Author + information
- Received November 26, 2001
- Revision received March 20, 2002
- Accepted April 5, 2002
- Published online July 3, 2002.
- Robert J Applegate, MD, FACC*,* (, )
- Mark A Grabarczyk, MD*,
- William C Little, MD, FACC*,
- Timothy Craven, MSPH†,
- Michael Walkup, MS†,
- Frederic R Kahl, MD, FACC*,
- Gregory A Braden, MD, FACC‡,
- Kevin M Rankin, MD, FACC§ and
- Michael A Kutcher, MD, FACC*
- ↵*Reprint requests and correspondence:
Dr. Robert J. Applegate, Section of Cardiology, Wake Forest University School of Medicine, Medical Center Boulevard, Winston-Salem, North Carolina 27157-1045USA.
Objectives The study assessed clinical outcomes of closure device use following percutaneous coronary revascularization using current standards of anticoagulation and antiplatelet therapy.
Background Evaluation of the outcomes of patients by use of vascular closure devices during coronary interventions employing current standards of anticoagulation and glycoprotein (GP) IIb/IIIa inhibitor therapy is limited.
Methods We evaluated outcomes of 4,525 consecutive patients who underwent percutaneous coronary intervention between July 1997 and April 2000. All patients received anticoagulation with heparin and GP IIb/IIIa inhibitor therapy with abciximab. The closure method was manual in 1,824 patients, Angioseal in 524 patients and Perclose in 2,177 patients. Procedural and hospital vascular outcomes were evaluated.
Results Closure device success was 97.1% Angioseal and 94.1% Perclose (p < 0.05). Minor vascular complications occurred in 1.8% of manual patients, 1.1% of Angioseal patients and 1.2% of Perclose patients (p = NS); major complications occurred in 1.3% of manual patients, 1.1% of Angioseal patients and 1.0% of Perclose patients (p = NS). Multivariate logistic regression identified only closure device failure as an independent predictor of a vascular complication. In patients with successful closure with a device, minor complications (0.8% vs. 1.8%, p < 0.05) and any complication (1.5% vs. 2.5%, p < 0.05) were reduced compared to manual compression.
Conclusions Arterial closure following coronary interventions using anticoagulation and GP IIb/IIIa inhibitor therapy can be safely and effectively performed, with vascular complication rates similar to or lower than with manual pressure. Additionally, vascular complication rates using GP IIb/IIIa inhibitor therapy regardless of the method of arterial closure are equivalent to or lower than previously published rates of vascular complications.
Femoral artery hemostasis can be achieved following sheath removal after the completion of percutaneous coronary interventions (PCI) using manual pressure or recently developed arterial closure devices (1–4). Two general types of closure devices are available: suture-based devices (4–6)and those that utilize a plug (7–9). Early studies examining the safety and efficacy of these devices demonstrated effectiveness in obtaining hemostasis in selected patients (4,8,10). A recent report of a large clinical experience, however, indicates that there is a higher rate of vascular complications using closure devices (11)following coronary interventions. Whether this is attributable to intrinsic limitations of the closure devices themselves, or a result of a greater degree of anticoagulation used during the coronary intervention, is unknown. Use of potent antiplatelet therapy with reduced anticoagulation is associated with lower bleeding complications than standard-dose heparin during PCI (12,13). Because only 6% of the patients in the previous study (11)received glycoprotein (GP) IIb/IIIa inhibitor therapy, it is possible that use of closure devices in patients treated with combined antiplatelet and reduced anticoagulant therapy may not be associated with increased vascular complications, or that they may even be beneficial. As such potent platelet inhibitor has become the standard of care in patients undergoing PCI, it is important that the relative efficacy and safety of manual and closure device hemostasis be determined in patients undergoing PCI receiving GP IIb/IIIa inhibitors.
Patients at our institution undergoing percutaneous revascularization using the femoral artery approach were included in the study. Patients were not excluded based on any clinical or procedural characteristics. A total of 3,643 patients undergoing 4,525 PCI procedures using the femoral approach between July 1997 and April 2000 form the basis for this study. The study was approved by the Institutional Review Board.
Percutaneous revascularization procedure
Patients underwent percutaneous revascularization procedure using standard techniques. Arterial sheath size ranged from 6 to 10F. Anticoagulation after sheath insertion was accomplished using intravenous (IV) unfractionated heparin to achieve a target activated clotting time (ACT) of 200 to 250 s. Additional heparin was administered approximately every 30 min to maintain the ACT in this range. Each patient in the study received GP IIb/IIIa receptor inhibition with abciximab according to usual protocol: 0.25 mg/kg bolus followed by 0.125 mg/kg each hour IV infusion for 12 h. All patients were treated with aspirin 281 to 325 mg a day. Patients receiving intracoronary stents received Ticlid 250 mg b.i.d., or Plavix 300 mg as a loading dose followed by 75 mg/day.
