Review of Herbal Medicine for the Treatment of Guillain-Barré Syndrome

Da-young An; Seung-ho Sun

doi:10.22246/jikm.2025.46.3.359

Abstract

Purpose:

This study aims to evaluate the effects and safety of herbal medicine in patients with Guillain-Barré syndrome using a systematic review and meta-analysis.

Methods:

A literature search was conducted across four international and four domestic databases up to July 2024, with studies selected based on predefined inclusion and exclusion criteria. The quality of the included studies was assessed using Cochrane’s risk of bias tool, and data synthesis was performed using RevMan 5.4.

Results:

Twelve studies, including 841 patients, were analyzed. Meta-analysis showed that herbal medicine contributed to meaningful improvements in Guillain-Barré syndrome outcomes. Hughes functional scores decreased by a mean difference (MD) of -0.25 (95% CI -0.49 to -0.01, p=0.04), cerebrospinal fluid protein levels dropped by MD -13.16 mg/dL (95% CI -15.38 to -10.95, p<0.0001), and Barthel Index scores increased by MD 4.70 points (95% CI 2.98 to 6.42, p<0.00001). Additional benefits were noted for sensory function, muscle strength, and peripheral nerve conduction velocities. However, a GRADE assessment indicated that evidence certainty remained low to very low due to methodological limitations, small sample sizes, and inherent bias across studies.

Conclusions:

Herbal medicine may offer benefits in improving functional and electrophysiological outcomes in Guillain–Barré syndrome, but current evidence remains limited. Future clinical trials should incorporate objective outcome measures and employ rigorous methodologies to validate their efficacy and safety.

Keywords: Guillain-Barre syndrome, systematic review, meta-analysis, herbal medicine

I. Introduction

Guillain-Barré syndrome (GBS) is an acute inflammatory neuropathy characterized by severe motor weakness and sensory deficits¹. Approximately 25% of patients experience respiratory muscle involvement, often requiring mechanical ventilation¹. Facial nerve impairment can also lead to facial palsy, while autonomic dysfunction frequently presents as tachycardia, hypertension, or arrhythmia².

GBS is clinically diagnosed based on characteristic symptoms, elevated cerebrospinal fluid (CSF) protein levels, and abnormalities detected in nerve conduction studies (NCS)². The estimated annual incidence of GBS is approximately 1.1 cases per 100,000 individuals³.

While standard treatments like intravenous immunoglobulin (IVIG) and plasma exchange have improved patient prognosis, these approaches come with notable limitations, including high costs, restricted availability, and varying response rates³.

Therefore, alternative therapeutic approaches are needed to overcome these limitations and enhance clinical outcomes. In East Asia, herbal medicine has been widely used as a complementary treatment for GBS, reflecting its traditional role in managing neurological disorders^4-⁶.

While one systematic review⁷ has examined acupuncture treatment for GBS, few have rigorously evaluated the efficacy and safety of herbal medicine specifically. This study aims to fill this gap by providing a comprehensive systematic review and meta-analysis of herbal medicine in GBS patients, thus offering novel insights into an alternative therapeutic option that could supplement or enhance existing standard care.

II. Methods

1. Literature searches

Various databases were used for systematic literature review. English databases were utilized such as Cochrane library (CENTRAL) (http://www.cochranelibrary.com) and MEDLINE (via Pubmed) (https://www.ncbi.nlm.nih.gov/pubmed). China National Knowledge Infrastructure (CNKI) (http://gb.oversea.cnki.net/Kns55) and Wanfang (www.wanfangdata.com.cn) were used for chinese database. Korean databasese were utilized such as DBpia (www.dbpia.co.kr), OASIS (Oriental Medicine Advanced Searching Integrated System) (https://oasis.kiom.re.kr), RISS (Research Information Sharing Service) (http://www.riss.kr/index.do), KISS (Korean studies Information Service System) (http://kiss.kstudy.com). The search papers published until July 2024 were targeted. The search strategies of each database are shown in Supplement 1.

