Akshay B. Jain, MD: Welcome from sunny San Diego, where we are at the 83rd American Diabetes Association Conference. With me is Professor Bruce Perkins from the University of Toronto.
Professor Perkins, you’ve been doing a large amount of work in people with type 1 diabetes in your clinical trials, where you’re looking at the possibility of predicting diabetic ketoacidosis (DKA) episodes. Can you share more about that?
Bruce A. Perkins, MD, MPH: This all came from this idea that people with type 1 diabetes struggle with glycemic control, reaching evidence-based targets, along with many other challenges, but that there are these transformative drugs in type 2 diabetes that work so well. If we have a population struggling on insulin alone, why aren’t we considering adding drugs that can help smooth out blood sugar control?
It came out of mechanistic studies I’ve been doing with collaborators in Toronto years ago that understand mechanisms of sodium-glucose–linked transporter inhibitors, for example, sodium-glucose cotransporter (SGLT) inhibitors. We moved to international, phase 3 clinical trials, and while they have a really nice metabolic benefit, they increase DKA risk. That’s kept them from receiving approvals in North America, though they’ve received approvals in Europe and in Japan. I realize that DKA is a risk in that context, but it’s also not well handled clinically, even in non-SGLT users.
I started a pump program years ago at Toronto General. Everyone who received a pump (because a pump is a risk factor for DKA) would get training to understand what ketones are and what DKA is. Inevitably, a few months later, I’d see someone who had a sick day with gastroenteritis. I’d ask about ketones, and they would ask me, ”What’s a ketone?” It’s hard. People with diabetes have to manage so much information.
In these large-scale, phase 3 trials, some of them involve people doing capillary ketone measurements even on well days. We just wanted to have week by week what’s going on with a fasting ketone. Then of course, if they got an illness, they’d use ketones for sick-day management. I thought why not look at the distribution of ketones on well days so we could understand if there’s a threshold that tells us that someone in the next month would develop DKA.
In other words, I have type 1 diabetes. What if my ketones do this and someone else’s do that? Am I more at risk in some setting where I have another illness to go into DKA?
We could study this in a general type 1 population because this protocol was used in a randomized placebo-controlled trial. The placebo group never got SGLT inhibitors — they thought maybe they did, but they didn’t — and we could see thresholds. It turns out that they predict quite well the risk for DKA — not perfectly, but pretty well.
Then in the SGLT inhibitor users, we thought they’d have a threshold that’s even higher because they tend to have a bit of an increase in ketones. It turns out the prognostic threshold also works in that group, but it’s almost the same.
– There’s like 0.8 mmol in the non-SGLT users and 0.9 mmol in the SGLT users, so we can probably use a common threshold.
This grant that I’m doing is actually a patient-oriented research project, where the final step is going to be co-creating educational tools for people living with diabetes, but the tools will be co-created with people living with diabetes and our team. They’re going to look at existing protocols; they’re going to look at data where we can do something like a weekly ketone, use those strips to predict who might generally be at more risk, and then we can have new education or improve basal insulin delivery, for example, to help prevent DKA.
I think the biggest implication of this is that in development now are continuous ketone meters likely to harmonize with continuous glucose meters. If one is going to be using such a thing, why only get alerted on a sick day?
– Why not get alerted that they’re the kind of person who might be generally at more DKA risk in the next month? That’s the concept behind that research.
Jain: Predictive ketone monitoring. I love that concept. Do you think this will possibly open the way to do more research in heart failure in type 1 using SGLT2s or progression of nephropathy now that we have a way to predict and possibly attenuate the risk?
Perkins: I think so. I was part of an FDA advisory committee promoting a safer dose of an SGLT inhibitor in type 1 diabetes. The advisory committee was certain that we need to have certainty about the approaches not having more DKA risk, but to run a trial like that is like doing postmarketing cardiovascular outcomes safety because the rates are similar.
– DKA is not that common. We’d be doing trials in like 15,000 to 20,000 people with type 1 diabetes to see if there’s safety, not even efficacy. It’s just not going to happen in type 1. That’s not possible in the type 2 world.
Now, we have an agent approved in the US, an SGLT inhibitor for heart failure. It does not exclude people with type 1 diabetes, so this is a start.
Secondly, we will be initiating a SGLT inhibitor trial in people with stage III chronic kidney disease so nephrologists and endocrinologists can learn if they can be used in people with type 1 because they have proven in type 2 to have such a kidney-protective effect.
There are many studies I’m involved in, some using SGLT inhibition to help improve the function of automated insulin delivery, closed-loop systems.
– I think we’re going to have a critical mass of these sorts of studies to help introduce better risk-benefit. I think this may move us forward with the possibility of adjunct drugs, insulin, and type 1.
Jain: Very well said. The future looks promising.
– SGLT2s could potentially be a standard of care for high-risk individuals with type 1. Thanks for joining us, Dr Perkins.
From www.medscape.com
Nyhetsinfo
www red DiabetologNytt