Prognostic value of incremental lactate elevations in emergency department patients with suspected infection

Abstract

Objectives: Previous studies have confirmed the prognostic significance of lactate concentrations categorized into groups (low, intermediate, high) among emergency department (ED) patients with suspected infection. Although the relationship between lactate concentrations categorized into groups and mortality appears to be linear, the relationship between lactate as a continuous measurement and mortality is uncertain. This study sought to evaluate the association between blood lactate concentrations along an incremental continuum up to a maximum value of 20 mmol/L and mortality.

Methods: This was a retrospective cohort analysis of adult ED patients with suspected infection from a large urban ED during 2007-2010. Inclusion criteria were suspected infection evidenced by administration of antibiotics in the ED and measurement of whole blood lactate in the ED. The primary outcome was in-hospital mortality. Logistic and polynomial regression were used to model the relationship between lactate concentration and mortality.

Results: A total of 2,596 patients met inclusion criteria and were analyzed. The initial median lactate concentration was 2.1 mmol/L (interquartile range [IQR] = 1.3 to 3.3 mmol/L) and the overall mortality rate was 14.4%. In the cohort, 459 patients (17.6%) had initial lactate levels >4 mmol/L. Mortality continued to rise across the continuum of incremental elevations, from 6% for lactate <1.0 mmol/L up to 39% for lactate 19-20 mmol/L. Polynomial regression analysis showed a strong curvilinear correlation between lactate and mortality (R = 0.72, p < 0.0001).

Conclusions: In ED patients with suspected infection, we found a curvilinear relationship between incremental elevations in lactate concentration and mortality. These data support the use of lactate as a continuous variable rather than a categorical variable for prognostic purposes.