Monday, August 11, 2014

More evidence against initiation of perioperative beta blockers in non-cardiac surgeries

- Jennifer Middleton, MD, MPH

Patients undergoing intermediate- or high-risk non-cardiac surgeries often get a perioperative beta blocker, and the American College of Cardiology Fellows (ACCF) and the American Heart Association (AHA)'s 2007 guideline advises as such. The ACCF/AHA's 2009 update, though, recommends perioperative beta blockers only in patients with a history of coronary artery disease (CAD) and/or at least 2 CAD risk factors. A 2014 meta-analysis additionally found that perioperative beta blockade in patients not chronically on beta blockers increased mortality -- even if they had CAD or at least 2 CAD risk factors.

JFP reviewed this 2014 meta-analysis recently in an online article. This article discussed serious methodological problems regarding a group of Dutch studies that heavily informed the ACCF/AHA 2007 guideline and 2009 update. These studies failed to consistently obtain written informed consent for subjects, recruited inappropriate patients, and could not produce their raw data when requested by a review panel. These flawed studies are the backbone for the ACCF/AHA's current recommendation to add a perioperative beta blocker to patients with at least 2 CAD risk factors and/or a history of CAD*.

Perioperative beta blockers were certainly controversial even prior to these studies' public disgrace; a 2012 AFP article on Perioperative Cardiac Risk Reduction cited several of the studies included in the 2014 meta-analysis and advised caution in using perioperative beta blockers in patients without known cardiac disease.

The 2014 meta-analysis examined post-operative outcomes for patients who were not taking a beta blocker prior to surgery (including patients with CAD and/or risk factors) and who were undergoing non-cardiac surgeries. The researchers included 9 high quality randomized controlled trials (RCTs) and excluded the suspect data from the Dutch trials. They found that beta blockers did decrease perioperative non-fatal myocardial infarctions (RR = 0.73 [0.61-0.88]), but they also increased the risk of stroke (RR = 1.73 [1.00-2.79]), hypotension (RR = 1.51[1.37-1.67]), and 30-day all cause mortality (RR = 1.27 [1.01-1.60]). For patients not previously on a beta blocker prior to intermediate- or high-risk surgery, adding one causes more harm than good (NNH for 30-day all cause mortality = 160).

When newer studies conflict with previously published guidelines, physicians have an important decision to make. Assuming the new study is methodologically sound, should we abandon the earlier guideline because of the updated evidence? Or, should we wait until the original recommending body assesses the new evidence?

The answer probably involves multiple variables: the amount of trust in the earlier guidelines, the degree to which our colleagues are changing their practice, and, perhaps, even our level of comfort in assessing the study's quality. Maybe it feels like the stakes are too high to change our practice based on one study, especially if larger recommending bodies have yet to weigh in.

Our specialty is fortunate to have high quality journals whose editorial experts sift through new studies and assist us with interpreting and applying them. AFP recently reviewed the ACCF/AHA guidelines for management of congestive heart failure, for example, and handily summarized the high points of this 300 page document into a brief article geared toward busy family docs. Although we should use our evidence-based medicine skills to reach our own independent conclusions about newly published studies, we can also rely on AFP and others to help confirm our assessments.

How comfortable are you changing your practice when new studies conflict with prior recommendations?

* Admittedly, most patients with CAD should be on a beta blocker chronically, but I suspect I'm not the only family doc with a couple of patients who either can't tolerate them or refuse to take one. Several of the studies included in this 2014 meta-analysis included this group of patients in their RCTs.