It is postulated that the unstable hydroxylamine metabolite of sulphamethoxazole is responsible for the adverse reactions to co-trimoxazole and in HIV infection systemic glutathione deficiency leads to a reduced capacity to counteract the hydroxylamine toxicity. This hypothesis has been investigated by studying the metabolism of sulphamethoxazole and assessing glutathione status in HIV infection in order to explore the modification of treatment. It is concluded that the toxicity of plasma sulphamethoxazole hydroxylamine is counteracted by normal glutathione concentrations as is the case in HIV-seropositive patients, but that increased oxidation within certain cells in HIV infected individuals may possibly give rise to increased concentrations of reactive intermediates of sulphamethoxazole. Sulphametrole and sulphamethoxazole have similar half-lives but are metabolized differently: in vivo no oxidised metabolites of sulphametrole could be detected. In a retrospective study the rate of adverse reactions to trimethoprim-sulphametrole appeared to be in the lower range of those reported for trimethoprim-sulphamethoxazole indicating that the combination of trimethoprim-sulphametrole may be more favourable. The ratio of trimethoprim:sulphonamide is 1:5, but in-vitro studies with Toxoplasma gondii indicate that because of the synergic effect of both agents the dose of sulphonamide is possibly unnecessarily high.