POISE published, debate on perioperative beta blockers continues

May 13, 2008

Hamilton, ON - Publication of the landmark Perioperative Ischemic Evaluation (POISE) trial online May 12, 2008 in the Lancet has triggered another heated debate about the pros and cons of using beta blockers perioperatively in noncardiac surgery [1]. In one corner are the authors of POISE, which was first reported at the AHA 2007 Scientific Sessions and which found that beta blockers do more harm than good in this setting; and in the other, two accompanying editorialists, who suggest that it is the POISE protocol—specifically, the doses of beta blocker used and timing of initiation of therapy—that explains the findings and that a different protocol might shift the risk/balance back in favor of using beta blockers in this setting [2].

"What POISE says is that in the dosing we used, we see beta blockers have substantial risk in the perioperative setting," Dr Philip J Devereaux (McMaster University, Hamilton, ON) told heart wire . "And until someone demonstrates with a clear and large randomized controlled trial that an alternative dose is both effective and safe, it's just not rational, not in people's best interests, to be assuming—that's how we got into this trouble in the first place."

The POISE study is absolutely monumental in terms of its scope and its intentions.

Sitting on the fence, but somewhat closer to the POISE investigators than the editorialists, is leading anesthesiologist Dr Martin London (University of California, San Francisco and San Francisco VA Medical Center), who has recently penned an editorial on this very subject in the April 2008 issue of Anesthesia and Analgesia[3]. He told heart wire : "The POISE study is absolutely monumental in terms of its scope and its intentions. These guys deserve a lot of credit for getting to this point when nobody else even got close. I think what we are going to see is any kind of wild enthusiasm about making sure everyone is on beta blockers—that's clearly going to go away. I think the POISE study will be very powerful in that respect."

However, London says he will continue to use beta blockers in certain patients, and he believes that the underlying, chronic risk of the patient is key to determining who should get beta blockers. "You can do risk stratification. The problem is no one has put together a good risk-stratification routine that deals with all the outcomes. We need a meeting of many minds to sort out this problem."

Deaths caused by shock

As background, Devereaux explains that the guidelines that recommend use of beta blockers in the perioperative setting were based on previous studies that were inconclusive and small—only a few hundred patients—and showed conflicting results. "In those trials, beta blockers looked like a panacea, so there's been a big push to make this happen."

At times we are convinced by small trials that something does benefit, but lo and behold, we do a large trial and discover that, rather than preventing, we are causing.

But he warns: "If even only 10% of physicians followed these guidelines—which incidentally in the US are used in quality assessments, where you have people going around ranking hospitals in terms of whether or not they are giving perioperative beta blockers—and if the POISE data are true, then in the past decade 800 000 people would have died prematurely and 500 000 would have had a major stroke perioperatively because we gave beta blockers. I think this is a very similar signal to WHI on [hormone replacement therapy]—at times we are convinced by small trials that something does benefit, but lo and behold, we do a large trial and discover that, rather than preventing, we are causing."

Devereaux first reported POISE—a randomized controlled trial in more than 8000 patients undergoing noncardiac surgery who were not on beta blockers, randomized to either the beta blocker metoprolol or placebo—at the AHA meeting last year. The results showed that the beta blocker reduced the risk of MI but increased the risk of severe stroke and overall death in patients undergoing noncardiac surgery. It suggested that for every 1000 patients treated, metoprolol would prevent 15 MIs, but there would be an excess of eight deaths and five severe disabling strokes.

Devereaux told heart wire that the new analysis featured in the Lancet "helps to explain why death went up and stroke went up [with metoprolol]. Death was clearly driven by a hypotensive state, leading to shock, which we've recognized is so common in the perioperative setting, and that's what tipped the balance and why we saw the excess death. Also it's not that simple to predict who will develop shock, and many people who are going to get it are the same people who are going to get a heart attack also."

He and his coinvestigators conclude: "Patients are unlikely to accept the risk associated with perioperative extended-release metoprolol. Current perioperative guidelines that recommend beta-blocker therapy to patients undergoing noncardiac surgery should reconsider their recommendations in light of these findings."

Beta blockers not radioactive, but perioperative environment is special

At the time of the presentation, as reported by heartwire , many experts agreed with Devereaux's conclusions. However, others argued that there were many issues to be resolved with regard to the dosing used in the study, titration of drug, etc.

