Upper extremity deep venous thrombosis: reassessing the risk for subsequent pulmonary embolism.

BACKGROUND: Although the incidence of upper extremity deep venous thrombosis (UEDVT) diagnoses has increased, anticoagulation therapy for UEDVT remains inconsistent and of variable duration. This study sought to analyze our institution's current treatment practices for UEDVT and assess the risk for subsequent pulmonary embolism (PE).

METHODS: Between April 2005 and July 2007, 200 consecutively encountered patients with UEDVTs were identified in the Peripheral Vascular Laboratory. Then, UEDVT location and sonographic characteristics, patient demographics, anticoagulation treatment, and PE incidence and mortality were examined.

RESULTS: Among the 200 patients with UEDVT, 156 (78%) had UEDVTs identified as clearly acute or acute on chronic, based on sonographic appearance. In all, 85% of the patients were symptomatic (n = 171). Among the patients, 71 (36%) had documented malignancy, 58 (29%) were postoperative or suffering from trauma, and 52 (26%) were obese (body mass index: >30). In addition, 153 (76%) had associated current or previous indwelling lines or leads. A total of 73 patients (36%) were put on anticoagulation therapy for variable periods. Younger age of the patient, duplex evidence of an acute deep venous thrombosis, and involvement of multiple named upper extremity venous segments were independent predictors of the decision to initiate anticoagulation therapy for patients with UEDVT. Two patients (1%) suffered PE, most likely the consequence of their UEDVTs. An additional two patients with UEDVT treated with coumadin died months after hospital discharge from intracranial bleedings after minor falls.

CONCLUSION: Currently, the clinical decision to initiate anticoagulation therapy for patients with UEDVT is most associated with the documented acute nature of the deep venous thrombosis by duplex ultrasonography, involvement of multiple venous segments, and younger age of the patient. The incidence of PE attributable to previously documented UEDVT is very small (1%), regardless of anticoagulant therapy. Anticoagulation therapy for UEDVT is most likely to be best suited to address the symptoms of UEDVT; its proposed use to decrease the very small risk of PE may be rarely indicated, and must be carefully weighed with the risks associated with therapeutic anticoagulation in this patient population.