ESPE Abstracts

Background and aims: In an artificial pancreas, the meal management depends on the meal size and time being reported to the system. Because of the delay of insulin action, it is recommended to declare a meal in advance to allow insulin to limit the glycaemic peak. We aim to assess if the post-prandial glycaemic balance improves as meals are reported in advance.

Clinical dataset: As part of Diabeloop DBLUS clinical trial (NCT04190277), 37 teenagers (age [14-18]) with type 1 diabetes were equipped with an artificial pancreas close-loop (CL) during 12 weeks in real life conditions. Data include CGM and meal reports. We select patients with more than 80% of CL in total, and data of days with more than 70% CL.

Frequency: First, the frequency and time intervals of early meal reporting (EMR) events are computed for every patient. We consider a meal to be reported early if the reporting is done at least 5 min before the meal.

Based on the 822 selected meals, the 26 teenagers tend to rarely report their meal in advance, with a median frequency of 3.2% (min = 0%, max=20.4%). The time interval between EMR and meal times is short, with a median of 10 min (min = 5 min, max= 35min).

Impact: Then, we compare the glycaemic post-prandial balance after an EMR or a standard meal reporting (SMR), during a window of 1 hour after meals. We select the eight patients with at least two EMR and two SMR, leaving a dataset of 47 EMR and 214 SMR. The meal size remains overall similar, with a median null variation over the individual median meal sizes (min = -10g, max = +13g). The median percentage of time in normal range ([70-140] mg/dL) of each patient globally increases with EMR, with a median increase of 8% (min = -10%, max=+63%) compared to SMR. The individual median average glycaemia during post-prandial periods generally decreases, with a median variation of -17 mg/dL (min = -87, max=+33) compared to SMR.

Conclusions: In this study, we show that teenagers who report meals in advance to their artificial pancreas benefit from a decrease of median glycaemic post-prandial peaks and an increase of the time in normal range. The possibility to report meals early could be encouraged by targeted patient education and customized reminders.