The method of arterial closure was chosen by the interventionist. In patients in whom arterial closure was contemplated, a femoral arteriogram was performed via the arterial sheath. Patients generally did not undergo arterial closure if: 1) the arteriotomy site was at or below the femoral bifurcation; 2) the common femoral artery was <5 mm in diameter; and 3) extensive calcification or plaque formation was present in the common femoral artery. The type of arteriotomy closure was left to the interventionist. Closure was performed using first-generation closure devices including either a collagen-based suture closure device (Angioseal, Daig, Minnetonka, Minnesota) or a suture-based closure device (ProStar, TecStar; Perclose, Redwood City, California). Placement of either closure device was performed according to standard technique. In the case of the Angioseal device a second bailout or “buddy” wire was placed in the last 399 patients to ensure vascular access in the event of device failure and was removed after successful Angioseal deployment. The femoral region around the arteriotomy site was infiltrated with lidocaine with 1% epinephrine. Arterial closure was performed in the cardiac catheterization laboratory using the arterial closure device at the completion of the study. In patients in whom arterial closure was not performed in the laboratory, the sheath was pulled when the ACT was ≤180 s. Hemostasis out of the laboratory was achieved by either manual pressure alone, or in conjunction with a c-clamp. A pneumatic compression device, the Femostop (USCI, Billerica, Massachussetts), was used when hemostasis was not adequate despite manual pressure, or after an arterial closure device was placed and hemostasis was not adequately achieved. Ambulation was initiated in general 2 h after an arterial closure device was placed, and 6 to 8 h following manual compression. A groin check was routinely made postprocedure and prior to discharge, and recorded in the chart.
Prior to hospital discharge the patient’s chart was evaluated by a clinical research nurse. The hospital outcomes were entered into an American College of Cardiology compatible database (CAOS, IBS, Winston-Salem, North Carolina), and these formed the basis for the outcomes in this study. Procedural outcomes were reviewed at the completion of each intervention procedure and entered into the database as well.
Descriptive statistics (means and SD of continuous factors, frequency counts and relative frequencies of categorical factors) were computed by closure type (manual, Perclose or Angioseal) and overall. Univariate associations between treatment type and various complications or patient outcomes were examined using two-way contingency tables and, where appropriate, associated odds ratios (ORs) with 95% confidence intervals. These tables were constructed using all available data (i.e., some patients contributed more than one observation to the analysis). Significance of associations between patient characteristics and types of closure procedure were assessed using either the Fisher exact test for two-way tables or, in the instances when within-subject correlation was nonignorable (e.g., patient age, history of hypertension and other fixed preprocedure characteristics), generalized estimating equations (GEEs). To estimate associations between closure type and complication outcomes in a setting that adjusted for possible within-subject correlation, as well as important concomitant risk factors, GEE multivariate logistic regression models were employed. The GEE models use the technique of Liang and Zeger (14)to control for within-subject correlation while allowing for inclusion of both continuous and categorical factors as covariates. All statistical analyses were performed using SAS statistical analysis software (15).
Of the 4,525 patients in the study, 524 patients underwent arterial closure with the Angioseal device, 2,177 patients underwent closure with the Perclose device and the remaining 1,824 patients underwent closure by manual pressure. Table 1displays the clinical characteristics of the patients in each of the study groups. Not unexpectedly, there were more patients with peripheral vascular disease in the manual pressure group than in either of the arterial closure groups, although this represented a minority of patients in the manual pressure group itself (10%). The traditional risk factors for coronary artery disease are also shown in Table 1, and these were commonly present in patients receiving closure devices. Diabetes was present in approximately 20% to 25% of patients in all of the closure groups. The procedural characteristics are shown in Table 2. Use of 10F arterial sheaths was more commonly seen in the manual pressure and Perclose groups (16% and 20%, respectively, p < 0.05 vs. Angioseal), whereas 9F sheaths were more commonly observed in the Angioseal and Perclose groups (37% and 44% vs. 22% manual, p < 0.05). The type of coronary intervention, number of vessels treated and procedural success are also shown in Table 2. Number of vessels treated and lesion characteristics were similar among the three groups. However, atherectomy procedures were more frequently performed in patients closed with the Perclose and Angioseal devices.
