Abstract
Endovenous recanalization with percutaneous transluminal angioplasty and stenting in post-thrombotic syndrome patients with iliocaval obstruction is a treatment modality quickly gaining popularity. Studies show good patency and clinical success rates. If the obstruction extends distally, below the inguinal ligament, stenting remains controversial. Without adequate inflow, the patency of stented iliocaval segments drops dramatically. This suggests that treatment of diseased common femoral, femoral and profunda femoral veins is required to ensure adequate inflow. Endophlebectomy, the removal of synechiae and septae from the common femoral vein, is a viable option in these cases. Another option, which can be done concurrently with the endophlebectomy, is the creation of an arteriovenous fistula. Selecting patients for these interventions however remains difficult, as precise preoperative prediction of inflow into the stented segments is difficult. In this paper we describe our experience in using duplex ultrasonography, magnetic resonance venography and conventional venography to assess the patency of the inflow trajectory. We believe this approach is essential in dealing with cases of complex post-thrombotic disease extending below the inguinal ligament. There is a great need to establish criteria to accurately assess pre- and postinterventional flow through treated vein segments.
Introduction
Although arterial percutaneous transluminal angioplasty (PTA) and stenting were first described in 1964 by Charles Dotter and in 1985 by Palmaz et al.1–3 respectively, and the first endovenous stenting already described in 1986, endovascular recanalization and stenting of the deep venous system has only recently gained widespread interest and use. This is partly due to disappointing early results, the lack of interest in the post-thrombotic syndrome (PTS) and the unawareness of the relationship between PTS severity and the location and extend of deep venous thrombosis. 4 Currently the number of patients treated with deep venous PTA and stenting is increasing exponentially, because of published good clinical results, improved patency rates and low mortality and morbidity rates.5–9 Neglén et al. 7 showed good overall long-term results, but lower patency rates in cases of post-thrombotic disease compared with non-thrombotic disease. This might be explained by the changed anatomy of post-thrombotic veins, characterized by rigidly scarified vessel walls and intraluminal synechiae and septae, which also create a thrombogenic environment. 10 This thrombogenic environment can be treated, both pre and poststenting, by adequate anticoagulation, either with Coumadins, low-molecular-weight heparins or new oral anticoagulants.11–13 In order address the changed anatomy one needs to recanalize a post-thrombotic diseased segment. Various guidewires are needed to pass the synechiae and septae after which an angioplasty, mostly with high pressure balloons, is needed before stent placement. 14 After stenting the restored patency of the treated caval and/or iliac vein segments should redirect flow through this normal anatomical route and decreases or resolve flow via previously developed collateral veins. This technique has been described previously. 14 However, in case of femoral involvement in post-thrombotic disease, venous inflow into the to-be-stented iliocaval segments is frequently inadequate. Obstruction or occlusion in the common femoral vein can impair outflow of the femoral, deep femoral and great saphenous vein into the stented segment, causing a low flow through the stented segment. As low flow conditions relate strongly with thrombosis and re-occlusion, this needs to be addressed. However, stent placement in the common femoral and femoral segments is still under debate. 15 Too effectively, minimally invasively guarantee adequate inflow from all three, femoral, deep femoral and great saphenous, veins into the stented segment above the saphenofemoral junction (SFJ) a branched stent has to be developed. Without this theoretical opportunity there is always the chance of not being able to cross the different intraluminal planes created by the septae in the common femoral vein. In such a case one would, e.g. only recanalize a portion of the vein and inflow might only be restored from one of the three veins connected to the common femoral vein. Besides this technical problem, we know from arterial stenting across joints that stent fractures may occur, 16 although Neglén et al. 15 showed that stent elongation across the inguinal ligament is safe using Wall stents (Boston Scientific, Natick, MA, USA). Other authors are generally less aggressive in stent placement across joints, especially when Nitinol stents are used, which have a greater tendency to fracture when stressed.
