This article is mentioned in:

Introduction

Acute ischemic stroke (AIS) is one of the main reasons for death and disability, particularly in an aging society (1). In 61% of stroke patients, administering intravenous thrombolysis with different agents [medical management; urokinase (2.7%), streptokinase (36.8%) and alteplase, recombinant tissue plasminogen activator] within 3-4.5 h of symptom onset has a favorable outcome (2,3). Furthermore, previous studies suggested that medical management is safe and may benefit patients presenting mild AIS symptoms with a low National Institutes of Health Stroke Scale (NIHSS) score (NIHSS score ≤8) and who are also found to have an intracranial large vessel occlusion (LVO) (4-6).

Endovascular thrombolysis (ET), focusing on relieving vessel occlusion in stroke, has been developed as an alternative for medical management or as an adjunct in administration over the past few years (7,8). However, there is limited knowledge about the role of ET in patients with mild symptoms caused by LVO (9). Nonetheless, some studies also proposed that ET may be favorable in patients with severe symptoms (10).

Patients with mild AIS symptoms with an NIHSS score ≤8 also have an intracranial LVO presentation, requiring remarkably complex therapeutic management (11). Unfortunately, LVO in those patients is associated with poor outcomes; they present with mild deficits (12) and because of weak collateral conditions, they can worsen rapidly or progressively (13). Thus, the need for any acute management is often debated since these patients ‘only’ have mild defects and any therapy has related risks (11).

Given these conflicting observational reports and the lack of randomized data, the present study aimed to evaluate the safety and effectiveness of medical management vs. ET therapy among patients with mild AIS symptoms (NIHSS score ≤8) accompanied by LVO.

Materials and methods

Search strategy

The study searched the comparative articles involving mild AIS and LVO, thrombectomy, and medical management through electronic databases, including the Cochrane Library, Medline (1980-2022.12), PubMed (1980-2022.12), and EMBASE (1980-2022.12). Preferred reporting items for systematic reviews and meta-analyses were applied for establishing protocol and manuscript design (14). The keywords ‘mild acute ischemic stroke’, ‘large vessel occlusion’ and ‘thrombectomy/endovascular thrombolysis and medical management’, which are present in the MeSH list, were used for database search.

Selection of studies

Two of the reviewers (GF and VEG) independently extracted data from the included articles, following the guidelines of the epidemiology of meta-analysis. The following basic information was captured: Main authors; year of publication; total case number in the thrombectomy and medical management groups; study type; outcome indicator; diabetes; hypertension; atrial fibrillation; location; prior ischemic stroke; NIHSS of admission. The extracted data were input into a table designed and standardized according to the Cochrane Handbook. Fig. 1 depicts the flow of the study selection process. In case of disagreement on the information to be retained, another author with authority (this authority was decided upon by all authors prior to commencing the research) had the final say.

Figure 1 Flow chart.

Inclusion and exclusion criteria. If an article satisfied the following PICOS criteria, it was considered for inclusion in the present meta-analysis: i) Population limited to patients with mild acute ischemic stroke (NIHSS score ≤8) and LVO on CT angiography [middle cerebral artery M1/M2, intracranial carotid artery (ICA) or vertebrobasilar occlusion]; ii) intervention limited to ET or medical management; iii) studies comparing the outcomes between ET and medical management; iv) one of the primary outcomes, such as morbidity and mortality, was involved. To avoid publication bias, the final aim was to collect a homogenous pool of manuscripts, including articles that compare only two intervention modalities, i.e. ET and medical management. Table I contains detailed data on these articles.

Table I Design and baseline characteristics of the included trials.

Table I

Design and baseline characteristics of the included trials.

