An Outpatient Pump Shutoff Pilot Feasibility and Safety Study

Reduction of Nocturnal Hypoglycemia by Using Predictive Algorithms and Pump Suspension: An Outpatient Pilot Feasibility and Safety Study

Status

Completed

Phases

Phase 2

Study type

Interventional

Source

ClinicalTrials.gov

Registry ID

NCT01736930

Acronym

PSO1

Enrollment

Registered

2012-11-29

Start date

2012-01-31

Completion date

2012-06-30

Last updated

2016-08-30

For informational purposes only — not medical advice. Sourced from public registries and may not reflect the latest updates. Terms

Conditions

Type 1 Diabetes

Keywords

Type 1 Diabetes, Hypoglycemia, Continuous Glucose Monitoring, Pump Suspension

Brief summary

The purpose of this study is to evaluate an overnight system that will turn off the insulin pump automatically if the system predicts that a low blood sugar is likely. The study system includes a combination continuous glucose monitor (CGM)/ insulin pump made by Medtronic MiniMed, Inc and a regular laptop computer that runs a computer program that predicts low blood sugar. It works by (1) measuring the glucose levels under the skin with a continuous glucose monitor, (2) using a computer program on a laptop to predict what will happen to the glucose level over the next 35-55 minutes, and (3) turning off the insulin pump when the computer program predicts that low blood sugar will occur. We have tested this system overnight in the hospital and are ready to test the system in the home environment to learn more about how well it will work and to make sure that the blood sugar does not go too high when the pump shuts off. This study has several phases and will take about a month or a little more for a patient to complete. Patients will use the study system for about 5 days at home to show that the patient is able to use it correctly. After that, the patient will be asked to use the study system each night for an additional 3-4 weeks. During this time, the system will be active for two-thirds of the nights and not active for one-third of the nights. When the system is active and predicts that your blood sugar will become low, the insulin pump will shut off for up to 2 hours. The study will include 2 clinical centers in the United States.

Detailed description

Patients who meet all eligibility criteria will use the study system overnight at home for about 5 nights to demonstrate their ability to use the system and submit study data to the Coordinating Center. Patients who successfully demonstrate their ability to use the system as described above will be eligible for the randomized trial phase. This phase consists of use of the full system as an outpatient for approximately 21 nights: * Each night the blood glucose level will be checked with the study blood glucose (BG) meter and used to perform a calibration of the CGM. The calibration must occur no more than 90 minutes prior to activation of the system. NOTE: Patients will be instructed to calibrate the CGM per manufacturer guidelines. * Then the system will be activated, linking the CGM and insulin pump to the computer at the bedside. * A randomization schedule on the laptop will be used to determine whether the 'pump shut off' application will be active that night or not. * Patients will be blinded as to whether the pump shut off is active when a session is initiated each night. * There will not be an alarm if the pump shuts off. The CGM alarm will be set to 60 mg/dL. When a CGM alarm occurs, the patient will be asked to measure the blood glucose with a BG meter, if he/she is aware of the alarm. * The time period for outcome assessment each night will be from the time the system is activated until it is turned off in the morning. * Pump shut off, when it occurs, will be for up to 2 hours. Multiple instances of pump suspension can occur if there are recurrent predictions of hypoglycemia during the night. * Patients will be asked to check blood glucose with the study BG meter, blood ketones with the study ketone meter, and urine ketones with a ketone strip each morning prior to breakfast and enter the results on the study laptop. The patient will be instructed to contact the study physician if the blood glucose or ketone readings are out of an expected range. Patients will be contacted if these morning safety values are not reported as required or are out of range. * Patients will be asked to record all overnight carbohydrate intake on the study laptop. * Patients will be asked to perform periodic data uploads using the study laptop. Monitoring processes will ensure that the patient is contacted if these uploads do not occur as required, or if review of an upload reveals any extreme, prolonged episodes of low or high blood glucose readings, or elevated morning blood glucose, blood ketone, or urine ketone values. Upon completion of the study, patients as well as study clinicians will be asked to complete a questionnaire regarding use of the study system. There will a follow-up visit after completion of 21 successful nights of study system use. A successful night of study system use is defined as use of the system for at least four hours. Phone contacts with the patients will be made once a week.

Interventions

DEVICEPump suspension

The study laptop will communicate to the pump causing suspension based on output from the algorithm which predicts hypoglycemia based on the continuous glucose sensor trend.

