© 2002 American Thoracic Society
Does Blocking Cytokines in Sepsis Work?Divisions of Infectious Diseases and Critical Care Medicine University of Colorado Health Sciences Center Denver, Colorado
Analyzing controlled, mortality-based intervention trials for sepsis syndrome is a valid approach for improving the design of clinical studies and determining which subgroups of patients, if any, may have demonstrated benefit from novel therapeutic agents. The report by Eichacker and colleagues (1) in this issue of AJRCCM (pp. 1197–1205) expands on a previous article by this group of National Institutes of Health scientists (2). Together with the analysis by Knaus and coworkers on the administration of the interleukin-1 (IL-1) receptor antagonist to sepsis patients (3) and analyses of trials using anti–tumor necrosis factor (TNF
Although we laud the study of Eichacker and coworkers (1) and agree with many of the conclusions, we believe that there are several issues that remain unresolved in their analysis of the various clinical trials and animal studies. Eichacker and coworkers (1) lump together such unrelated agents as anticytokines (anti-TNF The numbers of patients who have been randomized to receive agents that reduce the biologic consequences of TNF or IL-1 provide the largest groups for analysis, and also these two cytokines have been targeted in the largest number of animal trials. Therefore, we will focus our comments on anticytokine-based therapies. There is another reason why close attention should be paid to anticytokine therapies: over 250,000 patients are currently receiving the very same anticytokine therapies for the treatment of autoimmune diseases (rheumatoid arthritis, Crohn's disease, psoriasis, and ankylosing spondylitis). There is no question that IL-1 and TNF have pathologic roles in these chronic diseases, and there is no question that anticytokine therapies are highly effective in that they reduce the biologic consequences of these two cytokines. Why did they fail to reduce 28-day, all-cause mortality in sepsis? Notably, the soluble TNF receptor p75 fused to the Fc of Immunolglobulin G led to a dose-dependent increase in 28-day, all-cause mortality (7), and yet this same agent (etanercept; Enbrel, Seattle, WA) is highly effective in treating autoimmune diseases (8). Although the conclusions of Eichacker and coworkers (1) that better responses to antiinflammatory agents occur in the high-risk patients appear to be valid, the contention that failed clinical trials were due to inclusion of large numbers of patients not sick enough to benefit from reducing IL-1 and/or TNF activities is more questionable. In their analysis, treatment with an anticytokine agent in patients with a low risk of death was thought to be harmful. The "harm" is thought to be due to exogenously administered anticytokine agents interfering with the intrinsic benefit of cytokines in stimulating inflammatory mechanisms, such as neutrophil emigration, which aids in combating infections. In fact, the objective of anticytokine therapy is precisely to reduce the detrimental effect of the systemic inflammation at the cost of reducing its benefits. The case for anticytokine therapy increasing mortality, however, is complex and certainly different for IL-1 versus TNF. TNF kills cells and IL-1 does not. Blocking IL-1 did show benefit in patients with proven bacteremia but worsened mortality in patients without bacteremia, independent of severity score at entry (9). This observation suggests that risk of death is not the only, or even the most important, factor in determining survival among patients treated with such anticytokine agents. Although not always reported, nearly all clinical trials of anticytokines for sepsis did examine the relationship between acute physiology and chronic health evaluation or other severity-of-illness scores and response to therapy but did not consistently link efficacy to high acute physiology and chronic health evaluation scores on entry. Still unexplored is the issue that, unlike in animal studies, many patients entering sepsis trials are ill for days (even weeks) with various comorbidities. To our knowledge, no one has analyzed this variable in the outcomes of the anticytokine studies. Another issue in clinical trials of anticytokine therapies is the duration of treatment. Is it possible that a treatment period of 1–3 days (most anticytokines were used for short-term treatment) was insufficient to impact survival after 28 days? Not only short duration of treatment, but also monotherapy, may be at fault. Blocking IL-1 plus TNF in patients with rheumatoid arthritis appears to be more effective than either agent alone (10), and the basis for dual therapy has ample support in animal models of rheumatoid arthritis as well as in animal models for live bacterial infection using cecal ligation and puncture (11). What remains to be done? Surely, the thorough and detailed analysis by Eichacker and coworkers (1) will help in the design of future trials of agents to reduce mortality in patients at risk from an infectious process turned systemic. But how and when to rescue these patients will not come from knowing only the severity of their process once the downward spiral is initiated. We do not believe trials of anticytokine therapy (therapies) should be abandoned in these patients. Trials using more homogeneously defined entry criteria, including those that directly show evidence of the intrinsic (or endogenous) antiinflammatory pathway being modulated, may reveal who will benefit from anticytokine or other antisepsis therapies and who will not. REFERENCES
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
) therapies, the conclusions of Eichacker and coworkers (