	Sample size, n		Mean age (years)		No. of males		Diabetes		Hypertension		Atrial fibrillation
First author/s, year	Throm	Medic	Throm	Medic	Throm	Medic	Throm	Medic	Throm	Medic	Throm	Medic	(Refs.)
Urra et al, 2014	34	44	64	71	24	26	6	6	21	29	9	14	(18)
Haussen et al, 2016	10	22	60	68.2	5	15	2	4	6	13	3	8	(19)
Haussen et al, 2017	30	88	63.5	73	15	41	9	10	21	63	10	16	(20)
Nagel et al, 2018	80	220	65.3	69.5	44	117	13	36	56	159	35	62	(21)
Sarraj et al, 2018	124	90	65.8	65.4	72	51	36	27	87	72	NR	NR	(22)
Kastrup et al, 2018	145	160	71	72	65	67	32	35	122	129	57	74	(23)
Wang et al, 2020	23	24	53.9	64.4	17	17	4	6	17	17	2	2	(24)
Xue et al, 2022	123	449	69	68	82	307	21	78	75	299	25	61	(25)
			Location
	Prior to ischemic stroke		ICA		M1		M2		Vertebrobasilar		NIHSS of admission
First author/s, year	Throm	Medic	Throm	Medic	Throm	Medic	Throm	Medic	Throm	Medic	Throm	Medic	(Refs.)
Urra et al, 2014	5	2	0	1	13	13	7	16	12	12	4	3	(18)
Haussen et al, 2016	2	7	0	3	6	10	1	5	2	3	4	2	(19)
Haussen et al, 2017	NR	NR	2	17	11	17	9	17	7	11	4	3	(20)
Nagel et al, 2018	NR	NR	10	26	33	48	23	108	6	15	4	3	(21)
Sarraj et al, 2018	22	12	21	12	68	21	33	44	0	0	4	3	(22)
Kastrup et al, 2018	NR	NR	32	19	62	76	51	65	0	0	8	7	(23)
Wang et al, 2020	4	9	6	8	10	5	2	2	5	9	3	3	(24)
Xue et al, 2022	19	77	34	110	59	213	20	83	12	12	3	3	(25)
	mRS 0-2 after 3 months		Bleeding		Morbidity		Mortality
First author/s, year			Throm	Medic	Throm	Medic	Throm	Medic	Throm	Medic			(Refs.)
Urra et al, 2014			1	1	4	0	14	14	4	2			(18)
Haussen et al, 2016			10	17	0	0	1	5	0	3			(19)
Haussen et al, 2017			29	64	2	0	NR	NR	NR	NR			(20)
Nagel et al, 2018			68	154	4	3	8	36	3	20			(21)
Sarraj et al, 2018			78	61	17	9	35	28	11	1			(22)
Kastrup et al, 2018			86	89	NR	NR	19	22	NR	NR			(23)
Wang et al, 2020			16	18	2	0	12	5	1	1			(24)
Xue et al, 2022			84	331	10	9	NR	NR	11	24			(25)

[i] M, middle cerebral artery; Throm, thrombectomy; medic, medical management; mRS, modified Rankin scale; NR, not reported.

All prospective and retrospective studies that evaluated at least one of the two modalities were included, whereas editorials, reviews, case reports and articles focusing on the pediatric population, unrelated outcomes, co-morbidities, experimental techniques, or one of the two modalities from that article pool, were excluded. In addition, in the case of studies with mixed or unclear results, the corresponding information was removed or was placed into groups accordingly, although without affecting the ratio.

Outcomes' definition

The primary outcomes involved in the present study included mortality and post-intervention morbidity (Karnofsky Performance Score-KPS <80). In addition, to find out the association between ET and medical management, the present collected outcome measurements (secondary outcomes) such as the rates of modified Rankin scale (mRS) scores 0-2 (indicating a good outcome) at discharge and long-term follow-up (3 months for the interventional cohort and bleeding after the intervention). The outcomes reported in the included articles were assessed at least three months after the ET or medical management for large vessel occlusion strokes with minimal symptoms. Additionally, a quality assessment tool was used to decrease the risk of bias in our poor articles (Newcastle-Ottawa Scale; Table II) (15).

Table II NOS quality assessment of final article pool.

Table II

NOS quality assessment of final article pool.

