Références bibliographiques

Neovascularisation : quelle est la responsabilite du chirurgien ?

Catégorie : Resultats Complications Recidives
Phlebologie 2008;37:134-141
Denis Creton, MD, EC A Paré, rue A Paré, 54100 F, Nancy, France - Telephone: 33 3 83 95 54 00; fax: 33 3 83 95 54 23; e-mail: Cette adresse e-mail est protégée contre les robots spammeurs. Vous devez activer le JavaScript pour la visualiser..


Plusieurs causes induisent la néovascularisation: facteur de croissance endothélial vasculaire, inflammation, hypertrophie de vaisseaux pré-existant (veine lymphoganglionnaire), revascularisation d’hématomes et différence de pression. Les 3 dernières dépendent directement de la façon dont est réalisée l’opération, donc du chirurgien.

Concernant les varices primaires l’interposition d’un patch ne supprime pas la néovascularisation. Il est très important de ne faire une crossectomie que lorsqu’il y a une incontinence des valves terminales et préterminales. Dans ce cas l’incision doit être la plus petite possible et réalisée avec une dissection minimum, avec une tumescence afin de réduire le saignement et les hématomes. Quand la valve terminale est continente il est préférable d’utiliser une technique endovasculaire.

Concernant la récidive il est important d’éviter les reprises chirurgicales inutiles, le meilleur choix est la sclérothérapie peropératoire. Je suis convaincu que la néovascularisation est produite par les complications, elle-même induites par les dissections extensives. L’interposition d’un patch n’est certainement pas aussi nécessaire que nous l’avons prétendu 10 ans auparavant. Eviter tout traumatisme pendant l’intervention est certainement beaucoup plus important : « faire moins au niveau de l’aine pour faire mieux dans le risque de récidive et de re-récidive »


There are different causes of neovascularisation : vascular endothelial growth factor, inflammatory process, hypertrophy of pre-existant vessels (lymph node veins), haematomas revascularisation, and high difference in pressure. The latter 3 depend directly on the way the operation is performed hence on the surgeon.

Concerning the primary varices patch saphenoplasty do not abolish neovascularisation. It is of first importance to perform a flush ligation only in case of incontinent terminal and pre-terminal valves. In that case an incision as small as possible must be carried out with minimal dissection and with tumescence to limit the bleeding and hematomas. When the terminal valve is competent it is better to choose endovascular technique.

Concerning re-do it is mandatory to avoid useless re-do, the best choice is foam sclerotherapy. I am convinced that neovascularisation are produced by the complications which are induced by a large dissection. The barrier technique are, probably, not as useful as we was told 10 years ago. The lack of aggressiveness during the operation is certainly far more important: “Doing less in the groin to do better for the recurrence and re-recurrence”

In the context of varicose recurrence after great saphenous vein (GSV) surgery, the term neovascularisation describes the phenomenon of formation of new venous channels between the saphenous stump on the common femoral vein (CFV) and the residual GSV or its tributaries.

Despite the fact that this frustrating phenomenon is frequently encountered by each vascular surgeon, even by those well trained, after apparently having performed a correct and complete operation its nature and pathophysiology (hence its prevention) are poorly understood and the subjects of intensive controversial discussion [1, 2]. Neovascularisation is frustrating for a good well trained and skilled vascular surgeon, although surgeon is sometimes responsible for the development of neovascularisation. The surgeon is not always an “innocent bystander”

The problem is to answer these 2 questions: how and why does neovascularisation happen and how can this development be avoided or limited ?


In 1861, Langenbeck [3] described for the first time in detail the formation of new veins after ligation. In his article we can read a description of neovascularisation: “ln one case of very large varix of the great saphena in a young man I had extirpated the enlarged vein in the length of three inches and ligated the upper and lower ends. One year later I found, in the region of the scar tissue of the extirpation, a new vein channel of the thickness of the quill of a crow's feather, which again joined the both ends of the fully functioning saphena”

Twenty years ago a description was done by Glass in 1987 [4]. In his work a series of patients with venous ulceration due to GSV insufficiency were treated in stages. First, a transection of GSV was performed in the lower part of the thigh. In a second stage, individually timed for each patient by healing of the ulcer, the sapheno femoral junction (SFJ) was ligated in combination with stripping of the trunk. During the same operation the ends of the vein at the site of the original transection in the lower thigh were excised together with the tissue intervening between them. At 6 week follow-up (FU) he could see a recanalisation of the thrombus, at 18 week new vessels and at 40 week the continuity was completely restored. He hypothesized that one of the important triggers for restoration of continuity was a large pressure difference in the proximal and distal parts of the vein.


