Dr Philip J Devereaux

Orlando, FL - A landmark trial presented at the American Heart Association (AHA) 2007 Scientific Sessions of the continued-release (CR) beta blocker metoprolol (Toprol XL, AstraZeneca) in patients undergoing noncardiac surgery has shown that although the drug reduced the risk of MI, it increased the risk of severe stroke and overall death. The POISE study suggests that for every 1000 patients treated, metoprolol CR would prevent 15 MIs, but there would be an excess of eight deaths and five severe disabling strokes.

Lead investigator of the late-breaking study, Dr Philip J Devereaux (McMaster University, Hamilton, ON), said that physicians should weigh the potential risks and benefits before deciding whether to use beta blockers perioperatively. He admitted to heartwire that his team was unable to pinpoint who would be at risk and who would benefit from perioperative beta-blocker use at this stage, but added, "I certainly would not recommend it to my mother."

I certainly would not recommend it to my mother.

A discussant of the study, Dr Judith Hochman (Cardiovascular Clinical Research Center at the New York University School of Medicine), said, "POISE is the first robust randomized trial to look at this subject." She added that American College of Cardiology (ACC)/AHA guidelines for beta-blocker use in noncardiac surgery are based on previous studies that were inconclusive and small and showed conflicting results. "I believe POISE will change practice. In my opinion, beta-blocker therapy should not be initiated perioperatively as routine therapy to reduce cardiac events," she stated.

Other experts agreed with Hochman. Dr Mariell Jessup (University of Pennsylvania, Philadelphia) told heartwire that "the POISE trial was very surprising, because people routinely use beta blockers in the noncardiac-surgery setting. I think I'd really give pause in this general population of just routinely giving beta blockers, which has been the recommendation. I think this is potentially going to change the guidelines. It's really big news. Very unexpected." Chair of the AHA Scientific Sessions program committee, Dr Gordon Tomaselli (Johns Hopkins University, Baltimore, MD), said that "unless someone comes in on a beta blocker, I don't think you can increase the level of assurance that this is the right thing to do based on this trial."

The POISE trial was very surprising. It's really big news. Very unexpected.

However, others were concerned that people would take these results to mean that they should not use beta blockers perioperatively at all. "That would be premature. There are national practice guidelines on this," Dr Raymond Gibbons (Mayo Clinic, Rochester, MN) told heartwire. Gibbons is troubled about a number of issues related to the POISE trial. "I'm worried about titration going in, the dose selection, and the parameters to discontinue the drug, and I believe they will have to provide details about the deaths."

Dr Roger Blumenthal (Johns Hopkins University) concurred: "Perhaps they gave too high a dose two hours beforehand and within six hours after surgery, particularly in susceptible people. It will be interesting to see what post hoc analysis shows. This may dim enthusiasm for giving beta blockers to everyone who undergoes surgery, but people with known risk factors—for example, those with known vascular disease and a positive stress test—will still benefit."

The moderator of the Forum on theheart.org, Dr Melissa Walton-Shirley (TJ Samson Community Hospital, Glasgow, KY), said that "the take-home message from this study is clearly that we should not treat these patients to the point of hypotension."

Despite her overall conclusion, Hochman also stressed that there was much information still to be gleaned from POISE that might shed more light on the issues involved.

By way of background, Devereaux explained that 100 million adults worldwide undergo noncardiac surgery annually, and one million of these suffer a major perioperative cardiovascular event. The 2006 ACC/AHA guidelines for beta blocker use in noncardiac surgery suggest that beta blockers be started before elective surgery—particularly in high-risk patients—with the dose titrated to achieve a resting heart rate between 50 and 60 bpm.

POISE randomized 8351 patients 45 years or older undergoing noncardiac surgery with or at risk of atherosclerotic disease. Patients had to have a history of coronary artery disease, peripheral artery disease, stroke, or congestive heart failure within the last three years; be undergoing major vascular surgery; or have three of the following seven risk factors: undergoing high-risk surgery, having a history of CHF, having diabetes mellitus, having renal insufficiency, being 70 years of age or older, having a history of transient ischemic attack, or undergoing urgent/emergent surgery.

