(Updated on August 4, August 12, and August 14)
With the release today of updated European and US guidelines the ongoing controversy regarding beta-blockers appears to be resolved. But that doesn't necessarily mean there will be an outbreak of guideline peace and harmony. The critics who helped ignite the controversy over beta blockers now say new statin recommendations contained in the guidelines are based on deeply flawed evidence.
The previous incarnation of the European guideline on perioperative evaluation and treatment of people undergoing noncardiac surgery was the subject of intense criticism due to the scandal discrediting Don Poldermans, a Dutch researcher widely published in the field. To address the current uncertainty US and European medical societies earlier today released updated versions of these guidelines.
“Given the recent publication of several large-scale trials, including POISE-II, and new risk calculators, as well as the controversy regarding the use of beta blockers related to the DECREASE trials, the writing committee felt it was necessary to reevaluate all of the data on cardiovascular care for the patient undergoing noncardiac surgery,” said US Writing Committee Chair Lee Fleisher, in a press release.
Regarding beta-blockers the US and European guidelines now do not recommend routine use in patients who undergo non-cardiac surgery, though people who are already taking beta-blockers should continue taking them. (Previously the European guideline but not the US guideline did support routine use of beta-blockers.) Both guidelines state that beta blocker therapy may be initiated prior to surgery in carefully selected higher risk patients.
Statin Recommendation Comes Under Fire
Both the new European and US guidelines say that preoperative initiation of statin therapy may be considered in patients undergoing vascular surgery and that people already taking statins should continue taking them. Now some of the same critics who attacked the reliability of the beta blocker guideline say that this recommendation is not supported by the evidence.
The new recommendation is based on several observational studies and one randomized controlled trial. (Two other randomized trials were not considered because they were performed by Poldermans' group and have been discredited.) The critics, UK cardiologists and researchers Darrel Francis and Graham Cole, say that the one trial by Durazzo et al has fatal flaws that make it completely unreliable. (Durazzo, it may be worth noting, had been a frequent co-author of Polderman's and had been a co-author of several of the controversial or retracted studies.)
In an analysis published earlier this year, Cole, Francis, and co-authors wrote that the Durazzo study
...was a double-blind randomised trial of 100 patients undergoing vascular surgery, with a 45-day course of atorvastatin or placebo. It sought reduced perioperative events in the atorvastatin arm, which indeed was what was found: 8% versus 26% (p = 0.031) at 6 months [22].
This study has serious failings, which make it an unsound basis for recommending therapy. First, its sample size calculation is stated to have been based on a 22% event rate at 6 months in a previous paper [23]. In reality, the source article states that the rate was 12% at 6 months. Such a transcription error would cause a study to be approximately 4-fold undersized.
Second, the authors indicate that they designed their study to detect a relative risk reduction of 95%. This study design is not credible as no therapy has ever been so effective in preventing myocardial infarction. If the true effect size was, for example, half of this, this overestimate would have contributed a further ~ 4-fold undersizing of the study.
Third, the survival data published cannot be correct. The paper reports that of the 50 patients in each arm, none were lost to follow-up. Therefore, every patient surviving to each displayed time point should be exactly 2%. With this in mind, in the Kaplan–Meier graphs, almost all the numerical values in the survival follow-up figure contradict the graphical values shown.
Finally, for 50-patient groups with no loss to follow-up, event-free survival rates must again be multiples of 2%. They are quoted as 91.4% and 73.5%, values that are not possible.
Francis and Cole sent the following comment about the new guidelines:
We are very sad that over 100 world authorities were forced to sign the European guideline without all having had time to read the papers on which their recommendations were based, and without being able to openly voice dissent.
Their awful predicament is easiest to see for a therapeutic idea whose road has been very “bumpy” indeed: the perioperative course of statins. The outcome data of the key trial by Durazzo et al has for over 6 months been publically known to be impossible. This is buttressed by Don Poldermans’ now notorious DECREASE III and IV trials, whose own university’s investigation revealed extensive fictionalisation, and by meta-analyses whose events arose mostly or entirely amongst these extraordinary pieces of science.
There are important lessons to learn. First, guidelines must in future have the right to say that all the major trials have now been discredited, so there is no longer a recommendation. If we fail to recognise this, we have truly failed our patients.
Second, never again should we pretend that all the experts have agreed on recommendations. It was always unlikely, and in this case clearly ridiculous.
In an interview, Fleisher, the chair of the US guideline, defended the committee's recommendation of statins. He said that the committee was aware of the limitations of the Durazzo study and that the recommendation was based on the totality of the evidence, including the observational studies. He agreed that there was a significant need for more high quality studies.
