Objective: To investigate the safety and efficacy of the addition of a trust index to enhanced Model Predictive Control (eMPC) Artificial Pancreas (AP) that works by adjusting the aggressiveness of the controller’s insulin delivery based on the confidence intervals around predictions of glucose trends.
Methods: After one week of sensor-augmented pump (SAP) use, subjects completed a 48-hour AP admission that included 3 meals/day of at least 30 g CHO per meal, 1 hour of unannounced exercise and two overnight periods. Endpoints included sensor glucose percentage time 70-180 mg/dL, <70 mg/dL, >180 mg/dL, number of hypoglycemic events, and assessment of the trust index vs. standard eMPC glucose predictions.
Results: Baseline characteristics for the 15 subjects (mean±SD) were age 46.1±17.8 years, HbA1c 7.2±1.0%, diabetes duration 26.8±17.6 years and TDD 35.5±16.4 units/day. Glycemic outcomes are reported in the table. Mean percent time 70-180 mg/dL (88.0±7.7%), <70 mg/dL (1.5±1.8%), and number of hypoglycemic events (0.6±0.6%) all showed statistically significant improvement during AP (p<0.001). On average, the trust index enhanced controller responsiveness to predicted hyper- and hypoglycemia by 26% (p<0.005).
Glycemic Outcomes. eMPC Artificial Pancreas with Trust Index vs. Sensor-Augmented Pump.
| Glycemic Outcomes (Mean±SD) | 48-Hour AP Session | SAP Run-In Week | p-value |
| Percentage Time 70-180, mg/dL | 88.0±7.7 | 74.6±9.0 | <0.001 |
| Percentage Time 80-140, mg/dL | 60.1±13.4 | 45.2±10.3 | 0.004 |
| Percentage Time < 70, mg/dL | 1.5±1.8 | 7.8±5.8 | <0.001 |
| Percentage Time > 180, mg/dL | 10.5±7.7 | 17.6±10.4 | 0.081 |
| Mean CGM, mg/dL | 130.1±14.7 | 135.1±18.8 | >0.5 |
| Standard Deviation, mg/dL | 32.9±6.1 | 50.3±10.5 | <0.001 |
| Number of daily hypoglycemic events (CGM <70 mg/dL for ≥ 15 minutes) | 0.6±0.6 | 1.7±0.9 | <0.001 |
| Glycemic Outcomes (Mean±SD) | 48-Hour AP Session | SAP Run-In Week | p-value |
| Percentage Time 70-180, mg/dL | 88.0±7.7 | 74.6±9.0 | <0.001 |
| Percentage Time 80-140, mg/dL | 60.1±13.4 | 45.2±10.3 | 0.004 |
| Percentage Time < 70, mg/dL | 1.5±1.8 | 7.8±5.8 | <0.001 |
| Percentage Time > 180, mg/dL | 10.5±7.7 | 17.6±10.4 | 0.081 |
| Mean CGM, mg/dL | 130.1±14.7 | 135.1±18.8 | >0.5 |
| Standard Deviation, mg/dL | 32.9±6.1 | 50.3±10.5 | <0.001 |
| Number of daily hypoglycemic events (CGM <70 mg/dL for ≥ 15 minutes) | 0.6±0.6 | 1.7±0.9 | <0.001 |
Conclusions: eMPC AP with trust index additions achieved nearly 90% time in the target glucose range. Additional studies will further validate these results.
J.E. Pinsker: Research Support; Self; Insulet Corporation, Dexcom, Inc., Tandem Diabetes Care, Inc.. A.J. Laguna Sanz: None. M. Church: None. J. Lee: Employee; Self; Insulet Corporation. L. Lindsey: None. C.C. Andre: None. F.J. Doyle: Research Support; Self; Insulet Corporation, Roche Diagnostics Corporation, Xeris Pharmaceuticals, Inc., Dexcom, Inc., DreaMed Diabetes, Ltd., LifeScan, Inc.. Other Relationship; Self; ModeAGC. E. Dassau: Consultant; Self; Insulet Corporation. Research Support; Self; Insulet Corporation. Consultant; Self; Animas Corporation. Research Support; Self; Dexcom, Inc.. Speaker's Bureau; Self; Roche Diabetes Care Health and Digital Solutions. Research Support; Self; Roche Diabetes Care Health and Digital Solutions, Xeris Pharmaceuticals, Inc.. Consultant; Self; Eli Lilly and Company. Research Support; Self; Tandem Diabetes Care, Inc.. Other Relationship; Self; ModAGC. Research Support; Self; LifeScan, Inc., DreaMed Diabetes, Ltd..
