Different Spectrum of Vascular Complications After Angio-Seal Deployment or Manual Compression

Abstract: Background. Reported complication rates after vascular closure device deployment or femoral manual compression (MC) are similar. However, the features and severity of such complications have never been thoroughly evaluated. Methods and Results. A consecutive series of 1241 patients treated from 2008 to 2010 with Angio-Seal (AS) was prospectively evaluated for vascular complications (VC). As control group, we used a consecutive series of 672 patients treated with MC in the 7 months preceding AS adoption at our institution. VC were observed in 88 patients, 55 with AS and 33 with MC (relative risk, 0.90; 95% confidence interval, 0.59-1.38; P=.63). The clinical profile of complications observed in the 2 groups was different. Groin hematomas were more frequent with MC (100% vs 65.5%; P=.0005) and retroperitoneal bleedings were more common with AS (41.8% vs 6.1%; P=.0005). AS complications required more frequently transfusions (49.1% vs 18.2%; P=.006), while MC complications significantly delayed hospital discharge, in comparison to AS (4.3 ± 4.0 days vs 2.7 ± 1.9 days; P=.01). Differences in groin hematoma and retroperitoneal bleeding rates were confirmed after propensity score matching. Finally, a different allocation of diagnostic/therapeutic resources was observed in the 2 groups. Conclusion. AS and MC were associated with similar incidences of VC, with a higher prevalence of severe complications (retroperitoneal hemorrhages and transfusions) after using AS. However, complications after MC were associated with significantly prolonged hospital stay. Comparison between different hemostatic strategies should consider the logistic burden imposed by different vascular complications.

J INVASIVE CARDIOL 2012;24:90–96

Key words: vascular closure device, cardiac catheterization, groin hematoma, vascular complications

____________________________________________

Vascular closure devices (VCD) have been created to achieve rapid hemostasis and early mobilization^1-3 in patients undergoing transfemoral cardiac catheterization, even when treated with potent anticoagulant and antiplatelet drugs.^4-7 However, their use is not free from possible life-threatening complications,^8,9 thus justifying the position expressed by the American Heart Association that places VCDs in Class IIa when employed to achieve faster hemostasis and improve patient comfort, but in Class III when used with the purpose to reduce vascular complications.¹⁰ More recent information from the National Cardiovascular Data Registry suggested that VCDs could be associated with significantly lower bleeding rates, particularly among patients at greatest risk for bleeding,¹¹ justifying the position of Dauermann et al stating that VCDs are a key factor to improve bleeding complications in patients treated by cardiac catheterization and coronary interventions.¹²

Previous randomized trials and meta-analyses^13-26 reported similar incidences of vascular complications after VCD deployment and femoral manual compression (MC) hemostasis, so that the decision to use such devices is often left to operator discretion. However, the spectrum of such complications is highly heterogeneous, ranging from disturbing groin hematoma to severe retroperitoneal bleeding and limb-threatening ischemia. Thus, the knowledge of the rate and severity of vascular complications seems to be crucial in the choice of hemostasis strategy. Therefore, in the present study, we investigated whether VCD deployment or MC were associated with different types of vascular complications. Moreover, we evaluated if such differences could have translated into different discharge delays and different allocation of diagnostic and therapeutic resources.

Methods

Study population. Patients who underwent transfemoral cardiac catheterization and/or percutaneous coronary intervention (PCI) with a 6 Fr arterial sheath at our institution from May 2007 to May 2010 were screened (Figure 1). Of them, from May 2007 to December 2007, 672 consecutive patients were treated with MC. From January 2008 to May 2010, a policy of systematic use of VCD was adopted and 1281 consecutive patients were treated with a 6 Fr VCD (677 Angio-Seal STS and 604 Angio-Seal Evolution). Differences in cath lab throughput, with a progressive reduction of the monthly femoral procedures rate, are explained by the steady increase of radial approach at our institution. To take into account the learning curves, the first 20 patients treated with Angio-Seal STS and Evolution were excluded from the analysis, so the final cohort counted 1913 patients. Among them, we identified 88 patients with vascular complications after VCD deployment or MC, who represent the final study population.

