Effect of astragalus membranaceus on neurological function in acute aneurysmal subarachnoid hemorrhage patients with high inflammation: A preliminary randomized, double-blind, placebo-controlled clinical trial

Abstract

Background and aim

Astragalus membranaceus (AM) is a traditional Chinese herb. Our previous study revealed that AM can enhance neurological function in patients with acute intracerebral hemorrhage. The aim of this study was to investigated the effects of AM on patients with acute aneurysmal subarachnoid hemorrhage (aSAH).

Experimental procedure

Eighty-eight patients experiencing acute aSAH were randomly allocated to either the treatment group (TG) comprising 44 patients, who received 3 g of AM orally thrice daily for 14 days, or the control group (CG) with 44 patients, who received 3 g of a placebo.

Results

Eighty-three patients (41 in CG and 42 in TG) completed the trial. Stratified analyses revealed serum interleukin-6 (IL-6) median ≥7.28 pg/mL at baseline. The percentage of good GOS scores (GOS 4 or 5) at two weeks (W2) and four weeks (W4) was significantly higher in TG than in CG (W2: 35.3 % vs. 7.7 %, p = 0.042; W4: 62.5 % vs. 30.8 %, p = 0.044). Moreover, a higher percentage of Barthel index scores (>60) was observed in TG than in CG at W2 (35.3 % vs. 7.7 %, p = 0.042) after AM or placebo administration.

Conclusion

Administering AM for 14 days has shown potential in enhancing neurological function four weeks post-aSAH onset, especially in patients with a serum IL-6 level median ≥7.28 pg/mL. Therefore, further research is warranted to explore the anti-inflammatory role of AM. However, this study's limitations include a small sample size and the single-center design, signifying its status as a preliminary investigation.

Graphical abstract

Highlights

• •Astragalus membranaceus (AM) administration in acute aneurysmal subarachnoid hemorrhage (aSAH).
• •Patients with serum IL-6 levels median ≥ 7.28 pg/mL, AM can increase the percentage of good GOS scores at W2 and W4, and BI scores > 60 at W2.
• •This preliminary study underscores the need for larger, multicenter investigations.
• •AM's anti-inflammatory role in acute aSAH warrants further exploration.

Abbreviations

aSAH

aneurysmal subarachnoid hemorrhage

AM

Astragalus membranaceus

B

baseline

BI

Barthel index

CG

control group

CMUH

China Medical University Hospital

CONSORT

Consolidated Standards of Reporting Trials

CSF

cerebrospinal fluid

D7

seven days

D14

fourteen days

DCI

delayed cerebral ischemia

DIND

delayed ischemic neurological deficit

ELISA

enzyme-linked immunosorbent assay

ELSD

evaporative light scattering detector

FIM

functional independence measure

GMP

good manufacturing practice

GOS

Glasgow outcome scale scores

HPLC

high-performance liquid chromatography

ICH

intracerebral hemorrhage

IL-6

interleukin-6

IL-1β

interleukin-1β

mRS

modified Rankin scale

PAD

photodiode array detector

rpm

revolutions per minute

S100B

S100B calcium-binding protein B

SAH

spontaneous subarachnoid hemorrhage

TG

treatment group

TNF-α

tumor necrosis factor-α

W1

one week

W2

two weeks

W4

four weeks

W12

twelve weeks

1. Introduction

Acute aneurysmal subarachnoid hemorrhage (aSAH) is a significant contributor to cerebrovascular morbidity and mortality, particularly among younger patients. aSAH is a significant global public health threat. The overall worldwide incidence of aSAH is ≈ 6.1 per 100 000 person-years,¹ with a global prevalence of 8.09 million (95 % uncertainty interval, 7.02–9.72 million) cases^.2 However, the incidence of aSAH is highly variable by region, with the highest incidence in Japan and Finland at 28 and 16.6 per 100 000 person-years, respectively,¹ and the highest age-standardized prevalence in Japan and Andean Latin America.² aSAH is a devastating condition with severely morbid and often deadly condition. Prehospital mortality rates from aSAH have been reported to be 22 %–26 %,³ hospital inpatient mortality rates from aSAH have report from 13 % to 20 % (13.7 % in 2006 to 13.1 % in 2018 [United States]⁴ and 19%–20 % in 2021 [global]⁵).

Among the primary complications associated with SAH is a distinctive clinical syndrome marked by secondary neurological impairment. This syndrome, traditionally termed “vasospasm,” is observed angiographically and is clinically referred to as delayed cerebral ischemia (DCI) or delayed ischemic neurological deficit (DIND)^.6, 7, 8 Pathophysiological mechanisms in the first 72 h after aSAH that drive early brain injury also influence secondary complications and overall outcomes.⁹ DCI is now hypothesized to be caused by the combined effects of large-vessel cerebral vasospasm and multiple brain injury processes triggered by aneurysm rupture and early brain injury. Mechanisms involving arteriolar constriction and cerebral microthrombosis, cortical spreading depolarization/ischemia, blood-brain barrier breakdown, cerebral autoregulation impairment, and capillary transit time heterogeneity are hypothesized to play a role in the pathophysiology of DCI and related cerebral infarction.,^10,11 Neuroinflammation, either independently or as a consequence of early brain injury, has been identified as a potential target for intervention.^12,13

