- Email: [email protected]
Long-term results of a multicenter randomized study on direct versus crossover bypass for unilateral iliac artery occlusive disease
From the Society for Vascular Surgery
Long-term results of a multicenter randomized study on direct versus crossover bypass for unilateral iliac artery occlusive disease Jean-Baptiste Ricco, MD, PhD,a and Hervé Probst, MD, PhD,b on behalf of the French University Surgeons Association (AURC), Poitiers, France; and Lausanne, Switzerland Objective: To compare late patency after direct and crossover bypass in good-risk patients with unilateral iliac occlusive disease not amenable to angioplasty. Methods: Between May 1986 and March 1991, 143 patients with unilateral iliac artery occlusive disease and disabling claudication were randomized into two surgical treatment groups, ie, crossover bypass (n ⴝ 74) or direct bypass (n ⴝ 69). The size of the patient population was calculated to allow detection of a possible 20% difference in patency in favor of direct bypass with a one-sided alpha risk of 0.05 and a beta risk of 0.10. Patients underwent yearly follow-up examinations using color flow duplex scanning with ankle-brachial systolic pressure index measurement. Digital angiography was performed if hemodynamic abnormalities were noted. Median follow-up was 7.4 years. Primary endpoints were primary patency and assisted primary patency estimated by the Kaplan-Meier method with 95% confidence interval. Secondary endpoints were secondary patency and postoperative mortality and morbidity. Results: Cardiovascular risk factors, preoperative symptoms, iliac lesions TASC class (C in 87 [61%] patients and D in 56 [39%] patients), and superficial femoral artery (SFA) run-off were comparable in the two treatment groups. One patient in the direct bypass group died postoperatively. Primary patency at 5 years was higher in the direct bypass group than in the crossover bypass group (92.7 ⴞ 6.1% vs 73.2 ⴞ 10%, P ⴝ .001). Assisted primary patency and secondary patency at 5 years were also higher after direct bypass than crossover bypass (92.7 ⴞ 6.1% vs 84.3 ⴞ 8.5%, P ⴝ .04 and 97.0 ⴞ 3.0% vs 89.8 ⴞ 7.1%, P ⴝ .03, respectively). Patency at 5 years after crossover bypass was significantly higher in patients presenting no or low-grade SFA stenosis than in patients presenting high-grade (>50%) stenosis or occlusion of the SFA (74.0 ⴞ 12% vs 62.5 ⴞ 19%, P ⴝ .04). In both treatment groups, patency was comparable using polytetrafluoroethylene (PTFE) and polyester grafts. Overall survival was 59.5 ⴞ 12% at 10 years. Conclusion: This study showed that late patency was higher after direct bypass than crossover bypass in good-risk patients with unilateral iliac occlusive disease not amenable to angioplasty. Crossover bypass should be reserved for high-risk patients with unilateral iliac occlusion not amenable to percutaneous recanalization. ( J Vasc Surg 2008;47:45-54.)
Crossover femorofemoral bypass described by Freeman and Leeds1 in 1952 was first used as an alternative to direct aortofemoral bypass in high-risk patients with critical ischemia due to unilateral iliac artery occlusive disease.2 Having since gained broad acceptance, crossover bypass is now widely used to treat complex vascular problems associated with failed aortofemoral bypass or unilateral long-segment iliac occlusion graded TASC-C or –D.3 Unilateral iliac artery stenosis graded TASC-A or -B is now treated by angioplasty. Outcome of crossover bypass has varied widely in published series. The most likely explanation for this variability involves patient selection. Some authors use crossover bypass exclusively for high-risk patients with limb threatening ischemia while others extend indications to low-risk patients with claudication. Investigators using ex-
tended selection criteria have reported optimal outcomes with 5-year patency rates ranging from 82% to 90%.4-6 Variation in outcome has led to confusion regarding the utility of crossover bypass in comparison to direct bypass and the impact on late patency of variables such as the donor iliac artery disease,7 femoral artery run-off, and type of graft material.8-10 To gain more insight into these issues, the French University Surgeons Association (French acronym, AURC) undertook a multicenter randomized trial11 to compare the patency and safety of crossover bypass and direct bypass in good-risk patients with symptomatic unilateral iliac occlusive disease not amenable to endovascular treatment. The purpose of this report is to describe long-term follow-up results of that trial. PATIENTS AND METHODS
From the Vascular Surgery Service, University Hospital of Poitiers,a and Thoracic and Vascular Department, University Hospital, CHUV.b Competition of interest: none. Presented at the 2007 Vascular Annual Meeting, Baltimore, Md, Jun 6-10, 2007. Correspondence: Jean-Baptiste Ricco, MD, PhD, Vascular Surgery Service, University Hospital of Poitiers, Avenue Jacques Coeur, 86021, Poitiers, France (e-mail: [email protected]). 0741-5214/$34.00 Copyright © 2008 by The Society for Vascular Surgery. doi:10.1016/j.jvs.2007.08.050
This prospective multicenter trial comparing direct and crossover bypasses was conducted from May 1986 to March 1991 at 20 hospitals in France (see list in Appendix, online only). The study design was approved by the ethics committee of the University Hospital of Poitiers, and all patients provided written informed consent. Random assignment of patients to the two treatment groups was done independently of participating centers in a one-to-one ratio. The randomization sequence was gen45
JOURNAL OF VASCULAR SURGERY January 2008
46 Ricco and Probst
erated by a computer program and supplied to centers using sealed opaque envelopes generated in blocks of five. Endpoints. Primary endpoints were primary patency and assisted primary patency. Secondary endpoints were secondary patency and postoperative morbidity and mortality. This study used a strict definition for primary patency12 that included persistent patency without repeat intervention not only on the bypass and its anastomoses but also on the donor iliac artery of crossover bypasses. Assisted primary patency was defined as patency achieved with minor reintervention including dilation or anastomotic revision to prevent graft failure. Secondary patency was defined as patency obtained by restoration after occlusion. In this study, choice of prosthetic graft with regard to material, external support, and diameter was not randomized and left to the discretion of the surgeon. Patient selection criteria. Patients were eligible for inclusion if they had (1) unilateral disabling intermittent claudication with a walking distance of less than 300 meters and significantly reduced ankle-to-brachial systolic blood pressure index (ABI) ⱕ0.6 and (2) occlusion or diffuse stenosis involving the common iliac artery, external iliac artery, and/or common femoral artery considered by the surgeon as not amenable to percutaneous angioplasty or recanalization (Table I). An additional criterion for inclusion was demonstration of normal contralateral common and external iliac arteries on anteroposterior and oblique angiographic views with normal duplex scan findings including a peak systolic velocity ⬍1.2 m/s in the common and external iliac arteries. Patients over the age of 75 years, presenting comorbid conditions that impaired short-term survival or requiring hemodialysis were excluded. Patients with general risk factors for laparotomy including resting angina, recent myocardial infarction, severe obesity with a body-mass-index ⱖ35, or respiratory insufficiency defined as hypoxia ⬍75 mm Hg, hypercapnia ⬎50 mm Hg at rest, and one-second forced expiratory volume less than 50% of the calculated value were also excluded. Finally, patients with a history of aortofemoral or femorofemoral bypass, major abdominal wound hernia, more than two laparotomies, or pelvic radiation therapy were excluded. Immediate postoperative assessment. Bypass patency was assessed by color flow duplex scan. Respiratory complications were defined as either occurrence of pulmonary infection or need for respiratory support longer than 24 hours. Ischemic cardiac complications were defined as the presence of electrocardiogram (ECG) abnormalities. Groin incisions were regularly checked to detect healing complications including lymphocele, lymphorrea, and superficial or deep infection. Duration of hospitalization was recorded. Late follow-up assessment. Follow-up examinations were carried out at 1 month and yearly thereafter. Outpatient visits included clinical examination with measurement of walking distance and Doppler study with determination of ABI. Angiography was performed if
Table I. Baseline characteristics of patients randomized for direct and crossover bypass
Characteristics*
Direct bypass N ⫽ 69
Crossover bypass N ⫽ 74
Age -y (range) 54 (41-74) 55 (40-75) Male sex - no of patients (%) 58 (84%) 60 (81%) Vascular risk factors Hypertension 13 (19%) 22 (30%) Myocardial infarction 5 (7%) 7 (9%) Diabetes 4 (6%) 10 (13%) Hypercholesterolemia 11 (16%) 11 (15%) Tobacco use 28 (41%) 29 (39%) Body-mass-index† 28.3 ⫾ 4.2 28.7 ⫾ 3.8 Preoperative symptoms Claudication 57 (83%) 54 (73%) Rest pain 9 (13%) 14 (19%) Gangrene 3 (4%) 6 (8%) Erectile dysfunction 5/59 (8.5%) 6/67 (8.9%) Iliac lesions TASC class‡ TASC C 46 (67%) 41 (56%) TASC D 23 (33%) 33 (44%) SFA lesions (symptomatic leg)§ SFA with stenosis ⬍50% 44 (64%) 44 (59%) SFA with stenosis ⱖ50% 25 (36%) 30 (41%) or occluded
P value .89 .66 .17 .85 .16 .98 .98 .67 .50 .50 .50 .94 .18 .18 .59 .59
SFA, Superficial femoral artery. *Proportions, means, and medians were compared using the Fisher exact test, Student t-test, and the Wilcoxon nonparametric test, respectively. † The body-mass-index is the weight in kilograms divided by the square of the height in meters. Plus-minus values are means ⫾ standard deviation. ‡ According to TASC II [3], TASC C: Unilateral external iliac artery (EIA) stenosis or occlusion extending into the common femoral artery (CFA). TASC D: Unilateral occlusion of both the common iliac artery (CIA) and EIA or diffuse stenoses involving CIA, EIA, and CFA.
hemodynamic deterioration was found, ie, 15% decrease between two successive ABI measurements or if duplex scanning peak systolic velocity in the graft or in the donor iliac artery was greater than 2.5 m/s. The median duration of follow-up was 7.4 years with 112 (78%), 82 (57%), and 20 (14%) patients being followed up 5, 7, and 10 years, respectively. Statistical analysis. A meta-analysis of published crossover bypass series13-16 was performed to determine the size of the study population necessary to show a possible difference in the 3-year primary patency of 20% with a one-sided alpha of 0.05 and beta risk of 0.10. Findings demonstrated that the mean actuarial 3-year primary patency for crossover bypasses was 67%. Accordingly, it was calculated that the number of patients needed to show a 3-year patency of 87%, in favor of the direct bypass, was 70 per group. Proportions, means and medians were compared using the Fisher exact test or the Wilcoxon nonparametric test as appropriate. Primary and secondary patency rates were calculated using the Kaplan-Meier method with 95% confidence interval (95% CI).17 Groups were compared using the log-rank test with calculation of the hazard ratio with 95% CI.
JOURNAL OF VASCULAR SURGERY Volume 47, Number 1
Fig 1. Study flowchart. Two patients declined and one patient was withdrawn after a severe stroke that occurred 3 days after randomization and prior to surgery.
