Residual Venous Thrombosis as a Predictive Factor of Recurrent Venous Thromboembolism

Context Doctors typically treat deep venous thrombosis with anticoagulants for 3 to 6 months to prevent recurrence. Patients at high risk for recurrence may benefit from longer treatment. Can we identify them with ultrasonography that detects persistent thrombosis? Contribution Three hundred thirteen patients with deep venous thrombosis had ultrasonography every 6 to 12 months for 3 years. Recurrent thromboembolism, assessed over 6 years, was more frequent among those showing persistent residual thrombosis rather than early vein recanalization. Implications We now need trials that evaluate prolonged anticoagulation in patients with and without residual thrombosis to see whether tailoring treatment on the basis of serial ultrasonography is beneficial. The Editors Patients with deep venous thrombosis (DVT) of the lower extremities are usually treated with an initial course of unfractionated or low-molecular-weight heparin followed by at least 3 months of oral anticoagulant therapy. This treatment regimen reduces the risk for short-term thromboembolic complications to approximately 5% (1, 2). Although the short-term outcome of this disease has been extensively documented, only a few studies have addressed the risk for late recurrent venous thromboembolism. These studies suggest that this risk persists for years and is related to patient characteristics at initial presentation (1, 2). It has been documented that patients with continuous risk factors, such as cancer or the antiphospholipid antibody syndrome, and those with idiopathic thrombosis have a two- to threefold increased risk for recurrence compared with patients who developed a thrombotic event in association with a transient risk factor (3-5). As a consequence, longer anticoagulant treatment (up to several years) has been evaluated in patients with idiopathic DVT, with the aim of decreasing recurrences in patients with idiopathic DVT (6, 7). Although long-term anticoagulant therapy is effective in preventing recurrences, it is inconvenient and carries a risk for bleeding (8-11). It would be beneficial to further improve our ability to identify patients who are at higher risk for recurrent events. The absolute incidence of recurrent venous thromboembolism decreases over time (3, 5). Residual thrombus mass also decreases over time in patients with proximal venous thrombosis, according to recent studies that used repeated ultrasonographic imaging (12, 13). If vein recanalization indicates a lower risk for recurrent disease, ultrasonographic imaging might help clinicians adjust the duration of oral anticoagulant therapy on an individual basis in patients who have had an episode of DVT. To estimate whether the risk for recurrent venous thromboembolism is higher in the presence of residual thrombosis, we performed repeated ultrasonography in a large number of patients with proximal venous thrombosis who were followed prospectively for up to 6 years. Methods Study Design We performed a prospective cohort follow-up study to assess the potential effect of residual venous thrombosis on the risk for recurrent venous thromboembolism in patients with a first episode of symptomatic DVT. The institutional review board of Padua University, Padua, Italy, approved the study. Inception Cohort All consecutive outpatients who were referred to the Department of Medical and Surgical Sciences of the University of Padua between 1993 and 1996 for clinical suspicion of a first episode of DVT and who met inclusion criteria were eligible. Inclusion criteria were proximal venous thrombosis on compression ultrasonography, absence of diseases requiring indefinite anticoagulation (such as atrial fibrillation, active cancer, chronic medical illnesses, or other permanent risk factors for venous thrombosis), life expectancy of more than 6 months, and ability to return to the study center for follow-up visits. Eligible patients received conventional anticoagulation (full doses of unfractionated or low-molecular-weight heparin followed by 3 months of oral anticoagulant therapy). Patients who completed the initial 3 months of treatment without a recurrent thrombotic episode and who gave informed consent were recruited for the current study. On the basis of laboratory results and clinical characteristics, patients were divided into three categories: those who had thrombophilia, those who had secondary DVT, and those who had idiopathic DVT. Tests for thrombophilia (antithrombin, protein C or S defect, factor V Leiden mutation, prothrombin G20210 gene mutation, and lupus-like anticoagulants) were performed before anticoagulation was started or at least 2 weeks after its conclusion. On the basis of test results, patients were categorized as being with or without thrombophilia. Patients without thrombophilia were further classified as having idiopathic DVT or DVT secondary to transient risk factors, according to a standardized form for clinical data