Encephaloduroarteriosynangiosis Averts Stroke in Atherosclerotic Patients With Border-Zone Infarct: Post Hoc Analysis From a Performance Criterion Phase II Trial

Abstract

BACKGROUND

Intracranial atherosclerotic disease (ICAD) is one of the leading causes of stroke worldwide. Patients with ICAD who initially present with ischemia in border-zone areas and undergo intensive medical management (IMM) have the highest recurrence rates (37% at 1 yr) because of association with hemodynamic failure and poor collaterals.

OBJECTIVE

To evaluate the effect of encephaloduroarteriosynagiosis (EDAS) on stroke recurrence in patients with ICAD and border-zone stroke (BDZS) at presentation.

METHODS

A phase II clinical trial of EDAS revascularization for symptomatic ICAD failing medical management (EDAS Revascularization for Symptomatic Intracranial Atherosclerosis Steno-occlusive [ERSIAS]) was recently concluded. We analyze the outcomes of the subgroup of patients with BDZS at presentation treated with EDAS vs the previously reported Stenting versus Aggressive Medical Management for Preventing Recurrent stroke in Intracranial Stenosis (SAMMPRIS) IMM subgroup with BDZS at presentation.

RESULTS

Of 52 patients included in the ERSIAS trial, 35 presented with strokes at baseline, and 28 had a BDZ pattern, including 15 (54%) with exclusive BDZS and 13 (46%) with mixed patterns (BDZ plus other distribution). Three of the 28 (10.7%) had recurrent strokes up to a median follow-up of 24 months. The rate of recurrent stroke in ICAD patients with BDZS at presentation after EDAS was significantly lower than the rate reported in the SAMMPRIS IMM subgroup with BDZS at presentation (10.7% vs 37% P = .004, 95% CI = 0.037-0.27).

CONCLUSION

ICAD patients with BDZS at presentation have lower rates of recurrent stroke after EDAS surgery than those reported with medical management in the SAMMPRIS trial. These results support further investigation of EDAS in a randomized clinical trial.

Graphical Abstract

Graphical Abstract.

ABBREVIATIONS

ARR

absolute risk reduction

ASITN/SIR

American Society of Intervention and Therapeutic Neuroradiology/Society of Interventional Radiology

BDZS

border-zone stroke

COSS

Carotid Occlusion Surgery Study

CTP

computed tomographic perfusion

EDAS

encephaloduroarteriosynagiosis

ERSIAS

EDAS Revascularization for Symptomatic Intracranial Atherosclerosis Steno-occlusive

EC-IC

external carotid to internal carotid

ICA

internal carotid artery

ICAD

Intracranial atherosclerotic disease

IMM

intensive medical management

IQR

interquartile range

IRB

institutional review board

MCA

middle cerebral artery

MRI

magnetic resonance imaging

mRS

modified Rankin Scale

OPC

objective performance criterion

PC

performance criterion

QVSS

Questionnaire to Verify Stroke-Free Status

SAMMPRIS

Stenting versus Aggressive Medical Management for Preventing Recurrent stroke in Intracranial Stenosis

STA

superficial temporal artery

Intracranial atherosclerotic disease (ICAD) is one of the most common causes of stroke in the world and carries the worst prognosis compared to other stroke etiologies.^1,2 Medical management has improved and shown superiority to angioplasty, stenting, and direct bypass surgery in several randomized clinical trials such as Stenting versus Aggressive Medical Management for Preventing Recurrent stroke in Intracranial Stenosis (SAMMPRIS), Vitesse Intracranial Stent Study for Ischemic Stroke Therapy, Carotid Occlusion Surgery Study (COSS), and the external carotid to internal carotid (EC-IC) Bypass Study.^3-5

Despite the improvement in medical management outcomes, the rates of recurrent stroke in ICAD remain elevated, close to 15% for all the spectrum of patients with ICAD, and particularly high in certain populations.³ A recent subgroup analysis of SAMMPRIS patients treated with intensive medical management (IMM) showed that the risk of recurrent stroke for patients who presented with border-zone stroke (BDZS) as qualifying event was as high as 37%.⁶