RESULTS Between May 1986 and March 1991, 147 patients presenting unilateral iliac artery stenosis or occlusion were randomized into the two treatment groups (Fig 1). The two study groups (Table I) were comparable with regard to vascular risk factors, preoperative symptoms, iliac artery lesions, and superficial femoral artery (SFA) run-off. Surgical techniques. Direct revascularization consisted of unilateral aortofemoral bypass in 36 cases and common iliac to femoral bypass in 33. These procedures were performed through a transperitoneal approach in 26 cases and a retroperitoneal approach in 43. Distal implantation was performed on either the common femoral artery (n ⫽ 45) or deep femoral artery (n ⫽ 24) depending on the extent of the femoral lesions. Five patients underwent femoropopliteal or femorotibial bypass during the same procedure. Crossover revascularization consisted of femorofemoral bypass in 41 cases including 28 in which the graft was routed subcutaneously and 13 in which the graft was routed through the Retzius space. In the remaining 33 patients, crossover bypass consisted of iliofemoral bypass with exposure of the external iliac artery via a supra-inguinal route and graft placement in the Retzius space. Distal implantation was performed on the common femoral artery in 46 cases and on the deep femoral artery in 28. Nine
Ricco and Probst 47
patients underwent femoropopliteal or femorotibial bypasses during the same procedure. In the direct bypass group, the graft material was polytetrafluoroethylene (PTFE) in 29 cases and polyester in 40. In the crossover bypass group, the graft material was PTFE in 53 cases and polyester in 21. Externally supported grafts were used for crossover bypass in 30 cases. Graft diameter was 8 mm in 102 cases, 7 mm in 35, and 6 mm in 6. Prophylactic antibiotics and intravenous heparin therapy (0.5 mg/kg) were routinely used before clamping in both groups. Postoperative complications. As shown in Table II, general complications were more frequent after direct bypass (7.1%) than crossover bypass (2.6%) but the difference was not significant (P ⫽ .26, relative risk: 0.54 with 95% CI: 0.16 to 1.76). One patient in the direct bypass group died due to myocardial infarction within the first 30 postoperative days. Healing complications in the groin including one graft infection were more frequent after crossover bypass (13.4%) than direct bypass (4.3%) but this difference was not significant (P ⫽ .08 relative risk: 0.32 with 95% CI: 0.09 to 1.12). Duration of hospitalization was significantly shorter in the crossover bypass group than in the direct bypass group. Sexual function was analyzed in 59 patients who underwent direct bypass and in 67 patients who underwent crossover bypass. Preoperative erectile dysfunction was reported by five patients (8.5%) in the direct bypass group and six patients (8.9%) in the crossover bypass group. At 6 months postoperatively, erectile dysfunction was reported by two patients (3.4%) in the direct group and three patients (4.5%) in the crossover bypass group. In addition, four patients (6.8%) in the direct group reported postoperative ejaculatory disorder (Table II). Primary patency. Primary patency rates at 5 and 10 years were 71.8% and 55.6%, respectively in the crossover bypass group compared with 92.7% and 82.9%, respectively in the direct bypass group (Fig 2). These differences were significant (P ⫽ .001, hazard ratio: 4.1 with 95% CI: 1.8 to 6.7). The 30 primary graft failures in the crossover bypass group were due to occlusion of the crossover bypass in 14 cases, stenosis of the femoral anastomosis in four, and stenosis of the donor iliac artery in 12. The eight primary graft failures in the direct bypass group were due to graft occlusion in six cases and stenosis of the femoral anastomosis in two. In the direct bypass group, 14 patients (20.2%) developed significant stenosis of the contralateral iliac artery requiring angioplasty in six and crossover bypass in two. These procedures did not affect patency of the direct bypass. Assisted primary patency and secondary patency. Assisted primary patency rates at five and 10 years were 84.3% and 74.8%, respectively in the crossover bypass group compared with 92.7% and 86.1%, respectively in the direct bypass group (Fig 3). These differences were significant (P ⫽ .04, hazard ratio: 2.5 with 95% CI: 1.1 to 5.8). Secondary patency rates at five and 10 years were 89.8% and 82.9%, respectively in the crossover bypass group vs 97.0% and 94.9%, respectively in the direct
JOURNAL OF VASCULAR SURGERY January 2008
48 Ricco and Probst
Table II. Postoperative complications and treatment-related outcomes after direct and crossover bypass Outcome* Postoperative general complications Death Myocardial infarction Myocardial ischemia Acute respiratory failure Postoperative femoral complications Hematoma Lymphocele Superficial infection Graft infection Length of hospitalization (d)† Sexual dysfunction Erectile dysfunction Ejaculatory disorder Bypass patency at 5 y‡ Primary patency Assisted primary patency Secondary patency Survival at 10 y
Direct bypass N ⫽ 69
Crossover bypass N ⫽ 74
P value
5 (7.1%)
2 (2.6%)
.26
1 (1.4%) 1 (1.4%) 0 3 (4.3%) 3 (4.3%)
0 1 (1.3%) 1 (1.3%) 0 10 (13.4%)
.08
0 2 (2.9%) 1 (1.4%) 0 7 (4-10)
4 (5.4%) 4 (5.4%) 1 (1.3%) 1 (1.3%) 4 (2-7)
.03
2/59 (3.4%) 4/59 (6.8%)
3/67 (4.5%) 0/67 (0%)
.95 .04
92.7 ⫾ 6.1% 92.7 ⫾ 6.1% 97.0 ⫾ 3.0% 59.3 ⫾ 17%
71.8 ⫾ 10% 84.3 ⫾ 8.5% 89.8 ⫾ 7.1% 61.2 ⫾ 9%
.001 (HR: 4.1) .04 (HR: 2.5) .03 (HR: 3.7) .59 (HR: 1.2)
HR, Hazard ratio. *Proportions, means, and medians were compared using the Fisher exact test, Student t-test, and Kaplan-Meier method with log-rank and hazard ratios, respectively. † Expressed as medians with the interquartile range. ‡ Five and 10-year results calculated by the Kaplan-Meier method with 95% confidence interval method (⫾95% CI) and compared by the log-rank test with hazard ratio.
Fig 2. Primary patency of 69 direct (D) and 74 crossover (C) bypass procedures analyzed according to the Kaplan-Meier method. The number of patients at risk in each group at various intervals is indicated at the bottom of the figure. Results are expressed as percentage with 95% confidence interval (95% CI). Primary patency rates at 5 and 10 years were 71.8 ⫾ 10% and 55.6 ⫾ 12%, respectively in the crossover bypass group compared with 92.7 ⫾ 6% and 82.9 ⫾ 13%, respectively in the direct bypass group (P ⫽ .001, hazard ratio: 4.1 with 95% CI: 1.8 to 6.7).
Fig 3. Assisted primary patency of the 69 direct and 74 crossover procedures analyzed according to the Kaplan-Meier method. The number of patients at risk in each group at various intervals is indicated at the bottom of the figure. Results are expressed as percentage with 95% confidence interval (95% CI). Assisted primary patency rates at 5 and 10 years were 84.3% and 74.8%, respectively in the crossover bypass group and 92.7% vs 86.1%, respectively in the direct bypass group (P ⫽ .04, hazard ratio: 2.5 with 95% CI: 1.1 to 5.8).
bypass group (Fig 4). These differences were significant (P ⫽ .03, hazard ratio: 3.7 with 95% CI: 1.1 to 9.8). As shown in Fig 5, all primary crossover bypass graft failures (n ⫽ 30) required further revascularization. In 30 cases
of primary crossover bypass failure, flow was successfully maintained or restored by patch placement on the femoral anastomosis in four of four cases, angioplasty of the donor iliac artery in 10 of 12, and thrombectomy of the
JOURNAL OF VASCULAR SURGERY Volume 47, Number 1
Fig 4. Secondary patency of the 69 direct and 74 crossover procedures analyzed according to the Kaplan-Meier method. The number of patients at risk in each group at various intervals is indicated at the bottom of the figure. Results are expressed as percentage with 95% confidence interval (95% CI). Secondary patency rates at 5 and 10 years were 89.8% and 82.9%, respectively in the crossover bypass group vs 97.0% and 94.9%, respectively in the direct bypass group (P ⫽ .03, hazard ratio: 3.7 with 95% CI: 1.1 to 9.8).
crossover graft in six of 14. In five of the eight primary direct bypass failures, patency was successfully restored or maintained by femoral patch angioplasty in two of two and by thrombectomy in three of six. Two patients, one in each group (n ⫽ 2), required major lower limb amputation. Impact of femoral artery run-off on bypass patency. Patency after crossover and direct bypass was analyzed in function of SFA run-off in the recipient leg. In the crossover bypass group, primary patency rates at five and 10 years were 71.9% and 64.3%, respectively in patients presenting no or low-grade (⬍50%) stenosis of the SFA compared with 62.5% and 42.3%, respectively in patients presenting high-grade stenosis (ⱖ50) or occlusion of the SFA (Fig 6). This difference was significant (P ⫽ .04, hazard ratio: 2.0 with 95% CI: 1.04 to 5.0). In the direct bypass group, primary patency rates were not significantly correlated with SFA run-off with a 10-year patency rate of 95.8% in patients presenting no or low-grade (⬍50%) stenosis of the SFA compared with 90.4% in patients presenting highgrade stenosis (ⱖ50) or occlusion of the SFA (P ⫽ .94, hazard ratio: 0.98 with 95% CI: 0.2 to 5.8). In this analysis, SFA run-off was considered as normal in patients who underwent concomitant infrainguinal revascularization. Impact of technique and prosthetic material. In the crossover bypass group, primary patency at 5 years was not significantly different for polyester and PTFE grafts: 76.2% vs 65.6%, respectively (P ⫽ .24, hazard ratio: 0.61 with 95% CI: 0.29 to 1.37). Similarly in the direct bypass group, patency at 5 years was comparable for polyester and PTFE grafts: 95.0% vs 89.4% (P ⫽ .98, hazard ratio: 0.98 with
Ricco and Probst 49
95% CI: 0.22 to 4.4). Primary patency was not significantly different between aortofemoral direct bypass and commoniliac-to-femoral direct bypass or between femorofemoral crossover bypass and external iliac-to-femoral crossover bypass. Finally, no significant difference in patency was observed within each group according to the graft diameter and patency was comparable for externally supported grafts and unsupported grafts in the crossover group. Hemodynamic outcome on the symptomatic side after direct and crossover bypass. To evaluate hemodynamic outcome, ABI was measured in the symptomatic leg before and after revascularization. As shown in Table III, preoperative and postoperative ABI was comparable in the two treatment groups, thus, suggesting that the hemodynamic outcome of direct and crossover bypass was comparable. Hemodynamic assessment on the donor artery side after crossover bypass. To study the possible hemodynamic consequences of crossover bypasses on flow in the donor iliac artery used as the take-off vessel, ABI was measured on the donor side before and after crossover bypass in function of the SFA run-off. As shown in Table IV, no significant difference was found between preoperative and postoperative ABI on the donor side. Survival. Patient survival was 94.8% at 5 years and 59.5% at 10 years. Thirty-three patients died during followup. The main causes of death were cardiovascular disease (n ⫽ 18) and cancer (n ⫽ 6). There was no significant difference in survival between the direct and crossover bypass groups (P ⫽ .58, hazard ratio: 1.2 with 95% CI: 0.6 to 2.4). DISCUSSION This randomized study shows that primary patency, assisted primary patency, and secondary patency were significantly better after direct bypass than after crossover bypass. Since progression of atherosclerosis in the donor artery is a frequent cause of crossover bypass failure, this study used a strict definition for the end of primary patency12 that included not only development of significant stenosis in the crossover bypass graft itself but also in the contralateral donor iliac artery. The risk of disease progression in the contralateral donor iliac artery was the main criteria used by Piotrowski et al18 to indicate aortofemoral bypass instead of crossover bypass even in young patients. After a median follow-up of 7.4 years in our study in which presence of a normal contralateral iliac donor artery was a condition for inclusion, stenosis in the artery that served or would have served as the donor artery for crossover bypass, was observed in 12 patients (16.2%) randomized for crossover bypass and in 14 patients (20.2%) randomized for direct bypass. The high potential for development of significant atherosclerotic disease in previously healthy iliac arteries underlines the need for surveillance using color flow duplex imaging. In the 12 patients with crossover bypasses, color flow duplex surveillance allowed detection of stenosis and restoration of patency by iliac angioplasty.
