Bacillus Calmette-Guérin Therapy for Bladder Cancer and Risk of Alzheimer's Disease in Denmark

Background: Intra-vesical Bacillus Calmette-Guérin (BCG) instillation has since the 1970'ies been used in non-muscle invasive bladder cancer (NMIBC) patients. Studies from Israel and USA have found that BCG instillation was associated with reduced rate of Alzheimer's disease in bladder cancer patients.

Objectives: To examine the hypothesis that BCG instillation is associated with a decreased rate of Alzheimer's disease in Danish patients with NMIBC.

Methods: We conducted a cohort study using information from Danish nationwide registries linked using a unique identifier. We included all Danish patients aged ≥40 years with incident NMIBC between 2002 and 2016 and without cystectomy within 6 months after diagnosis. For patients still alive and without any registration of dementia, mild cognitive impairment, or amnestic syndrome one year after bladder cancer diagnosis, we assessed their exposure status since diagnosis as either BCG instillation, mitomycin instillation (active comparator) or no bladder instillation and followed them until incident Alzheimer's disease, death, migration or end of follow-up (December 31, 2018). We used Cox proportional hazards regression to estimate hazard ratios (HRs) of Alzheimer's disease adjusting for age, sex, calendar period of diagnosis, and co-morbidities.

Results: We included 17,099 patients, of which 3,409 (19.9%) had BCG instillations (median 6 treatments; interquartile range=6-8), 1,558 (9.1%) had mitomycin instillations, and 12,132 (71.0%) had no bladder instillations. The patients were followed for a median of 4.9 years (interquartile range=2.4-8.4). In the BCG group, 61 patients were diagnosed with Alzheimer's disease during follow-up. The adjusted HRs (aHRs) were 1.09 (95% confidence interval [CI]=0.65-1.82) compared with mitomycin instillation and 1.01 (95% CI=0.73-1.39) compared with no bladder instillation. To limit the possibility for reverse causation and ensure greater temporal distance between exposure and outcome, we made a sensitivity analysis starting follow-up 3 years after bladder cancer diagnosis, which showed similar results (aHRs were 1.23 [95% CI=0.64-2.36] compared with mitomycin and 1.00 [95% CI=0.69-1.46] compared with no bladder instillation). In addition, sensitivity analyses using the Fine and Gray subdistribution hazard model accounting for competing risk from death did not alter results (aHRs were 1.08 [95% CI=0.65-1.82] compared with mitomycin and 1.06 [95% CI=0.77-1.46] compared with no bladder instillation). There was no indication of effect measure modification by sex, age, or calendar period.

Conclusions: We did not identify an association between BCG instillations in the bladder and diagnosis of Alzheimer's disease in Danish patients with NMIBC.