Objective: A serum lactate level >2 mmol/L has been chosen as the preferred cut-off value for screening of patients with suspected sepsis. In patients with suspected sepsis presenting to the ED, we aimed to determine the outcomes of patients with initial lactate levels ≤2 mmoL/L, but abnormal bicarbonate or anion gaps (AGs). Methods: This prospective cohort study enrolled patients from an adult tertiary referral hospital who presented with suspected sepsis. The predictive value of lactate, bicarbonate and the AG for intensive care unit (ICU) admission and death at hospital discharge were evaluated using area under the receiver operating characteristic curves (AUROC). Results: There were 441 patients with suspected sepsis enrolled from February 2016 to June 2017. There were 96 (22.0%) patients who were admitted to the ICU and at hospital discharge, 42 (9.6%) patients had died. There was no statistically significant difference between the AUROCs of lactate or bicarbonate level or AG to predict ICU admission (P = 0.17). There was no statistically significant difference between the AUROCs of lactate or bicarbonate level or AG to predict mortality at hospital discharge (P = 0.44). Among the 73 patients with normal lactate levels, but abnormal bicarbonate or AG, there were seven (9.6%) deaths. Conclusions: A normal lactate level alone should not be used to exclude life-threatening sepsis. Patients with metabolic acidosis characterised by low bicarbonate or high AG levels, but with normal lactate levels, have high rates of ICU requirement and mortality and should also be considered for early, aggressive therapy.
- emergency department
- severe sepsis