Microemboli After Successful Thrombectomy Do Not Affect Outcome But Predict New Embolic Events

Abstract

Background and Purpose

We aimed to determine if microemboli after endovascular thrombectomy correlate with unfavorable outcomes despite successful recanalization.

Methods

This is a prospective multicenter study of consecutive patients with ischemic stroke and occlusion of anterior circulation vessels (terminal internal carotid or M1/M2 segments of middle cerebral artery) after successful thrombectomy (modified Treatment In Cerebral Ischemia grades 2b-3). Microembolic signals (MES) were assessed by 30-minutes of transcranial Doppler (TCD) monitoring within 72 hours of the last-seen-well time. Major outcomes included modified Rankin scale at 90 days and infarct volume on head CT at 24 hours. We also assessed early outcomes based on NIHSS variation and recurrence of stroke, TIA, or systemic embolism within 90 days.

Results

Among 111 patients, MES were detected in 43 (39%), with a median rate of 4 counts/hour (interquartile range 2 – 12). The occurrence of MES was not associated with a significant difference in mRS (ordinal shift analysis, adjusted odds ratio 1.06 (95% CI 0.48 – 2.34), p = 0.85) nor in functional independence (mRS 0 – 2: adjusted odds ratio = 0.52 (95% CI 0.19 – 1.39), p = 0.19). Patients with and without MES had similar infarct volumes (adjusted beta = 11.2 (95% CI – 46.6 – +22.9), p = 0.51) on 24-hour CT. MES did predict new embolic events (adjusted Cox hazard ratio 6.78 (CI 95% 1.63 – 27.8), p=0.01).

Conclusions

Microembolic signals detected by TCD following endovascular treatment of anterior circulation occlusions do not predict clinical or radiological outcome. However, such emboli are an independent marker of recurrent embolic events within 90 days.

Introduction

Endovascular thrombectomy has proven to be the most effective treatment for patients with acute ischemic stroke with large vessel occlusion in the anterior circulation.¹ A pooled analysis by age, neurological severity, time from symptom-onset, and pretreatment infarct area showed consistent benefits in properly selected patients.¹ However, approximately half of treated patients fail to show signs of early recovery from the initial neurological deficits, and roughly half fail to reach functional independence within 90 days.¹ Poor collateralization has been shown to correlate with poor outcome, but this does not explain all outcome variability.² To improve outcomes it will be important to identify other causes underlying the lack of benefit despite successful recanalization. One possibility is that continuous embolization to downstream microvasculature may occur as a result of vessel wall injury³ caused by the endovascular procedure, or repeated cerebral embolization may originate from the same source that caused the index stroke, such as upstream atherosclerotic lesions^4–6 or the heart,^7–9 and may result in a larger final infarct volume or even in new clinical events.

Transcranial Doppler ultrasound (TCD) is routinely used to assess the burden of embolization by detection and quantification of microembolic signals (MES). Emboli-detection TCD may be used to select patients at high risk of cerebrovascular events across various stroke etiologies.^10–12 Our objective in this study was to assess the prognostic value of MES after successful endovascular thrombectomy of acutely occluded large anterior circulation vessels. We hypothesized that MES would predict worse outcomes.

Methods

The data and SPSS syntax code that support the findings of this study are available from the corresponding author upon reasonable request (Prof. Pedro Castro).

Patients and participating centers

This is a prospective multicenter study with inclusion of consecutive ischemic stroke patients who underwent endovascular thrombectomy with successful recanalization. Patients were recruited from the Brigham and Women’s Hospital in Boston, MA, USA and the Centro Hospitalar Universitário São João, Porto, Portugal, between September 1, 2017 and November 15, 2018. Eligible patients had a large vessel occlusion of the proximal middle cerebral artery (M1 or M2 segment) and/or of the internal carotid artery (ICA) terminus, were ≥18 years of age, and had achieved grade 2b-3 according to the modified Thrombolysis in Cerebral Infarction scale (mTICI)¹³ after thrombectomy. Patients were excluded if they were unable to tolerate TCD or had insufficient temporal bone windows.

