Divi V, Proctor MC, Axelrod DA, Greenfield LJ. Thoracic Outlet Decompression for Subclavian Vein ThrombosisExperience in 71 Patients. Arch Surg. 2005;140(1):54-57. doi:10.1001/archsurg.140.1.54
Copyright 2005 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2005
There is a difference in outcomes when patients have neurogenic thoracic outlet syndrome in addition to subclavian vein thrombosis.
Analysis of a prospectively developed database, medical record review, and a patient questionnaire were used to summarize clinical experience from December 1990 to December 2001 on the basis of the patient’s original evaluation. Patients were stratified on the presence (group 1) or absence (group 2) of additional neurogenic pathologic features.
Of 928 patients evaluated for thoracic outlet syndrome, 71 underwent 73 operative procedures for subclavian vein obstruction. Men predominated (55%), and the mean age was 32 years. Group 1 (41%) had more preoperative disability, a higher incidence of persistent pain (24%), and less likelihood of returning to full activity compared with group 2 (67% vs 93%; P = .01). Catheter-directed thrombolysis was used in 65% of veins. Preoperative balloon angioplasty was used selectively (34%), and only 4% required stents. Supraclavicular decompression and venolysis were usually delayed 3 weeks to allow for healing of the venous endothelium. Complications included wound infection (3%) and postoperative hematoma (8%).
Patients with isolated subclavian vein obstruction have a more favorable outcome relative to those with combined neurogenic and venous pathologic features. Decompression following thrombolysis should be delayed to reduce the incidence of postoperative complications.
More than 100 years ago, a syndrome of spontaneous subclavian vein thrombosis in otherwise healthy patients was described.1 This condition was subsequently found to result from costoclavicular compression of the vein and became known as “effort thrombosis” or Paget-Schroetter syndrome.2,3 As the venous variant of the thoracic outlet syndrome (TOS), it is less controversial than the neurogenic variant and can be readily confirmed by ultrasound evaluation or venography. However, controversies do exist relative to the need for operative thoracic outlet decompression, the timing and type of procedure performed, the indications for balloon angioplasty or stent placement, the indications for and timing of perioperative thrombolytic or anticoagulant therapy, and the indications for decompression of the uninvolved extremity.
In a recent retrospective review of patients managed in a multidisciplinary clinic for neurogenic TOS, we demonstrated that concurrent depression, single marital status, and limited formal education were associated with a less favorable outcome of operative treatment.4 To evaluate determinants of outcome for venous TOS with or without a neurogenic component, we conducted a review of 71 consecutive patients managed operatively from December 1990 to December 2001.
Of the 928 patients seen for symptoms of TOS from December 1990 to December 2001, 71 were identified with subclavian vein compression and/or thrombosis and underwent 73 operative procedures. Only 9 patients (13%) had venous compression alone with disabling symptoms of engorgement without thrombosis, while the 10th patient had thrombosis on one side and symptomatic evidence of compression without thrombosis on the opposite side. Venograms in these patients confirmed the presence of external compression at the level of the first rib equivalent to the compression seen in the patients with thrombosis.
The diagnosis was made by contrast venography in 41 patients (58%), duplex ultrasonography in 25 patients (35%), or other modalities in 5 patients (7%), and thrombolytic therapy was used in 40 patients (56%). Balloon angioplasty was used selectively by the interventional radiology staff, but no intravascular stents were placed preoperatively.
All thoracic outlet decompressions were performed from a supraclavicular approach, supplemented by an infraclavicular incision for more extensive first-rib removal on 4 occasions. The infraclavicular approach alone was used on only 1 occasion in association with a Fogarty thrombectomy. Claviculectomy was used on 5 occasions to facilitate exposure for jugular vein transposition to the subclavian vein or because it was contributing to neurogenic symptoms and displaced following a previous Mumford procedure. The customary TOS decompression consisted of anterior scalenectomy, division of the middle scalene and any bands identified including a scalene minimus, and partial excision of the first rib from the vicinity of the vein followed by venolysis of scar tissue. Only the anterior 2 to 3 cm of the rib was resected to remove the base of venous compression. The venolysis allowed expansion of approximately two thirds of the venous circumference that could be visualized beneath the clavicle. Patients with neurogenic symptoms also underwent brachial plexus neurolysis. Partial sternotomy with preservation of the clavicle as described by Molina5 was used on 2 occasions for venous reconstruction. Cervical ribs found in 2 patients were excised.
Patients were followed up at intervals ranging from 1 to 60 months (mean, 10 months) by interview and physical examination. Only patients with persisting symptoms were followed up beyond 6 months. Long-term outcomes were obtained in 13 patients by a written questionnaire designed to assess current functional status and level of satisfaction.4 Patients were stratified into 2 groups before treatment based on the presence of neurogenic or shoulder symptoms in addition to venous findings.
Statistical analysis was conducted using Microsoft Excel 97 (Redmond, Wash) and SAS Version 8.0 (Cary, NC). The t test was used for analysis of continuous variables, and the Fisher exact test was used for dichotomous variables. A P value less than.05 was considered significant.
