New Meta-Analysis Reassuring on ARBs and Cancer, But Questions Remain

November 30, 2010

November 30, 2010 (New York, New York) — A new meta-analysis of almost 325 000 individuals from 70 clinical trials has found no excess risk of cancer or cancer death with any single antihypertensive drug [1]. The results counter the conclusions of a review published earlier this summer--by Dr Ilke Sipahi (University Hospitals Case Medical Center, Cleveland, OH) and colleagues--which indicated that angiotensin-receptor blockers (ARBs) were associated with a modest increased risk of developing cancer.

"Our new paper refutes the findings from the analysis by Sipahi et al," lead author Dr Sripal Bangalore (New York University School of Medicine, New York) told heartwire , "and it extends [this reassurance] to other classes of antihypertensive agents."

Our new paper refutes the findings from the analysis by Sipahi et al.

However, a statistically significant increase in cancers among patients taking an ARB in combination with an ACE inhibitor of roughly 10% means the authors "cannot rule out" an increased risk of cancer with the combination of these two medications.

In an accompanying comment [2], hypertension experts Drs Lars H Lindholm and Bo Carlberg (Umea University, Sweden) say: "Bangalore and colleagues' comprehensive analyses do not lend support to the hypothesis that any of the blood-pressure–lowering drug classes promote cancer." With regard to the possible signal with the concomitant use of an ACE inhibitor and ARB, they note that this "is not a preferred combination for long-term treatment of high blood pressure."

Others feel differently. Sipahi says there are methodological reasons why this new study comes up with different findings from his own, and Dr Steven E Nissen (Cleveland Clinic, OH), who wrote an editorial accompanying the Sipahi paper, told heartwire : "It hasn't completely put to rest the question about these drugs. We are in an area where the evidence isn't strong enough to warrant regulatory action but where it does warrant looking very carefully." He urged regulatory bodies to complete their review of this issue "promptly."

The Sipahi paper prompted a safety review by the European Medicines Agency and the US FDA [3,4]. FDA spokesperson Sandy Walsh told heartwire : "We have asked for data from the ARB drug sponsors. Companies have submitted data to us and the data are currently under review. We will review all of the data in total and then communicate any follow-up information. There is not yet a certain timeframe for the completion of that review."

Follow-Up May Not Be Long Enough to See Cancer Signal

The meta-analysis by Sipahi et al found an 8% to 11% increased risk of cancer with ARBs, which was mainly driven by a significant 25% increased risk of lung cancer. At the time, Sipahi acknowledged that this risk was small for the individual but could be important at a population level. Dr Michael D Peake (Glenfield Hospital, Leicester, UK), a spokesperson for the UK Lung Cancer Coalition, agreed; at the time, he told heartwire : "It's a relatively small risk, but if applied to large numbers of people, it might be quite important."

It hasn't completely put to rest the question about these drugs.

The Sipahi paper caused a huge furor in the hypertension community, with leading figures calling it deeply flawed, and some arguing that it should not even have been published. There were grave concerns that the media attention surrounding the article would lead to patients stopping their antihypertensive therapy, putting themselves at increased risk of cardiovascular and renal events.

The new meta-analysis does not specify cancer types but is much bigger than the one performed by Sipahi et al. "Although [Bangalore et al] did not replicate the ARB evidence that Sipahi et al reported, they do show evidence that the combination of ACE inhibitor and ARB may be associated with an increased risk of cancer," says Nissen. "So now you've got a second, much larger, study that raises some questions about the class, since ACE inhibitors alone are pretty definitively known not to induce an increase in malignancies."

Nissen says it's also important to remember that "patients may take these drugs for decades, so if there is a risk, it may emerge later on."

It's very difficult to know what this means. We don't know the exact latency time for lung cancer.

Bangalore agrees. He told heartwire the mean follow-up period in this meta-analysis was 3.5 years, with a treatment duration range of one to nine years. "These drugs are given for many decades, and I think we should be really careful in interpreting the results, because cancer is not going to develop overnight." However, he says he believes the findings "provide reassurance to patients and physicians."

Peake, too, has reservations about the time frame, both of this study and the meta-analysis of Sipahi et al, which had an average follow-up of four years. "It's very difficult to know what this means. We don't know the exact latency time for lung cancer, but it is probably in the region of 15 to 25 years," he commented.

No Evidence of Increase in Relative Risk of Cancer With Any Class

Bangalore et al include 70 randomized controlled trials with 324 168 participants in their new meta-analysis, which Bangalore points out uses three different methods of analysis and is therefore "more robust" than the one Sipahi et al employed.

In the network meta-analysis (fixed-effect model), they found no difference in the risk of cancer with any antihypertensive agent vs placebo. They did, however, find an increased risk of cancer with the combination of ACE inhibitor plus ARB (odds ratio 1.14 compared with placebo), but this risk was not apparent in the random-effects model, they note.