50 Ricco and Probst
JOURNAL OF VASCULAR SURGERY January 2008
Fig 5. Flowchart representing primary and secondary failures occurring in patients with crossover and direct bypass grafts. There were 30 primary failures of crossover bypass and eight primary failures of direct bypasses. Arterial flow was successfully maintained or restored by donor iliac angioplasty, thrombectomy, or femoral patch angioplasty in 20 failed crossover bypasses and in five failed primary direct bypasses. Secondary failures required 10 aortobifemoral grafts and one new crossover femorofemoral graft. Two major amputations were required in patients with failed direct or crossover bypass and unreconstructable distal arterial disease.
Determination of the status of the donor iliac artery is thus a key element for successful crossover bypass. However, assessment is difficult. Many authors consider angiography alone as unreliable.19,20 Archie et al19 stated that direct measurement of femoral artery pressure at rest and after injection of papaverine lacked sufficient sensitivity and specificity for preoperative decision-making. In our study, normal aspect of the potential donor iliac artery on both angiography and duplex scan was a prerequisite for inclusion. Although this criterion is arguably overly restrictive, our intention was to eliminate preoperative status of the donor iliac artery as a confounding factor for comparison of direct and crossover bypass. The requirement for a normal contralateral iliac artery was the main reason that it took 5 years to randomize 143 patients in the 20 surgical centers participating in this study. Indeed many patients with extensive unilateral iliac lesions have some form of contralateral iliac disease. As early as 1973, Porter et al21 acknowledged the frequency of some degree of contralateral iliac disease in patients with extensive unilateral iliac disease and became one of the first groups to recommend use of donor iliac angioplasty in combination with crossover bypass. Not surprisingly use of endovascular techniques that can provide excellent long-term results in selected iliac artery lesions has improved the outcome of crossover bypass in
patients with a suboptimal donor iliac artery.22-24 The experience of several authors25-27 has supported this view. In nonrandomized studies comparing crossover femoral grafts with or without donor iliac balloon angioplasty, both Schneider et al25 and Perler et al24 concluded that patency of the crossover bypass in patients who underwent preliminary stenting of the iliac artery was comparable to that of patients whose donor iliac artery was normal. These findings clearly support the use of angioplasty with or without stenting before or at the same time as crossover bypass in eligible patients with donor iliac lesions. Recently, AbuRahma et al28 reported that successful crossover bypass after angioplasty was more likely if the dilated iliac lesion was short and located in the common iliac artery. Another reported cause of crossover femorofemoral graft failure is progression of outflow arterial disease in the recipient limb.4,29,30 In our study, occlusion or significant stenosis of the SFA in the recipient leg was associated with significantly lower patency after crossover bypass but not after direct bypass. However, the clinical value of this finding is subject to caution because the impact of SFA status on bypass patency was not a primary endpoint of the study, and also because of the small size of these subgroups. Another concern expressed by many investigators about crossover bypass involves the ability of one iliac artery to supply blood flow to both legs and the possibility of
JOURNAL OF VASCULAR SURGERY Volume 47, Number 1
Ricco and Probst 51
Fig 6. Primary patency of crossover bypasses and direct bypasses analyzed in function of superficial femoral artery (SFA) run-off in the symptomatic leg. The number of patients at risk in each group at various intervals is indicated at the bottom of the figure. Results are expressed with 95% confidence interval (95% CI). In the crossover bypass group, primary patency rates at 5 and 10 years were 71.9% and 64.3%, respectively in patients presenting no or low-grade (⬍50%) stenosis of the SFA compared with 62.5% and 42.3%, respectively in patients presenting high-grade stenosis (ⱖ50) or occlusion of the SFA (P ⫽ .04, hazard ratio: 2.0 with 95% CI: 1.04 to 5.0). In the direct bypass group, primary patency rates were not significantly correlated with SFA run-off with a 10-year patency of 95.8% in patients presenting no or low-grade stenosis of the SFA compared with 90.4% in patients presenting high-grade stenosis or occlusion of the SFA (P ⫽ .94, hazard ratio: 0.98 with 95% CI: 0.2 to 5.8).
Table III. Hemodynamic assessment of the revascularized leg by measurement of the pre- and postoperative ankle-brachial systolic pressure index (ABI) in function of superficial femoral artery (SFA) run-off in the crossover and direct bypass groups Preoperative ABI SFA occluded
SFA patent
Postoperative ABI SFA occluded
SFA patent
Crossover bypass 0.43 ⫾ 0.16 0.56 ⫾ 0.16 0.75 ⫾ 0.17 0.97 ⫾ 0.17 Direct bypass 0.41 ⫾ 0.15 0.55 ⫾ 0.17 0.72 ⫾ 0.21 1.02 ⫾ 0.15 P value 0.89* 0.82* 0.79† 0.87‡ Results were expressed as means value ⫾ standard deviation. *Comparison of preoperative ABI in function of the SFA run-off did not show any significant difference in each category between the direct and crossover bypass groups. † Comparison of the postoperative ABI between direct and crossover bypass did not show any significant difference in each category, thus suggesting that hemodynamic outcome of direct and crossover bypass is comparable.
hemodynamic steal from the donor limb. Using an experimental model, Ehrenfeld31 was the first to demonstrate that the flow in the donor iliac artery increased after crossover bypass with no reduction in downstream iliac flow. Sumner and Strandness32 showed that crossover bypass had no significant deleterious effect on the donor limb and that the recipient limb was well perfused provided that there was no hemodynamically significant lesion in the donor iliac
Table IV. Hemodynamic assessment of the donor side in patients who underwent crossover bypass Ankle-brachial indices*
Patent SFA Occluded SFA
Preoperative
Postoperative
P value
1.02 ⫾ 0.14 0.72 ⫾ 0.17
1.00 ⫾ 0.15 0.72 ⫾ 0.21
0.89 0.96
SFA, Superficial femoral artery. Pre- and postoperative mean ABI values on the donor side were compared in patients undergoing crossover bypass in function of superficial femoral artery (SFA) lesions. No significant difference was found. Results were expressed as means value ⫾ standard deviation. *Ankle-brachial indices (ABI) expressed as means ⫾ standard deviation.
artery. Our follow-up findings are consistent with these studies since the only significant ABI reduction observed in the donor limb after crossover bypass involved the 12 patients who developed stenosis of the donor iliac artery. Measurement of ABI also enabled us to compare resting hemodynamics after crossover and direct bypass. This comparison indicated that the hemodynamic results of the two procedures were comparable. Literature data on this point have been contradictory. The Veterans Affairs Cooperative Study No 1419 including 340 patients with crossover femorofemoral grafts showed significant improvement in postoperative ABI. Conversely, based on a nonrandomized study comparing direct and crossover bypasses, Schneider et al25 concluded that resting recipient limb pressure in patients with crossover bypass was abnormal
JOURNAL OF VASCULAR SURGERY January 2008
52 Ricco and Probst
even if femoral outflow was strictly normal. Further study using exercise testing will be needed to settle this issue. Our results also provide some insight into the impact of graft material and external support on long-term results of direct and crossover bypass. Polyester and PTFE grafts performed equally well in both treatment groups. Though this finding is subject to bias since choice of type and diameter of the prosthetic graft was left to the discretion of the surgeon, it is consistent with two randomized studies showing no difference in patency outcome using different graft material for femorofemoral crossover bypass.8,9 Bypass-related complications including graft infection and seroma formation were also comparable for the two materials. Like Johnson et al,9 we could not confirm the findings of one nonrandomized retrospective study33 showing that externally supported grafts provided better results than unsupported grafts for femorofemoral crossover bypass. The use of different techniques of bypass in the two patient groups in this study could have had a confounding effect on the findings of our analysis although it should be noted that comparable patency was achieved using the two crossover bypass techniques, ie, from the external iliac artery or common femoral artery. Similarly no difference in patency was observed between direct bypass from the aorta and from the common iliac artery. Based on these observations, we do not think that the nonhomogeneity of the groups with regard to surgical techniques had a biasing effect in our study. Regarding graft diameter, our results were concordant with those of Schneider et al25 since there was no significant difference in patency using 6-, 7-, and 8-mm-diameter grafts. However, these subgroups were small and, thus, subject to type 2 error. This study also confirmed the low general morbidity of crossover bypass. This finding appears logical since superficial crossover bypass is less invasive than direct aorto- or iliofemoral bypass even when performed by a retroperitoneal approach. The absence of a significant difference in morbidity between the crossover (2.6%) and direct (7.1%) bypass group in our study is undoubtedly due to our selection criteria that excluded high-risk patients and to our relatively small patient population. Local complications were more frequent in patients undergoing crossover bypass. This difference was probably due to the fact that 41 of the 74 crossover bypasses were femorofemoral procedures requiring bilateral access to the common femoral arteries thus doubling the risk of complications in the groin area. The results of crossover bypass in this study were concordant with those reported in the literature. Table V summarizes the results of recent crossover bypass series that presented 5-year follow-up results using the actuarial or Kaplan-Meier method. Our long-term follow-up data were comparable to those reported by Brener et al29 and Criado et al.26 The good results of unilateral aortofemoral or iliofemoral direct bypass in our study are also in line with those reported by Couch et al44 and Kalman et al.45
Table V. Primary patency rate of crossover femorofemoral bypasses in previously published series with more than 40 procedures with 5 years of follow-up Primary patency at 5-year* First author Mannick 34† Flanigan 14 Sheiner 35 Dick 30 Devolfe 36 Plecha 37 Lamerton 38 Rutherford 4 Piotrowski 18 Farber 39 Perler 40 Harrington 41 Criado 26 Brener 29 Johnson 9 Mingoli 33 Purcell 42 Kim 43 This study
Year of publication
Number of bypasses
%
Bypasses at risk at 5-year
1978 1978 1979 1980 1983 1984 1985 1987 1988 1990 1991 1992 1993 1993 1999 2000 2005 2005 2007
53 80 73 133 99 119 54 60 47 71 50 162 110 228 340 228 144 192 74
80 74 73 73 71 72 60 62 55 82 57 64 60 55 49 70 74 65 72
na na na na na 39 (33%) 12 (22%) 5 (8%) 5 (11%) 21 (30%) 2 (4%) 31 (19%) 21 (19%) 54 (24%) 51 (15%) 89 (39%) 20 (14%) Na 51 (69%)
na, Data not available in the study. *Cumulative patency at 5 years according to life-table analysis. † Denotes references number.
Objective review of data from our study and the literature indicates that direct iliac revascularization offers the best long-term patency in patients with extensive unilateral iliac artery occlusion, but currently with the development of endovascular technology, this is likely done by interventional techniques. AUTHOR CONTRIBUTIONS Conception and design: JBR Analysis and interpretation: JBR, HP Data collection: JBR, HP Writing the article: JBR, HP Critical revision of the article: JBR Final approval of the article: JBR, HP Statistical analysis: JBR Obtained funding: JBR Overall responsibility: JBR, HP REFERENCES 1. Freeman NE, Leeds FH. Operations on large arteries: application of recent advances. Calif Med 1952;77:229-33. 2. Vetto RM. The femorofemoral shunt: an appraisal. Am J Surg 1966; 112:162-5. 3. Norgren L, Hiatt WR, Dormandy JA, Nehler MR, Harris KA, Fowkes FGR on behalf of the TASC II working group. Inter-Society Consensus for the Management of Peripheral Arterial Disease (TASC II). Eur J Vasc Endovasc Surg 2007;33:S54-7. 4. Rutherford RB, Patt A, Pearce WH. Extra-anatomic bypass: a closer view. J Vasc Surg 1987;6:437-46. 5. Fahal AH, McDonald AM, Marston A. Femorofemoral bypass in unilateral iliac artery occlusion. Br J Surg 1989;76:22-5.
