Cardiac interventions are widely accepted as a EFFECTIVENESS OF MECHANICAL COMPRESSION DEVICES IN ATTAINING HEMOSTASIS AFTER FEMORAL SHEATH REMOVAL

Transcription

1 EFFECTIVENESS OF MECHANICAL COMPRESSION DEVICES IN ATTAINING HEMOSTASIS AFTER FEMORAL SHEATH REMOVAL By Tina Jones, RN, BN, MNSc, Critical Care Cert, and Helen McCutcheon, RN, RM, BA, MPH, PhD. From the Department of Clinical Nursing, Royal Adelaide Hospital (TJ) and Adelaide University (TJ, HM), Adelaide, South Australia, Australia. BACKGROUND Cardiac interventions are widely accepted as a practical treatment option for coronary artery disease. However, few changes have occurred in the techniques used for percutaneous arterial cannulation and for attaining hemostasis after cardiac interventions. To date, researchers have focused on techniques to achieve optimal hemostasis at the time of removal of the arterial catheter and to minimize the impact and complications of arterial puncture. OBJECTIVE To summarize the best available evidence on the effectiveness of mechanical compression devices used to obtain hemostasis following femoral sheath removal after cardiac interventional procedures. METHOD An attempt was made to detect both published and unpublished reports of research evaluations of mechanical compression techniques used to attain hemostasis after femoral sheath removal. Methodological quality was assessed by using predesigned criteria. Data were extracted from information on randomized controlled trials and were statistically combined in meta-analysis where possible. Evidence was also synthesized by using narrative summaries. RESULTS Twelve studies met the inclusion criteria; however, only 3 were included in the metaanalysis. The results of meta-analysis indicated that the mechanical compression technique was the most effective for preventing formation of hematomas. The prevalence of bleeding did not differ significantly for different methods of compression. CONCLUSION A gap exists in the literature on quality randomized controlled trials of various devices used to attain hemostasis after femoral sheath removal. (American Journal of Critical Care. 2002;11: ) Cardiac interventions are widely accepted as a practical treatment option for coronary artery disease. Femoral artery puncture and cardiac catheterization are widely used diagnostic and interventional procedures, 1 and, increasingly, cardiac catheterization and related procedures are performed safely in an outpatient setting. 2 Although formerly removal of femoral sheaths after cardiac interventional procedures was done by To purchase reprints, contact the InnoVision Group, 101 Columbia, Aliso Viejo, CA Phone, (800) or (949) (ext 532); fax, (949) ; , reprints@aacn.org. physicians, more recently removal of the sheaths has become a component of the expanding role of cardiac nurses. 3 Nurses need information on the effectiveness and reliability of various techniques used to achieve hemostasis after femoral sheath removal. This information will enable cardiac nurses to choose the appropriate technique to facilitate a reduction in groin complications after femoral sheath removal. Background Despite major advances in coronary revascularization techniques in recent years, few changes have AMERICAN JOURNAL OF CRITICAL CARE, March 2002, Volume 11, No

2 occurred in the techniques used for percutaneous arterial cannulation and for attaining hemostasis after cardiac interventions. 4,5 The use of larger cannulation devices and the need for more aggressive anticoagulation have increased the prevalence of peripheral vascular complications. 5 Complications that may occur after cardiac interventions are costly, increase patients length of stay in the hospital, and affect morbidity. 6 Inadequate hemostasis can lead to one or all of the following: significant blood loss, patient s discomfort, vessel occlusion, thrombosis, formation of arteriovenous fistula, and pseudoaneurysm requiring surgical intervention. 4,7,8 Complications at the access site due to arterial cannulation occur in 1% to 5% of cases but may be as high as 14% with some cardiac interventional procedures. 9 The potential high prevalence of complications has stimulated investigation of techniques that achieve optimal hemostasis at the time of removal of the arterial catheter and minimize the impact and complications of arterial puncture. The femoral sheath provides support at the puncture site and reduces potential arterial trauma if multiple catheter exchanges are required. 1 At the end of the cardiac procedure, the femoral sheath is removed, and the femoral artery is compressed to control bleeding until hemostasis occurs. Although manual compression is commonly used to attain hemostasis, this technique has limitations. Application of manual compression may be required for up to 20 minutes or longer in order to control bleeding and allow coagulation to occur, 10 and inconsistent pressure due to hand and arm fatigue can lead to the formation of hematoma and/or thrombus. 