advertisement

  1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

Is a "closed-loop" CGM and Pump a Cure?

Discussion in 'Stickies' started by Sarah Maddie's Mom, Jan 6, 2010.

?

Is a "closed-loop" system a cure?

Poll closed May 6, 2010.
  1. Yes

    4 vote(s)
    1.7%
  2. No

    224 vote(s)
    96.1%
  3. I don't know

    5 vote(s)
    2.1%
  1. BrokenPancreas

    BrokenPancreas Banned

    Joined:
    Aug 9, 2009
    Messages:
    1,517
    I'll take it, I'll take it in a hearbeat.
    I think of this way...
    If there was something that a terminally ill cancer patient would have to wear to stay alive, no chemo, no radiation, and just pass on when they are old, I think lots of people would consider that a cure.
    If this keeps my baby from seizures, kidney damage, etc, then they call call it a "Goldmine" or "SuperKewl" or anything they want.
    I want this AP, and I want it NOW!!:D:D
     
  2. hawkeyegirl

    hawkeyegirl Approved members

    Joined:
    Nov 15, 2007
    Messages:
    13,151
    If we were talking a fully-functioning artificial pancreas that only required a site change every 3 days and a sensor change every 7 days, I'd be excited too, even though it wasn't a cure. But I have zero hope for any such device being commercially available in the next 15 years. Zero. I have more hope that we'll get a cure in the next 15 years, and my hope for that is next to zero. :(
     
  3. BrokenPancreas

    BrokenPancreas Banned

    Joined:
    Aug 9, 2009
    Messages:
    1,517

    Karla -
    Okay, totally confused now.
    What will be available in four year then?
    I read Dr. Kowalski's <sp> info, is that not happening?
     
  4. hawkeyegirl

    hawkeyegirl Approved members

    Joined:
    Nov 15, 2007
    Messages:
    13,151
    Nah. We MIGHT see a "first generation" AP in 4 years, but I'm not holding my breath. I think a first generation device will be something like I saw at the CWD seminar a year ago. Basically, it will regulate blood sugar at night to keep it in the 80-180 range. Based on the slides that I saw, we personally already do better 90-95% of the time than that device did. It was not even close to being blow your socks off great.
     
  5. Flutterby

    Flutterby Approved members

    Joined:
    Nov 11, 2006
    Messages:
    14,623
    I agree.. I personally feel they are hyping this way to much.. there is NO WAY they will have something ready in 4 years that EVERYONE will have access too.. absolutely no way.. Minimed is having a hard enough time just going through the FDA for a pump UPGRADE.. they can't get a pump/sensor combo through the FDA with an automatic shut off for a bg below 50.. how is a company going to get through product design, testing and trials AND the FDA in 4 years.:confused:
     
  6. Ellen

    Ellen Senior Member

    Joined:
    Oct 22, 2005
    Messages:
    8,240
    Interesting timing of this published article

    Diabetes Technol Ther. 2010 Jan;12(1):35-9.

    Parental attitudes towards overnight closed-loop glucose control in children with type 1 diabetes.

    Elleri D, Acerini CL, Allen JM, Hayes J, Pesterfield C, Wilinska ME, Dunger DB, Hovorka R.

    Department of Paediatrics, University of Cambridge, Cambridge CB2 0QQ, United Kingdom.

    BACKGROUND: This study examined parental attitudes towards overnight closed-loop (CL) glucose control in children and young people with type 1 diabetes (T1D). METHODS: Twelve families recruited by the INsulin PUmp Therapy group (INPUT), a U.K. patient/carer-led support group for people with T1D, attended a focus meeting on CL. The concept of CL was explained, and clinical results and plans regarding the Artificial Pancreas Project at Cambridge were presented. Participants completed a questionnaire to evaluate parental feelings about T1D management and attitudes towards overnight CL insulin delivery. RESULTS: Nineteen parents (12 mothers and seven fathers) anonymously completed the questionnaire. Main concerns about diabetes were related to long-term complications (84%) and hypoglycemia (16%). Achieving good glucose control represented major challenge for carers (83%) with nighttime being the most difficult period to manage (56%), worrying most parents (71%) particularly because of fear of hypoglycemia (33%). The development of CL to manage diabetes was welcomed by all parents (100%) with 95% happy for their child to wear a continuous glucose monitor together with an insulin pump. All parents were ready to respond to additional alarms at night with 90% parents not worried about their child's overnight insulin delivery being controlled by a computer. CONCLUSIONS: Selected parents of children with T1D who are already insulin pump users express implied trust in the technology for overnight CL insulin delivery. Nighttime blood glucose control is the biggest challenge and concern for parents, and the development of a commercially available overnight CL system may be an important goal.

    PMID: 20082583 [PubMed - in process]

    Publication Types
     
  7. Sarah Maddie's Mom

    Sarah Maddie's Mom Approved members

    Joined:
    Sep 23, 2007
    Messages:
    12,510
    A study of 12 parents...

