Association between aortic coverage and spinal cord ischemia after endovascular repair of type B aortic dissection

Abstract

Objectives

To examine the association between aortic coverage and occurrence of spinal cord ischemia (SCI) after thoracic endovascular aortic repair (TEVAR) for type B aortic dissection.

Methods

Thirty-eight consecutive patients (mean age 52 years; 35 men) who underwent TEVAR for type B aortic dissection at our centre were included. Patients were stratified into two groups based on stent graft length (SGL): group I (≤ 200 mm; n = 19) and group II (> 200 mm; n = 19). All the procedures were performed under strict blood pressure monitoring. Preoperative cerebrospinal fluid (CSF) drain was instituted in two patients.

Results

Mean SGLs were 200 mm in group I and 277 ± 27 mm in group II. The number of segmental arteries covered was significantly different between the two groups (p < 0.001). There was no significant difference in vertebral artery dominance between groups I and II (p = 0.99). One patient in group II, who also had bilateral internal iliac artery occlusion, developed postprocedural neurological deficit referrable to SCI, which resolved completely after institution of CSF drainage. However, the incidence of SCI was not significantly different between group I and group II (p = 0.5).

Conclusion

Based on our experience, we recommend longer aortic coverage (beyond 200 mm) in type B aortic dissection (TBAD) for better aortic remodelling, provided that the mean arterial pressure of > 90 mm Hg is maintained perioperatively to avoid SCI.

Introduction

The pathogenesis of spinal cord ischemia (SCI) after thoracic endovascular repair (TEVAR) is multifactorial. The two theories which are implicated include inadequate remodelling of collateral blood supply and athero-embolism of aortic plaques into segmental arteries [1]. Uchida reported the incidence of SCI to be 0–10.3% (average 4.5%) after TEVAR [2]. The main blood supply of the spinal cord is from the anterior and posterior spinal arteries, the inflow vessels being left subclavian artery (LSCA), intercostal, lumbar, and internal iliac arteries (IIA) [3]. Coverage of the LSCA during TEVAR compromises collateral circulation, including the vertebral arteries. Similarly, extensive coverage of long segments of the thoracic aorta using multiple stent grafts compromises intercostal (T7–L1) and lumbar segmental arteries that supply the anterior spinal artery [2]. Hence, extensive aortic coverage theoretically is associated with increased risk of SCI particularly with concomitant LSCA coverage without revascularization. However, limiting the extent of aortic coverage in type B aortic coverage can increase the likelihood of persistent false lumen flow and the need for secondary interventions.

In another study (n = 67), length of aortic coverage was an independent predictor of SCI with 205 mm as a threshold for increased risk [4]. Feezor et al. concluded that risk of SCI increases by 30% for every 2-cm coverage of thoracic aorta [5]. The lacuna in existing literature is that most of the published studies are pertaining to thoracic aortic aneurysms. However, there is difference between TEVAR in aneurysmal aorta and type B aortic dissection (TBAD). In aneurysmal aorta, there is high risk of athero-embolism from aortic plaque, thus occluding the segmental arteries. In addition, prompt sealing after stent graft placement leads to rapid sac thrombosis and cessation of flow through the intercostal arteries before collaterals have developed [1]. On the contrary, in TBAD, there is lower burden of mural thrombus with lower risk for athero-embolism. In addition, the persistence of false lumen perfusion after TEVAR maintains perfusion through intercostal arteries and provides time for collateralization.

There is lack of authentic data pertaining to the extent of aortic coverage and occurrence of SCI for TBAD. Hence, the aim of the present study was to examine the association between aortic coverage and occurrence of SCI after TEVAR for TBAD.

Material and methods

Study population

The medical records, radiographic studies and database of all consecutive patients who underwent TEVAR for TBAD between January 2014 and December 2020 at our centre were retrospectively reviewed. The Institutional Ethics Committee approved the study and waived off the need for informed consent. The indications of TEVAR were as follows: aortic/impending rupture, refractory pain, persistent hypertension, visceral/limb mal-perfusion and unfavourable remodelling. Patients with connective tissue and inflammatory disorders were excluded from this study. The sizing of the stent graft was determined using pre-procedural computed tomography angiography (CTA). CTAs were also reviewed for vertebral artery and internal iliac artery.

