Discussion

In our cohort, we address, for the first time, the impact of different anatomical aspects, particularly the relative inter-branch distances, on the outcome of surgically repaired CoA. Our study found that a shorter LCCA-IA distance was associated with a higher incidence of recoarctation, persistent hypertension, and re-elevation of blood pressure. Conversely, a longer LCCA-LSCA distance was also associated with more adverse outcomes. There is considerable interaction between the anatomical structure of the aortic arch and the expected arch anomalies. ^{Reference Dodge-Khatami, Ott, Bernardo and Berger17} One theory (anatomical theory) suggests that CoA may result from tissue from the ductus arteriosus invading the descending aorta distal to the aortic isthmus, leading to narrowing. However, this theory cannot explain cases of coarctation occurring outside the aortic isthmus. ^{Reference Hamdan3}

Another theory (haemodynamic theory) attributes the development of COA to altered fetal haemodynamics, specifically reduced blood flow through the aortic arch and isthmus. ^{Reference Buyens, Gyselaers and Coumans20}

We hypothesised that the displacement of left-common carotid artery towards innominate artery might cause stretching of the transverse arch; therefore, the constrictive effect of the invading ductal tissue may produce recoarctation and consequently re-elevation of blood pressure. The same effect could occur because the left subclavian artery is more inferiorly displaced and therefore the (LSCA-IA) distance increases.

The determinants of COA prognosis after surgery are still an area of research. While previous studies have demonstrated the impact of the anatomical shape and diameters of the arch on the prognosis regarding recoarctation, ^{Reference Turek, Conway and Cavanaugh21,Reference Ashcraft, Jones and Border22} persistent HTN, or re-elevation of blood pressure, ^{Reference Ou, Celermajer and Raisky23} we studied the relative branch anatomy and the impact of their inter-relationship on the prognosis after COA repair.

Meyers et al studied the distance between arch branches and sinotubular junction in both normal & bicarotid arch in patients with COA indexed to BSA; however, they did not provide clinical data about the impact of these measurements on the outcome such as recoarctation or HTN. ^{Reference Meyer, Turek and Froud24}

Additionally, some authors have noted the difficulty in consistently classifying the various morphologies of the aortic arch, which may limit the clinical utility of these classifications. ^{Reference Lashley, Curtin, Malcolm, Clark and Freeman25}

Arch geometry has been previously reported as a risk factor for residual HTN. ^{Reference Ou, Bonnet and Auriacombe26,Reference Iriart, Laïk and Cremer27}

Iriart et al found that gothic arch could contribute to persistent HTN after COA stenting despite the absence of residual COA. This may be due to shortened or absent transverse aortic arch that could alter the function of baroreceptors. ^{Reference Iriart, Laïk and Cremer27} This geometry could also affect arterial distensibility and increase central aortic stiffness. ^{Reference Ou, Celermajer and Raisky23}

Besides, the aortic angle in a gothic arch would lead to early and large systolic wave reflection. The resultant vascular remodelling in the aortic wall with increased wall stress and decrease compliance may finally lead to high central aortic pressure. ^{Reference London28}

Patients with a laterally displaced LSCA and a short aortic isthmus may develop a Gothic arch configuration after repair, leading to acute angulation of the aorta. ^{Reference Ou, Mousseaux and Celermajer29}

Indeed, the extended aortic arch anastomosis increases the size of the segment between the LCCA and the LSCA, but does not usually enlarge the segment between the innominate and the LCCA, and may, therefore, lead to a gothic geometry of the aorta, ^{Reference Siewers, Ettedgui, Pahl, Tallman and Delnido30} this may in part support our findings.

A previous study proved a negative correlation between the distance LCCA- LSCA and the duration before recoarctation with early reintervention for recoarctation in patients with long LCCA-LSCA segment. ^{Reference El Tahlawi and Tharwat31}

The common brachiocephalic trunk is another geometry that might affect the result of repaired COA. It is commonly referred to as a ‘‘bicarotid or bovine arch’’ where the left-common carotid artery arises more proximally on the arch, so as to originate immediately close to or as a branch of the IA.