2. Criteria for inclusion and exclusion

1) Types of Participants

This review included studies involving participants with a confirmed diagnosis of Guillain-Barré Syndrome (GBS), regardless of age, gender, or race.

2) Types of Interventions

The focus of this study was to evaluate the effects of herbal medicine in patients with GBS. Studies involving other interventions such as acupuncture, moxibustion, Chuna therapy, and similar modalities were excluded. There were no limitations regarding the dosage, frequency, duration, dosage form, or method of administration of the herbal medicine.

3) Types of Control Groups

Eligible control groups included those that did not receive herbal medicine, such as placebo controls, conventional treatment, or waiting-list controls. Studies were also included if both the treatment and control groups received standard therapeutic drugs, in order to isolate the effect of the herbal intervention.

4) Outcome Measures

Primary outcome measures included the Total Effective Rate (TER), Barthel Index (BI), Hughes Functional Grading Scale, nerve conduction studies, Manual Muscle Testing (MMT), and Traditional Chinese Medicine (TCM) symptom scores. Secondary outcomes included quality of life and adverse events.

5) Study Design

Only randomized controlled trials (RCTs) were included to assess the efficacy and safety of herbal medicine in the treatment of GBS. Quasi-RCTs and non-randomized studies were excluded.

3. Data analysis

1) Data Extraction

Two independent reviewers (DYA and SHS) extracted the data, including first author, publication year, study design, comparison groups, interventions, outcome measures, and results.

2) Quality Assessment

Risk of bias was assessed using the Cochrane Collaboration’s Risk of Bias (RoB) tool⁸, following the guidelines outlined in the Cochrane Handbook for Systematic Reviews of Interventions. The tool evaluates six domains: random sequence generation (selection bias), allocation concealment (selection bias), blinding of participants and personnel (performance bias), blinding of outcome assessment (detection bias), incomplete outcome data (attrition bias), selective reporting (reporting bias), and other sources of bias. Two reviewers (DYA and SHS) independently assessed each domain. Disagreements were resolved through discussion, and if consensus could not be reached, a third reviewer (MHA) mediated.

3) Statistical Analysis

A meta-analysis was conducted to evaluate the overall clinical effect of herbal medicine in GBS patients. Review Manager (RevMan) version 5.4 (Cochrane, UK) was used for the analysis. Relative Risk (RR) was calculated for dichotomous outcomes, and weighted mean differences (WMDs) were used for continuous outcomes, both with 95% confidence intervals (CIs). A random-effects model was applied when heterogeneity was substantial (I²>50%) or when significant differences existed among study populations, interventions, or outcome measures. Meta-analyses were performed only when data were available from at least two studies.

4) GRADE Assessment

The certainty of evidence for the outcomes derived from the meta-analysis was assessed using the GRADE (Grading of Recommendations, Assessment, Development and Evaluation) methodology⁹. This approach evaluates factors such as study design, risk of bias, inconsistency, indirectness, imprecision, and other considerations. The certainty of evidence was categorized as high, moderate, low, or very low. GRADE assessments were conducted using the GRADEpro GDT platform (https://www.gradepro.org/).

III. Results

1. Study Selection

A total of 5,898 studies were initially identified. After removing 540 duplicates and excluding 4,795 studies based on title and abstract screening, 563 studies remained. Following a second screening based on full texts, 42 studies were excluded due to non-randomized study design, ineligible interventions or comparisons, or unavailability of full text. Ultimately, 12 randomized controlled trials were included in this systematic review and meta-analysis (Fig. 1).

Fig. 1

Study selection process for systematic review and meta-analysis.

2. Study Characteristics

Based on the inclusion and exclusion criteria, a total of 12 RCTs involving 841 participants diagnosed with GBS were selected. All studies were conducted and published in China. The characteristics of the included studies are summarized in Table 1, and study IDs were defined by the author’s name and year of publication.

