Is SGLT2 Inhibitor-Associated Ketoacidosis Treated With Enough Insulin?

Patients with SGLT2 inhibitor-associated diabetic ketoacidosis (DKA) received significantly lower doses of insulin in the first 24 hours of treatment compared with patients with type 1 diabetes DKA, according to a retrospective cohort study.

Median intravenous or subcutaneous insulin dose during the initial 24 hours of treatment was 44 units for patients with SGLT2 inhibitor-associated DKA versus 87 units for those with type 1 diabetes DKA (P=0.01), reported Mahesh M. Umapathysivam, DPhil, of Royal Adelaide Hospital in Australia, and colleagues.

This was the case even though patients with SGLT2 inhibitor-associated DKA had a significantly higher median body weight compared with patients with type 1 diabetes DKA (81.8 kg [180.3 lb] vs 67.7 kg [149.3 lb]), the researchers detailed in a JAMA Network Open research letter.

While the SGLT2 inhibitor group had milder episodes compared with the type 1 diabetes patients based on median ketone peak (5.3 vs 6.5 mmol/L), they also had significantly delayed resolution, defined as normalization of both ketone and bicarbonate levels (median time 36 vs 18 hours, P=0.002).

Umapathysivam told MedPage Today that he has observed the pattern of a delayed resolution of SGLT2 inhibitor-associated DKA versus type 1 diabetes DKA in clinical practice, and therefore wasn't too surprised by that finding. However, he noted that he was surprised to see that SGLT2 inhibitor patients were receiving so much less insulin as part of treatment.

"The study was conducted across two tertiary centers with a research interest in SGLT2 inhibitor-associated DKA, with input from the endocrine team in all cases," Umapathysivam explained. "Despite this, the tendency toward normoglycemia in SGLT2 inhibitor patients was associated with substantially reduced insulin administration and prolonged time to resolution."

"We are seeing presentations with SGLT2 inhibitor-associated DKA at increasing frequency," he added. "Our retrospective study provides a time frame for expected resolution and raises concerns that at least across the two tertiary centers included -- but presumable more widely -- that DKA protocols developed for use in type 1 diabetes DKA do not perform as well in SGLT2 inhibitor DKA and that dedicated protocols that use more aggressive dextrose administration may be required."

The researchers advised that "it would be reasonable, based on the evidence presented and the safety profile of intravenous dextrose, to increase dextrose infusion rates and concentration to allow increased insulin administration and suppression of ketosis."

DKA cases were identified from endocrine consultation requests in electronic medical records for two South Australian hospitals from January 2019 to December 2021. All 37 patients in the SGLT2 inhibitor group had type 2 diabetes, and included patients with SGLT2 inhibitor-associated DKA (ketones 3 mmol/L and bicarbonate <15 mEq/L) and clinically significant SGLT2 inhibitor-associated ketosis (ketones >3 mmol/L and bicarbonate 15-20 mEq/L).

The SGLT2 inhibitor and type 1 diabetes groups were matched for age and sex. All patients in the SGLT2 inhibitor group had this therapy discontinued.

Among the patients with SGLT2 inhibitor-associated DKA and ketosis (mean age 62, 54.1% men) included in the analysis, mean body mass index was 27.5 compared with 24.5 for the 20 patients with type 1 diabetes DKA (mean age 62, 50% men). Median admission plasma glucose level was 12.5 mmol/L and 36.8 mmol/L, respectively, baseline HbA1c level was 9% and 12.6%, and bicarbonate nadir was 11 mEq/L and 6.5 mEq/L.

During the first 24 hours of treatment for SGLT2 inhibitor-associated DKA, change in ketone concentration was significantly linked with baseline insulin therapy (β=6.67, 95% CI 2.71-10.63, P=0.002) and lower bicarbonate nadir (β = -0.56, 95% CI -0.89 to -0.23, P=0.02), and trended towards higher admission plasma glucose levels (β=0.30, 95% CI 0.04-0.55, P=0.24).

Hypoglycemia during the first 24 hours of treatment only occurred in 2% of the SGLT2 inhibitor group and 1% of the type 1 diabetes group.

Umapathysivam said it's important to keep in mind that this was only a small retrospective study. "The next steps for the research team are to conduct a prospective interventional trial of additional dextrose," he said.

Comment