Study design

Allocation

RANDOMIZED

Intervention model

PARALLEL

Primary purpose

PREVENTION

Masking

NONE

Eligibility

Sex/Gender

ALL

Age

18 Years to No maximum

Healthy volunteers

Inclusion criteria

* Clinical diagnosis of type 1 diabetes and using daily insulin therapy for at least one year and an insulin infusion pump for at least 6 months * Age \>=18.0 years * HbA1c \<=8.0% * Availability of internet access for periodic upload of study data * Experience using the Medtronic pump and CGM and investigator is confident that subject will use CGM on a near-daily basis during the study * Nocturnal hypoglycemia meeting the following criterion based on CGM download: during the most recent 15 nights with CGM glucose data (must be within the past 42 days), one or more nights with a sensor glucose value \<=70 mg/dL.

Exclusion criteria

* Diabetic ketoacidosis in the past 3 months * Hypoglycemic seizure or loss of consciousness in the past 6 months * History of seizure disorder (except for hypoglycemic seizure) * Coronary artery disease or heart failure * Cystic fibrosis * Current use of oral/inhaled glucocorticoids, beta-blockers or other medications, which in the judgment of the investigator would be a contraindication to participation in the study. * History of ongoing renal disease (other than microalbuminuria). Creatinine level to have been obtained within the last year if subject has diabetes of \>10 years duration or is over 50 years of age. If creatinine is \> 1.5 mg/dL, the subject is excluded. * History of liver disease * Medical or psychiatric condition that in the judgment of the investigator might interfere with the completion of the protocol such as: inpatient psychiatric treatment in the past 6 months, uncontrolled adrenal disorder, and/or abuse of alcohol * Pregnancy: A negative urine pregnancy test will be required for all premenopausal women who are not surgically sterile. Subjects who become pregnant will be discontinued from the study.

Design outcomes

Primary

Measure	Time frame	Description
Mornings With Urine Ketones Characterized as Moderate or Large - Algorithm 3	21 days	The primary safety outcome will be evaluated by comparing the intervention and control nights and nights with vs. without actual shut-off of the pump for several measures of hyperglycemia such as number of mornings with urine ketones characterized as moderate or large based on measurement results of a urine dipstick test taken using Ketostix. Moderate is considered approximately 30 - 40 mg/dL and Large \>80 mg/dL.
Percent Morning Blood Glucose >250 mg/dL - Algorithm 3	21 days	The primary safety outcome will be evaluated by comparing the intervention and control nights and nights with vs. without actual shut-off of the pump for several measures of hyperglycemia such as overall percentage of mornings glucose measured with home glucose meter \>250 mg/dL.
Mornings With Blood Ketones >0.6 mmol/L - Algorithm 1	21 days	The primary safety outcome will be evaluated by comparing the intervention and control nights and nights with vs. without actual shut-off of the pump for several measures of hyperglycemia such as number of mornings with blood ketones \>0.6 mmol/L.
Mornings With Blood Ketones >0.6 mmol/L - Algorithm 2	21 days	The primary safety outcome will be evaluated by comparing the intervention and control nights and nights with vs. without actual shut-off of the pump for several measures of hyperglycemia such as number of mornings with blood ketones \>0.6 mmol/L.
Mornings With Blood Ketones >0.6 mmol/L - Algorithm 3	21 days	The primary safety outcome will be evaluated by comparing the intervention and control nights and nights with vs. without actual shut-off of the pump for several measures of hyperglycemia such as number of mornings with blood ketones \>0.6 mmol/L.
Mornings With Urine Ketones Characterized as Moderate or Large - Algorithm 1	21 days	The primary safety outcome will be evaluated by comparing the intervention and control nights and nights with vs. without actual shut-off of the pump for several measures of hyperglycemia such as number of mornings with urine ketones characterized as moderate or large based on measurement results of a urine dipstick test taken using Ketostix. Moderate is considered approximately 30 - 40 mg/dL and Large \>80 mg/dL.
Mornings With Urine Ketones Characterized as Moderate or Large - Algorithm 2	21 days	The primary safety outcome will be evaluated by comparing the intervention and control nights and nights with vs. without actual shut-off of the pump for several measures of hyperglycemia such as number of mornings with urine ketones characterized as moderate or large based on measurement results of a urine dipstick test taken using Ketostix. Moderate is considered approximately 30 - 40 mg/dL and Large \>80 mg/dL.
Mean Morning Blood Glucose (mg/dL)- Algorithm 3	21 study nights	The primary safety outcome will be evaluated by comparing the intervention and control nights and nights with vs. without actual shut-off of the pump for several measures of hyperglycemia such as mean morning blood glucose (mg/dL). The hypoglycemia prediction horizon was reduced further in algorithm 3 to 30 minutes. A total of 114 study nights (37 Control nights and 77 Intervention nights) using algorithm 3.
Mean Morning Blood Glucose (mg/dL)- Algorithm 1	21 study nights	The primary safety outcome will be evaluated by comparing the intervention and control nights and nights with vs. without actual shut-off of the pump for several measures of hyperglycemia such as mean morning blood glucose (mg/dL). An objective was to evaluate and refine the control algorithm. The data were reviewed periodically during the study with the pre-stated goal of determining whether any changes should be made in the control algorithm. Algorithm 1 was used for the first 105 nights of the study (38 Control nights and 67 Intervention nights). The horizon prediction time of algorithm 1 was set at 70 minutes.
Mean Morning Blood Glucose (mg/dL)- Algorithm 2	21 study nights	The primary safety outcome will be evaluated by comparing the intervention and control nights and nights with vs. without actual shut-off of the pump for several measures of hyperglycemia such as mean morning blood glucose (mg/dL). Algorithm 1 was modified to reduce the hypoglycemia prediction horizon from 70 minutes to 50 minutes, to suspend the pump only when the continuous glucose monitor sensor glucose value was ≤ 230 mg/dl, not suspend if there was a drop of \>40 mg/dl in consecutive sensor glucose readings, and to resume insulin delivery at the first rise in sensor glucose following a suspension. There was 156 nights of study data collected (48 Control nights and 108 Intervention nights) using algorithm 2.
Percent Morning Blood Glucose >250 mg/dL - Algorithm 1	21 days	The primary safety outcome will be evaluated by comparing the intervention and control nights and nights with vs. without actual shut-off of the pump for several measures of hyperglycemia such as the overall percentage of mornings glucose measured with home glucose meter \>250 mg/dL.
Percent Morning Blood Glucose >250 mg/dL - Algorithm 2	21 days	The primary safety outcome will be evaluated by comparing the intervention and control nights and nights with vs. without actual shut-off of the pump for several measures of hyperglycemia such as overall percentage of mornings glucose measured with home glucose meter \>250 mg/dL.

