Frequency of BBB Disruption Post Endovascular Therapy and Multiple Thrombectomy Passes in Acute Ischemic Stroke Patients

Abstract

Background and Purpose

The high prevalence of hyperintense acute reperfusion injury marker (HARM) seen following endovascular therapy (EVT), is suggestive of blood brain barrier (BBB) disruption and hemorrhage risk and may be attributable to multiple thrombectomy passes needed to achieve recanalization.

Methods

Acute stroke patients were included if they were screened from January 2015 through February 2019, received an acute ischemic stroke diagnosis involving the anterior circulation, treated with or without IV tPA, consented to the NINDS natural history study and imaged with a baseline MRI prior to receiving EVT. Consensus image reads for HARM and hemorrhagic transformation (HT) were performed. Good clinical outcome was defined as 0–2 using the latest available modified Rankin score (mRS).

Results

Eighty patients met all study criteria and were included in the analyses. Median age was 65 years, 64% female, 51% Black/African-American, median admit NIHSS=19, 56% treated with IV tPA, and 84% achieved TICI of 2b\3. Multiple pass patients had significantly higher rates of severe HARM at 24-hours (67% versus 29%, p=0.001), any HT (60% versus 36%, p=0.04) and poor clinical outcome (67% versus 36%, p=0.008). Only age (OR 1.1 95% CI [1.01–1.12], p=0.022) and severe HARM at 24-hours post-EVT were significantly associated with multiple passes (OR 7.2 95% CI [1.93–26.92], p=0.003).

Conclusions

In this exploratory study, multiple thrombectomy passes are independently associated with a significant increase in BBB disruption detected at 24-hours. Patients with HARM post-EVT had a more than seven-fold increase in the odds of having multiple versus single pass thrombectomy.

Clinical Trial Registration Information

URL: https://www.clinicaltrials.gov. Unique identifier: .

Introduction

Recently Zaidat et al¹ found that usage of a single pass of the Solitaire thrombectomy device was the most important predictive factor in clinical outcome rather than achievement of revascularization based on Thrombolysis In Cerebral Infarction (TICI) score of 3 with multiple passes.¹

Shi et al² found that patients with multiple attempts at removing clot had higher rates of blood brain barrier (BBB) disruption. Moderate BBB was independently associated with higher rates of hemorrhagic transformation (HT) and poor clinical outcome, defined as mRS 4–6, even after successful reperfusion.² A study by Bourcier et al, found that there was an increased risk of parenchymal hematoma at 24-hours in patients receiving 3 or more passes in the stent retriever group.³

Since Warach S and Latour LL⁴ first coined the term “hyperintense acute reperfusion marker (HARM),” numerous studies have included this marker as both a sign of injury and reperfusion in patients with successful recanalization post intervention.^5–9 HARM is visualized as increased signal intensity on T2-weighted Fluid Attenuated Inversion Recovery (FLAIR) MRI following injection of gadolinium based contrast agents.⁴

In the setting of multiple positive endovascular therapy (EVT) trials^10–14 and their impact on increased thrombectomy procedures in a larger stroke population worldwide, understanding of the imaging markers post-EVT aside from hemorrhagic complications has not been thoroughly investigated. The utility of MRI for sensitivity to both HT and BBB disruption post-EVT is advantageous.⁶

We hypothesized that multiple thrombectomy passes to achieve recanalization may lead to higher rates of BBB breakdown as evidenced by HARM on FLAIR at 24-hours post-EVT.

Methods

Patient Population

The data that support the findings of this study are available from the corresponding author upon reasonable request. Patients were included in this study if they were screened from January 2015 to February 2019 and received a clinical diagnosis of ischemic stroke involving large vessel occlusion (LVO) of the anterior circulation. Patients included received embolectomy with or without standard IV tPA. All patients were imaged with a baseline multimodal MRI prior to receiving EVT. Follow-up MRI was obtained at 24-hours post EVT. Patients provided written consent to the NIH Natural History of Stroke Study. The appropriate Ethics and Institutional Review Boards (NINDS/NIH IRB for Suburban Hospital, Johns Hopkins Medicine, Bethesda, MD; and Medstar Washington Hospital Center, Washington Hospital Center, Washington, DC IRB) approved the study (NCT00009243).

Image Acquisition

Patients were scanned with one of two clinical 3T scanners depending on the hospital site (Siemens Skyra, Siemens AG, Munich, Germany and Philips Achieva, Philips Healthcare, Best, the Netherlands) using established acute stroke imaging protocols.¹⁵ A dosage of 0.1 mmol/kg of Gd-DTPA was administered if the PWI sequence was obtained at baseline, 2-hours, or 24-hours.

