Pro/Con: Should continuous glucose sensors be routinely used?

Continuous glucose monitors are approved for use in adults; should they be recommended for children?

Correction In this article, it was incorrectly reported that the GlucoWatch blood glucose monitor was no longer available. According to Animas, the GlucoWatch G2 is available in the United States by prescription only. “It is intended for detecting trends and tracking patterns in glucose levels in adults aged 18 and older and children/adolescents (aged 7 to 17) with diabetes. Parents should supervise the use of the G2 Biographer by their children/adolescents. The G2 Biographer is intended to supplement, not replace, conventional blood glucose monitoring,” Animas told Endocrine Today. For more information, go to www.glucowatch.com or call 1-866-GLWATCH.

Studies in children with first generation, retrospective-use sensors have shown use of the devices is associated with improvements in glycemic control. However, sensors may not provide sufficiently accurate information when it comes to detecting hypoglycemia.

At the Pediatric Academic Societies’ 2006 Annual Meeting, Stuart Alan Weinzimer, MD, FAAP, assistant professor of pediatrics at Yale University, said the sensors could be valuable tools to keep children within target HbA1c ranges. Darrell M. Wilson, MD, chief of pediatric endocrinology at Stanford University Medical Center, Calif., agreed the sensors are important, but they may not do the job of warning of nighttime hypoglycemia — and routine use in children is premature at this point, he said.

Each physician presented evidence in a debate at the meeting held in San Francisco.

Trend analysis

Continuous glucose monitoring could help children stay within their target HbA1c range, Weinzimer said.

“Our kids with diabetes spend a lot of time outside their target range ... and even the first-generation devices have been able to show reductions in HbA1c,” he said.

Routine monitoring of blood glucose is not enough, he said.

“Even with up to nine finger sticks a day, fewer than 30% of routine tests are in goal ranges,” Weinzimer said. “Almost a third are above 180 mg/dL, and people with diabetes spend two hours a day hypoglycemic.”

In a study Weinzimer and colleagues conducted at Yale University, only 10% of children with type 1 diabetes (mean HbA1c 7.5%) had peak postprandial blood glucose at 100 mg/dL to 180 mg/dL; half had peak blood glucose greater than 300 mg/dL, he said.

“This was a group of well-controlled diabetic children, so you can imagine what this would look like if we chose children who had HbA1c levels of 8% or 9%,” he said.

Continuous monitoring can also provide a broader range of information, Weinzimer said.

“The advantages of continuous glucose monitoring are that it not only shows the four or six or eight times a day that you test your blood [glucose], but it gives a full range of glucose excursions and the trend analysis. This change over time is very valuable information that you can use in diabetes management,” he said.

There is also an interactive component to continuous monitoring.

“With a real-time device, you can adjust therapy at the time based on your glucose levels, and more importantly, based on the glucose trend,” he said. “If, for example, you read your device and your blood [glucose] is 150 mg/dL, you might not do anything about that, but if your blood [glucose] is 150 mg/dL and you’ve been trending up over the last hour, you can give yourself extra insulin. This trend analysis is particularly important.”

In a study using GlucoWatch (no longer available), children who wore the device were able to reduce HbA1c levels over time vs. controls who did not use the device. Controls who then used the device were able to see similar results, Weinzimer said.

Proactively, the most powerful use of continuous monitors would be an alarm function to prevent hyperglycemia or hypoglycemia, Weinzimer said.

In a recent study, use of a device from DexCom demonstrated a small reduction of time spent in both the hypoglycemic and hyperglycemic ranges, and more time spent in the euglycemic range, Weinzimer said. These trends were evident particularly at night, he said.

“These devices are not as accurate as hand-held glucose meters. As time goes on, the devices are getting more accurate, although they do suffer some in the hypoglycemic range. In the range where our patients with diabetes typically live, they are certainly accurate enough to be used for real-time decision-making,” Weinzimer said. “To improve the use of these devices, we should develop some algorithms so that we can help our patients. We also need to show the cost-effectiveness of using them. And, of course, we should be reimbursed for using them.”

Use with insulin pumps

Because they are not accurate in low ranges, sensors do not provide good hypoglycemia detection, Wilson said. Still, they could be valuable for other use.

“The point is that although these don’t make very good hypoglycemia detectors, they line up fairly well with what you need for the accuracy criteria and duration of action for pump control,” he said. “Ironically, you don’t need a sensor to be as accurate for pump control as you do for hypoglycemia detection. Paradoxically, the accuracy needs go down as you become closer to a closed system.”

For children as well as families, use of continuous glucose sensors is too preliminary for several reasons, Wilson said.

“There are too many unanswered questions at this time to recommend their routine use in children. We need to know how best to use them and how to minimize any risk,” he said.

For patients to see too much of what’s going on, they may be more likely to overreact and give themselves too much insulin too frequently, Wilson said.

This so-called “stacking” of insulin on insulin can happen when patients do not understand the time course that it takes for insulin to get ingested, absorbed and to work, he said.

Clinical studies have not necessarily proven the efficacy of continuous monitoring, Wilson said.

“There have been approaches in the clinical studies that kind of vary between the polar extremes of laissez-faire, or, as I like to refer to it, ‘free range studies,’ where you give somebody a direct reading sensor and say ‘go have fun with it,’” Wilson said. “You have to be very careful reading these studies because some of them have done that. Some studies have only been conducted for seven, eight or 10 days. Some of these studies are very small, and some of them are uncontrolled.”

Other questions for use of continuous monitoring surround an algorithm-driven approach, Wilson said. Such an approach means giving people advice on what to do with their blood glucose testing device.

“So, if you have an up-slope at a time that’s before a meal, how much should you increase your bolus? Should that vary by the insulin regimen you use?” he asked.

Which settings to best use the sensors on is another question, Wilson said.

“Many of our families are desperate for something that will warn them and help them prevent nocturnal hypoglycemia. That’s the big push and, unfortunately, that’s an area where the sensors do the least-good job,” he said.

“We need to see that the duration of improvement is longer, and we need to look at the appropriate outcomes, which are not just going to be HbA1c. We now have a number of patients who are pretty well controlled and to expect that you’re going to get them even better controlled may be a bit naive.

“We should be looking at things like increased variability of blood [glucose], easier time managing blood [glucose] levels and more flexibility,” Wilson said. “The false-positive and false-negative rates for hypoglycemia detection still remain too high to be a practical device. Currently, continuous glucose monitors are not going to be the tool that we want to prevent nocturnal hypoglycemia.” – by Rachel Eskenazi

Dr. Weinzimer has received grant support from Medtronic to study a closed-loop artifical pancreas, which includes a glucose sensor component. He has also received honoraria for speaking engagements from Medtronic.

For more information:

• Weinzimer SA, Wilson DM. Should continuous glucose monitors be routinely used? Presented at: Pediatric Academic Societies’ 2006 Annual Meeting; April 29-May 2, 2006; San Francisco.