Trial Outcomes & Findings for Effect of Gain on Closed-Loop Insulin (NCT NCT02065895)

Results Overview

Glucose Area Under the Curve (AUC) Breakfast defines the total exposure to glucose during breakfast. Breakfast is typically considered the most difficult meal to control; low AUC is desirable.This outcome measure was analyzed for each of the three calibration error values (high error, no error and low error).

Recruitment status

COMPLETED

Study phase

NA

Target enrollment

8 participants

Primary outcome timeframe

On day #1, day #2 and day #3 (each day could be 24 hours to 7 days apart from prior one, and completed within 6 week period) 8:00 AM to 2:00 PM on day following admission, with samples obtained every 10-15 minutes, for each sequence of calibration errors

Results posted on

2018-05-17

Participant Flow

Written informed consent was obtained from all 8 participants at the Joslin Diabetes Center. All study related activities were conducted at the Center for Clinical Investigations at the Beth Israel deaconess Medical Center.

Participant milestones

Participant milestones
Measure	HIGH Error First, LOW Error Second, Then NO Error Third In this arm, subjects were randomized to undergo in-clinic closed-loop control of nighttime and breakfast glucose on three occasions: first with the glucose value used for control calculated to be 33% higher than the true glucose value (HIGH error), then second with the value calculated to be 20% lower than the true value (LOW error), then third with the value equal to the true value (NO error). These conditions reflect real-life conditions that would be expected if closed-loop artificial pancreas control is effected with different sensor calibration error. No meal announcement was provided (simulates condition where patient forgets to announce meal), and all subjects were controlled with the same closed-loop gain (simulates conditions where an individual's insulin sensitivity changes). Intervention period: 17 hours; minimum washout period was 7 hours, maximum was 62 days (varied among subjects, allowed up to 3 months to complete all 3 interventions).	HIGH Error First, NO Error Second, Then LOW Error Third In this arm, subjects were were randomized to undergo in-clinic closed-loop control of nighttime and breakfast glucose on three occasions: first with the glucose value used for control calculated to be 33% higher than the true glucose value (HIGH error), then second with the value calculated to equal the true value (NO error), and then third with the value calculated to be 20% lower than the true value (LOW error). These conditions reflect real-life conditions that would be expected if closed-loop artificial pancreas control is effected with different sensor calibration error. No meal announcement was provided (simulates condition where patient forgets to announce meal), and all subjects were controlled with the same closed-loop gain (simulates conditions where an individual's insulin sensitivity changes). Intervention period: 17 hours; minimum washout period was 7 hours, maximum was 62 days (varied among subjects, allowed up to 3 months to complete all 3 interventions).	NO Error First, LOW Error Second, HIGH Error Third In this arm, subjects were were randomized to undergo in-clinic closed-loop control of nighttime and breakfast glucose on three occasions: first with the glucose value used for control equal the true value (NO error), then second with the value calculated to be 20% lower than the true value (LOW error), and then third with the value calculated to be 33% higher than the true glucose value (HIGH error). These conditions reflect real-life conditions that would be expected if closed-loop artificial pancreas control is effected with different sensor calibration error. No meal announcement was provided (simulates condition where patient forgets to announce meal), and all subjects were controlled with the same closed-loop gain (simulates conditions where an individual's insulin sensitivity changes). Intervention period: 17 hours; minimum washout period was 7 hours, maximum was 62 days (varied among subjects, allowed up to 3 months to complete all 3 interventions).	NO Error First, HIGH Error Second, LOW Error Third In this arm, subjects were were randomized to undergo in-clinic closed-loop control of nighttime and breakfast glucose on three occasions: first with the glucose value used for control calculated equal the true value (NO error), and then second with the value calculated as 33% higher than the true glucose value (HIGH error), and then third with the value calculated to to be 20% lower than the true glucose value (LOW error). These conditions reflect real-life conditions that would be expected if closed-loop artificial pancreas control is effected with different sensor calibration error. No meal announcement was provided (simulates condition where patient forgets to announce meal), and all subjects were controlled with the same closed-loop gain (simulates conditions where an individual's insulin sensitivity changes).Intervention period: 17 hours; minimum washout period was 7 hours, maximum was 62 days (varied among subjects, allowed up to 3 months to complete all 3 interventions).	LOW Error First, NO Error Second, HIGH Error Third In this arm, subjects were were randomized to undergo in-clinic closed-loop control of nighttime and breakfast glucose on three occasions: first with the glucose value used for control calculated to be 20% lower than the true glucose value (LOW error), then second with the value calculated to equal the true value (NO error), and then third with the value calculated to be 33% higher than the true value. These conditions reflect real-life conditions that would be expected if closed-loop artificial pancreas control is effected with different sensor calibration error. No meal announcement was provided (simulates condition where patient forgets to announce meal), and all subjects were controlled with the same closed-loop gain (simulates conditions where an individual's insulin sensitivity changes). Intervention period: 17 hours; minimum washout period was 7 hours,maximum was 62 days (varied among subjects, allowed up to 3 months to complete all 3 interventions).	LOW Error First, HIGH Error Second, NO Error Third In this arm, subjects were were randomized to undergo in-clinic closed-loop control of nighttime and breakfast glucose on three occasions: first with the glucose value used for control calculated to be 20% lower than the true glucose value (LOW error), then second with the value calculated to be 33% higher than the true value (HIGH error), and then third with the value calculated to be equal the true value (NO error). These conditions reflect real-life conditions that would be expected if closed-loop artificial pancreas control is effected with different sensor calibration error. No meal announcement was provided (simulates condition where patient forgets to announce meal), and all subjects were controlled with the same closed-loop gain (simulates conditions where an individual's insulin sensitivity changes). Intervention period: 17 hours; minimum washout period was 7 hours, maximum was 62 days (varied among subjects, allowed up to 3 months to complete all 3 interventions).
Overall Study STARTED	1	3	1	1	1	1
Overall Study COMPLETED	0	2	1	0	1	1
Overall Study NOT COMPLETED	1	1	0	1	0	0

