A pre-clinical animal study of a novel mechanical-only ablation device treatment for superficial venous disease

Introduction

Since the introduction and adoption of minimally invasive endovascular thermal techniques, in the late 1990’s for the treatment of chronic venous reflux disease, these procedures are now considered the standard of care. They revolutionized the treatment by shifting away from hospital-based stripping procedures to office-based endovenous procedures. In addition to the obvious benefits of reduced pain and faster recovery times with thermal procedures, the durability of both laser and radiofrequency methods have been proven in long-term follow up studies.^1–3

Injected chemical agents for the treatment of varicose veins has a long history of use. ⁴ Improvements in this original non-thermal technique over the past number of decades have included the use of synthetic agents, adjunctive compression, and foamed preparations. Ultrasound-guided foam ablation (USGFA) is particularly advantageous for treating tortuous venous anatomy, veins superficial to fascial layers and varicose veins. As a truly office-based procedure, the use of USGFA has gained widespread adoption by healthcare practitioners worldwide and is a recommended treatment in many clinical guidelines.^5–7

Despite the improvements in both these techniques, thermal and foam ablative procedures still have deficiencies and are associated with some significant risks to patients. To reduce the risk of thermal injury to surrounding skin and nerves, treatments are preceded by the time consuming, technically challenging and pain inducing delivery of tumescent anesthesia. Patients may be deterred from accessing or returning for repeat treatments due to needle phobias or the painful tumescent anesthesia. While USGFA does not usually require tumescent anesthesia for truncal ablation, this type of procedure suffers from lower efficacy rates in randomized comparison studies. ⁸ In addition, USGFA carries specific risks associated with chemical sclerosants entering the circulation including deep venous thrombosis (DVT), transient ischaemic attacks (TIAs), and allergic reactions. ⁹ While serious adverse events are rare, both heat and chemical induced thrombosis with subsequent risk of potentially fatal Pulmonary Embolism (PE) have been reported.^10–13

Efforts at improving the action of chemical sclerosants by the addition of a mechanical component have been introduced with the development of mechano-chemical ablation (MOCA). As shown by data from pre-clinical studies, the efficacy of MOCA is completely reliant on the coinciding use of chemical and mechanical ablation.^14,15 Comparative studies of MOCA with thermal ablation have demonstrated mixed efficacy, with lower occlusion rates yet similar clinical improvement scores.^16,17 The difficulty of mechanically effecting the venous system with its one-way valves and compliant walls has also been highlighted in reports of catching and stripping of target veins. ¹⁸

Cyanoacrylate adhesives represent an effective non-thermal treatment for venous disease but their implantation as a foreign body carries the risk of hypersensitivity reactions causing both local and systemic side effects. ¹⁹ Although the true rate of hypersensitivity reactions is unknown, some national venous societies have recommended against the use of cyanoacrylate in patients with a history of inflammatory disorders which limits their usage as a widespread treatment option. ²⁰

InVera Medical (Galway, Ireland) has developed a mechanical only ablation (MOA) catheter called EnVena for the treatment of superficial venous reflux disease, eliminating the need for tumescent anesthesia, chemical sclerosants and permanent implantation of foreign materials. The developers hypothesized that by causing an appropriately extensive level of cellular disruption in the vein wall, while avoiding excess damage or inadvertently catching in vein valves, a mechanical treatment with both high efficacy and safety could be achieved. The EnVena device is designed to be navigated using standard ultrasound guided techniques and is compatible with currently available 5 Fr sheaths. At its distal end a constrained nitinol helical ablation coil is deployed by retraction of the outer catheter (Figure 1). The ablation coil engages circumferentially with the vein wall and is simply withdrawn through the vein without the requirement for spinning or rotational movements.

OPEN IN VIEWER

Figure 1.

The EnVena device catheter with nitinol helical ablation coil in the deployed position.

The aim of this feasibility study was to evaluate the safety, efficacy and handling of the EnVena device in a large animal model with chronic follow up to assess histopathological response of veins to exclusively mechanical treatment.

