Sodium–glucose cotransporter type 2 inhibitors: managing the small but critical risk of diabetic ketoacidosis

To the Editor: We would like to highlight some points arising from the discussion by Hamblin and colleagues regarding euglycaemic diabetic ketoacidosis associated with sodium–glucose cotransporter type 2 (SGLT2) inhibitors.1

First, clinicians should be aware that this condition occurs not only in the perioperative context but also in systemically unwell patients with medical problems. Apart from the periprocedural insult, four categories of precipitating factors are recognised: intercurrent illness; dietary modifications (eg, prolonged fasting, very low calorie diet); medication changes (especially reducing or stopping insulin); and health system factors (eg, use in misdiagnosed type 2 rather than type 1 diabetes, and lack of patient education on the handling of SGLT2 inhibitors perioperatively).2,3,4

Second, diabetic ketoacidosis is more frequently reported with major surgery; for example, cardiothoracic, bariatric and abdominal surgery (postoperative ileus contributing).

Third, clinicians should be aware that the current recommendations5 are based on low quality evidence and are potentially subjective. For example, the use of glycated haemoglobin levels < 75 mmol/mol (9%) as one factor to stratify lower patient risk, while intuitive, is not an unequivocal finding in the literature.4 We are in agreement that one should not overreact to capillary ketone levels in the perioperative period; these should be interpreted in conjunction with other acidosis markers (pH, bicarbonate and base excess). We differ regarding the authors’ statement that blood ketone testing is warranted only in unwell or symptomatic patients. In our clinical experience, we have encountered asymptomatic presentations with ketone levels > 2.0 mmol/L and acidosis before colonoscopy, despite the cessation of SGLT2 inhibitors on the day of the scheduled colonoscopy, necessitating deferral and inpatient treatment.6 Bowel preparation, diet modification and changes in diabetes medications are possible contributory factors for a minor procedure such as colonoscopy.

Further, not all patients attend a pre‐assessment clinic and on the day of their procedure may be unable to recollect their diabetic medications. For these reasons, as recommended by the Australian Diabetes Society, it is prudent to check capillary ketones (using a single glucose strip) on admission for all patients with type 2 diabetes regardless of symptoms in the periprocedural period.5

Finally, in patients who have not held their SGLT2 inhibitors sufficiently or who have ketosis, the decision to proceed should depend on a nuanced appraisal integrating the complexity of the procedure, precipitating factors, and degree of acidosis.