Question 11.1

College answer

Not available.

Discussion

Empagliflozin? They are basically throwing this at you.

No, wait, wait. Let us dissect these results systematically.

1. It does not take a spark of genius to recognise that there is nothing wrong with the patient's oxygenation, but let's look at it anyway for the exercise. The A-a gradient is not raised:
   PAO₂ = (0.21 × 713) - (11 × 1.25) = 136
   Thus, A-a = (138 - 136 = 2mmHg.
2. There is severe acidaemia
3. The PaCO₂ is appropriately depressed
4. The SBE is -31.3, suggesting a profound metabolic acidosis.
5. The respiratory compensation is probably as good as one could wish for. The expected CO₂ is 8.7 (40-31.3) by one method, or (1.5 × 2 ) + 8 = 11 by another method.
6. The anion gap is elevated:
   (142) - (116 + 2) = 24;
   and this is not explained by the paltry lactate of 3.2 
   The delta ratio is 12/22 = 0.54, suggesting a mixed metabolic acidosis with both a normal and a high anion gap component.
7. The osmolal gap is already given to us, and it is normal, which narrows the list of MUDPILES down to the causes of HAGMA which are not associated with excess osmoles.

So, this high anion gap metabolic acidosis in a euglycaemic diabetic. What could this possibly be? 

Euglycaemic ketoacidosis comes to mind. It is the natural conclusion in this situation, where the stem clearly gives a history of an SGLT2 inhibitor. It is possible that the college would have wanted more detail, as they asked to "explain" rather than "list" or "give" the most likely diagnosis. In that case, one could go into the mechanism, where:

• Glycosuria by the SGLT2 inhibitor leads to lower blood glucose
• Lower blood glucose leads to decreased insulin secretion
• Lower insulin levels lead to a lower insulin/glucagon ratio
• The dehydration also produces the release of cortisol and catecholamines
• Glucagon, cortisol and catecholamines drive lipolysis and ketogenesis

Or at least that's the shortest version of a mechanism described by Bui & Nawathe (2018). Now, to "outline how you would investigate this further". The diagnosis of EDKA rests on the finding of a high anion gap acidosis with raised ketones, where the BSL is below 200 mg/dL, which is 11.1 mmol/L in local terms (Barski et al, 2019). So... a blood ketone level is "how you would investigate this further". However, the college have attributed 20% of the marks to this question, which suggests they might have expected something more than just a one-liner. If one felt compelled to write more, one could hold forth as follows:

• Primarily a diagnosis of exclusion
• Other causes of HAGMA would have to be excluded
• Thus:
  • Blood or urine ketones (to confirm the ketosis)
  • Drug levels (to exclude toxicological cause of  acidosis, eg. salicylate toxicity)
  • Urinary oxalate level (late presentation of ethylene glycol ingestion)
  • Urea and creatinine (renal failure/uraemia)
  • 5-oxoproline level (pyroglutamic acidosis)

But there is also a NAGMA here, which would remain unexplained with the HAGMAcentric explanation above.  Where did the chloride come from? A reader (thank you, Daniel Chung) has made some excellent suggestions for plausible explanations, including saline (possibly even hypertonic saline) given the neurocritical nature of the presentation, and a renal tubular acidosis due to the use of perindopril. Another potential explanation is the much-spoken-of tendency of ketoacids to be eliminated renally alongside with cations, which would decrease the strong ion difference and give rise to a normal anion gap metabolic acidosis. Or, if you prefer a more traditional approach to acid-base, the NAGMA develops because ketones are bicarbonate precursors and their renal excretion represents a loss of potential bicarbonate as in Adrogué et al (1982). Unfortunately a lot of the data that traditionalists like Adrogué use to support this "excreted with cations because electroneutrality" hypothesis comes from ancient studies such as Oh et al (1978) or Guest et al (1947!) where the authors seemed genuinely puzzled that their patients all ended up with hyperchloraemic acidosis after "only four litres of isotonic saline". In other words, this phenomenon may be entirely the effect of the intervention, and nothing to do with the ketones whatsoever. Still, this explanation remains sufficiently embedded in human thinking that even highly respected and authoritative resources (eg. Palmer and Emmett from UpToDate) continue to repeat it, with the same ancient references. 

Another reader (many thanks, Rashmi) has rightly pointed out that the presence of a normal osmolar gap brings the ketoacidosis hypothesis into question. Basic chemistry suggests that the presence of ketones should change the measured osmolality. On the other hand, ketoacidosis is not usually one of the top five, when you think of the causes of a raised osmolar gap. There are a couple of reasons for this:

• A normal osmolar gap is less than 10 and this range is wide enough to encompass some  ketones, i.e. modestly raised levels.
• Some authors contend that the ketoacids cannot increase the osmolality because of buffering and CO₂ elimination ("When lactic acid or a ketoacid accumulates in the extracellular fluid the hydrogen ions from the acid combine with bicarbonate, creating carbonic acid, which is then expired as carbon dioxide. Thus, multiplying the sodium concentration by 2 includes all the osmoles derived from these acids.")
• Ketones diminish rapidly with therapy, which further reduces their contribution to the measurement of osmolality (if it is delayed)

A high osmolar gap is therefore usually not the expected finding in ketoacidosis, to the extent that finding one is grounds for a case report; or at least that is the conventional teaching. Reputable resources, including LITFL, typically do not include ketones in the list of causes for a high osmolar gap, and from this it follows that the CICM examiners would also not consider this in their thinking. As always, the evidence is not entirely in agreement with this stance. Ketoacidosis patients frequently have elevated osmolar gaps in case series and some investigators have attributed the occasional incident of paediatric cerebral oedema to the rapid correction of ketones without attention to their osmolar effects.