Author + information
- Received February 1, 2008
- Revision received May 5, 2008
- Accepted May 18, 2008
- Published online August 1, 2008.
- Sunil V. Rao, MD, FACC⁎,⁎ (, )
- Fang-Shu Ou, MS⁎,
- Tracy Y. Wang, MD, MS⁎,
- Matthew T. Roe, MD, MHS, FACC⁎,
- Ralph Brindis, MD, MPH, FACC†,
- John S. Rumsfeld, MD, PhD, FACC‡ and
- Eric D. Peterson, MD, MPH, FACC⁎
- ↵⁎Reprint requests and correspondence:
Dr. Sunil V. Rao, Durham VA Medical Center, 508 Fulton Street (111A), Durham, North Carolina 27705.
Objectives Our goal was to compare trends in the prevalence and outcomes of the radial and femoral approaches to percutaneous coronary intervention (PCI) in contemporary clinical practice.
Background There are few current data on the use and outcomes of the radial approach to PCI (r-PCI) in clinical practice.
Methods Data from 593,094 procedures in the National Cardiovascular Data Registry (606 sites; 2004 to 2007) were analyzed to evaluate trends in use and outcomes of r-PCI. Logistic regression was used to evaluate the adjusted association between r-PCI and procedural success, bleeding complications, and vascular complications. Outcomes in elderly patients, women, and patients with acute coronary syndrome were specifically examined.
Results Although the proportion of r-PCI procedures has recently increased, it only accounts for 1.32% of total procedures (n = 7,804). Compared with the femoral approach, the use of r-PCI was associated with a similar rate of procedural success (adjusted odds ratio: 1.02 [95% confidence interval: 0.93 to 1.12]) but a significantly lower risk for bleeding complications (odds ratio: 0.42 [95% confidence interval: 0.31 to 0.56]) after multivariable adjustment. The reduction in bleeding complications was more pronounced among patients <75 years old, women, and patients undergoing PCI for acute coronary syndrome.
Conclusions The use of r-PCI is rare in contemporary clinical practice, but it is associated with a rate of procedural success similar to the femoral approach and with lower rates of bleeding and vascular complications, even among high-risk groups. These results suggest that wider adoption of r-PCI in clinical practice may improve the safety of PCI.
Percutaneous coronary intervention (PCI) can be performed via the femoral, brachial, or radial arteries. The femoral approach traditionally has been the primary approach for many operators, but is associated with a bleeding or vascular complication rate of up to 10% in some series (1). Given the decline in post-PCI ischemic events over time (2), reduction in bleeding risk has taken on new significance (3,4). A prior study using data from a large registry found that access site hematomas large enough to require transfusion are associated with an increased risk of adverse events, including mortality (5). In this context, randomized trials comparing the femoral approach to percutaneous coronary intervention (f-PCI) with the radial approach (r-PCI) have shown that the rate of vascular and bleeding complications is significantly lower with the radial approach (6–8). Despite these data, it is unclear how often the radial approach is used in clinical practice and whether the benefits seen in clinical trials translate to the wider population of patients undergoing PCI. Accordingly, we used data from the National Cardiovascular Data Registry (NCDR) to examine the prevalence of r-PCI and to compare procedural success and in-hospital complications between r-PCI and f-PCI. As prior studies have shown that elderly patients, women, and patients with acute coronary syndrome (ACS) are at higher risk for bleeding complications (9,10), we specifically explored trends in the use of r-PCI over time and corresponding outcomes among patients age <75 years versus ≥75 years, female versus male patients, and patients with stable angina versus those with ACS.
The NCDR, which is cosponsored by the American College of Cardiology (ACC) and the Society for Cardiovascular Angiography and Interventions, has been described previously (11). The NCDR catalogs clinical data and outcomes in PCI procedures that are gathered from over 600 sites across the U.S. Data are entered into NCDR-certified databases at participating institutions and exported in a standard format to a common database at Heart House (Washington, DC); only institutions whose submissions meet quality criteria for data reporting are included. The definitions of all variables are prospectively defined by a committee of the NCDR. An auditing program ensures the validity of collected data.
