Comparison of three clinical prediction rules among patients with suspected pulmonary embolism

Certain clinical findings raise the suspicion of pulmonary embolism (PE) and may be useful in selecting patients for further diagnostic testing. Three prediction rules for PE have been described recently: Wells' rule (WR), Geneva rule (GR) and Miniati' rule (MR). The aim of present study is to compare the predictive accuracy of the three methods on the basis of our patients' results. Eighty-five patients admitted to our department with suspicion of PE were included into the study. Sixtythree patients were discharged with the diagnosis of PE, whereas in 22 patients, the initial PE diagnosis was ruled out. The three methods for assessing the clinical probability of PE classified similar proportions of patients into the low, intermediate and high clinical probability categories. The frequencies of PE in each method (WR, GR and MR) were 5%, 64% and 14% in the low category, 90%, 80% and 75% in the intermediate category and 100%, 100% and 94% in the high category (p = 0.001, 0.064, 0.001) respectively. When we compared the performances of WR and GR, including all possible total score values, the area under the ROC curve (AUC) was 0.99 for the WR (p=0.001) and 0.74 for the GR (p=0.001). When we used only the three probability categories (low, intermediate, high), AUC was 0.96 for the WR (p=0.001), 0.64 for the GR (p=0.04), and 0.7 for the MR (p=0.005). In conclusion, the present study indicates that clinical assessment is a fundamental step in the diagnostic work-up of PE. The Wells' method performs better than other two methods.