The early hemorrhagic transformation after reperfusion therapy of acute ischemic stroke

Full Text

Introduction

Reperfusion therapy — intravenous thrombolytic therapy (IVT) and endovascular thrombectomy (EVT) — is the most effective treatment for ischemic stroke (IS) patients when administered within the first 24 hours of the disease onset in carefully selected patients [1, 2].

The thirty-year history of IVT implementation into clinical practice from an evidence-based medicine perspective began in 1995 following the publication of results from the NINDS randomized placebo-controlled trial (RCT), which demonstrated the safety and efficacy of IVT with recombinant tissue plasminogen activator (alteplase, rt-PA) within the first 3 hours of symptom onset [3]. The ECASS III study demonstrated the safety and efficacy of IVT with rt-PA in the 3.0–4.5 hour window after ischemic stroke onset [4]. Subsequent meta-analyses have shown that earlier therapy initiation proportionally correlates with greater IVT benefit. IVT significantly increases the likelihood of absence of disabling IS symptoms at 3 months follow-up [5, 6].

The successfully completed randomized WAKE-UP trial and a meta-analysis of EPITHET, ECASS-4, and EXTEND studies demonstrated the feasibility of extending the therapeutic window for IVT with rt-PA up to 9 hours and/or administering IVT to patients with unknown symptom onset time (including wake-up strokes) when specific neuroimaging protocols are followed [7, 8].

In 2020, Russia authorized a domestic agent for IVT in patients within the first 4.5 hours of IS — a recombinant protein containing the amino acid sequence of non-immunogenic staphylokinase (nSta), which demonstrated equivalent efficacy in achieving good functional outcomes and comparable safety profile to alteplase in the randomized multicenter FRIDA trial [9].

The last decade has seen rapid development of endovascular treatment technologies for IS. Results from five multicenter RCTs published in 2015 demonstrated the advantage of EVT over conservative therapy [10–14]. The main indications for selecting IS patients for EVT, identical to the inclusion criteria of RCTs, are:

• Age > 18 years;
• Occlusion of a major intracranial artery in the carotid basin confirmed with angiography;
• Neurological deficit severity ≥6 points on the NIHSS stroke scale;
• Preserved brain parenchyma with ASPECTS score ≥6 on non-contrast CT;
• Time window ≤6 hours from symptom onset;
• No significant pre-stroke disability.

It is important that patients admitted within the 0–4.5 hour therapeutic window without contraindications received intravenous IVT with rt-PA, even if EVT was planned.

The publication of results from the DAWN and DEFUSE 3 trials expanded the therapeutic window for endovascular interventions to 6–24 hours from symptom onset in IS patients meeting inclusion criteria [15, 16].

The Russian scientific and practical community has developed algorithms for reperfusion therapy in IS using both IVT and EVT (and their combination), reflected in the clinical guidelines “Ischemic Stroke and Transient Ischemic Attack”¹.

The rate of IVT administration in Russia has increased more than 5-fold since 2015, reaching 10.2% of all hospitalized IS patients in Q1 2025. Concurrently, the number of EVT procedures increased 26-fold during 2015–Q1 2025, with 2,982 (2.8%) procedures performed in Q1 2025 [17].

Hemorrhagic transformation (HT) of cerebral infarction remains one of the most frequent complications of both IVT and EVT [18, 19], warranting its inclusion as a safety endpoint in RCTs of reperfusion therapies. HT is classified by timing (early [≤ 24 hours] vs delayed) and clinical correlation (asymptomatic vs symptomatic). The ECASS II definition of symptomatic HT — any hemorrhage associated with ≥4-point NIHSS worsening or death — remains most widely used [21]. However, HT relationship to neurological deterioration may be unclear, particularly following EVT.

According to reperfusion therapy protocols, repeat computed tomography (CT) or magnetic resonance imaging (MRI) of the brain is performed 24 hours after the procedure to assess HT. P. Trouillas et al. proposed classifying HT into hemorrhagic infarctions (HI) and parenchymal hematomas (PH) [22]. In the ECASS II study, this classification was refined: isolated petechial hemorrhages within the ischemic brain tissue area were designated as HI type 1 (HI1), confluent petechiae as HI type 2 (HI2); PH occupying ≤ 30% of the infarct volume as PH type 1 (PH1), and PH occupying > 30% of the infarct volume and/or extending beyond the ischemic zone as PH type 2 (PH2) [21]. The Heidelberg classification, particularly relevant for EVT patients, defines three HT classes: HI or PH confined to the infarct region; PH within and beyond the infarct zone; PH outside the infarct area or intraventricular/subarachnoid/subdural hemorrhage [23].

Literature reports HT incidence of 6.45–31.77% (symptomatic: 1.27–15.75%) after IVT with rt-PA and 7.60–49.55% (symptomatic: 1.52–20.89%) after EVT [24], varying by study population and HT definition [24–27], with high-field MRI verification potentially increasing detection rates [25].

The relationship between HT and functional outcome in IS remains debated. In systemic IVT-treated patients, only PH (predominantly PH2) occurrence correlated with poor functional outcomes [28, 29], while HI clinical significance remains contentious. HI is traditionally viewed as a marker of successful reperfusion and predictor of favorable outcome [30], yet its development may adversely affect long-term prognosis [18].

