Early neurological deterioration in patients with acute ischemic stroke is linked to unfavorable cerebral venous outflow

Abstract

Introduction:

Early neurological deterioration (END) is associated with poor outcomes in patients with acute ischemic stroke due to large vessel occlusion (AIS-LVO). Causes of END after mechanical thrombectomy (MT) include unsuccessful recanalization and reperfusion hemorrhages. However, little is known about END excluding the aforementioned causes. We aimed to investigate factors associated with unexplained END (END_unexplained) with regard to the cerebral collateral status.

Patients and Methods:

Multicenter retrospective study of AIS-LVO patients with successful MT (mTICI 2b-3). On admission CT angiography (CTA), pial arterial collaterals and venous outflow (VO) were assessed using the modified Tan-Scale and the Cortical Vein Opacification Score (COVES), respectively. END_unexplained was defined as an increase in NIHSS score of ⩾ 4 within the first 24 hours after MT without parenchymal hemorrhage on follow-up imaging. Multivariable regression analyses were performed to examine factors of END_unexplained and unfavorable functional outcome (modified Rankin Scale score 3-6).

Results:

A total of 620 patients met the inclusion criteria. END_unexplained occurred in 10% of patients. While there was no significant difference in pial arterial collaterals, patients with END_unexplained exhibited more often unfavorable VO (81% vs. 53%; P < 0.001). Unfavorable VO (aOR [95% CI]; 2.56 [1.02-6.40]; P = 0.045) was an independent predictor of END_unexplained. END_unexplained was independently associated with unfavorable functional outcomes at 90 days (aOR [95% CI]; 6.25 [2.06-18.94]; P = 0.001).

Discussion and Conclusion:

Unfavorable VO on admission CTA was associated with END_unexplained. END_unexplained was independently linked to unfavorable functional outcomes at 90 days. Identifying AIS-LVO patients at risk of END_unexplained may help to select patients for intensified monitoring and guide to optimal treatment regimes.

Graphical abstract.

Introduction

Based on landmark trials, mechanical thrombectomy (MT) became the standard therapy for patients with acute ischemic stroke due to large vessel occlusion (AIS-LVO).^1–3 However, some patients who underwent MT suffer from early neurological deterioration (END), commonly defined as an increase of four or more points in the National Institutes of Health Stroke Scale (NIHSS).^4–6 Previous studies observed a strong association between END and unfavorable long-term functional outcomes in AIS-LVO patients.^7–10 Evident causes of END include the presence of reperfusion hemorrhage, unsuccessful recanalization, embolization in new vascular territories, early recurrent ischemic stroke and malignant edema.^11,12 The pathophysiology, risk factors and treatment strategies of END including these causes have been discussed in the literature.^10,13–15 However, other potential causes are insufficiently understood and in more than half of all END patients, no clear cause can be identified.^11,16 Identifying these patients is important to guide to optimal therapy selection and periprocedural management. In this regards, recently published studies focused on END of unexplained cause (END_unexplained) and observed several predictors including age, diabetes mellitus, number of passes and admission-to-groin puncture time.^5,9 The concept of fast infarct progression may further contribute to END and current research suggest a correlation between fast infarct progression and poor collaterals. ¹⁷ Yet, studies with respect to END did not provide data on imaging biomarkers, reflecting the cerebral microperfusion and collateral status.

The presence of favorable pial arterial collaterals prior to treatment has recently been validated for thrombectomy triage in the MR-CLEAN-LATE trial, ¹⁸ and was linked to reduced ischemic tissue damage and improved functional status after MT.^19,20 Beyond that, cerebral perfusion is not solely determined by arterial inflow, but also by its downstream venous outflow (VO). VO profiles were found to be correlated with favorable tissue-level microperfusion, less edema formation and improved functional outcomes.^21–24 However, there is still a paucity of research with regard to the impact of these collateral imaging biomarkers on END_unexplained.

We aimed to investigate whether pial arterial collaterals and VO correlate with END_unexplained and how END_unexplained affects long-term functional outcomes in AIS-LVO patients who were successfully treated by MT. We hypothesized that unfavorable pial arterial collaterals and VO profiles would increase the odds of END_unexplained. Moreover, we assumed that END_unexplained may be associated with unfavorable long-term functional outcomes.

