Long-term thrombotic recurrence after nonoperative management of Paget-Schroetter syndrome

Transcription

1 From the Western Vascular Society Long-term thrombotic recurrence after nonoperative management of Paget-Schroetter syndrome Jason T. Lee, MD, a John K. Karwowski, MD, a E. John Harris, MD, a Jason S. Haukoos, MD, MS, b,c and Cornelius Olcott IV, MD, a Stanford, Calif; and Denver, Colo Background: The purpose of this study was to determine the clinical predictors associated with long-term thrombotic recurrences necessitating surgical intervention after initial success with nonoperative management of patients with primary subclavian vein thrombosis. Methods: Sixty-four patients treated for Paget-Schroetter syndrome from 1996 to 2005 at our institution were reviewed. The standardized protocol for treatment includes catheter-directed thrombolysis, a short period of anticoagulation, and selective surgical decompression for patients with persistent symptoms. First-rib resection was performed in 29 patients (45%) within the first 3 months, with a success rate of 93%. The remaining 35 patients (55%) were treated nonoperatively and constitute this study s population. Results: Of the 35 patients with successful nonoperative management, 8 (23%) developed recurrent thrombotic events of the same extremity at a mean follow-up time of 13 months after thrombolysis (range, 6-33 months). These eight patients subsequently underwent first-rib resection with a 100% success rate without further sequelae at a mean follow-up time of 51 months (range, months). The other 27 patients remained symptom free at a mean follow-up interval of 55 months (range, months). Bivariate analyses determined that the use of a stent during the initial thrombolysis was associated with thrombotic recurrence (P.05). The recurrence group was also significantly younger than the asymptomatic group (22 vs 36 years; P.01). Sex, being a competitive athlete, a history of trauma, whether the dominant arm was affected, time of delay to lysis, initial clot burden, response to original lysis, use of adjunctive balloons or mechanical thrombectomy devices, residual stenosis on venography, length of time on warfarin, and patency of the vein on follow-up duplex examination were all characteristics not associated with long-term recurrence after nonoperative management. Conclusions: Conservative nonoperative management of primary subclavian vein thrombosis can be successfully used with acceptable long-term results. A younger age (<28 years old) and the use of a stent during initial thrombolysis are factors associated with long-term recurrent thrombosis. Younger patients should be offered early surgical decompression, and the use of stents without thoracic outlet decompression is not indicated. (J Vasc Surg 2006;43: ) 1 Since the initial descriptions by Paget and von Schroetter 2 more than a century ago, the optimal therapy of effort-related primary subclavian vein thrombosis has been debated. Although Paget-Schroetter syndrome (PSS) occurs infrequently and accounts for only approximately 1% 3 to 4% of all deep vein thromboses, affected patients are From the Division of Vascular Surgery, Stanford University Medical Center, a Department of Emergency Medicine, Denver Health Medical Center, b and Department of Emergency Medicine, University of Colorado Health Sciences Center. c Competition of interest: none. Presented at the Twentieth Annual Meeting of the Western Vascular Society, Park City, Utah, Sep 24-27, Reprint requests: Cornelius Olcott IV, MD, Division of Vascular Surgery H3600, Stanford University Medical Center, 300 Pasteur Dr, Stanford, CA ( olcott@stanford.edu) /$32.00 Copyright 2006 by The Society for Vascular Surgery. doi: /j.jvs often young and otherwise healthy. Major morbidity can develop in this patient population and can lead to longterm disability if PSS is not promptly diagnosed and treated. Early treatment algorithms involving arm elevation and rest coupled with anticoagulation alone resulted in significant recurrence rates and persistent upper extremity symptoms in more than 50% of cases. The advent of 4,5 thrombolytic therapy and the recognition of extrinsic venous compression in the pathophysiology of PSS have led to the more modern treatment paradigm of operative thoracic 6-13 outlet decompression after successful thrombolysis. Even more recently, hybrid procedures with endovascular techniques and open surgery have been proposed to potentially improve long-term vein patency. What remain 14,15 controversial are the timing of surgery and even the need for surgery. Various operative strategies described in the literature range from immediate first-rib resection during the same hospitalization to some delayed time interval 1 to 16,17 3 months later to no surgery at all. At our institution, a more conservative approach has been adopted to the treatment of PSS. Patients with resolution of symptoms after initial thrombolytic therapy and a short period of anticoagulation do not immediately undergo thoracic outlet decompression. Indications for early surgery after thrombolysis and the trial of outpatient anticoagulation include (1) the persistence or recurrence of venous hypertensive symptoms, (2) any evidence for recurrent or new thrombus, (3) obstruction of perivenous collaterals with abduction/external rotation in the face of subclavian vein occlusion, and (4) evidence of persistent or recurrent vein injury, as evidenced by significant wall thickening.