10,11 The increase in catheterizations in outpatient settings, the desire for earlier mobilization of patients, and an effort to decrease vascular injury and complications led to the development and use of alternative hemostatic devices such as mechanical clamps, inflatable pressure devices, implantable collagen plugs, and manual pressure aids. 9 These devices can be used as an adjunct to or can provide an alternative to manual compression to attain hemostasis. Some devices have transparent domes that provide direct visualization of the puncture site during compression. Others that use pneumatic or clamp pressure to compress the femoral artery may be considered less labor-intensive than is manual compression. 12 As with manual compression, use of mechanical devices may have limitations, including patients discomfort and prolonged immobilization. More recently, collagen vascular hemostatic sealing devices were developed to attain hemostasis after femoral sheath removal. With these devices, an absorbable bovine collagen plug is instilled into the tract from the femoral artery to the skin s surface. Hemostasis occurs after formation of a fibrin clot within this tract, sealing the puncture site. 5 In preliminary clinical trials, these devices were effective in achieving hemostasis with significantly reduced groin compression time, but their use in routine procedures such as coronary angiography may be limited because of the associated costs. Methods In a systematic review, a concise methodical investigation of a subject is done by using a predetermined plan that includes the summary, appraisal, and synthesis of multiple primary studies. 16,17 The findings of well-conducted systematic reviews can help define the evidence on a topic through the assessment and integration of large volumes of relevant information. In addition, the findings may explain variations in practice and resolve conflicting evidence on a topic, thereby assisting health professionals in making rational decisions about healthcare practices. 18 The aim of this systematic review was to determine the best available evidence on the effectiveness of mechanical compression devices in achieving hemostasis and reducing hematoma formation after removal of femoral artery sheaths. Selection and Search Strategies We examined primary research studies in which the investigators compared any mechanical compression device with (1) manual compression, (2) another mechanical compression device, or (3) any other form of compression device used to attain hemostasis in adult patients after cardiac investigational procedures in which a femoral sheath approach was used. Outcomes included the time to achieve hemostasis and the prevalence of bleeding, bruising, hematoma formation, inadequate distal blood flow, arteriovenous fistula, and pseudoaneurysm formation. Bleeding was defined as any ooze, leaking, or frank blood drainage from the puncture site. A bruise was defined as any discoloration of the subcutaneous tissue around the puncture site. A hematoma was defined as any swelling, palpable mass, or newly formed bruit, a sound heard during auscultation suggesting nonlaminar flow through the femoral artery. 4 Inadequate distal blood flow was defined as the absence of distal foot pulses. Randomized controlled trials were of primary interest because these studies provide the highest level of recognized evidence. 19,20 For other studies, such as uncontrolled clinical trials and descriptive studies, that met the inclusion criteria, we incorporated the results 156 AMERICAN JOURNAL OF CRITICAL CARE, March 2002, Volume 11, No. 2

3 of the studies in a narrative review. Although the level of evidence in these studies is considered less rigorous than that of randomized clinical trials, 20 the inclusion of the narrative summary will help identify current approaches used to attain hemostasis after femoral sheath removal. In the literature search, we sought to detect both published and unpublished studies. Although the unpublished studies we selected may not have been peer reviewed, the rigorous requirements for data inclusion should have reduced the possible negative effects of the inclusion of unpublished work. Databases searched for published studies included MEDLINE (1966 through June 1999), CINAHL (1982 through January 1999), HEALTHSTAR ( ), EBM Reviews-Evidence Based Database (1991 through June 1999), Current Contents ( ), EMBASE, DARE, and the Cochrane Library (1999, issue 1). Dissertation Abstracts International database ( ), Proceedings First Database ( ), and hand searching of cardiac conference proceedings were used to detect unpublished research. A search of MEDLINE, CINAHL, Current Contents, and the Cochrane Library for the period June 1999, when the systematic review was done, through September 2001 did not reveal any other published studies comparing manual and mechanical compression techniques. A 2-step search process based on the searching strategy outlined by Dickersin et al 19 was used. Key terms used to frame the search process were cardiac catheterization, femoral sheath removal, mechanical compression, manual compression, device, pneumatic clamp, hemostatic techniques, hemostasis, and bleeding. The reference lists and/or bibliographies of all retrieved articles were checked for additional studies. Assessment of Methodological Quality The methodological quality of studies that satisfied the inclusion criteria were critically appraised. A checklist based on the work of the Cochrane Collaboration 21 and the Centre for Reviews and Dissemination at the University of York 17 was developed to critically appraise the research (Table 1). The checklist was pilot tested on 5 research articles by one of us (T. J.), who had no previous experience in the critical appraisal used in systematic reviews, and by another person who did have such experience. The findings of the 2 reviewers were identical for all 5 articles. Randomized controlled studies were assessed for inclusion in a meta-analysis. The narrative summary Table 1 Randomized controlled trial critical appraisal form Author Year Record number Questions 1 to 5 must be answered yes for the study to be included in the meta-analysis. If any questions are answered no, the study may then be considered in the systematic review as a narrative summary. Were the participants randomly allocated to the study groups? Other than the research intervention, were participants in each group treated the same? Were the study outcomes measured in the same manner for all participants, regardless of the treatment groups? Were the study groups comparable at entry? Was there adequate follow-up of participants? (ie, was the attrition rate explained?) (>20% not followed up) Was allocation of participants to the treatment groups concealed from the allocator? Were those assessing the objective study outcomes blinded to treatment allocation? SUMMARY Decision Use in the meta-analysis Include as a narrative summary Reject from the systematic review Level of evidence Comments included research in which other methods, such as uncontrolled clinical trials and descriptive studies, were used. Studies were categorized according to the strength of their evidence by using a scale published by the Australian Quality of Care and Health Outcomes Committee 20 (Table 2). The quality of each study varied with respect to bias and error. Studies were excluded from meta-analysis if they had an inadequately defined randomization technique (selection bias); apart from the study intervention, had a difference in care within the study groups (performance bias); had different treatment groups because participants withdrew or dropped out from the study (attrition bias); or had different outcome assessment measures (detection bias). AMERICAN JOURNAL OF CRITICAL CARE, March 2002, Volume 11, No

4 Table 2 Randomized categories used to determine levels of evidence 20 I II Level III.1 III.2 III.3 IV Description of category Evidence obtained from a systematic review of all relevant randomized controlled trials Evidence obtained from at least one properly designed randomized controlled trial. Evidence obtained from well-designed controlled trials without randomization Evidence obtained from well-designed cohort or case control analytic studies, preferably from more than a single center or research group Evidence obtained from multiple time series with or without the intervention Dramatic results in uncontrolled experiments Opinion of respected authorities, based on clinical experience, descriptive studies, or reports of expert committees Data Extraction The components of data extracted from each study included in the meta-analysis are listed in Table 3. Data were extracted by using a data extraction tool. The tool was developed and tested for reliability by the Joanna Briggs Institute for Evidence-Based Nursing and Midwifery (Adelaide, Australia). The tool consists of 3 sections; each section pertains to a specific area of research: methodology, methods, and analysis. In order to add rigor to the data extraction, data were extracted independently by one of us (T. J.) and by another reviewer in a small sample of studies. When the data were compared between the 2 reviewers for 5 articles that met the inclusion criteria, the findings were identical. Data Analysis When sufficient