The recent article by von Kodolitsch et al1 demonstrates the poor discriminatory ability of individual clinical symptoms toward the diagnosis of acute aortic dissection. The recent emphasis on the use of likelihood ratios to emphasize symptoms, signs, and diagnostic tests has greatly enhanced appreciation of clinical data.2
Using the data presented in von Kodolitsch and colleagues' Table 1, the positive likelihood ratios (LRs) of symptoms and signs that do have a high sensitivity, ie, intense severity of pain, tearing or ripping pain, and mediastinal and/or aortic widening, were 1.56, 10.3, and 3.4, respectively. Alternatively, the negative LRs of these symptoms were 0.3, 0.4, and 0.3. The relatively high negative LRs indicate that in patients who are at intermediate and high risk for aortic dissection, absence of these symptoms alone would not effectively rule out the diagnosis.