Clinical, laboratory and radiological assessment

Demographic data and medical history, including vascular risk factors, medications, and previous cardiovascular disease were recorded. Blood pressure (BP), serum glucose, and cholesterol levels were recorded at admission. National Institutes of Health Stroke Scale (NIHSS) scores were obtained at baseline and after 24 hours from recanalization. Stroke type was classified by the Trial of Org 10172 in Acute Stroke Treatment (TOAST) scale.

A member of the stroke team recorded the Alberta Stroke Program Early Computed Tomography Score (ASPECTS) for the admission noncontrast computed tomography (CT)¹⁴ Initial CT angiography (CTA) was inspected to determine the location of vessel occlusion and collaterals grade.¹⁵ The presence of hemorrhage¹⁶ and infarct volumes calculated by the ABC/2 method were based on a head CT done approximately 24 hours after the procedure.¹⁷

Microembolic signals detection

The presence and rate of MES were assessed by 30 minutes of TCD monitoring (Doppler Box X, DWL, Sipplingen, Germany and San Juan Capistrano, CA) within 72 hours of thrombectomy. Bilateral M1 segments were insonated at a single depth with 2-MHz transducers secured by a probe-holder with sample volume of 4.5 mm, low gain, and detection threshold of 9 dB.¹⁸ MES was defined according to international consensus¹⁸ as a signal with 1) duration <300 ms, 2) amplitude >3 dB with respect to underlying flow signal, 3) unidirectionality, and 4) typical sound (“snap” and/or “chirp”). Records were analyzed by a single experienced observer (PC). Patients with ≥1 MES detected were assigned to the MES-positive group.

Clinical endpoints

The major endpoints were functional outcome at 90 days, measured by modified Rankin scale score (mRS), and infarct volume on a 24-hour CT. We compared the functional outcome across the whole scale (ordinal shift analysis) and by dichotomization of the mRS: 0–2 (independent) versus 3–6 (dependent or dead). An additional endpoint characterized the initial response to mechanical thrombectomy: early neurological recovery at 24h, defined as a decline in NIHSS score of at least 8 points or a score of 0 or 1.

Finally, we compared ischemic stroke, TIA, or systemic embolism within 90 days. To detect these events, we interviewed patients and/or their caregivers and examined electronic medical records. Events were defined by clinical history, and all strokes were confirmed by head CT. Outcome ascertainment was blinded to the MES results. We received approval by the ethics committees of both institutions. Written informed consent was obtained from all participants/guardians.

Statistical Analysis

Normality of continuous variables was inferred by the Shapiro-Wilk test. The baseline characteristics of patients in the MES-positive and MES-negative groups were compared using chi-square/Fisher exact tests for categorical variables and Student’s t/Mann-Whitney tests for continuous variables as appropriate. Bonferroni correction was used when categorical variables had more than one class. Clinical outcomes in the MES-positive and MES-negative groups were compared using mRS scores at 90 days with ordinal shift analysis modeling across all levels of the scale after verification of conformity to the assumption of a common proportional odds. Logistic regression was used to generate the odds ratios (OR) and 95% confidence intervals (CI) for 1) functional independence (mRS 0 to 2 versus 3 to 6) at 90 days and early neurological recovery at 24 hours. Infarct volume at 24 hours was compared with general mixed linear models to estimate standardized B-coefficients and their 95% CI.

We adjusted analyses for all outcomes with two multivariate models. In the first, we adjusted to the independent variables associated with outcome in univariate analysis (Supplemental Table I). This included age, NIHSS, LSW-to-recanalization time, ASPECTS, and mTICI grade. In the second, we adjusted for the variables that were unbalanced between MES groups (Table 1 with p<0.10) by propensity score matching approach: we computed a propensity of being associated with MES by incorporating age, smoking, NIHSS, cholesterol, TOAST, ASPECTS, and LSW-to-recanalization time in a single propensity score between 0 and 1.

Table 1.

