A (8:01)

JOHN so, Neil, I find this to be always kind of very fascinating. So kind of this pancreatitis risk. So the GLP1s come on the scene, the DP4s come on the scene, and this talk of pancreatitis, and if you looked in the package insert, it would talk about caution in someone who ever had pancreatitis or caution for someone at risk for pancreatitis. Well, who's at risk for pancreatitis? People with diabetes. So if you look at multiple papers over the last 20 years, the hazard ratio roughly for acute pancreatitis in someone who has diabetes versus someone who doesn't have diabetes is about 1.74. So just having diabetes puts you at increased risk for pancreatitis. So the question is, and you know, and this paper really doesn't support this, really comparing, you know, the DPP4s to the SGLT2s, but maybe they all do a little bit, right? And you're comparing it, you're not necessarily, there's not a control group there in this study. So, you know, we don't completely know that someone with diabetes who is on none of those drugs, someone who's on insulin or whatever, what would that kind of risk be? So I would feel okay using these medicines in someone at risk for pancreatitis, because I would use these in someone who has diabetes who has a 75% higher risk than someone without diabetes in these studies. So I would feel good in that risk. The person with a history of pancreatitis kind of depends on what it was. Right? Did someone have a necrotizing pancreatitis that was in the ICU and was on event and all this other things, or someone who had some abdominal pain and had a lipase that was 15 points above normal and had a, you know, a pretty normal looking CAT scan? I'm not sure I would make that a contraindication forever. Although in our EMR it's going to say that person had acute pancreatitis and it's probably going to pop something up. So I think that that's interesting. The increased case of biliary disease, one per thousand patient years doesn't seem like a whole lot, right? So you know, you have 100 patients who are on these medicines for 10 years. One more of them will have, you know, some biliary disease which, you know, for the most part doesn't tend to have horrible end of life kind of consequences and things like that. So I would be reassured if you look in some of the older articles though around pancreatitis, it might say that people with pancreatitis do a little better with insulin and in this, in this with the biliary disease, maybe it's a pinch better, 15% lower and some someone an SGLT2 and someone with some biliary disease. So I think this should be kind of reassuring on both fronts. Our next article is from jama, Open Network and Diabetes and Endocrinology and looked at continuous glucose monitoring, frequency and glycemic control in patients with type 2 diabetes. So this study was to look at the association of CGM frequency with glycemic status over a 12 month period versus no CGM use. So it was a retrospective propensity scored matched cross sectional study that used Optum defined market clarity data obtained between January 1st, 2019 and December 31st, 2023 including data from 6 months prior to each participant's index date with 12 months follow up. They were participants with type 2 diabetes who were over the age of 18 and had a hemoglobin A1C between 7 and 15 at baseline. So they looked at the number of days using the CGM during the 12 month post index period and the frequency was 1 to greater to 1 than less than 90 days. Frequency 2 greater than 90 to less than 180 days. Frequency 3180 to 270, frequency 4 greater than 270 and then there was the control group. The primary outcome was change in A1C relative to the frequency of CGM versus no CGM. So the analysis included over 9200 patients of which 4200 were female. The mean age was 55.9 years and there were 4600 plus patients in the control group. There were 1,081 in frequency 1, 1523 in frequency 2, 2540 in frequency 3 and 2485 in frequency 4. High CGM use, which would be frequency 4 was associated with greater reduction in A1C, a decrease in 1.52 versus no CGM use that also had a decrease but 0.63 percentage points. CGM users experienced the greatest reduction approximately 3 months, so frequency 1 had a decrease of 0.5 percentage points, frequency 2 decrease in 0.57, frequency 3 0.79 and frequency 4 0.91 compared with control patients. No further glycemic improvement was observed with frequency in 2 and 3 after 6 months. Improvements in patients with frequency 1 and 4 groups were sustained for the duration of the post index period. The addition of a GLP1 receptor agonist in the frequency 4 group was associated with an A1C treatment difference of a decrease in 1.13 percentage points versus controls at approximately 12 months Neil John, this.