2 JOURNAL OF VASCULAR SURGERY Volume 43, Number 6 Lee et al 1237 Recurrent symptoms 1 st rib resection n=29 + Duplex Ultrasound n=64 Venogram/Thrombolysis 52/64 (81%) IV heparin warfarin n=64 Remain symptom-free n=27 Outcomes from our algorithm have been previously reported and demonstrated that 41% (9 of 22) of patients did not demonstrate the above-mentioned indications and safely avoided surgery. 16 This previous study was admit - tedly limited by a relatively small number of patients and short-term follow-up, similar to the few other studies that observed patients treated nonoperatively after successful thrombolytic therapy. 17,18 The potential benefit of a more conservative approach to thoracic outlet decompression is to avoid the possible operative complications, such as brachial plexopathies, phrenic nerve paresis, subclavian arterial injuries, and lymphatic leaks, with incidences ranging from 15% to 20%. 19 The purpose of this study was to examine the long-term results in our updated series of PSS patients treated by thrombolytic therapy and outpatient anticoagulation without thoracic outlet decompression. We sought to determine whether there were any clinical or radiographic predictors of symptomatic recurrence after initial nonoperative therapy. Detecting such factors might help to more accurately determine which subset of PSS patients would benefit most from conservative management without thoracic outlet decompression. PATIENTS AND METHODS Symptoms resolved warfarin x3 months n=35 Recurrent thrombosis 1 st rib resection n=8 Fig 1. Summary of interventions of patients (n 64) treated for Paget-Schroetter Syndrome at Stanford University Medical Center from A review of a prospective registry of PSS patients treated at a single institution from 1996 to 2005 was conducted. Sixty-four patients were consecutively treated with an algorithm of selective surgical therapy after thrombolysis and full anticoagulation with warfarin for at least 1 month (Fig 1). Follow-up examinations of these patients were completed through August Clinical diagnosis of primary axillosubclavian venous thrombosis was on the basis of a history and physical examination that typically included an antecedent episode of physical exertion involving the affected extremity, the absence of any secondary mechanical or physiologic risk factors, and the characteristic arm pain, cyanosis, and swelling. The diagnosis was confirmed with color-flow venous duplex ultrasonography in all cases (n 64). Contrast venography was obtained in 88% of patients, and 81% underwent thrombolytic therapy. The workup and management were initiated at an outside referring institution in 29% of patients. The remaining patients either presented to our institution or were promptly transferred to us at the time of diagnosis. Patients included in this series who did not undergo initial thrombolysis or venography were all from outside institutions where there was a delay of longer than 7 days from diagnosis to consultation or where symptoms had already begun to significantly improve on anticoagulation regimens initiated at the referring institution. All of the nonvenography patients (12%) had imaging of the thoracic outlet by magnetic resonance imaging or computed tomography at the referring institution. Details of the venography technique and thrombolytic regimens performed on patients at our own institution have been previously reported. 16 Available venograms or their reports were reviewed, and initial thrombus load, use of balloons or mechanical thrombectomy devices, postlysis improvement, and residual stenoses were recorded. Postlysis improvement was characterized as complete, partial, or minimal/occluded. Clinical improvement was measured on a subjective basis as minimal, moderate, or severe symptomatology. At discharge after the initial hospitalization, patients were transitioned from intravenous heparin to warfarin. Achieving an international normalized ratio of 2.0 to 3.0 allowed for discharge, and this therapeutic level was monitored and maintained closely on an outpatient basis by a dedicated anticoagulation clinic. Discharged patients were instructed to resume activities of daily living but to avoid any strenuous, repetitive overhead activity, especially any that may have precipitated the patient s initial thrombotic event. Patients were routinely set up with physical therapy as outpatients to maintain the range of motion in the affected arm. All patients were reevaluated in the vascular surgery clinic at 1 month with a repeat duplex ultrasound interrogation of the innominate, subclavian, axillary, internal jugular, and brachial venous segments. Of the 64 patients, 29 patients were noted to have recurrent or persistent symptoms of venous hypertension (n 20) or demonstrated ultrasound evidence of recurrent thrombus (n 5) or significant wall thickening (n 4). This group was offered surgical decompression of the thoracic outlet, and all 29 patients underwent first-rib resection within the first 6 months after thrombolysis. Our surgical approach is via a single supraclavicular incision without any attempt at venous reconstruction or venoplasty during thoracic outlet decompression. Success rates of the operative approach in terms of relief of symptoms occurred in 93% of these 20 patients, and they have all returned to work successfully. The clinical characteristics of the entire cohort of 64 pa-

