Between the derivation of bacille Calmette-Guérin (BCG) vaccine in 1921 and the lyophilization of BCG daughter strains in the 1960s, a number of clinical trials were performed looking at the protective efficacy of BCG vaccination against tuberculosis. These trials differed from one another in a number of ways: they employed different methodologies for delivering the vaccine and interpreting outcomes; they were performed on populations with different genetic backgrounds and different levels of exposure to environmental Mycobacteria; and, finally, they used different BCG vaccine strains. The results of these trials were estimates of protective efficacy against pulmonary tuberculosis ranging from about 80% to nil. Because of the differences in outcomes and confounding variables, it is difficult to conclude whether differences in interventions alone may have contributed to the remarkably variable results. Analysis of BCG vaccines used in clinical trials suggests a trend towards decreasing efficacy with increased passage in the laboratory; however, trials that used relatively "older" BCG strains were generally performed at different sites than trials which used "younger" BCG strains. Genomic analysis of BCG vaccines demonstrates that during the half-century of ongoing passage of BCG vaccines in vitro there have been numerous genetic changes, comprising single nucleotide polymorphisms, duplications and deletions. The impact of these changes in the BCG genome on the protective efficacy observed in field trials remains to be determined.