- Email: [email protected]
Nonemergent percutaneous coronary interventions in a veterans affairs medical center without onsite cardiac surgery
Nonemergent percutaneous coronary interventions in a veterans affairs medical center without onsite cardiac surgery Oleg Roussanov, MD, Greta Estacio, FNP, Maribeth Capuno, ANP, Linda Archuleta, RN, Judith Hill, FNP, and Nabil F. Jarmukli, MD Salem, VA
Background Percutaneous coronary intervention (PCI) performed in centers without onsite cardiac surgery remains controversial. Advances in PCI techniques and medical therapy have markedly decreased postprocedural complications. Our aim was to assess the efficacy and safety of performing PCI in the Veterans Affairs patient population in a hospital without onsite cardiac surgery. Methods We prospectively evaluated 401 consecutive patients who underwent elective PCI or PCI after admission for acute coronary syndrome. Patients who had ST-elevation myocardial infarction or were hemodynamically unstable were classified as emergent and had their PCI performed elsewhere and were therefore excluded from our analysis. Our cardiac surgery backup was a community hospital 8 miles away. Results The patient’s mean age was 65.6 F 10 years, and most were men (99.5%). Patients had high-risk clinical and angiographic profiles, with diabetes mellitus in 44%, prior myocardial infarction in 41%, comorbid conditions in 45% and type B or C angiographic lesions in 83%. Of 401 patients, 338 (84%) received a stent, and 86 (21%), a drug eluting stent. Percutaneous coronary intervention success rate was 97%. There were no periprocedural or inhospital deaths, and no patients required emergency transfer for cardiac surgery. At 1 and 6 months of follow-up, total mortality was 1.5% and 3.5%, respectively; target vessel revascularization rate was 0% and 1.7%. Conclusions
Nonemergent PCI can be performed effectively and safely in patients with higher clinical and angiographic risk without onsite backup cardiac surgery. This has significant implications for most hospitals that have an invasive but not an interventional program. (Am Heart J 2006;152:9092 13.)
The number of percutaneous coronary interventions (PCIs) performed annually in the United States and worldwide continues to grow. Parallel to this has been the improvement in technology, which allows for increasing safety and success of PCIs. This adds new fuel to the ongoing debate whether PCIs should be performed in centers without onsite cardiac surgical capability. Many such hospitals are providing care to the patients living in rural areas of the country who have limited access to medical services. If safety and outcomes of nonemergent PCIs performed without onsite surgical backup were shown to be comparable with ones done in centers with onsite cardiac surgery,
From the Salem Veterans Administration Medical Center, Salem, VA. Submitted October 31, 2005; accepted May 22, 2006. Reprint requests: Nabil Jarmukli, MD, FACC, Cardiology, Salem VAMC, Cardiology111B, Salem, VA 24153. E-mail: [email protected] 0002-8703/$ - see front matter Published by Mosby, Inc. doi:10.1016/j.ahj.2006.05.016
a large number of smaller hospitals with catheterization laboratories could improve patient care and remain competitive. To date, no randomized clinical trials to address this issue have been performed. Most of the data accumulated so far come from single institution reports and retrospective registry reviews. Although the recent comparative analysis of Medicare files from a large number of hospitals1 suggested a higher risk of adverse outcomes in centers without onsite cardiac surgery (for nonprimary/rescue PCI at hospitals performing b50 Medicare PCIs per year), another recent report by Foster et al2 found that on-site surgical backup was not a predictor of lower inhospital mortality after PCI in a heterogeneous group of ST-elevation myocardial infarction (STEMI), non–ST elevation myocardial infarction (NSTEMI), and non–myocardial infarction (MI) patients derived from the American College of Cardiology National Cardiovascular Data Registry (ACC-NCDR) registry. Several individual center reports3-6 seem to indicate an excellent safety and efficacy profile. A report by Turgeman et al5 from a single site in Israel with no
American Heart Journal November 2006
910 Roussanov et al
Table I. Clinical characteristics of patients who underwent elective PCI and PCI in the setting of ACS
Table II. Coronary lesion types in patients who underwent elective PCI and PCI in the setting of ACS Coronary lesion type (%)
No. of patients (%) Characteristic
Elective PCI ACS with PCI All patients
Age (y) (mean F SD) N65 years Hypertension Diabetes mellitus Cerebrovascular disease Peripheral vascular disease Chronic renal insufficiency Severe pulmonary disease Prior coronary artery disease Prior MI Prior CABG Prior PCI Prior congestive heart failure Left ventricular ejection fraction (%) (mean F SD)
65.7 F 9.7 114 (53) 188 (87) 97 (45) 27 (12)
65.5 F 10.4 93 (50) 162 (88) 78 (42) 19 (10)
65.6 F 10 207 (52) 350 (87) 175 (44) 46 (11)
60 (28)
36 (19)
96 (24)
39 (18)
44 (24)
83 (21)
11 (5)
8 (4)
19 (5)
164 (76)
121 (65)
285 (71)
91 65 81 41
(42) (30) (37) (19)
48 F 11
75 50 64 29
(40) (27) (35) (16)
45 F 13
166 (41) 115 (29) 145 (36) 70 (17) 47 F 12
onsite backup surgery showed that referral for emergency surgery was 1.3% among 235 patients who had balloon angioplasty and 0.6% among 1016 patients who had coronary stenting. Similarly, another single center report by Zavala-Alarcon et al6 from Phoenix, AZ, showed that the referral for emergency cardiac surgery was 0.7% among 1000 patients who had elective coronary intervention with a periprocedural death of 0.2% and MI of 2.1%. A potential criticism of the singlecenter reports is the bias to limit PCIs to selected group of very low-risk patients. Yet, these articles appropriately emphasize the importance of careful case selection at hospitals with off-site surgical backup. We had an opportunity to prospectively follow up a large higher-risk cohort of patients typical of a Veterans Affairs hospital. This allowed us to take advantage of minimal outside referral and thorough observation practices and obtain a reliable data on immediate and long-term safety and outcomes of PCIs performed without onsite cardiac surgery.
