Perioperative Implication of Sodium-glucose Cotransporter-2 Inhibitor in a Patient Following Major Surgery

Sodium-glucose cotransporter-2 inhibitors (SGLT2i) are being hailed for their cardiovascular and renal protective effects. In view of evidence of a reduction in hospitalization for heart failure, cardiovascular-related deaths, and all-cause mortality, they are increasingly prescribed in patients with diabetes mellitus type 2 with cardiovascular disease. Diabetic ketoacidosis (DKA) with lower-than-anticipated glucose levels is an important SGLT2i-related adverse effect in postoperative patients. Decisions regarding discontinuation and restarting medications in the perioperative period are complex issues. Herein we highlight the need for increased postoperative surveillance of patients on this group of medications.

A 56-year-old male, known hypertensive, sleep apnea, and diabetes mellitus type 2 (Acarbose—since a year and Metformin, Liraglutide, Lantus insulin, and Dapagliflozin—since 5 years) was hospitalized for coronary artery bypass grafting. His body mass index was at 41 kg/m². Acarbose was taken 72 hours prior, and Metformin and Dapagliflozin were last taken 48 hours prior to the day of surgery. Preoperative blood glucose control was managed by intravenously administered customized scaled short-acting insulin. Intraoperative course was uneventful. He did not receive any steroids. Postoperative hemodynamics were stable. He complained of excessive thirst postoperatively. Arterial blood gas revealed metabolic acidosis. Given no other causes of metabolic acidosis and a medication history of SGLT2i, urine ketones were checked and reported “large.” His capillary blood glucose levels ranged from a minimum 107 mg/dL to maximum of 280 mg/dL. (Table 1) This was managed as DKA with lower-than-anticipated glucose levels. Postoperative urinary glucose was 2245 mg/dL. Subsequently, his beta-hydroxybutyrate levels came positive at 0.81 mmol/L. Stabilized, he was discharged on subcutaneous short and long-acting insulin. On follow-up, his fasting blood glucose and postprandial blood glucose have been under control with negative urine ketones.

Inadequate clearance of SGLT2i and the stress of major surgery are precipitating factors for DKA with lower-than-normal glucose levels. An important point to remember is that SGLT2i aid reabsorption of ketone bodies. Consequent ketoacidosis may not be accompanied by ketonuria.¹ SGLT2i causes glucosuria with a resultant reduction in insulin secretion and use. This leads to a reduction in insulin to glucagon ratio. Insulin requirements are masked. Hence, reduction in dose or discontinuation of insulin after major surgery should be avoided.² SGLT2i-related urinary glucose losses may persist for several days.³ All oral hypoglycemic agents are conventionally withheld on the day of surgery. This may not hold true for the SGLT2i group of medication. American Association of Clinical Endocrinologists and American College of Endocrinology recommend withholding SGLT2i for at least 24 hours prior to any event that may precipitate DKA, to be restarted once the patient has resumed normal diet, while Food and Drug Administration⁴ recommends withholding for 3 days prior to surgery. Half-life of SGLT2i is 11 to 13 hours. However, their pharmacodynamic effects may persist beyond the five half-lives² with glucosuria and ketonemia seen up to 9 to 10 days post-SGLT2i discontinuation.⁵ Hence, it would be prudent to withhold this group of medication for more than 5 days preoperatively, requiring coordination among the surgical, anesthesia, endocrinology, and intensive care unit teams. Perioperative monitoring of acid–base and ketone status is mandatory, with monitoring to be continued until the postoperative stress is resolved. SGLT2i may be reintroduced once the oral intake is resumed to normal.

Awareness of DKA with lower-than-anticipated glucose levels is an important clinical challenge, an entity that can be confused in the setting of major and complex surgeries. The frequency of this arcane and underreported diagnosis in the perioperative setting is unknown. Patients with diabetes mellitus on SGLT2i, who develop metabolic acidosis, need to be knowingly investigated for DKA with lower-than-anticipated glucose levels. The optimal timing of discontinuation of SGLT2i’s preoperatively is undetermined but definitely needs to be longer than several of their elimination half-lives. We suggest withholding SGLT2i for at least 7 days prior to major surgery. Knowledge of the existence of this entity in the “real world” clinical practice with vigilance in its diagnosis by astute postoperative surveillance will help to identify more cases, thus influencing outcomes.