It is estimated that 10% to 20% of the world s population

Transcription

1 Mechanical Inhibition of Angiogenesis at the Saphenofemoral Junction in the Surgical Treatment of Varicose Veins Early Results of a Blinded Randomized Controlled Trial André M. van Rij, MD, FRACS; Gregory T. Jones, PhD; B. Geraldine Hill, BSc(Hon); Mohammad Amer, MB, ChB; Ian A. Thomson, MB, ChB, FRACS; Ross A. Pettigrew, MB, ChB, FRACS; Stephen G.K. Packer, MB, ChB, FRACS Background Saphenofemoral junction (SFJ) ligation has been a major component of surgical intervention for varicose veins; however, recurrence occurs in as many as 40%. Neovascularization with reconnection of the venous channels at the transected SFJ has been identified as the major cause of this recurrence. This randomized controlled study sought to evaluate mechanical suppression of neovascularization at the SFJ, with the use of a synthetic patch, to prevent recurrence after ligation surgery. Methods and Results A total of 389 limbs (from 292 patients) were randomized into either control (SFJ ligation surgery) or patch (SFJ ligation with polytetrafluoroethylene patch of the transected SFJ) groups. All patients underwent clinical assessment, duplex imaging, and air plethysmography studies preoperatively and at 1, 6, 12, and 36 months postoperatively. The patch consistently halved the recurrence rate to 3 years postoperatively in all clinical subgroups. In those patched SFJs that still developed recurrence, evidence of neovascularization circumventing the polytetrafluoroethylene patch was observed by both ultrasound and histology. Conclusions This study demonstrates that use of a polytetrafluoroethylene patch is an effective mechanical suppressant of neovasculogenesis at the SFJ and can be safely used as a strategy to improve long-term outcome of varicose vein surgery. (Circulation. 2008;118:66-74.) Key Words: angiogenesis follow-up studies surgery varicose veins veins It is estimated that 10% to 20% of the world s population has varicose veins in the lower extremities. 1 Although surgery has been the standard treatment for varicose veins for many years, recurrence rates remain as high as 40% at 5 years. 2,3 Consequently, 20% of all varicose vein operations are performed for recurrent disease. 4 Clinical Perspective p 74 More than 90% of recurrent varicose veins are associated with recurrent incompetence at the previously ligated saphenofemoral junction (SFJ). 5 Various suggestions have been made for the cause of this high rate of recurrence, including poor surgical technique, particularly by junior surgeons in training 6 ; anatomic variability and complexity of the SFJ 7 ; and persistent progressive venous disease. 1 However, it is now believed that the most common cause of this recurrence is angiogenesis with new vessel formation and reconnection of the venous channels. This is commonly referred to as neovascularization in the surgical literature and has been demonstrated in several clinicopathological and histological studies Prevention or inhibition of this form of angiogenesis would be a major step toward reducing the rate of recurrence. Suppression of angiogenesis has generally focused on pharmacological measures with the use of vascular growth factor inhibitors or their receptors. 14 Although this approach has considerable potential, significant potential for adverse side effects also exists, particularly with regard to wound healing. 15 Another possible approach is the introduction of a physical barrier across the site of angiogenic focus to prevent new vessels crossing back and reconnecting the venous system. The use of a mechanical barrier to prevent angiogenesis has been described in other settings such as the treatment of exudative age-related macular degeneration, 16 in which a macular buckle of Gore-Tex buttonholed through a silicone Received July 10, 2007; accepted April 14, From the Department of Surgery, Dunedin School of Medicine, University of Otago, Dunedin, New Zealand. Clinical trial registration information URL: Unique identifier: The online-only Data Supplement is available with this article at Correspondence to André van Rij, MD, FRACS, Department of Surgery, Dunedin School of Medicine, University of Otago, PO Box 931, Dunedin, New Zealand. andre.vanrij@stonebow.otago.ac.nz 2008 American Heart Association, Inc. Circulation is available at DOI: /CIRCULATIONAHA

2 van Rij et al Mechanical Inhibition of Angiogenesis 67 sponge is used to act as a barrier against the growth of choroidal neovascular membranes. This results in a marked reduction in the rate and degree of visual loss compared with the natural history of the disease. Similarly, Hunziker et al 17 used a Gore-Tex patch to inhibit unwanted neovascular effects of growth factors in full-thickness articular cartilage repair in a miniature pig model. In varicose vein surgery, the transected and ligated great saphenous vein presents an appropriate setting to test the concept of whether a mechanical barrier can significantly prevent angiogenesis. The occurrence of neovascularization in this situation can be readily monitored by ultrasound. Several studies have suggested that physical inhibition of angiogenesis, at the ligated SFJ, may be effective. The first such report was published well before the biology of angiogenesis was