JOURNAL OF VASCULAR SURGERY Volume 47, Number 1
6. Brief DK, Brener BJ, Alpert J, Parsonnet V. Crossover femorofemoral grafts followed up 5 years or more. Arch Surg 1975;110:1294-9. 7. Levinson SA, Levinson HJ, Halloran LG. Limited indications for unilateral aortofemoral or iliofemoral vascular grafts. Arch Surg 1973;107: 791-6. 8. Eiberg JP, Roder O, Stahl-Madsen M, Eldrup N, Qvarfordt P, Laursen A, et al. Fluoropolymer-coated Dacron versus PTFE grafts for femorofemoral crossover bypass: a randomized trial. Eur J Vasc Endovasc Surg 2006;32:431-8. 9. Johnson WC, Lee KK, with members of Veterans Affairs Cooperative Study No 141. Comparative evaluation of externally supported Dacron and polytetrafluoroethylene prosthetic bypasses for femorofemoral and axillofemoral arterial reconstructions. J Vasc Surg 1999;30:1077-83. 10. Lau H, Cheng SW. Is the preferential use of ePTFE grafts in femorofemoral bypass justified? Ann Vasc Surg 2001;15:383-7. 11. AURC & Ricco JB. Unilateral iliac artery occlusive disease: a randomized multicenter trial examining direct revascularization versus crossover bypass. Ann Vasc Surg 1992;6:209-19. 12. 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. 13. Brief DK, Brener B, Alpert J, Parsonnet V. Crossover femorofemoral grafts followed up 5 years or more: an analysis. Arch Surg 1975;110: 1294-9. 14. Flanigan P, Pratt DG, Goodreau JJ, Burnham SJ, Yao JST, Bergan JJ. Hemodynamic and angiographic guidelines in selection of patients for femorofemoral bypass. Arch Surg 1978;113:1257-60. 15. Livesay JJ, Atkinson JB, Baker JD, Busuttil RW, Barker WF, Machleder HI. Late results of extra-anatomic bypass. Arch Surg 1979;114:1260-7. 16. Eugene D, Goldstone J, Moore WS. Fifteen-year experience with subcutaneous bypass grafts for lower extremity ischemia. Ann Surg 1977;186:177-83. 17. Peto R, Pike MC, Armitage P. Design and analysis of randomized clinical trials requiring prolonged observation of each patient. Br J Cancer 1977;35:1-39. 18. Piotrowski JJ, Pearce WH, Jones DN, Whitehill T, Bell R, Patt A, et al. Aortobifemoral bypass: the operation of choice for unilateral iliac occlusion. J Vasc Surg 1988;8:211-8. 19. Archie JP. The value of donor iliac artery pressure gradients in predicting the outcome of femorofemoral bypass. J Vasc Surg 1996;23:38393. 20. Moore WS, Hall AD. Unrecognized aortoiliac stenosis: physiologic approach to the diagnosis. Arch Surg 1971;103:633-8. 21. Porter JM, Eidemiller LR, Dotter CT, Rosch J, Vetto RM. Combined arterial dilatation and femorofemoral bypass for limb salvage. Surg Gynecol Obstet 1973;137:409-12. 22. Walker PJ, Harris JP, May J. Combined percutaneous transluminal angioplasty and extra-anatomic bypass for symptomatic unilateral iliac occlusion with contralateral iliac artery stenosis. Ann Vasc Surg 1991; 5:209-17. 23. Shah RM, Peer RM, Upson JF, Ricotta JJ. Donor iliac angioplasty and crossover femorofemoral bypass. Am J Surg 1992;164:295-8. 24. Perler BA, Williams GM. Does donor iliac artery percutaneous transluminal angioplasty or stent placement influence the results of femorofemoral bypass. Analysis of 70 consecutive cases with long-term followup. J Vasc Surg 1996;24:363-70.
Ricco and Probst 53
25. Schneider JR, Besso SR, Walsh DB, Zwolack RM, Cronenwett JL. Femorofemoral versus aortofemoral bypass. Outcome and hemodynamic results. J Vasc Surg 1994;19:43-57 26. Criado E, Burnham SJ, Tinsley EA, Johnson G, Keagy BA. Femorofemoral bypass graft: analysis of patency and factors influencing longterm outcome. J Vasc Surg 1993;18:495-505. 27. Brewster DC, Cambria RF, Darling RC. Long-term results of combined iliac balloon angioplasty and distal surgical revascularization. Ann Surg 1989;210:324-31. 28. AbuRahma AF, Robinson PA, Cook CC, Hopkins ES. Selecting patients for combined femorofemoral bypass grafting and iliac balloon angioplasty and stenting for bilateral iliac disease. J Vasc Surg 2001;33: S93-9. 29. Brener BJ, Brief DK, Alpert J, Goldenkranz RJ, Eisenbud DE, Huston J, et al. Femorofemoral bypass: a 25-year experience. In: Yao JST, Pearce WH, eds. Long-term results in vascular surgery. East Norwalk, (CT): Appleton and Lange; 1993. p.385-93. 30. Dick LA, Brief DK, Alpert J, Brener BJ, Goldenkranz R, Parsonnet V. A 12-year experience with femorofemoral crossover grafts. Arch Surg 1980;115:1359-65. 31. Ehrenfeld WK, Harris JD, Wylie EJ. Vascular steal phenomenon: an experimental study. Am J Surg 1968;116:192-7. 32. Sumner DS, Strandness DE. The hemodynamics of the femorofemoral shunt. Surg Gynecol Obstet 1972;134:629-36. 33. Mingoli A, Sapienza P, Feldhaus RJ, di Marzo L, Burchi C, Cavallaro A. Femorofemoral bypass grafts: factors influencing long-term patency rate and outcome. Surgery 2001;129:451-8. 34. Mannick JA, Maini BS. Femorofemoral grafting: indications and late results. Am J Surg 1978;136:190-2. 35. Sheiner NM, Ashby D, Zeltzer J. Treatment of unilateral iliac artery disease by femorofemoral bypass grafting. Can J Surg 1979;22:474-6. 36. Devolfe C, Adeleine P, Henrie M, Violet F, Descotes J. Iliofemoral and femorofemoral crossover grafting. J Cardiovasc Surg 1983;24:634-40. 37. Plecha FR, Plecha FM. Femorofemoral bypass grafts. Ten-year experience. J Vasc Surg 1984;1:555-61. 38. Lamerton AJ, Nicolaides AN, Eastcott HHG. The femorofemoral graft. Arch Surg 1985;120:1274-8. 39. Farber MA, Hollier LH, Eubanks R, Ochsner JL, Bowen JC. Femorofemoral bypass: a profile of graft failure. South Med J 1990;83:1437-43. 40. Perler BA, Burdick JF, Williams GM. Femorofemoral or ilio-femoral bypass for unilateral inflow reconstruction? Am J Surg 1991;161:42630. 41. Harrington ME, Harrington EB, Haimov M, Schanzer H, Jacobson JH. Iliofemoral versus femorofemoral bypass: the case for an individualized approach. J Vasc Surg 1992;16:841-54. 42. Pursell R, Sideso E, Magee TR, Galland RB. Critical appraisal of femorofemoral crossover grafts. Br J Surg 2005;92:565-9. 43. Kim YW, Lee JH, Kim HG, Huh S. Factors affecting the long-term patency of crossover femorofemoral bypass graft. Eur J Vasc Endovasc Surg 2005;30:376-80. 44. Couch NP, Clowes AW, Whittemore AD. The iliac origin arterial graft: a useful alternative for iliac occlusive disease. Surgery 1985;97:83-7. 45. Kalman PG, Hosang M, Johnston KW, Walker PM. Unilateral iliac disease: the role of iliofemoral bypass. J Vasc Surg 1987;6:139-43. Submitted May 31, 2007; accepted Aug 21, 2007.
DISCUSSION Dr George Andros (Encino, Calif). I would like to congratulate you not only for presenting a wonderful series but also for coming to us with 10-year clinical and hemodynamic data. That is something we must encourage from our colleagues who favor endovascular techniques and I applaud you for doing that. I have two questions. In view of the fact that 20% of your patients developed contralateral occlusive disease after they had an aorto or iliofemoral bypass, would you recommend that the patients with extensive unilateral iliac disease undergo primary aortobifemoral bypass and thereby solve the problem once and forever?
Second, since the patients with compromised outflow on the recipient side did less well, would you recommend that they are watched more closely with a low threshold for receiving a distal bypass graft? Not only would this policy relieve symptoms but also preserve the cross femoral bypass. Dr Jean-Baptiste Ricco. Concerning your first question and looking at the results of our study, you may be right, however, at the present time, we will certainly proceed first with an iliac recanalization or a hybrid procedure with iliac recanalization and femoral reconstruction. If this does not work, we will then certainly
54 Ricco and Probst
consider a bilateral revascularization particularly if the life expectancy of the patient is high. Concerning the femoral outflow, occlusion or significant stenosis of the SFA in the recipient leg was associated in our study with a significantly lower patency after crossover bypass but not after direct bypass, and we will certainly recommend in these cases a distal revascularization to enhance patency of the crossover bypass. Dr Vikram Kashyap (Cleveland, Ohio). Two questions. One, in our series, looking at endovascular recanalization of iliac occlusions, 30% of groin vessels needed either an endarterectomy or a profundaplasty. Is that similar in your series? Second, when this study started 20 years ago, all of these patients were excluded from endovascular techniques that were deemed not feasible. But in the end, if I understood correctly, approximately 16% ended up getting an angioplasty on the ipsilateral iliac artery. Have endovascular techniques changed your approach to iliac occlusions? Dr Jean-Baptiste Ricco. Concerning your first question, we have the same experience, in this series, 52 out of 143 patients (36%) had a profundaplasty with or without a femoral endarterectory. Considering your second question, we did 12 angioplasties of the donor iliac artery in the crossover group mainly for TASC A or
JOURNAL OF VASCULAR SURGERY January 2008
B lesions that developed some years after construction of the crossover bypass. In the direct bypass group, 14 patients (20.2%) developed significant stenosis of the contralateral iliac artery requiring angioplasty in six and crossover bypass in two. Our practice has certainly changed in the last 10 years. Not surprisingly, endovascular techniques can provide excellent long-term results in selected iliac artery lesions and have also improved the outcome of crossover femoral bypass in patients with suboptimal donor iliac artery. Dr Subodh Arora (Washington, DC). I noticed that quite a number of patients in the iliofemoral group, you used the transperitoneal approach. Was there any particular reason why that approach was used over the retroperitoneal? Dr Jean-Baptiste Ricco. Two different techniques were used in the direct bypass group: 36 patients had an aortofemoral bypass and 33 patients had a common iliac-femoral bypass. Forty-three patients had a retroperitoneal approach and 26 patients had a transperitoneal approach. All direct iliofemoral bypasses and 10 aortofemoral bypasses were approached by a retroperitoneal route. Transperitoneal route was used exclusively in the remaining 26 patients with an aortafemoral bypass. In this study, the choice of the operative technique was left to the discretion of the surgeon.
JOURNAL OF VASCULAR SURGERY Volume 47, Number 1
APPENDIX Participating centers University Hospital, Poitiers (J.B. Ricco), University Hospital Charles Nicolle, Rouen (J. Testart, J. Watelet), University Hospital Saint-Eloi, Montpellier (H. Mary), Hôpital Saint-Philibert, Lomme (P. Puppinck), University Hospital, Caen (D. Maiza), Hôpital Pasteur, Langon (P. Plagnol), University Hospital Edouard Herriot, Lyon (J.M. Chevalier), University Hospital Sainte-Marguerite, Marseille (P. Piquet), Clinique Poirier, Chambery (B. Habozit), University Hospi-
Ricco and Probst 54.e1
tal Pitié Salpêtrière, Paris (E. Kieffer, F. Koskas), Hôpital Saint-Michel, Paris (P. Lagneau), University Hospital Henri Mondor, Créteil (J.P. Becquemin), University Hospital du Bocage, Dijon (M. David, R. Brenot), American Hospital, Neuilly (F. Bacourt), University Hospital, Besancon (G. Camelot, J. Combe), University Hospital, Nancy (G. Fieve), University Hospital, Rennes (Y. Kerdiles), Clinique Bizet, Paris (J.M. Fichelle), University Hospital Laennec, Nantes (P. Patra), University Hospital de la Timone, Marseille (A. Branchereau).
