Slashdot Mirror
Artificial Pancreas Shows Promise In Diabetes Test
An anonymous reader writes A cure for Type 1 diabetes is still far from sight, but new research suggests an artificial "bionic pancreas" holds promise for making it much more easily manageable. From the article: "Currently about one-third of people with Type 1 diabetes rely on insulin pumps to regulate blood sugar. They eliminate the need for injections and can be programmed to mimic the natural release of insulin by dispensing small doses regularly. But these pumps do not automatically adjust to the patient's variable insulin needs, and they do not dispense glucagon. The new device, described in a report in The New England Journal of Medicine, dispenses both hormones, and it does so with little intervention from the patient."
While I appreciate healthy skepticism about medical advances, "google it" is a phrase I associate more with pseudoskepticism than the real deal.
Why are we screwing around with artificial organs when we can have the real deal?
Yep. I'm a *big* fan of my insulin pump, but the included "Constant Glucose Monitoring" device that the pump's company touts as "FDA-approved artificial pancreas!!one11eleven!!" is anything but. It's measuring interstitial fluid, which is *randomly* accurate in *random* people at *random* times. It can neither be trusted, nor should be. And I've since stopped using the CGM side of the pump.
The tech that they're talking about in this article is the same idea: measure interstitial fluid and make insulin decisions based on that. Bad. Ju. Ju.
We need some way of measuring blood glucose levels from, ya know, actual *blood*, without the risk of causing infections. Until we get that, no bueno. Just pass on it.
Jason Van Patten
FTFA:
The system consists of an iPhone 4S with an attached glucose monitoring device, two pumps, and reservoirs for insulin and glucagon.
How is this in any way an improvement over the continuous monitoring setup that multiple pump suppliers already offer? All you're doing is adding an additional pump with glucagon. Nothing about this even remotely resembles what anyone would fairly call an "artificial pancreas."
Click bait. Move along.
Dude's kid is a certified loser. At 15, with 14 years of T1 diabetes, you should be able to remember your fucking insulin pump. Kudos to the father for doing the work but, damn, that's one helicopter parent who has seriously removed all sense of responsibility from his son's upbringing.
To be successful, this kind of a device will need substantial improvements in Continuous Glucose Monitoring (CGM) devices. I used one of these 2 to 3 years ago, and it required a finger-stick reading to "calibrate" it at a minimum once every 12 hours, but recommended 4 times a day. Even with this calibration, the algorithm in their software didn't adjust to this as truth data, and would continue to read quite different values. Many times this was in the 60-80 point (mg/dL) range. When you're trying to control blood glucose into a range of 80-120 mg/dL, having an error so great is a significant challenge. Granted, this was likely 1 generation old technology, but from what my endocrinologist (who's also a pump wearing diabetic) tells me, the newest generation isn't much better.
I can't imagine what the device would do when you factor this error in along with the algorithm trying to account for situations such as eating, without having additional input from the user.
Oh, and one last hurdle: A newly placed sensor for the CGM devices generally take a period of 1 to 2 hours to acclimate, then need a "calibration", before the data is useful. What does a diabetic do during this time period (which needs to occur once every 3 days)?
In a very long list titled - Stupid Things I Have Done ---- ignoring my doctors warnings on my blood sugar levels ranks right around #1. To be sure at this time we had really bad family things going on along with the usual work BS and financial issues. But once you cross that line there is no going back to a normal life.
Pretty much if you are facing pre-diabetes this is what you must do:
- Stop eating fast food and high carb foods
- Drop around 40lbs
- start moderate exercise
The 40lbs sounds hard and it is - but with a healthy(er) diet and modest exercise the weight will come off. My doctor was not quite this clear or blunt with me but I wish he had.been.
Over the past four decades, we've seen squat in the form of treatment for diabetes other than improving the delivery of insulin delivery for diabetics, which has been around since the 1920s. Honestly, it almost seems as if the insulin market is just too lucrative to allow a real cure for Type 1 diabetes. We march on continuing to watch little children struggle with this disease through adulthood and often succumb to an early death because of it. C'mon scientific community. Get your collective heads our of your arses and curse this.
And those marketeering people need paid, along with the other medical middlemen. A diabetic needs insulin, most advances of the last 20 years make it easier to dispense, but not cheaper. I can't afford insulin this month, but yay! there's a new pump on the way. Modern insulin is fermented from yeast like beer, so why is it 1000 times more expensive than beer? Because manufacturing cost savings get passed along to the stock holders not the patient.
see Hicks - go kill yourselves...
