The natural history and outcomes of line-associated upper extremity deep venous thromboses in critically ill patients

Abstract

Objective: Anticoagulation remains the standard of care for line-associated upper extremity deep venous thrombosis (UEDVT). This treatment carries the risk of hemorrhagic complications, possibly more so in surgical patients. Considering the low-risk profile of UEDVT-which is associated with fewer, less severe pulmonary emboli than lower extremity deep venous thrombosis-current UEDVT treatment guidelines may be overly aggressive. The goal of this study was to review outcomes of line-associated UEDVT in critically ill patients and to define the efficacy of current treatment protocols in pulmonary embolism (PE) prevention while avoiding hemorrhagic complications.

Methods: A retrospective review was performed of 193 consecutive patients admitted to the medical and surgical intensive care unit (ICU) at a tertiary care hospital between 2009 and 2014 diagnosed with acute line-associated UEDVT by duplex ultrasound. The examined treatment arms included anticoagulation with intent to reach therapeutic levels, prophylactic or subtherapeutic anticoagulation, and no anticoagulation. Primary outcomes included major hemorrhage (defined as any intracranial hemorrhage or any hemorrhage resulting in transfusion, ICU readmission, or death), PE, in-hospital mortality, total hospital length of stay (LOS), and ICU LOS.

Results: Of the 10,907 patients, 161 (1.48%) were diagnosed with acute line-associated UEDVT, 81 of 6027 in the medical ICU (1.34%) and 80 of 4880 in the surgical ICU (1.64%), after exclusion of 32 patients with concurrent lower extremity deep venous thrombosis. In total, 122 patients (75.8%) received anticoagulation with intent to reach therapeutic levels, 23 (14.3%) received prophylactic anticoagulation, and 16 (9.94%) received no anticoagulation. Major hemorrhage was significantly more common than symptomatic PE in all patients (15.5% vs 4.97%; P < .001). A single fatal PE and two fatal hemorrhages were recorded across the series. There was no significant difference in in-hospital mortality (34.8% vs 16.7%; P = .726) or ICU LOS (33.1 vs 18.3 days; P = .739) for patients who developed major hemorrhage or symptomatic PE. On multivariate analysis, incidence of symptomatic PE was not significantly related to Acute Physiology and Chronic Health Evaluation III score (P = .963), anticoagulation regimen (P = .940), catheter type (P = .313), or bacteremia (P = .833).

Conclusions: Major hemorrhagic complications are more common than symptomatic PE after anticoagulation for line-associated UEDVT in ICU patients, raising concern that current treatment guidelines are too aggressive. These data necessitate further prospective investigation to determine the optimal treatment protocol for line-associated UEDVT in this critically ill cohort.