A previous study by Reinshagen et al. highlighted the relatively high prevalence of bicarotic trunk in patients with coarctation of the aorta. This anatomical variation may influence the hemodynamics of the aortic arch and the risk of complications after surgical repair. ^{Reference Reinshagen, Vodiskar, Mühler, Hövels-Gürich and Vazquez-Jimenez32}

Turek et al ^{Reference Turek, Conway and Cavanaugh21} found that there were significantly higher recoarctation rates in the cohort with bovine arch (bicarotid) anatomy (P < 0.03). These results may be in concordant with ours, as we found that the possibility for recoarctation and persistence of HTN increases when LCCA is close to left IA.

The extended end-to-end anastomosis technique, commonly used for the surgical repair of coarctation of the aorta (COA), is designed to address a range of anatomical challenges, particularly in infants. ^{Reference Vasile, Laforest, Bulescu, Jalal, Thambo and Iriart33} However, its effectiveness may be influenced by certain anatomical variations, such as a shorter distance between the LCCA and the IA. ^{Reference Backer, Mavroudis, Zias, Amin and Weigel34} In cases where this inter-branch distance is reduced, the surgical repair may be more challenging, as the aortic arch may be less flexible, leading to less optimal surgical outcomes. A shorter LCCA-IA distance could contribute to a more acute angulation or a less favourable geometry of the aortic arch after the anastomosis, potentially predisposing patients to complications like recoarctation or persistent pressure gradients.

Turek et al. suggested that a bovine arch configuration limits the extent of possible aortic arch reconstruction due to a shortened transverse aortic arch, making extended end-to-end anastomosis technically challenging because of the insufficient clampable arch segment. ^{Reference Turek, Conway and Cavanaugh21}

In the same context, a systematic review concerning the COA repair in neonates, the authors linked Gothic arch anatomy to postoperative hypertension and bicarotid arch to restenosis and increased mortality in neonatal coarctation repair. ^{Reference Margarint, Youssef and Robu35}

However, Lashley et al ^{Reference Lashley, Curtin, Malcolm, Clark and Freeman25} studied the arch geometry in an adult population that had undergone surgical correction for COA, taking into account the A/T index (ratio of height/transverse diameter, which is higher in theory when the arch is gothic). They did not find a link with residual HTN (at rest and during exercise).

Similarly, in the study of Liu et al ^{Reference Liu, Jones and Cheung36} using ambulatory blood pressure monitoring in patients with corrected COA, they could not correlate adverse arch geometry with any adverse cardiovascular outcomes.

In addition, some investigators concluded that bovine arch anatomy did not influence recoarctation. However, the cohort prevalence of bovine arch anatomy in these studies was only 3 to 5%. ^{Reference McElhinney, Yang and Hogarty37,Reference Burch, Cowley and Holubkov38}

In a recent study that may support our hypothesis, Huang et al found that the relative inter-branch anatomy may influence aortic dissection, as the distance from the left-common carotid artery to the left subclavian artery was significantly stretched and greater in the patient group with acute Type A aortic dissection compared with the control group (P < 0.05). ^{Reference Huang, Liu, Wu, Lai and Cai39}

In our study, we found no association between the presence of hypoplastic arch and recoarctation.

A recent study conducted by Lehnert et al reported no association between hypoplasia of the aortic arch and the type of surgical repair on one side and the risk of recoarctation in young infants on the other side. ^{Reference Lehnert, Villemain, Gaudin, Méot, Raisky and Bonnet40}

Previous studies have reported conflicting findings regarding the association between aortic arch hypoplasia and the risk of recoarctation. While McElhinney ^{Reference McElhinney, Yang and Hogarty37} found that an absolute transverse arch diameter less than 3.5 mm was an independent predictor of earlier recoarctation necessitating reintervention in infants less than 3 months old. In contrast, however, Kaushal ^{Reference Kaushal, Backer and Patel41} failed to find any association between transverse arch hypoplasia and recoarctation in their cohort of infants and children. Similarly, Truong et al ^{Reference Truong, Tani, Minich, Burch, Bardsley and Menon42} and Callahan et al ^{Reference Callahan, Saudek, Creighton, Kuhn, Mitchell and Tweddell43} reported no association between the occurrence of recoarctation and arch Z-score in patients who underwent successful arch repair.