For the purpose of this analysis, we included the first PCI procedure performed in any individual patient during a qualifying hospitalization between January 2004 and March 2007. The dataset comprised 725,549 procedures from 637 hospitals. From this, we excluded nonindex PCIs (n = 18,745 procedures); any PCI involving an access site other than the femoral or radial artery (n = 6,991 procedures); any emergency or salvage procedures, defined as procedures performed for ongoing myocardial ischemia or infarction, pulmonary edema requiring intubation, or shock; or if cardiopulmonary resuscitation was being performed on a patient en route to the catheterization laboratory (n = 106,407); and procedures from any hospitals performing fewer than 30 PCIs during the study period due to inability to obtain stable estimates of the proportion of r-PCI procedures at these institutions (n = 312 procedures).
Definitions and end points
Vascular access site (radial or femoral) is defined in the NCDR as the site of successful vascular entry; failed attempts are not captured. The primary outcomes examined were the incidence of procedural success (defined as residual stenosis ≤50% with Thrombolysis In Myocardial Infarction [TIMI] flow grade ≥2, and ≥20% decrease in stenosis severity in all lesions attempted), bleeding complications (defined as access site bleeding, retroperitoneal bleeding, gastrointestinal bleeding, genitourinary bleeding, or other bleeding), and vascular complications (defined as access site occlusion, peripheral embolization, arterial dissection, arterial pseudoaneurysm, or arteriovenous fistula). All bleeding end points in the NCDR are further defined as requiring transfusion and/or prolonging the hospital stay, and/or causing a drop in hemoglobin >3.0 g/dl. Hematomas >10 cm for femoral access or >2 cm for radial access also qualify as access site bleeding. Access site occlusion is defined in the database as total obstruction of the artery, typically by thrombus (but may have other causes), usually at the site of access requiring surgical repair. Occlusions may be accompanied by absence of palpable or Doppler pulse. Peripheral embolization is defined as a loss of distal pulse, pain, and/or discoloration of the extremities (especially the toes). Dissection is defined as a disruption of an arterial wall resulting in splitting and separation of the intimal layers; pseudoaneurysm is defined as the occurrence of a disruption and dilation of the arterial wall without identification of the arterial wall layers at the site of the catheter entry demonstrated by arteriography or ultrasound. Arteriovenous fistula is defined as a connection between the access artery and the accompanying vein that is demonstrated by arteriography or ultrasound.
The prevalence of r-PCI was calculated for the overall population as well as for each hospital. Hospitals then were grouped by their percentage of r-PCI procedures. For descriptive analyses, we compared baseline characteristics, treatment profiles, procedure use, and clinical outcomes between r-PCI and f-PCI. Continuous variables are presented as medians with interquartile percentiles; categorical variables are expressed as percentages. To test for independence of a patient's baseline characteristics, in-hospital care patterns, and outcomes with respect to receiving r-PCI, Mann-Whitney Wilcoxon nonparametric tests were used for continuous variables and Pearson chi-square tests were used for categorical variables.
In order to determine trends in the use of r-PCI over time, the study period was divided into quarters and the rates of r-PCI were calculated for each quarter. Similarly, the rates of r-PCI usage were also calculated in different patient subgroups to demonstrate the differences in trend over time. Subgroups considered were age <75 years versus ≥75 years, women versus men, and different PCI indications (stable angina, non–ST-segment elevation acute coronary syndrome [NSTE ACS], and ST-segment elevation myocardial infarction [STEMI]).