described. In 1978, Sheppard 5 sutured a reflected flap of pectineus fascia over the femoral vein and the saphenous vein stump, thereby physically separating the femoral vein from the plane of superficial veins. Good results were claimed but never well documented. Almost 20 years later, Gibbs et al 18 performed a randomized controlled trial of this procedure involving 40 limbs and failed to demonstrate any benefit. Subsequently, a small number of prospective studies have been conducted, applying various materials to act as a physical barrier at the ligated SFJ site, including Dacron, 8 Gore-Tex, 19 and silicone The results have been generally favorable, although the designs of these studies have had various significant limitations, including small numbers and short follow-up periods. A more critical evaluation, in the form of a suitably powered prospective randomized controlled trial, is required to demonstrate the efficacy of mechanical inhibition of SFJ angiogenesis. This article reports on our early findings of a randomized controlled trial of the use of a polytetrafluoroethylene patch as a mechanical suppressant of angiogenesis at the SFJ to prevent recurrence. It also describes both the acute and chronic histological characteristics of the neovascular response to the patch barrier. Methods Patients Patients were a nonselected consecutive group who were referred for varicose vein surgery. All patients gave informed consent for the trial, which was approved by the Otago Ethics Committee. Patients were assessed preoperatively with the use of a standard clinical, etiologic, anatomic, and pathological (CEAP) classification as recommended by the Society for Vascular Surgery and the International Society for Cardiovascular Surgery 23 and the Venous Clinical Severity Score (VCSS). 24 All patients were CEAP clinical score 2. General demographics, a clinical evaluation of medical and surgical history, risk factors for venous disease, and other relevant data were recorded. A full venous duplex ultrasound assessment was performed with the use of an ATL 5000 scanner and 7-4 MHz/12-5 MHz linear array probe with color and spectral Doppler modalities (Philips Medical Systems), as detailed elsewhere. 25 Physiological testing with air plethysmography was performed with a digital multichanneled MacLab/8 system (Analog Digital Instruments, Dunedin, New Zealand), as previously described. 12 Independent randomization (by patient) was conducted from a remote site, immediately before commencement of the surgery, in blocks of 10 for each of the 4 participating surgeons, with the use of cards in preprepared sealed envelopes to undergo standard SFJ ligation and great saphenous vein Table 1. Baseline Patient Demographics, Limb Characteristics, and Withdrawals Control Patch P Patient demographics Patients, n Patients with both 45 (30) 52 (37) 0.40 limbs in study Female 84 (56) 81 (57) 0.60 Age Body mass index Body mass index 46/130 (35) 39/115 (34) Limb characteristics Limbs Left limbs 102 (52) 103 (53) 0.87 Previous deep vein 18 (9) 18 (9) 0.98 thrombosis history Previous varicose vein 80 (41) 63 (33) 0.09 surgery CEAP score (44) 67 (35) 0.06 VCSS (total) VCSS S VCSS S 5 71 (36) 58 (30) 0.17 Active ulceration 21 (11) 15 (8) 0.30 Deep venous disease 34 (18) 34 (18) 0.98 (Kistner 2 4) Venous filling index, ml/s Ejection fraction Withdrawals Patients 9 (6) 8 (6) Limbs 10 (5) 10 (5) Deceased 1 2 Leg amputated 2 0 Moved out of town 2 3 Patient request 4 3 Age, body mass index, venous filling index, and ejection fractions are expressed as mean 1 SD. Other values are expressed as n (%). stripping with or without the insertion of a 2 3-cm polytetrafluoroethylene patch (Gore-Tex, WL Gore & Associates, Inc) over the ligated SFJ stump. Follow-Up After surgery, all patients were recalled for clinical assessment, duplex imaging, and air plethysmography studies at 1, 6, 12, 36, and 60 months postoperatively. The 5-year follow-up is yet to be done and is not included in this report. Once successful surgical intervention was confirmed by ultrasound, no interventions for recurrence were performed during the 3-year follow-up period. Study Assessment End Points Incidence of neoreflux (on duplex scanning) at the previously confirmed obliterated SFJ was the primary end point and constituted the definition of recurrence. Recurrence at the SFJ was recognized on duplex ultrasonography by the reappearance of reflux for 0.5 seconds at the site of previous surgery first by Valsalva and then by calf augmentation and confirmed on a subsequent reassessment. For

