Background and Aims

We evaluated the feasibility and safety of the interoperable artificial pancreas system (iAPS) utilizing zone model predictive control in a pediatric outpatient study (NCT04255381).

Methods

Pediatric subjects (n=20, 8F) with type 1 diabetes completed 48 hours of automated insulin delivery (AID) using the iAPS on an unlocked smartphone in three sequential age-based cohorts: 12-<18 years (n=8, 5F), 8-<12 years (n=7, 2F) and 5-<8 years (n=5, 1F). Subjects consumed larger-than-usual meals of their choice and engaged in unannounced physical activities such as escape room, trampoline park and ropes course. Primary outcomes using fingerstick blood glucose (BG) assessed safety while secondary outcomes using continuous glucose monitoring (CGM) compared performance with sensor augmented pump (SAP) preceding AID.

Results

During AID, there were no instances of more than one confirmed BG < 50 mg/dL, two instances of more than two confirmed BG ≥ 300 mg/dL longer than 2 hours, and no adverse events. Overnight during AID (Table), the low blood glucose index decreased by 0.3 (p=0.006) and glucose variability decreased by 14.9 mg/dL (p=0.009). In the youngest cohort, percent time in 70-140 mg/dL range increased by 12.2% (p=0.015). Pre-pubertal subjects (1F) were significantly more sensitive to exercise-induced glycemic changes (Figure).

Table. Comparison of outcomes (mean (standard deviation)) during iAPS AID over SAP for all subjects. ‡ indicates p < 0.05.

Figure. Pre-pubertal (Tanner I) vs pubertal/post-pubertal (Tanner II-V) during physical activities.

Conclusions

The smartphone-based iAPS AID system was feasible and safe in pediatric subjects, even with larger meals and unannounced physical activities.