PURPOSE: The paradox in the diagnosis of pulmonary embolism (PE) is that it is both under-diagnosed and over-investigated. The widespread availability and excellent accuracy of CT Pulmonary Angiography (CTPA) may lower the threshold for performing this study and result in its over-use. On the other hand studies incorporating prediction rules have shown that PE can be excluded in patients with low clinical probability and normal d-dimer levels without an imaging study. The aim of our study was to determine if CTPA was being over-utilized.
METHODS: We retrospectively reviewed the clinical information of 231 consecutive ED patients suspected of PE from January 2005 to December 2005. We retrospectively applied the Wells’ simplified clinical prediction model score without prior knowledge of results of the d-dimer levels and the CTPA. We reviewed all information obtained from the CTPA studies.
RESULTS: The mean age of our patients was 53 years and 58.4% were women. PE was diagnosed by CTPA in 20% of suspected cases. The proportion of patients with low, moderate and high clinical probability of PE was 59%, 31% and 3.5% respectively. Among patients with low clinical probability, 54% had a d-dimer sample sent, and the d-dimer test was normal in 25%; none of these patients was diagnosed with PE. The most common alternative diagnoses on the CTPA were emphysema (7.6%), pneumonia (7%), atelectasis (5.5%), bronchiectasis (3.8%), and congestive heart failure (3.3%).
CONCLUSION: The 20% prevalence of PE in our series is consistent with prospective series, even though quantification of risk and d-dimer levels were not incorporated into the evaluation. If a more stringent approach had been undertaken, PE could have been excluded without a CTPA in roughly 1 out of 4 patients. In the majority (65%) of patients without PE, the CTPA was not useful in providing an alternative diagnosis.
CLINICAL IMPLICATIONS: In the practical setting of an urban ED, the diagnostic approach to suspected PE can be refined in patients who have a low clinical probability and normal d-dimer.
DISCLOSURE: Adil Shujaat, None.