Physicians fine-tune use of diabetes drugs after recent CV safety results

In 2008, the FDA challenged developers of new diabetes drugs to prove that these agents do not increase cardiovascular risk. Results from those studies are now coming in, and beyond demonstrating safety, several are suggesting that drug classes developed over the past decade can reduce overall cardiovascular mortality and mitigate the heart failure and kidney disease that often result from diabetes.

Last year, Endocrine Today assembled a distinguished panel of experts to discuss the topic of CV effects of diabetes medications and what these effects might mean for treating patients (see “Beyond safety, diabetes drugs offer CV benefits” in the August 2016 issue). Since then, findings have been published from key SGLT2 inhibitor studies, including CANVAS and CVD-REAL, that lend evidence to a claim for a class effect with these agents, and from the LEADER trial demonstrating a similar CV benefit from a GLP-1 receptor agonist. Based on LEADER results, an FDA advisory panel in June voted in favor of an expanded indication for liraglutide (Victoza, Novo Nordisk), supporting language in the prescribing information that the drug can reduce CV risk in patients with type 2 diabetes. Similar language was added to the label for empagliflozin (Jardiance, Boehringer Ingelheim) last summer.

The expert discussion continues in this issue with commentary on the possible use of these agents by physicians other than endocrinologists and in patients without diabetes. The roundtable participants, assembled at the American Association of Clinical Endocrinologists annual meeting in May, explore the mechanisms of action of the different agents, the likelihood that the observed CV benefits are a class effect, and more.

Appropriate patients, prescribers

Alan J. Garber, MD, PhD, FACE: The FDA has approved an indication for empagliflozin (Jardiance, Boehringer Ingelheim) for prevention of death in patients with diabetes and established CV disease. This indication could extend to the entire diabetes population. To some extent, it is being promoted as a drug independent of glycemia, which leads me to this question: How dangerous is it to promote empagliflozin as an add-on treatment, as a CVD preventive, without regard to glycemic control? Are we creating situations where a practitioner adds on a drug to prevent CVD in a patient taking a sulfonylurea or insulin who then experiences hypoglycemia? Do we have to warn people about that combination?

Ralph A. DeFronzo, MD: It’s an issue. The FDA overstated the indication: If you say empagliflozin is for use in people who have type 2 diabetes and CVD, that’s virtually every patient who has had diabetes for several years, which is not the intended population. Here we have a great drug that we, as diabetes providers, have not been able to convince prescribers to use, and now it has been taken almost out of our hands; cardiologists are using the drug for CV protection.

The one thing I feel comfortable about is this: Anyone taking this drug will excrete 80 g of glucose in the urine, but fasting glucose will not change. There is a whole new renal physiology in which glucose in the urine sends a signal to the liver to produce glucose, so I’m not worried about precipitating hypoglycemia with the drug or with the drug being used with GLP-1 agonist. However, a patient taking a sulfonylurea will secrete insulin independently of the plasma glucose concentration. A lot of patients the cardiologists are going to be seeing will be taking sulfonylureas and insulin. Prescribers need to understand this particular interaction because this is the one that can cause trouble, ie, hypoglycemia.

Yehuda Handelsman, MD, FACP, FNLA, MACE: Cardiologists have jumped to too many conclusions based on one study. They would like to say that the mechanism of action is all hemodynamic, hence the whole effect on chronic heart failure (HF), and since cardiologists manage people with HF, they should start a program with empagliflozin in HF patients without diabetes. However, whether it will work in people without diabetes is still a question regardless of its mode of action.

Guillermo E. Umpierrez, MD, CDE: There are clinical trials underway investigating the use of SGLT2 inhibitors in patients without diabetes, using dapagliflozin (Farxiga, AstraZeneca). For example, DECLARE-TIMI 58 is a multicenter, multinational study of dapagliflozin in an estimated 17,000 patients. The researchers are randomizing patients with and without diabetes to observe the CV effect on patients who have low ejection fraction. The question is whether the CV effect of SGLT2 inhibitors is independent of glucose lowering.