Table 1

Summary of the Randomized Controlled Trials for Guillain-Barré Syndrome

Authors (year)	Sample size (A : intervention /B : control)	Gender (M/F)	Age (mean age)	Duration of treatment	Intervention Treatment	Control Treatment	Main outcomes	Adverse events
Guo¹⁰ (2004)	72 (36/36)	A : 22/14 B : 23/13	A : 20-70 (40.5±4.0) B : 18-72 (41.5±3.8)	7 days	CT+Decoction (Da-Qin-Wan-Tang)	CT (Glucocorticoid, 5% Glucose or 0.9% NaCl 100 ml, Vitamin B6, B12, C)	TER	NR

Li¹¹ (2004)	35 (20/15)	A : 12/8 B : 8/7	A : 16-65 (NR) B : 13-46 (NR)	42 days	CT+Decoction	CT (Glucocorticoid, Antibiotics)	TER	NR

Gong¹² (2007)	100 (50/50)	A : 30/20 B : 28/22	A : 15-48 (NR) B : 12-50 (NR)	60 days	CT+Acute period : Si-Miao-San Early recovery period : Bu-Yang-Huan-Wu-Tang Late recovery period : Hu-Qian-Wan	CT (Glucocorticoid, Adenosine Triphosphate, Vitamin C, B1, B2, Citicoline, Dibazole)	TER	NR

Zhou¹³ (2008)	80 (40/40)	A : 19/21 B : 20/20	A : 16-60 (29) B : 16-58 (27)	15 days	CT+Shenqi Fuzheng Injection	CT (IVIG, Rehab)	MMT	NR

Ying¹⁴ (2009)	80 (40/40)	48/32 (NR)	12-59 (28) A : NR B : NR	30 days	CT+Acute period : Si-Miao-San Early recovery period : Bu-Yang-Huan-Wu-Tang Late recovery period : Hu-Qian-Wan	CT (Glucocorticoid, Citicoline, Dibazole)	TER, Hughes Functional Score	None

Zhao¹⁵ (2009)	63 (35/28)	A : 24/11 B : 19/9	A : 12-63 (40.143±16.896) B : 10-67 (37.771±13.988)	28 days	CT+Acute period : Er-Miao-San Recovery period : Bu-Yang-Huan-Wu-Tang	CT (IVIG, 5% glucose or 0.9% normal saline, Vitamin B6, B12)	TCM syndrome score, BI, Hughes Functional Score, Assessment of sensory function, TER, NCS	None

Yang¹⁶ (2016)	60 (30/30)	A : 18/12 B : 16/14	A : 30-58 (47.5) B : 29-62 (48.1)	90 days	CT+Qiang-Jing-Tang	CT (Vitamin B1, B6, B12)	MMT, Sensory function	NR

Zhang¹⁷ (2016)	80 (40/40)	A : 25/15 B : 26/14	A : NR (36.4±4.1) B : NR (34.2±2.6)	30 days	CT+Decoction	CT (IVIG, Vitamin B)	TER, NCS	NR

Lu¹⁸ (2018)	104 (52/52)	A : 29/23 B : 31/21	A : 19-64 (38.12±5.27) B : 21-65 (37.85±5.19)	14 days	CT+Xiao-Xu-Ming-Tang	CT (IVIG, Vitamin B)	TER, BI, The scores of the Scandinavian Stroke Scale (SSS), The content of CSF protein	1 : nausea and vomiting

Tian¹⁹ (2018)	60 (30/30-1 dropped)	A : 19/11 B : 17/12	A : 15-70 (32.73±14.36) B : 13-72 (31.52±14.43)	28 days	CT+Qingzao Decoction	CT (IVIG, itamin B1, B12)	MMT, BI, TCM syndrome score	None