Image Analysis

Two independent raters evaluated the pre-contrast FLAIR images at 24-hours post-EVT, compared to the baseline pre-contrast FLAIR images, for any enhancement in the CSF spaces, especially in the sulci and on the cortical surface. This enhancement, noted as HARM, was rated as: none; present but not severe; severe occupying 10 or more slices but focal to the acute DWI ischemic lesion; or severe occupying 10 more slices but diffuse and not limited to just the territory where the acute DWI ischemic lesion was present. HARM reads were performed using the pre-contrast FLAIR images at 24-hours on patients that had received at least one dosage of Gd-DTPA at baseline or 2-hours. Two independent raters evaluated the GRE at 24-hours post-EVT, compared to the baseline GRE, for any HT by applying the ECASS-II criteria. Discrepancies in the HARM and HT reads were resolved by an independent expert rater with consensus reads used for all study analyses.¹⁵ Complete recanalization was defined in the IR suite as TICI scores of 2B or 3. Symptomatic intracranial hemorrhage (sICH) was defined as an increase of at least 4 points on the NIHSS at 24-hours coincident with any HT. Early neurologic improvement (ENI) was defined as a ≥8-point decrease or a 0–1 value on the NIHSS at 24-hours. Good clinical outcome was defined as modified Rankin Scale (mRS) of 0–2 using the latest available follow-up data from discharge, 5 days, 30 days or 90 days.

A fully automated algorithm was used to calculate the core lesion volume using the baseline DWI and ADC images obtained prior to EVT. Brain and lesion masks were derived from the DWI. The voxels within the lesion mask were interrogated to apply a thresholding with a cutoff of ADC ≤ 620 μm²/s. The largest object meeting criteria was designated as the core lesion and grown in three dimensions to calculate the core lesion volume.

Statistical Analysis

Descriptive statistics were tabulated for the patient population including demographic, clinical, procedural, and imaging variables. Chi-square test was used. Multiple logistic regression analysis was performed to identify independent variables associated with multiple passes as well as the incidence of any HARM and severe HARM. Variables with P<0.10 from the univariate analysis were entered into the regression models. Spearman correlation coefficients were calculated for the HARM reads with significance level of 0.01. The SPSS Statistics software v19 (IBM) was used.

Results

In the study period across both hospital sites, a total of 224 patients received EVT. One hundred and eight (48%) of the 224 patients were excluded since they did not consent to the NIH Natural History Stroke study and 36 patients were excluded due to posterior circulation LVO or no pre-EVT MRI. Eighty patients (n=80) met all study criteria. Median age was 65 years [54–75], 64% female, 51% Black/African-American, admit NIHSS=19 [12–22], baseline mRS=0, core volume=14mL [7–31], 56% treated with IV tPA, 84% achieved TICI of 2b\3, onset to recanalization of 292 minutes [230–430], and time from baseline to 24-hour MRI of 25 hours [22–27]. The interobserver reliability for the HARM reads was ρ=0.733 for any HARM and ρ=0.805 for severe HARM. Patients received an average of 1.7 dosages of Gd-DTPA prior to the acquisition of the pre-contrast FLAIR at 24-hours.

Patients treated with multiple thrombectomy passes (n=52) compared to patients treated with single pass (n=28) were older (69 versus 57 years, p=0.003), had more severe outcome mRS (3 versus 2, p=0.001) and slightly higher creatinine (1.1 versus 0.85, p=0.003), see Table 1. Multiple pass patients had significantly higher rates of severe HARM at 24-hours (67% versus 29%, p=0.001), any HT at 24-hours (60% versus 36%, p=0.04) and higher rates of poor clinical outcome (67% versus 36%, p=0.008) as illustrated in Figure 1. There were no significant differences in recanalization (81% versus 89%, p=0.36) or sICH rates (10% versus 0%, p=0.09).

Table 1:

Patient characteristics stratified according to single versus multiple thrombectomy passes

Patient Characteristic^	Single pass (n=28)	Multiple passes (n=52)	p-value
Age	57 [47–69]	69 [59–76]	0.003
Sex (Female)	20 (71%)	31 (60%)	0.33
Race (Black or African-American)	17 (61%)	24 (46%)	0.51
Ethnicity (Hispanic)	1 (4%)	3 (6%)	0.83
Creatinine	0.85 [0.68–1.02]	1.1 [0.9–1.4]	0.003
GFR	95 [60–102]	69 [45–89]	0.05
Admit NIHSS	19 [13–22]	19 [11–23]	0.79
Preadmit mRS	0 [0–0]	0 [0–0]	0.43
IV tPA received	19 (68%)	26 (50%)	1.0
Core lesion volume (mL)	15 [7–36]	12 [6–29]	0.25
Core >70mL	3 (11%)	6 (12%)	0.89
M1 occlusion	19 (68%)	40 (77%)	0.39
Onset to groin puncture	278 [221–327]	213 [168–640]	0.19
Onset to recanalization	297 [242–348]	290 [223–729]	0.98
Complete recanalization (TICI of 2b\3)	25 (89%)	42 (81%)	0.36
Any HT at 24-hours	10 (36%)	31 (60%)	0.04
sICH	0 (0%)	5 (10%)	0.08
Any HARM at 24-hours	15 (54%)	37 (71%)	0.13
Severe HARM at 24-hours	8 (29%)	35 (67%)	0.001
NIHSS at 24-hours	9 [3–14]	10 [4–20]	0.31
ENI at 24-hours	19 (68%)	17 (33%)	0.003
Latest available mRS	2 [1–3]	3 [2–5]	0.001
Good clinical outcome	18 (64%)	16 (31%)	0.005
Poor clinical outcome	10 (36%)	35 (67%)	0.008

