Exclusive breastfeeding is promoted by the World Health Organization (WHO) with the aim of improving infant health. The specific objective of the study was to examine the effect of exclusive breast-feeding on morbidity and mortality in a randomised prospective study. Children born by mothers recorded as living in the study area during pregnancy was randomised at birth. WHO recommendations to postpone introduction of water and weaning food were told to the mother by 2-weekly home visits. All study children were followed from birth till 1 year of age according to morbidity, hospitalisation and mortality.

Examine in a randomised trial whether postponing of introduction of water and weaning food as recommended by WHO reduces infant morbidity and mortality. Results obtained from observational studies are likely to be due to reverse causality as the mother her self chooses when to introduce water and weaning food. Only a few randomised studies have examined the effect on infant morbidity, and no study has investigated the impact on infant mortality in a low-income country. The study was conducted at Bandim Health Project (BHP) which has conducted a demographic health surveillance system since 1978. Mothers and children were identified by the BHP routine registration. All children born between March 1 2000 until February 28th 2001 in the study area and whose mothers were recorded as living in the area during pregnancy were eligible for inclusion. Children who died within 7 days of birth were excluded in order to avoid birth related morbidity and mortality leaving a total of 1300 children for the study. At birth all children were allocated to an intervention group or a control group. All mother and child pairs were visited bi-weekly by a local, special trained health worker. By individual health education the mothers were motivated to delay introduction of water and weaning food to 6 months of age according to recommendations given by WHO. All mother and child pairs in the control group were visited by the same health assistant. By these visits only information on feeding status was obtained and no intervention given. Further more all study children were followed by assistants blinded to the randomisation with: bi-weekly diarrhoea morbidity interview, monthly anthropometry, including weight, height and arm circumference. All children were followed up to 12 months of age. Demographic as well as socio-economic data, data on hospitalisation, deaths and migrations were obtained from the BHP routine registration system. These data will be collected until the age of 3 years of the child.

Inclusion Criteria:

• Children born in the study area in the period 1st March 2000 till 28th February 2001 by mothers recorded as living in the area during pregnancy.

Exclusion Criteria:

• Death before 7 days of age