The unadjusted rates of the primary outcomes between r-PCI and f-PCI were calculated in the overall population as well as in the subgroups of age <75 versus ≥75 years, women versus men, and stable angina versus NSTE ACS versus STEMI. In examining the association between r-PCI and outcomes, a multivariable logistic regression was used to estimate the marginal effects of r-PCI. The generalized estimating equations method (12) was used to account for within-hospital clustering, as patients at the same hospital are more likely to have similar responses relative to patients in other hospitals (i.e., within-center correlation for response). This method produces estimates similar to those from ordinary logistic regression, but the estimated variances of the estimates are adjusted for the correlation of outcomes within each hospital. Due to the low number of bleeding events and unsuccessful PCI procedures in the r-PCI group, we avoided overfitting the logistic regression models by including ACC-NCDR mortality risk score as a covariate in the models for procedure success and bleeding complications (13). The ACC-NCDR risk score consists of cardiogenic shock, age, salvage/urgent/emergent PCI, pre-procedure intra-aortic balloon pump insertion, left ventricular ejection fraction, presentation with acute myocardial infarction, diabetes mellitus, renal failure, chronic lung disease, thrombolytic therapy, use of nonstent devices, and lesion characteristics including the left main artery, proximal left anterior descending disease, and Society for Cardiac Angiography and Interventions lesion classification (c-index 0.89). Additional covariates included in the model for procedure success are ACC/American Heart Association lesion risk, bifurcation disease, chronic total occlusion, and pre-procedure TIMI flow. Additional variables included in the model for bleeding outcome are gender, body mass index, glycoprotein IIb/IIIa inhibitor use, unfractionated heparin use, direct thrombin inhibitor use, history of congestive heart failure, and peripheral vascular disease. The effects of patient age (<75 vs. ≥75 years), patient gender, and PCI indication on the relationship between r-PCI and the outcomes were assessed by including interaction terms between arterial entry location (radial or femoral) and the groups of interest in the models adjusted for NCDR risk score. We repeated the analysis after excluding centers that did not perform any r-PCI procedures during the study period. A p value <0.05 was considered significant for all tests. All statistical analyses were performed by the Duke Clinical Research Institute using SAS software (version 9.0, SAS Institute, Cary, North Carolina).
Of the 725,549 procedures entered into the NCDR during the study period, 593,094 PCI procedures remained after applying the exclusion criteria. Of these, 7,804 (1.32%) procedures were performed via the radial artery approach. Figure 1 displays the prevalence of r-PCI across hospitals; the vast majority of centers performed <10% of PCI procedures via the radial artery approach. However, there were 7 centers in the database that performed ≥40% of PCI procedures via the radial approach.
Table 1 displays the baseline characteristics of r-PCI versus f-PCI procedures. r-PCI procedures were performed in slightly younger patients and in patients with significantly higher body mass index, and with a higher prevalence of peripheral vascular disease compared with f-PCI. There was a significantly lower prevalence of prior coronary artery bypass graft surgery, prior renal failure, and NSTE ACS or STEMI among r-PCI patients. In terms of procedure characteristics, r-PCI procedures had longer fluoroscopy times, but there was no significant difference between r-PCI and f-PCI in terms of total volume of contrast used (median [25th, 75th percentiles] r-PCI 200 cc [140, 250] vs. f-PCI 200 cc [140, 260]). Unfractionated heparin was more commonly used for r-PCI procedures, while bivalirudin and glycoprotein IIb/IIIa inhibitors were more commonly used for f-PCI procedures. r-PCI procedures were performed more commonly at university hospitals.
Trends in r-PCI over time
Figures 2A to 2D display the prevalence of r-PCI over time in the overall dataset as well as in the key subgroups of age, gender, and PCI indication. The use of r-PCI remained stable until the first quarter of 2007 when the proportion increased, although the sample size was relatively smaller during the first quarter of 2007. This trend was also present among the subgroups defined by age, gender, and PCI indication; however, the use of r-PCI in patients age ≥75 years, women, and patients with ACS (both NSTE ACS and STEMI) was lower than among patients <75 years, men, and patients with stable angina.
Figure 3 displays the unadjusted rates of the primary outcomes between r-PCI and f-PCI. Importantly, there were only 15 r-PCI patients who developed a vascular complication (0.19%). Table 2 shows the association between r-PCI and the primary outcomes after multivariable adjustment. There was no significant association between the arterial approach used and procedural success; however, there were significantly lower adjusted odds for the occurrence of bleeding complications with r-PCI. As stated in the preceding text, there were too few vascular complications among r-PCI patients to perform multivariable adjustment.