data were available, the Peto odds ratio (for categorical outcome data) or standardized mean differences (for continuous outcome data) and the respective 95% CIs were calculated for each study. When statistical pooling of results was not appropriate or possible, the findings were summarized in narrative form. The Review Manager software program (RevMan 4.0), developed by the Cochrane Collaboration, was used to graphically present these results and pool statistical data when appropriate in a meta-analysis. For odds ratios (categorical data), CIs that did not cross 1.0 were considered statistically significant; for mean differences (continuous data), CIs that did not cross zero were statistically significant. Results A total of 53 articles met the initial inclusion criteria. Of these, 12 were randomized controlled trials or descriptive cohort studies and were of acceptable methodological quality for inclusion in the systematic review. These 12 studies are listed chronologically in Table 4, which outlines the study design, study interventions, and outcome measures. A total of 41 articles were excluded from the systematic review because they were overviews or reports of various hemostatic techniques, did not evaluate mechanical compression devices, were single or case series reports, had deficiencies in study design, or had insufficient reporting of the details of the study design. Although meta-analysis was initially confined to studies of the same intervention that measured conceptually similar outcomes, if sufficient data had been available from studies of different types of mechanical devices, the data would have been synthesized in a meta-analysis. Consequently, meta-analysis was limited because of inadequate reporting of results or missing data. For this reason, much of our results are in narrative form. We acknowledge that for the purpose of meta-analysis, a reviewer would typically contact the investigator of the original study if data were insufficiently reported or missing; however, financial and time constraints meant that this step was not possible. Table 3 Components of data extraction General demographic details of the study participants A description of the study institution(s) Diagnostic and/or interventional treatment group(s) The mechanical device being compared and its method of application Outcome measures such as Bleeding Hematoma formation Ecchymosis Pseudoaneurysm Arteriovenous fistula formation Patients discomfort associated with the compression technique Time to effect hemostasis Study design, including method of randomization, allocation, and blinding Sample size Results: dichotomous (numbers of patients in treatment and control groups) and continuous (mean and SD) data Narrative data, such as expert opinion or anecdotal evidence 158 AMERICAN JOURNAL OF CRITICAL CARE, March 2002, Volume 11, No. 2

5 Table 4 Studies included in the systematic review Reference (evidence level) Study design Sample Interventions Outcome(s) Bogart (II) 503 patients: 332 men, 171 women; Group A: n=173; group B: n=168; group C: n=162 Group A: manual compression with a pressure dressing (a roll of 4 x 4 gauze sponges) Group B: mechanical compression with the Compressar clamp; group C: manual compression with stasis button and 0.9-kg weight after hemostasis Time to effect hemostasis, bleeding Gibson (II) 45 patients: 31 men; 14 women; Group A: n=23; group B: n=22 Mechanical compression clamp for at least 20 minutes the FemoStop device for at least 20 minutes Arteriovenous fistula, hematoma, pseudoaneurysm, patient s comfort, thromboemboli, time to effect hemostasis, vasovagal reaction Homuth (II) 51 patients: 40 men, 11 women; Group A: n=26; group B: n=25 the Compressar clamp Hematoma, rebleeding, patient s comfort Janerot- Sjoberg et al (II) 979 patients; Group A: n=433, 75% men; group B: n=546, 71% men Mechanical compression with the ClampEase device for at least 15 minutes the FemoStop device for at least 2 hours Pseudoaneurysm, hematoma or any other complications Lehmann et al (II) 397 patients: 389 men, 8 women; Group A: n 107; group B: n=105; group C: n=96; group D: n=107 Group A: mechanical compression with the HOLD device; group B: 4.5-kg sandbag placed directly over the site Group C: pressure dressing (six 10 x 10-cm sterile gauze sponges); group D: no compression technique Hematoma, ecchymosis, rebleeding, patient s discomfort Nordrehaug et al (II) Multicenter 213 patients: 167 men, 46 women; Group A: n=109; group B: n=104 for at least 8 minutes the FemoStop device for at least 1 hour Hematoma, arteriovenous fistula, pseudo-aneurysm Pracyk (II) 778 patients: 516 men, 262 women; Group A: n=396; group B: n=382 for at least 15 minutes the ClampEase device for at least 15 minutes Any groin complication Continues