OUP user menu

Is myocardial revascularization safe in trainees' hands?

Thierry C. Gillebert MD, PhD, FESC, Benny Drieghe MD, Marc L. De Buyzere MSc
DOI: http://dx.doi.org/10.1093/eurheartj/eht177 2859-2861 First published online: 31 May 2013

This editorial refers to ‘Clinical outcomes after myocardial revascularization according to operator training status: cohort study of 22 697 patients undergoing percutaneous coronary intervention or coronary artery bypass graft surgery’, by D.A. Jones et al., on page 2887

This issue of the European Heart Journal includes an analysis, which shows that there is no evidence that outcome of patients after myocardial revascularization (coronary artery bypass surgery (CABG) or percutaneous coronary intervention (PCI) would depend on primary operator grade, trainee versus consultant.1 This is the largest surgical CABG dataset with the longest follow up looking at trainee grade and outcome, which coupled with similar and still unpublished findings in PCI procedures results in a large and important revascularization dataset. Great news, trainees are as good as experienced cardiac surgeons or experienced interventional cardiologists! Hurray for the quality of the trainees, hurray for the quality of the training programmes! The results of the investigation however may not be lumped to this simple message and should be carefully analysed in order to develop the right insights leading to adequate measures and decisions with regard to training in myocardial revascularization.

Who are these trainees performing CABG and PCI as a primary operator? In the UK system the MD has to complete Foundation Training before starting his specialization (Figure 1). The specialization in Cardiology comprises three modules: Core Medical Training (or Acute Care Common Stem), Cardiology Core training and Advanced Specialist Area Modules.2 This training takes up 7 full-time years, of which 5 years are devoted to cardiology, after which the trainee is expected to be entirely autonomous as a specialist practitioner. The specialization programme in the UK is in accordance with the requirements of the ESC Core Curriculum for the General Cardiology (2013), which includes a module of Internal Medicine and General Acute Care (2 years) and a module of General Cardiology (4 years). This training takes up at least 6 full-time years, after which the trainee is expected to be autonomous as a general cardiologist. The trainee may start subspecialty training such as interventional cardiology during the last year of General Cardiology training. In order to develop the knowledge and skills corresponding to Interventional Cardiology he still has to extend his training by at least one year. This means that the cardiology trainee does not start to be first operator in an intervention until year 6 and 7 of training. The trainee as a first operator is a 30 years old physician already experienced and skilled in clinical medicine and cardiology. Similarly, trainees in cardiac surgery would not undertake operations as a primary operator before the sixth year of surgical training at the earliest.

Figure 1

Training in cardiology in the UK and the ESC

The report of Jones et al. is not a prospective randomized trial; it is a retrospective analysis of the database from a single cardiovascular centre between 2003-2011. In this period 7,100 patients underwent CABG with 16,984 undergoing PCI. This should correspond to 835 first time isolated CABG (excluding redo surgery and concomitant valve surgery) and 2000 PCI per year. This is a high volume tertiary cardiac centre. The provided outcome data (in-hospital mortality, MACE, short-term and long-term mortality) are low and in the range of published registries in other countries with well organized cardiac care. For example, the pooled mortality at 1-year follow-up after PCI was 5.8%. The Swedish registry reports for the period 2007-2008 (in the middle of the sampling of the present study) a global mortality at 1-year follow up after PCI of 5.7%, with a progressive increase in mortality over the years.3 The progressive increase in mortality after PCI is due to the increasing inclusion of complex anatomies and unstable patients.

Only cardiac centres that are big enough are likely to provide optimal care for myocardial revascularization and to obtain lower MACE, lower short- and long-term mortality. For CABG the most common benchmark is a threshold of ≥200 operations per year,4 although some reimbursement groups use a higher cut off of 450 operations per year to select providers.5 The Belgian CABG Registry (2004 data) indicates that centres with less than 200 operations per year had an in-hospital mortality rate that was 80% higher than centres with more than 200 operations per year and that this was not attributable to more severe disease.6 Similarly for PCI, the most frequently applied benchmark is an upper threshold of ≥400 PCI per year.4,5 The AHA/ACC recommend more than 400 PCI per centre and per year (class IB recommendation).7 A multivariate analysis from the Belgian PCI Registry 2006-2008 indicates that lowering of in-hospital mortality after PCI as a function of hospital volume is even observed until a cutoff of 800 PCI per year.8

