How to Eliminate Carb Counting for Type 1 Pump Users 

Could adding a SGLT-2 like empagliflozin to a closed-loop automated insulin delivery system reduce the need for carbohydrate counting? 

Counting carbs can help people with diabetes manage their blood sugar. The ADA (American Diabetes Association) guidelines recommend 45-65% of an individual’s daily calories should come from carbs. Previous trials have been conducted with SGLT inhibitors (SGLTis), showing a reduction in HbA1c, blood pressure, and weight without increasing hypoglycemia. An open-label, crossover trial assessed whether adding empagliflozin, a highly selective SGLTi, to a closed-loop automated insulin delivery could reduce or eliminate the need for carbohydrate counting in type 1 diabetes (T1D). 

Between July 2018 and November 2019, 30 adults with T1D were recruited from the Mount Sinai Hospital in Toronto, Ontario, Canada. Five study days were excluded from analysis because of technical errors. The study was conducted for a total of 2 months in which participants were randomly assigned three sequences of prandial insulin strategy days: carbohydrate counting, simple meal announcement (no carbohydrate counting), and no meal announcement. Participants were age 40 ± 15 years, with a mean of 27 ± 15 years diabetes duration, HbA1c of 7.6% ± 0.7% (59 ± 8 mmol/mol), average BMI of 27.7±6.4, and 57% of the patients were women. All participants provided written informed consent. The exclusion criteria included: 

• Renal insufficiency
• Use of loop diuretics
• Use of non-insulin antihyperglycemic drugs (e.g., glucagon-like peptide-1 [GLP-1] analogs)
• Pregnancy
• Diabetic ketoacidosis in the last three months
• History of lower limb amputation
• A recent history of genital or urinary infection
• Recent history of leg or foot infection

Throughout the study, the participants used their insulin pumps; 70% used a Medtronic pump, 17% used Omnipod, and 13% used an Animas pump. The participants installed Dexcom G5, a glucose sensor, at least 24 to 48 h before each automated insulin delivery intervention. Before breakfast, a research staff member arrived at the participant’s location and applied an automated insulin delivery intervention for 9 hours (Toronto site) and 14 hours (Montreal site). The difference in hours between sites was accounted for in the statistical sensitivity analysis. Participants were instructed to maintain their regular life routine during the 9 to 14 hours, and this included attending work/school, driving, and participating in any activity without restriction. Blood ketone levels were measured on the morning of each intervention using a capillary ketone meter. The amount and timing of carbohydrate intake were not standardized or restricted. The study staff held the sensor receiver throughout the interventions. Readings were entered manually every 10 minutes by the staff, from the receiver into a tablet computer, which ran the dosing algorithm that calculated insulin delivery. There was only one study staff member present at a time with the study participants during interventions. 

The comparison of glucose and insulin safety outcomes between carb counting alone, empagliflozin with simple meal announcement lowered mean glucose (95% CI;−0.9 to 0.6 mmol/L; P= .71), total insulin (95% CI, -9.4 to -4.1 units; P <.001) and increased ketone levels at the start of the day (95% CI, 0.06-0.19 mmol/L; P <.001); empagliflozin with carbohydrate counting lowered mean glucose (95% CI, -1.9 to -0.1 mmol/L; P =.005), total insulin (95% CI, -5.9 to -0.4 mmol/L; P =.023) and increased ketone levels at the start of the day ( 95% CI, 0.04 to 0.18 mmol/L; P =.002). During this study, there was one mycotic infection reported. Eighteen hypoglycemic events occurred in the group that received empagliflozin and the carbohydrate counting, and 10 in the group that received empagliflozin and the simple meal strategy, versus 18 in the group that received the carbohydrate counting strategy alone. No diabetic ketoacidosis events occurred despite the increased ketone level at the start of the day. This study was limited by the small sample size and short treatment durations. 

Although the findings indicated that empagliflozin incorporated with an automated insulin delivery system might allow patients with type 1 diabetes to eliminate carbohydrate counting, there is still a need for a meal announcement system.  

Practice Pearls: 

• Adding empagliflozin with no meal announcement to an automated insulin delivery is inferior to the automated insulin delivery system with carbohydrate counting and no empagliflozin.
• Empagliflozin eliminates the need for carbohydrate counting if a strategy of simple meal announcement is implemented.
• While this study mentions the benefits of adding empagliflozin to automated insulin, further studies are needed with a longer duration of treatment and larger sample size.

Haidar, Ahmad et al. Reducing the need for carbohydrate counting in type 1 diabetes using closed-loop automated insulin delivery (artificial pancreas) and empagliflozin: A randomized, controlled, non-inferiority, crossover pilot trial. Diabetes, obesity & metabolism; February 2, 2021. 

Brenda Oppong, PharmD Candidate, LECOM School of Pharmacy 

See more about closed loop insulin systems.