Risks of breastfeeding not confined to HIV transmission

This article is more than 22 years old.

According to a study involving HIV-positive women in Nairobi, maternal survival may be adversely affected by breastfeeding their newborn baby compared to bottle feeding. In a large randomised, controlled trial designed to compare the risk of mother to baby HIV transmission in women who did not receive antiretroviral treatment, women who breastfed were three times more likely to die during the two years following delivery than women who used formula feed.

This stark finding emerged from a new analysis of a Kenyan trial for which the main results were published earlier this year in the Journal of the American Medical Association.

The study randomised 425 women to either breastfeed (n=212) or formula feed (n=213) their baby. At entry, average CD4 count was around 400 cells in each arm, and average viral load around 40,000 copies. This new analysis of the impact of breastfeeding on maternal health included 200 women who bottle fed, and 197 who breastfed. Of these, 44 women were lost to follow-up during the course of the study. Women lost to follow-up tended to be of lower socio-economic status.


loss to follow up

In a research study, participants who drop out before the end of the study. In routine clinical care, patients who do not attend medical appointments and who cannot be contacted.


Abnormal bowel movements, characterised by loose, watery or frequent stools, three or more times a day.


An alternative term for ‘adherence’.

Compliance to formula feed was 71% and over 90% to breastfeeding. This latter rate fell over the course of the study however as children were weaned. By 12 months, 80% of women in the breastfeeding arm were still breastfeeding their children. Breastfeeding continued for 17 months on average, and the average time at which supplements were introduced was 3.8 months.

After 24 months, 18 of 197 (10.5%) breastfeeding women had died, and 6 of 200 (3.8%) women who bottle fed. The difference in survival rates between the two arms was significant as early as six months after delivery however. Women who breastfed lost more weight than women who used formula feed, and weight loss was associated with an increased risk of death.

Infant survival was found to be dependent on maternal survival. Whilst any maternal mortality was significantly associated with infant mortality, maternal death was associated with a 7.8 times increase in the risk of subsequent infant death even when controlling for the infant’s HIV infection status.

Impact of breastfeeding on children

A separate presentation, also delivered at the 13th International AIDS Conference in Durban today, focussed on the health and survival of the infants born to mothers participating in this same Kenyan study. This analysis found no difference in the mortality rate between formula and breastfed infants over the two years of follow-up.

There were no differences in diarrhoea incidence, or in the incidence of other morbidities, though there was more diarrhoea and dehydration in the formula feed arm during the first three months of life.

Whilst the data provide encouragement over the use of formula feeding in similar resource-poor settings, healthcare workers and policy makers continue to face a dilemma over advice on breastfeeding. Amongst uninfected infants, over the first three months of life, the risk of death was greater in those children who were bottle fed compared to breastfed children.


Mbori-Ngacha D et al. Morbidity and mortality in breastfed and formula fed infants of HIV-1 infected women: results of a randomized clinical trial. 13th International Conference on AIDS, abstract WeOrC494, Durban, 9-14 July 2000.

Nduati R et al. Impact of breastfeeding on maternal mortality among HIV-1 infected women: Results of a randomized clinical trial. 13th International Conference on AIDS, abstract WeOrC495, Durban, 9-14 July 2000.

Nduati R et al. Effect of breastfeeding and formula feeding on transmission of HIV-1: a randomized clinical trial. Journal of the American Medical Association 283(9):1167-1174, 2000.