Peri‐procedural ketoacidosis risk with sodium‐glucose cotransporter 2 inhibitor use in people without diabetes

We wish to inform readers of the potential peri-procedural risk of ketoacidosis in association with sodium-glucose cotransporter 2 inhibitors (SGLT2i) use in patients without diabetes. These are increasingly prescribed for patients with heart failure and chronic kidney disease, irrespective of their diabetic status. Until now, the risk of ketoacidosis with SGLT2i use has been reportedly limited to individuals with type-2 diabetes. However, the Australian Product Information for empagliflozin has recently changed and now states “cases of ketoacidosis have also been reported in patients without diabetes” [1]. Amongst the heart failure and renal SGLT2i outcome studies, there was one case of ketoacidosis in a patient without diabetes randomised to empagliflozin treatment in the EMPA-KIDNEY study [2]. There has been one published case report of ketoacidosis in a normoglycaemic patient taking an SGLT2i for heart failure [3]. The case was an 83-year-old female who had fasted from midnight for a scheduled transcatheter aortic valve replacement which took place in the early afternoon. She had taken dapagliflozin that morning. She had stage-2 chronic kidney disease and experienced diarrhoea 6 days before the procedure. Of note, she was hypoglycaemic immediately post-procedure and the following morning when ketoacidosis was diagnosed. There are multiple potential mechanisms for periprocedural ketoacidosis in the population of patients who do not have diabetes (Fig. 1). The use of SGLT2i decreases the insulin-to-glucagon ratio even in people without diabetes [4]. Additionally, extended fasting in the peri-operative period may contribute, as suggested in the case report. In a study examining ketone levels at the time of colonoscopy, seven of 151 normoglycaemic participants (not taking SGLT2i) had beta-hydroxybutyrate concentrations > 1.5 mmol.l-1 [5]. Starvation before a procedure would be exacerbated by glycosuria caused by SGLT2i use (this occurs even in normoglycaemic individuals). Other potential mechanisms include increased catecholamines and cortisol associated with surgery and dehydration, all of which may change substrate utilisation from carbohydrates to fatty acids, resulting in ketogenesis. This was demonstrated in rat studies, where dehydration and insulinopenia were necessary for developing ketoacidosis in rats (without diabetes) treated with dapagliflozin [6]. Many heart failure and chronic kidney disease patients are also taking diuretics and are prone to dehydration in the peri-operative period. Finally, kidney disease could theoretically contribute to ketoacidosis, given the potential for reduced excretion of ketones. Hence, continuing the use of SGLT2i until surgery in people without diabetes may carry a rare, but not insignificant, risk of ketoacidosis. We suggest that anaesthetists be mindful of the fasting duration of the patient (both with respect to carbohydrate and water intake), check glucose and ketone levels before, during and after the procedure, minimise dehydration and correct hypoglycaemia peri-procedurally. If ketone level is > 1.5 mmol.l-1, acidosis should be excluded.