Doughnuts aren't going to eat the selves.
The Kruger Dunning explains most post on
Let me count the issues here:
1. This device seems to "do a bit better" than conventional treatments. How much better? A lot or almost none at all?
3. When you eat - it can take (minimum 20 minutes, maximum much longer) for the carbohydrates you eat to be broken down into glucose, detectable by a CGM. This can be MUCH longer for fatty foods which can often result in the liver secreting Glucose. Commercially available insulin can take up to 2 ours to reach peak affect. This means that by the time you eat and your CGM begins to notice it - it is too late to take any meaningful affect and keep your blood sugar under reasonabily control (for the next several ours).
4. There are devices now (by Medtronic) that will shut OFF your Insulin supply if your CGM says your blood glucose is too low - but aside from problems with poor CGM readings, this could be too late. (Furthermore, it's a minor firmware tweak on an existing pump). 5. There have been other project out there for years in which pumps can inject glucogon when BG levels are low. In fact, I credit my 8 year-old daughter for first coming up with the idea a few years ago - at least that't the first time that I personally heard it! Either way - no novily there.
So in short - nothing spectacular here, but I bet if they made a snazzy "solar roadways" type video, made it on a 3D printer and accepted BitCoins for payments, they'd monopolize the front-page headlines for weeks to come!
Ditto. I'm a type 1 who has used a pump for the past 7 years. I tried the CGM device for a few months about two years ago, and was really disappointed. The readings were widely inaccurate (sometimes over 100 mg/dl). I also didn't see much point in it if I still had to manually check my blood sugar levels at least 4 times a day to calibrate it. Having an additional piece of equipment stuck in your body all day was also another turn-off.
But the biggest downside? The $35 that each sensor cost out-of-pocket after my insurance fees. When these need to be changed every 3-6 days, that adds up pretty quickly.
I believe that these devices will eventually reach the point of convenience where you'll seldom need to think about type 1 diabetes (outside of filling the pump and changing sensors), but the price is a huge barrier to entry.
If you post as Anonymous Coward, don't expect a reply.
Funny, I was about to write the same thing. My wife has a pump + a continuous monitor, and her experience is just like yours.
It's an interesting idea, but the implementation isn't quite there yet.
My daughter has been using the DexcomG4 for a year (off and on) and it is more accurate then 60-80. I've been very pleased with the results. Early models were certainly used for trends rather then reads. I look forward to the joining of CGM and pump with limits (just like pumps have now). Now if they could just fix the adhesive. Too many hours in the pool and the tape starts to peel off.
I currently have the CGM and I understand some of what you are saying but I believe things have improved a lot since then. Currently my CGM gives me a calibration within about 10-15% of my blood glucose. This is normally good enough to monitor trends and keep my blood glucose in the proper range without finger pricks. The one huge advantage of it is the fact that when I am hungry I can look at it and see if I am hungry because of low blood sugar or just being hungry.
About your complaints with the CGM, I have seen numbers jump to be about a 100 point difference, but this is normally followed by a calibration failure. This causes me to have to recheck my blood glucose (I agree this is annoying, but at least it recognizes it is off). Currently the sensor lasts 6 days (I can normally get it to go a 7th day before the battery dies). I still normally test my blood glucose 3-4 times a day, but that is better than my previous 8-12 times a day.
I think I do better with the CGM, but then again I am just happy not to have to go back to using constant needle injections and finger pricks.
When you say "the CGM device", which one do you mean? There are several, and most especially they've *all* become more accurate over the last couple of years. The one that my son is wearing, the Dexcom G4 (I have ZERO affilication with them except that my son uses one), is leaps and bounds more accurate than anything you could've been using two years ago. I don't say that to invalidate your experience, but technology has progressed in the interim. We rarely see inaccuracies that aren't explained by the time lag between blood glucose levels and interstitial fluid levels (which is actually what a CGM measures), and calibration is usually only required every other day. That said, it's still important for people with T1D to test their BG at every meal due to the physiological limitations of CGMs.
For us, the most important aspect of CGM usage is the ability to see trends during the day and based on activity levels. It's huge to know in advance that my son's BG is trending toward low *before* it happens and give him additional carbohydrates to prevent him from doing so. This, in effect, is the information that the 'bionic pancreas' research is basing its dosage of insulin and glucagon (which, in turn, is essentially how a non-T1D pancreas works).
