The Long-Term Clinical Course of Acute Deep Venous Thrombosis

Methods

Identification of Inception Cohort

The Department of Internal Medicine of the University of Padua, Padua, Italy, is a diagnostic facility for outpatients with clinically suspected venous thromboembolism in a community of approximately 350 000 persons. All consecutive outpatients with a first episode of clinically suspected deep venous thrombosis who were referred by their general practitioners between January 1986 and December 1991 had noninvasive testing [8]. Patients were potentially eligible for the study if confirmatory venography showed deep venous thrombosis. Patients were excluded from the study if they had been referred because of recurrent venous thrombosis, were geographically inaccessible for follow-up, or refused to give informed consent. The Institutional Review Board of the hospital of the University of Padua approved the study.

Baseline Assessment

At the time of referral, demographic characteristics were recorded and a medical history was taken; information was elicited on the period between the onset of symptoms and presentation to the thrombosis service (patient–physician delay), the presence of risk factors for thrombosis (that is, cancer, surgery, trauma or fracture, immobilization for more than 7 days, pregnancy or childbirth, or estrogen use), and symptoms of pulmonary embolism. Information was also obtained on the history of venous thromboembolism in first-degree relatives. Antithrombin, protein C and S, and lupus-like anticoagulant levels were subsequently measured. Assays were done, and previously described criteria for abnormality and deficiency were used [9].

The venograms obtained at baseline were divided into those representing proximal venous thrombosis (with or without concurrent venous thrombosis of the calf) and those indicating isolated venous thrombosis of the calf. Proximal venous thrombosis was defined as thrombosis located above the trifurcation of the calf veins that involved at least the popliteal vein, superficial femoral vein, common femoral vein, or iliac vein. The location and occlusiveness of proximal thrombi were also determined. A patient was considered to have nonocclusive deep venous thrombosis if contrast material was seen between the thrombus and the vessel wall along the entire thrombus.

Treatment

Patients were admitted to the hospital and treated with an initial course of high-dose intravenous standard heparin (a bolus of 5000 U followed by continuous infusion of 30 000 U/d, subsequently adjusted to maintain an activated partial thromboplastin time between 1.5 and 2.5 times the normal value) or subcutaneous low-molecular-weight heparin (90 U of anti-factor Xa/kg of body weight twice daily). Therapy with oral anticoagulant agents (warfarin) was started on day 5 to 7 of treatment and was continued for 3 months. The oral anticoagulant dose was adjusted daily to maintain an international normalized ratio between 2.0 and 3.0. Treatment with low-molecular-weight heparin was discontinued on day 10 or later if the international normalized ratio was less than 2.0. This treatment strategy deviated in the following groups of patients: those with cancer, protein deficiencies, or lupus anticoagulant, in whom oral anticoagulation therapy was prolonged; those with small isolated venous thrombosis of the calf, who received oral anticoagulation alone; those with contraindications to anticoagulant treatment, who received no treatment or an inferior caval-vein filter; those who refused to be hospitalized, who received low-dose heparin and oral anticoagulant agents; and those with threatened viability of the leg, who received thrombolytic therapy. The actual type and duration of treatments were recorded. All patients were instructed to wear elastic graduated compression stockings (providing 40 mm Hg of pressure at the ankle) for at least 2 years.

Follow-up

All patients were seen 3 and 6 months after the initial referral and thereafter returned to the study center every 6 months for follow-up assessments. Patients were asked to return to the thrombosis center immediately if symptoms suggestive of recurrent venous thromboembolism developed. Follow-up was continued for as long as 8 years or until July 1995. To avoid diagnostic suspicion bias, the medical history on general health, symptoms of recurrent venous thromboembolism, and the post-thrombotic syndrome was obtained by using a standardized form. Patients who could not attend the follow-up sessions were visited at home. For all patients who died during follow-up, the date and cause of death were documented.

