Propensity Score Analysis of Vascular Complications after
Diagnostic Cardiac Catheterization and Percutaneous Coronary Interven

Cardiac catheterization and PCI are usually performed by percutaneous access using the femoral artery. Following completion of the procedure, hemostasis is traditionally performed by manual compression, followed by 6 hours of bed rest. Hemostatic patches have been used as an adjunct for manual compression for management of femoral access site sheath removal. Use of hemostasis patches has been shown in small studies to reduce the time to hemostasis and ambulation compared to manual compression after percutaneous procedures involving femoral artery access.^1–5 However, only limited data are available examining their efficacy and safety in broader clinical practice.^6,7 Recently, the risk of local adverse events following the use of 2 patches, Chito-Seal (Abbott Vascular Devices, Redwood City, California) and the Syvek NT Patch (Marine Polymer Technologies, Inc., Dankers, Massachusetts) for access site hemostasis after cardiac catheterization procedures has been reported.⁷ Neither patch was associated with a lower risk of vascular complications compared to manual compression alone, although there was a trend towards a higher risk of entry site bleeding with the Syvek patch compared to manual compression.
Topical thrombin has been used for local hemostasis for a variety of surgical and nonsurgical procedures with great success.^8,9 It has been incorporated into a bandage for topical hemostasis, including use for cardiac catheterization procedures. However, the safety and efficacy of a thrombin hemostatic patch for access site hemostasis after diagnostic cardiac catheterization and PCI procedures in clinical practice is not known. Accordingly, we compared vascular complications in a cohort of patients with femoral artery access management by manual compression to a cohort of consecutive patients with femoral artery access site managed with thrombin hemostatic patch-facilitated manual compression, adjusting for potential baseline differences using propensity score analysis.

Methods

All patients at our institution undergoing percutaneous cardiac catheterization (CATH) and coronary revascularization (PCI) using the femoral artery approach between April 2002 and November 2005 were evaluated for this study, which was approved by the Institutional Review Board. Nine patients with in-lab vascular complications were excluded from analysis. A total of 4,371 patients undergoing 2,956 CATH and 1,415 PCI procedures using the femoral approach between April 2002 and March 2004 formed the control group for this study. Thrombin hemostatic patch (D-Stat dry; Vascular Solutions; Minneapolis, Minnesota) facilitated manual compression (THP-MC) was introduced in April 2004, replacing simple manual compression for obtaining hemostasis after femoral artery sheath removal. A total of 3,464 procedures including 2,464 consecutive CATH and 1,000 PCI procedures in which THP-facilitated manual compression was performed form the treatment group of this study. CATH and PCI procedures were performed according to standard techniques.
Anticoagulation after sheath insertion for PCI patients was obtained using unfractionated heparin or bivalirudin. Patients in the study received glycoprotein (GP) IIb/IIIa receptor inhibition according to usual protocol with abciximab or eptifibatide.10 Post-PCI patients were treated with aspirin (81–325 mg a day) and clopidogrel (300 to 600 mg as a loading dose followed by 75 mg/day) if stents were placed.
Femoral artery access management. The method of arterial access management was chosen by the cardiologist performing the procedure.¹¹ The femoral artery sheath was removed immediately after the procedure for cardiac catheterization procedures, and when the ACT was ≤180 seconds in patients who received heparin or ≥2 hours after bivalirudin infusion was completed, for PCI procedures. Guidelines for hemostasis were as follows: Manual compression was achieved using standard techniques. For all CATH procedures, occlusive pressure was initially held for 5 minutes; for 4–5 Fr sheaths, subocclusive pressure was held for an additional 5 minutes or until hemostasis was achieved; for 6 Fr sheaths, subocclusive pressure was held for an additional 10 minutes or until hemostasis was achieved. For all PCI procedures regardless of sheath size, occlusive pressure was held for 10 minutes, and subocclusive pressure was held for an additional 10 minutes or until hemostasis was achieved. For THP-MC cases, an initial trial of subocclusive compression at 1 minute per French size for CATH, and 2 minutes per French size for PCI procedures was completed. After this initial period, a simpler guideline with longer compression was employed: the THP bandage was placed over the access site and subocclusive manual compression applied for 10 minutes after all CATH procedures, for 20 minutes after all PCI procedures, or until satisfactory hemostasis was obtained. C-clamps were used in a minority of the PCI patients similarly for MC or THP-MC. By protocol, the THP was removed in 6 hours. Ambulation was permitted 2 hours after the THP was placed for both CATH, and 3 hours after PCI, and 3–6 hours following manual compression alone for CATH (3 hours for 4 Fr, 3–4 hours for 5 Fr and 4–6 hours for 6 Fr), and 6–8 hours for PCI.
Femoral access site evaluation was routinely made postprocedure and prior to discharge by the medical team caring for the patient. The presence of vascular complications was recorded on the chart. Prior to hospital discharge, the patient’s chart was evaluated by a clinical research nurse.¹⁰ Outcomes measures collected conformed to the ACC database definitions for vascular complications.¹² Minor vascular complications were defined as any of the following: hematoma >10 cm, arteriovenous fistulae or pseudoaneurysm. Major vascular complications were defined as: death due to vascular complications, vascular repair, major vascular bleeding (>3 gm Hgb drop due to access site bleeding, required transfusion, and/or prolonged hospital stay or retroperitoneal bleeding), vessel occlusion or loss of pulse.^{10, 12}
Statistical methods. Baseline covariates were compared by Chi-square testing (categorical) or the Wilcoxon rank sum test (continuous). Vascular complications were initially compared with multivariate logistic regression models using generalized estimating equations to account for repeated measures.¹³ Propensity scores indicating the likelihood of receiving a THP were calculated for each patient based on a nonparsimonious logistic regression model¹⁴ constructed with THP-MC as the dependent variable. All the variables shown in Table 1 were included in the model except for closure success, as well as significant interactions. The c-statistic for the propensity score model was 0.59, indicating similar distribution of baseline covariates between the two groups. The relative incidence of vascular complications was evaluated by logistic regression models using the propensity score as a single covariate.¹⁵ SAS, version 8.02 statistical software package (SAS Institute, Cary, North Carolina) was used for all statistical analysis.

