Bacillus Calmette–Guérin (BCG) vaccine against tuberculosis (TB) is recommended at birth in TB-endemic areas. Currently, BCG vaccination programmes use “BCG vaccination coverage by 12 months of age” as the performance indicator. Previous studies suggest that BCG-vaccinated children, who develop a scar, have better overall survival compared with BCG-vaccinated children, who do not develop a scar. We summarized the available studies of BCG scarring and child survival. A structured literature search for studies with original data and analysis of BCG scarring and mortality were performed. Combined analyses on the effect of BCG scarring on overall mortality. We identified six studies covering seven cohorts, all from Guinea-Bissau, West Africa, with evaluation of BCG scarring amongst BCG-vaccinated children and follow-up for mortality. Determinants of BCG scarring were BCG strain, intradermal injection route, size of injection wheal, and co-administered vaccines and micronutrients. In a combined analysis, having a BCG scar vs. no BCG scar was associated with a mortality rate ratio (MRR) of 0.61 (95% CI: 0.51–0.74). The proportion with a BCG scar varied from 52 to 93%; the estimated effect of a BCG scar was not associated with the scar prevalence. The effect was strongest in the first (MRR = 0.48 (0.37–0.62)) and second (MRR = 0.63 (0.44–0.92)) year of life, and in children BCG-vaccinated in the neonatal period (MRR = 0.45 (0.36–0.55)). The effect was not explained by protection against TB. Confounding and genetic factors are unlikely to explain the strong association between BCG scarring and subsequent survival. Including “BCG scar prevalence” as a BCG vaccination programme performance indicator should be considered. The effect of revaccinating scar-negative children should be studied.