Of the 71 patients diagnosed with venous TOS, 29 (41%) also had symptoms of shoulder abnormality or neurogenic TOS consisting of pain and paresthesias (group 1) elicited at the time of original evaluation. The remainder had venous obstructive symptoms alone (group 2). The average age in group 1 (33 years) was comparable with group 2 (31.5 years). Patient characteristics are listed in Table 1. In contrast to the previously reported patients with neurogenic TOS,4 men were more frequently affected in this series (38 [54%]), and in group 2, the male preponderance was greater (47 [66%]). The dominant upper extremity was also more frequently affected (43 [61%]), and 41 (57%) reported a precipitating event that was more likely to be repetitive strenuous motion than direct trauma. Patients in group 1 had more preoperative disability affecting driving (52% vs 10%; P<.001) and personal care activities (58% vs 23%; P = .008). They were also more likely to have had preoperative physical therapy (43% vs 12%; P = .005). There were 61 patients with venographic evidence of thrombosis in 62 subclavian veins, and 3 patients (5%) had evidence of pulmonary embolism. Of the 62 thrombosed veins, only 40 (65%) were treated with lytic agents, and in 5 (8%) of them it was more than 2 weeks after onset. Patients in group 1 were less likely to have undergone preoperative thrombolysis (40% vs 64%; P = .05) or angioplasty (17% vs 41%; P = .03). In 5 patients, rethrombosis led to a second round of lytic therapy. One patient had a stent placed following lytic therapy at another hospital and was followed up until he developed hoarseness from stent fragmentation and penetration of his recurrent nerve, requiring operative removal, subclavian artery repair, and venoplasty. A second patient had a stent placed following angioplasty without sequelae. Two patients had prior first-rib resection by the axillary approach. The median time between onset of symptoms and our evaluation was 82 days, ranging between 4 and 788 days. Most patients were working prior to the development of symptoms (71%), but only 50% were working at the time of evaluation with no difference between groups. None of the patients in group 2 were involved in litigation, but 7% in group 1 were. Nine percent were receiving workers’ compensation with no difference between groups.
The interval between onset of symptoms and operative treatment ranged from 4 days to 84 months and did not differ between groups. The recommended interval was 3 to 6 weeks to allow for endothelial healing except in 2 cases where operative treatment followed thrombolysis.
Postoperative thromboprophylaxis was used in 31 patients (44%), consisting of low-molecular-weight heparin in 27 and aspirin in 4. Postoperative physical therapy was prescribed for 18 patients (26%), predominantly in group 1, to treat the residual neurogenic symptoms or shoulder abnormality (13 patients [72%]). The only intraoperative complication occurred when the tip of a Fogarty catheter broke during an attempted thrombectomy in a patient in whom thrombolysis failed. The distal fragment was not retrieved because of patient refusal. The most common postoperative complication was persistent pain in 7 patients, all in group 1 (24%), 4 of whom had neurogenic symptoms while 3 had preexisting shoulder abnormality. There were 6 postoperative hematomas (8%) in the series, all in patients who received perioperative anticoagulation. Only 2 required operative drainage. Both patients were operated on immediately following thrombolysis and developed hematomas. Two patients (3%) developed recurrent thrombosis, each at an interval of 2 months, and both responded well to anticoagulation. Only 1 patient had a documented prothrombotic condition (anticardiolipin syndrome). Two patients (3%) had postoperative wound infections, consisting of cellulitis in 1 and an infected sternotomy following postoperative septic arthrosis of the knee in the other. Preoperative angioplasty was performed in 24 patients (34%) and repeated in 2 (3%). Three patients had stents placed postoperatively, 1 of which thrombosed, requiring lytic therapy, and led to 1 of the hematomas reported as a complication.
The functional results from the venous standpoint were satisfactory in all but 1 patient in whom thrombolysis failed; the patient had persistent edema requiring an elastic garment.Table 2 summarizes patient outcomes. The patients who underwent claviculectomy reported no alteration in extremity function and little concern regarding the cosmetic effects. There was improvement in pain in 85% of patients (71% in group 1 and 93% in group 2; P = .01) and paresthesias in 94%, with 77% able to return to full presymptomatic activity. In group 1, 67% returned to full activity while 93% did so in group 2 (P = .01). Overall, 81% returned to work without restrictions, and an additional 9% were able to work with restrictions. In group 1, 14% were unable to work, while only 2.5% were so limited in group 2 (P = .002). A change in jobs was required in 33% in group 1 and none in group 2 (P = .03).
The etiology of primary subclavian vein thrombosis has been assumed to be secondary to mechanical damage to the vein by bony or muscular compression at the thoracic outlet, although a small percentage of patients will be found to be prothrombotic.6 None of our patients had a history of other thromboses and only 1 was prothrombotic with the anticardiolipin antibody. Since the disorder occurs in otherwise healthy, young patients, typically male and following vigorous activity with the upper extremities, we have not routinely screened for inherited or acquired prothrombotic disorders but that may be appropriate for recurrent thrombosis. Acute swelling and pain, usually in the dominant arm and accompanied by dilated superficial venous collaterals over the shoulder and chest, should make the diagnosis obvious. However, delays in diagnosis are common with a tendency to attribute the edema to cellulitis or an allergic reaction to a puncture wound. The diagnosis can be readily confirmed by contrast venography or by duplex examination; however, the latter cannot interrogate vein beneath the clavicle.