Handelsman: The problem is we don’t know. When we look at the data in EMPA-REG OUTCOME, there was reduction in both glucose and blood pressure. I’m not saying that this necessarily accounts for the CV benefit, but in ADVANCE and also in ACCORD, as presented in ACCORDION, people who had a reduction in glucose and BP had a reduction in mortality. In SUSTAIN-6 with the GLP-1 receptor agonist semaglutide (Novo Nordisk), there was a significant 1.5% reduction in HbA1c, and a significant reduction in weight and some changes in BP. Could that be part of the CV benefit?

In a recent study from Denmark, researchers looked at people with early metformin treatment who got their HbA1c to less than 7.5%. They found that the earlier treatment group got glycemic control, and the lower the HbA1c, the lower the CV mortality risk over 10 to 12 years. We don’t know enough to reach a full conclusion.

Lawrence Blonde, MD, FACP, MACE: In the long-term follow-up of the United Kingdom Prospective Diabetes Study (UKPDS), in the sulfonylurea-insulin group, relative risk reductions persisted at 1 year for any diabetes-related endpoint and microvascular disease, and significant risk reductions occurred for myocardial infarction and all-cause mortality in spite of an early loss of HbA1c differences from the conventional group. So, you’re right; it may take longer to see the effects of glucose reduction.

DeFronzo: The point you’re making is that whether the prescriber is a cardiologist or a primary care physician or an endocrinologist, glucose lowering is important. Empagliflozin lowers glucose, and that is the primary reason it was approved. Glucose control should be the primary indication for using empagliflozin, and if your patients have that group of CV risk factors used in the EMPA-REG OUTCOME study, then this is clearly a major advantage. In other groups, such as those without established CVD, I’m skeptical as to whether we’re going to get the same results.

Impact of HF, kidney disease

Blonde: One of the underrecognized issues that EMPA-REG OUTCOME called attention to is that congestive HF is a major problem for people with diabetes. HF is often not identified early enough, and SGLT2 inhibitors appear to have a remarkable benefit in reducing hospitalization for HF. Moreover, HF in some studies is the most common reason — more common than MI or stroke — that people with diabetes end up being hospitalized.

DeFronzo: You can throw in renal protection as well, even though we don’t have definitive proof in an FDA indication.

Umpierrez: Many guidelines have already included EMPA-REG OUTCOME data and empagliflozin — the American Diabetes Association, the Canadian Diabetes Association and the European Society of Cardiology — and the ADA came on strong. We definitely need more data, but this is positive, and they’re on the right track.

Handelsman: Looking at the three trials EMPRA-REG OUTCOME, LEADER and SUSTAIN-6, the one thing they have in common is improvement in kidney function. Kidney function is a risk factor for heart disease. In the 2017 American Association of Clinical Endocrinologists lipid guideline, kidney disease is recognized as a very high-risk factor for CVD. Improving kidney function may also affect CV outcome results.

Effect of older diabetes agents

Garber: How much of this effect may not be drug specific, but the absence of the older drugs that really caused more problems? How much trouble is caused by a sulfonylurea or even low-dose insulin used where it’s not necessary? Maybe by using more modern, more effective treatments, we’re avoiding complications and seeing better results.

DeFronzo: There was more sulfonylurea and insulin use in the placebo group in EMPA-REG OUTCOME, and the difference may be accounted for by the excess sulfonylurea or insulin use.

Handelsman: There are some other trials using more recent drugs, like a DPP-IV inhibitor or a short-acting GLP-1 inhibitor, for example, that showed CV safety, but not superiority. Similarly, with thiazolidinediones, there was potential superiority with a TZD, presumably pioglitazone; however, its effect on CVD may be masked by HF, thus using it in the right patient may be beneficial.

Umpierrez: In EMPA-REG OUTCOME, empagliflozin was an ongoing therapy. Do we know if the CV effects on protection were different among the different drugs that were being taken?

Garber: It’s hard to resolve because you need a fair amount of statistical power to resolve subgroups of one drug vs. subgroups of another drug.

DeFronzo: Every time you do these multivariate analyses where you have hundreds of variables, it’s difficult. You can build a model, but it’s still a model that is trying to sort out maybe 10 different components that are changing. The CAROLINA trial of the DPP-IV inhibitor linagliptin (Tradjenta, Boehringer Ingelheim) may give us some answers to this because in one group the control group is placebo and in the other it’s a sulfonylurea. If you go through the literature where people have looked at sulfonylureas and insulin, there are many more papers finding that insulin and sulfonylureas are associated with an increased incidence of CVD than there are papers saying they’re neutral. And when the drug is neutral, the HR is usually above 1.

