Endovascular thrombectomy for large vessel occlusion acute ischemic stroke after cardiac surgery

Abstract

Introduction

Acute ischemic stroke (AIS) can be a catastrophic complication of cardiac surgery previously without effective treatment. Endovascular thrombectomy (EVT) is a potentially life‐saving intervention. We examined patients at our institution who had EVT to treat AIS post cardiac surgery.

Methods

We retrospectively reviewed a stroke database from January 1, 2016 to October 31, 2021 to identify patients who had undergone EVT to treat AIS following cardiac surgery. Demographic data, operation type, stroke severity, imaging features, management and outcomes (mortality and modified Rankin Score (mRS)) were assessed.

Results

Of 5022 consecutive patients with AIS, 870 underwent EVT. Seven patients (0.8%) had EVT following cardiac surgery. Operations varied: two coronary artery bypass grafting (CABG), two transcatheter AVR, one redo surgical aortic valve replacement (AVR), one mitral valve repair and one patient with combined aortic and mitral valve replacements and CABG. Meantime postsurgery to stroke symptoms onset was 3 days (range 0–9 days). Median NIHSS was 26 (range 10–32). Five patients had middle cerebral artery occlusion and two internal carotid artery (n = 2). Median time between onset of symptoms and recanalization was 157 min (range 97–263). Two patients received Intra‐arterial Thrombolysis. All patients survived and were discharged to another hospital (n = 3), home (n = 2), or rehabilitation facility (n = 2). Median 3‐month mRS was 3 (range 0–6).

Conclusion

We report the largest case series of EVT after cardiac surgery. EVT can be associated with excellent outcomes in these patients. Close neurological monitoring postoperatively to identify patients who may benefit from intervention is key.

1. INTRODUCTION

Prevalence of stroke in the postoperative period after cardiac surgery is estimated to be between 1% and 2%. ¹ , ² , ³ Stroke after cardiac surgery is associated with high mortality and low quality of life. ¹ Large vessel occlusion (LVO) acute ischemic stroke (AIS) is a catastrophic complication of cardiac surgery. As thrombolysis is usually contraindicated in these situations due to significant risk of intrathoracic bleeding, historically no effective treatment existed. Recently, endovascular thrombectomy (EVT) has emerged as a potential life‐saving intervention for these patients, and can be the intervention of choice to treat LVO AIS at centers where resources and specialist staff are available. Catheter‐based Thrombectomy devices are used to manually extract the intra‐arterial clot, with high‐level evidence demonstrating the superiority of this approach over best medical therapy alone in AIS patients with favorable perfusion imaging who present within 24 h of onset of symptoms. ⁴ , ⁵

Successful EVT may significantly reduce postoperative morbidity and mortality following post‐cardiac surgery AIS. While EVT presents a potentially life‐saving intervention, limited information exists in the peer‐reviewed literature on its safety and outcome in this setting. To help inform the management of cardiac patients worldwide, we examined a consecutive series of patients from our institution who underwent EVT to treat AIS after cardiac surgery.

2. METHODS

We performed a retrospective review of a prospectively collected database in a tertiary referral hospital located in Adelaide, Australia. Ethical authorization was obtained from the Central Adelaide Local Health Network Human Research Ethics Committee (Reference Number 14647). Patients who had undergone EVT to treat LVO AIS following cardiac surgery between January 1, 2016 to October 31, 2021 were included. Our definition of LVO occlusion was demonstratable perfusion deficit on radiography of either the terminal internal carotid artery (ICA), proximal middle cerebral artery (MCA; M1 or M2), or basilar artery. The decision to proceed with endovascular clot retrieval was a joint clinical decision made between a Stroke Neurologist and Interventional Neuroradiologist Proceduralist incorporating clinical symptoms using National Institutes of Health Stroke Scale (NIHSS) score, patient characteristics and identifiable clot burden on imaging, demographic data, operation type, severity of stroke and multimodal neuroimaging features (perfusion imaging processed by MiSTAR (Apollo, Inc). The decision to use intra‐arterial thrombolysis (IAT) was made in conjunction with the cardiac interventional and surgical teams. In our series, all cases of open cardiac surgery were deemed to high risk for fatal bleeding, and so IAT was contraindicated in these cases. Management and outcomes (24 h NIHSS, mortality, and modified Rankin Scale (mRS) scores at 3 months) were extracted and analyzed.