Diagnosis of Recurrent Venous Thromboembolism and Hemorrhage

Contrast venography of the symptomatic leg or legs was done as described previously [10]. The criteria for deep venous thrombosis were a constant intraluminal filling defect confirmed in at least two different projections or nonvisualization of a vein or a segment thereof, despite adequate technique and repeated injections with contrast material. The presence or absence of venous thrombosis was assessed by a panel of independent observers who were unaware of the patient's other clinical features or previous test results. If a patient presented with clinically suspected recurrent venous thrombosis of the leg, venography was done. The criterion for recurrent venous thrombosis of the leg was a new intraluminal filling defect on the venogram. If the venogram was not diagnostic, recurrent venous thrombosis was diagnosed on the basis of an abnormal ¹²⁵I-fibrinogen leg scan or results of noninvasive tests that had changed from normal to abnormal [11, 12]. Patients with suspected pulmonary embolism had venography if they had concurrent leg symptoms or perfusion lung scanning in the absence of leg symptoms. Pulmonary embolism was excluded if the perfusion scan was normal. Because ventilation lung scanning was not available during the first years of the study and because pulmonary angiography could not routinely be done, we could not definitively diagnose pulmonary embolism in some patients. If a definitive diagnosis could not be made, patients were classified as not having recurrent venous thromboembolism. Perfusion lung scanning and pulmonary angiography were done and their results were interpreted according to standard procedures [13]. Hemorrhagic episodes were classified as major or minor, as reported previously [14]. The documentation of all patients suspected of having a recurrent venous thromboembolic or bleeding event was reviewed by a three-member adjudication committee that was unaware of further clinical details of the patient.

Criteria for the Post-Thrombotic Syndrome

Presence of the post-thrombotic syndrome was assessed by investigators who were unaware of previous post-thrombotic manifestations and further clinical details of the patient. The presence of leg symptoms (pain, cramps, heaviness, pruritus, and paresthesia) and signs (pretibial edema, induration of the skin, hyperpigmentation, new venous ectasia, redness, and pain during calf compression) was scored. For each item, the investigators assigned a score of 0 (not present or minimal) to 3 (severe). The presence of a venous ulcer of the lower limb was recorded. In patients with bilateral thrombosis, the higher score was used. A total score of 15 or more on two consecutive visits or the presence of a venous ulcer indicated severe post-thrombotic syndrome, and a total score of 5 to 14 on two consecutive visits indicated mild post-thrombotic syndrome. This score has been shown to have good reproducibility, and it correlates well with the patient's perception of the interference of leg symptoms with daily life [15].

Statistical Analysis

We calculated Kaplan-Meier estimates and 95% CIs for a visual assessment of survival and calculated the risk for recurrent venous thromboembolism and mild and severe post-thrombotic syndrome. Using the stepwise Cox proportional-hazards model, we calculated the hazard ratios for death, recurrent venous thromboembolism, and mild and severe post-thrombotic syndrome associated with various clinical features [16]. In each case, we used the duration of oral anticoagulant treatment as a time-dependent variable for recurrent venous thromboembolism and death. The occurrence of a recurrent event in the same leg was used as a time-dependent variable for mild and severe post-thrombotic syndrome. We expressed the results of these analyses as hazard ratios and 95% CIs. Because death is a competing risk for recurrent venous thromboembolism and the post-thrombotic syndrome, our estimated hazard ratios for these outcomes refer to the hazards that would exist if death could be removed as a competing risk. Our analysis also assumes that death is statistically independent of these outcomes.

Results

Three hundred fifty-five consecutive patients with a first episode of venography-confirmed deep venous thrombosis were included in our study. The demographic and clinical characteristics and prevalence of potential risk factors are presented in Table 1. Two hundred forty-five patients (69%) completed 5 years of follow-up, and 148 patients (42%) completed 8 years of follow-up. Thirteen patients (3.7%) were lost to follow-up after 2 to 7 years.

Recurrent Venous Thromboembolism

Of the 355 patients, 78 had one or more documented recurrent venous thromboembolic events. Thirty-five first recurrences (44.9%) occurred in the leg that had been initially involved, 28 (35.9%) occurred in the contralateral leg, and 15 (19.2%) were pulmonary emboli. The pulmonary emboli were fatal in 9 patients (11.5%); in 7 of these 9 patients, the diagnosis was based on autopsy findings. The cumulative incidence of recurrent venous thromboembolism was 4.9% after 3 months (95% CI, 2.6% to 7.1%) and 8.6% after 6 months (CI, 5.6% to 11.6%). This incidence gradually increased to 17.5% (CI, 13.6% to 22.2%) after 2 years of follow-up, 24.6% (CI, 19.6% to 29.7%) after 5 years of follow-up, and 30.3% (CI, 23.6% to 37.0%) after 8 years of follow-up (Figure 1). Among the evaluated potential risk factors and clinical characteristics, the presence of cancer and of impaired coagulation inhibition increased the risk for recurrent venous thromboembolism (hazard ratios, 1.72 [CI, 1.31 to 2.25] and 1.44 [CI, 1.02 to 2.01], respectively). In contrast, surgery and recent trauma or fracture were associated with a decreased risk for recurrent venous thromboembolism (hazard ratios, 0.36 [CI, 0.21 to 0.62] and 0.51 [CI, 0.32 to 0.87], respectively).

