We showed in experiments in our labs that the effects of two anti- inflamatory agents, tyrphostin AG556 and tumor necrosis factor soluble receptor, were influenced by the severity of the infectious challenge in animal models of sepsis. To deter-mine whether a similar relationship existed with other mediator-specific anti-inflamatory agents, we performed a meta-analysis of all preclinical studies cited in published clinical sepsis trials of anti-tumor necrosis factor (TNF) monoclonal antibodies, soluble TNF recep-tors, interleukin-1 receptor antagonist, platelet activtating receptor antagonist, bradykinin antagonist, and prostaglandin antagonist. For both anti-TNF monoclonal antibodies and soluble TNF receptor, which were the two most frequently studied agents, there was a significant association between control mortality and the effects of the agents on survival. These agents were most beneficial at high control mortalities. With decreasing control mor-tality each was less beneficial. This relationship was similar to the one we had noted in our experiments. For the remaining four agents, this relationship was present but not significant. However, combining the data for all six agents from published preclinical studies, showed that a common slope describing this relationship fit the data for each agent individually very closely. These findings have suggested that the effects of each of these anti-inflamatory agents may be influenced by the severity of the underlying infectious challenge.