Demographic, clinical and radiological characteristics of all patients and the differences between MES-positive and MES-negative subgroups.

	All (n = 111)	MES-positive (n = 43)	MES-negative (n = 68)	P value*
Age, years – mean (SD)	70 (13)	63 (14)	74 (10)	<0.01
Male sex – n (%)	57 (51)	22 (51)	35 (52)	0.98
Hypertension – n (%)	72 (65)	25 (58)	47 (69)	0.31
Diabetes mellitus – n (%)	24 (22)	9 (21)	15 (22)	0.88
Dyslipidemia – n (%)	60 (54)	21 (51)	39 (56)	0.47
Obesity – n (%)	21 (19)	7 (16)	14 (22)	0.62
Smoker – n (%)	18 (16)	12 (28)	6 (9)	0.02
Atrial Fibrillation – n (%)	53 (48)	14 (33)	39 (57)	0.01
Ischemic Heart Disease – n (%)	12 (16)	6 (14)	12 (18)	0.75
Heart Failure – n (%)	18 (16)	5 (12)	12 (18)	0.56
Chronic medication – n (%)
Statin	57 (52)	21 (49)	36 (53)	0.85
Antiplatelet	30 (27)	9 (21)	20 (29)	0.32
Anticoagulant	31 (28)	7 (16)	24 (38)	0.03
Antihypertensive	70 (63)	25 (58)	46 (68)	0.31
Systolic BP – mean (SD)	142 (22)	145 (21)	141 (22)	0.42
Diastolic BP – mean (SD)	77 (17)	79 (15)	76 (18)	0.28
Glucose – median (IQR)	129 (104 – 160)	132 (100 – 158)	128 (107 – 163)	0.62
Total cholesterol – median (IQR)	157 (129 – 194)	181 (139 – 209)	147 (126 – 175)	0.06
TOAST Classification – n (%)				0.02†
Large Artery Atherosclerosis	13 (12)	†9 (21)	†4 (6)
Cardioembolism	61 (55)	19 (44)	42 (62)
Undetermined	35 (32)	13 (30)	22 (32)
Other (Carotid Dissection)	2 (2)	2 (5)	0 (0)
Baseline NHISS – median (IQR)	15 (9 – 18)	12 (6 – 16)	16 (13 – 18)	0.01
Occlusion site – n (%)				0.76†
Terminal ICA	14 (12)	5 (2)	9 (13)
M1	84 (76)	34 (79)	50 (74)
M2	13 (12)	4 (9)	9 (13)
ASPECTS Score – median (IQR)	9 (7 – 10)	8 (7 – 10)	9 (8 – 10)	0.05
Collateral grade	2 (2 – 3)	2 (2 – 3)	2 (2 – 3)	0.87
IV Thrombolysis – n (%)	45 (40)	18 (42)	27 (40)	0.82
Grade 3 in mTICI scale – n (%)	70 (63)	27 (63)	43 (63)	0.96
No. of passages – median (IQR)	1 (1 – 2)	1 (1 – 2)	2 (1 – 2)	0.81
Hemorrhage (PH1-PH2)	11 (10)	4 (9)	7 (10)	0.87
LSW to TPA time, hours – median (IQR)	2.42 (1.75 – 3.77)	2.38 (1.90 – 3.33)	2.75 (1.75 – 4.38)	0.53
LSW to Recanalization time, hours – median (IQR)	6.40 (5.05 – 8.20)	6.85 (5.33 – 9.67)	5.77 (5.00 – 7.85)	0.08
LSW to TCD time, hours – median (IQR)	21.0 (12.0 – 36.8)	20.5 (10.8 – 37.5)	21.5 (12.7 – 36.5)	0.41

Abbreviations: ASPECTS = The Alberta Stroke Program Early Computed Tomography Score (ASPECTS); BP = blood pressure; CI = confidence interval; ICA = internal carotid artery; IQR = interquartile range; LSW = last-seen-well time; M1 = main trunk of the middle cerebral artery; M2 = first-order branch of the main trunk of the middle cerebral artery; MES = microembolic signals; mTICI = modified Thrombolysis in Cerebral Infarction; NHISS = Scores on the National Institutes of Health Stroke Scale; SD = standard deviation; TIA = transient ischemic attack; TCD = transcranial Doppler; TOAST = Trial of Org 10172 in Acute Stroke Treatment.

