PURPOSE: Evaluation of the indications for spiral chest CT and application of the pretest probability model in patients with suspected pulmonary embolism in the emergency department.
METHODS: We retrospectively reviewed the medical record of 383 consecutive patients in our institution, from May 2005 to April 2006, who underwent spiral chest CT to rule out pulmonary embolism.
RESULTS: Common presenting symptoms were dyspnea (67%), chest pain (39%) and palpitations (11%). Hypoxemia was the most frequent sign at 13%. Risk factors for thromboebolism were malignancy (16%), history of thromboembolism (7%) and immobilization (7%). A high d-dimer level preceded a spiral chest CT angiogram in 49 cases (13%). Pulmonary embolism was diagnosed is 39 of 383 cases (10%). The patients with pulmonary embolism (PE group), were older than the group without the disease (non-PE group), median age of 64 years versus 58 years (p=0.0094). In addition the PE group exhibited chest pain less frequently (23%vs41%, p=0.0373) and complained of palpitations more often ((23%vs9%, p=0.0237). The PE group had a higher incidence of malignancy (33%vs14%, p=0.0054) and recent surgery (15%vs4%, p=0.01). The spiral chest CT revealed an alternate diagnosis in 99 cases (29%). The Wells score was calculated retrospectively to evaluate the pretest probability of pulmonary embolism. Low Wells score correlated with negative spiral chest CT in 222 of 226 cases (98%). Patients with moderate probability Wells score had positive CT in 24 out of 138 cases (17%) and high probability Wells score in 11 of 19 cases (58%).
CONCLUSION: Spiral chest CT diagnosed pulmonary embolism in 1 out of 10 suspected cases and established an alternate diagnosis in 3 out of 10 cases. Patients with pulmonary embolism tended to be older and have more risk factors for thromboembolism. Ninety eight percent of patients with low Wells score did not have pulmonary embolism.
CLINICAL IMPLICATIONS: Spiral chest CT is used widely in the emergency department. Careful selection of patients and use of a pretest probability model may reduce the overuse of spiral chest CT.
DISCLOSURE: Vladimir Sabayev, No Financial Disclosure Information; No Product/Research Disclosure Information