- Email: [email protected]
Collagen- versus gelatine-coated Dacron versus stretch PTFE bifurcation grafts for aortoiliac occlusive disease: Long-term results of a prospective, randomized multicenter trial
Collagen- versus gelatine-coated Dacron versus stretch PTFE bifurcation grafts for aortoiliac occlusive disease: Long-term results of a prospective, randomized multicenter trial Manfred R. Prager, MD, Thomas Hoblaj, MD, Josif Nanobashvili, MD, Emanuel Sporn, MD, Peter Polterauer, MD, Otto Wagner, MD, Hans-Jörg Böhmig, MD, Harald Teufelsbauer, MD, Meinrad Ploner, PhD, and Ihor Huk, MD, Vienna, Austria
Background. In this prospective randomized multicenter trial, knitted gelatine-coated Dacron, knitted collagen-coated Dacron, and stretch polytetrafluoroethylene (PTFE) aortic bifurcation grafts were compared for their long-term results. Methods. Between 1991 and 1998, 149 patients undergoing elective revascularization for aortoiliac occlusive disease were prospectively randomized at 3 tertiary referral centers of vascular surgery. The patients received either gelatine-coated Dacron (GEL-D) grafts (n = 52), collagen-coated Dacron (COLD) grafts (n = 49), or stretch PTFE grafts (n = 48). Results. No intraoperative deaths were recorded. The 30-day mortality was 4%. The mean follow-up time was 97 months. Primary patency rates were 77% for GEL-D, 78% for COL-D, and 79% for PTFE at 8 years. The differences were not different (P > .8). Secondary corrected 8-year patency rates were also not significantly different (P > .5): 91% for GEL-D, 96% for COL-Dm and 90% for PTFE. Five Dacron and 1 PTFE grafts were affected by infections. Conclusions. Bifurcation grafts for revascularization of aortoiliac occlusive disease using these 3 materials were comparable in terms of primary and secondary patency and long-term complication rates. (Surgery 2003;134:80-5.) From the University of Vienna Medical School and the Ludwig Boltzmann Research Institute for Interdisciplinary Vascular Medicine, Vienna, Austria
THE HISTORY OF REPLACING THE AORTOILIAC VESSELS with synthetic grafts spans almost 50 years and began with Charles Dubost1 and Michael DeBakey. Dubost was the first to replace an abdominal aortic aneurysm with a homograft. When DeBakey first grafted the aortoiliac vessels some time later, he called the procedure Dubost’s operation.2,3 The original grafts were of Vinyon-N, a nylon fabric used for parachutes.4,5 Few studies compared bifurcation grafts currently in use. Those studies that are available are of limited value because they were either retrospective or nonrandomized, or their Accepted for publication March 20, 2003. Reprint Requests: Manfred R. Prager, MD, Department of Vascular Surgery, University of Vienna Medical School-AKH, Waehringer Guertel 18-20, Wien-Vienna A-1090, Austria. © 2003 Mosby, Inc. All rights reserved. 0039-6060/2003/$30.00 + 0 doi:10:1067/msy.2003.179
80 SURGERY
follow-up time was too short. Prospective randomized trials comparing long-term patency of bifurcation grafts for aortoiliac occlusive disease by statistically valid methods are scarce. A PubMed search, which was conducted until July 2002 to identify relevant prospective randomized trials comparing Dacron and polytetrafluoroethylene (PTFE) bifurcation grafts for aortoiliac occlusive disease, produced no more than 4 statistically valid trials.6-9 In the present study, the long-term results of a series of patients with aortoiliac occlusive disease are presented. Immediate, early, and intermediaterange data collected showed the 3 material classes tested to have comparable primary and secondary patency rates and similar complications rates.9 PATIENTS AND METHOD Between September 1991 and December 1998 in a multicenter trial, 315 elective patients were
Prager et al 81
Surgery Volume 134, Number 1
randomized prospectively for bifurcation grafting according to Pocock and Simon.10 The patients received either a knitted gelatine-coated doublevelour Dacron graft (Uni-Graft DV; B. Braun Melsungen AG, Melsungen, Germany), a knitted collagen-coated double-velour Dacron graft (Hemashield Microvel; Meadox Medicals Inc, Oakland, NJ), or a stretch PTFE graft (Stretch GoreTex; Gore-Tex, W L Gore & Associates Inc, Elkton, Md, subsequently called PTFE) at one of the three cooperating vascular surgery centers. In Austria, in 1999, a GEL-D bifurcation graft cost (in United States dollars) $607, a COL-D graft $610, and a PTFE graft $789. All patients on record in the prospective randomized data bank who underwent surgery for aortoiliac occlusive disease, (n = 149, 104 [70%] males, 45 [30%] females) were included in this extension study. Their median age was 59 years (range, 3684).They were operated for lifestyle-limiting claudication (14%), disabling claudication (53%), rest pain (15%), and gangrene (18%). Of the grafts, 52 were GEL-D, 49 were COL-D, and 48 were PTFE. The patients were well matched for the risk factors used for randomization (Table I). All operations were performed under a general anesthesia. Antibiotics (amoxicillin and clavulanic acid) were administered from the initiation of surgery and continued for 72 hours. Before aortic cross-clamping, systemic heparin (70 IU/kg body weight) was administered intravenously without routine reversal by protamine. Postoperatively, fraxiparine, 100 mg/kg once daily, was administered subcutaneously until discharge. Patients with anastomoses on the deep femoral arteries received 100 mg/kg fraxiparine twice a day. Grafts were not soaked in antibiotic solutions before implantation. Overall, 88% (n = 131) of the patients were smokers, and 22% (n = 33) had diabetes mellitus. One third of the diabetics (n = 11) were on a dietary regimen, another third received oral antidiabetics (n = 11), and the remaining third needed insulin for glycemic control (n = 11). Run-off was good in 75 (50%) patients and poor in 74 (50%). It was thought to be poor if one superficial femoral (profunda) artery was occluded so that at least one distal anastomosis had to be made to the deep femoral artery. All proximal anastomoses were sideto-end. Of the 27 patients with ischemic trophic lesions, 22 (15% of 149) presented with infected gangrene of the lower extremity confirmed by physical findings and microbiological evidence of infection. Access was transperitoneal in 133 patients (90%) and retroperitoneal in 15 (10%). Concomitant renal artery repair was performed in
