INTRODUCTION: In the emergency department (ED), lactate is routinely used for risk stratification. Whether venous or arterial lactate measured on blood gas is interchangeable is not known. We hypothesized that venous lactate can be used instead of arterial lactate for the evaluation of acute patients in the ED. PATIENTS AND METHODS: This was a prospective single-center study. All patients requiring a lactate measurement were enrolled and we simultaneously drew arterial and venous blood. We followed up all patients to hospital discharge. Our primary aim was to evaluate agreements between the two measurements using Bland and Altman plots with the report of bias (mean difference) and limits of agreements. We also aimed to determine the rate of misclassification (defined as one measurement<1.8 mmol/l and the other>2.2). Our secondary aim was to evaluate their respective prognostic value to predict in-hospital death or admission in the ICU longer than 72 h. RESULTS: The mean age of the 132 analyzed patients was 62 years (SD: 18 years), and 59% were men. The mean difference (bias) between arterial and venous lactate was -0.6 mmol/l (limits of agreement: -1.7 to 0.6 mmol/l). The rate of misclassification was 8% (95% confidence interval: 3-2%). Both methods present similar performances for the prediction of poor outcomes, with an area under the receiving operator characteristic curves of 0.67 for both. Results were similar when focused only on septic patients. CONCLUSION: Venous and arterial lactates do not agree well, and there is a high misclassification rate. Venous lactate does not appear to be interchangeable with arterial sampling.
INTRODUCTION: In the emergency department (ED), lactate is routinely used for risk stratification. Whether venous or arterial lactate measured on blood gas is interchangeable is not known. We hypothesized that venous lactate can be used instead of arterial lactate for the evaluation of acute patients in the ED. PATIENTS AND METHODS: This was a prospective single-center study. All patients requiring a lactate measurement were enrolled and we simultaneously drew arterial and venous blood. We followed up all patients to hospital discharge. Our primary aim was to evaluate agreements between the two measurements using Bland and Altman plots with the report of bias (mean difference) and limits of agreements. We also aimed to determine the rate of misclassification (defined as one measurement<1.8 mmol/l and the other>2.2). Our secondary aim was to evaluate their respective prognostic value to predict in-hospital death or admission in the ICU longer than 72 h. RESULTS: The mean age of the 132 analyzed patients was 62 years (SD: 18 years), and 59% were men. The mean difference (bias) between arterial and venous lactate was -0.6 mmol/l (limits of agreement: -1.7 to 0.6 mmol/l). The rate of misclassification was 8% (95% confidence interval: 3-2%). Both methods present similar performances for the prediction of poor outcomes, with an area under the receiving operator characteristic curves of 0.67 for both. Results were similar when focused only on septic patients. CONCLUSION: Venous and arterial lactates do not agree well, and there is a high misclassification rate. Venous lactate does not appear to be interchangeable with arterial sampling.