Objective: Mortality in sepsis is believed to be associated with exaggerated inflammatory responses, but recent evidence suggests that poor outcome is associated with reduced inflammation. To test this hypothesis, we measured several inflammatory markers to determine whether any of them or any combinations are associated with mortality or organ dysfunction.
Design: Clinical study.
Setting: School of medicine.
Patients: Thirty-five patients with severe sepsis.
Interventions: Markers of endothelial, platelet, and leukocyte activation were measured on days 1, 2, and 3 after enrollment. The markers were a) endothelial microparticles (EMPs) and their conjugates with monocytes (EMP/MONO); b) platelet microparticles (PMPs) and platelet activation marker CD62P; c) platelet-leukocyte conjugates (PLT/LEU) and leukocyte activation marker CD11b; and d) intracellular nitric oxide in leukocytes.
Measurements and main results: The 28-day mortality rate was 51% (18 of 35). Significant differences between survivors and nonsurvivors on day 1 were found in PLT/LEU (p = .001), CD11b (p = 0.02), and EMP/MONO (p = .02) groups. Using logistic regression to assess if these markers predict mortality on day 1, we found that PLT/LEU had the best predictive value among the markers used (area under receiver operating characteristics curve = 0.82). All markers of cell activation and inflammation were significantly higher among survivors on days 2 and 3, except nitric oxide, which was lower. This marker showed significant negative correlation with the Sequential Organ Failure Assessment score throughout the study.
Conclusions: Our data support the hypothesis that early increased, not decreased, inflammatory response as measured by our markers is associated with improved survival rate. A high negative correlation was found between some of these markers and Sequential Organ Failure Assessment score.