Early lactate clearance for death prediction in severe sepsis or septic shock patients presenting to the Emergency Department

Background: As part of a sepsis mortality reduction initiative at a major health system, a sepsis database to evaluate processes and outcomes related to sepsis care management was developed. 

Objective: To analyze the relationship between lactate clearance and in-hospital mortality.

Methods: Retrospective observational study from January 2012 to June 2014. Data abstractors entered clinical and demographic data into database via a web tool created by health system’s Quality Management Institute.

Setting: A multi-hospital health system in New York Metro area with over 500,000 emergency department (ED) visits annually.

Patients: A cohort of patients presenting to eleven emergency departments with severe sepsis or septic shock.

Results: There were 13,925 severe sepsis and septic shock patients during the cohort period. An initial lactate was recorded for 13,196 of these patients, 98% of which had a discharge status listed. The mean initial lactate level was higher for patients that died than for survivors (4.52 +/- 3.66 vs. 2.93 +/- 2.09, respectively; P<.001).  3,901 of the patients with an initial lactate had a repeat lactate within 6 hours of which 3,797 had a discharge status recorded.  Mean lactate clearance was 0.24 +/-0.48 for survivors whereas for non-survivors mean lactate clearance was 0.04 +/-0.83 (P<.001). We calculated lactate clearance by the equation (lactate _initial – lactate _repeat)/lactate _initial. Logistic regression was used with lactate clearance as a predictor of mortality adjusting for age, weight, admission unit, hospital, and initial lactate level. The adjusted odds ratio for lactate clearance was 0.50 (95% CI=0.43, 0.59; P<.001).  The unadjusted area under the Receiver Operating Characteristic Curve was computed for lactate clearance and in-hospital mortality at 0.62 (95% CI=0.60, 0.64).

 Conclusion: Results suggest that improvement in 6 hour lactate clearance among patients admitted from the ED with severe sepsis and septic shock is associated with decreased in-hospital mortality. This information may be useful in assessing the response to short term emergency treatment of severe sepsis in the acute care setting.