EMA to Review ARBs and Cancer, Infuriating Experts, Who Point to Missing Data and Adverse Consequences

June 25, 2010

June 25, 2010 (London, United Kingdom and Oslo, Norway)— The European Medicines Agency (EMA) has just announced plans to review the possible risk of cancer in patients taking angiotensin-receptor blockers (ARBs) [1], based on a meta-analysis in Lancet Oncology last week, which suggested a "modest but significant" link between ARBs and cancer [2].

At the same time, hypertension experts are up in arms about the paper, calling it deeply flawed. They contend that it is extremely unlikely that ARBs are associated with an increased risk of cancer--one physician even says he has data that, if added to the analysis, would wipe out any cancer signal. Overall, they say, the media splash this paper has made could cause irreparable damage to the reputation of ARBs, which they consider vital tools in their armamentarium, and may unfortunately prompt many patients to stop taking the medicines, putting themselves at increased risk of cardiovascular and renal events.

Specifically, the meta-analysis found an increased risk of cancer of between 8% and 11% associated with ARBs, based primarily on data with telmisartan (Micardis, Boehringer Ingelheim), which included a significant 25% increased risk of lung cancer, but no link to breast or prostate cancer.

The lead author of the meta-analysis, Dr Ilke Sipahi (University Hospitals Case Medical Center, Cleveland, OH), for his part, acknowledges that any cancer risk, if proven, would be small for the individual, but on a population level could be associated with up to 100 000 cancers, given that around 10 million patients are taking ARBs, in a market worth $25 billion, he says. "The bottom line is that our data are solid, and we are calling for a review by the FDA because there is so much at stake here. The FDA has to look into this, they have the ball now," he told heartwire .

To Publish or Not to Publish, That Is the Question

heartwire asked a number of attendees at the European Society of Hypertension (ESH) European Meeting on Hypertension 2010 in Oslo this week about the Lancet Oncology paper; several responded with expletives. Many questioned whether the study should have been published at all and were critical of the methodology; in addition, there is no clear pathophysiology to explain any ARB-cancer link, they say.

I'm lacking words here, it's a very bad example of science.

Dr Hermann Haller (Hannover Medical School, Germany) said: "I'm lacking words here, it's a very bad example of science, and Lancet [Oncology] should not have accepted the paper. Basic questions are not answered; we do not know about the other risk factors; it's an analysis that is very disturbing."

ESH secretary Dr Sverre Kjeldsen (University of Oslo, Norway) told heartwire : "It's a very bad paper; it selected only a few articles. It was rejected by the Lancet, it went to Lancet Oncology, and nobody can explain why they accepted it. . . . This is very strange."

Kjeldsen says he even has suspicions that the paper was planted by competitors who may have a vested interest in knocking ARBs from their pedestal, given that they are "becoming generic and cheap," with losartan the first ARB to have come off patent. People on the editorial board of many primary journals "are bombarded with so-called meta-analyses and reviews all the time," he told heartwire , and "their network for catching and refusing noise and rubbish occasionally leaks."

Dr George Bakris

American Society of Hypertension president Dr George Bakris (University of Chicago Pritzker School of Medicine, IL) called the analysis "very skewed. One has to be extremely cautious, and the truth is, I found it very disturbing that the person who is the first author--Sipahi--was the student of the person who wrote the editorial--Dr Steven E Nissen [Cleveland Clinic, OH]--and they are both in the same town. . . . I am a little concerned and would fault Lancet [Oncology] for not being more discriminating in who they got to write their editorial."

Others, however, feel that the information in the meta-analysis has a right to be in the public domain. Dr Lars Lindholm (Umea University Hospital, Sweden) told heartwire : "In contrast to many others, I don't want to ridicule the journal or anyone. The Lancet didn't want it and so sent it off to one of its daughter journals, which is a good journal. I know it was reviewed by four reviewers and sailed through, and they put a lot of caveats into it. There is a report on the table; data like this should always be taken seriously and investigated; we ought to look at them."

Dr Ingrid Os

And Dr Murray Esler (Baker Heart Research Institute, Melbourne, Australia) said: "It's caused a huge splash for obvious reasons. There's a concerted effort to just wipe the board with it and say it means nothing, but in my view it was actually well conducted, because it was a meta-analysis and the emphasis was on trials in which cancer was a specified end point, which is very important. And the data are very worrying. An 11% higher incidence [of cancer] is not trifling if it's true."

Nephrologist Dr Ingrid Os (Ullevaal University Hospital, Oslo, Norway) told heartwire : "They need to do some more studies to see if this is a true finding. But you cannot ignore that they have found it, so you have to respond [to] it in some way."

Adding VALUE Would Completely Cancel Out Cancer Signal

Kjeldsen told heartwire that in his opinion, the authors of the paper "messed up" by omitting several important ARB trials. "I am now copied on emails from people who will do a proper meta-analysis just to rest this case, like previous debates about diuretics, beta blockers, [calcium-channel blockers] CCBs--[which] we put to rest with cancer data from NORDIL, INSIGHT, STOP-2, and ASCOT--and ACE inhibitors."

Kjeldsen himself has data to add to the mix. He was the coordinator of one large ARB trial, VALUE--with valsartan (Diovan, Novartis)--which was not included in the current meta-analysis. In fact, in VALUE, there were 81 fewer cancers in the ARB (valsartan) arm than in the CCB arm, Kjeldsen noted. He and two colleagues have now submitted a letter to Lancet Oncology [3] stating that if the cancer cases from VALUE are added to the equation, any cancer signal with ARBs is abolished, with a new cancer diagnosis rate of 7.7% in the ARB monotherapy group vs 7.7% with other drugs. "In other words, the [cancer] effect is completely canceled out," he stated.

