The controversy about using beta blockers in the perioperative period for non-cardiac surgery just heated up again. Results of the POISE (Perioperative Ischemic Evaluation) trial (1) revealed this month at the American Heart Association 2007 Scientific Sessions, showed that perioperative beta blockers did reduce the frequency of perioperative MI but increased frequency of stroke and actually increased overall mortality.
Use of perioperative beta blockers to reduce morality in high risk patients undergoing noncardiac surgery has been in vogue for the past 10 years. A number of clinical studies had shown beneficial effect of beta-blockers in myocardial ischemia in other circumstances, and since most post-op cardiac events were felt to be likely mediated by myocardial ischemia, extension of their use to the perioperative setting was logical.
The actual clinical evidence for use of beta-blockers in the perioperative period was based upon results of only a few clinical trials, each having very few patients enrolled. Mangano et al. (2) randomized 200 high-risk patients undergoing major noncardiac surgery to either atenolol or placebo. They found a 50% reduction in perioperative ischemia detected by Holter in the atenolol group but no difference in in-hospital mortality or MI occurrence. Surprisingly, though, those in the atenolol group had a significant reduction in cardiac events 6-8 months later.
The Dutch DECREASE trial (3) looked at 112 vascular surgery patients who had evidence of ischemia on pre-op dobutamine stress echocardiography. This was a randomized controlled trial but was not blinded. Those randomized to bisoprolol had a 10-fold decrease in perioperative MI and cardiac death compared to placebo.
An AHRQ evidence report (4) in 2001 identified perioperative use of beta blockers as one of eleven evidence-based patient safety practices that should be more widely adopted. Many have now extended the use of periperative beta blockers to those intermediate and high risk patients not previously on beta blockers.
The ACC/AHA Guidelines on Perioperative Beta Blocker Therapy (5) were updated late in 2006. They provide a class I indication for continuation of beta blockers in patients taking them for any class I indication (eg. angina, MI, CHF, arrhythmias, hypertension) or in any patient whose pre-op evaluation demonstrates ischemia. They add class IIa indications (probably recommended) for patients with known coronary artery disease or multiple risk factors and class IIb (“may be considered”) in patients with a single risk factor undergoing intermediate or high risk procedures. Class IIa evidence means there is conflicting evidence, weighted in favor, and class IIb means there is conflicting evidence, with efficacy less well established.
Several national quality improvement programs, including SCIP (the Surgical Care Improvement Project) (6) and the IHI’s 5 Million Lives Campaign (7), have included use of beta blockers in attempt to reduce perioperative mortality. However, their programs have focused only on those patients previously on beta blockers, for which there is class I evidence of efficacy.
But not everyone has been convinced of the benefit of perioperative beta blockers in patients not already on them. There has been considerable evidence on the negative side as well. The POBBLE randomized trial (8) published in 2005 showed no benefit of beta-blockade in patients undergoing noncardiac vascular surgery. However, it was a small study (only 100 patients total in the study). Likewise, the MaVS study (9) published in 2006 failed to show a statistically significant benefit of beta blockade in high risk patients undergoing vascular surgery.
A retrospective cohort study published in 2005 by Lindenauer et al (10) showed a reduction of in-hospital deaths on perioperative beta blockers in those patients deemed high risk but not in those deemed low risk. Though this was not a randomized controlled trial, it lent further impetus to the routine use of beta blockers perioperatively in high risk patients.
A sizeable multicenter randomized controlled trial of long-acting forms of metoprolol in diabetic patients undergoing major noncardiac surgery (DIPOM) (11)subsequently showed no improvement in overall mortality or cardiac morbidity in the treated group.
Several meta-analyses have subsequently been done. Devereaux et al. (12) analyzed data from 22 RCT’s (2437 patients) and concluded that beta blockers produced no statistically significant improvement in any single outcome measure but did produce a nominally significant improvement in a composite measure. Wiesbauer et al. (13) found no reduction in perioperative MI or mortality in patients on beta blockers.
Now the POISE trial seems to answer many of the controversial issues. POISE was a randomized, controlled multicenter trial in high-risk patients undergoing major noncardiac surgery who were not on beta blockers prior to their planned surgery. They were randomized to receive either the long-acting beta blocker metoprolol-CR or placebo before surgery and throughout the perioperative period. The primary outcome measure was a composite of cardiovascular death, nonfatal MI, and cardiac arrest at 30 days after randomization. There was a statistically significant reduction in the primary outcome in the metoprolol CR group, driven primarily by a reduction in nonfatal MI. However, overall mortality and frequency of stroke were increased in the metoprolol CR group. For every 1000 patients treated, 15 MI’s were prevented but there was an excess of 8 deaths in the metropolol group.
Some have criticized the POISE protocol for a number of reasons, including the dosage used and the lack of a titration period. The metoprolol group did have a higher incidence of significant hypotension and bradycardia. Others have criticized the degree of hypotension tolerated by the investigators before intervention. And the study has only been presented in abstract form so far. But the bottom line is that this is the highest level of evidence to date and it strongly suggests that we need to change our current approach.
So where does this all leave us? There is still general agreement that those patients previously on beta blockers should have them continued through the perioperative period. In fact, there is likely a detrimental effect from beta blocker withdrawal in such patients. It seems to us that this is the only group currently for which we can comfortably recommend perioperative beta blockers. For those not previously on beta blockers there is currently no clearcut answer at this time. The POISE trial certainly casts doubt on the benefit and safety of prophylactic beta blockers in the perioperative period. Undoubtedly, there will be analysis of subgroups in the POISE trial attempting to identify characteristics of those patients who might benefit. However, those will be post-hoc analyses and they serve only to raise hypotheses for further testing. Only randomized controlled trials are likely to provide the answer to this dilemma.
One very difficult issue is what to do with patients who were not previously on beta blockers but have a condition for which beta blockers are indicated long-term (eg. prior MI, CHF, etc.). This group would have been started on beta blockers in the perioperative period according to ACC/AHA guidelines. However, even though POISE excluded patients in whom physicians planned to start beta blockers within 30 days, we think that the results of the POISE trial even raise the question as to whether those patients would be better off having beta blockers started electively at a time well after the perioperative period. Some physicians will probably continue to consider use of perioperative beta blockers in patients in whom preoperative evaluation shows evidence of ischemia. But they will probably begin the beta blockers well prior to surgery and titrate the dosage gradually and be very careful to avoid hypotension. The results of future studies, such as the DECREASE V study currently being undertaken in the Netherlands, may clarify some of the remaining issues. But for now POISE should significantly reduce the number of patients receiving perioperative beta blockers.
Hospitals and ambulatory surgery facilities still need protocols and procedures in place to ensure that patients who should receive perioperative beta blockers do, in fact, receive them. Obviously, good medication reconciliation is required in ensuring those who have been on beta blockers are identified and appropriately continued on them through the perioperative period.
Sometimes things that make a lot of sense turn out to be wrong when subjected to the type of scientific scrutiny we need to apply to patient safety practices as well as therapeutic or diagnostic practices.
Update: See our November 4, 2008 Patient Safety Tip of the Week “Beta Blockers Take More Hits” and our November 2010 What’s New in the Patient Safety World column “More Perioperative Beta Blocker Controversy”