Michael is a professor and head of the Department of Diabetes at Bad Lauterberg in Germany. He has contributed to the development of new therapeutic principles of treatment of diabetes, particularly focusing on gut hormones and GLP-1 receptor agonists. He has published extensively in this field. He's very excited that he could participate in this program, and he is going to update us on GLP-1 receptor-based strategies early in the natural history of the disease, and also focus on some safety issues. Welcome Michael.

Let's start with a question. I'll read the question to you. Which of the following statements is true with regard to GLP-1 receptor agonists and DPP-4 inhibitors? Answer 1, DPP-4 inhibitors lower HbA1c levels more than do GLP-1 receptor agonists. Second, GLP-1 receptor agonists result in more weight lost than DPP-4 inhibitors. Third, DPP-4 inhibitors and GLP-1 receptor agonists share the same mode of action. And fourth, DPP-4 inhibitors are based on the enzyme found in the saliva of the Gila monster.

Okay, we'll talk about the correct answer.

Let me now introduce the 2 classes of GLP-1, or incretin-based glucose-lowering medications: DPP-4 inhibitors and GLP-1 receptor agonists.

They are both derived from biological activities of the parent compound, GLP-1, a gut hormone that stimulates insulin secretion, suppresses glucagon, has motility effects, especially slowing gastric emptying. It reduces appetite and food intake. It may have specific activity in the cardiovascular system, and, sure enough, when you enhance insulin secretion and suppress glucagon secretion, this will effect liver metabolism, hepatic glucose output, and the way muscle and adipose tissue handle the glucose.

Incretin mimetics are analogs of GLP-1. They act through the same receptor. They should not be susceptible to degradation by DPP-4. They should be resistant, and they should have a much longer half-life than the natural product, GLP-1 itself, which has a half-life of 1 to 2 minutes. As you know, pharmaceutical companies have succeeded in providing drugs with very long half-lives so that you only have to inject them once a week.

DPP-4 inhibition is acting on an enzyme, a proteolytic enzyme that degrades GLP-1 very actively. When you administer a DPP-4 inhibitor, that means that it is the endogenously secreted GLP-1 that reduces glucose through the mechanisms I have outlined.

Some more information about GLP-1 receptor agonists.

Here you see the molecular structures of GLP-1, and exenatide is in the second row, and lixisenatide is a more novel analog to exenatide. Liraglutide is the first designer molecule that was not just isolated from the saliva of the Gila monster, but designed to be very similar to the original GLP-1 molecule and, then, with a side chain of a free fatty acid that promotes binding to albumen and prolongs half-life. A novel molecule from the same company is making use of the same trick, if you like. Taspoglutide is no longer very important because the development has been stopped because of side effects.

Some more novel approaches, and especially regarding dulaglutide and albiglutide, the full phase 3 program is being presented at this ADA conference. These are all very large molecules, for example, with a backbone of human albumen or with some mock peptides that you can let grow to such a size. They have a very long half-life, once weekly injection schedule, and we'll see which of them will be best.

I am talking about the early-stage diabetes and treatment with GLP-1 receptor agonists. These are patients who mostly have been treatment naive and receive the new treatment in monotherapy.

This is a monotherapy study for exenatide twice daily versus placebo, and what you can see is over these 24 weeks, HbA1c is reduced by .8 to .9%. With placebo, you don't see it to the same degree.

Weight also goes down by 3 kilograms, more so with a larger dose of exenatide, but even a little bit with the placebo treatment.

If I now move to liraglutide, this is not a placebo-controlled study, but it was compared to glimepiride and 2 doses of liraglutide were used. If you look at all participants, liraglutide was more potent in reducing HbA1c than was the sulfonylurea, glimepiride. This was especially apparent with large drops in HbA1c in the previously drug-naive patients, whereas previously treated patients had a little bit smaller responses, but the same hierarchy.