    I'm the first to admit that stats was NOT my best subject in grad school... but a sample survey of 12?:rolleyes:
     
  8. Darryl

    Darryl Approved members

    Joined:
    May 8, 2008
    Messages:
    4,313
    Dr. Kowalski,

    I have great admiration and appreciation for your lifelong pursuit of a cure, as well as the work and studies you undertook in recent years to prove that CGM's are beneficial for people with T1. Thank you for posting here recently about the AP project.

    Support for the AP project among parents is mixed, for reasons including the misleading name, and the virtual impossibility that it will ever be a "fully automated" closed loop control system - that is, not requiring human input or manual override.

    I would like to address a more fundamental question, and hope that you can provide an answer that we can understand.

    As you are most likely aware from your work so far, closed loop control systems consists of three components: A sensor, a control mechanism, and a feedback algorithm. Loop stability requires that both the sensor and the control mechanism have stable and reliable gain and time delay parameters.

    Of the two physical components, the sensor is by far the most crucial: If the pump fails, the closed loop algorithm can alarm when BG reaches a threshold. However, should the sensor read innacurately, or in the worst case reads the opposite of the current BG trend, the loop may become dangerously unstable.


    Current sensor technology has the following limitations:
    • The worst case scenario mentioned above (flat or falling ISIG while BG is rising, and vice-versa) is fairly common after a bad calibration, when there is pressure on a sensor, if the sensor is bumped, or if it becomes partially dislodged.
      .
    • The ISIG:BG ratio (in engineering terms, the sensor's gain parameter) can vary signficantly during the first 12 hours and the last full day of the sensor's life. Since these gain variations are not predictable in the same way that IC ratio is predictable by time of day, there is no way the AP algorithm could be calibrated to compensate automatically; therefore, the "fully automated" control concept is virtually impossible to conceive during the first and last day of a sensor's life.
      .
    • Each sensor has it's own inherent ISIG:BG ratio - which can be determined through careful calibration - but in addition, other sensor-specific gain and time delay parameters that can not be easily determined. One of these parameters is the sensor's gain compression curve (a range of BG's where the sensor's gain becomes highly nonlinear). Some sensors respond at a faster rate than other sensors, (i.e., the sensor's time delay), and this delay can also vary by the site location and level of hydration. It will be virtually impossible to calibrate the algorithm for these types of gain and time delay parameter variations that will be different in every sensor.
    Current pump technology is somewhat reliable and accurate, but even if the pump fails, at least the sensor would know it after a while. But if the sensor fails or has signfiicant changes to its gain, compression curve, or time delay, all bets are off. Problems would present symptomatically or at a BG check - and these presentations could be magnified for the worse if the sensor is reading the opposite of the actual BG trend under "fully automated" control.

    This all comes down to one thing: The sensor. Until a sensor is developed that has repeatable and reliable transfer gain and phase delay, closed loop control (whether semi-automatic or fully automatic) will not work in realistic, large-scale clinical settings.

    So my question for you comes down to one thing as well: Of all of the money, time, and effort being invested into the partnerships with J&J, Dexcom, and BD, is anything being done to develop a new sensor that will be suitable for closed loop control? If so, have the issues listed above been considered in the new sensor specification, and has a technology even yet been identified or tested that can provide for such a sensor?


    The answer to this question is the key to our understanding of whether a FDA approval of closed loop system is on the 4-year horizon.
     
  9. SarahKelly

    SarahKelly Approved members

    Joined:
    Nov 14, 2009
    Messages:
    1,147
    so...not reading back, but wanted to add that this made me cry. No, it's not a cure. Makes me hopeful though that they're not stopping at one level of treatment and that they'll hopefully keep making treatment easier as they're also researching better ways to prevent and alleviate the disease. I have two family members that have T1D and it would be an amazing gift to have them both able to just live without anymore fingerpricks, shots, infusion sites, CGMS...anything...an absolute dream. I am hopeful.
     
  10. OSUMom

    OSUMom Approved members

    Joined:
    Sep 10, 2006
    Messages:
    5,324
    Twelve families recruited by the INsulin PUmp Therapy group (INPUT)

    CONCLUSIONS: Selected parents of children with T1D who are already insulin pump users...

    Sarah the survey sample set selection is not representative of the entire community of parents either. Shouldn't the study title be "Some parental attitudes towards overnight closed-loop glucose control in children with type 1 diabetes"?
     
  11. BrokenPancreas

    BrokenPancreas Banned

    Joined:
    Aug 9, 2009
    Messages:
    1,517

    HAHA! Love your response!!:D
     
  12. My_Dana

    My_Dana Approved members

    Joined:
    Apr 25, 2006
    Messages:
    596
    Nice description.
    You are correct in that the feedback of the control loop (the sensor) needs to be much better. Consistent most of all.
    Inaccuracies can be compensated such as non-linearities as long they are repeatable.