Reasons for withdrawal

Reasons for withdrawal
Measure	HIGH Error First, LOW Error Second, Then NO Error Third In this arm, subjects were randomized to undergo in-clinic closed-loop control of nighttime and breakfast glucose on three occasions: first with the glucose value used for control calculated to be 33% higher than the true glucose value (HIGH error), then second with the value calculated to be 20% lower than the true value (LOW error), then third with the value equal to the true value (NO error). These conditions reflect real-life conditions that would be expected if closed-loop artificial pancreas control is effected with different sensor calibration error. No meal announcement was provided (simulates condition where patient forgets to announce meal), and all subjects were controlled with the same closed-loop gain (simulates conditions where an individual's insulin sensitivity changes). Intervention period: 17 hours; minimum washout period was 7 hours, maximum was 62 days (varied among subjects, allowed up to 3 months to complete all 3 interventions).	HIGH Error First, NO Error Second, Then LOW Error Third In this arm, subjects were were randomized to undergo in-clinic closed-loop control of nighttime and breakfast glucose on three occasions: first with the glucose value used for control calculated to be 33% higher than the true glucose value (HIGH error), then second with the value calculated to equal the true value (NO error), and then third with the value calculated to be 20% lower than the true value (LOW error). These conditions reflect real-life conditions that would be expected if closed-loop artificial pancreas control is effected with different sensor calibration error. No meal announcement was provided (simulates condition where patient forgets to announce meal), and all subjects were controlled with the same closed-loop gain (simulates conditions where an individual's insulin sensitivity changes). Intervention period: 17 hours; minimum washout period was 7 hours, maximum was 62 days (varied among subjects, allowed up to 3 months to complete all 3 interventions).	NO Error First, LOW Error Second, HIGH Error Third In this arm, subjects were were randomized to undergo in-clinic closed-loop control of nighttime and breakfast glucose on three occasions: first with the glucose value used for control equal the true value (NO error), then second with the value calculated to be 20% lower than the true value (LOW error), and then third with the value calculated to be 33% higher than the true glucose value (HIGH error). These conditions reflect real-life conditions that would be expected if closed-loop artificial pancreas control is effected with different sensor calibration error. No meal announcement was provided (simulates condition where patient forgets to announce meal), and all subjects were controlled with the same closed-loop gain (simulates conditions where an individual's insulin sensitivity changes). Intervention period: 17 hours; minimum washout period was 7 hours, maximum was 62 days (varied among subjects, allowed up to 3 months to complete all 3 interventions).	NO Error First, HIGH Error Second, LOW Error Third In this arm, subjects were were randomized to undergo in-clinic closed-loop control of nighttime and breakfast glucose on three occasions: first with the glucose value used for control calculated equal the true value (NO error), and then second with the value calculated as 33% higher than the true glucose value (HIGH error), and then third with the value calculated to to be 20% lower than the true glucose value (LOW error). These conditions reflect real-life conditions that would be expected if closed-loop artificial pancreas control is effected with different sensor calibration error. No meal announcement was provided (simulates condition where patient forgets to announce meal), and all subjects were controlled with the same closed-loop gain (simulates conditions where an individual's insulin sensitivity changes).Intervention period: 17 hours; minimum washout period was 7 hours, maximum was 62 days (varied among subjects, allowed up to 3 months to complete all 3 interventions).	