Again, in this study, there is weight loss associated with treatment with liraglutide, the GLP-1 receptor agonist, and weight gain as expected with the sulfonylurea glimepiride.

A similar study has been published with a new preparation of exenatide that is changed so that you inject material encapsulated in a polymer only once weekly. From this depot, the exenatide is slowly released. It takes weeks before you reach fully active drug levels, about 10 weeks. Then you see a nice drop in HbA1c. This is the open symbols down here.

In this study, that was compared to metformin, pioglitazone, and to sitagliptin, and it's a surprise that all of these oral drugs, with the exception of sitagliptin, almost were as good as exenatide once weekly. I attribute this to the early stage of diabetes, which you can easily treat with these oral agents and that also means that you get similar percentages of these patients to the targets of less than 7 and equal or less than 6.5% HbA1c.

Weight goes down with exenatide once weekly and is reduced with metformin as well. Sitagliptin is weight neutral, and with pioglitazone, the patients put on some weight.

The next stage of diabetes is when the patients mostly have been initiated on metformin, but you find out that metformin is no longer satisfying you with regard to treatment results.

This is a panel showing different GLP-1 receptor agonists. You see the color code down here, and they uniformly reduce HbA1c. The more short-acting ones, such as exenatide and lixisenatide, which is shown here, are somewhat weaker in their ability to reduce HbA1c. This is especially true when you look at fasting glucose, or the blue columns, and with placebo you see little changes. All the GLP-1 receptor agonists, to a similar degree, reduce body weight by 2.63 kilograms. None of them, in monotherapy or with metformin background, produce hypoglycemia.

I can even go a little bit further and compare the effect of GLP-1 receptor agonists. It's again, quite a panel of different compounds that are depicted here, and in each case, it was compared to insulin, either insulin glargine, the dark green, or the light green is premixed insulin. If you look for the pairs of data that were compared head-to-head, then you always see that the treatment result was very similar with insulin and with a GLP-1 receptor agonist. Sometimes with a better result for the GLP-1 receptor agonist, for example, here with the exenatide once weekly.

That was not quite the case with fasting glucose, which usually was reduced a little bit better with the insulin preparation, but insulin made the patients gain weight, and all the GLP-1 receptor agonists made the patients lose some weight, as previously shown. In some of these studies, it was not only metformin as a background, but also sulfonylureas, and that explains why there is hypoglycemia, not only with the green columns representing insulin therapy, but, to a lesser degree, with the GLP-1 receptor agonist therapy.

That means, if you use them without sulfonylureas or insulin, then you can be pretty sure that they do not provoke hypoglycemic episodes. If, however, you use them together with sulfonylureas or insulin, there may be some hypoglycemia caused by these added agents.

Since DPP-4 inhibitors and GLP-1 receptor agonists are both incretin-based medications and they work through stimulating GLP-1 receptors, it's of interest to compare GLP-1 receptor agonists and DPP-4 inhibitors.

In most studies, sitagliptin was used. As you will see, very uniformly, sitagliptin reduced HbA1c, but not quite to the degree that the GLP-1 receptor agonists did.

DPP-4 inhibitors induced little weight loss. Most people would say they are weight-neutral, do not change anything there, but you have the substantial weight loss to over 3 kilograms in patients treated with GLP-1 receptor agonists.

Again, a comparison between sitagliptin and, in this case, exenatide once weekly. Exenatide once weekly gives you the better glycemic response, minus 1.4, 1.5% HbA1c. You do not quite see this with sitagliptin, although, even with this study, there was a drop by .9%. Pioglitazone was of intermediate effectiveness here. That is reflected also by the percentage of patients achieving targets below 7 or equal or below 6.5%.

If you look for fasting glucose, it mirrors what you have seen for HbA1c, so the weakest reduction with sitagliptin, intermediate response to pioglitazone, and the best reduction by 2 millimolar, 36 milligrams per deciliter with exenatide once weekly. That also gives you the best weight result, with sitagliptin again being more or less weight neutral and pioglitazone making the patient gain a little bit of weight.

