In insulin-treated type 1 diabetes, canagliflozin increased diabetic ketoacidosis

Taking canagliflozin, at a dose of either 100 mg or 300 mg per day, was associated with risk of diabetic ketoacidosis (DKA), according to a trial that randomized patients with type 1 diabetes to either of the doses or a placebo. At baseline, all of the patients were receiving multiple daily insulin injections or continuous subcutaneous insulin infusion and had an HbA1c between 7% and 9%. None of the patients in the placebo group had DKA during the 18-week trial, compared to 4.3% of those on the 100-mg dose and 6% of those on the 300-mg dose.

The study was published in the April Diabetes Care. The following commentary by Jad G. Sfeir, MD, and Victor M. Montori, MD, MSc, FACP, was published in the ACP Journal Club section of the July 19 Annals of Internal Medicine.

To date, pramlintide and insulin are the only medications approved to treat patients with type 1 diabetes. Since their introduction, sodium–glucose cotransporter 2 (SGLT2) inhibitors have drawn substantial interest as an add-on to insulin in type 1 diabetes. There have, however, been unexpected reports of increased incidence of ketoacidosis in patients with type 2 diabetes using these agents. Peters and colleagues used data from the pivotal trial of canagliflozin in type 1 diabetes to evaluate the association between SGLT2 inhibitors and DKA. The results are limited by lack of precision and inadequate information about insulin dosing around each event.

Aside from the reduction in daily insulin dose for patients receiving canagliflozin, 2 factors may contribute to the observed increased risk for ketoacidosis. First, despite a reduction in fasting glucose, SGLT2 inhibitors increase endogenous glucose production and glucagon levels in patients with type 2 diabetes. The amplitude of this paradoxical physiologic response has yet to be determined in type 1 diabetes. However, it does have a direct effect on hepatic ketone production, thus contributing to the development of ketoacidosis. Second, patients and clinicians may not recognize DKA without severe hyperglycemia, and this could lead to delays in diagnosis and treatment. This is especially true if patients do not have good guidance about when to look for ketonuria or when to seek medical care in the absence of hyperglycemia. In all, for most patients with type 1 diabetes, the modest reduction in HbA1c level conferred by SGLT2 inhibitors may not outweigh the cost and potential risk for DKA requiring hospitalization.