The Post-Thrombotic Syndrome

Eighty-four patients developed the post-thrombotic syndrome. Of these, 25 (30.2%) had severe post-thrombotic manifestations. The cumulative incidence of the post-thrombotic syndrome was 17.3% (CI, 14.6% to 23.5%) after 1 year of follow-up and 22.8% (CI, 18.0% to 27.5%) after 2 years of follow-up. The cumulative incidence increased gradually to 28.0% (CI, 22.7% to 33.3%) after 5 years but did not substantially change thereafter (29.1% at 8 years; CI, 23.4% to 34.7%) (Figure 2). When only severe post-thrombotic manifestations are considered, a different pattern is seen in the first 5 years of follow-up, because the cumulative incidence increased gradually from 2.6% (CI, 0.8% to 4.4%) after 1 year to 9.3% (CI, 5.6% to 12.8%) after 5 years. Thereafter, the cumulative incidence of severe post-thrombotic manifestations did not increase (Figure 2).

The development of ipsilateral recurrent deep venous thrombosis was associated with an increased risk for the post-thrombotic syndrome (hazard ratio, 6.4; CI, 3.1 to 13.3). No significant associations were seen between the occurrence of the post-thrombotic syndrome and the presence of thrombi in the popliteal vein (hazard ratio, 1.2), occlusive thrombi (hazard ratio, 0.8), or the extent of thrombosis (hazard ratio, 1.1). When ipsilateral recurrent deep venous thrombosis was included in the analysis, none of the other clinical features showed a significant association with the risk for the post-thrombotic syndrome.

Other Clinical Events

Forty-seven patients developed a hemorrhagic complication. The complication was major in 17 of these patients and fatal in 2 of the 17. Nineteen of the hemorrhagic episodes (6 of which were major) occurred during the initial 14-day treatment period, for a cumulative incidence of 5.1% (CI, 2.8% to 7.4%). Another 16 hemorrhages (5 of which were major) occurred between days 14 and 90, for a cumulative incidence at 3 months of 9.8% (CI, 6.7% to 13.0%). Of the 47 patients who continued to receive treatment with oral anticoagulants after 3 months, 12 developed a hemorrhage during long-term follow-up (6 of these 12 had a major episode).

Malignant disease became apparent during follow-up in 26 of the 297 patients who had no cancer at baseline. Cancer developed primarily in patients with unexplained deep venous thrombosis at baseline, as reported previously [17]. Sixteen patients had an ischemic stroke during follow-up; the stroke was fatal in 10 of these patients.

Mortality

Ninety patients died during follow-up. Fifty-four of the 84 patients in whom cancer was diagnosed at baseline or during follow-up died. Thirty-four of the 58 patients with cancer at baseline died during the first year after the thrombotic episode. The causes of death included cancer (n = 52), ischemic stroke (n = 10), acute myocardial infarction (n = 2), pulmonary embolism (n = 9), heart failure (n = 3), anticoagulant-related hemorrhage (n = 2), and miscellaneous conditions (n = 6). In 6 patients who died suddenly, a definite diagnosis could not be made.

Survival was 83.3% (CI, 79.4% to 87.3%) after 1 year of follow-up, 80.1% (CI, 75.9% to 84.3%) after 2 years, 74.6% after 5 years, and 70.2% after 8 years (Figure 3). The presence of cancer increased the risk for death (hazard ratio, 8.1; CI, 3.6 to 18.1). No other clinical features were associated with death.