^*P values are shown for differences between MES-positive and MES-negative groups obtained from chi-square test/ Fisher’s exact test for categorical variables, or Student’s t/Mann-Whitney test for continuous variables.

^†Multiple comparisons were adjusted by Bonferroni method (new P value was set to P<0.05 divided by the number of comparisons)

For the analysis of ischemic stroke/TIA recurrence or systemic embolism, the Kaplan-Meier method and log-rank Mantel-Cox test were used. Hazards ratios and 95% CI were estimated by Cox regression models. Due to the low number of these events, we only performed adjusted analysis by one covariate (propensity score).

All statistical analyses were performed with IBM SPSS Statistics for Windows, version 25. Statistical significance was inferred at p<0.05.

Results

Patient characteristics

From September 1, 2017 through November 15, 2018, 138 patients fulfilled the inclusion criteria. Of these, 1 patient was excluded due to re-occlusion at the time of TCD assessment, 3 were excluded due to poor quality TCD recordings, and 23 due to insufficient temporal bone windows for TCD. Thus, we enrolled 111 patients (Figure 1).

Figure 1:

Flow chart of the study’s recruitment

Abbreviations: LOCF = last observation carried forward; MES = microembolic signals; mTICI = modified Thrombolysis in Cerebral Infarction; NICU = neurocritical care unit; TCD = transcranial Doppler

MES were detected on at least one side in 43 of 111 patients (39%). These patients constituted the MES-positive group. MES detection ranged from 1 to 21 counts with a median rate of 4 (interquartile range [IQR] 2 – 12 counts/hour). In the MES-positive subgroup (n=43), MES were detected in a single unilateral probe in 33 patients (77%). In 26 of these 33 patients (79%), the side of the MES corresponded to the affected hemisphere. Although there was a trend for MES-positive patients to have an increased LSW-to-recanalization time (Table 1: median [IQR] 6.85 [5.33 – 9.67] versus 5.77 [5.00 – 7.85] hours, p=0.08), there was no significant correlation between the rate of MES detection with relevant time points (Supplemental Figure I).

Comparison of baseline characteristics between MES-positive and MES-negative groups are shown in Table 1. MES-positive patients were younger (63±14 versus 74±10 years-old, p<0.01) and more frequently smokers (28% versus 9%, p=0.02). They tended to have higher total cholesterol (median [IQR] 181 [139 – 209] versus 147 [126 – 175] mg/dl, p=0.08). MES detection differed according to the TOAST classification (table 1, p=0.02; Supplemental Figure II). Excluding the two MES-positive cases of carotid dissection, symptomatic large artery atherosclerosis (all high-grade cervical ICA stenosis), showed the highest frequency of MES positivity (9 of 13 cases, 69%). In comparison, only 31% (19 of 61) of those with cardioembolism were MES-positive. Bilateral detection was more frequent in cardioembolic stroke (32%) and embolic stroke of undetermined source (30%) than in other stroke types, although the difference did not reach statistical significance (chi-squared test p=0.373). MES-positive patients showed less severe neurological deficits at baseline (NIHSS, median [IQR] 12 [6–16] versus 16 [13–18], p=0.01) but more evidence of early ischemia on admission CT (ASPECTS, median [IQR] 8 [7–10] versus 9 [8–10], p=0.05).

Functional and radiological outcome

Table 2 shows the results from the comparison of outcomes between the MES-positive and MES-negative groups. There was no difference in functional outcomes between the groups. One can visually compare the relative proportion of patients in each mRS category by MES group in Figure 2. The main functional end-point, the ordinal shift across all categories of mRS at 90 days, showed no significant difference between the MES-positive and MES-negative groups (adjusted OR = 1.06, 95% CI 0.48 – 2.34, p = 0.85) (Table 2). The proportions of patients who were functionally independent at 90 days (mRS 0 – 2) were also not significantly different (adjusted OR = 0.52, 95% CI 0.19 – 1.39, p = 0.19). The presence of MES did not significantly correlate with the odds of significant early neurological recovery (adjusted OR = 1.09, 95% CI 0.43 – 2.79, p = 0.86).