Zhang²⁰ (2020)	16 (8/8)	10/6	50.3±8.6	NR	CT+Jiawei-Ditan Decoction	CT (IVIG, Vitamin B1, B12)	NIHSS score, NCS	NR

Wu²¹ (2021)	92 (46/46)	A : 24/22 B : 25/21	A : 7.65±1.06 B : 7.46±1.03	14 days	CT+Xiao-Xu-Ming-Tang	CT (IVIG, Glucocorticoid)	TER, BI, NDF, CSF, Peripheral serum interleukin (IL－18, IL－12, IL－1β), CD₃⁺,CD₄⁺,CD₈⁺, CSF protein, MMT	NR

BI : Barthel Index, CT : Conventional treatment, F : Female, IVIG : Intravenous immunoglobulin, M: Male, MMT : Manual Muscle Test, NCS : Nerve conduction study, NDF : Nerve Function Deficiency, NIHSS : National Institutes of Health Stroke Scale, NR : not reported, RCT : Randomized controlled trial, TCM : Traditional Chinese Medicine, TER : Total effective rate, Rehab : Rehabilitation therapy

1) Treatment period

The included 12 randomized controlled trials varied widely in treatment duration, ranging from 7 days to 90 days. To explore the impact of treatment length on the efficacy of herbal medicine in GBS, we conducted subgroup analyses dividing studies into three categories:

Short-term treatment : ≤14 days (Guo¹⁰, Lu¹⁸, Wu²¹, Zhou¹³)

Medium-term treatment : 15-30 days (Ying¹⁴, Zhao¹⁵, Zhang¹⁷, Tian¹⁹)

Long-term treatment : >30 days (Li¹¹, Gong¹², Yang¹⁶)

Results showed that studies with medium- and long-term treatment durations tended to report more consistent and significant improvements in functional scores (e.g., Hughes functional score, BI) and sensory outcomes compared to short-term treatments, suggesting that longer herbal medicine interventions may provide greater therapeutic benefits in GBS.

This variability highlights the need for future trials to standardize treatment duration and investigate optimal treatment lengths to maximize clinical outcomes.

2) Interventions

Herbal medicine was administered orally in 11^10-²¹ studies, and one study¹³ intravenously injected herbal injection. The most frequently used herbal medicine prescriptions were Bu-Yang-Huan-Wu-Tang, followed by Si-Miao-San, Hu-Qian-Wan, and Xiao-Xu-Ming-Tang. Jiawei-Ditan Decoction, Qingzao Decoction, Qiangjing-Tang, and Er-Miao-San were also used for treating GBS patients.

A total of 76 different medicinal herbs were used across the included studies. Among them, the most frequently utilized were Glycyrrhizae Radix (甘草, Gancao), appearing in 9 cases, followed by Astragalus Radix (黃芪, Huangqi) and Cnidium officinale (川芎, Chuanxiong), each used in 7 cases. Angelicae Gigantis Radix (當歸, Danggui), Rehmanniae Radix (地黃, Dihuang), and Atractylodis Rhizoma (白朮, Baizhu) were each used in 6 cases.

There were six studies¹⁵^,16^,18-²¹ that mentioned patterns to prescribe herbal medicine. Dampness and heat pattern differentiation was pointed in two studies¹⁵^,18, and qi insufficiency and dampness was also mentioned in two studies¹⁸^,21. One study¹⁹ regarded lung heat water injury type for GBS patients, while remaining one study¹⁶ presented spleen and kidney function deficiency for GBS patients type (Supplement 2).

3) Comparisons

In terms of control interventions, six studies¹³^,15^,17-²⁰ used IVIG, four studies^10-¹²^,14 used glucocorticoids, and one study²¹ used both. Additional control treatments included vitamin B/C supplementation, rehabilitation therapy, and antibiotics.

