At the AADE 2013 Conference in Philly, I was thrilled to meet Molly McElwee Malloy, a CDE and clinical nurse trial coordinator at the Center for Diabetes Technology at the University of Virginia School Of Medicine. She’s a type 1 herself who works with the famous Dr. Boris Kovatchev running cutting-edge Artificial Pancreas trials. They even have a “research house” with four bedrooms where patients can sleep for overnight studies.
Molly told me about how they’re testing an AP system made up of the Dexcom and various pump models (they’re “pump agnostic”), plus software and a receiver they’ve developed at her clinic in both children and adults – and how they’ll start an adult study using a Medtronic pump next month. The research team, also associated with Dr. Howard Zisser’s work at the Sansum Diabetes Research Institute in Santa Barbara, has published a number of cutting-edge research reports, including a 2012 report on how “Closed-Loop Glucose Control Maintains Near Normoglycemia,” and a new study in Diabetes Care July 2013 demonstrating “Feasibility of Outpatient Fully Integrated Closed-Loop Control.”
Molly told me how they “treat to goal” of 112.5 mg/dL, because they find that “shooting for a broader range leads to suboptimal results… you tend to have patients hover at the high end.”
She told me how they’re experimenting to see how the system can react to a missed bolus, or an unannounced 30-gram-carb snack, and how they’re conducting a “sub-study” in which parents can watch their kids’ results in real-time using sophisticated AP remote monitoring.
But most importantly, Molly told me about a feature they’re testing that quite possibly could have saved her life. It’s something they call a Safety Supervision System (SSS), which works inside the artificial pancreas designed to continuously monitor hypoglycemic risk and intervene as necessary: “The SSS is currently implemented in a cell phone which has been modified to become a medical device, and, in contrast to existing hypoglycemia alerts that use a CGM signal alone, is informed by glucose and insulin-on-board data processed by a model of glucose-insulin kinetics.” The cell phone can also alert a loved one or emergency contact if you are in trouble and not responding to alerts.
They’ve also tested something referred to as a “breaks algorithm” that tells the AP to sloooow dowwwn insulin delivery in certain situations, like during exercise. Basically, automated insulin delivery can be “sensitized 10-fold” for exercise or other activity by the patient just by hitting a button at the start and end of the exercise session. It’s kind of like the Low-Glucose-Suspend feature, except that it drastically reduces insulin delivery rather than shutting it off entirely, and it’s activated and deactivated manually by the patient. When their integrated, closed-loop system was tested last year in two “randomized cross-over” studies, “time spent in near normoglycemia increased significantly overall … to 74.4%, with a maximal effect overnight.”
Molly thinks the SSS feature could have averted disaster for her on the fateful day of Oct. 12, 2011 — when she “almost became a (dead in bed) statistic.”
She had a severe nighttime low that led to a seizure — in which she bit nearly all the way through her tongue (eww!)
She was rushed to the Emergency Room after her husband couldn’t revive her, and the physicians apparently told her a few minutes’ more delay in getting treatment could have spelled catastrophe.
So what did this scientist-type do? Thoroughly research her own case, of course, to illustrate the power of the SSS in the artificial pancreas!
“My team ran my data through a simulation to show what the artificial pancreas WOULD have done in the same situation,” she explains. According to their calculations, the SSS would have activated a “red light” at 32.7 minutes before her CGM readings dropped to 70 mg/dL.
Molly, along with researchers Colleen Hughes-Karvetski and Boris Kovatchev, even wrote an abstract that was published at the ADA 72nd Scientific Sessions on how this breaks system could have thwarted her emergency.
“Figure 1 shows the course of the event and how the SSS would have intervened: 60 minutes prior to severe hypoglycemia (SH) the system would have attenuated and discontinued insulin delivery; 38 minutes before the event, the SSS would have alerted that, without any additional insulin, hypoglycemia is imminent. This alert would have been presented to the patient and caregiver via the system’s remote monitoring capability. Automatic safety measures provide opportunity to reduce the incidence of SH, particularly overnight.”
“It’s like ADT for diabetes,” Molly told me. Wow.
Finally, reviewing the various papers Molly shared with me, I noticed this passage about developing a working AP system:
“To cope with the changing environmental conditions and with the physiological/behavioral changes of the patient, the future ambulatory artificial pancreas will have to adapt to the changes in an individual’s biobehavioral parameters over time. Possible methods to cope with changing daily conditions include individual controller calibration strategies and run-to-run control algorithms, as well as behavioral analysis and profiling of patient lifestyle.”
What? A “smart” Artificial Pancreas with a built-in security system that actually takes your lifestyle habits into consideration?!
“Bring it on!” was all I could think of in response.