Clinical experience with antiinflammatory agents in patients with sepsis has been disappointing to date. Tumor necrosis factor soluble receptor (TNFsr) is one such agent. In early preclinical work utilizing intravenous endotoxin or gram-negative bacteria challenges, TNFsr was protective. However, in clinical sepsis trials this agent appeared to have little beneficial effect. In some cases of gram-positive infection it may have been harmful. These finding suggest that factors not identified in preclinical studies may have altered the effects of this antiinflamatory agent in clinical trials. Such factors might include the type, site, and severity of infecting bacteria. In order to study the influence of these factors on the effects of TNFsr in sepsis, we have employed a rat model of sepsis we previously developed which incorporates a multifactorial study design and allows simultaneous study of multiple variables. The results from this study show that in this rat model, severity of infection and to a lesser degree, type of bacterial infection, significantly influence the effects of TNFsr. This anti-inflammatory agent was beneficial with doses of E. coli resulting in a high lethality and it was harmful with doses of S. aureus resulting in a low lethality. Further studies are underway to determine mechanisms underlying these differences. These studies suggest that TNFsr may only be beneficial in patients with a high likelihood of dying due to their underlying infection, especially with gram-negative bacteria like E. coli. It may be harmful in patients with a low likelihood of dying from their underlying infection, especially with gram-positive bacteria like S. aureus. Additional investigations are underway to determine the basis for the variable effects of TNFsr in these studies.