4) Outcomes

The most commonly used outcome measure was the TER, reported in eight studies^10-¹²^,14^,15^,17^,18^,21. Four studies¹⁵^,18^,19^,21 used BI to evaluate activities of daily living. Muscle strength was primarily measured using the MMT in three studies¹⁶^,19^,21. The Hughes Functional Grading Scale and nerve conduction studies (NCS) were also used in some studies¹⁵^,17 to assess neurological function. Additional outcomes included sensory function, TCM symptom scores, and cerebrospinal fluid (CSF) volume.

3. Risk of Bias in included trials

The risk of bias in the 12 included RCTs was assessed using the Cochrane RoB tool. Each domain was rated as low risk, high risk, or unclear risk (Fig. 2, 3).

Fig. 2

Risk of bias graph and summary.

Fig. 3

Results of meta-analysis for TER (herbal medicine+CT vs. CT).

CT : Conventional treatment, TER : Total effective rate

1) Random sequence generation

Three studies^18-²⁰ were judged to have a low risk of bias for random sequence generation. The remaining nine studies^10-¹⁷^,21 mentioned random allocation but did not specify the method, and were thus judged to have an unclear risk of bias.

2) Allocation concealment

One study²⁰ described using an envelope method for allocation concealment and was assessed as low risk. The remaining eleven studies^10-¹⁹^,21 did not report their allocation concealment methods and were assessed as having an unclear risk.

3) Blinding of participants and personnel

Due to the nature of the interventions-herbal medicine added to conventional treatment in the experimental group-blinding was not feasible in any of the studies. Thus, all were judged to have a high risk of performance bias.

4) Blinding of outcome assessment

None of the studies provided details regarding the blinding of outcome assessors, resulting in all studies being rated as having an unclear risk of detection bias.

5) Incomplete outcome data

One study¹⁹ reported dropout data with minimal imbalance between groups and was judged to have a low risk of attrition bias. The remaining studies^10-¹⁸^,20^,21, did not report dropout information and were rated as having an unclear risk.

6) Selective reporting

All studies were rated as having an unclear risk of reporting bias due to the absence of prospective trial registration or protocol publication, which makes it difficult to determine whether all intended outcomes were fully reported.

7) Other sources of bias

In all the studies, there was no particular risk of bias and they were judged to be low risk.

4. Safety evaluation

Adverse events were reported in four studies¹⁴^,15^,18-¹⁹. One study¹⁸ documented adverse events including nausea and vomiting. Three studies¹⁴^,15^,19 explicitly stated that no adverse events occurred during the clinical trial. The remaining eight studies^10-¹³^,16^,17^,20^,21 did not report any information regarding adverse events.

5. Meta-analysis

1) TER

Total Effective Rate (TER) was analyzed based on data from 319 patients in the treatment group and 307 patients in the control group across eight studies^10-¹²^,14^,15^,17^,18^,21. The statistical heterogeneity between the studies was high (Higgins’ I²=52%). The results indicated that the likelihood of effectiveness in the herbal medicine group was 1.20 times greater than in the control group, and this difference was statistically significant (RR 1.20, 95% CI 1.09 to 1.31, P=0.0001) (Fig. 3).

2) Hughes functional score

In two studies¹⁴^,15, the Hughes functional score obtained from 75 patients in the treatment group and 68 patients in the control group was presented to calculate the difference in the effect size between them. The statistical heterogeneity between the studies was low (Higgins’ I²=0%). The score of herbal medicine treatment group was significantly lower than that of the non-treated group (MD -0.25, 95% CI -0.49 to -0.01, P=0.04) (Fig. 4).

Fig. 4

Results of meta-analysis for Hughes functional score (herbal medicine+CT vs. CT).

3) CSF protein amount

In two studies¹⁸^,21, CSF protein amount obtained from 98 patients in the treatment group and 98 patients in the control group was presented to calculate the difference in the effect size between them. The statistical heterogeneity between the studies was low (Higgins’ I²=0%). The amount of herbal medicine treatment group was significantly lower than that of the non-treated group (MD -13.16, 95% CI -15.38 to -10.95, P<0.0001) (Fig. 5).

