Evaluation of the Utility of the Wells Score in Predicting Pulmonary Embolism in Patients Admitted to a Spine Surgery Service

Abstract

Study Design: A retrospective medical chart review of 4179 patients admitted to the spine surgery service. Objective: To evaluate the utility of the Wells score in predicting pulmonary embolism (PE) in patients admitted to a spine surgery service. Summary of Background Data: The decision to perform computed tomography pulmonary angiography (CTPA) to diagnose PE in patients who have undergone spine surgery requires consideration of multiple factors: false-positive CTPA results may lead to unnecessary anticoagulation treatment, and computed tomography scans are costly and expose patients to ionizing radiation. The Wells score was developed to assign risk categories to patients with suspected PE and thereby indicate the need for CTPA. However, the utility of the Wells score in predicting the likelihood of PE, specifically in spine surgery patients, has not been described to date. We identified all patients who were admitted to the spine surgery service at our institution from January 1, 2001 to December 31, 2011 and underwent CTPA. Each patient's CTPA result was classified as positive or negative for PE, and the reason for ordering the CTPA was recorded. The Wells score was calculated retrospectively for each patient, and risk categories were assigned by using the traditional and alternative interpretations of the Wells score. The reason for the CTPA, the Wells score, and Wells risk category were compared for patients who were classified as being positive or negative for PE. Results: Sixty-six of the 4179 patients who were admitted to the spine surgery service underwent CTPA for suspected PE. Nineteen of the 66 patients (28.8%) were diagnosed with acute PE, and the overall PE rate was 0.45% (19 of 4179 patients). The mean Wells score for patients diagnosed with PE was 5.3, whereas the mean score for the remaining patients was 4.9 (P = 0.793). Neither the traditional nor the alternative interpretation of the Wells score was predictive of PE (P = 0.394 and P = 0.178, respectively). Our study examined the utility of the Wells score in predicting PE in spine surgery patients. Conclusion: The results of the CTPA did not show a significant correlation with the Wells score or the reason for the test. Our findings indicate the need to develop a predictive scoring system that assesses the risk of PE and assists in the decision-making process for ordering CTPA in spine surgery patients.