B (13:54)

Is an interesting study in that not that it's surprising, but I think it is interesting and basically what does it tell us? It tells us that using CGM matters. It helps people get better control and unlike what some of us, I think sometimes think that if you use it now and then so you get motivated, understand your diabetes, that's enough. What this says is that continuous use, using it for much of the year is better than using it now and then and after that it's speculation. Why does it help? Well, I think the two reasons are 1 it allows more regular changes of medicine based on regular updating of data. You don't have to wait that three months or more for your A1C to see a sufficient change to change a medicine and two it creates an internal feedback loop that patients get when they eat wrong. They see the effect very quickly on their blood sugar. If they exercise, they see that blood sugar going down and we all respond to natural feedback that is close in time to the behavior that caused that feedback to occur. And we've talked about that on other podcasts. So take home point here. CGM Good Worth using more Next article from JAMA is titled An AI Powered Lifestyle Intervention versus Human Coaching in the Diabetes Prevention Program A randomized clinical trial oh boy, I know who I was rooting for in this one. So we know prediabetes is common. We know the Diabetes Prevention Program is one of the great trials of all time showing lifestyle intervention decreases progression from prediabetes to diabetes. But it is a resource intensive program. So determine whether referral to an exclusively artificial intelligence AI led lifestyle intervention based on the DPP4 program was non inferior to referral to a human led DPP in achieving recommended thresholds for weight loss and a 1C reduction as well as weekly physical activity among adults with prediabetes and overweight or obesity. The authors developed a randomized trial. It was a phase three parallel group pragmatic non inferiority randomized clinical trial conducted from 10-1-21 to 12-16-24 at two US clinical sites. That was in Baltimore and reading it is worth talking a moment about the interventions to which participants were randomized in a one to one ratio. They either got referral to an AI AI powered DPP4 lifestyle intervention delivered by a mobile app with Bluetooth enabled digital scale and they also used some step counters or referral to a human coach led lifestyle intervention. It's important to understand that human coach was delivered remotely. This was started during part of the COVID pandemic and the decision was made to deliver the lifestyle interventions in a remote manner. It's worth discussing what the AI based program looked like because we're not real familiar with that sort of thing. The participants received a digital health kit within 8 to 12 days of signing up. It had a Bluetooth scale that connected with an app and what was fascinating here is the app delivered personalized push notifications for weight management, physical activity and nutrition and it was was informed by the data that it collected, meal logging, weight measurements as well as things like geolocation which was fascinating that was tracked through the phone. So for instance if you were walking into the supermarket, you would get a text saying let's make some good choices while we're shopping here. And those texts would give you the choices based on what you're eating has been so that it was personalized to you. If you hadn't for instance exercised on a day where you had planned to, you would get a text saying might be a good idea to exercise today. And even if you can't do your full exercise program, even if you do part of that program, it's a win. So you have a sense of what they were doing. A total of 368 participants were included, 71% were female the median BMI was 32. After referral, 93% of the people initiated the AI led program and 82% initiated the human led program. So AI was winning right out of the starting gate. The primary outcome was achieved by the 31% in the AI group and 31% in the human group meeting the criteria for non inferiority. They were about equal in their outcomes. Oh boy, John.

A (19:38)