Table I. Distribution of randomization criteria (age, gender, diabetes, smoking, and run-off) in 149 patients receiving GEL-D, COL-D, or PTFE aortic bifurcation grafts according to the random plan of Pocock et al 10 Risk factors Age >65 (%) Female (%) Diabetes (%) Smoking (%) Poor run-off (%)
PTFE
GEL-D
COL-D
Chi-square P
23
33
29
0.5
25 23 92 46
33 23 83 52
33 20 90 51
0.9 0.9 0.3 0.8
PTFE, polytetrafluoroethylene
7 (5%) patients. The median aortic cross-clamp time was 31 minutes (range, 9-59). Of these procedures, 128 (81%) were performed by experienced senior residents and 28 (19%) by junior residents supervised by an attending surgeon. The graft diameter was 20/10 mm in 1%, 18/9 mm in 22%, 16/8 mm in 57%, and 14/7 mm in 20%. The median follow-up time was 97 months. Follow-up studies at 6 monthly intervals included ultrasonography of the graft. If this was inconclusive or unclear, angiography or computed tomography (CT) was performed. Survival and patency were estimated with the Kaplan-Meier method11 with statistical analysis computed with the Breslow test.12 Patencies were calculated as recommended by the Vascular Surgery/International Society for Cardiovascular Surgery ad hoc Committee on Reporting Standards.13 Data were stored electronically and processed on an IBM computer (IBM Instruments, Inc, Danbury, Conn) with access to BMDP and SAS software packages for statistical analysis. Dichotomous response variables were tested with Fisher’s exact test and the likelihood ratio test for significance, the latter test being less conservative for rare events. RESULTS Mortality and survival. Thirty-day mortality was 4% (6/149). Two patients died of myocardial infarction on day 10 postoperatively, 1 to a fatal stroke on day 11, 2 to pulmonary failure on days 8 and 21, and 1 of multiorgan failure after aspiration pneumonia on day 14. The median overall survival after surgery was 96 months. No differences were found in median survival across the 3 material groups. Patency. Overall patency. Overall (n = 149) the primary 8-year patency rate was 79%. The secondary 8-year patency rate was 92% (Fig 1).
82 Prager et al
Surgery July 2003
Fig 1. Life table analysis of overall primary and secondary graft patency, estimated according to the KaplanMeier method.
Patency by materials. Primary patency rates were 92% for GEL-D, 89% for COL-D, and 88% for PTFE at 5 years and 77%, 78%, and 79% at 8 years (Fig 2). These differences were not significant (P > .8). Secondary corrected 8-year patency rates (Fig 3) were also not different (P > .5): 91% for GEL-D, 96% for COL-D, and 90% for PTFE. Of a total of 21 graft limb occlusions, 8 could not be repaired, and cross-over bypasses or redo bifurcation grafting (2 patients) were required. In 5 patients, thrombectomy failed. Three graft limbs had to be explanted due to graft infection. In one patient in the PTFE group, a purulent graft infection occurred 110 months after the initial procedure in conjunction with perforated sigmoid diverticulitis, which necessitated graft limb explantation and replacement of the total graft by a homologous aortic bifurcation. Another patient with infection had graft limb explantation, intraoperative angioplasty, and profundoplasty, and one had explantation followed by cross-over bypass 8 weeks later. Factors underlying late graft limb occlusion were progression of the underlying disease in 14 patients. In 2, the cause was unidentifiable. In another 2, technical factors were responsible, that is, graft limb kinking and superficial femoral artery dissection at the distal anastomosis. In three patients, infection caused the graft occlusion. A breakdown by the graft materials used showed seven late occlusions in the GEL-D group (5%), six in the COL-D group (4%), and eight in PTFE patients (5%) (log-rank P = .99).
Patency and poor run-off. Subgroup analysis of patients with poor run-off (n = 74 overall; GEL-D n = 27, COL-D n = 26, PTFE n = 23) failed to show any difference between the 3 material groups (log-rank P = .8). Of the 10 occlusions (GEL-D n = 3, COL-D n = 3, PTFE n = 4), 4 were successfully thrombectomized. Graft infection. The severity of graft infection was classified according to Szilagyi et al.14 Graft infections (Szilagyi stage III) were observed in 6 patients (4%), all but 1 in Dacron grafts (COL-D, n = 3 and GEL-D, n = 2, PTFE, n = 1). The time to the onset of infection necessitating revision was 2 weeks to 25 months after implantation (median, 4.5 months). In 2 patients graft patency was restored by debridement or a sartorius muscle flap. In another 2 patients, one of the limbs of the graft had to be sacrificed and was replaced later with eventual limb salvage. One patient ultimately required amputation. The one PTFE graft infection seen occurred 110 months after implantation as the result of a perforated sigmoid diverticulitis with an associated abscess. Through a transabdominal approach, the graft was removed, an aortic bifurcation homograft was implanted, and the sigmoid colon was resected. The patient finally succumbed to multiorgan failure on day 21 after the initial revision. Two patients developed infections during the initial hospitalization for graft placement. One developed an extensive subcutaneous tissue necrosis (GEL-D), the other initially had an inguinal wound seroma (COL-D) as a possible source of infection. In
Surgery Volume 134, Number 1
Prager et al 83
Fig 2. Life table analysis of primary patency: GEL-D (gelatine-coated Dacron) versus COLD-D (collagen-coated Dacron) versus stretch II PTFE.