In addition to hospital volume, operator volume could prove to be important. For most procedures, the mortality rate is higher among patients of low-volume surgeons than among those of high-volume surgeons, regardless of the surgical volume of the hospital in which they practiced.9 The effect seems to be of limited magnitude for isolated CABG surgeons, but could be more important for PCI operators. Available evidence from observations and registries however lead to contradictory results and there is an obvious interaction between centre volume and operator volume, i.e. operator volume seems less important in high volume centres.10 Current ACC/AHA/SCAI guidelines nevertheless advise that interventional cardiologists would perform >75 PCI yearly to maintain their procedural proficiency (recommendation class IB).7 The SCAI (US Society for Cardiovascular Angiography and Intervention) challenges this point of view and argues that in the stent era there would be no relation anymore between operator volume and in-hospital mortality.11 For the SCAI the importance is to work in a centre of excellence with optimized quality control. The SCAI acknowledges the need for validated risk-adjustment models, which include all major complications, not just mortality, as the standard of quality. In our interpretation volume remains a prerequisite but is not sufficient (see below). In real life there is a considerable skew in the distribution of per-operator volume. Maroney et al. assessed the per-operator volume PCI procedures among US cardiologists in the 2008 database of Medicare.12 A significant majority of interventional cardiologists (61%) performed 40 or fewer PCI in 2008. Although these (at least remarkable) data might not be fully representative of the US situation, it illustrates the need for better surveillance regarding the clinical scenarios under which PCI procedures should be performed. Such observations are not without clinical implications. It may be reasonably assumed that less experienced operators predominantly perform lower risk procedures in lower risk patients. Surprisingly some registry data show the opposite, i.e. higher pathology scores for low volume operators. In addition, the cut offs are variable and essential data concerning anatomical complexity of coronary lesions are sparse in most older registries and observations. It may be a very sensitive issue for the organization of revascularization, but the relation between volume and quality of care may not be ignored when considering the quality of training. The Leapfrog group indicates for operators with volumes <75 per year a 35% increase in hospital mortality after case-mix correction (referred in ref.6). When compared to the entire Belgian PCI registry 2006-2008, operators with <75 interventions per year have an observed unadjusted hospital mortality that is 51% higher than the predicted mortality (Belgian PCI registry, unpublished data).

A high-quality PCI programme requires appropriately trained and skilled physician operators. However, the operator does not work in a vacuum. An operator needs a well-maintained, high-quality cardiac catheterization facility to practice effectively. In addition, the operator depends on a multidisciplinary institutional infrastructure for support and response to emergencies. Thus, to provide quality PCI services, the institution must ensure that its catheterization facility is properly equipped and managed and that all of its necessary support services are of high quality and are readily available.7 This editorial is not meant to trigger a major discussion on CABG and PCI volumes, but mentions these sensitive issues in order to emphasize that the results of the study under scrutiny by Jones et al. might only be valid in a centre that is big enough and that allows trainees to perform a sufficient volume of operations or interventions.

A training centre in interventional cardiology should provide on top structured and progressive education allowing the trainee to start with simple cases and gradually progress toward interventions in vulnerable patients, patients with unstable syndromes and patients with higher SYNTAX scores. In unadjusted analyses from Jones et al.,1 consultants perform CABG and PCI with higher mortality and morbidity than trainees. Primary operator status is no longer associated with outcome in adjusted analyses, which implies that trainees preferentially perform easier cases. Both for CABG and PCI, the authors analysed long-term mortality in patients stratified by propensity score and operator grade and showed no difference between consultants and trainees. The database unfortunately didn't include a measure of complexity of the anatomical coronary lesions, such as the SYNTAX score. The propensity analysis confirms that trainees do less high-risk procedures than consultants. Of note, in the highest quintile of propensity score, trainees still perform 13.1% of the CABG and 38.5% of the PCI. Unfortunately the database does not identify the year of training for the trainee operators. The data do not allow telling, but we just can presume that these more complex interventions by trainees are performed at the end of the training.

Some particular aspects of the database published by Jones et al. deserve to be noticed. Consultants do much more cases in cardiogenic shock, which might suggest that, even in emergencies, waiting for the boss still is (too) common. We also note the high use of glycoprotein IIb/IIIa inhibitors (44% for consultants and 35% for trainees). This presumably will no more be so after the publication of more recent guidelines.1315

In conclusion, myocardial revascularization is safe in trainees' hands, provided that the interventions are performed in a high volume centre, allowing the trainee to perform sufficient interventions under the umbrella of a structured and continuously supervised training programme. In the end this will deliver a cardiac surgeon or an interventional cardiologist having been trained to work in a well functioning Heart Team13,16 in order to provide optimal and global care to the cardiac patient, rather than solely revascularization.

Conflict of interest: Thierry Gillebert was leader of the task force “Core Curriculum 2013” of the ESC Committee for Education and is corresponding author of the manuscript ESC Core Curriculum for the General Cardiologist (2013).


  • doi:10.1093/eurheartj/eht161.


View Abstract