Long-term results of a multicenter randomized study on direct versus crossover bypass for unilateral iliac artery occlusive disease Jean-Baptiste Ricco, MD, PhD,a and Hervé Probst, MD, PhD,b on behalf of the French University Surgeons Association (AURC), Poitiers, France; and Lausanne, Switzerland Objective: To compare late patency after direct and crossover bypass in good-risk patients with unilateral iliac occlusive disease not amenable to angioplasty. Methods: Between May 1986 and March 1991, 143 patients with unilateral iliac artery occlusive disease and disabling claudication were randomized into two surgical treatment groups, ie, crossover bypass (n ⴝ 74) or direct bypass (n ⴝ 69). The size of the patient population was calculated to allow detection of a possible 20% difference in patency in favor of direct bypass with a one-sided alpha risk of 0.05 and a beta risk of 0.10. Patients underwent yearly follow-up examinations using color flow duplex scanning with ankle-brachial systolic pressure index measurement. Digital angiography was performed if hemodynamic abnormalities were noted. Median follow-up was 7.4 years. Primary endpoints were primary patency and assisted primary patency estimated by the Kaplan-Meier method with 95% confidence interval. Secondary endpoints were secondary patency and postoperative mortality and morbidity. Results: Cardiovascular risk factors, preoperative symptoms, iliac lesions TASC class (C in 87 [61%] patients and D in 56 [39%] patients), and superficial femoral artery (SFA) run-off were comparable in the two treatment groups. One patient in the direct bypass group died postoperatively. Primary patency at 5 years was higher in the direct bypass group than in the crossover bypass group (92.7 ⴞ 6.1% vs 73.2 ⴞ 10%, P ⴝ .001). Assisted primary patency and secondary patency at 5 years were also higher after direct bypass than crossover bypass (92.7 ⴞ 6.1% vs 84.3 ⴞ 8.5%, P ⴝ .04 and 97.0 ⴞ 3.0% vs 89.8 ⴞ 7.1%, P ⴝ .03, respectively). Patency at 5 years after crossover bypass was significantly higher in patients presenting no or low-grade SFA stenosis than in patients presenting high-grade (>50%) stenosis or occlusion of the SFA (74.0 ⴞ 12% vs 62.5 ⴞ 19%, P ⴝ .04). In both treatment groups, patency was comparable using polytetrafluoroethylene (PTFE) and polyester grafts. Overall survival was 59.5 ⴞ 12% at 10 years. Conclusion: This study showed that late patency was higher after direct bypass than crossover bypass in good-risk patients with unilateral iliac occlusive disease not amenable to angioplasty. Crossover bypass should be reserved for high-risk patients with unilateral iliac occlusion not amenable to percutaneous recanalization. ( J Vasc Surg 2008;47:45-54.)
Crossover femorofemoral bypass described by Freeman and Leeds1 in 1952 was first used as an alternative to direct aortofemoral bypass in high-risk patients with critical ischemia due to unilateral iliac artery occlusive disease.2 Having since gained broad acceptance, crossover bypass is now widely used to treat complex vascular problems associated with failed aortofemoral bypass or unilateral long-segment iliac occlusion graded TASC-C or –D.3 Unilateral iliac artery stenosis graded TASC-A or -B is now treated by angioplasty. Outcome of crossover bypass has varied widely in published series. The most likely explanation for this variability involves patient selection. Some authors use crossover bypass exclusively for high-risk patients with limb threatening ischemia while others extend indications to low-risk patients with claudication. Investigators using ex-
tended selection criteria have reported optimal outcomes with 5-year patency rates ranging from 82% to 90%.4-6 Variation in outcome has led to confusion regarding the utility of crossover bypass in comparison to direct bypass and the impact on late patency of variables such as the donor iliac artery disease,7 femoral artery run-off, and type of graft material.8-10 To gain more insight into these issues, the French University Surgeons Association (French acronym, AURC) undertook a multicenter randomized trial11 to compare the patency and safety of crossover bypass and direct bypass in good-risk patients with symptomatic unilateral iliac occlusive disease not amenable to endovascular treatment. The purpose of this report is to describe long-term follow-up results of that trial. PATIENTS AND METHODS
From the Vascular Surgery Service, University Hospital of Poitiers,a and Thoracic and Vascular Department, University Hospital, CHUV.b Competition of interest: none. Presented at the 2007 Vascular Annual Meeting, Baltimore, Md, Jun 6-10, 2007. Correspondence: Jean-Baptiste Ricco, MD, PhD, Vascular Surgery Service, University Hospital of Poitiers, Avenue Jacques Coeur, 86021, Poitiers, France (e-mail: [email protected]). 0741-5214/$34.00 Copyright © 2008 by The Society for Vascular Surgery. doi:10.1016/j.jvs.2007.08.050
This prospective multicenter trial comparing direct and crossover bypasses was conducted from May 1986 to March 1991 at 20 hospitals in France (see list in Appendix, online only). The study design was approved by the ethics committee of the University Hospital of Poitiers, and all patients provided written informed consent. Random assignment of patients to the two treatment groups was done independently of participating centers in a one-to-one ratio. The randomization sequence was gen45
JOURNAL OF VASCULAR SURGERY January 2008
46 Ricco and Probst
erated by a computer program and supplied to centers using sealed opaque envelopes generated in blocks of five. Endpoints. Primary endpoints were primary patency and assisted primary patency. Secondary endpoints were secondary patency and postoperative morbidity and mortality. This study used a strict definition for primary patency12 that included persistent patency without repeat intervention not only on the bypass and its anastomoses but also on the donor iliac artery of crossover bypasses. Assisted primary patency was defined as patency achieved with minor reintervention including dilation or anastomotic revision to prevent graft failure. Secondary patency was defined as patency obtained by restoration after occlusion. In this study, choice of prosthetic graft with regard to material, external support, and diameter was not randomized and left to the discretion of the surgeon. Patient selection criteria. Patients were eligible for inclusion if they had (1) unilateral disabling intermittent claudication with a walking distance of less than 300 meters and significantly reduced ankle-to-brachial systolic blood pressure index (ABI) ⱕ0.6 and (2) occlusion or diffuse stenosis involving the common iliac artery, external iliac artery, and/or common femoral artery considered by the surgeon as not amenable to percutaneous angioplasty or recanalization (Table I). An additional criterion for inclusion was demonstration of normal contralateral common and external iliac arteries on anteroposterior and oblique angiographic views with normal duplex scan findings including a peak systolic velocity ⬍1.2 m/s in the common and external iliac arteries. Patients over the age of 75 years, presenting comorbid conditions that impaired short-term survival or requiring hemodialysis were excluded. Patients with general risk factors for laparotomy including resting angina, recent myocardial infarction, severe obesity with a body-mass-index ⱖ35, or respiratory insufficiency defined as hypoxia ⬍75 mm Hg, hypercapnia ⬎50 mm Hg at rest, and one-second forced expiratory volume less than 50% of the calculated value were also excluded. Finally, patients with a history of aortofemoral or femorofemoral bypass, major abdominal wound hernia, more than two laparotomies, or pelvic radiation therapy were excluded. Immediate postoperative assessment. Bypass patency was assessed by color flow duplex scan. Respiratory complications were defined as either occurrence of pulmonary infection or need for respiratory support longer than 24 hours. Ischemic cardiac complications were defined as the presence of electrocardiogram (ECG) abnormalities. Groin incisions were regularly checked to detect healing complications including lymphocele, lymphorrea, and superficial or deep infection. Duration of hospitalization was recorded. Late follow-up assessment. Follow-up examinations were carried out at 1 month and yearly thereafter. Outpatient visits included clinical examination with measurement of walking distance and Doppler study with determination of ABI. Angiography was performed if
Table I. Baseline characteristics of patients randomized for direct and crossover bypass
Characteristics*
Direct bypass N ⫽ 69
Crossover bypass N ⫽ 74
Age -y (range) 54 (41-74) 55 (40-75) Male sex - no of patients (%) 58 (84%) 60 (81%) Vascular risk factors Hypertension 13 (19%) 22 (30%) Myocardial infarction 5 (7%) 7 (9%) Diabetes 4 (6%) 10 (13%) Hypercholesterolemia 11 (16%) 11 (15%) Tobacco use 28 (41%) 29 (39%) Body-mass-index† 28.3 ⫾ 4.2 28.7 ⫾ 3.8 Preoperative symptoms Claudication 57 (83%) 54 (73%) Rest pain 9 (13%) 14 (19%) Gangrene 3 (4%) 6 (8%) Erectile dysfunction 5/59 (8.5%) 6/67 (8.9%) Iliac lesions TASC class‡ TASC C 46 (67%) 41 (56%) TASC D 23 (33%) 33 (44%) SFA lesions (symptomatic leg)§ SFA with stenosis ⬍50% 44 (64%) 44 (59%) SFA with stenosis ⱖ50% 25 (36%) 30 (41%) or occluded
P value .89 .66 .17 .85 .16 .98 .98 .67 .50 .50 .50 .94 .18 .18 .59 .59
SFA, Superficial femoral artery. *Proportions, means, and medians were compared using the Fisher exact test, Student t-test, and the Wilcoxon nonparametric test, respectively. † The body-mass-index is the weight in kilograms divided by the square of the height in meters. Plus-minus values are means ⫾ standard deviation. ‡ According to TASC II [3], TASC C: Unilateral external iliac artery (EIA) stenosis or occlusion extending into the common femoral artery (CFA). TASC D: Unilateral occlusion of both the common iliac artery (CIA) and EIA or diffuse stenoses involving CIA, EIA, and CFA.
hemodynamic deterioration was found, ie, 15% decrease between two successive ABI measurements or if duplex scanning peak systolic velocity in the graft or in the donor iliac artery was greater than 2.5 m/s. The median duration of follow-up was 7.4 years with 112 (78%), 82 (57%), and 20 (14%) patients being followed up 5, 7, and 10 years, respectively. Statistical analysis. A meta-analysis of published crossover bypass series13-16 was performed to determine the size of the study population necessary to show a possible difference in the 3-year primary patency of 20% with a one-sided alpha of 0.05 and beta risk of 0.10. Findings demonstrated that the mean actuarial 3-year primary patency for crossover bypasses was 67%. Accordingly, it was calculated that the number of patients needed to show a 3-year patency of 87%, in favor of the direct bypass, was 70 per group. Proportions, means and medians were compared using the Fisher exact test or the Wilcoxon nonparametric test as appropriate. Primary and secondary patency rates were calculated using the Kaplan-Meier method with 95% confidence interval (95% CI).17 Groups were compared using the log-rank test with calculation of the hazard ratio with 95% CI.
JOURNAL OF VASCULAR SURGERY Volume 47, Number 1
Fig 1. Study flowchart. Two patients declined and one patient was withdrawn after a severe stroke that occurred 3 days after randomization and prior to surgery.
