By: Patrick Totty
The Holy Grail pursued by all diabetes researchers is a complete cure for both the type 1 and type 2 forms of the disease. But until then, the “artificial pancreas,” a combination of glucose monitoring and insulin dosing technology, may be as close as they get to a final breakthrough in treating diabetes.
So far, two of the three elements necessary to mimic a normal pancreas’s insulin production and distribution are in place: continuous glucose monitors and insulin pumps. Each is attached to the user via its own entry point. The CGM provides a continuous stream of data about blood sugar levels to the user, who can then send a command to the pump to deliver a correcting dose of insulin if those levels are too high.
Now a European research project, “AP@home,”* (AP = Artificial Pancreas) is looking to create a CGM/insulin pump device that requires only one entry point. Via that single entry point, the monitor will read blood glucose levels and the pump will dispense insulin.
The “single-port” device, which the project’s participants hope to have on the market within four years, would reduce the wear and tear on users’ bodies as they periodically reposition CGM and pump entry points, as well as lessen the risks for opportunistic infections posed by multiple entry points.
As much as the single-port variation offers, the European group knows that its ultimate success as a true artificial pancreas depends on a third element, a small computer that can analyze CGM data and then give an appropriate instruction to the insulin pump. Users wearing this combination meter/computer/pump, or artificial pancreas, would be freed from constantly having to track their blood sugar levels and make potentially risky decisions about their dosing levels.
But for now, that crucial bridge between data and response is provided by users. Based on their past experiences and knowledge of their own bodies, as well as data from their CGMs, they decide whether to dose themselves and with what quantities.
The problem, of course, is that no computer can bring to bear the mass of insights, experiences, hunches, facts, and other elements that people use when they make a decision. The algorithm-or series of specialized instructions-used by the computer in an artificial pancreas would have to be very complex.
In fact, given how each person with diabetes has a distinctly individual pattern of highs, lows, and anomalies, the ideal algorithm will have to be customized. The software will have to be “trained” to anticipate events.
For example, a triathlete with diabetes may deliberately want her blood glucose reading at the start of a competition or intense training session to register 140 mg/dl, which under normal circumstances would be considered very high. But she knows from experience that the rigors of a race or workout will reduce her blood sugar levels considerably as she goes along.
So the algorithm for her artificial pancreas will have to take into account race or training day conditions versus days when she’s not engaging in robust physical activity. People like her who have unusual needs may wind up having more than one algorithm for operating their artificial pancreases-each depending on the need of the day. (No doubt this will inspire the development of little digital minders that ask, “Are you sure you want Algorithm A, or were you thinking of Algorithm B today?”)
Another problem is that the ideal blood sugar level currently aimed for by pump and CGM users is not very far above hypoglycemic. (The median “normal” blood sugar level is about 72 mg/dL, although that figure temporarily fluctuates upward after meals. Most experts warn people with diabetes to be alert if their readings reach 70 mg/dL.) The technology has to be able to keep users’ levels as close to ideal as possible without letting them drop the crucial last few deciliters that spell the difference between strict control and a debilitating hypoglycemic episode.
What might have to happen is a combination of two things: 1) Development of an incredibly sophisticated algorithm that allows users to safely skirt the lower end of healthy blood sugar levels (housed in a fail-safe mechanism that is constantly self-diagnosing and has a loud, attention-getting way of signaling when it’s having a breakdown); and 2) A redefinition of ideal blood glucose levels that allows for somewhat higher readings-a “cushion” between strict control and hypoglycemia.
In the meantime, the Europeans will push for a single-port device that significantly improves user comfort and moves the artificial pancreas concept one large step closer to reality.
* Seven European universities and five companies in the fields of metabolism research, product design, diabetes, R&D, computer technology, and engineering are participating in the EU Artificial Pancreas System, AP@home, research project. Among the partners, the Swiss company Sensile Medical AG is contributing the central innovative technology of the single-port-system.
* * *
Source: Artificial Pancreas Project press release