EVT was usually performed through cannulation of the femoral artery. Under angiographic guidance. suction thrombectomy, stent‐retriever clot removal, non‐retrievable stenting and/or intra‐arterial thrombolysis were performed as appropriate. Reperfusion efficacy was graded by the modified Thrombolysis in Cerebral Infarction (TICI) 2c scale (mTICI 2c). ⁶ Following‐up imaging to quantify infarct volume, hemorrhagic transformation, and durable recanalization was routinely performed.

3. RESULTS

Of 5022 consecutive patients with AIS, 870 underwent EVT. Seven (0.8%) patients (5 maless) had EVT following cardiac surgery and were included. Complete information on these patients can be seen in Table 1, along with accompanying imaging (Figures 1, 2, 3, 4, 5, 6, 7, 8, 9, 10). Cardiac interventions varied: two coronary artery bypass grafting (CABG), two transcatheter AVR (TAVR), one minimally invasive mitral valve repair (MiMVr), and one patient with combined AVR and mitral valve replacement (MVR) and CABG. Four patients were noted to have atrial fibrillation (AF) postoperatively, of which three were known to have pre‐existing AF. Left ventricular ejection fraction was preserved in all patients. Median duration post‐surgery to stroke symptoms onset was 3 days (range 0–9). Out of the cohort, two patients received Intra‐arterial Thrombolysis in addition to EVT; both of these patients initially underwent TAVR. Median NIHSS was 26 (range 10–32), with a mean of 23. Five patients had middle cerebral artery occlusion and two internal carotid artery (n = 2). Median time between onset of symptoms and recanalization was 157 min (range 97–263). TICI score was 0 in 3 patients, 2B in one patient, 2 C in two patients, and 3 in one patient. Median NIHSS at 24 h was 5 (range 0–29) with a mean of 11. All patients survived and were transferred downstream to another hospital (n = 3), discharged home (n = 2) or to a rehabilitation facility (n = 2). One of the patients who was discharged to another hospital subsequently died from hospital‐acquired pneumonia. Median 3‐month mRS score was 3 (range 0–6). There were no direct procedural or bleeding complications from EVT in any of the patients described in our series. Antithrombotic therapy after EVT was instituted as appropriate on the basis of stroke etiology. Two patients had dual antiplatelet therapy, one had single antiplatelet therapy, three had therapeutic anticoagulation and one had therapeutic anticoagulation with single antiplatelet therapy. We present two cases to highlight good outcomes from EVT in this setting, and one case which resulted in subsequent death of a patient after EVT.

Table 1.

Summary of case series

Case	Gender	Age	Surgery	Time from surgery to stroke (days)	Prior cardiac surgery	Preop AF	Postop AF	Type 2 diabetes	Known ICA disease	Preop mRS	NIHSS at Code Stroke	Vessel(s) occluded	Stroke etiology	Time from LKW to recanalization (mins)	rCBF < 30% Core (ml)	DT > 3 s lesion volume (ml)	Recanalisation	24 h NIHSS	Medical therapy after EVT	Disposition	3‐month mRS
#1	M	54	CABG	1	No	No	No	No	Yes	0	20	Left ICA	Atherosclerosis	139	92	383	TICI 3	0	Aspirin and Ticagrelor	Other hospital	1
#2	M	74	TAVR	0	No	No	No	No	No	0	32	Left M1, Right Basilar Tip Embolus	Cardioembolic	97	14	80	TICI 2 C	0	Aspirin and Clopidogrel	Home	0
#3	M	66	CABG	5	No	Yes	Yes	Yes	No	1	29	Right ICA	Cardioembolic	225	0	334	TICI 0	29	Aspirin	Other hospital	6
#4	M	64	Redo AVR	3	Yes	No	No	No	No	0	26	Left ICA	Cardioembolic	210	2	69	TICI 0	3	Warfarin	Home	3
#5	F	83	MiMVr	6	No	Yes	Yes	Yes	No	0	27	Left M1	Cardioembolic	157	3	73	TICI 0	23	Apixaban	Rehab	4
#6	F	83	AVR + MVR + CABG	9	No	No	Yes	No	Yes	0	20	Left M1	Cardioembolic	125	6	76	TICI 2B	18	Aspirin (with addition of Apixaban on Day 7)	Other hospital	3
#7	M	88	TAVR	1	No	Yes	Yes	No	No	1	10	Right M2	Cardioembolic	263	1	56	TICI 2 C	5	Apixaban	Rehab	2

Abbreviations: AF, atrial fibrillation; CABG, coronary artery bypass grafting; LKW, last known well; mRS, modified Rankin Scale; NIHSS, National Institutes of Health Stroke Scale.