3 1238 Lee et al JOURNAL OF VASCULAR SURGERY June 2006 Table I. Clinical and venographic features of all 64 Paget-Schroetter patients treated from 1996 to 2005 Variable Total (n 64) Early operation (n 29) Nonoperative (n 35) Mean age (y) Male sex (%) Competitive athlete (%) Right arm involvement (%) Dominant arm (%) Delayed therapy (%)* Total occlusion (%) Thrombolytic therapy (%) Balloon venoplasty (%) Mechanical thrombectomy (%) Stent placement (%) Complete response to thrombolysis Residual stenosis after thrombolysis Patent last duplex (%) Follow-up time (mo) The early-operation group was treated with thoracic outlet decompression within the first 6 months of thrombolysis. The nonoperative group all had resolution of symptoms and completed a regimen of outpatient anticoagulation. P values comparing the two groups for all variables were.05. Statistical analyses were performed by using Systat 8.0 (SPSS Inc, Chicago, Ill), SAS Version 9.1 (SAS Institute, Inc, Cary, NC), and classification and regression tree Version 5.0 (Salford Systems, San Diego, Calif). A P value.05 was considered statistically significant, and no adjustment was made for multiple comparisons. RESULTS Between June 1996 and August 2005, 35 patients with PSS were treated under a conservative algorithm without operative decompression. The mean age of this nonoperative cohort was years (range, years), and 20 (57%) were male. Eleven (31%; 95% CI, 17%-49%) of the patients were competitive athletes, including three collegiate-level swimmers, two high-school baseball pitchers, one college football player, one college water polo player, and four weightlifters. The dominant arm was affected in 23 (66%; 95% CI, 48%-81%) of the 35 patients. Twenty-five patients (71%) were engaged in sports- or work-related activities involving repetitive, heavy upper extremity physical movement. Thrombolytic therapy was initiated in 29 (83%) of the 35 patients and yielded complete clot dissolution in 22 (76%; 95% CI, 57%-90%) cases, partial residual thrombus in 6 (21%; 95% CI, 8%-40%), and only 1 (3%; 95% CI, 0%-18%) failed thrombolysis. The patient whose thrombolytic therapy failed was a 22-year-old soccer player who tients are shown in Table I. It is interesting to note that underwent the obligate 1-month anticoagulation regimen, there were no significant differences clinically or radiographically when the early operation group was compared with the nonoperative group. The remaining 35 patients who constituted the nonoperative group had complete or near-complete resolution of symptoms along with duplex ultrasound scans showing no new evidence of thrombus or subclavian vein thickening. This nonoperative cohort, representing the focus of this study, was subsequently maintained on warfarin for a minimum of 3 months (range, 3-12 months; mean, 5.2 months). These patients were reevaluated once again at 3 months with symptom inventory, physical examination, and duplex ultrasonography. If they continued to do well clinically at that point, warfarin was discontinued, and all activity restrictions were lifted. Follow-up consisted of repeat duplex ultrasonography at 6 months and yearly physical examinations thereafter. Pretreatment clinical data, radiographic imaging data, and long-term follow-up data on all 35 patients treated nonoperatively were entered into an electronic database and his 1-month duplex examination showed significant collateral development, but the subclavian vein remained occluded. His symptoms, however, were significantly improved at this follow-up, and he therefore was not offered thoracic outlet decompression, completed 3 months of warfarin therapy, and was able to return to his normal physical activity. Of the patients who did not receive thrombolytic therapy (6 of 35), their care was initiated at outside institutions that lacked thrombolytic capabilities. These six patients were all placed on heparin and converted to warfarin and were subsequently referred to our institution for definitive management or a second opinion. Subclavian vein narrowing or impingement with provocative maneuvers was seen on venography, usually at the location of the subclavian vein crossing the first rib, in 25 (86%; 95% CI, 68%-96%) of the 29 patients who underwent thrombolytic therapy. Balloon venoplasty was attempted in 48% of patients, and mechanical thrombectomy devices were used in 16% of patients. Indications for the use of these devices were at the discretion of the interventionalist (Microsoft Excel; Microsoft Corporation, Redmond, performing the procedure and typically included focal residual Wash). The Wilcoxon rank sum test or Fisher exact test was used to test for statistical differences between groups, where appropriate. Measurements are presented as means SD, and 95% confidence intervals (CIs) are reported when applicable. Because of the small sample size, classification and regression tree analysis, a form of binary recursive partitioning, was used instead of a logistic regression analysis to assess for multivariable relationships between all potential 21 predictor variables and whether the disease recurred. narrowing of the vein, lack of rapid improvement with thrombolytic agents, or personal experience. No patient received thrombolytic therapy for longer than 48 hours, and they were all converted to intravenous heparin therapy and transitioned to oral warfarin. This group of patients was treated with warfarin for a mean time of months (range, 3-12 months), with resolution of symptoms to baseline according to each patient. All 35 patients were discontinued from warfarin therapy, and their

