Aortic elongation in aortic aneurysm and dissection: the Tübingen Aortic Pathoanatomy (TAIPAN) project†

Abstract

OBJECTIVES

To study the lengths and diameters of aortic segments in healthy and diseased aortas and to assess the role of aortic elongation in Type A aortic dissection (TAD) prediction.

METHODS

Ectasia and aneurysm were defined by ascending aorta diameters of 45–54 mm and ≥55 mm, respectively. Computed tomography angiography studies of 256 healthy, 102 ectasia, 38 aneurysm, 17 pre-TAD and 166 TAD aortas were analysed using curved multiplanar reformats.

RESULTS

The study groups were structurally equal. The diameter of the ascending aorta was 35 mm in the control group and was larger (P < 0.001) in the pre-TAD (43 mm) and TAD (56 mm) groups. The length of the ascending aorta from the aortic annulus to the brachiocephalic trunk was 92 mm in the control group, 113 mm in the ectasia group, 120 mm in the aneurysm group and 111 mm and 118 mm in the pre-TAD and TAD groups (all P < 0.001 compared with the control group). An ascending aorta length of 120 mm was exceeded in 2% of the control group, 31% of the ectasia group, 50% of the aneurysm group, 24% of the pre-TAD group and 48% of the TAD group. The correlation between the diameter and the length of the ascending aorta was r = 0.752; therefore, both parameters must be examined separately. A score considering both parameters identified 23.5% of pre-TAD patients, significantly more than the diameter alone, and 31.4% of ectasia aortas were elongated.

CONCLUSIONS

Patients with ectatic (45–54 mm diameter) and elongated (≥120 mm) ascending aortas represent a high-risk subpopulation for TAD.

INTRODUCTION

Ascending aorta dilatation is the only broadly accepted morphological risk factor for Stanford Type A aortic dissection (TAD), and a 55-mm diameter is the threshold for prophylactic ascending aorta replacement in the absence of connective tissue disorders [1]. However, it is well known that most TADs develop in aortas with diameters below this threshold [2–4]. Consequently, effective TAD prophylaxis is currently not possible. In recent studies, we showed that dissected aortas and aortas before dissection are elongated compared with healthy controls [3–5]. We hypothesized that aortic elongation may be a morphological risk factor for TAD, which may be usable in risk stratification and prognostication. We proposed a score involving the diameter and length of the ascending aorta for the establishment of a high-risk morphology [3]. Subsequently to our previous study [3], in this study, we evaluate the aortic length in aneurysmatic aortas exceeding the threshold diameter for prophylactic surgery (≥55 mm) and in ectatic aortas (45–54 mm), which do not reflect an indication for prophylactic surgery. We compare these aortic dimensions with the respective values in healthy subjects and in patients before and after TAD. The aim of this study was to assess the diagnostic value of the length of the ascending aorta to identify patients with a risk for TAD. We intend to define a high-risk subpopulation within a group of ectasia patients who may be considered for prophylactic surgery in the future.

METHODS

Study groups and clinical data

We compared 5 groups of patients retrospectively. We identified all patients treated for TAD in our centre between January 2006 and March 2017. We excluded patients who had a genetic or clinical diagnosis of a connective tissue disorder, iatrogenic or traumatic dissection. Within this cohort, we identified patients who had received adequate computed tomography angiography (CTA) within 24 months before the actual dissection occurred. These patients (n = 17) formed the pre-TAD group and the remaining 166 patients formed the TAD group (statistically independent). In the pre-TAD group, the median time between the pre-TAD CTA and the actual dissection was 5.1 months [quartile 1(Q1)–Q3 1.8–9.4; range 0.6–23.2 months).

Complementary to previous reports [3], we identified patients who were diagnosed with ascending aortic aneurysms ≥55 mm (n = 38) or ascending aorta ectasia of 45–54 mm (n = 102) in our institution from January 2009 to March 2017. All patients from the TAD and aneurysm groups, and some from the ectasia group, underwent surgery. However, for this study, only preoperative data were relevant.

We studied all our emergency department patients diagnosed with adequate CTA between March 2014 and March 2016 because of non-aortic emergencies (n = 271). The youngest patient in the TAD group was 23 years old. For age homogenization, we excluded all patients below the age of 23 from this emergency department group (n = 12). Furthermore, in this emergency department group, 3 individuals with ascending aorta diameters ≥45 mm (maximum diameter 47 mm) were excluded to avoid overlap and interference with our ectasia group. We refused to transfer them to the ectasia group to maintain the integrity of the emergency department cohort. This resulted in a healthy control group of 256 patients.

