Intravesical Bacillus Calmette-Guérin induction followed by maintenance remains the standard adjuvant treatment for patients with high-risk non–muscle-invasive bladder cancer (NMIBC). However, up to 40% of patients may not respond adequately, prompting investigation of systemic immune checkpoint inhibitors, additional intravesical agents, and modified BCG schedules.
A systematic review and meta-analysis has now compared the efficacy and safety of these strategies in patients with BCG-naïve high-risk non–muscle-invasive bladder cancer. The findings suggest that adding systemic immune checkpoint inhibition to adequate BCG induction and maintenance may modestly reduce recurrence, but this benefit is sensitive to differences among trials and is accompanied by a substantial increase in treatment-related toxicity.
The article, titled “Novel BCG Combination or Modified BCG Regimens in BCG-Naïve High-risk Non–muscle-invasive Bladder Cancer: A Systematic Review and Meta-analysis,” was published online as an article in press in European Urology Focus on July 15, 2026.
Authors: Maxime Pattou, Morgan Rouprêt, Ashish Kamat, Stephen A. Boorjian, Paolo Gontero, Shahrokh F. Shariat, and David D’Andrea.
Study Design
The investigators searched PubMed, Embase, and Web of Science through October 20, 2025. The review followed PRISMA 2020 guidelines, and the protocol was registered in PROSPERO (CRD420251079219).
Eligible studies were phase 3 randomised controlled trials enrolling adults with BCG-naïve high-risk non–muscle-invasive bladder cancer. Trials compared standard BCG induction and maintenance with BCG-based combination treatment or a modified BCG schedule and were required to report recurrence-free, event-free, or disease-free survival outcomes, together with adverse events graded according to the Common Terminology Criteria for Adverse Events.
The investigators performed random-effects meta-analyses and a frequentist network meta-analysis. The primary outcome was tumour recurrence, with secondary outcomes including event-free or disease-free survival, adverse events, treatment discontinuation, and BCG maintenance completion.
Five phase 3 trials involving 3,486 patients met the eligibility criteria. Three evaluated systemic immune checkpoint inhibitors combined with BCG, one compared mitomycin C plus BCG with BCG alone, and one assessed a reduced-frequency BCG schedule.
The included trials were CREST, POTOMAC, ALBAN, ANZUP 1301, and NIMBUS.
Overall, 82% of participants were male, and the median age was 68 years. High-grade pTa and pT1 disease accounted for 43% and 46% of cases, respectively, while 31% of patients had concurrent carcinoma in situ. The sample size–weighted average of the study-reported median follow-up durations was 45 months, with an interquartile range of 26–54 months.
Event-free and Disease-free Survival
Across the evaluated combination strategies, pooled fixed-time estimates of event-free or disease-free survival were numerically higher than with standard BCG induction and maintenance.
At 1 year, the pooled event-free or disease-free survival estimate was 87% with combination treatment (95% CI 85–89) and 84% with BCG induction and maintenance alone (95% CI 82–87). At 2 years, the corresponding estimates were 82% (95% CI 79–85) and 78% (95% CI 75–81).
However, the trials used different definitions of event-free and disease-free survival and had different follow-up durations. A pooled hazard-ratio analysis was therefore not considered feasible.
High-grade Recurrence
Among the three trials evaluating immune checkpoint inhibition plus BCG, the absolute risk difference at a weighted median follow-up of 45 months was 5%, with a 95% confidence interval of 2–9%. Because the trials differed in recurrence definitions and treatment delivery, recurrence results from all five studies were described rather than pooled into a single overall estimate.
In trials using centrally adjudicated high-grade recurrence, immune checkpoint inhibition combined with BCG induction and maintenance reduced the risk of high-grade recurrence compared with BCG alone, with a relative risk of 0.63 and a 95% confidence interval of 0.41–0.98.
In the network meta-analysis, immune checkpoint inhibition plus BCG induction and maintenance ranked highest for preventing high-grade recurrence, with a P-score of 0.98. However, the estimate was not statistically significant, with a relative risk of 0.75 and a 95% confidence interval of 0.55–1.03.
When the separately reported ALBAN high-grade recurrence subgroup was incorporated, the effect remained directionally favourable but was attenuated, with greater heterogeneity. The pooled relative risk was 0.75, with a 95% confidence interval of 0.50–1.12.
The investigators identified ALBAN as the principal source of between-study variability because it used a different primary endpoint and a shorter BCG maintenance schedule than CREST and POTOMAC.