In direct-comparison meta-analyses, similar results were recorded for all antihypertensive drug classes, except for an increased risk of cancer with the ACE-inhibitor/ARB combination (odds ratio 1.14; p=0.004) and with calcium-channel blockers (CCBs; OR 1.06; p=0.02). However, they noted no significant differences in cancer-related mortality.

Risk of Cancer and Cancer Death With Different Antihypertensives, Placebo and Controls

Drug class Cancer proportion, % (fixed-effect model) Odds ratio* Cancer-related death, % (fixed-effect model) Odds ratio*
Placebo 2.02 1.00 1.32 1.00
ARBs 2.04 1.01 1.33 1.00
ACE inhibitors 2.03 1.00 1.25 0.95
Beta blockers 1.97 0.97 1.23 0.93
CCBs 2.11 1.05 1.27 0.96
Diuretics 2.02 1.00 1.30 0.98
ACE inhibitors+ARBs 2.30 1.14 1.45 1.10
Other controls 1.95 0.97 1.43 1.08
*vs placebo

In a third look, on the basis of trial sequential analysis, "our results suggest no evidence of even a 5% to 10% relative risk [RR] increase of cancer and cancer-related deaths with any individual class of antihypertensive drug studied," the researchers note.

But for the ACE-inhibitor/ARB combination, the figures, "driven largely by the ONTARGET trial," suggest "firm evidence" of at least a 10% RR increase in cancer risk, they note. This finding "has to be interpreted in the context of the short duration of follow-up of these trials," they note. Bangalore added to heartwire : "It is a signal that needs to be followed up."

In their comment, Lindholm and Carlberg say that although the ACE-inhibitor/ARB combination is little used in hypertension, it is "frequently prescribed for patients with severe heart failure." But the expected survival time of such patients "is short," they observe, and "ruling out cancer risk is a difficult task to undertake, since elderly patients are likely to die of either cancer or of cardiovascular disease."

Cancer Underreported in Some Trials; Indirect Meta-Analysis Inappropriate

Sipahi says there a number of reasons that Bangalore et al got different results: inclusion of trials where cancer was not a prespecified end point, which in some cases led to "massive underreporting of cancers, [which] skews the findings"; the addition of the VALUE trial, because nobody in this study was taking valsartan at the clinically recommended target dose of 320 mg, and "to study the safety of a drug, patients actually have to be taking the drugs . . . at recommended doses"; and finally, Bangalore and colleagues mainly use the approach of indirect meta-analysis. The latter takes as a major assumption the fact that all trials have to be homogenous in several regards. "The trials that they include span a period from 1974 to 2008, with huge variations in most study characteristics. Indirect meta-analysis is not an appropriate method in this context to say that ARBs are safe," Sipahi asserts.

Indirect meta-analysis is not an appropriate method in this context, to say that ARBs are safe.

In response, Bangalore told heartwire that while he and his colleagues have "clearly identified some of these [same issues] in our study limitations, we do not completely agree with the viewpoint of Sipahi. Our analyses and conclusion are not based on one single analysis--as we have clearly pointed out, we looked for consistency of effect with the [three] different methodologies used--as well as analysis based on trials where cancer was a prespecified end point. These analytic techniques offer complementary information.

No matter how you slice or dice the data, we did not see a consistent increase in the risk of cancer with . . .  antihypertensive agents.

"No matter how you slice or dice the data, we did not see a consistent increase in the risk of cancer with the currently used antihypertensive agents," Bangalore stresses. "While there is underreporting of cancer in trials where cancer was a not a prespecified end point, the biases and underreporting are likely similar across the arms of a randomized trial. And direct comparison analysis is also not without any drawbacks," he notes.

Nissen concludes: "The FDA needs to look at everything, they are the only people that have all of the data; not every trial is published, and not all trials report cancer data--these authors were able to include only those studies that they had access to data for, and that's not everything. The burden is now on the regulatory community to answer the question of whether or not there is a signal here."

Bangalore has no conflicts of interest. Disclosures for the coauthors are listed in the paper; none of the authors received any compensation for their work on this report. Lindholm is past president of the International Society of Hypertension; Carlberg has no disclosures. Sipahi has received an educational grant and lecture honoraria from Pfizer and lecture honoraria from AstraZeneca and Ranbaxy. Peake reports no conflicts of interest. Nissen reports receiving support for clinical trials from Pfizer, AstraZeneca, Novartis, Novo Nordisk, Roche, Daiichi-Sankyo, Takeda, Sanofi-Aventis, Resverlogix, and Eli Lilly. He consults for many pharmaceutical companies but requires them to donate all honoraria or consulting fees directly to charity so that he receives neither income nor tax deductions.


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.