First of all, the physiopathology of neovascularisation seems to have something in common with the physiopathology of the varicose veins [5]. The physiopathology of varicose veins involves genetic mechanisms, reduced apoptosis, increasing in collagen and decrease in elastin, decreased endothelin β receptor, vascular endothelial growth factor (VEGF) and inflammatory process. Only the 2 last one are concerned in the neovascularisation.


It is known that plasma levels of VEGF are higher in chronic venous disease (CVD) patients than in normal controls [6]. Such stump-related neovascularisation might originate from hypoxia-induced activation of endothelial cells distal to the stump ligature, which could be mediated by growth factors [7]. VEGF are the strongest stimulators for angiogenesis. In a study of Rewerk [8] patients were distributed into a control group, a primary varicose vein group and a recurrent varicose vein group. Thin-walled veins embedded in scar tissue (neovascularisation) were stained immunohistochemically with antibodies against VEGF. Presence and role of VEGF was demonstrated by a higher staining intensity seen in the recurrence group.

Inflammatory process

Strong evidence has now been gathered to indicate that chronic venous insufficiency (CVI) is an inflammatory disease. Actually, ultrastructural and immunohistochemical studies of valves and the venous wall have revealed leukocytes adhering and transmigrating into the venous wall. The suggestion that inflammation may be involved in CVD comes from the evidence for elevated endothelial permeability [9]. Inflammation that is involved in postoperative cicatrisation is also involved in CVD. An analysis may lead us toward the trigger mechanisms for inflammation in CVD. An inflammatory process may be involved particularly at the cellular and molecular levels and may be the origin of the earliest manifestations of varicose veins

Role of the venous pressure in neovascularisation development

The venous pressure could be an inductor of neovascularisation by the way of difference in pressure, variation in pressure, irregular displacement of blood and specially by the fluid shear stress which is the tangential force induced by moving blood acting on the endothelial surface and a function of the velocity gradient of blood near the endothelial surface [10].

The difference in pressure already induces vascular remodelling as demonstrated by El Wajeh [11]. The venous channels that develop at the previously ligated SFJ may represent adaptive dilatation of pre-existing venous channels (vascular remodelling), probably in response to abnormal haemodynamic forces. Tissue samples from the region of the previously ligated SFJ were taken from 14 limbs with recurrent varicose veins and 9 control limbs. Tissue samples were analysed histologically, and immunohistoligically for S100, a neural marker, and Ki-67 (Mib1), a marker of endothelial proliferation. No significant difference was found between the venous recurrence and control groups in respect to histological features.

The difference in pressure induces also inflammation as described by Takase [12] who demonstrated that short-term venous pressure elevations caused elevated microvascular permeability (blue escape across the endothelium was evidence for elevation of venular permeability and extravasation of red cells across postcapillary venules was evidence of inflammation)

Shear stress could be an inductor of neovascularisation. Owatverot [13] demonstrated that low flow or flow disturbances (that is typically located in the groin area), especially if they involve instances of reversed flow direction with forward and backward shear, make cells more susceptible to inflammatory mediators. Stretching of endothelial cells and smooth muscle cells also has a direct effect on many aspects of the endothelial biology including synthesis and release of many inflammatory molecules such as leukotrienes, prostaglandin, bradykinin, free oxygen radical, and cytokines. It is typically what happens in the area of the groin.

Shear stress increases the inflammatory activity. Many authors demonstrated that both VEGF expression and its receptor expression are up-regulated under variations of blood shear stress and with the inflammatory reaction [14-16]

Role of pre-existant vessels : lymph node veins

The hypertrophy of pre-existing vessels which provide a connection between the femoral vein, the abdominal sub cutaneous vein, pelvic veins, inguinal veins and the residual saphenous trunk is obvious. These developments are a part of the neovascularisation [17].

