Radiofrequency Closure

The first theories of endovenous ablation were based on the belief that specifically directing relatively omnidirectional RF energy into vein walls to cause their destruction was potentially safer, easier to engineer and more controllable than other mechanisms for doing so. Initial designs involved a mechanism by which RF current heated tissue by resistive (or ohmic) heating of a narrow rim (less than 1 mm) of tissue in direct contact with an electrode. Deeper tissue planes could be slowly heated by conduction from the small-volume region of heating, although heat was typically dissipated by conduction into surrounding normothermic tissue.⁷ By carefully regulating the degree of heating with microprocessor control, subtle gradations of either controlled collagen contraction or total thermocoagulation of the vein wall could be achieved.

The initial design was such that when the RF catheter was pulled back, a feedback-controlled loop regulated by readings from a thermocouple enabled the operator to heat a section of vein wall to a specified preset temperature. This was chosen for its relative safety since the temperature increase remained localized around the active electrode. This necessitated the maintenance of close, stable contact between the active electrode and the vessel wall without coagulum formation. It was believed by strictly limiting the temperature to 85°C, boiling, vaporization, and carbonization of the tissues could be avoided.⁸ It was also believed that heating the endothelial wall to 85°C resulted in heating the vein media to no more than 65°C, the minimal temperature at which collagen contracts.

Ex-vivo studies by Reich-Schupke et al⁹ investigated histological changes following radiofrequency ablation at various powers and application times. When low power (5 W) and an application time up to 400 ohms was applied, histological changes were not uniform. Necrosis was limited to the endothelium in the majority of vein segments and rarely reached the media. This would most likely not result in complete vein shrinkage and occlusion. At 20 W and an application time up to an impedance of 400 ohms, histological changes included widespread necrosis of the initima and media and collagen bundle coagulation. The authors concluded that with increased power and application time, there was a more homogenous and extensive heating of the vein wall, which was thought to lead to a more successful outcome.

Vessel wall ablation using electrode-mediated RF is a self-limiting process. As coagulation of tissue occurs, there is a marked decrease in impedance that limits heat generation.¹⁰ Alternatively, if a clot builds up on the electrodes, blood is heated instead of tissue and there is a marked rise in impedance (resistance to RF). The RF generator can be programmed to rapidly shut down when impedance rises, thus assuring minimal heating of blood but efficient heating of the vein wall. The problem is that the electrodes must be manually debrided of coagulum, which requires the removal of the catheter, cleaning by the operator and then reinsertion – which is problematic during tumescent anesthesia.

The initial system introduced with electrode-mediated RF ablation was the Closure System (VNUS Medical Technologies, Sunnyvale, Calif., now Covidien, North Haven, Conn.). With the Closure catheter system, bipolar electrodes are deployed by spring action and placed in contact with the vein wall. As the vein wall contracts, the electrodes are able to retract somewhat within the vein allowing vein wall narrowing. Selective insulation of the electrodes results in a preferential delivery of the RF energy to the vein wall and minimal heating of the blood within the vessel.

The initial catheter designs included collapsible catheter electrodes and a central lumen to allow a guidewire and/or fluid delivery structured within the 5-French (1.7-mm) catheter. This permits treatment of veins as small as 2 mm and as large as 8 mm. A larger 8-French catheter allowed treatment of saphenous veins up to 12 mm in diameter. Both catheters had thermocouples on the electrodes embedded in the vein wall which measured temperature and provided feedback to the RF generator for temperature stabilization. The control unit displayed power, impedance, temperature, and elapsed time so that precise control could be obtained. The unit delivered the minimum power necessary to maintain the desired electrode temperature. For safety, if a coagulum formed on the electrodes, the impedance rises would cut off the RF generator.

