Should SGLT2 inhibitors be held prior to elective surgery?

Sodium-glucose co-transporter 2 inhibitors (SGLT2i) are commonly prescribed to treat diabetes, slow the progression of chronic kidney disease, and reduce cardiovascular events in heart failure. They work by blocking the reabsorption of glucose in the kidneys, leading to lower blood glucose levels. This reduced glucose leads to relatively lower levels of insulin and higher levels of glucagon.

During fasting or stressful events such as an infection, ketone levels in the blood can increase to meet the body’s energy demands. Normally, insulin acts to decrease ketone production by inhibiting lipolysis, but SGLT2 inhibitors reduce this negative feedback. Simultaneously, the increased glucagon promotes further ketogenesis, leading to an increased risk of ketoacidosis. Because of the glycosuria effect, this ketoacidosis is usually euglycemic, although diabetic ketoacidosis (requiring blood glucose ≥ 200 mg/dL) can also occur.

In 2020, the FDA issued guidance recommending SGLT2i be held for three days prior to surgery to reduce the risk of euglycemic ketoacidosis. Since that time, research has questioned the benefit of this recommendation. As with any decision, the benefits of reducing the harms of an uncommon side effect should be weighed against the downsides of having poorly controlled chronic disease (e.g. blood glucose) in the perioperative setting.

A group of researchers set out to determine if patients with type 2 diabetes who were taking SGLT2i at the time of emergency surgery had a higher incidence of diabetic ketoacidosis than diabetic controls not taking SGLT2i. They performed a retrospective cohort study including 34,671 patients, of whom 2,607 were using SGLT2i. Inclusion criteria included undergoing one of 13 different emergency surgeries. Patients prescribed SGLT2i were only included if a recent prescription would have lasted through the date of surgery. Because of the emergency nature, it was assumed these patients were taking their SGLT2i; however, exact timing remains a limitation of this study. The primary outcome was diabetic ketoacidosis or acidosis (defined by ICD-10 codes) within 14 days postoperatively.¹

They found the unadjusted incidence of acidosis to be 4.9% in those taking SGLT2i and 3.5% in the unexposed group. After adjusting for covariates, SGLT2i use was not associated with an increased risk of acidosis (3.8% vs 3.5% in the unexposed group). The major limitations of this study were its retrospective, non-random nature, the exclusion of non-diabetic patients taking SGLT2i, and the lack of access to laboratory data to further inform the primary endpoint.

How should guidelines be changed?

Multiple medical groups in Europe recommend holding SGLT2i for only 24 hours and/or only holding them for patients undergoing major surgery or who are hospitalized with serious illness. The French Diabetology and Anesthesiology society recommends continuing SGLT2i for ambulatory surgeries and only holding them on the day of surgery for minor and major surgeries.

It may be reasonable to only hold SGLT2i for major surgeries or when patients are expected to continue fasting into the postoperative period. Considering that people routinely fast overnight even when not undergoing anesthesia and many also skip breakfast each day without negative consequences, it seems unreasonable to hold SGLT2i in patients undergoing minor procedures when they will resume eating the same day.

If a patient fails to hold their SGLT2i, measuring serum or urine ketones could help guide decision-making. In the U.S., Northwell Health System created a protocol to test serum β-hydroxybutyrate (BHB) in patients who did not hold SGLT2i. Patients with BHB concentrations < 1.5 mM can proceed with surgery. For patients with BHB concentrations of 1.5-3.0 mM, anion gap is also assessed. At the University of Pennsylvania, a BHB cutoff of 2.0 mM with anion gap > 12 leads to postponement of surgery. While research has yet to determine the optimal BHB level to mitigate risk, these protocols have helped avoid cancellations in low-risk instances.

Until the FDA changes its recommendation, it may be prudent for institutions to create their own guidelines weighing the risks of ketoacidosis against those of cancelled surgeries and poorly controlled disease resulting from held medications.

1. There is no ICD-10 code for euglycemic ketoacidosis