The following parameters were collected: date of birth, date of CTA, height, weight, gender and diagnosis of hypertension. The presence of 3 or more drugs from the groups, namely beta-blockers, angiotensin-converting enzyme inhibitors, angiotensin II receptor antagonists, vasodilators, calcium channel blockers groups, but not diuretics among chronic medications, indicated that patients had massive hypertension.

We checked for the presence of a bicuspid aortic valve in the pathological groups, not in the control group, because in the latter, only a minority of patients received echocardiography.

This study was approved by our local ethics committee (no. 076/2015R). Written informed consent was not necessary because of the retrospective observational nature of the study.

Image acquisition, processing and analysis

Technical details of CTA acquisition and of image processing were described elsewhere [3, 5]. Curved multiplanar reformats [6, 7] were created by manually defining the aortic central line with a 3D Bezier path in frontal, sagittal and transversal CTA reconstructions using the OSIRIX-MD (PIXMEO, Bernex, Switzerland) software package. At defined landmarks, orthogonal to this centre line, cross-sectional reconstructions were produced (Figure 1). The definition of the landmarks adhered to international guidelines [1, 6] with slight modifications:

• D1: aortic valve annulus (AV)

• D2: sinus of Valsalva (halfway between D1 and D3)

• D3: sinotubular junction (STJ)

• D4: mid-ascending aorta (halfway between D3 and D5)

• D5: orifice of the brachiocephalic trunk (BCT)

• D6: mid-aortic arch (halfway between D5 and D7)

• D7: distal aortic arch (directly downstream from the left subclavian artery)

• D8: descending aorta (at the height of the pulmonary artery bifurcation)

• D9: thoraco-abdominal (at the orifice of the coeliac trunk)

• D10: mid-abdominal (halfway between D9 and D11)

• D11: distal abdominal (directly proximal to the aortic bifurcation)

The aortic segments and length values were defined as follows, also corresponding to predefined conventions [6]:

• L1: aortic root (AV to STJ, D1 to D3)

• L2: ascending aorta (STJ to BCT, D3 to D5)

• L3: aortic arch (BCT to distal of subclavian artery, D5 to D7)

• L4: distal aortic arch (subclavian to the pulmonary artery bifurcation, D7 to D8)

• L5: descending aorta (pulmonary artery bifurcation to the coeliac trunk, D8 to D9)

• L6: abdominal aorta (coeliac trunk to the bifurcation, D9 to D11)

The length of the aortic segments (L1–L6) was measured along the centre line after identifying the mentioned landmarks (D1–D11). We measured the aortic perimeters and calculated the mean derived diameters to minimize failure resulting from the non-exact placement of length-measuring tools and from non-circularly shaped aortas. The measuring tool was placed within the aortic wall, retracing the shape of the aorta including true and false lumens and thrombus if present.

We also analysed the aortic arch morphology: In the Type I arch configuration, the most cranial point of the arch is between the supra-aortic vessels, and in the Type II arch configuration (‘gothic’), it is distal to the left subclavian artery [8].

Statistical analysis

There were no formal sample size calculations prior to the study, and we included all available patients. Because most, but not all, continuous variables fulfilled the criteria for normality (analysing skewness and kurtosis, the Shapiro–Wilk test), we waived the presentation of parametrical statistics. As previously reported [3], continuous data were described with the median, the first (Q1) and third (Q3) quartiles and the range (minimum and maximum) and are presented with box-and-whisker plots. Categorical data are reported as percentages. For inferential statistical comparisons of continuous variables between the 3 study groups, we performed a closed testing procedure: we first ran a Kruskal–Wallis test. In the case of a significant result, we performed pairwise Mann–Whitney–Wilcoxon rank-sum tests. Similarly, categorical variables were compared using the χ² tests. All reported P-values were 2-sided, and P-values ≤0.001 were considered to indicate statistical significance. The Spearman’s rank correlation coefficient was used to determine the correlations between continuous variables.

SSPS 23.0 (IBM Corp., Armonk, NY, USA) and MS Excel 2010 (Microsoft, Redmond, WA, USA) were used for the analyses and data presentation. Statistical analysis was performed in accordance with the guidelines of our association [9].

RESULTS

Demographic variables and hypertension

Table 1 presents the demographic variables in our study cohort. The groups were structurally equivalent with respect to gender, age and body dimensions.

