University of California, San Francisco Logo

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

Home > Global Health Literature Digest > Relationship between mortality and feeding
Relationship between mortality and feeding modality among children born to HIV-infected mothers in a research setting: the Kesho Bora Study
Global Health Sciences Literature Digest
Published June 17, 2013
Journal Article

Cournil A, de Vincenzi I, Gaillard P, Cames C, Fao P, Luchters S, Rollins N, Newell M-L, Bork K, Read JS, Kesho Bora Study Group. Relationship between mortality and feeding modality among children born to HIV-infected mothers in a research setting: the Kesho Bora Study. AIDS 2013; 27: 1621-30.


To assess the association between infant feeding practices and mortality by 18 months of age among children born to HIV-infected mothers in a three-country trial setting (Burkina-Faso, Kenya and South Africa).


Five clinical trials study sites in Bobo-Dioulasso, Burkina Faso; Mombasa and Nairobi, Kenya; and Durban and Somkhele, South Africa).

Study Design

Cohort study nested within a randomized controlled trial of extended maternal antiretroviral prophylaxis through cessation of breastfeeding compared to either breastfeeding exclusively up to 6 months of age or formula-feeding from birth without extended antiretroviral prophylaxis. Women were randomized to initiate either triple antiretroviral prophylaxis (combination of zidovudine (ZDV), lamivudine and lopinavir/ritonavir) until cessation of breastfeeding or the South African standard antiretroviral prophylaxis for MTCT (ZDV until delivery with single dose nevirapine [sdNVP] to women at the onset of labor and to the newborn infant). The authors note that non-exclusive breastfeeding was common, that many women breastfed for less than 6 months and that feeding practices within the first six months were highly heterogeneous, presumably in both arms.


Pregnant HIV-infected women presenting for antenatal care at <34 weeks gestation, World Health Organization (WHO) clinical stage 1, 2, or 3, and a CD4 cell count of 200-500 cells/µL and their infants. This analysis includes all women who were randomized, regardless of arm.

Intervention (Predictor Variable)

Breastfeeding practices defined as (1) never breastfed, weaned (stopped all breastfeeding) or still breastfed and (2) if still breastfeeding at the time of the interview, exclusively, predominantly or partially breastfed.

Main Outcome Measures

Infant mortality.


The authors conducted three analyses. First, they compared never breastfed with a fixed (non-time-varying) variable of ever breastfed infants. Secondly, they compared feeding practices (still breastfeeding, weaned and never breastfed) as a time-varying variable, recording mothers' reports at each visit. In a third analysis, they divided the category "still breastfeeding" into three subcategories: exclusively, predominantly or partially still breastfed.


There were 824 women enrolled in the Kesho Bora trial; 805 delivered live born infants. Of these 805 infants, six died before feeding initiation, two were lost to follow-up before recording information on the first feeding, and two died without a PCR test, leaving 795 mother-infant pairs for this analysis. Of these 795, 177 (22.3%) mothers never breastfed.

Sixty-three children died within 18 months of birth (4.5% by six months, 7.4% by 12 months, and 8.3% by 18 months). There was no significant difference in estimated cumulative mortality rate by 18 months (10.5% in the never breastfed group vs. 7.7% in the ever breastfed group, p=0.20 by log-rank test). HIV infection status of the child was strongly associated with the adjusted hazard of death (adjusted hazards ratio [aHR] 8.0, 95% CI 4.5 to 14.2).

Using time-dependent variables and comparing weaned versus still breastfeeding among HIV-uninfected children the aHR of mortality was 8.7 (95% CI 2.7 to 28.5), and the aHR of never breastfeeding versus still breastfeeding was 7.8 (95% CI 2.3 to 26.7). Among HIV-infected children the adjusted hazard of mortality for weaned versus still breastfeeding was 4.8 (95% CI 1.1 to 20.3); the adjusted hazard of mortality in never breastfed versus still breastfeeding was marginally significant (aHR 5.5, 95% CI 0.9 to 31.6, p=0.06). In the ZDV/sdNVP arm the excess of mortality among weaned (aHR 8.9, 95% CI 2.7 to 29.8) or never-breastfed children (aHR 8.7, 95% CI 2.4 to 31.1) compared with still breastfeeding was also observed but in the triple ARV arm only among weaned children (aHR 4.4, CI 1.1 to 17.8).

