This study sought to examine the safety of percutaneous coronary intervention (PCI) before and during de novo establishment of a transradial (TR) program at a teaching hospital. TR access remains underused in the United States, where cardiology fellowship programs continue to produce cardiologists with little TR experience. Establishment of TR programs at teaching hospitals may affect PCI safety. Starting in July 2009 a TR program was established at a teaching hospital. PCI-related data for academic years 2008 to 2009 (Y1) and 2009 to 2010 (Y2) were prospectively collected and retrospectively analyzed. Of 1,366 PCIs performed over 2 years, 0.1% in Y1 and 28.7% in Y2 were performed by TR access. No major complications were identified in 194 consecutive patients undergoing TR PCI, and combined bleeding and vascular complication rates were lower in Y2 versus Y1 (0.7% vs 2.0%, p = 0.05). Patients treated in Y2 versus Y1 and by TR versus transfemoral approach required slightly more fluoroscopy but similar contrast volumes and had similar procedural durations, lengths of stay, and predischarge mortality rates. PCI success rates were 97% in Y1, 97% in Y2, and 98% in TR cases. TR PCIs were performed by 13 cardiology fellows and 9 attending physicians, none of whom routinely performed TR PCI previously. In conclusion, de novo establishment of a TR program improved PCI safety at a teaching hospital. TR programs are likely to improve PCI safety at other teaching hospitals and should be established in all cardiology fellowship training programs.
Since the first description of transradial (TR) coronary intervention in 1993, several advantages of the TR approach have been described. Decreased bleeding and vascular complications, decreased procedural cost, shorter length of stay, decreased nursing needs, and earlier ambulation with improved patient comfort and satisfaction have driven an increase in its global popularity. Despite these advantages the TR approach has been used in only 1.3% of all percutaneous coronary interventions (PCIs) performed in the United States from 2004 through 2007. Reasons for the reluctance of many United States cardiologists to adopt the TR approach are not well established but likely contributors include familiarity with the transfemoral (TF) approach and concern for longer procedure times and increased radiation exposure. The substantial learning curve for TR procedures is another deterrent, 1 that is perhaps best overcome by formal TR training during fellowship. Although such training programs should ultimately improve PCI safety, the TR learning curve may decrease PCI safety as training programs are established. Effects of de novo establishment of TR training programs on PCI safety at teaching hospitals are unknown. The purpose of this study was to assess the impact of a newly established TR training program on PCI safety at a previously femoral-only teaching hospital.
Methods
In July 2009 a physician-initiated programmatic transition toward routine TR coronary arteriography and PCI was initiated at the Medical University of South Carolina, where cardiology fellows are the primary operators for almost all cardiac catheterizations and are assisted by attending cardiologists. None of the involved physicians had any significant previous experience with TR procedures. Formal TR training was not part of the transition; rather, physicians’ learning was self-directed. Staff was actively involved in the establishment of the program including the management of its effect on laboratory workflow and patient preparation and recovery.
Attending cardiologists were encouraged at the same time (beginning of the TR study period) to start performing TR procedures, but there was no formal stepwise transition to TR access. The TR approach was recommended as the default approach except in patients with inadequate Allen tests and/or Barbeau grade D perfusion by plethysmography; the TF approach remained the default approach for these patients. Use of the TR approach was especially encouraged in patients who were thought to be at high risk of bleeding and vascular complications. However, the TR approach was not mandated: the final decision to use a TR or TF approach was made by the attending cardiologist.
Decisions on periprocedural treatment with antithrombotic therapy were made by the attending cardiologist. An intravenous bolus of unfractionated heparin 3,000 to 5,000 U was usually given at the initiation of TR procedures. Bivalirudin or additional heparin was administered if PCI was performed. The radial sheath was flushed at the time of all catheter exchanges with nicardipine solution or a “radial cocktail” including verapamil, nitroglycerin, and lidocaine. A TR Band (Terumo Medical Corporation, Somerset, New Jersey) was applied at the end of every TR procedure. Femoral arteriotomy closure devices were used at the discretion of the attending cardiologist. Dual antiplatelet therapy with aspirin and a thienopyridine was prescribed after nearly every PCI.