Mechanisms of action

Handelsman: After EMPA-REG OUTCOME, many people said that the congestive HF benefit was the result of empagliflozin’s hemodynamic effect and not an atherosclerotic effect, and after LEADER, people said the benefit resulted from liraglutide’s (Victoza, Novo Nordisk) atherosclerotic effect, based perhaps on a post hoc analysis combining nonfatal MI with fatal MI that was not even part of the original endpoint. I am not sure that I have enough information that identifies the exact mechanism of action for either of these drugs. What do you think, Ralph?

DeFronzo: The mechanisms are different. I would say that with 70% to 80% authority, which means there is still a 20% or 30% question in my mind, that the effect of GLP-1 receptor agonists is not all atherogenic or that the effect of SGLT2 inhibitors is not all hemodynamic. But I think that these are likely to be the major mechanisms.

Handelsman: If it’s atherogenic, how does it exert this effect?

DeFronzo: Daniel J. Drucker, MD, professor of medicine at the Lunenfeld-Tanenbaum Research Institute at Mount Sinai Hospital in Toronto, wrote a review in Cell Metabolism on the potential anti-atherogenic mechanisms. There are a host of inflammatory cytokines that the GLP-1 receptor agonist class inhibits: It reduces C-reactive protein, improves endothelial dysfunction, has a modest natriuretic effect and decreases BP. The mechanism may not be any one of these, or it could be all of these collectively.

It’s going to be difficult to sort out. It’s more likely we’ll sort out what SGLT2 does in a relatively short period than we’re going to be able to sort out what liraglutide does. There are at least studies you can do in the cath lab, and we’re doing them now. I can tell you what’s happening hemodynamically and volume-wise. Can I tell you that’s what produced the benefit in EMPA-REG OUTCOME? No, but at least for the HF story, I can give you some insights. I suggest that the long-term beneficial effects of empagliflozin could be more than just hemodynamic.

Garber: What was seen in LEADER?

Handelsman: The HF data flatten after 3 months, and then there is a separation of the mortality rate. That most likely points to other mechanisms beyond hemodynamics. For example, an intraplaque hemorrhage can cause rupture. This is typically BP-related, so as there is an effect on BP it can lead to a reduction in the potential for plaque hemorrhage; this is in essence an atherosclerotic effect. Still, why do we see a trend toward strokes in all these compounds?

DeFronzo: One could argue that there are people who have these watershed lesions, where they are marginally profusing in certain areas of the brain. With a BP reduction, you’re getting a little decrease in intravascular volume. You could put together a scenario in which this might precipitate a stroke. However, it should be remembered that the increase in stroke was not statistically significant, so we should not overinterpret this result.

Handelsman: Why would this be more the case with an SGLT2 inhibitor than with any diuretic?

DeFronzo: I don’t know how carefully that’s been looked at with the diuretics. Diuretics don’t have major effects on atherosclerotic-type issues. But if you go back to the old literature, there was always this issue that if you over-diuresis people, or if you lower their BP too much, you’ll cause them to have a stroke. I do have this concern about stroke not declining, because from the data I’ve seen, with all three drug classes, the stroke rate is going up or at least is neutral.

Umpierrez: The area that needs to be studied for SGLT2 inhibitors is the effect on sympathetic activity and parasympathetic activity. SGLT2 inhibitors decrease BP, but heart rate doesn’t rise.

Handelsman: Indeed, this is very interesting with the SGLT2 inhibitors. There is BP reduction while the heart rate, which should go up, actually goes down. This points to a potential parasympathetic effect.

Determination of a class effect

Garber: A clinical trial does not tell you the mechanism of how something occurred. What you’re looking at is a complete inversion of the usual order of business, if you will, because a large-scale clinical trial used to be expensive to mount and would not have been done unless there had been an accumulation of basic science data that mandated that a clinical impacts study be done to figure out whether it was meaningful or not. Now we’re all asking in retrospect, “How did this happen? Why does this happen this way?” because the FDA mandated that we prove that these drugs don’t kill people by CV results.

