From Unexpected CV Benefits to Potential in Heart Failure: Insights and Outlook for SGLT2 Inhibitors
In September 2015, results from the EMPA-REG OUTCOME trial stunned the medical world: for the first time, a treatment for type 2 diabetes (T2D), empagliflozin, was found to have cardiovascular (CV) benefits.1 The good news about the sodium glucose co-transporter-2 (SGLT2) inhibitor did not end there, however. Researchers have continued to pore over data, finding evidence of additional benefits.
In April, The American Journal of Managed Care® convened its second Diabetes Stakeholders Summit. Moderator Dennis P. Scanlon, PhD, professor of health policy and administration and director of the Center for Health Care Policy and Research in the College of Health and Human Development at Pennsylvania State University, University Park, Pennsylvania, led the Peer Exchange™ panel discussion, “Diabetes Therapy and Cardiovascular Outcomes: An Update.” Joining him were Silvio Inzucchi, MD, medical director for Yale Diabetes Center, New Haven, Connecticut; Zachary Bloomgarden, MD, clinical professor in the Division of Endocrinology, Diabetes, and Bone Disease of the Department of Medicine at Mount Sinai, New York, New York; Robert A. Gabbay, MD, PhD, FACP, senior vice president and chief medical officer, Joslin Diabetes Center, Boston, Massachusetts; and Kenneth Snow, MD, MBA, medical director, Aetna.
Inzucchi explained that for years, diabetes providers had been frustrated by the fact that correcting a fundamental feature of the disease—hyperglycemia—had little or no effect on CV outcomes. “We can reduce retinopathy, and nephropathy, and probably neuropathy,” he said. “But when you look at studies over many decades, it’s been very difficult to demonstrate that lowering glucose with a specific strategy or any drug actually benefits the heart.”
EMPA-REG OUTCOME didn’t set out to find a CV benefit. The trial’s purpose was to show that empagliflozin was safe, in the wake of events in the mid-2000s that suggested rosiglitazone caused heart attacks. While the FDA ultimately cleared rosiglitazone, the saga paved the way for new protocols that require diabetes and obesity therapies to demonstrate safety for those at high risk of heart attack or stroke.2
Inzucchi made an important distinction about what the trial did and did not find. “I think what EMPA-REG OUCOME showed us is that you can improve cardiovascular outcomes, perhaps not through lowering glucose, but through using a glucose-lowering therapy,” he said. This was the first time a diabetes drug was shown to have a benefit for CV mortality, and it was associated with a 38% reduction.
“I must say, when I saw these results—and I was on the steering committee for the trial—I almost fell out of my chair,” Inzucchi shared with the panel. He was struck that a diabetes therapy that was effective, but not hugely powerful, in lowering blood glucose, could bring such a result in reducing CV death.
It’s important, he said, not to overinterpret the results, as EMPA-REG OUTCOME involved patients at high risk or established CV disease (CVD). “Primary prevention, in terms of patients without prior history of CVD, has not been demonstrated,” he said.
As much as researchers are still gleaning information from EMPA-REG OUTCOME—and still learning about SGLT2 inhibitors generally—the trial is a breakthrough and has changed the thinking about treating diabetes in many ways, Gabbay said.
Snow agreed. “Certainly, one of the major driving forces for why we treat diabetes to begin with, and why payers pay for the treatment of diabetes is not so much because we want to see lower blood sugar, but because we want folks to live longer, healthier lives,” he said. “And ultimately, these types of outcomes trials, particularly if we are seeing reductions in major cardiovascular events, are exciting.”
What Do We Know About SGLT2 Inhibitors?
The SGLT2 inhibitor drug class has a completely different mechanism from other antidiabetic therapies. The drugs target a protein that normally reabsorbs glucose in the kidney, and instead blocks this function and causes excess glucose to be expelled in the urine. Scanlon asked Gabbay what researchers have learned about the mechanism of action of SGLT2 inhibitors that explains the results found in EMPA-REG OUTCOME.