the remaining 4 patients, infections occurred after discharge 5 to 110 months after graft implantation. All patients were claudicants, but none had lower limb infection or gangrene at the time of surgery. Four patients were diabetics. In 4 of the patients with graft limb infection, the graft was anastomosed to the deep femoral arteries. Comparison of the 5 infected Dacron grafts with the single infected PTFE graft did not show any significant difference (P > .05). Late morphology. False aneurysms. One patient was reoperated because of a false aneurysm at the distal anastomoses (PTFE). The underlying cause was believed to be the result of the endarterectomy at the time of bifurcation grafting. Duplex ultrasonography of the graft limbs and body. One hundred thirty-two patients (89%) were available for follow-up studies, which included 6 monthly studies of the suture line and the graft bodies and limbs. These studies did not show major graft dilatation or aneurysms. DISCUSSION The present study is based on data from a prospective, randomized multicenter trial. The early and midterm results of this trial after a median follow-up time of 42 months were published elsewhere.9 The bifurcation grafts of all 3 materials
were comparable in primary and secondary patency rates, incidence of false aneurysms, and rate of perioperative complications. To strengthen the validity of the conclusions drawn from this previous study,9 patients with aortoiliac occlusive disease from this trial were selected for further analysis because it was felt that a homogeneous patient population well matched for the primary disease would be essential for comparing long-term patencies. So far, no other prospective randomized study has compared bifurcation grafts of different materials—that is, 2 Dacron grafts and a stretch PTFE graft—in patients with aortoiliac occlusive disease after a median follow-up time of 97 months. But long follow-up times are needed to detect shortcomings inherent in specific materials.15 Currently only 4 statistically valid studies comparing Dacron and PTFE are available.6-9 In 1988, Lord et al6 compared Dacron with conventional PTFE, both tube and bifurcation grafts, in patients with abdominal aortic aneurysms and aortoiliac occlusive disease. Lord found no difference between the 2 materials in perioperative mortality, morbidity, blood loss, cross-clamping time, and operating time. Long-term results have so far not been reported. In a prospective, randomized trial of 60 patients, Friedman et al7 compared the long-
84 Prager et al
Surgery July 2003
Fig 3. Life table analysis of secondary patency: GEL-D (gelatine-coated Dacron) vs COLD-D (collagen-coated Dacron) vs stretch II PTFE.
term results of PTFE and Dacron grafts for aortoiliac occlusive disease at a mean follow-up time of 57 months. These authors did not find any differences in early postoperative morbidity, mortality, and long-term results. In another prospective, randomized trial of 165 patients, Polterauer et al8 found the 2 graft materials to be comparable in secondary patency rates, but conventional PTFE—the thickwalled precursor of the current stretch PTFE—had a higher rate of complications. In 1987, Burke et al16 published a nonrandomized comparison of Dacron and PTFE in 42 patients with hypoplastic aortas. Patency rates 4 years postoperatively were 95% for PTFE and 52% for Dacron. In our study, none of the patients had hypoplastic vessels and needed grafts smaller than 14/7 mm. Several nonrandomized, retrospective studies argued in favor of PTFE for aortoiliac repair. Extensive experience with conventional PTFE17-23 led to the development of stretch PTFE, which is thought to have hemodynamic advantages due to the omission of the flow divider24 and better compliance because of the reduced wall thickness.24 In the present long-term trial with a median follow-up time of 97 months, no difference was seen in primary and secondary patency rates. This contrasts
with experimental data obtained with 111In-labeled thrombocytes25,26 and complement activation, which suggested that PTFE was less thrombogenic than Dacron27 because of its electronegative and hydrophobic surface.28 The 4% rate of graft infections in our study is comparatively high. In large series of aortic grafts, the risk of graft infection was reported to be between 0.7% and 2.0%.14,29 In our study, 5 of 6 patients developed graft infections within the first 2 years after implantation (2 weeks to 25 months; median 4.5 months). These may have been caused by chronic long-term infection secondary to intraoperative contamination/inoculation or by hematogenous spread of microorganisms from a distant source. The infection of the one PTFE graft 110 months after implantation was the result of sigmoid perforation and abscess. The graft infection caused by it in our study cannot necessarily be attributed to the graft material. Comparison of the Dacron grafts with the single infected PTFE graft did not show any significant difference (P > .05). Thus, long-term results presented here fail to underline the trend seen in our earlier trial.9 Today, financial aspects govern medical routine, and cost effectiveness influences our decisions.
