This article is more than 5 years old. Information may no longer be current.
Expert panel discusses potential link between angiotensin receptor blockers, cancer
Click Here to Manage Email Alerts
A study published in 2010 in Lancet Oncology suggested an association between angiotensin receptor blockers and cancer. The results generated a great deal of discussion in both the cardiology and oncology communities.
To further explore the results of this study and their potential implications, Carl J. Pepine, MD, professor of medicine, division of cardiovascular medicine, University of Florida, moderated a round table discussion that included Rita Redberg, MD, of the University of California-San Francisco; Dan Simon, MD, of Case Western Reserve University; Udho Thadani, MD, of the University of Oklahoma and VA Medical Center; Joseph Alpert, MD, of the University of Arizona; Mark Pfeffer, MD, of Harvard Medical School; and Rhonda Cooper-DeHoff, PharmD,of the University of Florida, during the American Heart Association Scientific Sessions 2010 in Chicago this past November.
|
|
In part one of this two-part round table, the group discussed the results of the study and what it meant to their practices.
Carl J. Pepine, MD: We’re here today to gather your thoughts on the topic of cancer and angiotensin receptor blockers (ARBs), or perhaps to the broader area of RAAS inhibitors, and cancer. The reason for this was that Dr. Dan Simon’s group published a paper several months ago (Lancet Oncology) suggesting that there might be evidence that ARBs use is linked with the development of cancer. Their paper was met with a tremendous reaction from patients, particularly for those of us who have practices; in terms of questions about what should we do with the fact that our patients may be taking drugs that cause cancer. Dan, tell us about yourself and the report.
Dan Simon, MD: I’m the director of the Harrington McLaughlin Heart and Vascular Institute at University Hospital and the Herman K. Hellerstein Professor of Cardiovascular Research at Case Western Reserve University in Cleveland, and it is an honor and privilege to be speaking on behalf of Dr. Sipahi and Dr. Fang. The background for this study came from three data sets that were suggestive of a link between ARBs and cancer: the CHARM data set: one from ON-TARGET, and another from TRANSCEND. Overall we found an absolute increase of 1.2%; the genesis of this article was basically examining those three data sets.
Udho Thadani, MD: I’m Udho Thadani, professor emeritus (active) of medicine, division of cardiology at the University of Oklahoma, and consultant cardiologist at the OU Medical Center and VA Medical Center, Oklahoma City. I’ve been interested in CV drug therapy, and I was fascinated to see this article. My curiosity increased further when I started receiving calls from my patients who were taking these medications, so I read more about the topic.
Rita Redberg, MD: Rita Redberg, professor of medicine in the division of cardiology at UCSF and the editor of the Archives of Internal Medicine. I thought the Sipahi study was a well-done article, and it raises important questions.
|
|
Joseph Alpert, MD: Joseph Alpert, professor of medicine, University of Arizona, editor-in-chief of the American Journal of Medicine. I’m skeptical. First of all, association is not causality. Just because A is followed by B doesn’t mean A caused B.
Secondly, I have friends who are clinical scientists who say that, “Meta-analysis is to analysis as meta-physics is to physics,” so you have to be skeptical. The meta-analysis points you in a direction, but it’s not definitive proof.
Rhonda Cooper-DeHoff, PharmD: I’m Rhonda Cooper-DeHoff, associate professor at the University of Florida in the Colleges of Pharmacy and Medicine. When I saw this article, I was intrigued, to say the least. I did a literature search and it took me back 20 years to data looking at ACE inhibitors being protective from cancer. I thought, did I read that wrong? What did that say? Did it say ‘causes’ or ‘is associated with’? So that got me to look back at a lot of that basic science literature and I have to say, like Dr. Alpert, I’m a little bit skeptical. But I think these data should cause us to pause and think about the translation and all of the data in toto and what it really does mean for our patients.
Marc Pfeffer, MD: I’m Marc Pfeffer. I’m the Dzau Professor of Medicine at Harvard Medical School and a cardiologist at Brigham and Women’s Hospital, and I’ve been inhibiting the renin angiotensin system since I was a graduate student. And, of course, I’m interested in this and also presented the CHARM data in 2003 where we raised and thought we had answered the question. I didn’t hear anything about it for another 7 years and I have the greatest respect for the senior authors on this paper, both Dan Simon and Jim Fang. I know their integrity and their honesty and their robustness for science. But I also know that the field is bigger than they captured because my first glance at the article I said, “Well, where is VALIANT? Where is this trial?” So I knew that it wasn’t a complete story.
Pepine: So, let’s go back to Marc. You’ve had a long time to think about this. What could be the biologic mechanism by which ARBs might cause cancer?