The diagnoses of hypertension and massive hypertension were unequally distributed between the groups: In the ectasia, aneurysm, pre-TAD and TAD groups, the prevalence of hypertension was significantly (P < 0.05) higher compared with the healthy control group. However, the differences between the pathological groups were not significant.

In the ectasia and aneurysm groups, the prevalence of bicuspid aortic valve was significantly higher compared with the TAD group. In 2 TAD patients, no entry was found during surgery, while the other patients had an entry in the ascending aorta.

Aortic diameters

The diameter of the mid-ascending aorta (Fig. 2) in our control group had a median of 35 mm and ranged from 22 to 45 mm. In 75% of the healthy controls, this diameter measured ≤38 mm, and in 90% it was ≤40 mm. The diameter of the mid-ascending aorta was used to define the ectasia and aneurysm groups, and in both groups, in almost all patients, the ascending diameter exceeded the root diameter. In the ectasia group, diameters between 45 and 54 mm were evenly distributed, but in the aneurysm group, the distribution was asymmetric: 50% of aneurysm patients had ascending diameters of 55 or 56 mm, and 70% had diameters <60 mm. The largest ascending aortic aneurysm in our cohort measured 95 mm. In both the ectasia and aneurysm groups, the largest differences in diameter compared with the control group were found in the mid-ascending aorta, but for all other diameters from the aortic annulus to the aortic bifurcation, the diameters in the ectasia and the aneurysm groups were significantly (P < 0.001) larger compared with the control group. In the pre-TAD group, the median diameter of the ascending aorta was 43 mm and was 50 mm in the TAD group, which were both significantly (P < 0.001) larger than in the control group. However, 90% of the pre-TAD patients and nearly 70% of the TAD patients had ascending diameters <55 mm. At the landmarks from the STJ (D3) to the distal arch (D7), dissected aortas (TAD) had larger diameters than pre-TAD aortas (P < 0.05).

From the STJ (D3) to the brachiocephalic trunk (D5), the median aortic diameters in the pre-TAD group were very similar to those of the ectasia group and significantly (P < 0.05) smaller than those of the aneurysm group.

Aortic length

Figure 3 depicts the length values of different aortic segments in the study groups. The median aortic root length ranged between 23 mm (control group), 25 mm (ectasia and aneurysm groups) and 26 mm (TAD group). The median ascending aorta length (L2) was 69 mm in the control group and it was significantly (P < 0.001) longer in the ectasia (87 mm), aneurysm (98 mm), pre-TAD (87 mm) and TAD (92 mm) groups. We measured the length of the root and the ascending aorta individually, but because the STJ had been difficult to identify in individual cases, especially in aneurysmatic aortas, we found it advantageous to calculate the distance between the aortic valve and the brachiocephalic trunk as the complete length of the ascending aorta (L1 + L2). The median length of the complete ascending aorta was 92 mm in the control group, 113 mm in the ectasia group, 120 mm in the aneurysm group and 111 mm and 118 mm in the pre-TAD and TAD groups (all P < 0.001 compared with the control group). In the control group, 75% of complete ascending aortas measured ≤100 mm and 90% measured ≤110 mm. A value of 120 mm was exceeded in only 2% of the control group patients, but in 31% of ectasia patients, 50% of aneurysm patients, 24% of pre-TAD patients and 48% of TAD patients.

In the aortic arch, measured from the brachiocephalic trunk to the left subclavian artery, the length values of the ectasia (44 mm), aneurysm (46 mm), and TAD patients (44 mm) exceeded that of the control group (36 mm) significantly (P < 0.001) and the arch length in the pre-TAD group (39 mm) also exceeded (P < 0.05) that of the controls. Similar results were found for the segment of the distal arch, with values of 76 mm in the ectasia group, 83 mm in the aneurysm group, 80 mm in the pre-TAD group and 73 mm in the TAD group (all P < 0.001) compared with 61 mm in the control group.

The length of the descending aorta, measured from the pulmonary artery bifurcation to the coeliac trunk, did not differ significantly between the groups. In the abdominal aorta, the length differences between the control group (128 mm) and the pathological groups (between 136 mm and 139 mm) showed significance, but the absolute numerical differences were comparably small (data not shown).

Arch morphology

In the control group, most patients (77.3%) had Type I arch morphology and 22.7% had Type II arch (‘gothic’) morphology. In all pathological groups, the frequency of Type II arch morphology was significantly higher (P < 0.001), with 45.1% in the ectasia group, 60.5% in the aneurysm group, 58.8% in the pre-TAD group and 45.2% in the TAD group.

