NEW YORK – Insulin pumps, continuous glucose monitors, and controller algorithms are all available off the shelf, but the reality of assembling them into a "set-and-forget" artificial pancreas is still some time away, according to Dr. William Tamborlane.
Postprandial hyperglycemia and nocturnal hypoglycemia excursions remain obstacles to the full implementation of "closed-loop" artificial pancreas systems that closely mimic the function of the human pancreas, said Dr. Tamborlane, professor of pediatrics at Yale University in New Haven, Conn.
Dr. William Tamborlane
In the past 2 decades, there have been substantial improvements in diabetes care, with the advent of insulin analogues, pramlintide as an aid in the absorption of glucose, smart insulin pumps, and continuous glucose monitoring systems, Dr. Tamborlane said at the annual advanced postgraduate course held by the American Diabetes Association.
However, as long-term follow-up data from the Type 1 Diabetes Exchange consortium show, "the findings are particularly disheartening for pediatrics, in that adolescents still have a mean hemoglobin A1c at the best centers of 8.7%. Too many type 1 diabetes patients fail to achieve target A1c goals," he commented.
In addition, while the relative risk of severe hypoglycemia with good glucose management has declined, it is an ever-present danger, and new diabetes technologies implemented in clinical practice have only added to the already high burden of care for both patients and practitioners.
"Certainly we would be delighted if there was a biological solution to type 1 diabetes that we could just cook up in our laboratories: [Grow] large numbers of beta [islet] cells based on stem-cell technology, infuse them into patients, and take away the problem of type 1 diabetes entirely. But the truth is that current transplant technologies, both whole pancreas and islet transplants, have their challenges and are only limited to a small segment of the population," he said.
Parts of the whole
The best current alternative, therefore, is a closed-loop artificial pancreas consisting of proven technologies combined into a single system.
For example, insulin pumps have been in use for decades, are continually being improved, and are now used by many children, adolescents, and young adults with type 1 diabetes. Continuous glucose monitors have been commercially available for more than 10 years, and have been shown to work in open-loop systems. Controller algorithms for dispensing insulin according to varying physiologic demands are also available, including PID (proportional-integral-derivative), MPC (Model Predictive Control), and fuzzy logic.
"Airplanes run on controller algorithms [as do] electrical generating plants, so there is plenty of information that could be applied to running an insulin pump based on basic engineering controller algorithms," Dr. Tamborlane said.
But early experience with a closed-loop system (Diabetes 2006;55:3344-50) showed exaggerated postmeal hyperglycemia excursions and a tendency toward late postprandial hypoglycemia from time lags in carbohydrate absorption, insulin absorption from the cutaneous infusion site (rather than the normal anatomic site, the portal vein of the liver), and from late increases in interstitial glucose concentrations.
"Overnight control was excellent, but you always worry that when you have thousands of patients on this treatment, you’re going to get some problems with sensors that are overreading which can cause hypoglycemia," he said.
Possible solutions to these problems include a hybrid design, for a semiautomatic control that will allow the patient to "prime" the pump with a conventional premeal bolus to account for some of the carbohydrate in a meal, and adjusting the sensor to a slightly higher-than-normal target glucose to prevent nocturnal hypoglycemia.
"Instead of trying to shoot for a morning serum glucose of 80 [mg/dL], why don’t we try to shoot for 120? Would you mind if, on average, your patients woke up every morning with a mean glucose of 120 mg/dL? I don’t think so."
Real-life conditions
Current artificial pancreas research is focused on testing the devices used in real-world situations but in a controlled inpatient testing environment.
In one such study, Dr. Tamborlane and his colleagues investigated whether the use of a closed-loop system can reduce the risk of delayed nocturnal hypoglycemia following exercise during the day. They enrolled 12 adolescents with type 1 diabetes and mean HbA1c levels of 7.4%, and admitted them to the research unit on two separate occasions for two overnight evaluations on both a closed-loop system and open-loop pump therapy.
The teens were evaluated on the night after an afternoon exercise period, and on a second night after being sedentary during the same time of day as the exercise. Investigators found that with the closed-loop system, there were only three episodes of hypoglycemia requiring treatment during all test nights, and only one on a night following exercise. In contrast, there were 22 total hypoglycemic excursions requiring treatment among the patients when they were on the open-loop system, including 14 on the night following exercise. (This study has been submitted for publication.)