Whether to hold and what to measure: perioperative euglycemic ketoacidosis in patients taking SGLT2 inhibitors

Sodium-glucose cotransporter 2 (SGLT2) inhibitors are widely used for type 2 diabetes, heart failure, and chronic kidney disease, including in patients without diabetes. Their ketogenic effects may be amplified perioperatively by fasting, surgical stress, and volume loss, resulting in ketoacidosis without marked hyperglycemia. Reported frequencies depend on ascertainment. Where ketones were measured against a diagnostic threshold, cohorts reported 0–1.8% in non-cardiac and mixed populations and 0–15.1% in cardiac surgery; one prospective study using a lower threshold (≥ 0.6 mmol/L) reported a far higher figure, and studies without ketone measurement reported surrogate metabolic-acidosis endpoints of 30–41%. None of these are comparable incidence estimates. Severe and fatal cases have been reported. Because plasma glucose may be normal, diagnosis requires attention to acid–base status and ketones. Quantitative blood β-hydroxybutyrate is preferred when SGLT2 inhibitor (SGLT2i)-associated ketoacidosis is suspected; urine ketone tests may screen when blood testing is unavailable, but a negative result does not exclude it. Perioperative guidance remains inconsistent: some bodies recommend withholding SGLT2i for 3 days (4 days for ertugliflozin), whereas others support risk-stratified continuation in selected patients under specified fasting and dietary conditions. No randomized trial has compared continuation with discontinuation using perioperative ketoacidosis as an endpoint; observational evidence remains limited and conflicting, and ketone thresholds and monitoring intervals lack prospective validation. This review summarizes evidence on recognition, diagnosis, withholding, monitoring, and postoperative restart.