7 Quick Takeaways That Will Hopefully Help You Better Manage Diabetic Ketoacidosis (DKA)

The pathophysiology that occurs during any given disease process may not change, but how we approach this must constantly be refined. We live in an age where we can no longer rely on what we knew yesterday or what worked back then. The classics may have their place in the arts, but in the art of medicine, we must build on what worked yesterday, with what we know today.

In 2009, the wonderful work of Kitbachi et al helped us form a foundation for the treatment of DKA that we still use to this day, but I think it’s time to revisit this tried-and-true approach with some updates from the 2024 and some from 2026.

In 2024, five professional societies — the ADA, EASD, AACE, JBDS, and Diabetes Technology Society — put their heads together and reissued the international consensus on hyperglycemic crises. Some changes are subtle. Others will make you rewrite your order sets. The 2026 ADA Standards of Care now reflect all of it, and a few changes — like dropping the anion gap from resolution criteria — are still missing from many residency teaching files.

Here are seven takeaways that deserve real estate in your working memory. None require a new drug, a new monitor, or a new line on the budget. They are mostly mental adjustments — but the kind that change what your next patient’s hospital course looks like.

 

1. The glucose bar is lower than you remember.

Old rule: DKA requires glucose >250 mg/dL. New rule: anything >200 — or any glucose at all if the patient has known diabetes. That isn’t a typo. The 2024 consensus deliberately widened the diagnostic net because roughly one in ten DKA patients now presents euglycemic, and the old threshold was missing them.

This costs nothing to implement. It’s a mental adjustment. A known T1DM patient with glucose 184, pH 7.18, and β-hydroxybutyrate 4.6 has moderate DKA — full stop. Don’t anchor on the sugar. Use the sugar to adjust your drips. The goal is to prevent hypoglycemia during treatment. Personally, I shoot for 140-180 as we resolve the underlying metabolic disturbance.

2. β-hydroxybutyrate is your best friend. Urine ketones are your gossipy cousin.

The nitroprusside reaction detects acetoacetate and acetone, not β-hydroxybutyrate (BHB) — which happens to be the dominant ketone in DKA. Early in the illness, that ratio tilts up to 8:1, meaning the urine dipstick under-reports. Late in treatment, BHB metabolizes back to acetoacetate, so the dipstick stays falsely positive for up to 36 hours after the patient is metabolically resolved.

Point-of-care capillary BHB strips are >90% sensitive and specific at the ≥3.0 mmol/L threshold and cost a couple of dollars. They’re the same strips endocrinologists send home with T1DM patients. If your hospital doesn’t have them yet, this is the cheapest quality improvement on your list, but this isn’t mandatory. In rural medicine we can get around this by focusing on the pH and the bicarb level.

3. The anion gap broke up with DKA. Stop chasing it.

This is the change that catches the most attendings off guard.

The 2024 consensus no longer recommends the anion gap as a criterion for DKA resolution. Why? Because after a few liters of 0.9% saline (chloride 154 mmol/L vs. plasma 100), patients may develop a hyperchloremic, non-anion-gap acidosis (side note: we won’t cover this here, but we have to drop the saline and address the free water deficit and also use more balanced solutions). The gap stays open even as the ketones clear, and well-meaning teams hold patients on insulin drips for hours of unnecessary care. Furthermore, the gap doesn’t truly measure the resolution of the underlying pathophysiology. We need to shift the patient out of a dysregulated state of rampant lipolysis and accumulation of ketoacids which subsequently cause the metabolic acidosis. DKA resolves with the resolution of this physiology and the anion gap doesn’t properly measure this.

Resolution is now: BHB <0.6 mmol/L, and either pH ≥7.3 or bicarbonate ≥18. If the gap is still 13 but the BHB is 0.4 and the pH is 7.32, your patient’s DKA has resolved. The stickiness of the gap is iatrogenic, and it heals as the kidneys excrete the chloride load. But practically speaking, I have yet to have a patient whose gap is open while their bicarb and acidosis resolved. Usually, the gap is the first thing to resolve but the patient still needs more time.

4. Yes, basal insulin during the IV drip is allowed. (You probably should.)

Old teaching: never overlap between basal insulin and insulin drip due to risk of hypoglycemia. New evidence: multiple RCTs show that co-administering 0.15–0.3 U/kg of basal insulin (typically glargine) alongside the standard 0.1 U/kg/h IV drip, speeds DKA resolution, shortens length of stay, and prevents the rebound hyperglycemia that haunts the transition off the drip. There was no increased incidence of hypoglycemia in the studied range.

This is now explicitly endorsed in both the 2024 consensus and the 2026 ADA Standards. It’s especially valuable in T1DM, where the drip can be paused for a transport, a scan, or a lost line – and the patient can flip back into DKA before you’ve ordered the next BMP. Think of it as a safety net for the drip. A reasonable starting dose is 0.25 U/kg of glargine, given when the drip is started and continued daily until discharge. This is something I personally incorporate in every DKA/HHS patient, and it has proven to be effective.