There has been a recent publication, about 1 year old, prepared by a panel of delegates, both from the European Diabetes Association, I was a member on that panel, and from the American Diabetes Association.

They have come up with a concept of individualized or personalized treatment. That means you ask the patient, "What are your preferences?" You know the patient. You know the characteristics that drives your treatment decisions. Then, within the process of shared decision-making, you pick the optimum treatment for this given patient, although it usually starts with education to lead to a healthy lifestyle. Then you add metformin if the patient tolerates metformin and has no contraindications. Then you switch to the dual therapy, maybe later triple therapy.

This gives you all the choices that are supported by approval of these drugs. But this system can be used to ask specific questions.

For example, you can just not show those treatment options that would produce hypoglycemia. So the residue shows you the medications that would not produce hypoglycemia. Of course, then the choice is not as wide as it was before.

You can ask the question, "What would be my medications when I want to avoid weight gain?" Then the number is even smaller.

You can as well ask the question, if you live in an environment where resources are not there to pay for costly medications, then you can use metformin, diet is always cheap, sulfonylurea, and insulin preparations.

Let me now show you a patient that I have treated about a year ago.

He was an engineer, pretty young, 37 years. He was of Turkish origin. He had a mother and a sister with diabetes, so some genetic risk there. He was diagnosed at the age of 35 with type 2 diabetes. He hadn't found the time to undergo structured patient education. His treatment was quite okay, metformin and liraglutide, but he admitted that he did not regularly inject and so the result was pretty bad, with an HbA1c of 9.3%. He was obese and he had psoriasis, including psoriatic arthritis. He needed intermittent treatment with glucocorticoids, and he had a chronic treatment for that condition as well.

This is just to show you that this is a real case with psoriatic plaques. You also see the swollen fingers and joints of his right hand.

This was his chest x-ray, indicating some goiter.

This was his electrocardiogram, just to prove that this is a real patient and it was not just made up.

He had a fasting glucose of 181 milligrams per deciliter. C-peptide clearly indicated type 2 diabetes. He was very much insulin resistant, and psoriasis is associated with insulin resistance. He had lipid abnormalities, with a slightly elevated LDL and hypertriglyceridemia. His liver enzyme activities were normal, but he had signs of chronic inflammation related to his arthritis, with CRP elevated to more than 10-fold, and thrombocytosis.

A question. What might be the priorities of this patient regarding treatment of his diabetes? Is he interested or should he be interested in glycemic control? Second, in weight reduction? Should the treatment be compatible with his chaotic meal rhythm? Or should he be interested in reducing the risk for diabetic complications? Or does he want little disturbance of stressful professional activities, which take his full attention?

Okay, so we'll see the results. Of course, there is some possibility that all of these suggestions are true. You had to make a choice, and your priorities were glycemic control and weight reduction. I think this is a fair choice among these 5 answers.

Which medication would not fully serve these purposes? Number 1, sulfonylureas; second, glitizones; third, insulin glargine; fourth, DPP-4 inhibitors; and fifth, alpha-glucosidase inhibitors.

So, yes. Sulfonylureas would cause hypoglycemia and weight gain. Glitizones, weight gain. Glargine also has both these risks, and this is least the case with DPP-4 inhibitors. Very correct choices that you've made.

Let's continue. He had steatosis, as you can see from the hyperdensity of his liver upon ultrasound.

This was how he arrived at our hospital with a low dose of metformin, which was then increased. It clearly did not control his fasting glucose, which was 181, and he had postprandial peaks above 200 milligrams per deciliter.

We increased his metformin dose, and we initiated liraglutide, as recommended, with a low dose for the first week. You see already some improvement.

The week after we increased to 1.2 milligrams, and the fasting, as well as the profile of glucose concentrations were pretty much in the target range afterward. Probably this was a pretty wise choice for this patient.

Let me now discuss a number of safety aspects that have received a lot of attention in recent years and weeks.