Methods Hospital and PCI program Salem Veterans Affairs Medical Center (VAMC) is a level I, 182-bed acute care medical/surgical/psychiatric teaching hospital affiliated with the University of Virginia, Charlottesville. It is located in southwest Virginia and is quite distant from several VAMCs, which serve as receiving centers for coronary intervention candidates: 150 miles from Durham VAMC, NC,
Type A (low risk) Type B (moderate risk) Type C (high risk)
Elective PCI
ACS with PCI
All PCIs
14 61 25
20 58 22
17 60 23
169 miles from Richmond VAMC, VA, and 228 miles from Asheville VAMC, NC. Salem VAMC has had an invasive program for N20 years but no interventional program. The cardiology section at Salem VAMC is equipped with 2 cardiac catheterization laboratories (Phillips Poly C and Integris Allura, Shelton, CT), both with digital capabilities. In March of 2002, Salem VAMC obtained formal approval from the Veterans Affairs Central Office to initiate a PCI program. Approval was restricted to elective PCIs and patients with acute coronary syndrome (ACS), including NSTEMI. Approval was not given for patients with STEMI or hemodynamically unstable patients because they were considered emergent. In July of 2002, the first patient underwent PCI. All procedures were performed by 1 of 3 coronary interventional board-certified cardiologists with an experience of at least 300 diagnostic cardiac catheterizations and 85 coronary interventional cases per operator per year in a cardiac catheterization laboratory (CCL) with at least 600 diagnostic cases and 170 interventional cases per year. This was preceded by formal training of CCL and cardiac recovery nurses and technicians in preoperative, intraoperative, and postoperative care. Our surgical backup is a private hospital with an established cardiac surgery program located 8 miles from Salem VAMC. Before the start of every case, a point person from Salem VAMC CCL would contact the surgical team at the receiving hospital to get approval for the time of the procedure. A fully equipped advanced cardiac life support–capable ambulance was on standby at Salem VAMC emergency room door with its personnel stationed in a room next to the CCL for the duration of the case. Before the initiation of the program, and biyearly thereafter, a bdry runQ to ensure PCI failure safe transport procedure was performed with door-to-door ambulance transport time range of 10 to 15 minutes.
Patients and procedures All consecutive patients who were referred for nonemergent PCI between July 2002 and November 2004 were included in the study. Although primary PCI for STEMI was not performed in our institution, both patients who were scheduled for PCI electively (from cardiology and primary care clinic referrals as well as referrals from the nuclear and treadmill stress laboratories), and those admitted to the hospital with ACS including NSTEMI were included. Percutaneous coronary interventions on patients who were hemodynamically unstable or had active symptoms at the time of procedure were not performed at Salem VAMC and were directly referred to one of the local private hospitals.
American Heart Journal Volume 152, Number 5
Roussanov et al 911
Table III. Clinical outcomes of patients who underwent elective PCI and PCI in the setting of ACS No. of patients (%) Elective PCI Characteristic Total mortality Noncardiac death Myocardial infarction ACS PCI Coronary bypass surgery Cardiac hospitalizations Target vessel revascularization
1m 3 2 0 1 0 0 3 0
(1.4) (0.9) (0.5)
(1.4)
ACS with PCI 6m
6 (2.8) 2 (0.9) 0 7 (3.3) 4 (1.5) 1 (0.5) 12 (5.6) 4 (1.5)
1m 3 2 0 0 0 1 1 0
(1.6) (1.1)
(0.5) (0.5)
All patients
6m 8 3 1 7 6 2 11 3
(4.3) (1.6) (0.5) (3.8) (3.2) (1.1) (5.9) (1.6)
1m 6 4 0 1 0 1 4 0
(1.5) (1) (0.2) (0.2) (1)
6m 14 5 1 14 10 3 23 7
(3.5) (1.2) (0.2) (3.5) (2.5) (0.7) (5.7) (1.7)
Standard techniques were used for PCI, and a variety of intracoronary stents (bare metal and drug eluting) were deployed when appropriate. Low- to high-risk coronary lesions in patients with low- to high-risk clinical profiles were treated, provided that it could be safely and effectively performed using available equipment and techniques. Rare patients requiring specialized intracoronary procedures (laser, brachytherapy, directional coronary atherectomy, or rotablator) were referred to outside centers and excluded from this study. Standard medical therapy was used pre, intra, and post procedure. Patients were then followed up prospectively with regular clinical visits within the Veterans Affairs system. Baseline clinical information, including demographics, comorbid conditions, mode of presentation, lesion complexity, and outcome data, including procedural complications, mortality, hospitalizations, and repeat revascularization, were analyzed and presented below as percentages or mean F 1SD. The definitions of periprocedural PCI complications and lesion types were used as presented in ACC–National Cardiovascular Data Registry (http://www.acc.org/ncdr/cathlab.htm). All other definitions used were also based on the same registry.
