The natural history of upper extremity deep venous thromboses in critically ill surgical and trauma patients: What is the role of anticoagulation?

Background: The natural history and optimal treatment of upper extremity (UE) deep venous thromboses (DVT's) remains uncertain as does the clinical significance of catheter-associated (CA) UE DVT's. We sought to analyze predictors of UE DVT resolution and hypothesized that anticoagulation will be associated with quicker UE DVT clot resolution and that CA UE DVT's whose catheters are removed will resolve more often than non-CA UE DVT's. Methods: All patients on the surgical intensive care unit service were prospectively followed from January 2008 to May 2010. A standardized DVT prevention protocol was used and screening bilateral UE and lower extremity duplex examinations were obtained within 48 hours of admission and then weekly. Computed tomography angiography for pulmonary embolism was obtained if clinically indicated. Patients with UE DVT were treated according to attending discretion. Data regarding patient demographics and UE DVT characteristics were recorded: DVT location, catheter association, occlusive status, treatment, and resolution. The primary outcome measure was UE DVT resolution before hospital discharge. Interval decrease in size on the subsequent duplex after UE DVT detection was also noted. UE DVTs without a follow-up duplex were excluded from the final analysis. Univariate and multivariate analyses were used to identify independent predictors of UE DVT resolution. Results: There were 201 UE DVT's in 129 patients; 123 DVTs had a follow-up duplex and were included. Fifty-four percent of UEDVTs improved on the next duplex, 60% resolved before discharge, and 2% embolized. The internal jugular was the most common site (52%) and 72% were nonocclusive. Sixty-four percent were CAUEDVT's and line removal was associated with more frequent improvement on the next duplex (55% vs. 17%, p = 0.047, mid-P exact). Sixty-eight percent of UEDVTs were treated with some form of anticoagulation, but this was not associated with improved UE DVT resolution (61% vs. 60%). Independent predictors of clot resolution were location in the arm (odds ratio = 4.1 compared with the internal jugular, p = 0.031) and time from clot detection until final duplex (odds ratio =1.052 per day, p = 0.032). Conclusion: A majority of UE DVT's are CA, more than half resolve before discharge, and 2% embolize. Anticoagulation does not appear to affect outcomes, but line removal does result in a quicker decrease in clot size.