I've been blogging on research aimed at curing type-1 diabetes for over 15 years, and in that entire time, rarely have I rushed to blog on a new piece of research. That is because research progresses slowly, and it rarely matters if findings become public this month or next month. However, the results of this research initially felt so good, that I started rushing. However, as I've worked more on the posting, I've become less sure of how important it is. I do think that adults who are in the early part of their honeymoon may want to discuss this with their endocrinology team and consider taking action quickly. The big, open, unknown question is, should teenagers and children do the same?
The drug in question is Semaglutide, which is sold under several trade names: Ozempic, Wegovy and Rybelsus. It was approved for use by people with type-2 diabetes in 2017 and for overweight people in 2021. It is widely prescribed for both of these conditions. Semaglutide is an example of a class of drugs called Glucagon-like peptide-1 (GLP-1) receptor agonists. As a receptor agonists it attaches to the same part of the cell as glucagon. This class includes seven approved drugs, sold under at least twelve different brand names by several "big pharma" companies.
Semaglutide is thought to work in four different ways: beta cells generate more insulin in response to food, the liver releases less sugar, digestion is slowed down, and appetite is suppressed. The last three can be expected to help anyone with T1D, and the first one to help T1D honeymooners. Of course, Semaglutide could have other actions in addition to these.
This was not a randomized, controlled, clinical trial. Rather several endocrinologists knew that Semaglutide had good results in causing beta cells to generate more insulin, and thought this would help their patients, especially those in the early honeymoon phase. So therefore, they chose to prescribe it "off label". After having done this for 10 patients for a year, they went back and reviewed their charts. The data was very positive and they published a correspondence in the New England Journal of Medicine. There was no external funding for this; they spent their own time.
This blog is long because the research has many implications, but also some weaknesses, all of which require discussion.
- What Was Found: The Good News
- What Are The Weaknesses: The Bad News
- What Should Honeymooners Do Now
- Next Research Steps
What Was Found: The Good News
Semaglutide was given to 10 adults (18 years or older: average age 27) during the first 100 days after they were diagnosed with T1D ("stage 3" in Trialnet terminology). This was a weekly injection. They also got standard treatment for T1D in the honeymoon phase. For this practice, that included daily injections of long acting insulin, short acting insulin with meals, and general advice to eat fewer carbs. The key results were (these are all quotes from their publication, which I have edited slightly):
- Mealtime insulin was eliminated in all the patients within 3 months.
- Long acting insulin was eliminated in 7 patients within 6 months.
- These doses were maintained until the end of the 12-month follow-up period.
- The mean A1c level fell to 5.9 at 6 months and to 5.7 at 12 months. [Normal for non-T1D is below 5.7]
- The fasting average C-peptide level increased from 0.65 at diagnosis to 1.05 [Normal for non-T1D is between 0.5 and 2.0.]
- The time-in-range was 89% according to continuous glucose monitoring.
Obviously, these are some big results.
During the honeymoon phase, people with T1D continue to generate some insulin, so it is quite common to have fluctuating insulin requirements during this time. Some people go weeks, even months, without needing to inject insulin, and this is even more true of adults. The results above are much better than seen on average in teenage honeymoons, but I'm not sure how they compare to average adult honeymoons. I discuss this much more below, in the "weaknesses" section.
I do think it is important that the fasting C-peptide numbers went up for the year of the honeymoon. (Remember that C-peptide is generated by the same process that generates natural insulin. So generating more is better because it means the body is also generating insulin.) In the past, I have seen honeymoon studies where treatments would stop the decline of C-peptide. The researchers would then compare them to the control group where the C-peptide number did decline, and see a statistically significant difference. They would say that their treatment "preserved beta cell function". That's true, but these people already had T1D, and so preserving that did not help them much, and these treatments have not entered widespread use.
But these C-peptide numbers show improvement, not just preservation. At the end, these people were generating 61% more fasting C-peptide than when they started, they went from 0.65 up to 1.05. For comparison, non-T1Ds are typically between 0.5 and 2.0.
This study was published as a correspondence in the New England Journal of Medicine (NEJM), which is an important medical research journal. The authors all work at the University of New York at Buffalo, and the authors have worked on dozens of clinical trials.
What Are The Weaknesses: The Bad News
This study has several weaknesses. I want to stress that none of these show a mistake or invalidate the results. They do show the need to run a larger clinical trial, and also more/different clinical trials. Larger trials validate that this was not just an accidental result and more/different trials show more about who is likely to be helped and exactly what part of the treatment is helpful.