Discussion

The short-term outcome of patients with acute deep venous thrombosis of the lower extremity has been extensively documented, but little is known about the long-term clinical course of the disease. We assessed the incidence of complications after a first episode of documented deep venous thrombosis in a large cohort of consecutive patients during long-term follow-up. We found a surprisingly high risk for recurrent venous thromboembolic disease that persisted after anticoagulant treatment; this risk resulted in a cumulative 30% incidence of these complications after 8 years of follow-up. The cumulative incidence after 2 years of follow-up was consistent with the results of the DURAC (Duration of Anticoagulation) study [6]. Patients with underlying cancer or defects that impaired coagulation inhibition had a statistically significantly higher risk for recurrences than patients without these features. Because in patients with cancer the risk for death competes with the risk for recurrent venous thromboembolism, the observed relative risk may be an underestimation. As expected, many patients developed deep venous thrombosis after surgery or trauma. The finding that these patients had a significantly lower risk for recurrent venous thromboembolism indicates that these conditions are transient risk factors for deep venous thrombosis. One of every five recurrent episodes was pulmonary embolism (which was fatal in half of them), and 40% of the recurrent venous thromboses of the leg occurred in the previously asymptomatic leg.

The post-thrombotic syndrome occurred in approximately 30% of patients. However, the cumulative incidence of severe post-thrombotic manifestations after 8 years of follow-up was less than 10%. This is in contrast with the results of small studies, in which post-thrombotic sequelae were seen in as many as 90% of patients [18-24]. However, the systematic use of graduated compression stockings could have contributed to this relatively low incidence, as shown by a recent controlled study [25]. In many patients (65%), manifestations of the post-thrombotic syndrome became apparent within the first 2 years after the acute thrombosis. These findings, which challenge the general view that the symptoms of the post-thrombotic syndrome require a long time to become manifest [24], suggest that the duration of follow-up in our patients might be adequate to give a valid estimate of the overall incidence of the syndrome. Although we had expected that the extent of the initial deep venous thrombosis and the degree of occlusiveness would be related to the risk for developing the post-thrombotic syndrome [3, 24, 26], we could not show such a relation. However, patients with recurrent ipsilateral deep venous thrombosis had a significantly increased risk for developing the post-thrombotic syndrome.

Mortality was high (30% after 8 years), and most deaths occurred during the first year in patients with underlying cancer. These data are fully consistent with the results of a similar study among patients with pulmonary embolism [27].

Because our center is a diagnostic facility for all general practitioners in the area, we believe that our observations reflect the true clinical course of symptomatic deep venous thrombosis in outpatients. Venography, the reference standard, was used to diagnose deep venous thrombosis. The diagnosis was confirmed in approximately 35% of patients referred for suspected deep venous thrombosis. The demographic and clinical characteristics of our patients are similar to those in other large series of patients with symptomatic deep venous thrombosis. Although treatment regimens slightly varied, patients were treated according to standard practice. Furthermore, patients were followed prospectively, and few patients were lost to follow-up. Finally, predefined criteria were strictly applied to the diagnosis of recurrent venous thromboembolism, and a validated scale was used to assess the post-thrombotic syndrome.

What do our findings imply for the management of patients with deep venous thrombosis? The high incidence of recurrent venous thromboembolism after cessation of anticoagulant therapy suggests that prolongation of this treatment could be considered in selected patients, depending on the presence of risk factors for recurrent venous thromboembolism. However, the recommendation to use prolonged anticoagulation in these patients can only be based on the results of a large trial addressing the reduction of the number of cases of venous thromboembolism relative to the increased risk for warfarin-related bleeding. Because recurrent venous thrombosis strongly predicted the development of the post-thrombotic syndrome, the prevention of recurrent deep venous thrombosis might be a key to decreasing the incidence of this condition.

We conclude that deep venous thrombosis carries a high risk for recurrent venous thromboembolism that persists for many years, especially in patients without transient risk factors for deep venous thrombosis. The post-thrombotic syndrome occurs in almost one third of patients with deep venous thrombosis and is strongly related to ipsilateral recurrent deep venous thrombosis. Our findings challenge the widely adopted use of short-course anticoagulant therapy in patients with symptomatic deep venous thrombosis.

Dr. Lensing: Center for Hemostasis, Thrombosis, Atherosclerosis and Inflammation Research, F-4, Academic Medical Center, Meibergdreef 9, Amsterdam 1105 AZ, the Netherlands.

Dr. Prins: Department of Clinical Epidemiology and Biostatistics, Academic Medical Center, B-2 Meibergdreef 9, Amsterdam 1105 A2, the Netherlands.