Primary subclavian vein thrombosis and its long-term effect on quality of life

Introduction

Primary (spontaneous and effort related) deep arm vein thrombosis is defined as an acute and painful swelling of the arm resulting from thrombosis in the subclavian or axillary vein in the absence of central venous catheters or malignancy and is also known as the Paget–Schroetter syndrome. The estimated incidence of upper extremity deep vein thrombosis in the general population is approximately 2–3 per 100,000 persons per year, occurring predominantly in men (2:1). These figures include both secondary (e.g. caused by indwelling catheters) and primary thrombosis and are probably an underestimation, because upper extremity deep vein thrombosis is frequently asymptomatic. ^1–4 Several modalities such as thrombolysis – with or without first rib resection – and anticoagulants have been evaluated for the treatment of primary subclavian vein thrombosis (PSVT); however, optimal therapy remains controversial as most strategies result in varying degrees of post-thrombotic morbidity (PTS). ^1,5–12

This diversity in results may be due to use of non-standardized and poorly reproducible scales along with the lack of an unequivocal definition of PTS.

Thrombolysis, even if technically successful, does not guarantee symptom relief in the long run. This is supported by the finding that no association was found between the occurrence of PTS and ultrasonographic findings after thrombolytic therapy. ⁸ Therefore, in this study, we have focused on clinical occurrence of PTS rather than on initial roentgenological success.

In this study, we have compared the effect of either thrombolysis followed by anticoagulation, thrombolysis combined with first rib resection followed by anticoagulation, or anticoagulant therapy alone on the long-term outcome in terms of pain, swelling, fatigue, functional limitations and general quality of life (QoL) in patients with PSVT.

Methods

We reviewed all records of consecutive patients with spontaneous or effort-related axillary–subclavian vein thrombosis from 1994 to 2008. Upon admission, symptoms were present for less than two weeks in all patients. Patients had been referred to either the Internal Medicine Department or the Vascular Surgery Department, based on their general practitioner's preference. Duplex scanning was performed in all patients. In instances of high clinical suspicion and a negative or inconclusive duplex scan, selective contrast venography was performed.

All patients who were initially admitted to the Internal Medicine Department were given oral anticoagulants only, for three months (group 1). Those referred to the Vascular Surgery Department were treated with either thrombolysis and anticoagulants (group 2) or thrombolysis combined with first rib resection (when considered appropriate based on clinical and venographic findings, group 3), followed by anticoagulant therapy for three months.

Thrombolytic therapy consisted of catheter-guided thrombolysis with a 250,000 IU bolus of urokinase, followed by 100,000 IU per 24 hours. Serum fibrinogen was measured at six-hour intervals, aiming at 1.0 g/L. Heparin was given at 10,000 IU per 24 hours to prevent thrombosis along the catheter. Daily venography was performed to evaluate the thrombolytic effect. Any decrease in thrombus mass was considered a success and thrombolysis was continued until complete lysis, or until no further lysis could be reached, to a maximum of 72 hours. Patients in whom severe stenosis (>50% at venography) of the subclavian vein at the costoclavicular junction was demonstrated at completion venography, were offered a transaxillary first rib resection, usually within two to six weeks after thrombolysis. However, some patients refused surgery for a variety of personal reasons. Positional stenosis was not regarded as an indication for surgery. Coumadin was administered to all patients at discharge for a three-month period. Patients were scheduled for one outpatient clinic visit combined with a follow-up duplex scan at six weeks. Patients showing residual stenosis at duplex had additional venography, and if necessary, balloon angioplasty.

Data on early PTS (swelling, pain, fatigue, functional impairment) were retrieved retrospectively by means of chart review (outpatient clinic visit, six weeks after discharge). Symptoms at the end of follow-up (February, 2009) were collected via a questionnaire (by mail and telephone). This questionnaire consisted of four questions regarding the presence of fatigue, pain, swelling and limitation of function, which had to be answered with ‘yes’ or ‘no’. Patients were also asked to complete a EuroQol (EQ-5D) form, which is a validated generic tool for QoL assessment (summary score between −0.590 and 1.000, with the latter representing optimal QoL). This figure is calculated from the weighted factors mobility, self-care, normal daily activities, pain or discomfort and anxiety or depression, combined with a visual analogue scale which assesses self-graded health on a scale from 0 to 100, with 100 representing the best imaginable health state (EQ-VAS). ^13,14 All patients were scheduled only once for a duplex follow-up scan and a visit to the outpatient clinic at six weeks, except when individual patients had recurrent or persisting symptoms, urging renewed visits.

