Implantation of Cryopreserved Allograft Pulmonary Monocusp Patch

Abstract

We implanted cryopreserved allograft monocusp patches made from cadaveric pulmonary arteries to correct nonthrombotic valvular insufficiency of the common femoral vein in patients with chronic venous ulcers. Thirty-eight patients with 40 ulcers of longer than 3 years' duration underwent valvular repair of the common femoral vein from September 1995 through November 2001. We studied ulcer healing and competence of the monocusp patches using duplex ultrasonography.

Three patients with a total of 4 ulcers were lost to follow-up; therefore, 35 patients with 36 ulcers were available for examination. Twenty-four ulcers healed, and 12 did not. Of the 24 healed ulcers, 5 recurred. Four of the 5 ulcers were treated, 3 by monocusp patch replacement and 1 by iliac vein dilation and stenting. All such recurrent ulcers healed. The 5th recurrent ulcer was not treated, because the patient refused therapy. Therefore, of the 24 ulcers that healed initially, 23 were healed at the end of the study.

Of the 12 unhealed ulcers, 4 were treated: 2 by ligation of incompetent perforating veins, 1 by saphenous vein ligation and partial stripping, and 1 by monocusp replacement. All 4 ulcers healed. The other 8 patients, all of whom had severe monocusp insufficiency, refused therapy. Therefore, of the 12 ulcers that did not heal initially, 4 were healed after additional treatment.

No implanted monocusp patch developed clots. The ulcers remained healed when the prostheses remained competent.

We conclude that monocusp patches can restore valvular competence to the femoral vein in patients who have primary valve insufficiency. In such patients, venous ulcers heal rapidly after prosthesis implantation if the monocusp remains competent and if there is no severe incompetence of the superficial venous system or of perforating veins. (Tex Heart Inst J 2002;29:92–9)

The ingenuity of many investigators has made possible the surgical treatment of venous valvular insufficiency in the lower extremities. In 1975, Kistner ¹ emphasized the importance of chronic venous insufficiency caused by single-valve incompetence and described a method to correct insufficiency of the superficial femoral vein valve. Other strategies have been described to treat venous valvular incompetence, ^2–8 the most recent of which is the transplantation of cryopreserved allograft valved venous segments by Dalsing and colleagues. ⁹ Those authors reported the use of such prostheses as interpositional grafts in the venous circulation. Graft thrombosis occurred in 33% of grafts in the superficial femoral veins and in 22% of those in the popliteal veins. Nonetheless, 66% of the ulcers healed. ⁹ These results were similar to those in Kistner's original report of ulcer healing after ligation of the superficial femoral vein in cases of imperfect correction of insufficient venous valves. ¹

In the past, we treated common femoral vein incompetence by implanting xenograft monocusp patches made of glutaraldehyde-fixed bovine pericardium. The valves were fashioned after the sinus of Valsalva, which has an anatomical structure that causes it to continuously cleanse itself of stagnant blood. ⁸ These xenograft patches were effective in healing venous ulcers. However, bovine pericardial MCPs are no longer available from suppliers; therefore, we explored the use of MCPs made from cryopreserved cadaveric pulmonary arteries. Herein, we summarize our initial experience with this bioprosthesis in patients who had severe nonthrombotic valvular insufficiency of the common femoral vein and venous ulcers for longer than 3 years. We discuss changes in implantation techniques and our responses to the problems that we encountered as we acquired experience with the monocusp patch (MCP).

Patients and Methods

Thirty-eight patients underwent implantation of 40 allograft pulmonary MCPs in the common femoral vein to treat venous valvular insufficiency during the 63 months from September 1995 through November 2001. During this study, we performed concomitant allograft implantation at another site in 12 of the 38 patients: 10 of these underwent ipsilateral greater saphenous implantation for severe superficial vein incompetence, and 2 underwent ipsilateral popliteal implantation for associated popliteal vein incompetence.

The use of more than 1 implant occurred when the supplier of the homografts had available only entire, 3-leaflet pulmonary artery trunks, which we cut into 3 monocusp patches.

The ages of the patients ranged from 49 to 83 years (average, 67.8 years). Nineteen of the patients were women. By the CEAP classification, ¹⁰ all patients had clinical findings of skin changes with active ulceration (C6), an etiology of primary disease (Ep), anatomic findings in the deep veins (Ad), and the pathophysiologic component of reflux (Pr). Patients who had saphenous implants were also given the classification of As (anatomic findings in the superficial venous system).