Bleeding Data on bleeding could be synthesized from the studies by Nordrehaug et al 12 and Pracyk et al 4 because the outcome measures in these studies were homogeneous. The meta-analysis of data collected from these studies (Figure 1) indicated that bleeding from the femoral puncture site after femoral sheath removal did not differ significantly when either a mechanical compression device or manual compression was used to attain hemostasis. Other studies 2,22-24,26 included bleeding as an outcome measure. The prevalence of bleeding after femoral sheath removal did not differ significantly between any interventions in these studies. Hematoma Formation Outcome data on hematoma formation after femoral sheath removal were extracted from the randomized controlled trials reported by Pracyk et al 4 and Homuth. 24 When the data were pooled from these studies, the meta-analysis for hematoma formation indicated that the mechanical compression technique was the most effective; no incidents of hematoma formation occurred in either study (Figure 2). AMERICAN JOURNAL OF CRITICAL CARE, March 2002, Volume 11, No

6 Table 4 Continued Reference (evidence level) Study design Sample Interventions Outcome(s) Rudisill et al (II) 32 patients; Group A: n=16; group B: n=16 PTCA patients: manual compression for 20 minutes; direct coronary atherectomy and rotoblator PTCA patients: manual compression and mechanical compression for 1 hour PTCA patients: mechanical compression for 40 minutes; direct coronary atherectomy and rotoblator PTCA patients: mechanical compression for 1 hour 20 minutes Any groin complication; no specific outcomes stated Semler (III.1) Multicenter non patients; Group A: n=2250; group B: n=1005 Mechanical compression with the Compressar clamp Mean compression time, hematoma and lower extremity ischemic signs and symptoms or complications Simon (IV) Descriptive study 200 patients; Group A: n=100, retrospective; group B: n=100, prospective Mechanical compression with the ClampEase device for at least 15 minutes Any hemostatic complication, including hematoma, arterial occlusion, ischemia, and traumatic neuropathy Simon et al (III.1) Non- 720 patients: 478 men, 242 women; Group A: n=377; group B: n=377 for at least 10 minutes Mechanical compression with the Compressar clamp for at least 10 minutes Bleeding, bruising, patient s discomfort, activity level Sridhar et al (IV) Descriptive, retro-spective study 99 patients: 84 men, 15 women; Group A: n=58; group B: n=41 PTCA indicates transluminal percutaneous angioplasty;, randomized controlled trial. Mechanical compression with the FemoStop device for at least 30 minutes Hematoma, arteriovenous fistula, pseudoaneurysm As with bleeding, hematoma formation was a widely reported clinical outcome. We did find some variation in the prevalence of hematoma formation, 4,12,24,28 but the prevalence was low, and in one study, 28 formation of hematoma occurred significantly more often in the manual compression group than in the group in which a mechanical device was used. Complications of Femoral Artery Sheath Removal Ecchymosis was reported as a clinical outcome in only 2 studies. 2,4 The prevalence of ecchymosis did not differ between study groups. The femoral puncture site was assessed for clinical evidence of a pulsatile mass in only 1 study. 4 The prevalence varied between the studies. Pseudoaneurysm is a more serious vascular complication; it is often difficult to recognize and may be masked by the presence of a hematoma. 8 Pseudoaneurysm was a clinical outcome in 4 studies. 2,4,12,25 Although the prevalence of pseudoaneurysm formation varied between studies, when manual compression was compared with the FemoStop mechanical device, the prevalence was significantly higher with manual compression. 28 However, data were collected from the manual compression intervention group retrospectively in this study and were accumulated during an unknown period; we therefore considered the study poorly controlled. When 2 mechanical compression devices were compared by Janerot-Sjoberg et al, 25 pseudoaneurysm, detected by Doppler examination, occurred in the ClampEase compression group significantly more often than in the FemoStop compression group. 160 AMERICAN JOURNAL OF CRITICAL CARE, March 2002, Volume 11, No. 2