Revascularisation of hematomas

Several recent studies emphasized the role of the hematomas.
The first study concerned 70 strippings [18]. At 1 week 54 legs presented with small haematomas and 16 legs with large haematomas. At 1 year only the 16 who presented previously with large hematomas presented with revascularisations. Among these, 12 had a partial revascularisation (6 distal third, 5 distal half, 1 almost complete) and 4 a complete revascularisation shown as multiple channels, in continuity with the distal segment of native SV and with reflux.

The second study concerned 102 cases [19]. At 2 month-FU, 6 patients had low velocity and long duration reflux in the haematomas in the saphenous tract which was seen to fill from either a thigh perforator or tributary. At 1 year-FU this haematoma tract had extended proximally towards the groin in 3 patients, 2 having connected to the CFV.

The aggressiveness of the operation could be an inductor of the neovascularisation by means of hematomas. In a study of Earnshaw [20], at one year-FU there were a higher rate of new neovascularisation in the group of patients operated on for re-do surgery than in the group operated on for primary varices. It is known that re-do surgery is more difficult and more aggressive than primary operation. Hematomas are more likely encountered in re-do surgery than in primary operation.

Among these different causes of neovascularisation (Table I) there are, at last, 3 which depend directly on the way the operation is performed that is to say it depends on the surgeon. Firstly there are “inflammatory cicatrisation” and “hematomas revascularisation” and secondly the “high difference in pressure” which depends on the surgical decision if for example a flush ligation alone is proposed without the ablation of the incompetent trunk which is an obvious cause of neovascularisation.


Macroscopic examination [21-23] revealed several lumens in an irregular mass of vein tissue and cords or bands traversing the lumen, which suggested that the vessels were newly formed and not pre-existing. Large lymph nodes were often in close proximity to them. Neovascular channels were covered by a simple squamous endothelium overlying a medial layer consisting of two to five layers of vascular smooth muscle. They lacked elastic fibers, had no distinct intimal medial boundary, and no distinct adventitia. The histology confirmed the microscopic findings: an irregular vessel wall with a varying thickness at different points of the circumference, often with several lumens. Also typical was the presence of many small vessels close to the newly formed vessel and in neighbouring lymph nodes. A resin casting provided by van Rij [23] showed a connecting network of aboundant tortuous vessels.

Histological examination of the venous tissue blocks, excised during groin re-explorations, showed neovascularisation in 27 blocks, characterized by vein tortuosity, small size, and mural asymmetry and lack of intramural nerves on immunohistologically S 100 stained sections. Nyamekye [24] had defined 3 negative criteria: no intramural nerves, no three layered wall structure and lack of lumen regularity. Stücker [25] performing 91 histological examinations of excised venous tissue block can demonstrate the presence of neovascularisation and defined a positive criterion as scar tissue that must always surround a newly formed vein. He clearly defined the characteristics of the different elements seen in the area of neovascularisation.

To resume :


Primary varices

Unintentionnally the study of Nyamekye [24] gives us a part of the answer to this question. Actually, in these 28 histological analyses the 27 neovascularisations was observed only in the 27 previous flush ligations, no neovascularisation was observed in the previously untouched SFJ. The conclusion may be “to avoid neovascularisation don’t touch the SFJ”

a) The relationship between the residual stump and neovascualriwation has long been considered as evidence and historically dogma has maintained that to remove the stump would do away with the need for neovascularization, which has been the subject of numerous studies.