This study aims to analyze early HT frequency following IVT (administered within 0–4.5 hours post-IS) and EVT (performed within 6 hours post-IS) in carotid territory IS patients, and evaluate its impact on functional outcomes after reperfusion therapies.

Materials and methods

Medical data of 50 patients with IS were retrospectively analyzed. Systemic IVT with rt-PA (0.9 mg/kg, 10% as intravenous bolus, 90% as intravenous infusion over 1 hour) or nSta (10 mg intravenous bolus) was administered to 204 patients (126 males (61.8%) and 78 females (38.2%); mean age 65.1 years) hospitalized at the Russian Center of Neurology and Neurosciences (Moscow) between 2008–2013 (n = 69) and at the Regional Vascular Center of Central Clinical Medical Sanitary Unit (Magnitogorsk) between 2022–2025 (n = 135). Endovascular EVT or combination of IVT and EVT were performed in 316 patients (168 males (53.2%) and 148 females (46.8%); mean age 68.6 years) treated at the Regional Vascular Center of Central Clinical Medical Sanitary Unit (Magnitogorsk) between 2019–2025. Considering study objectives, data from several patients were excluded from analysis (Fig. 1). The algorithm for selecting reperfusion methods in patients included in the study is presented in Fig. 2.

 

Fig. 1. Scheme for the selection of patients with ischemic stroke for inclusion in the study. One patient in the TLT group and six patients in the TE group had no data for the 90-day visit.

 

Fig. 2. Algorithm for selecting reperfusion therapy methods in examined patients. 1Age exceeded limits per clinical guidelines but was accepted due to licensing requirements of the healthcare facility where treatment was performed. 2MRI was performed in standard sequences, including diffusion-weighted and T2* sequences. 3Disabling IS was defined as presence of disorders that, if persistent, would prevent patient from performing self-care activities (bathing, moving, toileting, personal hygiene, and eating) or returning to work, or any focal symptoms considered disabling for individual patients based on occupation and lifestyle specifics. 4IVT with rt-PA. 5For patients treated before 2015, IVT with rt-PA was performed in case of no contraindications, regardless of presence or absence of large vessel occlusion on angiographic imaging. 6Additional criteria included neurological deficit severity ≥ 6 NIHSS points, ASPECTS ≥ 6 on CT scan, and pre-stroke mRS score 0–2. 7IVT with rt-PA or recombinant protein containing nSta amino acid sequence. 8Thrombectomy was performed in case of contraindications for IVT (e.g., use of direct oral anticoagulants)..

 

EVT was performed using stent-retriever thrombectomy in 174 patients (69.3%), aspiration techniques in 18 patients (7.2%), and a combination of both in 59 patients (23.5%). Recanalization during EVT was assessed using the modified Thrombolysis in Cerebral Infarction (mTICI) scale for intracranial arterial blood flow restoration [31]. The technical EVT goal was to achieve reperfusion of mTICI 2b–3 (achieved in 192 patients (76.5%)). In 59 patients (23.5%), lower reperfusion grades (mTICI 0–2a, hereafter referred to as mTICI < 2b) were observed.

All patients underwent clinical and instrumental examination according to current standards of medical care and clinical guidelines. The IS subtype was classified using the SSS-TOAST criteria [32].

Neurological deficit severity was assessed with the National Institutes of Health Stroke Scale (NIHSS) at hospital admission; functional outcomes were evaluated using the modified Rankin Scale (mRS) at 10–14 days of hospitalization (at discharge/transfer to secondary rehabilitation) and at day 90 after IS (via in-person consultation or structured telephone interview). An mRS score of 0–2 at discharge and 0–1 at day 90 defined a good functional outcome, while 4–6 at discharge and 5–6 at day 90 defined a poor outcome.

HT was assessed during repeat neuroimaging (CT or MRI) on day 2 post-IS using the HT classification from the ECASS II RCT [21].

Statistical analysis was performed using R programming language v.4.4.1 in RStudio v. 2025.05.1. Median, upper and lower quartiles were calculated for continuous variables, and frequencies for discrete variables. The Mann–Whitney U test was used for continuous variables and Pearson’s χ² test for discrete variables when comparing two independent groups. Predictor significance analysis employed univariate and multivariate logistic regression with odds ratio (OR) and 95% confidence interval (CI) calculations. The significance level was set at 0.05, and all tests were two-sided.

Results

The study included 191 patients with IS within the first 4.5 hours of neurological symptom onset who received systemic IVT, and 251 patients with IS within the first 6 hours of neurological symptom onset who underwent endovascular EVT. Key clinical characteristics of the study participants are presented in Table 1. Patients who underwent EVT were older than those receiving IVT, and the groups also differed in sex distribution.