Figure 1.

Digital subtraction cerebral angiography showing an occluded left ICA (Patient #1). ICA, internal carotid artery.

Figure 2.

Computed tomography perfusion showing favorable characteristics and penumbra map with irreversible brain ischemia volume (cerebral blood volume<30%) of 92 ml and total ischemic brain volume (delay time>3 s) of 471 ml (Patient #1)

Figure 3.

Deployment of proximal left ICA stent with TICI 3 reperfusion (Patient #1). ICA, internal carotid artery.

Figure 4.

CTA showing left M1 occlusion (Patient #2). CTA, computed tomography angiography.

Figure 5.

Computed tomography perfusion showed favorable imaging characteristics with irreversible brain ischemic volume (cerebral blood volume<30%) of 14 ml and total ischemic brain volume (delay time>3 s) of 94 ml (Patient #2).

Figure 6.

Postprocedural CTA showing TICI 2c reperfusion (Patient #2). CTA, computed tomography angiography.

Figure 7.

MRI at 24 h showing multifocal small left MCA territory and small right medial thalamic infarcts (Patient #2). MCA, middle cerebral artery; MRI, magnetic resonance imaging

Figure 8.

Digital subtraction angiography showing an occluded right ICA (Patient #3). ICA, internal carotid artery.

Figure 9.

Computed tomography perfusion demonstrating favorable imaging characteristics with irreversible brain ischemia volume (cerebral blood volume<30%) of 36 ml and predicted total ischemic brain tissue volume (delay time>3 s) of 320 ml (Patient #3).

Figure 10.

Computed tomography brain showed a large right hemispheric infarction complicated by severe cerebral edema with subfalcine and uncal herniation (Patient #3).

3.1. Case 1

A 54‐year‐old man with non‐ST elevation Myocardial Infarction (NSTEMI) and newly diagnosed triple vessel disease was admitted electively for CABG surgery. He had a pre‐existing left internal carotid artery stenosis that he was on aspirin monotherapy to treat. Vascular imaging also demonstrated the presence of an azygous left anterior cerebral artery. Upon weaning of sedation in ICU, he exhibited a complete left MCA syndrome with initial NIHSS of 20 at 54 min post‐last known well (LKW) time. Following the activation of a Code Stroke, initial CT cerebral angiography showed an occluded left ICA and CT head perfusion demonstrated a large severe perfusion abnormality of the left ICA territory and the right ACA territory with predicted irreversible brain ischemia volume (Cerebral Blood Volume <30%) of 92 ml and total ischemic brain tissue volume (Delay Time>3 s) of 471 ml (Figures 1 and 2). Time from LKW to groin puncture was 2 h and 19 min. TICI 3 reperfusion was achieved within 30 min with a proximal left ICA stent deployed (Figure 3). CT angiography (CTA) at 6 h and 15 min post‐procedurally showed a patent left ICA stent. Due to early recanalization, despite the large volume of severely ischemic tissue, his NIHSS at 1 week was 0 and he was discharged home. His mRS score at 3 months was 1.

3.2. Case 2

A 74‐year‐old man with severe aortic stenosis was admitted for TAVR. Post‐TAVR, he was difficult to rouse with a complete left MCA syndrome (NIHSS 32) at 1 h and 37 min post‐LKW time. CT cerebral angiography showed left M1 occlusion (Figure 4). CT head perfusion showed favorable imaging characteristics with predicted irreversible brain ischemia volume (Cerebral Blood Volume<30%) of 14 ml and total ischemic brain tissue volume (Delay Time>3 s) of 94 ml (Figure 5). Time from LKW to groin puncture was 3 h. After three passes with suction thrombectomy, TICI 2c reperfusion was achieved within 30 min (Figure 6). Pathological examination of the embolus showed thrombus. MRI at 24 h showed multifocal small left MCA territory and small right medial thalamic infarcts. (Figure 7). NIHSS at 24 h improved to 0. He was discharged home and his mRS at 3 months was 1.

