Since 1976, when Morales, Eidinger and Bruce first published on the treatment of human urothelial bladder cancer, bacillus Calmette-Guérin has gained worldwide acceptance as a therapy against bladder carcinoma recurrences. However, there are many uncertainties with regard to patient selection, treatment protocol, reduction of side effects, mechanisms of action, and thus room enough for laboratory researchers and clinicians to contribute to the improvement of BCG therapy. Initiation of mycobacterial binding to urothelial cells may be promoted by methods which increase fibronectin exposure on target cells, reduce inhibitory factors in the instillate, or by identification of more avidly binding substrains of BCG. Quantification of the ensuing immune response focuses on the local rather than the systemic immune reaction of the host and may give the opportunity to predict a favourable response, or even the endpoint of therapy. For this purpose, assessment of local cytokine production seems to be a valuable tool. Ultimately, revealing the mode of action(s) of live BCG may render the use of potentially dangerous live bacilli unnecessary and lead to the application of derivatives with better results.