The concept was to obtain the elimination of the GSV stump with a vascular side clamp on the SFJ with a short venotomy that can be closed with a two-layer running suture. The incidence of recurrent reflux in the above-described technique in the long term has not yet been reported. Up to now, only preliminary results have been published. Some having had results, such a serious intervention can’t be reasonably suggested for a first cross-section.

b) Based on the idea of hiding or destroying the stump endothelium, it has been proposed to ligate the stump with a non resorbable suture and then bury the stump with a running polypropylene suture [26]. Hundred and fourteen duplex imagings at 2-year FU shown that recurrent reflux in the groin was significantly reduced from 11 to 3% by oversewing the ligated SFJ with a running polypropylene suture. But in a prospective randomized study with follow-up up to 5 years after surgery, Haas [27] could not confirm any beneficial effect of such inverting suture of the stump endothelium. On duplex ultrasound control, they still found neovascular vessels in 9% of 279 limbs operated on with this technique.

c) Other investigators have chosen to focus only on the stump endothelium, destroying it with chemical or heat cauterization, or in some instances reducing the amount of endothelium exposed by placing a second ligature near the free end of the GSV stump, all without conclusive results.

d) Spatial separation of the stump is probably the most important way to reduce the rate of neovascularisation and it is something the surgeon can easily do. Actually, it also seems important to achieve greater spatial separation between the deep veins underlying the fossa ovalis and the superficial venous system, by careful ligation of all tributaries in the groin beyond their junction and stripping the thigh incompetent portion of the GSV. The bad role of the residual incompetent trunk is proved by a study of Jones [28]. Two groups of patients were re-explored at 2-year follow-up. The 2 groups presented with quite the same rate of inguinal neovascularisation but only the group with residual incompetent trunk had clinical recurrences. The neovascularisation becomes incompetent and associated with varicose vein recurrences only when associated with an incompetent residual trunk.

Duplex and operative findings were recorded prospectively in a consecutive series of 500 limbs undergoing surgery for recurrent varicose veins [29]. Neovascularisation was identified on Duplex scanning in 41 limbs. In 27 of these, surgical exploration revealed a residual GSV stump with 1 or more significant tributaries. Each of the remaining 14 limbs had a residual incompetent thigh GSV. The conclusion was: neovascularisation is either associated with persistent reflux in the GSV stump or thigh GSV, or both. Neovascularisation alone is not a cause of recurrent varicose veins. There was no isolated neovascularisation.

e) Barrier techniques
Interposition of a physical barrier (prosthetic or anatomical) between the ligated GSV stump on the CFV and the surrounding superficial veins is a great old idea which could theoretically prevent tiny neovascular veins developing at the stump from connecting with superficial veins in the groin and thigh.

f) Avoid the useless flush ligation.
It is of first importance to perform a flush ligation only in case of incontinent terminal and pre-terminal valves. Saphenous trunk reflux with continent SFJ represents 50% of the cases [37, 38].
When a flush ligation is necessary an incision as small as possible must be carried out with tumescence to limit the bleeding and hematomas with minimal dissection, no aggressiveness and a running suture to close the fascia cribriformis. When the terminal valve and the tributaries must be protected a section below the triburaries is performed (Photo 1) or only a section of the anterior accessory saphenous vein as well (Photo 2).
The endovascular ablation of the GS trunk with radiofrequency obliteration (RFO) (ClosurePlus®) reduced the rate of neovascularisation. In this comparative study [39], 6 of 55 (11%) legs in the open surgery group showed clear evidence of tortuous refluxing veins related to the SFJ. None of the 55 in the RFO group showed any neo-vascularisation at the SFJ (Fischer exact test P = 0.028).
Although a significant difference could not be demonstrated in the randomized EVOLVeS study [40] (endovascular ablation of the GS trunk with RFO ClosurePlus® versus stripping) at 2-year FU, 36 limbs with RFO show one neovascularisation and 29 limbs with stripping show 4 neovascularisations

Redo surgery

In the re-do surgery, because of the destruction of the cribriform fascia only prosthetic barriers are usable. Because of some particular difficulties of re-do surgery it is of the highest importance to try to diminish the risk of re-neovascularisation during re-do operation. Effectively, another re-do in the groin could be quite impossible. Consequently we must keep in mind a re-do must be the last one.

In a recent study De Maeseneer [41] found that patch saphenoplasty significantly improved the clinical and duplex ultrasound results five years after repeat surgery. Recurrent thigh varicosities were observed in 58% of limbs without a patch and in 26% with a patch. After 5 years Duplex scans revealed important neovascular vessels in 45% in the group without a patch versus 9% in those with a patch (P=.0007). The use of a silicone patch at repeat operation to treat recurrent varicose veins is associated with a lower incidence of neovascularisation. In re-do procedures, it is of particular importance to fix the patch well to the CFV, as there is usually no cribriform fascia left to cover the patch and keep it where it should stay (covering the saphenous stump).