3.3. Case 3

A 66‐year‐old man with abnormal screening ECG and newly diagnosed triple vessel disease was admitted electively for CABG surgery. 5 days post‐CABG, he experienced left hemiparesis with an initial NIHSS of 29. Cerebral angiography demonstrated right ICA occlusion (Figure 8). CT head perfusion demonstrated favorable imaging characteristics with predicted total ischemic brain tissue volume (Delay Time>3 s) of 320 ml and irreversible brain ischemia volume (cerebral blood volume<30%) of 36 ml. (Figure 9). Time from LKW to groin puncture was 4 h and 16 min. After 8 passes with suction thrombectomy and rescue stenting with a stent retriever, TICI 0 reperfusion was achieved within 69 min. Pathological examination of the embolus showed thrombus. NIHSS at 24 h was 18. He passed away on Day 7 post‐CABG from large right hemispheric infarction complicated by severe cerebral edema with subfalcine and uncal herniation (Figure 10).

4. DISCUSSION

Our consecutive patient cohort series demonstrates that EVT to treat LVO for AIS after cardiac surgery can be associated with excellent functional outcome, especially with early excellent reperfusion. Two of our seven cases were left free of disability, however despite EVT, some patients were still left with significant neurological deficits. To our knowledge, this is the largest case series in the literature and supports the significant benefits in certain patients of EVT for AIS after cardiac surgery. ⁷ , ⁸ , ⁹

EVT is indicated in patients with large vessel occlusion (ICA, M1, proximal M2) within 24 h of symptom onset. Meticulous patient selection for EVT based on advanced multimodal CT perfusion imaging to determine CT perfusion mismatch ratio, presence of established cerebral infarction on plain CT brain, neurological deficits at time of thrombectomy and baseline functional status is required to achieve a better NIHSS score and functional independence at 90 days. ⁵ In previous studies, patients with favorable CT perfusion imaging characteristics (such as high perfusion mismatch ratio with low CBV < 30% or “core” infarct volume and lack of establish cerebral infarction on plain CT brain), low baseline NIHSS score and low baseline modified Rankin Score were associated with higher rates of successful reperfusion and good functional outcome at 3 months. ¹⁰ In our case series, shorter duration from LKW time to EVT being performed and favorable perfusion imaging characteristics correlated with good functional outcome at 3 months.

Previous studies demonstrated that higher mTICI scores were associated with better functional outcome at 90 days. ¹¹ In our study, case 4 had no reperfusion post‐EVT but achieved a 23‐point improvement in NIHSS. This could be due to spontaneous thrombus dissolution which may possibly be enhanced by perioperative antiplatelet and heparin therapy. Conversely, Case 5 had a mild 2‐point improvement in NIHSS despite TICI2B reperfusion. This may be due to rapid progression of cerebral infarction between time of initial CT perfusion imaging and successful endovascular clot retrieval. 3 out of 7 patients (43%) in our study had TICI 0 revascularization despite multiple attempts at EVT. This was inconsistent with previous studies demonstrating higher efficacy of EVT in cardioembolic LVO (85.6%). ¹² This may be due to operator‐dependent factors such as EVT technique as well as patient‐dependent factors such as thrombus composition.