LOW Error First, NO Error Second, HIGH Error Third In this arm, subjects were were randomized to undergo in-clinic closed-loop control of nighttime and breakfast glucose on three occasions: first with the glucose value used for control calculated to be 20% lower than the true glucose value (LOW error), then second with the value calculated to equal the true value (NO error), and then third with the value calculated to be 33% higher than the true value. These conditions reflect real-life conditions that would be expected if closed-loop artificial pancreas control is effected with different sensor calibration error. No meal announcement was provided (simulates condition where patient forgets to announce meal), and all subjects were controlled with the same closed-loop gain (simulates conditions where an individual's insulin sensitivity changes). Intervention period: 17 hours; minimum washout period was 7 hours,maximum was 62 days (varied among subjects, allowed up to 3 months to complete all 3 interventions).	LOW Error First, HIGH Error Second, NO Error Third In this arm, subjects were were randomized to undergo in-clinic closed-loop control of nighttime and breakfast glucose on three occasions: first with the glucose value used for control calculated to be 20% lower than the true glucose value (LOW error), then second with the value calculated to be 33% higher than the true value (HIGH error), and then third with the value calculated to be equal the true value (NO error). These conditions reflect real-life conditions that would be expected if closed-loop artificial pancreas control is effected with different sensor calibration error. No meal announcement was provided (simulates condition where patient forgets to announce meal), and all subjects were controlled with the same closed-loop gain (simulates conditions where an individual's insulin sensitivity changes). Intervention period: 17 hours; minimum washout period was 7 hours, maximum was 62 days (varied among subjects, allowed up to 3 months to complete all 3 interventions).
Overall Study Protocol Violation	1	0	0	1	0	0
Overall Study Scheduling conflicts	0	1	0	0	0	0

Baseline Characteristics

One subject withdrew

Baseline characteristics by cohort

Baseline characteristics by cohort
Measure	Closed-loop Control n=8 Participants Closed-loop control was performed from 9:00 PM to 2:00 the day following admission on three occasions: once with a glucose-value-used-for control calculated to be higher than the true blood glucose (analogous to a sensor glucose signal that is miss-calibrated); once with the value equal to blood glucose (analogous to a sensor signal with no calibration), and once with the value calculated to be lower than blood glucose. Six different arms were utilized, defined by the differing sequences of each of the three values used for control (no error, high error and low error previously defined). On each occasion control was separated into the nighttime period (midnight to 08:00 AM) and breakfast period (8:00 AM to 2:00 PM).
Age, Continuous	57 years n=99 Participants
Sex: Female, Male Female	0 Participants n=99 Participants
Sex: Female, Male Male	8 Participants n=99 Participants
Region of Enrollment United States	8 participants n=99 Participants • One subject withdrew
Median duration of diabetes	47 years n=99 Participants
Daily insulin use	35.4 units/day n=99 Participants
HbA1c	7.7 % n=99 Participants

PRIMARY outcome

Timeframe: On day #1, day #2 and day #3 (each day could be 24 hours to 7 days apart from prior one, and completed within 6 week period) 8:00 AM to 2:00 PM on day following admission, with samples obtained every 10-15 minutes, for each sequence of calibration errors

Glucose Area Under the Curve (AUC) Breakfast defines the total exposure to glucose during breakfast. Breakfast is typically considered the most difficult meal to control; low AUC is desirable.This outcome measure was analyzed for each of the three calibration error values (high error, no error and low error).