0.8 stents per lesion. Drug-eluting stents (all Taxus) were used in 86 (21%) patients. The proportion of coronary lesion types is presented in Table II. Eightythree percent of all lesions were higher-risk (types B and C). Clopidogrel was given as 300 mg loading and 75 mg daily maintenance dose to all patients undergoing stent placement. All patients received aspirin 81 to 324 mg unless contraindicated. h-Blockers, statins, and angiotensin-converting enzyme inhibitors were used in 77%, 78%, and 61% of patients, respectively, as of the procedure day. Platelet IIb/IIIa inhibitors were administered during the procedure in 75 (19%) patients. Vascular access closure device was deployed in 40% of cases. During this period, 31 patients were excluded from the study: 21 with STEMI who underwent reperfusion with primary PCI off site or fibrinolytic treatment on site, 7 because of refractory to medical therapy ACS requiring emergent intervention, and 11 due to complex anatomy requiring atherectomy, rotablator, or other specialized procedures.
Results
Clinical outcomes Seven patients (1.7%) of the total group had periprocedural complications. One patient developed transient renal failure after elective PCI. The remaining 6 complications (3%) occurred in patients with ACS. No patients required transfer for urgent surgical treatment. No periprocedural/inhospital deaths or cerebrovascular accidents were recorded. One patient had NSTEMI because of no-reflow after stenting venous bypass graft with mild enzyme elevation and no clinical sequelae. The incidence of major vascular complication, arrhythmias, and renal failure were 0.75%, 0.5%, and 0.5%, respectively. One patient developed cardiogenic shock in the setting of coronary ischemia because of occluded venous bypass graft, recurrent ventricular tachycardia, and ventricular fibrillation. Patient quickly recovered after PCI and placement of intra-aortic balloon pump.
Patients and procedures A total of 401 patients underwent PCI between July 2002 and November 2004. The procedure was successful in 97% of cases. Success was defined as b20% residual stenosis associated with TIMI III flow. Patients were almost exclusively men (99.5%). Ages ranged from 39 to 90 years. Two hundred seven (52%) were elderly (N65). Demographic and clinical characteristics are given in Table I. Of the patients, 25% and 22% had multivessel and multistage procedures, respectively. Two hundred sixteen (54%) patients had PCI as an elective procedure. One hundred eighty-five (46%) were admitted to the hospital with ACS (including NSTEMI) and assigned to early invasive strategy (within 24-48 hours). A total of 608 coronary lesions were treated with PCI. Stents were used in 338 patients (84%) with a ratio of
912 Roussanov et al
Follow-up data were obtained in all patients because of the fully computerized medical record of the Veterans Affairs system in place for N6 years now and computer connectivity with the Veterans Affairs satellite clinic network. Data on 1- and 1-month clinical outcome are presented in Table III. Total mortality was 1.5% and 3.5% at 1 and 6 months, respectively. Myocardial infarction occurred in only 1 patient, whereas ACS occurred in 3.4% of all patients at 6 months of follow-up. Requirement for PCI at 6 months was 2.5%, and for coronary artery bypass graft (CABG) surgery, 0.7%. There were no documented cerebrovascular accidents on follow-up.
Discussion There is continued controversy regarding the performance of PCI in hospitals without onsite backup cardiac surgery. The most recent 2005 American College of Cardiology/American Heart Association guidelines restrict class I recommendation to centers with onsite cardiac surgical capability. As for centers without onsite cardiac surgery, guidelines label primary PCI for acute STEMI as class IIb (acceptable) and elective PCI as class III (not acceptable).7 Yet the guidelines clearly state that several centers have reported satisfactory results based on careful case selection with well-defined arrangements for immediate transfer to a surgical program. However, it is quite evident from the literature that over the past 5 years, continued advances in coronary stenting techniques as well as aggressive periprocedural antiplatelet therapy and enhanced operator experience has led to significant reduction in periprocedural complications requiring emergency bypass surgery. A published report by the Cleveland Clinic foundation on their interventional program showed a significant decline (N10-fold) in the prevalence of emergency CABG from 1.5% in 1992 to 0.14% in 2000 ( P = .0010).8 Ting et al3 reported the results of low-risk PCI on type A and B1 lesions in low-risk patients performed in a community hospital without onsite cardiac surgery with a backup hospital 85 miles away. Procedural success was achieved in 99.5% of patients with 0.5% inhospital deaths. Yang et al9 reported the changing trend at Mayo Clinic for emergency cardiac surgery referral post coronary intervention over a period of 24 years. Although during the prestent era the incidence of their emergency surgical referral was 2.9%, it dropped to 0.7% during the early-stent era and even lower to 0.3% during the later-stent era. This drop in emergency surgical referral was despite the increasing trend of performing PCI on a higher-risk population who were older and had a higher prevalence of diabetes mellitus, lower left ventricular systolic function, higher incidence of prior revascularization, and more complex coronary anatomy.