Alternatively the obstruction or occlusion of the common femoral vein can be addressed in a surgical manner, which is termed endophlebectomy, endovenectomy or desobstruction.10,17,18 This technique was first described by Gloviczki in 1999 in relation to venous bypass surgery. Raju et al. 19 showed feasibility in 1999 in cases of axillary vein transfer and Puggioni et al. 10 in 2004 in cases of deep venous reconstruction. Summarized endophlebectomy entails: the post-thrombotic damaged common femoral vein is longitudinally opened and the intraluminal senechiae and masses are removed. Any post-thrombotic lesions obstructing outflow from the femoral, deep femoral or great saphenous vein need to be removed and inflow into the common femoral vein restored. The venotomy can be primarily closed or by using a patch, in case of a narrowed lumen. This should guarantee physiological inflow into the stented segments. It is vital that the distal end of the stent should at least extend into the ‘cleaned’ common femoral vein segment. After restoration of the physiological flow combined with adequate anticoagulation, there still might be a highly thrombogenic local environment at the level of endophlebectomy or stents, caused by both surgical damage to the vessel wall and the stent material. This is why an endophlebectomy is frequently combined with the creation of an arteriovenous fistula (AVF) in order to increase flow in the treated area and theoretically reduce the chance of thrombosis.10,18,20–22 In our center a loop with a 6 mm externally supported polytetrafluoroethylene (PTFE) graft of about 6–8 cm long is created between the common femoral artery and vein. We learned that native fistulae are too small or when created with a diameter larger than 5–6 mm they need to be closed (surgically) after six months because of persistent swelling of the leg. The advantage of using a PTFE loop is that it guarantees a high flow rate in the first months and that it can be percutaneous closed, e.g. with an Amplatzer Vascular Occluder (AGA Medical Corporation, Plymouth, MN, USA).23,24
When PTA, stenting, endophlebectomy, AVF and adequate anticoagulation are combined in these complex and extensive cases of PTS, all three facets of Virchow's triad (blood, flow and vessel) are addressed.
To our knowledge no effective and validated diagnostic tests, and their cut-off values, are currently available to pre- or peri-intervention quantify haemodynamic parameters associated with patency rates after deep venous stenting in order to help us decide if there is an indication for an endophlebectomy with or without an AVF. Ideally such a test should measure flow at the in- and outflow site of the stented segment. At this moment only anatomical parameters, acquired by or duplex ultrasonography (DUS), ascending venography, magnetic resonance venography (MRV), computed tomography venography (CTV) or intravascular ultrasonography (IVUS) are used to indicate if patients are eligible for endophlebectomy and/or AVF25,26 These parameters include location and extend of post-thrombotic lesions, percentage of lumen reduction, and number and extend of collateral veins. In this paper we described current literature and our own experience in the challenge of diagnosing (in)adequate venous flow in post-thrombotic deep veins and indicating patients for ancillary surgery combined with or after PTA and stenting.
Imaging characteristics of patients undergoing endophlebectomy and arteriovenous fistula
Patients with chronic venous disease in which a deep vein obstruction is suspected require adequate analyses of the deep venous system. Besides clinical examination, e.g. abdominal wall venous collaterals and in some centers functional tests, e.g. plethysmography, a crucial part of this analysis is imaging.
27
The initial imaging modality of choice should be duplex ultrasound.28,29 The primary objective is the assessment of the presence and the extent of any deep venous reflux, signs of recanalization and/or flow impairment, luminal narrowing and vein wall thickening. The veins that need to be examined are the popliteal, femoral, deep femoral and common femoral vein. Additionally, collaterals in the groin and signs of obstruction in the pelvis should be identified. When there are signs of obstruction with collateralization in the groin and/or pelvis, additional imaging should be performed. The aim of this additional imaging should be to accurately assess the extent of the obstruction, reconstruct the anatomy of the inferior vena cava and the common and external iliac veins and assess the routes of collateralization to assess the extent of the diseased tract. Furthermore, imaging should quantify the chronic, mostly post-thrombotic, vein abnormalities in both the inferior vena cava and iliac veins, as well as the common femoral, deep femoral and femoral vein. In a non-invasive setting, the choice of examination is either MRV or CTV. In our experience, both techniques are capable of identifying obstruction and collateralization, but only MRV is capable of accurately detecting intraluminal post-thrombotic changes, which affect our treatment decisions as we described with the LOVE score (Figures 1 and 2).