Correlation between ascending aorta dilatation and elongation

Figure 4 shows the scatter plot of all the individual values of the diameter and length of the ascending aorta . The values of the non-dissected aortas of the control, ectasia, aneurysm and pre-TAD patients are represented by grey circles. The correlation between the diameter and the length of the ascending aorta in the non-dissected aortas was significant (P < 0.001), and the Spearman’s rho correlation coefficient was r = 0.752. The arrays within the scatter plot represent the discrete definitions of normal, ectatic, aneurysmatic, elongated and non-elongated aortas. Red circles represent the respective values of TAD patients.

Indications for prophylactic surgery and TAIPAN score

Considering the recommendation for prophylactic ascending aorta replacement at a diameter of ≥55 mm [1], none of the patients in our control and ectasia groups and all patients in the aneurysm group would have had an indication for surgery. In the pre-TAD group, only 1 of the 17 (4.9%) patients exceeded the 55-mm threshold and another patient was close, with a diameter of 54 mm. In the TAD group, only 32.5% of dissected aortas met the 55-mm diameter threshold.

Recently, we proposed a score [3] to define the aortic risk morphology for TAD consisting of the diameter and length of the ascending aorta (Table 2). The diameter of the ascending aorta ≥55 mm was assigned 2 points and the diameter between 45 and 54 mm was assigned 1 point. A central line length of the complete ascending aorta, from the aortic valve to the BCT, of ≥120 mm was assigned 1 point (Table 2). A total of at least 2 points would indicate a high-risk aortic morphology that warrant prophylactic surgery.

In the actual data set, this score was not positive in any of our control group patients (specificity = 1.0) but in all aneurysm patients. The Tübingen Aortic Pathoanatomy (TAIPAN) score was positive in 23.5% of the pre-TAD patients (sensitivity = 0.24) and would have identified at least twice as many pre-TAD patients compared with the diameter alone. Within the ectasia group, 31.4% of the patients would have had a positive TAIPAN score. The mid-upper array in the scatter plot of Fig. 4 represents the subgroup of patients with ectatic and simultaneously elongated aortas. Although not applicable, the score would have also been positive in 52.4% of the TAD patients.

DISCUSSION

Study limitations: computed tomography angiography acquisition and processing

The exactness of ascending aorta measurements on CTA is challenged by motion artefacts. To minimize this problem, most CTA scans were recorded in flash technique or electrocardiography-triggering mode. All CTA data sets were intended for routine diagnostics and accordingly had adequate quality.

Measuring aortic diameters with a linear measuring tool is routine, but it is problematic because of projection artefacts and in cases of non-circularly shaped aortas. We measured the aortic perimeter in short-axis views and calculated the ideal diameter to overcome this problem.

Some of the landmarks used for the definition of aortic segments can be difficult to identify. This is especially true for the STJ, which appears flattened in aneurysmatic or dissected aortas. The aortic annulus and the offspring of the brachiocephalic trunk were easily identified in all cases, which is why we chose these landmarks for the definition of the complete ascending aorta.

The inclusion of the central line length in the indication for prophylactic ascending surgery would require 3D reconstruction of the CTA data set in routine diagnostics, which admittedly requires a significant additional expenditure of work and time, but is becoming standard in the era of aortic valve and aortic interventions.

Study limitations: study design and analysis

The retrospective nature of the study makes it unsuitable for risk factor identification: The formal proof of aortic elongation as a risk factor for TAD can only be done prospectively, which is particularly complex and ethically problematic because of inevitably exposing subjects to radiation and the risk of TAD. The actual approach must be considered a pragmatic compromise to investigate aortic elongation without these study-associated risks.

The pre-TAD group was evidently smaller than the other groups because patients who received a CTA before a TAD event were quite rare in our experience. However, to our knowledge, this is the largest reported cohort of such patients and we plan to increase the number in a multicentre approach.

Although the diameter of the ascending aorta is a continuous variable, it is usually examined and discussed in a discrete manner to enable thresholds, but the absolute difference between an ectatic and an aneurysmatic aorta may be evanescent. Other authors proposed indices referring to aortic diameters and body size [10]. We recently showed a significant correlation between age and aortic dimensions [3] and suggested the establishment of age-adjusted nomograms for aortic dimensions.