Cardiac catheterization, PCI, and VCD deployment. Cardiac catheterization and PCI were performed using standard techniques. As routine practice, at the end of the procedure, a femoral angiogram via the arterial sheath was obtained. According to current recommendations, patients did not undergo VCD deployment 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 burden was present in the common femoral artery. If none of those angiographic contraindications were present, a VCD was deployed to achieve hemostasis. Deployment of any VCD was performed according to the standard recommended techniques. Routine antibiotic prophylaxis was performed by oral amoxicillin 2 g administered 1 hour before cardiac catheterization. In patients with a history of previous allergic reaction to beta-lactamic drugs, intravenous gentamicin 1.5 mg/Kg was used. In case of urgent cardiac catheterization, intravenous ampicillin 2 g was administered immediately before the procedure. During diagnostic catheterization, the administration of intravenous heparin was left to the operator’s discretion. During PCI, a 5000 U bolus of heparin was administered at the beginning of the procedure followed by additional bolus(es) to achieve an activated clotting time ranging between 200 and 300 seconds. The use of intravenous glycoprotein IIb/IIIa inhibitors was left to the operator’s discretion.

Clinical follow-up. As is routine at our institution, a groin check was routinely made 1 hour and 6 hours following the procedure and prior to discharge. In case of groin hematoma or new femoral bruit detection, an ultrasound evaluation of the groin was performed. Presence and B-mode measures of hematomas, pseudoaneurysms, or arteriovenous fistulas were assessed and recorded. The patient charts were evaluated by a clinical research nurse and outcomes were entered into a dedicated database.

The following complications were registered: groin bleeding (any uncontrolled bleeding from the femoral stick site); groin hematoma >5 cm (any groin hematoma with at least one of its EchoDuplex 2-D dimensions >5 cm); groin pseudo-aneurysm (any groin pulsatile hematoma with evidence of a patent communication to the femoral artery at the EchoDuplex examination); groin arteriovenous fistula (any iatrogenic communication between femoral artery and vein as demonstrated by the EchoDuplex examination); groin infection (any groin swelling, tenderness, and pain associated to fever, chills, and/or increased white blood cell count); retroperitoneal bleeding (any hemorrhagic infiltration of the retroperitoneal structures, as assessed by CT scan); limb-threatening ischemia (any loss of leg pulses with evidence of major leg vessel occlusion at the EchoDuplex and Color Doppler examination); femoral surgical/interventional repair (any surgical or interventional radiology repair of any of the previously reported vascular complications); and the need for transfusion (any packed red cell transfusion due to severe hemorrhagic anemia).

Study endpoints. In the 88 patients with complicated hemostasis following transfemoral procedure, the occurrence of the following endpoints was noted:

1. Individual vascular complications occurring in the 24 hours following cardiac catheterization: groin bleeding, groin hematoma >5 cm, pseudoaneurysm, arteriovenous fistula, groin infection, retroperitoneal hemorrhage, limb-threatening ischemia, surgical/interventional repair, hemoglobin loss >3 g/dL, need for transfusion as a consequence of any vascular complication.
2. Hospital discharge delay, defined as the prolongation of hospital stay due to any vascular complication (in patients with prolonged hospitalization due to causes unrelated to vascular complication of cardiac catheterization, delay to achieve mobilization was considered).
3. Need for vascular surgeon consultation, echo-Duplex scanning or computed tomography scan.

Femostop (RADI Medical Systems, Inc.), a pneumatic compression device, was used when hemostasis was not adequate despite manual pressure or ASE deployment. Ambulation was routinely initiated 2 to 4 hours following successful VCD deployment and 6 to 8 hours following manual or Femostop compression.

The study was approved by the Local Institutional Review Board and all patients gave a written informed consent to the investigation.

Statistical analysis. Descriptive statistics (means and standard deviations for continuous factors, frequency counts and relative frequencies for categorical factors) were computed for the whole population and by closure type (VCD vs MC). Comparison between groups for continuous variables was performed by unpaired t-test (in case of parametric distribution) or Mann-Whitney U-test (in case of non-parametric distribution), as appropriate. Univariate associations between treatment allocation and patient features and outcomes were examined using two-way contingency tables; significance of associations were assessed using the Chi-squared test or the Fisher exact test, as appropriate. Risk of specific vascular complication was expressed as relative risk (RR) and 95% confidence interval (CI). A P-value of <.05 was required to reject the null hypothesis. Statistical analyses were performed using the SPSS statistical software package version 16.0 (IBM Corporation).

Propensity score matching. The logistic model, which included patient background characteristics (age, sex, obesity, diabetes, hypertension, high blood cholesterol, smoking, peripheral artery disease, presence of acute coronary syndrome, PCI) and drug classes affecting coagulation and platelet function (aspirin, clopidogrel, sodium-heparin, low-molecular-weight heparin, IIb/IIIa inhibitors, thrombolytics), was used to generate a propensity score for each individual in the data set. Propensity score matching was performed using a 5-digit, greedy 1:1 matching algorithm.^27,28 This statistical analysis was performed with the SAS statistical software package version 9.1.

