You could be reading the full-text of this article now if you...

If you have access to this article through your institution,
you can view this article in

15: CYTOKINE LEVELS IDENTIFY CLINICAL SUBGROUPS OF SEPSIS PATIENTS

Fjell, Christopher; Boyd, John; Walley, Keith; Russell, James

Critical Care Medicine:
doi: 10.1097/01.ccm.0000424271.05136.89
Oral Abstract Session
Author Information

University of British Columbia

The University of British Columbia

St. Paul’s Hospital

The University of British Columbia

Abstract

Introduction: The VASST clinical trial compared use of vasopressin versus norepinephrine infusion in patients with septic shock. From this study, cytokine blood levels were measured in a total of 362 patients at both baseline and 24 hours in addition to approximately 300 clinical features (such as age, sex, blood pressure, temperature, dose of norepinephrine to treat).

Hypothesis: Inflammatory effects are believed to underlie aspects of sepsis, so we hypothesized that cluster analysis of cytokines would independently identify subgroups of patients at risk of death or other clinical outcome such renal failure.`

Methods: Cytokines were measured using a Luminex cytokine array. Hierarchical clustering was performed on cytokine values to identify patients groups. Enrichment analysis identified clinical features occurring more often than expected in each patient group. Independently, classification and regression tree analysis was performed to assess the importance of clinical features and cytokines on predicting patient outcome. Significance was assessed using over-representation analysis and survival curves.

Results: Cytokine levels at baseline and 24 hours produced four similar clusters of patients with differing risks of adverse outcome or previous clinical features. With p-value < 0.05 and enrichment >50%, one cluster of 47 patients showed enrichment for severe septic shock, coagulopathy, renal failure, and risk of death at 28 days. A separate cluster of 95 patients was enriched for chronic obstructive pulmonary disease and recent surgical history but not increased risk of death. Independent of cluster analysis, tree analysis of survival indicated that IL8 levels at 24 hours had greatest predictive content for survival. Combined with IL8, patient age and dose of norepinephrine needed to treat gave the greatest predictive performance for survival rate. We noted that IL8 at was weakly but significantly correlated (spearman correlation of 0.52) with dose of norepinephrine.

Conclusions: Cytokine levels in sepsis patients significantly predict risk of death and other clinical outcomes. Further analysis of cytokines may lead to enhanced understanding of the disease process in sepsis though clinical interventions may confound interpretation.

© 2012 by the Society of Critical Care Medicine and Lippincott Williams & Wilkins