collection completed at referral. Secondary thrombosis was defined as that occurring during pregnancy or childbirth; during estrogen use for contraception or hormone replacement therapy (ongoing or interrupted for <1 month); or after recent trauma, fracture, or surgery (within <3 months). If thrombophilia screening could not be performed, patients were classified as having idiopathic or secondary DVT on the basis of clinical presentation. Ultrasonographic Assessments The first ultrasonographic assessment was performed 3 months after the initial event. Patients found to have residual thrombosis were scheduled for repeated assessment 6, 12, 24, and 36 months after the initial event. Independent experts who were unaware of patients' clinical details or of previous ultrasonographic findings performed assessments according to a standardized procedure. Only the common femoral vein at the saphenofemoral junction and the popliteal vein in the midpopliteal fossa were scanned for residual venous thrombosis. Vein compression was performed in the transverse plane; vein diameter was measured during maximal compression and was expressed in millimeters. Veins were considered recanalized if they were 2.0 mm or less in diameter on a single test or 3.0 mm or less in diameter on two consecutive tests. This definition was based on ultrasonographic findings from a separate group of 145 patients with proven proximal DVT who were followed prospectively at our institution in the early 1990s; fewer than 2.0% of these patients developed recurrent thromboembolism in the 2 years after vein recanalization (12). Follow-up and Recurrent Venous Thromboembolism All recruited patients were instructed to interrupt oral anticoagulant therapy when they were included in the current study. Patients were followed to document the incidence of symptomatic recurrent DVT or pulmonary embolism. Patients were educated about the main signs and symptoms of recurrent venous thromboembolism and received a card with the telephone numbers of the thrombosis clinic. They were instructed to return to the study center if they noted clinical manifestations suggestive of recurrent venous thrombosis (edema, redness, tenderness, pain, or swelling) in either leg or suggestive of pulmonary embolism (dyspnea, chest pain, or tachycardia). Patients were also seen at the time of ultrasonographic assessments and were contacted at least twice yearly to ascertain whether signs and symptoms had occurred. If they had, patients were invited to come to the study center for additional diagnostic procedures. Recurrent DVT was diagnosed by compression ultrasonography, followed by ascending phlebography in case of indeterminate findings, or a strong discrepancy between clinical suspicion and negative results on ultrasonography (14). Patients with suspected pulmonary embolism had ventilationperfusion lung scanning, which was followed by pulmonary angiography if findings were inconclusive (15). Fatal pulmonary embolism was diagnosed on the basis of autopsy findings or the opinion of an independent physician. Statistical Analysis The first 3 months of initial anticoagulant treatment were not included in any of our analyses. For the remaining months, KaplanMeier estimates and 95% CIs were calculated to assess the risk for recurrent venous thromboembolism. Time-dependent multivariate Cox proportional-hazards models were then used to calculate hazard ratios for recurrent venous thromboembolism in patients with residual thrombosis versus those without. For this analysis, patients who did not have a recurrence were censored at the end of the available follow-up or at death. In addition, the clinical categorization of patients was evaluated by using two dummy variables to indicate the presence of idiopathic DVT or thrombophilia. Secondary thrombosis was therefore used as the reference group, and in each case, the duration of oral anticoagulant treatment was used as a time-dependent covariate. Finally, all of these variables, as well as age and sex, were introduced in a stepwise hierarchical Cox model (16, 17). The results of these analyses were expressed as risk ratios and 95% CIs. In addition, incidence density of recurrent venous thromboembolism (events per observation-year) was calculated for the various combinations of clinical characteristics. All calculations were performed by using SAS software, version 6.10 (SAS Institute, Inc., Cary, North Carolina). Results Patients Of 1250 patients who were referred for clinical suspicion of first DVT, 389 (31.1%) received a diagnosis of proximal venous thrombosis on compression ultrasonography (Figure 1). Forty-nine of these 389 patients were excluded because of poor life expectancy (n = 29), diseases requiring permanent anticoagulation (n = 14), or inability to attend follow-up visits (n = 6). Of the remaining 340 eligible patients, 25 developed a recurrent thromboembolic event during initial anticoagulation. Therefore, 315 patients completed an uneventful 3-month period of anticoagulation.