3 68 Circulation July 1, 2008 Table 2. Ultrasound-Detected SFJ Recurrence at Postoperative Follow-Up in Control Versus Patch Groups 1 Month 6 Months 1 Year 3 Years Valsalva-detected ultrasound recurrence Control 1.0 (2/195) 22.8 (37/162) 37.6 (67/178) 45.8 (54/118) Patch 0 (0/194) 7.3 (12/164) 16.3 (27/166) 22.9 (23/105) Adjusted odds ratio 0.27* ( ) 0.32* ( ) 0.33* ( ) Calf augmentation detected ultrasound recurrence Control 0.5 (1/195) 6.8 (11/162) 10.7 (19/178) 14.3 (17/118) Patch 0 (0/194) 0.6 (1/164) 2.4 (4/166) 4.7 (5/105) Adjusted odds ratio 0.08 ( ) 0.21 ( ) 0.30 ( ) Values are % (n/total). Data represent percentage of limbs, assessed at each time point, with ultrasound-detected recurrence (No. recurrent/group size). *P 0.001, P , P 0.003, P Adjusted logistic regression model, shown as odds ratio (95% CI), includes bilateral/unilateral surgery, baseline CEAP, VCSS S, previous varicose vein surgery, and surgeon. Odds ratios were not calculated for 1 month because of lack of events. Valsalva, patients were asked to bear down as hard as possible (rather than simply blow into their cheeks) to maximize their response. Calf augmentation was standardized by the use of a pneumatic compression cuff (6 kg/cm 2 ). Secondary end points included a VCSS( S) 3, that is, a VCSS excluding the stocking compressive therapy component 24 ; clinical reappearance of varicosities reported by the patient and confirmed at the vascular laboratory follow-up (with the use of the VCSS subscale for varicosities); the number of limbs with active ulceration; and a venous reflux index, venous filling index 3 ml/s. Other markers of venous disease were also compared including pain, heaviness, swelling, sclerosis, edema, pigmentation, inflammation, and induration (with the use of VCSS subscales as appropriate). Vascular technicians with extensive experience in these types of studies performed all assessments. Both technicians and patients were blinded to the treatment group applied to each limb. The vascular technicians could not reliably determine the presence of a patch in the postoperative setting. The surgeons, while aware of the procedure at the time, did not record this in the clinical record and were unaware of the patient group thereafter and did not participate in the study follow-up assessments of their patients. All subjects were invited for follow-up evaluation, regardless of preceding recurrence outcome, and no assumptions of continued recurrence were applied. Recurrence rates reported at each time interval are the number of patients who demonstrated recurrence at that specific time divided by all the subjects assessed at the same interval. Surgery Standard surgical procedures were performed, consisting of SFJ ligation, great saphenous vein stripping to the knee, ligation of Table 3. Ultrasound-Detected SFJ Recurrence at Postoperative Follow-Up in Control Versus Patch Groups Subdivided by Preintervention VCSS Preintervention VCSS 1 Month 6 Months 1 Year 3 Years VCSS S 1 3 Control 0 (0/89) 16.9 (33/77) 32.1 (27/84) 41.1 (23/56) Patch 0 (0/115) 7.4 (7/95) 14.6* (14/96) 19.0 (11/58) VCSS S 3 Control 1.9 (2/106) 28.2 (24/85) 42.6 (40/94) 50.0 (31/62) Patch 0 (0/75) 6.2 (4/65) 18.2 (12/66) 25.0 (11/44) Values are % (n/total). Data represent percentage of limbs, assessed at each time point, with ultrasound-detected recurrence (No. recurrent/group size). *P 0.006, P 0.001, P , P 0.002, P incompetent perforators (marked preoperatively), and multiple stab avulsions. Ligation of the saphenopopliteal junction was performed as required. The patch was secured with a 3-0 Prolene suture at each corner to adjacent tissue to lie comfortably over the stump and the adjacent common femoral vein. Four experienced senior surgeons performed all the surgery. Patients were provided with compression stockings for the postoperative period and were encouraged to mobilize promptly. None of the patients experienced any postoperative complications related to the patch. Histology A histological sample of the patch was obtained from a case in which a thigh epigastric vein was mistaken for the SFJ and ligated instead, and a patch was placed over the resulting stump. At the 1-month follow-up, this error was recognized, a decision was made to reoperate, the original patch was removed and used for histological analysis, and a new patch was inserted. The limb was therefore retained in the study, but data pertaining to the initial failed operation were excluded from analysis. In addition, a subject who developed recurrence was recommended for further surgery, at which time a patch was removed 5.5 years after insertion. These specimens were carefully dissected and fixed in 10% formalin before being embedded in paraffin wax and prepared for histology. Sections (4 m thick) were cut and stained with either Verhoeff elastic stain with van Gieson counterstain, phosphotungstic acid hematoxylin, or Hoechst DNA stain and examined on an Olympus BX51 light microscope. Separate dewaxed sections were immunostained for endothelial cells Table 4. Ultrasound-Detected SFJ Recurrence at Postoperative Follow-Up in Control Versus Patch Groups Subdivided by CEAP Severity Clinical Severity 1 Month 6 Months 1 Year 3 Years Baseline CEAP 0 3 Control 0 (0/109) 17.0 (15/88) 31.0 (31/100) 43.8 (28/64) Patch 0 (0/127) 8.4 (9/107) 17.3* (19/110) 23.5 (16/68) Baseline CEAP 4 6 Control 2.3 (2/86) 29.7 (22/74) 46.2 (36/78) 48.2 (26/54) Patch 0 (0/67) 5.3 (3/57) 14.3 (8/56) 18.9 (7/37) Values are % (n/total). Data represent percentage of limbs, assessed at each time point, with ultrasound-detected recurrence (No. recurrent/group size). *P 0.02, P 0.02, P , P , P