Materials and methods

Study design

This retrospective multicenter cohort study was conducted in accordance with the STROBE guideline for observational studies. ²⁵ The aim of this study was to analyze the occurrence of END_unexplained in AIS-LVO patients who were treated at two comprehensive stroke centers between October 2013 and January 2021: University Medical Center Hamburg-Eppendorf in Germany and Stanford University School of Medicine in California, United States. Data collection included information on patient characteristics, imaging results, and treatment details. Approval for the study protocol was obtained from the Institutional Review Boards of both centers (Ethikkommission der Universität Hamburg: ID 689–15 and Ethics Committee of Stanford University: ID 37209) and all procedures adhered to the guidelines set by the Health Insurance Portability and Accountability Act (HIPAA) and the Declaration of Helsinki. Given the retrospective nature of the study, patient informed consent was waived by the review boards.

Patient inclusion and clinical data

All AIS-LVO patients were evaluated for mechanical thrombectomy (MT). Inclusion criteria were as follows: [1] thrombectomy triage within 16 hours after stroke onset ²⁶ ; [2] patients who underwent computed tomography (CT), including baseline non-contrast head CT and single-phase CT angiography (CTA); [3] patients with occlusion in the anterior circulation: internal carotid artery (ICA) or first (M1) or second segment (M2) of the middle cerebral artery (MCA); [4] patients treated with MT. Exclusion criteria were: [1] unsuccessful recanalization, defined as a modified Thrombolysis in Cerebral Infarction (mTICI) score of 0–2a; [2] missing data on NIHSS on admission or after 24 h; [3] occlusion of the M3 (or further distal) segment or in the posterior circulation; [4] NIHSS increase of ⩾4 within the first 24 h, but occurrence of parenchymal hemorrhage or missing data on follow-up imaging. Detailed information on patient inclusion and exclusion criteria can be found in Supplemental Figure S1.

Imaging analysis

The Cortical Vein Opacification Score (COVES) was used to determine VO profiles on admission CTA. The contrast opacification of three major cortical veins (superficial middle cerebral vein, sphenoparietal sinus and inferior anastomotic vein; Figure 1), reflecting the main venous drainage of the middle cerebral artery territory, was graded on a scale of 0–2 points (0 = not visible; 1 = moderate opacification; 2 = full opacification). ²¹ The total COVES ranges from 0 to 6 points. According to previous studies, unfavorable VO was considered as a score of ⩽2. ²⁷

Figure 1.

Patients with acute ischemic stroke due to an occlusion of the right middle cerebral artery (a–e). The upper two images (a and b) display the assessment of pial arterial collaterals, while the lower images show the evaluation of venous outflow. The first patient presented with favorable pial arterial collaterals (a) and the second patient exhibited unfavorable pial arterial collaterals (b). Images (c–e) display patients with unfavorable venous outflow profiles on admission CTA and unexplained early neurological deterioration after 24 h. Dotted green arrows indicate full, dotted red arrows missing and the dotted orange arrow moderate venous contrast opacification of the (c) superficial middle cerebral vein, (d) vein of Labbé, and (e) sinus sphenoparietalis.

Pial arterial collaterals were scored using the modified Tan scale at baseline CTA. ²⁸ Favorable pial arterial collaterals were defined as a filling of ⩾50% of the middle cerebral artery territory of the affected hemisphere (Figure 1). Computed tomography perfusion imaging was analyzed using the software platform RAPID (iSchemaView, Menlo Park, California, USA). Baseline ischemic core volume was identified by RAPID as tissue with a relative cerebral blood flow of <30% compared to the mean cerebral blood flow measured in regions of both hemispheres that do not have T_max delays as the control.

The degree of recanalization was assessed by using the modified Thrombolysis in Cerebral Infarction (mTICI) scale on digital subtraction angiography (DSA) images.

All ratings were performed by two experienced neuroradiologists (T.D.F. and J.J.H.) with 12 and 17 years of experience, respectively. Discrepancies in grading were settled by consensus.