25
In our treatment approach, conventional venography and IVUS are reserved for those patients in whom we have decided to perform an endovascular recanalization and stenting, since these are invasive procedures that require a hospital admission.
Magnetic resonance venography image in the transverse plane showing the chronically diseased common femoral vein (CFV) in a patient with a post-thrombotic syndrome with an occluded right external and common iliac vein. The arrow marks trabeculations within the CFV as a result of poor recanalization which suggest impaired inflow to the iliac veins Magnetic resonance venography image in the coronal plane showing a chronically diseased common femoral vein and external iliac vein with trabeculations. Multiple trabeculations can be identified suggesting flow impairment over these segments

In chronic deep venous obstruction, roughly four patterns of disease can be identified. The first pattern is isolated obstruction at the level of the common iliac vein, in virtually all cases associated with compression (May-Thurner syndrome). In these patients there are usually no, or minimal post-thrombotic changes in the external iliac and femoral veins. The second pattern is obstruction of both the common and external iliac vein. In these patients the patterns of collateralization are more extensive and usually the groin shows signs of collaterals with relatively mild or unscarred femoral veins. The third pattern shows occlusion of the common and external iliac vein with involvement of the femoral veins. Collateralization is prominent in these cases, but more importantly severe scarring of the common femoral vein, often combined with impaired recanalization of the femoral and deep femoral vein, is seen. The fourth pattern shows involvement of the inferior vena cava, usually with extensive paralumbar collateralization and hyperplastic azygos and hemiazygos veins. In these cases disease in the iliac and femoral veins can either be symmetrical showing pattern 1, 2 or 3 on both sides or asymmetrical, usually related to the occurrence of deep vein thrombosis in the past.
Patients who present with isolated common iliac vein obstruction can be effectively treated with endovascular recanalization and stenting.5–8,14,21,22 Also in patients with a common and/or external iliac vein obstruction, without common femoral vein involvement below the SFJ endovascular recanalization and stenting can be performed without additional surgical intervention.
In patients with a common and/or external iliac vein obstruction with common femoral vein involvement below the SFJ we perform an endophlebectomy In most cases, in addition to the involvement of the common femoral vein, the femoral and deep femoral vein are also involved in which case we perform an endophlebectomy with an AVE Lack of inflow into the stented segment, due to the pre-existing inflow impairment from the deep femoral and femoral vein, combined with an increased thrombogenetic post-thrombotic common femoral vein cause these patients to be at increased risk for in-stent thrombosis due to lack of flow.
In patients where the IVC is involved we decide on our treatment approach based on the extent of the involvement of the vena cava as well as the iliac and femoral axis. In isolated infrarenal IVC involvement an antegrade approach from either the groin or femoral vein usually suffices. In those cases where the suprarenal IVC is involved, access from the jugular vein can be helpful in which case patient positioning and procedure planning is adjusted. By utilizing all our preoperative imaging information as described above we can accurately select patients for what we believe will prove to be the most optimal approach of treating chronic venous obstructions.
At the end of the procedure it is crucial to assess the preferential outflow pattern. In patients with isolated iliac disease, collaterals identified preintervention need to show signs of decreased flow and preferably stasis with primary outflow through the recanalized iliac vein. If collateral outflow persists after common iliac vein stenting, this usually is a sign of inadequate stent extension caudally The same applies in patients with endophlebectomy with or without AVF creation. Primary outflow should lead into the recanalized iliac tract, not retrograde into the femoral veins or collaterals. It is our conviction that this should also be treated with stent extension, if needed, across the inguinal ligament into the common femoral vein.
Follow-up imaging is primarily aimed at assessing stent and AVF patency, which is most easily done with duplex. Additionally we perform abdominal radiographs in four projections of the stented trajectory (Anterio-Posterior, lateral, left and right anterior oblique). These radiographs are to identify stent kinking, migration, fracture or changes in orientation. Only in selective cases additional imaging is required. Since haemodynamic information is crucial, the next step after duplex is conventional venography which allows us to not only visualize, but also treat any remaining stenosis or obstruction as well as complications in order to increase patency.
