Heart-smart diabetes drugs: Pipe dream or possibility?

March 28, 2013


London, UK - Everyone remembers rosiglitazone (Avandia, GlaxoSmithKline), a diabetes drug finally withdrawn in Europe and severely restricted in the US in 2010 after it emerged that it could increase the risk of heart attack and cardiovascular death.

Debate about rosiglitazone had rumbled on for years, and prior to its demise, the US Food and Drug Administration (FDA) had already decreed in 2008 that, going forward, new type 2 diabetes drugs should be subject to more stringent clinical trials, to rule out cardiovascular toxicity rather than being approved on the basis of studies that simply showed than an agent lowered blood sugar. Last year, the European Medicines Agency (EMA) followed suit.

Given that CVD is the biggest cause of morbidity and mortality among those with diabetes, most diabetologists seem to agree that this decision is a "no brainer." But some had concerns that these regulatory requirements might have drawbacks. These included curtailment of drug development—deterring all but the biggest pharmaceutical companies from investing in diabetes programs—increasing cost of medicines to patients and the concern that focusing overly on cardiovascular risk might detract from other outcomes of interest in patients with diabetes.

Last month, at the Excellence in Diabetes meeting in Istanbul, doctors discussed what has happened in the five years since the FDA introduced this new requirement, and what they think the impact has been on diabetes drug development. Other key opinion leaders were also polled to see how they think this whole issue is panning out.

Was this regulation necessary?

Prior to the 2008 requirement, all that was needed for approval of a new diabetes drug was to show that it lowered an established surrogate marker for glycemic control, HbA1c. Better glycemic control is known to reduce the microvascular complications of diabetes such as neuropathy, retinopathy, and nephropathy, but unfortunately the impact of glucose lowering on cardiovascular outcomes is not as clear.

Dr Michael Nauck

Dr Michael Nauck (Diabeteszentrum Bad Lauterberg, Germany) said in an interview that it was time the diabetes world caught up with other disciplines. "If you look at cardiovascular medications to improve prognosis, every intervention is supported by one or two major trials. You would not give heparin, or a platelet agent, or a beta-blocker without knowing there is a very obvious benefit associated with this intervention. But this tradition has never been there in the field of diabetology, although basically we are dealing with the same topic—the number-one consequence of diabetes is cardiovascular disease—and in my view this has been dangerous."

Dr Richard Kahn (University of North Carolina, Chapel Hill) agrees. "You are not lowering glucose for glucose's sake, you are lowering glucose to prevent the complications of diabetes, and the most serious complication of diabetes is cardiovascular disease. If you have a pill that could worsen CVD, that is the worst of all worlds," he said.

You are lowering glucose to prevent the complications of diabetes, and the most serious . . . is cardiovascular disease.

Add to this the fact that it has been the rule, rather than the exception, that clinical trials in diabetes have produced findings that were not expected, and it is clear that this new guidance is necessary, said Nauck. "The [majority of] diabetes drugs were introduced into the market based [on the fact they] reduced HbA1c, but does that tell us that the long-term benefits are there? And that they outweigh the potential risks? I don't think we can be sure. So I think [these new requirements] are a good move."

Dr Michael Lincoff (Cleveland Clinic, OH), who is involved in diabetes drug development, has a similar take on the issue. "The requirement to do large-scale trials in populations that are huge and where the primary source of morbidity and mortality is cardiovascular disease, such as diabetes, is not really onerous," he said in an interview.

Were concerns about impact of changes justified?

Although most doctors agree that the regulatory changes are necessary, some have expressed concern about possible often unintended effects.

Dr Jennifer Green

"Obviously this is going to increase the cost of drug development, there's no question about that," Dr Jennifer Green (Duke University, Durham, NC) told the Excellence in Diabetes meeting. "And some worry that this may perhaps be feasible only for a small number of drug companies and in turn may have an adverse impact on drug development." It is also "very likely," she noted "that this will increase the direct cost to the patient of use of these medications."

Dr David Matthews (Oxford University, UK) said: "CV outcome is important, and I do believe that is absolutely what we should do. But the problem is you need huge investment in time, money, and [number of] patients, and there's a tendency to recruit patients who are going to have a higher risk of CV outcomes, which may not be the generality of the patients you want to treat. So then you've got a problem that very few pharmaceutical companies can do that and . . . [soon] you're going to only have one pharma company that can do this, which will be called 'Mega Pharm.' "

Dr David Matthews

Nauck said he agrees, to an extent, explaining that it may now take 10 years and cost hundreds of millions of euros to develop a new diabetes drug. He and Matthews suggest that possible extensions to the patent life of diabetes drugs could be considered as a way of encouraging companies to continue to develop new therapies for diabetes. "If we want better drugs, we will have to offer something to the companies, because investment is higher," said Nauck.