The initial experience, dating back to 1998, demonstrated an efficacy equal to or better than that of ligation and stripping, with few, if any, adverse sequelae.^11–23 Early experience directly comparing RF Closure with ligation and stripping procedures, even with RF performed under general anesthesia, noted equal efficacy with less pain, shorter ‘sick leave’, and faster return to normal activities.¹⁸

When performed by us, the procedure was entirely under local tumescent anesthesia, with over 90% of patients resuming normal activities 1 to 2 days postoperatively. Its main drawbacks were the high cost of single-use catheters and the necessity to withdraw the catheter manually at a speed of 2 to 3 cm per minute and frequent cleaning of coagulum on the electrodes, which made the procedure tedious at times. To speed up the procedure, Goldman¹² recommended that only the most proximal 20 cm of the GSV be treated with RF and the remaining varicose GSV be treated with ambulatory phlebectomy, but this technique has not found wide acceptance. Goldman believes that the addition of ambulatory phlebectomy minimizes the possibility of recurrence from distal perforators. Proebstle et al²⁴ found that up to 30% of tributary veins do not resolve with laser ablation of the GSV alone, thereby necessitating removal with ambulatory phlebectomy.

However, treatment of the GSV or its tributaries below the knee may not be entirely necessary, as others have shown that ligation and stripping procedures from the groin to the knee add little to the procedure’s efficacy.^15,25 In addition, others have also demonstrated equal effectiveness with less than 2 years follow-up when only the proximal 30 to 40 cm of the GSV is treated without treating distal varicose tributaries.^{13–15,17,26}

Weiss and Weiss¹⁷ evaluated patients treated with a percutaneous approach allowing access of the Closure catheter to treat the proximal GSV. Patients (mean age, 47.2 ± 12.6 years; 76% female) had symptomatic saphenous reflux with a saphenous vein diameter of 2 to 12 mm (mean, 7.4 mm). Most of the veins treated were above-knee great saphenous (73%), some entire great saphenous (21%), with the remaining including below-knee great saphenous, small saphenous, and accessory saphenous. Adjunctive procedures performed at the time of treatment were phlebectomy on more distal branches in 61% and high ligation in 21%, but the adjunctive procedures did not affect outcome.

Vein occlusion at 1 week was documented by duplex ultrasound in 300 out of 308 legs, or a success rate of 97%. Occlusion persisted at 6 weeks in 95% and at 6 months in 92%. In this report, if the saphenous vein was closed at 6 months it was noted by duplex ultrasound to remain closed to 12 months and beyond. Subsequent follow-up for up to a decade by duplex ultrasound indicates that any vein noted to have been eliminated at 12 months by RF will never recur. Typically when the GSV is treated there is closure or elimination of major tributaries at the SFJ except for the superficial or superior epigastric vein, which, intentionally not treated, continues to empty superiorly into the common femoral vein. We believe that there is a high margin of safety by maintaining flow through this tributary. The high flow rate appears to diminish the possibility of extension of any thrombus (in the unlikely event that this would occur) from the GSV and has the additional benefit of allowing normal venous flow from the lower abdominal wall into its proper drainage into the common femoral vein. By leaving the superior epigastric vein intact, thrombus in the GSV following this procedure has not been observed.¹³

Long-term efficacy with the RF ablation has been documented by Merchant et al²⁷ investigating 1222 limbs (great saphenous, small saphenous, and accessory saphenous veins). Occlusion rates (evaluated via duplex ultrasound) of 96.8%, 89.2%, 87.1%, 88.2%, 83.5%, 84.9%, and 87.2% were found at 1 week, 6 months, 1 year, 2 years, 3 years, 4 years, and 5 years, respectively. Body mass index greater than 25 was associated with an increased incidence of nonocclusion, groin reflux, and recanalization. A pullback speed above 3 cm/minute at 85°C was more likely to result in nonocclusion and recanalization. In the study by Vasquez et al,²⁸ factors associated with improved occlusion rates included increasing age, female sex, and volumes greater than 250 ml of tumescent anesthesia. The authors theorized that increased failure rates associated with male sex and younger age are secondary to variations in collagen and inflammation in these populations.