Findings According to Disease Subgroup
Results differed according to disease characteristics and individual trial. For patients with pT1 disease, combination treatment was associated with improved outcomes in POTOMAC and CREST but not in ALBAN. The pooled hazard ratio for high-grade recurrence in patients with pT1 disease was 0.58, with a 95% confidence interval of 0.43–0.79.
Findings were less consistent among patients with carcinoma in situ. CREST demonstrated an event-free survival benefit, whereas POTOMAC and ALBAN did not show statistically significant improvements. The pooled hazard ratio for high-grade recurrence among patients with carcinoma in situ was 0.75, with a 95% confidence interval of 0.40–1.40.
Complete response findings also varied across the trials. At 1 year, investigator-assessed complete response rates in CREST were 90% with sasanlimab plus BCG induction and maintenance, 88% with sasanlimab plus BCG induction only, and 85% with BCG induction and maintenance.
In POTOMAC, the corresponding rates were 93% with durvalumab plus BCG induction and maintenance, 83% with durvalumab plus BCG induction only, and 93% with BCG induction and maintenance.
Increased Toxicity With Systemic Immunotherapy
The potential recurrence benefit was accompanied by substantially greater toxicity. In the detailed analysis restricted to the immune checkpoint inhibitor trials, all-grade treatment-related adverse events occurred in 85% of patients receiving immune checkpoint inhibition plus BCG and 72% of those receiving BCG alone. The pooled relative risk was 1.14, with a 95% confidence interval of 1.06–1.23. Grade 3 or higher treatment-related adverse events occurred in 22% of patients receiving immune checkpoint inhibition plus BCG and 6% of those receiving BCG alone. The pooled relative risk was 3.15, with a 95% confidence interval of 1.94–5.12.
Within the subgroup receiving an immune checkpoint inhibitor together with BCG induction and maintenance, the relative risk of grade 3 or higher treatment-related adverse events was 3.98, with a 95% confidence interval of 2.54–6.24, compared with standard BCG induction and maintenance.
In the overall analysis reported in the abstract, grade 3 or higher treatment-related adverse events occurred in 25% of combination-treated patients, compared with 6% of patients receiving BCG alone.
Treatment discontinuation due to treatment-related adverse events was also more frequent with treatment intensification. Discontinuation rates were 33% with immune checkpoint inhibition plus BCG induction and maintenance, 18% with immune checkpoint inhibition plus BCG induction only, 25% with mitomycin C plus BCG, and 20% with standard BCG induction and maintenance.
Among patients receiving an immune checkpoint inhibitor, the overall rate of immune checkpoint inhibitor withdrawal because of treatment-related adverse events was 18%, with a 95% confidence interval of 13–23%.
Importance of BCG Maintenance
The analysis reinforced the importance of adequate BCG induction and maintenance. Reduced or omitted BCG instillations were associated with inferior oncological outcomes, even when systemic immune checkpoint inhibition was added. In CREST and POTOMAC, the benefit of combination treatment was observed in the arms that retained BCG maintenance, whereas the arms using BCG induction without maintenance did not demonstrate the same benefit.
The findings therefore suggest that systemic immune checkpoint inhibition cannot compensate for inadequate local immune priming with BCG.
Study Limitations
The analysis was limited by differences in treatment regimens, endpoint definitions, follow-up durations, diagnostic criteria, and BCG maintenance schedules.
Several important prognostic factors were also incompletely reported, including variant histology, lymphovascular invasion, PD-L1 expression, and the quality of transurethral resection of the bladder tumour. These limitations restricted direct comparisons across trials and contributed to between-study heterogeneity.
Clinical Implications
The results do not support routine systemic immunotherapy intensification in unselected patients with BCG-naïve high-risk non–muscle-invasive bladder cancer.
Although adding systemic immune checkpoint inhibition to adequate BCG induction and maintenance may provide a modest, heterogeneity-sensitive recurrence benefit, this must be weighed against a marked increase in severe treatment-related adverse events and treatment discontinuation. The authors also highlighted the additional treatment burden and higher costs, while no overall survival advantage has been demonstrated.
The authors concluded that standard BCG induction and maintenance should remain the reference regimen. Immune checkpoint inhibitor–BCG combinations should be reserved for clinical trials or select patients with the highest-risk disease biology, pending biomarker-guided patient selection and mature survival data.
The full article is available on the European Urology Focus.