Follow-up imaging (non-contrast head CT or magnetic resonance imaging) was performed within 48 h of MT. Reperfusion hemorrhage was assessed on follow-up imaging according to the European Cooperative Acute Stroke Study (ECASS) I criteria. ²⁹ Total infarct volumes (in ml) were manually outlined on non-contrast head CT or magnetic resonance follow-up imaging.

Assessment of clinical outcome characteristics

The NIHSS scores on admission and after 24 h were assessed by stroke neurologists. The modified Rankin Scale scores at 90 days were either determined by a stroke neurologist or a registered study nurse by a telephone or face-to-face assessment.

Outcome measures

Primary outcome was an early neurological deterioration of unexplained cause (END_unexplained), defined as an NIHSS score increase of ⩾4 points within the first 24 h after MT without evidence of a parenchymal hemorrhage on follow-up imaging, according to previous studies.^5,6,9 Secondary outcome was an unfavorable functional outcome at 90 days (modified Rankin Scale [mRS] score of 3–6).

Statistical analysis

Patient characteristics, imaging findings, treatment details, and clinical outcomes were compared between two subgroups of patients: those with END_unexplained and those without. The Mann-Whitney U test was used for continuous variables, while the chi-square test was used for categorical variables. The normality of data distributions was assessed using Shapiro-Wilk tests. Continuous variables are presented as medians with interquartile ranges (IQR), while categorical variables are described in counts and percentages.

We performed multivariable logistic regression analyses to identify independent factors associated with END_unexplained and unfavorable functional outcome at 90 days. The regression model with END_unexplained as the dependent variable was adjusted for baseline Alberta Stroke Program Early CT Score (ASPECTS), admission National Institutes of Health Stroke Scale (NIHSS) score, time from symptom onset to recanalization, final mTICI grade, VO, occlusion site, baseline ischemic core volume, and total infarct volume. The regression model with unfavorable functional outcome at 90 days as the dependent variable was adjusted for age, sex, intravenous thrombolysis, admission NIHSS, final mTICI grade, pial arterial collaterals, VO, occlusion site, END_unexplained, baseline ischemic core volume, and total infarct volume. In addition, we employed multiple imputation by chained equations to replace missing data in our multivariable logistic regression analyses. The imputation models were conducted with 200 iterations and 20 imputations. The proportion of imputed observations for each variable is provided in the supplementary table legends. Adjusted odds ratios (aOR) with associated p-values and 95% confidence intervals (CI) were reported for each independent variable. Statistical significance was considered at a p-value of less than 0.05. The variance inflation factor for each independent variable was calculated to reduce the possibility of multicollinearity in the regression model. The data analysis was conducted using Stata 16.0 (College Station, Texas, USA).

Data availability

Data supporting the findings of this study are available from the corresponding author upon reasonable request.

Results

Baseline characteristics of patients

A total of 620 patients met the inclusion criteria. Across all patients, median age was 75 (IQR: 63–82) and the sex ratio was balanced (50%). The median NIHSS on admission was 15 (IQR: 9–19) and the median ASPECTS was 8 (IQR: 6–9). Unfavorable VO profiles were observed in 56% of all patients and unfavorable pial arterial collaterals in 31%. The median time from symptom onset to recanalization was 325 min (IQR: 207–578) and complete recanalization [mTICI 2c-3] was achieved in 58% of patients. In terms of long-term outcomes, the median mRS score at 90-day follow-up was 3 (IQR: 1–5). Please refer to Table 1 for further details on baseline-, imaging-, and treatment characteristics.

Table 1.

Patients’ baseline, procedural and outcome characteristics.

