Despite recent advances in the prospective identification of the patient with sepsis who may benefit from anti-inflammatory or antithrombotic therapies, successful treatment regimens have been fairly modest. We have explored whether determination of several proinflammatory cytokine or mediator concentrations can complement physiologic scoring systems to identify patients with severe sepsis who will survive or expire within 28 days. The design of the study included an exploratory analysis performed in conjunction with a prospective, randomized, double-blind, placebo-controlled, multicenter, clinical trial and involved 33 academic institutions in the United States. One hundred twenty-four patients with severe sepsis with or without septic shock were included in this analysis. Blood samples were obtained at baseline and on days 1 through 4, and were evaluated for proinflammatory and anti-inflammatory cytokine concentrations, as well as for procalcitonin and total protein C levels. Baseline concentrations and changes in the concentrations of these mediators were evaluated in relationship to the Acute Physiology and Chronic Health Evaluation (APACHE) II and multiple organ dysfunction (MOD) scores, and 28-day all-cause mortality. Using univariate logistic regression analyses, APACHE II and MOD scores, age (but not gender), and baseline plasma interleukin (IL)-6 and soluble tumor necrosis factor receptor (sTNFR) 1 (log transformed) concentrations were all predictive of increased 28-day all-cause mortality (P < 0.01). Baseline total protein C, IL-8, IL-10, TNF-alpha, and procalcitonin concentrations, and the change in plasma cytokine concentrations from baseline over the initial 4 days were not useful in predicting outcome. Selected baseline proinflammatory cytokine concentrations and APACHE II score were correlated (P < 0.01). IL-6 concentration is a strong candidate for predicting clinical outcome in patients with severe sepsis alone, or when combined with the APACHE II or MOD scores. The potential usefulness of the combination of cytokine measurements and prognostic scores to identify patients who may benefit from treatment with anti-inflammatory or antithrombotic therapies should be further evaluated.