Astragalus membranaceus (AM) is a traditional Chinese herb which can tonify Qi, and it is common used in stroke treatment in China and Taiwan. AM is believed to offer a range of benefits including immune enhancement, fatigue reduction, anti-inflammatory and antioxidative properties, as well as potential anticancer effects. Several in vitro and animal model studies suggest that specific components in AM, such as flavonoids and saponins, may have neuroprotective qualities beneficial against stroke. Such as the saponin astragaloside IV (AS-IV), a 3-O-beta-d- xylopyranosyl-6-O-beta-D-glucopyranosylcycloastragenol, is a main component. of AM can inhibit the nuclear factor-κB pathway generates an anti-inflammatory effect in vivo.¹⁴ And also has an antioxidant effect to prevent glucose-Induced podocyte apoptosis.¹⁵ Furthermore, AM is posited to contribute to cardiovascular health, a factor that could facilitate recovery following cerebrovascular events such as strokes.16, 17, 18, 19

Our prior research established that AM can improve neurological function in patients with acute intracerebral hemorrhage (ICH).²⁰ Consequently, the aim of the current study was to further evaluate the therapeutic impact of AM on acute spontaneous aSAH. We hypothesized that AM's anti-inflammatory properties could mitigate the inflammatory responses triggered by acute spontaneous aSAH, thereby aiding neurological recovery. Within the context of a preliminary double-blind, randomized, placebo-controlled study, our objective was to assess the effects of AM on neurological function in patients diagnosed with acute spontaneous aSAH.

2. Materials and methods

2.1. Subjects

Patients were enrolled from the neurosurgery and emergency departments of China Medical University Hospital (CMUH), Taichung City, Taiwan, between September 14, 2017, and December 24, 2021. All patients were included within 24 h of SAH onset. The study adhered to the ethical principles of clinical trials as outlined in the Declaration of Helsinki. The Research Ethics Committee of CMUH approved the study protocol (CMUH106-REC1-074; NIH registration NCT 03271697, 2017). Patients were provided with a detailed explanation of the study procedure and purpose, and they provided signed consent before participating. The study also adhered to the International Conference on Harmonization and Good Clinical Practice guidelines. Inclusion criteria were as follows: (1) both males and females, (2) aged between 20 and 80 years, (3) admission within 24 h of SAH onset, (4) imaging evidence (head computer tomography scan or cerebral angiography) confirming intracranial aneurysm rupture, (5) Hunt & Hess grade 1–4, and (6) signed informed consent from patients or legal representatives. Exclusion criteria were: (1) traumatic or mycotic aneurysm rupture; (2) severe complications from heart, liver disease, infection, renal failure, or malignant tumors and more; (3) patients considered inappropriate by the principal investigator, such as history of drug abuse or psychiatric disorders; (4) pregnant or breastfeeding women; (5) patients who had enrolled in or not completed other investigational drug trials within the month prior to screening; and (6) patients with a previous stroke and a modified Rankin scale (mRS) score of 3 or more.

3. Preparation of AM

Both the AM extract and the placebo (starch) were prepared by Sun Ten Pharmaceutical Co. Ltd. (Taiwan), a company certified in Good Manufacturing Practice (GMP). Crude AM [Astragali Radix as the radix of Astragalus membranaceus (Fisch.) Bunge. Var. mongholicus (Bunge.) P.K. Hsiao; Huang Qi] was sourced from Northwest China. The authenticity of AM was verified by the Non-Profit Organization Brion Research Institute of Taiwan, in accordance with the Chinese Herbal Pharmacopeia²¹ and the Pharmacopeia of the People's Republic of China.²² The AM extract underwent rigorous testing for microorganisms, heavy metals, and pesticides to meet the Taiwan Government's accepted GMP standards.

Three grams of dry Astragali Radix yielded 0.66 g of dry extract (3.0:0.66 = 4.5:1). The freeze-dried AM extracts were authenticated using high-performance liquid chromatography (HPLC; Pump: Waters 600; Autosampler: Waters 717 plus, Milford, Massachusetts, USA)/Evaporative light scattering detector (ELSD, Sedex75, Olivet, France) with astragaloside IV (Biotic Chemical Co. Ltd., Shanghai, China) and HPLC/Photodiode array detector (PAD) with calycosin 7-O-glucoside as the standard markers (Fig. 1). Each 3 g of AM extract and placebo were sealed in indistinguishable aluminum foil sachets and stored in the Clinical Trials Pharmacy of CMUH to be administered by a dedicated pharmacist.

Fig. 1.

Chromatographic fingerprint of Astragalus membranaceus This was obtained using high-Performance liquid chromatography (HPLC) coupled with evaporative light scattering detector (ELSD) for astragaloside IV and HPLC coupled with photodiode array detector (PAD) for calycosin 7-O-glucoside.