American Heart Journal November 2006
Lotfi et al10 reported on the Canadian experience from the University Health Network in Toronto and noted that although b1 in 100 patients needed emergency cardiac surgery post PCI, 1 in 4 patients referred for emergency cardiac surgery would be placed at increased risk of harm if delays to surgery were encountered. Consequently, we initiated our PCI program after obtaining formal approval from the Veterans Affairs central office to test the feasibility, efficacy, and safety of performing PCI without onsite cardiac surgical capability in the Veterans Affairs hospital system. Our results clearly showed that PCI was successful in 97% of patients enrolled with an excellent safety profile because there were no deaths or need for urgent cardiac surgery. The low rate of MI, arrhythmic, vascular, and renal complications are comparable with rates in patients undergoing PCI with onsite cardiac surgery backup1. Our results of 1-month 1.5% mortality and 0.2% CABG requirement compare favorably to the 2.8% 30-day mortality and 1.1% CABG requirement in 16 184 Medicare patients with nonprimary/nonrescue PCI performed at centers with surgical backup. It is noteworthy that our study included patients with higher clinical and angiographic risk profiles, with ACS in 46%, age N65 years in 52%, diabetes mellitus in 44%, comorbid conditions (cerebrovascular, peripheral vascular, renal, pulmonary disease) in 45%, prior MI in 41%, and type B and C angiographic lesions in 83%. Our results are very similar to the Polish experience where Dudek et al11 reported on the efficacy and safety of performing nonemergent PCI on NSTEMI or patients with stable angina without onsite surgical backup. In a cohort of 479 patients (160 with NSTEMI and 319 with stable angina), the inhospital outcome was equally favorable in both groups with MI in 2.5% versus 1.6% and urgent referral for surgery in 0.6% versus 0.3%. At 36 months of follow-up, the incidence of MI was 0.7% and death, 2.0% in both groups. Several reports have elucidated an inverse relationship between operator procedural volume, as well as hospital CCL volume and periprocedural complication rates. The Duke Research Group examined Medicare patient registry and showed improving outcomes seen at a threshold of 75 cases per operator and 200 cases per hospital.12 Current American College of Cardiology/ American Heart Association guidelines for PCI7 recommend a minimum operator volume of 75 interventions per year (class I) at high volume center with N400 yearly onsite elective interventions. Our operator and hospital interventional volume per year were at least 85 and 170 cases respectively. Our results have significant impact on a large number of Veterans Affairs and community hospitals that have an active cardiac catheterization laboratory but no interventional program. By meeting certain safety standards, such hospitals can serve their patients more effectively
American Heart Journal Volume 152, Number 5
and expediently, in lieu of transferring patients for intervention after cardiac catheterization. It should be clear that our results do not apply to the higher-risk patients such as acute STEMI or hemodynamically unstable patients. The principal limitation of this study is its cohort and not a randomized controlled design. A control group from the receiving hospital could not be used for comparison because the receiving hospital cares for a nonveteran population with different demographics and risks. In addition, the fact that patients with STEMI or unstable hemodynamics were not enrolled could have skewed the data to a better outcome. Furthermore, serial troponin or MB fraction of creatine phosphokinase were not routinely measured post intervention, and therefore, the number of asymptomatic MIs is probably underestimated. Finally, the delay in transfer after a failed or complicated angioplasty is critical to the outcome, and we had no patients who required emergency transfer. We conclude that if rigorous benchmarks of operator experience, safety standards, patient screening, proper communication, intense hospital and personnel commitment and adherence to high program standards are met, then nonemergent PCI can be performed effectively and safely in a hospital without onsite cardiac surgery, even in patients with ACS and type B and C angiographic lesions. Prospective randomized clinical trial is needed to answer the bigger question of whether outcomes of nonemergent PCI are equivalent at hospitals with and without onsite cardiac surgery. Our data demonstrate that it is safe to launch such a trial. We thank Dr Nelson Bernardo and Dr Milan Kothari for initiating our PCI program and performing most the interventions. We also thank the following for their participation in the care of those patients as well as gathering some of the data and assisting in the analysis: Kathy Henley, NP, Brenda Dogan, RN, Brenda Hart, RN, and Jeannie Wilson, RN.
Roussanov et al 913
References 1. Wennberg DE, Lucas FL, Siewers AE, et al. Outcomes of percutaneous coronary interventions performed at centers without and with onsite coronary bypass graft surgery. JAMA 2004;292: 1961 - 8. 2. Foster JK, Klein LW, Veledar E, et al. On-site Surgical Backup as a Predictor of Mortality Among PCI Patients with STEMI, NSTEMI, or NO AMI. Circulation 2005;112:II-737. 3. Ting HH, Garratt KN, Singh M, et al. Low-risk percutaneous coronary interventions without on-site cardiac surgery: two years’ observational experience and follow-up. Am Heart J 2003;145: 278 - 84. 4. Paraschos A, Callwood D, Wightman MB, et al. Outcomes following elective percutaneous coronary intervention without on-site surgical backup in a community hospital. AJC 2005;95:1091 - 3. 5. Turgeman Y, Atar S, Suleiman K, et al. Diagnostic and therapeutic percutaneous cardiac interventions without on-site surgical backup—review of 11 years experience. Isr Med J Assoc 2003; 5:89 - 93. 6. Zavala-Alacron E, Cecena F, Ashar R, et al. Safety of elective— including bhigh riskQ—percutaneous coronary interventions without on-site cardiac surgery. Am Heart J 2004;148:676 - 83. 7. Smith SC, Feldman TE, Hirshfield JW, et al. ACC/AHA/SCAI 2005 Update for Percutaneous Coronary Intervention—Summary Article. A Report of the ACC/AHA Task Force on Practice Guidelines Writing Committee to update the 2001 Guidelines for Percutaneous Coronary Intervention. Circulation 2006;113:1 - 20. 8. Seshadri N, Whitlow PL, Acharya N, et al. Emergency coronary artery bypass surgery in the contemporary percutaneous coronary intervention era. Circulation 2002;106:2346 - 50. 9. Yang EH, Gumina MD, Lennon RJ, et al. Emergency coronary artery bypass surgery for percutaneous coronary interventions: changes in the incidence, clinical characteristics, and indications from 1979 to 2003. J Am Coll Cardiol 2005;46:2004 - 9. 10. Lotfi M, Mackie K, Dzavik V, et al. Impact of delays to cardiac surgery after failed angioplasty and stenting. J Am Coll Cardiol 2004;43:337 - 42. 11. Dudek D, Rakowski T, Legutko J, et al. Efficacy and safety of percutaneous coronary interventions in patients with non ST segment elevation acute coronary syndrome in catheterization laboratory without ON-SITE surgical back-UP. Kardiol Pol 2003;58:356 - 60. 12. Jollis JG, Peterson ED, Nelson CL, et al. Relationship between physician and hospital coronary angioplasty volume and outcome in elderly patients. Circulation 1997;95:2485 - 91.
