Occult hypoperfusion and mortality in patients with suspected infection

Objective: To determine, in the early stages of suspected clinically significant infection, the independent relationship of the presenting venous lactate level to 28-day in-hospital mortality. Design: Prospective, observational cohort study. Setting: Urban, university tertiary-care hospital. Patients: One thousand two hundred and eighty seven adults admitted through the emergency department who had clinically suspected infection and a lactate measurement. Measurements and results: Seventy-three [5.7% (95% CI 4.4-6.9%)] patients died in the hospital within 28 days. Lactate level was strongly associated with 28-day in-hospital mortality in univariate analysis (p < 0.0001). When stratified by blood pressure, lactate remained associated with mortality (p < 0.0001). Normotensive patients with a lactate level >= 4.0 mmol/l had a mortality rate of 15.0% (6.0-24%). Patients with either septic shock or lactate >= 4.0 mmol/l had a mortality rate of 28.3% (21.3-35.3%), which was significantly higher than those who had neither [mortality of 2.5% (1.6-3.4%), < 0.0001]. In a model controlling for age, blood pressure, malignancy, platelet count, and blood urea nitrogen level, lactate remained strongly associated with mortality. Patients with a lactate level of 2.5-4.0 mmol/l had adjusted odds of death of 2.2 (1.1-4.2); those with lactate >= 4.0 mmol/l had 7.1 (3.6-13.9) times the odds of death. The model had good discrimination (AUC = 0.87) and was well calibrated. Conclusions: In patients admitted with clinically suspected infection, the venous lactate level predicts 28-day in-hospital mortality independent of blood pressure and adds significant prognostic information to that provided by other clinical predictors.