	Total	No ENDunexplained	ENDunexplained	p-Value
	N = 620	N = 558	N = 62
Baseline patient characteristics
Age	75 (63–82)	75 (63–82)	76 (67–85)	0.15
Male sex	312 (50%)	283 (51%)	29 (47%)	0.56
Hyperlipidemia	181 (33%)	165 (33%)	16 (32%)	0.90
Diabetes mellitus	139 (23%)	128 (23%)	11 (18%)	0.34
Blood glucose on admission (mg/dL)	122 (105–150)	121 (105–148)	126 (110–164)	0.14
Arterial hypertension	425 (69%)	377 (68%)	48 (77%)	0.13
Atrial fibrillation	261 (42%)	236 (43%)	25 (41%)	0.81
Antiplatelet or anticoagulant treatment	286 (47%)	257 (47%)	29 (47%)	0.95
Admission NIHSS	15 (9–19)	15 (9–19)	15 (11–17)	0.74
Imaging characteristics
Left hemispheric stroke	329 (53%)	293 (53%)	36 (58%)	0.41
Proximal occlusion site a [ICA/proximal M1]	340 (55%)	293 (53%)	47 (76%)	<0.001
ICA	114 (18%)	94 (17%)	20 (32%)	0.003
MCA − M1	385 (62%)	351 (63%)	34 (55%)	0.21
MCA − M2	120 (19%)	112 (20%)	8 (13%)	0.17
ASPECTS	8 (6–9)	8 (7–9)	7 (6–9)	0.039
Unfavorable pial arterial collaterals [Tan scale]	191 (31%)	169 (31%)	22 (35%)	0.43
Unfavorable VO [COVES ⩽ 2]	343 (56%)	293 (53%)	50 (81%)	<0.001
Baseline ischemic core volume [CBF < 30%] (ml)	9 (0–30)	8 (0-28)	16 (3–43)	0.036
Treatment characteristics
Administration of tPA	316 (52%)	286 (52%)	30 (48%)	0.59
Time from symptom onset to tPA (min)	105 (70–150)	96 (70–150)	124 (109–195)	0.029
Time from symptom onset to recanalization (min)	325 (207–578)	320 (205–535)	415 (244–858)	0.019
Complete recanalization [mTICI 2c-3]	309 (58%)	283 (59%)	26 (45%)	0.038
Follow-up imaging
Any reperfusion hemorrhage on follow-up imaging	167 (33%)	144 (33%)	23 (37%)	0.49
PH1	33 (7%)	33 (8%)	0 (0%)
PH2	26 (5%)	26 (6%)	0 (0%)
Total infarct volume (ml)	21 (7–58)	20 (7–55)	46 (19–90)	<0.001
Outcome characteristics
24-h NIHSS	10 (4–18)	9 (3–16)	21 (18–24)	<0.001
mRS90	3 (1–5)	3 (1–5)	5 (4–6)	<0.001
Unfavorable functional outcome [mRS90 3–6]	342 (58%)	287 (55%)	55 (90%)	<0.001
mRS90 5–6	178 (30%)	141 (27%)	37 (61%)	<0.001

END_unexplained: early neurological deterioration of unexplained cause; NIHSS: National Institutes of Health Stroke Scale; ICA: Internal Carotid Artery; ASPECTS: Alberta Stroke Program Early CT score; VO: Venous Outflow; COVES: Cortical Vein Opacification Score; CBF: cerebral blood flow perfusion map; tPA: tissue plasminogen activator; mTICI: modified Thrombolysis in Cerebral Infarction; PH: parenchymal hematoma; mRS90: modified Rankin Scale score at 90 days follow-up.

Data are presented as median (IQR) for continuous measures, and n (%) for categorical measures. Characteristics were compared between 558 and 62 patients by using either Mann-Whitney U test (1) for continuous variables or a chi-square test (2) for categorical variables. Statistical significance: p < 0.05.

^aA proximal occlusion site was defined as an occlusion of the internal carotid artery or proximal first segment of the middle cerebral artery. The proximal first segment of the middle cerebral artery was defined as the first third before the lenticulostriate branches emerge.

Primary and secondary outcome

Of all included patients, 10% suffered from END_unexplained. These patients were significantly more likely to have a proximal vessel occlusion site [ICA or M1] (76%vs 53%; p < 0.001) and median ASPECTS on admission was slightly lower (7 vs 8; p = 0.039). In patients with END_unexplained, median times from symptom onset to the admission of intravenous thrombolysis (124 min vs 96 min; p = 0.029) and to recanalization (415 min vs 320 min; p = 0.019) were significantly longer. Complete recanalization [mTICI 2c-3] was more often achieved in patients without END_unexplained (59%vs 45%; p = 0.038). Baseline ischemic core volume (16 ml vs 8 ml; p = 0.036) and total infarct volume (46 ml vs 20 ml; p < 0.001) were higher in patients with END_unexplained. Patients with END_unexplained exhibited significantly more often unfavorable VO profiles (81%vs 53%; p < 0.001) and were more often affected by an unfavorable functional outcome at 90 days (90%vs 55%; p < 0.001). Please refer to Figure 2 illustrating the distribution of mRS at 90 days stratified by END_unexplained.