4 van Rij et al Mechanical Inhibition of Angiogenesis 69 Table 5. Ultrasound-Detected SFJ Recurrence at Postoperative Follow-Up in Control Versus Patch Groups Subdivided by History of Previous Varicose Vein Surgery History of Previous Surgery 1 Month 6 Months 1 Year 3 Years No previous varicose vein surgery (primary surgery) Control 0.9 (1/115) 25.7 (29/113) 41.9 (52/124) 48.6 (35/72) Patch 0 (0/130) 9.4* (11/117) 16.2 (19/117) 21.1 (15/71) Previous varicose vein surgery Control 1.3 (1/80) 17.5 (8/49) 27.8 (15/54) 41.3 (19/46) Patch 0 (0/63) 2.2 (1/46) 14.6 (7/48) 21.9 (7/32) Values are % (n/total). Data represent percentage of limbs, assessed at each time point, with ultrasound-detected recurrence (No. recurrent/group size). *P 0.002, P , P , P 0.02, P 0.10, P (anti-human von Willebrand factor, DAKO, Carpenteria, Calif) or vascular progenitor cells (anti-human CD34, 26 Zymed Laboratories Inc, No ). The primary antibody was bridged with a biotinylated secondary antibody and ExtrAvidin Peroxidase (Sigma- Aldrich, St Louis, Mo), then visualized with 3,3 -diaminobenzidine tetrahydrochloride. Immunohistochemical sections were counterstained with Hoechst and examined with the use of Normarski differential interference contrast bright-field imaging and fluorescence microscopy. Statistical Analysis All data were held in a FileMaker Pro 6.0 (FileMaker Inc, Santa Clara, Calif) database created by the Vascular Laboratory, Department of Medical and Surgical Sciences, Dunedin School of Medicine, University of Otago. Statistical analysis was performed with the use of StatView (Abacus Concepts Inc, Berkeley, Calif) and Stata Table 6. (StataCorp LP). The Student t test was used for continuous variables, and the 2 test was used for nominal variables. Multiple logistic regression was used to test interactive effects of variables on time point outcome incidence. A Huber-White robust variance estimator was used to account for patient clustering on limb-level recurrence. Statistical significance was set at P The authors had full access to and take full responsibility for the integrity of the data. All authors have read and agree to the manuscript as written. Results Demographics A total of 292 patients were recruited to the study, giving a total of 389 limbs (195 in the control arm and 194 in the patch Secondary End Points at Postoperative Follow-Up in Control Versus Patch Groups 1 Month 6 Months 1 Year 3 Years No. of limbs Control Patch CEAP classes 4 6 (%) Control 83 (43) 69 (43) 73 (41) 50 (42) Patch 66 (34) 56 (34) 54 (33) 37 (35) VCSS S Control Patch * Active ulceration (%) Control 9 (5) 5 (3) 6 (3) 3 (3) Patch 6 (3) 4 (2) 0 0 Deep venous disease, Kistner score 2 4 Control 29 (15) 25 (15) 25 (14) 18 (14) Patch 30 (16) 23 (14) 23 (14) 19 (18) Venous filling index, ml/s Control Patch Limbs with venous filling index 3 Control 23 (12) 21 (13) 23 (13) 20 (17) Patch 16 (8) 14 (9) 20 (12) 16 (15) Values with parentheses are n (%). Only VCSS S and active ulceration rates showed any differences between groups. VCSS S values adjusted for baseline scores are shown in Figure 1. *P 0.05, P 0.03, P (Mann Whitney U test); P 0.02 ( 2 test).

5 70 Circulation July 1, 2008 Figure 1. Changes in VCSS from baseline. Follow-up VCSS less the stocking compliance component (VCSS S) divided by preoperative VCSS S was used to compare patch and control limbs. By 3 years, the patch group had a significantly smaller increase in their severity scores. Bars are 95% CIs. arm) that were assessed preoperatively, operated on according to the randomization protocol, and followed up with at least 1 postoperative assessment at the time of this report. Patient demographics were comparable at baseline between the 2 arms of the study (Table 1). Limb characteristics in each group were also similar except for the frequency of previous varicose vein surgery and severity of the CEAP and VCSS scores (Table 1), although the discrepancy in the latter was found to be entirely attributable to a difference in the swelling and pigmentation subscales. Patient Withdrawals A total of 17 patients (20 limbs) withdrew from the study (Table 1), with comparable numbers and reasons in each group. These all occurred after at least 1 postoperative assessment had been performed, and data pertaining to these limbs up to the time of withdrawal were retained and included in the analysis. SFJ Recurrence There was a highly significant reduction in the incidence of ultrasound-detectable SFJ recurrence in the patch group compared with the controls (Table 2), which was already apparent at 6 months and more so after 1 year (16.3% versus 37.6%; P ). This difference remained at 3 years. If reflux with calf augmentation was the criterion, then recurrence at 1 year was 2.4% versus 10.7% (P 0.003), and a 3-fold difference persisted for 3 years (Table 2). This apparent benefit of placing a patch of Gore-Tex to mechanically impair neovasculogenesis at the SFJ ligation site was observed irrespective of gender, age, history of deep vein thrombosis, or the presence of deep reflux. More importantly, this effect was observed in limbs with more severe disease as measured by venous filling index and in particular VCSS (Table 3) and clinical CEAP (Table 4), which were more commonly randomized to the control group. Although those having had previous surgery did not reach statistical significance (P 0.07), at 3 years the trend mirrored the significant difference observed in patients after primary surgery (Table 5). Secondary End Points Compared with the ultrasound findings, the clinical impact of the insertion of the patch was much more modest at the 3-year follow-up (Table 6). Nevertheless, VCSS scores were significantly lower in patched versus