Bicuspid aortic valve (BAV) hasn’t also any association with recoarctation in our study. This result was in accordance with Lehnert et al study, ^{Reference Lehnert, Villemain, Gaudin, Méot, Raisky and Bonnet40} which could not associate between BAV and the risk of recoarctation.

The subgroup analysis of infants undergoing extended end-to-end anastomosis revealed significant differences in adverse outcomes based on the anatomical ratio of LCCA-IA to LSCA-IA (Group 1). Infants with a shorter ratio demonstrated a notably higher incidence of recoarctation. Additionally, Group 1 had a higher mean pressure gradient and a greater incidence of re-elevation of pressure gradient, suggesting that a smaller LCCA-IA/LSCA-IA ratio may be associated with more challenging aortic arch anatomy and worse long-term outcomes. Although the difference in hypertension rates did not reach statistical significance, the trend towards higher hypertension in Group I further supports the notion that anatomical features could influence post-surgical vascular function. These findings suggest that the LCCA-IA to LSCA-IA ratio may serve as an important predictor of surgical outcomes in infants, particularly in predicting recoarctation and pressure gradients, and could help refine pre-operative assessments and surgical planning.

We hypothesise that specific anatomical features, such as a shorter LCCA-IA distance, may contribute to difficulties in achieving optimal aortic arch geometry, which could predispose patients to adverse outcomes. The surgical technique itself, while tailored to the patient’s specific anatomy, might not fully correct these inherent anatomical challenges, leading to residual pressure gradients or recoarctation, as demonstrated in Group 1 of our subgroup analysis.

In our cohort, age at surgery did not significantly impact the outcome. However, other studies have suggested that older age may be associated with a higher risk of residual hypertension after COA stenting, potentially due to age-related changes in aortic compliance. ^{Reference Iriart, Laïk and Cremer27}

Multidetector CT is a valuable tool for three-dimensional assessment of the aortic root and arch anatomy. ^{Reference Sakrana, Nasr, Ashamallah, Abuelatta, Naeim and El Tahlawi44} In our cohort, we used maximum intensity projection image to draw the measurements. While maximum intensity projection images are widely used in clinical practice due to their ease of interpretation and automated measurement capabilities, they can be subject to limitations, particularly when the aortic arch curves significantly out of the plane of the image. This can lead to inaccuracies in measuring inter-vascular distances. Maximum intensity projection projections can potentially distort spatial relationships and may introduce variability in measurement, especially when the vessel plane deviates from the imaging plane. Curved plane reformatted images, on the other hand, allow for visualisation of the aortic arch along its centreline, potentially providing a more accurate assessment of these distances. However, they may require more complex image processing techniques and manual measurements. ^{Reference Bartolozzi, Neri and Caramella45}

More precise techniques—such as confirming measurements on thin-slice reconstructions or using dedicated oblique planes for each segment (IA–LCCA and LCCA–LSCA)—can reduce variability. Alternatively, employing a single oblique or curved multiplanar reformation that includes the origin of all three arch branches could enhance anatomical fidelity.

While the LSCA–IA distance was measured as a straight line, which may introduce error in markedly curved (“gothic”) aortic arches, we did not find any overtly gothic arches in our study cohort, thereby minimising the likelihood of this source of bias affecting our results. Nonetheless, we acknowledge that in future studies, a more accurate approach would be to measure both segments along the aortic centreline or to compare IA–LCCA to the total curved distance (IA–LCCA + LCCA–LSCA), which would better account for variations in arch curvature.

In our study, we primarily utilised a left posterolateral thoracotomy approach, which is a well-established technique for coarctation repair. The specific surgical technique employed was tailored to the individual patient’s anatomy and clinical condition. Tulzer et al concluded that, For patients with severe aortic arch hypoplasia (z-score < −4.59), a median sternotomy approach with cardiopulmonary bypass is recommended to minimise the risk of postoperative complications, such as recurrent coarctation. ^{Reference Tulzer, Mair, Kreuzer and Tulzer46}