EDITOR:

We assayed lactate levels in plasma using point-of-care analysers and obtained a fallaciously high value when compared to the value obtained from the central laboratory. The divergence in the lactate values suggested the possibility of ethylene glycol poisoning, but due to the limited information valuable time was lost in initiating treatment.

A 36-yr-old male was brought to the hospital in an unconscious state with a core temperature of 32.9°C. Glasgow coma score on arrival in the accident and emergency department was 3. Clinical evaluation and an urgent computed tomography (CT) scan of the head ruled out intracranial pathology. The patient was moved to the critical care unit for further management.

Blood gas analysis breathing 50% oxygen showed a pH of 6.8, PCO₂ 1.4 kPa, PO₂ 34 kPa, HCO₃⁻ 1.0 mmol L⁻¹, base excess −26 mmol L⁻¹ and lactate 33 mmol L⁻¹. Routine blood tests showed a lactate of 15.8 mmol L⁻¹. Past medical history revealed a suicidal tendency with previous admission to the hospital with paracetamol overdose. Toxicology screening was sent. A portable ultrasound of the abdomen was performed and contrast CT was planned to rule out intra-abdominal conditions. Fluid resuscitation, haemodynamic support, respiratory support and later renal replacement therapy were initiated. There was no improvement in acidosis despite treatment.

All the arterial blood gas samples analysed from the critical care unit using a Radiometer ABL 725 (Radiometer Medical A/S, Bronshoj, Denmark) blood gas analyser showed consistently high levels of lactate despite a sodium bicarbonate infusion and haemofiltration. Thus the cause of the metabolic acidosis remained obscure. The lactate levels from the clinical chemistry laboratory taken at the same time did not match those from the ICU analyser. These differing results aroused the suspicion of ethylene glycol poisoning and further samples were sent for methanol/ethylene glycol detection to the regional toxicology laboratory. It was not before another 6 h that 333 mg L⁻¹ of ethylene glycol was detected in the blood and appropriate treatment (4-Methylpyrazole) was instituted. Within an hour of this treatment the acidosis began to improve and the cardiovascular support could be reduced. The patient’s condition progressively improved and the lactate had fallen to 0.5 mmol L⁻¹ by the 4th day. He was discharged to a tertiary care unit for further management of his renal failure.

We analysed some samples of plasma with a known quantity of glycolic acid and compared the results obtained from our ICU blood gas analyser with those from our central clinical chemistry laboratory (Cobas Integra 400 plus analyser; Roche Diagnostics, Basel, Switzerland). The levels of glycolic acid added and the discrepancies in the two sets of observations are shown in Table 1. This clearly demonstrates the lactate gap in all the readings, which increases linearly as the glycolic acid levels in plasma increased.

Table 1 Glycolic acid measured as lactic acid.

A known concentration of glycolic acid was added to a sample of plasma and then measured as lactate by the two systems. Data are all mmol L⁻¹.

The delay in diagnosing this case of ethylene glycol poisoning could have been possibly averted by awareness of this artifactual elevation of lactate levels by our point-of-care analyser and facilities for the prompt detection of ethylene glycol in blood. The lactate gap has been described as divergent lactate levels obtained from a single sample when measured using two different modalities. Recently an erroneous reading by a Radiometer 700 analyser (Radiometer Limited, Crawley, UK) resulted in the patient having an emergency laparotomy [Reference Brindley, Butler, Cembrowski and Brindley2]. Our patient presented to us approximately 12 h after ingestion of anti-freeze, an interval corresponding with high levels of toxic metabolites. Clearly, ethylene glycol metabolites were causing falsely elevated lactate levels. This was attributed to the large dose of ethylene glycol consumed and delay in the time of arrival at the hospital.

The turning point in the diagnosis and subsequent management in our case were two different lactate values measured from the analyser in our critical care unit and that in the clinical chemistry laboratory. The method for lactate measurement in the latter equipment utilises a lactate oxidase method [3]. This enzyme converts lactate to pyruvate and produces hydrogen peroxide (H₂O₂). The peroxide reacts with 4-aminoantipyrine and other unspecified reactants to form a coloured product that is quantified colorimetrically. This method has the practical advantage of having improved reagent stability when compared to alternative methods based on lactate dehydrogenase. However, lactate oxidase may be less specific for the substrate lactate than lactate dehydrogenase. The l-lactate analyser such as that in our ICU, which is widely used to monitor lactate levels in critical care units, measures lactate using the l-lactate oxidase method. However, it uses an electrochemical principle. Lactate determination is accomplished by the enzymatic reaction of lactate oxidase and the detection of H₂O₂ [Reference Mansouri4]. It seems that most lactate oxidase-based systems respond to glycolate. The difference in response probably depends on the way in which the reaction is monitored. The false-positive results from the ICU equipment occur because ethylene glycol metabolites are substrates for l-lactate oxidase. In contrast, ethylene glycol metabolites cause minimal lactate elevation with the Bayer, iSTAT and Vitros devices [3].

Metabolites of ethylene glycol cause a worse acidosis than the parent compound itself. These may continue to remain in blood for a variable period of time. Increased glycols are measured as lactate in blood [Reference Morgan, Clark and Clague1]. However, despite the equipment being used so frequently for blood gas analysis in the critical care unit, it has not been an object of attention for its erroneously high lactate readings in the presence of glycolic acid. Glycolic acid may account for as much as 96% of the anion gap in patients poisoned with ethylene glycol [Reference Gabow, Clay, Sullivan and Lepoof5]. Point-of-care test systems may not mark the reaction course as atypical or erroneous. Therefore, having the local laboratory check a high lactate value is prudent, particularly if the diagnosis is not firmly established. In this regard it is useful to have previously established the response to glycolate of both point-of-care and laboratory systems. This case has highlighted the fact that more understanding of equipment and its mechanisms of action are critical in interpreting the data.