Leroy V, Viho I, Manigart O, et al. 18-Month mortality and perinatal exposure to zidovudine in west Africa.AIDS 2001; 15:771–779.
22. Eshlemann S, Guay L, Fleming T, Mwatha A, Mracna M, Becker-Pergola G, et al. Survival of Ugandan infants with subtype A and D HIV-1 infection (HIVNET 012).J Acquired Immune Defic Syndr 2002; 31:327–330.
23. Mbori-Ngacha D, Nduati R, Reilly M, Richardson BA, Mwatha A, Ndinya-Achola J, et al. Morbidity and mortality in breastfed and formula-fed infants of HIV-1-infected women.JAMA 2001; 286:2413–2420.
24. Taha ET, Kumwenda N, Hoover D, Biggar R, Broadhead R, Cassol S, et al. Association of HIV-1 load and CD4 lymphocyte count with mortality among untreated African children over one year of age.AIDS 1999; 14:453–459.
25. 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.Lancet 2001; 357:1651–1655.
26. Newell ML. Does breastfeeding really affect mortality among HIV-1 infected women?Lancet 2001; 357:1634–1635.
27. Coutsoudis A, Coovadia H, Pillay K, Kuhn L. Are HIV-infected women who breastfeed at increased risk of mortality?AIDS 2001; 15:653–655.
28. Killewo C, Karlsson K, Swai A, Massawe A, Lyumuya E, Mhalu F, Biberfeld G. Mortality during the first 24 months after delivery in a cohort of HIV-1 seropositive women in Dar es Salaam, Tanzania. Vol. II. In: XIVth International AIDS Conference. Barcelona, Spain, July 7–12 2000 [Abstract WePeC6066].
29. Newell M-L, Read J, Leroy V, Dabis F. Mortality among HIV-infected mothers and children's feeding modality: the Breastfeeding and HIV International Transmission Study (BHITS). In: 2nd IAS Conference on HIV Pathogenesis. Paris, France, July 2003 [Abstract 221].
30. Nakiying JS, Whitworth JAG, Ruberantwari A, Busingye J, Mbulaiteye SM, Bracher M, Zaba B. Child survival in relation to mother's HIV infection: evidence from a Ugandan cohort study.AIDS 2003; 17:1827–1834.
31. Ng'weshemi J, Urassa M, Isingo R, Mwaluko G, Ngalula J, Boerma T, et al. HIV impact on mother and child mortality in rural Tanzania.J Acquired Immune Defic Syndr 2003; 33: 393–404.
32. Crampin A, Floyd S, Glynn J, Madise N, Nyondo N, Khondowe M, et al. The impact of HIV on the mortality and well-being of children in rural Malawi.AIDS 2003; 17:389–398.
33. Brahmbhatt H, Kigozi G, Wabwire-Mangen F, Wawer MJ, Gray R. Two year survival of HIV-positive and HIV-negative children born to HIV-positive and HIV-negative mothers; evidence from Rakai, Uganda. In: Empirical Evidence for the Demographic and Socioeconomic Impact of AIDS. Durban, South Africa, 2003.
34. Brahmbhatt H, Bishai D, Wabwire-Mangen F, Kigozi G, Wawer M, Gray RH, et al. Polygyny, maternal HIV status and child survival: Rakai, Uganda.Soc Sci Med 2002; 55:585–592.
35. Ota M, O'Donovan D, Alabi A, Milligan PJM, Yamuah L, N'Gom P, et al. Maternal HIV-1 and HIV-2 infection and child survival in the Gambia.AIDS 2000; 14:435–439.
36. Lepage P, Spira R, Kalibala S, Pillay K, Giaquinto C, Castetbon K, et al. Care of human immunodeficiency virus-infected children in developing countries.Pediatric Infect Dis J 1998; 17:581–586.
37. Fawzi WW, Msamanga GI, Hunter D, Renjifo B, Antelman G, Bang H, et al. Randomized trial of vitamin supplements in relation to transmission of HV-1 through breastfeeding and early child mortality.AIDS 2002; 16:1935–1944.
38. Newell ML, for the IHMA study group. Mortality in children born to HIV infected mothers in Africa: Infants, HIV and Mortality in Africa (IHMA) Study Group. Oral presentation. ICASA. Nairobi 2003, 21–26 September [Abstract 1098982].
39. Walker N, Stanecki KA, Brown T, Stover J, Lazzari S, Garcia-Calleja JM, et al. Methods and procedures for estimating HIV/AIDS and its impact: the UNAIDS/WHO estimates for the end of 2001.AIDS 2003; 17:2215–2225.
40. The UNAIDS Reference Group on Estimates. Modelling and projections. Improved methods and assumptions for estimation of the HIV/AIDS epidemic and its impact: recommendations of the UNAIDS Reference Group on Estimates, Modelling and Projections.AIDS 2002; 16:W1–W14.
41. Zaba B, Marston M, Floyd S. The effect of HIV on child mortality trends in sub-Saharan Africa. Available at: [Accessed 19 August 2003]

Africa; breastfeeding; child mortality; models; timing of acquisition

© 2004 Lippincott Williams & Wilkins, Inc.