The investigators aim to conduct a randomised controlled trial comparing two Bacille Calmette-Guérin (BCG) strains currently used in Guinea-Bissau, the Danish and the Russian, in terms of prevention of neonatal and early life morbidity and mortality, immune responses and adverse events related to BCG vaccination. The primary outcome will be hospital admissions within 6 weeks of age.

BACKGROUND At the Bandim Health Project (www.bandim.org) in Guinea-Bissau, we observed that Bacille Calmette-Guérin (BCG) vaccination at birth is associated with survival benefits which cannot be explained by prevention of tuberculosis (TB), which is rare during the first year of life. There is mounting evidence that the beneficial effects of BCG on neonatal mortality stems from non-specific beneficial immune training, also termed heterologous immunity. This way, BCG immunisation possibly confers enhanced protection against a broad range of severe infections such as pneumonia and septicemia.

In addition, observational studies from Guinea-Bissau suggest that BCG-vaccinated children who produced a scar upon immunisation have significantly better survival than BCG-vaccinated children who did not produce a scar. The pattern is the same with regard to PPD (tuberculin) skin test. While correct BCG-vaccination technique is undoubtedly important for scarring, several studies have shown that the strain of BCG might be important as well.

Globally, BCG coverage exceeds 90%, making BCG the most widely used vaccine in the world. According to UNICEF, the annual demand for the BCG vaccine was 130 million doses in 2013. Several different vaccine strains are being used, yet few studies have compared the different strains. However, several animal and human studies have shown that these genetically diverse vaccine strains induce different protective efficacy against TB, risk of adverse events and susceptibility to anti-TB drugs. Recently, a large observational study based in Uganda showed that there were significant differences between BCG strains concerning the response to specific mycobacteria and non-specific immune responses. In particular, BCG-Denmark was associated with a much higher rate of scarring (93%) than BCG-Russia (52%), but also a higher rate of adverse events in terms of ulcers and abscesses (1.8% versus 0.3%).

OBJECTIVE We aim to compare the effect of two strains of BCG (BCG-Denmark, BCG-Russia) provided at birth to children born at the National Hospital Simão Mendes (HNSM) on subsequent hospitalisations. Furthermore, mortality in the first 6 weeks and adverse events at 2 and 6 months of age will be reported for all children, and BCG scar frequency (2 and 6 months) and PPD response (6 months) will be reported for a proportion of the infants.

METHODS The study will be conducted by the Bandim Health Project (BHP). The BHP maintains a health and demographic surveillance system (HDSS) in 6 districts of the capital in Guinea-Bissau covering approximately 102,000 inhabitants. All houses in the HDSS are visited monthly and all pregnancies and births are registered.

Study participants will be enrolled at the National Hospital and followed up through the HDSS and by telephone interviews as well as through hospital records. The study will be individually randomised, with 1:1 randomisation between the two strains (BCG-Denmark, BCG-Russia). During a two year inclusion period, we expect to be able to include at least 12,000 children, i.e. 6,000 in each BCG strain group. Of these, around 25% will reside within the HDSS.

Inclusion Criteria:

Newborn infants at the HNSM maternity ward.

Exclusion Criteria:

Infants included in another randomized trial of BCG. Infants with a severe congenital abnormality.