Despite the advances in supportive care and the availability of potent antimicrobial agents, mortality from sepsis, a leading cause of death in intensive care units, has not improved. Over the last decade, clinical trials with numerous adjunctive therapies, including antiendotoxin antibodies and inhibitors of the inflammatory response, have yielded disappointing results. Recently, treatment with recombinant human activated protein C reduced mortality 6% compared with controls. Given the likelihood that many processes in the complex pathophysiology of sepsis are simultaneously activated, it is unlikely that therapy directed at any one of them, as has been done in the past, will dramatically improve survival. Rather, a combination of therapies directed at many arms of the septic process, much like the strategy used for cancer and HIV infection, is required. Given the likelihood that sepsis represents an excessive innate immune response to microbial products, vigorous attempts must be made to develop rapid assays that reflect the level of innate immune activation. Such assays could be used to identify patients who would benefit from therapy and to monitor their response so that overtreatment does not completely abrogate host defense mechanisms and render these patients susceptible to fatal infection. It is now time to test a new therapeutic paradigm based on an improved understanding of the pathophysiology of the septic process and the recognition that we may have reached the limits of adjunctive monotherapy.