Cerebrospinal fluid drainage

We put prophylactic spinal drain only in patients at high risk of SCI [6]. Preoperative cerebrospinal fluid (CSF) drain was placed in two patients who had hypoplastic/atretic right vertebral artery the night prior to the procedure by cardiac anesthesiologist [7]. The pressure of CSF was maintained up to 10 mm Hg perioperatively and the catheter was removed after 48–72 h.(1).

Intraprocedural and postprocedural management

The goal was to cover the primary tear and redirect blood flow into the true lumen. The procedure was performed under general anaesthesia. Spinal cord perfusion pressure is mean arterial pressure (MAP) minus CSF pressure [8]. Strict blood pressure (BP) monitoring was done and MAP of > 90 mm Hg was maintained intraoperatively and for first 48 h after the procedure. Vasoactive medications were used, if needed. Patients were admitted postoperatively to the cardiac intensive care unit. Neurological evaluation was done hourly for first 48 h.

Definitions

SCI was defined as any new lower limb motor or sensory deficit, or both, in the absence of any documented intracerebral hemispheric events [5]. It was considered transient if it resolved at the time of discharge. When the preoperative functional status could not be restored, it was considered permanent [9].

The length of aortic coverage was determined by stent graft length (SGL). SGL was the absolute standard pre-deployment length of the implanted stent graft (manufacturer specified) [10, 11]. In case of multiple stent grafts, it was calculated as the length of all implanted segments minus the overlap. When multiple stent grafts were used, overlap of at least 4 cm in straight segment and 6 cm in curved segment was aimed for. Patients were stratified into two groups based on SGL: group I (SGL ≤ 200 mm) and group II (SGL > 200 mm). The number of segmental arteries covered was also evaluated by comparison between preoperative and postoperative CTA (Fig. 1). In cases where postoperative CTAs were not available, the number of segmental arteries covered was estimated by evaluating the distance between the lower extent of stent graft and celiac artery (fixed vascular landmark) on digital subtraction angiography (DSA) and extrapolating it to preoperative CTA.

Fig. 1.

Cinematic volume rendered image depicting segmental arteries covered (marked with white arrow) after multiple (A) and single (B) stent graft deployment

Stastical analysis

Statistical analysis was done using SPSS (v24). Continuous variables were presented as mean ± standard deviation and Student’s t test/Mann–Whitney U test was used to compare values between the groups depending upon normality. Categorical variables were presented as frequencies and compared using chi-square test. A p value < 0.05 was considered statistically significant. The main outcome measure in the study was aortic coverage and occurrence of SCI.

Results

Patient demographics

Thirty-eight consecutive patients (35 men) underwent TEVAR for TBAD. The mean age of the patients was 52 ± 15 years. There were 19 patients each in group I (SGL ≤ 200 mm) and group II (SGL > 200 mm). The various comorbidities that were present were as follows: renal insufficiency (n = 5), chronic obstructive pulmonary disease (n = 3), hypertension (n = 17) and anaemia (n = 3). Acute TBAD dissection was present in 15 patients. The dissection flap was confined to thoracic aorta in 4, whereas in 34 patients, it extended to the abdominal aorta. The major entry point was proximal descending thoracic aorta (DTA) in 32, mid-DTA in 5 and distal DTA in one patient. The additional entry points, if present, were mainly seen in the abdominal aorta. The median length of stay in hospital was 4 days (2–10 days).

Procedural and anatomical characteristics

Hybrid procedures (neoarch formation) were performed in 5 patients in cases of insufficient proximal landing zone. No significant difference in vertebral artery dominance was observed between groups I and II (p = 0.99). None of the patients developed hypotension during and after the procedure. Technical success was achieved in 100%. The comparison of patient-related and procedural risk factors between the two groups is shown in Table 1.

Table 1.

