Critical Illness
Conditions
Keywords
hyperglycemia, hypoglycemia, insulin, critical illness, intensive insulin therapy, continuous glucose monitoring, CGM, interstitial fluid glucose, intensive care unit, Blood glucose
Brief summary
We hypothesize that measurements of interstitial fluid (ISF) glucose by the FreeStyle Navigator (Abbot Diabetes Care), a continuous glucose monitoring system, will correlate with blood glucose (BG) values in surgical and ICU patients with a clinically useful degree of accuracy.
Detailed description
Critically ill patients treated with intensive insulin therapy will be monitored for up to 72 hours with the Freestyle Navigator continuous glucose monitor. The device display will be disabled so that continuous glucose monitoring results will not be available to care givers, although low (\<60 mg/dL) and high (\>250 mg/dL) threshold audio alarms will be enabled. Blood glucose monitoring will be performed according to the usual practice for patients treated with intensive insulin therapy in each intensive care unit, except that additional blood glucose measurements will be performed in response to Freestyle Navigator threshold alarm. All blood glucose management decisions will be made according to intensive insulin therapy protocol in each unit. Freestyle Navigator data will be downloaded from the device after each subject has completed the monitoring period. Blood glucose values obtained in the course of usual care, or in response to Freestyle Navigator threshold alarms, will be compared with time-matched Freestyle Navigator values to assess device accuracy and alarm sensitivity and specificity.
Interventions
DEVICEFreestyle Navigator
Continuous glucose monitoring with the Freestyle Navigator for 72 hours or until discharge from the ICU
Sponsors
Abbott Diabetes Care
Massachusetts General Hospital
Study design
Allocation
NA
Intervention model
SINGLE_GROUP
Primary purpose
DIAGNOSTIC
Masking
NONE
Eligibility
Sex/Gender
ALL
Age
18 Years to No maximum
Healthy volunteers
No
Inclusion criteria
* Patients admitted to the general surgical ICU (SICU), medical ICU (MICU), cardiac surgery intensive care unit (CSICU), or neurological intensive care unit (NeuroICU), are being treated with IIT, and have an arterial line in place * Patients scheduled for open cardiac surgery or craniotomy that have diabetes mellitus or will be given high dose glucocorticoids. All of these patient will have an arterial line placed at the time of surgery as a part of their usual care
Exclusion criteria
* Age less than 18 years * Enrollment in another clinical trial that modifies their glucose management (e.g. trial of modified IIT algorithm or modified glucose monitoring regimen) In order to obtain adequate representation of patients with physiological alterations that might decrease correspondence between blood and ISF glucose, we will recruit 50 subjects in each of five groups: 1. Patients undergoing craniotomy who are either diabetic or will be given high dose corticosteroids 2. Diabetic patients undergoing open cardiac surgery 3. Patients admitted to one of the participating ICUs with an arterial line in place and receiving IIT who are vasopressor dependent, defined by continuous infusion of at least 60 mg/min phenylephrine, 5 mg/min norepinephrine, or 10 mg/kg/min of dopamine, or any combination of two agents for at least 90 continuous minutes in the last 8 hours 4. Patients admitted to one of the participating ICUs with an arterial line in place and receiving IIT with pitting peripheral or dependent edema, defined by the persistence of pitting to at least 5 mm depth for 5 seconds after pressure is applied to a dependent area of the trunk or two limbs. 5. Critically ill patients admitted to one of the participating ICUs with an arterial line in place and receiving IIT who do not fall into groups 1-4.
Design outcomes
Primary
| Measure | Time frame |
|---|---|
| Sensitivity and Specificity of Navigator Threshold Alarms for Hypoglycemia (<60 mg/dl). | 72 hours |
Secondary
| Measure | Time frame | Description |
|---|---|---|
| Sensitivity and Specificity of Navigator Projected Low Blood Sugar Alarm Criteria Calculated for Events Defined by a Low Reference BG Value (< 60 mg/dL) | 72 hours | — |
| Sensitivity and Specificity of Navigator Projected High Blood Sugar Alarm Criteria Calculated for Events Defined by a High Reference BG Value (> 250 mg/dl) | 72 hours | — |
| Accuracy of Navigator CGM as Compared to Reference Blood Glucose Values | 72 hours | Mean Absolute Relative Difference (MARD) of Navigator CGM compared with reference blood glucose values |
| Percentage of Readings in Different Blood Sugar Ranges as Shown by Point of Care Testing: | 72 hours | Percentage of readings in different blood sugar ranges as shown by point of care testing: \< 60 mg/dl 61-120 mg/dl 121-180 mg/dl 181-240 lmg/dl \>240 mg/dl |
| Navigator CGM Accuracy by Category | 72 hours | Mean Absolute Relative Difference (MARD) between the Navigator CGM and reference blood glucose values by category (1, Neurosurgery; 2, Cardiac Surgery; 3, Hypotensive/Vasopressor; 4, Edema; 5, None of the Above) |
| Sensitivity and Specificity of Navigator Threshold Alarms for Hyperglycemia (>240 mg/dl). | 72 hours | — |
Countries
United States
Participant flow
Participants by arm
| Arm | Count |
|---|---|
| Freestyle Navigator Continuous monitoring with the Freestyle Navigator for 72 hours or until discharge from the ICU
Freestyle Navigator: Continuous glucose monitoring with the Freestyle Navigator for 72 hours or until discharge from the ICU | 69 |
| Total | 69 |
Baseline characteristics
| Characteristic | Freestyle Navigator |
|---|---|
| Age, Categorical <=18 years | 0 Participants |
| Age, Categorical >=65 years | 24 Participants |
| Age, Categorical Between 18 and 65 years | 45 Participants |
| Age, Continuous | 59 years STANDARD_DEVIATION 15.2 |
| Region of Enrollment United States | 69 Participants |
| Sex: Female, Male Female | 31 Participants |
| Sex: Female, Male Male | 38 Participants |
Adverse events
| Event type | EG000 affected / at risk |
|---|---|
| deaths Total, all-cause mortality | 1 / 69 |
| other Total, other adverse events | 5 / 69 |
| serious Total, serious adverse events | 0 / 69 |
Outcome results
Primary
Sensitivity and Specificity of Navigator Threshold Alarms for Hypoglycemia (<60 mg/dl).