Background Percutaneous coronary intervention (PCI) performed in centers without onsite cardiac surgery remains controversial. Advances in PCI techniques and medical therapy have markedly decreased postprocedural complications. Our aim was to assess the efficacy and safety of performing PCI in the Veterans Affairs patient population in a hospital without onsite cardiac surgery. Methods We prospectively evaluated 401 consecutive patients who underwent elective PCI or PCI after admission for acute coronary syndrome. Patients who had ST-elevation myocardial infarction or were hemodynamically unstable were classified as emergent and had their PCI performed elsewhere and were therefore excluded from our analysis. Our cardiac surgery backup was a community hospital 8 miles away. Results The patient’s mean age was 65.6 F 10 years, and most were men (99.5%). Patients had high-risk clinical and angiographic profiles, with diabetes mellitus in 44%, prior myocardial infarction in 41%, comorbid conditions in 45% and type B or C angiographic lesions in 83%. Of 401 patients, 338 (84%) received a stent, and 86 (21%), a drug eluting stent. Percutaneous coronary intervention success rate was 97%. There were no periprocedural or inhospital deaths, and no patients required emergency transfer for cardiac surgery. At 1 and 6 months of follow-up, total mortality was 1.5% and 3.5%, respectively; target vessel revascularization rate was 0% and 1.7%. Conclusions
Nonemergent PCI can be performed effectively and safely in patients with higher clinical and angiographic risk without onsite backup cardiac surgery. This has significant implications for most hospitals that have an invasive but not an interventional program. (Am Heart J 2006;152:9092 13.)
The number of percutaneous coronary interventions (PCIs) performed annually in the United States and worldwide continues to grow. Parallel to this has been the improvement in technology, which allows for increasing safety and success of PCIs. This adds new fuel to the ongoing debate whether PCIs should be performed in centers without onsite cardiac surgical capability. Many such hospitals are providing care to the patients living in rural areas of the country who have limited access to medical services. If safety and outcomes of nonemergent PCIs performed without onsite surgical backup were shown to be comparable with ones done in centers with onsite cardiac surgery,
From the Salem Veterans Administration Medical Center, Salem, VA. Submitted October 31, 2005; accepted May 22, 2006. Reprint requests: Nabil Jarmukli, MD, FACC, Cardiology, Salem VAMC, Cardiology111B, Salem, VA 24153. E-mail: [email protected] 0002-8703/$ - see front matter Published by Mosby, Inc. doi:10.1016/j.ahj.2006.05.016
a large number of smaller hospitals with catheterization laboratories could improve patient care and remain competitive. To date, no randomized clinical trials to address this issue have been performed. Most of the data accumulated so far come from single institution reports and retrospective registry reviews. Although the recent comparative analysis of Medicare files from a large number of hospitals1 suggested a higher risk of adverse outcomes in centers without onsite cardiac surgery (for nonprimary/rescue PCI at hospitals performing b50 Medicare PCIs per year), another recent report by Foster et al2 found that on-site surgical backup was not a predictor of lower inhospital mortality after PCI in a heterogeneous group of ST-elevation myocardial infarction (STEMI), non–ST elevation myocardial infarction (NSTEMI), and non–myocardial infarction (MI) patients derived from the American College of Cardiology National Cardiovascular Data Registry (ACC-NCDR) registry. Several individual center reports3-6 seem to indicate an excellent safety and efficacy profile. A report by Turgeman et al5 from a single site in Israel with no
American Heart Journal November 2006
910 Roussanov et al
Table I. Clinical characteristics of patients who underwent elective PCI and PCI in the setting of ACS
Table II. Coronary lesion types in patients who underwent elective PCI and PCI in the setting of ACS Coronary lesion type (%)
No. of patients (%) Characteristic
Elective PCI ACS with PCI All patients
Age (y) (mean F SD) N65 years Hypertension Diabetes mellitus Cerebrovascular disease Peripheral vascular disease Chronic renal insufficiency Severe pulmonary disease Prior coronary artery disease Prior MI Prior CABG Prior PCI Prior congestive heart failure Left ventricular ejection fraction (%) (mean F SD)
65.7 F 9.7 114 (53) 188 (87) 97 (45) 27 (12)
65.5 F 10.4 93 (50) 162 (88) 78 (42) 19 (10)
65.6 F 10 207 (52) 350 (87) 175 (44) 46 (11)
60 (28)
36 (19)
96 (24)
39 (18)
44 (24)
83 (21)
11 (5)
8 (4)
19 (5)
164 (76)
121 (65)
285 (71)
91 65 81 41
(42) (30) (37) (19)
48 F 11
75 50 64 29
(40) (27) (35) (16)
45 F 13
166 (41) 115 (29) 145 (36) 70 (17) 47 F 12
onsite backup surgery showed that referral for emergency surgery was 1.3% among 235 patients who had balloon angioplasty and 0.6% among 1016 patients who had coronary stenting. Similarly, another single center report by Zavala-Alarcon et al6 from Phoenix, AZ, showed that the referral for emergency cardiac surgery was 0.7% among 1000 patients who had elective coronary intervention with a periprocedural death of 0.2% and MI of 2.1%. A potential criticism of the singlecenter reports is the bias to limit PCIs to selected group of very low-risk patients. Yet, these articles appropriately emphasize the importance of careful case selection at hospitals with off-site surgical backup. We had an opportunity to prospectively follow up a large higher-risk cohort of patients typical of a Veterans Affairs hospital. This allowed us to take advantage of minimal outside referral and thorough observation practices and obtain a reliable data on immediate and long-term safety and outcomes of PCIs performed without onsite cardiac surgery.