Table 2.

Outcome measures based on MES positive or negative detection

Outcome measures	MES- positive n = 43	MES- negative n=68	Univariate Model	*Multivariate Model 1	†Multivariate Model 2
			Unadjusted Odds ratio (CI 95%); P value	Adjusted Odds ratio (CI 95%); P value	Adjusted Odds ratio (CI 95%); P value
‡mRS score shift – median (IQR)	2 (1 – 3)	3 (1 – 4)	0.61 (0.31 – 1.21); p = 0.16	1.04 (0.46 – 2.34); p = 0.92	1.06 (0.48 – 2.34); p = 0.85
§mRS 0 – 2 at 90 days – n (%)	25 (58)	34 (50)	1.38 (0.64 – 3.00); p = 0.40	0.46 (0.15 – 1.40); p = 0.17	0.52 (0.19–1.39); p = 0.19
§Early neurological recovery at 24 hours – n (%)	16 (38)	25 (39)	1.04 (0.47 – 2.30); p = 0.92	2.15 (0.64 – 7.12); p = 0.24	1.09 (0.43 – 2.79); p = 0.86
			Unadjusted Beta (CI 95%); P value	Adjusted Beta (CI 95%); P value	Adjusted Beta (CI 95%); P value
||Infarct volume at 24 hours, milliliters – median (IQR)	14 (1 – 67)	14 (7 – 47)	–0.08 (–29.9 – 29.9); p = 1.00	–7.90 (–37.3 – 21.4); p = 0.59	11.2 (–46.6 – 22.9); p = 0.51

Abbreviations: CI = confidence interval; MES = microembolic signals; M1 = main trunk of the middle cerebral artery; M2 = first-order branch of the main trunk of the middle cerebral artery; ICA = internal carotid artery; NHISS = scores on the National Institutes of Health Stroke Scale. Early neurological recovery defined as scoring a NIHSS of 0 or 1 or a drop of ≥8 points in NIHSS from baseline to 24 hours.

^*Multivariate model 1 = adjusted to predictors of outcome – age, NIHSS, Last-seen-well to canalization time (hours), the Alberta Stroke Program Early Computed Tomography Score (ASPECTS), modified Thrombolysis in Cerebral Infarction (mTICI).

^†Multivariate model 2 = adjusted to unbalanced baseline differences between MES groups by probability score matching – age, smoking, baseline NIHSS, cholesterol level, Trial of ORG 10172 in Acute Stroke Treatment (TOAST) classification, ASPECTS, last-seen-well to recanalization time.

^‡Univariate and multivariate analysis with ordinal shift analysis to predict outcome

^§Univariate and multivariate analysis with logistic regression to predict outcome

^||Univariate and multivariate analysis with generalized linear mixed model to predict infarct volume

Figure 2:

Distribution of scores for disability on the modified Rankin scale at 90 days (which ranges from 0 to 6, with higher scores indicating more severe disability) among patients in the subgroup with detected microembolic signals (MES positive) and those with no detection (MES negative). The numbers in the bars are percentages of patients who had each score. For the statistical analysis please see Table 2.

MES-positive and MES-negative subgroups had similar infarct volumes on 24-hour CT scan (median [IQR] 14 [1 – 67] versus 14 [7 – 47] ml, adjusted beta = –11.2, 95% CI –46.6 – +22.9, p = 0.51).

