[Pulmonary embolism: analysis of cases without initial suspicion and sensitivity of three clinical models]

Pulmonary embolism (PE) is a frequent and severe condition. Although clinical models do exist, many patients are wrongly diagnosed. Our objective was to evaluate the clinical characteristics of patients with PE who are admitted at a hospital without initial clinical suspicion of it. We also evaluated the sensitivity of three clinical models in order to categorize the pre-test probability of PE.

Retrospective review of clinical cases of PE diagnosed in a teaching hospital during one year. We compared the clinical features of patients whose PE was not initially suspected with those of patients in whom PE was a first diagnosis. The three clinical models were applied in all cases.

58 patients were identified. In 15 out of them (26%), PE was not an initial diagnosis. Patients without an initial PE suspicion were older, and previous surgery was less frequent among them. The clinical models exhibited sensitivities (60, 85, 89%) which were lower than those previously reported but higher than the doctor's initial clinical suspicion (74%). The models showed bad agreement between them (kappa = 0.06-0.1).

An initial misdiagnosis of PE is not justified by atypical clinical characteristics. The use of clinical models would improve the clinical suspicion.