American College of Cardiology

A. Quality Assurance

A mechanism for valid peer review must be established and ongoing at each institution performing PCI. Interventional cardiology procedures are associated with complications that in general are inversely related to operator and institutional volume ^(43,180-183). The mechanism for institutional review should provide an opportunity for interventionalists as well as physicians who do not perform angioplasty, but are knowledgeable about it, to review overall results of the program on a regular basis. The responsible supervising authority should monitor the following issues as outlined in Table 12.

The institutional credentialing committee should document that an interventionalist wishing to start practice meets the established training criteria, including those of the ACC Task Force on Training in Cardiac Catheterization and Interventional Cardiology ^(21,185,186). The ACC Training Statement ⁽¹⁸⁶⁾ for coronary invasive training requires a 3-year comprehensive cardiac program with 12 months of training in diagnostic catheterization during which the trainee performs 300 diagnostic catheterizations with 200 of those being the primary operator. The interventional training requires a fourth year of fellowship during which the trainee should perform more than 250 interventional procedures, but not more than 600 per year ⁽¹⁸⁶⁾. To be eligible for the American Board of Internal Medicine (ABIM) certifying examination in interventional cardiology, a trainee must be actively involved in at least 250 interventional procedures during a 4th year of interventional cardiology fellowship. Only one trainee may receive credit for the intervention on a given patient. Until 2003, the practicing interventionalist can qualify for the examination by active involvement in interventional cardiology, including the performance of at least 150 interventions over the prior 2 year period ⁽¹⁸⁷⁾. Credentials committees should evaluate the physicians’ outcomes to be certain that volume and results meet the current standards or benchmarks for successful management ⁽²¹⁾. These benchmarks refer to procedural rates of unadjusted mortality (0.9%) and emergency coronary artery bypass surgery (3.0%). It should be noted that these benchmarks are derived from PTCA performed in New York State on all procedures including those for complicated acute MI and that they were gathered before the use of stents and platelet GP IIb/IIIa inhibitors. Thus, the standard for benchmark complication rates will be subject to future revision as newer data emerge. It is important that institutions assist with these efforts by participating in active database efforts to track clinical and procedural information for individual operators amd their institutions. In the future, certification by the ABIM in Interventional Cardiology should be required.

This Writing Committee agrees with the ACC Task Force recommendations for the Assessment and Maintenance of Proficiency in Coronary Interventional Procedures ⁽²¹⁾. Institutions performing PCI should meet the following standards as outlined in Tables 13 and 14 ^(21,184,186).

B. Operator and Institutional Volume

The proliferation of small angioplasty or small surgical programs to support such angioplasty programs is strongly discouraged. Several studies have identified procedural volume as a determining factor for frequency of complications with PCI ^{(43,182,183,188-191)}. Kimmel, using data from the Society of Cardiac Angiography and Interventions (SCA&I), found that an inverse relationship existed between the number of angioplasty procedures performed at a hospital and the rate of major complications ⁽¹⁸¹⁾. These results were risk-stratified and independent of the patient-risk profile. Significantly fewer complications occurred in laboratories performing 400 angioplasty procedures per year. Conversely, low-volume hospitals were associated with higher rates of emergency coronary artery bypass surgery and death ⁽¹⁸²⁾. Improved outcomes were identified with a threshold volume of 75 Medicare angioplasties per physician and 200 Medicare angioplasty procedures per hospital. Using a 35 to 50% ratio of Medicare patients, the threshold value was 150 to 200 angioplasty procedures per cardiologist and 400 to 600 angioplasty procedures per institution ⁽⁴⁰⁾. Other studies have also supported the relationship of complications to procedural volume ^(43,180,183). Although some investigators have suggested that low procedure volume does not contribute to poor outcomes ^(188,192), these studies are small in number and underpowered for analysis ⁽¹⁸⁹⁾. Development of small cardiovascular surgical programs to support angioplasty is a poor use of resources that will likely lead to suboptimal results ⁽¹⁹⁰⁾.

Given the concerns regarding operator volume and surgical standby, it is recommended that PCI be performed by higher volume operators (75 cases/year) with advanced technical skills (e.g., subspecialty certification) at institutions with fully equipped interventional laboratories and experienced support staff. This setting will most often be in a high-volume center (>400 cases/year) associated with an on-site cardiovascular surgical program ⁽¹⁹³⁾. Similar concerns have been identified and supported by the Task Force for Practice Guidelines for Coronary Angiography ⁽¹⁹⁴⁾.