4 JOURNAL OF VASCULAR SURGERY Volume 43, Number 6 Lee et al 1239 Table II. Summary of bivariate analyses of clinical and venographic predictors of long-term thrombotic recurrence after initial nonoperative treatment Variable Initial nonoperative (n 35) Group A recurrent (n 8) Group B stable (n 27) P value Mean age (y) Male sex (%) Competitive athlete (%) Right arm involvement (%) Dominant arm (%) Delayed therapy (%)* Total occlusion (%) Thrombolytic therapy (%) Balloon venoplasty (%) Mechanical thrombectomy (%) Stent placement (%) Complete response to thrombolysis Residual stenosis after thrombolysis Duration of anticoagulation (mo) Patent last duplex scan (%) Follow-up time (mo) Group A includes patients in long-term follow-up who developed recurrent thrombosis and subsequently underwent thoracic outlet decompression. Group B patients had no thrombotic recurrences and remained clinically asymptomatic in follow-up. *Defined as greater than 1 week. activity restrictions were lifted. Review of the computerized charts indicated that all 35 patients returned to a satisfactory activity level in the short term without further venous hypertensive symptoms. In long-term follow-up, however, 8 (23%) of the 35 patients (95% CI, 10%-40%) had thrombotic recurrences in the ipsilateral extremity at a mean postlytic time of 13 8 months (range, 6-33 months) and constitute group A. The remaining 27 patients had no evidence of recurrent symptoms at a mean follow-up time of months (range, months) and constitute group B. Both groups had returned to work at the latest follow-up. A comparison of the clinical and radiographic features of the two groups is listed in Table II. All of the patients in group A (n 8) were treated again with thrombolysis and, because of their second thrombotic episode, were all recommended to undergo thoracic outlet decompression with subclavian venolysis. Two patients (25%) in this group initially had stents placed at an outside referring institution and in follow-up were found to have stent fractures and subsequent rethrombosis (Fig 2). All eight patients had successful operations with complete resolution of symptoms. This group currently remains asymptomatic without any further recurrences at a mean follow-up time of months (range, months) since their operation. At last follow-up, they have all returned to work without disability. Table III outlines the characteristics of these eight patients who had long-term thrombotic recurrences. Postoperative complications for group A included three small pneumothoraces that were treated expectantly and were self-resolving over 48 hours without the need for chest tube thoracostomy. No lymphatic leaks or brachial nerve plexus injuries were documented during follow-up. Fig 2. A crushed stent is seen in the subclavian vein, with damage in the area of compression by the clavicle over the first rib. Bivariate analyses determined that the use of a stent during the initial thrombolysis was associated with thrombotic recurrence (P.05). The recurrence group was also significantly younger than the asymptomatic group (22 vs 36 years; P.01). Sex, being a competitive athlete, a history of trauma, whether the dominant arm was affected, the time delay to thrombolysis, the initial clot burden, the response to initial thrombolysis, the use of adjunctive balloons or mechanical thrombectomy devices, residual stenosis on venography, the length of time on warfarin, and patency of the vein on follow-up duplex examination were

5 1240 Lee et al JOURNAL OF VASCULAR SURGERY June 2006 Table III. Clinical characteristics of patients who demonstrated a second episode of effort thrombosis after initial nonoperative therapy Patient No. Age (y) Sex Athlete? Initial venogram Lysis response Residual stenosis Months of warfarin Time to recurrence (mo) 1 23 F No Partial occlusion/venoplasty/stent Complete Yes M No Total occlusion/venoplasty/stent Partial Yes M Baseball Total occlusion Complete Yes F No Total occlusion/venoplasty Complete Yes F Weightlifting Total occlusion Complete No M No Total occlusion Complete No F Swimming Total occlusion/mechanical Complete Yes 3 6 thrombectomy F No Total occlusion/mechanical thrombectomy/venoplasty Complete Yes 3 8 all characteristics not associated with long-term thrombotic recurrence. The multivariable analysis identified age as the only predictor of recurrence. All eight patients in group A were 28 years old or younger. Of the 16 patients 28 years or younger, 8 (50%; 95% CI, 25%-75%) had thrombotic recurrence. Of the remaining 19 patients, all greater than 28 years old, 0 (0%; 95% CI, 0%-18%) had recurrence. Long-term results of the early-operation group (n 29) reveal a clinical success rate of 93% at a mean follow-up time of months (range, 4-98 months). One of the adverse outcomes was in a 19-year-old baseball pitcher who, after successful thrombolysis by angiographic criteria, still had arm heaviness and fatigue clinically and on 1-month duplex scan had severe thickening of the subclavian vein and therefore underwent early decompression. Three months after the operation, however, he re-presented with arm swelling, and duplex scanning showed recurrent thrombosis at the operative site which required repeat lysis, mechanical thrombectomy, and long-term anticoagulation. Still, the overall duplex patency rate at last follow-up is 97% for the entire early-operation cohort, thus indicating the durable results of operative decompression. DISCUSSION PSS is an infrequent condition that portends significant