Tuberculin reaction, BCG scar, and lower female mortality.

BACKGROUND: Recent studies have suggested that bacille Calmette-Guérin (BCG) immunization may have a nonspecific beneficial effect on infant survival and that the effect may be more pronounced among girls. In a prospective birth cohort, we examine whether a positive tuberculin skin test and BCG scar in response to BCG immunization were related to better overall survival in Guinea-Bissau and, if so, whether the effect was sex-specific. METHODS: Skin tests and BCG scarring were monitored at ages 2 months (n = 2332) and 6 months (n = 1817) in children born from March 2000 to July 2002. A tuberculosis (TB) surveillance system allowed us to exclude from the analysis children with likely TB exposure. The children were followed for survival until 18 months of age. RESULTS: Among children with a tuberculin skin test at 2 and 6 months of age, the mortality rate ratio for skin test reactors (>1 mm) versus nonreactors (0-1 mm) was 0.54 (95% confidence interval = 0.30-0.99). Comparing children with and without a BCG scar, the ratio was 0.55 (0.31-0.96). The effect of a skin test reaction or a BCG scar seemed stronger among girls; for those with positive reaction, the mortality ratio was 0.31 (0.11-0.88) among girls and 0.84 (0.39-1.82) among boys; and for BCG scar, the results were 0.41 (0.21-0.82) and 0.88 (0.34-2.30), respectively. CONCLUSIONS: A good response to BCG vaccination is related to lower child mortality. The effect seems most pronounced among girls. The findings may have implications for future vaccine trials and policy.

Vaccination technique, PPD reaction and BCG scarring in a cohort of children born in Guinea-Bissau 2000-2002.

The rates of positive tuberculin skin test (TST) reactions and BCG scarring after BCG vaccination vary between studies and populations. Tuberculin reactivity and BCG scarring may be related to better child survival in low-income countries. We therefore studied determinants for TST reaction and scarring in Guinea-Bissau. In a cohort of children born in suburban Bissau from March 2000 to July 2002, we assessed a Mantoux test with Purified protein derivative (PPD) (SSI, 2 T.U.) at 2 (2689 children), 6 (N=2148) and 12 months (N=1638) of age, and BCG scar was assessed at 2 (N=2698) and 6 months (N=2225) of age. In a subgroup of the children the vaccination technique was monitored by direct observation of post-vaccination wheal and route of administration. Three different types of BCG vaccine supplied by the local Extended Programme on Immunization were used. At 6 months of age the rate of PPD reactors (>1mm) after BCG vaccination was 25% and the rate of scarring was 89%. One BCG strain was associated with fewer PPD reactors (OR=0.54 (0.31-0.91)) and BCG scars (OR=0.13 (0.05-0.37)) and larger post-vaccination wheals produced more PPD reactions (OR 1.21 (95% CI 1.02-1.43)) and BCG scars (OR 1.66 (1.24-2.21)). In the multivariable analyses of BCG-vaccinated children assessed at 6 months of age, monitoring of vaccination technique and type of BCG vaccine were important. This was not changed by control for other determinants, including sex, season, vaccination place, birthplace, ethnic group, low birth weight, place of residence, education and civil status of mother. We reason that vaccination technique and BCG strain are important for PPD reaction and scarring in response to BCG vaccination. Considering that these responses are associated with better infant survival, the importance of monitoring vaccination technique and of different BCG strains should be evaluated with respect to infant mortality.