 

Table 1. Key clinical characteristics of the patients examined

Parameter	IVT				EVT
	rt-PA	nSta	р	IVT (all patients)	mTICI 2b–3	mTICI < 2b	р	EVT (all patients)	р
	n = 101	n = 90		n = 191	n = 192	n = 59		n = 251
Age, years (Ме [Q1; Q3])	64 [57; 72]	68 [59; 75]	0.075	66 [58; 73]	70 [63; 77]	73 [67.5; 77.5]	0.083	70 [64.5; 77.0]	< 0.001
Male, n (%)	66 (65.3)	53 (58.9)	0.358	119 (62.3)	101 (52.6)	24 (40.7)	0.101	125 (49.8)	0.009
Left hemisphere affected, n (%)	59 (58.4)	44 (48.9)	0.186	103 (53.9)	98 (51.0)	32 (54.2)	0.668	130 (51.8)	0.654
NIHSS at admission (Ме [Q1; Q3])	12 [8; 16]	9 [5; 12]	0.003	10.0 [6.5; 15.0]	16 [13; 19]	17 [14.0; 18.5]	0.99	16.0 [14.0; 19.0]	< 0.001
Distribution by IS subtype (SSS-TOAST), n (%)
Atherothrombotic	27 (26.7)	18 (20.0)	0.274	45 (23.6)	54 (28.1)	13 (22.0)	0.355	67 (26.7)	0.457
Cardioembolic	49 (48.5)	37 (41.1)	0.305	86 (45.0)	97 (50.5)	32 (54.2)	0.617	129 (51.4)	0.185
Lacunar	5 (5.0)	6 (6.7)	0.611	11 (5.8)				0	< 0.001*
IS of another established etiology	10 (9.9)	13 (14.4)	0.336	23 (12.0)	10 (5.2)	6 (10.2)	0.172	16 (6.4)	0.048
IS of unspecified etiology	10 (9.9)	16 (17.8)	0.113	26 (13.6)	31 (16.1)	8 (13.6)	0.631	39 (15.5)	0.571

Note. Mann–Whitney U test; Pearson’s chi-squared test; *for comparison of indicators with frequency of 0% or 100%, Fisher’s exact test was used.

 

The clinical presentation in examined patients included impaired consciousness leveland focal neurological symptoms caused by lesions in one of the cerebral hemispheres. Compared to patients receiving intravenous IVT, those who underwent EVT had higher NIHSS scores, greater frequency of consciousness impairment, more severe motor deficits (hemiplegia present in more than half of cases), deep sensory disturbances, and oculomotor disorders (gaze palsy) either isolated or combined with forced head turning toward the affected hemisphere (Table 2).

 

Table 2. Principal clinical IS manifestations, n (%)

Symptom	IVT (n = 191)	EVT (n = 251)	р
Symptom	49 (25.7)	98 (39.0)	0.003
Decreased level of consciousness (from stupor to sopor)	183 (95.8)	251 (100.0)	0.002
Central hemiparesis	49 (25.7)	132 (52.6)	< 0.001
Hemiplegia	3 (1.6)	0	0.06*
Central monoparesis of the arm	188 (98.4)	251 (100.0)	0.03
Central paresis of facial muscles	98 (51.3)	128 (51.0)	0.95
Aphasia of varying severity	93 (48.7)	123 (49.0)	0.95
Dysarthria	98 (51.3)	140 (55.8)	0.356
Disorders of superficial sensitivity (hemihypalgesia, hemianalgesia)	68 (35.6)	119 (47.4)	0.0126
Disorders of deep sensitivity, neglect syndrome	23 (12.0)	66 (26.3)	< 0.001
Oculomotor disorders (gaze restriction, gaze paresis)	19 (9.9)	32 (12.7)	0.363
Hemianopsia

Note. Mann–Whitney U test; Pearson's chi-squared test; *for comparison of indicators with frequency of 0% or 100%, Fisher's exact test was used.

 

The distribution of patients by IS subtypes was comparable between groups receiving both types of reperfusion therapy. The exception was the lacunar IS subtype, for which EVT was not performed.

Patients who underwent IVT with rt-PA (n = 101) had more severe neurological deficit compared to patients receiving IVT with nSta (90 patients), with comparable distributions by sex, age, brain infarction lateralization, and IS subtypes.

Patients undergoing endovascular EVT showed varying degrees of intracerebral artery blood flow restoration: 192 patients achieved optimal reperfusion (mTICI 2b–3), while 59 patients had poorer angiographic recanalization outcomes. M1 segment occlusion of the middle cerebral artery predominated in both EVT subgroups. The most common EVT method was stent-retriever thrombectomy with single pass. Patients with mTICI <2b more frequently underwent simultaneous internal carotid artery stenting compared to those with mTICI 2b–3 (12 (20.3%) vs 6 (3.1%) patients respectively, p < 0.001).

Functional outcomes of examined patients are shown in Fig. 3. Patients receiving IVT with rt-PA showed a trend toward better functional outcomes at discharge and at day 90 compared to IVT with nSta, though statistically significant differences were only demonstrated for mRS 4–6 scores at discharge (p = 0.029). EVT patients with mTICI 2b–3 also showed better functional outcomes both at discharge and at day 90 compared to those with mTICI <2b, though these differences did not reach statistical significance.

 

Fig. 3. Functional outcomes in patients with IS among examined groups.

 

Early HT during IVT is reported in 34 (17.8%) patients, including symptomatic HT in 10 (5.2%). Endovascular EVT was associated with more frequent HT compared to IVT, occurring in 79 (31.5%) patients (p = 0.001).

HT structure is shown in Figure 4. Parenchymal hematomas were more frequently recorded in the EVT group compared to IVT (p < 0.001).

 

Fig. 4. Frequency of early HT (according to ECASS II classification) in the examined patients. *p = 0.045 (Pearson’s χ2 test) when comparing with the IVT with rt-PA group; **p < 0.001 (Pearson’s χ2 test) when comparing IVT and EVT groups.

