To the Editor:—
London et al. 1state that the “evidence for the efficacy of perioperative β-adrenergic receptor blockade is strong” and Kertai et al. 2state that the perioperative β-blocker “data provide solid evidence for their efficacy.” Both groups advocate perioperative β-blockade for high-risk patients undergoing noncardiac surgery. However, they are recommending perioperative β-blockade for millions of patients, annually, worldwide on the basis of randomized controlled trials that include only 866 patients, with only 15 cardiac deaths and 18 nonfatal myocardial infarctions. We believe that these results, although promising, do not justify such enthusiasm. More definitive evidence from large-scale randomized controlled trials is required before strong recommendations can be made.3Such a trial, the PeriOperative ISchemic Evaluation (POISE) trial, is currently recruiting patients in six countries to evaluate the effectiveness of perioperative β-blockade in 10,000 moderate-risk and high-risk patients undergoing noncardiac surgery.
Despite their conclusions, London et al. 1acknowledge several limitations in the perioperative β-blocker randomized controlled trial data. Among the 866 patients randomized most of the events (i.e. , 11 deaths and nine nonfatal myocardial infarctions) occurred in the trial of Poldermans et al. ,4which is limited by lack of generalizability and blinding and which was stopped early because of an unexpectedly high risk reductions (100% for nonfatal myocardial infarction and 80% for cardiovascular death). In Kertai et al.' s2editorial (in fig. 1), a reduction of death and myocardial infarction from 18 of 38 (47%) in the control patients to 11 of 40 (25%) in treated patients is quoted. If Polderman et al.' s4data were true, one would expect fewer than three events, not 11, among these patients receiving perioperative β-blocker therapy. This suggests that the data in Poldermans et al.' s trial are too good to be true. Indeed, these results are inconsistent with everything we know about β-blockade from randomized trials of tens of thousands of patients with myocardial infarction and heart failure that demonstrate relative risk reductions of 25–30% rather than 80–100%.5,6
Similarly, a constellation of problems limits the reliability of Mangano et al. 's results.7For example, only events that occurred after the patients stopped taking the study drug were counted in their analyses. In fact, if all the deaths are included, then the results are no longer statistically significant. As London et al. 1report, the limitations of Mangano et al. 's data resulted in a class IIa recommendation for this type of cohort in the American College of Cardiologists/American Heart Association guidelines.8
London et al. 1state that “it can be argued that given strong efficacy in secondary prevention after myocardial infarction, additional perioperative studies are unnecessary.” However, as they themselves point out, these data may not be applicable to patients who merely have risk factors for coronary artery disease. These patients make up the majority of patients for whom perioperative β-blockade is advocated. The numbers needed to treat may be substantially higher in this group, which may mean that patients are exposed to the risks of perioperative β-blockade without much chance of benefit.
Despite the reassurance of Kertai et al. ,2the safety of perioperative β-blockade is not well established among the 866 patients that have been randomized. Safety data are lacking on the impact of age, acute β-blockade, blood loss, and other pharmacologic interventions both acute and chronic.1In addition, as acknowledged by London et al. 1and Kertai et al. 2, there is no consensus about the logistics of perioperative blockade (i.e. , how much, how long, by what route?). Along with the lack of definitive evidence, these factors have limited the current enthusiasm of clinicians for perioperative β-blockade.9,10
The evidence for perioperative β-blockade in patients undergoing noncardiac surgery is encouraging but limited. The current evidence does not justify strong recommendations for the routine use of perioperative β-blockers. We, as anesthesiologists, need to heed the lessons learnt in internal medicine over several decades: large trials are essential to confirm promising (but potentially incorrect) results from small trials.3,11
* Royal Melbourne Hospital and University of Melbourne, Parkville, Victoria, Australia. firstname.lastname@example.org