Data including baseline patient characteristics, procedural characteristics, and procedural outcomes were prospectively collected and retrospectively analyzed using the American College of Cardiology/National Cardiovascular Data Registry Cath Lab Module 3.04 (for procedures completed from July 15, 2008 through June 30, 2009 [Y1]) and CathPCI Registry 4.3 (for procedures completed from July 1, 2009 through June 30, 2010 [Y2]). No cases were excluded from analysis. The primary outcome was the composite of bleeding and vascular complications. Secondary outcomes included components of the primary outcome and procedural success, length of stay, and predischarge mortality. Procedural characteristics of interest included arterial access site(s), procedural duration, maximum sheath size, closure device use, fluoroscopy time, contrast volume, number of lesions treated, and treatment with anticoagulants and inhibitors of glycoprotein IIb/IIIa. Baseline patient characteristics of interest included age, gender, weight, prothrombin time, and platelet count; levels of hemoglobin, serum urea nitrogen, and creatinine; presence or absence of a history of diabetes, hypertension, or dyslipidemia; and indications for PCI.
Procedural success was defined by the passage of any interventional device across a target lesion. In cases where PCI was attempted on >1 lesion the procedure as a whole was considered successful if ≥1 lesion was crossed successfully. In TR cases conversion to a TF approach was counted as procedural failure. Predischarge mortality was defined as death before discharge from the hospitalization with which the PCI was associated. In American College of Cardiology/National Cardiovascular Data Registry Cath Lab Module 3.04 and CathPCI Registry 4.3, bleeding from the access site or retroperitoneal, gastrointestinal, genitourinary, or other sources constituted a complication when associated with a hematocrit decrease ≥10% or a hemoglobin decrease ≥3 g/dl, transfusion of whole blood or packed red blood cells, or need for procedural intervention at the bleeding site. Acute anemia with hemoglobin decrease ≥3 g/dl attributable to procedure-related blood loss without an obvious alternative source was considered bleeding at the access site. In American College of Cardiology/National Cardiovascular Data Registry Cath Lab Module 3.04, vascular complications were defined as any access site arterial occlusion, peripheral embolization, dissection, pseudoaneurysm, or arteriovenous fistula. For American College of Cardiology/National Cardiovascular Data Registry CathPCI Registry 4.3, the definition of vascular complications was clarified to include only those complications that required a procedural intervention.
Continuous variables were summarized as means and dichotomous variables were summarized in absolute numbers as percentages. Generalized mixed models were used to evaluate for differences in procedural outcomes, procedural characteristics, and baseline patient characteristics in Y2 versus Y1 and with TR versus TF PCI. These models account for correlated observations. For comparisons with a 0 cell Fisher’s exact test was used. Fisher’s exact test does not take into account correlated observations; however, models with a 0 cell did not converge. Significance was set at a 2-sided alpha level of 0.05 and all analyses were completed using SAS 9.2 (SAS Institute, Cary, North Carolina).
Results
In total 1,366 PCIs were performed in 1,249 patients at the Medical University of South Carolina from July 15, 2008 through June 30, 2010. Six hundred ninety-three PCIs were performed in Y1 and 673 in Y2. TR access was used in 1 PCI in Y1 and 193 PCIs (28.7%) in Y2. TR PCIs in Y2 were performed by 13 different cardiology fellows and 9 different attending cardiologists, none of whom routinely performed TR PCI previously.
Baseline patient characteristics and indications for PCI are presented in Table 1 . Patients treated in Y2 versus Y1 were more likely to have a history of diabetes or hypertension, and patients undergoing TR versus TF PCI were heavier and had minimally lower serum urea nitrogen and creatinine levels. Patients treated by the TR versus TF approach presented less commonly with ST-segment elevation myocardial infarction. Patients treated in Y2 versus Y1 and by a TR versus TF approach presented less commonly with non–ST-segment elevation myocardial infarction and more commonly with angina.
