From Evidence-Based Medicine to the Heart Team Approach

Clinical judgment is based on both available relevant data and the decision-making process. The shift in recent decades to “evidence-based medicine” demands considerable time to navigate through the proliferating data. The need to identify and categorize data that influence clinical decision making has become a Sisyphean struggle for the practicing physician. As a result, a hierarchical “pyramid of evidence” has been formulated to help categorize data quality ( Figure 1 ). To further assist and direct health-care decisions in a wide range of clinical settings, the hierarchical data are processed into recommendations in “practice guideline” (PG) statements issued by working groups, representing authoritative professional organizations.

Pyramid of evidence. RCT = randomized controlled trial.

PGs on the management of cardiovascular diseases by the American College of Cardiology (ACC), the American Heart Association (AHA), and the European Society of Cardiology (ESC) have been available for >20 years. The PGs use a grade system to classify the strength of recommendation (from class 1 to 3) and level (quality) of evidence (from A to C) (available online, e.g., http://www.acc.org ). In recent years, PGs are perceived to be the unbiased epitome of evidence-based medicine. In the words of Sniderman and Fruberg: “… (the) guidelines are generally assumed to have the same level of certainty and security as conclusions generated by the conventional scientific method. For many clinicians, guidelines have become the final arbiters of care ”. In fact, PGs are not part of the pyramid of evidence and too often suffer from limited scientific evidence to support their conclusions. From the introduction of level of evidence in 1998 to September 2008, the ACC/AHA Joint Task Force issued 33 PGs; of which, 27 adopted the level of evidence classification. The PGs that report the level of evidence comprise a total of 2,711 recommendations. Of these recommendations, only median of 11% are supported by high-quality scientific data (level of evidence A).

Furthermore, most PGs include >50% recommendations supported by low-quality data and consensus of experts (level of evidence C). Similarly, only a minority of the recommendations for coronary revascularization are supported by the level of evidence A, 11% of recommendations for percutaneous coronary intervention (PCI) and 19% of recommendations for coronary artery bypass grafting (CABG). We suggest that the recent PG recommendation on the heart team approach is an example of a recommendation with potential profound impact on clinical practice supported by very limited data and limited expert opinion.

The most recent PG by the ESC on myocardial revascularization calls for the formation of “heart teams” (consisting of cardiologists and cardiothoracic surgeons) to optimize revascularization strategy. Recently, the ACC/AHA/SCAI 2011 guidelines for PCI embraced the heart team approach as the preferred method to optimize revascularization decision making in cases of complex coronary anatomy. The proposed heart team approach is based on the assumption that this method “… serves the purpose of a balanced multidisciplinary decision process”. Both ACC/AHA/SCAI and ESC PGs adopt a class 1 recommendation for the heart team approach, supported by level of evidence C. As a result, a change in the decision-making process on revascularization strategy, supported by low quality data, could lead to a major shift in patient care.

The theoretical basis for this major change in the decision-making process for revascularization lies in the SYNTAX trial, which studied the optimal revascularization strategy (CABG vs PCI with drug-eluting stent) in patients with complex coronary artery disease (CAD). In the SYNTAX trial, a heart team, consisting of an interventional cardiologist and a cardiac surgeon at each study site, prospectively evaluated patients’ eligibility at Heart Team Conferences, not essentially dissimilar to the way patients were screened in the past in similar studies. Hence, the objectives of the heart team in the SYNTAX trial setting were different from the objectives of the heart team in a clinical practice setting. This extrapolation of a research method used in clinical trials, to optimize patient selection and recruitment, to the broad clinical spectrum has potential limitations. We suggest that both the need for a new method to optimize patient triage for the various revascularization strategies and the optimal method to optimize decision making should be rigorously studied with meaningful outcomes (e.g., major adverse cardiac events, system-based such as patient/physician resource utilization) before wide spread to implementation of the heart team approach.

Is There a Need to Optimize the Revascularization Decision-Making Process?

In the last decade, “the PCI era”, PCI has evolved into the most common in hospital procedures followed by a dramatic reduction in mortality in CAD, mostly in patients with acute myocardial infarction.

Two retrospective analyses of large registries have examined adherence by current methods of decision making to PGs. Anderson et al, in a retrospective analysis of 412,617 PCIs in the ACC National Cardiovascular Registry, found that only 8% of the PCIs were performed in class 3 indication whereas 85% were performed in either class 1 or class 2 indication. Hannan et al reported similar proportions of adherence to PGs using data on 10,333 patients from the New York State Cardiac Diagnostic Catheterization Database. Of the 7,984 patients who underwent PCI, in only 716 (9%) of the PCI cohort, the procedure was performed when the ACC/AHA PGs indicated CABG or “neither CABG nor PCI”. In the CABG cohort, CABG was performed in 20% when the guidelines indicated that they belong to the PCI or “neither CABG nor PCI” groups.

The observed minor deviations from guideline-based indications may represent “real-life” complexities and uncertainties when PG recommendations are applied to clinical setting, in which individualized patient-oriented judgment must be made.

The power of the current methods of decision making and clinical judgment is further highlighted by the long-term results from the Bypass Angioplasty Revascularization Investigation (BARI) trial. The BARI trial included 4,039 patients with multivessel CAD. The long-term clinical outcome in the trial registry’s patients (n = 2,010) was compared with the randomized patients (n = 1,829). Both study groups had similar baseline characteristics. Of note, in the BARI registry, nearly twice as many patients were selected for percutaneous transluminal coronary angioplasty (PTCA). Yet, the 7-year mortality rate was lower in the patients assigned for PTCA by clinical judgment in the registry arm compared with the patients assigned for PTCA in the randomized trial. Furthermore, the patients chosen for PTCA in the registry had lower CAD burden, a higher PTCA success rate, and a lower rate of complications. Thus, the data suggest that traditional clinical judgment may result in adequate decision making for revascularization strategy.