EDAS is an indirect revascularization technique that has been used in patients with moyamoya disease, with success in both adults and children.⁷ Compared with direct bypass surgery, such as the superficial temporal artery (STA)-middle cerebral artery (MCA) bypass technique tested in previous trials,^5,8 the EDAS surgery approach has multiple advantages including (1) being less technically demanding; (2) having no requirement for temporary occlusion of intracranial vessels during surgery; (3) lacking a high pressure retrograde flow producing counterbalanced flow stasis in stenotic intracranial arteries, thus accelerating progression of native vessels to complete occlusion; and (4) resulting in gradual development of revascularization only where the brain demands it, thus deterring hyperperfusion with risk of edema and hemorrhage.

We report a subgroup analysis of the outcomes of patients with BDZS as qualifying event at presentation enrolled in the recently concluded EDAS Revascularization for Symptomatic Intracranial Atherosclerosis (ERSIAS) performance criterion (PC) phase II trial.

METHODS

ERSIAS PC Phase II Trial

ERSIAS PC phase II trial was a prospective, single-arm study of EDAS plus IMM vs an objective performance criterion (OPC) derived from SAMMPRIS patients treated with IMM alone. ERSIAS was designed as an intermediate phase, PC with a futility threshold design.⁹ The study was conducted in 2 institutions and the operations were done by 4 experienced vascular neurosurgeons.

ERSIAS entry criteria were similar to that of SAMMPRIS, except for a requirement to have poor collaterals in angiography and the inclusion of patients with occlusions. Patients ages 30 to 80 yr with symptoms of nondisabling stroke (modified Rankin Scale [mRS] 0-2) or transient ischemic attack (within 30 d of enrollment) unequivocally attributable to ICAD of more than 70% of the internal carotid artery (ICA) or MCA despite optimized medical management (at least 1 antiplatelet agent and optimal medical management of the primary and secondary stroke risk factors) underwent cerebral angiography.

We used the population from SAMMPRIS that underwent collateral evaluation using the American Society of Intervention and Therapeutic Neuroradiology/Society of Interventional Radiology (ASITN/SIR) collateral flow classification (grades 0-2) to derive the OPC against which the ERSIAS population was tested. All patients in ERSIAS underwent cerebral angiogram and cerebral perfusion assessment (computed tomographic perfusion (CTP), perfusion magnetic resonance imaging [MRI]) at enrollment and follow-up.

All patients received IMM, following the SAMMPRIS algorithm, except for 2 time-limited modifications: (1) systolic blood pressure management that was set to a minimum threshold at which patients were asymptomatic perioperatively, and (2) patients received only aspirin or clopidogrel alone 7 d before and 15 d after the procedure. Otherwise, the management was done exactly like in SAMMPRIS, including a lifestyle modification program.⁹

The primary clinical endpoint of the study was the same defined in SAMMPRIS: any stroke or death within 30 d after enrollment or any ischemic stroke in the territory of the qualifying artery. Clinical endpoints in ERSIAS were adjudicated by the stroke neurologists of the participating institutions, independently of the surgical and decision-making team and confirmed by an independent centralized adjudication panel, based on the review of the Questionnaire to Verify Stroke-Free Status (QVSS)¹⁰ and imaging confirmation of stroke.

Patient Selection

From the ERSIAS trial population, patients who had as the qualifying event a BDZS were included in this post hoc study. All patients had poor collaterals at baseline angiography (ASITN/SIR 0-2) and reduced perfusion in MRI or CTP (Figure 1). All patients enrolled in ERSIAS had signed informed consent, approved by the institutional review board (IRB) of each institution (IRB #12-000439, Pro00044887). This consent added permission to use the data collected in the study for future research.

FIGURE 1.

CONSORT flow diagram of patients analyzed in this study.

