Predictive Accuracy of Revised Geneva Score in the Diagnosis of Pulmonary Embolism


1 Department of Pulmonary Medicine

2 Tobacco Prevention and Control Research Center

3 Department of Chronic Respiratory Research Center,

4 Lung Transplantation Research Center,

5 Department of Radiology, NRITLD, Shahid Beheshti University M.C., TEHRAN-IRAN

6 Department of Chronic Respiratory Research Center


Background: Revised Geneva score is a clinical prediction rule used in determining the pre-test probability of pulmonary embolism (PE). It has been recently introduced and is independent of the doctor's experience applying the rule. This study aimed to evaluate the predictive accuracy of revised Geneva score in the diagnostic protocol of pulmonary embolism and its role in decreasing the need for pulmonary imaging studies. Materials and Methods: In this study, we evaluated the medical records of 242 patients suspected for pulmonary embolism who underwent CT scan of the lung as part of their diagnostic protocol from October 2007 to February 2009. Six patients were excluded from the study due to their indeterminate CT scan results. The mean age of patients was 58 yrs and 62% of patients were males. Results: The overall prevalence of pulmonary embolism was 24%. By increased scoring, the clinical probability of pulmonary embolism increased as well (P=0.011). According to the classification of revised Geneva score, clinical probability of pulmonary embolism was evaluated to be low in 25% of patients, intermediate in 72% and high in 2%. Prevalence of pulmonary embolism based on the CT scan results was 7.7% ranged (0.5-14.9) in the low probability category, 22.5% ranged (15.6-29.4) in the intermediate, and 50% ranged (0.01-0.99) in the high-probability category which were comparable with the rates reported in the derivation set except for the prevalence rate for high probability patients (9%, 27.5% and 71.7%, respectively). The area under the ROC curve was calculated based on continuous scoring to be 0.675. Conclusion: Revised Geneva score had an acceptable predictive accuracy in low and intermediate-probability groups. We could not reach a conclusion regarding high probability patients due to the small number of such cases in this study. (Tanaffos 2009; 8(4): 7-13)