Surgery Volume 134, Number 1
PTFE grafts were about one third more expensive in 1999 than GEL-D and COL-D grafts. Their manufacturers have justified the higher price by stressing their superior qualities. Our overall data suggest to us that PTFE grafts, although more expensive, may be justified in patients carrying an increased risk of poor wound-healing and graft infection. In conclusion, in this study of patients with aortoiliac occlusive disease, none of the 3 materials was superior in long-term patency, long-term complications, or risk of infection. This paper is dedicated to our coworker Professor Otto Wagner, a pioneer and teacher of vascular surgery, who died in February 2002. REFERENCES 1. Dubost C. Resection of an aneurysm of the abdominal aorta. Arch Surg 1952; 64:405-8. 2. Kram HB, Dietzek AM, Veith FJ. Optimal synthetic grafts for aortic replacement. In: Greenhalgh RM, ed. The cause and management of aneurysms. Philadelphia: WB Saunders; 1990. p 340-50. 3. Friedman SG. The 50th anniversary of abdominal aortic reconstruction. J Vasc Surg 2001; 33:895-8. 4. Voorhees AB. How it all began. In: Sawyer PN, ed. Vascular Grafts. New York: Appleton-Century-Crofts; 1978. p 3. 5. Voorhees AB, Jaretzki A, Blakemoore AH. The use of tubes constructed from Vinyon N cloth in bridging arterial defects. Ann Surg 1952; 135:332-6. 6. Lord RS, Nash PA, Raj BT, Stary DL, Graham AR, Hill DA, et al. Prospective randomized trial of polytetrafluoroethylene and Dacron aortic prosthesis. I. Perioperative results. Ann Vasc Surg 1988; 2:248-54. 7. Friedman SG, Lazzaro RS, Spier LN, Moccio C, Tortolani AJ. A prospective randomized comparison of Dacron and polytetraflouroethylene aortic bifurcation grafts. Surgery 1995; 117:7-10. 8. Polterauer P, Prager M, Hölzenbein T, Karner J, Kretschmer G, Schemper M. Dacron versus polytetrafluoroethylene for Y-aortic bifurcation grafts: a six-year prospective, randomized trial. Surgery 1992; 111:626-33. 9. Prager M, Polterauer P, Böhmig HJ, Wagner O, Fugl A, Kretschmer G, et al. Collagen versus gelatin-coated Dacron versus stretch polytetrafluoroethylene in abdominal aortic bifurcation graft surgery: results of a seven-year prospective, randomized multicenter trial. Surgery 2001; 130:408-14. 10. Pocock SJ, Simon R. Sequential treatment assignment with balancing for prognostic factors in the controlled clinical trial. Biometrics 1975; 31:103-11. 11. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc 1958; 53:457-65. 12. Breslow A. A generalized Kruskal-Wallis test for comparing k-samples to unequal patterns of sensorship. Biometrika 1970; 57:579-83.
Prager et al 85
13. Rutherford RB, Baker JD, Ernst C, Johnston KW, Porter JM, Ahn S, et al. Recommended standards for reports dealing with lower extremity ischemia: revised version. J Vasc Surg 1997; 26:517-38. 14. Szilagyi DE, Smith RF, Elliott JP, Vrandecic MP. Infection in arterial reconstruction with synthetic grafts. Ann Surg 1972; 176:321-33. 15. Szilagyi DE, Elliott JP Jr, Smith RF, Reddy DJ, McPharlin M. A thirty-year survey of the reconstructive surgical treatment of aortoiliac occlusive disease. J Vasc Surg 1986; 3:421-36. 16. Burke PM, Jr., Herrmann JB, Cutler BS. Optimal grafting methods for the small abdominal aorta. J Cardiovasc Surg (Torino) 1987; 28:420-6. 17. Cintora I, Pearce DE, Cannon JA. A clinical survey of aortobifemoral bypass using two inherently different graft types. Ann Surg 1988; 208:625-30. 18. Avramov S, Petrovic P, Fabri M. Bifurcated grafts (Dacron and PTFE) in aorto-iliac reconstruction: Five year follow-up. J Cardiovasc Surg 1987; 28:33-6. 19. Avramov S, Petrovic P, Plau J, Fabri M. Synthetic vascular grafts (Dacron and PTFE) in the treatment of abdominal aortic aneurysms. Five year follow-up. J Cardiovasc Surg 1988; 29:6-10. 20. Polterauer P, Wagner O, Kretschmer G, Piza F. The PTFE Y graft: one year experience in twenty-one patients. Int J Artif Organs 1982; 5:263-7. 21. Polterauer P, Wagner O, Kretschmer G, Piza F. Arterial grafting with PTFE in the pelvic region. Vasc Surg 1982; 14:390-6. 22. Polterauer P, Kretschmer G, Wagner O, Waneck R, Piza F, Lechner G. Die PTFE Y-Bifurkationsprothese: Frühergebnisse. Chirurg 1984; 55:106-10. 23. Karner J, Polterauer P, Kretschmer G, Piza F, Schemper M. Bifurkationsprothese: Dacron vs Teflon (PTFE). Acta Chir Austriaca 1987; 1:8-10. 24. Prager M, Polterauer P, Huk I, Trubel W, Nanobashvili J, Claeys L. Should Dacron or PTFE be used for aorto-iliac reconstruction ? In: Greenhalgh RM, ed. Trials and tribulations of vascular surgery. London: WB Saunders Company LTD; 1994. p 313-24. 25. Wakefield TW, Shulkin BL, Fellows EP. Platelet reactivity in human aortic grafts: a prospective, randomized midterm study of platelet adherence and release products in Dacron and polytetraflouroethylene conduits. J Vasc Surg 1986; 3:732-6. 26. Dewanjee MK. Cardiac and vascular imaging with labeled platelets and leukocytes. Semin Nucl Med 1984; 14:154-87. 27. Shepard AD, Gelfand JA, Callow AD, O’Donnell TF Jr. Complement activation by synthetic vascular prostheses. J Vasc Surg 1984; 1:829-38. 28. Chiesa R, Melissano G, Castellano R, Astore D, Castrucci M, del Maschio A, et al. A new ePTFE stretch graft for aortoiliac reconstructions. Surgical evaluation and one year follow-up with Magnetic Resonance Imaging. J Cardiovasc Surg 1995; 36:135-41. 29. Goeau-Brissoniere O, Pechere JC, Leport C. Comment prevenir les infections de prothese. In: Kieffer E, editor. Les aneurysmes de láorte abdominale sous-renale. Paris: AERCV; 1990. p 143.