morbidity if not managed appropriately. Long-term sequelae can be severely debilitating, compounded by the generally young and highly active nature of the afflicted patient population. Poorly managed, these patients may experience chronic arm swelling, pain, and early exercise fatigue that can severely limit function. The exact pathophysiology of this condition is poorly understood, but several well-established features have lent practitioners at least a theory. The typical history includes a strenuous, repetitive overhead activity, most often involving the dominant arm. Reproducibly observed are venographic and surgical evidence of chronic scarring around and within the vein where it courses between the clavicle, first rib, and anterior scalene muscle. The mechanical, traumatic nature of PSS involves a misalignment of musculoskeletal elements surrounding the subclavian vein such that a characteristic motion of these elements results in chronic constrictive trauma to the vein, with attendant endothelial damage, progressive stenosis, and ultimately a thrombotic tendency. Thus, over the last 20 years, a treatment paradigm has gained acceptance that rests on two principles: (1) eliminating or at least reducing the thrombus load to reestablish patency and (2) decompressing the thoracic outlet. Before aggressive venography and thrombolysis, conservative measures such as arm rest, elevation, and anticoagulation left more than 50% of patients with residual chronic disability. 4,5 Catheter-directed thrombolysis has proven very effective at eliminating acute thrombus, the development and persistence of which leads to the characteristic acute and chronic pain, swelling, and fatigue. Urschel and Razzuk s 10 unparalleled experience with 294 patients over 30 years highlights the paradigm shift toward expeditious thrombolytic therapy followed by prompt first-rib resection to obtain optimal results. Many other institutions have used this treatment algorithm with high rates of symptom resolution and freedom from recurrence. 6,7,9,11,12 Operative decompression is not, however, without its complications, which in many of the aforementioned series are reported up to 20% of the time and occasionally include devastating brachial plexopathies, arterial injuries, or significant hemorrhage. In fact, even in our series, operative success can be tempered by recurrent thromboses. Although surgery for arterial thoracic outlet syndrome is often associated with the highest complication rates, both neurogenic and venous thoracic outlet syndrome operations carry problems whether the approach is transaxillary or supraclavicular. 19 In a review of our operative patients, we had an 18% total complication rate, including wound hematomas and seromas (6%), phrenic nerve palsies (6%), 20 and small pneumothoraces (6%). This finding formed the impetus for our desire to delineate the subgroup of patients who truly require operative rib resection after thrombolytic therapy. The design of our algorithm used since 1996 contrasts with established published protocols that dictate mandatory surgical decompression of the thoracic outlet and even others that recommend that it be performed within the same hospitalization after thrombolytic therapy.

6 JOURNAL OF VASCULAR SURGERY Volume 43, Number 6 Lee et al 1241 We do not dispute that a substantial proportion of patients will require early thoracic outlet decompression to optimally alleviate symptoms, restore functionality, and prevent recurrence. In our series, 45% of patients required surgery within the first 1 to 3 months because of a lack of improvement of symptoms or persistent thrombus loads. These patients returned to the clinic still complaining of arm swelling, heaviness, fatigue, and cyanosis. Five patients were offered early surgery because of early rethrombosis after successful lytic therapy. We propose, however, that after expeditious thrombolysis and some period of anticoagulation, a substantial percentage can enjoy durable vein patency, a lack of symptoms, and a return to premorbid activity with a modest risk of rethrombosis. Gaining a better understanding of which patients are at the highest risk for recurrence is thus paramount to our strategy and was the focus of this study. Younger patients ( 28 years old) in this study were found to have a higher chance of recurrence, nearly 50%, and those older than 28 years old had a 0% recurrence rate at long-term follow-up. The reasons for this are likely multifactorial and probably difficult to prove. We postulate that returning to what is considered normal activity is likely more strenuous in the younger patients. It is also possible that the younger patients may have been less reliable in terms of limiting their activity to range-of-motion exercises with physical therapy during the healing phase after successful thrombolysis. Our patients were followed up closely in the first 3 months after initial thrombolysis, with regimented visits to physical therapy and teaching regarding activity restrictions. The involvement with competitive athletics did not affect the long-term recurrence in our nonoperative cohort; 38% of the recurrence group were athletes, compared with 30% in group B. In fact, a collegiate-level football player, baseball player, and swimmer all were in the nonoperative group that underwent thrombolysis and outpatient warfarin therapy with resolution of symptoms. These three patients avoided operative intervention and were able to continue their collegiate careers the