^{^}Values reported as: median IQR [25–75] or n (%)

Abbreviations used: IQR: interquartile range, n: number, SD: standard deviation, NIHSS: National Institutes of Health Stroke Scale, mRS: modified Rankin scale, M1: M1 branch of middle cerebral artery (MCA); HT: hemorrhagic transformation; symptomatic intracranial hemorrhage (sICH); hyperintense acute reperfusion marker (HARM); early neurological improvement (ENI)

Figure 1:

Top panel: female patient, 75 years of age, L MCA stroke, pre-admit mRS=0, admit NIHSS=21, small core=9mL, large visual mismatch (PWI>DWI, white oval), achieved TICI=3 with 3 passes of EVT, onset to recanalization of 244 minutes, no HT at 24-hours but severe focal HARM at 24-hours (red arrows), 24-hour NIHSS=9 but follow-up mRS=3 (poor outcome). Bottom panel: male patient, 75 years of age, L MCA stroke, pre-admit mRS=0, admit NIHSS=13, small core=9mL, large visual mismatch (PWI>DWI, white oval), achieved TICI=2b with single pass, onset to recanalization of 263 minutes, no HT at 24-hours and no HARM at 24-hours, 24-hour NIHSS=0, and follow-up mRS=0 (good outcome).

Only age (OR 1.1 95% CI [1.01–1.12], p=0.022) and severe HARM at 24-hours post-EVT were significantly associated with multiple passes (OR 7.2 95% CI [1.93–26.92], p=0.003) while creatinine (OR 3.91 95% CI [0.83–18.46], p=0.09) and any HT at 24-hours (OR 3.4 95% CI [0.92–12.5], p=0.07) failed to show significant associations.

When including the clinical outcome variables that were significantly different between single and multiple pass subgroups (Table 1), age (OR 1.07 95% CI [1.01–1.13], p=0.02) and severe HARM at 24 hours (OR 8.52 95% CI [2.23–32.48], p=0.002) were significantly associated with multiple pass patients. The other variables failed to reach significance. Considering just patients with severe HARM at 24-hours, only 26% (9/35) multiple pass had a good outcome compared to 88% (7/8) single pass patients (p=0.001).

Patients with any HARM at 24-hours had higher 24-hour NIHSS (p=0.026), 10 [5–20], versus those with no HARM, 6 [2–11]. The 24-hour NIHSS (OR 1.1 95% CI [1.00–1.15], p=0.043) was significantly associated with any HARM at 24-hours. Patients with severe HARM at 24-hours had a smaller decrease in NIHSS at 24-hours (p=0.005), −3 [−9-(1)], versus those without severe HARM, −9 [−14-(−3)]. Change in NIHSS at 24-hours (OR 1.1 95% CI [1.02–1.20], p=0.015) was significantly associated with severe HARM.

Discussion

The main finding of this exploratory study is that multiple thrombectomy passes are independently associated with a significant increase in BBB disruption detected on FLAIR at 24-hours post-EVT. This association was independent of age, creatinine, and HT. Patients with HARM post-EVT had a more than seven-fold increase in the odds of having multiple versus single pass thrombectomy. Post-EVT MRI is overall widely available in sites performing endovascular therapy. Although this study showed an association of ENI and good clinical outcome with single pass (Table 1), it was not powered to test if presence of HARM at 24-hours was an independent predictor of poor outcome. In a recently completed study, we have shown that second phase BBB disruption on day 5 was a predictor of poor clinical outcome.¹⁵ It is possible that mechanical BBB disruption resulting from multiple passes and reperfusion injury are variables limiting good clinical outcome in these patients.^5–8 The presence of reperfusion injury as visualized on MRI post-EVT is valuable to understand the pathophysiologic mechanisms associated with rapid revascularization in these patients. HARM may serve as a secondary injury target for future adjunctive therapies.

There are several limitations to this exploratory study including limited sample size relative to the number of variables considered, the variability in the amount of gadolinium and when given across study patients and its impact on HARM reads in particular, and the potential selection bias based on requirement of enrollment into the NIH Natural History of Stroke Study. The utility of multimodal MRI prior to EVT is limited to select sites like ours and not the standard imaging approach at most hospital sites screening for EVT. However, the ability to obtain MRI post-EVT at 24-hours for evaluation of secondary injury is feasible at most hospital sites.