control limbs (Figure 1). Other general markers of venous disease including pain, heaviness, swelling, sclerosis, edema, pigmentation, inflammation, and induration (data not shown) also showed some improvement in the patch arm. Although this was not significant at 3 years, follow-up at 5 years may reveal more obvious clinical differences. The physiological measures of venous function were not different at 3 years. The characteristics of limbs that developed SFJ recurrence in both groups were similar, with number of incompetent perforators, deep reflux, and other markers of venous disease elsewhere in the limb being the same. The presence of a patch did not appear to directly inhibit the neovascular response at the SFJ ligation site. This was apparent histologically at 1 month (Figure 2), with the patch appearing to separate between 2 distinct neovascular networks, one originating superficially from the wounded subcutaneous tissue (Figure 2A and 2C) and the other on the deep

6 van Rij et al Mechanical Inhibition of Angiogenesis 71 aspect from the ligated femoral vein stump (Figure 2B and 2D). There was minimal cellular infiltration and no evidence of vessel penetration into the polytetrafluoroethylene material. All of the channels within these networks immunostained positively for von Willebrand factor. Quantification of 15 consecutive ( 20) microscope fields of the tissue surrounding the patch indicated that 61 of 217 vessels (28%) contained CD34-positive cells (Figure 2E and 2F). The CD34-positive cells were predominantly observed within smaller channels ( 50 m in diameter) and formed both the epithelial lining and mural components of the wall. In the specimen removed 5.5 years after initial patch placement from a recurrent limb, these networks had persisted, enlarged, and more obviously reconnected around the mechanical barrier (Figure 3C) to form recurrent venous channels in continuity with overlying superficial veins (Figure 3A). These channels were now visible on ultrasound. Although all channels were lined by von Willebrand factor positive endothelial cells, CD34-positive cells were extremely rare and when observed were present only within small channels 40 m in diameter. Discussion After a previous demonstration that neovascularization of the SFJ is a major cause of recurrence of venous reflux in the groin, 13 this study investigated the use of a polytetrafluoroethylene patch as a mechanical suppressant of angiogenesis at the SFJ to prevent recurrence. This was prompted by mixed reports of benefits in this approach in venous surgery and the absence of any controlled randomized trials ,27 The use of Figure 2. Polytetrafluoroethylene patch removed 1 month after SFJ vein ligation. Wound (A) and vein stump (B) sides of the patch showing well-formed granulation tissue containing abundant small ( 100- m) blood vessel channels (arrowheads) are shown. The channels become progressively smaller as they approach either side of the patch surface, indicating that these vessels originated from both the wound site and the vein stump. C and D, Note the thinwalled structure of the channels forming the neovascular network. E, All channels were von Willebrand factor positive (arrows). F, Approximately 10% also contained CD34-positive cells. A to D, Verhoeff elastic stain and van Gieson counterstain. Scale bars 250 m (Aand B), 100 m (CtoE),50 m (F). mechanical suppression was also more attractive than the use of potent antiangiogenic agents with their associated risks of nonhealing and wound complications. 15 In a study of venous surgery, the choice of end points for recurrence is critical. Clinical, physiological, and ultrasound recurrences are quite different measures. They have different time frames and variability and do not correlate well for early disease. In this study, the primary end point was the demonstration of any reflux by ultrasonography at the site of SFJ ligation, whether reflux was shown by Valsalva or calf augmentation. Valsalva induces reflux earlier through smaller higher-resistance vessels, whereas calf augmentation reflux becomes apparent when the vessels become larger and have lower resistance to reflux. If the definition of reflux is based on calf augmentation, then the 3-year recurrence is considerably less (Table 2). In that instance, the patch also delayed and reduced recurrence by two thirds (from 14.3% to 4.7%; P 0.02). At present, it is not clear whether the progression from Valsalva to calf-augmented reflux results in more clinically relevant reflux. Therefore, the demonstration of either form of reflux was chosen as the primary end point at 3 years. Because this randomized controlled trial was designed with care, it was surprising to observe some differences in baseline demographics between the 2 study arms given the sample size and the fact that randomization was performed in the strictest of terms. There was no obvious explanation for these disparities apart from random chance. Notably, there was a difference between the groups with regard to whether the limb had previously undergone varicose vein surgery and a significant

