University of California, San Francisco Logo

University of California, San Francisco | About UCSF | Search UCSF | UCSF Medical Center

Child mortality according to maternal and infant HIV status in Zimbabwe
Global Health Sciences Literature Digest
Published July 2, 2007
Journal Article

Marinda E, Humphrey JH, Iliff PJ, Mutasa K, Nathoo KJ, Piwoz EG, et al. Child mortality according to maternal and infant HIV status in Zimbabwe. Pediatr Infect Dis J 2007 Jun;26(6):519-26.


To evaluate the influence of maternal and infant HIV status, and the influence of the time of infant infection, on mortality and causes of death, by comparing survival among five groups of children: those infected in utero, intrapartum, and postnatally; those born to uninfected mothers; and those born to HIV-infected mothers who remain uninfected.

Study Design

A prospective cohort design. This study was part of the ZVITAMBO trial, in which >14,000 mother-infant pairs were enrolled in the post-partum period and followed. Enrolment occurred between November 1997 and January 2000. Follow-up visits were conducted at six weeks and three months, and then every three months for up to two years.


Harare, Zimbabwe.


Fourteen thousand, one hundred ten post-partum mothers and their infants were enrolled within 96 hours of delivery. Mothers and infants were eligible if neither had an acutely life-threatening condition, the infant was a singleton with birth weight >= 1500 grams, and the mother planned to stay in Harare after delivery. The average age of mothers was 25 years; and >90% were married.


There was no intervention. Medical care and HIV and infant-feeding counseling were offered throughout the trial. However, the trial preceded the availability of ARV prophylaxis for antenatal women, and of WHO recommendations for cotrimoxazole prophylaxis for exposed infants.

Primary Outcomes

The primary outcome was infant mortality and cause of death. Mortality was estimated using Kaplan-Meier analyses and calculation of incidence density. Cox-proportional hazard modeling was used to estimate the effect of HIV infection and timing on mortality.

Baseline data were collected by questionnaire, hospital record review, and direct measurement. At baseline, all mothers were tested for HIV and CD4 cell counts, if sero-positive. Baseline-negative mothers were retested at every follow-up visit. Infants <=18 months of age were tested using a qualitative DNA PCR assay, and with an EIA test at >18 months of age. At the end of follow-up, the last available sample from each infant was tested. If positive, earlier stored samples were tested to determine timing of HIV infection. For infants who died, a study pediatrician blinded to treatment and maternal and infant HIV status assigned the cause of death.

Infants were classified as follows: NE (not exposed)-mother negative at baseline (BL) and did not seroconvert at follow-up; NI (not infected)-mother sero-positive at BL and infant's last available test was negative; IU (intrauterine infection)-infant tested PCR-positive at BL; IP (intrapartum infection)-infant tested negative at BL, but PCR-positive at six weeks; PN (postnatal infection)-infant tested negative at BL and six weeks, but positive at subsequent time-point.


At baseline, 4495 (32%) mothers were HIV-infected, and 9562 (68%) were HIV-negative. Of infants born to HIV-infected mothers, 1336 (29.7%) became HIV-infected by 24 months of age: 381 (8.5%) IU; 508 (11.3%) IP; 258 (5.7%) PN; 189 (4.2%) uncertain timing. Two-year mortality was 2.9% among NE infants, 9.2% among NI, 67.5% among IU, 65.1% among IP, and 33.2% among PN. Fifty percent (84/167) of unexposed (NE) infants who died and 38% (81/ 212) of uninfected (NI) infants who died did so within the first eight weeks after birth; this was compared to only 8% (17/ 207) of infants infected in utero (IU) who died. However, between eight weeks and six months, the risk of death declined by 50% and 73% among uninfected and unexposed infants, respectively, but increased four-fold for those who were infected in utero.

Among all infants who survived to six months, those born to HIV-infected mothers (NI, IU, IP, PN) were 16 times more likely to die in the second half of infancy compared to unexposed (NE) infants (12.3/100 children-years versus 0 .8/100 child-years). The earlier an infant was infected, the more likely he was to die: median time from infection to death was 208, 380, and >500 days for IU, IP, and PN infants, respectively. Major causes of all infant deaths were acute respiratory infection (ARI) (67%), malnutrition (19%) and acute diarrhea (18%). Death due to malnutrition (46%) and acute diarrhea (30%) were especially high among infants infected postnatally.

Low birth weight and maternal death were independent predictors of death among all groups of children. Among uninfected children, maternal illness and among unexposed children, birth weight, maternal death, and mother's marital status were independent predictors of death.