Patient Selection

The patients selected for inclusion in this study had active ulcers of more than 3 years' duration that had failed to heal with supervised conservative therapy, such as leg elevation, compression therapy, or skin grafting. The disorder had to be classified as primary valvular insufficiency of the common femoral vein that was both chronic and nonthrombotic.

Patients specifically excluded were those who had a history of previous or active deep venous thrombosis or documented hypercoagulopathy. Patients were also excluded if they had untreated, symptomatic arterial disease, congestive heart failure, lymphedema of any cause, or a major or progressive noncardiac disorder.

All patients underwent supine and standing duplex ultrasound to determine the presence or absence of, as well as the degree or severity of, femoral vein valvular insufficiency, incompetent perforating veins, saphenous vein incompetence, and acute or chronic venous thrombosis. ^11–13

We evaluated venous reflux by color-flow duplex ultrasonography and Valsalva's maneuver. The degree of valvular insufficiency was determined with the patient in a reverse Trendelenburg position. We used the following formula to calculate the reflux volume: volume (mL/min) = average velocity (cm/sec) × area (cm²) × 60 (sec). We classified the degree of reflux as mild (<10 mL/min), moderate (10–30 mL/min), or severe (>30 mL/min).

Incompetent perforating veins were identified by color-flow duplex ultrasonography; we then applied distal compression to study the normal augmentation of those veins. Reflux, if present, was found with use of Valsalva's maneuver.

We did not perform any other venous hemodynamic evaluation because, at that time, we did not have the necessary equipment available.

Monocusp Pulmonary Patch Preparation

Initially, the allografts were supplied as entire cadaveric cryopreserved main pulmonary trunks. Not only were these expensive, but they also required a great deal of last-minute preoperative preparation to resect a single cusp from the pulmonary trunk. Later, the supplier began resecting single-valved sections of main pulmonary trunks and cryopreserving them before shipment. Their techniques are described, as follows, by one of the authors (PT). A longitudinal incision is made in order to resect 1 commissural post from the myocardial skirt to the point of distal transection. Both of the remaining commissural posts can be resected in the same manner if additional cuspal patches are to be produced. The patch is then placed in a solution of RPMI 1640 with L-glutamine, HEPES buffer, and fetal bovine serum (FBS). Dimethyl sulfoxide (DMSO) is added until the concentration reaches 10%. The MCP grafts are subsequently sealed for cryopreservation at a controlled rate near 1 °C per minute (between +4 and −40 °C per min) until they reach –100 °C.

The patches were received from the supplier in dry ice. When they were needed, the patches were thawed and washed to remove the cryopreservation solution. During surgery, excess tissue on the inflow and outflow aspects of the bioprosthesis was removed.

These monocusp patches were a simpler and less expensive alternative to entire pulmonary trunks. The supplier provided the largest individual MCPs available. Unfortunately, we could not request an MCP of a specific size to fit the vein of a particular patient.

In the first 15 patients, we implanted 16 monocusp patches that were not blood typed. In December of 1998, we identified 2 patients in whom the cuspal tissue had disintegrated, an outcome that we attributed to an immunologic reaction. For this reason, we began using blood-type-specific patches after that date.

Operative Management

At the time of surgery, epidural anesthesia was administered to all patients. A sufficient length of the common femoral vein was exposed for patch implantation. The superficial femoral, profunda femoral, and saphenous veins were identified, and vascular control was obtained using Dacron tape. The femoral veins were manipulated as little as possible. Circumferential mobilization was performed only at the sites selected for proximal (cephalad) and distal (caudal) vascular clamp placement. If by visual inspection we could find the insufficient valve, we placed the monocusp patch at this site. Heparin was administered (1 mg/kg), and the activated clotting time was monitored to ensure proper anticoagulation, particularly during the period of interrupted venous flow.

We modified our implantation techniques as we learned more about the performance of the allografts in restoring femoral vein competence. For example, we found the cadaveric allograft cusps to be smaller than those of the xenograft bovine pericardial patches that we had used previously, and had to adjust accordingly.

We achieved intraoperative venous competence in all patients. Some of the patients (although improved), had some degree of postoperative incompetence when they were in an upright position or during Valsalva's maneuver. We were able to achieve venous competence in 3 patients by banding the corrected femoral vein. To reduce the possibility of postoperative incompetence, we modified our technique in later patients by maximally dilating the host vein before performing partial resection of the anterior wall of the vein. We achieved maximal dilation by first applying a clamp in the cephalad position of the vein and then compressing the thigh vigorously. When the vein was distended maximally, we applied the caudal clamp. The surgical procedure was then completed on the dilated vein in an attempt to compensate for subsequent changes in venous diameter.