following season without further thrombotic sequelae. This contrasts to Feugier et al, 22 who reported excellent results after immediate operative first-rib resection after successful thrombolytic therapy in 10 athletes treated in France. Our results indicate that one can safely treat a competitive athlete without thoracic outlet decompression and that being an athlete does not obligate operative intervention. Still, the longterm effects of PSS on the career of a competitive athlete are poorly reported in the literature, and recommendations regarding the treatment of these patients are difficult to generalize. The other factor by bivariate analysis that predicted longterm thrombotic recurrence was the use of subclavian stents. In two patients treated initially at an outside institution, stents were placed for residual stenoses after successful thrombolysis, and both of these patients had recurrences at 12 and 14 months with stent fracture and thrombosis (Fig This 2). finding corroborates what has reported in the literature. 8,14,23 We believe strongly that there is no role for stenting of the subclavian vein in the region of the thoracic outlet before operative decompression and that the persistent mechanical forces in this region likely lead to stent malfunction and failure, thus making the situation worse. The use of balloon venoplasty or mechanical thrombectomy devices did not have any effect on the likelihood of thrombotic recurrence in our series. Balloon venoplasty was used in approximately 50% of our patients without any correlation to long-term thrombotic recurrence. This finding highlights again the pathophysiology of PSS and indicates that temporary short-term therapies including balloons or so-called roto-rooter devices really are not effective in reversing the underlying physical compression of the thoracic outlet. Some have even suggested that this further traumatizes the vein and makes it more likely to subsequently thrombose, 6 although we did not observe this fact. Although most agree that stents and balloon venoplasty have a limited role in predecompression subclavian vein stenoses, an interesting hybrid type of intervention has recently been recommended after operative intervention. 14,15 We have not found this to be necessary in our series of patients undergoing first-rib resection, and at latest follow-up we have not had any thrombotic recurrences. In Kreienberg and associates 14 series, subclavian stents were deployed in 14 of 23 patients undergoing first-rib resection for residual stenoses. At a mean follow-up of 4 years, 5 of 14 were noted to have occluded stents; the patients noted mild swelling, and no further intervention was offered. We suspect that those patients may have been just as well off 15 without a stent. Schneider et alreported on 25 patients undergoing first-rib resection and subsequent balloon venoplasty in 16 patients with residual stenoses seen on intraoperative venogram. They concluded that this may reduce the incidence of postoperative recurrent thrombosis, although long-term results are pending. Venographic measures were not found in our series to be predictors of long-term recurrence. A similar percentage in the recurrent and stable group presented with complete occlusion (88% vs 78%), demonstrated a full response to original lysis (88% vs 71%), and had residual stenoses with provocative maneuvers on venography (86% vs 86%). Machleder 6 proposed that the full response to lysis and demonstration of stenosis on venography predicted higher rates of success after operative decompression. Certainly most of our patients would have been offered first-rib resection in many other institutions. Our findings suggest that relying on venographic factors cannot reliably distinguish those who benefit from operative intervention. Even in the previously mentioned large series of patients who underwent obligate thoracic outlet decompression, there were always a few patients who refused surgery or were not operated on, and follow-up on those patients echoes our results in that many of them do remarkably well 6 with low rates of recurrence. In Machleder s series, five patients underwent successful thrombolysis but avoided surgery by either refusing or failing to meet criteria. They all did well in follow-up, with no recurrent thromboses and

7 1242 Lee et al JOURNAL OF VASCULAR SURGERY June 2006 minimal to no symptoms, and returned to full activity. The clinical and radiographic characteristics of these five patients were not separately described in that article, so it is difficult to know whether those patients were similar to our nonoperative cohort. He further acknowledges that nine patients underwent unsuccessful lytic therapy who had residual stenoses and refused further treatment and that in follow-up four (44%) were free of symptoms. Again, these patients clinical and radiographic factors were not specifically delineated, but it highlights the fact that there is a population of patients that does not require thoracic outlet decompression. In the only other large series of nonoperative therapy 18 after initial thrombolysis for PSS, Lokanathan et al reported on 28 consecutive patients, of whom only 2 underwent first-rib resection in follow-up. Rethrombosis occurred in three (11%) patients at 8, 16, and 17 months after the initial thrombotic event. One patient with disease recurrence had residual symptoms after thrombolysis, which by our algorithm would have undergone thoracic outlet decompression. The second recurrence