Methods Hospital and PCI program Salem Veterans Affairs Medical Center (VAMC) is a level I, 182-bed acute care medical/surgical/psychiatric teaching hospital affiliated with the University of Virginia, Charlottesville. It is located in southwest Virginia and is quite distant from several VAMCs, which serve as receiving centers for coronary intervention candidates: 150 miles from Durham VAMC, NC,
Type A (low risk) Type B (moderate risk) Type C (high risk)
Elective PCI
ACS with PCI
All PCIs
14 61 25
20 58 22
17 60 23
169 miles from Richmond VAMC, VA, and 228 miles from Asheville VAMC, NC. Salem VAMC has had an invasive program for N20 years but no interventional program. The cardiology section at Salem VAMC is equipped with 2 cardiac catheterization laboratories (Phillips Poly C and Integris Allura, Shelton, CT), both with digital capabilities. In March of 2002, Salem VAMC obtained formal approval from the Veterans Affairs Central Office to initiate a PCI program. Approval was restricted to elective PCIs and patients with acute coronary syndrome (ACS), including NSTEMI. Approval was not given for patients with STEMI or hemodynamically unstable patients because they were considered emergent. In July of 2002, the first patient underwent PCI. All procedures were performed by 1 of 3 coronary interventional board-certified cardiologists with an experience of at least 300 diagnostic cardiac catheterizations and 85 coronary interventional cases per operator per year in a cardiac catheterization laboratory (CCL) with at least 600 diagnostic cases and 170 interventional cases per year. This was preceded by formal training of CCL and cardiac recovery nurses and technicians in preoperative, intraoperative, and postoperative care. Our surgical backup is a private hospital with an established cardiac surgery program located 8 miles from Salem VAMC. Before the start of every case, a point person from Salem VAMC CCL would contact the surgical team at the receiving hospital to get approval for the time of the procedure. A fully equipped advanced cardiac life support–capable ambulance was on standby at Salem VAMC emergency room door with its personnel stationed in a room next to the CCL for the duration of the case. Before the initiation of the program, and biyearly thereafter, a bdry runQ to ensure PCI failure safe transport procedure was performed with door-to-door ambulance transport time range of 10 to 15 minutes.
Patients and procedures All consecutive patients who were referred for nonemergent PCI between July 2002 and November 2004 were included in the study. Although primary PCI for STEMI was not performed in our institution, both patients who were scheduled for PCI electively (from cardiology and primary care clinic referrals as well as referrals from the nuclear and treadmill stress laboratories), and those admitted to the hospital with ACS including NSTEMI were included. Percutaneous coronary interventions on patients who were hemodynamically unstable or had active symptoms at the time of procedure were not performed at Salem VAMC and were directly referred to one of the local private hospitals.
American Heart Journal Volume 152, Number 5
Roussanov et al 911
Table III. Clinical outcomes of patients who underwent elective PCI and PCI in the setting of ACS No. of patients (%) Elective PCI Characteristic Total mortality Noncardiac death Myocardial infarction ACS PCI Coronary bypass surgery Cardiac hospitalizations Target vessel revascularization
1m 3 2 0 1 0 0 3 0
(1.4) (0.9) (0.5)
(1.4)
ACS with PCI 6m
6 (2.8) 2 (0.9) 0 7 (3.3) 4 (1.5) 1 (0.5) 12 (5.6) 4 (1.5)
1m 3 2 0 0 0 1 1 0
(1.6) (1.1)
(0.5) (0.5)
All patients
6m 8 3 1 7 6 2 11 3
(4.3) (1.6) (0.5) (3.8) (3.2) (1.1) (5.9) (1.6)
1m 6 4 0 1 0 1 4 0
(1.5) (1) (0.2) (0.2) (1)
6m 14 5 1 14 10 3 23 7
(3.5) (1.2) (0.2) (3.5) (2.5) (0.7) (5.7) (1.7)
Standard techniques were used for PCI, and a variety of intracoronary stents (bare metal and drug eluting) were deployed when appropriate. Low- to high-risk coronary lesions in patients with low- to high-risk clinical profiles were treated, provided that it could be safely and effectively performed using available equipment and techniques. Rare patients requiring specialized intracoronary procedures (laser, brachytherapy, directional coronary atherectomy, or rotablator) were referred to outside centers and excluded from this study. Standard medical therapy was used pre, intra, and post procedure. Patients were then followed up prospectively with regular clinical visits within the Veterans Affairs system. Baseline clinical information, including demographics, comorbid conditions, mode of presentation, lesion complexity, and outcome data, including procedural complications, mortality, hospitalizations, and repeat revascularization, were analyzed and presented below as percentages or mean F 1SD. The definitions of periprocedural PCI complications and lesion types were used as presented in ACC–National Cardiovascular Data Registry (http://www.acc.org/ncdr/cathlab.htm). All other definitions used were also based on the same registry.
0.8 stents per lesion. Drug-eluting stents (all Taxus) were used in 86 (21%) patients. The proportion of coronary lesion types is presented in Table II. Eightythree percent of all lesions were higher-risk (types B and C). Clopidogrel was given as 300 mg loading and 75 mg daily maintenance dose to all patients undergoing stent placement. All patients received aspirin 81 to 324 mg unless contraindicated. h-Blockers, statins, and angiotensin-converting enzyme inhibitors were used in 77%, 78%, and 61% of patients, respectively, as of the procedure day. Platelet IIb/IIIa inhibitors were administered during the procedure in 75 (19%) patients. Vascular access closure device was deployed in 40% of cases. During this period, 31 patients were excluded from the study: 21 with STEMI who underwent reperfusion with primary PCI off site or fibrinolytic treatment on site, 7 because of refractory to medical therapy ACS requiring emergent intervention, and 11 due to complex anatomy requiring atherectomy, rotablator, or other specialized procedures.