Figure 2.

Distribution of the modified Rankin Scale (mRS) at 90 days stratified by early neurological deterioration of unexplained cause (END_unexplained). Patients with END_unexplained suffered significantly more often from unfavorable functional outcome at 90 days (mRS 3–6; dotted line; 90%vs 55%; p < 0.001).

Independent predictors of END_unexplained and unfavorable functional outcome at 90 days

We conducted multivariable logistic regression analyses to assess independent predictors of END_unexplained and unfavorable functional outcomes at 90 days. Unfavorable VO (adjusted odds ratio [95% CI], 2.56 [1.02–6.40]; p = 0.045) was an independent predictor of END_unexplained. Predicted probabilities are illustrated in Figure 3 showing the association between VO and END_unexplained adjusted for the variables displayed in Table 2. After multiple imputation has been performed (Supplemental Table 1), unfavorable VO was still independently associated with END_unexplained (aOR [95% CI], 2.90 [1.35–6.22]; p = 0.006). In the imputed model, NIHSS score on admission was independently associated with END_unexplained (aOR [95% CI], 0.95 [0.90–1.00]; p = 0.032).

Figure 3.

Predicted probabilities for the occurrence of unexplained early neurological deterioration (END_unexplained) stratified by the Cortical Vein Opacification Score (COVES). The model was adjusted for the variables displayed in Table 2. Shaded bars in light blue indicate the 95% confidence intervals.

Table 2.

Multivariable logistic regression analysis to assess independent predictors of END_unexplained in successfully recanalized (mTICI 2b-3) AIS-LVO patients.

Independent variables	Adjusted odds ratio	95% Conf. interval		p-Value
ASPECTS [per point]	0.99	0.80	1.21	0.904
Admission NIHSS [per point]	0.95	0.89	1.01	0.113
Time from symptom onset to recanalization [per 30 min]	1.03	0.99	1.06	0.112
Complete recanalization [mTICI 2c-3]	0.62	0.30	1.27	0.188
Unfavorable VO [COVES ⩽ 2]	2.56	1.02	6.40	0.045
Proximal occlusion site a [ICA/proximal M1]	1.59	0.67	3.74	0.291
Baseline ischemic core volume [CBF < 30%] (per ml)	1.00	0.99	1.01	0.542
Total infarct volume (per ml)	1.00	1.00	1.01	0.767

END_unexplained: early neurological deterioration of unexplained cause; mTICI: modified Thrombolysis in Cerebral Infarction; AIS-LVO: acute ischemic stroke due to large vessel occlusion; ASPECTS: Alberta Stroke Program Early CT score; NIHSS: National Institutes of Health Stroke Scale; VO: Venous Outflow; COVES: Cortical Vein Opacification Score; ICA: Internal Carotid Artery; M1: M1-segment of middle cerebral artery; CBF: cerebral blood flow perfusion map.

Three hundred ninety-seven patients included in multivariable logistic regression model. Statistical significance: p < 0.05.

^aA proximal occlusion site was defined as an occlusion of the internal carotid artery or proximal first segment of the middle cerebral artery. The proximal first segment of the middle cerebral artery was defined as the first third before the lenticulostriate branches emerge.

The adjusted odds ratios and confidence intervals of independent variables associated with unfavorable functional outcome at 90 days are displayed in Table 3. Please note that unfavorable VO (aOR [95% CI], 4.09 [2.29–7.32]; p < 0.001) and END_unexplained (aOR [95% CI], 6.25 [2.06–18.94]; p = 0.001) were strongly correlated with unfavorable functional outcome at 90 days. The adjusted odds ratios and confidence intervals between Table 3 and the imputed Supplemental Table 2 were highly comparable, with the exception that a proximal vessel occlusion was no longer significant (aOR [95% CI], 1.64 [0.99–2.70]; p = 0.053) after imputing values.