Background. In this prospective randomized multicenter trial, knitted gelatine-coated Dacron, knitted collagen-coated Dacron, and stretch polytetrafluoroethylene (PTFE) aortic bifurcation grafts were compared for their long-term results. Methods. Between 1991 and 1998, 149 patients undergoing elective revascularization for aortoiliac occlusive disease were prospectively randomized at 3 tertiary referral centers of vascular surgery. The patients received either gelatine-coated Dacron (GEL-D) grafts (n = 52), collagen-coated Dacron (COLD) grafts (n = 49), or stretch PTFE grafts (n = 48). Results. No intraoperative deaths were recorded. The 30-day mortality was 4%. The mean follow-up time was 97 months. Primary patency rates were 77% for GEL-D, 78% for COL-D, and 79% for PTFE at 8 years. The differences were not different (P > .8). Secondary corrected 8-year patency rates were also not significantly different (P > .5): 91% for GEL-D, 96% for COL-Dm and 90% for PTFE. Five Dacron and 1 PTFE grafts were affected by infections. Conclusions. Bifurcation grafts for revascularization of aortoiliac occlusive disease using these 3 materials were comparable in terms of primary and secondary patency and long-term complication rates. (Surgery 2003;134:80-5.) From the University of Vienna Medical School and the Ludwig Boltzmann Research Institute for Interdisciplinary Vascular Medicine, Vienna, Austria
THE HISTORY OF REPLACING THE AORTOILIAC VESSELS with synthetic grafts spans almost 50 years and began with Charles Dubost1 and Michael DeBakey. Dubost was the first to replace an abdominal aortic aneurysm with a homograft. When DeBakey first grafted the aortoiliac vessels some time later, he called the procedure Dubost’s operation.2,3 The original grafts were of Vinyon-N, a nylon fabric used for parachutes.4,5 Few studies compared bifurcation grafts currently in use. Those studies that are available are of limited value because they were either retrospective or nonrandomized, or their Accepted for publication March 20, 2003. Reprint Requests: Manfred R. Prager, MD, Department of Vascular Surgery, University of Vienna Medical School-AKH, Waehringer Guertel 18-20, Wien-Vienna A-1090, Austria. © 2003 Mosby, Inc. All rights reserved. 0039-6060/2003/$30.00 + 0 doi:10:1067/msy.2003.179
80 SURGERY
follow-up time was too short. Prospective randomized trials comparing long-term patency of bifurcation grafts for aortoiliac occlusive disease by statistically valid methods are scarce. A PubMed search, which was conducted until July 2002 to identify relevant prospective randomized trials comparing Dacron and polytetrafluoroethylene (PTFE) bifurcation grafts for aortoiliac occlusive disease, produced no more than 4 statistically valid trials.6-9 In the present study, the long-term results of a series of patients with aortoiliac occlusive disease are presented. Immediate, early, and intermediaterange data collected showed the 3 material classes tested to have comparable primary and secondary patency rates and similar complications rates.9 PATIENTS AND METHOD Between September 1991 and December 1998 in a multicenter trial, 315 elective patients were
Prager et al 81
Surgery Volume 134, Number 1
randomized prospectively for bifurcation grafting according to Pocock and Simon.10 The patients received either a knitted gelatine-coated doublevelour Dacron graft (Uni-Graft DV; B. Braun Melsungen AG, Melsungen, Germany), a knitted collagen-coated double-velour Dacron graft (Hemashield Microvel; Meadox Medicals Inc, Oakland, NJ), or a stretch PTFE graft (Stretch GoreTex; Gore-Tex, W L Gore & Associates Inc, Elkton, Md, subsequently called PTFE) at one of the three cooperating vascular surgery centers. In Austria, in 1999, a GEL-D bifurcation graft cost (in United States dollars) $607, a COL-D graft $610, and a PTFE graft $789. All patients on record in the prospective randomized data bank who underwent surgery for aortoiliac occlusive disease, (n = 149, 104 [70%] males, 45 [30%] females) were included in this extension study. Their median age was 59 years (range, 3684).They were operated for lifestyle-limiting claudication (14%), disabling claudication (53%), rest pain (15%), and gangrene (18%). Of the grafts, 52 were GEL-D, 49 were COL-D, and 48 were PTFE. The patients were well matched for the risk factors used for randomization (Table I). All operations were performed under a general anesthesia. Antibiotics (amoxicillin and clavulanic acid) were administered from the initiation of surgery and continued for 72 hours. Before aortic cross-clamping, systemic heparin (70 IU/kg body weight) was administered intravenously without routine reversal by protamine. Postoperatively, fraxiparine, 100 mg/kg once daily, was administered subcutaneously until discharge. Patients with anastomoses on the deep femoral arteries received 100 mg/kg fraxiparine twice a day. Grafts were not soaked in antibiotic solutions before implantation. Overall, 88% (n = 131) of the patients were smokers, and 22% (n = 33) had diabetes mellitus. One third of the diabetics (n = 11) were on a dietary regimen, another third received oral antidiabetics (n = 11), and the remaining third needed insulin for glycemic control (n = 11). Run-off was good in 75 (50%) patients and poor in 74 (50%). It was thought to be poor if one superficial femoral (profunda) artery was occluded so that at least one distal anastomosis had to be made to the deep femoral artery. All proximal anastomoses were sideto-end. Of the 27 patients with ischemic trophic lesions, 22 (15% of 149) presented with infected gangrene of the lower extremity confirmed by physical findings and microbiological evidence of infection. Access was transperitoneal in 133 patients (90%) and retroperitoneal in 15 (10%). Concomitant renal artery repair was performed in
