Frequent Use of CT Angiography, Low Diagnostic Yield, and Rare Utilization of Clinical Decision-Making Algorithms in Emergency Department Patients With Suspected Acute Aortic Syndrome

Aortic dissection and other forms of AAS are life-threatening and time-dependent emergencies. The rate of aortic CT angiography (CTA) is believed to be increasing in the US, while adoption of clinical decision-making algorithms (CDAs) has remained uncommon, despite multiple evidence-based guidelines recommending their use. Our objectives were 1. to evaluate the frequency of CTA use in patients with suspected AAS in our academic ED serving a high-volume aortic surgery center, 2. to determine the rate of documented CDA use by ED providers, 3. to estimate the frequency of low-value imaging based on predicted very low pre-test probability, and 4. to determine the diagnostic yield for AAS.

Trained abstractors reviewed charts of patients who underwent CTA from 1/1 to 5/1/25. Duplicate patients, those with a prior history of AAS, those imaged for reasons other than suspected AAS, and those transferred to our ED with known AAS were excluded. Abstractors recorded documented CDA use, d-dimer testing, components of the AD Detection Risk Score (ADD-RS), and CTA results. Each data element was abstracted twice, with disagreements adjudicated by third review by an EM physician. Simple descriptive statistics were used for analysis.

356 unique patients underwent CTAs over 125 days (2.84/day). Of these, 284 (80%) meet inclusion criteria. Use of a CDA (ADD-RS plus d-dimer) was documented in just one case. D-dimer testing was performed in an additional 22/284 (7.7%) cases, but as a means of ruling out PE rather than AAS. Estimation of the ADD-RS based on abstraction of its components indicated that 247/284 (87%) and 124/284 (44%) of cases would have been classified as low (score < 2) or very low (score < 1) pre-test probability, respectively. AAS was diagnosed in 5/284 (1.8%) of patients. The sensitivity and NPV of the estimated ADD-RS were 60% and 98.8% using a cutoff of <2, and 100% and 100% using a cutoff of < 1.

Our results indicate a rate of aortic CTA use nearly 5 times that previously reported in the literature. Conversely, documented use of a guideline-recommended CDA was rare (<0.5%), and 44% of CTAs were performed in patients with very low predicted pre-test probability. Future studies aimed at defining barriers to CDA use and interventions designed to address these barriers may increase diagnostic yield for AAS without increasing the rate of missed or delayed diagnosis.