We made a venotomy of sufficient length to accommodate the bioprosthesis. To perform a partial resection of the anterior wall of the vein, we marked with a pen the boundaries that constituted 50% of the distended vein's diameter. The resection of the anterior aspect of the vein was then begun.

The patch was sewn from cephalad to caudal with a continuous 6-0 polypropylene suture (Fig. 1A), from inside-out on the vein. To achieve this, we sutured 1 side about 80% of the distance. The bioprosthesis was then inspected from the inside to ensure that the leaflet was free and fully mobile. We sutured the other side of the patch from the outside-in on the vein until the suture line was 80% complete (Figs. 1B and 1C). After releasing the cephalad vascular clamp to allow blood to enter the bioprosthesis and distend the leaflet, we evaluated intraoperative competence (the Rinaldi MCP competency test) (Fig. 2). The suture line was completed without the cephalad clamp.

Fig. 1 Implantation of the monocusp patch: A) the anastomosis is sutured along 80% of the medial or lateral aspect; B) the anastomosis is sutured on the opposite side with an outside-in suture to protect the cusp; and C) the anastomosis is 80% complete.

Fig. 2 The monocusp patch (MCP) is tested intraoperatively with the Rinaldi MCP competency test. The cephalad clamp is released to allow blood to enter the bioprosthesis and distend the cusp. The cusp must have unrestricted movement and prevent backflow.

Protamine was given for complete reversal of heparinization. An intraoperative Doppler ultrasonographic examination was performed during Valsalva's maneuver to confirm forward venous flow and cusp competence.

Patients were discharged the day after surgery. They were all given a 300-mg loading dose of clopidogrel. Thereafter, they were on a daily regimen of aspirin (81 mg) or clopidogrel (75 mg). Patients with an intolerance to aspirin took only the clopidogrel. No unfractionated heparin, low-molecular-weight heparin, or warfarin anticoagulation was used in any patient.

The postoperative treatment protocol included absolute cleanliness of the ulcer and elevation of the extremities twice a day for 30 minutes and during sleep. Patients began wearing light compression stockings (20–30 mmHg) as soon as it was practical. In 2 patients with especially large ulcers (>10 cm in diameter), split-thickness, meshed grafts were applied when granulation tissue became evident at the site of the ulcer.

At the outset, we had underestimated the degree to which incompetent perforating veins prevented healing of the ulcer. Analysis of the data confirmed that incompetence of the perforators persisted and prevented ulcers from healing. Therefore, we began ligating the incompetent perforators either at the time of monocusp implantation or before.

Similarly, our follow-up confirmed that severe superficial venous incompetence should not be ignored. We therefore began treating superficial venous incompetence before or at the time of deep venous correction.

Postoperative Evaluation

Duplex ultrasonography was performed within 30 days after surgery. The examination was repeated 6 months and 1 year postoperatively. Thereafter, yearly duplex scans were scheduled unless the patient's condition on clinical examination warranted earlier study. Patients with nonhealing ulcers or grossly incompetent monocusp patches were encouraged to undergo descending venography with the addition of ascending venography if appropriate.

Results

During the last 4 months of the designated follow-up period, we called all patients to return for clinical evaluations and duplex ultrasonography. Re-evaluation proved to be a laborious process, because most patients initially refused. Follow-up venography was performed in only 4 patients. There was no death or serious adverse health event in any patient. Three patients (with a total of 4 ulcers) were lost to follow-up. When those patients were last seen, the ulcers were healed in 3 and unchanged in 1.

To calculate the results, we used the total number of ulcers rather than the number of patients (Fig. 3). Originally, there were 38 patients with 40 ulcers included in this study. After 3 patients were lost to follow-up, only 36 ulcers were available for analysis.

Fig. 3 Diagram of the ulcers, treatments, and outcomes.

Twenty-four ulcers (67%) healed rapidly (within 45 days). Two of those ulcers required skin grafts because of their large size. Figure 4 shows the descending venogram of a patient with a monocusp patch 58 months after implantation. The monocusp is fully competent and the ulcers remain healed. Five of the 24 ulcers (21%) recurred after healing. One patient with a recurrent ulcer and an incompetent monocusp refused venography or any further surgical therapy. The other 4 ulcers were treated. In 3 of those, the cusp had disintegrated; however, the ulcers healed when new patches were matched with the patients' blood types (ABO system) before placement. The 4th patient had new monocusp insufficiency, but the ascending venogram revealed ipsilateral external iliac vein stenosis (Fig. 5A). Dilation and stenting (Fig. 5B) resolved the stenosis (Fig. 5C), the MCP became competent, and the ulcer healed rapidly. As of December 2001, 23 of the 24 ulcers remained healed.

