Although symptomatic treatments have been optimized, mortality in severe infectious conditions still exceeds 50% in many instances. Excessive activation of inflammation mediators, the endogenous relay of the initial stimulus, may account for the secondary occurrence of polyvisceral failure. This "internal" conception of severe infectious conditions opens new therapeutic possibilities including the modulation of the host's response. However, this new treatment of severe septic conditions, which complements antibiotic therapy and symptomatic measures, encounters several difficulties: a) proinflammatory mediators also have beneficial effects; b) as the mediator cascade is activated within a few hours, ordering immunity-modulating treatments early appears to be a logical step; c) monoclonal antibodies and recombining molecules are very expensive. Costs may be further increased by the probable need to use "cocktails" of antiinflammatory molecules. All these reservations show how technically blameless future clinical trials have to be. Cost efficiency and cost/benefit studies are also required. We may wonder whether society will make the financial choice to use these new treatments.