Umpierrez: We are 10 years on that road now. Do you guys think that it’s going to be a class effect?

DeFronzo: I would drop dead on the spot if this is not a class effect in the population that they studied for EMPA-REG OUTCOME. If you start to study people earlier in their disease, I’m not convinced you’d see the same results; but in this high-risk population, I would be shocked, particularly in light of the real-world data from the THIN study and CVD-REAL. The analysis of CANVAS, the phase 2b studies of dapagliflozin, the THIN studies and CVD-REAL, collectively, looked at all three drugs in six different countries, and all showed exactly the same trends in CV outcomes trials and in the real-world situation.

Umpierrez: Also, DECLARE-TIMI 58 with dapagliflozin is estimated to be published in 2019 or 2020.

Blonde: Many people think the results of EMPA-REG OUTCOME are probably a class effect, but it is possible that some subsequent CV outcome trials with other SGLT2 inhibitors may not demonstrate all of the same results because of differences in the patient population or the study design. However, the recently reported CVD-REAL data were impressive and informative. This was a real-world observational study of more than 300,000 patients with type 2 diabetes, with and without established CVD, from clinical practice databases in six countries. The study evaluated hospitalization for HF and all-cause mortality in those treated with SGLT2 inhibitors compared with those treated with other antihyperglycemic agents. Use of specific SGLT2 inhibitors differed in different countries (predominantly canagliflozin in the United States and dapagliflozin in Europe), but the results were the same, suggesting an association with the class rather than any single agent. Initiation of SGLT2 inhibitors was associated with a lower incidence of hospitalization for HF, a lower rate of all-cause death and a lower rated of the combined endpoint of hospitalization for HF or all-cause death. Most patients did not have established CVD, suggesting that the benefits may apply to lower-risk patients than those enrolled in the CV outcomes trials that have been reported so far.

DeFronzo: We have three drugs in six countries, all trending in the same direction, with the same magnitued of reductions and looking exactly like EMPA-REG OUTCOME, with stroke rate tending to go up a little, MI down a little, and all-cause mortality — mostly CV mortality — driving the reduction in major adverse cardiac events. I would be shocked if they don’t all turn out to be positive

FDA DPP-IV decision

Garber: What do you think of the FDA’s complete response letter refusing to grant an exception to Merck for the DPP-IV inhibitor sitagliptin (Januvia) on the HF issue, in light of the TECOS data? In other words, TECOS fairly clearly shows no excess hospitalization for HF.

Blonde: But the latest sitagliptin label doesn’t include HF under warnings and precautions as does the prescribing information for saxagliptin (Onglyza, AstraZeneca) and alogliptin (Nesina, Takeda).

DeFronzo: I can’t prove this, but I don’t think any of these drugs cause HF. The SAVOR-TIMI 53 trial was just bad luck.

Umpierrez: The FDA recently reported results from a database with 358 million person-years of use of DPP-IV inhibitors, and it concluded there is no increased risk of HF with sitagliptin or saxagliptin.

Handelsman: That’s terrific as these results support clinical practice, because in real life when we prescribe a sulfonylurea, we see hypoglycemia. When we give insulin, we see hypoglycemia and weight gain. When prescribing a TZD, we often witness HF and weight gain. Yet I have not seen any of my patients on a DPP-IV inhibitor who had HF.

Umpierrez: The study showed increased risk for hospitalization in the U.S. and Canada, but not in the rest of the world. Here, if you have shortness of breath and you have diabetes, it’s called HF.

Garber: It’s difficult to refute an admission diagnosis of HF, and so that’s the No.1 reason for HF being the most frequent Medicare readmission diagnosis.

DeFronzo: That was the case with pioglitazone in PROACTIVE. Most of those people who were admitted with so-called HF didn’t, in my opinion, have HF, they had edema. In fact, in the “HF” group, there were fewer MIs and fewer strokes, and the patients lived longer than the group who didn’t have HF, which to me says that these people really didn’t have HF.

Garber: Some of these people may not have had any cardiac-related underlying problem but were admitted because the easiest diagnosis under which to admit is HF.