Patients were recruited from 193 centers in 23 countries and randomized to receive either metoprolol CR or placebo started two to four hours preoperatively and continued for 30 days. The dose of metoprolol given was 100 mg in the preoperative period, 100 mg in the six-hour postoperative period, 200 mg 12 hours later, and 200 mg daily thereafter out to 30 days. Doses were not titrated, and the drug was stopped only if blood pressure dipped below 100 mm Hg.

The primary outcome was a composite outcome of cardiovascular death, nonfatal MI, and nonfatal cardiac arrest at 30 days after randomization. Secondary outcomes included total mortality, cardiovascular death, MI, cardiac revascularization, clinically significant atrial fibrillation, clinically significant bradycardia, clinically significant hypotension, and stroke.

"The cumulative evidence of POISE suggests a clear reduction in MI and a decrease in coronary revascularization and atrial fibrillation, but an accompanying increase in death and stroke and an increase in hypotension and bradycardia," said Devereaux

He noted that the reduction in MI with metoprolol drove the reduction in the primary outcome and that most people who suffered an MI had one in the first few days after surgery. The majority of strokes also occurred in these first few postoperative days, "and stroke was a strong determinant of death," he noted, with a hazard ratio (HR) of 12.74. Clinical hypotension was another robust indicator of mortality, with a HR of 4.32. There were also twice as many deaths in metoprolol-treated patients who had sepsis or infection as in those with sepsis who received placebo.

Hochman said it is important to clarify that these results do not apply to patients already taking beta blockers undergoing surgery (such patients were excluded from POISE) or to patients in whom a physician had planned to prescribe a beta blocker within 30 days of surgery.

Jessup stressed this too: "It doesn't mean that we should stop beta blockers in patients who are already on them." Tomaselli concurred: "This study does not say you should stop beta blockers in those patients if they come into surgery on them. In fact, I would say that is absolutely the wrong thing to do."

Hochman said a number of issues still need to be addressed, including dosing and titration. The full fixed dose of metoprolol CR used in POISE is at the upper range of those used in previous trials, she said. She also stated that "perhaps a systolic pressure of a 100 mm Hg is too low for hypertensive patients, the elderly, and those with a critical carotid or coronary stenosis."

There are established national practice guidelines. Before we say they are wrong, this study really needs to provide more details.

Gibbons agrees. "I need to see more details. All the issues in my mind have not been addressed. There are established national practice guidelines. Before we say they are wrong, this study really needs to provide more details."

"It does not appear that there was a period of any titration of the dose, which the guidelines show is specifically called for because the serum levels of beta blockers at a given dose are very variable in a given population. And in elderly patients, we wouldn't start at this dose of 100 mg. The titration issue needs to be addressed; this is supported by the heart rates recorded," he continued.

Gibbons also pointed out that "they only stopped beta blockers for a BP of less than 100 mm Hg—most of us would view that criteria as not strict enough in an elderly patient with [transient ischemic attack] TIA. I would argue that the threshold for stopping the drug should be higher; they are going to get in trouble."

More details about the deaths are also needed, he noted. "If the deaths are related to hypotension, because the drug was continued and then led to a stroke, that raises some questions about whether the guidelines need to be changed. Also, the mortality diverged after nine days, and this raises a whole bunch of separate issues in my mind. Is it really perioperative?"

Jessup said, "We are going to have to look at the details of the study. There was no signal as to who benefited and who did not—if they couldn't really identify a profile, it's really difficult to know who to recommend beta blockers for anymore. We'll have to see how everybody else slices and dices the patient population."

Hochman concluded: "We await information as to whether we can accurately identify those at risk for hypotension and adverse events based on baseline characteristics and hemodynamic response to dosing. Further randomized trial data are needed. Is there a better beta-blocker dosing regimen? What about genetic polymorphisms? And are there other therapies to reduce perioperative risk, such as statins or antithrombotics?"