Gastroenteritis and pneumonia accounted for 71% of all causes of death in weaned and formula fed children, but gastroenteritis was never reported as a cause of death in children who were still breastfed at the time of death. Among ever breastfed children, weaning before 3 months was associated with a 3.5-fold increase in mortality (95% CI 1.7 to 7.3), whereas weaning at 3 to 4.9 months was marginally associated with a 2.2-fold increase (95% CI 0.9 to 5.4, p=0.08) compared with weaning after five months.


While the cruder analysis of ever breastfed versus never breastfed did not find a difference in mortality, using time-dependent variables for infant feeding practices children who had been weaned or never breastfed had about a 7-fold higher risk of death than those who were still being breastfed. This hazard was most pronounced in the first three months of life and then less so in the fourth and fifth. Importantly, breastfeeding is as protective in HIV-infected children as in uninfected children and is protective regardless of whether the mother receives ongoing antiretroviral prophylaxis throughout breastfeeding or not.

Risk of Bias

The risk of bias in this study is moderate. While the Kesho Bora trial was an RCT, this analysis is an observational sub-study in which women elected their own feeding practices. The investigators did an excellent job of examining both fixed (ever vs. never breastfed) and time-depending variables and for controlling for confounding through a variety of multivariable adjustments. Follow-up in this cohort was high.

In Context

Previous studies have examined type of feeding and mortality among HIV-exposed children and have shown that, at best, formula feeding results in no benefit in survival(1, 2, 3, 4, 5, 6) and, at worst, increases mortality 6-fold by 12 months of age.(7, 8) The findings in this study confirm earlier observations of a continuous adverse effect of receiving replacement feeding after cessation of breastfeeding.(9) Of particular note is that aHRs were similar in weaned and never breastfed children.

Programmatic Implications

Breastfeeding is an important intervention for HIV-exposed infants (and indeed all infants), particularly in the first six months of life, and replacement feeding or early weaning during this period carries a significant risk of increased mortality. The clinical trials setting of Kesho Bora provided extensive counseling and support for breastfeeding mother, and, while antenatal clinics may not be able to offer similar services, interventions to support HIV-infected mothers to exclusively breastfeed their children, especially as the provision of antiretroviral prophylaxis during lactation becomes increasingly accessible, will be key to promoting HIV-free survival among these vulnerable infants.


  1. Kuhn L, Aldrovandi GM, Sinkala M, et al. Effects of early, abrupt weaning on HIV-free survival of children in Zambia. N Engl J Med 2008; 359:130-41.
  2. Becquet R, Bequet L, Ekouevi DK, et al. Two-year morbidity-mortality and alternatives to prolonged breast-feeding among children born to HIV- infected mothers in Cote d'Ivoire. PLoS Med 2007; 4:e17.
  3. Mbori-Ngacha D, Nduati R, John G, et al. Morbidity and mortality in breastfed and formula-fed infants of HIV-1-infected women: a randomized clinical trial. JAMA 2001; 286:2413-20.
  4. 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; 364:1236-43.
  5. Peltier CA, Ndayisaba GF, Lepage P, et al. Breastfeeding with maternal antiretroviral therapy or formula feeding to prevent HIV postnatal mother- to-child transmission in Rwanda. AIDS 2009; 23:2415-23.
  6. Rollins NC, Becquet R, Bland RM, Coutsoudis A, Coovadia HM, Newell ML. Infant feeding, HIV transmission and mortality at 18 months: the need for appropriate choices by mothers and prioritization within programmes. AIDS 2008; 22:2349-57.
  7. Homsy J, Moore D, Barasa A, et al. Breastfeeding, mother-to-child HIV transmission, and mortality among infants born to HIV-Infected women on highly active antiretroviral therapy in rural Uganda. J Acquir Immune Defic Syndr 2010; 53:28-35.
  8. Kagaayi J, Gray RH, Brahmbhatt H, et al. Survival of infants born to HIV- positive mothers, by feeding modality, in Rakai, Uganda. PLoS One 2008; 3:e3877.
  9. Kuhn L, Sinkala M, Semrau K, et al. Elevations in mortality associated with weaning persist into the second year of life among uninfected children born to HIV-infected mothers. Clin Infect Dis 2010; 50:437-444.