Table 3.

Multivariable logistic regression analysis to predict an unfavorable functional outcome (mRS90 3–6) at 90 days in AIS-LVO patients.

Independent variables	Adjusted odds ratio	95% Conf. interval		p-Value
Age [per 10 years]	1.66	1.34	2.04	<0.001
Male sex	0.92	0.54	1.57	0.769
Administration of tPA	0.84	0.49	1.44	0.533
Admission NIHSS [per point]	1.12	1.07	1.17	<0.001
Complete recanalization [mTICI 2c-3]	0.66	0.39	1.11	0.116
Unfavorable pial arterial collaterals	1.45	0.77	2.73	0.247
Unfavorable VO [COVES ⩽ 2]	4.09	2.29	7.32	<0.001
Proximal occlusion site a [ICA/proximal M1]	1.72	1.00	2.96	0.049
ENDunexplained	6.25	2.06	18.94	0.001
Baseline ischemic core volume [CBF < 30%] (per ml)	0.99	0.98	1.00	0.182
Total infarct volume (per ml)	1.02	1.01	1.03	<0.001

AIS-LVO: acute ischemic stroke due to large vessel occlusion; tPA: tissue plasminogen activator; NIHSS: National Institutes of Health Stroke Scale; mTICI: modified Thrombolysis in Cerebral Infarction; VO: Venous Outflow; COVES: Cortical Vein Opacification Score; ICA: Internal Carotid Artery; M1: M1-segment of middle cerebral artery; END_unexplained: early neurological deterioration of unexplained cause; CBF: cerebral blood flow perfusion map.

Four hundred seventy-three patients included in multivariable logistic regression model. Statistical significance: p < 0.05.

^aA proximal occlusion site was defined as an occlusion of the internal carotid artery or proximal first segment of the middle cerebral artery. The proximal first segment of the middle cerebral artery was defined as the first third before the lenticulostriate branches emerge.

Discussion

In this retrospective multicenter cohort study, we investigated the association of collateral imaging biomarkers on END_unexplained in AIS-LVO patients. We observed an END_unexplained incidence of 10% in patients with successful recanalization. Patients with END_unexplained suffered significantly more often from unfavorable functional long-term outcomes. In multivariable logistic regression analysis, unfavorable VO profiles on admission CTA were strongly correlated with END_unexplained.

To date, the scope of most studies with respect to END was on AIS-LVO patients who were treated with intravenous thrombolysis alone. The incidence of END in this patient population varied between 7% and 14%.^30–33 However, there are only a few studies that have focused on END after MT. In these studies, the incidence of END rises up to 40%.^8,10,34 The discrepancy to our findings may be explained by a broader selection criteria of the aforementioned studies, including patients with unsuccessful recanalization and reperfusion hemorrhages. As we aimed to assess factors of END that are less well known, we have excluded patients with unsuccessful recanalization and reperfusion hemorrhages. In this context, research with respect to END_unexplained remains scarce. Recently published studies observed END_unexplained in 4%–7% of AIS-LVO patients treated with MT and highlighted several independent clinical predictors.^5,9 However, no information was provided on the cerebral collateral circulation, which may encompass valuable pathophysiological implications.

The strength of our study is the distinctive evaluation of collateral imaging biomarkers with regard to END_unexplained. While two Chinese studies demonstrated a significant association between unfavorable pial arterial collaterals and END in AIS-LVO patients after MT,^4,35 there was no significant association between pial arterial collaterals and END_unexplained in our study. One explanation may be the slightly different inclusion criteria as the Chinese studies included patients with unsuccessful recanalization and reperfusion hemorrhage, both of which are associated with the pial arterial collateral status.^36,37