Small Study Size: This study has 10 people, which is common for a phase-I or pilot study, but larger studies are needed to be sure of the results. Phase-II studies are often 100 people, phase-III studies for T1D cures are usually 300 people, and in other diseases, they are often even larger.
No Control Group: This study has no control group. That is common for a phase-I trial (especially the first phase-I trial, a "pilot" trial), but it is still a weakness. The researchers did compare their treatment group to the control groups from 5 recently completed clinical trials, 4 of which were done in adults. You could say they "borrowed" control groups from these other studies.
They compared A1c data from their treatment group to the other study's control groups. This technique is better than nothing, but it is not a true control group, and (for me) using A1c data is a second choice to using C-peptide data. (Because C-peptide means the body is making insulin, but A1c numbers are influenced by how much people eat and how careful they are in injecting insulin.) With all those limitations, basically they found that A1c numbers in all the groups dropped for the first 6 months, but then went up in the the control groups. However, in their treated group, it continued to go down (slightly) for the whole year.
Retrospective: This was a retrospective study, meaning the researchers gathered the data, and then realized there was something important in it, and created the study "backwards" while already holding the data. Most clinical trials that I follow are prospective. The researchers design the study, and then recruit the patients, gather the data, and analyze it according to the design done at the start. There are a large number of potential problems which can occur in a retrospective study that are avoided with a prospective design. It would take a book chapter, or maybe a whole book, to describe all the potential problems, so I can't even list them here. In any case, there is no way for me to know if this study actually encountered any of them, and therefore no reason to worry about them. But this is why a prospective study confirming these results is important.
Impact of LADA: This study enrolled people between the ages of 18 and 49 (average age was 27 and standard deviation was 6 years), so many of these people probably had LADA rather than classic type 1 diabetes. LADA stands for Latent Autoimmune Diabetes in Adults; you can think of it as juvenile diabetes diagnosed in adults, but we don't call it juvenile diabetes any more. LADA has a honeymoon, just as T1D does, but it is not the same. It is well known to be longer and stronger. Since there was no control group, the researchers compared their patients to patients in other studies of adults in their honeymoon phase of T1D. I think that is the best comparison available, but it is important that people not compare these results to standard T1D honeymoons seen in teenagers or younger kids. That is not a good comparison.
Correspondence vs. Article: This publication was not a journal article. Instead it was a correspondence, a shorter work. I don't know what the NEJM's review and editorial processes is for correspondence and how it differs from other articles. I don't know how much peer review NEJM correspondence get, if any.
Only Fasting C-peptide Data: I would have preferred to also have C-peptide data from after a meal, but this data was never collected because it is not part of standard care, and these patients were mostly getting standard care (just with Semaglutide in addition). I like to see both fasting and meal based C-peptide data, because the fasting number just measures the body's ability to make insulin "at rest". The meal number measures the body's ability to make insulin in response to eating food. They are both important in different ways, which is why I like to see both.
No Clinical Trial Registration: This study was not registered
with the FDA. That is normal for this kind of "chart review" /
retrospective style of research. But it still means I don't have as
much data on the study as I usually have before blogging.
Restricted Carb Diet: The letter stated "Carbohydrate intake was restricted in all the patients" which initially made me very nervous, because carb restrictions can significantly lower insulin requirements, especially if taken to extreme. However, in a discussion with one of the authors, I was told that this was simply part of "standard of care" for newly diagnosed T1Ds at their practice. That all such people are told not to eat excessive carbs. They were not told to limit carbs to X grams per day, or Y grams per meal, or anything like that. It was just suggested that they eat fewer carbs as part of a healthy T1D diet.
Inclusion Criteria: The publication does not say how
these 10 people were selected to be included in this study. However, in discussion with one of the authors, I was told that the 10 people included everyone who had been treated with Semaglutide at the practice and followed for a year. There was no "selection" of people. Everyone who they had data for was included. Of course, it is still possible that something about the people who agreed to take Semaglutide or something about how the doctors decided to prescribe it, impacted the findings.
What Should Honeymooners Do Now
I think that adults (especially those with LADA) who are early in their honeymoon phase have two choices:
The first is to do nothing. I think that is a reasonable thing to do in response to a 10 person, no control group, pilot study such as this one. I've gone over the pros and the cons of this treatment above, but it is also important to remember that most research is not successful. In fact, less than 1% of the successful phase-I clinical trials ever become successful treatments. So the long term odds of this being successful are low, and doing nothing is a reasonable course of action. There are some risks both because Semaglutide has known side effects, but also because it has not been fully tested in people with T1D.