We repeated the analysis after excluding procedures reported by centers that did not perform any r-PCI. After excluding these centers, 343,467 procedures remained, of which 7,804 (2.27%) were performed via radial artery approach. Baseline patient and procedure characteristics of r-PCI and f-PCI in this subset were similar to those seen in the overall study population; in addition, the adjusted outcomes were nearly identical to those seen in the overall population (data not shown).
Outcomes among key subgroups
Figures 4A to 4C display the incidence of bleeding and vascular complications with r-PCI and f-PCI in the key subgroups. The incidence of both complications was lower for r-PCI among all patient subgroups examined. Notably, there were no reported bleeding or vascular complications among patients with STEMI treated with r-PCI.
The interaction terms of age, gender, and PCI indication were significant in the adjusted analysis of bleeding, such that the protective effect of r-PCI on bleeding complications was more pronounced among patients age <75 years, women, and patients with NSTE ACS (Table 3). The interaction terms were not significant in the adjusted analysis of procedural success, demonstrating that r-PCI and f-PCI had similar associations with procedural success across the subgroups examined.
This analysis of a national contemporary multicenter PCI registry demonstrates several findings as they relate to the radial artery approach to PCI in clinical practice. First, although the proportion of PCI procedures performed via the radial approach increased during the last quarter of the time period studied, it was very uncommon overall, with most centers performing <10% of cases via the radial artery. Second, r-PCI is used less frequently among patients at risk for PCI-related complications, such as the elderly patients, women, and patients with ACS. Third, r-PCI is associated with a significantly lower risk for bleeding and vascular complications without sacrificing procedural success or involving more use of contrast agents. Finally, when r-PCI was used in patients >75 years, women, and patients with ACS, the rate of bleeding and vascular complications was lower compared with that in f-PCI. These data support the efficacy and safety of r-PCI in clinical practice and suggest that wider application of the radial approach may enhance the safety of coronary intervention.
Randomized trials have previously shown an advantage of r-PCI over f-PCI with respect to access site complications (including bleeding), early ambulation, and costs (6–8). Mann et al. (6) randomly assigned 142 patients with ACS to PCI via the radial or femoral artery approach and observed similar rates of PCI success with both approaches, but noted significantly fewer access site complications with the radial approach, which was associated with a 15% decrease in hospital costs by reducing length of stay. Similarly, Kiemeneij et al. (8) randomly assigned 900 patients undergoing PCI to either the femoral, brachial, or radial artery approach and found that the incidence of vascular complications was significantly lower among patients assigned to the radial approach; in fact, there were no complications in the radial artery group, compared with 2.3% in the brachial artery group and 2.0% in the femoral artery group (p = 0.035). Agostoni et al. (14) performed a meta-analysis of 12 studies, both randomized and observational, and found that the radial approach was associated with a 71% to 85% decrease in the odds of entry site complications. The aforementioned studies by Mann et al. (6) and Kiemeneij et al. (8) were conducted at a single center, and the studies included in the Agostoni et al. (14) meta-analysis were also either single center or included fewer than 7 sites. Our results confirm the outcomes seen in these prior studies and extend them to contemporary clinical practice by utilizing a database of over 700,000 procedures from more than 600 hospitals. This reduction in PCI-related complications may ultimately confer significant survival advantages in patients who are at high risk for vascular complications after PCI (15,16).