 

The frequency of early HT in patients receiving IVT with rt-PA and nSta was similar, with no significant differences observed in the rate of symptomatic HT (5 (5%) and 5 (5.6%) patients in both groups, respectively; p = 0.851). However, patients undergoing IVT with rt-PA showed more frequent development of HI1. Achieving different recanalization levels during EVT did not demonstrate differences in early HT rates (including separately assessed HI and PH).

The predictive significance of HT for functional outcomes in IS was analyzed using binomial logistic regression (Table 3).

 

Table 3. Association of HT following reperfusion and IS functional outcomes (regression analysis)

Parameter	IVT						EVT
	weight	SE	Z	odds ratio	CI	p	weight	SE	Z	odds ratio	CI	p
At discharge
No HT and mRS < 2	0.91	0.44	2.09	2.48	1.06–5.53	0.037	1.80	0.49	3.67	6.07	2.31–15.90	< 0.001
HT and mRS 4–6	0.39	0.38	1.02	1.47	0.70–3.10	0.308	1.32	0.36	3.66	3.76	1.85–7.62	< 0.001
HT and death	1.41	0.51	2.75	4.08	1.50–11.12	0.006	0.78	0.30	2.62	2.19	1.22–3.94	0.009
PH2 and mRS 4–6	2.05	1.09	1.88	7.80	0.92–66.10	0.06	NA					0.06
PH2 and death	3.41	0.88	3.87	30.36	5.39–170.89	< 0.001	1.92	0.52	3.66	6.79	2.43–18.96	< 0.001
PH and mRS 4–6	0.95	0.63	1.51	2.59	0.75–8.91	0.131	1.70	0.62	2.73	5.45	1.61–18.37	0.006
PH and death	2.13	0.65	3.29	8.42	2.36–30.00	0.001	1.20	0.38	3.18	3.31	1.58–6.90	0.001
HI2-PH1 and mRS 4–6	0.28	0.57	0.66	1.46	0.47–5.53	0.51	0.76	0.47	1.59	2.13	0.84–5.39	0.111
HI2-PH1 and death	1.11	0.71	1.57	3.04	0.76–12.18	0.117	0.88	0.39	2.28	2.41	1.13–5.15	0.023
HI and mRS 4–6	0.02	0.46	0.04	1.02	0.42–2.49	0.966	0.79	0.42	1.89	2.20	0.97–4.99	0.059
HI and death	0.41	0.68	0.61	1.51	0.40–5.66	0.541	0.09	0.23	0.82	1.09	0.52–2.27	0.819
Day 90
HT and mRS 5–6	1.20	0.43	2.82	3.32	1.44–7.66	0.005	0.82	0.28	2.94	2.27	1.31–3.92	0.003
No HT and mRS < 1	0.77	0.46	1.69	2.16	0.88–5.28	0.073	2.06	0.54	3.81	7.84	2.72–22.61	< 0.001
HT and death	1.50	0.46	3.29	4.50	1.84–10.99	< 0.001	0.79	0.29	2.78	2.21	1.26–3.86	0.005
PH2 and mRS 5–6	3.50	1.10	3.19	33.21	3.85–286.55	0.001	1.58	0.59	2.68	4.84	1.53–15.31	0.007
PH2 and death	2.96	0.87	3.41	19.29	3.52–105.74	< 0.001	2.00	0.59	3.39	7.38	2.33–23.44	< 0.001
HI and mRS 5–6	0.34	0.55	0.63	1.41	0.48–4.13	0.531	0.25	0.33	0.75	1.28	0.67–2.47	0.456
HI and death	0.72	0.56	1.29	2.06	0.69–6.17	0.197	0.26	0.34	0.76	1.30	0.66–2.55	0.449
HI2-PH1 and mRS 4–6	0.78	0.63	1.23	2.18	0.63–7.54	0.219	0.84	0.38	2.20	2.32	1.09–4.91	0.028
HI2-PH1 and death	1.14	0.64	1.78	3.13	0.89–11.04	0.076	0.58	0.38	1.52	1.78	0.85–3.74	0.13
PH and mRS 5–6	2.06	0.62	3.31	7.83	2.32–26.48	< 0.001	1.20	0.39	3.05	3.32	1.54–7.16	0.002
PH and death	2.07	0.62	3.31	7.90	2.32–26.85	< 0.001	1.08	0.38	2.86	2.93	1.40–6.130	0.004

Note. Here and in Tables 4 and 5: NA indicates that no statistically processed data are available.

 

In the IVT group, absence of HT increased the likelihood of favorable functional outcomes at discharge, though this pattern was not observed at 90-day follow-up. This contrasted with the EVT group, where absence of HT post-endovascular intervention increased the probability of favorable functional outcomes both at discharge and at 90-day follow-up.

Early HT detection following reperfusion therapies increased the probability of unfavorable outcomes at discharge only in the EVT group, while this pattern became evident in both IVT and EVT groups at 90-day assessment. HT also served as a predictor of lethal outcomes in both groups.

