Incorporation of empagliflozin in a closed-loop, automated insulin pump for patients with type 1 diabetes may allow for the elimination of carbohydrate counting. These findings, from an open-label, crossover, noninferiority trial ( Identifier: NCT03510000), were published in Diabetes, Obesity, and Metabolism.

Patients (N=30) with type 1 diabetes who had been using an insulin pump for more than 3 months were recruited at Mount Sinai Hospital in Toronto, Ontario, Canada between 2018 and 2019. All patients participated in the study for a total of 2 months during which time they had a median of 8 days between treatment combination (on or off empagliflozin with 3 different prandial insulin strategies). The insulin strategies included carbohydrate counting, meal announcements, or no counting/announcement. The sequence of drug/insulin strategy combinations were randomly assigned.

Patients were 57% women, aged 40±15 years, with an average BMI of 27.7±6.4, glycated hemoglobin A1c of 7.6%±0.7%, and total daily insulin consumption of 0.63±0.21 U/kg. Participants used Medtronic (70%), Omnipod (17%), or Animas (13%) pumps.

Continue Reading

At the end of the run-in periods with and without empagliflozin, the average basal rates were 0.97±0.56 and 1.00±0.54 U per hour, carbohydrate-to-insulin ratios were 10.1±3.9 and 9.7±3.5 g/U, and starting glucose levels were 6.4±2.0 and 8.6±3.1 mmol/L, respectively.

The treatment combination of empagliflozin with meal announcement had significantly lower glucose levels than carbohydrate counting alone (mean difference [MD], -0.2; 95% CI, -0.9 to 0.6 mmol/L; P =.007). Empagliflozin without a meal announcement did not significantly lower glucose than carbohydrate counting alone (MD, 1.4; 95% CI, 0.6-2.3 mmol/L; P =.94).

Compared with carbohydrate counting alone, empagliflozin with meal announcement lowered insulin bolus (MD, -7; 95% CI, -9.8 to -4.2 units; P <.001) and total insulin (MD, -6.8; 95% CI, -9.4 to -4.1 units; P <.001) and increased morning ketone levels (MD, 0.12; 95% CI, 0.06-0.19 mmol/L; P <.001); empagliflozin with carbohydrate counting lowered mean glucose (MD, -1.0; 95% CI, -1.9 to -0.1 mmol/L; P =.005) and total insulin (MD, -3.2; 95% CI, -5.9 to -0.4 mmol/L; P =.023) and increased morning ketone level (MD, 0.11; 95% CI, 0.04-0.18 mmol/L; P =.002).

There were 28 hypoglycemic events; 18 in the group that received empagliflozin and the carbohydrate counting and 10 in the group that received empagliflozin and the simple meal strategy vs 18 in the group that received the carbohydrate counting strategy alone. Despite higher morning ketone during empagliflozin periods, no diabetic ketoacidosis events occurred. A single mycotic infection was recorded. This study was limited by its small sample size and short treatment durations (9-14 hours).

These data indicated empagliflozin incorporated with an automated insulin pump may allow for patients with type 1 diabetes to forgo carbohydrate counting, however, meal announcement would still be required.

Disclosure: Several study authors declared affiliations with the pharmaceutical industry. Please see the original reference for a full list of authors’ disclosures.


Jaidar A, Yale J-F, Lovblom L E, et al. Reducing the need for carbohydrate counting in type 1 diabetes using closed-loop automated insulin delivery (artificial pancreas) and empagliflozin: a randomised controlled non-inferiority crossover pilot trial. Diabetes Obes Metab. Published online February, 2 2021. doi:10.1111/dom.14335