Update (August 4):
Here is a response from the ESC:
The 2014 guidelines writing committee conducted a comprehensive review of relevant clinical papers and meta-analyses available at the time of the guideline development. Several original papers (including the Durazzo et al. paper), observational studies and meta-analyses were cited which have evaluated the peri-operative use of statins in patients undergoing non-cardiac surgery.
We were aware of the limitations of the Durazzo study and the guidelines clearly indicate that, with respect to the use of statins, the class of recommendation (IIa B) is based on ‘conflicting evidence and/or a divergence of opinion about the usefulness/efficacy of the given treatment or procedure’. The recommendation itself states that ‘pre-operative initiation of statin therapy should be considered in patients undergoing vascular surgery’ and not that it is recommended.
Specific recommendations within the guidelines serve as a vital starting point for the care of an individual patient. It is important to stress that they are created with the voluntary input of over 115 experts, who took the time needed to properly review and evaluate several hundred papers.
Update (August 12):
Two members of the ESC guideline panel, Christian Funck-Brentano and Bernard Iung, sent the following response to Darrel Francis's remarks:
Dear Sir,
Following the highly debatable number of deaths allegedly caused by the 2009 European Society of Cardiology / European Society of Anaesthesiology (ESC/ESA) guideline regarding the use of beta-adrenergic blockers for the prevention of cardiovascular accidents during the perioperative period of non-cardiac surgery (1, 2), Darrel P Francis and colleagues raised severe criticisms on the ESC/ESA revised guideline published on August 1, 2014 and now question the use of statins in perioperative non-cardiac surgery (3). Francis and coworkers write that they are “very sad that over 100 world authorities were forced to sign the European guideline without all having had time to read the papers on which their recommendations were based, and without being able to openly voice dissent”. They add that “never again should we pretend that all the experts have agreed on recommendations. It was always unlikely, and in this case clearly ridiculous.”
We, as contributors to the 2014 ESC/ESA Non-cardiac surgery guideline, feel insulted by such statements. None of the members of the Task Force who contributed to the guideline have been forced to sign the final version of the document. We were invited to participate as unpaid volunteers because we were experts in this field and have written, discussed and revised the guideline document during the last 18 months. Guidelines production, like other scientific writing, is team work and all recommendations have been reviewed collectively with “section experts” presenting their arguments during face to face meetings and teleconferences. The whole document has been reviewed 3 rounds by a large number of reviewers and thousands of comments have been made. The draft guideline was revised accordingly in each round. Ultimately, all co-authors have approved and signed-off the final version of the guideline. We are confident that this final version is finely shaded and well balanced – including the section on the use of statins – and although we recognize the right of Francis and coworkers to disagree with us, we would like to remind them that, before making such extreme and offensive statements, they should gather proofs of their defamatory allusions. We hope that, in the future, they will use the appropriate academic style and moderation when expressing their right to a disagreement.
--Christian Funck-Brentano
--Bernard Iung
Update (August 14):
Darrel Francis and Graham Cole sent the following statement in response to Christian Funck-Brentano and Bernard Iung:
We are very sorry for offending two of Europe's top experts, Dr Funck-Brentano and Dr Iung. We should have emphasised that we were querying only the choice of trials used to support perioperative courses of statins. We share their views that statins provide huge cardiovascular prevention benefits for patients.
We chose our words poorly, but were intending to sympathise with the guideline signatories in their unenviable and Herculean task. Their bold declaration that results from the DECREASE studies should not be included in the guideline is a landmark moment for scientific integrity.
We apologise for speculating (incorrectly, it seems) that time pressure may have prevented some of the >115 signatories from reading all of the hundreds of articles on which the guideline depends. The ESC statement appears to us to confirm that all signatories read each article. We are humbled by the enormity of this achievement.
Dr Funck-Brentano and Dr Iung, despite the hurt we inadvertently caused them, remain charitably open-minded in recognising that other opinions may exist. Of course we respect the opinions of >115 experts in such unanimity. Presumably, the group resolved the four queries listed above on the statin trial, during the process of all agreeing to rely on the resulting data. We encourage them to share the resolutions, if it is permissible.
Finally, we are especially sorry if we wrote our first sentence under a misapprehension. We had believed that signatories were not allowed to discuss scientific disagreements and voice dissent outside the group while developing the guideline. If the 2014 guidelinists have indeed ended this secrecy policy, it is another massive achievement.
We hope Dr Funck-Brentano and Dr Iung, and any other guideline developers who were offended, will accept our unreserved apology. We are all only human, and striving to the common goal of promoting reliable clinical science for the benefit of patients.