It is surprising to compare my personal results using a very small patch (1/1.5 cm) (4% of re-neovascularisation at 5 year follow-up) [43] with those of other good surgical teams [41, 42]. It is known of their comprehensive large experience in re-do surgery, (31% at 2 year follow-up for Winterborn, 9% and 22.3% for De Maesseneer). The differences are also in the rate of complications (0.7% in my personal study with no lymphatic and hematoma complication and only one postoperative abscess, 32.5 and 19% for Winterborn and De Maesseneer respectively). In fact, in these studies many complications are reported, 13%, 6% infections, 6.5%, 3% hematomas, 13%, 10% lymphatic complications and 0.9% stenosis of the common femoral vein [44]. Regarding these studies and Earnshaw's [35] as well and specially the pictures and photographs of the operations we can see that the dissection was always large with a wide exposure of the femoral vein. This can explain exceptional late postoperative stenosis of the common femoral vein reported in recent series [44].

I am convinced that neovascularisation are produced by the complications which are induced by a large dissection. It is likely in my study [43] the low rate of neovascularisation (4%) was not due to the patch but to the low rate of complications, itself due to the small dissection.

Consequently since 2000 we have still reduced the aggressiveness of re-do operations by doing only a ligature of the stump with non absorbable suture through the same small lateral approach (Photo 4). Usually the ligature of the stump is easy because there is always continuity between the adventitia of the femoral vein and the residual stump. Finally, intra-operatively, I securely inject foam in the distal part of the ligated stump from top to bottom (Photo 5) or by the way of a thigh vein from bottom to top (Photo 6). From top to bottom the puncture of the vein must be performed carefully with a 25 gauge needle and a prolongator in order not to move. The reflux is verified. During the injection we can see the whitening of the injected vein. The foam is prepared preoperatively according to Tessari method [45] with 2 sterile syringes without silicon and a 3-way stopcock. The aim of the surgical ligature of the stump is to secure and make the foam sclerotherapy effective avoiding the passage of the foam in the common femoral vein.


Regarding the neovascularisation the surgeon is not an innocent bystander. There are some gestures, habits which can be unfortunately responsible for the appearance of neovascularisation. Do not touch the SFJ if we want to avoid any risk of neovascularisation! Firstly, after a comprehensive ultrasound scanning of the SFJ doing nothing when there are no incompetence of terminal and pre-terminal valves. When there is only an incompetent preterminal valve it is better to choose endovascular technique like EVL or RFO. When a flush ligation is decided it is mandatory to avoid useless dissection, wrenching, bruising, retractor gapping, bleeding, electrocoagulation, postoperative haematomas and other complications. The larger the incision the worse the neovascularisation. On the contrary, extensive resection of all incompetent residual trunks and varices on the thigh are mandatory. Concerning re-do surgery it is mandatory to avoid useless re-do, the best choice is foam sclerotherapy. The barrier technique are, probably, not so useful as we was told 10 years ago. The lack of aggressiveness during the operation is certainly far more important: “Doing less in the groin to do better for the recurrence and re-recurrence”


Table I

The role of the surgeon. Among the different causes of neovascularisation there are 3 which depend directly on the surgeon (2, 3, and 6).

Photo 1

When the terminal valve and the tributaries must be protected a section below the triburaries is performed.

Photo 2 & 3

In the case of incompetent anterior accessory saphenous vein with competent terminal and preterminal valves only a section of the anterior accessory saphenous vein is necessary.

Photo 4

the dissection space was made as obliquely and as small as possible.

Photo 5

A ligature of the stump with non absorbable suture is carried out through the same small lateral approach.

Photo 6

The injection of foam in the distal part of the ligated stump from top to bottom is carried out carefully with a 25 gauge needle and a prolongator in order not to move.

Photo 7

The injection of foam is carried out by the way of a thigh vein from bottom to top.