NEJM discussion of ethylene glycol poisoning misses crucial points

February 3, 2015, 1:51 pm


Case 4-2015: A 49-Year-Old Man with Obtundation Followed by Agitation and Acidosis. Cooper CM, Baron JM. N Engl J Med 2015 Jan 29;372:465-473.


This episode of the Journal‘s “Case Records of the Massachusetts General Hospital” series is the latest of their amazing inept discussions of medical toxicology cases. Yet again, the discussants do not include a medical toxicologist or emergency physician. Given the level of toxicological awareness evidenced by this case presentation, it’s fortunate that the patient did well.

A unresponsive 49-year-old man was found outside with a half-empty bottle of cloudy liquid and an empty bottle of trazodone that had been filled earlier that day. On arrival in the emergency department, he would open his eyes and move his extremities in response to painful stimuli, but did not speak (Glasgow Coma Score = 8.) His vital signs were unremarkable, his serum glucose was 98 mg/dL, bicarbonate 23.9 mom/liter and anion gap 21. Other basic labs, including serum calcium and creatinine, as well as salicylate and acetaminophen levels, were unremarkable.

I have simplified the details of the case somewhat, but let’s stop here. We have an intermittently unresponsive and agitated man with a metabolic acidosis (evidenced b the increased anion gap) who had been found with a half-empty bottle of liquid. What would be the next steps in managing this patient?

Given in elevated anion gap along with the apparent ingestion of a liquid, this should be considered a case of toxic alcohol ingestion until proven otherwise. Since the anion gap is only mildly elevated, this seems to be an early acute ingestion. An immediate dose of fomepizole will minimize conversion of methanol or ethylene glycol to their much more dangerous metabolites. Specimens should be sent for toxic alcohol levels, but the initial fomepizole dose should not be delayed waiting for the lab tests. For a further clue a serum osmolality can be measured and the osmolal gap calculated. An elevated osmolal gap supports the diagnosis of toxic alcohol ingestion, but a normal result should not dissuade the clinician from administering fomepizole.

In this case, none of the above was done. A serum trazodone level came back supra-therapeutic. Seven-and-a-half hours after presentation, lab tests were repeated. At this time the serum bicarbonate was 8.3 mmol/liter and the anion gap 34. Ten hours (!!) are presentation another set of laboratory tests showed at serum bicarbonate of 5.7 mmol/liter and an anion gap of 35. At that point: “A diagnostic test was performed.”

At 10 hours after presentation, the measured serum osmolality was 394 mOsm/kg with a calculated osmolality of 300, yielding an osmolal gap of 94 (“normal” < 10-15.) The ethylene glycol level was 351 mg/dL (toxic level > 20 mg/dL.) Note that since the increasing acidosis resulted from metabolism of ethylene glycol to glycol acid and oxalic acid, the initial level would certainly have been much higher.

In my opinion, the key lesson in this case is the need to administer the antidote fomepizole early when there is reasonable suspicion of toxic alcohol poisoning. this crucial point goes virtually unmentioned in the discussion. Although the times at which appropriate interventions were initiated are a little unclear from this case presentation, the fact that acidosis progressed for 10 hours and that a serum osmolality was not measured until that 10 hour mark suggests that these were not started until quite late. And of course, there is no evidence that consultation with medical toxicology or the poison center was obtained.

One interesting point made in the discussion: Glycolate, a metabolite of ethylene glycol, can cross-react and give a falsely elevated serum lactate level. In this patient, the plasma lactate was measured as 9.0 mmol/ligter.
Related posts:

Surely the NEJM could do a better review of delirium tremens

CPC: alcohol withdrawal with delirium tremens and a significant missed diagnosis

More criticism of New England Journal CPC on alcohol withdrawal


  1. Dan P Says:

    The NEJM Cases always involve over-the-top testing for every conceivable etiology. Are these tests billed to the patient?
    Does obtaining a trazodone level with metabolites provide any clinically useful information?

  2. Leon Says:


    Great points! You are correct. By history the patient had taken an overdose of trazodone, which is probably what caused the urinary retention from anticholinergic effects. However, trazodone is relatively safe even in overdose, and a level adds absolutely nothing to the clinical care of the patient.

    The entire discussion involved academic aspects of the differential diagnosis without at all focusing on critical points of patient care. I found it astonishing that they didn’t have a toxicologist discussing the case, or at least adding comments at the end. Then again, that’s what post-publication review on social media is for!

  3. Eric Strong Says:

    Nice critique of a case that seems to be more straight-forward than the NEJM and MGH is implying with the length of their discussion. When a patient presents with obvious overdose of an unknown substance and has an elevated AG, serum lactate and osmolal gap are essential and should have been checked at presentation. It’s unfortunate that not only was that not done, but no one even brings up the obvious oversight in the case discussion. In addition, the case discussion does in fact refer to contacting poison control – they should be even more embarrassed if it took them 10 hrs to figure this out…

  4. Mike Emmett MD Says:

    Letter I wrote to the NEJM which they rejected for publication:

    “Case 4-2015 describes a man with suicidal ethylene glycol ingestion. The correct diagnosis was initially obscured by co-ingestion of other medications and it took over 10 hours for a correct tentative diagnosis and administration of Fomepizole, an antidote that should be given as soon as ethylene glycol ingestion is strongly suspected. It is important to point out that the urine sediment, showing calcium oxalate crystals, was not examined until 12 hours after presentation. Although routine urinalysis is no longer a screening test, it is a very useful and cheap laboratory test when clinically indicated. An obtunded patient found with a “half-filled bottle of cloudy fluid” whose anion gap is 21 mEq/1 should have had a complete urinalysis sent immediately. Identification of calcium oxylate crystals would have raised the very strong possibility of ethylene glycol ingestion. The plasma osmolal gap would have then been quickly determined and a dose of Fomepizole would have been promptly administered. It is unfortunate that very simple and useful tests such as a urinalysis may be forgotten or markedly delayed”

  5. Leon Says:


    Thanks for forwarding your letter. I agree that the delay to treatment was unfortunate and not really discussed in the CPC. But I’d argue that given the unexplained anion gap and the cloudy liquid, one would not need to check an osmolal gap and urine for calcium oxalate crystals before starting fomepizole and waiting for the toxic alcohol levels to come back.