Intuitively, it is clear that it would be best for the rare patient requiring surgery after elective PCI to remain in the same hospital rather than have the patient and family undergo the confusion, stress, and anxiety of emergency transfer. Given the widespread availability of sophisticated interventional/surgical programs in the United States, it is difficult to demonstrate a need for additional low-volume programs to do elective angioplasty except in underserved areas that are geographically far removed from major centers. This Committee acknowledges that not every cardiologist desiring to do PCI should perform these procedures and not every hospital anxious to have an interventional program should start one ⁽¹⁹¹⁾. This caveat is particularly true where there are high-volume programs and operators nearby. In these situations, operators should be subspecialty board certified.

The Committee, therefore, recommends that angioplasty is best done by high-volume operators in high-volume institutions. Any change in this recommendation awaits further data confirming the comparable safety and outcomes for patients treated in an alternative manner. The Committee cannot recommend angioplasty by low-volume operators (<75 cases/year) working in low-volume institutions (<200 cases/year) with or without on-site surgical coverage. As noted earlier, ongoing investigational experience and clinical data are mandatory if these recommendations are to be modified.

Recommendations for PCI Institutional and Operator Volumes at Centers With Onsite Cardiac Surgery ^(21,186) (Table 14)

Class I
1. PCI done by operators with acceptable volume (75) at high-volume centers (>400). (Level of Evidence: B)

Class IIa
1. PCI done by operators with acceptable volume (75) at low-volume centers (200-400). (Level of Evidence: C)

2. PCI done by low-volume operators (<75) at high-volume centers (>400). Note: Ideally operators with an annual procedure volume <75 should only work at institutions with an activity level of >600 procedures/year.* (Level of Evidence: C)

Class III
1. PCI done by low-volume operators (<75) at low-volume centers (200-400). Note: An institution with a volume <200 procedures/year, unless in a region that is underserved because of geography, should carefully consider whether it should continue to offer service.* (Level of Evidence: C)

*Operators who perform <75 procedures/year should develop a defined mentoring relationship with a highly experienced operator who has an annual procedural volume >150 procedures/year.

C. On-Site Cardiac Surgical Backup

Cardiac surgical backup for PCI has evolved from the formal surgical standby in the 1980s to an informal arrangement of first available operating room and, in some cases, off-site surgical backup ^(40,195-199). With the advent of intracoronary stenting, there has been a decrease in the need for emergency coronary artery bypass, ranging between 0.4 and 2% ^(200-202). Not surprisingly, emergency coronary artery bypass for a patient with an occluded or dissected coronary artery is associated with a higher mortality than elective surgery ^(203-208). Emergency procedures are also associated with high rates of perioperative infarction and less frequent use of arterial conduits. Complex CAD intervention, hemodynamic instability, and prolonged time to reperfusion are contributing factors to the increased risk of emergency bypass surgery.

1. Primary PCI Without On-Site Cardiac Surgery. Although thrombolytic trials demonstrated that early reperfusion saves myocardium and reduces mortality ^(209-212), the superiority and greater applicability of primary PCI for the treatment of acute MI has raised the question of whether primary PCI should be performed at institutions with diagnostic cardiac catheterization laboratories that do not perform elective PCI or have on-site cardiac surgery. For this reason, the establishment of PCI programs at institutions without on-site cardiovascular surgery has been promoted as necessary to maintain quality of care ^{(195-197,213-220)}. In those patients where there is a contraindication to thrombolytic therapy, or when there are complications such as cardiogenic shock, catheter-based therapy may limit infarct size ^(221,222). It must be realized that PCI in the early phase of an acute MI can be difficult and requires even more skill and experience than routine PCI in the stable patient. The need for an experienced operator and experienced laboratory technical support ⁽²²³⁾ with availability of a broad range of catheters, guidewires, stents, and other devices (e.g., IABP) that are required for optimum results in an acutely ill patient is of major importance (Table 15). If these complex patients are treated by interventionalists with limited experience at institutions with low volume, then the gains of early intervention may be lost because of increased complications. In such circumstances, transfer to a center that routinely performs complex PCI will often be a more effective and efficient course of action ⁽¹⁶⁾. Thrombolysis is still an acceptable form of therapy ⁽²²⁴⁾ and is preferable to acute PCI by an inexperienced team ^(224,225).