Stroke Pattern Allocation

Infarct patterns of the qualifying event were determined on patients’ most recent MRI before surgery, by 2 independent adjudicators blinded to the primary endpoint outcomes. The topography of the ischemic infarcts by vascular territories was determined with reference to published templates¹¹ previously validated in ICAD patients.^12,13

The infarct patterns were classified as follows: (1) pure border-zone pattern, one or more lesions in the internal (corona radiata or centrum semiovale) and/or the cortical border-zone region (between the MCA and the anterior cerebral artery or the MCA and the posterior cerebral artery); (2) perforator pattern, subcortical lesions in the distribution of perforating vessels that originate at the site of stenosis; (3) territorial pattern, one or more lesions located distal to the stenotic vessel (cortical, subcortical, or both) that are restricted to the territory supplied by a single intracranial artery; and (4) mixed pattern, a combination of any of the previous patterns.¹² For this study, BDZS definition included pure border-zone pattern and mixed patterns in which there was also a border-zone territory involvement (Figure 2).

FIGURE 2.

Schematic representation of the BDZ vascular territory. Represented in red are the territories corresponding to the BDZ areas, which includes the internal BDZ (corona radiata or centrum semiovale), and the cortical BDZ (between the middle cerebral artery and the anterior cerebral artery or the middle cerebral artery and the posterior cerebral artery).

Patient Follow-up

After enrollment in the ERSIAS trial, patients were followed up at 4 d, 30 d, 3 mo, and every 3 mo for 2 yr. All patients were scheduled for MRI every 3 mo to evaluate for silent strokes. During follow-up visits, patients were interviewed using the QVSS, and any positive response triggered a review of the occurrence of a clinical endpoint. ERSIAS mandated that any patient with symptoms for ≥1 h have brain MRI.⁹

Statistical Analysis

The hypothesis of this study was that the rate of recurrent stroke in the ERSIAS subgroup with BDZS at presentation (ERSIAS_BDZS) was different from the rate observed in the medically treated SAMMPRIS subgroup with poor collaterals and BDZS at presentation (SAMMPRIS_BDZS).

The hypotheses of the study can be expressed as

where H₀ and H_a are the null and alternative hypotheses, respectively, is the rate of recurrent stroke for the ERSIAS_BDZS group, and is the rate of recurrent stroke for the SAMMPRIS_BDZS, which is 0.37.

Demographic characteristics of the population are reported using median and interquartile range (IQR) for continuous variables and percentages for categorical variables. Inferential analysis was conducted using the one proportion Z-test of the statistical R software (© The R Foundation).

RESULTS

All 52 patients enrolled in ERSIAS had poor collaterals in angiography and reduced perfusion in CTP or perfusion MRI. Thirty-five patients presented with stroke as a qualifying event, defined in brain MRI or CT. Thirty-three patients had MRI before surgery and 2 patients had CT scans. In patients with preoperative MRIs, the predominant pattern was BDZS in 28 patients (85%). Four (12%) had pure territorial, and one (3%) had pure perforator strokes. Among patients with BDZS, 15 had pure BDZS and 13 had a mixed pattern (Figure 3).

FIGURE 3.

Examples of pure BDZ and mixed BDZ qualifying events at presentation. Brain MRI and DWI sequences for 2 patients with stroke qualifying events. A, Shown is an isolated acute DWI restriction in the corona radiata that corresponds to a pure BDZ stroke. B, Shown is DWI restriction in a mixed pattern with combined BDZ (arrow) and territorial (arrowhead) acute strokes. DWI, diffusion-weighted imaging.

Table 1 summarizes the demographic characteristics of patients with BDZS and non-BDZS. In the BDZS at presentation group, median age was 54.5 (IQR 43-67.5), 14 (50%) were females, the majority were white (53.6%), and non-Hispanic (89%). The most common location of the lesion was in the MCA (53.6%), and 22 (78.6%) patients presented stenosis with a median degree of 99% (IQR 90%-99%), whereas 6 patients (21.4%) had occlusions. The rates of hypertension and hyperlipidemia were very similar, but there were fewer patients with diabetes type II in the BDZS at presentation group.

