Potential pitfalls in the diagnosis of aortic dissection.

STUDY OBJECTIVES: The primary objective of our study was to identify clinical issues contributing to delay in the diagnosis of nontraumatic aortic dissection. DESIGN: Retrospective observational study. SETTING: A 425-bed community based, university-affiliated teaching hospital. PATIENTS: Medical records were analyzed for physiologic, clinical, and outcome variables in 32 consecutive patients who presented to our hospital with non-traumatic aortic dissection over a three-year period (1995-1997). MEASUREMENTS AND RESULTS: The mean time to diagnosis was 10.7 hours (range 0.5-72 hours). The time to diagnosis and clinical variables did not differ between type A and type B aortic dissections (8.7 vs 11.4 hours; P = 0.54). The diagnosis of aortic dissection was not entertained initially in 44% (14/32) of patients, leading to a significant delay in diagnosis compared to patients with suspected aortic dissection on admission (15.0 vs 4.5 hours; P = 0.008). Lack of a widened mediastinum on chest roentgenography (18.7 vs 6.6 hours; P = 0.026) and lack of hypertension (< 140/90 mmHg) at presentation (14.9 vs 7.1 hours; P = 0.03) were associated with a delayed diagnosis of aortic dissection as compared to presence of either finding. The presence or absence of chest or back pain, resting ECG changes, and a past medical history of hypertension did not correlate with time to diagnosis. There was a trend toward earlier diagnosis when the initial diagnostic modality was transesophageal echocardiography as compared to computed axial tomography (CAT) scan (7.7 vs 10.1 hours; P = 0.20). CONCLUSIONS: Patients with aortic dissection are frequently normotensive and may lack typical chest roentgenographic findings at presentation. A high index of clinical suspicion for aortic dissection is essential to avoid potentially fatal delays in its diagnosis.