Scannographic Study of Risk Factors of Abdominal Aortic Aneurysm Rupture

Abstract

Background: Current decision about when to operate abdominal aortic aneurysms (AAAs) is based only on the maximum aneurysm diameter (MAD). However, small aneurysms still rupture and we can observe very large AAA without any symptom. A simple morphologic analysis could be a tool to assess the risk of rupture. The main objective of this study was to assess the relevance of ratios between MAD and healthy aorta on computed tomography (CT) as a risk factor of AAA rupture. The secondary objective was to evaluate CT signs as risk factors of AAA rupture.

Methods: Retrospective observational bicentric study comparing CT scans of a ruptured AAA group and a control group treated electively was conducted. Appariement 1:1 based on MAD was applied. Ratios between healthy aorta diameters at several levels, celiac trunk (CTR), superior mesenteric artery (SMA), highest renal artery (RA), and the MAD were calculated. The presence of blebs, crescent signs, ruptures of calcifications of the aneurysm sack, and draped aorta were notified.

Results: From 2010 to 2016, 38 ruptured AAA and 38 controls were included. Ratios were superior in the rupture group, respectively: MAD/CTR [2.77 (±0.5) versus 2.58 (±0.4) P < 0.095], MAD/SMA [2.92 (±0.7) versus 2.74 (±0.5) P < 0.194], and MAD/RA [3.02 (±0.70) versus 2.76 (±0.5) P < 0.054] but not significatively. Receiver operating characteristic curve analysis demonstrated optimal threshold to detect rupture at 2.8 for the ratio MAD/CTR (area under the curve (AUC) 0.593, sensitivity 47.4%, specificity 78.9%), at 3.3 for the ratio MAD/SMA (AUC 0.564, sensitivity 31.6%, specificity 92.1%), and at 3.3 for the ratio MAD/RA (AUC 0.591, sensitivity 31.6%, specificity 94.7%). Bivariate analysis for rupture risk factor showed significance for the three ratios (MAD/CTR > 2.8 [OR = 11 (1.42; 85.20) P < 0.0217], MAD/SMA > 3.3 [OR = 10 (1.28; 78.12) P < 0.0281], and MAD/RA >3.3 [OR = 11.00 (1.42; 85.20) P < 0.0217]). One scannographic sign was more present in the rupture group: crescent sign 36.8% versus 5.3%, P = 0.0007, as well in bivariate analysis [OR = 7 (1.59; 30.80) P < 0.0326].

Conclusions: In our experience, specific ratios when they exceed calculated threshold, seem to be more prone to rupture. We could consider that these measures, easy to apply in clinical practice, would be complementary keys for rupture risk individual assessment.