But Kahn believes most pharmaceutical companies can take CV-outcomes trials in their stride and notes that small companies have rarely developed diabetes drugs; most come from big pharma.

"The profit margin on these drugs is huge, so it's not like they can't afford a CV trial; such studies are routinely done for medicines related to heart disease. The pharmaceutical industry's net profit margin is probably the highest in the world with the possible exception of software," he asserts. And the idea that the price of drugs will need to go up or that patents should be extended, is a smokescreen, he said.

"The truth is they can lower their profit margin and still be quite successful, but they don't want to do that." Resisting these changes is corporate nature, he said: "You can't blame them for not wanting to spend more money to develop a drug when the ultimate goal is increasing shareholder profit."

I recognize that [the guidance] is not universally loved. But I think that this is somewhat of an alarmist view.

Lincoff added: "There are some who believe that the guidance is an overreaction; I recognize that it is not universally loved. But I think that this is somewhat of an alarmist view—that there are not going to be new drugs [for diabetes] developed. If this were a small orphan population or even an infrequent population, that might be the case, but this is a huge and growing population of patients, and there is clearly unmet need."

And what about the concern that the focus on CV risk might detract attention from other outcomes of interest in patients with diabetes? Nauck said there are other safety concerns with diabetes drugs—including their effects on malignancy rates, risk of pancreatitis, and bone metabolism/risk of fracture—but none is as important as cardiovascular outcomes in terms of the number of people that could be affected. Randomized clinical trials are likely unfeasible to look at these issues, because of the large number of patients that would be needed, he said, noting that other surveillance systems, such as registries, are needed to learn about these rarer side effects.

Initial signs: No letup in development of diabetes drugs

Initial indications, however, are that the pharmaceutical industry has not been deterred from developing diabetes drugs by these new regulations, rather the contrary.

"Resistance on the side of the pharmaceutical companies has been low; as soon as the FDA requested it, everybody did it, although it needed to be a requirement for that to happen," Nauck said.

Figures presented by Green in Istanbul also bear this out. "If you look at the number of ongoing trials designed to assess the CV safety of diabetes medications, there's a fairly robust field out there." She noted that there are currently 35 large trials under way with diabetes agents, both phase 3 and phase 4, which have major cardiovascular end points as primary outcomes.

The profit margin on these drugs is huge, so it's not like they can't afford a CV trial.

"Of these, only ORIGIN has been completed, so we've a large number of other trials designed to rigorously assess CV safety of diabetes medications. So this seems, at least at present, to have been embraced by the community."

Recent publications also suggest there are over 200 medications currently in development (clinical trials or awaiting approval by FDA) for the treatment of diabetes and related conditions, she added.

With diabetes affecting 10% of the world's population, and up to 20% of people in some countries, it would appear that the potential market is so huge and the lure so great to the pharmaceutical industry that these new regulatory requirements will do little to dampen enthusiasm.

Dr Michael Lincoff

Evidence of this can already be seen in sales of newer classes of diabetes agents, which appear very healthy, Green said: the DPP-4 inhibitor class ("gliptins") increased by almost 200% in 2009 - 2010, and glucagonlike peptide-1 (GLP-1) GLP-1 analog sales grew by 85% over the same period, she noted.

Lincoff said the requirement will also potentially have a positive side effect: "We still don't have any [diabetes] drug up until now that we know reduces cardiovascular events. This provides some impetus for companies to look for classes of drugs that, in addition to their glucose-lowering effect, might focus on the parts of diabetes that relate to atherogenicity. The regulatory requirement will drive a whole lot of informing about what we are doing with CV risk."

Any company that does hit on a glucose-lowering drug that also reduces cardiovascular outcomes—and some are looking at superiority in their CV-outcomes trials—will really hit the hit the jackpot, Lincoff says. "It will become a standard of care for reasons beyond the diabetes."