3.1. Study design and sample size evaluation

The present study was a preliminary single-center, double-blind, placebo-controlled, randomized phase II/III study. The study aimed to assess whether AM increased the percentage of good outcome scores (GOS 4 or 5) in ten dimensions. Sample size was calculated according to the hypothesis that AM can increase a patient's good GOS scores percentage. According to our previous results,²³ if the difference of the percentage of good GOS scores changed between two groups were 34 %–37 %, a sample size of 108 patients, with 54 in each group, was needed to achieve a statistical power of 80 % with a significant level at 0.05. Considering a follow-up rate of 0.9, 120 patients were required for this study.

3.2. Randomization and grouping

Computer-generated random numbers using block randomization (block size 2 or 4) determined subject allocation. Sealed envelopes with pre-generated random numbers were assigned serial numbers based on the allocation sequence. Envelopes were opened sequentially per subject admission, with the envelope's number determining the subject's group allocation. The study was divided into two groups: the treatment group (TG) and the control group (CG). In the TG, patients received oral or nasopharyngeal administration of AM (3 g thrice daily for 14 days) within 24 h of SAH onset alongside standard treatment as per Guidelines for the Management of Spontaneous SAH in Adults.¹³ In the CG, patients followed the same schedule as in the TG but received oral or nasopharyngeal administration of placebos.

3.3. Assessment

Subjects, investigators, and clinical assessor were blinded to group allocation. The password for the randomization envelope was known solely to a designated researcher. The assessment of GOS and mRS scores, and Functional Independence Measure (FIM) and Barthel Index (BI) scores was performed by a fully trained clinical assessor.

Primary outcome measures were the percentage change in good GOS (GOS scores were 4 or 5) and good mRS (mRS scores were 0, 1 or 2), while GOS scores 1, 2, or 3 represented bad GOS, and mRS score 3, 4, or 5 indicated bad mRS. Both GOS and mRS are a 6-point disability scale with possible scores ranging from 0 to 5, is a single item, global outcomes rating scale for post-stroke patients. It is used to categorize level of functional independence with reference to pre-stroke activities rather than on observed performance of a specific task. And the mRS is the most widely used outcome measure in stroke clinical trials. The assessment was conducted before AM or placebo administration (baseline, B), and one week (W1), two weeks (W2; 14 ± 2 days), four weeks (W4; 28 ± 2 days), and twelve weeks (W12, 84 ± 7 days) after AM or placebo administration.

Secondary outcome measures consisted of percentage changes in FIM and BI scores, assessed at the same time points. Both FIM and BI are an instrument that was developed as a measure of disability for a variety of populations and is not specific to any diagnosis. Changes in interleukin-6 (IL-6), interleukin-1β (IL-1β), tumor necrosis factor-α (TNF-α), and S100B calcium-binding protein B (S100B) levels in serum and cerebrospinal fluid (CSF) were also measured at baseline, Day 7 (D7), and Day 14 (D14) following AM or placebo administration. The level pf IL-6), IL-1β, TNF-α), and S100B was associated with acute inflammation which was used to measure post-aSAH acute neuroinflammation.

3.4. Measurement of IL-6, IL-1β, TNF-α, and S100B

Peripheral blood (4 mL) was drawn from the cubital vein, and CSF (4 mL) was collected from the ventriculoperitoneal shunt at baseline, and on D7 and D14 after AM or placebo administration. Both blood and CSF were sent to the neuroscience laboratory room of China Medical University for centrifugation for 10 min at 3600 revolutions per minute (rpm), and the supernatant was collected and stored at −20 °C until analysis.

Commercial enzyme-linked immunosorbent assay (ELISA) kits IL-1β (ThermoFisher, Catalog Number 88–7261; sensitivity: 2 pg/mL), IL-6 (ThermoFisher, Catalog Number 88–7066; sensitivity: 2 pg/mL), TNF-α (ThermoFisher, Catalog Number 88–7346; sensitivity: 4 pg/mL) and S100B (Elabscience, Catalog Number: E-EL-H1297; sensitivity: 18.75 pg/mL) and an ELISA reader (BioTek Instruments, USA) were used for measurements. Assays were duplicated, with no cross-reactivity or interference observed.

The checklist of Consolidated Standards of Reporting Trials (CONSORT) was completed (Committee, 2015)²⁴

3.5. Statistical analysis

Baseline variables were compared using t-tests for continuous variables (for example, age) and chi-squared tests for categorical variables (for example, sex). An intention-to-treat analysis was employed. Efficacy outcomes, including mRS, GOS, FIM, and BI scores between the two groups at various time points were compared using two-sample t-tests and non-parametric Mann-Whitney tests for non-normally distributed data when data presented as continuous variables. And the efficacy outcomes between the two groups at various time points were compared using chi-squared tests or fisher exact test when data presented as categorical variables. The Bonferroni correction was used when consider multiple testing.

A stratified analysis was conducted based on serum IL-6 median levels at baseline. All analyses were performed using SAS version 9.2 (SAS Institute Inc., Cary, NC) and SPSS (v21.0; IBM, Armonk, NY, USA). All tests were two-sided, and a p-value <0.05 indicated statistical significance.