TABLE 1.

Demographic Characteristics of Patients With BDZ and Non-BDZ Strokes

	BDZ		Non-BDZ
	Median	IQR	Median	IQR
Age	54.5	43-67.5	52	39.5-68.5
Sex	n	%	n	%
Female	14	50	3	60
Race
Asian	7	25	1	20
Black	6	21.4	1	20
White	15	53.6	3	60
Ethnicity
Hispanic	3	10.7	0	–
Non-Hispanic	25	89.3	5	100
Location
MCA	15	53.6	2	40
ICA	9	32.1	2	40
ICA and MCA	4	14.3	1	20
Hypertension (yes)	25	89.3	4	80
Hyperlipidemia (yes)	23	82.1	4	80
Diabetes type II (yes)	15	53.6	4	80
Stenosis (yes)	22	78.6	4	80
	Median	IQR	Median	IQR
Degreea	99	90-99	95	80.5-99.5

^aDegree of stenosis for only those with stenosis.

Demographic characteristics were comparable between the BDZS subgroups from ERSIAS and SAMMPRIS, with a slightly higher proportion of patients with factors known to be associated with higher risk of stroke recurrence (females, diabetes mellitus), and a worst degree of stenosis in the ERSIAS group (Table 2).

TABLE 2.

Demographic Characteristics of Patients With BDZ Strokes at Presentation From ERSIAS and SAMMPRIS

	ERSIAS		SAMMPRIS
	Median	IQR	Median	IQR
Age	54.5	43-67.5	55.1	46.9-64.5
Sex	n	%	n	%
Female	14	50	19	44.2
Race
Asian	7	25	0	–
Black	6	21.4	15	34.9
White	15	53.6	27	62.8
Native Hawaiian	0	–	1	2.3
Ethnicity
Hispanic	3	10.7	1	2.3
Non-Hispanic	25	89.3	42	97.7
Location
MCA	15	53.6	20	46.5
ICA	9	32.1	23	53.5
ICA and MCA	4	14.3	0	–
Hypertension (yes)	25	89.3	36	83.7
Hyperlipidemia (yes)	23	82.1	36	83.7
Diabetes type II (yes)	15	53.6	20	46.5
Stenosis (yes)	22	78.6	43	100
	Median	IQR	Median	IQR
Degreea	99	90-99	78	75-84

^aDegree of stenosis for patients without occlusion.

Stroke Recurrence

During a median follow-up of 24 mo (IQR = 11-41.5, range = 1-160 mo), stroke recurrences occurred in 3 patients at 2, 5, and 7 d after EDAS. All of them had presented with BDZS at presentation and had stenosis of the MCA. The recurrent strokes in those patients were 2 BDZS and 1 territorial. Although the mechanism of stroke cannot be certainly determined, the events likely occurred associated to hypotension in 2 cases related to excessive fluid extraction during dialysis in patients with chronic renal failure. The imaging studies performed during these patients’ strokes (computed tomography angiography, CTP, MRI, and magnetic resonance angiography) demonstrated no large vessel occlusions or progression of the stenotic segment narrowing. The initial NIHSS during the recurrence for those patients ranged from 4 to 6. The 3 patients were treated following stroke protocols at the participating institutions, but none had large vessel occlusions amenable for endovascular treatment. They were discharged to rehabilitation units and their mRS at last follow-up were 1, 2, and 4, indicating that only one had a permanent functional deficit.

There were no differences in their baseline ASITN/SIR, which were grades 0 to2. The strokes occurred before the study prescribed follow-up perfusion imaging.