Regulation is to establish lack of cardiovascular toxicity

In Istanbul, Green outlined the FDA guidance. The actual requirement is for new diabetes drugs to establish lack of cardiovascular toxicity, and this is broken down into a two-stage process. In order to file a new drug application (NDA) for a diabetes drug, companies need to show that the upper bound of the two-sided 95% confidence interval for the estimated increased risk (hazard ratio) for CV events is <1.8.

If this number comes up between 1.3 and 1.8, and the overall risk/benefit analysis supports approval, a postmarketing outcomes trial will be necessary to definitively demonstrate that this figure is <1.3.

If, however, clinical data in the NDA show this number is <1.3 to start with, a postmarketing CV trial may not be necessary. Greater details on the requirements are provided in the FDA guidance and are also outlined in a May 2011 article in Diabetes Care.

Put simply, "you are talking about a maximum of 122 [CV] events to get to the first <1.8; it's a very small number, although, granted, the [larger CV-outcomes] trial has to be under way," Lincoff explains.

"If you come in with <1.8, you can apply for your NDA and then you would be approved provisionally with the idea that you are in the process of carrying out the larger trial, which needs to be large enough to rule out greater than 1.3, so for that you would need 600 to 650 events [if the effect of the diabetes drug is neutral on CV outcomes, ie, hazard ratio of 1.0]. If the drug is CV protective, you could get away with that earlier."

Overall, therefore, the FDA guidance translates into a larger and longer phase 3 clinical development program for diabetes drugs.

Asked if the drug companies are clear on what they need to do, Lincoff replied: "Yes, they are very clear, at least any of the ones I've dealt with. It's not that complicated from a statistician's standpoint."

What would you do with a drug that you knew was safe only in low-risk patients?

Another consequence of the new requirement is the shift in patient selection, with trials now needing to include individuals at high or higher CV risk than have traditionally been involved in diabetes trials, Green explained. The studies also need to go on long enough (minimum of two years) to assess the impact of such therapies on CV events, and the latter must be adjudicated by a designated committee, she noted.

Lincoff is unconcerned about this. "We use high-risk patients not because we think that that's all patients with diabetes, but because you couldn't do a trial practically if you used a low-risk group." But also, "What would you do with a drug that you knew was safe only in low-risk patients? At what point do you say, Ah, you've become too high risk for me to treat with this drug? I have to take you off it'?"

No precedents set yet

Examining diabetes drugs that have recently come onto the market also gives some indication of the impact, so far, of these new regulations. Saxagliptin (Onglyza, AstraZeneca), a DPP-4 inhibitor, was the first new diabetes drug to receive FDA approval after the issuance of the new guidelines, in July 2009. Its clinical development program had been completed before the guidance, but because of the new rule, the company had to perform a postmarketing trial to demonstrate lack of CV toxicity.

In this case, AstraZeneca decided to aim high. Its SAVOR-TIMI 53 study, the first example of a postapproval commitment under the new guidance, is a five-year trial examining whether saxagliptin compared with placebo reduces CV events in patients with type 2 diabetes. It has enrolled approximately 12 000 patients and is expected to complete this year.

In fact, said Green in Istanbul, a meta-analysis of DPP-4 inhibitors to date suggests a protective effect of this drug class in cardiovascular terms, with an odds ratio of 0.71 for major adverse cardiovascular events (MACE), so the results of SAVOR-TIMI 53 are eagerly awaited.

Similarly, analysis of phase 2 and 3 trials for the GLP-1 agonist liraglutide (Victoza, Novo Nordisk), which was approved in the US in January 2010, showed it met the standard for ruling out unacceptable CV risk, but a postapproval safety study was still required. The resulting LEADER trial is comparing liraglutide with placebo for time to first occurrence of MACE and is expected to complete in 2016.

The most recent diabetes medication to be considered by the FDA is canagliflozin (Invokana, Janssen Research & Development), the first in a new drug class, inhibitors of sodium glucose cotransporter 2 (SGLT2).

Although recommended for approval by the FDA's advisory committee in January, panel members also voted 8 to 7 that they had concerns about the cardiovascular safety of the drug. Although the hazard ratio for MACE-plus (a composite end point consisting of the following adjudicated events: cardiovascular death, nonfatal MI, nonfatal stroke, and hospitalization for unstable angina) was 0.91 in the data the panel considered, there was a concern about stroke, risk of which was greatest during the first 30 days.

I don't think there is enough precedence to know how the FDA is going to handle these . . . it's going to depend upon how clean the numbers look in terms of the CV risk.

