Kai Swenson

Date of Award

January 2015

Document Type

Open Access Thesis

Degree Name

Medical Doctor (MD)



First Advisor

Charles R. Wira

Subject Area(s)



HYPOTHESIS AND SPECIFIC AIMS: The aim of this study is to characterize the incidence and outcomes of various groups within a novel classification system of severe sepsis and septic shock, for the purpose of informing more accurate risk prediction in the proximal phases of care. Our primary hypothesis is that an early classification system of septic patients categorized by organ dysfunction, initial emergency department (ED) serum lactate, and ED vasopressor utilization will offer accurate mortality prognostication in patients with severe sepsis and septic shock.

METHODS: We performed a retrospective analysis of a prospectively- gathered registry of severe sepsis and septic shock patients presenting to a dual-site academic emergency department (ED). In the primary analysis, registry subjects were categorized into five groups by initial ED serum lactate level and vasopressor requirement in the ED: dysoxic shock (vasopressor use + lactate >4 mmol/L), vasoplegic shock (vasopressor use + lactate <4 mmol/L), cryptic shock major (no vasopressor use + lactate >4 mmol/L), cryptic shock minor (no vasopressor use + lactate >2 and <4 mmol/L), and severe sepsis without lactate elevation (no vasopressor use + lactate <2 mmol/L + evidence of organ dysfunction). For each group, the 28-day mortality rate was evaluated by logistic regression controlling for specific factors associated with sepsis severity.

RESULTS: Of 521 registry subjects, 85.6% (n=446) met inclusion criteria. 4.9% (n=22) subjects presented in dysoxic shock, 11.7% (n=52) in vasoplegic shock, 12.1% (n=54) in cryptic shock major, 30.9% (n=138) in cryptic shock minor, and 40.4% (n=180) in severe sepsis without lactate elevation. The 28-day mortality rates for these groups were 50.0%, 21.1%, 18.5%, 12.3%, and 7.2%, respectively; this stepwise trend was paralleled by metrics of critical care utilization such as ICU admission, mechanical ventilation, and vasopressor use within 72 hours of admission. After controlling for known risk factors for sepsis severity, the odds ratios for death before 28 days were 15.06 for dysoxic shock, 3.61 for vasoplegic shock, 3.77 for cryptic shock major, and 1.93 for cryptic shock minor, as compared to severe sepsis without lactate elevation.

CONCLUSION: This study suggests that high-risk subgroups of severe sepsis and septic shock patients can be identified at presentation and during the emergency department stay. We show that in severe sepsis and septic shock, a proximal-phase classification system based on vasopressor requirement in the ED and initial ED lactate level predicts 28-day in-hospital mortality and may inform prognostication, triage decisions and future sepsis clinical trial design.


This is an Open Access Thesis.

Open Access

This Article is Open Access