RESULTS Between May 1986 and March 1991, 147 patients presenting unilateral iliac artery stenosis or occlusion were randomized into the two treatment groups (Fig 1). The two study groups (Table I) were comparable with regard to vascular risk factors, preoperative symptoms, iliac artery lesions, and superficial femoral artery (SFA) run-off. Surgical techniques. Direct revascularization consisted of unilateral aortofemoral bypass in 36 cases and common iliac to femoral bypass in 33. These procedures were performed through a transperitoneal approach in 26 cases and a retroperitoneal approach in 43. Distal implantation was performed on either the common femoral artery (n ⫽ 45) or deep femoral artery (n ⫽ 24) depending on the extent of the femoral lesions. Five patients underwent femoropopliteal or femorotibial bypass during the same procedure. Crossover revascularization consisted of femorofemoral bypass in 41 cases including 28 in which the graft was routed subcutaneously and 13 in which the graft was routed through the Retzius space. In the remaining 33 patients, crossover bypass consisted of iliofemoral bypass with exposure of the external iliac artery via a supra-inguinal route and graft placement in the Retzius space. Distal implantation was performed on the common femoral artery in 46 cases and on the deep femoral artery in 28. Nine
Ricco and Probst 47
patients underwent femoropopliteal or femorotibial bypasses during the same procedure. In the direct bypass group, the graft material was polytetrafluoroethylene (PTFE) in 29 cases and polyester in 40. In the crossover bypass group, the graft material was PTFE in 53 cases and polyester in 21. Externally supported grafts were used for crossover bypass in 30 cases. Graft diameter was 8 mm in 102 cases, 7 mm in 35, and 6 mm in 6. Prophylactic antibiotics and intravenous heparin therapy (0.5 mg/kg) were routinely used before clamping in both groups. Postoperative complications. As shown in Table II, general complications were more frequent after direct bypass (7.1%) than crossover bypass (2.6%) but the difference was not significant (P ⫽ .26, relative risk: 0.54 with 95% CI: 0.16 to 1.76). One patient in the direct bypass group died due to myocardial infarction within the first 30 postoperative days. Healing complications in the groin including one graft infection were more frequent after crossover bypass (13.4%) than direct bypass (4.3%) but this difference was not significant (P ⫽ .08 relative risk: 0.32 with 95% CI: 0.09 to 1.12). Duration of hospitalization was significantly shorter in the crossover bypass group than in the direct bypass group. Sexual function was analyzed in 59 patients who underwent direct bypass and in 67 patients who underwent crossover bypass. Preoperative erectile dysfunction was reported by five patients (8.5%) in the direct bypass group and six patients (8.9%) in the crossover bypass group. At 6 months postoperatively, erectile dysfunction was reported by two patients (3.4%) in the direct group and three patients (4.5%) in the crossover bypass group. In addition, four patients (6.8%) in the direct group reported postoperative ejaculatory disorder (Table II). Primary patency. Primary patency rates at 5 and 10 years were 71.8% and 55.6%, respectively in the crossover bypass group compared with 92.7% and 82.9%, respectively in the direct bypass group (Fig 2). These differences were significant (P ⫽ .001, hazard ratio: 4.1 with 95% CI: 1.8 to 6.7). The 30 primary graft failures in the crossover bypass group were due to occlusion of the crossover bypass in 14 cases, stenosis of the femoral anastomosis in four, and stenosis of the donor iliac artery in 12. The eight primary graft failures in the direct bypass group were due to graft occlusion in six cases and stenosis of the femoral anastomosis in two. In the direct bypass group, 14 patients (20.2%) developed significant stenosis of the contralateral iliac artery requiring angioplasty in six and crossover bypass in two. These procedures did not affect patency of the direct bypass. Assisted primary patency and secondary patency. Assisted primary patency rates at five and 10 years were 84.3% and 74.8%, respectively in the crossover bypass group compared with 92.7% and 86.1%, respectively in the direct bypass group (Fig 3). These differences were significant (P ⫽ .04, hazard ratio: 2.5 with 95% CI: 1.1 to 5.8). Secondary patency rates at five and 10 years were 89.8% and 82.9%, respectively in the crossover bypass group vs 97.0% and 94.9%, respectively in the direct
JOURNAL OF VASCULAR SURGERY January 2008
48 Ricco and Probst
Table II. Postoperative complications and treatment-related outcomes after direct and crossover bypass Outcome* Postoperative general complications Death Myocardial infarction Myocardial ischemia Acute respiratory failure Postoperative femoral complications Hematoma Lymphocele Superficial infection Graft infection Length of hospitalization (d)† Sexual dysfunction Erectile dysfunction Ejaculatory disorder Bypass patency at 5 y‡ Primary patency Assisted primary patency Secondary patency Survival at 10 y
Direct bypass N ⫽ 69
Crossover bypass N ⫽ 74
P value
5 (7.1%)
2 (2.6%)
.26
1 (1.4%) 1 (1.4%) 0 3 (4.3%) 3 (4.3%)
0 1 (1.3%) 1 (1.3%) 0 10 (13.4%)
.08
0 2 (2.9%) 1 (1.4%) 0 7 (4-10)
4 (5.4%) 4 (5.4%) 1 (1.3%) 1 (1.3%) 4 (2-7)
.03
2/59 (3.4%) 4/59 (6.8%)
3/67 (4.5%) 0/67 (0%)
.95 .04
92.7 ⫾ 6.1% 92.7 ⫾ 6.1% 97.0 ⫾ 3.0% 59.3 ⫾ 17%
71.8 ⫾ 10% 84.3 ⫾ 8.5% 89.8 ⫾ 7.1% 61.2 ⫾ 9%
.001 (HR: 4.1) .04 (HR: 2.5) .03 (HR: 3.7) .59 (HR: 1.2)
HR, Hazard ratio. *Proportions, means, and medians were compared using the Fisher exact test, Student t-test, and Kaplan-Meier method with log-rank and hazard ratios, respectively. † Expressed as medians with the interquartile range. ‡ Five and 10-year results calculated by the Kaplan-Meier method with 95% confidence interval method (⫾95% CI) and compared by the log-rank test with hazard ratio.
Fig 2. Primary patency of 69 direct (D) and 74 crossover (C) bypass procedures analyzed according to the Kaplan-Meier method. The number of patients at risk in each group at various intervals is indicated at the bottom of the figure. Results are expressed as percentage with 95% confidence interval (95% CI). Primary patency rates at 5 and 10 years were 71.8 ⫾ 10% and 55.6 ⫾ 12%, respectively in the crossover bypass group compared with 92.7 ⫾ 6% and 82.9 ⫾ 13%, respectively in the direct bypass group (P ⫽ .001, hazard ratio: 4.1 with 95% CI: 1.8 to 6.7).
Fig 3. Assisted primary patency of the 69 direct and 74 crossover procedures analyzed according to the Kaplan-Meier method. The number of patients at risk in each group at various intervals is indicated at the bottom of the figure. Results are expressed as percentage with 95% confidence interval (95% CI). Assisted primary patency rates at 5 and 10 years were 84.3% and 74.8%, respectively in the crossover bypass group and 92.7% vs 86.1%, respectively in the direct bypass group (P ⫽ .04, hazard ratio: 2.5 with 95% CI: 1.1 to 5.8).
bypass group (Fig 4). These differences were significant (P ⫽ .03, hazard ratio: 3.7 with 95% CI: 1.1 to 9.8). As shown in Fig 5, all primary crossover bypass graft failures (n ⫽ 30) required further revascularization. In 30 cases
of primary crossover bypass failure, flow was successfully maintained or restored by patch placement on the femoral anastomosis in four of four cases, angioplasty of the donor iliac artery in 10 of 12, and thrombectomy of the
JOURNAL OF VASCULAR SURGERY Volume 47, Number 1
Fig 4. Secondary patency of the 69 direct and 74 crossover procedures analyzed according to the Kaplan-Meier method. The number of patients at risk in each group at various intervals is indicated at the bottom of the figure. Results are expressed as percentage with 95% confidence interval (95% CI). Secondary patency rates at 5 and 10 years were 89.8% and 82.9%, respectively in the crossover bypass group vs 97.0% and 94.9%, respectively in the direct bypass group (P ⫽ .03, hazard ratio: 3.7 with 95% CI: 1.1 to 9.8).
crossover graft in six of 14. In five of the eight primary direct bypass failures, patency was successfully restored or maintained by femoral patch angioplasty in two of two and by thrombectomy in three of six. Two patients, one in each group (n ⫽ 2), required major lower limb amputation. Impact of femoral artery run-off on bypass patency. Patency after crossover and direct bypass was analyzed in function of SFA run-off in the recipient leg. In the crossover bypass group, primary patency rates at five and 10 years were 71.9% and 64.3%, respectively in patients presenting no or low-grade (⬍50%) stenosis of the SFA compared with 62.5% and 42.3%, respectively in patients presenting high-grade stenosis (ⱖ50) or occlusion of the SFA (Fig 6). This difference was significant (P ⫽ .04, hazard ratio: 2.0 with 95% CI: 1.04 to 5.0). In the direct bypass group, primary patency rates were not significantly correlated with SFA run-off with a 10-year patency rate of 95.8% in patients presenting no or low-grade (⬍50%) stenosis of the SFA compared with 90.4% in patients presenting highgrade stenosis (ⱖ50) or occlusion of the SFA (P ⫽ .94, hazard ratio: 0.98 with 95% CI: 0.2 to 5.8). In this analysis, SFA run-off was considered as normal in patients who underwent concomitant infrainguinal revascularization. Impact of technique and prosthetic material. In the crossover bypass group, primary patency at 5 years was not significantly different for polyester and PTFE grafts: 76.2% vs 65.6%, respectively (P ⫽ .24, hazard ratio: 0.61 with 95% CI: 0.29 to 1.37). Similarly in the direct bypass group, patency at 5 years was comparable for polyester and PTFE grafts: 95.0% vs 89.4% (P ⫽ .98, hazard ratio: 0.98 with
Ricco and Probst 49
95% CI: 0.22 to 4.4). Primary patency was not significantly different between aortofemoral direct bypass and commoniliac-to-femoral direct bypass or between femorofemoral crossover bypass and external iliac-to-femoral crossover bypass. Finally, no significant difference in patency was observed within each group according to the graft diameter and patency was comparable for externally supported grafts and unsupported grafts in the crossover group. Hemodynamic outcome on the symptomatic side after direct and crossover bypass. To evaluate hemodynamic outcome, ABI was measured in the symptomatic leg before and after revascularization. As shown in Table III, preoperative and postoperative ABI was comparable in the two treatment groups, thus, suggesting that the hemodynamic outcome of direct and crossover bypass was comparable. Hemodynamic assessment on the donor artery side after crossover bypass. To study the possible hemodynamic consequences of crossover bypasses on flow in the donor iliac artery used as the take-off vessel, ABI was measured on the donor side before and after crossover bypass in function of the SFA run-off. As shown in Table IV, no significant difference was found between preoperative and postoperative ABI on the donor side. Survival. Patient survival was 94.8% at 5 years and 59.5% at 10 years. Thirty-three patients died during followup. The main causes of death were cardiovascular disease (n ⫽ 18) and cancer (n ⫽ 6). There was no significant difference in survival between the direct and crossover bypass groups (P ⫽ .58, hazard ratio: 1.2 with 95% CI: 0.6 to 2.4). DISCUSSION This randomized study shows that primary patency, assisted primary patency, and secondary patency were significantly better after direct bypass than after crossover bypass. Since progression of atherosclerosis in the donor artery is a frequent cause of crossover bypass failure, this study used a strict definition for the end of primary patency12 that included not only development of significant stenosis in the crossover bypass graft itself but also in the contralateral donor iliac artery. The risk of disease progression in the contralateral donor iliac artery was the main criteria used by Piotrowski et al18 to indicate aortofemoral bypass instead of crossover bypass even in young patients. After a median follow-up of 7.4 years in our study in which presence of a normal contralateral iliac donor artery was a condition for inclusion, stenosis in the artery that served or would have served as the donor artery for crossover bypass, was observed in 12 patients (16.2%) randomized for crossover bypass and in 14 patients (20.2%) randomized for direct bypass. The high potential for development of significant atherosclerotic disease in previously healthy iliac arteries underlines the need for surveillance using color flow duplex imaging. In the 12 patients with crossover bypasses, color flow duplex surveillance allowed detection of stenosis and restoration of patency by iliac angioplasty.
50 Ricco and Probst
JOURNAL OF VASCULAR SURGERY January 2008
Fig 5. Flowchart representing primary and secondary failures occurring in patients with crossover and direct bypass grafts. There were 30 primary failures of crossover bypass and eight primary failures of direct bypasses. Arterial flow was successfully maintained or restored by donor iliac angioplasty, thrombectomy, or femoral patch angioplasty in 20 failed crossover bypasses and in five failed primary direct bypasses. Secondary failures required 10 aortobifemoral grafts and one new crossover femorofemoral graft. Two major amputations were required in patients with failed direct or crossover bypass and unreconstructable distal arterial disease.