Results

Baseline clinical and procedural characteristics are shown in Table 1. The two groups were reasonably similar with respect to baseline clinical characteristics, although there were small differences in some variables between the two groups. Approximately two-thirds of patients in each group underwent CATH, while the remainder underwent PCI. Average sheath size was 5.5 ± 0.9 Fr in the MC CATH group, and 5.7 ± 0.7 Fr in the THP-MC CATH group, p <0.001; 6.9 ± 0.5 Fr in the MC PCI group and 6.8 ± 0.6 Fr in the THP-MC PCI group, p <0.001. Among patients undergoing PCI, GP IIb/IIIa inhibitors were used similarly, 86.7% in the MC group, and 88.2% in the THP group, p = 0.280.
Time to hemostasis is shown in Table 2 for a subgroup of CATH and PCI patients. In a subgroup of 196 CATH procedures hemostasis times were 13.0 ± 3.3 minutes for the THP-MC, and 14.4 ± 5.7 minutes for MC, p = 0.51. However, for 6 Fr CATH procedures, the hemostasis times were shorter for THP-MC, 13.5 ± 3.3 minutes, than for MC, 17.0 ± 6.4 minutes, p = 0.008. In a subgroup of 69 PCI procedures, hemostasis times were 14.2 ± 5.4 minutes for THP-MC and 20.1 ± 5.4 minutes for MC, p <0.001. Initiation of ambulation was guided by protocol as outlined above, and time to ambulation was not measured.
The incidence of vascular complications for both groups is shown in Figure 1. Overall, any vascular complication occurred in 1.0% of the MC group and 0.6% of the THP-MC group, p <0.05. The incidence of major vascular complications was similar, while minor vascular complications occurred less frequently in the THP-MC group, 0.4%, than the MC group, 0.7%, p <0.05. The incidence of individual vascular complications is shown in Table 3. The incidence of hematoma >10 cm occurred less frequently in THP-MC vs. MC patients, 0.03% vs. 0.37%, p = 0.001. This was due almost entirely to a decrease in the incidence of hematoma >10 cm with THP-MC compared to MC alone in CATH procedures using 6 Fr sheaths. There were no other significant differences in the incidence of the individual vascular complications between the two groups for either CATH or PCI.
The incidence of vascular complications stratified by clinical and procedural variables is shown in Figure 2. The risk of vascular complication favored THP-MC for almost all of these variables, although none was statistically significant because of wide confidence intervals. Interestingly, there were no vascular complications with 4 Fr CATH, and only 2 vascular complications with 5 Fr CATH using either method of hemostasis.
The incidence of vascular complications for patients receiving femoral artery access management with vascular closure devices (VCDs) for the two time periods of the study are shown in Table 4 for comparison. Although there was a trend towards a decrease in vascular complications from the earliest time period to the more recent time period with VCD use, these were not statistically significant. Importantly, the incidence of minor vascular complications including hematomas >10 cm was identical after VCD use in both time periods.
The unadjusted odds ratio (OR) for any vascular complication THP-MC compared to MC was 0.57 (0.34–0.97) (Table 5). The multivariate-adjusted OR for any vascular complication THP-MC versus MC alone was 0.58 (0.34–0.99), while the propensity score-adjusted OR was 0.58 (0.34–0.99). The OR of any, major, and minor vascular complications for each procedure type comparing THP-MC to MC is shown in Figure 3. The OR for CATH was 0.42 (0.20–0.87), primarily due to lower risk of minor vascular complications with THP-MC, 0.30 (0.11–0.80). Also shown in Table 5, the results of the multivariate analysis for the entire study group identified THP-MC as associated with decreased risk of any vascular complication, and older age female gender, and larger sheath size as independently predictive of higher risk of any vascular complication.