Take-home messages

Garber: What’s your take-home message? It seems to me the data are clear that modern antidiabetic drugs are beneficial for patients to a far greater extent than older antidiabetic drugs, particularly sulfonylureas, which should be displaced from usage formularies in the United States.

Umpierrez: The initial studies — EMPA-REG OUTCOME, LEADER and SUSTAIN-6 — suggest the beneficial effects of SGLT2 inhibitors and GLP-1 receptor agonists on CV outcomes, but more data are needed. We’re hoping that the next few trials that are coming out in the next 2 years are going to confirm these preliminary data.

Blonde: The results that we’ve seen so far with the CV outcomes trials have been in sync with the recommendations in the AACE algorithm for the management of people with type 2 diabetes.

DeFronzo: We now have moved into a new era of diabetes treatment. We have newer drugs. The SGLT2 inhibitors and the GLP-1 receptor agonists and at least one older drug, pioglitazone, where these drugs not only provide important glycemic benefits, but also give CV protection. Something even less appreciated is that they provide renal protection. We need to use these new drugs, and, unfortunately, we’re often driven by cost constraints and are forced to use the older drugs, such as metformin and sulfonylureas.

Handelsman: We are in an exciting time. The CV safety and outcome trials showed that drugs for diabetes are safe. A lot of cardiologists had proposed that people who have diabetes managed with the diabetes drugs are actually at increased CV risk because of those drugs. Now the conversation has changed. These newer drugs have demonstrated safety, and some of them have demonstrated improved CV outcomes as well. I’m glad to see that a lot of cardiologists are finally recognizing that it is indeed important to treat people who have diabetes with approved diabetes medications. – by Jill Rollet

• For more information:
• Lawrence Blonde, MD, FACP, MACE, can be reached at Ochsner Medical Center, 1514 Jefferson Highway, Jefferson, LA 70121; email: lblonde@ochsner.org.
• Ralph A. DeFronzo, MD, can be reached at University of Texas Health Science Center at San Antonio, 7703 Floyd Curl Drive, San Antonio, TX 78229; email: defronzo@uthscsa.edu.
• Alan J. Garber, MD, PhD, FACE, can be reached at Baylor Clinic, 6620 Main St., Suite 1100, Houston, TX 77030; email: agarber@bcm.edu.
• Yehuda Handelsman, MD, FACP, FNLA, MACE can be reached at Metabolic Institute of America, 18372 Clark St. #212, Tarzana, CA 91356; email: info@yehudahandelsman.com.
• Guillermo E. Umpierrez, MD, CDE, can be reached at Emory University School of Medicine, 49 Jesse Hill Jr. Drive SE, Atlanta, GA 30303; email: eumpie@emory.edu.

Disclosures: Blonde reports receiving grant or research support from AstraZeneca, Janssen, Lexicon, Merck, Novo Nordisk and Sanofi, and serving as a consultant or speaker for AstraZeneca, GlaxoSmithKlein, Intarcia, Janssen, Merck, Novo Nordisk and Sanofi. DeFronzo reports serving on advisory boards for AstraZeneca, Boehringer Ingelheim, Elcelyx, Intarcia, Janssen and Novo Nordisk and on the speakers bureau for AstraZeneca and Novo Nordisk and receiving support from AstraZeneca, Boehringer Ingelheim, Janssen and Takeda. Garber reports serving on the advisory board for Janssen, Kowa, Lexicon, Merck, Novo Nordisk, Takeda, Viking Therapeutics and Vivus and serving as a consultant for Janssen, Kowa, Lexicon, Merck, Novo Nordisk, Takeda, Viking Therapeutics and Vivus. Handelsman reports receiving consultant, speaker fees and research grants from various pharmaceutical companies, including Amarin, Amgen, AstraZeneca, Boehringer Ingelheim, Bristol-Myers Squibb, diaDeux, Daiichi Sankyo, Eisai, Gilead, GlaxoSmithKline, Grifols, Hamni, Halozyme, Intarcia, Janssen, Lexicon, Lilly, LipoScience, Merck, Novo Nordisk, Regeneron, Sanofi, Salix, Takeda and Vivus. Umpierrez reports no relevant financial disclosures.