The outcome of the arterial closure is shown in Table 3. Failure to achieve adequate hemostasis or need for Femostop closure device as part of the hemostasis process occurred in 15 of the 524 Angioseal patients (2.9%), in 64 of 1,824 manual pressure patients (3.5%) and 128 of 2,177 (5.9%) Perclose patients (p < 0.05 vs. Angioseal and manual). Vascular complications associated with arterial hemostasis are also shown in Table 3, while the OR and 95% confidence limits for outcomes are shown in Table 4. Vascular repair was required in 15 of 1,824 (0.8%) manual pressure patients, 1 of 530 (0.2%) Angioseal patients and 12 of 2,177 Perclose patients (0.6%) (p = NS). Significant bleeding was defined as blood loss resulting in any of the following: ≥3 g/dl drop in hemoglobin, blood transfusion, prolonged hospital stay, hematoma ≥10 cm or retroperitoneal hematoma. Significant bleeding was equally distributed among all three groups and ranged from 1.1% to 1.7% (p = NS among groups) Occlusion or loss of pulse occurred in 8 of 1,824 (0.4%) of manual-pressure patients, 1 of 524 (0.2%) Angioseal patients and 3 of 2,177 (0.1%) Perclose patients (p = NS). Overall mortality was lower in the Angioseal (0.2%) and Perclose (1.2%) groups than the manual group (3.2%) (p < 0.05). However, deaths directly attributed to vascular complications occurred in two of the manual-pressure group, 0 of the Angioseal and 2 of the Perclose group (p = NS). The outcomes of any closure device use are shown in Table 5. Complication rates tended to be lower with closure device use, but the differences were not statistically significant.
Multivariate logistic regression (Table 6) was performed using the clinical and procedural characteristics identified in Tables 1 and 2. Only in laboratory closure device, failure was identified as independently predictive of any minor or major complication. Thus, although baseline differences existed with respect to clinical and procedural characteristics among the groups, they did not seem to have affected the outcomes. We then re-examined the outcome data, excluding those patients who experienced failure of the closure device in the catheter laboratory (Table 7). The incidence of any minor complication excluding these closure-device failures was 1.8% manual and 0.8% combined closure device (p < 0.05); of any major complication 1.3% versus 0.9% (p = NS); and of any minor or major complication 2.5% versus 1.5% (p < 0.05). Thus, successful closure appeared to be associated with reduced overall events.
The major finding of this study is that arterial closure can be performed safely and effectively with currently available vascular closure devices in patients treated with anticoagulation and potent antiplatelet GP IIb/IIIa inhibitor therapy during coronary intervention procedures. Excluding those patients with closure-device failure, our data suggest that successful arterial closure may be associated with decreased vascular complication rates compared to manual compression in patients receiving anticoagulant and GP IIb/IIIa inhibition during PCI. Finally, vascular complication rates of both the manual and closure-device groups are equivalent to or lower than previously reported vascular complication rates in patients undergoing coronary revascularization using heparin as the sole method of anticoagulation during the procedure (4,11,16–18). These data suggest that use of the arterial-closure devices, Angioseal and Perclose, should be considered when early ambulation or continued anticoagulation is indicated. Whether routine use of closure devices will reduce the complication rate associated with sheath removal remains to be determined.
Comparison to prior studies
Our findings differ from a recently published large observational study suggesting that closure devices are associated with increased vascular complication rates following coronary interventions (11). There may be several reasons for these apparent differences. First, only 6% of their patients received GP IIb/IIIa inhibitor therapy, with ACTs averaging 277 s (closure group) and 268 s (manual group). In our study ACTs were generally in the 200 to 225 s range, substantially lower than in the study of Dangas et al. (11). Thus, a greater degree of anticoagulation may have contributed to the higher complication rates seen in their study with closure-device use. Second, patients chosen for arterial closure may have been different. Only 516 of 5,609 patients (9%) in the Dangas et al. (11)study underwent arterial closure, whereas 2,701 of 4,525 patients (60%) underwent closure in our study. Thus, higher complication rates may occur in patients chosen because they are perceived to be “higher risk” patients than if routine closure is the practice, as was the case in our laboratory. Third, we routinely performed femoral arteriograms prior to closure device use, which may have prevented higher-risk patients from receiving closure devices, and which may not have been a practice in previous studies. Finally, the strategies of “ad hoc” closure for higher-risk patients versus planned closure for all patients (the standard at our institution) may have had an effect on outcomes based on operator experience. The extent of training and experience in vascular closure devices has been identified in a recent editorial as one factor that may influence outcomes associated with closure-device use (19). In our laboratory the closure-device technique is considered part of the intervention procedure, and thus both extensive training and experience were provided for each operator using these devices. Whether this significantly influenced the results remains to be determined.