| Characteristic | Y2 | Y1 | p Value ⁎ | TR | TF | p Value ⁎ |
|---|---|---|---|---|---|---|
| (n = 673) | (n = 693) | (n = 194) | (n = 1,172) | |||
| Baseline characteristics | ||||||
| Age (years) | 63.2 ± 11.9 | 63.6 ± 11.9 | 0.4 | 62.0 ± 11.0 | 63.6 ± 12.0 | 0.1 |
| Weight (kg) | 88.1 ± 20.7 | 86.6 ± 19.6 | 0.2 | 91.1 ± 23.6 | 86.7 ± 19.5 | 0.01 |
| Men | 434 (64%) | 447 (65%) | >0.9 | 119 (61%) | 762 (65%) | 0.3 |
| Diabetes mellitus | 284 (43%) | 254 (37%) | 0.04 | 69 (36%) | 469 (40%) | 0.3 |
| Hypertension | 619 (92%) | 581 (84%) | <0.001 | 176 (91%) | 1,024 (87%) | 0.2 |
| Dyslipidemia | 587 (88%) | 602 (87%) | 0.6 | 173 (91%) | 1,016 (87%) | 0.2 |
| Hemoglobin (g/dl) | 12.8 ± 2.0 | 13.0 ± 1.9 | 0.2 | 13.2 ± 1.9 | 12.9 ± 2.0 | 0.1 |
| Platelet count (thousands/mm 3 ) | 214 ± 67.4 | 218 ± 63.7 | 0.4 | 216 ± 65.0 | 216 ± 65.7 | >0.9 |
| Serum urea nitrogen (mg/dl) | 18.8 ± 13.3 | 17.7 ± 10.9 | 0.1 | 16.5 ± 10.4 | 18.6 ± 12.4 | 0.05 |
| Creatinine (mg/dl) | 1.3 ± 1.4 | 1.3 ± 1.4 | 0.4 | 1.1 ± 0.5 | 1.4 ± 1.5 | 0.01 |
| Prothrombin time (s) | 14.8 ± 2.3 | 14.6 ± 2.0 | 0.2 | 14.5 ± 1.8 | 14.7 ± 2.2 | 0.4 |
| Indications for percutaneous coronary intervention | ||||||
| ST-segment elevation myocardial infarction | 72 (11%) | 58 (8%) | 0.1 | 8 (4%) | 122 (10%) | 0.008 |
| Non–ST-segment elevation myocardial infarction | 113 (17%) | 149 (22%) | 0.03 | 25 (13%) | 237 (20%) | 0.02 |
| Unstable angina pectoris | 234 (35%) | 209 (30%) | 0.07 | 75 (39%) | 368 (32%) | 0.05 |
| Stable angina pectoris | 161 (24%) | 103 (15%) | <0.001 | 65 (34%) | 199 (17%) | <0.001 |
| Atypical chest pain | 20 (3%) | 19 (3%) | 0.8 | 7 (4%) | 32 (3%) | 0.5 |
| Other † | 73 (11%) | 149 (22%) | <0.001 | 14 (7%) | 208 (18%) | <0.001 |
⁎ The p values consider the fact that some observations were made on the same subjects.
† Other indications for percutaneous coronary intervention included exertional dyspnea, cardiac allograft vasculopathy, spontaneous or traumatic coronary artery dissection, and ventricular tachycardia/fibrillation.
Procedural characteristics are listed in Tables 2 and 3 . Nearly all TR procedures were completed through the right radial artery. Patients with left internal mammary arterial bypass grafts or acute ST-segment elevation myocardial infarction were more likely to be treated from a TF approach. Procedures completed in Y2 versus Y1 and by the TR versus TF approach were similar in total duration, required slightly more fluoroscopy time, and required similar volumes of iodinated contrast. Unfractionated heparin was used in >80% of all procedures. Glycoprotein IIb/IIIa inhibitors were used less commonly and bivalirudin was used more commonly for procedures completed in Y2 versus Y1 and by the TR versus TF approach. Enoxaparin and fondaparinux were used infrequently. Sheath sizes >6Fr were used only for TF PCIs and femoral arteriotomy closure devices were used less frequently in Y2 versus Y1.