4. Results

4.1. Baseline demographic characteristics

A total of eighty-eight patients with acute spontaneous aSAH were enrolled in this study and randomized into either the TG, consisting of 44 patients, or the CG with 44 patients. Five patients withdrew during the trial procedure: one patient died, and one patient had a history of cancer (Because the patient had received cancer therapy for 10 years, he had no obvious clinical symptoms, therefore, who was included the present study) in the TG; and one patient died, one patient withdrew, and one patient's family refused treatment due to a Do-Not-Resuscitate (DNR) order in the CG. Therefore, a total of 83 patients (42 in the TG and 41 in the CG) completed the trial (Fig. 2).

Fig. 2.

Flow chart of study enrollment AM, Astragalus Membranaceus; GOS, Glasgow outcome scale; mRS, modified Rankin scale; FIM, functional independence measures; BI, Barthel index; CSF, cerebrospinal fluid; IL-6, interleukin-6; IL-1β, interleukin-1β; TNF-α, tumor necrosis factor-α (TNF-α); S100B, S100B calcium-binding protein B (S100B); W1, one week after AM or placebo administration; W2, two weeks after AM or placebo administration; W4, four weeks after AM or placebo administration; W12, twelve weeks after AM or placebo administration; t. i.d., three times per day.

Baseline characteristics, including sex, age, mRS, GOS, FIM, and BI scores did not significantly differ between the TG and CG (all p > 0.05; Table 1).

Table 1.

Baseline demographic characteristics.

	CG (n = 44)	TG (n = 44)	p-value
Sex
Female	27 (61.4 %)	31 (70.5 %)	0.368
Male	17 (38.6 %)	13 (29.5 %)
Age (years)	53.98 (9.04)	53.14 (11.85)	0.709
mRS
Good (score 0–2)	0 (0 %)	1 (2.3 %)	1.000
Bad (score 3–5)	44 (100 %)	43 (97.7 %)
GOS
Good (score 4–5)	0 (0 %)	1 (2.3 %)	1.000
Bad (score 1–3)	44 (100 %)	43 (97.7 %)
FIM (Score 18–71)
Median (Q1, Q3)	18 (18, 18)	18 (18, 18)	0.941a
Score ≥55	3 (6.8 %)	0 (0.0 %)	0.241b
Score <55	41 (93.2 %)	44 (100.0 %)
BI (Score 0–55)
Median (Q1, Q3)	0 (0, 0)	0 (0, 0)	0.230
Score >60	0 (0.0 %)	0 (0.0 %)	–
Score ≤60	44 (100.0 %)	44 (100.0 %)
GCS (SD)	11.25 (4.93)	13.21 (3.04)	0.028
Intervention			1.000b
TAE	2 (4.5 %)	2 (4.5 %)
clipping	42 (95.5 %)	42 (95.5 %)
Location			–
A-Com	20 (36.4 %)	13 (27.1 %)
P-com	10 (18.2 %)	13 (27.1 %)
VA	6 (10.9 %)	4 (8.3 %)
MCA	4 (7.3 %)	6 (12.5 %)
ICA	13 (23.6 %)	11 (22.9 %)
BA	2 (3.6 %)	1 (2.1 %)
Comorbidity			–
CHF	2 (22.2 %)	1 (11.1 %)
DM	4 (44.4 %)	5 (55.6 %)
Liver disease	1 (11.1 %)	1 (11.1 %)
uremia	2 (22.2 %)	2 (22.2 %)

Data are presented as n (%) or mean (standard deviation, SD).

Abbreviation: CG, control group; TG, treatment group; mRS, modified Rankin scale; GOS, Glasgow outcome scale; GCS, Glasgow coma scale; FIM, functional independence measure; BI, Barthel index; TAE, transcatheter arterial embolization; A-Com, anterior communicating artery; P-Com, posterior communicating artery; VA, vertebrate artery; MCA, middle cerebral artery; IVA, internal carotid artery; BA, basilar artery; CHF, congestive heart failure; DM, diabetic mellitus.

p-values are from independent sample t-test or Chi-squared test for continuous or categorical variables.

^aMann-Whitney Test.

^bFisher exact test.

4.2. Effect of AM on good and bad GOS and mRS scores in patients with acute aSAH

The percentages of good and bad GOS scores at Baseline (B), W1, W2, W4, and W12 did not demonstrate significant differences between the TG and CG (all p > 0.05; Table 2). Similarly, the percentages of good and bad mRS scores at B, W1, W4, and W12 did not differ significantly between the groups (all p > 0.05; Table 3).

Table 2.