HI showed no association with increased probability of unfavorable outcomes or mortality regardless of reperfusion method. Conversely, PH increased unfavorable outcome probability at discharge and 90-day follow-up in EVT patients, while in the IVT group this trend reached statistical significance only for functional outcomes at day 90. Notably, PH significantly increased mortality risk in IS patients regardless of reperfusion method, with PH2 specifically increasing the probability of unfavorable functional outcomes at day 90 as well as mortality at discharge and 90-day follow-up.

HT predicted mortality at discharge and 90-day follow-up, while also increasing the probability of unfavorable outcomes at 90-day follow-up regardless of thrombolytic agent used (Table 4). HI showed no association with increased unfavorable outcomes or mortality, whereas PH increased both unfavorable outcome probability at day 90 and mortality risk at discharge/90-day follow-up when using either IVT agent. PH2 significantly increased unfavorable functional outcome probability at day 90 in rt-PA IVT patients, and elevated mortality risk at discharge/90-day follow-up when using both IVT agents.

 

Table 4. Association between HT after IVT with different thrombolytic agents and IS functional outcomes (regression analysis)

Parameter	IVT rt-PA						IVT nSta
	weight	SE	Z	odds ratio	CI	p	weight	SE	Z	odds ratio	CI	p
At discharge
No HT and mRS < 2	0,82	0,53	1,53	2,26	0,80–6,42	0,125	1,26	0,80	1,56	3,51	0,73–16,95	0,118
HI1 and mRS < 2	0,09	0,64	0,13	1,09	0,31–3,83	0,893	NA
HT and mRS 4–6	0,28	0,50	0,56	1,32	0,50–3,50	0,579	0,79	0,64	1,22	2,19	0,62–7,73	0,222
HT and death	1,50	0,76	1,98	4,45	1,02–19,98	0,048	1,49	0,72	2,07	4,44	1,08–18,22	0,038
PH2 and mRS 4–6	NA						1,01	1,18	0,86	2,73	0,27–27,33	0,392
PH2 and death	3,42	1,30	2,64	30,67	2,42–388,28	0,008	3,38	1,21	2,78	29,25	2,71–315,18	0,005
PH and mRS 4–6	1,99	1,14	1,75	7,29	0,78–67,89	0,081	0,17	0,8	0,21	1,18	0,25–5,61	0,834
PH and death	2,30	1,01	2,29	10,00	1,39–71,77	0,022	1,95	0,85	2,30	7,03	1,33–37,17	0,022
HI2/PH1 and mRS 4–6	0,24	0,79	0,30	1,26	0,267–5,98	0,767	0,60	0,89	0,67	1,82	0,32–10,47	0,503
HI2/PH1 and death	1,77	0,94	1,89	5,87	0,94–36,81	0,059	0,39	1,15	0,34	1,48	0,16–13,99	0,732
HI and mRS 4–6	–0,33	0,58	–0,57	0,72	0,23–2,25	0,567	1,56	1,12	1,40	4,77	0,53–42,56	0,162
HI and death	0,63	0,87	0,73	1,88	0,34–10,28	0,466	0,39	1,15	0,34	1,48	0,16–13,99	0,732
Day 90
HT and mRS 5–6	1,27	0,61	2,08	3,33	1,07–11,74	0,038	1,35	0,63	2,14	3,86	1,12–13,26	0,032
No HT and mRS < 1	0,73	0,56	1,29	2,07	0,69–6,22	0,196	0,97	0,81	1,21	2,64	0,54–12,84	0,228
HT and death	1,78	0,67	2,66	5,92	1,60–21,94	0,008	1,43	0,66	2,16	4,19	1,14–15,37	0,031
PH2 and mRS 5–6	2,65	1,26	2,10	14,17	1,19–168,39	0,036	NA
PH2 and death	2,97	1,27	2,33	19,56	1,61–237,41	0,02	2,92	1,20	2,44	18,50	1,78–192,81	0,015
HI1 and mRS < 1	0,02	0,67	0,03	1,02	0,28–3,74	0,079	NA
HI and mRS 5–6	0,43	0,72	0,59	1,53	0,38–6,25	0,553	0,61	0,91	0,67	1,83	0,31–10,82	0,503
HI and death	0,79	0,74	1,06	2,19	0,51–9,36	0,289	1,00	0,92	1,10	2,73	0,45–16,48	0,273
HI2/PH1 and mRS 5–6	1,72	0,83	2,08	5,59	1,1–28,36	0,038	–0,38	1,13	–0,34	0,68	0,08–6,22	0,736
HI2/PH1 and death	2,08	0,85	2,45	7,97	1,51–42,04	0,014	NA
PH and mRS 5–6	2,44	0,97	2,52	11,46	1,72–76,31	0,012	1,72	0,81	2,12	5,58	1,14–27,47	0,034
PH and death	2,79	0,99	2,84	16,31	2,27–112,4	0,005	1,49	0,83	1,81	4,44	0,88–22,36	0,071

 

Table 5 presents the results of the analysis of the predictive significance of HT following EVT in patients who achieved different levels of recanalization. Of note, only in patients with IS who achieved mTICI 2b-3 reperfusion during EVT, the absence of HT was associated with a favorable functional outcome, while its detection increased the likelihood of an unfavorable functional outcome and death both at discharge and at day 90. Furthermore, exclusively for patients with IS and mTICI 2b–3 after EVT, an association was demonstrated between HI and increased risks of unfavorable outcomes at discharge, and PH was linked to higher probabilities of unfavorable functional outcomes and death at discharge and at day 90. The same trend was observed for PH2.