The second is to talk to your endocrinologist about it. Semaglutide is a prescription medicine, so if you want to use it, you will need a doctor to write you a prescription. For people in the early part of their honeymoon (the first 100 days) it is a very reasonable conversation to have with your doctor. New medicines and treatments are one of the reasons you have a doctor (and not only a pharmacist). Semaglutide has been widely used for many years, so there is a lot of safety data available. However, it is not approved for type-1 diabetes, so a discussion with your doctor about off label use is appropriate. Doctors have the discretion to use medicine off label. They have the knowledge of T1D in general and the information about their patients in particular to make good decisions even with uncertainty.
Obviously, one major question is: this study was done on adults, does it apply to teenagers or children? No one knows the answer to this with certainty, and that is why I would talk to your endocrinologist if you are considering it.
What about non-honeymooners? Semaglutide has been given to people with T1D for years, usually to treat being overweight. There is no question it helps lose weight, lowers the amount of insulin used, and generally improves A1c and time in range numbers for most people who use it. You can read about cases here:
https://t1dexchange.org/semaglutide-type-1-diabetes/
https://www.diabetes.co.uk/diabetes-medication/semaglutide.html
https://www.oatext.com/impact-of-adding-semaglutide-in-a-person-with-type-1-diabetes-a-case-report.php
Next Research Steps
For me, there are several different research programs that this study should motivate, and I think we (as a society) should work on all of them at the same time.
Follow Those 10 People For Years
One of the big questions is, how long will these people avoid injecting insulin. Obviously a year delay is valuable to some people. The Teplizumab value proposition is based on a 2-3 year delay. By simply following these 10 people for years in the future, we can get some data on the longer term effects of this treatment.
I was able to learn that a couple of the people in the study were no longer being followed because they had moved out of the area or out of the practice. Also, a few more had stopped taking Semaglutide because they had continually lost weight on it, to the point where they needed/wanted to stop. But even with this shrinking study size, I'm still interested in what happens to people in the coming years.
One of the worries of a treatment like this is that it is somehow "burning out" beta cells early and that people who get it might do better in the short term, but worse in the long term. That is the kind of fear that can be removed (or confirmed) by following people for more years. With less than 10 people, the data will be approximate, but I'm a big believer that some data is better than none.
Run A Larger Phase-II Study on Semaglutide In Honeymooners
I listed a bunch of weaknesses in this research, but the three biggest weaknesses are (a) small size, (b) no control group, and (c) only adults were included. All of these issues could be resolved by running a larger, randomized, controlled, clinical study which included younger people. The authors are looking into doing this. The first step is (always) getting the money.
JDRF: are you listening? Fund this larger study! 😁
Run A Study On Semaglutide In People "At Risk" of T1D
(By "at risk" I mean "stage 2" in Trialnet terminology.)
One thing that came through very clearly is that C-peptide levels improved with this treatment. Therefore, it makes sense to try the treatment earlier, when people are still generating more C-peptide naturally. This is the same progression that Teplizumab went through, and TrialNet is perfectly set up to run a similar clinical trial for Semaglutide.
TrialNet: are you listening? Organize this longer, stage 2 study! 😁
Survey Honeymooners Taking Semaglutide "Off Label" Now
Finally, I suspect that some younger people are going to get Semaglutide "off label" during their honeymoon in the hopes that they will see results like those seen in this study of adults. I have no idea if that will be successful or not, but I hope that the doctors who prescribe it in this situation publish the results. That would be very valuable no matter if it is successful or not. One of the things I'm worried about is that many people try it, it does not work, but no one publishes the negative results, so people continue to try it.
Of course, if someone runs the larger phase-II study discussed above, that would give better results than a survey, but a survey might be quicker to organize. Of course, we could do both, a quick survey and a slower but more definitive phase-II clinical trial.
Lots More To Read
https://www.cnn.com/2023/09/08/health/ozempic-diabetes-semaglutide-study/index.html
https://www.buffalo.edu/news/releases/2023/09/UB-research-semaglutide-Type-1-diabetes-insulin.html
https://www.nejm.org/doi/full/10.1056/NEJMc2302677
https://en.wikipedia.org/wiki/Semaglutide
https://www.novomedlink.com/semaglutide.html
...and you might want to read my previous blogging on Teplizumab, too.
