Risk-benefit analyses are a prerequisite for a rational decision about therapies, e.g. drugs. Ingredients of a risk-benefit analysis are: 1. quality and quantity of the benefit; 2. quality and quantity of the harm; 3. benefit and harm of the natural history of the disease; 4. benefit and harm of therapeutic alternatives; 5. commensurability of harm and benefit; 6. consideration of all types of adverse drug reactions together. The available evidence was not suitable for a sound risk-benefit analysis of the use of trimethoprim-sulfamethoxazole. However, there seems to be enough evidence to restrict the use of trimethoprim-sulfamethoxazole in diseases with a benign prognosis and/or when there are therapeutic alternatives with a smaller risk/benefit ratio. Reliable rules for risk-benefit analyses have to be developed and the necessary information has to be gained also in methodologically sound phase IV research.