Let me first talk about thyroid cancer. As you know, there are different histological types, and they have a very much different biology. Here I want to focus on medullary thyroid carcinoma, which is a carcinoma of calcitonin-secreting C-cells. In both women, upper panels, and men, lower panels, only a very small percentage of thyroid carcinoma is of that histological type, and you call it medullary thyroid carcinoma.

The rest of the story is mainly on findings in animals, more specifically rodents. We are talking about both mice and rats. If you concentrate on male rats receiving no drug treatment, then you see that these male rats have C-cell hypoplasia. They have C-cell adenomas. A few of them develop C-cell carcinomas. There is a spontaneous tendency in this species to develop these things. The numbers increase when they are exposed to increasing doses of liraglutide. In females, it's similar but not quite as bad. In mice, the numbers are smaller, but there also is an induction of C-cell proliferation with exposure to liraglutide.

One might be concerned that, if the same thing happens in humans, then this treatment could induce C-cell carcinoma. However, if you examine C-cells in tissue culture in rats, then you can stimulate with any agent that stimulates GLP-1 receptors, exenatide, liraglutide, GLP-1 itself, and you will have an increment in cyclic AMP production and in calcitonin secretion. If you take human cell lines, there is no rise in cyclic AMP. The notable exception is forskolin, that has nothing to do with GLP-1 receptors. It just raises cyclic AMP. There is also no rise in calcitonin secretion.

In long-term studies, and this is a 1-year study using very high doses of liraglutide for the treatment of obesity, you see that the calcitonin levels remain low throughout, if you calculate mean values and ranges, well below the upper limit of normal for both female and male patients.

I think the best way to summarize these data is that C-cells in rodents express a lot of GLP-1 receptors. They can be stimulated by liraglutide, and probably also by exenatide. That leads to cyclic AMP production and calcitonin release, which goes along with a proliferative response, giving rise to hyperplasia, adenomas, and carcinomas.

But in humans, there is very little GLP-1 receptor expressed on C-cells, and even with high concentrations, you cannot provoke cyclic AMP production. There certainly is not calcitonin release, even with prolonged exposure. We tend to use this as an indicator that there will also not be a proliferative response leading to tumors.

Another topic is pancreatitis and pancreatic cancer.

I will try to interpret a recent study that was published by Singh and coworkers in JAMA Internal Medicine. This gives you the results as presented in this paper. They say current exposure; recent exposure, so this means within the last 30 days; this is within the last year, or any exposure; elevates the risk by 2-fold. This was significant because the confidence intervals did not cut this line, indicating a relative risk of 1. However, this is a mixed group of patients that received either exenatide or sitagliptin.

If you simply separate the groups and just look for what exenatide does and what sitagliptin does, then the elevation in risk is much smaller and it is no longer significant. Current exposure to exenatide, according to this analysis, almost reduces the risk. So that is not compatible with the way the authors have published this.

I recalculated, because they gave all the data you need to do that, the rates for the combined group, I did not have the possibility to adjust for confounders. My relative risk is 1.5-fold rather than 2-fold.

An additional point that they reported was that, in this case-control study, cases with pancreatitis had more risk factors for pancreatitis. There were patients with hypertriglyceridemia, more patients who used alcohol. I don't believe this number; it must be greater than that. Gallstones, tobacco abuse, obesity, and some other confounders. This means to me that the pancreatitis was not only caused by the drugs that were examined by this study but also by these confounders.

If you control for these confounders, I expect that the relative risk should be reduced, but the opposite was the case. I personally suspect that there is something wrong with the way, how they took account of these confounding factors.

Lastly, other histological changes.