Table I. Distribution of randomization criteria (age, gender, diabetes, smoking, and run-off) in 149 patients receiving GEL-D, COL-D, or PTFE aortic bifurcation grafts according to the random plan of Pocock et al 10 Risk factors Age >65 (%) Female (%) Diabetes (%) Smoking (%) Poor run-off (%)
PTFE
GEL-D
COL-D
Chi-square P
23
33
29
0.5
25 23 92 46
33 23 83 52
33 20 90 51
0.9 0.9 0.3 0.8
PTFE, polytetrafluoroethylene
7 (5%) patients. The median aortic cross-clamp time was 31 minutes (range, 9-59). Of these procedures, 128 (81%) were performed by experienced senior residents and 28 (19%) by junior residents supervised by an attending surgeon. The graft diameter was 20/10 mm in 1%, 18/9 mm in 22%, 16/8 mm in 57%, and 14/7 mm in 20%. The median follow-up time was 97 months. Follow-up studies at 6 monthly intervals included ultrasonography of the graft. If this was inconclusive or unclear, angiography or computed tomography (CT) was performed. Survival and patency were estimated with the Kaplan-Meier method11 with statistical analysis computed with the Breslow test.12 Patencies were calculated as recommended by the Vascular Surgery/International Society for Cardiovascular Surgery ad hoc Committee on Reporting Standards.13 Data were stored electronically and processed on an IBM computer (IBM Instruments, Inc, Danbury, Conn) with access to BMDP and SAS software packages for statistical analysis. Dichotomous response variables were tested with Fisher’s exact test and the likelihood ratio test for significance, the latter test being less conservative for rare events. RESULTS Mortality and survival. Thirty-day mortality was 4% (6/149). Two patients died of myocardial infarction on day 10 postoperatively, 1 to a fatal stroke on day 11, 2 to pulmonary failure on days 8 and 21, and 1 of multiorgan failure after aspiration pneumonia on day 14. The median overall survival after surgery was 96 months. No differences were found in median survival across the 3 material groups. Patency. Overall patency. Overall (n = 149) the primary 8-year patency rate was 79%. The secondary 8-year patency rate was 92% (Fig 1).
82 Prager et al
Surgery July 2003
Fig 1. Life table analysis of overall primary and secondary graft patency, estimated according to the KaplanMeier method.
Patency by materials. Primary patency rates were 92% for GEL-D, 89% for COL-D, and 88% for PTFE at 5 years and 77%, 78%, and 79% at 8 years (Fig 2). These differences were not significant (P > .8). Secondary corrected 8-year patency rates (Fig 3) were also not different (P > .5): 91% for GEL-D, 96% for COL-D, and 90% for PTFE. Of a total of 21 graft limb occlusions, 8 could not be repaired, and cross-over bypasses or redo bifurcation grafting (2 patients) were required. In 5 patients, thrombectomy failed. Three graft limbs had to be explanted due to graft infection. In one patient in the PTFE group, a purulent graft infection occurred 110 months after the initial procedure in conjunction with perforated sigmoid diverticulitis, which necessitated graft limb explantation and replacement of the total graft by a homologous aortic bifurcation. Another patient with infection had graft limb explantation, intraoperative angioplasty, and profundoplasty, and one had explantation followed by cross-over bypass 8 weeks later. Factors underlying late graft limb occlusion were progression of the underlying disease in 14 patients. In 2, the cause was unidentifiable. In another 2, technical factors were responsible, that is, graft limb kinking and superficial femoral artery dissection at the distal anastomosis. In three patients, infection caused the graft occlusion. A breakdown by the graft materials used showed seven late occlusions in the GEL-D group (5%), six in the COL-D group (4%), and eight in PTFE patients (5%) (log-rank P = .99).
Patency and poor run-off. Subgroup analysis of patients with poor run-off (n = 74 overall; GEL-D n = 27, COL-D n = 26, PTFE n = 23) failed to show any difference between the 3 material groups (log-rank P = .8). Of the 10 occlusions (GEL-D n = 3, COL-D n = 3, PTFE n = 4), 4 were successfully thrombectomized. Graft infection. The severity of graft infection was classified according to Szilagyi et al.14 Graft infections (Szilagyi stage III) were observed in 6 patients (4%), all but 1 in Dacron grafts (COL-D, n = 3 and GEL-D, n = 2, PTFE, n = 1). The time to the onset of infection necessitating revision was 2 weeks to 25 months after implantation (median, 4.5 months). In 2 patients graft patency was restored by debridement or a sartorius muscle flap. In another 2 patients, one of the limbs of the graft had to be sacrificed and was replaced later with eventual limb salvage. One patient ultimately required amputation. The one PTFE graft infection seen occurred 110 months after implantation as the result of a perforated sigmoid diverticulitis with an associated abscess. Through a transabdominal approach, the graft was removed, an aortic bifurcation homograft was implanted, and the sigmoid colon was resected. The patient finally succumbed to multiorgan failure on day 21 after the initial revision. Two patients developed infections during the initial hospitalization for graft placement. One developed an extensive subcutaneous tissue necrosis (GEL-D), the other initially had an inguinal wound seroma (COL-D) as a possible source of infection. In
Surgery Volume 134, Number 1
Prager et al 83
Fig 2. Life table analysis of primary patency: GEL-D (gelatine-coated Dacron) versus COLD-D (collagen-coated Dacron) versus stretch II PTFE.