You know, Neil, not surprising really. So in March of this year article in the New England Journal looking at psychotherapy, right? If someone could be referred to therabot and could get kind of AI driven psychotherapy, and it showed, it did. Okay, so, but some of the caveats kind of written up in a recent edition of Psychology Today about that the AI did better with protocol driven type things. So the, you know, the undifferentiated and you and I kind of talk about this, you know, in, in primary care is really that seasoned clinician. Really what differentiates them from a protocol driven system is someone who can, you know, who can kind of take that undifferentiated patient and kind of unpack it. So the DPP for coaching I think is probably pretty structured, right? If, if A, then B, if B, then C. And you know, instead of kind of viewing this as the boogeyman, maybe this becomes kind of a great kind of tool for us, for our patients, which is multitudes who kind of fall in that. Boy, I need to eat a little better, I need to exercise a little bit more. And maybe if I got something to my phone, I would be. Because people are doing an awful lot of stupid stuff based on what's coming across their phone. Maybe something coming smart and helpful coming across their phone. Wouldn't that be revolutionary instead of another TikTok or whatever people are wasting. So I find this to be interesting. I don't think I would want this to kind of replace kind of everything else, but I think instituting a protocol, and we're on a diabetes podcast right now, you could imagine that in the years ahead that insulin dosing recommendations will go through some artificial intelligence. You could imagine that warfarin dosing and kind of something that kind of loops in with, with some artificial intelligence to tell people to, you know, adjust their dose things. You know, we just talked about congestive heart failure. If someone's weight goes up. You could see a lot of this stuff that is very protocol driven kind of using this AI technology. So I, I don't find it surprising. I find it interesting. Came from diabetes and obesity and metabolism. And it looked at two tier screening approach for liver fibrosis stratification in outpatients with type 2 diabetes A multicenter cross sectional study so this study retrospectively enrolled over 1200 older Italian outpatients who had type 2 diabetes who underwent a vibration controlled transient elastography which we will refer to as VCTE with liver stiffness measurement and controlled attenuation parameter. People with metabolic liver disease was defined as having a cap greater than 248, significant liver fibrosis was defined as having greater than 8kPa, compensated advanced chronic liver disease as liver stiffness measurement greater than 10 and clinically significant portal hypertension, that number being greater than 25k, pa or LSM greater than 20 and platelet count under 150. A FIB index was calculated in all participants. The prevalence of having metabolic dysfunction associated liver disease, significant liver fibrosis, chronic advanced chronic liver disease or chronic systemic portal hypertension were 71.3, 21.1, 11.7 and 1.7 respectively. A two tiered screening strategy for fibrosis using the Fib4 index and the VCTE showed that among patients with a normal FIB4 index, 83.3 had stiffness under 8 and 126 had it greater than 8. Sensitivity, specificity, negative and positive predict value of the FIB4 index for detecting a level greater than 8 was 50.4, 66.3, 83.3 and 28.6 respectively. Increased body weight adjusted odds ratio of 3.34 and elevated AST levels adjusted odor ratio 1.54 were the strongest predictors of significant liver fibrosis.

B (24:30)

Neil John, let me put this article into context. So this, the, the authors said that this supported the guideline approach and the guideline approach is that we should be screening everyone with diabetes and obesity with a fib 4 and if they have a score less than 1.3 they fall out of the need to do any further evaluation. The issue though that this study really nicely delineates is that the FIB4 doesn't have anywhere near perfect negative predictive value. In fact, that would miss about 17% of people with significant liver fibrosis. And I just want to focus on that particular issue because clinically that's what we're looking for now people with F2 and F3 fibrosis, because those are the people who are eligible for the FDA approved treatments for mash being rosmetiron and semaglutide. At this point there are plenty more in the pipeline but it means that you be missing about 17% of the people who have MASH that would be eligible for treatment if you did nothing in addition to the fib 4. And I find that a bit concerning and the reason for that in this study, the prevalence of MASLD and MASH was the same as we see in other studies. 71% of the people had mastled. About 21% had significant liver fibrosis. So what do we do with this information? Because we wouldn't want to jump and do elastography on everybody. So what I take from, from this and here I'm using the part that you said last that people who had particularly increased body weight were at higher risk of fibrosis. Same with high alt levels. Well, we do the fib 4, but we don't adhere to it religiously. We then think, and if someone has class 3 obesity or elevated LFT's and they have a fib 44 less than 1.3, maybe we go on and do that transient elastography as well. The other group I worry about are people who are younger because remember, the FIB4 is driven by age. You get a higher number, the higher your age is in a way that's very similar to our ASCVD calculator. And if nothing else changes, you're a decade older, your score goes up. So if someone has a score near that borderline score and they were say 30 or 40, I might go ahead and check them anyway. I think this is an important article and will likely influence future guidelines. Our final article of this issue is from BMJ Open Diabetes Research and Care and is titled Change in Urine Albumin to Cranin Ratio and Clinical Outcomes in Patients with Chronic Kidney Disease and and type 2 diabetes. The study aimed to investigate the association between the change in urine albumin and creatinine ratio and clinical outcomes in patients with chronic kidney disease and type 2 diabetes. They looked at adult patients with an elevated UACR greater than 30 in initial testing after the diagnosis of type 2 diabetes and CKD and it was identified from an Optum electronic health records database. We're hearing about this data set a lot because it is one that encompasses somewhere around 70 million people's information. A UACR change from initial to last test from 6 to 24 months was categorized as either a greater than 30% decrease stable or a greater than 30% increase in UACR. And what they found was that compared with patients who had a stable UACR, there were 35,000 of them. Those who had greater than a 30% decrease in their UACR, 89,000 people had a lower risk of all cause mortality, 7% lower risk and a lower risk of composite cardiovascular outcomes as well as CKD progression. And patients who had an increase UACR of greater than 30% had a higher risk of each of those endpoints.