The comparison of patient and procedural risk factors among two groups

	Group I (n = 19)	Group II (n = 19)	p value
Patient factors
Age (years)	48.8 ± 14	55.15 ± 16	0.58
Renal insufficiency	2	3	0.63
COPD	2	1	0.61
Hypertension	9	8	0.75
Diabetes mellitus	2	3	0.63
Coronary artery disease	2	1	0.61
Smoking	3	2	0.63
Concomitant abdominal aortic aneurysm	0	1	0.50
Anemia (Hb < 12 g/dl)	2	1	0.61
Procedural risk factors
Urgency of procedure (acute)	7	8	0.75
Internal iliac artery stenosis	1	1	0.99
Atresia/hypoplasia of right vertebral arteries	1	1	0.99

Stent graft characteristics between two groups

The zone of landing was zone 3 (distal to left subclavian artery) in 3, zone 2 (between left common carotid artery and left subclavian artery) in 30 and zone 0 in 5 patients where hybrid repair was performed. Mean SGLs were 200 mm in group A and 277 ± 27 mm in group B. The number of segmental arteries covered were significantly different between the two groups (p < 0.001) (Table 2). LSCA was covered in all patients with the stent graft. When right vertebral artery was atretic, revascularization of LSCA was performed. A single stent graft was deployed in group I, whereas the mean number of stent grafts deployed in group II was 2.26. Celiac artery was not covered in any patient. None of the patients developed arm ischemia. After implantation of endograft, partial false lumen filling on follow-up computed tomography (CT) angiography was observed in 7 patients, 3 in group I and 4 in group II. Postoperative CT angiography was not available for 8 patients.

Table 2.

Comparison of stent graft characteristics and occurrence of SCI among two groups

	Group I (n = 19)	Group II (n = 19)
Stent graft length	≤ 200 mm	> 200 mm Mean length: 277 ± 27 mm
No. of segmental arteries involved
< 8 9–12	15 4	0 19 (p < 0.001)
No. of stent grafts	One: 19	Two: 15 Three: 3 Four: 1
LSCA coverage	Covered in all patients	Covered in all patients
Occurrence of spinal cord ischemia	0	1 (p = 0.50)

Spinal cord ischemia

One patient in group II, who also had bilateral internal iliac artery occlusion, developed postprocedural neurological deficit referrable to SCI. The procedure was performed urgently due to aortic rupture, so spinal drain could not be placed. The patient developed lower limb paraplegia and the onset was within 15 h. Spinal drain was immediately placed along with BP augmentation. Spinal drainage catheter was promptly inserted and was allowed to drain for 72 h. The drainage was kept to less than 350 ml/24 h to prevent cerebral herniation or subdural hematoma. SCI was transient, i.e. resolved within 72 h. Rest of the in-hospital course was uneventful and the patient was discharged on postoperative day 10. However, the incidence of SCI was not significantly different between group I and group II (p = 0.50). None of the patients developed delayed SCI. Posterior circulation stroke was not observed in any patient.

Discussion

In the present study, we observed that despite extensive aortic coverage during TEVAR (mean SGL: 277 ± 27 mm) and concomitant LSCA coverage in group II, there was low risk of SCI. Hence, the critical length of coverage of 205 mm, beyond which the risk of paraplegia significantly increases, is debatable in patients undergoing TEVAR for TBAD [4].

Similarly, in a study by Zipfel et al., where stent grafts were implanted in 406 patients, 164 had TBAD. The length of aorta covered was 204 mm (75–584 mm) and was calculated in a method similar to our study. Prophylactic CSF drain was put in 4 patients. Eleven patients (2.7%) had new-onset spinal cord events. The authors concluded that incidence of SCI was low despite two-thirds of cases being treated in emergency and despite extensive aortic coverage [10]. In another study by Chiesa et al., the intentional occlusion of the LSCA, as well as coverage of extensive thoracic aortic segment, was not predictors of paraplegia [12].

The occurrence of SCI after TEVAR is a dreaded complication and occurs due to multiple factors. The blood supply of spinal cord is derived from multiple arteries. Hence, disruption of perfusion in any of these theoretically increases the risk of SCI [3]. Impairment of perfusion by intercostal arteries is considered a major mechanism. Stent grafts exclude the intercostal arteries in TBAD, but the occlusion is not abrupt with continued perfusion of false lumen through intercostal arteries providing time for collaterization [1]. Hence, association of coverage of long thoracic segment and SCI is more pertinent in cases of thoracic aortic aneurysms where there is rapid sac thrombosis and cessation of flow through intercoastal vessels, unlike TBAD.