Previous studies directed greater attention toward the venous drainage as the terminal segment of cerebral perfusion, where VO profiles can be considered as an integral part of collective tissue perfusion.^22,23 In this respect, we found a strong association between deteriorated VO and END_unexplained. Current research highlighted ischemia progression as a conceivable etiology of END. ⁸ VO may serve as a sensitive indicator of ischemia progression by an indirect reflection of impaired cerebral microperfusion and a collective marker of blood supply through the brain tissue. Consequently, patients with impaired cerebral microperfusion, reflected by reduced VO, may be more vulnerable to ischemia progression within the first 24 h after stroke onset. Interestingly, a recently published study using follow-up perfusion imaging observed that approximately one quarter of AIS-LVO patients treated by MT have incomplete microcirculatory reperfusion despite complete macrovascular recanalization (eTICI 2c-3). ³⁸ This so called no-reflow phenomenon is proposed to be due to a microvascular dysfunction and was associated with worse functional outcomes. ³⁸ Unfortunately our study does not provide follow-up perfusion imaging. However, previous research highlighted admission perfusion imaging in the assessment of tissue-level collaterals and found poor VO profiles to be correlated with unfavorable tissue-level microperfusion prior to treatment. ²² Further studies including follow-up perfusion imaging are warranted to investigate the association between unfavorable VO profiles on admission CTA and the no-reflow phenomenon on post-treatment perfusion imaging.

In addition, VO profiles were found to be correlated with baseline ischemic core volume, infarct growth, edema formation and long-term functional outcomes, which supports the aforementioned deliberations.^23,24,39 Another aspect that has been raised with respect to END is an infarct growth beyond the initial penumbra caused by a decrease in local perfusion pressure through distal migration of the thrombus and new embolic events.^5,11,40,41 In this regards, increased interstitial pressure due to edema formation, impaired cerebral autoregulation or thrombosis of the capillary bed and/or draining venules may be indicated by reduced VO and may subsequently result in higher odds of END_unexplained.^24,42,43 Further research is needed to validate these findings.

Numerous studies demonstrated a significant correlation between END and unfavorable functional long-term outcomes.^7,8,10,11 However research on END_unexplained with respect to long-term outcomes remains scarce. ⁹ In line with the aforementioned study, our work reveals a significant association between END_unexplained and unfavorable functional outcomes at 90 days.

Limitations

This study is subject to certain limitations. Our study does not report recognized causes of END such as malignant swelling, early recurrent ischemic stroke due to new embolization and new ischemic infarction, which could have changed patients from END_unexplained to explained END. Malignant swelling has been described as a cause of END in 6%–25% of patients who were treated by intravenous thrombolysis alone. ¹¹ However, the definition of END varied and studies suggest that malignant swelling is more likely a cause of neurological deterioration after more than 24 h ⁵ as it tends to develop later.^44,45 Still, it is conceivable that some patients in the END_unexplained group may have experienced clinical deterioration due to malignant swelling. This also applies to early recurrent ischemic stroke due to new embolization and new ischemic infarction, although the low incidence observed in previous studies makes a substantial impact on our results unlikely.^30,46 Second, the retrospective study design and complete case analysis are prone to selection bias and may reduce generalizability. Third, venous opacification was assessed on single-phase CTA which is known to be highly dependent on bolus and acquisition timing. ²¹ Although we only included patients with full opacification of the sigmoid sinuses, rating on single-phase CTA might curtail the comprehensive evaluation of venous outflow during the entire venous phase. In addition, the assessment of VO reflects the main venous drainage of the middle cerebral artery territory and does not account for more regionally specific effects. Early-onset comorbidities such as infectious diseases or cardiorespiratory adverse events that could affect the NIHSS score were not recorded. As NIHSS data were only available after 24 h, no statement can be made regarding END at a later time, although END has been reported up to 72 h after treatment. Our study did not specifically assess complications of endovascular treatment, even though they may contribute to END regardless of final mTICI grade.

Conclusion

AIS-LVO patients with unfavorable VO profiles on admission CTA exhibit higher odds of suffering from END_unexplained after successful MT. Furthermore, the presence of END_unexplained after MT is significantly associated with worse functional outcomes after 90 days. These findings may have important pathophysiological implications for the occurrence of END_unexplained (e.g. through ischemia progression beyond the initial penumbra via hampered cerebral microperfusion, reflected by VO). In addition, the assessment of VO profiles, may help to determine patients at risk of END_unexplained in order to triage intensified monitoring and guide to optimal treatment regimes.