Effect of Astragalus Membranaceus (AM) on Glasgow outcome scale (GOS) in patients with acute subarachnoid hemorrhage.

	n	CG	n	TG	p-value
Baseline	44		44		1.000a
Good		0 (0 %)		1 (2.3 %)
Bad		44 (100 %)		43 (97.7 %)
W1	43		43		1.000a
Good		4 (9.3 %)		3 (7 %)
Poor		39 (90.7 %)		40 (93.0 %)
W2	42		42		0.314
Good		8 (19.0 %)		13 (31.0 %)
Poor		34 (81.0 %)		29 (69.0 %)
W4	42		40		0.261
Good		20 (47.6 %)		24 (60.0 %)
Poor		22 (52.4 %)		16 (40.0 %)
W12	42		39		0.838
Good		26 (61.9 %)		25 (64.1 %)
Poor		16 (38.1 %)		14 (35.9 %)

Abbreviation: CG, control group; TG, treatment group; Good, GOS score 4 or 5; Bad, GOS score 1, 2, or 3; W1, one week after AM or placebo administration; W2, two weeks after AM or placebo administration; W4, four weeks after AM or placebo administration; W12, twelve weeks after AM or placebo administration.

Differences in categorical variables were tested using the Chi-squared test.

^aFisher's exact test.

Table 3.

Effect of Astragalus Membranaceus (AM) on modified Rankin scale (mRS) in patients with acute subarachnoid hemorrhage.

	n	CG	n	TG	p-value
Baseline	44		44		1.000a
Good		0 (0 %)		1 (2.3 %)
Bad		44 (100 %)		43 (97.7 %)
W1	43		43		1.000a
Good		1 (2.3 %)		1 (2.3 %)
Bad		42 (97.7 %)		42 (97.7 %)
W2	42		42		1.000
Good		6 (14.3 %)		6 (14.3 %)
Bad		36 (85.7 %)		36 (85.7 %)
W4	42		40		0.382
Good		18 (42.9 %)		21 (52.5 %)
Bad		24 (57.1 %)		19 (47.5 %)
W12	42		39		0.947
Good		24 (57.1 %)		22 (56.4 %)
Bad		18 (42.9 %)		17 (43.6 %)

Abbreviation: CG, control group; TG, treatment group; Good, mRS scale score 0, 1, or 2; Bad: mRS scale score 3, 4, or 5; W1, one week after AM or placebo administration; W2, two weeks after AM or placebo administration; W4, four weeks after AM or placebo administration; W12, twelve weeks after AM or placebo administration.

Differences in categorical variables were tested using the Chi-squared test.

^aFisher's exact test.

4.3. Effect of AM on FIM and BI scores in patients with acute aSAH

The percentages of FIM scores at B, W1, W2, W4, and W12 did not significantly differ between the CG and TG (all p > 0.05; Table 4). The percentages of BI scores at B, W1, W2, W4, and W12 also showed no significant differences between the groups (all p > 0.05; Table 5).

Table 4.

Effect of Astragalus Membranaceus (AM) on functional independence measure (FIM) in patients with acute subarachnoid hemorrhage.

	n	CG	n	TG	p-valuea	p-valueb
Baseline	44		44
mean (SD)		22.64 (12.55)		20.23 (6.27)	0.259	0.241c
Score ≥55		3 (6.8 %)		0 (0.0 %)
Score <55		41 (93.2 %)		44 (100.0 %)
W1	43		43
mean (SD)		37.21 (27.80)		37.60 (24.51)	0.944	0.451
Score ≥55		9 (20.9 %)		12 (27.9 %)
Score <55		34 (79.1 %)		31 (72.1 %)
W2	42		42
mean (SD)		47.31 (34.37)		52.52 (39.96)	0.523	0.498
Score ≥55		14 (33.3 %)		17 (40.5 %)
Score <55		28 (66.7 %)		25 (59.5 %)
W4	42		40
mean (SD)		70.26 (46.66)		80.25 (49.60)	0.350	0.632
Score ≥55		23 (54.8 %)		24 (60.0 %)
Score <55		19 (45.2 %)		16 (40.0 %)
W12	42		39
mean (SD)		85.24 (44.79)		93.64 (40.98)	0.382	0.596
Score ≥55		29 (69.0 %)		29 (74.4 %)
Score <55		13 (31.0 %)		10 (25.6 %)

Data are presented as n (%) or mean (standard deviation, SD).

CG, control group; TG, treatment group; W1, one week after AM or placebo administration; W2, two weeks after AM or placebo administration; W4, four weeks after AM or placebo administration; W12, twelve weeks after AM or placebo administration.

^aIndependent sample t-test.

^bChi-squared test.

^cFisher's exact test.

Table 5.