the remaining 4 patients, infections occurred after discharge 5 to 110 months after graft implantation. All patients were claudicants, but none had lower limb infection or gangrene at the time of surgery. Four patients were diabetics. In 4 of the patients with graft limb infection, the graft was anastomosed to the deep femoral arteries. Comparison of the 5 infected Dacron grafts with the single infected PTFE graft did not show any significant difference (P > .05). Late morphology. False aneurysms. One patient was reoperated because of a false aneurysm at the distal anastomoses (PTFE). The underlying cause was believed to be the result of the endarterectomy at the time of bifurcation grafting. Duplex ultrasonography of the graft limbs and body. One hundred thirty-two patients (89%) were available for follow-up studies, which included 6 monthly studies of the suture line and the graft bodies and limbs. These studies did not show major graft dilatation or aneurysms. DISCUSSION The present study is based on data from a prospective, randomized multicenter trial. The early and midterm results of this trial after a median follow-up time of 42 months were published elsewhere.9 The bifurcation grafts of all 3 materials
were comparable in primary and secondary patency rates, incidence of false aneurysms, and rate of perioperative complications. To strengthen the validity of the conclusions drawn from this previous study,9 patients with aortoiliac occlusive disease from this trial were selected for further analysis because it was felt that a homogeneous patient population well matched for the primary disease would be essential for comparing long-term patencies. So far, no other prospective randomized study has compared bifurcation grafts of different materials—that is, 2 Dacron grafts and a stretch PTFE graft—in patients with aortoiliac occlusive disease after a median follow-up time of 97 months. But long follow-up times are needed to detect shortcomings inherent in specific materials.15 Currently only 4 statistically valid studies comparing Dacron and PTFE are available.6-9 In 1988, Lord et al6 compared Dacron with conventional PTFE, both tube and bifurcation grafts, in patients with abdominal aortic aneurysms and aortoiliac occlusive disease. Lord found no difference between the 2 materials in perioperative mortality, morbidity, blood loss, cross-clamping time, and operating time. Long-term results have so far not been reported. In a prospective, randomized trial of 60 patients, Friedman et al7 compared the long-
84 Prager et al
Surgery July 2003
Fig 3. Life table analysis of secondary patency: GEL-D (gelatine-coated Dacron) vs COLD-D (collagen-coated Dacron) vs stretch II PTFE.
term results of PTFE and Dacron grafts for aortoiliac occlusive disease at a mean follow-up time of 57 months. These authors did not find any differences in early postoperative morbidity, mortality, and long-term results. In another prospective, randomized trial of 165 patients, Polterauer et al8 found the 2 graft materials to be comparable in secondary patency rates, but conventional PTFE—the thickwalled precursor of the current stretch PTFE—had a higher rate of complications. In 1987, Burke et al16 published a nonrandomized comparison of Dacron and PTFE in 42 patients with hypoplastic aortas. Patency rates 4 years postoperatively were 95% for PTFE and 52% for Dacron. In our study, none of the patients had hypoplastic vessels and needed grafts smaller than 14/7 mm. Several nonrandomized, retrospective studies argued in favor of PTFE for aortoiliac repair. Extensive experience with conventional PTFE17-23 led to the development of stretch PTFE, which is thought to have hemodynamic advantages due to the omission of the flow divider24 and better compliance because of the reduced wall thickness.24 In the present long-term trial with a median follow-up time of 97 months, no difference was seen in primary and secondary patency rates. This contrasts
with experimental data obtained with 111In-labeled thrombocytes25,26 and complement activation, which suggested that PTFE was less thrombogenic than Dacron27 because of its electronegative and hydrophobic surface.28 The 4% rate of graft infections in our study is comparatively high. In large series of aortic grafts, the risk of graft infection was reported to be between 0.7% and 2.0%.14,29 In our study, 5 of 6 patients developed graft infections within the first 2 years after implantation (2 weeks to 25 months; median 4.5 months). These may have been caused by chronic long-term infection secondary to intraoperative contamination/inoculation or by hematogenous spread of microorganisms from a distant source. The infection of the one PTFE graft 110 months after implantation was the result of sigmoid perforation and abscess. The graft infection caused by it in our study cannot necessarily be attributed to the graft material. Comparison of the Dacron grafts with the single infected PTFE graft did not show any significant difference (P > .05). Thus, long-term results presented here fail to underline the trend seen in our earlier trial.9 Today, financial aspects govern medical routine, and cost effectiveness influences our decisions.