Regarding clinical symptoms, a successful RF (or laser) endovenous occlusion procedure rapidly reduces patient pain, fatigue, and aching, correlating with a reduction in the CEAP clinical class for symptoms and clinical severity of disease. When patients have had simultaneous surgical stripping on the opposite leg, the degree of pain, tenderness, and bruising have been far greater on the leg treated by stripping. Side effects of the Closure technique have included thrombus extension from the proximal GSV in 0.8%, with one case of pulmonary embolus; also, skin burn (prior to the tumescent anesthesia technique) in 2.5%, clinical phlebitis at 6 weeks in 5.7%, and temporary quarter-sized areas of paresthesia in 18%, with most of these occurring immediately above the knee and resolving within 6 months to a year. Thus, compared with most techniques – but, in particular, traditional surgery of ligation and stripping of similar size saphenous veins – the effectiveness of endovenous RF occlusion is quite high.

In a study by Goldman and Amiry,¹⁶ closure of the GSV with endoluminal RF thermal heating in combination with ambulatory phlebectomy was equally as effective as closure of the GSV as described above. The first 47 sequential, nonrandomized patients having an incompetent GSV from an incompetent SFJ and painful varicosities in 50 legs were treated with the VNUS Closure procedure. The varicose veins were marked with the patient standing and again with the patient lying down in the operative position with a venoscope (LLC, Lafayette, La.), as previously described (Fig. 11.1).^12,29,30 After appropriate marking, the area surrounding the GSV and distal tributaries to be treated was infiltrated with 0.1% lidocaine tumescent anesthesia. The amount of tumescent fluid averaged 800 mL with a lidocaine dose of 8 mg/kg. The GSV was then accessed through a 2- to 3-mm incision in the medial midthigh, usually 20 cm inferior to the SFJ. The proximal portion of the GSV was then treated with VNUS Closure and the distal portion, including all varicose tributaries, was removed with a standard ambulatory phlebectomy technique.

Figure 11.1 Premarkings. Varicose veins were marked with the patient standing and again with the patient lying down in the operative position with a Venoscope (LLC, lafayette, La.).

Thirty-nine patients with 41 treated legs were available for evaluation at the longest follow-up period. Six patients (nine treated legs) could not be located for re-evaluation after 6 months because of change in location (often out of the state).

The average time to access the GSV in the medial thigh was 7 minutes (range, 1–30 minutes). Twenty-seven patients had the GSV accessed in approximately 1 minute. The average catheter pullback rate was 2.76 cm/minute over an average length of treated GSV of 19 cm (range, 6–42 cm). Complete surgical time, including the phlebectomy portion of the procedure, was approximately 20 minutes (range, 13–35 minutes).

Ninety-five percent of all patients could resume all preoperative activities within 24 hours. The other two patients could resume all activities within 48 hours. Every patient had complete elimination of leg pain and fatigue. Twenty-one of 22 patients who presented with ankle edema had resolution of ankle edema. All patients said that they would recommend this procedure to a friend.

Adverse sequelae were minimal, with four patients complaining of heat distal to the SFJ during the procedure which resolved with additional tumescent anesthesia. Twenty-eight of 50 treated legs had some degree of purpura lasting 1 to 2 weeks. Five patient legs developed mild erythema over the GSV closure site that lasted 2 to 3 days. Eight legs had an indurated fibrous cord over sites of ambulatory phlebectomy that lasted up to 6 months.

Clinical and duplex evaluation performed by an independent laboratory and/or physician at 6, 9, 12, 18, and 24 months disclosed 90% abolition of reflux. No new varicose veins were noted to appear in three patients with recurrent reflux in the GSV. One patient who developed reflux had the development of new veins at 1 year post-treatment.

Other surgeons have had a different experience with the use of VNUS Closure in the treatment of incompetent GSV. The reason for the difference in results is likely to be secondary to the anesthesia used, as well as the technique, as described below.

Three separate papers detail a similar cohort of patients treated in multicenter studies encompassing from 16 to 31 clinics, 210 to 324 patients, and 6 to 12 month follow-up.^13,14,31 The vein occlusion rate at 1 year examination was 91.6% from nine centers and 81.9% from fourteen centers. Forty-nine patients were followed at 2 years with duplex scans and showed an 89.8% closure rate. There was a 3% incidence of paresthesia which was decreased to 1.6% when treatment was confined to the thigh. Two limbs (0.8%) developed scarring from skin burns and three patients developed a deep vein thrombosis (DVT) with one embolism. The reason for the increase in adverse effects appears to be the use of general anesthesia without tumescent anesthesia by a majority of the surgeons.