Pfeffer: Well, you could say that about any molecule that somebody takes in their body. Yesterday, I talked about statins and I have faced this same question, which is important for the readers. Any drug we give can do anything. Statins have raised the same question with breast cancer, when these were in small numbers, but the signal went away with larger numbers. The lesson I had learned from that is if you over-interpret that potential risk, you might stop a sub-population from receiving the benefit of a therapy. So in that case, (in 1996), we would have called an end to women receiving statins because we would have said they caused breast cancer. Think of the 14 years since then and how women wouldn’t have benefited.
Pepine: So, Dan, is there a biologic plausibility? Is it possible that angiotensin-II is a naturally occurring anticancer agent that kills cells that are genetically awry before they can multiply, and that in using an ARB we destroy or upset that mechanism?
Simon: I would say two things. First is that one cannot lump ACE inhibition and ARBs together, so there’s no question in our own meta-analysis, which is under review for ACE inhibitors, that they are mutual in cancer in 69,000 patients. So, yes, there is biologic plausibility for unopposed angiotensin-II stimulation. Lung cancer is loaded with AT2 receptors and unopposed angiotensin-II stimulation is pro-angiogenic and pro-proliferative. But I have to say, biologic plausibility as a way to refute or agree with this paper is irrelevant because one can come up with biologically plausible mechanisms for anything. The point is the observation in made in those 69,000 patients.
|
|
In response to Marc, 2,057 ARB studies were reviewed for our analysis. If you look at the data for studies that were more than 1 year in duration, that had more than 100 patients, and that had cancer as a pre-specified endpoint (or as an endpoint that was later gathered), [the data] led to the number of 60,000 patients with cancer occurrence and 90,000 patients in cancer death. This wasn’t limited to published studies. It was exhaustive searches of FDA websites, e-mails to every PI of every trial which, as you know, led to some very unsavory interchanges regarding whether cancer data was collected or not. The fact of the matter is that this is the reason why we need the FDA and the European Medicine Agency (EMA) to collect patient-level data. Marc is right, there are a lot more data that have been filed, but you need to do time-to-event analyses, and the only people who can do that are at the FDA or the EMA. When you publish something like this, nobody is going to do a 100,000-patient randomized trial of ARBs vs. something else to see if there is a cancer link. The only better data that we’re going to get is the FDA and EMA to do a time-to-event analysis, and hopefully, that’s where we’ll come out. As you know, both, in the first 2 weeks of July, the FDA and the EMA announced that they were exactly doing that.
Pepine: Dan, did you have patient-level data in your meta-analysis?
Simon: No. I would say the only difference beyond publications is that we exhaustively searched FDA databases, so a lot of the data actually come out of company-filed databases that are not published.
Pepine: What do you think, Rita?
Redberg: This was a well-done and exhaustive review that raised a question about a commonly used drug class and not one that we don’t have alternatives for. We have ACE inhibitors; we have other drugs that treat most of what we’re using ARBs for. When a question is raised (and this is a lot of patients and from a lot of data), I feel like it’s my responsibility as a physician to raise the question with patients. I would seriously reconsider using a drug that has a serious unanswered question about risk when there are alternatives to these drugs. This points to the importance of having real post-marketing surveillance data, because we have so many people on these drugs and we really don’t know what the long-term effect is. We can’t answer this because in general our database is poor.
Pepine: Joe, what have you told your patients and has it changed your prescribing patterns?
Alpert: It has not changed it. I’ve seen people who are on ACE inhibitors with swollen faces and with microvasculitis and a host of stuff like that. I put them all on ARBs because they don’t feel a thing and it controls their BP effectively. My old boss, Lou Dexter, used to say, “It’s hard to make an asymptomatic patient feel better.” When you put somebody on antihypertensives and they feel worse, they don’t say thank you. So with the ARB, there are almost no side effects.
|
|
Redberg: What about hydrochlorothiazide (HCTZ)?
Alpert: There are a lot of side effects with HCTZ, particularly in males. There’s hypokalemia and erectile dysfunction, for example. There’s gout if they already have an elevated uric acid. If they’re a little fat, I worry about their glucose levels. People don’t like HCTZ.
I worry about their glucose intolerance, so I’m not changing. I need to see more data, which, if when the FDA and the European regulatory agencies review their data, comes out like that, then I’m going to have a heart-to-heart talk with my patients each time they come in on an ARB and tell them that there’s a small increased risk of cancer. I would then ask them if they want to try something else or do they want to stay with what they’ve got.
Pepine: Udho, what did you do with your patients?