Secondary

Measure	Time frame
Mean Sensor Glucose Overnight - Algorithm 2	Overnight from system activation to deactivation in the morning upon awakening for 21 nights of system use
Mean Sensor Glucose Overnight - Algorithm 3	Overnight from system activation to deactivation in the morning upon awakening for 21 nights of system use
Percentage of Sensor Glucose Values 71 to 180 mg/dL - Algorithm 1	Overnight from system activation to deactivation in the morning upon awakening for 21 nights of system use
Percentage of Sensor Glucose Values 71 to 180 mg/dL - Algorithm 2	Overnight from system activation to deactivation in the morning upon awakening for 21 nights of system use
Percentage of Sensor Glucose Values 71 to 180 mg/dL - Algorithm 3	Overnight from system activation to deactivation in the morning upon awakening for 21 nights of system use
Mean Sensor Glucose Overnight - Algorithm 1	Overnight from system activation to deactivation in the morning upon awakening for 21 nights of system use

Countries

United States

Participant flow

Recruitment details

The study was conducted at Stanford University and the Barbara Davis Center. A total of 20 subjects were enrolled between January 1, 2012 and June 30, 2012. Nineteen participants 18-56 years old with type 1 diabetes (HbA1c 6.0%-7.7%) completed the trial.

Pre-assignment details

After the run-in phase, a 21-night randomized trial was conducted in which each night was randomly assigned 2:1 to have either the predictive low glucose suspend system active (intervention) or inactive (control). Participants were current users of the MiniMed Paradigm® REAL-Time Revel™ System and Sof-sensor® glucose sensor at time of enrollment.

Participants by arm

Arm	Count
Randomized Nights- Treatment or Control Each study night will be randomized to have either Predictive Low Glucose Suspend or to be inactive (control). On nights randomized to the intervention treatment, the study laptop will be running actively during the night and suspending the patient's pump if the algorithm predicts hypoglycemia based on the patient's continuous glucose sensor trend. (Pump suspension : The study laptop will communicate to the pump causing suspension based on output from the algorithm which predicts hypoglycemia based on the continuous glucose sensor trend.) On control nights, the algorithm will run passively and not recommend suspensions or resumption to the patient's pump.	20
Total	20

Arm

Count

Randomized Nights- Treatment or Control

Each study night will be randomized to have either Predictive Low Glucose Suspend or to be inactive (control). On nights randomized to the intervention treatment, the study laptop will be running actively during the night and suspending the patient's pump if the algorithm predicts hypoglycemia based on the patient's continuous glucose sensor trend. (Pump suspension : The study laptop will communicate to the pump causing suspension based on output from the algorithm which predicts hypoglycemia based on the continuous glucose sensor trend.) On control nights, the algorithm will run passively and not recommend suspensions or resumption to the patient's pump.