5. Euglycemic DKA: the SGLT2 ambush.

SGLT2 inhibitors are FDA-approved and have quietly become foundational therapy for HFrEF and diabetic kidney disease. They’ve also pushed roughly one in ten DKA cases into the looks-fine-on-the-glucometer zone. About 35% of SGLT2i-associated DKA presents with glucose under 200 mg/dL. The osmotic glucosuria masks the polyuria and polydipsia; the patient just feels lousy and a little short of breath.

The fix: any unwell patient on an SGLT2i gets a β-hydroxybutyrate and a venous blood gas, right off the bat.

Treatment parallels classic DKA except: start dextrose-containing fluids at presentation (because glucose is already low), stop the SGLT2i, and watch for relapse for 24 hours after resolution. Also of note, for patients undergoing surgery, your pre-op instructions should include: hold the SGLT2i three days before scheduled surgery.

6. The pseudohyponatremia trap.

Severe DKA frequently produces gorgeous, milky lipemic serum – courtesy of insulin deficiency knocking out lipoprotein lipase, with triglycerides routinely climbing past 1,000 and sometimes past 3,000 mg/dL (I recently had a patient whose triglycerides were so high our machine couldn’t read it). Cue the lab artifact: indirect ion-selective electrode (ISE) analyzers, which most basic metabolic panels use, dilute the sample and measure sodium per total plasma volume. When lipid takes up a chunk of that volume, the reported sodium reads falsely low. Same trick with severe hyperproteinemia – myeloma, IVIG, Waldenström macroglobulinemia.

Direct ISE (the blood gas analyzer, the point-of-care meter) measures ion activity in plasma water itself and is unaffected. So, if your BMP sodium is 125 and your VBG sodiums 135 that isn’t a glitch – it’s pseudohyponatremia. Don’t frantically call your nephrologist, call for the 3%, or hold your fluids; treat the metabolic state, and the lipemia clears with insulin within hours. Also, let’s not forget, severe hyperglycemia will also falsely lower the sodium reading on the BMP. I recently had a patient whose blood glucose was in the 900s, and the sodium came back around 120. This was a textbook example of pseudohyponatremia.

7. Potassium first, insulin second, every single time.

This take-away isn’t new, but it remains the most lethal sequence error in DKA. The total-body potassium deficit is 3–6 mmol/kg even when the serum number looks reassuringly normal – because acidosis pushes potassium out of cells. When there’s a drop in insulin, that potassium slams back in. If the starting potassium level was 3.0, you’ve just given your patient a one-way ticket to a peaked-T-wave rhythm strip. You must address this before the patient’s ECG learns French (Torsades de Pointes).

Hard rule: serum K+ ≥3.3–3.5 mmol/L before insulin. If lower, hold the drip, replace IV, monitor on telemetry, and recheck. Severe hypokalemia (≤2.5 mmol/L) during DKA treatment is associated with a threefold increase in mortality. There is no scenario in which it is appropriate to “fix the potassium with maintenance fluids.” In my practice the ONLY time I hold the insulin drip is relative to the potassium level. I can get creative with maintaining the blood sugar between 140-180 (D50, increase rates of d5, switching to d10, etc) but the potassium is one that gives me a full stop – if it’s low then the drip is on hold until we are in a safe range.

Put it together now: Modern DKA care is less about the heroic protocol and more about the small choices: the BHB at the bedside instead of the urine cup, the basal insulin co-administered alongside the drip, the dextrose started before the glucose drops, the calm refusal to chase a gap that’s been replaced by chloride. The 2024 consensus didn’t reinvent the disease – it sanded the rough edges off a protocol you already know. Is this everything? No. But hopefully, this can help optimize your treatment of DKA and maybe help update your protocols.

 

Sources: Umpierrez GE, Davis GM, ElSayed NA, et al. Hyperglycaemic crises in adults with diabetes: a consensus report. Diabetologia. 2024;67(8):1455–1479. • American Diabetes Association. Diabetes Care in the Hospital: Standards of Care in Diabetes — 2026. Diabetes Care. 2026;49(Suppl 1):S339–S355. • Chow E, Clement S, Garg R. Euglycemic diabetic ketoacidosis in the era of SGLT-2 inhibitors. BMJ Open Diab Res Care. 2023;11(5):e003666. • Hirsch IB, Emmett M. Diabetic ketoacidosis (adults): Clinical features, diagnosis, treatment. UpToDate, accessed 2026.

For this article and more like it, download our FREE Digital Magazine, Fields of Care.

To learn more about our service lines, explore partnership opportunities, or stay up to date on RPG news, visit www.RuralPhysiciansGroup.com and follow us on Instagram and Facebook @RuralPhysiciansGroup.