Statistical analysis was conducted with Statistical Package for Social Sciences version 17.0 (SPSS, Chicago, IL, USA). Baseline characteristics and outcome parameters were compared in univariate models (analysis of variance [ANOVA], chi-squared test). Prediction models were made with logistic regression analysis. The Hosmer and Lemeshow test was applied to confirm model fit. Linear regression analysis was performed for the outcome QoL (EQ-5D and EQ-VAS) at the end of follow-up. A P value of ≤0.05 was considered statistically significant.

Results

Forty-five patients were diagnosed with PSVT (38% women, 62% men). The mean age was 40 years (range 18–63). Patients presented with complaints of spontaneous onset pain, swelling, fatigue or impaired function in the afflicted arm. In 42% (n = 19), symptoms occurred in the left arm. There were no significant differences between groups at presentation (Table 1). One patient in group 2 was found to have a cervical rib. This patient, however, refused surgery. Complete clot dissolution was reached in eight (57%) patients in group 2 and in 16 (94%) in group 3. A summary of follow-up imaging is presented in Table 2.

Six of seven patients who had an unsuccessful attempt of thrombolysis had relief of symptoms and were administered coumadin for three months.

Based on clinical manifestations, recurrence was diagnosed in five patients, four who initially had successful thrombolytic therapy and one who was treated with anticoagulants. Suspicion of recurrence was always confirmed with venography. One asymptomatic recurrence was found at a periodical duplex scan. In five cases of recurrent thrombosis, lifelong anticoagual therapy was prescribed. Four of them previously had decompressive surgery.

Three patients suffered complications. One patient had a wound hematoma after first rib resection, one patient developed a hemothorax after first rib resection and another patient receiving thrombolytic therapy developed pulmonary embolism.

The response rate to the post-thrombotic symptoms and QoL questionnaires was 64%. Responder baseline characteristics were not significantly different from those of each entire initial treatment group, indicating that the responders were representative of their treatment groups. The mean follow-up was 57 months (range 2–176, SD ± 46).

Univariate analysis showed that, at six weeks after treatment, patients in group 1 had significantly more pain, swelling and fatigue in the affected limb compared with those in groups 2 and 3 (Table 3), whereas no significant differences in symptoms were present between groups 2 and 3. Logistic regression analysis confirmed the association of thrombolysis with improvement of pain at six weeks (P = 0.01, odds ratio = 0.061, 95% confidence interval = 0.007–0.517). Treatment was not significantly associated with fatigue and swelling at six weeks. At a mean follow-up of 57 months, univariate analysis showed no differences in residual symptoms between groups, leaving treatment strategy insignificant to the presence of any symptoms that might contribute to PTS, including functional impairment. Treatment strategy was not significantly associated with the presence of pain, fatigue, swelling and functional limitation at the end of follow-up (logistic regression). Recurrent thrombosis was relatively frequent in those having undergone decompressive surgery, but this difference was not significant (P = 0.10, Table 3). Comparison of patients with and without residual stenosis did not show a significant difference in late symptoms and QoL at the end of follow-up.

Linear regression analysis (ANOVA) did not reveal any variable significantly associated with QoL or self-graded health, including treatment strategy, at the end of follow-up (Tables 4a and b). This finding was confirmed by additional univariate comparison of EQ-5D and EQ-VAS results between groups at the end of follow-up (ANOVA) not showing any significant difference (P = 0.25 and 0.27, respectively) (Table 3).

Both QoL and the presence of late symptoms were similar between patients with successful and unsuccessful thrombolysis. No difference in demographics, follow-up, QoL, self-graded health, and the presence of short- and long-term residual symptoms existed between patients with successful and unsuccessful thrombolysis (Table 5).

Discussion

In this study we demonstrated a short-term relief of pain following thrombolytic therapy in patients with primary deep arm vein thrombosis. However, at the end of a mean follow-up of 57 months (range 2–176, SD ± 46), this early advantage had disappeared, leaving patients who had thrombolysis, successful or not, with residual symptoms of the same proportions as those treated with anticoagulants only. Additionally, there is no difference in QoL at the end of follow-up. Remarkably indeed, subgroup analysis revealed no significant difference in symptom relief and QoL following successful versus unsuccessful thrombolysis. The difference in thrombolysis success rate between groups 2 and 3 obviously is explained by the fact that those with unsuccessful thrombolysis were considered not eligible for decompressive surgery. Although not significant, there was a tendency towards recurrent thrombosis in those treated with first rib resection. First rib resection following thrombolysis was not associated with recurrent thrombosis, QoL or symptom relief.