The editorial accompanying the paper, by Dr Lee A Fleis her (University of Pennsylvania School of Medicine, Philadelphia) and Dr Don Poldermans (Erasmus Medical Center, Rotterdam, the Netherlands), suggests that a lower dose of a beta blocker—started a week before surgery rather than the two to four hours before, as in POISE—may still help. They recommend that beta blockers can still be used cautiously "in high-risk patients with proven or suspected coronary artery disease, preferably supervised by physicians who have experience with perioperative hemodynamics."

But this conclusion is based almost entirely on the results of one small trial in the DECREASE series—which is Poldermans's own trial—says Devereaux. Also, he adds, Fleisher and Poldermans were involved in writing the guidelines on perioperative beta-blocker use. "I have empathy for them," Devereaux says. "Obviously, it's an uncomfortable situation to have been the one making these recommendations and then along come the POISE data, but I think there are things people might confuse when they read this editorial. The real message from POISE is that we should learn from what just happened."

The Lancet paper also includes an updated meta-analysis of all trials in the field, says Devereaux. "The crucial point from this is that all the other trials were showing the same trends despite different dosing. The meta-analysis did show some heterogeneity, but this was explained entirely by the DECREASE trial."

London says Devereaux "has to be careful. I don't think you can say that beta blockers are radioactive." Nevertheless, London believes the editorial is "very biased. Poldermans [who has been very heavily involved in the field] ignores most of the other research outside of his own studies, and his approach all along has been very myopic. His work has been mainly observational, done primarily only on vascular surgery patients and in only one country in the world."

Also, London takes issue with the editorialists' assertions that beta blockers should be given only by someone who's really experienced in hemodynamics. "What does that mean?

This is very sensational—they are making a big deal of how we are killing people.

"You're going to hear people arguing back and forth about little details of this study, but the overriding thing is that the perioperative environment is nothing like the medical outpatient standard environment, it's the total opposite. The perioperative environment is like putting somebody in a racing car—there are tons of different types of surgeries, tons of different types of stresses, and all very, very unique even within their own specialties." London believes the POISE results should and will change practice.

However, Fleisher told heart wire : "We actually concluded high-dose beta blockade definitely causes harm, so in the end, we said, treat all the underlying causes but if people still have tachycardia, then treat the patient. This is very sensational—they are making a big deal of how we are killing people.

"I'm concerned that people who are taking beta blockers will stop taking them as a result of this or that physicians will stop treating tachycardia [in the perioperative setting] with beta blockers."

Beta blockers can be used, sparingly and carefully

Fleisher stressed to heart wire that US recommendations currently state that perioperative use of beta blockers is a class 1 indication only for those already on beta blockers [who were not studied in POISE] and, second, for patients with a positive stress test undergoing vascular surgery. For most other patients, the indications for beta-blocker use are class 2a or class 2b, he notes.

With regard to the second group for whom this is a class 1 recommendation—patients with a positive stress test undergoing vascular surgery—Fleisher says the committee on guidelines has already discussed changes: "I will be reviewing this and sending it out today."

"What I think will happen is that it will become a class 2b indication—possibly effective but based on limited data," says London. "Like it or not, this is a bombshell in the whole area. What it means is that hospitals that have jumped on the beta-blocker bandwagon fairly aggressively, in large respect to try to boost their performance measures, will have to reconsider."

Nevertheless, there will always be patients in whom it is necessary to use beta blockers, London concludes. "I do a lot of high-risk surgery anesthesia, and I know if I can't control that stress period with an anesthetic drug, I will get a beta blocker out and use that sparingly and carefully. Most of the time, I don't see any big drops in blood pressure or heart rate."

London says this is one of the problems with POISE, however. "We just don't know to what extent this happened in POISE. We'll never know, because it's just too large a study for them to report. That's the thing with a large, simple trial, you don't collect a whole lot of data, you just collect the essential, that's the only way you can get it done. So it's unlikely we will ever hear exactly what the BP was at what time in 8000 patients."

The study was funded by the Canadian Institutes of Health Research , the Australia National Health and Medical Research Council , the Instituto de Salud Carlos III ( Ministerio de Sanidad y Consumo ) in Spain, the British Heart Foundation , and AstraZeneca . Disclosures for the authors are listed in the paper.




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.
Post as: