Too thin a beam of light in thick fog

Van Puyvelde et al. presented a histological investigation of aortic medial thickness in patients undergoing aortic surgery, reporting the evidence of thinner aortic media in dissection patients compared with aneurysm patients, but only in the subgroup with an aortic diameter exceeding 49 mm [ 1 ]. As current guidelines recommend to consider a lower threshold for surgical indication (50 mm instead of 55 mm) in patient subsets with adjunctive risk factors for dissection [ 2 ], the evidence that among patients with aortic diameter >49 mm those with dissection have a thinner aortic media seems to argue that aortic wall thickness might be among those factors to consider. However, the light that this study sheds on the issue of the unpredictability of aortic dissection may actually be weaker than it may seem, and the above argument fails to be persuasive, after taking into account some considerations that we will present below.

Firstly, the statistical strength of the analysis was poor: the mean medial thickness was 1528 μm in aortic aneurysm patients versus 1504 μm in aortic dissection patients ( P = NS), and only excluding those with a diameter <49 mm, the result of such comparison barely reached statistical significance (1412 vs 1525 μm, P = 0.0499). From the scatterplots illustrating the paper, one can appreciate that the difference is made fundamentally by those patients having a diameter in the 50- to 60-mm range. However, nearly 60% of dissection patients had a diameter ≤49 mm [ 1 ]; thus, the relatively rare dissections occurring on an overly dilated aorta drove the result, which therefore does not apply to the majority of dissections. It is now well known that, in most cases, aortic dissection occurs at a diameter smaller than the threshold defining an aneurysm [ 3 , 4 ]; therefore, the critical part of the aortopathy patient population is represented by the patients with normal or mildly dilated diameter, on whom research should focus on to identify new risk predictors. Thus, the authors introduced the ratio of the diameter to medial thickness, which was found to be significantly higher in dissection patients with a diameter >45 mm (actually barely significant again, with P = 0.0498). Frankly this looks like a mere statistical trick: the real ‘meat’ of the analysis was the weak ( r = −0.26) though significant negative correlation between diameter and thickness found only in dissection patients, not in aneurysm patients. As a consequence, dividing diameter by thickness artificially amplifies the difference between the two groups.

Secondly, the authors failed to perform further analysis that could have provided some more elements for the interpretation of their finding: for example, although they performed histological grading of ‘cystic medial necrosis’, elastin fragmentation, inflammation and atherosclerosis, they did not evaluate possible correlations between thickness and any of these pathology features. Why was a thinner media associated with dissection only in patients with an enlarged or aneurysmal aorta? Recent studies have estimated that an average ascending aortic diameter increase of ∼30% occurs acutely as the dissection develops [ 4 ]: what happens to wall thickness when the aorta is going to dissect? The remodelling of the media that occurs with chronic aneurysm development involves an increase in total medial mass, since the progressive thinning of the media is ‘slower’ than the increase in diameter [ 5 , 6 ]: put in mechanistic terms, this may be an attempt to compensate for the increased stress entailed by dilatation. It would be interesting to investigate whether this mechanism is lacking or inadequate in the aorta that will dissect before reaching a large size, or if it is just the rapid diameter increase accompanying (or heralding) dissection [ 4 ] that prevents such compensation. When considering the pathogenetic consequences of altered medial thickness in a proportion of aortopathy patients, we should also bear in mind that the Laplace law often advocated in this setting is currently questioned inasmuch as incorrectly estimating wall stress in biological vessels [ 7 ]. Considering only diameter and thickness when inferring on the forces driving mechanical failure of the aortic wall is surely limitative. If future studies will address the relation between aortic diameter, medial thickness and dissection risk again, they should also take individual blood pressure values into due account.

Thirdly, in such a ‘foggy’ topic as the one of the mechanisms driving dissection, there is a primary need to gain scientific bases for the development of clinically useful risk criteria and predictive tools: by now, it is hard to imagine an exploitation of the diameter/thickness ratio to these purposes. The fact that the association of a thinner media with the occurrence of dissection only applies to the patients with greater diameters limits the translational potential of the study. Aortic wall thickness (or intimal-medial thickness) is measurable by imaging methods, although this measurement is not in the clinical practice today; however, how does wall thickness relate to medial thickness? In both other authors' experiments [ 8 ] and ours [ 9 ], variable degrees of intimal thickening have been observed to occur during aneurysm development, thus the imaging-measured thickness of the wall could be not always a reliable surrogate for media thickness. Moreover, how thin must an aortic wall be to constitute an increased risk for aortic dissection? To answer this question, crucial for possible future clinical applicability, the normal range of wall thickness should be known: there are currently no reference data on large healthy populations. Without further insights, the only application of Dr Van Puyvelde's method would be by intraoperative histology. Nevertheless, if patients are being operated on, we can have two scenarios: if the aorta is smaller than 45 mm (a range where no significant difference in wall thickness was found between dissected and non-dissected aortas) [ 1 ], then wall thickness cannot be of help in the decision, whereas if it is >45 mm, then prophylactic aortic replacement is reasonable irrespective of wall thickness.

In conclusion, a candidate risk marker for aortopathy or dissection, should fulfil at least the following requirements: it should (i) be unrelated to aortic diameter, so that combined with diameter measurement it may have an incremental prognostic value; (ii) be significantly different between healthy subjects and aortopathy/dissection patients; (iii) have a reasonable pathogenetic justification (a plausible explanation of why it would represent a risk factor), unless its relation with the outcome/event is prospectively verified; (iv) be measurable in clinical practice.

REFERENCES