Bacillus Calmette-Guérin (BCG) has been part of the treatment algorithm for non-muscle invasive bladder cancer (NMIBC) for a number of years. In this single centre, prospective, randomised trial, two of the most commonly available strains, BCG Connaught and BCG Tice, were compared. An experimental murine trial to compare in vivo response was also run. One hundred and forty-two patients with high-risk NMIBC were recruited and randomised to receive six instillations of one of the two strains. Maintenance therapy was not given as the study was commenced prior to the benefits of maintenance being known. The patients were followed up with regular cystoscopy and wash cytology over a five-year period. The groups were matched for age, gender, tumour stage and grade. A significantly better five-year recurrence free survival was found in those receiving BCG Connaught (74%) compared with BCG Tice (48%). Progression-free survival was also better in the BCG Connaught group (94.1% vs. 87.9%), with significantly fewer patients experiencing dysuria as a side-effect. In vivo experimentation involved BCG instillation into mice, with subsequent bladder and lymph node harvest. A higher level of BCG was found in the regional lymph nodes in those receiving BCG Connaught, with greater T-cell recruitment to the bladder. Although in the UK there is currently only one licensed BCG preparation, this paper suggests consideration needs to be made to the strain we use. Further research, with larger population recruitment, would help guide us to select the optimum treatment.