Stroke Recurrence Compared With Patients in the IMM Arm of SAMMPRIS

The rate of recurrent stroke among patients in the ERSIAS study that had BDZS at presentation was 10.7% after EDAS revascularization. This proportion was significantly lower than the rate of recurrent stroke in patients with BDZS at presentation treated with IMM in SAMMPRIS, 37% (P = .004, 95% CI = 0.037-0.27). The ERSIAS patients faired significantly better than those in SAMMPRIS in IMM with an absolute risk reduction (ARR) of 26.3%. In addition, there was only one disabling stroke in the patients treated with EDAS.

Surgical Complications and Adverse Events

Only 2 patients of the complete ERSIAS cohort (3.8%, 2/52) had surgery-related complications (wound dehiscences that required debridement and closure). There were no intracranial hemorrhages or postoperative cerebral edema in any cases.

Imaging Follow-up

Twenty-two patients (22/28) completed angiographic follow-up at 6 mo from surgery, and 17(77%) had improvement in at least 1 point in their ASITN/SIR grades. None of the patients with stenosis who completed imaging follow-up developed progression of stenosis after surgery.

Anonymized data not published within this article will be made available by request from any qualified investigator. Researchers interested in working with the data should contact the corresponding author.

DISCUSSION

In this subgroup analysis of patients with BDZS at presentation of the ERSIAS population, EDAS averted stroke recurrences to a greater proportion than IMM in the SAMMPRIS trial.

After the results of the SAMMPRIS trial, IMM has become the standard treatment for patients with ICAD. Other promising alternatives of treatment, such as submaximal angioplasty, are currently under study but have not been validated in randomized clinical trials.¹⁴ IMM is very effective in mitigating progression of endothelial damage, plaque growth and stability, perforator vessel occlusion, in situ thrombosis, and artery-to-artery embolisms.¹⁵ However, IMM does not correct hemodynamic aspects, which are critical determinants of the high recurrence rates of stroke in ICAD patients.^6,16,17

A previous post hoc analysis from the medical arm of SAMMPRIS analyzed a subgroup of patients with BDZS and poor collaterals.⁶ They reported that more than half of the patients in the medical arm group of SAMMPRIS who have poor collaterals presented with BDZS (52.4%) as qualifying events.

Key Results

Similarly, the rate of BDZS as qualifying event in the ERSIAS PC trial was 53.8%. Patients with BDZS at presentation had the highest rate of recurrent stroke in both studies, 37% in SAMMPRIS and 10.7% in ERSIAS, which emphasize the importance of hemodynamic factors in stroke recurrence. However, the rate was significantly lower in the ERSIAS patients.

These findings suggest that a treatment strategy directed to correct those hemodynamic factors with EDAS in conjunction with IMM may play an important role in improving the outcomes of patients with ICAD beyond medical management alone.

In this study, we have observed that all patients who had stroke recurrences had stenoses rather than occlusions. Patients with occlusions may develop conditions that stimulate even more the growth of collaterals from neighbor territories and the donor EDAS arteries, likely related to stimulation of angiogenesis and arteriogenesis.^18,19 However, even in the absence of that additional stimulation in patients with stenosis and occlusions, EDAS reduced the risk of recurrent stroke with an ARR of 26.3% compared with IMM.

The EDAS surgery utilizes branches from the external carotid circulation, mainly the STA and the middle meningeal artery to procure the new vessels that end up connecting with the intracranial vessels. This process does not occur immediately after surgery, although signs of neovascularization in catheter angiography have been reported as early as 7 d postop.²⁰