Results
Clinical outcomes Seven patients (1.7%) of the total group had periprocedural complications. One patient developed transient renal failure after elective PCI. The remaining 6 complications (3%) occurred in patients with ACS. No patients required transfer for urgent surgical treatment. No periprocedural/inhospital deaths or cerebrovascular accidents were recorded. One patient had NSTEMI because of no-reflow after stenting venous bypass graft with mild enzyme elevation and no clinical sequelae. The incidence of major vascular complication, arrhythmias, and renal failure were 0.75%, 0.5%, and 0.5%, respectively. One patient developed cardiogenic shock in the setting of coronary ischemia because of occluded venous bypass graft, recurrent ventricular tachycardia, and ventricular fibrillation. Patient quickly recovered after PCI and placement of intra-aortic balloon pump.
Patients and procedures A total of 401 patients underwent PCI between July 2002 and November 2004. The procedure was successful in 97% of cases. Success was defined as b20% residual stenosis associated with TIMI III flow. Patients were almost exclusively men (99.5%). Ages ranged from 39 to 90 years. Two hundred seven (52%) were elderly (N65). Demographic and clinical characteristics are given in Table I. Of the patients, 25% and 22% had multivessel and multistage procedures, respectively. Two hundred sixteen (54%) patients had PCI as an elective procedure. One hundred eighty-five (46%) were admitted to the hospital with ACS (including NSTEMI) and assigned to early invasive strategy (within 24-48 hours). A total of 608 coronary lesions were treated with PCI. Stents were used in 338 patients (84%) with a ratio of
912 Roussanov et al
Follow-up data were obtained in all patients because of the fully computerized medical record of the Veterans Affairs system in place for N6 years now and computer connectivity with the Veterans Affairs satellite clinic network. Data on 1- and 1-month clinical outcome are presented in Table III. Total mortality was 1.5% and 3.5% at 1 and 6 months, respectively. Myocardial infarction occurred in only 1 patient, whereas ACS occurred in 3.4% of all patients at 6 months of follow-up. Requirement for PCI at 6 months was 2.5%, and for coronary artery bypass graft (CABG) surgery, 0.7%. There were no documented cerebrovascular accidents on follow-up.
Discussion There is continued controversy regarding the performance of PCI in hospitals without onsite backup cardiac surgery. The most recent 2005 American College of Cardiology/American Heart Association guidelines restrict class I recommendation to centers with onsite cardiac surgical capability. As for centers without onsite cardiac surgery, guidelines label primary PCI for acute STEMI as class IIb (acceptable) and elective PCI as class III (not acceptable).7 Yet the guidelines clearly state that several centers have reported satisfactory results based on careful case selection with well-defined arrangements for immediate transfer to a surgical program. However, it is quite evident from the literature that over the past 5 years, continued advances in coronary stenting techniques as well as aggressive periprocedural antiplatelet therapy and enhanced operator experience has led to significant reduction in periprocedural complications requiring emergency bypass surgery. A published report by the Cleveland Clinic foundation on their interventional program showed a significant decline (N10-fold) in the prevalence of emergency CABG from 1.5% in 1992 to 0.14% in 2000 ( P = .0010).8 Ting et al3 reported the results of low-risk PCI on type A and B1 lesions in low-risk patients performed in a community hospital without onsite cardiac surgery with a backup hospital 85 miles away. Procedural success was achieved in 99.5% of patients with 0.5% inhospital deaths. Yang et al9 reported the changing trend at Mayo Clinic for emergency cardiac surgery referral post coronary intervention over a period of 24 years. Although during the prestent era the incidence of their emergency surgical referral was 2.9%, it dropped to 0.7% during the early-stent era and even lower to 0.3% during the later-stent era. This drop in emergency surgical referral was despite the increasing trend of performing PCI on a higher-risk population who were older and had a higher prevalence of diabetes mellitus, lower left ventricular systolic function, higher incidence of prior revascularization, and more complex coronary anatomy.