Time frame: 72 hours
Population: This measure was not calculated as we did not have sufficient data to calculate sensitivity (not enough hypoglycemia) and discordant BG check times to calculate specificity.
Secondary
Accuracy of Navigator CGM as Compared to Reference Blood Glucose Values
Mean Absolute Relative Difference (MARD) of Navigator CGM compared with reference blood glucose values
Time frame: 72 hours
| Arm | Measure | Value (MEAN) |
|---|---|---|
| Freestyle Navigator | Accuracy of Navigator CGM as Compared to Reference Blood Glucose Values | 14.5 percent absolute relative difference |
Secondary
Navigator CGM Accuracy by Category
Mean Absolute Relative Difference (MARD) between the Navigator CGM and reference blood glucose values by category (1, Neurosurgery; 2, Cardiac Surgery; 3, Hypotensive/Vasopressor; 4, Edema; 5, None of the Above)
Time frame: 72 hours
Population: MARD by category (1, Neurosurgery; 2, Cardiac Surgery; 3, Hypotensive/Vasopressor; 4, Edema; 5, None of the Above)
| Arm | Measure | Group | Value (MEAN) | Dispersion |
|---|---|---|---|---|
| Freestyle Navigator | Navigator CGM Accuracy by Category | Neurosurgery | 15.1 percent absolute relative difference | Standard Deviation 12.7 |
| Freestyle Navigator | Navigator CGM Accuracy by Category | Cardiac surgery | 14.3 percent absolute relative difference | Standard Deviation 14.6 |
| Freestyle Navigator | Navigator CGM Accuracy by Category | Hypotensive/vasopressor | 18.8 percent absolute relative difference | Standard Deviation 18.6 |
| Freestyle Navigator | Navigator CGM Accuracy by Category | Edema | 9.7 percent absolute relative difference | Standard Deviation 4.1 |
| Freestyle Navigator | Navigator CGM Accuracy by Category | None | 10.1 percent absolute relative difference | Standard Deviation 10 |
Secondary
Percentage of Readings in Different Blood Sugar Ranges as Shown by Point of Care Testing:
Percentage of readings in different blood sugar ranges as shown by point of care testing: \< 60 mg/dl 61-120 mg/dl 121-180 mg/dl 181-240 lmg/dl \>240 mg/dl
Time frame: 72 hours
| Arm | Measure | Group | Value (NUMBER) |
|---|---|---|---|
| Freestyle Navigator | Percentage of Readings in Different Blood Sugar Ranges as Shown by Point of Care Testing: | <60mg/dl | 10 percentage of readings |
| Freestyle Navigator | Percentage of Readings in Different Blood Sugar Ranges as Shown by Point of Care Testing: | 61-120mg/dl | 23 percentage of readings |
| Freestyle Navigator | Percentage of Readings in Different Blood Sugar Ranges as Shown by Point of Care Testing: | 121-180mg/dl | 41 percentage of readings |
| Freestyle Navigator | Percentage of Readings in Different Blood Sugar Ranges as Shown by Point of Care Testing: | 181-240mg/dl | 12 percentage of readings |
| Freestyle Navigator | Percentage of Readings in Different Blood Sugar Ranges as Shown by Point of Care Testing: | >240mg/dl | 14 percentage of readings |
Secondary
Sensitivity and Specificity of Navigator Projected High Blood Sugar Alarm Criteria Calculated for Events Defined by a High Reference BG Value (> 250 mg/dl)
Time frame: 72 hours
Population: This measure was not calculated as we did not have sufficient data to calculate sensitivity and inappropriate BG check times to calculate specificity and few trends towards hyperglycemia
Secondary
Sensitivity and Specificity of Navigator Projected Low Blood Sugar Alarm Criteria Calculated for Events Defined by a Low Reference BG Value (< 60 mg/dL)
Time frame: 72 hours
Population: This measure was not calculated as we did not have sufficient data to calculate sensitivity (not enough hypoglycemia) and discordant BG check times to calculate specificity.
Secondary
Sensitivity and Specificity of Navigator Threshold Alarms for Hyperglycemia (>240 mg/dl).
Time frame: 72 hours
Population: This measure was not calculated as we did not have sufficient data to calculate sensitivity and inappropriate BG check times to calculate specificity and few trends towards hyperglycemia