The relative persistence capacity in mouse spleen of 10 and 9 BCG substrains from liquid and dried vaccines, respectively, was evaluated in two studies. Recoverable BCG colony counts from mouse spleen were determined at given days on solid medium in the two studies during a period of 1–360 and 1–345 days, respectively, after the intravenous BCG vaccination, performed with two different viable units. From 36 000 (study 1) and 21 600 (study 2) recoverable BCG colony counts, 180 and 108 mean relative persistence capacity values were estimated to test the residual virulence during the follow-up time, using computerized statistical analysis. The early and late trends of mean relative persistence capacity of the BCG substrains in mouse spleen were tested by linear regression analysis and analysis of variance and covariance; then with ranked adjusted group mean relative persistence capacity, Gabriel's simultaneous test procedure was performed for multiple comparison to diminish type 1 error in statistical inference and in objective interpretation of the experimental results. The associations of the ranked mean relative persistence capacity of the BCG substrains at the different sacrifice days of mice were also analyzed by Kendall's test of concordance. The early, late, and overall relative persistence capacity reflects the residual virulence of the BCG substrains and provides information on the required protective efficacy (immunogenicity) and adverse reactions (reactogenicity), allowing the appropriate vaccination dose, expressed in viable units of the substrain used, to be determined. Key words: BCG substrains, residual virulence, relative persistence in mouse spleen, multiple comparison.
REINFORCER EFFICACY IN A DELAYED MATCHING-TO-SAMPLE TASK