Panel members said data from the ongoing cardiovascular-outcomes trial, CANVAS , will be needed to determine the true risk, but this trial is not expected to complete until 2018. Whether the FDA will approve canagliflozin in the meantime remains to be seen; a decision was expected as this article was published.

"I don't think there is enough precedence to know how the FDA is going to handle these independently, I think it's going to depend upon how clean the numbers look in terms of the CV risk they've got," Lincoff said. "None of the ones that are currently under way with the CV trials have been knocked out yet, and the meta-analysis for the DPP-4 inhibitors looks pretty good," he added.

But still a way to go . . . 

However, a compound from the same drug class as canagliflozin, dapagliflozin (Bristol-Myers Squibb/AstraZeneca), had a hazard ratio of 0.67 for MACE but was rejected by the FDA in January 2012 because of concerns about a cancer signal. Dapagliflozin was nevertheless approved in the European Union later that year, in November 2012.

Matthews said this illustrates what he believes has become "a very defensive view" from regulators. "Given the same data, the FDA and EMA come to different conclusions about what the advice should be, and I think that highlights to us that a lot of what goes on in regulation comes back to opinion. We have to make some sort of decision, we hope that it's underpinned by data, but the problem is the data never tell us exactly and definitively what we want to know.

"What one needs is some kind of equipoise, whereby we keep a proper view about the balance of probabilities of getting one problem against the benefits you will accrue in the opposite direction," he said.

I think we need to move away from this black-and-white determination of drug safety.

Green concurred: "We've seen a tendency lately for regulators to become overconservative so that they don't suffer the repercussions of approving a medication that may have unanticipated consequences. I think we need to move away from this black-and-white determination of drug safety. It's almost like [drugs in development] are in a race and then at some point they cross the finish line and they are available for use, and that's the end of the story."

Attitudes need to change, she said. And although there is encouraging news in terms of the number of CV-outcomes trials under way with diabetes drugs, the vast majority of trials of these agents still fall short, she noted.

For example, of the approximately 2500 studies registered on clinical trials.gov for diabetes and diabetes-related outcomes between 2007 and the end of 2010, most don't enroll many patients, they take place at a single site, they do not last for very long, they exclude extremes of age, and they are not really focused upon complications of significant or emerging clinical interest, she explained.

"So I'm not sure whether they will really enhance our understanding of drug safety. For the vast majority of classes of diabetes medications, the impact of therapies on cardiovascular-event rates for the most part remains largely unknown, and there is general recognition that this is not acceptable.

"We need to move toward a more comprehensive life-cycle approach, to the assessment of drug safety, both in diabetes and other areas, where there are periodic assessments that occur in a way that is standardized and fair to all the drugs that are available for use, and I'm not sure we really understand how to do that at present," she concluded.

Green reports institutional grant support from Merck and Amylin, and receiving compensation for lectures from Merck and the American Association of Diabetes Educators. Kahn reports no conflicts of interest. Matthews reports advisory board consulting fees or honoraria from Novo Nordisk, GlaxoSmithKline, Novartis, Eli Lilly, Sanofi, Johnson & Johnson, and Servier. He also currently receives research support from Johnson & Johnson and lectures for Novo Nordisk, Servier, Sanofi, Eli Lilly, and Novartis. Nauck reports being on advisory boards for Amylin Pharmaceuticals, Berlin Chemie, Boehringer Ingelheim, Eli Lilly, Hoffman-La Roche, Intarcia Therapeutics, Janssen, Merck, Sharp & Dohme, Novo Nordisk, Sanofi, Takeda, and Versartis. He reports consulting fees from Amylin Pharmaceuticals, AstraZeneca, Berlin Chemie, Boehringer Ingelheim, Bristol-Myers Squibb, Diartis Pharmaceuticals, Eli Lilly, Hoffman-La Roche, GlaxoSmithKline, MannKind, Merck, Sharp & Dohme (Germany and US), Novo Nordisk, Sanofi, Takeda, and Wyeth Research. He receives research support from AstraZeneca, Berlin Chemie, Boehringer Ingelheim, Eli Lilly, GlaxoSmithKline, Merck, Sharp & Dohme, MetaCure, Novartis, Novo Nordisk, Roche Pharma, and Tolerx. Lincoff reports he is principal investigator for two trials of diabetes agents: the ALECARDIO trial with aleglitazar (Roche) and the GRAND-306 study with TAK-875 (Takeda).


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.