 

Table 5. Association of HT after EVT with achievement of different reperfusion level (regression analysis)

Parameter	EVT with reperfusion mTICI 2b–3						EVT with reperfusion mTICI 0–2a
	weight	SE	Z	odds ratio	CI	p	weight	SE	Z	odds ratio	CI	p
At discharge
No HT and mRS < 2	1,88	0,55	3,41	6,53	2,22–19,21	< 0,001	1,50	1,10	1,37	4,50	0,52–38,92	0,172
HT and mRS 4–6	1,50	0,42	3,59	4,50	1,98–10,23	< 0,001	0,71	0,72	0,98	2,02	0,49–8,33	0,329
HT and death	0,91	0,35	2,61	2,48	1,25–4,90	0,009	0,43	0,59	0,73	1,54	0,48–4,92	0,468
PH2 and mRS 4–6	NA						NA
PH2 and death	1,89	0,59	3,22	6,63	2,10–20,95	0,001	2,08	1,19	1,74	8,00	0,77–83,02	0,082
PH and mRS 4–6	1,54	0,63	2,44	4,68	1,36–16,17	0,015	NA
PH and death	1,20	0,43	2,82	3,31	1,44–7,62	0,005	1,29	0,83	1,56	3,62	0,72–18,24	0,119
HI2-PH1 and mRS 4–6	1,08	0,57	1,91	2,95	0,97–8,98	0,056	–0,29	0,09	–0,32	0,75	0,13–4,37	0,749
HI2-PH1 and death	1,17	0,44	2,67	3,21	1,36–7,56	0,008	–0,11	0,90	–0,12	0,90	0,16–5,14	0,906
HI and mRS 4–6	1,10	0,51	2,13	2,99	1,09–8,19	0,033	–0,09	0,75	–0,12	0,92	0,21–3,99	0,908
HI and death	0,23	0,44	0,52	1,26	0,54–2,95	0,6	–0,34	0,74	–0,46	0,71	0,17–3,01	0,643
Day 90
HT and mRS 5–6	0,77	0,32	2,41	2,16	1,16–4,05	0,016	0,98	0,59	1,66	2,68	0,84–8,53	0,096
No HT and mRS < 1	2,22	0,62	3,58	9,23	2,73–31,17	< 0,001	1,40	1,11	1,26	4,07	0,46–35,75	0,206
HT and death	0,90	0,33	2,74	2,45	1,29–4,66	0,006	0,46	0,58	0,78	1,58	0,05–4,92	0,434
PH2 and mRS 5–6	1,74	0,68	2,57	5,69	1,51–21,44	0,01	1,06	1,19	0,89	2,89	0,28–29,58	0,371
PH2 and death	2,05	0,68	3,02	7,74	2,05–29,24	0,003	1,85	1,19	1,55	6,35	0,62–65,66	0,121
HI and mRS 5–6	0,20	0,39	0,52	1,22	0,57–2,64	0,606	0,29	0,66	0,44	1,34	0,37–4,85	0,657
HI and death	0,38	0,40	0,96	1,46	0,67–3,20	0,339	–0,10	0,69	–0,14	0,91	0,24–3,48	0,886
HI2-PH1 and mRS 4–6	0,85	0,43	1,98	2,34	1,01–5,43	0,047	0,92	0,88	1,04	2,50	0,44–14,12	0,3
HI2-PH1 and death	0,82	0,43	1,91	2,26	0,98–5,22	0,056	-0,34	0,89	–0,38	0,71	0,13–4,05	0,701
PH and mRS 5–6	1,12	0,43	2,59	3,05	1,31–7,09	0,01	1,87	1,12	1,68	6,50	0,73–57,99	0,094
PH and death	1,09	0,42	2,57	2,97	1,29–6,81	0,01	1,04	0,82	1,27	2,83	0,57–14,18	0,205

 

Discussion

In this study, a retrospective cohort of patients with IS who underwent reperfusion therapy demonstrated the frequency and prognostic significance of early hemorrhagic transformation in cerebral infarction under real-world clinical practice conditions.

In our study, the efficacy of IVT with rt-PA at day 90 was 36%, with mortality observed in 11% of cases. In the meta-analysis by J. Emberson et al., IVT was effective in 34.4% of patients, with mortality recorded in 18.8% [6]. Thus, IVT with rt-PA efficacy parameters were comparable with RTC results, while the mortalityfrequencyy was slightly lower.

In our study, the efficacy of IVT with nSta at day 90 was 30%, with mortality observed in 16.7% of cases. In the FRIDA study, these rates were 50% and 10%, respectively [9]. Thus, our data show inferior outcomes, which may be explained by the small sample sizes in both our study and the FRIDA study, as well as differences between real-world clinical practice and RCT setting.

In our study, the efficacy of EVT in patients with IS in the carotid system at day 90 was 21.8%, with mortality observed in 34% of cases. In the meta-analysis of key RCTs on EVTA by M. Goyal et al., 26.9% of patients achieved a functional outcome of 0–1 on the mRS scale, with 15.3% mortality, despite comparable mean age and severity of neurological deficits [35]. Thus, the efficacy and safety of EVT in our study were inferior to those in the meta-analysis of 5 EVT RCTs. The differences in post-EVT mortality rates in the real-world clinical practice (even when adjusted for variations in cerebral artery occlusion patterns between our study and the reference meta-analysis) largely served as the rationale for this study.