Methods

With prior approval by the Ethics Committee responsible for overseeing the animal care and use program at Semmelweis University, Budapest, Hungary, the study was performed at the George Berci Research Laboratory, Department of Surgical Research, Semmelweis University. Four sheep (Ovis Aries) received from a local farm in Budapest, Hungary, were studied with different versions of the EnVena device which varied in ablation coil surface roughness finish. The ovine model was chosen as the anatomy of the venous system and clotting behaviour of blood is closely related to humans. ²¹ It is also a widely used test system for endovenous devices. ²²

Four sheep each weighing >50 kg at time of treatment, underwent bilateral hind leg endovascular procedures of the right and left, lateral saphenous veins (LSVs). Before, the mechanical vein ablation procedure and while the animal was in the lateral decubitus position under general anesthesia, the selected veins were examined with duplex ultrasound imaging (Innosite™, Phillips, The Netherlands) for size and patency, the latter defined on ultrasound as a vein with an anechoic lumen and smooth wall.

Procedures were performed under a general anesthetic in a surgical operating room. All procedures were carried out by the authors as follows, Veins 1 to 4 by N.P., Veins 5 and 6 by L.S.K. as referenced in Table 1. Induction of general anesthesia was performed with an intravenous injection of medetomidine Hydrochloride 0.01 mg/kg and ketamine 5 mg/kg. Each sheep was intubated, and general anaesthesia was maintained with isoflurane 0.5% to 2% delivered in oxygen through a rebreathing system.

OPEN IN VIEWER

Table 1.

Summary of Histological findings in treated veins.

Follow-up(days)	Vein no.	No. of segments analysed	Classification of lumen occlusion			Vessel diameter (mm)	Lumen area	Vein wall fibrosis (1 to 4)
			Occluded	Impinged	Open
51	1	5	3	1	1	1.8 (0.7–2.6)	0.04 (0.01–0.08)	2.2
51	2	5	2	2	1	1.4 (0.5–2.5)	0.26 (0.02–0.77)	2.6
51	3	8	5	1	2	2.4 (1.9–3.1)	0.33 (0.03–1.72)	2.1
51	4	7	4	0	3	2.5 (1.7–3.4)	0.24 (0.04–0.91)	2.1
89	5	11	8	0	3	2.5 (1.8–3.5)	0.12 (0.01–0.48)	1.8
89	6	9	4	3	2	3.2 (2.1–5.3)	0.56 (0.12–0.89)	2.6

Following preparation of the hind limb and creation of a sterile field, percutaneous access to the LSV at the ankle level (cranial branch) was achieved with a 20-gauge intravenous cannula. A 4 Fr introducer sheath (Pinnacle™, Terumo, Japan) was inserted using a Seldinger technique. The EnVena device was inserted and advanced under ultrasound guidance approximately 25 cm from the access site at the ankle to within 5cms of the saphenofemoral junction. The oversized helical coil was unsheathed to exert an outward radial force on the vein wall and achieve circumferential contact. Treatment was initiated by withdrawing the device at a rate of approximately 5 seconds/cm. Treatment was terminated within 1 cm of the insertion site by withdrawing the introducer sheath. To investigate device handling under different conditions in one instance (Vein 1) the procedural steps were repeated by recapturing the coil, advancing and repeating the withdrawal of the MOA catheter. At the end of all procedures the EnVena device was removed completely leaving no implanted material behind. Technical success was defined as successful delivery to the target vessel, deployment, withdrawal and recapture of the coil element without snagging or catching. After the procedure, the entry site was covered with cotton wool and tape; however, no compression wrapping was applied to any legs. Postoperative analgesia of a non-steroidal (meloxicam 0.4 mg/kg s.c.) and prophylactic antibiotic (amoxicillin 15 mg/kg i.m.) was administered to each animal. After post-operative recovery, the animals were placed in group housing for the respective follow-up period.

Venous explantation was performed at 51 days in two animals and at 89 days in one animal to assess the histopathological response in sub-chronic and chronic states. One sheep died at 23 days post procedure from ruminal tympany (bloat) with characteristic findings in the gastrointestinal system as confirmed on necropsy. Veins in this animal were not harvested for histological analysis. Follow up procedures and vein extraction protocols were the same in each animal. Under general anesthesia, the sheep’s treated veins were examined with ultrasound prior to removal. All vein segments were fixed in formalin following excision. Vein length measurements were repeated following fixation to estimate the level of tissue contraction. All veins were cut into approximately 4 mm segments and processed into standard paraffin blocks with each block containing 2 adjacent segments. Each block was numbered starting peripherally and all even numbered blocks were used for histological analysis to form a representative sample of the entire treated vein. Slides were sectioned into 4 µm sections and stained with hematoxylin and eosin for microscopic assessment. In slides containing multiple sections from adjacent vein segments, the section with the highest degree of injury was scored.