7 Comparison: 01 versus mechanical compression devices Outcome: 01 Bleeding Manual Mechanical Peto odds ratio Weight Peto odds ratio Study n/n n/n (95% CI fixed) % (95% CI fixed) Year Nordrehaug et al 12 9/109 7/ [ ] 1996 Pracyk et al 4 9/307 10/ [ ] 1998 Total (95%CI) 18/416 17/ [ ] χ (df = 1) P =.85 Z = 0.01 P = Favors manual Favors mechanical Figure 1 Manual versus mechanical compression: bleeding The prevalence of arteriovenous fistula formation after femoral sheath removal was reported in only 2 studies. 4,28 The prevalence was not significantly different in either study. Time to Effect Hemostasis Factors differed between studies for the time required to attain hemostasis after application of the hemostatic technique. In 3 studies, 22,24,27 the Compressar device took significantly longer to effect hemostasis than did manual compression. Only Semler 10 found that manual compression required more time than did the Compressar device. Meta-analysis was not possible because of insufficient data. Gibson 23 reported that significantly more time was required to attain hemostasis with the FemoStop than with the ClampEase device. However, in this small sample, a few patients might have skewed the data in favor of the ClampEase device, and so this finding should be interpreted with caution. No significant difference was found in the time required to effect hemostasis after sandbag pressure, pressure dressing, or no compression when these 3 methods were compared with the HOLD compression device. 2 Implications for Nursing Practice and Research Removal of the femoral sheath after cardiac interventional procedures can be a time-consuming procedure. Although manual compression has been the favored method of attaining hemostasis after removal of the sheath, alternative techniques, such as mechanical compression devices, are being used more often in clinical practice. From a clinical point of view, the development of protocols and procedures for femoral sheath removal must be based on evidence that indicates which practices are the most efficient. Although we found articles on primary research related to attaining hemostasis after femoral sheath removal, several of the articles did not address mechanical compression techniques specifically Comparison: 01 versus mechanical compression devices Outcome: 02 Hematoma formation Manual Mechanical Peto odds ratio Weight Peto odds ratio Study n/n n/n (95% CI fixed) % (95% CI fixed) Year Homuth 24 3/26 0/ [ ] 1990 Pracyk et al 4 4/307 0/ [ ] 1998 Total (95%CI) 7/333 0/ [ ] χ (df = 1) P =1 Z = 2.59 P = Favors manual Favors mechanical Figure 2 Manual versus mechanical compression: hematoma formation AMERICAN JOURNAL OF CRITICAL CARE, March 2002, Volume 11, No

8 or contained insufficient primary research data for inclusion. The limited depth and breath of our review are due to the quality of the research findings available. Although Sridhar et al 28 reported a significant difference between manual compression and a mechanical device in the prevalence of hematoma and pseudoaneurysm formation, their findings should be cautiously interpreted because of the poorly controlled retrospective study design. In the study by Janerot-Sjoberg et al, 25 the prevalence of pseudoaneurysm formation was lower with the FemoStop device than with the ClampEase device. This finding is important, because this large randomized controlled trial was well designed, and the formation of pseudoaneurysms was detected by using Doppler examination. In those studies in which the time required to effect hemostasis was measured, some significant differences between study interventions occurred, but the conflicting findings make it difficult to draw clear conclusions about this clinical outcome. When these findings are considered, it is not possible to make clear recommendations that a particular mechanical device is more effective than another mechanical device. However, the meta-analysis that indicated no difference between techniques in regard to bleeding and a statistically reduced prevalence of hematoma formation after mechanical compression make it possible to make a clinical recommendation for the use of mechanical devices. The few nursing studies related to techniques used to attain hemostasis after femoral sheath removal highlight the need for further studies on this topic. We recommend that more benchmark randomized controlled trials be undertaken. The protocols for these trials should include a rigorous randomization process, inclusion criteria, and establishment of the power of the sample. These requirements may be challenging because the low frequency of complications may necessitate trials with large numbers of participants and careful control strategies to accommodate the variability between manual compression skills and application techniques for various compression devices. Conclusion In this systematic review, we used a rigorous preplanned process to detect primary research on mechanical compression devices used to attain hemostasis after femoral sheath removal after cardiac interventional procedures. Our results highlight the lack of quality research on this topic. Nevertheless, results do highlight particular compression techniques used to attain hemostasis after femoral sheath removal. This information may help cardiac nurses make informed decisions about particular devices and techniques used in clinical practice. REFERENCES 1. Schickel S, Cronin SN, Mize A, Voelker C. Removal of femoral sheaths by registered nurses: issues and outcomes. Crit Care Nurse. April 1996;16: Lehmann K, Ferris S, Heath Lange S. Maintenance of hemostasis after invasive cardiac procedures: implications for outpatient catheterization. J Am Coll Cardiol. 