Early breast-feeding cessation and infant mortality in low-income countries: workshop summary.

The importance of breast-feeding for infant and child health and survival in less developed countries has been the subject of a number of studies over the last 25 years. However the epidemic of AIDS in low-income countries together with the discovery of an important risk of mother-to-child transmission of HIV-1 through breast-milk has prompted renewed interest in this subject. Indeed in order to weigh the mortality risks associated with different infant feeding practices against the risk of mother-to-child transmission of the virus it is essential to have precise estimates of relative risk of death of non-breastfed infants within narrow age intervals. Available decision guides mostly use relative mortality risks published by a WHO working group based on 6 retrospective or prospective studies among which 3 contributed to risk estimations during infancy. The mortality risk was not analysed by mode of breast-feeding (i.e. exclusive predominant or partial) because such data were not available for all studies included. In addition exclusive breast-feeding is rare in many settings and sample sizes are therefore insufficient to assess mortality risks associated with this mode of feeding. This is unfortunate since a cohort study in South Africa suggested that exclusive breast-feeding up to 3 months postpartum may be associated with a lower risk of postnatal HIV-1 transmission to the child than when other fluids or foods are added. As for all observational epidemiological studies those dealing with the association between child mortality and breast-feeding may be subject to a number of methodological problems. The objective of this workshop was to illustrate some of these using examples from the literature and mostly consider confounding reverse causality and bias by indication. (excerpt)