Poor Patient Adherence May Undermine Aim of Continuous Glucose MonitoringPoor Patient Adherence May Undermine Aim of Continuous Glucose Monitoring

Chicago—The approval in 2006 of continuous glucose monitoring (CGM) devices brought hopes that providing patients with type 1 diabetes with real-time information from a sensor implanted just under the skin would allow them to achieve better glycemic control. However, researchers have had a tough time showing efficacy of CGM in the clinic due to such factors as problems with device calibration and a lack of rigorous trial data.

A study presented at the American Diabetes Association’s 67th Scientific Sessions, held here in June, suggests that yet another element may limit how well CGM translates into glycemic control: patients who are indifferent to taking action to adjust their blood glucose level based on the information generated by CGM.

In a 6-month randomized multicenter trial, 138 experienced insulin pump wearers with initial hemoglobin A_1c (HbA_1c) levels above 7.5% (7.0% or lower is the recommended target level for patients with diabetes) were randomized to CGM or the conventional fingerstick method to monitor and act on adjusting their glucose levels. At 6 months, HbA_1c levels for both the CGM and control groups decreased from a mean of about 8.5% to 7.8%, reported Irl B. Hirsch, MD, medical director of the University of Washington Diabetes Care Center in Seattle.

In a subset analysis of the data, Hirsch and colleagues found a linear relationship between CGM compliance and lower HbA_1c levels. Those with 100% or more adherence (defined as using the sensor at least 6 days per week) saw their average HbA_1c drop from 8.56% to 7.69%. Those with less than 60% adherence saw their baseline HbA_1c levels actually rise from a mean of 9.45% to 9.63%.

“These people with the higher [HbA_1c levels] were not taking care of their diabetes before the trial, and if they're not interested and focused, this won't help them,” Hirsch said.

For future trials, Hirsch said, investigators should enroll motivated patients if they expect to see any impact of CGM on glycemic control. “We used the wrong population because diabetes is lifestyle—patients have to be interactive with their sugars, and if they're not involved, we can't help them.”

For patients who are motivated to play an active role, CGM is the kind of technological advance that warrants studies to evaluate its efficacy, said Aaron Kowalski, PhD, director of strategic research projects with the Juvenile Diabetes Research Foundation International (JDRF), in New York City. “Glucose control has been very difficult for patients with diabetes, and a major contributor has been inadequate tools,” said Kowalski, who has type 1 diabetes and has used CGM himself. “But now there's interest in a new generation of technology that many of us in the field think [is] functional . . . but now we want to demonstrate through studies that it works.”

To that end, the JDRF is sponsoring a randomized controlled trial, the Randomized Study of Real-Time Continuous Glucose Monitors in the Management of Type 1 Diabetes (http://clinicaltrials.gov/ct/show/NCT00406133?order=1), which began enrolling 450 patients at 10 sites in December 2006. Trial participants are randomly assigned to either CGM or the fingerstick method of glucose monitoring for 6 months, with frequent follow-up visits and telephone contact to review their diabetes management. After the initial study period, those in the CGM group will be monitored for an additional 6 months with less intensive contact (fewer follow-up visits and telephone contacts) to see if any benefits from the first 6 months are sustained; those in the control group will initiate CGM with less intensive contact with clinicians.

The study's primary end points at 6 and 12 months are HbA_1c levels, episodes of severe hypoglycemia, and percentage of blood glucose sensor values in the range of 70 mg/dL to 180 mg/dL (3.89 mmol/L to 9.99 mmol/L). Six-month results are expected by the middle of 2008, with final results by the end of that year.

Given the lack of efficacy data (and of long-term studies showing that CGM reduces morbidity and mortality associated with diabetes), it is not surprising that payers have been reluctant to reimburse for CGM, which can cost thousands of dollars for the initial purchase of the devices and hundreds of dollars annually for replaceable sensors. Another potential barrier to its use is physicians fearing information overload as they try to incorporate the new data provided by CGM into clinical practice settings where they may have only 20 minutes to work with a patient.

Kowalski is hopeful the JDRF-sponsored trial will show benefit for CGM. “When we ask payers how we can get [CGM] into the hands of people with diabetes, they say they need to see [Hb]A_1c reduction,” he said.