occurred in a patient who later had an unrelated clavicular fracture and required claviculectomy. Lokanathan et al, unfortunately, did not report what the comparable clinical and radiographic features were in this rethrombotic group, so we could not compare them with our current series. We believe that Lokanathan and associates algorithm is too conservative. By not offering decompression to any patient, they had 21 nonoperative patients with long-term follow-up via a validated questionnaire. Only 6 (28%) of the 21 patients had complete resolution of symptoms, and 62% still had mild symptoms. Still, their series points to the fact that there are patients who can be successfully treated without an operation. There are several limitations to our study: namely, the review of the venograms was performed in a retrospective fashion. The follow-up, although able to be assessed in all 35 patients in the study, was based on a review of the computerized patient charts. Also, functional outcome in these patients is difficult to compare between series unless a 18 more standardized evaluation tool is used. We have begun to prospectively evaluate all of our patients with venous, arterial, and neurogenic thoracic outlet syndrome before and after intervention with a rapid and reproducible symptom-scoring scale that has been validated in the hand surgical literature. 24 The small sample size in terms of the recurrent group of patients also predisposes to a type II error, where we might not discover a venographic finding that would be an important predictor of long-term recur- 18 rence. W e agree with Lokanathan et al that a prospective, randomized multicenter trial would provide the most insight into which patients truly require first-rib resection. For now, we are limited to reports from the various institutions that have taken an interest in the treatment of these patients. In summary, the optimal treatment of PSS patients certainly requires timely diagnosis, the initiation of systemic anticoagulation, thrombolytic therapy with positional venography, and close follow-up. An obligate period of operative delay allows the selection of patients who may not require first-rib resection. Patients who respond to outpatient warfarin therapy and demonstrate complete resolution of symptoms can be offered a nonoperative approach with acceptable rates of success. Younger patients who are potentially more active and might engage more rapidly in strenuous activities in the near future are perhaps the subgroup that should be offered surgical decompression. There do not seem to be consistent venographic findings that necessitate surgical intervention. Not all patients with PSS require operative decompression. REFERENCES 1. Paget J. Clinical lectures and essays. In: Anonymous. London: Longmans Green; Von Schroetter L. Erkrankungen der Gefasse. In: Nothangels Handbuch der Speciellen Pathologie und Therapie. Vol 15. Vienna: Holder; p Hurlbert SN, Rutherford RB. Basic data underlying clinical decision making primary subclavian-axillary vein thrombosis. Ann Vasc Surg 1995;9: Tilney NL, Griffiths HJG, Edwards EA. Natural history of major venous thrombosis of the upper extremity. Arch Surg 1970;101: Donayre CE, White GH, Mehringer SM, Wilson SE. Pathogenesis determines late morbidity of axillosubclavian vein thrombosis. Am J Surg 1986;152: Machleder HI. Evaluation of a new treatment strategy for Paget- Schroetter syndrome: spontaneous thrombosis of the axillary-subclavian vein. J Vasc Surg 1993;17: Adelman MA, Stone DH, Riles TS, Lamparello PJ, Giangola G, Rosen RJ. A multidisciplinary approach to the treatment of Paget-Schroetter Syndrome. Ann Vasc Surg 1997;11: Beygui RE, Olcott C, Dalman RL. Subclavian vein thrombosis: outcome analysis based on etiology and modality of treatment. Ann Vasc Surg 1997;11: Lee MC, Grassi CJ, Belkin M, Mannick JA, Whittemore AD, Donaldson MC. Early operative intervention after thrombolytic therapy for primary subclavian vein thrombosis: an effective treatment approach. J Vasc Surg 1998;27: Urschel HC, Razzuk MA. Paget-Schroetter syndrome: what is the best management? Ann Thorac Surg 2000;69: AbuRahma AF, Robinson PA. Effort subclavian vein thrombosis: evolution of management. J Endovasc Ther 2000;7: Angle N, Gelabert HA, Farooq MM, Ahn SS, Caswell DR, Freschlag JA, et al. Safety and efficacy of early surgical decompression of the thoracic outlet for Paget-Schroetter syndrome. Ann Vasc Surg 2001; 15: Divi V, Proctor MC, Axelrod DA, Greenfield LJ. Thoracic outlet decompression for subclavian vein thrombosis. Arch Surg 2005;140: Kreienberg PB, Chang BB, Darling RC, Roddy SP, Paty PSK, Lloyd WE, et al. Long-term results in patients treated with thrombolysis, thoracic inlet decompression, and subclavian vein stenting for Paget- Schroetter syndrome. J Vasc Surg 2001;33:S Schneider DB, Dimuzio PJ, Martin ND, Gordon RL, Wilson MW, Laberge JM, et al. Combination treatment of venous thoracic outlet syndrome: open surgical decompression and intraoperative angioplasty. J Vasc Surg 2004;40: Lee WA, Hill BB, Harris EJ, Semba CP, Olcott C. Surgical intervention is not required for all patients with subclavian vein thrombosis. J Vasc Surg 2000;32: Wilson JJ, Zahn CA, Newman H. Fibrinolytic therapy for idiopathic subclavian-axillary vein thrombosis. Am J Surg 1990;159: Lokanathan R, Salvian AJ, Chen JC, Morris C, Taylor DC, Hsiang YN. Outcome after thrombolysis and selective thoracic outlet decompression for primary axillary vein thrombosis. J Vasc Surg 2001;33:783-8.

