Aortic Arch Variations and Supra-aortic Arterial Tortuosity in Stroke Patients Undergoing Thrombectomy

Endovascular thrombectomy has been shown to be an effective treatment in acute ischemic stroke due to large vessel occlusion. Nonetheless, successful recanalization is not possible in every case. According to the Highly Effective Reperfusion evaluated in Multiple Endovascular Stroke Trials (HERMES) meta-analysis, successful recanalization could not be achieved in about 29% in the thrombectomy cohort [1]. The Analysis of Pooled Data from Randomized Studies of Thrombectomy More Than 6 Hours After Last Known Well (AURORA) meta-analysis similarly reported that the rate of unsuccessful recanalization in the extended time window was about 20% [2].

One of the reasons for unsuccessful recanalization is difficult arterial access to the target vessel occlusion [3‐7]. Especially unfavorable vascular anatomy of the aortic arch and the supra-aortic vessels is associated with a prolonged thrombectomy procedure [8]. Kaymaz et al. reported that supra-aortic vessel tortuosity leads to prolonged time until access to the internal carotid artery is achieved and to lower rates of successful recanalization [9]. Ribo et al. also reported longer procedure times and lower rates of recanalization leading to poor outcome due to difficult arterial access [10].

Unfavorable anatomy is a frequent phenomenon among patients with acute ischemic stroke. In aortic arch types II and III, the supra-aortic arteries arise with a steep angle making it more difficult to navigate catheters [8]. A wide range of prevalences between 35% and 86% have been reported for these aortic arch types [5, 8, 11‐14]. Similarly, variant branching patterns, such as the so-called bovine aortic arch can be challenging when trying to achieve access to the left internal carotid artery [8, 9]; however, these branching patterns have been reported with inconsistent nomenclature in the past as pointed out by Layton et al. [15].

Therefore, this present large-scale retrospective study aimed at systematically assessing the prevalence of aortic arch types, branching patterns and supra-aortic arterial tortuosity in stroke patients who underwent thrombectomy at a high-volume comprehensive stroke center over a period of 11 years. These data may help training programs and manufacturers of anatomical flow models and angiography simulators to provide more realistic training scenarios.

Supra-aortic vessel tortuosity was first described and categorized by Metz et al. in 1961 and later revised by Weibel and Fields [17, 18]. Anatomical analyses of aortic arch branching patterns even date back to 1933 [21]. Now that endovascular thrombectomy is an evidence-based treatment in acute ischemic stroke due to large vessel occlusion, neurointerventionalists frequently face challenging cases especially in older patients due to unfavorable anatomy [9, 22]. The present large-scale anatomical analysis therefore aimed at systematically assessing the prevalence of aortic arch variants and types as well as supra-aortic arterial tortuosity in stroke patients undergoing thrombectomy.

The reported range of prevalence of type II aortic arches is 20–41% and 15–39% for type III (Supplemental Table 1). Hence, our results of 34.3% for type II and 19.9% for type III aortic arches are within that range. With about 20%, type III aortic arches make up a relevant proportion of thrombectomy cases. Especially in combination with carotid tortuosity, type III aortic arches can be challenging for neurointerventionalists [8].

Older age was associated with aortic arch type III confirming prior studies that the aorta undergoes changes with increasing age [12, 13, 23]. Also, women seem to be more susceptible to this change of aortic geometry compared to men [24]. Some authors such as Demertzis et al. however, reported that there is no correlation with age and gender [11].

Among 1705 analyzed aortic arches in this study, 3 cases with right-sided aortic arches were present. The formation of the aortic arch is indeed a complex process. During the third week of gestation, two aortae are formed which are later connected by several aortic arches [25]. Some arches regress, progress or develop from cranial to caudal forming the final aortic arch and variations [26]. Through regression of the right fourth arch and the right dorsal aorta, the left-sided aortic arch is formed, which is the most common variant. The right-sided aortic arch with mirror configuration develops from regression of the left dorsal aorta and persistence of the right fourth arch [27‐29].

Frequently observed branching pattern variants are i) a common origin of the brachiocephalic trunk and the left common carotid artery and ii) the left common carotid artery arising from the brachiocephalic trunk. In our study, these variants had a prevalence of 15.1% and 12.3%, respectively, comparable to results of prior studies (Supplemental Table 2). As pointed out by Layton et al., the umbrella term bovine aortic arch is often falsely used to describe both these variants [15]. In the true bovine anatomy, one single great vessel arises from the aortic arch and gives off the right and left subclavian arteries and then continues as the bicarotid trunk [20]. This variant was not present in any of our cases. Snelling et al. and Rosa et al. developed scores based on the anatomy of the aortic arch and arterial tortuosity to predict procedural time and success of thrombectomy. Both used the term bovine arch and did not differentiate between the two abovementioned variants [8, 22].

Instead of bovine aortic arch, we suggest using other terms such as radix communis (common origin of brachiocephalic trunk and left common carotid artery) and truncus communis (common trunk due to left common carotid artery arising from the brachiocephalic trunk) for these variants as these describe the vascular anatomy more precisely [21]. Interestingly, the truncus communis variant is associated with a higher proportion of vessel tortuosity compared to the radix communis variant. Our data do not offer any explanation for that, though.

The prevalence of supra-aortic arterial tortuosity in our study reflects our observations in the daily practice of thrombectomy. Elongations make up to nearly 40% in our study cohort. Kinking is less frequent with about 10%. Coiling of the internal carotid artery can be seen in nearly 8%. Whether a vessel is kinked or coiled can often only be determined if it is viewed in all dimensions. The prevalence values of some authors differ tremendously, especially regarding kinking and coiling (Supplemental Table 3). This might be related to the study population and how these types of tortuosity were defined [3, 30‐33].

Beigelmann et al. and Togay-Isikay et al. stated that carotid tortuosity is a result of alterations in embryological development and not related to age and risk factors [30, 34]; however, our results indicated that older age is associated with supra-aortic vessel tortuosity suggesting that vessels undergo age-related changes over time. Female gender was also associated with tortuosity of the internal carotid artery. This has been similarly reported by Martins et al. and Sacco et al. [32, 35].

Although dyslipidemia is known to be a cardiovascular risk factor, it was inversely correlated with tortuosity of the internal carotid artery in our study. Our data do not offer any explanation for this finding, though. Interestingly, in the abovementioned study by Togay-Isikay et al., hyperlipidemia was slightly less frequent in patients with carotid tortuosity (45.9% vs. 53.1%) [34]. Others such as Del Corso et al. reported significantly higher rates of dyslipidemia in patients with carotid tortuosity [36].

The presented prevalence data may help training programs and manufacturers of anatomical flow models and angiography simulators to provide more realistic training scenarios. Especially fellows who are in the beginning stage of their neurointerventional career, may benefit from a deeper understanding of prevalence and configuration of supra-aortic artery variants and might also benefit from more realistic anatomic models and simulations. The relatively high proportion of type III aortic arches may also encourage device manufacturers to develop catheters, which are especially suited for this type of anatomy.

The main limitation of this study is the retrospective design. It also needs to be stated that this study was performed at a single center in southwest Germany. The majority of our patients were European with only few African and Asian patients. Ethnicity was not assessed in this study, though. The results might have been different if the study had been performed in another country. Besides, approximately 10–15 out of 2026 cases were repeated thrombectomies due to recurrent large vessel occlusion. Each thrombectomy case is individually documented in our registry. This might have led to a slight bias.

With about 20%, type III aortic arches represent an important proportion in stroke patients eligible for endovascular stroke treatment. Nearly half of the stroke patients presented with supra-aortic arterial tortuosity, mostly arterial elongation.