Variation in emergency department use of computed tomography for investigation of acute aortic dissection

Introduction Acute aortic dissection (AAD) is a life-threatening condition making early diagnosis critical. Although 90% present with acute pain, the myriad of associated symptoms can make diagnosis a challenge. Our objective was to assess how we are using computed tomography to rule out acute aortic dissection specifically rate of ordering, diagnostic yield, and variation in practice. Methods We included consecutive adult patients presenting to two tertiary academic care emergency departments over one calendar year presenting with non-traumatic chest, back, abdominal, or flank pain. Primary outcome was rate of CT thorax/abdomen ordered to rule out AAD. Secondary outcome was variation in CT ordering, measured comparing number of CTs ordered per physician. Sample size of 12 per group was calculated based on an expected delta in mean CT ordered of 5 and a within group SD of 3. Results Thirty-one thousand two hundred one patients presented with truncal pain during the study period, 22,729 were included (mean 47 years, SD 18.5 years, 56.2% female); prevalence of AAD (N = 4) was 0.02%. CT was ordered to rule out AAD in 175 (0.7%) patients (mean 62 years, SD 16.5, 50.6% female). Significant variation between physicians ordering was found, with individual physicians ordering varying from 0.6 to 12%. Conclusions Current rate of imaging for acute aortic dissection is low and potentially inefficient, with a large variation in practice. These findings suggest potential for more standardized and efficient use of CT for the diagnosis of acute aortic dissection.