| Characteristic | Y2 | Y1 | p Value ⁎ | TR | TF | p Value ⁎ |
|---|---|---|---|---|---|---|
| (n = 673) | (n = 693) | (n = 194) | (n = 1,172) | |||
| Duration (min) | 76.7 ± 39.7 | 72.8 ± 44.6 | 0.1 | 75.8 ± 43.3 | 74.5 ± 42.1 | 0.7 |
| Lesions treated | 1.4 ± 0.7 | 1.4 ± 0.7 | 0.2 | 1.4 ± 0.7 | 1.4 ± 0.7 | >0.9 |
| Contrast volume (ml) | 156 ± 68.3 | 150 ± 68.4 | 0.1 | 159 ± 66.8 | 151 ± 68.6 | 0.2 |
| Fluoroscopy (min) | 18.6 ± 12.9 | 17.2 ± 12.8 | 0.05 | 20.4 ± 12.9 | 17.5 ± 12.8 | 0.01 |
| Largest Sheath Size (Fr) | 6.3 ± 0.7 | 6.4 ± 0.8 | 0.1 | 6.0 ± 0.4 | 6.4 ± 0.8 | <0.001 |
| Treatment with glycoprotein IIb/IIIa inhibitors | ||||||
| Any IIb/IIIa inhibitor | 16 (2%) | 107 (15%) | <0.001 | 1 (0.5%) | 122 (10%) | 0.002 |
| Abciximab | 13 (2%) | 38 (5%) | 0.001 | 1 (0.5%) | 50 (4%) | 0.03 |
| Eptifibatide | 3 (0.5%) | 69 (10%) | <0.001 | 0 (0%) | 72 (6%) | <0.001 † |
| Treatment with anticoagulants | ||||||
| Bivalirudin | 189 (28.1%) | 140 (20.2%) | <0.001 | 61 (31.4%) | 268 (22.9%) | 0.01 |
| Enoxaparin | 2 (0.3%) | 29 (4.2%) | <0.001 | 2 (1.0%) | 29 (2.5%) | 0.2 |
| Fondaparinux | 12 (1.8%) | 1 (0.1%) | 0.02 | 1 (0.5%) | 12 (1.0%) | 0.5 |
| Unfractionated heparin | 588 (87.4%) | 616 (88.9%) | 0.4 | 175 (90.2%) | 1,029 (87.8%) | 0.3 |
⁎ The p values consider the fact that some observations were made in the same subjects.
† For comparisons with a 0 cell, Fisher’s exact tests were used.
| Closure Method | Y2 | Y1 | p Value ⁎ |
|---|---|---|---|
| (n = 481) | (n = 691) | ||
| Manual pressure | 277 (58%) | 221 (32%) | <0.001 |
| Perclose ProGlide (Abbott Vascular, Abbott Park, Illinois) | 151 (31%) | 334 (48%) | <0.001 |
| Angioseal (St. Jude Medical, St. Paul, Minnesota) | 7 (1%) | 15 (2%) | 0.4 |
| Mynx (Access Closure, Mountain View, California) | 46 (10%) | 121 (18%) | <0.001 |
| Any closure device | 204 (42%) | 470 (68%) | <0.001 |
⁎ The p values consider the fact that some observations were made in the same subjects.
Outcomes are presented in Table 4 . Major bleeding and vascular complications over the 2-year study period were confined to TF procedures. No bleeding or vascular complications were identified in 194 consecutive TR cases and the combined incidence of bleeding and vascular complications was significantly lower in Y2 versus Y1 (0.7% vs 2.0%, p = 0.05; Figure 1 and Table 4 ). There were no cases of symptomatic radial arterial occlusion that required postprocedure imaging or intervention. Several cases of radial arterial spasm occurred periprocedurally but all were resolved with vasodilators and/or additional sedative medication. Combined bleeding and vascular complication rates were similar with the TF approach in Y2 versus Y1 (2.0% vs 1.0%, p = 0.2). Predischarge mortality was statistically similar in Y2 versus Y1 and with the TR versus TF approach. PCI was successfully accomplished in 97% of patients in Y1, in 97% of patients in Y2, and in 98% of all TR cases. Thus, procedural success rates were similar in Y2 versus Y1 and with the TR versus TF approach. Of 194 TR PCIs 3 (1.5%) were converted to the TF approach and were counted as failures for TR PCI. Mean length of stay after TR versus TF PCI was shorter (2.5 vs 3.1 days) but the difference was not statistically significant.