Outcome measures

Outcome measures
Measure	HIGH Error n=5 Participants Glucose sensor errors leading to sensor glucose reading higher than that true blood glucose result in closed-loop control systems (artificial pancreas) behaving as if the control gain is increased and the target is lowered. In this arm we study closed-loop control with a sensor signal reading 33% higher than blood glucose.	NO Error n=5 Participants Glucose sensor errors leading to sensor glucose reading higher than that true blood glucose result in closed-loop control systems (artificial pancreas) behaving as if the control gain is increased and the target is lowered. In this arm we study closed-loop control with a sensor signal reading equal to blood glucose.	LOW Error n=5 Participants Glucose sensor errors leading to sensor glucose reading higher than that true blood glucose result in closed-loop control systems (artificial pancreas) behaving as if the control gain is increased and the target is lowered. In this arm we study closed-loop control with a sensor signal reading 20% lower than blood glucose.
Glucose Area Under the Curve (AUC) Breakfast	66.8 mmol/l/min Interval 53.3 to 80.3	48.8 mmol/l/min Interval 41.1 to 56.4	37.4 mmol/l/min Interval 33.4 to 41.5

SECONDARY outcome

Timeframe: On day #1, day #2 and day #3 (each day could be 24 hours to 7 days apart from prior one, and completed within 6 week period) 8:00 AM to 12:00 PM on day following admission, with samples obtained every 10-15 minutes, for each sequence of calibration errors

Population: Analysis was limited to the 5 subjects completed all aspects of the study per protocol (i.e. 5 subjects who completed all three scheduled breakfast meals).

Highest and lowest glucose concentrations obtained during breakfast meal.

Outcome measures

Outcome measures
Measure	HIGH Error n=5 Participants Glucose sensor errors leading to sensor glucose reading higher than that true blood glucose result in closed-loop control systems (artificial pancreas) behaving as if the control gain is increased and the target is lowered. In this arm we study closed-loop control with a sensor signal reading 33% higher than blood glucose.	NO Error n=5 Participants Glucose sensor errors leading to sensor glucose reading higher than that true blood glucose result in closed-loop control systems (artificial pancreas) behaving as if the control gain is increased and the target is lowered. In this arm we study closed-loop control with a sensor signal reading equal to blood glucose.	LOW Error n=5 Participants Glucose sensor errors leading to sensor glucose reading higher than that true blood glucose result in closed-loop control systems (artificial pancreas) behaving as if the control gain is increased and the target is lowered. In this arm we study closed-loop control with a sensor signal reading 20% lower than blood glucose.
Peak and Nadir Postprandial Glucose Concentration Peak glucose concentration	13.3 mmol/l Interval 11.3 to 15.4	11.8 mmol/l Interval 9.8 to 14.0	11.3 mmol/l Interval 10.2 to 12.4
Peak and Nadir Postprandial Glucose Concentration Nadir glucose concentration	4.5 mmol/l Interval 3.8 to 5.2	4.8 mmol/l Interval 3.7 to 6.5	6.6 mmol/l Interval 5.6 to 7.5

OTHER_PRE_SPECIFIED outcome

Timeframe: On day #1, day #2 and day #3 (each day could be 24 hours to 7 days apart from prior one, and completed within 6 week period) 12:00 AM to 6:00 AM on day following admission, with samples obtained every 10-15 minutes, for each sequence of calibration errors

Population: Analysis was limited to the 6 subjects completing all 3 nighttime periods per protocol

Night-time in target range 5.0-8.33, following the 3 hour controller initialization period blood glucose remained at or near target.