A retrospective comparison between clinical experience with trimethoprim-sulfamethoxazole (TMP-SMZ) and predictive laboratory investigations demonstrates that the most accurate predictions concern antibacterial activity of the drug against individual organisms. With regard to therapy for some relatively severe infections, clinical experience is limited and further research is needed. Some such studies that were not feasible in the past are now possible as a result of the availability of trimethoprim (TMP) as a single agent and of parenteral forms of both TMP and TMP-SMZ. The infections reviewed in this light herein include septicemia, endocarditis, meningitis, neonatal sepsis, anaerobic infections, ophthalmic infections, and infections due to "higher" bacteria and fungi. Based on current knowledge of in vitro drug activity, the clinical picture in such serious infections, and the experience with TMP-SMZ so far, it is predicted that TMP will be used either alone or in combination with other antibacterial agents (such as the polymyxins, metronidazole, and various sulfonamides in different ratios) for treatment of some of these infections.