Here I'm referring to a recent publication by Alexandra and Peter Butler. What they found by examining donor pancreases that were taken out of brain-dead patients in order to be transplanted, but they couldn't be transplanted for various reasons and were used for research. They had nondiabetic patients, and they had patients with diabetes mellitus, and some of them were treated with incretin-based drugs: mainly sitagliptin, 1 patient with exenatide. Then they see differences between the 2 groups in pancreatic weight, in beta cell area, a little bit in beta cell diameter, and certainly in beta cell mass. They also find differences in alpha cell area, with much more glucagon expressed in the incretin-treated patients. A little bit of a change in the diameter, and certainly a major change in alpha cell mass.

In addition, they described little nodules, which turned out to be microadenomas, mostly also expressing glucagon. However, this study has been criticized on the following grounds.

Here is the non--incretin-based medication control group. They were called type 2 diabetic. You can retrieve data from the database supplying the pancreas samples, and then you can calculate the age at diagnosis of these patients. It was 16, 18, 20, 34, 15, 23, which to me suggests that these must have been type 1 diabetic patients, which is supported by low C-peptide concentrations, by the presence of auto-antibodies, which is not depicted here, and by the fact that they had diabetic ketoacidosis at diagnosis. This indicates that a majority of patients in that group were not unequivocally type 2 diabetic, but type 1 diabetic. This is one of the confounding factors that may explain much of their findings.

Also, if you are on an intensive care unit, mechanical respiration intensive care unit for more than 3 days, that increases the proliferation rate in your pancreas by 30-fold to 25-fold. Data regarding intensive unit treatment in these patients were not reported.

There are reasons to believe that you have to do a better study in order to support these findings.

Lastly, let me talk about some even earlier stages of diabetes, impaired glucose tolerance. You have the glucose tolerance a little bit worse. You have insulin being secreted at the same level, despite this difference in glucose concentrations. These subjects also have a GLP-1 response after a meal.

What Maria Byrne published in 1998 was that, if you induce variations in glucose rhythmic changes, then these IGT subjects do not respond by secreting insulin in parallel. Only if you infuse GLP-1 you can induce parallel changes in insulin secretion that mirror the ups and downs in glucose concentrations.

Robert Ritzel in my group did a similar study. In this case, no glucose was infused. This was the spontaneous variation in glucose concentrations when you draw blood every minute. You have a little pulsatile insulin secretion. But if you infuse GLP-1, glucose goes down a little bit and you suddenly have these pulses of insulin secretion that go very much parallel to the changes in C-peptide, and then you can use deconvolution to quantify the changes.

Both in patients with type 2 diabetes, which I have not shown, and in the IGT subjects, you have a higher amplitude of insulin pulses with the infusion of GLP-1.

Now I have my last question for you. You have seen that before, so let's see whether your opinion has been changed through my talk.

Which of the following statements is true with regard to GLP-1 receptor agonists and DPP-4 inhibitors? First, DPP-4 inhibitors lower HbA1c levels more than GLP-1 receptor agonists. Second, GLP-1 receptor agonists result in more weight loss than DPP-4 inhibitors. Third, DPP-4 inhibitors and GLP-1 receptor agonists share the same mode of action. And last, DPP-4 inhibitors are based on the enzyme found in the saliva of the Gila monster.

So, which was the true answer? It was answer number 2.

DR. FONSECA: In view of all these controversies that we've heard about on side effects, is it still appropriate to use GLP-1 receptor agonists and DPP-4 inhibitors early in type 2 diabetes?

DR. NAUCK: Well many specialists, and those familiar with the details of these studies, have come to the conclusion that there is no reason to change clinical practice. These medications should be used, as of last year or half a year ago. I think, psychologically speaking, the only change would be if a patient has received information that scares him so much that he just doesn't feel well doing this. Then, of course, we are happy that there are alternative treatments. But I would not be prepared to recommend to patients that they should be taken off incretin-based medications for these reasons.

DR. FONSECA: Just to remind the audience that none of these agents are approved for the treatment of IGT and IFG or prediabetes. That data that Michael showed you was purely a mechanistic thing, which is very important. There are no clinical trials ongoing, as far as I know, looking at prediabetes treatment.