Discussion

In unselected patients undergoing CATH and PCI we observed that THP-MC use reduced manual compression times necessary to achieve hemostasis modestly compared to conventional manual compression, ranging from 29% (7 Fr PCI) to 0% (4 Fr CATH). Following a strategy of ambulation permitted at 2–3 hours post procedure with THP-MC, we observed a low incidence of vascular complications after both CATH and PCI procedures. The adjusted odds ratio of any vascular complication after THP-MC, compared to manual compression alone, in a cohort of CATH and PCI procedures was 0.58 (0.34–0.99). The lower hazard of any vascular complication was almost entirely due to a lower incidence of hematoma >10 cm with THP-MC compared to MC after 6 Fr CATH procedures, 0.03% vs. 0.4%, p <0.05. Thus, a strategy of THP-MC for femoral artery access site hemostasis reduced manual compression times, permitted early ambulation, and was as safe as conventional manual compression. The cost effectiveness of this hemostasis strategy was not evaluated in this study, however, and remains to be determined in future studies.
Issues related to understanding the role of patch use for hemostasis after arterial access include identifying the mechanism of action by which local hemostasis is achieved, and demonstrating whether patch use adds any benefit beyond manual compression itself. Unfortunately, the evidence base available to help address these issues is limited. Hemostatic patches have been available for commercial use for years, but only recently have gained specific FDA approval for use as an external closure device for management of femoral artery access sites. While multiple types of patches have been evaluated, including chitines, marine algae and thrombin, the exact mechanism by which they achieve hemostasis in patients is unclear.^16–18In vitro and ex vivo studies indicate that all of the hemostatic patches stimulate coagulation. For superficial wounds or oozing, topical thrombin promotes rapid anticoagulation, which has been well documented, and accounts for its frequent use in surgical procedures.^8,9 After application on the skin over a femoral artery puncture site, topical thrombin appears to promote coagulation within the upper aspects of the access tract, as the topical thrombin itself does not likely diffuse to the surface of the artery. Presumably, thrombus formation also occurs at the arteriotomy site with final hemostasis being achieved by a combination of these two phenomena. Future studies will be needed to more precisely define the mechanism by which topical thrombins, and other agents used in hemostatic patches, achieve hemostasis.
Because each patch requires a period of manual compression in addition to application of the patch, it has not been clear that there are hemostasis and clinical benefits beyond the manual compression itself. The extent to which manual compression contributes to hemostasis has not been specifically studied. In this experience, after an initial roll-in phase where an effective strategy of achieving hemostasis was achieved, we demonstrated shorter compression times with THP-MC than with MC alone with 6 Fr and larger sheaths. The compression times however, were longer than the manufacturers’ recommended times of 1 minute per sheath Fr size. This experience suggests, but is by no means conclusive, that manual compression is an essential component of hemostasis with patches. However, THP itself appears to be an active component contributing to hemostasis as well.