In other words, the [cancer] effect is completely canceled out.

At the ESH meeting, Kjeldsen and others made much of the fact that Sipahi et al did not contact the VALUE investigators for their cancer data, a fact they repeat in their letter to the Lancet Oncology. But Sipahi, who showed heartwire the emails to prove it, says: "Although I did not systematically contact the investigators of all the ARB trials, VALUE was the exception because it was so large that it was difficult to ignore, and I was curious. I contacted the primary investigator, Dr Stevo Julius [University of Michigan, Ann Arbor], who replied that the incidence of neoplasms was not monitored in VALUE."

Dr Stevo Julius

Julius told heartwire he does not recall this specific correspondence, but he does receive numerous such requests. Reporting cancer data is now mandatory by agencies such as the FDA, he says, but he noted, "I did not have the data; the study files were kept by Novartis."

Sipahi says: "If companies or other people have favorable results [with regard to cancer and ARBs] they can disclose these data, and if it's not favorable they are not going to disclose it." He adds that that his analysis focused on the only three ARB trials in which cancer was a prespecified end point, which did not include VALUE.

But Kjeldsen maintains that Sipahi and colleagues should have tried much harder to obtain the cancer statistics from VALUE--which he says would have been provided--and from other ARB trials as well.

Dr Michael A Weber

"The conduct of a meta-analysis demands a rigorous search for all the relevant data. This obligation is not satisfied by a cursory review of the literature but instead requires that those performing the meta-analysis proactively make every effort to obtain information known to be germane to the process," he, Julius, and Dr Michael A Weber (State University of New York, Brooklyn) say in their letter to Lancet Oncology.

"It is our strong understanding that investigators responsible for published clinical trials will respond positively to requests for additional information when important clinical issues are being explored," they add.

Is the Cancer Risk With ARBs Feasible?

Most doctors speaking with heartwire at the ESH meeting, even if they did not object to the publication of the meta-analysis, said there are methodological problems with it, including the fact that the trials included were designed to look at cardiovascular disease outcomes, not cancer; the latter limitation was acknowledged by the Lancet Oncology authors and the editorialist, however.

Also, from a pathophysiological viewpoint, the finding in the meta-analysis of increased cancer with ARBs in such a short time period--two to four years--does not make sense: it takes around 10 years to develop lung and other cancers from smoking, the hypertension doctors argue. Some said they had discussed the findings with oncologists, who think the concept is farfetched.

Lindholmtold heartwire : "Always be careful when you have studies aimed at studying something, and all of a sudden you decide to look at something different. . . . There is no pathophysiology, really, and when you speak to oncologists, they laugh at us, and say 'Do you really believe that between two and four years after taking a tablet you will have a full-size lung cancer?' "

There is no pathophysiology, really, and when you speak to oncologists, they laugh at us.

A lung-cancer expert who spoke with heartwire on publication of the data last week, Dr Michael D Peake (Glenfield Hospital, Leicester, UK), said the risk of lung cancer is around 200% to 300% higher in a smoker than in a nonsmoker and in this context, the increase in risk of lung carcinoma observed with ARBs in this meta-analysis--25%--is relatively small. Nevertheless, Peake agreed that even such a small risk, if applied to large numbers of people, "might be quite important."

Fears That Patients Will Stop Therapy

ARBs and ACE inhibitors are really a must for the type of patients I am treating, and if they stop taking the medication, [it] is frustrating.

Most of all, clinicians treating hypertension are concerned that patients will stop taking ARBs because of this scare, despite the fact that Sipahi and coauthors urged people to stay on these medications, pending further research and discussion with their doctors.

Os says: "ARBs and ACE inhibitors are really a must for the type of patients I am treating, and if they stop taking the medication, of course that is frustrating, especially if you try to argue against it, because what is in the newspapers is 'always true,' " she quipped. "It's difficult to fight against it; it's a worry."

Dr Björn Dahlöf

Dr Björn Dahlöf (Sahlgrenska University Hospital, Ostra, Sweden) said he fears unnecessary scaremongering and a rerun of past scenarios, such as the panic with calcium-channel blockers during the 1990s, "which it was calculated caused around 10 000 events or deaths because people stopped treatment, so I would urge people not to do that."

Esler said: "I think a lot of what happens in the short term will be driven by patient fears, because it's very worrying to patients, and they are going to ask their doctors or cardiologists, 'Should I stop this?' And I would say, 'It's somewhat worrying, I'm not sure about it.' Is this going to be definitive? If it is, it will obviously influence prescribing. It concerns me, but I haven't quite made my mind up about it.

"This is a matter that is going to be determined by the regulators; they are not going to neglect this, and they are not going to take any immediate denial of the importance of this, they are going to review it," Esler added.

More Cautious Use of ARBs?

I think that's a reasonable approach, to say we should be using ACE inhibitors more.

In their meta-analysis, Sipahi et al suggest that physicians use ARBs more cautiously until additional research is performed. Nissen concurred in his editorial [4], suggesting more judicious use of this drug class, which he said is "often overprescribed as a result of aggressive marketing" and should be used "with greater caution," reserved for patients with intolerance to ACE inhibitors.

Esler said Nissen's advice is sensible: "I think that's a reasonable approach, to say we should be using ACE inhibitors more than we have been."

But Kjeldsen disagrees: "The ARBs are becoming the most preferred antihypertensive drugs because they are exceptionally well tolerated; now, when becoming generic and cheap, they will undoubtedly be the number-one drugs on the market worldwide."

Bakris, Haller, Esler, and Kjeldse have consulted for or received grants, research support, or speakers' fees from companies that market angiotensin-receptor blockers. Julius reports no conflicts of interest. Dahlof, Lindholm, and Os had not responded to heartwire 's requests for disclosures when this story was published.