7 72 Circulation July 1, 2008 Figure 3. Polytetrafluoroethylene patch removed 5.5 years after vein ligation. Duplex (A) and B-mode (B) ultrasound of the femoral vein (FV) and femoral artery (FA) showing the patch in situ. The patch is most clearly observed in the B-mode image (arrows). Note the recurrence of channels from the ligated vein stump (*), which run along the posterior surface of the patch to its margins. By histology (C), the inferior surface of the patch is shown (approximately matching the region between the arrows in A and B). Note the numerous large (100- to 300- m-wide) vessels (*) running between the patch and remnants of the femoral vein adventitial elastic tissues (stained black, toward the bottom of the image). Toward the mid patch region (C, left side), there was no clear pattern of vessel orientation, whereas at the patch periphery the vessels appeared to align perpendicular to the patch edge. These vessels formed communications with a network of similar-sized channels within the connective tissue overlying the anterior surface of patch. Verhoeff elastic stain and van Gieson counterstain. Scale bar 500 m (C). difference with regard to baseline clinical CEAP grades and VCSS. These differences necessitated subgroup analysis to counter any possible influences of these baseline differences. Nevertheless, it was clear that the overall effect of inserting the patch was similar across all subgroups in that the use of a polytetrafluoroethylene patch approximately halved the rate of ultrasound-detected recurrence of venous reflux in the groin. This randomized controlled study has now shown that there is a highly significant decrease in neovascular recurrence at the SFJ when a polytetrafluoroethylene patch is interposed between the ligated vein stump and the overlying soft tissue. The 3-year recurrence was halved, from 46% in the control arm, which is similar to the recurrence rates reported by Fischer et al 2 and Perrin et al 3 and our earlier work, 12 to just 23%. This reduction was apparent as early as 6 months and continued to 3 years. Such a strongly favorable outcome in an adequately powered prospective randomized trial is difficult to discount even with the imbalance of more severely affected limbs in the control group. If only the more severe limbs in each group (baseline CEAP clinical grades 4 to 6) are compared, then the recurrence after the use of the patch is reduced by 60%. Similarly, 3-year clinical severity scores (adjusted for baseline scores) were significantly reduced in patch-treated limbs. Functional physiological air plethysmography values, in particular the measures of reflux, changed dramatically with surgery in both groups from before surgery to after surgery, and although such tests are useful in determining the severity of reflux and the immediate response to surgery, 28 they do not appear to be sufficiently discriminatory to be useful in monitoring postoperative outcome in the following months. A more complete set of results in the longer future may yet, however, reveal some physiological differences between the 2 groups. Vascular corrosion casts indicate that the neovascular channels associated with SFJ recurrence originate from both the ligated stump of the femoral vein and the surrounding connective tissues within the surgical wound. 13 The histological observations of the 2 tissue specimens recovered from patch-treated limbs in this study support this view, with 2 distinct networks being observed on either side of the patch at 1 month postoperatively. These vessels had a primitive structure, lacking distinct tunica and associated elastic tissue. Many of the newly formed channels surrounding the patch removed after 1 month contained CD34-positive cells, indicating the possible involvement of circulating vascular progenitors in the formation of these networks. 26,29,30 Although the polytetrafluoroethylene patch appeared to at least halve the rate of ultrasound-detected SFJ recurrence, it is interesting to note that 20% of patch limbs still developed SFJ-associated reflux. The patch specimen removed from a patient 5 years postoperatively provides a valuable insight into the probable mechanism of the reflux within the polytetrafluoroethylene patch group. Vessels were observed, by both ultrasound and histology, that originated at the femoral vein ligation site (mid posterior surface of the patch) and extended toward and around the patch margin. These in turn formed connections with vessels on the anterior surface extending to the superficial veins. It is clear that in a subset of patients, neovascularization can circumvent the presence of a physical obstruction in the form of a synthetic patch. That the

8 van Rij et al Mechanical Inhibition of Angiogenesis 73 formation of this more tortuous path delays the appearance of ultrasound-detectable reflux is suggested by the relative catch up between the 6- and 12-month follow-ups. The functional importance of these vessels circumventing the patch may only become clear with longer follow-up. Although they remain multiple and small, reflux will be modest and less likely to be associated with severe venous insufficiency in later years. 28 Other measures that increase the distances to achieve neovascular reconnection, such as stripping of the great saphenous vein, also reduce clinical recurrence. 