“It’s a great question,” Gabbay said, adding that the results surprised many. There’s been a lot of “thinking backward,” to truly understand how SGLT2 inhibitors work, and therefore, how they achieve what they do. “What we do know about SGLT2 inhibitors is that they result in a little bit of diuresis and volume contraction, and that, certainly, could be one of the factors [particularly in terms of congestive heart failure incidences and hospitalizations for congestive heart failure], for which they saw a benefit. There’s also a small amount of weight loss, which could also be a factor.”
As he explained, regression models using data from the EMPA-REG OUTCOME trial estimated that about half the effects could be related to volume. Another correlation that merits further study involves uric acid levels.
“There’s another finding of the empagliflozin trial, which is fascinating and may shed light on this—the effect on renal disease,” said Bloomgarden. EMPA-REG OUTCOME showed that empagliflozin was not simply a diuretic, but also acted on sodium secretion; it worked in the kidney “in a lovely way with angiotensin blocking agents,” Bloomgarden said.
“So, at the level of the macula densa, delivering more sodium to that part of the kidney seems to then potentiate the benefit of not having so much angiotensin action on board,” he said. “Well, this fits very nicely into a lot of our clinical knowledge of what’s good for heart failure and our theoretical ideas of what’s good for the heart and what’s good for the kidneys.”
A New Indication for Empagliflozin
Scanlon asked the panel to discuss an FDA decision to add a new indication to empagliflozin, to reduce CV death in patients with T2D.3 What, he asked, are the clinical decision-making implications?
Payers face challenges, Snow said, in deciding whether the effect is just for empagliflozin or a class effect that applies to other SGLT2 inhibitors. “Is it in all patients or only those with preexisting heart disease?” Snow asked. “These are research questions that are still in the process of being answered, and somehow, in the process, there needs to be a decision on coverage.”
Bloomgarden said that if other trials do not show benefits, it would be difficult for any payer to not provide empagliflozin to patients with known heart disease, “of which there are so many.”
“At the very least, there are now data out there about the population of folks who clearly got a benefit with a particular agent,” Snow said. “And now, really the question is, is it unique and is it unique to that population?”
CV Benefits in the GLP-1 Class
Scanlon asked Bloomgarden to comment on the LEADER trial, which found that liraglutide, a glucagon-like peptide-1 (GLP-1) receptor agonist, also had cardioprotective benefit: results presented at the 76th Scientifiic Sessions of the American Diabetes Association (ADA) in June 2016 showed it reduced CV death in high-risk patients by 22%.4 Results presented at ADA the year prior for another GLP-1, lixisenatide, had shown only that the drug was safe, not that it had demonstrated any CV benefit.5
“LEADER was fascinating, coming as it did, immediately after the EMPA-REG OUTCOME trial,” Bloomgarden said. “The strategy was a little bit different. The centers were asked to try to achieve good glycemic control in the patients in the control group and the patients in the liraglutide group, so that there was an up-titration of nonliraglutide, non-GLP-1 receptor agonist therapies in the control group.” From his vantage point, liraglutide’s benefit was that it allowed clinicians to avoid the harms of older medications like sulfonylureas.
Inzucchi agreed. He discussed results for SUSTAIN-6, which found a CV benefit for semaglutide, a once-weekly GLP-1 that is not yet available.6 “It seemed that the most potent effect was actually on stroke, which was surprising because that’s not what you see with atherosclerosis trials,” Inzucchi said. “You typically see it on nonfatal (myocardial infarction) and, maybe, cardiovascular mortality.”
For Inzucchi, the differences point up the importance of not assuming anything. “It leads us to understand that you cannot appreciate the effects of these individual drug categories until you get at least 3 or 4 trials under your belt so you can see the overall effect of the class,” he said.
SGLT2 Inhibitors and Heart Failure
Recently, increased attention has been given to the idea that SGLT2 inhibitors could have a role in primary prevention of heart failure (HF) for the broader population with diabetes, not just patients at high risk for CV events. The presentation of results for the CVD-REAL trial in March 2017 at the American College of Cardiology7 (which were followed by a paper presented in June at the 77th Scientific Sessions of the ADA8), involved the use of more than 300,000 patient records from 6 countries to study the effect of SGLT2 inhibitors. Researchers found a 39% reduction in rate of hospitalization for HF and a 51% reduction in death for any cause.7
Bloomgarden explained the relationship between diabetes and HF. “Diabetes is associated with increased likelihood of atherscrotic cardiovascular disease,” he said. “And certainly, individuals who have myocardial damage have decreased heart function and are at risk of heart failure.”
What is less appreciated, he explained, is how diabetes leads to fibrosis and decreased heart function, and endocrinologists are aware of 2 types of HF in these patients: HF with reduced ejection fraction (HFrEF) and HF with preserved ejection fraction (HFpEF). “These are extremely important causes of morbidity and mortality,” Bloomgarden said. “People with heart failure feel tired. They have less energy, and then, eventually they progress to peripheral edema, dyspnea, and all the classic things we learned.
When more and more patients with diabetes come in with these symptoms, he said, “it becomes very attractive to say, we may have a specific drug that could be useful.”
Inzucchi noted that a secondary endpoint in EMPA-REG OUTCOME showed a 35% risk reduction for hospitalization for HF among patients with T2D with known coronary artery disease.1 The number of people with high-risk T2D and existing heart disease is “a real epidemic,” Inzucchi said. “I think our cardiology colleagues are getting so good at saving people during their acute coronary syndromes that many patients are now living with somewhat damaged ventricles. So, the 35% risk reduction, we saw that and we wondered whether this was something that occurred in patients with established heart failure or whether it was preventing heart failure episodes. And I think the answer is both.”
As a result, the EMPEROR studies are now under way for empagliflozin, which will examine the effect of the SGLT2 inhibitor specifically on patients with HFpEF and HFrEF.9,10 Results will come in 3 to 4 years. “These are heart failure trials being driven by heart failure experts,” Inzucchi said. “The heart failure community is very interested in this class because of the EMPA-REG signal, »
but the question is, will the benefits be seen again in the diabetic population?”
“Or in prediabetes?” Bloomgarden asked.
Big Data and CVD Benefits
Of course, clues to what the EMPEROR results may reveal were contained in CVD-REAL. Scanlon asked Gabbay to discuss these findings, as well as the importance of using data to gain these insights.
“It’s the beginning of what we’ll see a lot more of—using big data to try to answer some of these questions,” Gabbay said. “There are a lot of hypotheses. Is it a class effect? What’s the effect on congestive heart failure? We have some data, but not all the questions are answered. There are a number of studies that are ongoing. Some, we’ll get results soon, but some will still take several years. What do we do in the meantime?”
Gabbay explained that while CVD-REAL seemed to suggest a signal for reducing congestive heart failure across all SGLT2 inhibitors, the “challenge with a retrospective analysis is you’re not randomizing people to therapy.” Gabbay and Inzucchi agreed there was some value in observational studies—as statisticians can use propensity matching to make up for the loss of randomization—but Inzucchi said these studies should be taken “with a grain of salt.”
Bloomgarden, too, noted the need to watch for “channeling bias,” in observational studies—when patients ask to be put on the “new drug.” But he said that even the largest clinical trials have event rates that are so low there are questions that can’t be answered. Inzucchi said the question is whether patients in trials are different from those in the real world. “I think it’s great when the randomized clinical trials and the observational data sets point in the same direction, but when they don’t, I think it’s really confusing.”
Designing Better CV Outcomes Trials
CV outcomes trials started with one idea, “First, do no harm.” But, Scanlon asked, is it time for a redesign? Are they powered sufficiently? As large as they are, are they large enough? How long should patients be followed? Can data be collected retrospectively, and if so, for how long?
Bloomgarden said as “hugely expensive” as it would be to follow large numbers of patients for a decade or more, this must be balanced against the 650 million individuals worldwide who will develop diabetes by 2040. And trials shouldn’t just examine the effect of therapies or strategies on the highest-risk patients. “Let’s try to figure out how all people developing diabetes should be treated going forward,” he said.