The low incidence of SCI despite extensive aortic coverage can be explained by the concept of collateral network. Intraspinal system that feeds the anterior spinal artery via the segmental arteries has multiple connections with longitudinally oriented continuous epidural arcade. Similarly, preformed collaterals in erector spinae and psoas muscles (paraspinous network) maintain perfusion to spinal cord despite occlusion of intercostal arteries [13, 14]. Hence, after deployment of stent graft, collateral network remodelling provides important alternate source of blood supply. Etz et al. described this concept in a porcine model and observed that the various vessels involved in the recovery of spinal cord perfusion start dilating within 24 h and the structural remodelling of intraspinal vessels and collateral system is completed in about 5 days [14]. Colman et al. proposed that the non-occluded segmental arteries play an important role in maintaining perfusion to ischemic spinal cord segments by reversal of direction of flow in anterior spinal artery [15].

In addition, just one SCI event occurring in our series despite extensive aortic and LSCA coverage can be attributed to maintenance of strict MAP which is extremely important adjunctive measure. In study by Zipfel et al., significant number of cases developed paraplegia due to hypotension either due to hypovolemia, systemic inflammation or medication [10]. Chiesa et al. proposed that periprocedural hypotension (MAP < 70 mm Hg) was significant predictor of SCI [12]. Hence, careful monitoring and prompt correction of hypotension are important factors to prevent paraplegia.

In TBAD, the primary entry tear is just distal to LSCA; hence, intentional coverage of LSCA is done for better proximal seal. The need for routine LSCA revascularization when the vessel is covered during TEVAR for TBAD is controversial. In a previous meta-analysis, LSCA coverage without revascularization increased the risk of arm and vertebrobasilar ischemia, but there was non-significant increase in SCI [16]. However, in a recent meta-analysis, LSCA revascularization was associated with similar risk of SCI and stroke when compared to no LSCA revascularization [17]. Extensive aortic coverage is only a relative indication for preoperative LSCA revascularization. Hence, LSCA was covered without revascularization in the present study when the right vertebral artery was dominant/co-dominant. One patient who developed SCI had occluded bilateral internal iliac arteries. This emphasizes the importance of IIA in providing collateral supply to spinal cord via radicular lumbosacral arteries [18].

Extensive stent graft coverage is required in cases of acute aneurysmal false lumen, contained rupture, to promote false lumen thrombosis, improve aortic remodelling and reduce long-term aneurysmal degeneration (Figs. 2 and 3). In a study by Xue et al., in 201 patients (mean age 52.4 ± 11.5 years; 178 men), the patients were divided into lower and higher groups on the basis of percentage stented descending aorta (PSDA). In higher group, where PSDA was more than 31.3%, the SGL was 174.82 ± 38.46 mm. None of the patients in their series developed SCI. Instead, it was observed that large percentage of stented descending aorta is independently associated with lower risk of thoracic aortic expansion and more chance of completely thrombosed thoracic false lumen. The authors postulated that the cause of favourable remodelling in higher group can be due to the fact that extensive coverage increases the distance through which dissection lamellae are supported while simultaneously compressing false lumen. The small false lumen size reduces backflow and promotes thrombus formation. In addition, this also allows the exclusion of large number of re-entry tears throughout the length of the aorta [19].

Fig. 2.

Oblique sagittal computed tomography image (A) and volume rendered image (B) in a 50-year-old man with acute refractory chest pain showing type B aortic dissection with aneurysmal dilatation of false lumen until the juxta celiac level. Volume rendered image (C) after two stent grafts’ deployment with the covered portion distal to the left common carotid artery until just above the origin of celiac axis

Fig. 3.

Oblique sagittal computed tomography image volume rendered image (A) in a 48-year-old man with refractory hypertension depicts type B aortic dissection with aneurysmal false lumen. Digital subtraction image (B) after deployment of two stent grafts with the covered portion distal to the left common carotid artery until just above the origin of celiac axis

Limitations

Retrospective nature and bias regarding patient selection were the main limitations. Having one event in the number of patients treated has limited the power of our study. The sample size was limited for any definite conclusion. Hence, the results should be interpreted with caution.

Conclusions

Choice of the length of stent graft in TEVAR should be made by weighing the benefits of extensive stent graft coverage against the related risk of SCI. Based on our experience, we recommend longer aortic coverage (beyond 200 mm) in TBAD for better aortic remodelling, provided that the mean arterial pressure of > 90 mm Hg is maintained perioperatively, to avoid SCI.

Funding

None.

Declarations

Ethics approval

Obtained.

Informed consent

Obtained from all participants.

Conflict of interest

None.

Ethical statement

All procedures performed in studies involving human participants were in accordance with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.