A (29:12)

John so there was a, there are a lot of things methodologically I liked about this study and starting off that better or worse had to be more than 30%. So you know, these, these changes that are just a couple points that would be say, well, it's better or it's worse. I kind of liked that. It really defined better or worse in a nice way. One of the things though, in this particular study, I think the average time between microabumins was 18 months. Maybe this is real, complete real world data and that's what's going on out there. I think this is just yet another reminder to us as we are going to hear more and more and more and more and more about this kind of cardiorenal link. So if we can make people's kidneys better in the setting of diabetes here, we're probably going to decrease their chance of having cardiovascular disease. Because really if you look at the difference for that person whose microalbumin is 30% worse, but maybe their creatinine has not changed or their GFR hasn't changed, their chance of ending up on dialysis is the relative change in those numbers might not be huge or their need for being on a transplant list or whatever might not be huge, but it clearly accelerates the risk of having cardiovascular disease. And that is what kills most people with renal disease. One of the things this is probably something very easy for electronic records to follow. I think once upon a time, you and I in primary care, we were judged did our patient have a microalbumin. But I really didn't see ever really that kind of judgment. Well, what did you do about that microalbumin? Or maybe if it did, did you start an ACE or did you start an R? And that was it. And I think a lot of people might just check it and not do anything about it. I hope not. But if it is elevated, you know, starting off with putting someone on an ACE or an arb, but pushing the dose of that ACE or ARB to really try to get that 30% reduction and I think that really should be a little bit of our charge. If you have that person with some increased microalbumin and I started an acorn arb, that I can get that down to 30%. If that doesn't work that maybe I'm adding on one of the other categories. I'm adding an SGLT2. I'm adding, you know, one of the MRA type medicines to kind of protect people's kidneys. But I think this is going to be more and more of the challenge for those of us in primary care, for those people who are in that CKD3, that CKD4 to make those differences so someone never becomes CKD5.

B (32:10)

John before we close, I want to share some important news with our listeners. As you know, in addition to diabetes, we've been covering the full range of the cardiometabolic conditions. And so our name is going to be changing to meet that consistent mission. Our podcast will be rebranded as Diabetes, Obesity and Cardiometabolic Update. I know, I know you're already thinking and you're right. The new name is Doc Update. Diabetes, Obesity and Cardiometabolic Update. You'll notice this when the change is made on your subscriber feed will have a new title and branding. Same great information. For more information and links to the articles that we discussed in this issue, just go to diabetesjournal.org until next month. Keep listening and keep learning.

A (33:12)

It.