Median effect analysis of efficacy versus adverse effects of immunosuppressants

Background A rigorous model to describe concentration‐effect relations—the median effect analysis—was applied to quantitate immunosuppressive versus adverse effects in human renal transplantation. Methods The median effect equation was used to analyze data collected from three clinical studies, including the two phase III blinded, placebo‐controlled trials (n = 1295 patients) of sirolimus versus azathioprine or placebo treatment added to a cyclosporine (INN, ciclosporin)/prednisone regimen and a sirolimus/azathioprine/prednisone (in the absence of cyclosporine) phase II cohort (n = 41 patients). Results The clinical effects correlated with drug concentrations as expressed by the median effect equation. Sirolimus or cyclosporine alone permitted drug concentrations that were 5‐fold and 2.2‐fold lower, respectively, to render 90% of patients rejection‐free, suggesting a synergistic interaction between the two drugs. Further, the sirolimus concentrations to render 50% of patients rejection‐free were about 200‐fold and 60‐fold less, respectively, than the concentration that caused 50% of patients to experience thrombocytopenia or hypertriglyceridemia. The correlation coefficient of the median effect analysis for the occurrence of hypercholesterolemia was more robust for sirolimus than for cyclosporine. Although the concentrations for 50% of patients rendered rejection‐free versus 50% affected by hypercholesterolemia were similar, a 7‐fold difference was calculated between the concentrations at which 90% of patients were free of rejection versus patients who were affected by hypercholesterolemia. Conclusion The median effect analysis proffers a useful tool to assess both drug interactions and the windows between therapeutic versus toxic effects of immunosuppressive agents. The current analysis suggests a synergistic interaction between sirolimus and cyclosporine. Clinical Pharmacology & Therapeutics (2001) 70 , 74–81; doi: 10.1067/mcp.2001.116309