Surgery Volume 134, Number 1
PTFE grafts were about one third more expensive in 1999 than GEL-D and COL-D grafts. Their manufacturers have justified the higher price by stressing their superior qualities. Our overall data suggest to us that PTFE grafts, although more expensive, may be justified in patients carrying an increased risk of poor wound-healing and graft infection. In conclusion, in this study of patients with aortoiliac occlusive disease, none of the 3 materials was superior in long-term patency, long-term complications, or risk of infection. This paper is dedicated to our coworker Professor Otto Wagner, a pioneer and teacher of vascular surgery, who died in February 2002. REFERENCES 1. Dubost C. Resection of an aneurysm of the abdominal aorta. Arch Surg 1952; 64:405-8. 2. Kram HB, Dietzek AM, Veith FJ. Optimal synthetic grafts for aortic replacement. In: Greenhalgh RM, ed. The cause and management of aneurysms. Philadelphia: WB Saunders; 1990. p 340-50. 3. Friedman SG. The 50th anniversary of abdominal aortic reconstruction. J Vasc Surg 2001; 33:895-8. 4. Voorhees AB. How it all began. In: Sawyer PN, ed. Vascular Grafts. New York: Appleton-Century-Crofts; 1978. p 3. 5. Voorhees AB, Jaretzki A, Blakemoore AH. The use of tubes constructed from Vinyon N cloth in bridging arterial defects. Ann Surg 1952; 135:332-6. 6. Lord RS, Nash PA, Raj BT, Stary DL, Graham AR, Hill DA, et al. Prospective randomized trial of polytetrafluoroethylene and Dacron aortic prosthesis. I. Perioperative results. Ann Vasc Surg 1988; 2:248-54. 7. Friedman SG, Lazzaro RS, Spier LN, Moccio C, Tortolani AJ. A prospective randomized comparison of Dacron and polytetraflouroethylene aortic bifurcation grafts. Surgery 1995; 117:7-10. 8. Polterauer P, Prager M, Hölzenbein T, Karner J, Kretschmer G, Schemper M. Dacron versus polytetrafluoroethylene for Y-aortic bifurcation grafts: a six-year prospective, randomized trial. Surgery 1992; 111:626-33. 9. Prager M, Polterauer P, Böhmig HJ, Wagner O, Fugl A, Kretschmer G, et al. Collagen versus gelatin-coated Dacron versus stretch polytetrafluoroethylene in abdominal aortic bifurcation graft surgery: results of a seven-year prospective, randomized multicenter trial. Surgery 2001; 130:408-14. 10. Pocock SJ, Simon R. Sequential treatment assignment with balancing for prognostic factors in the controlled clinical trial. Biometrics 1975; 31:103-11. 11. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc 1958; 53:457-65. 12. Breslow A. A generalized Kruskal-Wallis test for comparing k-samples to unequal patterns of sensorship. Biometrika 1970; 57:579-83.
Prager et al 85
13. Rutherford RB, Baker JD, Ernst C, Johnston KW, Porter JM, Ahn S, et al. Recommended standards for reports dealing with lower extremity ischemia: revised version. J Vasc Surg 1997; 26:517-38. 14. Szilagyi DE, Smith RF, Elliott JP, Vrandecic MP. Infection in arterial reconstruction with synthetic grafts. Ann Surg 1972; 176:321-33. 15. Szilagyi DE, Elliott JP Jr, Smith RF, Reddy DJ, McPharlin M. A thirty-year survey of the reconstructive surgical treatment of aortoiliac occlusive disease. J Vasc Surg 1986; 3:421-36. 16. Burke PM, Jr., Herrmann JB, Cutler BS. Optimal grafting methods for the small abdominal aorta. J Cardiovasc Surg (Torino) 1987; 28:420-6. 17. Cintora I, Pearce DE, Cannon JA. A clinical survey of aortobifemoral bypass using two inherently different graft types. Ann Surg 1988; 208:625-30. 18. Avramov S, Petrovic P, Fabri M. Bifurcated grafts (Dacron and PTFE) in aorto-iliac reconstruction: Five year follow-up. J Cardiovasc Surg 1987; 28:33-6. 19. Avramov S, Petrovic P, Plau J, Fabri M. Synthetic vascular grafts (Dacron and PTFE) in the treatment of abdominal aortic aneurysms. Five year follow-up. J Cardiovasc Surg 1988; 29:6-10. 20. Polterauer P, Wagner O, Kretschmer G, Piza F. The PTFE Y graft: one year experience in twenty-one patients. Int J Artif Organs 1982; 5:263-7. 21. Polterauer P, Wagner O, Kretschmer G, Piza F. Arterial grafting with PTFE in the pelvic region. Vasc Surg 1982; 14:390-6. 22. Polterauer P, Kretschmer G, Wagner O, Waneck R, Piza F, Lechner G. Die PTFE Y-Bifurkationsprothese: Frühergebnisse. Chirurg 1984; 55:106-10. 23. Karner J, Polterauer P, Kretschmer G, Piza F, Schemper M. Bifurkationsprothese: Dacron vs Teflon (PTFE). Acta Chir Austriaca 1987; 1:8-10. 24. Prager M, Polterauer P, Huk I, Trubel W, Nanobashvili J, Claeys L. Should Dacron or PTFE be used for aorto-iliac reconstruction ? In: Greenhalgh RM, ed. Trials and tribulations of vascular surgery. London: WB Saunders Company LTD; 1994. p 313-24. 25. Wakefield TW, Shulkin BL, Fellows EP. Platelet reactivity in human aortic grafts: a prospective, randomized midterm study of platelet adherence and release products in Dacron and polytetraflouroethylene conduits. J Vasc Surg 1986; 3:732-6. 26. Dewanjee MK. Cardiac and vascular imaging with labeled platelets and leukocytes. Semin Nucl Med 1984; 14:154-87. 27. Shepard AD, Gelfand JA, Callow AD, O’Donnell TF Jr. Complement activation by synthetic vascular prostheses. J Vasc Surg 1984; 1:829-38. 28. Chiesa R, Melissano G, Castellano R, Astore D, Castrucci M, del Maschio A, et al. A new ePTFE stretch graft for aortoiliac reconstructions. Surgical evaluation and one year follow-up with Magnetic Resonance Imaging. J Cardiovasc Surg 1995; 36:135-41. 29. Goeau-Brissoniere O, Pechere JC, Leport C. Comment prevenir les infections de prothese. In: Kieffer E, editor. Les aneurysmes de láorte abdominale sous-renale. Paris: AERCV; 1990. p 143.