Results

Clinical features and overall complication rate. Clinical features of the overall study population are summarized in Table 1. Vascular complication rates are summarized in Table 2. Complications were observed in 88 patients, 55 after AS placement and 33 after MC (4.4% vs 4.9%; P=.65). Patients with complications after AS and MC differed only for age, which was older in the MC group (71.3 ± 8.4 years vs 63.8 ± 10.4 years, respectively; P=.0006). Patients treated with AS had lower risk of groin hematoma (RR, 0.59; 95% CI, 0.37-0.94; P=.028), and pseudoaneurysm formation (RR, 0.27; 95% CI, 0.08-0.89; P=.03), but higher rates of retroperitoneal bleeding (RR, 6.23; 95% CI, 1.47-26.33; P=.013) with more frequent need for transfusions (RR, 2.53; 95% CI, 1.05-6.07; P=.038).

Rates of specific vascular complications. Rates of specific vascular complications in patients with complicated hemostasis after AS or MC are summarized in Table 3 and Figure 2. The rates of major vascular complications (47.3% vs 18.2%; RR, 2.60; 95% CI, 1.18-5.65; P=.016), retroperitoneal bleeding (41.8% vs 6.1%; RR, 8.43; 95% CI, 2.14-33.3; P=.002), hemoglobin loss >3 g/dL (63.6% vs 39.4%; RR, 1.61; 95% CI, 1.01-5.58; P=.045) and the need for transfusions (49.1% vs 18.2%; RR, 2.70; 95% CI, 1.25-5.84; P=.01) were higher after complicated AS hemostasis (Table 3 and Figure 2), while groin hematomas (100% vs 65.5%; RR, 0.66; 95% CI, 0.54-0.79; P=.0005) and femoral pseudoanurysms (24.2% vs 7.3%; RR, 0.30; 95% CI, 0.10-0.92; P=.035) were more common after complicated MC hemostasis (Table 3 and Figure 2). Surgical or interventional vascular repair rates were similar in complicated cases after AS or MC (7.3% vs 12.1%; RR, 0.60; 95% CI, 0.16-2.24; P=.45; Table 3 and Figure 2).

After propensity score matching, groin hematomas remained more frequent after MC (100.0% vs 80.1%; RR, 1.62; 95% CI, 1.18-2.22; P=.003; Table 3) with a borderline significant trend for more pseudoaneurysms (27.6% vs 3.9%; RR, 7.17; 95% CI, 0.96-53.55; P=.055; Table 3). Similarly to unmatched analysis, retroperitoneal bleeding rates were significantly higher after AS (34.6% vs 9.9%; RR, 5.02; 95% CI, 1.19-21.14; P=.026; Table 3).

Allocation of diagnostic/therapeutic resources. Use of a vascular surgeon consultation (0.84 ± 0.37 vs 0.91 ± 0.77 consultations/patient; P=.55) and echo-duplex scans (0.87 ± 0.39 vs 0.94 ± 0.79 echo-duplex scans/patient; P=.60) was similar for complicated cases after AS and MC (Table 4). However, patients with complications after AS underwent significantly more abdominal angio-CT scans (0.47 ± 0.50 vs 0.15 ± 0.36; P=.002; Table 3). On the other hand, patients with complications after MC were discharged with significantly more delay than patients with complications after AS (4.3 ± 4.0 days vs 2.7 ± 1.9 days; P=.01; Table 4). Similar results were observed after propensity score matching, with more angio computed tomography scans per patient used after AS failure (0.6 ± 0.4 vs 0.2 ± 0.4; P=.001), but more discharge delay after MC failure (2.7 ± 1.8 days vs 5.8 ± 6.2 days; P=.02; Table 4).

Discussion

The main finding of the present study is that, in a large single-center series, different hemostasis strategies, namely AS and MC, were affected by different vascular complications. In fact, retroperitoneal bleedings were more frequent after using AS, while groin hematomas prevailed after MC. As a consequence, the allocation of diagnostic and therapeutic resources and the discharge delay were different in relation to the hemostasis strategy chosen.

