Critical Failure in Perioperative SGLT2 Inhibitor Management Leading to Life-Threatening Euglycemic DKA
What Went Wrong: The Core Failure
The fundamental error was continuing canagliflozin (an SGLT2 inhibitor) up to the day of surgery without proper preoperative discontinuation, combined with prolonged fasting (20 hours NPO), which created the perfect storm for euglycemic diabetic ketoacidosis (euDKA) that was missed until it became life-threatening. 1
The FDA label for canagliflozin explicitly states: "Withhold INVOKANA at least 3 days, if possible, prior to surgery or procedures associated with prolonged fasting" 1. This was not done. The American College of Cardiology/American Heart Association 2024 guidelines recommend withholding canagliflozin ≥3 days before scheduled surgery 2, 3. Even the more conservative UK guidelines recommend stopping SGLT2 inhibitors at minimum the day before surgery 4, 3.
The Cascade of Missed Opportunities
1. Preoperative Planning Failure
- The patient should have stopped canagliflozin 3-4 days before surgery 2, 3, 1
- Emergency surgery carries significantly higher risk of perioperative ketoacidosis (1.1%) versus elective surgery (0.17%), but this was not emergency surgery—it was appendicitis diagnosed at 10:30 PM with surgery delayed until 10:40 PM the next day, providing ample time to recognize the SGLT2 inhibitor risk 2, 3
- The patient had multiple risk factors for SGLT2i-induced euDKA: acute illness (appendicitis), fever, reduced oral intake for days, volume depletion, and upcoming surgery 1
2. Failure to Recognize Early Metabolic Acidosis
The initial presentation already showed metabolic acidosis that was completely missed:
- pH 7.24, bicarbonate 17, anion gap 15, base excess -9 on admission 5
- No ketone testing was performed despite clear acidosis in a diabetic patient on an SGLT2 inhibitor 1
- The FDA label explicitly warns: "Assess for ketoacidosis regardless of presenting blood glucose levels in patients who present with signs and symptoms consistent with severe metabolic acidosis" 1
- Blood glucose was likely normal or only mildly elevated (not documented initially), which is characteristic of euDKA and makes it easy to miss 2, 1
3. Prolonged Fasting Without Glucose-Containing Fluids
- The patient was kept NPO for approximately 20 hours with only normal saline bolus 5
- The American Society of Anesthesiologists recommends considering glucose-containing IV fluids in cases of unavoidable prolonged fasting to mitigate ketone generation 2, 5
- Prolonged fasting is a known precipitating factor for SGLT2i-induced ketoacidosis 1
- The patient received 1000 mL NS bolus but no ongoing glucose-containing maintenance fluids despite being NPO for nearly a full day
4. Progressive Acidosis Ignored
The morning labs on the day of surgery showed worsening acidosis that was again ignored:
- Bicarbonate dropped from 17 to 13 (worsening) 5
- Chloride rose from 103 to 108 (hyperchloremic component from NS) 5
- Sodium rose to 146 (dehydration) 5
- Still no ketone testing despite progressive metabolic acidosis in a patient on canagliflozin 1
5. Catastrophic Post-Operative Discharge Decision
The most egregious error was discharging the patient with severe metabolic acidosis:
- Post-operative labs showed bicarbonate of 6, anion gap 21, potassium 5.5 5
- A bicarbonate of 6 with AG 21 represents severe, life-threatening metabolic acidosis requiring immediate ICU-level intervention, not discharge 5
- The patient was breathless with high pulse—clear signs of Kussmaul respirations attempting to compensate for severe acidosis 5, 1
- The American Society of Anesthesiologists recommends abbreviated surgery and ICU resuscitation when pH < 7.2 or base deficit > 8 5
The Mechanism: Why SGLT2 Inhibitors Cause This
SGLT2 inhibitors create euglycemic DKA through a distinct pathophysiologic mechanism:
- They alter the insulin/glucagon ratio, decreasing insulin secretion and relatively increasing glucagon, creating a hormonal environment that promotes ketogenesis despite normal glucose levels 2
- Surgical stress amplifies this effect 2
- The effects of SGLT2 inhibitors persist beyond their plasma half-life, with clinical effects continuing for 3-4 days after discontinuation 2
- Urinary glucose excretion persists for 3 days after discontinuing canagliflozin, and there have been postmarketing reports of ketoacidosis lasting greater than 6 days and up to 2 weeks after discontinuation 1
What Should Have Been Done
Preoperatively:
- Stop canagliflozin at least 3 days before surgery 2, 3, 1
- Check baseline ketones given the patient's reduced oral intake and acute illness 1
- Ensure adequate hydration before surgery 2, 3
Intraoperatively:
- Use glucose-containing IV fluids during prolonged fasting 2, 5
- Use balanced crystalloids (lactated Ringer's or Plasma-Lyte) rather than normal saline to avoid worsening hyperchloremic acidosis 5
Postoperatively:
- Serial ketone monitoring in any patient who was on SGLT2 inhibitors perioperatively 2, 1
- When bicarbonate dropped to 13, immediate ketone testing and aggressive treatment should have been initiated 5, 1
- When bicarbonate was 6 postoperatively, the patient required immediate ICU admission, not discharge 5
- Do not restart canagliflozin until the patient is eating and drinking normally AND capillary ketones are <0.6 mmol/L 2, 3
The Final Presentation: Severe Euglycemic DKA
The patient returned within 2 hours with textbook severe euglycemic DKA:
- pH 6.99 (profound acidosis, incompatible with life if not immediately corrected) 5, 1
- Bicarbonate 4, anion gap 23 5
- β-hydroxybutyrate 6.51-6.77 mmol/L (markedly elevated, confirming ketoacidosis) 1
- Respiratory rate ~40/min with Kussmaul breathing (compensatory hyperventilation) 5, 1
- Drowsiness/confusion (cerebral effects of severe acidosis) 5, 1
This represents a near-fatal complication that was entirely preventable with proper perioperative SGLT2 inhibitor management 2, 3, 1.
Critical Pitfalls to Avoid
- Never assume normal glucose excludes DKA in patients on SGLT2 inhibitors—euglycemic DKA presents with glucose <250 mg/dL despite severe ketoacidosis 2, 1
- Always check ketones when evaluating metabolic acidosis in diabetic patients, especially those on SGLT2 inhibitors 1
- Never discharge a patient with bicarbonate <10-12 without investigating the cause and ensuring stability 5
- Remember that SGLT2 inhibitor effects persist for days after discontinuation—ketoacidosis can develop even after the drug is stopped 2, 1
- Prolonged fasting + SGLT2 inhibitors + surgical stress = high risk for euDKA 2, 1