What do we know? The dilemma of modelling risk for aortic dissection

We read with great interest the article by Tozzi et al. [1] from Lausanne. The authors focused on the important topic of estimating the aortic diameter prior to the occurrence of an acute type A aortic dissection. By simply subtracting 7 mm (previously identified by the Yale group retrospectively [2]) from the post-dissection diameter in 102 non-syndromic patients, they determined that 99% of their patients would not have reached the threshold for prophylactic ascending replacement. Thus, the authors concluded that the indications for such prophylactic surgery should be changed to lower diameters, a conclusion that we generally support. Two major issues, however, have to be considered concerning this kind of reasoning. (1) The prevalence of non-dissecting, non-syndromic aortic aneurysmal disease at any specific size and thus the denominator of the equation, are completely unknown! Aortic aneurysms represent a silent disease whose presence as well as progression is often an incidental finding. Furthermore, the ‘normal’ aortic diameter at any age and for any sex may be estimated fairly accurately using established formulas [3, 4]. (2) The numerator for calculating risk models and a reliable number-needed-to-treat, however, also remains unknown! Epidemiological studies emphasize a significant number of unrecorded dissection cases at autopsy [5]. The only proven model in non-syndromic disease is thus still the ‘predicted probability for risk of complication’ based on a retrospective analysis of the post-dissection diameter [6]. The Yale group recently revisited their fundamental work from 1997 and found a second ‘hinge point’ of increased natural risk at 52.5 mm clearly warranting in principle and based on real life data, what the authors of the above work promote—a diameter shift to a lower aortic size threshold for prophylactic intervention [6]. Conflict of interest: none declared.