		NOS
First author/s, year	Study design	Selection	Comparability	Exposure	Total scores	(Refs.)
Urra et al, 2014	Prosp	3	3	3	9	(18)
Haussen et al, 2016	Prosp	3	3	3	9	(19)
Haussen et al, 2017	Retro	3	2	2	7	(20)
Nagel et al, 2018	Prosp	3	3	3	9	(21)
Sarraj et al, 2018	Retro	3	2	2	7	(22)
Kastrup et al, 2018	Retro	3	2	2	7	(23)
Wang et al, 2020	Retro	3	2	2	7	(24)
Xue et al, 2022	Retro	3	2	2	7	(25)

[i] Retro, retrospective; Prosp, prospective; NOS, Newcastle-Ottawa Scale.

Evaluation of the risk of bias

The Cochrane Collaboration tool was used by two reviewers (GF and VEG) to assess the risk of bias in each study (16). The evaluation included random sequence generation, allocation concealment, blinding of participants and assessors, blinding of outcome assessment, incomplete outcome data, selective reporting and other biases (such us criteria of ‘unclear risk’ of bias; this was in the case of insufficient available information to assess whether an important risk of bias exists or insufficient evidence that an identified problem will introduce bias. The assessment results were classified into low, high and unclear risk levels. The third reviewer arbitrated any disagreements.

Data synthesis and assessment of heterogeneity

All analyses were carried out using Review Manager Software v5.4 (RevMan; https://training.cochrane.org/online-learning/core-software/revman). Heterogeneity across trials was identified using I2 statistics; considering I2 >50% as high heterogeneity, a meta-analysis was conducted using a random-effect model according to the Cochrane Handbook for Systematic Reviews of Interventions (v5.1.0) (17). Otherwise, a meta-analysis using the fixed-effect model was performed. The continuous outcomes were expressed as a weighted mean difference with 95% confidence intervals (CIs). For discontinuous variables, odds ratios (OR) with 95% CIs were applied for the assessment. To test the sensitivity, the ‘leave out one’ model was used by removing one study at a time. P<0.05 was considered to indicate a statistically significant difference.

Results

Articles meeting criteria

A total of eight articles met the eligibility criteria (18-25). The total number of patients was 569 who underwent ET and 1097 with medical management for LVO strokes with minimal symptoms. The study sample was based on eight studies (Table II). Among these, five were retrospective and three were prospective observational studies.

Epidemiological and clinical features

The mean age of the patients was 70.46 years (range, 53-72 years; 64.06 and 76.93 years for the ET and the medical management sample, respectively). The male-to-female ratio was 1:1.7. Diabetes was diagnosed in 325/1666 (19.5%) of patients (123/569 and 202/1097 in the ET and the medical management group, respectively), hypertension in 1186/1666 (71.1%) of patients (405/569 and 781/1097 in the ET and medical management groups, respectively), and atrial fibrillation in 378/1666 (22.6%) of patients (141/569 and 237/1097 in the ET and medical management groups, respectively) (Table I).

Diabetes mellitus

Information regarding diabetes mellitus was available in eight articles (18-25) and demonstrated no statistical result (OR 1.07, CI 95% 0.58-1.40; P=0.61) with no heterogeneity (P=0.56 and I2=0%) (Fig. S1). Diabetes mellitus was found in 325 of the 1666 (19.5%) patients in the total group. In the ET group, diabetes mellitus was diagnosed in 123 of 569 (21.6%) patients, compared with 202 of 1097 (18.4%) in the medical management group. Looking at the funnel plot of the same parameter, no publication bias was found.

Hypertension

As regards hypertension, information was available in eight articles (18-25) and demonstrated no statistical result (OR 1.87, CI 95% 0.68-1.10; P=0.23) with no heterogeneity (P=0.83 and I2=0%) (Fig. S2). Hypertension was found in 1186 of 1666 (71.1%) patients in the total group. In the ET group, hypertension was diagnosed in 405 of 569 (71.1%) patients, compared with 781 of 1097 (71.1%) in the medical management group. Looking at the funnel plot of the same parameter, no publication bias was found.