The data on HT frequency after IVT with rt-PA (see Fig. 4) align with international findings [24, 26]. However, according to the Chinese CASE II thrombolysis registry of over 13,000 IS patients, the incidence of specific HT types was lower: 4.1% for HI and 3.4% for PH [36]. Conversely, the HT frequency with IVT with nSta was lower than in the hospital registry of IS patients by A.A. Kulesha et al., where HI developed in 8% and PH in 7% of cases during TNT [37].

An interesting observation in our study was the lower incidence of HI1 in patients receiving IVT with nSta compared to IVT with rt-PA. This may be explained by both the longer exposure time of the thrombolytic agent in the area of cerebral ischemic damage during IVT with rt-PA (administered intravenously over 1 hour versus bolus administration of nSta) and the specific properties of rt-PA, which has additional mechanisms of blood-brain barrier disruption through activation of matrix metalloproteinases [38].

A key finding of our research is the negative impact of early HT on functional outcomes in IS patients undergoing IVT (see Table 3), regardless of the thrombolytic agent used (see Table 4). The prospective monocenter study by M. Annan et al. revealed similar results (with the caveat of broader definitions of “unfavorable outcome” and HT verification methods in French colleagues): HT increased the likelihood of unfavorable outcome (OR = 4.6; 95% CI 1.9–11.4; p = 0.001) [25].

HI showed no association with either increased probability of adverse outcomes/mortality in I patients receiving IVT with both agents, nor with achieving good functional outcomes. Earlier studies hypothesized HI role as a recanalization marker during systemic IVT with rt-PA [39]. A key study limitation was its sample of 32 IS patients with proximal MCA occlusion receiving 3-hour IVT with rt-PA. Notably, not all patients showed cerebral artery occlusion pre-IVT, suggesting diverse reperfusion mechanisms beyond arterial recanalization [40].

Both PH and PH2 increased adverse outcome risks and mortality regardless of thrombolytic used, with minimal differences in functional outcomes at discharge/day 90 (Table 3). CASE II registry data confirm PH negative impact on post-IVT outcomes: this HT type correlated with poor functional recovery (OR = 3.61; 95% CI 2.54–5.13; p < 0.001) across definition variations [36]. Conversely, D. Gill et al. observational study linked only PH2 to neurological deterioration post-IVT with rt-PA [41].

Our study found HI incidence of 17.6% and PH rate of 14% EVT. Comparable HT rates were reported in Jazayeri meta-analysis: 33.6% with EVT alone, rising to 37.9% with EVT+IVT combination (OR = 1.1; 95% CI 1.00–1.21; p = 0.52) [42]. Symptomatic HT post-EVT was intentionally not assessed, as clinical deterioration (including fatal outcomes) after endovascular intervention involves more complex mechanisms than post-IVT cases.

The frequency of HT in patients with IS after EVT was higher than after IVT, which is also supported by literature data [24]. These differences may be attributed, first, to patient selection criteria for EVT, such as occlusion of a major cerebral artery and the presence of severe neurological symptoms, which in most cases reflect a larger volume of brain tissue damage [17]. Indeed, in our study, the severity of neurological impairments was significantly higher in the EVT group than in IS patients who underwent IVT. Second, the instruments used for EVT (primarily stent retrievers) may cause micro-injuries to the vascular wall, especially during repeated tractions [43]. The most common EVT technique in our study was also thrombectomy using stent retrievers. Third, EVT is performed later than IVT [44]. Fourth, the use of general anesthesia in some cases, as well as simultaneous carotid stenting [45], may also contribute to HT risks. Thus, 18 patients (7.2%) in our study underwent simultaneous internal carotid artery stenting with EVT.

An important finding of our study is the demonstration of early HT impact on functional outcomes in IS patients who underwent EVT. Specifically, absence of HT after endovascular intervention increased the likelihood of favorable functional outcomes; conversely, HT detection increased the odds of unfavorable functional outcomes and death by day 90. HI did not affect negative IS outcomes, whereas PH and especially PH2 increased the probability of unfavorable functional outcomes and death by day 90 (see Table 2).

The literature data on the impact of HT on IS outcomes after EVT remain limited. X. Yu et al. note that HT serves as a substrate for additional (to ischemic) brain tissue damage, thereby contributing to a vicious cycle of complications and increasing risks of adverse outcomes and patient mortality [46]. At the same time, Y.B. Lee et al. indicate that only PH is an independent predictor of poor functional outcome in IS after EVT (OR = 10.15; 95% CI 3.26–31.63; p < 0.001) [47]. Of particular interest are the results from the TITAN database analysis investigating EVT outcomes in tandem occlusions, which demonstrated no impact of HI on IS outcomes, unlike PH that was associated with increased 90-day mortality rates in IS (adjusted OR = 2.63; 95% CI 1.05–6.59; p = 0.039) [48].

The most significant finding in our study appears to be the assessment of HT impact on IS outcomes at different reperfusion achievement levels following EVT: despite comparable detection rates of both HT and separately HI and PH in IS patients with varying reperfusion grades, associations between HT and functional IS outcomes were demonstrated only in patients achieving mTICI 2b–3 recanalization levels (see Table 4).