Access site bleeding and GP IIb/IIIa inhibitor therapy
Vascular complications associated with arterial access have long been recognized to occur during cardiac catheterization and percutaneous intervention procedures (6,16–18,20). Factors associated with increased risk of complications have been identified including the presence of vascular disease, concomitant anticoagulation, older age, and at least in some studies, multiple procedures (16–18,20). Studies have also shown the level of anticoagulation with unfractionated heparin had a substantial bearing on the bleeding rates (13); with modification of the anticoagulation regimen, vascular complications at the access site were significantly reduced. An interesting observation from this study is that major vascular complication rates were 0.9% for the combined closure-device groups, and 1.3% for the manual-pressure group, including a large number of patients with 10F sheaths and complex coronary procedures. Major vascular complication rates have been reported to range from 1.0% (2)to over 3% (12)in intervention series, most of whom did not receive GP IIb/IIIa inhibitor therapy. The reason for this apparent low rate of major vascular complications is not clear, but it does indicate that routine abciximab use itself is not associated with vascular complications, as has been recently suggested (21).
Limitations of closure devices
Although vascular devices theoretically should minimize or eliminate complications associated with hemostasis of the vascular access site, existing data on closure-device outcomes in the aggregate suggest that this may not be the case. Several reasons may explain this, including incomplete closure, previous sticks or multiple sticks at sites other than the closure site, and a small but real incidence of spontaneous bleeding secondary to the use of anticoagulant and antiplatelet therapy itself. Our observation that successful arterial closure was associated with reduced vascular complication rates is intriguing, and this suggests that development of more effective closure devices should be pursued. Whether future generations of arterial closure devices can substantially reduce or completely eliminate the complications associated with arterial access site hemostasis remains to be determined.
Several limitations of the study merit further discussion. This was not a randomized study; thus, it may have inherent selection biases, which might have influenced the results (19). The decision not to perform arterial closure may have been made because of the presence of vascular disease; thus, patients treated by manual pressure may represent a higher-risk group of patients. However, manual closure of the arterial access site in these patients was done when the effect of anticoagulation was much lower than in the closure-device patients, which might have offset any potential increased risk introduced by the presence of vascular disease. Second, vascular complications experienced outside the hospital were not formally monitored. Anecdotal experience in recent published series indicate that late infections may occur in approximately 0% to 3% of patients undergoing closure with the Perclose device (2,5,6,10,21). This study did not allow us to formally measure this complication rate in our patient cohort. Although certain biases may have been introduced, with respect to the outcomes of the closure devices, the results reflect the use of interventionists who felt most comfortable with the technique and, thus, likely the most optimal outcomes from their use. Moreover, this experience represents “real world” experience and, thus, is directly applicable to clinical practice.
In conclusion, arterial closure by Angioseal or Perclose closure devices can be performed safely and effectively following coronary interventions using anticoagulation and GP IIb/IIIa inhibitor therapy, with vascular complication rates similar to or lower than with manual pressure. Moreover, vascular complication rates following GP IIb/IIIa inhibitor use appear low regardless of the method of closure.
We gratefully acknowledge the efforts of Aruna Joel, Nancy Salstrom and Robin Taylor for database acquisition and management; Sharon Alcom, Wendy Hobbs and Teresa Young for intervention research support; and Tammy Davis for expert secretarial service.
☆ Supported in part by Wake Forest University School of Medicine (Winston-Salem, North Carolina) Interventional Research Program and an unrestricted grant from Daig Corporation (Minnetonka, Minnesota).
- activated clotting time
- generalized estimating equation
- odds ratio
- percutaneous coronary intervention
- Received November 26, 2001.
- Revision received March 20, 2002.
- Accepted April 5, 2002.
- American College of Cardiology Foundation
- Sanborn T.A.,
- Gibbs H.H.,
- Brinker J.A.,
- Knopf W.D.,
- Kosinski E.J.,
- Roubin G.S.
- Brachmann J.,
- Ansah M.,
- Kosinski E.J.,
- Schuler G.C.
- Dangas G.,
- Mehran R.,
- Kokolis S.,
- et al.
- Blankenship J.C.,
- Hellkamp A.S.,
- Aguirre F.V.,
- Demko S.L.,
- Topol E.J.,
- Califf R.M.
- Liang K.Y.,
- Zeger S.L.
- ↵(1999) SAS/STAT User’s Guide Version (SAS Institute Inc, Cary, NC), 8 ed.
- Popma J.J.,
- Satler L.F.,
- Pichard A.D.,
- et al.
- Omoigui N.A.,
- Califf R.M.,
- Pieper K.,
- et al.
- Tavris D.,
- Gross T.,
- Gallauresi B.
- Duffin D.C.,
- Muhlestein J.B.,
- Allison S.B.,
- et al.