Sybrandy and Wittens³² from Rotterdam reported one year follow-up of 26 patients treated with VNUS Closure. They reported five patients with postoperative paresthesia of the saphenous nerve and one with a cutaneous burn, for an overall complication rate of 23%. One patient (3.8%) had total recurrence of the GSV. One patient (3.8%) could not be treated due to a technical failure. Eight patients (30.8%) had closure of the GSV, but with persistent reflux of the SFJ. Thirteen patients (50%) had closure of both the GSV and SFJ. Overall, 88% of patients had a totally occluded GSV.

One probable reason for the increase in adverse effects was their use of a spinal anesthesia instead of the recommended tumescent anesthesia. In addition, they treated all patients from the ankle proximally, which exposed the GSV within the calf to heat from the RF catheter. Their mean operating time was 67 minutes (range, 25–120 minutes).

Another report describes two episodes of DVT in 29 patients treated with the RF Closure.³³ Here, the surgeons treated the patient with a groin incision and passage of the catheter from the groin downward. The authors do not report the type of anesthesia used or the length of vein treated. It is presumed that patients were not ambulatory and were treated under general anesthesia.

The important information to come out of a review of various treatments of the GSV is that the use of tumescent anesthesia in awake patients who can ambulate immediately after the procedure is important in preventing skin burns and DVT. Treatment when limited to the GSV segment above the knee is also important in preventing paresthesia to the saphenous nerve.

In our experience using tumescent anesthesia in awake patients, two patients have developed focal numbness 4 cm in diameter on the lower medial leg. These resolved within 6 months. Since adopting the principles outlined above of tumescent anesthesia and moving the catheter rapidly from any points of sharp pain, no paresthesias have been noted. No skin injury or thrombus has been observed in any of our patients. Unfortunately, with both endoluminal RF and laser procedures, if patients are not ambulatory after the procedure and/or if tumescent anesthesia is not given, complications in the form of DVT, PE, or angiogenesis have been reported.^34–36 Tumescent anesthesia or the placement of large volumes of dilute anesthesia in a perivascular position serves several purposes:

Contrary to the report by Hingorani et al³⁴ we have never seen DVT in any of our patients treated with intravascular laser or RF. We believe that the reason for our lack of adverse sequelae is the use of tumescent anesthesia in awake patients with immediate ambulation and avoidance of occlusion of the superior epigastric vein. While we realize treating patients without general anesthesia is not standard practice for general or vascular surgeons,³⁷ some vascular surgeons who perform tumescent anesthesia on awake patients with immediate ambulation have reported similar results with virtually no DVT. A DVT was noted in a female patient treated using tumescent anesthesia while awake but she weighed more than 350 pounds and did not ambulate after the endoluminal RF procedure.^38,39

Salles-Cunha et al³⁵ reported on the development of angiogenesis and fibrotic tissue along the course of the GSV treated with RF Closure. Contrary to this report, our experience with tumescent anesthesia utilized is a complete lack of detection of small vessel networks (angiogenesis) by duplex ultrasound. We believe that the reason for our lack in detecting small vessel networks is not from a lack of trying to see them, but from the minimization of inflammation that occurs with tumescent anesthesia placed in the perivascular space during either RF or laser endothelial ablation.⁴⁰

We have not performed ligation of the SFJ in any of our over 1000 patients and question the accuracy of the findings of Salles-Cunha et al, who found a decreased incidence of small vessel networks in patients whose SFJs were ligated. We suspect that the small number of patients who were treated without ligation of the SFJ (13) versus the 93 patients who did have SFJ ligation produced falsely positive statistical significance. We question if inflammation is the most likely cause for small vessel networks since ligation should not increase or decrease the extent or time of inflammation.