There are only limited data evaluating outcomes after hemostatic patch use for femoral artery access site management. In small, randomized trials poly-N-acetyl glucosamine^4,5 and a hydrophilic wound pad³ reduced the time to hemostasis and ambulation compared to manual compression after cardiac procedures. In these studies, the incidence of vascular complications after hemostatic patch use was low and was comparable to manual compression. Nader et al examined vascular complications in 1,000 patients after CATH and PCI procedures and electrophysiologic studies.⁶ They observed a low incidence of vascular complications, but did not include a control group for comparison. The Chito-Seal and Syvek patches were included in the Phase II report of the ACC-NCDR,⁷ and appeared equally safe as manual compression alone. However, details concerning time to hemostasis and ambulation were not provided in this report. Moreover, the D-Stat Dry patch was not included in this analysis. There are extremely limited reports of the incidence of vascular complications after D-Stat Dry use for hemostasis of femoral artery access sites. Elian et al² reported no vascular complications in 41 patients using THP-MC with D-Stat Dry after diagnostic catheterization. In this study, the risk of vascular complications was essentially similar for MC and THP-MC for CATH and PCI. The risk of hematoma >10 cm was lower after THP-MC for CATH, almost entirely due to a significant decrease in hematoma with 6 Fr catheters. For both THP-MC and MC, however, the absolute rates of hematoma were low and it is unclear that this represents a clinically meaningful difference.
While the current study supports the conclusion that the risk of vascular complications after THP-MC and MC alone are comparable, there are limitations of the study that should be discussed. The study groups were obtained from sequential rather than contemporary cohorts. Thus, vascular complications rates in general may have been lower in the more contemporary cohort,¹⁹ favoring THP-MC. While the exact magnitude of this potential factor will remain unmeasured, several lines of reasoning suggest that this is unlikely to have significantly affected the findings. First, the incidence of factors known to affect the rate of vascular complication such as age, gender, history of vascular disease, BMI and sheath size were all similar between the two groups. Moreover, the incidence of upstream GP IIb/IIIa and clopidogrel use was also similar in both groups. We compared the incidence of vascular complications in patients in whom vascular closure devices were used in the two different time periods. The incidence of vascular complications was similar in these two groups. Thus, neither a reporting bias, nor a general improvement in vascular access management, would seem to have influenced the study results. Finally, ambulation by protocol after THP-MC was permitted at 2 hours postprocedure, compared to 4–6 hours postprocedure for the MC-alone group. Thus, the results should have been biased towards lower vascular complications after manual compression alone because of longer bed rest. Like any observational study, the current results may be subject to unrecognized biases in patient selection and/or procedures not accounted for in this analysis. Only appropriately powered randomized clinical trials will adequately answer whether THP-MC definitively improves safety for femoral artery access site management compared to manual compression alone.

Conclusion

In this large, single-center observational study, we evaluated the strategy of thrombin hemostatic-facilitated manual compression and ambulation at 2–3 hours after diagnostic cardiac catheterization and PCI procedures. We observed that the use of thrombin hemostatic patch facilitated manual compression for diagnostic cardiac catheterization and PCI procedures, reduced, but did not eliminate manual compression time to achieve hemostasis, permitted an early ambulation strategy, and was associated with an incidence of vascular complications similar to that of manual compression alone. Thus, thrombin hemostatic patch facilitated manual compression may allow an early ambulation strategy without compromising safety. The cost effectiveness of this strategy remains to be determined.

Acknowledgements. We gratefully acknowledge Tammy Davis for manuscript preparation; and Aruna Joel, Sabrina Smith and Robin Taylor for data collection and database entry.