The literature, mainly consisting of observational series and expert opinions, does not provide any consensus on a treatment of preference for patients with primary and effort-related thrombosis. Despite the absence of strong evidence, however, thrombolysis for initial thrombus dissolution has now gained almost universal acceptance for patients with PSVT. Also, decompression of the thoracic inlet is becoming more and more common. ^{1,9–12,15–26}

Sajid and colleagues ²⁷ published a review of 47 studies with a combined total of 2557 patients, leading them to conclude that there is no consensus on the optimal treatment for upper extremity thrombosis. It was concluded that anticoagulation is the minimum appropriate intervention in PSVT management, preventing clot propagation and pulmonary embolism. ^27–29

Catheter-guided thrombolysis has been shown to be a safe and effective method to achieve recanalization in several studies; however, dose and duration of thrombolytic therapy utilized are not uniform. ^{12,16,17,19,25,27,30} Sajid found that thrombolysis did not influence clinical outcome, recurrence rates and symptom improvement in patients with longstanding thrombosis. ²⁷

From their 2005 review of 41 studies with a total of 559 patients, Thomas et al. proposed thrombolysis as the first step in a staged therapy, combined with surgical thoracic inlet decompression in selected patients. However, they also found that it remains unclear whether the indication for surgical decompression should be based on the patient's symptoms or on venographic findings after thrombolysis, as several authors have noted a discrepancy between imaging and symptoms. In accordance with our study, they also found a higher recurrence rate in patients who had undergone additional decompressive surgery compared with those without decompression. ^26,31 Also, the timing of decompressive surgery remains a subject of debate. ^{16,22–24,32} Although commonly performed, the literature shows that surgical decompression in fact is still controversial and evidence is lacking to support its routine practice. ^18,31,32

Percutaneous angioplasty (PTA) of the axillary and subclavian veins is frequently practiced. PTA should preferably be performed after decompression, whereas PTA prior to decompression is less effective. ^16,24 PTA after decompressive surgery may decrease recurrence rates by reducing venous obstruction resulting from caliber loss due to vessel wall damage. As a single additional treatment after thrombolysis, PTA and stenting have shown rather disappointing results. ^22,24,30,33

The pivotal concern in the treatment of PSVT is persistence of residual symptoms and occurrence of the ill-defined post-thrombotic syndrome. In young, active patients, PTS constitutes a considerable handicap. This is the reason for the increasingly aggressive approach of PSVT, despite the lack of evidence. The results of our study, along with many previously published reports, show that prevention of PTS remains problematic in the long term. In their 2002 study of long-term outcomes in 95 patients with upper extremity thrombosis, Sabeti et al. ⁹ found no association between the use of thrombolysis and the incidence of PTS. This finding was confirmed by Vik et al. ⁵ in their 2009 report of 30 patients. Residual thrombosis may contribute to PTS. Prandoni et al. prospectively reviewed 53 patients and found that residual thrombosis was associated with an increased incidence of PTS (hazard ratio 4.0). In contrast with our findings, they also found that late symptoms or PTS negatively influence QoL. ¹⁰ Kahn et al. ³⁴ concluded accordingly.

The limitations of this study are inherent to its retrospective, non-randomized design. Selection bias may be present due to the fact that patients were treated in different departments. Also, treatment allocation was the result of the general practitioners' (more or less random) referral pattern, which is very likely to be based merely on habit rather than on actual and up-to-date insight in to the disease itself and the available treatment options. Although patients in the three treatment groups compared well with regard to their baseline characteristics, chronicity and symptoms, the decision regarding treatment strategy was at the discretion of the physician in charge. The difference in questionnaire response rates between groups is a potential source of bias, although no differences between responders and their initial treatment groups could be identified. In addition, the QoL test used is not disease specific but generic, and our study lacks a QoL assessment prior to therapy. Therefore, our present findings represent no basis for final recommendations on optimal therapy, but may initiate further discussion on the indications for different therapies in different types of patients (regarding age and daily activities), with respect to long-term outcomes and QoL. In spite of lacking evidence for long time benefit, we believe that in patients with otherwise untreatable pain and swelling, a more aggressive initial approach with thrombolysis may be indicated, if only for short-term symptom relief. Subsequently, in those with a severe residual stenosis at venography and continuing symptoms, decompressive surgery and PTA may be considered.

Conclusions

In this small study, thrombolysis led to short-term symptom relief but not to lasting symptom reduction and improvement of QoL in patients with acute subclavian vein thrombosis. Routine addition of transaxillary first rib resection for severe residual subclavian vein stenosis was not associated with improved late outcome (symptoms, QoL) and did not reduce recurrence rate.