Atrial fibrillation

Atrial fibrillation data were available in eight articles (18-25). Atrial fibrillation was reported for 378 of the 1666 (22.6%) patients in the total group and 141 of the 569 (28.2%) patients in the ET group. The statistical analysis demonstrated no statistically significant difference (OR 1.27, CI 95% 0.85-1.91; P=0.24), but indicated heterogeneity (P=0.07 and I2=48%) (Fig. 2A). By testing sensitivity (Table III), after removing the article by ‘Kastrup et al, 2018’ (23), a statistically significant difference result was found (OR 1.61, CI 95% 1.18-2.21; P<0.05), with no heterogeneity (P=0.52 and I2=0%) (Fig. 2B). The total sample size for this combination was 247 (18.1%) out of 1361, of which 84 (19.8%) out of 424 were from the ET group and 163 (17.3%) out of 937 were from the medical management group. Looking at the funnel plot of the same parameter, it was found that the study results without the Karstup et al, 2018 article (23) displayed an improved dispersion with no publication bias (Fig. 2C and D).

Figure 2 Atrial fibrillation ORs. (A) Forest plot of atrial fibrillation ORs demonstrated no statistically significant difference between the groups (total OR 1.27, CI 95% 0.85-1.91; P=0.24). (B) Forest plot of atrial fibrillation ORs obtained by excluding the ‘Kastrup et al, 2018’ (23) article demonstrated a statistically significant difference between the groups, (total OR 1.61, CI 95% 1.18-2.21; P<0.05). (C) Funnel plot of the atrial fibrillation between groups, with ‘Kastrup A et al 2018’ (23) article and with heterogeneity (P=0.07 and I2=48 %). (D) Funnel plot of the atrial fibrillation between groups, without ‘Kastrup A et al, 2018’ (23) article and without heterogeneity (P=0.52 and I2=0%). ET, endovascular thrombolysis; medical, medical management; OR, odds ratio.

Table III Meta-analysis of clinical parameters.

Table III

Meta-analysis of clinical parameters.

			Groups		Overall effect			Heterogeneity
Parameters	Excluded trials	Trials, n	Throm	Medic	Effect estimate	CI 95%	P-value	I2 (%)	P-value	(Refs.)
Atrial fibrillation	None	8	141	237	1.27	0.85-1.91	0.24	48	0.07	(18-25)
	Urra et al, 2014	7	132	223	1.36	0.87-2.12	0.18	52	0.06	(18)
	Haussen et al, 2016	7	138	229	1.31	0.85-2.03	0.22	55	0.05	(19)
	Haussen et al, 2017	7	131	221	1.18	0.76-1.82	0.46	49	0.08	(20)
	Nagel et al, 2018	7	106	175	1.13	0.73-1.73	0.58	36	0.17	(21)
	Sarraj et al, 2018	7	141	237	1.27	0.85-1.91	0.24	48	0.07	(22)
	Kastrup et al, 2018	7	84	163	1.61	1.18-2.21	<0.05	0	0.52	(23)
	Wang et al, 2020	7	139	235	1.28	0.83-1.98	0.26	56	0.04	(24)
	Xue et al, 2022	7	116	176	1.19	0.72-1.97	0.50	51	0.07	(25)
Prior to ischemic stroke	None	8	52	107	0.97	0.56-1.70	0.93	32	0.21	(18-25)
	Urra et al, 2014	7	47	105	0.90	0.60-1.36	0.63	14	0.32	(18)
	Haussen et al, 2016	7	50	100	1.03	0.55-1.96	0.92	45	0.14	(19)
	Haussen et al, 2017	7	52	107	0.97	0.56-1.70	0.93	32	0.21	(20)
	Nagel et al, 2018	7	52	107	0.97	0.56-1.70	0.93	32	0.21	(21)
	Sarraj et al, 2018	7	30	95	0.84	0.39-1.78	0.65	36	0.20	(22)
	Kastrup et al, 2018	7	52	107	0.97	0.56-1.70	0.93	32	0.21	(23)
	Wang et al, 2020	7	48	98	1.10	0.68-1.80	0.69	14	0.32	(24)
	Xue et al, 2022	7	33	30	1.01	0.41-2.49	0.99	46	0.13	(25)
Diabetes mellitus	None	8	123	202	1.07	0.58-1.40	0.61	0	0.56	(18-25)
Hypertension	None	8	405	781	0.87	0.68-1.10	0.23	0	0.83	(18-25)