VCD use, even if it facilitates rapid hemostasis and early patient ambulation,^1-3 is not free from severe and sometimes life-threatening complications.^8,9 Previous studies^13-26 reported substantially similar incidences of vascular complications after VCD deployment and femoral manual compression (MC) hemostasis, with sporadic reports of excessive minor complications after VCD deployment.²⁹ Consequently, a position paper by the American Heart Association placed VCDs in Class IIa when employed to achieve faster hemostasis and improve patient comfort, but in Class III when used with the purpose to reduce vascular complications.¹⁰ On the other hand, more recent information from the National Cardiovascular Data Registry suggested that VCDs could be associated with significantly lower bleeding rates, particularly among patients at greatest risk for bleeding,¹¹ and Dauermann et al stated that VCDs are a key factor to improve bleeding complications in patients treated by cardiac catheterization and coronary interventions.¹² Thus, the topic is highly debated and, in the majority of catheterization laboratories, the decision to use VCD or MC is left to the operator’s discretion, often depending on the preference and expertise of the single interventional cardiologist. However, the spectrum of vascular events complicating VCD or MC hemostasis is highly heterogeneous, ranging from disturbing groin hematomas to severe retroperitoneal bleeding and limb-threatening ischemia. As diagnostic and therapeutic needs can be critically influenced by the relative frequency of each kind of complication, the knowledge of the rate and severity of vascular complications after VCD and MC could be important in the decision process. Unfortunately, most clinical studies are largely underpowered to detect differences in low-frequency endpoints and consequently focused on composite endpoints.^{1,4,7,13,16,23,30,31}

Our study, according to previous studies, demonstrated that the vascular complication spectrum is quite different between the two strategies. In fact, an increased relative frequency of groin hematomas after MC in comparison to VCD was described in the analysis of Resnic at al³² and other studies reported similar trends.^6,13,14

The reason for the higher rates of retroperitoneal bleeding after VCD is not clear and could be related to high puncture site, which could offer a low-resistance way for blood oozing in case of VCD failure. However, femoral angiograms were not reviewed in our study and this hypothesis could not be verified. On the other hand, the longer discharge delay observed with MC could be a consequence of the high incidence of groin hematomas, some of them evolved in pseudoaneurysms treated by surgical repair. Actually, in the population studied, the pseudoaneurysm incidence was higher after MC, even if surgical/interventional repair was not statistically different according to the hemostasis strategy.

Our study did not evaluate other popular VCDs, such as Perclose or Starclose devices, as AS was the platform chosen by our cath lab for extensive routine use. Thus, caution should be taken before applying our results to patients treated by other VCDs.

Study limitations. First, this study is not a randomized trial; therefore, the results may be confounded by other factors. Although this is unlikely to occur, given the broad spectrum of characteristics evaluated in our study and which were taken into account by the propensity score matching, we cannot exclude residual confounding by unmeasured factors.³³

Moreover, the two consecutive series compared in the present study are not contemporary, but MC cases were collected in a period immediately antecedent to the adoption by our cath lab of a policy of systematic AS use. Our intention was to avoid a source of selection bias, as after AS introduction in our institution, MC was reserved for cases with angiographic contraindications to VCD. The comparison of non-contemporary cohorts can introduce bias due to changes in therapeutic practice over time. However, the rates of use of antiplatelets, anticoagulants, and their combinations, as well as the rates of PCI, were similar in the two groups. Moreover, after propensity score matching, groin hematomas remained the most frequent complication after MC and retroperitoneal bleeding the leading cause of trouble after AS deployment.

Another limitation could arise from the use of different VCDs, namely Angio-Seal STS and the novel Angio-Seal Evolution. However, subgroup analysis did not provide different results in comparison to the overall population analysis (data not shown), even if without sufficient statistical power due to the small number of events in subgroups.

Finally, the definition of “large groin hematoma” has been a subject of debate and the choice of a cut-off of ≥5 cm, as well as the methods of hematoma measurements, could be questionable. In a meta-analysis by Nickolsky et al,²⁰ the presence of a very wide-ranged criteria for the definition of groin hematoma was remarked and authors did not consider groin hematoma as a primary endpoint to avoid bias. Thus, in the present study, the cut-off was chosen at 5 cm, as in a large study by Bangalore et al.²²

Conclusion

Our study demonstrated that AS and MC were associated with different types of vascular complication, with a higher prevalence of retroperitoneal bleeding needing transfusions after AS deployment. However, hospital discharge was more delayed after MC, probably related to a higher incidence of pseudoaneurysms and the consequent surgical treatment. Comparison between strategies for achieving hemostasis after transfemoral procedures should take into account the different logistic and economic burdens imposed by different vascular complications.

