A team of Boston researchers has developed an artificial pancreas that prevented blood sugar from falling perilously low blood sugar (hypoglycemia) in a trial
Boston researchers have made a major step toward the development of an artificial pancreas that overcomes the bugaboo of most previous such attempts — dangerously low blood sugar caused by injection of too much insulin.
Their experimental device secretes two hormones normally produced by the pancreas — insulin and its counterbalancing hormone, called glucagon — and has been shown to control blood sugar levels in about a dozen people for at least 24 hours, they reported Wednesday.
The team is now planning longer trials as they gear up for what they hope will be approval by the Food and Drug Administration in as little as seven years.
“This is a very important proof-of-concept study,” said Dr. Irl B. Hirsch, an endocrinologist at the University of Washington School of Medicine, who was not involved in the research. “It was becoming obvious that if we were ever going to get [an artificial pancreas], we would have to use both hormones. . . . The fact that they have been able to do so successfully is very big and very exciting news.”
Most people know that Type 1 diabetes, which affects more than a million Americans, is caused by the loss of insulin-secreting beta cells in the pancreas.
Few realize, however, that the disease also affects alpha cells of the pancreas, which secrete glucagon to raise the level of sugar in the blood. Together, the two hormones help the body in the delicate balancing act of maintaining blood sugar levels that are neither too high nor too low.
Researchers have made tremendous advances in controlling blood sugar levels with continuous monitors and insulin pumps, “but one of the challenges is that we have an accelerator but not a brake,” which means blood sugar levels can fall too much, said molecular biologist Aaron Kowalski, a vice president of the Juvenile Diabetes Research Foundation, which partially sponsored the new research. The glucagon research, reported in the journal Science Translational Medicine, may provide that brake.
The brake is crucial, Kowalski added, because low blood sugar, or hypoglycemia, can cause seizures and even be fatal. For people who have been on insulin for long periods, the risk of hypoglycemia appears to grow over time, but it is also a problem in young children. “That’s what keeps us awake at night,” said Hirsch, who has been diabetic for 46 years.
What the Boston team has invented is a computer algorithm that responds to changes in blood sugar and computes how much insulin or glucagon to inject. Biomedical engineer Edward Damiano of Boston University began developing it a decade ago when his then-11-month-old son, David, was diagnosed with diabetes. He put the algorithm in a laptop and paired it with off-the-shelf insulin pumps — which could also be used to inject glucagon — and glucose monitors that are implanted under the skin.
After Damiano proved the system would work in diabetic pigs — which are remarkably similar to humans physiologically — Dr. Steven Russell of Massachusetts General Hospital suggested that it be tested in humans. Their first trials were conducted with 11 diabetics, who were hospitalized for 27 hours for the tests.
It was “a really rigorous test,” Russell said. “We fed them three very-high-carbohydrate meals, which is the most challenging part of automated control.”
Six of the patients experienced no hypoglycemia, but five had episodes that required they drink orange juice to recover. Subsequent analysis showed that those patients absorbed and metabolized insulin more slowly than normal. When Damiano adjusted the algorithm to account for this slowed absorption, all passed a repeat of the trial with flying colors — as did the six who did well the first time around.
The next test, perhaps beginning next month, will involve a larger group who will be on the devices for at least two days in the hospital. They will use a portable form of the system that will allow them to walk around and exercise on a treadmill, which is another challenge for automated systems.
A key problem is developing a stable solution of glucagon that can be used in the pump without decomposing, an effort that is being pursued by several companies.
But because that may produce delays in commercializing a pump that uses both hormones, Damiano and Russell think an insulin-only device will be first to market, in about five years. A device that uses both hormones might follow within a couple of years.
“I am committed to trying to get something along fast enough that it could develop into a commercial product before my kid goes to college,” Damiano said.