References

1. Gruchevsky M, Manubens C, Hardy S, et al. Reduced time to ambulation in the diagnostic catheterization patient. Am J Cardiol 2004;94(Suppl 6A):202E.
2. Elian D, Guetta V, Alcalai R, et al. Early mobilization after diagnostic cardiac catheterization with the use of a hemostatic bandage containing thrombin. Cardiovasc Revasc Med 2006;7:61–63.
3. Mlekusch W, Dick P, Haumer M, et al. Arterial puncture site management after percutaneous transluminal procedures using a hemostatic wound dressing (Clo-Sur P.A.D.) versus conventional manual compression: A randomized controlled trial. J Endovasc Ther 2006;13:23–31.
4. Najjar SF, Healey NA, Healey CM, et al. Evaluation of poly-n-acetyl glucosamine as a hemostatic agent in patients undergoing cardiac catheterization: A double-blind, randomized study. J Trauma 2004;57:S38–S41.
5. Palmer BL, Gantt DS, Lawrence ME, et al. Effectiveness and safety of manual hemostasis facilitated by the SyvekPatch with one hour of bedrest after coronary angiography using six-French catheters. Am J Cardiol 2004;93:96–97.
6. Nader RG, Garcia JC, Drushal K, Pesek T. Clinical evaluation of the SyvekPatch in consecutive patients undergoing interventional, EPS and diagnostic cardiac catheterization procedures. J Invasive Cardiol 2002;14:305–307.
7. Tavris DR, Dey S, Albrecht-Gallauresi B, et al. Risk of local adverse events following cardiac catheterization by hemostasis device use: Phase II. J Invasive Cardiol 2005;17:644–650.
8. Holcomb JB, Pusateri AE, Harris RA, et al. Dry fibrin sealant dressing reduce blood loss, resuscitation volume, and improve survival in hypothermic coagulopathic swine with grade V liver injuries. J Trauma 1999;47:233–240.
9. Froud T, Yrizarry JM, Alejandro R, Ricordi C. Use of D-STAT to prevent bleeding following percutaneous transhepatic intraportal islet transplantation. Cell Transplant 2004;13:55–59.
10. Applegate RJ, Sacrinty M, Kutcher MA, et al. Propensity score analysis of vascular complications after diagnostic cardiac catheterization and percutaneous coronary intervention 1998-2003. Catheter Cardiovasc Interv 2006;67:556–562.
11. Applegate RJ, Draughn T, Yarbrough WD, Little WC. Restenosis of a sirolimus drug-eluting stent: Aorto-ostial involvement of the proximal right coronary artery. Catheter Cardiovasc Interv 2006;67:391–395.
12. American College of Cardiology Website. American College of Cardiology-National Cardiovascular Data Registry Module Version 3.0. American College of Cardiology 2003.
13. Liang KY, Zeger SL. Longitudinal data analysis using generalized linear models. Biometrika 1986;73:13–22.
14. D'Agostino RB. Tutorial in biostatistics — Propensity score methods for bias reduction in the comparison of a treatment to a non-randomized control group. Stat Med 1998;17:2265–2281.
15. Brener SJ, Lytle BW, Casserly IP, et al. Propensity analysis of long-term survival after surgical or percutaneous revascularization in patients with multivessel coronary artery disease and high-risk features. Circ 2004;109:2290–2295.
16. Fischer TH, Connolly R, Thatte HS, Schwaitzberg SS. Comparison of structural and hemostatic properties of the poly-n-acetyl glucosamine Syvek patch with products containing chitosan. Microsc Res Tech 2004;63:168–174.
17. Thatte HS, Zagarins S, Khuri SF, Fischer TH. Mechanisms of poly-n-acetyl glucosamine polymer-mediated hemostasis: Platelet interactions. J Trauma 2004;57:S13–S21.
18. Fischer TH, Thatte HS, Nichols TC, et al. Synergistic platelet integrin signaling and factor XII activation in poly-n-acetyl glucosamine fiber-mediated hemostasis. Biomaterials 2005;26:5433–5443.
19. Konstance R, Tcheng JE, Wightman MB, et al. Incidence and predictors of major vascular complications after percutaneous coronary intervention in the glycoprotein IIb/IIIa platelet inhibitor era. J Intervent Cardiol 2004;17:65–70.