The Search for Effective Therapy for Sepsis: Back to the Drawing Board?

Abstract

NEW THERAPEUTIC APPROACHES FOR SEPSIS HAVE NOT fared well recently. In January, Eisai announced that its worldwide phase 3 randomized trial of a novel anti–Toll-like receptor (TLR)-4 compound, eritoran tetrasodium, had failed to demonstrate an improvement in the primary end point of 28-day all-cause mortality in a cohort of 2000 patients with severe sepsis. This news was disappointing, especially because the manipulation of TLR4 signaling would represent a new avenue of research and drug development. Perhaps ironically, only a few months later, the importance of TLR4 signaling was recognized with the award of the Nobel Prize in Physiology or Medicine to Jules Hoffmann and Bruce Beutler for their work in this area. In October, Eli Lilly announced it was withdrawing Xigris (drotrecogin alfa, a recombinant activated protein C) from the market following the failure of its worldwide trial, PROWESS Shock, to demonstrate improved outcome. Drotrecogin alfa, the only approved drug specifically indicated for the treatment of severe sepsis, had been approved on the basis of an earlier trial, PROWESS, which had demonstrated a large improvement in survival. That the findings were not confirmed by the subsequent trial was another major disappointment. So what now? At a minimum, researchers and clinicians need to rethink the therapeutic approach to sepsis. Some obvious questions come to mind. First, is the current understanding of the pathophysiology of sepsis flawed in some important way? Second, is the current approach to the discovery and evaluation of potential therapies in need of overhaul? Third, assuming the answer is yes to either of these questions, where should researchers go next? Sepsis is a broad term, with its roots in the writings of Hippocrates. At its heart is the concept of a patient fighting to survive a life-threatening infection. And it is the fight that is thought to be injurious. The invading pathogen can be directly toxic and destructive to tissue, but much of the pathology associated with sepsis is attributed to the host response. Host immune cells exposed to pathogenassociated molecular patterns (PAMPs), such as lipopolysaccharide (LPS), rapidly produce a broad array of cytokines, chemokines, and other proteins to sequester and eradicate invading pathogens. However, these same proteins can profoundly disturb and harm host tissue function and anatomy, a form of “friendly fire.” These findings led to 2 central tenets of current sepsis research. First, the host response in sepsis is unhelpfully exuberant, and thus agents that block or suppress the host response should improve outcome. Second, the host response represents a “final common pathway,” and thus agents that manipulate this pathway should work regardless of the source of infection. Neither tenet may be true. The host response does not appear to be ubiquitously overexuberant. Indeed, patients with similar signs and symptoms can have widely different cytokine profiles. As noted in the report by Boomer et al in this issue of JAMA, the host response can be dramatically suppressed rather than overly exuberant. Second, the discovery of PAMP-induced host response has helped highlight that host-pathogen interactions are much more sophisticated and nuanced than previously recognized. In short, there may not be a “final common pathway.” Current preclinical experiments and clinical trials of potential therapies account only partially for these observations. Preclinical animal experiments may be too simplistic. The animals are usually young and have no comorbidity. In contrast, patients with sepsis are often older and have underlying comorbidity, both of which are strong predictors of sepsis susceptibility and outcome. The insult in animal models is typically LPS, cecal ligation and puncture, or instillation of live bacteria into the lung to induce pneumonia, but formal evaluation of differences in response across these different infectious challenges is not usually done. Antibiotics are often not given in animal models, and there is little or no supportive care for the ensuing organ dysfunction. The ideal host response to fight infection in the absence of antibiotics and life support may be very different from that required in a modern intensive care unit setting. It would also be helpful to evaluate how a drug works in an animal model in terms of the effect on the immune system