Determination of the status of the donor iliac artery is thus a key element for successful crossover bypass. However, assessment is difficult. Many authors consider angiography alone as unreliable.19,20 Archie et al19 stated that direct measurement of femoral artery pressure at rest and after injection of papaverine lacked sufficient sensitivity and specificity for preoperative decision-making. In our study, normal aspect of the potential donor iliac artery on both angiography and duplex scan was a prerequisite for inclusion. Although this criterion is arguably overly restrictive, our intention was to eliminate preoperative status of the donor iliac artery as a confounding factor for comparison of direct and crossover bypass. The requirement for a normal contralateral iliac artery was the main reason that it took 5 years to randomize 143 patients in the 20 surgical centers participating in this study. Indeed many patients with extensive unilateral iliac lesions have some form of contralateral iliac disease. As early as 1973, Porter et al21 acknowledged the frequency of some degree of contralateral iliac disease in patients with extensive unilateral iliac disease and became one of the first groups to recommend use of donor iliac angioplasty in combination with crossover bypass. Not surprisingly use of endovascular techniques that can provide excellent long-term results in selected iliac artery lesions has improved the outcome of crossover bypass in
patients with a suboptimal donor iliac artery.22-24 The experience of several authors25-27 has supported this view. In nonrandomized studies comparing crossover femoral grafts with or without donor iliac balloon angioplasty, both Schneider et al25 and Perler et al24 concluded that patency of the crossover bypass in patients who underwent preliminary stenting of the iliac artery was comparable to that of patients whose donor iliac artery was normal. These findings clearly support the use of angioplasty with or without stenting before or at the same time as crossover bypass in eligible patients with donor iliac lesions. Recently, AbuRahma et al28 reported that successful crossover bypass after angioplasty was more likely if the dilated iliac lesion was short and located in the common iliac artery. Another reported cause of crossover femorofemoral graft failure is progression of outflow arterial disease in the recipient limb.4,29,30 In our study, occlusion or significant stenosis of the SFA in the recipient leg was associated with significantly lower patency after crossover bypass but not after direct bypass. However, the clinical value of this finding is subject to caution because the impact of SFA status on bypass patency was not a primary endpoint of the study, and also because of the small size of these subgroups. Another concern expressed by many investigators about crossover bypass involves the ability of one iliac artery to supply blood flow to both legs and the possibility of
JOURNAL OF VASCULAR SURGERY Volume 47, Number 1
Ricco and Probst 51
Fig 6. Primary patency of crossover bypasses and direct bypasses analyzed in function of superficial femoral artery (SFA) run-off in the symptomatic leg. The number of patients at risk in each group at various intervals is indicated at the bottom of the figure. Results are expressed with 95% confidence interval (95% CI). In the crossover bypass group, primary patency rates at 5 and 10 years were 71.9% and 64.3%, respectively in patients presenting no or low-grade (⬍50%) stenosis of the SFA compared with 62.5% and 42.3%, respectively in patients presenting high-grade stenosis (ⱖ50) or occlusion of the SFA (P ⫽ .04, hazard ratio: 2.0 with 95% CI: 1.04 to 5.0). In the direct bypass group, primary patency rates were not significantly correlated with SFA run-off with a 10-year patency of 95.8% in patients presenting no or low-grade stenosis of the SFA compared with 90.4% in patients presenting high-grade stenosis or occlusion of the SFA (P ⫽ .94, hazard ratio: 0.98 with 95% CI: 0.2 to 5.8).
Table III. Hemodynamic assessment of the revascularized leg by measurement of the pre- and postoperative ankle-brachial systolic pressure index (ABI) in function of superficial femoral artery (SFA) run-off in the crossover and direct bypass groups Preoperative ABI SFA occluded
SFA patent
Postoperative ABI SFA occluded
SFA patent
Crossover bypass 0.43 ⫾ 0.16 0.56 ⫾ 0.16 0.75 ⫾ 0.17 0.97 ⫾ 0.17 Direct bypass 0.41 ⫾ 0.15 0.55 ⫾ 0.17 0.72 ⫾ 0.21 1.02 ⫾ 0.15 P value 0.89* 0.82* 0.79† 0.87‡ Results were expressed as means value ⫾ standard deviation. *Comparison of preoperative ABI in function of the SFA run-off did not show any significant difference in each category between the direct and crossover bypass groups. † Comparison of the postoperative ABI between direct and crossover bypass did not show any significant difference in each category, thus suggesting that hemodynamic outcome of direct and crossover bypass is comparable.
hemodynamic steal from the donor limb. Using an experimental model, Ehrenfeld31 was the first to demonstrate that the flow in the donor iliac artery increased after crossover bypass with no reduction in downstream iliac flow. Sumner and Strandness32 showed that crossover bypass had no significant deleterious effect on the donor limb and that the recipient limb was well perfused provided that there was no hemodynamically significant lesion in the donor iliac
Table IV. Hemodynamic assessment of the donor side in patients who underwent crossover bypass Ankle-brachial indices*
Patent SFA Occluded SFA
Preoperative
Postoperative
P value
1.02 ⫾ 0.14 0.72 ⫾ 0.17
1.00 ⫾ 0.15 0.72 ⫾ 0.21
0.89 0.96
SFA, Superficial femoral artery. Pre- and postoperative mean ABI values on the donor side were compared in patients undergoing crossover bypass in function of superficial femoral artery (SFA) lesions. No significant difference was found. Results were expressed as means value ⫾ standard deviation. *Ankle-brachial indices (ABI) expressed as means ⫾ standard deviation.
artery. Our follow-up findings are consistent with these studies since the only significant ABI reduction observed in the donor limb after crossover bypass involved the 12 patients who developed stenosis of the donor iliac artery. Measurement of ABI also enabled us to compare resting hemodynamics after crossover and direct bypass. This comparison indicated that the hemodynamic results of the two procedures were comparable. Literature data on this point have been contradictory. The Veterans Affairs Cooperative Study No 1419 including 340 patients with crossover femorofemoral grafts showed significant improvement in postoperative ABI. Conversely, based on a nonrandomized study comparing direct and crossover bypasses, Schneider et al25 concluded that resting recipient limb pressure in patients with crossover bypass was abnormal
JOURNAL OF VASCULAR SURGERY January 2008
52 Ricco and Probst
even if femoral outflow was strictly normal. Further study using exercise testing will be needed to settle this issue. Our results also provide some insight into the impact of graft material and external support on long-term results of direct and crossover bypass. Polyester and PTFE grafts performed equally well in both treatment groups. Though this finding is subject to bias since choice of type and diameter of the prosthetic graft was left to the discretion of the surgeon, it is consistent with two randomized studies showing no difference in patency outcome using different graft material for femorofemoral crossover bypass.8,9 Bypass-related complications including graft infection and seroma formation were also comparable for the two materials. Like Johnson et al,9 we could not confirm the findings of one nonrandomized retrospective study33 showing that externally supported grafts provided better results than unsupported grafts for femorofemoral crossover bypass. The use of different techniques of bypass in the two patient groups in this study could have had a confounding effect on the findings of our analysis although it should be noted that comparable patency was achieved using the two crossover bypass techniques, ie, from the external iliac artery or common femoral artery. Similarly no difference in patency was observed between direct bypass from the aorta and from the common iliac artery. Based on these observations, we do not think that the nonhomogeneity of the groups with regard to surgical techniques had a biasing effect in our study. Regarding graft diameter, our results were concordant with those of Schneider et al25 since there was no significant difference in patency using 6-, 7-, and 8-mm-diameter grafts. However, these subgroups were small and, thus, subject to type 2 error. This study also confirmed the low general morbidity of crossover bypass. This finding appears logical since superficial crossover bypass is less invasive than direct aorto- or iliofemoral bypass even when performed by a retroperitoneal approach. The absence of a significant difference in morbidity between the crossover (2.6%) and direct (7.1%) bypass group in our study is undoubtedly due to our selection criteria that excluded high-risk patients and to our relatively small patient population. Local complications were more frequent in patients undergoing crossover bypass. This difference was probably due to the fact that 41 of the 74 crossover bypasses were femorofemoral procedures requiring bilateral access to the common femoral arteries thus doubling the risk of complications in the groin area. The results of crossover bypass in this study were concordant with those reported in the literature. Table V summarizes the results of recent crossover bypass series that presented 5-year follow-up results using the actuarial or Kaplan-Meier method. Our long-term follow-up data were comparable to those reported by Brener et al29 and Criado et al.26 The good results of unilateral aortofemoral or iliofemoral direct bypass in our study are also in line with those reported by Couch et al44 and Kalman et al.45
Table V. Primary patency rate of crossover femorofemoral bypasses in previously published series with more than 40 procedures with 5 years of follow-up Primary patency at 5-year* First author Mannick 34† Flanigan 14 Sheiner 35 Dick 30 Devolfe 36 Plecha 37 Lamerton 38 Rutherford 4 Piotrowski 18 Farber 39 Perler 40 Harrington 41 Criado 26 Brener 29 Johnson 9 Mingoli 33 Purcell 42 Kim 43 This study
Year of publication
Number of bypasses
%
Bypasses at risk at 5-year
1978 1978 1979 1980 1983 1984 1985 1987 1988 1990 1991 1992 1993 1993 1999 2000 2005 2005 2007
53 80 73 133 99 119 54 60 47 71 50 162 110 228 340 228 144 192 74
80 74 73 73 71 72 60 62 55 82 57 64 60 55 49 70 74 65 72
na na na na na 39 (33%) 12 (22%) 5 (8%) 5 (11%) 21 (30%) 2 (4%) 31 (19%) 21 (19%) 54 (24%) 51 (15%) 89 (39%) 20 (14%) Na 51 (69%)
na, Data not available in the study. *Cumulative patency at 5 years according to life-table analysis. † Denotes references number.
Objective review of data from our study and the literature indicates that direct iliac revascularization offers the best long-term patency in patients with extensive unilateral iliac artery occlusion, but currently with the development of endovascular technology, this is likely done by interventional techniques. AUTHOR CONTRIBUTIONS Conception and design: JBR Analysis and interpretation: JBR, HP Data collection: JBR, HP Writing the article: JBR, HP Critical revision of the article: JBR Final approval of the article: JBR, HP Statistical analysis: JBR Obtained funding: JBR Overall responsibility: JBR, HP REFERENCES 1. Freeman NE, Leeds FH. Operations on large arteries: application of recent advances. Calif Med 1952;77:229-33. 2. Vetto RM. The femorofemoral shunt: an appraisal. Am J Surg 1966; 112:162-5. 3. Norgren L, Hiatt WR, Dormandy JA, Nehler MR, Harris KA, Fowkes FGR on behalf of the TASC II working group. Inter-Society Consensus for the Management of Peripheral Arterial Disease (TASC II). Eur J Vasc Endovasc Surg 2007;33:S54-7. 4. Rutherford RB, Patt A, Pearce WH. Extra-anatomic bypass: a closer view. J Vasc Surg 1987;6:437-46. 5. Fahal AH, McDonald AM, Marston A. Femorofemoral bypass in unilateral iliac artery occlusion. Br J Surg 1989;76:22-5.