31 In the meantime, it is encouraging to note that the overall VCSS of patched limbs was significantly lower than that of nonpatch limbs at 3 years. We conclude that the reduction in both disease severity and progression occurs as a direct result of the interference of the polytetrafluoroethylene patch in the formation of a refluxing neovascular network at the site of SFJ ligation. We are aware that the placement of a foreign body has the added risk of infection or misplacement. This complication has been recorded with silicone used as the patch 27 ; silicone should therefore be avoided, and the use of polytetrafluoroethylene, in which this complication has not yet been described, should be preferred. Other approaches to reducing the neovasculogenic response at the SFJ have included occlusion of the lumen of the refluxing great saphenous vein with the use of endovenous procedures that avoid the transection of the vein. Early results seem promising with laser or radiofrequency but not with sclerosants, whether liquid or foam This study is the largest and most thorough of its kind, and these results at 3 years show that use of a polytetrafluoroethylene patch of suitable size is an effective mechanical suppressant of neovascularization at the SFJ and can be safely used to improve long-term outcome of varicose vein surgery. Use of a polytetrafluoroethylene patch is therefore the only method thus far proven clinically in a large randomized trial to improve neovascular SFJ recurrence as an outcome of varicose vein surgery. Acknowledgments The authors wish to thank all of the participants who took part in this study. The assistance of Samantha Dutton in the preparation of the manuscript and Andrew Gray for statistical support is much appreciated. Sources of Funding This study was supported by the Health Research Council of New Zealand. None. Disclosures References 1. Turton EP, Scott DJ, Richards SP, Weston MJ, Berridge DC, Kent PJ, Kester RC. Duplex-derived evidence of reflux after varicose vein surgery: neoreflux or neovascularisation? Eur J Vasc Endovasc Surg. 1999;17: Fischer R, Chandler JG, De Maeseneer MG, Frings N, Lefebvre-Vilarbedo M, Earnshaw JJ, Bergan JJ, Duff C, Linde N. The unresolved problem of recurrent saphenofemoral reflux. J Am Coll Surg. 2002;195: Perrin MR, Labropoulos N, Leon LR Jr. Presentation of the patient with recurrent varices after surgery (REVAS). J Vasc Surg. 2006;43: ; discussion Gruer R, Gordon DS, Gunn AA, Ruckley CV. Audit of surgical audit. Lancet. 1986;1: Sheppard M. A procedure for the prevention of recurrent saphenofemoral incompetence. AustNZJSurg. 1978;48: Negus D. Recurrent varicose veins: a national problem. Br J Surg. 1993;80: Royle JP. Recurrent varicose veins. World J Surg. 1986;10: Glass GM. Prevention of recurrent saphenofemoral and saphenopopliteal recurrence of varicose veins by forming a partition to contain neovascularisation. Phlebology. 1998;13: Jones L, Braithwaite BD, Selwyn D, Cooke S, Earnshaw JJ. Neovascularisation is the principal cause of varicose vein recurrence: results of a randomised trial of stripping the long saphenous vein. Eur J Vasc Endovasc Surg. 1996;12: Kostas T, Ioannou CV, Touloupakis E, Daskalaki E, Giannoukas AD, Tsetis D, Katsamouris AN. Recurrent varicose veins after surgery: a new appraisal of a common and complex problem in vascular surgery. Eur J Vasc Endovasc Surg. 2004;27: Nyamekye I, Shephard NA, Davies B, Heather BP, Earnshaw JJ. Clinicopathological evidence that neovascularisation is a cause of recurrent varicose veins. Eur J Vasc Endovasc Surg. 1998;15: van Rij AM, Jiang P, Solomon C, Christie RA, Hill GB. Recurrence after varicose vein surgery: a prospective long-term clinical study with duplex ultrasound scanning and air plethysmography. J Vasc Surg. 2003;38: van Rij AM, Jones GT, Hill GB, Jiang P. Neovascularization and recurrent varicose veins: more histologic and ultrasound evidence. J Vasc Surg. 2004;40: Herbst RS. Therapeutic options to target angiogenesis in human malignancies. Expert Opin Emerg Drugs. 2006;11: Thornton AD, Ravn P, Winslet M, Chester K. Angiogenesis inhibition with bevacizumab and the surgical management of colorectal cancer. Br J Surg. 2006;93: Peyman GA, Conway MD, Recasens MA, Peralta E, Kertes PJ, Greve MD, El-Dessouky ES. Use of a macular buckle in the treatment of exudative age-related macular degeneration. Ophthalmic Surg Lasers. 1999;30: Hunziker EB, Driesang IM, Saager C. Structural barrier principle for growth factor-based articular cartilage repair. Clin Orthop Relat Res. 2001;391:S182 S Gibbs PJ, Foy DM, Darke SG. Reoperation for recurrent saphenofemoral incompetence: a prospective randomised trial using a reflected flap of pectineus fascia. Eur J Vasc Endovasc Surg. 1999; 18: Bhatti TS, Whitman B, Harradine K, Cooke SG, Heather BP, Earnshaw JJ. Causes of re-recurrence after polytetrafluoroethylene patch saphenoplasty for recurrent varicose veins. Br J Surg. 2000;87: De Maeseneer MG, Giuliani DR, Van Schil PE, De Hert SG. Can interposition of a silicone implant after sapheno-femoral ligation prevent recurrent varicose veins? Eur J Vasc Endovasc Surg. 2002;24: De Maeseneer MG. The role of postoperative neovascularisation in recurrence of varicose veins: from historical background to today s evidence. Acta Chir Belg. 2004;104: De Maeseneer MG, Vandenbroeck CP, Van Schil PE. Silicone patch saphenoplasty to prevent repeat recurrence after surgery to treat recurrent saphenofemoral incompetence: long-term follow-up study. J Vasc Surg. 2004;40: Porter JM, Moneta GL; International Consensus Committee on Chronic Venous Disease. Reporting standards in venous disease: an update. J Vasc Surg. 1995;21: Rutherford RB, Padberg FT Jr, Comerota AJ, Kistner RL, Meissner MH, Moneta GL. Venous severity scoring: an adjunct to venous outcome assessment. J Vasc Surg. 2000;31: van Rij AM, Hill G, Gray C, Christie R, Macfarlane J, Thomson I. A prospective study of the fate of venous leg perforators after varicose vein surgery. J Vasc Surg. 2005;42: Torsney E, Mandal K, Halliday A, Jahangiri M, Xu Q. Characterisation of progenitor cells in human atherosclerotic vessels. Atherosclerosis. 2007;191:

9 74 Circulation July 1, De Maeseneer MG, Vandenbroeck CP, Lauwers PR, Hendriks JM, De Hert SG, Van Schil PE. Early and late complications of silicone patch saphenoplasty at the saphenofemoral junction. J Vasc Surg. 2006;44: Jiang P, van Rij AM, Christie RA, Hill GB, Thomson IA. Venous physiology in the different patterns of recurrent varicose veins and the relationship to clinical severity. Cardiovasc Surg. 2000;8: Asahara T, Murohara T, Sullivan A, Silver M, van der Zee R, Li T, Witzenbichler B, Schatteman G, Isner JM. Isolation of putative progenitor endothelial cells for angiogenesis. Science. 1997;275: Shi Q, Rafii S, Wu MH, Wijelath ES, Yu C, Ishida A, Fujita Y, Kothari S, Mohle R, Sauvage LR, Moore MA, Storb RF, Hammond WP. Evidence for circulating bone marrow-derived endothelial cells. Blood. 1998;92: Winterborn RJ, Foy C, Earnshaw JJ. Causes of varicose vein recurrence: late results of a randomized controlled trial of stripping the long saphenous vein. J Vasc Surg. 2004;40: Myers KA, Jolley D, Clough A, Kirwan J. Outcome of ultrasound-guided sclerotherapy for varicose veins: medium-term results assessed by ultrasound surveillance. Eur J Vasc Endovasc Surg. 2007;33: Pichot O, Kabnick LS, Creton D, Merchant RF, Schuller-Petroviae S, Chandler JG. Duplex ultrasound scan findings two years after great saphenous vein radiofrequency endovenous obliteration. J Vasc Surg. 2004;39: Theivacumar NS, Dellagrammaticas D, Beale RJ, Mavor AI, Gough MJ. Fate and clinical significance of saphenofemoral junction tributaries following endovenous laser ablation of great saphenous vein. Br J Surg. 2007;94: CLINICAL PERSPECTIVE The major frustration in the treatment of varicose veins is recurrence. After surgery, this occurs in 40% to 60% after 5 to 10 years and even earlier after sclerotherapy. Although there are a number of contributing factors, neovascularization with regeneration of venous channels is most common, as shown by ultrasound imaging and measurement of reflux. Prevention of neovascular venous reconnection by placing a mechanical obstacle to this process at the site of surgery has been suggested. This study tested this in a sizable surgical randomized controlled trial with 398 limbs, in which the active treatment group had a patch of polytetrafluoroethylene placed over the site of ligation of the saphenofemoral junction. Follow-up after 3 years showed the ultrasound-detectable recurrence at the saphenofemoral junction to be consistently reduced by 50% whether the surgery was for severe venous disease or for previous recurrence. However, not all neovascular reconnection was prevented. Ultrasound imaging in these instances demonstrated the newly formed vessels to skirt around the patch. Histological examination of the excised polytetrafluoroethylene patch, as shown in this study, confirmed the ability of neovascular vessels to track around the patch and reconnect to the superficial venous system. Whether these more circuitous neovascularizations will become clinically significant is not clear yet, but the potential clinical benefit of the reduced neovascularization is evident from the lower Venous Clinical Severity Score in the patch group after 3 years. However, the more reliable establishment of the clinical utility of the patch will require the 5-year assessment to be completed.

10 Mechanical Inhibition of Angiogenesis at the Saphenofemoral Junction in the Surgical Treatment of Varicose Veins: Early Results of a Blinded Randomized Controlled Trial André M. van Rij, Gregory T. Jones, B. Geraldine Hill, Mohammad Amer, Ian A. Thomson, Ross A. Pettigrew and Stephen G.K. Packer Circulation. 2008;118:66-74; originally published online June 16, 2008; doi: /CIRCULATIONAHA Circulation is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX Copyright 2008 American Heart Association, Inc. All rights reserved. Print ISSN: Online ISSN: The online version of this article, along with updated information and services, is located on the World Wide Web at: Data Supplement (unedited) at: Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published in Circulation can be obtained via RightsLink, a service of the Copyright Clearance Center, not the Editorial Office. Once the online version of the published article for which permission is being requested is located, click Request Permissions in the middle column of the Web page under Services. Further information about this process is available in the Permissions and Rights Question and Answer document. Reprints: Information about reprints can be found online at: Subscriptions: Information about subscribing to Circulation is online at: