Abstract: Objective: To analyze the clinical data of pulmonary embolism(PE) patients retrospectively for evaluating the performance of clinical scores(Wellsscore, revised Geneva scoreand their simplified versions) in assessing the probabi-lity of PE. Methods: A total of 78 patients diagnosed as PE and received treatment were recruited at Ruijin Hospital from June 2011 to June 2016 and at Ruijin Hospital North Branch from December 2012 to June 2016. All the patients were assessed by Wells score, revised Geneva score and simplified Wells score and simplified revised Geneva score, respectively. The performance of these scores were analyzed and compared. Results: For Wells score,15.38% patients had low clinical probability, 70.51% patients had moderate clinical probability and 14.10% patients had high clinical probability. By revised Geneva score, the percentages of low, moderate and high clinical probability were 20.52%, 70.51% and 8.98%, respectively. Bysimplified Wells score, the percentages of PE unlikely and PE likely were 33.33% and 66.67%, respectively. By simplifiedrevised Geneva score, the percentages of PE unlikely and PE likely were 55.13% and 44.87%, respectively; and the percentages of low, moderate and high probability were 19.23%, 74.36% and 1.28%, respectively, when the three levels scoring was used. Most patients had moderate or high clinical probability of PE when assessed by Wells score and revised Geneva score, with a percentage of moderate or high clinical probability of 84.61% and 79.49%, respectively; there was no significant difference in performance between Wells score and revised Geneva score (P=0.482). Among these 78 patients, the sensitivity of Wells score and simplified Wells score for PE were 84.61% and 66.67%, respectively(P<; 0.01). The percentages of PE patients diagnosed by simplified Wells score, simplified revised Geneva score-two levels scoring and simplified revised Geneva score-three levels scoring were 66.67%, 44.87% and 75.64%, respectively. The difference between the simplified revised Geneva score-two levels scoring and simplified revised Geneva score-three levels scoring was statistically significant(P<; 0.01). Significant difference was also found between simplified Wells score and simplified revised Geneva score-two levels scoring. Conclusions: For assessing the clinical probability of PE, the higher the clinical score, the higher the sensitivity of Wells score, revised Geneva score and simplified Wells score for diagnosing PE is. The Wells score has similar performance with revised Geneva score in diagnosing PE. Compared with simplified Geneva score-two level and simplied Wells score, Simplified revised Geneva score-three levels scoring could improve the vigilance of PE and decrease the rate of missed diagnosis.

Key words: Pulmonary embolism, Wells score, Revised Geneva score, Simplified Wells score, Simplified Geneva score