Great presentation! I do wonder, how does this bihormonal artifical pancreas manage the postprandial peaks without user input, what is the difference in the control algorithm compared to the closed loop devices that are currently available in clinical practice which still need user input for meal boluses (such as the 670G)?
Thank you! The control algorithm responds to the change in glucose value, together with the glucose value itself this will determine the size and frequency of insulin dosing. The closed loop devices currently available in clinical practice need carbohydrate input for the delivery of meal boluses, those control algorithms adapt the basal insulin delivery.
Wonderful news! These are really exciting and promising data!
– Strict glucose control is achieved with no or little patient interaction with the system. I understand that tpatients complained about alarms, the number and the size of devices. How many alarms a day did patients experience? Were most alarms during the day or during the noght? How long does it take a patient to ‘handle’ an alarm? Did you ask patients what system they preferred: closed loop or open loop?
– Why did you use different rtCGM devices in closed and open loop?
The number of alarms was highly variable and could be from 1 till 10 per day. These alarms relate for example to required maintenance operations, such as replacing the glucagon or need for an SMBG, or glucose sensor statuses. Most alarms occurred during the day, but also during the night SMBGs were requested. The time to handle an alarm can be very short till around 5 minutes (in case of an infusion set occlusion). There was no formal evaluation of their preference, but the majority was positive about the closed loop system.
The closed loop system has its own CGM component (including transmitter), which cannot be used in open loop for two weeks. We did not want to use an additional (third) glucose sensor during the closed loop because of the already high burden with two sensors and two infusion sets.
Top study ! Congrats !
I do have a Q: the open loop control period, on what pumps and sensors was that ?
The patients continued their usual care, so they used different pumps and sensors (5 patients used rtCGM, 12 FGM and 6 no sensor) . For data collection they used a masked Dexcom G6.
Very interesting. I wonder, what type of patients were they? highly educated, highly motivated? What efforts did they have to maken? Only twice a day fingermeasurement? How about stress, work, exercise, was that allowed?
The patients were indeed highly motivated, but their work/educational background was different. They had to perform maintenance operations regarding sensor, infusion sets and battery replacements. At least one SMBG is required every day, but often more were needed (3-4 per day). There were no restrictions, except some contact sports to prevent damage to the device and swimming was not allowed. The size of the device limited patients in some specific sport activities.
Thank you for your presentation of this very nice study with very interesting results!
I have a question. Did the patients do any physical exercise / sports during the closed loop period?
In patients that use an insulin-only (advanced) closed loop system, exercising is often quite problematic, resulting in having to shift back to manual mode. Since this bihormonal artificial pancreas does not have the opportunity for manual mode, I was wondering if you are expecting any issues in patients with (intensive) physical exercise?
And is there a specific reason for not including a manual mode as a fail-safe?
The patients did perform physical exercise. The exercising itself is not problematic for the closed loop, it also reduces the insulin administration based on accelerometer data and the glucagon prevents hypoglycemia. However, the size of the device (wearing it on the front side of the body) limited some patients in their sport activities. Adding a manual mode would make the system more complex (regarding user interface, training, software, risk analysis etc.)
Thank you for your answer.
I was just wondering since I just saw the comment about swimming. Is swimming never allowed for users of this device or was it just for the study period? Is the device easily removed/detached if a patient wants to go swimming?
The device itself can be easily disconnected. However, the transmitters attached to the glucose sensors are not approved/tested for swimming. This means that the glucose sensors and transmitters need to be removed from the body before swimming. This needs to be addressed in a subsequent version of the system.
I have several patients using an insuline only closed loop DIY/DIT system and the important thing about both exercise and stress is the algorithm; the algorithms of the commercially available pump/sensor combinations do not react as swift as the open source ones and thus exercise can lead to necessity of overrruling the algorithm. I guess the algorithm here is smarter and furthermore the glucagon works instantaneously as opposed to insulin, which requires more time
Top study, congratulations
Wonderful study, thank you for the clear presentation. I am positively surprised about the time in range all the participants achieved. However, do you think the postprandial peaks – as observed in the closed loop group – might be an issue?
We compared the postprandial periods, defined as 3 hours from the start time of the meal. For the lunch, the postprandial median glucose was lower for the closed loop than for the open loop. The postprandial median glucose values were not different for the breakfast and dinner.
Congratulations, Helga. Impressive results!