Aortic dimensions and risk morphology

The aneurysm and ectasia groups were defined by the diameter in the mid-ascending aorta. Within both groups, the aortic diameters upstream and downstream of the mid-ascending aorta were significantly larger compared with those of the control group. This suggests a generalized predisposition for aortic dilatation in patients with ascending aorta aneurysms even though other aortic segments are affected less substantially. This is remarkable because all patients with known connective tissue disorders have been excluded from this study. The higher prevalence of hypertension in the respective groups may be an explanation.

The diameter of the ascending aorta is traditionally considered the most important morphological risk factor for TAD, which seems to be debatable to us: The larger diameters in the TAD group compared with the pre-TAD group indicate a significant change in morphology during the acute process of dissection. This has been described previously [2] and reported that it is impossible to determine a TAD risk morphology from dissected aortas. Moreover, only a minority of TAD aortas, and an even smaller minority of pre-TAD aortas, exceeded the threshold value of 55 mm. Certainly, aneurysmatic dilatation is a risk factor for dissection as established in the landmark publications by Elefteriades et al. [11, 12], but the diameters in both our pre-TAD and TAD groups were much more akin to the values of the ectasia group, making these patients an interesting group to focus on. Generally, ascending aorta diameters between 45 and 54 mm are regarded as pathological, but the risk for complications in this cohort is reported to be comparably small [11, 12]; therefore, prophylactic surgery is not recommended. The prevalence of aortic ectasia in the general population is certainly higher than that of aneurysm, and aneurysm patients usually undergo surgery when they are diagnosed. Consequently, in the real world, most dissections occur at aortic diameters between 39 and 47 mm. These considerations raise the question of which morphological parameters define a high-risk subpopulation within the ectasia cohort. It seems plausible that the same structural changes that accompany aneurysmatic (transversal) dilatation, namely, a deficiency of elastic fibres, wall stiffening and reduced failure stress [13–16], also involve longitudinal material properties. The usually horizontal orientation of the dissection entry [17–19] is suggestive of longitudinal material failure of the aortic wall in TAD. The correlation between the diameter and the length of the ascending aorta in our non-dissected patients was significant but not perfect: some aortas were only mildly dilated but strongly elongated, which is why both parameters must be examined separately.

These aspects allow us to hypothesize that aortic elongation is another risk factor for TAD and the prominence of aortic elongation in pre-TAD and TAD aortas supports this assumption: a central line distance of ≥120 mm between the aortic annulus and the offspring of the BCT was found in more than a quarter of the pre-TAD and ectasia patients and in half of the TAD and aneurysm patients but only in 2% of the healthy control patients. Consequently, we regard this value as pathological.

We designed our TAIPAN score (Table 2) [3] so that aneurysm patients with aortic diameters ≥55 mm receive 2 points and are therefore, candidates for surgery. Ectatic (45–54 mm) and elongated aortas (≥120 mm) are rated with 1 point each, leading to 2 points for subgroup patients with combined pathologies. Within our pre-TAD group, the 55-mm threshold would have only identified one patient, if interpreted liberally, it may have identified two of the 17 patients. The TAIPAN-Score would have identified 4 of the 17 patients (23.5%). Nearly one-third (31.4%) of our ectasia patients had a positive TAIPAN score (mid-upper array in Fig. 4). According to the actual guidelines [1], these patients have no indication for surgery. However, we obtained evidence that prophylactic surgery may be advisable for this subgroup, which should at least be regarded as a high-risk population. Currently, the evidence is insufficient to provide a general recommendation, but in our centre, we changed our practice: we now test for aortic elongation in our routine diagnostics. Patients with ectatic and elongated aortas are followed up very closely and we tend to operate on them early. We will follow-up this interesting subgroup prospectively to evaluate the relative risk resulting from aortic elongation in ectasia patients.

CONCLUSION

Significant aortic elongation can be observed in aortas before and after dissection and in ectatic and aneurysmatic aortas. This and the pathophysiological plausibility of the concept are strong arguments that ascending aorta elongation may be a risk factor for dissection. The combination of ascending aorta ectasia and elongation enabled us to identify many more pre-TAD patients compared with diameter alone and renders the subpopulation of patients with ectatic (45–54 mm) and elongated (≥120 mm) ascending aortas a high-risk cohort for whom prophylactic surgery should be discussed.

Funding

This work was supported by the Dr Karl Kuhn-Stiftung, Tübingen, Germany.

Conflict of interest: none declared.

REFERENCES

Author notes