[i] Throm, thrombectomy; medic, medical management; mRS, modified Rankin Scale.

Prior ischemic stroke

As regards the prior ischemic stroke, the total number of patients was 159 (9.5%) out of 1666 patients, of which 52 out of 569 (9.1%) were from the ET group and 107 out of 1097 (9.7%) were from the medical management group. The statistical analysis demonstrated no significant difference (OR 0.97, CI 95% 0.56-1.70; P=0.93) but indicated heterogeneity (P=0.21 and I2=32%) (Fig. S3A). By testing the sensitivity (Table III), low heterogeneity (P=0.32 or and I2=14%) was achieved only after removing the articles by ‘Urra et al, 2014’ (18) or ‘Wang et al, 2020’ (24), as the combination with the removed article by ‘Urra et al, 2014’ (18) was of higher quality because the total sample was 152 compared with 146 in the combination without the article by ‘Wang et al, 2020’ (24) (Table III). However, it was found again no statistically significant difference (OR 0.90, CI 95% 0.60-1.36; P=0.63) (Fig. S3B). Looking at the funnel plot of the same parameter, it was found that the study results without the ‘Urra et al, 2014’ (18) article displayed an improved dispersion with low publication bias (Fig. S3C and D).

MRS 0-2 after 3 months

As regards the mRS 0-2 after 3 months, the total number of patients was 1,107 (66.4%) out of 1,666 patients in the total group, of which 372 (65.3%) out of 569 were from the ET group and 735 (67.0%) out of 1097 patients were from the medical management group. The statistical analysis demonstrated no significant difference (OR 1.23, CI 95% 0.79-1.93; P=0.37) but indicated some heterogeneity (P=0.03 and I2=55%) (Fig. 3A). By testing the sensitivity (Table IV), low heterogeneity (P=0.16 and I2=35%) was achieved only after removing the article by ‘Nagel et al, 2018’ (21) but it was found again no statistically significant difference (OR 1.01, CI 95% 0.68-1.50; P=0.96) (Fig. 3B). Looking at the funnel plot of the same parameter, it was found that the study results without the ‘Nagel S et al, 2018’ (21) article displayed an improved dispersion with low publication bias (Fig. 3C and D).

Figure 3 MRS 0-2 ORs after 3 months. (A) Forest plot of mRS 0-2 ORs after 3 months demonstrated no statistically significant difference between the groups (total OR 1.23, CI 95% 0.79-1.93; P=0.37). (B) Forest plot of mRS 0-2 ORs after 3 months obtained by excluding the ‘Nagel et al, 2018’ (21) article demonstrated again no statistically significant difference between groups (total OR 1.01, CI 95% 0.68-1.50; P=0.96). (C) Funnel plot of the mRS 0-2 ORs after 3 months between the groups, including the ‘Nagel et al, 2018’ (21) article and with heterogeneity (P=0.03 and I2=55%). (D) Funnel plot of the mRS 0-2 OR after 3 months between groups, without ‘Nagel et al, 2018’ (21) article and with low heterogeneity (P=0.16 and I2=35%). MRS, modified Rankin scale; OR, odds ratio.

Table IV Meta-analysis of clinical outcomes.

Table IV

Meta-analysis of clinical outcomes.