Despite the potential safety advantage of r-PCI, our study shows that the radial approach is very infrequently used in clinical practice among the sites represented in the NCDR; in particular, it is less frequently used among elderly patients, women, and patients with ACS. Potential reasons for this include the learning curve associated with this technique, unwillingness to adopt a new approach on the part of interventionalists, concerns over longer fluoroscopy times, or a lack of a concerted effort on the part of industry to encourage the radial approach through marketing of devices specifically designed for that application. Goldberg et al. (17) evaluated procedural success in an initial series of 27 patients undergoing PCI via the radial artery for an operator inexperienced with r-PCI. They found that the rate of successful PCI was 84%, with the most significant limiting factor being spasm of the radial artery, which occurred in 30% of cases. Current combinations of spasmolytic drugs injected into the radial artery during or immediately after sheath insertion virtually eliminate arterial spasm and facilitate procedural success (18,19). While we could not measure the number of failed attempts at radial arterial access, our data do show that once radial access is obtained, the rate of procedural success is statistically similar to that seen with the femoral approach.
Some limitations to our study should be considered. First, the present analysis is not a randomized trial, and, as such, unmeasured confounders could be present. However, we adjusted for a wide array of clinical and procedural variables and accounted for site clustering effects in our analysis. In addition, only a proportion of the collected data are audited; therefore, as in any large registry, there is a small potential for inaccurate data collection. Second, as mentioned in the preceding text, the ACC-NCDR only has information on successful arterial access and does not capture unsuccessful attempts at vascular access. Therefore, the r-PCI procedures in our analysis represent cases in which radial artery access was obtained successfully. Moreover, only the first PCI in each patient was considered, and reaccess of the radial artery was not examined. Despite these limitations, our results suggest that once vascular access is obtained in the radial artery, the rate of procedural success is high and the rate of bleeding or vascular complications is low. Third, we examined r-PCI volumes by site, not by operator; thus, high-volume r-PCI operators could account for our findings. Finally, although the definition of vascular complications is quite broad in the NCDR, it may still underestimate the rate of these complications among r-PCI patients, because occlusion of the radial artery may occur with a palpable radial artery pulse due to collateral circulation in the hand. In addition, the NCDR does collect data with granularity sufficient to examine bleeding complications according to various bleeding definitions.
Our examination of a large contemporary multicenter PCI registry demonstrates that there is marked variation in the use of r-PCI. Overall, it is infrequently used in clinical practice, but it is associated with a rate of procedural success similar to the femoral approach and with lower rates of bleeding and vascular complications. These findings were present even among patients at high risk for PCI-related complications such as elderly patients, women, and patients with ACS. These data, in the context of prior clinical trials, suggest that wider adoption of r-PCI in clinical practice may improve the safety of PCI.
This analysis was funded by the National Cardiovascular Data Registry.
- Abbreviations and Acronyms
- American College of Cardiology
- acute coronary syndrome
- femoral approach to percutaneous coronary intervention
- National Cardiovascular Data Registry
- NSTE ACS
- non–ST-segment elevation acute coronary syndrome
- percutaneous coronary intervention
- radial approach to percutaneous coronary intervention
- ST-segment elevation myocardial infarction
- Received February 1, 2008.
- Revision received May 5, 2008.
- Accepted May 18, 2008.
- American College of Cardiology Foundation
- Singh M.,
- Rihal C.S.,
- Gersh B.J.,
- et al.
- Yatskar L.,
- Selzer F.,
- Feit F.,
- et al.
- Mann T.,
- Cubeddu G.,
- Bowen J.,
- et al.
- Kiemeneij F.,
- Laarman G.J.,
- Odekerken D.,
- Slagboom T.,
- van der Wieken R.
- Alexander K.P.,
- Chen A.Y.,
- Newby L.K.,
- et al.
- Manoukian S.V.,
- Feit F.,
- Mehran R.,
- et al.
- Brindis R.G.,
- Fitzgerald S.,
- Anderson H.V.,
- Shaw R.E.,
- Weintraub W.S.,
- Williams J.F.
- Liang K.Y.,
- Zeger S.L.
- Shaw R.E.,
- Anderson H.V.,
- Brindis R.G.,
- et al.
- Agostoni P.,
- Biondi-Zoccai G.G.,
- de Benedictis M.L.,
- et al.
- Batchelor W.B.,
- Anstrom K.J.,
- Muhlbaier L.H.,
- et al.