Effect of Astragalus Membranaceus (AM) on Barthel index (BI) in patients with acute subarachnoid hemorrhage.

	n	Placebo	n	Treatment	p-valuea	p-valueb
Baseline	44		44
mean (SD)		2.73 (9.67)		0.57 (2.69)	0.160	–
Score >60		0 (0.0 %)		0 (0.0 %)
Score ≤60		44 (100.0 %)		44 (100.0 %)
W1	43		43
mean (SD)		14.65 (24.84)		12.91 (19.77)	0.720	1.000c
Score >60		3 (7.0 %)		2 (4.7 %)
Score ≤60		40 (93.0 %)		41 (95.3 %)
W2	42		42
mean (SD)		25.83 (34.52)		32.98 (38.83)	0.375	0.124
Score >60		7 (16.7 %)		13 (31.0 %)
Score ≤60		35 (83.3 %)		29 (69.0 %)
W4	42		40
mean (SD)		49.52 (45.78)		58.50 (46.80)	0.383	0.487
Score >60		22 (52.4 %)		24 (60.0 %)
Score ≤60		20 (47.6 %)		16 (40.0 %)
W12	42		39
mean (SD)		62.62 (43.82)		70.00 (39.02)	0.427	0.506
Score >60		25 (59.5 %)		26 (66.7 %)
Score ≤60		17 (40.5 %)		13 (33.3 %)

Data are presented as N (%) or mean (standard deviation, SD).

CG, control group; TG, treatment group; W1, one week after AM or placebo administration; W2, two weeks after AM or placebo administration; W4, four weeks after AM or placebo administration; W12, twelve weeks after AM or placebo administration.

^aIndependent sample t-test.

^bChi-squared test.

^cFisher's exact test.

Effect of AM on IL-6, IL-1β, TNF-α, and S100B serum and CSF in patients with acute aSAH.

The serum levels of IL-6, IL-1β, TNF-α, and S100B at baseline, D7, and D14 did not significantly differ between the CG and TG (all p > 0.05; Table 6). Similarly, the levels of IL-6, IL-1β, TNF-α, and S100B in CSF at baseline, D7, and D14 showed no significant differences between the groups (all p > 0.05; Table 7).

Table 6.

Effect of Astragalus Membranaceus (AM) on serum IL-6, IL-1β, TNF-α, and S100B levels in patients with acute subarachnoid hemorrhage (pg/mL).

	n	CG		n	TG		p-valuea	p-valueb
		Mean (SD)	Median (Q1, Q3)		Mean (SD)	Median (Q1, Q3)
IL-6
Baseline	43	11.65 (14.06)	8.85 (3.17, 14.11)	43	10.04 (11.11)	5.86 (3.22, 10.87)	0.56	0.442
D7	41	1.34 (2.12)	0 (0, 2.34)	42	1.01 (2.30)	(0, 1.52)	0.50	0.335
D14	41	1.04 (3.06)	0 (0, 0.62)	41	2.59 (7.12)	0 (0, 0.93)	0.21	0.682
IL-1β
Baseline	43	0.2298 (0.6689)	0 (0, 0)	43	0.1544 (0.6141)	0 (0, 0)	0.59	0.904
D7	42	0.1273 (0.8169)	0 (0, 0)	42	0.0036 (0.0162)	0 (0, 0)	0.33	0.679
D14	41	0.0095 (0.0350)	0 (0, 0)	40	0.1605 (0.7610)	0 (0, 0)	0.22	0.920
TNF-α
Baseline	43	0.3479 (1.1062)	0 (0, 0)	43	0.3564 (1.2918)	0 (0, 0)	0.97	0.774
D7	42	2.0689 (11.7679)	0 (0, 0)	42	0.1826 (0.6294)	0 (0, 0)	0.30	0.855
D14	41	2.8259 (10.9441)	0 (0, 0)	41	0.6739 (2.4186)	0 (0, 0)	0.23	0.189
S100
Baseline	43	406.16 (364.22)	313.31 (131.37, 524.07)	43	650.10 (1787.76)	280.14 (80.56, 537.20)	0.38	0.531
D7	42	246.91 (219.70)	161.68 (70.06, 364.43)	42	331.58 (558.65)	162.87 (51.75, 386.14)	0.36	0.720
D14	41	294.17 (379.46)	206.96 (81.57, 295.54)	41	257.91 (303.74)	201.10 (48.97, 333.58)	0.63	0.572

CG, control group; SD, standard deviation; TG, treatment group; IL-6, interleukin-6; IL-1β, interleukin-1β; TNF-α, tumor necrosis factor-α; S100B, S100 calcium binding protein B; D7, 7th after AM or placebo administration; D14, 14th after AM or placebo administration.

^aIndependent t-test.

^bMann-Whitney Test.

Table 7.

Effect of Astragalus Membranaceus (AM) on IL-6, IL-1β, TNF-α, and S100B levels in cerebrospinal fluid of patients with acute subarachnoid hemorrhage (pg/mL).