Histologic evaluation of each vein section was performed by analysing the overall level of lumen occlusion and the level of fibrosis in the layers of the wall. Quantitative measurements of vein sections were performed digitally (CellSens Entry™ Ver 2.1, Olympus, Japan) and included the area (mm²) of intimal hyperplasia or proliferation of fibrosis and the area of luminal patency if present. Luminal patency in fully occluded sections was measured as the largest area of white space created by micro-vessel invasion, if present.

If the vein was not fully occluded but the area of hyperplasia or fibrosis was greater than the area of luminal patency, the section was classified as impinged. This level of impingement was considered sufficient to have significantly affected blood flow. The level of lumen occlusion was classified as either: occluded, impinged or patent.

Histological analysis was performed at the National University of Ireland, Galway, by a consultant clinical pathologist with expertise in vascular disease. The pathologist performing the analysis was blinded from the other aspects of the study.

Results

Histological analysis

An average of 7.5 (range 5 to 11) vein segments were analyzed for each vein (Table 1). The majority of sections analyzed (Figure 2) were categorised as either occluded (26 sections) or impinged (7 sections). Of the 12 open sections recorded, 10 occurred in the most cranially located vein segments near the saphenofemoral junction, while 2 occurred most caudally at the beginning of treated segments. The average vessel diameter was 2.3 mm and the average lumen area was 0.25 mm². There were no perforations of the vein wall or damage to adventitia observed in any sections. Occluded sections demonstrated widespread intimal layer disruption and extensive early fibrosis (Figure 3). Impinged sections demonstrated large neointimal projections with evidence of both early fibrosis and re-endothelization (Figure 4).

OPEN IN VIEWER

Figure 2.

Summary of classification of vein lumens.

OPEN IN VIEWER

Figure 3.

Histology of vein section showing fibrotic occlusion following thrombus maturation in the vein lumen. There is widespread disruption of the original intimal area and endothelial layer. Capillary micro vessels are observed surrounded by extensive early fibrosis (thin black arrows). Hemosiderin deposition is also visible (heavy black arrows). Inflammatory changes extend deep into the media layer (asterisk).

OPEN IN VIEWER

Figure 4.

Histology of treated vein section showing impingement. The area of intimal hyperplasia (asterisk) is greater than the luminal area and there is moderate to severe early fibrosis in the media layer (heavy black arrows). The circumference is composed of large neointimal projections with evidence of reendothelialisation (thin black arrow). There is significant early fibrosis within some of these projections.

Discussion

This is the first report of a purely mechanical action caused by a minimally invasive endovascular device resulting in fibrotic occlusion of a vein in a pre-clinical setting. The extent of reduction in vein luminal diameter varied from complete fibrotic occlusion to partial occlusion and wall hyperplasia across treated vessels. This is to be expected given the early stage of development of the technology and treatment method.

Previous reports of purely mechanical treatments for vein ablation from pre-clinical studies revealed limited effects without occlusion.^14,15 In addition, the mechanical effect, in ex-vivo experiments, resulted in a patchy pattern of endothelial layer disruption, leaving large islands of endothelium intact. ²³ In contrast, the results of the present study show fibrotic occlusion in contiguous sections in all veins at follow up between 7 to 12 weeks. Occlusions up to a maximum of eight adjoining segments in one vein are reported which represents an estimated occluded length of 12 to 16 cm in vivo prior to shrinkage from spasm and fixation in formalin. This highlights the ability of the coil design with outer abrasive surface to achieve widespread, circumferential disruption to the endothelial and media layers, resulting in mature fibrotic transformation of thrombus.

Regarding safety, the EnVena device resulted in no serious adverse events in the periprocedural follow up period. The long-term follow up period recorded one early death attributed to a non-device related complication with no other complications noted. During ablation, all devices functioned correctly without mechanical failures. During venous navigation, deployment and pullback, the catheter tip and deployed helical ablation coil were both clearly visible by ultrasound. By the coil’s physical properties, exerting an outward radial force the device maintained contact with the vein wall; however, the coil responded appropriately on encountering valves and sections of reduced diameter by elongating to reduce the axial force and prevent snagging. Histological analysis showed no evidence of device related trauma to the adventitia or periadventitial tissue.