The authors conclude that perinatally infected infants are at very high risk of death between the postnatal period and six months of age. Uninfected infants of infected mothers have at least twice the mortality risk of infants born to uninfected mothers. Specific programs should be implemented (see programmatic implications) to address these issues.

Quality Rating

According to the Newcastle-Ottawa quality rating scale for prospective cohort studies, this study was of good quality. Some bias may have occurred when assessing the infants' cause of death due to the use of death records and verbal autopsy, which may have resulted in misclassification. However, bias was minimized by blinding a study determining cause of death. Additionally, there was some lack of precision in determining the timing of HIV infection, based on blood tests taken at three-month follow-up visits.

In Context

This study confirms the main findings of a previous pooled analysis of seven African trials.(1) Mortality is very high for all HIV-infected children, especially those infected perinatally. The mortality was even higher among the infected infants in this trial compared with the pooled analysis, primarily because IU and IP infants made up 77% of the infected infants in this study, compared with 39% in the pooled analysis.

Consistent with one,(2) but not a second(3) study, uninfected infants of HIV-infected mothers are at higher risk of death compared with infants born to HIV-negative mothers. This increased risk has been hypothesized to result from poor care given by a sick mother, low socioeconomic status due to a dead father and/or sick mother, or increased exposure to infections carried by immune-compromised parents. Additionally, this study is consistent with previous reports in which HIV-exposed infants whose mothers have advanced HIV disease are more likely to become infected, more likely to die , regardless of whether they become infected.(2,4) In a recent study among HIV-exposed Kenyan infants, advanced maternal HIV disease was associated with reduced ability to transfer protective maternal antibodies to measles.(5)

Programmatic Implications

Because of the risk of acute respiratory infection, particularly due to PCP, cotrimoxazole prophylaxis is now recommended by WHO for all HIV-exposed children from four to six weeks of age until HIV infection has been excluded or until the child is no longer exposed. The high mortality and short survival time of perinatally infected infants underlines the importance of early identification of HIV-infected antenatal women. Large numbers of infants who are infected in utero or during delivery will die unless HIV diagnosis and aggressive clinical management of infants are initiated early. Recent observational data indicate that starting ART during the early months of life provides virological, immunologic, and clinical benefits.(6,7) Because HIV-infected mothers with more advanced disease are more likely to infect their infants, and these children are more likely to die whether or not they are infected, universal access to ART for pregnant and lactating women and their infants would be highly beneficial. One-third of mothers in this hospital were HIV-infected, and this high proportion underscores the need for continued primary prevention, and for strengthening the capacity of extended family members and others to care for infants born to these women.


  1. Newell ML, Coovadia H, Cortina-Borja M, Rollins N, Gaillard P, Dabis F, et al. Mortality of infected and uninfected infants born to HIV-infected mothers in Africa: a pooled analysis. Lancet 2004 Oct 2-8;364(9441):1236-43.
  2. 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 Dec 1;41(11):1654-61.
  3. Spira R, Lepage P, Msellati P, Van De Perre P, Leroy V, Simonon A, et al. Natural history of human immunodeficiency virus type 1 infection in children: a five-year prospective study in Rwanda. Mother-to-Child HIV-1 Transmission Study Group. Pediatrics 1999 Nov;104(5):e56.
  4. Chearskul S, Chotpitayasunondh T, Simonds RJ, Wanprapar N, Waranawat N, Punpanich W, et al. Survival, disease manifestations, and early predictors of disease progression among children with perinatal human immunodeficiency virus infection in Thailand. Pediatrics 2002 Aug;110(2 Pt 1):e25.
  5. Farquhar C, Nduati R, Haigwood N, Sutton W, Mbori-Ngacha D, Richardson B, et al. High maternal HIV-1 viral load during pregnancy is associated with reduced placental transfer of measles IgG antibody. J Acquir Immune Defic Syndr. 2005 Dec 1;40(4):494-7.
  6. Berk DR, Falkovitz-Halpern MS, Hill DW, Albin C, Arrieta A, Bork JM, et al. Temporal trends in early clinical manifestations of perinatal HIV infection in a population-based cohort. JAMA 2005 May 11;293(18):2221-31.
  7. Chiappini E, Galli L, Tovo PA, Gabiano C, Gattinara GC, Guarino A, et al. Virologic, immunologic, and clinical benefits from early combined antiretroviral therapy in infants with perinatal HIV-1 infection. AIDS 2006 Jan 9;20(2):207-15. Erratum in: AIDS. 2006 Aug 22;20(13):1789. Baddato, Raffaele [corrected to Badolato, Raffaele].