    One thing that comes to mind on the control loop reliability/stability is loop update time.
    If we can get sensor technology that is in the 'seconds' range (still quite slow in the controls world) rather than 'minutes' it could be
    possible for the feedback algorithm to ignore an unrealistic number.
    So if I can read my BG every second, even if the the sensor gives a wacky erroneous value,
    the algorithm would say "that is not a possible reading based on the past
    trending data, therefore I will not act on it." (also called digital filtering).
    A reading every 5min would not allow this as it is quite possible to have a wacky number in 5min from now.
    Any AP (if that's what they want to call it) needs to act like the real thing, and that means responding to BG changes in seconds.

    And while I'm at it, a true AP should also be able to administer glucagon as well.
     
  13. Darryl

    Darryl Approved members

    Joined:
    May 8, 2008
    Messages:
    4,313
    Ed,

    You are correct on all of these points. To clarify -

    Sensor update time is not the limiting factor, it is the net pipeline delay (sensor time delay plus insulin action time delay) that are the primary limiting factors. The total pipeline delay from sensing (~30 minutes) plus insulin action time (~2 hours on average for humalog) translates to a minimum loop response time of 2.5 hours. Under closed loop control, this implies a minimum round-trip response time of 5 to 7.5 hours minimum to adjust to real-time glucose levels without significant overshoot. A worst case yet probably scenario would be glucose level changing in a roller-coaster fashion (such as post-prandial) with a 2.5 hour cyclical rate, in which case the closed loop would never be able to keep up. Honestly I have no idea how the AP designers are going to deal with post-prandials unless it is guaranteed that meals are eaten at least 5 hours apart. With a net 2.5 hours of pipeline time delay, closed loop control will do nothing but sit still if glucose levels vary significantly every hour or two due to meals. Which they will. Anyway back to the question of update time, even if the sensor update time was seconds instead of minutes, it would not make a singificant change to the net the pipeline delay unless insulin also acted in a matter of seconds as well.

    You are also correct that a positive steering function (i.e., glucagon) is typical in closed loop systems. Most closed loop systems do have balanced control in both directions. And to call the AP "fully automated" you would certainly need glucagon, or a feeding tube connected to a Pez dispensor. However from a technical perspective the loop can function with unbalanced steering as long as the algorithm is aware of the differences in upward and downward control, much in the way the cruise control in a car can function in a stable manner even though powered accelleration is more responsive than passive decelleration.

    I hope that either Aaron or the J&J moderator will take some of these questions to the AP engineers and fill us in on how these issues are going to be addressed... because looking at this from an engineering perspective, I can't see it no matter how hard I try.
     
    Last edited: Jan 21, 2010
  14. My_Dana

    My_Dana Approved members

    Joined:
    Apr 25, 2006
    Messages:
    596
    Good points, Darryl.
    The big issue l thought about but left out is the insulin response time - you mention as pipeline delay.
    In the real pancreas, we get insulin response in a few minutes when BG rise is detected.
    So the AP needs armed with a super fast acting 'flat' insulin with similar response speed.
     
  15. Darryl

    Darryl Approved members

    Joined:
    May 8, 2008
    Messages:
    4,313

    That's exactly right. Non-D people enjoy closed loop BG control because glucose sensing is instantaneous, insulin action time is 5 minutes, there are compensatory processes in both directions (insulin and glucagon), and reliability of all processes is 100%.
     
  16. wilf

    wilf Approved members

    Joined:
    Aug 27, 2007
    Messages:
    9,650
    If someone comes up with a closed loop system that works, even if it is just usable at night at first (because nights are less challenging), then you'll have a line-up down the street and around the corner - and never mind doing silly statistically meaningless surveys of whether this or that group might trust it.

    The bottom line is that it has to work. Until there is something, anything, available that works I'll happily go on with the MDI.
     
  17. sarahspins

    sarahspins Approved members

    Joined:
    May 5, 2009
    Messages:
    2,205
    Actually there is more than just glucagon and insulin involved, but I really really like this... I think the reliability of all my "processes" is about 0% :p

    wilf - I love how you say nights are less challenging. Again, reliability of my processes: 0%!!! :D Case in point.. we're all different :eek:
     
  18. Toni

    Toni Banned

    Joined:
    Sep 14, 2009
    Messages:
    2,882
    We have used both the Minimed and Dexcom cgms. That group of 12 parents should educate themselves further. I would definitely welcome computer analysis of basals in the absence of food. I would like daily computer analysis and suggestions from the computer regarding D management. I would like to be apprised by the computer of suggested overnight changes at 12 midnight and 2:30am and accept or deny the changes. Or be able to turn off automation on a night I could not stay up. I cannot trust the sensors; that is the biggest issue.
     
  19. CrystalK

    CrystalK Approved members

    Joined:
    Jul 25, 2008
    Messages:
    159
    Here's our chance!! The JDRF Advocacy page on Facebook is taking questions for Dr. Kowalski which he will answer in a live streaming video event on the APP on Wed at noon.

    Fire away CWD parents!!!
     
  20. Sarah Maddie's Mom

    Sarah Maddie's Mom Approved members

    Joined:
    Sep 23, 2007
    Messages:
    12,510

Share This Page

advertisement