The exact depth of vein wall cellular disruption required to reliably cause high rates of obliterative occlusion in the clinical setting is still unknown. Histologic studies of both laser and chemical sclerosant effects on the vein wall have led to different conclusions in relation to the extent of media damage required for successful remodelling.^24,25 A comprehensive review of the molecular pathways involved in endovenous laser ablation by Heger, M et al. concludes that chemotactic agents released by dead and dying cells in the intima and inner media of the vein wall provide the most potent trigger to inflammatory mediated vein remodelling. ²⁶ Heat denatured extracellular proteins likely amplify this effect but their true contribution to effective treatment is unknown. Based on a review of histopathological studies of venous ablation methods the authors conclude that a mechanism of MOA of the intima and inner media layers could be sufficient to cause fibrotic vein remodelling and long term occlusion. This hypothesis is further supported by long term histological analysis of veins treated with cyanoacrylate glue which reveal mature collagenized fibrous tissue despite the glue not causing a transmural injury to the wall. ²⁷

Reducing the luminal diameter of the target vein is an important factor that can influence success rates in vein ablation procedures. While efforts were made to raise the leg and place the sheep in the Trendelenburg position, the effect on reducing the vein diameter was limited due to the muscular drainage of the target vein. The levels of venospasm in sheep veins seen during the procedures were significantly less than would be expected in the clinical setting. There is a paucity of published literature available on the levels of venospasm in large animal models compared to humans. Compression bandaging was not used during or post any procedures in the study which may have resulted in exposure to increased flow rates in target veins.

Further complexity in the interpretation of the extent of occlusion is introduced by the venous anatomy of the sheep hind limb. Given the muscular drainage of the LSV of the hindleg, ²⁸ the position of adjoining tributaries in relation to the treatment site could have a significant impact on venous flow. In particular, the confluence of the caudal branch of the LSV at approximately the ankle level may significantly increase flow rates and as a result may be a contributing factor to the higher rate of patency in treated sections in the proximal leg. The two peripherally open sections adjacent to the introducer site may not have been treated or only partially treated by the device as to avoid excluding any treated segments a wide margin was taken around the estimated treatment length for histological analysis which may have resulted in this underestimation of treatment efficacy.

Previous pre-clinical studies of thermal, mechanochemical and cyanoacrylate devices have been reported in similar ovine, caprine, and porcine models. Unfortunately, there is generally a failure in these studies to report on the estimated length of vein occlusion achieved and its relationship to the confluence of major tributaries which limits comparative analysis. Future pre-clinical studies using these models must include such detail to allow for better interpretative results.

The current study is limited by the small number of veins treated and the heterogenous nature of the treatment protocols. This feasibility study was not hypothesis-driven and consequently performance goals were not defined for safety and efficacy. Procedures on animals were also performed at different timepoints and by practitioners with varying levels of clinical experience which may have affected the results. However, this study provides histological data showing consistent fibrotic transformation leading to occlusion of vessels in a representative animal model which has translated closely to the clinical setting with other previously developed technologies.

Intra-procedural pain assessment is not possible in the pre-clinical setting due to the general anaesthesia and lack of any reliable surrogate physiological markers. Clinical studies under local anaesthesia will be required to fully ascertain pain sensation. Other non-pain sensations such as pressure are unlikely given the limited radial force exerted on the vein wall by the MOA system.

A purely mechanical mechanism of action as described in this pre-clinical study inherently avoids the risks associated with thermal injury, chemical effects on non-target sites or permanent foreign body implantation. In addition, since MOA is free from thermal injury mechanisms, below knee treatment would effectively exclude nerve injury risk. However, these potential benefits can only be realised through further investigation in a clinical setting demonstrating safety and efficacy.

Conclusion

This report demonstrates early feasibility of the EnVena device and an exclusively mechanical ablation technique to safely achieve effective chronic vein closure in a representative pre-clinical animal model. Future studies in the clinical setting are needed to prove safety and efficacy of the MOA device for the treatment of superficial venous disease.