Background Recurrent TaT1 non-muscle-invasive bladder cancer (NMIBC) patients should be treated with immediate instillation of chemotherapy after transurethral resection of bladder tumour followed by instillation therapy. Objective To present long-term results of a study exploring the effect of initial mitomycin C (MMC) instillations followed by two types of immunotherapy for patients with frequently recurring NMIBC. Design, setting, and participants Between 1992 and 1996, 236 patients with frequently recurring TaT1 grade 1-2 NMIBC were enrolled in the prospective randomised multicentre FinnBladder-4 study. Intervention One perioperative plus four weekly instillations of MMC followed by monthly bacillus Calmette-Guérin (BCG) or alternating BCG and interferon (IFN)-α2b instillations for up to 1 yr. Outcome measurements and statistical analysis Primary end points were time to first recurrence and time to progression. Secondary end points were disease-specific mortality and overall survival. The principal statistical methods were the proportional subdistribution hazards model and Cox proportional hazards model plus cumulative incidence and Kaplan-Meier analyses. Results and limitations The median follow-up was 10.3 yr (maximum: 19.8 yr) in the MMC-BCG group and 8.6 yr (maximum: 19.8 yr) in the MMC-BCG/IFN group. The probability of recurrence was significantly lower in the MMC-BCG group than in the MMC-BCG/IFN group (43% vs 78% at 10 yr and 45% vs 80% at 15 yr, respectively; hazard ratio: 2.86; 95% confidence interval, 1.98-4.13; p < 0.001). There were no significant differences in the probability of progression, disease-free mortality, or overall survival. Conclusions Perioperative plus four weekly MMC instillations followed by monthly BCG, instead of alternating BCG and IFN-α2b instillations, significantly reduce long-term recurrence. Patient summary We demonstrated in non-muscle-invasive bladder cancer patients with exceptionally frequent recurrences that the risk of long-term recurrence was reduced from 78-80% to 43-45% if one perioperative plus four weekly mitomycin C instillations were followed by monthly bacillus Calmette-Guérin (BCG) instillations for 1 yr instead of alternating instillations of BCG and interferon-α2b. Trial registration The registration was not considered necessary at this stage of the follow-up because the study was initiated as early as in 1992 and the last randomisation took place in 1996, before the current requirements concerning study registrations were implemented.