Stroke or systemic embolism recurrence

Follow-up included 8505 person-days until 90 days or death/event occurrence. There were 8 recurrent ischemic strokes, and there was 1 TIA, representing an event in 8% of the entire cohort. Four of these events occurred within 7 days, 4 within 30 days, and 1 after 30 days. Most recurrences were in the MES-positive group (7/43 [16%] versus 2/68 [3%] in MES-negative group). Two patients suffered systemic embolization (one to the intestines and one to the lower limb), one in each MES group. In the 7 cases with cerebral ischemic recurrence in the MES-positive group, 4 (57%) repeated the stroke on the same side as the initial TCD detection. Cardioembolism accounted for the greatest number of recurrences (details in Supplemental Table II). The time-dependent change in risk showed significantly higher risk of embolic events (Figure 3A, log rank test p=0.02), and particularly ischemic stroke/TIA (Figure 3B, log rank test p=0.01) in the MES-positive group. MES-positive patients had eight times the risk of having an ischemic stroke/TIA at follow-up compared to the MES-negative patients (adjusted HR 8.22, CI 95% 1.55–43.9, p=0.01) and over 6 times the risk of the composite outcome of all embolic events (adjusted HR 6.73, CI 95% 1.63–27.8, p=0.01) (Table 3). Ten patients died during study follow-up; 2/43 (5%) in the MES-positive group and 8/68 (12%) in the MES-negative group. There was no correlation between the presence of MES and all-cause mortality in either the Kaplan-Meier survival analysis (log-rank test p = 0.21, Figure 3C) or the Cox-regression hazard model (Table 3).

Figure 3:

Kaplan-Meier survival curves for event recurrence within 90 days from index stroke. Composite outcome (Panel A; ischemic stroke, TIA, systemic embolism), only ischemic stroke recurrence or TIA (Panel B) and death (Panel C). In each panel MES-positive group versus MES-negative group are compared. MES-positive were at higher risk of recurrent embolic events but not death within a period of follow-up at 3 months from index stroke. The inset in each panel shows the same data on an enlarged segment of the y axis.

Table 3.

Comparison of the risk of composite outcome vascular event, ischemic stroke/TIA and all-cause mortality in patients with MES-positive compared with MES-negative using Cox proportional hazard regression.

	Recurrent Ischemic Stroke or TIA			*Composite Outcome			All-cause Mortality
	Unadjusted analysis
	n	HR (95% CI)	P value	n	HR (95% CI)	P value	N	HR (95% CI)	P value
MES-positive	7	5.78 (1.20–27.9)	0.03	8	4.50 (1.19–16.9)	0.03	2	0.38 (0.08–1.82)	0.29
MES-negative	2	Reference		3	reference		8	reference
	†Adjusted analysis
	n	HR (95% CI)	P value	n	HR (95% CI)	P value	N	HR (95% CI)	P value
MES-positive	7	8.22 (1.55 – 43.6)	0.01	8	6.73 (1.63–27.8)	0.01	2	0.40 (0.07–2.27)	0.30
MES-negative	2	reference		3	reference		8	reference

Abbreviations: CI = confidence interval; HR = Hazard Ratio; MES = Microembolic Signals; TIA = transient ischemic attack

^*Composite outcome: ischemic stroke, TIA, systemic embolism

^†Cox-Hazards regression model adjusted to unbalanced baseline differences between MES groups by probability score matching – age, smoking, National Institutes of Health Stroke Scale (NIHSS) scores, cholesterol level, Trial of ORG 10172 in Acute Stroke Treatment (TOAST) classification, ASPECTS, last-seen-well to recanalization time.

Discussion and conclusions

Among patients with anterior circulation stroke due to occlusion of the ICA or proximal MCA (M1/M2) who achieved an mTICI score of 2b-3 following mechanical thrombectomy, early neurological recovery, functional outcome at 90 days, and final infarct volume were not associated with the occurrence of MES within the first 72 hours after the procedure. However, the presence of MES increased the risk of recurrent ischemic stroke or systemic embolization by 6- to 8-fold during 90 days of follow-up.