JOURNAL OF VASCULAR SURGERY Volume 47, Number 1
6. Brief DK, Brener BJ, Alpert J, Parsonnet V. Crossover femorofemoral grafts followed up 5 years or more. Arch Surg 1975;110:1294-9. 7. Levinson SA, Levinson HJ, Halloran LG. Limited indications for unilateral aortofemoral or iliofemoral vascular grafts. Arch Surg 1973;107: 791-6. 8. Eiberg JP, Roder O, Stahl-Madsen M, Eldrup N, Qvarfordt P, Laursen A, et al. Fluoropolymer-coated Dacron versus PTFE grafts for femorofemoral crossover bypass: a randomized trial. Eur J Vasc Endovasc Surg 2006;32:431-8. 9. Johnson WC, Lee KK, with members of Veterans Affairs Cooperative Study No 141. Comparative evaluation of externally supported Dacron and polytetrafluoroethylene prosthetic bypasses for femorofemoral and axillofemoral arterial reconstructions. J Vasc Surg 1999;30:1077-83. 10. Lau H, Cheng SW. Is the preferential use of ePTFE grafts in femorofemoral bypass justified? Ann Vasc Surg 2001;15:383-7. 11. AURC & Ricco JB. Unilateral iliac artery occlusive disease: a randomized multicenter trial examining direct revascularization versus crossover bypass. Ann Vasc Surg 1992;6:209-19. 12. 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. 13. Brief DK, Brener B, Alpert J, Parsonnet V. Crossover femorofemoral grafts followed up 5 years or more: an analysis. Arch Surg 1975;110: 1294-9. 14. Flanigan P, Pratt DG, Goodreau JJ, Burnham SJ, Yao JST, Bergan JJ. Hemodynamic and angiographic guidelines in selection of patients for femorofemoral bypass. Arch Surg 1978;113:1257-60. 15. Livesay JJ, Atkinson JB, Baker JD, Busuttil RW, Barker WF, Machleder HI. Late results of extra-anatomic bypass. Arch Surg 1979;114:1260-7. 16. Eugene D, Goldstone J, Moore WS. Fifteen-year experience with subcutaneous bypass grafts for lower extremity ischemia. Ann Surg 1977;186:177-83. 17. Peto R, Pike MC, Armitage P. Design and analysis of randomized clinical trials requiring prolonged observation of each patient. Br J Cancer 1977;35:1-39. 18. Piotrowski JJ, Pearce WH, Jones DN, Whitehill T, Bell R, Patt A, et al. Aortobifemoral bypass: the operation of choice for unilateral iliac occlusion. J Vasc Surg 1988;8:211-8. 19. Archie JP. The value of donor iliac artery pressure gradients in predicting the outcome of femorofemoral bypass. J Vasc Surg 1996;23:38393. 20. Moore WS, Hall AD. Unrecognized aortoiliac stenosis: physiologic approach to the diagnosis. Arch Surg 1971;103:633-8. 21. Porter JM, Eidemiller LR, Dotter CT, Rosch J, Vetto RM. Combined arterial dilatation and femorofemoral bypass for limb salvage. Surg Gynecol Obstet 1973;137:409-12. 22. Walker PJ, Harris JP, May J. Combined percutaneous transluminal angioplasty and extra-anatomic bypass for symptomatic unilateral iliac occlusion with contralateral iliac artery stenosis. Ann Vasc Surg 1991; 5:209-17. 23. Shah RM, Peer RM, Upson JF, Ricotta JJ. Donor iliac angioplasty and crossover femorofemoral bypass. Am J Surg 1992;164:295-8. 24. Perler BA, Williams GM. Does donor iliac artery percutaneous transluminal angioplasty or stent placement influence the results of femorofemoral bypass. Analysis of 70 consecutive cases with long-term followup. J Vasc Surg 1996;24:363-70.
Ricco and Probst 53
25. Schneider JR, Besso SR, Walsh DB, Zwolack RM, Cronenwett JL. Femorofemoral versus aortofemoral bypass. Outcome and hemodynamic results. J Vasc Surg 1994;19:43-57 26. Criado E, Burnham SJ, Tinsley EA, Johnson G, Keagy BA. Femorofemoral bypass graft: analysis of patency and factors influencing longterm outcome. J Vasc Surg 1993;18:495-505. 27. Brewster DC, Cambria RF, Darling RC. Long-term results of combined iliac balloon angioplasty and distal surgical revascularization. Ann Surg 1989;210:324-31. 28. AbuRahma AF, Robinson PA, Cook CC, Hopkins ES. Selecting patients for combined femorofemoral bypass grafting and iliac balloon angioplasty and stenting for bilateral iliac disease. J Vasc Surg 2001;33: S93-9. 29. Brener BJ, Brief DK, Alpert J, Goldenkranz RJ, Eisenbud DE, Huston J, et al. Femorofemoral bypass: a 25-year experience. In: Yao JST, Pearce WH, eds. Long-term results in vascular surgery. East Norwalk, (CT): Appleton and Lange; 1993. p.385-93. 30. Dick LA, Brief DK, Alpert J, Brener BJ, Goldenkranz R, Parsonnet V. A 12-year experience with femorofemoral crossover grafts. Arch Surg 1980;115:1359-65. 31. Ehrenfeld WK, Harris JD, Wylie EJ. Vascular steal phenomenon: an experimental study. Am J Surg 1968;116:192-7. 32. Sumner DS, Strandness DE. The hemodynamics of the femorofemoral shunt. Surg Gynecol Obstet 1972;134:629-36. 33. Mingoli A, Sapienza P, Feldhaus RJ, di Marzo L, Burchi C, Cavallaro A. Femorofemoral bypass grafts: factors influencing long-term patency rate and outcome. Surgery 2001;129:451-8. 34. Mannick JA, Maini BS. Femorofemoral grafting: indications and late results. Am J Surg 1978;136:190-2. 35. Sheiner NM, Ashby D, Zeltzer J. Treatment of unilateral iliac artery disease by femorofemoral bypass grafting. Can J Surg 1979;22:474-6. 36. Devolfe C, Adeleine P, Henrie M, Violet F, Descotes J. Iliofemoral and femorofemoral crossover grafting. J Cardiovasc Surg 1983;24:634-40. 37. Plecha FR, Plecha FM. Femorofemoral bypass grafts. Ten-year experience. J Vasc Surg 1984;1:555-61. 38. Lamerton AJ, Nicolaides AN, Eastcott HHG. The femorofemoral graft. Arch Surg 1985;120:1274-8. 39. Farber MA, Hollier LH, Eubanks R, Ochsner JL, Bowen JC. Femorofemoral bypass: a profile of graft failure. South Med J 1990;83:1437-43. 40. Perler BA, Burdick JF, Williams GM. Femorofemoral or ilio-femoral bypass for unilateral inflow reconstruction? Am J Surg 1991;161:42630. 41. Harrington ME, Harrington EB, Haimov M, Schanzer H, Jacobson JH. Iliofemoral versus femorofemoral bypass: the case for an individualized approach. J Vasc Surg 1992;16:841-54. 42. Pursell R, Sideso E, Magee TR, Galland RB. Critical appraisal of femorofemoral crossover grafts. Br J Surg 2005;92:565-9. 43. Kim YW, Lee JH, Kim HG, Huh S. Factors affecting the long-term patency of crossover femorofemoral bypass graft. Eur J Vasc Endovasc Surg 2005;30:376-80. 44. Couch NP, Clowes AW, Whittemore AD. The iliac origin arterial graft: a useful alternative for iliac occlusive disease. Surgery 1985;97:83-7. 45. Kalman PG, Hosang M, Johnston KW, Walker PM. Unilateral iliac disease: the role of iliofemoral bypass. J Vasc Surg 1987;6:139-43. Submitted May 31, 2007; accepted Aug 21, 2007.
DISCUSSION Dr George Andros (Encino, Calif). I would like to congratulate you not only for presenting a wonderful series but also for coming to us with 10-year clinical and hemodynamic data. That is something we must encourage from our colleagues who favor endovascular techniques and I applaud you for doing that. I have two questions. In view of the fact that 20% of your patients developed contralateral occlusive disease after they had an aorto or iliofemoral bypass, would you recommend that the patients with extensive unilateral iliac disease undergo primary aortobifemoral bypass and thereby solve the problem once and forever?
Second, since the patients with compromised outflow on the recipient side did less well, would you recommend that they are watched more closely with a low threshold for receiving a distal bypass graft? Not only would this policy relieve symptoms but also preserve the cross femoral bypass. Dr Jean-Baptiste Ricco. Concerning your first question and looking at the results of our study, you may be right, however, at the present time, we will certainly proceed first with an iliac recanalization or a hybrid procedure with iliac recanalization and femoral reconstruction. If this does not work, we will then certainly
54 Ricco and Probst
consider a bilateral revascularization particularly if the life expectancy of the patient is high. Concerning the femoral outflow, occlusion or significant stenosis of the SFA in the recipient leg was associated in our study with a significantly lower patency after crossover bypass but not after direct bypass, and we will certainly recommend in these cases a distal revascularization to enhance patency of the crossover bypass. Dr Vikram Kashyap (Cleveland, Ohio). Two questions. One, in our series, looking at endovascular recanalization of iliac occlusions, 30% of groin vessels needed either an endarterectomy or a profundaplasty. Is that similar in your series? Second, when this study started 20 years ago, all of these patients were excluded from endovascular techniques that were deemed not feasible. But in the end, if I understood correctly, approximately 16% ended up getting an angioplasty on the ipsilateral iliac artery. Have endovascular techniques changed your approach to iliac occlusions? Dr Jean-Baptiste Ricco. Concerning your first question, we have the same experience, in this series, 52 out of 143 patients (36%) had a profundaplasty with or without a femoral endarterectory. Considering your second question, we did 12 angioplasties of the donor iliac artery in the crossover group mainly for TASC A or
JOURNAL OF VASCULAR SURGERY January 2008
B lesions that developed some years after construction of the crossover bypass. In the direct bypass group, 14 patients (20.2%) developed significant stenosis of the contralateral iliac artery requiring angioplasty in six and crossover bypass in two. Our practice has certainly changed in the last 10 years. Not surprisingly, endovascular techniques can provide excellent long-term results in selected iliac artery lesions and have also improved the outcome of crossover femoral bypass in patients with suboptimal donor iliac artery. Dr Subodh Arora (Washington, DC). I noticed that quite a number of patients in the iliofemoral group, you used the transperitoneal approach. Was there any particular reason why that approach was used over the retroperitoneal? Dr Jean-Baptiste Ricco. Two different techniques were used in the direct bypass group: 36 patients had an aortofemoral bypass and 33 patients had a common iliac-femoral bypass. Forty-three patients had a retroperitoneal approach and 26 patients had a transperitoneal approach. All direct iliofemoral bypasses and 10 aortofemoral bypasses were approached by a retroperitoneal route. Transperitoneal route was used exclusively in the remaining 26 patients with an aortafemoral bypass. In this study, the choice of the operative technique was left to the discretion of the surgeon.
JOURNAL OF VASCULAR SURGERY Volume 47, Number 1
APPENDIX Participating centers University Hospital, Poitiers (J.B. Ricco), University Hospital Charles Nicolle, Rouen (J. Testart, J. Watelet), University Hospital Saint-Eloi, Montpellier (H. Mary), Hôpital Saint-Philibert, Lomme (P. Puppinck), University Hospital, Caen (D. Maiza), Hôpital Pasteur, Langon (P. Plagnol), University Hospital Edouard Herriot, Lyon (J.M. Chevalier), University Hospital Sainte-Marguerite, Marseille (P. Piquet), Clinique Poirier, Chambery (B. Habozit), University Hospi-
Ricco and Probst 54.e1
tal Pitié Salpêtrière, Paris (E. Kieffer, F. Koskas), Hôpital Saint-Michel, Paris (P. Lagneau), University Hospital Henri Mondor, Créteil (J.P. Becquemin), University Hospital du Bocage, Dijon (M. David, R. Brenot), American Hospital, Neuilly (F. Bacourt), University Hospital, Besancon (G. Camelot, J. Combe), University Hospital, Nancy (G. Fieve), University Hospital, Rennes (Y. Kerdiles), Clinique Bizet, Paris (J.M. Fichelle), University Hospital Laennec, Nantes (P. Patra), University Hospital de la Timone, Marseille (A. Branchereau).