Fig. 4 Descending venogram of a patient with a monocusp patch 58 months after implantation. The femoral vein is fully competent.

Fig. 5 Ascending venogram shows A) severe stenosis of the ipsilateral external iliac vein; B) dilation and stenting of the external iliac vein; and C) the vein after stenting. Unobstructed flow is restored, and the monocusp is competent again.

Twelve ulcers (33%) did not heal. Nine of these patients had monocusp insufficiency, 8 of whom refused venography or surgical therapy. In the 9th patient, the MCP was replaced with an experimental glutaraldehyde-fixed bovine jugular vein MCP, and the large ulcer healed within 30 days. One patient had severe untreated saphenous vein incompetence, and 2 had incompetent perforating veins. Upon correction of the superficial incompetence in the 1st patient and ligation of the perforators in the other 2, the 3 ulcers healed.

In summary, 27 of the 36 ulcers (75%) available for examination were eventually healed. All 9 unhealed ulcers were in patients with severe monocusp insufficiency who had undergone monocusp implantation before we began using ABO-compatible bioprostheses.

Twelve of the patients underwent implantation of 2 monocusps in the same extremity, 10 of whom received saphenous vein implants and 2 of whom received popliteal vein implants. The first 7 of these patients underwent monocusp implantation before we began using ABO-compatible patches (6 saphenous, 1 popliteal). In 5 of the 7, the common femoral MCP became severely incompetent, and the ulcers did not heal. Nevertheless, the superficial or popliteal incompetence was much less than it had been preoperatively. In 2 of the 7 patients, both monocusps were competent, and the ulcers healed. The remaining 5 patients received implants that were ABO compatible: 4 saphenous and 1 popliteal. In all of these patients, both patches were competent and the ulcers healed.

No implanted monocusp patch was associated with local femoral vein thrombosis or pulmonary emboli. Two patients developed lymphoceles that had to be treated surgically.

Discussion

Venous valvular insufficiency (reflux) constitutes a major pathophysiologic cause of venous ulcers. ¹⁴ Various strategies to treat venous valvular incompetence include restricting the circumference of the veins, ² creating muscle slings, ³ performing internal or external valvuloplasty (with or without video assistance), ^4,5 transposing veins, ⁶ transferring valved vein segments, ⁷ transplanting glutaraldehyde-fixed xenograft monocusp pericardial patches, ⁸ and, more recently, transplanting cryopreserved allograft valved vein segments. ⁹

Currently, the most effective treatment for venous valvular insufficiency is valvuloplasty for incompetent valves. ¹⁵ However, valvuloplasty is a technically demanding operation. Moreover, many patients have severely damaged or absent valves, which often cannot be reconstructed. ¹⁶ For this group of patients, venous valved segments can be transferred from the upper extremities. ⁷ Unfortunately, such valves are frequently insufficient and smaller than those of the legs, and, after transfer, the transplanted veins tend to dilate and develop further insufficiency over time. ¹⁷

These difficulties have led to an intensive search for a venous valve that, according to DeLaria, ¹⁸ is easily implantable, has preserved leaflet flexibility and strength, and has open and nonthrombogenic leaflets at zero pressure. ¹⁸ Only tissue valves seem to fulfill these criteria. In 1986, we reported our experience with the use of glutaraldehyde-fixed monocusp bovine pericardial patches ⁸ in a patient with common femoral vein valvular insufficiency. These patches tended not to form thrombi, and the patient's ulcers healed. More recently, Dalsing and colleagues ⁹ used cryopreserved venous valve allografts as interpositional grafts in the venous circulation, after which 66% of the ulcers healed.

Herein, we discuss our study involving the implantation of cryopreserved monocusp pulmonary allograft patches in the common femoral vein. Our data support the hypothesis that implantation of a functioning valve can result in the healing of long-standing venous ulcers in patients with primary deep vein reflux of the common femoral vein.

Complete competence of the monocusp must be achieved at the time of implantation. In our experience, 50% of the anterior portion of the host vein must be resected to achieve coaptation of the cusp of the patch with the posterior wall of the vein and to make the vein competent. Our current techniques in human beings paralleled those in our laboratory model,^* in which we used intraluminal video techniques to study the closure characteristics of glutaraldehyde-fixed monocusp patches made from bovine jugular veins. The patches, which were implanted into fresh bovine jugular veins, required up to 70% anterior wall resection of the host vein for total competence to be achieved.

