The World Health Organization (WHO) currently recommends BCG vaccination at birth in developing countries. Pre-term infants should be vaccinated when they reach the chronological age of 40 weeks. Due to difficulties in establishing the correct gestational age, the vaccination policy for BCG in many developing countries is defined by birth weight rather than by gestational maturity. In the study area, low birth weight (LBW) infants (< 2500 g) are not supposed to be vaccinated at birth; instead the mother is asked to return for vaccination when the child has gained sufficient weight. BCG has marked immune stimulatory effects in both animal and human studies and observational studies suggest that BCG is associated with a non-specific reduction in mortality in areas with high infant and child mortality. The specific objective of the study is to examine the effect of early vaccination of LBW children for adverse events, purified protein derivative of tuberculin (PPD) reaction, scar size, morbidity, and mortality in a randomised prospective study of BCG vaccination at birth versus later (according to policy) among children 19 months of age in Guinea-Bissau. The hypothesis is that BCG vaccination of low birth weight (LBW) children at birth reduces infant mortality of this high-risk group by 25%.

Criteria for verification: 1,600 low birth weight (LBW) infants recruited at the national hospital and local health centres and enrolled in a randomised trial of BCG vaccination at birth versus later vaccination; information on potential adverse events following early BCG vaccination of LBW infants; follow-up of children to 12 months of age with home visits at 2, 6 and 12 months of age; assessment of tuberculin reaction and scar-formation at 2, 6 and 12 months of age; assessment of BCG vaccination coverage during infancy; and assessment of all-cause mortality during infancy.

Recruitment: All LBW children delivered at the maternity ward of the national hospital at the local Health Centre will be offered enrolment in the study. Furthermore, all children born at home in the study area will be weighed and referred for BCG vaccination at the health centres as quickly as possible. For mothers of LBW children, the current policy and the purpose of the study will be explained. If she accepts participation, she will draw a randomisation number indicating whether the children will be BCG vaccinated early or not. At the time of enrolment, children will be examined clinically with anthropometrics, and major risk factors will be documented (including TB exposure at home, recent infections, access to malaria drugs, ethnic group, schooling, housing conditions, mother's recent drug consumption). Newborns will only be included in the study if they are considered sufficiently healthy to be discharged from the hospital, or when they arrive at the health centre if born at home. Twins will be a large part of the LBW children as twinning is common in the study area (around 2% of births). They will be allocated to the same group, as there could be confusion for the mother as to who had been vaccinated. As is happening at the moment, mothers of children not receiving vaccination at birth will be recommended to attend a local health centre as soon as the child has gained weight. For ethical reasons, it will not be possible to use a placebo as this could mean that some children would not be BCG vaccinated if the mother believed that the child had in fact been vaccinated. The children who are not vaccinated initially will subsequently be vaccinated at the limited number of health centres in the capital, which administer BCG vaccination. The intradermal vaccination technique and vaccine storage will be supervised, and it will be monitored as to whether scar-formation and PPD reactivity is similar in children vaccinated at the different centres to assure that vaccinations have been administered in the same way.

Follow-up: To assure proper identification of the children, the mother and newborn child will be driven home to document the location of the home. The children included in the study will be visited at 2, 6, and 12 months of age, to identify subsequent BCG vaccinations by inspection of the vaccination card, to weigh the child and measure arm-circumference, to assess BCG scarring and to document morbidity, hospitalisations and survival. Sick children identified during these visits will be referred for treatment at the relevant health institution. If the child has still not been vaccinated, the mother will be encouraged to bring the child for vaccination. Should the child have died, a verbal autopsy will be conducted. The children will be followed for morbidity, hospitalisations, and mortality up to 12 months of age. Within the study area they may be followed longer. In order to assure detection of possible complications to BCG vaccination, LBW children enrolled in the study will be provided with a card identifying the child, and the mother will be told to bring the child for consultation at the local Health Centre where consultations and treatment will be free of charge. The expected very rare complications will be treated according to WHO's recommendations.

A subgroup of the children will be enrolled in immunological studies. To assess the effect of BCG vaccination on the immune profile, initially unvaccinated LBW children will have a blood sample collected when they turn up for BCG vaccination at a health centre. For each child examined, a BCG vaccinated LBW child matched for age and sex will be examined and a blood sample collected.

Inclusion Criteria:

• Must weigh less than 2500 g at birth
• Must be able to nurse
• Must be healthy enough to be discharged from the hospital

Exclusion Criteria:

• Overt illness
• Malformation