A step by step approach to uncovering the cause of an elevated anion gap metabolic acidosis.
We learn about the MUD PILES, the causes of anion gap acidosis, as medical students. And it gets even further drilled into us in residency. But sorting out a gap acidosis can be real challenge, even with a nifty mnemonic. To help us get smarter in understanding some of the nuance of gap acidosis, Sean Nordt, MD, PharmD.
Case: Alcoholic, diabetic with a blood glucose of 295, bicarbonate of 12, and an anion gap 28. Is this alcoholic ketoacidosis (AKA), diabetic ketoacidosis (DKA), toxic alcohol, something else? What is the cognitive process for sorting out this anion gap acidosis?
Nordt: Without additional history, send…
-Serum ketones (acetone and beta hydroxybutyrate) if possible
-Serum calcium- a good surrogate marker for ethylene glycol. Most hospitals have a volatile alcohol screen looking for methanol and isopropanol, but not ethylene glycol. To detect ethylene glycol, you’ll need to look at surrogate markers.
-Start IV fluids
Case continues: The patient has a normal mental status. Heart rhythm is sinus tachycardia in the low 100s. To treat this sinus tachycardia, he gets the sinus tachycardia antidote – 3 liters of normal saline. Since AKA (a starvation and volume depletion ketosis) is high on the differential diagnosis, he also gets a hamburger and apple juice. His labs are rechecked and few hours later and his bicarbonate is unchanged at 12 and anion gap drops slightly from 28 to 24.
How fast should the anion gap and serum bicarbonate to correct in AKA?
Nordt: It should start to improve in 1-2 hours and takes about 5-7 hours to reverse. If the anion gap and bicarbonate aren’t improving (or getting worse) in an hour or two, think about an alternate diagnosis.
Case continues: Since there’s not much improvement in the gap acidosis after several hours of re-feeding and fluid resuscitation, maybe this is not a case of AKA. Could it be DKA? Probably, but it could also be one of the many other causes of anion gap acidosis. With a persistent gap acidosis, or maybe one that’s hard to figure out, are there clues to look for or other tests that can help?
Let’s go through the different things that we use, or tests we order, to sort out the gap.
Urine Ketones: May or may not be helpful. IN AKA, the predominate ketone is beta-hydroxybutyrate. The UA only looks for acetone. Also not helpful with toxic alcohols. May he helpful in later states of AKA.
Osmolar gap: A challenge to use effectively. For toxic alcohols, it’s only high early in the ingestion. When the alcohol gets metabolized, the osmolar load is less because the toxic metabolites don’t contribute to the osmolar load. If you forget to account for ethanol in your equation, you can have an elevated osmolar gap. A confounding factor is that the normal osmolar gap ranges from -14 to +14. So if your patient drank ethylene glycol but had a non-intoxicated osmolar gap of 0, they might get their gap up to 10 which, for them, is high, but within the normal laboratory range.
Ethanol: If the level is higher than 100, it’s less likely (not impossible) that there’s a toxic alcohol on board. If there was enough ethanol on board before a toxic alcohol ingestion, the patient shouldn’t make many toxic metabolites. If the ethanol is less than 100, toxic alcohols should still be on the differential. A high ETOH level does not completely rule out a toxic alcohol, it just makes it less likely. In other words, there should not be a metabolic acidosis unless the toxic alcohol was ingested before ethanol. In that case, the toxic alcohol would already have been metabolized and the ETOH would not be protective.
Serum Lactate: Helpful in sepsis. Confounder- one of the metabolites of ethylene glycol can cause a falsely elevated lactate on the lab assay.
Urine Calcium Oxalate Crystals: Looking for dihydrate (envelope shaped) crystals. The presence of crystals raises concern for a toxic alcohol ingestion. Absence of crystals does not help. The limitation of crystals is that if your patient is early in the poisoning, they may be acidemic but not have crystals. It takes a while for crystals to form
Woods lamp urine: Mostly useless. Normal urine can fluoresce, especially with high urine phosphorous. What might help is using the Woods lamp on their clothing to look for spill patterns as would come from drinking antifreeze.
Serum Calcium: You can’t get calcium oxalate crystals without calcium. As more crystals form, the patient will become increasingly hypocalcemic. Diuresing your patient can also cause the calcium to drop, but not significantly.
Working through the MUD PILES on this patient…
Methanol – Still possible
Uremia- No. Renal function normal
DKA- Still possible
P- Propylene Glycol. Possibly. Propylene glycol gets metabolized to lactic acid. Excessive propylene glycol from medications is unlikely in the ED (unlike the ICU where patients get many injectable meds), but can be found in pet friendly antifreeze.
INH- No Iron- No Infection-No evidence of this
Lactic acidosis- Still possible. He takes metformin but renal function is normal
Ethylene Glycol – Still possible
Salicylates- Salicylate level negative
The remaining MUD PILES…
- Lactic Acidosis
- Propylene glycol
- Ethylene Glycol
How do we figure out the answer? We don’t always have to have the right answer in emergency medicine, but we shouldn’t be wrong.
Methanol- if you can’t get a level, it’s going to be a challenge because there aren’t good surrogate markers. Methanol is very osmotically active so, early on, it’s likely to elevate the osmolar gap.
Propylene glycol- if the lactate is getting worse or is severely elevated, think propylene glycol
Ethylene Glycol- when they get sick, they often have single digit bicarbonates. Watch their calcium level (hypocalcemia is a surrogate marker). Patients with toxic alcohol ingestions get sick quick and continue to get sicker. If in doubt, start fomepizole.
More on CPR (or no CPR) in trauma arrest with Scott Weingart