Angiopoietin-1 and angiopoietin-2 as clinically informative prognostic biomarkers of morbidity and mortality in severe sepsis*

OBJECTIVE: To determine the utility of angiopoietin-1 and angiopoietin-2 as potentially novel biomarkers of morbidity and mortality in patients with severe sepsis. DESIGN: Multicenter longitudinal cohort study. SETTING: Three tertiary hospital intensive care units in Hamilton, Ontario, Canada. PATIENTS: A total of 70 patients with severe sepsis were enrolled within 24 hrs of meeting the inclusion criteria for severe sepsis and followed until day 28, hospital discharge, or death. INTERVENTIONS: Clinical data and plasma samples were obtained at intensive care unit admission for all 70 patients and then daily for 1 wk and weekly thereafter for a subset of 43 patients. Levels of angiopoietin-1 and angiopoietin-2 in stored plasma samples were measured and compared with clinical characteristics, including the primary outcomes of 28-day mortality and morbidity measured by the Multiple Organ Dysfunction score. MEASUREMENTS AND MAIN RESULTS: Lower angiopoietin-1 plasma levels (≤ 5.5 ng/mL) at admission were associated with increased likelihood of death (relative risk 0.49 [95% confidence interval of 0.25-0.98], p = .046). Lower angiopoietin-1 levels remained a significant predictor of 28-day mortality in a multiple logistic regression model (adjusted odds ratio of 0.282 [95% confidence interval of 0.086-0.93], p = .037). Analysis of serial data using linear mixed models confirmed that sepsis survivors had higher levels of angiopoietin-1 (p = .012) and lower daily levels of angiopoietin-2 (p = .022) than nonsurvivors. Furthermore, survivors had higher peak angiopoietin-1 levels (median 13 vs. 10 ng/mL, p = .019) and lower nadir angiopoietin-2 levels (median 2.8 vs. 6.2 ng/mL, p = .013) than nonsurvivors. A score incorporating angiopoietin-1 and angiopoietin-2 and three other markers of endothelial activation discriminated with high accuracy between fatal and nonfatal cases (c-index of 0.80 [95% confidence interval of 0.69-0.90], p < .001). Plasma levels of angiopoietin-2 correlated with clinical markers of organ dysfunction and molecular markers of endothelial cell activation. CONCLUSIONS: Angiopoietin-1 levels at admission and both angiopoietin-1 and angiopoietin-2 levels measured serially correlated with 28-day mortality in severe sepsis. Angiopoietin-2 levels also correlated with organ dysfunction/injury and a validated clinical sepsis score. These results suggest the use of angiopoietins as clinically informative biomarkers of disease severity and patient outcome in severe sepsis.