Acknowledgments. We gratefully acknowledge the efforts of Maria Grazia D’Ulisse, Anna Federzoni, Chiara Clementi, Maria Teresa Bausani, Adele Mandrino, Pierangelo Rizzotti, and Roberta Sandri for their expertise and support in the cath lab and data collection.

References

1. Michalis LK, Rees MR, Patsouras D, et al. A prospective randomized trial comparing the safety and efficacy of three commercially available closure devices (Angio-Seal, VasoSeal and Duett). Cardiovasc Intervent Radiol. 2002;25(5):423-429.
2. Kapadia SR, Raymond R, Knopf W, et al. The 6 Fr Angio-Seal arterial closure device: results from a multimember prospective registry. Am J Cardiol. 2001;87(6):789-791.
3. Eggebrecht H, Haude M, von Birgelen C, et al. Early clinical experience with the 6 French Angio-Seal device: immediate closure of femoral puncture sites after diagnostic and interventional coronary procedures. Catheter Cardiovasc Interv. 2001;53(4):437-442.
4. Applegate RJ, Grabarczyk MA, Little WC, et al. Vascular closure devices in patients treated with anticoagulation and IIb/IIIa receptor inhibitors during percutaneous revascularization. J Am Coll Cardiol. 2002;40(1):78-83.
5. Kim HY, Choo SW, Roh HG, et al. Efficacy of femoral vascular closure devices in patients treated with anticoagulant, abciximab, or thrombolytics during percutaneous endovascular procedures. Korean J Radiol. 2006;7(1):35-40.
6. Cura FA, Kapadia SR, L’Allier PL, et al. Safety of femoral closure devices after percutaneous coronary interventions in the era of glycoprotein IIb/IIIa platelet blockade. Am J Cardiol. 2000;86(7):780-782.
7. Chevalier B, Lancelin B, Koning R, et al; HEMOSTASE trial investigators. Effect of a closure device on complication rates in high-local-risk patients: results of a randomized multicenter trial. Catheter Cardiovasc Interv. 2003;58(3):285-291.
8. Van Den Broek T, Liqui Lung PF, Suttorp MJ, Eefting FD, Schipper ME, Vink A. Vascular occlusion as a late complication of the Angio-Seal closure device. A review of literature. Minerva Cardioangiol. 2007;55(6):815-819.
9. Mehta SK, Frutkin AD. Correspondence to article: closure of the femoral artery after cardiac catheterization — a comparison of Angio-Seal, StarClose, and manual compression. Catheter Cardiovasc Interv. 2008;72(4):586.
10. Patel MR, Jneid H, Derdeyn CP, et al. Arteriotomy closure devices for cardiovascular procedures: a scientific statement from the American Heart Association. Circulation. 2010;122(18):1882-1893.
11. Marso SP, Amin AP, House JA, et al. National Cardiovascular Data Registry. Association between use of bleeding avoidance strategies and risk of periprocedural bleeding among patients undergoing percutaneous coronary intervention. JAMA. 2010;303(21):2156-2164.
12. Dauerman HL, Rao SV, Resnic FS, Applegate RJ. Bleeding avoidance strategies. Consensus and controversy. J Am Coll Cardiol. 2011;58(1):1-10.
13. Amin FR, Yousufuddin M, Stables R, et al. Femoral haemostasis after transcatheter therapeutic intervention: a prospective randomised study of the Angio-Seal device vs the Femostop device. Int J Cardiol. 2000;76(2-3):235-240.
14. Dangas G, Mehran R, Kokolis S, et al. Vascular complications after percutaneous coronary interventions following hemostasis with manual compression versus arteriotomy closure devices. J Am Coll Cardiol. 2001;38(3):638-641.
15. Koreny M, Riedmüller E, Nikfardjam M, Siostrzonek P, Müllner M. Arterial puncture closing devices compared with standard manual compression after cardiac catheterization: systematic review and meta-analysis. JAMA. 2004;291(3):350-357.