The key factor underlying HT following successful EVT with target cerebral blood flow restoration (mTICI 2b–3) is reperfusion injury: high-pressure arterial blood influx into extensive areas of developing cerebral infarction with compromised blood-brain barrier promotes intracranial hemorrhage. An additional damaging factor is the blood flow restoration velocity, significantly higher in mTICI 2b–3 cases causing abrupt pressure surges on weakened vascular walls, compared to mTICI 2a where gradual pressure increase during slow/incomplete recanalization occurs. This concept is supported by P. Steen et al. study demonstrating that patients with higher reperfusion grades (mTICI > 2b) had increased symptomatic HT rates (31.8% vs 10.3% in mTICI 2b group, p = 0.031) despite better post-stroke recovery trends [49]. Finally, patients achieving superior reperfusion often share similar baseline risk profiles with those having suboptimal reperfusion outcomes. These HT risk factors remain equally relevant regardless of mTICI scores: anticoagulant therapy (warfarin/DOACs), hyperglycemia (endothelial damage), hypertension (elevated reperfusion zone pressure), and large ischemic core volume at recanalization.

However, in cases of low recanalization (mTICI 0–2a), the primary mechanism of HT is ongoing ischemia triggering a cascade of pathological processes that disrupt the blood–brain barrier (activation of matrix metalloproteinases, oxidative stress, etc.). In these cases, collateral blood flow insufficient to restore brain tissue viability contributes to hemorrhagic imbibition in areas of extensive ischemia (with subsequent evolution into other forms of HT). Technical aspects of EVT procedures such as microinjuries to vascular walls from stent retrievers or simultaneous stenting in tandem occlusions may also play a role in HT within this group.

The observed differences in HT impact on ischemic stroke prognosis between mTICI 2b–3 and mTICI < 2b recanalization after EVT are explainable: successful recanalization restores blood flow in the penumbra zone containing viable brain tissue surrounding the infarct core. Blood components penetrating areas meant for recovery cause toxic and mechanical damage, negating the positive effects of successful reperfusion. Conversely, when recanalization fails during EVT, HT occurs in areas of irreversible ischemic damage without fundamentally exacerbating brain tissue injury.

Conclusion

The selection of early HT (at the expense of delayed HT) in our study as a prognostic factor for the efficacy and safety of reperfusion therapies is based on the fact that most symptomatic (including fatal) HTs occurring during IVT with rt-PA develop within the first 24 hours of IS [27]. On one hand, this constitutes a limitation of the present study since we did not account for the contribution of late HT to long-term IS outcomes; on the other hand, this choice can be considered justified based on real-world clinical practice setting.

The data presented in this study on the frequency of early HT and its association with functional outcomes in IS patients undergoing IVT and EVT are important both for understanding reperfusion mechanisms in these IS treatment methods, and for further exploration of ways to prevent hemorrhagic complications of IVT and EVT, as well as for developing adjuvant therapies for reperfusion injury.

 

¹ Ischemic Stroke and Transient Ischemic Attack: National Clinical Guidelines, 2024. URL: https://cr.minzdrav.gov.ru/preview-cr/814_1

About the authors

Maksim A. Domashenko

M.V. Lomonosov Moscow State University

Author for correspondence.
Email: mdomashenko@gmail.com
ORCID iD: 0009-0000-3630-6130

Cand. Sci. (Med.), Head, Department of neurology and neurosurgery, Faculty of Medicine

Russian Federation, Moscow

Mark A. Loskutnikov

M.V. Lomonosov Moscow State University

Email: losmark@mail.ru
ORCID iD: 0000-0003-0179-2735

Cand. Sci (Med.), Head, Department of neurology for acute stroke patients, Central Clinical Hospital, chief stroke neurologist

Russian Federation, Moscow

Viktor I. Konstantinov

Central Clinical Medical Sanitary Hospital

Email: bluesbart@mail.ru
ORCID iD: 0000-0003-1681-4773

Head, Department of endovascular diagnostics and treatment methods

Russian Federation, Magnitogorsk

Renata Iu. Zaliautdinova

Central Clinical Medical Sanitary Hospital

Email: renatika94@gmail.com
ORCID iD: 0000-0002-7009-735X

neurologist, 2^nd Neurology department

Russian Federation, Magnitogorsk

Yan V. Vishnevskii

Central Clinical Medical Sanitary Hospital

Email: Yanv95@mail.ru
ORCID iD: 0009-0008-9086-7842

neurologist, 1^st Neurology department

Russian Federation, Magnitogorsk

Marina Yu. Maximova

Russian Center of Neurology and Neurosciences

Email: ncnmaximova@mail.ru
ORCID iD: 0000-0002-7682-6672

D. Sci (Med.), Prof., Head, 2^nd Neurology department

Russian Federation, Moscow

Мarine M. Tanashyan

Russian Center of Neurology and Neurosciences

Email: mtanashyan@neurology.ru
ORCID iD: 0000-0002-5883-8119

D. Sci. (Med.), Prof., Full Member of RAS, Deputy director for scientific research work, Head, 1^st Neurology department, Institute of Clinical and Preventive Neurology

Russian Federation, Moscow

References

Supplementary files