Based on your pump + cgm statement, I'm going to assume you have one of the generations of Medtronic devices. The CGM functionality in these, even the newer 530G with Enlite sensors, simply isn't very good at all. If you ask any of the (credible) researchers in the field about them they will either stay entirely silent so as not to piss Medtronic off or they will drop enough data on you about them to scare you off from CGMs generally forever. They simply aren't very good. I do contest any assertions that interstitial fluid can't be trusted, however. It's not a trivial thing to do right (or Medtronic and everybody else would've made it work a long time ago), but when done right the results really are quite useful.
All I ask is that you don't judge the entirety of an approach based on your experience with one flawed implementation.
It seems as though the big problem with this technology is that it's not measuring blood directly. What are the barriers to placing a sensor more-or-less permanantly inside the body that can test blood directly and the send, via radio or whatever, commands to an external insulin pump to dispense insulin?
I'm guessing "blood clots" is the problem here, but I don't know.
...but it's being eaten...by some...Linux or something...
I had a Medtronic Paradigm. My pump is made by Medtronic (which I'm very happy with), so this CGM was designed to be used in conjunction with it.
Thanks for the advice, you're the second person in this thread to recommend the Dexcom. Looks like I will need to talk to my endocronologist :-).
If you post as Anonymous Coward, don't expect a reply.
Why do you assume the finger stick is "truth?" It also has a standard deviation. You don't usually do many finger sticks in a row, and therefore don't notice. But if you read the literature, you'll find it does. Further, if you burn some test strips doing a bunch of finger sticks in a row, on different fingers and hands, you'll find it's not just a disclaimer. It really does deviate quite a bit, as expected.
Their "algorithm" doesn't take your stated "truth" as truth, because it's not truth. Which is correct.
Does their "algorithm" do the right thing? Or a good thing? Or a better thing? Well, there are studies you should read to determine that. But your stated criticism is false.
I usually don't respond to AC's but what you are saying is absolutely true. My experience is completely anecdotal, but when the CGM would show a fluctuation of 100 within an hour and the test strips show a deviation of 10 during the same time frame when checked every 15 minutes, it definitely made me question the CGM.
I'd love to read some studies about the accuracy of different brands of test strips and CGM devices, as long as they weren't tainted by the manufactures and vendors of said devices.
If you post as Anonymous Coward, don't expect a reply.
]but when done right the results really are quite useful.
All I ask is that you don't judge the entirety of an approach based on your experience with one flawed implementation.
Show me it "done right" with years(!) of lab evidence, trials including hundreds (if not thousands) of individuals, and perhaps I'll believe you. Oh, and when you provide said data, don't be an "Anonymous Coward" about it, either.
No, blood tests aren't 100% accurate. They are, however, a far, *FAR* more accurate way to get an idea of levels than using interstitial fluid. And, as it turns out in this case: accuracy counts. A lot.
Jason Van Patten
for years, we have been pouring billions (literaly) into stem cells, without a whole lot to show for it
A few tens of millions, and a bionic pancreas is nearing usability
tell me again why the bandwagon for stem cells
I'm guessing you're talking about the medtronic CGM/pump... the CGM is worthless (plus the sensor is frankly a harpoon). Dexcom's recent models however have been pleasantly accurate- in particular, I've gotten quite good readings via using the arms for it.
Why? Data is either valid or it isn't; whether he has a cute little username at the top of his post has no bearing on that.
The artifical pancreas has been "just around the corner since before I became diabetic, 40 years ago. There has been no notable advancement except to make the electronics smaller, and to make insulin more pure. The sensors have *never, never, never* worked reliably, they always involve consuming chemical reagents or just don't *work*. I participated in artifical pancreas research several times, as subject and later as investigator.
Fortunately, a genuine "cure" for most Type 1 diabetes is gathering funding for Phase 2 of human trials at Mass. General Hospital, in Dr. Faustmann's lab.
http://www.faustmanlab.org/
The BCG vaccine, used worldwide for tuburculois for millions of people, is applied in small daily doses for one month while the diabetic maintains strict blood sugar control. This allows adult stem cells, not implanted tissue, to transform to insulin producing cells and cure the diabetes. This is the first fundamentally new treatment for Type 1 diabetes, one that's actually worked in test subjects, since insulin was refined and tested.
This is confusing: From my diabetic colleague: glucagon is dispensed in response to low blood sugar by the alpha cells of the pancreas, which apparently remain intact, not by the destroyed beta cells that are missing form the pancreas. If the diabetes is being treated well with insulin, why wouldn't the patient's normal glucagon response work well?