	n	CG		n	TG		p-valuea	p-valueb
		Mean (SD)	Median (Q1, Q3)		Mean (SD)	Median (Q1, Q3)
IL-6
Baseline	41	301.79 (186.60)	285.01 (183.07, 381.78)	43	808.53 (3136.82)	311.20 (240.99, 415.54)	0.305	0.219
Day7	38	448.33 (690.39)	315.33 (257.45, 410.93)	39	454.16 (695.06)	290.07 (218.66, 466.55)	0.971	0.603
Day14	32	831.81 (3164.62)	284.07 (146.07, 399.94)	33	236.35 (175.74)	247.05 (65.30, 335.29)	0.284	0.254
IL-1β
Baseline	42	2.3098 (6.9253)	0 (0, 0.2525)	43	1.7435 (5.7797)	0 (0, 0)	0.683	0.673
D7	38	9.2972 (27.3895)	1.04 (0, 6.4650)	39	3.8982 (5.8203)	0 (0, 6.53)	0.233	0.667
D14	31	4.6306 (21.9170)	0 (0, 0)	33	2.4309 (10.2287)	0 (0, 0)	0.605	0.860
TNF-α
Baseline	41	3.0855 (9.4621)	0 (0, 0.9150)	43	2.2817 (6.7125)	0 (0, 0.8750)	0.653	0.686
D7	38	14.0341 (44.7159)	0 (0, 2.8775)	39	7.5724 (25.7294)	0 (0, 0.1250)	0.438	0.185
D14	32	2.5014 (9.7507)	0 (0, 0)	33	10.3324 (27.1003)	0 (0, 0.5850)	0.127	0.172
S100
Baseline	41	20.5662 (23.8089)	11.06 (0.54, 33.365)	43	22.5909 (34.3016)	10.71 (0, 35.7)	0.755	0.615
D7	38	7.1120 (13.1432)	0 (0, 7.2525)	39	6.2299 (10.0888)	1.17 (0, 7.08)	0.742	0.700
D14	32	5.3647 (12.1376)	0 (0, 5.6925)	33	4.5278 (8.9198)	0 (0, 5.68)	0.752	0.757

CG, control group; TG, treatment group; SD, standard deviation; IL-6, interleukin-6; IL-1β, interleukin-1β; TNF-α, tumor necrosis factor-α; S100B, S100 calcium binding protein B; D7, 7th after AM or placebo administration; D14, 14th after AM or placebo administration.

^aIndependent t-test.

^bMann-Whitney Test.

4.4. Effect of AM on GOS and BI scores based on different baseline IL-6 serum levels

A stratified analysis based on baseline IL-6 serum levels was conducted. Among patients with higher IL-6 serum levels (median ≥7.28 pg/mL) at baseline, the percentage of good GOS scores at W2 and W4 was significantly higher in the TG than in the CG (W2: 35.3 % vs. 7.7 % for TG and CG, respectively, p = 0.042; W4: 62.5 % vs. 30.8 % for TG and CG, respectively, p = 0.044). Furthermore, a higher percentage of BI scores >60 in the TG than in the CG at W2 was observed (35.3 % vs. 7.7 %, respectively, p = 0.042) (Fig. 3a and b). No effect of AM on GOS or BI was observed among patients with lower IL-6 serum levels (median <7.28 pg/mL) at baseline. Additionally, no other significant effects of AM on outcomes were observed in the stratified analyses.

Fig. 3.

Stratified analysis based on baseline interleukin-6 (IL-6) serum levels GOS, Glasgow outcome scale; BI, Barthel index; CG, control group; TG, treatment group. (a) IL6 Median <7.28 pg/mL; (b) IL6 Median ≥7.28 pg/mL; *p < 0.05.

5. Adverse events and side effects

Since most patients are hospitalized while taking AM, the adverse events and side effects of AM could be observed by investigators and clinical assessors. Three patients died in this study, all of whom had diseases unrelated to administering AM. In addition, no noteworthy side effects or adverse events were found in this study.

6. Discussion

Spontaneous SAH is associated with high mortality and morbidity rates, primarily due to the development of vasospasm leading to delayed cerebral ischemia (DCI) after hemorrhage. Current research efforts concerning spontaneous SAH largely revolve around preventing and treating post-hemorrhagic vasospasm, aiming to reduce delayed cerebral ischemia and enhance prognosis²⁵ According the 2023 AHA guideline,²⁶ for patients who survive initial aneurysmal rupture, DCI represents a major concern.²⁷ An autopsy study finds the major cause of aSAH patients is due to undiagnosed and insufficiently treated cerebral arterial vasospasm.²⁸ However, the mechanisms that lead to loss of brain tissue viability secondary to DCI are more complex than vasospasm alone. Proposed mechanisms for DCI include blood-brain barrier disruption, microthrombosis, cortical spreading depolarization/ischemia, and failure of cerebral autoregulation.²⁹ Therapy directed at vasospasm alone has had limited impact on DCI or mortality in patients with aSAH. Clearly, effective risk stratification, prophylaxis, detection, and treatment of DCI represent an important frontier in the management of aSAH. Nimodipine,a dihydropyridine calcium channel blocker, was the only drug approved by the US Food and Drug Administration for clinical neurological improvement after aSAH.²⁶

This study aimed to evaluate the potential of AM in preventing post-SAH vasospasm and improving outcomes. The results in the present study showed that the percentages of good and bad GOS, and good and bad mRS scores, and FIM and BI scores at baseline, W1, W2, W4, and W12 did not demonstrate significant differences between the TG and CG, suggesting AM administration for two weeks did not improve neurological function in patients with acute spontaneous SAH.