Clinically, we observed some patients with total intraoperative competence whose ulcers healed but whose MCPs later developed minimal-to-moderate insufficiency. Because leaflet motion was preserved, we judged the problem to be due to dilatation of the host veins. For that reason, we began dilating the vein maximally and, presumably, “overcorrecting” at the time of MCP implantation.

We used the largest MCP available from the supplier, but we could not choose a specific size. This seriously impeded our ability to match the size of the patch with that of the host vein, which we would have preferred.

The importance of using ABO-compatible patches cannot be overstated. We identified 4 patients with severe monocusp insufficiency in whom the cusps had disintegrated. Replacement of the MCPs with ABO-compatible patches restored competence and healed their ulcers. Of 9 patients with severe monocusp insufficiency who refused treatment, 7 had undergone implantation before we began using ABO-compatible patches. In addition, blood typing had a positive effect in patients who underwent concomitant MCP implantation in other veins. Unfortunately, the high percentage of MCP incompetence that occurred in all cases before the use of ABO-compatible patches made it impossible for us to draw any significant conclusions about concomitant patch implantation.

Lack of thrombus formation in any of the monocusps was a welcome finding. However, our selection process may have encouraged such results: we restricted this study to patients who had nonthrombotic venous reflux and to those who had primary venous insufficiency. Thus, we may have selected a population of patients with a relative freedom from intravascular thrombosis. Ultimately, any technique to restore venous valvular competence, including this one, should be applicable to patients who have postthrombotic valvular insufficiency. Such patients are more difficult to treat, and valve repair is less durable than it is in patients with primary venous valvular insufficiency. ¹⁶ Our early clinical attempts to combine cryopreserved pulmonary allograft MCP placement with venous bypass surgery for thrombotic disease led to MCP malfunction and failure. This failure was primarily due to thrombosis of the venous bypass.^* Patients with thrombotic disease who undergo monocusp patch placement are likely to need long-term or lifetime anticoagulation.

Among the important issues that need to be resolved is whether the common femoral vein is the best site at which to correct venous valvular insufficiency. The technique of valvuloplasty described by Kistner was used to correct valves in the superficial femoral vein. ¹ It seems as though he chose the superficial femoral vein primarily for safety reasons: failure of the repair could be treated by ligation of that vein. Some investigators have expressed the opinion that the popliteal vein valve serves as the “gatekeeper” above the calf muscle venous pump. ¹⁹ This issue is still not settled.

Superficial venous insufficiency cannot be ignored and should be treated, together with incompetent perforating veins, before deep venous treatment. ^20,21 The question of whether the monocusp can be used to correct superficial insufficiency was not examined in our study. We did find, however, that when we underestimated the severity of the superficial venous incompetence or did not ligate the perforating veins, the ulcers did not heal. Multisystem venous valvular insufficiency is a very complicated syndrome that requires more complex therapy than just correction of the deep venous system.

We chose the common femoral vein for MCP implantation due to its size and easy surgical accessibility. No patient in whom we implanted a xenograft or allograft patch had any untoward result other than recurrent reflux. Therefore, we think that it is very safe to implant these patches in the common femoral vein if the patient has primary valvular insufficiency in this vein.

Another important issue that needs to be examined is whether more than 1 MCP needs to be implanted in patients who have deep venous valvular insufficiency at multiple sites. Two of our patients underwent concomitant common femoral and popliteal monocusp implantation. In one of these patients, the 2 monocusps were competent, and the ulcer healed. Conversely, the other patient had incompetence in both valves, and the ulcer did not heal. Further studies are needed to determine whether multiple-site implantation is efficacious.

Thus far, our experience with the allograft monocusp pulmonary patch has yielded these observations:

• Monocusp implantation is technically simpler than is internal or external valvuloplasty, and the MCP can be used in patients with severely damaged or absent valves.

• Monocusp competence can be achieved at the time of implantation in all cases.

  Overcorrection of the vein at the time of operation reduces the chance of monocusp insufficiency that can occur secondary to postoperative dilatation of the vein.

  The MCP offers resistance to reflux flow and allows the ulcer to improve or heal, even if mild-to-moderate insufficiency develops. ¹⁴

  The direct relationship between severe MCP insufficiency and ulcer recurrence, as seen with the glutaraldehyde-fixed bovine monocusp pericardial patch, is also present between ulcers and the cryopreserved allograft MCPs.

  If the MCP is replaced and competence is achieved, the ulcer heals.

The competent monocusp patch and its healing effect on venous ulcers can be a useful tool in the study of venous disorders. Controlled, randomized clinical trials that evaluate monocusp patch valvuloplasty and other methods of correction of venous valvular insufficiency, and compare them with conservative compression therapy, are needed in order to establish the relative therapeutic value of each technique.