The Spine Blog

Tuesday, October 15, 2013

Risk Factors for Venous Thrombosis Following Spine Surgery

Venous thrombolic events (VTEs) including deep venous thrombosis (DVT) and pulmonary embolism (PE) following spine surgery are poorly understood. While great progress has been made in preventing VTEs following joint replacement and extremity fracture surgery, the risks, benefits, and timing of chemoprophylaxis following spine surgery have not been well-defined. Post-operative hematoma in extremity surgery can lead to wound complications, however, epidural hematoma following spine surgery can have devastating and permanent consequences. Given the lack of clarity on the topic of VTE and its prevention in spine surgery, Dr. Schoenfeld and his colleagues from Texas queried the National Surgical Quality Improvement Program (NSQIP) database to evaluate VTE following spine surgery from 2005-2011 and included over 27,000 adult patients undergoing all types of spine surgery other than that for acute trauma. Their main objective was to determine patient and surgical characteristics that were risk factors for VTE in order to define populations who might benefit most from more aggressive prophylaxis. They found a 1% rate of VTE within 30 days of surgery, including a 0.7% rate of DVT and 0.4% rate of PE. Seventeen percent of DVT patients also developed a PE. Univariate analyses identified BMI over 40, age greater than 80, ASA score of 3 or higher, diabetes, OR time over 261 minutes, and posterior cervical or lumbar fusion were risk factors for DVT. Male sex was also associated with PE. In order to identify independent risk factors, multivariate logistic regression was performed that identified BMI over 40, age over 80, ASA score of 3 or higher, and OR time over 261 minutes as independent risk factors for DVT. Only BMI over 40, OR time over 261 minutes, and male sex remained significant independent predictors of PE. Other than prolonged OR time, surgical factors tended not to be identified as risk factors for VTE, though longer duration of surgery is a marker for more complex procedures including long fusions and anterior-posterior surgery. The most pronounced risk factor was BMI over 40, which increased the risk of DVT three-fold compared to those with BMI less than 25.


The authors should be applauded for performing what is one of the largest studies on this topic. Given that VTE is relatively uncommon following spine surgery, large database studies are likely the only viable study design to evaluate the topic. This paper does report a relatively low rate of VTE, which may be accurate given that procedures with very low VTE rates, such as lumbar discectomy and ACDF, were included. On the other hand, rates of VTE vary in the literature, most likely due to variability in definitions of VTE and surveillance methods employed. Given that this study defined VTE as reporting it in the database, these events most likely represent clinically significant occurrences that were detected using “typical” post-operative surveillance rather than a standardized screening protocol. Studies in which all patients undergo routine surveillance with CT scan or duplex ultrasound will obviously yield higher rates of DVT, but the clinical significance of these small, asymptomatic thromboses remains unclear. Somewhat surprisingly, the type of surgery (i.e. posterior lumbar fusion, anterior lumbar fusion, discectomy, etc.) did not end up being a significant predictor of VTE. Prior studies have shown increased rates for patients undergoing anterior or anterior-posterior (AP) thoracolumbar procedures, though this study did not appear to study AP procedures separately. While posterior cervical and posterior lumbar fusion were strongly associated with DVT in the univariate analysis, they were no longer significant in the multivariate model. It is likely that these were associated with other patient and surgical factors (i.e. duration of surgery) that were stronger predictors and remained significant while the type of surgery was a weaker predictor. It is likely that the number of levels fused would have also been associated with VTE, though this variable was not evaluated. This study has clearly identified some risk factors for VTE, but the spine community remains without any clear guidelines for when chemoprophylaxis is indicated in elective spine surgery, what agent should be used and in what dose, when it should be started, and the appropriate duration of treatment. The current study suggests that older, sicker, and obese patients undergoing long surgeries are at increased risk of VTE, but it is also likely that this population is at higher risk of complications of chemoprophylaxis. Much more work needs to be done on this topic, though the large number of patients required to study VTE makes this work challenging. Until more definitive studies are done, spine surgeons will have to continue to base their decisions about post-operative VTE prophylaxis on intuition rather than evidence.


Adam Pearson, MD, MS

Associate Web Editor