Thadani: When I read this article, I was impressed because this is one of those occasional articles one reads in which the authors went to the FDA website to look at the previously submitted databases submitted to the FDA for different ARBs by pharmaceutical companies. Most of the published meta-analyses don’t do that, so I have to give the authors the credit that they went to the FDA website and looked at all the available databases and also contacted the authors of the published studies.
|
|
The only thing I found missing were some of the newer trials, including NAVIGATOR, which has been published (probably when the Sipahi review was in print). NAVIGATOR was conducted, however, in prediabetic patients. It is hard to criticize Dr. Simon and colleagues for the conclusions made in their meta-analysis. Of course, there always remains a possibility of uncertainty in any meta-analysis conclusions, and one can’t be absolutely sure of the true facts. When there’s a noise out there challenging published data, you have to keep that in mind and keep your mind open for future developments. If I see a patient with hypertension (being an investigator of the ALLHAT study), I’ll probably use chlorthalidone rather than HCTZ. Chlorthalidone is probably a better drug. It’s long-acting, patients tolerate it very well and it is cheaper than other drugs of a different class (such as calcium channel blockers and ACE inhibitors used to treat high BP). If it’s a black patient, the BP response to ACE inhibitors is not as good, so in that instance, I would stay away from that class of drugs (including the ARBs, initially), and most likely will add a calcium channel blocker to the diuretic therapy.
If it’s a patient with HF, then treatment decisions are more critical because these drugs (ACE inhibitors and ARBs) have been shown to save lives and prevent rehospitalizations for HF. In patients with HF, you have to balance the judgment, between the fact that these drugs save lives and re-hospitalization, while accepting the fact that there may be a small possibility of causing cancer in the future with these life-saving drugs. This is what I tell my patients.
Pepine: Marc, what did you tell your patients?
Pfeffer: Well, first of all, it’s usually other people’s patients. Friends call me up and say, “I have a friend who is stopping taking their medicine,” or something to that effect. I then say, “Please tell them not to stop taking the medicine and that this is very preliminary.” And then I have to actually say do I agree with the finding? So it starts with the hypothesis that came from CHARM, and I said, in 2003 when I presented that data and we did cardiovascular death, you look at non-cardiovascular death. We were very proud of CHARM overall. There was a reduction in cardiovascular death. There was no increase in non-cardiovascular death, and that is a huge thing. Ok, that’s the first safety net.
Next were noncardiovascular sub-subgroups. We found, with a significant P-value, that we had more cancer deaths. Now, we had 13 minutes to present this, and we spent 3 minutes talking about it. What did we do? We looked at new cancers. They were not increased.
We then went to the sponsor and said, “I’m standing up here representing CHARM, but you had to have more data than just CHARM. Give us everything you know.” And we put that into the manuscript and we put that into our presentation that we didn’t have, so I put that to rest in 2003.
Pepine: What do you think, Rhonda?
Cooper-DeHoff: In preparation for this round table, I found a review article published last year that reviewed the data to date before the Sipahi paper was published. After studying it, I would comment that it all depends on what you choose to include in your paper. This paper basically says that ACE inhibitors and ARBs in general are protective from cancer, but none of these papers get mentioned in the Sipahi paper, nor does this review even get referenced in this. So I came away feeling like you really have to do your own intense review to understand everything that is in the whole picture. I came away feeling like this paper, although well done, didn’t adequately present all of the information. The basic science data really doesn’t support this clinical finding.
Simon: This is critical because it gets to the issue of private databases and not having access to databases. Some things need clarification. The first is that RENAAL reported only cancers that resulted in discontinuation of the study drug. The second is that cancer data was not included in IDNT. Third, VALUE did not collect cancer data according to their PI but may have collected cancer data in retrospect. In response to the reviewers to say that the result was dated, they only put in a subset of cancer data. They didn’t put all their cancer data. They included only the people who were on ARBs who weren’t also on ACE inhibitors. When you include everybody who was on an ARB in the trial, it doesn’t affect the overall picture. OPTIMAL collected only cancer death data and no new cancer data. So this is a very common misconception, which is that one can say 60,000 to 90,000 patients isn’t enough and you need to include more studies. The problem is that when you get under the hood, this is the fact of these studies.
I think what will finally happen is the FDA and the EMA will now have all the cancer data. It will have private databases and it will go back to drug companies and say, “Send us your cancer data.” The problem is the misinformation that can happen when someone tries to include private-database cancer data (from VALUE, for example) on a trial and then it turns out that they’re only actually citing a subset of their cancer data. It’s highly selective.
For more information:
- Sipahi I. Lancet Oncol. 2010;11:627-636.
Disclosure: The participants in this round table report no relevant financial disclosures.