Long term, the progression of diabetes makes it impractical to study a single therapy for an extended period, Inzucchi said. “It’s not clean like that,” he said, using an example of a study conducted in 2004 based on the approaches that were common at the time that resulted in “cross contamination” as patients needed additional therapies.
Snow, the payer, said there’s no chance that an analysis of big data will ever replace the role of the randomized clinical trial, “no matter how good it is.” However, he said, “we do know that there are certain situations where a randomized clinical trial just doesn’t work because the population is relatively homogenous. So, you’re stuck with the question of, ‘Well, can’t I expand this into other populations? Do I need a full, other randomized controlled trial to answer that question or not?’ ”
Observational studies can help with questions that would take a long time to answer, that would require studies of great complexity, or in cases in which there is great risk of patients dropping out of the study, Snow said.
“I totally agree,” Gabbay said. There are many questions
that need answers, and not every question will get a randomized controlled trial that collects data for 5 to 10 years. Practically speaking, there are patients who need treatment today.
“As big data [analysis] becomes more sophisticated,…and studies are done more accurately, we’re going to have to rely on that kind of data to answer some of the questions that there are unlikely to be clinical trials on,” Gabbay said. The challenge is that some will be well done and others will be poorly done, and the average provider reading an abstract won’t know the difference. The danger is that kind of data sways clinical care. I think a better arbitration of study technique for big data analysis will really help move the field forward.”
Clinical Decision Making and Cost–Effectiveness
Scanlon turned the discussion toward the future—of using data beyond 1 institution or health plan to mine data sets for insights from larger populations, so that clinicians gain a more balanced view than might otherwise happen if they are influenced by an outlier case. Snow said this requires cooperation between plans and providers.
“One of the hopes for the future is that we’ll be able to integrate that type of data, more effectively, into the data that we have through various relationships we have with the providers,” Snow said, “where we’re able to share that information and able to bring the power of the information that’s collected on the individual patient level—lab data, physical findings, etc—but also bring it to a level where we’re talking about not necessarily hundreds or thousands, but now, talking millions of folks that we’re looking at into the analysis.”
Coverage decisions, he said, start with the scientific evidence. “That’s separate, or divorced, from the cost (either the cost of the therapy or even the savings). Once the scientific data are established, that it’s effective, then the question (that) comes is, ‘What is that cost? And how is it going to fit into a benefits plan?’”
“And so, obviously, it’s about something that, in addition to being scientifically valid, also saves money. Well, that’s about as easy as it gets.
“Those that are scientifically valid and cost some money; those are very likely to still be approved. And those that are scientifically valid but cost a lot of money may still well be approved, but they may get more scrutiny to make sure they’re being utilized for the appropriate patient in the appropriate way.”
Snow acknowledged that the time element does enter the cost-effectiveness discussion—will the payer of today realize the benefit for an expensive therapy that may prevent costly events years into the future?
Gabbay said that is where “class effect” become important. Once a class of drugs is shown to have a benefit, payers may choose among different drugs based on price. “But if it turns out that there’s ambiguity there, and right now, we’re still in an area of some ambiguity, it makes that much more problematic.
“That’s really where I think we’ll have a sense, over the coming months, of whether studies now confirm that there’s a class effect or not. For most other drugs, that has been the case.”
The panelists concluded by discussing how this is an exciting time in diabetes care.
“Diabetes has always been one of those situations in medicine where there was just a very negative association with it,” Snow said. “It’s increasing in frequency. The prevalence of diabetes is increasing. Folks will develop microvascular complications…the news is always bad.”
But now, “We have slowly chipped away at the microvascular complications and we’ve chipped away at the macrovascular complications, and now we have even further agents that look like we’ll be able to chip away at this big chip much more. And so, we can really give our patients an upscale message that, yes, it’s diabetes, but you can live a long and healthy life despite having diabetes.”