This window of time may contribute to explain the time distribution of events early on after surgery. As a response to this pattern of events, some groups have suggested the use of direct bypass or a combination of direct-indirect bypass to overcome the delay in spontaneous collateral formation.^21,22 However, stroke recurrences in patients with symptomatic ICAD tend to occur early on after their initial presentation, independently of the form of treatment.^15,23 This has been extensively discussed in the literature and has affected patients under medical management, as well as those who underwent angioplasty and stenting²⁴ and bypass.^5,8 Our perioperative risk of stroke in the ERSIAS PC trial was 7.7%,⁹ in the subgroup analysis for patients with BDZS at presentation reported here was 10.7%, which compares favorably with the reported perioperative risk of stroke with direct bypass of 15% and 12% in COSS and the EC-IC bypass trial, respectively.^5,8 This points to the fact that the rapid correction of flow induced by direct bypass may not necessarily completely avert early ICAD stroke recurrence and overcome the risks associated with the direct bypass shortcomings, including the need for temporary occlusion of vessels on territories at critical risk of ischemia, and post-bypass hyperperfusion with the associated risk of edema and hemorrhage. Awad et al⁸ analyzed the effects of EC-IC direct bypass in intracranial stenotic lesions and found that the direct bypass can modify the hemodynamic parameters across stenotic lesions, predisposing them in some cases to worsening, contributing to the higher morbidity observed with EC-IC bypass surgery in the setting of arterial stenosis.

In the ERSIAS PC trial, to compensate for the expected interval in vessel formation, patients were managed with IMM, which has been proven in several randomized controlled trials to be superior to the direct bypass. Thus, ERSIAS patients received the best standard of care while the EDAS technique was being tested. The combined approach rendered a significant reduction in the risk of recurrent stroke, particularly significant in the patients at highest risk, those with BDZS and hemodynamic failure at presentation.

Interpretation

Our findings support the concept that a targeted intervention to modify sources of collateral circulation, such as EDAS, in association with IMM can produce long-term sufficient hemodynamic modifications to reduce the risk of stroke in patients with ICAD, playing a fundamental role in improving outcomes of these patients beyond medical management alone.

Limitations

This is a post hoc analysis of a subgroup of patients from the ERSIAS PC phase II trial, and as such, no claims of efficacy for the EDAS revascularization can be made. The nature of the study is subject to bias as the patients were not randomized.

CONCLUSION

This post hoc analysis highlights the critical role of hemodynamic factors in the stroke recurrence of patients with ICAD. Our data indicate that the EDAS technique should be further studied in randomized clinical trials to confirm the signs of efficacy observed in the ERSIAS PC trial and in the subgroup of patients with BDZS.

Funding

This research was supported by the National Institute of Neurological Disorders and Stroke of the NIH under award no. K23NS079477. The content is solely the responsibility of the authors.

Disclosures

The authors have no personal, financial, or institutional interest in any of the drugs, materials, or devices described in this article.

Notes

This study has been previously presented as an E-poster at the CNS Annual Meeting, Congress of Neurological Surgeons, on October 19-23, 2019, in San Francisco, California, and as an oral presentation at the International Stroke Conference, AHA, on February 19-21, 2020, in Los Angeles, California.

COMMENT

In this subgroup analysis of the phase II study ERSIAS (EDAS Revascularization for Symptomatic Intracranial Arterial Stenosis), the authors compared the rate of stroke and death at 30 days in patients presenting with borderzone infarcts who were treated with EDAS to the historical cohort from the SAMMPRIS trial treated with maximal medical therapy (Wabnitz et al, 2019). Of the 28 patients enrolled in ERSIAS who presented with borderzone infarcts, only three (11%) had recurrent stroke within two years. In comparison, 37% of the historical cohort had a recurrent stroke. While these are unmatched populations—one a prospective cohort and one a retrospective subgroup from a clinical trial—the results of this study support the need for a randomized controlled clinical trial comparing EDAS to medical management alone in patients presenting with borderzone infarcts.

A non-futility trial design like this has many limitations. Most importantly, one cannot claim efficacy based on these results. The patient populations are not matched and the medical intervention protocol differs between the two cohorts. What this study does accomplish, however, is justification for a randomized controlled clinical trial on revascularization for prevention of recurrent stroke in a subpopulation of ICAD patients. While much of the data on interventions in ICAD has not shown a significant benefit, narrowing our focus to identify that population in which it might actually help is clinically important. This study uses a cost-effective and efficient study design to build the evidence needed to support this ongoing work.

Kate T. Carroll

Louis J. Kim

Seattle, Washington