Once the diagnosis is confirmed, current management is directed to restoring venous patency by catheter-infused thrombolysis, which was used in 40 cases (65%) in our series. The subsequent venogram to confirm patency will usually show residual extrinsic compression of the vein from a structural abnormality that can be relieved by operative decompression. Failure to restore venous patency either because of delay in diagnosis or treatment by anticoagulation alone results in persisting symptoms in the majority of patients.7,8 Management of the residual defect has been more controversial, with some authors reporting satisfactory outcomes using balloon angioplasty alone for residual venous deformities.9 Review of their follow-up data, however, shows that 3 (27%) of 11 patients with objective studies had developed rethrombosis, albeit without disabling symptoms. This points to the importance of the individual’s upper extremity work and recreational circulatory requirements. Our experience with recurrent symptomatic rethrombosis after anticoagulation is discontinued has prompted us to recommend operative decompression. We have also favored a more conservative delay of 3 to 6 weeks in timing of the procedure to allow endothelial healing to occur. Although a legitimate case for immediate operative correction of the structural defect following thrombolysis can be made, our experience has been that there is a greater risk of postoperative hematoma formation. This has also been seen in other recent series with hematoma or intrathoracic bleeding in 17%,10 contrasting with the 8% incidence in this report, all of which were associated with anticoagulation or the need for thrombolysis. Others have reported more favorable outcomes with no mention of postoperative bleeding.11 Preoperative balloon angioplasty was performed in 23 patients (33%), more in the latter period of the study than the earlier cases. We believe that angioplasty is more likely to be of lasting value following thoracic outlet decompression than preoperatively, as has been reported by others.12 Postoperative angioplasty was used in 4 patients (6%), and stents were placed in 3, 1 of which thrombosed. We share the concern expressed by Green et al13 for the long-term fate of stents in this location. We failed to find a significant correlation between angioplasty and outcome, although the 1 patient with significant postoperative edema requiring elastic compression did not undergo angioplasty. Since access to perform angioplasty depends on the ability to pass a guidewire through the obstructed vein, the more extensive and fibrotic occlusions are less likely to be amenable and to remain functionally significant.
Although we performed more extensive rib resections in the earlier years of the study, our recent experience limiting rib resection to the anterior portion of rib beneath the vein has been favorable, and there did not seem to be any advantage in adding a second incision as recommended by Thompson et al,14 which was used in 4 cases. Once the scalene muscles are removed, eliminating rib elevation, the divided rib should be restrained by the remaining intercostal muscles to the advantage of the brachial plexus as well. The venolysis, removing scar tissue from the supraclavicular approach, was also satisfactory provided it was extended to the junction with the jugular vein where occasional bands are found. Since the supraclavicular approach allows both complete resection of the anterior scalene muscle and satisfactory venous release by venolysis and rib resection, it has been preferred to the alternative axial and infraclavicular options. Venous reconstruction was performed in 4 patients (6%), using an autologous vein patch in 2 and jugular transposition in 2. The indications for these included both irreversible scarring and flow demands, specifically an arteriovenous fistula for maintenance dialysis in 1 patient.
The limitations of this study include the absence of objective studies of vein patency and the limited long-term follow-up. Patients with satisfactory functional results are usually reluctant to undergo the time and expense of additional studies and often fail to return for follow-up examinations. The inclusion of patients with both shoulder and/or neurogenic symptoms and venous obstruction also makes interpretation of postoperative symptoms more difficult. Chronic pain problems in 24% of group 1 patients is typical.4 The role of postoperative physical therapy, which was used in 18 patients (26%), is dictated by the presence of myofascial pain, which differs from the neurogenic paresthesias and is more amenable to appropriate treatment consisting of muscle stretching and energy techniques rather than strengthening, which is detrimental. Only 1 patient developed symptoms in the opposite extremity requiring operation, and we have not felt that prophylactic procedures should be performed on the contralateral side in the absence of symptoms. Although some authors have lamented the lack of prospective randomized trials to formulate treatment protocols,15,16 the heterogenous nature of the patient population in terms of extent of involvement, initial treatment, and variable treatment options pose a formidable barrier.
Experience with this group of 71 patients leads us to recommend prompt initial treatment with lytic agents and anticoagulation and surgical decompression after a 3- to 6-week healing period. The role of angioplasty following decompression for persistent venous stenosis seems appropriate, but the use of vascular stents awaits long-term assessment. A multidisciplinary approach to the diagnosis and management of these complex patients offers the best chance for relief of symptoms and return to functional status.
Correspondence: Mary C. Proctor, MS, Department of Surgery, University of Michigan, 1327 Jones Dr, Suite 201, Ann Arbor, MI 48105 (email@example.com).
Accepted for Publication: September 10, 2004.
Previous Presentation: This study was presented at the 26th Annual Meeting of the Southern Association for Vascular Surgery; January 18, 2002; Miami Beach, Fla.