Are better insulins still needed?
Click Here to Manage Email Alerts
Better insulins are absolutely needed.
Many people with type 2 diabetes are only on a basal insulin and are more dependent on noninsulin therapies, and with the advent of additional noninsulin options, such as GLP-1 receptor agonists or SGLT2 inhibitors, their outcomes are improving. There is a community out there that feels that insulins are as good as they can be.
Developing a new insulin takes time and money, and researchers do not know, up front, whether a new insulin will be transformational. That, sadly, is tainting the argument of “maybe we don’t need better insulins.”
It is important to define what a better insulin is. One is speed of action — trying to get as close as possible to pancreatic insulin secretion. Recently, we have seen some advancements with this. Fiasp (rapid-acting insulin injection, Novo Nordisk), for example, is 12 minutes faster than Novolog (insulin aspart, Novo Nordisk) and Humalog (insulin lispro, Eli Lilly), and people with type 1 diabetes will gladly take 12 minutes of improvement. That said, those incremental developments are becoming insulin’s Achilles’ heel. That development may have some clinical relevance for people with type 1 diabetes, but it doesn’t make much of a difference for people with type 2 diabetes. We need to mimic physiology.
The second point is we are doing this wrong, because we don’t have any other choice. With the exception of inhaled rapid-acting mealtime insulin (Afrezza, MannKind), we are giving insulin under the skin, and that is not where insulin is made. That is not where insulin works. The field of liver-targeted insulin somehow hasn’t picked up traction at all. There have been some initial promises and failures that have thwarted innovation. A liver-targeted insulin with intelligent design would resolve many physiologic problems. That would be transformational, and it is a priority for JDRF.
Third, we still need glucose-responsive insulins. Technically, this is very challenging. Can we design a smart-acting insulin that can sense the demand for insulin and then adjust supply — it’s activity — accordingly? We are taking this research on, and we have had some successes in large animal models.
I would posit that people with type 1 diabetes are not able to have physiologic glucose control largely because insulin does not work the way it should work. This is a lifesaving medication that is close to 100 years old. Despite that, there has been such limited innovation. More work remains to be done.
Sanjoy Dutta, PhD, is vice president of research at JDRF. Disclosure: Dutta reports no relevant financial disclosures.
There are no developments needed for any long-acting insulins; however, short-acting insulins still have some limitations.
Long-acting insulins for people with diabetes have been exactly the advancement we hoped for from a pharmacokinetic and pharmacodynamic standpoint. Long-acting insulin works better than insulin glargine 100 (Lantus, Sanofi), it works better than insulin detemir (Levemir, Novo Nordisk) and it works better than NPH insulin. It has a longer duration and flattens curves, with less hypoglycemia in the morning. At the same time, it requires just one injection at any time during the day. That is what we want. From that perspective, we don’t need any further development whatsoever.
Short-acting insulins, however, do have certain limitations. It is impossible for any insulin on the market right now to imitate natural insulin secretion. Natural insulin secretion happens in a biphasic manner. There is an early phase, which takes about 10 to 15 minutes, followed by a late insulin phase. The only way to come close to recreating this is to use a diabetes device that can mimic these phases. Up until now, the available devices have also struggled with this — even hybrid closed-loop insulin delivery systems. They are not designed to be biphasic in any way.
We still have a long way to go with diabetes devices that can really help with optimizing insulin timing. Ideally, it would be great to have a real-time, short-acting insulin for use with an insulin infusion device, so the insulin can adapt quickly and rapidly to a person’s meal and can, in that scenario, simulate early-phase insulin secretion. But this is close to impossible. When you inject subcutaneously, you will always have problems with absorption.
What we really need is a real-time insulin — meaning very quick and very short duration — delivered intraperitoneally via a closed-loop insulin delivery system. This would allow for hepatic clearance of insulin at first pass, providing quicker absorption and superior glycemic control.
Osama Hamdy, MD, PhD, FACE, is medical director of the Obesity Clinical Program at Joslin Diabetes Center and associate professor of medicine at Harvard Medical School. Disclosure: Hamdy reports he serves on advisory committees for AstraZeneca and Sanofi Aventis, is a consultant for Abbott Nutrition and Merck, receives research support from Novo Nordisk and is a shareholder for Healthimation LLC.