While our results showed no significant difference in prognosis, we did find that IL-6 served as a prognostic factor, aligning with previous research.^30,31 During stratified analysis in the present study, we identified a significant association between AM administration and the GOS and BI outcomes within four weeks of follow-up, especially among those patients with higher baseline serum IL-6 levels (median ≥7.28 pg/mL). The percentage of good GOS scores at W2 and W4, and BI scores >60 at W2 was significantly higher in the TG than in the CG, suggesting serum IL-6 median ≥7.28 pg/mL levels in acute stage of patients with aSAH induces a severe neuroinflammatory response, and this severe inflammatory response promotes AM to produce a potential inhibitory effect on neuroinflammation. Therefore, decrease the development of vasospasm and DCI, thus, conducive the recovery of neurological function. However, this suggestion needs further study in the future. There are numerous studies have explored the connection between elevated IL-6 levels and later DCI. For example, Gaetani et al. reported increased IL-6 levels in aneurysm patients' CSF, suggesting a potential correlation with vasospasm.³² This trend has been supported by others, indicating correlations between IL-6 levels and both CSF and serum levels in relation to DCI.^33,34 IL-6's role as a key mediator in SAH complications is widely proposed. Observational studies have linked IL-6 and IL-1β increases shortly after SAH to subsequent vasospasm development.³⁵ Stable high IL-6 levels have also been associated with DCI, highlighting its potential as a tool for anticipating and mitigating vasospasm and brain injury.³⁶ Animal models have shown promising results, with IL-6 modulation resulting in reduced vasospasm rates and neuronal death.³⁶ Moreover, IL-1 receptor antagonists have demonstrated potential in reducing IL-6 expression, enhancing clinical recovery after SAH.³⁷

Our results also indicated that the levels of IL-6, IL-1β, TNF-α, and S100B in serum and CSF at baseline, D7, and D14 were similar between the CG and TG. Therefore, suggesting there is similar inflammatory response between TG and CG in the baseline, D7 and D14. IL-6, IL-1β, and TNF-α are pro-inflammatory cytokines that play a critical role in inflammation after stroke.³⁸ S100B is a biomarker for brain ischemia, and S100B elevation after aneurysmal SAH represents a poor prognosis.³⁹ Therefore, the anti-inflammatory role of AM in patients with acute spontaneous aSAH needs further study.

The present study found no significant adverse events or side effects in both the TG and CG. Therefore, oral or nasogastric tube administration of AM for two weeks was considered safe.

Some limitations in the present study include: 1) AM administration for only two weeks, and hence, longer duration of AM administration is needed in future studies; 2) single-center design and limited statistical power due to only 88 subjects in this study, and thus, future studies with multiple centers and larger sample sizes are needed; 3) cerebral blood flow (CBF) assessment was missing in this study, and additional assessment of CBF is needed in future studies; 4) the incidence of aSAH is very low and the restriction of inclusion and exclusion criteria in clinical trial resulting in difficulty recruiting subjects to achieve predetermined sample size. Therefore, longer term or multiple center study is need in the future; and 5) serum IL-6 levels are related to the prognosis of aSAH, therefore, serum IL-6 levels set up in inclusion criteria is very important in the future study.

7. Conclusions

The results of the present study showed the percentage of good GOS scores at W2 and W4, and BI scores >60 at W2 was significantly greater in the TG than in the CG in patients with aSAH, who serum IL-6 levels median ≥7.28 pg/mL. Therefore, suggesting AM can produce potential inhibitory effect on inflammation in high inflammatory aSAH patients. However, this suggestion needs further study in the future, such as multiple center and greater sample size, and animal experiments etc.

Ethics approval and consent to participate

The study protocol was approved by the office of the Human Protection Project of CMUH, Taichung, Taiwan. Patient informed consent was obtained prior to publication.

Consent for publication

Not applicable.

Availability of data

The data in this study are available to other researchers upon request.

CRediT authorship contribution statement

Chun-Chung Chen: Conceptualization, Methodology, Investigation, Data curation, Writing-original draft. Hung-Lin Lin: Conceptualization, Methodology. Jeng-Hung Guo: Conceptualization, Methodology. Xian-Xiu Chen: Conceptualization, Methodology. Der-Yang Cho: Conceptualization, Methodology. Wen-Ling Liao: Methodology, Data analysis, Conceptualization, Methodology, Supervision. Ching-Liang Hsieh: Conceptualization, Methodology, Supervision, Writing-review & editing.

Funding

This study received financial support from China Medical University Hospital (DMR-107-062; DMR-109-067) in Taiwan. Additionally, partial funding was provided by the Ministry of Health and Welfare Clinical Trial and Research Center of Excellence (MOHW106-TDU-B-212-113004) in Taiwan. Furthermore, this study received financial assistance from the Chinese Medicine Research Center of China Medical University as part of the Featured Areas Research Center Program under the framework of the Higher Education Sprout Project by the Ministry of Education in Taiwan (CMRC-CENTER-5).

Declaration of competing interest

The authors declare that there are no conflicts of interest associated with this manuscript, and no significant financial support was received that would influence the findings.

All the listed authors have read and approved the submitted manuscript.