			Groups		Overall effect			Heterogeneity
Outcomes	Excluded trials	Trials, n	Throm	Medic	Effect estimate	CI 95%	P-value	I2 (%)	P-value	(Refs.)
mRS 0-2	None	8	372	735	1.23	0.79-1.93	0.37	55	0.03	(18-25)
	Urra et al, 2014	7	371	734	1.24	0.77-1.99	0.37	61	0.02	(18)
	Haussen et al, 2016	7	362	718	1.18	0.76-1.85	0.46	58	0.03	(19)
	Haussen et al, 2017	7	343	671	1.10	0.75-1.61	0.63	43	0.10	(20)
	Nagel et al, 2018	7	304	581	1.01	0.68-1.50	0.96	35	0.16	(21)
	Sarraj et al, 2018	7	294	674	1.40	0.81-2.42	0.23	58	0.03	(22)
	Kastrup et al, 2018	7	286	646	1.32	0.72-2.42	0.36	61	0.02	(23)
	Wang et al, 2020	7	356	717	1.31	0.80-2.13	0.29	61	0.02	(24)
	Xue et al, 2022	7	288	404	1.43	0.84-2.43	0.19	50	0.06	(25)
Bleeding	None	7	39	21	3.13	1.65-5.95	<0.05	14	0.33
Mortality	None	6	30	51	1.37	0.55-3.41	0.50	45	0.10	(18,19,21-25)
	Urra et al, 2014	5	26	49	1.19	0.40-3.52	0.75	53	0.07	(18)
	Haussen et al, 2016	5	30	48	1.56	0.60-4.05	0.36	50	0.09	(19)
	Sarraj et al, 2018	5	19	50	1.07	0.46-2.48	0.87	33	0.20	(20)
	Nagel et al, 2018	5	27	31	1.94	1.02-3.69	0.04	2	0.40	(21)
	Wang et al, 2020	5	29	50	1.41	0.50-3.97	0.51	56	0.06	(22)
	Xue et al, 2022	5	19	27	1.27	0.34-4.75	0.73	53	0.08	(23)
Morbidity	None	6	89	110	1.02	0.64-1.63	0.94	39	0.14	(18,19,21-24)
	Urra et al, 2014	5	75	96	0.95	0.55-1.64	0.85	45	0.12	(18)
	Haussen et al, 2016	5	88	105	1.07	0.65-1.75	0.80	47	0.11	(19)
	Sarraj et al, 2018	5	70	88	1.07	0.56-2.01	0.84	51	0.08	(22)
	Kastrup et al, 2018	5	81	74	1.16	0.70-1.94	0.56	35	0.19	(23)
	Nagel et al, 2018	5	54	82	1.10	0.58-1.09	0.78	50	0.09	(21)
	Wang et al, 2020	5	77	105	0.87	0.61-1.24	0.45	0	0.57	(24)

Bleeding

Information regarding bleeding was available in seven out of eight articles (18-22,24,25) and demonstrated a statistical result (OR 3.13, CI 95% 1.65-5.95; P<0.05) with no heterogeneity (P=0.33 and I2=14%) (Fig. 4). Bleeding was found in 60 out of 1666 (3.6%) patients in the total group. In the ET group, bleeding was diagnosed in 39 out of 129 (30.2%) patients, compared with 21 out of 1097 (1.9%) in the medical management group. Looking at the funnel plot of the same parameter, it was found a very low publication bias.

Figure 4 Bleeding ORs. (A) Forest plot of the bleeding ORs demonstrated a statistically significant difference between groups (total OR 3.13, CI 95% 1.65-5.95; P<0.05). (B) Funnel plot of the bleeding ORs demonstrated very low heterogeneity (P=0.33, and I2=14%). Medical, medical management; OR, odds ratio.