Total

Baseline characteristics

Characteristic	Randomized Nights- Treatment or Control
Age, Categorical <=18 years	0 Participants
Age, Categorical >=65 years	0 Participants
Age, Categorical Between 18 and 65 years	20 Participants
Race (NIH/OMB) American Indian or Alaska Native	0 Participants
Race (NIH/OMB) Asian	0 Participants
Race (NIH/OMB) Black or African American	0 Participants
Race (NIH/OMB) More than one race	0 Participants
Race (NIH/OMB) Native Hawaiian or Other Pacific Islander	0 Participants
Race (NIH/OMB) Unknown or Not Reported	1 Participants
Race (NIH/OMB) White	19 Participants
Sex: Female, Male Female	11 Participants
Sex: Female, Male Male	9 Participants

Adverse events

Event type	EG000 affected / at risk	EG001 affected / at risk
deaths Total, all-cause mortality	— / —	— / —
other Total, other adverse events	0 / 20	0 / 20
serious Total, serious adverse events	0 / 20	0 / 20

Outcome results

Primary

Mean Morning Blood Glucose (mg/dL)- Algorithm 1

The primary safety outcome will be evaluated by comparing the intervention and control nights and nights with vs. without actual shut-off of the pump for several measures of hyperglycemia such as mean morning blood glucose (mg/dL). An objective was to evaluate and refine the control algorithm. The data were reviewed periodically during the study with the pre-stated goal of determining whether any changes should be made in the control algorithm. Algorithm 1 was used for the first 105 nights of the study (38 Control nights and 67 Intervention nights). The horizon prediction time of algorithm 1 was set at 70 minutes.

Time frame: 21 study nights

Arm	Measure	Value (MEAN)	Dispersion
Predictive Pump Suspension Algorithm	Mean Morning Blood Glucose (mg/dL)- Algorithm 1	158 mg/dl	Standard Deviation 52
Standard of Care	Mean Morning Blood Glucose (mg/dL)- Algorithm 1	125 mg/dl	Standard Deviation 53

Primary

Mean Morning Blood Glucose (mg/dL)- Algorithm 2

The primary safety outcome will be evaluated by comparing the intervention and control nights and nights with vs. without actual shut-off of the pump for several measures of hyperglycemia such as mean morning blood glucose (mg/dL). Algorithm 1 was modified to reduce the hypoglycemia prediction horizon from 70 minutes to 50 minutes, to suspend the pump only when the continuous glucose monitor sensor glucose value was ≤ 230 mg/dl, not suspend if there was a drop of \>40 mg/dl in consecutive sensor glucose readings, and to resume insulin delivery at the first rise in sensor glucose following a suspension. There was 156 nights of study data collected (48 Control nights and 108 Intervention nights) using algorithm 2.

Time frame: 21 study nights

Arm	Measure	Value (MEAN)	Dispersion
Predictive Pump Suspension Algorithm	Mean Morning Blood Glucose (mg/dL)- Algorithm 2	151 mg/dl	Standard Deviation 57
Standard of Care	Mean Morning Blood Glucose (mg/dL)- Algorithm 2	138 mg/dl	Standard Deviation 63

Primary

Mean Morning Blood Glucose (mg/dL)- Algorithm 3

The primary safety outcome will be evaluated by comparing the intervention and control nights and nights with vs. without actual shut-off of the pump for several measures of hyperglycemia such as mean morning blood glucose (mg/dL). The hypoglycemia prediction horizon was reduced further in algorithm 3 to 30 minutes. A total of 114 study nights (37 Control nights and 77 Intervention nights) using algorithm 3.

