CardioExchange Archive

A recent article in the European Heart Journal by Graham Cole and Darrel Francis explores the continuing implications of the Don Poldermans research misconduct case. In light of this continuing controversy, we asked Prashant Vaishnava, Vineet Chopra, and Kim Eagle — all of the University of Michigan Health System — to comment on the use of perioperative beta-blockade.

Evidence of research misconduct has discredited the Dutch Echocardiographic Cardiac Risk Evaluation Applying Stress Echocardiography (DECREASE) family of studies. All studies investigated in the DECREASE family were found to be insecure because of flaws ranging from fictitious methods to fabrication of data to no evidence of written informed consent.

While we would not deny the far-reaching harms resulting from “perioperative mischief,” we are cautious about suggesting that implementation of the current European guidelines (informed, in part, on the indicted data from the DECREASE family of studies) may partially account for the projected 160,000 excess deaths annually in Europe as suggested by Cole and Francis (Note: The European Heart Journal has removed this paper while it conducts peer review. For more context, see Larry Husten’s Forbes article.). In an editorial titled, “Is the panic about beta-blockers in perioperative care justified?” the editorial staff of the European Heart Journal write that the contribution by Cole and Francis “failed to undergo peer review that is required for opinion pieces, if scientific statements affecting clinical practices are involved.”  We share in these sentiments and would advocate for peer review of such potentially influential conclusions before broad dissemination. We offer the following comments:

1. The perioperative community is stuck between a rock and a hard place.  On one hand, there are now discredited data. On the other hand, credible data employ a peri-operative beta-blocker regimen that is not representative of clinical practice. Specifically, the randomized, double-blind, placebo-controlled PeriOPerative Ischemic Evaluation (POISE) trial implemented 100 mg oral extended-release metoprolol or matching placebo 2-4 hours prior to non-cardiac surgery in 8351 patients with, or at risk of, atherosclerotic vascular disease. As indicated in the European perioperative guidelines, “the total dose of metoprolol succinate in the 24 h [was] 400 mg, at least in a number of patients.” This intensity of perioperative beta-blockade is rarely used in clinical practice, even in those maintained on chronic beta-blockade, much less in patients naïve to such treatment. More commonly, perioperative beta-blockade is initiated weeks in advance of elective non-cardiac surgery.

We recognize the POISE data as being credible, noting both the cardiac benefit and excess mortality associated with those randomized to beta-blockade; however, the intensity of perioperative beta-blockade used in this trial simply dilutes its clinical relevance. In being stuck between a rock and a hard place, it is difficult for guideline-writing authorities to make meaningful changes in reaction to data that lack credibility when the alternative is simply not fully applicable to current clinical practice.

2. Cole and Francis draw attention to a meta-analysis of “credible” data. The pooled effect size found by Bouri et al, while a step forward in excluding flawed data, is almost completely influenced by the singular effects of the POISE trial. Initiation of a course of beta-blockade before surgery was associated with a significant 27% increase in mortality (relative risk [RR] 1.27, 95% confidence interval [CI] 1.01 to 1.60, p=0.04). The clinical and statistical impact of these data may be questionable, however, given the proximity of the CI to unity.  More importantly, the meta-analysis may do little more than corroborate the findings from POISE, a trial inherently limited by the protocol of beta-blockade used perioperatively.

3. Assuming that there are ~760,000 deaths/year following non-cardiac surgery, Cole and Francis suggest that as many as >400 excess deaths daily may be ascribed to the implementation of guideline-based care for perioperative beta-blockade. We suggest caution in attributing guideline-based care to actual mortality. We have long known that multiple barriers impact physician implementation of evidence-based care, and that guidelines are frequently not practiced. Such extrapolation, at best, may thus be inaccurate. Moreover, the attribution of such mortality is predicated on accepting that perioperative beta-blockade is indeed associated with adverse outcomes. Given the limitations of the POISE protocol and a meta-analysis that does little more than echo the findings of POISE, the suggestion that mortality is inextricably bound to perioperative beta blockade is dubious. Such an association may even be disingenuous when operating through a Bayesian lens of posterior probability.

For example, an appraisal of credible data must not overlook a recent retrospective cohort analysis by London et al. In this analysis of 140,000 patients treated at 104 U.S. Veterans Affairs medical centers, the authors found that the use of perioperative beta-blockade was associated with a reduction in 30-day mortality (RR 0.73, 95% CI 0.65 to 0.83, p<0.001) in patients with two or more Revised Cardiac Risk Factors. Though we acknowledge the inherent limitations of this retrospective analysis, we pause in reaction to an indictment of beta-blockade. We must thus remain mindful of the ambiguity involved and the possibility that perioperative beta-blockade may be a useful practice.

4. More than two years have passed since allegations of research misconduct justly discredited the DECREASE family of studies and investigators. It is time to move beyond finger-pointing.  Guideline writing authorities cannot be faulted for not reacting to negligent data when available evidence is difficult to apply to clinical practice.

We should leverage and advocate for what we know to be true: 1. The indiscriminant use of high doses of beta-blockers in the immediate hours before non-cardiac surgery is more likely to be dangerous than beneficial; and 2. Judicious use of titrated doses of beta-blockers in carefully selected patients with established coronary artery disease, may, in fact, reduce cardiovascular risk.  When titrating beta-blockade to a target heart rate, the physician must be mindful that it is important to avoid perioperative hypotension at all times.

Ultimately, we should search for more definitive answers by advocating for a randomized clinical trial that is conducted honestly and transparently and uses a regimen that mirrors clinical practice.