Outcome and the effect of mechanical thrombectomy

MES did not predict functional outcome after stroke nor the final infarct volume at 24 hours after thrombectomy. One possible explanation is that microemboli are adequately washed out from the cerebral microvasculature and, therefore, do not cause additional clinically relevant ischemic lesions. Animal models show that even when a large percentage of injected artificial microemboli remain in both pial and penetrating arteries, the emboli may not create significant ischemia.¹⁹ Also, the majority of patients who undergo cardiac procedures that are known to cause very high rates of microembolization do not suffer major neurological harm afterwards.²⁰ The presence of MES might just be a marker of a successful intracranial clot dissolution after recanalization therapy,²¹ and one study in the pre-recanalization therapy era suggested that there was no association with stroke outcome.²²

However, a recent study monitored 40 patients soon after mechanical thrombectomy and found that patients with high MES rate had worse disability and increased mortality.²³ Contrary to our study, which restricted inclusion to mTICI 2b-3, this study included 30% of patients with incomplete recanalization (mTICI scores 0 – 2a). Incomplete recanalization was itself related to the presence of MES, and incomplete recanalization and poor collateralization were both independent angiographic markers of poor prognosis.^{1, 2}

To our knowledge, this is the largest study of MES after endovascular treatment. The differences in baseline characteristics between MES groups could have biased our ability to predict outcome. MES-positive patients were younger and presented with less severe strokes with lower NIHSS scores.¹ Although we tried to compensate for this bias in the multivariate models by propensity score matching, it is possible that MES confer risk, however, that they do so in a lower risk subgroup of stroke patients, making it difficult to detect their harmful effects.

The source of MES and the cause of stroke

This study shows that MES are very common after successful endovascular therapy. We found rates of MES detection similar to those reported in prior studies done before the introduction of revascularization therapies, ^{10, 24, 25} suggesting that the source of MES might be related more to underlying ischemic stroke etiology than to vascular damage resulting from mechanical thrombectomy itself. For example, our MES rate was highest at 69% in symptomatic carotid stenosis and lower in cardioembolism at 31%. These rates are similar to those reported without endovascular therapies.²⁶ In support of this viewpoint, we found that the surrogates for procedure-related damage, such as number of passes required and post-thrombectomy hemorrhage procedure complications, were not associated with occurrence of MES (Table 1). However, we only monitored the M1 segment of MCA, so we may have missed device-related complications at more distal MCA branches.

Predictive role of MES in recurrent ischemic stroke, TIA and systemic embolism

The presence of MES was associated with a 6- to 8-fold increase in the risk of ischemic stroke/TIA or systemic embolization, suggesting that MES may be considered an independent risk factor for recurrent vascular events. This finding is in accordance with various studies which have demonstrated that MES predict increased risk of future cerebral ischemia after an acute stroke.^{10, 27} Most of these studies address the predictive value of MES in carotid stenosis.^{11, 28} In our study, all patients with high-grade carotid stenosis were symptomatic and were treated with endarterectomy. This would be expected to lower the stroke recurrence, and may have underestimated the stroke recurrence rate and may also explain the fact that the patients with recurrent events were almost all cardioembolic (Supplemental Table II).

These finding highlight the potential role for continuous neuromonitoring with MES detection as a tool for identifying patients with higher risk of ischemic stroke/TIA.

Limitations

The wide confidence intervals obtained for OR estimates for stroke / TIA recurrence suggest that the sample size was relatively small, limiting statistical power to predict clinical events at 90 days. Although shorter monitoring times improves patient compliance in the acute stroke setting, 30 minutes is half the time recommended to monitor MES.¹⁸ Longer TCD monitoring may have increased the detection rates, but we do not believe that it would have changed the observed relationship with the outcome substantially. We did not perform MRI, and thus, may have missed smaller ischemic lesions associated with MES.²⁹ The small number of recurrent events might also lead to type error II in estimates of risks.

Conclusions

In this study the occurrence of microembolization during the first 72 hours after successful endovascular therapy did not correlate with functional outcome or infarct volume at follow up. However, the presence of microemboli in the acute stage appears to be an important risk factor for further embolic events in the weeks and months after a stroke. This finding may be useful in risk stratification and patient management.