Morbidity

Morbidity data were available in six of the eight articles (18,20-22,24,25): A total of 199 (20.3%) from 976 included patients; and poor KPS in 89 (21.3%) of 416 patients from the ET group. The statistical analysis demonstrated no significant difference (OR 1.02, CI 95% 0.64-1.63; P=0.94) and indicated heterogeneity (P=0.14 and I2=39%) (Fig. 5A). By testing the sensitivity (Table IV), no heterogeneity (P=0.57 and I2=0%) was achieved only after removing the article by ‘Wang et al, 2020’ (24) but again without a statistically significant difference (OR 0.87, CI 95% 0.61-1.24; P=0.45) (Fig. 5B). Looking at the funnel plot of the same parameter, it was found that the study results without the ‘Wang et al, 2020’ (24) article displayed an improved dispersion with no publication bias (Fig. 5C and D).

Figure 5 Morbidity ORs. (A) Forest plot of morbidity ORs demonstrated no statistically significant difference between groups (total OR 1.02, CI 95% 0.64-1.63; P=0.94). (B) Forest plot of morbidity ORs obtained by excluding the ‘Wang et al, 2020’ (24) article demonstrated again no statistically significant difference between the groups (total OR 0.87, CI 95% 0.61-1.24; P=0.45). (C) Funnel plot of the morbidity ORs between groups, with the ‘Wang et al 2020’ (24) article and heterogeneity (P=0.14 and I2=39%). (D) Funnel plot of the morbidity ORs between groups, without the ‘Wang et al 2020’ (24) article and heterogeneity (P=0.57 and I2=0%). OR, odds ratio.

Mortality

Mortality data were available in six of the eight articles (18,19,21,22,24,25): A total of 81/1,243 (6.5%) patients, including 30 out of 394 (7.6%) from the ET group and 51 (6.0%) out of 849 from the medical management group. The statistical analysis demonstrated no statistically significant difference (OR 1.37, CI 95 % 0,55-3.41; P=0.50) but indicated heterogeneity (P=0.10 and I2=45 %) (Fig. 6A). By testing the sensitivity (Table IV), after removing the article by ‘Nagel S et al, 2018’ (21), it was found a statistically significant difference (OR 1.94, CI 95% 1.02-3.69; P<0.05) with no heterogeneity (P=0.40 and I2=2%) (Fig. 6B). Looking at the funnel plot of the same parameter, it was found that the study results without the ‘Nagel S et al, 2018’ (21) article displayed an improved dispersion with very little publication bias (Fig. 6C and D).

Figure 6 Mortality ORs. (A) Forest plot of mortality ORs demonstrated no statistically significant difference between groups (total OR 1.37, CI 95 % 0.55-3.41; P=0.50). (B) Forest plot of mortality ORs without the ‘Nagel et al, 2018’ (21) article demonstrated a statistically significant difference between the groups (total OR 1.94, CI 95% 1.02-3.69; P<0.05). (C) Funnel plot of the mortality ORs between the groups, with the ‘Nagel et al, 2018’ (21) article and with heterogeneity (P=0.10 and I2=45%). (D) Funnel plot of the mortality ORs between groups without the ‘Nagel et al, 2018’ (21) article and without heterogeneity (P=0.40 and I2=2%). ET, endovascular thrombolysis; medical, medical management; OR, odds ratio.

Discussion

The present study suggested that medical management for LVO strokes with minimal symptoms was associated with fewer complications compared with ET intervention. More precisely, atrial fibrillation was the only statistically significant parameter related to bleeding and mortality in patients with LVO strokes with minimal symptoms. Interestingly, the rates of the mRS score of 0-2 (indicating a good outcome) at discharge and the morbidity (KPS <80) showed no association with medical management or ET therapy in those patients.

The findings of the present meta-analysis pointed out that ET therapy may be associated with a high risk of bleeding and mortality in patients with LVO presenting with mild symptoms (NIHSS score ≤8).

The absence of benefits from ET therapy is consistent with the findings of several previous studies (26-29). However, a multicenter observational study reported that ET therapy was effective in recanalizing the occluded vessel but increased the risk of serious bleeding significantly without improving the functional outcome, suggesting that it was not justified routinely in minor strokes (18). The current study showed that ET therapy is related to a high risk of bleeding and mortality in patients with LVO presenting mild symptoms (NIHSS score ≤8). This may be attributed to several reasons, such as the acceptable collateral condition that developed in patients with LVO and mild symptoms, which predisposes to bleeding and reduces the advantages of ET.

