Dr. Siegel cites my concern that modulating the septic process may be a "two-edged sword," and he defends the use of all-cause mortality as the primary end point for determining the efficacy of a potential therapeutic agent. I agree that it is not desirable to have therapies that reverse relevant endotoxin-induced physiologic variables while increasing mortality by interfering with important defense mechanisms. We have been successful in identifying situations in which antisepsis therapy is deleterious (one edge of the "two-edged sword"). The CHESS (Centocor: HA-1A Efficacy in Septic Shock), the interleukin-1Ra, and the anti-tumor necrosis factor (TNF) monoclonal antibody (MAb) studies are some investigations that were stopped before completion because of increased mortality. Obviously, we must continue to design clinical trials that can detect such harmful effects early.

It was toward the other, potentially beneficial, edge of the sword that I questioned the wisdom of 28-day all-cause mortality. We have yet to show that any of the antisepsis candidates tested to date provide any significant benefit. Some agents (anti-TNF MAb and interleukin-1Ra) showed efficacy in preclinical studies and appeared to have the intended effect of modulating the cytokine response. Clinical benefit was also shown for a subpopulation of patients. We must improve our understanding of the septic process and accelerate the efforts that assist our ability to identify the appropriate patients who may benefit from therapy, as we must avoid those in whom such therapy would be deleterious. To that end, rapid, simple assays for the detection of circulating endotoxin are currently being tested. Clinically useful assays for other discriminatory markers, perhaps interleukin-6, should also be pursued.

More importantly, I am not optimistic that any single agent will provide the significant reduction in 28-day all-cause mortality that Dr. Siegel seeks. We should not expect such a "home run." Rather, given the multifactorial nature of sepsis, a combination of therapies directed at different stages of the septic process (not simply combination anticytokine therapy) may be required. Some experimental evidence suggests that immunotherapy directed toward the invading bacteria, endotoxin, and circulating cytokines is superior to any combination of monotherapy or duotherapy [1]. If these experimental conditions are relevant to human disease, we must ask what type of benefit from a single agent, short of a statistically significant reduction in all-cause mortality, is sufficient to proceed to studies of combination immunotherapy. Such a two-step approach need not delay the development process. Many of the agents already tested in pivotal studies were restudied in follow-up clinical trials. We may learn from our oncology colleagues, who recognized that the partial responses to therapy achieved using single agents may lead to complete responses when used together.

Although I cannot resolve the problem of how to assess the promise of an antisepsis agent, I am concerned that our current lack of understanding of the septic process may prevent the identification of better target populations and may relegate potentially useful drugs to therapeutic oblivion. Dr. Siegel's comments are a welcome initiation of such a dialogue.