Time frame: 21 study nights

Arm	Measure	Value (MEAN)	Dispersion
Predictive Pump Suspension Algorithm	Mean Morning Blood Glucose (mg/dL)- Algorithm 3	144 mg/dl	Standard Deviation 48
Standard of Care	Mean Morning Blood Glucose (mg/dL)- Algorithm 3	133 mg/dl	Standard Deviation 57

Primary

Mornings With Blood Ketones >0.6 mmol/L - Algorithm 1

Time frame: 21 days

Arm	Measure	Value (NUMBER)
Predictive Pump Suspension Algorithm	Mornings With Blood Ketones >0.6 mmol/L - Algorithm 1	0 number of mornings
Standard of Care	Mornings With Blood Ketones >0.6 mmol/L - Algorithm 1	0 number of mornings

Primary

Mornings With Blood Ketones >0.6 mmol/L - Algorithm 2

Time frame: 21 days

Arm	Measure	Value (NUMBER)
Predictive Pump Suspension Algorithm	Mornings With Blood Ketones >0.6 mmol/L - Algorithm 2	3 number of mornings
Standard of Care	Mornings With Blood Ketones >0.6 mmol/L - Algorithm 2	1 number of mornings

Primary

Mornings With Blood Ketones >0.6 mmol/L - Algorithm 3

Time frame: 21 days

Arm	Measure	Value (NUMBER)
Predictive Pump Suspension Algorithm	Mornings With Blood Ketones >0.6 mmol/L - Algorithm 3	1 number of mornings
Standard of Care	Mornings With Blood Ketones >0.6 mmol/L - Algorithm 3	0 number of mornings

Primary

Mornings With Urine Ketones Characterized as Moderate or Large - Algorithm 1

The primary safety outcome will be evaluated by comparing the intervention and control nights and nights with vs. without actual shut-off of the pump for several measures of hyperglycemia such as number of mornings with urine ketones characterized as moderate or large based on measurement results of a urine dipstick test taken using Ketostix. Moderate is considered approximately 30 - 40 mg/dL and Large \>80 mg/dL.

Time frame: 21 days

Arm	Measure	Value (NUMBER)
Predictive Pump Suspension Algorithm	Mornings With Urine Ketones Characterized as Moderate or Large - Algorithm 1	0 number of mornings
Standard of Care	Mornings With Urine Ketones Characterized as Moderate or Large - Algorithm 1	1 number of mornings

Primary

Mornings With Urine Ketones Characterized as Moderate or Large - Algorithm 2

The primary safety outcome will be evaluated by comparing the intervention and control nights and nights with vs. without actual shut-off of the pump for several measures of hyperglycemia such as number of mornings with urine ketones characterized as moderate or large based on measurement results of a urine dipstick test taken using Ketostix. Moderate is considered approximately 30 - 40 mg/dL and Large \>80 mg/dL.

Time frame: 21 days

Arm	Measure	Value (NUMBER)
Predictive Pump Suspension Algorithm	Mornings With Urine Ketones Characterized as Moderate or Large - Algorithm 2	7 number of mornings
Standard of Care	Mornings With Urine Ketones Characterized as Moderate or Large - Algorithm 2	0 number of mornings

Primary

Mornings With Urine Ketones Characterized as Moderate or Large - Algorithm 3

The primary safety outcome will be evaluated by comparing the intervention and control nights and nights with vs. without actual shut-off of the pump for several measures of hyperglycemia such as number of mornings with urine ketones characterized as moderate or large based on measurement results of a urine dipstick test taken using Ketostix. Moderate is considered approximately 30 - 40 mg/dL and Large \>80 mg/dL.

Time frame: 21 days

Arm	Measure	Value (NUMBER)
Predictive Pump Suspension Algorithm	Mornings With Urine Ketones Characterized as Moderate or Large - Algorithm 3	0 number of mornings
Standard of Care	Mornings With Urine Ketones Characterized as Moderate or Large - Algorithm 3	0 number of mornings

Primary

Percent Morning Blood Glucose >250 mg/dL - Algorithm 1

Time frame: 21 days

Arm	Measure	Value (NUMBER)
Predictive Pump Suspension Algorithm	Percent Morning Blood Glucose >250 mg/dL - Algorithm 1	3 percentage of mornings
Standard of Care	Percent Morning Blood Glucose >250 mg/dL - Algorithm 1	3 percentage of mornings

Primary

Percent Morning Blood Glucose >250 mg/dL - Algorithm 2

Time frame: 21 days

Arm	Measure	Value (NUMBER)
Predictive Pump Suspension Algorithm	Percent Morning Blood Glucose >250 mg/dL - Algorithm 2	6 percentage of mornings
Standard of Care	Percent Morning Blood Glucose >250 mg/dL - Algorithm 2	8 percentage of mornings

Primary

Percent Morning Blood Glucose >250 mg/dL - Algorithm 3

Time frame: 21 days

Arm	Measure	Value (NUMBER)
Predictive Pump Suspension Algorithm	Percent Morning Blood Glucose >250 mg/dL - Algorithm 3	1 percentage of mornings
Standard of Care	Percent Morning Blood Glucose >250 mg/dL - Algorithm 3	3 percentage of mornings

Secondary