Reports of emergency primary angioplasty programs from hospitals without established open-heart surgery or elective angioplasty, similar to those of most tertiary centers, have demonstrated generally favorable results. Such acceptable clinical results have been reported with intensive training, continuous oversight, and the combination of nearby, readily available bypass surgery support, a team of highly experienced interventionalists and support staff, and careful patient selection ⁽²¹⁴⁾. However, poor results of similar endeavors are rarely reported. Before the use of stenting and glycoprotein receptor blockers, primary angioplasty in certain hospitals has been associated with acute mortality rates greater than those reported from centers with established primary angioplasty programs. Overall, in-hospital mortality rates have ranged from 1.4 to 13% ^{(196,197,216)}.

Criteria have been suggested for the performance of primary PCI at hospitals without on-site cardiac surgery (Tables 15 and 16). Of note, large-scale registries have shown an inverse relationship between the number of primary angioplasty procedures performed and in-hospital mortality ^(226-228). The data suggest that both door-to-balloon time and in-hospital mortality are significantly lower in institutions performing more than 36 primary angioplasty procedures per year ⁽²²⁹⁾. Communities may identify a unique qualified and experienced center wherein the on-site intervention for acute MI could be performed. Suboptimal results may relate to operator/staff inexperience and capabilities and delays in performing angioplasty for logistical reasons ⁽²³⁰⁾. From clinical data and expert consensus, the Committee recommends that primary PCI for acute MI performed at hospitals without established elective PCI programs should be restricted to those institutions capable of performing a requisite minimum number of primary angioplasty procedures (36 per year) with a proven plan for rapid and effective PCI as well as rapid access to cardiac surgery in a nearby facility ⁽¹⁹³⁾ (Table 17).

2. Elective PCI Without On-Site Surgery. Technical improvements in interventional cardiology have led to the development of elective angioplasty programs without on-site surgical coverage. Several centers have reported satisfactory results based on careful case selection with well-defined arrangements for immediate transfer to a surgical program ^{(195-199,231-235)}. The studies of angioplasty without on-site surgical coverage have not identified significant differences in the outcomes, recalling the infrequent rate of complications ⁽²³⁶⁾. Despite many reported successful angioplasty series without on-site surgical backup and a very low percentage need for off-site surgery in failed angioplasty, some clinicians have expressed concern ^(237,238) about the appropriateness of elective angioplasty in centers without on-site surgical coverage. Caution is warranted before endorsing an unrestricted policy for PCI in hospitals without appropriate facilities. Several outstanding and critically important clinical issues, such as timely management of ischemic complications, adequacy of specialized post-interventional care, logistics for managing cardiac surgical or vascular complications and operator/laboratory volumes, and accreditation must be addressed. Mere convenience should not replace safety and efficacy in establishing an elective PCI program without on-site surgery.

At this time, the Committee, therefore, continues to support the recommendation that elective PCI should not be performed in facilities without on-site cardiac surgery (Table 17). As with many dynamic areas in interventional cardiology, these recommendations may be subject to revision as clinical data and experience increase.

Recommendations for PCI With and Without On-Site Cardiac Surgery (Table 17)

Class I
1. Patients undergoing elective PCI in facilities with on-site cardiac surgery. (Level of Evidence: B)

2. Patients undergoing primary PCI in facilities with on-site cardiac surgery. (Level of Evidence: B)

Class IIb
1. Patients undergoing primary PCI in facilities without on-site cardiac surgery, but with a proven plan for rapid access (within 1 h) to a cardiac surgery operating room in a nearby facility with appropriate hemodynamic support capability for transfer. The procedure should be limited to patients with ST-segment elevation MI or new LBBB on ECG, and done in a timely fashion (balloon inflation within 90 ± 30 min of admission) by persons skilled in the procedure (75 PCIs/year) ⁽¹⁹³⁾ and only at facilities performing a minimum of 36 primary PCI procedures per year ⁽²²⁹⁾. (Level of Evidence: B)

Class III
1. Patients undergoing elective PCI in facilities without on-site cardiac surgery. (Level of Evidence: C)

2. Patients undergoing primary PCI in facilities without on-site cardiac surgery and without a a proven plan for rapid access (within 1 h) to a cardiac surgery operating room in a nearby facility with appropriate hemodynamic support capability for transfer or when performed by lower skilled operators (<75 PCIs per year) in a facility performing <36 primary PCI procedures per year. (Level of Evidence: C)