In the immediate postoperative period following cardiac surgery patients are often heavily sedated, precluding accurate neurological assessment and identification of symptoms and signs of cerebral ischemia. Elevated risk of stroke is compounded by delayed identification. ¹³ Following uncomplicated cardiac surgery, it is recommended that patients be extubated within 6 h where possible. ¹⁴ Early and regular neurological assessment by early weaning of sedation after surgery, and early activation of code stroke with neurological deterioration can help identify stroke symptoms at a time when deficits are still reversible. Apart from some cases of transcatheter valve intervention, systemic thrombolysis is not a standard treatment option when strokes occur in the acute postoperative period due to the risk of fatal hemorrhage from high‐dose fibrinolysis. ¹⁵ , ¹⁶ Risk of hemorrhagic transformation in cases of large infarct volumes renders the use of antiplatelet and anticoagulation especially dangerous. EVT can potentially reduce the risk of disabling stroke associated with more cautious introduction of more potent antithrombotic medications. Postoperative AF (POAF) occurs frequently after coronary and valvular procedures, ¹⁷ and is associated with a higher incidence of stroke. ¹⁸ , ¹⁹ Major bleeding rates after tPA in the early postoperative period after cardiac surgery have been found to be as high as 23%. ¹⁶ In our series, one patient developed new‐onset POAF. Evidence is emerging for posterior left pericardiotomy to be considered for all cardiac surgical procedures to reduce the risk of POAF. ²⁰

Patients with prior history of cerebrovascular events or known vascular disease are at particularly high risk of postoperative stroke. Two patients in our series had pre‐existing intracranial atherosclerotic disease, of which one had a previous carotid artery revascularization. International guidelines recommend carotid duplex ultrasound in patients with previous stroke who have had an event within 6 months or have significant risk factors. ²¹ In select patients with severe cerebrovascular atherosclerosis, off‐pump CABG should be considered as the preferred approach to reduce risk of perioperative stroke. ²² , ²³ By avoiding the risks associated with aortic cannulation and cross‐clamping, off‐pump CABG has been demonstrated to improve neurological outcomes in patients with a history of previous cerebrovascular events. ²⁴ Similarly, there is emerging evidence for anaortic, total arterial techniques that avoid cardiopulmonary bypass and all manipulation of the ascending aorta. ²⁵ , ²⁶ , ²⁷ , ²⁸ The presented cohort demonstrates that EVT may be a safe intervention to treat AIS in this patient population. TAVR is demonstrated to have a significantly lower risk of early postoperative stroke within 30 days compared with Surgical AVR. ²⁹ Interestingly, the two patients in the series who underwent TAVR and then had EVT for AIS also had thrombolysis with Tenecteplase. Given that TAVR is a less invasive approach to open surgical AVR, the use of thrombolysis may be considered, although the supporting literature is limited to only case reports. ³⁰ , ³¹ , ³² Randomized controlled trials are warranted to investigate outcomes of thrombolysis and endovascular thrombectomy for AIS after TAVR.

Our study has several limitations. We did not prospectively identify the number of patients who suffer an interoperative or perioperative stroke that is recognized too late for EVT to be feasible. Although ours represents the largest case series in the literature, seven cases are a small number and statistically causative relationships cannot be demonstrated due to the inadequate power. In the sedate patient postoperatively, accurate neurological assessment can often be difficult to ascertain. Accordingly, it may be difficult to precisely assess the time of onset of stroke symptoms and initial NIHSS. Retrospective design prevented control over exposure and outcome assessment, thus making it difficult to assess temporal relationships. ³³ , ³⁴ , ³⁵ , ³⁶ , ³⁷ The data recorded relied on healthcare staff accuracy in collection during a potentially stressful situation while assessing critically unwell stroke patients. Given that patients were transferred from multiple hospitals performing cardiac surgery, we were unable to obtain intraoperative cardiac perfusion data including aortic cross‐clamp and cardiopulmonary bypass times.

Our case series illustrates how EVT can be associated with excellent outcomes in these patients and close neurological monitoring postoperatively to identify patients who may benefit from the intervention is imperative. However, some patients still incurred severe neurological deficit and even mortality following their postoperative stroke, highlighting that further work is required to attempt to improve outcomes in this patient cohort.

5. CONCLUSION

This study comprises the largest case series of EVT for AIS after cardiac surgery, and demonstrates that excellent outcomes can be achieved. Close neurological monitoring during the postoperative period is imperative to identify stroke patients early and minimize infarct growth before reperfusion. Further research involving larger cohorts is warranted.

CONFLICT OF INTEREST

The authors declare no conflict of interest.

Gupta AK, Sabab A, Goh R, et al. Endovascular thrombectomy for large vessel occlusion acute ischaemic stroke after cardiac surgery. J Card Surg. 2022;37:4562‐4570. 10.1111/jocs.17082