8 JOURNAL OF VASCULAR SURGERY Volume 43, Number 6 Lee et al Degeorges R, Reynaud C, Becquemin JP. Thoracic outlet syndrome surgery: long-term functional results. Ann Vasc Surg 2004;18: Karwowski JK, Arko FR, Harris EJ, Olcott C. Non-operative management of primary subclavian vein thrombosis. Presented at the 76th meeting of the Pacific Coast Surgical Association, Dana Point, CA, Feb 18-21, Zhang H, Singer B. Recursive partitioning in the health sciences. New York: Springer-Verlag; Feugier P, Aleksic I, Salari R, Durand X, Chevalier JM. Long-term results of venous revascularization for Paget-Schroetter syndrome in athletes. Ann Vasc Surg 2001;15: Urschel HC, Patel AN. Paget-Schroetter syndrome therapy: failure of intravenous stents. Ann Thorac Surg 2003;75: Beaton DE, Wright JG, Katz JN. Upper Extremity Collaborative Group. Development of the QuickDASH: comparison of three itemreduction approaches. J Bone Joint Surg Am 2005;87: Submitted Sep 16, 2005; accepted Feb 4, DISCUSSION Dr Hugh Gelabert (Los Angeles, Calif). This was a brave presentation to this particular audience. I think that Stanford s reports of the treatment of Paget-Schroetter syndrome have been followed closely by most of the surgeons who like to operate on these patients. You mentioned a couple of things which caught my ear, and I wanted to ask you a few questions. You mentioned that one of the indications for surgery might be persistent vein injury on duplex scan. I m not really sure I know what you mean by that. I was hoping you might explain that a little bit more. And then a second brief question. Getting to the issue of stents, you have alluded that in two patients, their stents were crumpled or deformed, and I m guessing that you probably didn t place them at Stanford and would concede that was a wrong move at this stage. Could you explain a little bit of your rationale for bothering to do the decompression now that the stent is deformed since it s not going to be reformed after you take the rib out? What would be your explanation of the benefit of proceeding with decompression at that stage? Dr Lee. Thank you for those questions. Regarding the slide documenting the indications for operation after a short period of anticoagulation, we want to underscore the main reason to operate at the 1-month period was persistence of symptoms of venous hypertension or the development of new symptoms. Persistent vein injury as an indication for surgery was rare but, when documented by our excellent vascular laboratory technologists, allowed us to visualize subtle findings on the duplex. Such imaging findings include worsening wall thickening when compared to baseline, visible luminal reduction by grayscale or color Doppler, or documentation of continuous flow by pulse-wave Doppler rather than phasic flow. Regarding the stented patients, both of them had their stents placed at outside institutions and were seen by us after they had recurred. In those patients, they underwent another round of thrombolytic therapy, the subclavian veins were opened up again, and then rather than leave this area as a potential problem for future thrombotic recurrence, we offered them thoracic outlet decompression after their recurrent thrombosis. Venolysis in this area was difficult due to scarring, but we felt the rib would need to be removed to prevent continued fracturing of the stent. Dr Mark Meissner (Seattle, Wash). Your group has done a very nice job describing the anatomic and demographic risk factors that lead to recurrence. However, as surgeons, I think we have neglected to realize that deep venous thrombosis is usually a multifactorial problem. We focus on the anatomic problem since we can address that, while admitting there is a coagulation problem requiring anticoagulation as well. There will never be enough of these patients to do a trial as to what the appropriate duration of anticoagulation is. Three months is entirely arbitrary, and we have not focused attention on what role activated coagulation plays in these recurrences. It is becoming clear in the lower extremities that if there is evidence of ongoing activated coagulation, particularly with an elevated D-dimer several months after an event, the patient is at high risk of recurrence. I think that s what we need to start focusing on rather than simply these anatomic factors. I d like to know your opinion regarding this. Dr Lee. Thank you. We actually looked at whether hypercoagulable states were checked in our patients, and unfortunately there was not a consistent workup performed on the entire cohort due to many of the patients initially being seen at outside institutions. In the eight patients that recurred, they were all checked for standard hypercoagulable states and were negative. We do agree with your statement that there must be some other factor that is outside of the anatomy that is related to why some of these patients recur. What we need is a larger registry of these patients or more institutions who might be treating this disease conservatively to report their findings. Dr John Lane (San Francisco, Calif). Certainly our groups have a different approach to this problem, and I m not going to touch that subject. I did have just one comment on your statistics. You did choose to do a univariate and multivariate analysis with an outcomes number of 8. Most statisticians will recommend that to carefully evaluate 1 predictor you will need at least 10 outcomes, so I think that it might be unwise to do or present univariate and multivariate analysis; however, with this small number of outcomes, just to show your audience the data in a descriptive way and let them draw their own conclusions and certainly with this having 25 problems with stent placements and 0 without, I think that s completely telling. I just wanted to make a comment on that. Dr Lee. We agree with your statements. Unfortunately, we are limited by the relative lack of literature regarding large series treating this uncommon problem conservatively. This again highlights the need for more a multicenter registry to track outcomes. Dr Carlos Donayre (Torrance, Calif). Jason, that was a very nice presentation. I basically have a comment because of our interest in imaging. The duplex and the venograms, I think, still underevaluate this disease, so I think one of the things that we may consider is that in young patients or competitive athletes that at 3 months present with an open vein, intravascular ultrasound has been very good at looking really at that wall status and whether you have significant injury. It would be interesting to submit these folks to an IVUS [intravascular ultrasound] examination because it may give you some answers about predicting the long-term outcome. Dr Lee. Thank you. That is a great idea to use intravascular ultrasound, and although we have not used that in these patients yet, we would be interested in trying this in the future.