Reports
Potential for an Artificial Pancreas: Addressing Continuous Glucose Monitoring
This report is based on medical evidence presented at sanctioned medical congress, from peer reviewed literature or opinion provided by a qualified healthcare practitioner. The consumption of the information contained within this report is intended for qualified Canadian healthcare practitioners only.
66th Annual Scientific Sessions of the American Diabetes Association
Washington, DC / June 9-13, 2006
Tight glycemic control has been demonstrated in several clinical trials to reduce the risk of micro- and macrovascular complications of diabetes. Blood glucose management has traditionally relied on the results of self-glucose monitoring from a relatively small number of fingerstick capillary-blood samples taken throughout the day. This method has several limitations due to the frequency and timing of the sampling.
Continuous glucose monitoring systems have been developed that measure glucose from the subcutaneous (s.c.) interstitial fluid glucose every five minutes over a period of up to 72 hours. Although the correlation between blood and interstitial fluid glucose is good, discrepancies between these two compartments have been reported, especially when glucose levels are changing rapidly.
Toward an Artificial Pancreas
An improved diabetes management system would create an “artificial pancreas” in which continuous s.c. insulin infusion (CSII), commonly referred to as insulin pump therapy, is combined with continuous glucose monitoring to allow real-time glucose monitoring with minimal patient effort, explained Dr. Richard Rubin, Department of Medicine and Pediatrics, The Johns Hopkins University, Baltimore, Maryland.
Insulin pump and glucose sensing technology have independently been advancing steadily, but the integration of real-time monitoring with insulin delivery would be a major step forward in treating patients with type 1 diabetes.
Enhancing Glycemic Control
Dr. John Pickup, Division of Diabetes and Metabolism, King’s College London School of Medicine, UK, enumerated four aspects of glycemic control considered important to normalize: elevated hemoglobin A1c (HbA1c) and mean blood glucose; elevated postprandial blood glucose; hypoglycemia; and blood glucose variability (both within-day and between-day).
With multiple daily insulin regimens, about 15% of patients have elevated HbA1c values, he observed, and about one in five patients have two or more episodes of severe hypoglycemia in a given year. “The distribution is very skewed so that 5% of the patients have more than 50% of the episodes of severe hypoglycemia,” Dr. Pickup told delegates. This is a target group for CSII, but even with continuous infusion, hypoglycemia is not abolished.
In a meta-analysis of 21 clinical trials, the rate of severe hypoglycemia was reduced from 46 episodes per 100 patient-years with multiple daily injections to 12 episodes per 100 patient-years with CSII, he reported. “Blood glucose variability is reduced but not totally abolished by switching from multiple daily injections to CSII,” he cautioned. With continuous insulin infusion, overall glycemic control is good but can “disguise tremendous day-to-day variability.”
“The potential benefits of an artificial pancreas are less blood glucose variability, less severe hypoglycemia and improved quality of life,” stated Dr. Rubin.
Assessing Intermittent Blood Glucose Monitoring
Intermittent blood glucose monitoring has several shortfalls, including discomfort, non-compliance and an inability to perform while sleeping or driving. Fluctuations in glucose values are often missed because a single fingerstick sample can only reflect the glycemic status at the particular time of sampling. Intermittent monitoring has been shown to miss episodes of hyper- or hypoglycemia; testing glucose four or five times a day has revealed episodes of severe hypoglycemia missed by intermittent testing, noted Dr. Pickup.
Artificial pancreas systems incorporate glucose sensing, insulin delivery and controllers (or algorithms). Fully closed-loop strategies combine continuous glucose sensing with automatic feedback control of insulin delivery. In a semi-closed loop system, the patient can activate an insulin boost at meals to better control glycemic excursions. A “closed-loop” system that would represent an artificial pancreas has not yet been achieved as patients must still perform twice-daily fingersticks to calibrate the glucose-monitoring system, and there is a lag time of seven to eight minutes in delivery of data to the patient.
The main reason an artificial pancreas is not in routine use is that the development of reliable glucose sensors has been slow, Dr. Pickup indicated. “Continuous glucose sensing is necessary for complete assessment of glycemic control… but perhaps not for closing the loop.” Computer simulation and clinical studies have shown that three-hourly glucose sampling produces reasonable glucose response after initial hyperglycemia, although one- to two-hourly sampling may be necessary in free-moving patients.
Real-time Monitoring
The main advantage of real-time continuous glucose monitoring is that patients can make adjustments to their insulin much earlier than with intermittent glucose testing, said Dr. Jan Bolinder, Department of Medicine, Karolinska University Hospital Huddinge, Stockholm, Sweden. In this way, they may be able to take preventive action to avoid glucose excursions, especially with hypo- and hyperglycemia alerts built into the system.
In a study of transcutaneous real-time continuous glucose monitoring, patients in the intervention group spent significantly less time in hypo- and hyperglycemia (P<0.0001) and significantly more time in the target glucose range of 5.0 to 7.7 mmol/L (P<0.0001) during the day compared with a control group that did not use continuous glucose monitoring (Garg et al. Diabetes Care 2006;29(1):44-50). There were also trends toward improved nighttime glucose control with continuous monitoring, he reported.
Dr. Bolinder described the GuardControl study, a European multicentre clinical trial of 162 patients with type 1 diabetes with inadequate glycemic control as evidenced by HbA1c >8.1% who were randomized to the Guardian RT (either continuous use or biweekly use for three days) or conventional self-monitoring of blood glucose using fingersticks (control group). The overall primary objective was to determine whether patients with poor glycemic control could achieve improved metabolic control using the real-time values of the Guardian RT compared with fingersticks after 12 weeks of continuous use.
The Guardian RT Telemetered Glucose Monitoring System is indicated for continuous or periodic monitoring of interstitial blood glucose values and has alarms for low and high blood glucose. The glucose values calculated by the device are used to trigger hypo- and hyperglycemia alerts in the form of an alarm and are displayed every five minutes. Data are stored for up to 21 days.
The group that wore the continuous glucose monitoring system had an HbA1c reduction of 1.1%, the group with biweekly use of the device experienced a 0.8% reduction and the controls had a 0.4% reduction (P=0.03 for both groups vs. control). Furthermore, patients assigned to continuous monitoring had significantly fewer glucose excursions >10.5 mmol/L or <3.8 mmol/L. “Three-fourths of the patients wanted to include continuous glucose monitoring as routine in their clinical care,” Dr. Bolinder reported. “One-third wanted to do it all the time or once weekly.” Patients also indicated that the alarm system was beneficial in alerting them to high or low glucose values.
The study demonstrated that real-time continuous glucose monitoring facilitated self-adjustment of diabetes management and led to improvement in glycemic control without increasing the risk of hypoglycemia.
Clinical Experience with Real-time Systems
Other continuous glucose monitoring devices with real-time glucose monitoring are the Paradigm 722 System, Dexcom STS and the Navigator. Clinical experience with these systems indicate that the low-alarm setting needs to be patient-specific, with a higher setting encouraged for hypoglycemic unawareness, said Dr. Bruce Buckingham, Department of Pediatrics, Stanford University School of Medicine, California.
From a study of 30 children wearing the Navigator for three months with an average HbA1c of 6.8%, an appropriate initial setpoint for a high-glucose alarm appears to be 13.3 mmol/L. Appropriate candidates for real-time systems are patients who want to improve their diabetes management and those willing to test at least four times a day (twice per day to calibrate and before any therapy adjustment), Dr. Buckingham told delegates. He noted that populations with high potential benefit are those with hypoglycemia unawareness, poor metabolic control, possible over-insulinization, fear of hypoglycemia and a history of nocturnal seizures.