BACKGROUND: It is not clear whether maintenance Bacillus Calmette-Guerin (BCG) is necessary for intermediate- or high-risk non-muscle-invasive bladder cancer (NMIBC). This systematic review and meta-analysis aimed to illustrate the effects of maintenance BCG for intermediate- or high-risk NMIBC. METHODS: A comprehensive literature search of PubMed, EMBASE, Cochrane Library, ClinicalTrials.gov databases and International Clinical Trials Register (ICTRP) Search was conducted to identify relevant randomized controlled trials (RCTs) that have assessed the efficacy of maintenance or non-maintenance BCG therapy for patients with NMIBC. The maintenance group first received induction BCG instillations, and subsequently received BCG intravesical instillations regularly for at least 1 year, while the control group only received induction BCG instillations. Systematic review and meta-analysis were performed according to Preferred Reporting Items for Systematic Reviews and Meta-analysis Criteria. RESULTS: Ten RCTs were eligible in this systematic review. The meta-analysis showed that induction BCG followed by maintenance BCG instillation after transurethral resection (TUR) could reduce the risk ratios of tumor recurrence by 21% (RR = 0.79; 95% CI 0.70-0.89; P < 0.0001) and prolong recurrence-free survival (RFS) by 33% (HR: 0.67; 95% CI, 0.54-0.82; P < 0.001), compared with non-maintenance BCG. It could also reduce the risk ratios of tumor progression (RR = 0.81; 95% CI 0.68-0.97; P = 0.02). However, these pooled results should be considered with caution since the quality of evidences for outcomes ranged low. Subgroup analysis implied that different durations of maintenance BCG instillations might be one of the sources of potential clinical heterogeneity of included studies. Begg's funnel plot and Egger's test did not reveal any evidence of publication bias in this meta-analysis. CONCLUSIONS: Induction BCG followed by maintenance BCG instillation after TUR, compared with induction BCG along, can reduce the risk ratios of tumor recurrence and tumor progression, and prolong RFS. However, these results with a lower level of evidence should be treated with caution. The optimal maintenance schedule has yet to be determined and a large multi-institutional study in intermediate- and high-risk patients is also needed to determine the optimal maintenance schedule.