We analyzed single drugs and combinations of drugs used clinically in the treatment of opportunistic infections and other conditions for their activities against Pneumocystis carinii pneumonia in immunosuppressed rats. When they were used alone, atovaquone, rifabutin, and dapsone were more active than clarithromycin or trimethoprim. Drug combinations were evaluated for synergistic activity by an analysis of variance model for two-way factorial experiments and a response surface model. Atovaquone combined with trimethoprim trimethoprim and some combinations of dapsone and clarithromycin was synergistic; however, the activities of combinations of atovaquone and rifabutin, atovaquone and clarithromycin, and atovaquone and dapsone were simply additive. Lovastatin, which inhibits 3-hydroxy-methylglutaryl coenzyme A reductase, was inactive whether it was used alone or in combination with other agents. None of the synergistic drug combinations was as effective as trimethoprim-sulfamethoxazole. We conclude that the rat model can be used to test combinations of anti-P. carinii agents for synergistic activity by well-established statistical techniques. While some combinations of clinically used antimicrobial drugs have enhanced anti-P. carinii activity, further studies are needed before clinical trials can be contemplated.