Low Incidence of Pulmonary Embolism Associated With Upper-Extremity Deep Venous Thrombosis

Background: Most recent Chest 2008 guidelines counsel at least 3 months of anticoagulation for acute upper-extremity deep venous thrombosis (UEDVT). These guidelines are inconsistently followed, perhaps owing to relatively limited information regarding clinical outcomes among patients with UEDVT. Our institution maintains an UEDVT registry of consecutively encountered patients with sonographically confirmed UEDVT. We analyzed patient characteristics, treatment, and outcomes among these patients. Methods: Between April 2005 and November 2008, 300 consecutively encountered peripheral vascular laboratory patients with UEDVTs were identified. Data on UEDVT sonographic characteristics, patient demographics, anticoagulation treatment, pulmonary embolism (PE) incidence and diagnostic modality, hemorrhagic complications, and mortality were then extracted. Results: Among the 300 patients, there was deep venous obstruction in the distal innominate (n = 69), internal jugular (n = 146), subclavian (n = 161), axillary (n = 107), and brachial (n = 91) veins. Two hundred forty-six patients (82%) had UEDVTs identified as clearly acute or acute on chronic, based on sonographic appearance. Most patients with UEDVTs were symptomatic (n = 265, 88%). One hundred six patients had documented malignancy (35%), 92 were postoperative or trauma patients (31%), and 76 patients were obese (body mass index: >30, 25%). Additionally, 240 patients had associated or previous indwelling central venous lines or leads (80%). One hundred twenty-eight patients (43%) were initially anticoagulated with heparin, whereas 121 of these patients were converted to warfarin therapy (40%) for variable lengths of time. One hundred sixty-seven patients were not treated with anticoagulation (56%), of whom 16 had documented contraindication to anticoagulation. Although the anticoagulated subset of patients tended to be younger, the decision to anticoagulate patients correlated significantly with the sonographically documented acute nature of the deep venous thrombosis, and its extent. Six patients (2%) suffered PE in association with their UEDVT diagnosis. There was no PE-related mortality. However, among anticoagulated UEDVT patients, there were four patients who suffered intracranial hemorrhage resulting in three deaths, and an additional three patients who required rehospitalization for upper gastrointestinal (n = 2) or stomal (n = 1) hemorrhage. Conclusions: Anticoagulation therapy is inconsistently used to treat UEDVT and most often used for patients with multiple vein segments involved and with sonographically acute UEDVT components. However, regardless of the decision to anticoagulate, the incidence of PE attributable to UEDVT is small (2%), regardless of treatment with anticoagulation. Given the observed mortality associated with anticoagulation in this fragile patient cohort, the risk-benefit analysis for anticoagulation does not favor routine anticoagulation among these patients.