1997;30: Peet GI, McGrath MA, Brunt JH, Hilton JD. Femoral arterial sheath removal after PTCA: a cross-canada survey. Can J Cardiovasc Nurs. 1995;6: Pracyk JB, Wall TC, Longabaugh JP, et al. A randomized trial of vascular hemostasis techniques to reduce femoral vascular complications after coronary intervention. Am J Cardiol. 1998;81: Gibbs H, Sanborn T. Percutaneous vascular hemostasis devices for arterial sealing after interventional procedures. In: Topol E, ed. Textbook of Interventional Cardiology. Philadelphia, Pa: WB Saunders Co; 1994: Simon AW. Use of a mechanical pressure device for hemostasis following cardiac catheterization. Am J Crit Care. 1994;3: Spies JB, Berlin L. Complications of femoral artery puncture. AJR Am J Roentgenol. 1998;170: Jones C, Holcomb E, Rohrer T. Femoral artery pseudoaneurysm after invasive procedures. Crit Care Nurse. August 1995;15: Kussmaul WG III, Buchbinder M, Whitlow PL, et al. Rapid arterial hemostasis and decreased access site complications after cardiac catheterization and angioplasty: results of a randomized trial of a novel hemostatic device. J Am Coll Cardiol. 1995;25: Semler HJ. Transfemoral catheterization: mechanical versus manual control of bleeding. Radiology. 1985;154: Beyer-Enke SA, Soldner J, Zeitler E. Immediate sealing of arterial puncture site following femoropopliteal angioplasty: a prospective randomized trial. Cardiovasc Intervent Radiol. 1996;19: Nordrehaug JE, Chronos NAF, Priestley KA, Buller NP, Foran J. Randomized evaluation of an inflatable femoral artery compression device. J Intervent Cardiol. 1996;9: Ernst J, Kloos R, Schrader R, et al. Immediate sealing of arterial puncture sites after catheterization and PTCA using a vascular hemostasis device with collagen: an international registry [abstract]. Circulation. 1991;84(suppl II): Gibbs J, Slade A, Nordrehaug J, et al. The use of a collagen plug for femoral arterial haemostasis following cardiac catheterization: single institution experience [abstract]. Eur Heart J. 1992;13(suppl): Ernst S, Gin M, Mast G, et al. Percutaneous application of collagen seal femoral arterial puncture sites in fully anticoagulated patients. J Am Coll Cardiol. 1992;19(suppl A):217A. 16. Cook DJ, Mulrow CD, Hayes RB. Systematic reviews: synthesis of best evidence for clinical decisions. Ann Intern Med. 1997;126: Deeks J, Glanville J, Sheldon T. Undertaking Systematic Reviews of Research on Effectiveness: CRD Guidelines for Carrying Out or Commissioning Reviews. York, United Kingdom: National Health System Centre for Reviews and Dissemination, The University of York, York Publishing Service Ltd; CDS Report No Jones T, Evans D. Conducting a systematic review. Aust Crit Care. May 2000;13: Dickersin K, Scherer R, Lefebvre C. Identifying relevant studies for systematic reviews. BMJ. 1994;309: Quality of Care and Health Outcomes Committee. Guidelines for the Development and Implementation of Clinical Guidelines. Canberra, Australia: Australian Government Publishing Service; Oxman A. The Cochrane Collaboration Handbook. Oxford, England: The Cochrane Collaboration; Bogart M. Time to hemostasis: a comparison of manual versus mechanical compression of the femoral artery. Am J Crit Care. 1995;4: Gibson C. A Comparison of Compression Devices for Femoral Artery Haemostasis [master s thesis]. Spokane, Wash: Gonzaga University; Homuth G. Relationship of Methods for Removal of Femoral Sheaths to the Incidence of Hematomas and Patient Comfort [master s thesis]. De Kalb, Ill: Northern Illinois University; Janerot-Sjoberg B, Broqvist M, Fransson SG. Femoral artery haemostasis with a pneumatic compression device versus a clamp after coronary angiography. Scand Cardiovasc J. 1998;32: Rudisill PT, Williams LB, Craig S, Schopp P. Study of mechanical versus manual/mechanical compression following various interventional cardiology procedures. J Cardiovasc Nurs. January 1997;11: Simon A, Bumgarner B, Clark K, Israel S. Manual versus mechanical compression for femoral artery hemostasis after cardiac catheterization. Am J Crit Care. 1998;7: Sridhar K, Fischman D, Goldberg S, et al. Peripheral vascular complications after intracoronary stent placement: prevention by use of a pneumatic vascular compression device. Cathet Cardiovasc Diagn. 1996;39: AMERICAN JOURNAL OF CRITICAL CARE, March 2002, Volume 11, No. 2