In addition, the recent meta-analysis showed that ET therapy, compared with medical management alone, is associated with higher rates of bleeding and mortality outcomes in patients with atrial fibrillation. A possible explanation is the higher proportion of atrial fibrillation in the ET group, an extensively reported risk factor for intracerebral haemorrhage (30,31).

In agreement with previous cohorts of patients presenting with mild symptoms, the treatment modality had a high percentage of good early clinical outcomes and an excellent early clinical outcome (23); however, even after excluding patients with contraindications for an ET, the clinical outcome was not significantly different between ET and medical management in patients with an admission NIHSS ≤8. The aforementioned results are similar to the ones shown in the present study, with mRS 0-2 and morbidity having no association with medical management or ET therapy in those patients.

Prior studies reported on outcomes achieved using ET therapy in milder stroke patients but were limited because of a single-center study design (20), a small number of patients (18-20) or the lack of a similar medical management group (32). These uncertain results highlight the need for higher-quality data for this population.

There are several limitations in the present study. First, most of the eligible reports that were included were retrospective. These retrospective studies are affected by low accuracy and data loss. Additionally, the study design differed significantly among the included studies, e.g. the duration of follow-up (e.g., 30-90 days).

In conclusion, the routine use of ET was not significantly favorable for the early clinical or radiological outcome in patients with mild strokes and LVO compared with medical management alone. However, the present results suggested that ET could be beneficial only in patients with NIHSS ≥8. It is also undisputable that a randomized trial is needed to determine the difference in outcome between these two treatment modalities in these patients.

Supplementary Material

Diabetes mellitus ORs. (A) Forest plot of diabetes mellitus ORs demonstrated no statistically significant difference between the groups (total OR 1.07, CI 95% 0.58-1.40; p 0.61). (B) Funnel plot of the diabetes mellitus ORs demonstrated no heterogeneity in the groups (P=0.56, and I2=0%). ET, endovascular thrombolysis; medical, medical management; OR, odds ratio.

Hypertension ORs. (A) Forest plot hypertension ORs demonstrated no statistically significant results (total OR 1.87, CI 95% 0.68-1.10; P=0.23). (B) Funnel plot of the hypertension ORs demonstrated no heterogeneity in the groups (P=0.83 and I2=0%). ET, endovascular thrombolysis; medical, medical management; OR, odds ratio.

Prior ischemic stroke ORs. (A) Forest plot of prior ischemic stroke ORs demonstrated no statistically significant difference groups (total OR 0.97, CI 95% 0.56-1.70; P=0.93). (B) Forest plot of prior ischemic stroke ORs without ‘Urra et al 2014’ (18) article demonstrated no statistically significant difference between the groups (total OR 0.90, CI 95% 0.60-1.36; P=0.63). (C) Funnel plot of the prior ischemic stroke ORs between groups, with ‘Urra et al, 2014’(18) article and with heterogeneity (P=0.21 and I2=32%). (D) Funnel plot of the prior ischemic stroke ORs between groups, without ‘Urra et al, 2014’(18) article, and very low heterogeneity (P=0.32 and I2=14%). Medical, medical management; OR, odds ratio.

Acknowledgements

Not applicable.

Funding

Funding: No funding was received.

Availability of data and materials

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Authors' contributions

GF and VEG conceptualized the study. VEG, IT, GF, PP, ISA, DAS and NT analyzed the data and wrote and prepared the draft of the manuscript. VEG and GF provided critical revisions. All authors contributed to the manuscript revision. All authors read and approved the final manuscript. GF and VEG confirm the authenticity of all the raw data.

Ethics approval and consent to participate

Not applicable.

Patient consent for publication

Not applicable.

Competing interests

DAS is the Editor-in-Chief for the journal, but had no personal involvement in the reviewing process, or any influence in terms of adjudicating on the final decision, for this article.

References