Discussion
Although high grade evidence is still lacking in the emerging field of endovascular treatment of chronic venous obstructive disease, dedicated centres are quickly generating great amounts of clinical experience and more and more patients are treated for complex post-thrombotic aberrations. Iliofemoral obstructions reaching into the common femoral vein are of specific interest as endovascular treatment of the common femoral vein has not been well established, but intervention at the level of the common femoral vein is required to guarantee adequate inflow into to stented iliocaval segments. In this paper we describe our experience in assessing iliofemoral flow and indications for ancillary interventions combined with PTA and stenting by use of modern imaging modalities. It is noteworthy that even though techniques like DUS, MRV and IVUS can clearly show the intraluminal post-thrombotic aberrations, the actual aspect of the vein after venectomy with the naked eye is always worse than on imaging, stressing the need for surgical interventions in these patients (Figure 3).
Surgical image showing the extensive intraluminal changes in the common femoral vein as a result of incomplete recanalization after a deep venous thrombosis more than one year ago. These changes correspond with the trabeculations visualized with magnetic resonance venography
Safety and effectiveness of endophlebectomy and AVF in venous disease have been described in a small number of studies. Our own experience will be described elsewhere. As already mentioned Raju et al. 19 described an endophlebectomy technique in 1999 in relation to axillary vein transfer for deep venous reflux correction in post-thrombotic limbs. They performed this operation in 83 limbs, in which the transplant patency rate was 83% after 10 years. They however did not specifically asses parameters of the endophlebectomy itself. Furthermore, their indications for treatment were based on severe PTS symptoms in the presence of venographic intraluminal post-thrombotic lesions, and treatment was aimed at deep venous reflux relief. And therefore the occlusive component of PTS in these patients was not addressed as it is done in current literature. Puggioni et al. 10 described endophlebectomy of 23 vein segments in 13 patients in 2004. They did focus on the obstructive component of PTS, as indications were either to improve inflow into reconstructed vein segments, to allow for valve repair or to increase calf outflow. They reported a primary patency rate of 77% after a median of eight months. Lastly, Vogel et al. 18 recently described endophlebectomy of the common femoral vein combined with PTA and stenting in 10 patients, of five of whom six-month follow-up data were available. In two patients an adjunctive AVF was created. Patency was 80% at six months, and clinical scores (VCSS and Villalta) and quality of life all improved. Interestingly these authors noted that complications rates because of aggressive anticoagulation might have been reduced if AVFs were more frequently used instead. Despite these very favourable results it is important to realize that when combining endophlebectomy and AVF creation with PTA and stenting, outcome assessment is difficult as the individual contributions of these techniques to the final effect are intertwined. With regard to the technique it has not yet been established to what level the stents should be placed, specifically if there is a need to keep stents limited to the region cranial of the endophlebectomy, or to extend the stents into this region. And as new dedicated venous stents are now becoming available, the whole issue of femoral stenting might need to be revised.
Ideally we would like to have a parameter giving us the information in whom we should perform an endophlebectomy and/or an AVF creation to increase patency. To adequately address this still a lot of research has to be done and registries from al centers performing these procedures might give us more answers in the future.
Conclusion
Endovenous recanalization of post-thrombotic vein segments by PTA and stenting is a treatment quickly gaining popularity. Endovascular techniques suffice for iliocaval disease; however in case of femoral involvement adequate inflow from the common femoral vein into the iliac tract needs to be guaranteed. This can be done by performing an endophlebectomy and/or creating an AVF. Indications for these interventions are however at this moment not well established. In our experience careful preoperative assessment with DUS and MRV is needed and helpful. When intraluminal processes are present at the level of the common femoral vein below the SFJ the common femoral, femoral and deep femoral vein outflow are generally impaired, indicating the need for an additional surgical intervention.
Funding
This research received no specific grant from any funding agency in the public, commercial or not for-profit sectors.
Footnotes
None.
