The evidence for prognostication using lactate is often based on arterial lactate (AL). Arterial sampling is painful and difficult, and carries risks. Studies comparing peripheral venous lactate (PVL) with AL showed little difference but predominantly included patients with normal lactate. The objective of this study was to measure agreement between PVL and AL in patients with elevated venous lactate.

This is a retrospective cross-sectional study. Inclusion criteria: ED patients age ≥16, attending from October 2010 to June 2011 inclusive, with PVL ≥2.0mmol/L and AL taken within 1 hour. Exclusion criteria: intravenous fluid prior to or between initial venous and arterial sampling. Primary endpoint: agreement between PVL and AL defined as mean difference ±95% limits of agreement (LOA). The misclassification rate was assessed.

N=232. VL median 3.50mmol/L, range 2.00 to 15.00mmol/L. AL median 2.45mmol/L, range 1.0 to 13.2mmol/L. The mean difference±SD between PVL and AL for all patients was 1.06±1.30mmol/L (95%LOA −1.53 to 3.66mmol/L). Using a cut-off of 2mmol/L and 4mmol/L, 36.2% and 17.9% of patients respectively were incorrectly classified as having elevated lactate.

We report greater bias between VL and AL with broader LOA than previously documented. This may partly be due to the fact that we studied only patients with abnormal venous values, for whom close agreement would confer greatest clinical significance. The agreement between abnormal PVL and AL is poor and the high rate of misclassification may suggest that PVL is not a good substitute for AL if the venous lactate is abnormal.