American Heart Journal November 2006
Lotfi et al10 reported on the Canadian experience from the University Health Network in Toronto and noted that although b1 in 100 patients needed emergency cardiac surgery post PCI, 1 in 4 patients referred for emergency cardiac surgery would be placed at increased risk of harm if delays to surgery were encountered. Consequently, we initiated our PCI program after obtaining formal approval from the Veterans Affairs central office to test the feasibility, efficacy, and safety of performing PCI without onsite cardiac surgical capability in the Veterans Affairs hospital system. Our results clearly showed that PCI was successful in 97% of patients enrolled with an excellent safety profile because there were no deaths or need for urgent cardiac surgery. The low rate of MI, arrhythmic, vascular, and renal complications are comparable with rates in patients undergoing PCI with onsite cardiac surgery backup1. Our results of 1-month 1.5% mortality and 0.2% CABG requirement compare favorably to the 2.8% 30-day mortality and 1.1% CABG requirement in 16 184 Medicare patients with nonprimary/nonrescue PCI performed at centers with surgical backup. It is noteworthy that our study included patients with higher clinical and angiographic risk profiles, with ACS in 46%, age N65 years in 52%, diabetes mellitus in 44%, comorbid conditions (cerebrovascular, peripheral vascular, renal, pulmonary disease) in 45%, prior MI in 41%, and type B and C angiographic lesions in 83%. Our results are very similar to the Polish experience where Dudek et al11 reported on the efficacy and safety of performing nonemergent PCI on NSTEMI or patients with stable angina without onsite surgical backup. In a cohort of 479 patients (160 with NSTEMI and 319 with stable angina), the inhospital outcome was equally favorable in both groups with MI in 2.5% versus 1.6% and urgent referral for surgery in 0.6% versus 0.3%. At 36 months of follow-up, the incidence of MI was 0.7% and death, 2.0% in both groups. Several reports have elucidated an inverse relationship between operator procedural volume, as well as hospital CCL volume and periprocedural complication rates. The Duke Research Group examined Medicare patient registry and showed improving outcomes seen at a threshold of 75 cases per operator and 200 cases per hospital.12 Current American College of Cardiology/ American Heart Association guidelines for PCI7 recommend a minimum operator volume of 75 interventions per year (class I) at high volume center with N400 yearly onsite elective interventions. Our operator and hospital interventional volume per year were at least 85 and 170 cases respectively. Our results have significant impact on a large number of Veterans Affairs and community hospitals that have an active cardiac catheterization laboratory but no interventional program. By meeting certain safety standards, such hospitals can serve their patients more effectively
American Heart Journal Volume 152, Number 5
and expediently, in lieu of transferring patients for intervention after cardiac catheterization. It should be clear that our results do not apply to the higher-risk patients such as acute STEMI or hemodynamically unstable patients. The principal limitation of this study is its cohort and not a randomized controlled design. A control group from the receiving hospital could not be used for comparison because the receiving hospital cares for a nonveteran population with different demographics and risks. In addition, the fact that patients with STEMI or unstable hemodynamics were not enrolled could have skewed the data to a better outcome. Furthermore, serial troponin or MB fraction of creatine phosphokinase were not routinely measured post intervention, and therefore, the number of asymptomatic MIs is probably underestimated. Finally, the delay in transfer after a failed or complicated angioplasty is critical to the outcome, and we had no patients who required emergency transfer. We conclude that if rigorous benchmarks of operator experience, safety standards, patient screening, proper communication, intense hospital and personnel commitment and adherence to high program standards are met, then nonemergent PCI can be performed effectively and safely in a hospital without onsite cardiac surgery, even in patients with ACS and type B and C angiographic lesions. Prospective randomized clinical trial is needed to answer the bigger question of whether outcomes of nonemergent PCI are equivalent at hospitals with and without onsite cardiac surgery. Our data demonstrate that it is safe to launch such a trial. We thank Dr Nelson Bernardo and Dr Milan Kothari for initiating our PCI program and performing most the interventions. We also thank the following for their participation in the care of those patients as well as gathering some of the data and assisting in the analysis: Kathy Henley, NP, Brenda Dogan, RN, Brenda Hart, RN, and Jeannie Wilson, RN.
Roussanov et al 913
References 1. Wennberg DE, Lucas FL, Siewers AE, et al. Outcomes of percutaneous coronary interventions performed at centers without and with onsite coronary bypass graft surgery. JAMA 2004;292: 1961 - 8. 2. Foster JK, Klein LW, Veledar E, et al. On-site Surgical Backup as a Predictor of Mortality Among PCI Patients with STEMI, NSTEMI, or NO AMI. Circulation 2005;112:II-737. 3. Ting HH, Garratt KN, Singh M, et al. Low-risk percutaneous coronary interventions without on-site cardiac surgery: two years’ observational experience and follow-up. Am Heart J 2003;145: 278 - 84. 4. Paraschos A, Callwood D, Wightman MB, et al. Outcomes following elective percutaneous coronary intervention without on-site surgical backup in a community hospital. AJC 2005;95:1091 - 3. 5. Turgeman Y, Atar S, Suleiman K, et al. Diagnostic and therapeutic percutaneous cardiac interventions without on-site surgical backup—review of 11 years experience. Isr Med J Assoc 2003; 5:89 - 93. 6. Zavala-Alacron E, Cecena F, Ashar R, et al. Safety of elective— including bhigh riskQ—percutaneous coronary interventions without on-site cardiac surgery. Am Heart J 2004;148:676 - 83. 7. Smith SC, Feldman TE, Hirshfield JW, et al. ACC/AHA/SCAI 2005 Update for Percutaneous Coronary Intervention—Summary Article. A Report of the ACC/AHA Task Force on Practice Guidelines Writing Committee to update the 2001 Guidelines for Percutaneous Coronary Intervention. Circulation 2006;113:1 - 20. 8. Seshadri N, Whitlow PL, Acharya N, et al. Emergency coronary artery bypass surgery in the contemporary percutaneous coronary intervention era. Circulation 2002;106:2346 - 50. 9. Yang EH, Gumina MD, Lennon RJ, et al. Emergency coronary artery bypass surgery for percutaneous coronary interventions: changes in the incidence, clinical characteristics, and indications from 1979 to 2003. J Am Coll Cardiol 2005;46:2004 - 9. 10. Lotfi M, Mackie K, Dzavik V, et al. Impact of delays to cardiac surgery after failed angioplasty and stenting. J Am Coll Cardiol 2004;43:337 - 42. 11. Dudek D, Rakowski T, Legutko J, et al. Efficacy and safety of percutaneous coronary interventions in patients with non ST segment elevation acute coronary syndrome in catheterization laboratory without ON-SITE surgical back-UP. Kardiol Pol 2003;58:356 - 60. 12. Jollis JG, Peterson ED, Nelson CL, et al. Relationship between physician and hospital coronary angioplasty volume and outcome in elderly patients. Circulation 1997;95:2485 - 91.