16. Martin JL, Pratsos A, Magargee E, et al. A randomized trial comparing compression, Perclose Proglide and Angio-Seal VIP for arterial closure following percutaneous coronary intervention: the CAP trial. Catheter Cardiovasc Interv. 2008;71(1):1-5.
17. Veasey RA, Large JK, Silberbauer J, et al. A randomised controlled trial comparing StarClose and Angio-Seal vascular closure devices in a district general hospital — the SCOAST study. Int J Clin Pract. 2008;62(6):912-918.
18. Deuling JH, Vermeulen RP, Anthonio RA, et al. Closure of the femoral artery after cardiac catheterization: a comparison of Angio-Seal, StarClose, and manual compression. Catheter Cardiovasc Interv. 2008;71(4):518-523.
19. Rastan A, Sixt S, Schwarzwälder U, et al. VIPER-2: a prospective, randomized single-center comparison of 2 different closure devices with a hemostatic wound dressing for closure of femoral artery access sites. J Endovasc Ther. 2008;15(1):83-90.
20. Nikolsky E, Mehran R, Halkin A, et al. Vascular complications associated with arteriotomy closure devices in patients undergoing percutaneous coronary procedures: a meta-analysis. J Am Coll Cardiol. 2004;44(6):1200-1209.
21. Vaitkus PT. A meta-analysis of percutaneous vascular closure devices after diagnostic catheterization and percutaneous coronary intervention. J Invasive Cardiol. 2004;16(5):243-246.
22. Bangalore S, Arora N, Resnic FS. Vascular closure device failure: frequency and implications: a propensity-matched analysis. Circ Cardiovasc Interv. 2009;2(6):549-556.
23. Park Y, Roh HG, Choo SW, et al. Prospective comparison of collagen plug (Angio-Seal) and suture-mediated (the Closer S) closure devices at femoral access sites. Korean J Radiol. 2005;6(4):248-255.
24. Abando A, Hood D, Weaver F, Katz S. The use of the Angio-Seal device for femoral artery closure. J Vasc Surg. 2004;40(2):287-290.
25. Upponi SS, Ganeshan AG, Warakaulle DR, et al. Angio-Seal versus manual compression for haemostasis following peripheral vascular diagnostic and interventional procedures — a randomized controlled trial. Eur J Radiol. 2007;61(2):332-334.
26. Hermanides RS, Ottervanger JP, Dambrink JH, et al. Closure device or manual compression in patients undergoing percutaneous coronary intervention: a randomized comparison. J Invasive Cardiol. 2010;22(12):562-566.
27. Parsons LS. Reducing bias in a propensity score matched-pair sample using greedy matching techniques. In: The Twenty-Sixth Annual SAS Users Group International Conference. Cary, NC: SAS Institute, 2004;214-226.
28. Black PE. Greedy algorithm. In: Black PE, ed. Dictionary of Algorithms and Data Structures [online]. Gaithersburg, MD: US National Institute of Standards and Technology; February 2, 2005. https://www. itl.nist.gov/div897/sqg/dads/HTML/greedyalgo.html. 
29. Lupi A, Lazzero M, Plebani L, Sansa M, Bongo AS. Safety and efficacy of the new Angio-Seal Evolution closure device: a single-center experience. J Invasive Cardiol. 2011;23(4):150-155.
30. Behan MW, Large JK, Patel NR, Lloyd GW, Sulke AN. A randomised controlled trial comparing the routine use of an Angio-Seal STS device strategy with conventional femoral haemostasis methods in a district general hospital. Int J Clin Pract. 2007;61(3):367-372.

31. Assali AR, Sdringola S, Moustapha A, et al. Outcome of access site in patients treated with platelet glycoprotein IIb/IIIa inhibitors in the era of closure devices. Catheter Cardiovasc Interv. 2003;58(1):1-5.
32. Resnic FS, Arora N, Matheny M, Reynolds MR. A cost-minimization analysis of the Angio-Seal vascular closure device following percutaneous coronary intervention. Am J Cardiol. 2007;99(6):766-770.
33. Rubin DB. Estimating causal effects from large data sets using propensity scores. Ann Intern Med. 1997;127(8 Pt 2):757-763.

____________________________________________

From ¹Hospital Cardiology, “Maggiore della Carità” Hospital, Novara, Italy, ²Division of Cardiology, University of Eastern Piedmont, “Maggiore della Carità” Hospital, Novara, Italy, ³Division of Cardiology, “S. Andrea” Hospital, Vercelli, Italy, and 4Division of Cardiology, “Castelli” Hospital, Verbania, Italy.
Disclosure: The authors have completed and returned the ICMJE Form for Disclosure of Potential Conflicts of Interest. The authors report no conflicts of interest regarding the content herein.
Manuscript submitted August 29, 2011, provisional acceptance given September 27, final version accepted December 20, 2011.
Address for correspondence: Alessandro Lupi, MD, Ospedale Maggiore della Carità, Cardiologia 2, Cso Mazzini 18, 28100, Novara, Italy. Email: lupialessandro1@tin.it