Meta-analysis was used to determine whether maintenance intravesical bacillus Calmette-Guerin (BCG) could reduce recurrence after transurethral resection of tumor 1 grade 3 (T1G3) superficial bladder cancer. All available published data of randomized clinical trials comparing transurethral resection plus intravesical BCG to either resection alone or resection plus another treatment on the treatment results in patients with superficial bladder cancer of T1G3 were selected for analysis. Both the fixed effects model and random effects model were applied, and the odds ratio (OR) with its 95% confidence interval (CI) was used as the effect size estimate. Sensitivity analysis and publication bias determination were performed by funnel plots and comparing ORs of different models. Within the follow-up period, 375 of 915 (41.0%) BCG-treated patients and 332 of 733 (45.3%) non-BCG-treated patients developed tumor recurrence. In the combined results, a statistically significant difference in the ORs for tumor recurrence between the two treatment groups was found (randomized model combined effect OR 0.58, 95% CI 0.41 to 0.83, P 5 0.003). The stratified meta-analysis did not show any statistically significant confounding effects on the results when stratified by BCG strains. The randomized model combined effect OR of Pasteur F and other strains were 0.50 (95% CI 0.26 to 0.95, P 5 0.04) and 0.63 (95% CI 0.40 to 0.99, P 5 0.04), respectively. Therefore, we came to the conclusion that adjuvant maintenance instillation BCG combined with transurethral resection of bladder tumor (TURBT) is an effective conservative treatment for preventing recurrence of T1G3 bladder cancer.