From my colleague reading over my shoulder: many diabetics lose their glucagon sensitivity, but apparently due to overall blood sugar control. They still have the relevant alpha cells, and my colleague would expect the glucagon sensitivity to recover with otherwise good diabetes control from manipulating the insulin alone..
The developers tested the device over five days in two groups of patients, 20 adults and 32 adolescents, comparing the results with readings obtained with conventional insulin pumps that the participants were using.
The artificial pancreas performed better than the conventional pump on several measures. Among the adolescents, the average number of interventions for hypoglycemia was 0.8 a day with the experimental pump, compared with 1.6 a day with the insulin pumps. Among adults, the device significantly reduced the amount of time that glucose levels fell too low.
http://www.nejm.org/doi/full/10.1056/NEJMoa1314474#Results=&t=articleResults
Mine is extremely accurate when I'm at my desk working or sleeping, usually within 1 or 2 points (mg/dL).
The trouble is when I am active. When I play basketball, the lag between blood sugar and the interstitial fluid the CGM monitors can be extreme, like my blood sugar could be 40 when the CGM thinks it's still 120. The other problems is that in cases like this, I'm not sure how fast either insulin or glucagon can be injected. I mean, it's usually easy to deal with if I play moderately intense sports like tennis or racquetball, but full court basketball or weight lifting can lower my sugar so fast, I even don't always get adjusting for it (by lowering my basal rate and eating proteins and carbs ahead and keeping my basal lower for the next 24 hours) right, even after 36 years of practice.
I've been using the medtronic CGM & pump for four years, and going to insulin pump support group meetings for slightly longer. The reps have been pretty honest about the CGMs being neither accurate nor precise, as one would expect from a system that's calibrated via another inaccurate system. Still, most of the group most of the time gets at least enough accuracy to determine the direction glucose is trending or if it's stable. Everyone agrees that the Dexcom CGMS beats Medtronic's CGMS on most metrics except maybe for pump integration. It seems like there's a ton of different ways to put the sensors in that affect accuracy too. Some people have scarred up areas that get bad readings, some have to change the angle of insertion to get better results, and a few of us put in the new sensor a few hours to a day before the old sensor is done because the early hours of sensor use seem less accurate. Almost everyone seems to be using the sensors at least six days and some up to three weeks. For me, ten sensors can last about 3 months.
JVP, I'm not the same person as the AC, but I have been on both MiniMed and Decom CGMs. I've also done some CGM medical trials for MiniMed.
Without a doubt, I can agree with you that MiniMed CGMs absolutely suck and that the "artificial pancreas" marketing from MiniMed is crap. I used the MiniMed CGM on my pump for about 2 years and it was often way off my actual blood sugar. I talked with MiniMed reps several times and they would tell me the same crap: don't calibrate when your blood sugar is rising / dropping, don't calibrate more often than every 8 hours, etc., etc. I stuck with it mainly because of the convenience factor of having only one device to carry around.
My doctor convinced me to try a Dexcom because of the issues I was having, and I can tell you that it's a world of difference. The Dexcom has been so damn accurate, and a hell of a lot more comfortable than any of the MiniMed CGMs I have used. I can let it run past calibration (>12 hours) and it still gives readings. When I go to calibrate it (even if it's been >12 hours) it's very rarely off by more than 20 mg/dL unless my blood sugar is shooting up like crazy.
Maybe you have tried it, and maybe it didn't work for you.. but don't write off all CGMs. They do work. The only thing that sort of sucks is having to carry around another device.. For what it's worth, last time I saw the study in TFA, they were using a Dexcom CGM.
Yep. I'm a *big* fan of my insulin pump, but the included "Constant Glucose Monitoring" device that the pump's company touts as "FDA-approved artificial pancreas!!one11eleven!!" is anything but. It's measuring interstitial fluid, which is *randomly* accurate in *random* people at *random* times. It can neither be trusted, nor should be. And I've since stopped using the CGM side of the pump.
The tech that they're talking about in this article is the same idea: measure interstitial fluid and make insulin decisions based on that. Bad. Ju. Ju.
We need some way of measuring blood glucose levels from, ya know, actual *blood*, without the risk of causing infections. Until we get that, no bueno. Just pass on it.
I concur ... even the pumps are failure prone even if it is hidden in the press when they kill people. So, how can you trust two electronic devices and a scientific algorithm PLUS $160 a day glucagon (stupid stupid stupid) that takes control out of our hands.