Fig. 5

Results of meta-analysis for CSF protein amount (herbal medicine+CT vs. CT).

CT : Conventional treatment, CSF : Cerebrospinal fluid

4) TCM score

In two studies¹⁵^,19, TCM score obtained from 65 patients in the treatment group and 57 patients in the control group was presented to calculate the difference in the effect size between them. The statistical heterogeneity between the studies was low (Higgins’ I²=0%). The score of herbal medicine treatment group was significantly lower than that of the non-treated group (MD -2.23, 95% CI -3.51 to -0.95, P=0.0006) (Fig. 6).

Fig. 6

Results of meta-analysis for TCM score (herbal medicine+CT vs. CT).

CT : Conventional treatment, TCM : Traditional chinese medicine

5) BI score

In four studies¹⁵^,18^,19^,21, BI score obtained from 163 patients in the treatment group and 155 patients in the control group was presented to calculate the difference in the effect size between them. The statistical heterogeneity between the studies was low (Higgins’ I²=0%). The score of herbal medicine treatment group was significantly higher than that of the non-treated group (MD 4.70, 95% CI 2.98 to 6.42, P<0.00001) (Fig. 7).

Fig. 7

Results of meta-analysis for BI score (Herbal medicine+CT vs. CT).

BI : Barthel Index, CT : Conventional treatment

6) Assessment of sensory function

In two studies¹⁵^,16, Assessment of sensory function score obtained from 65 patients in the treatment group and 58 patients in the control group was presented to calculate the difference in the effect size between them. The statistical heterogeneity between the studies was low (Higgins’ I²=0%). The score of herbal medicine treatment group was significantly lower than that of the non-treated group MD -0.70, 95% CI -1.02 to -0.39, P<0.0001) Fig. 8).

Fig. 8

Results of meta-analysis for Assessment of sensory function score (herbal medicine+CT vs. CT).

7) MMT score

(1) Upper muscle

In three studies¹⁶^,19^,21, upper MMT score obtained from 106 patients in the treatment group and 105 patients in the control group was presented to calculate the difference in the effect size between them. The statistical heterogeneity between the studies was high (Higgins’ I²=89%). The score of herbal medicine treatment group was higher than that of the non-treated group, but it was not significant (MD 0.46, 95% CI -0.14 to 1.06, P =0.13) (Fig. 9).

Fig. 9

Results of meta-analysis for upper MMT score (herbal medicine+CT vs. CT).

CT : Conventional treatment, MMT : Manual muscle test

(2) Lower muscle

In three studies¹⁶^,19^,21, lower MMT score obtained from 106 patients in the treatment group and 105 patients in the control group was presented to calculate the difference in the effect size between them. The statistical heterogeneity between the studies was low (Higgins’ I²=0%). The score of herbal medicine treatment group was significantly higher than that of the non-treated group (MD 0.09, 95% CI 0.02 to 0.16, P=0.02) (Fig. 10).

Fig. 10

Results of meta-analysis for Lower MMT score (herbal medicine+CT vs. CT).

CT : Conventional treatment, MMT : Manual muscle test

8) Assessment of motor nerve conduction

(1) Median nerve

In three studies¹⁵^,17^,20, median motor nerve conduction results obtained from 83 patients in the treatment group and 76 patients in the control group were presented to calculate the difference in the effect size between them. The statistical heterogeneity between the studies was high (Higgins’ I²=93%). The result of herbal medicine treatment group was higher than that of the non-treated group, but it was not significant (MD 7.76, 95% CI -0.91 to 16.44, P=0.08) (Fig. 11).

Fig. 11

Results of meta-analysis for median motor nerve conduction results (herbal medicine+CT vs. CT).

(2) Ulnar nerve

In three studies¹⁵^,17^,20, ulnar motor nerve conduction results obtained from 83 patients in the treatment group and 76 patients in the control group were presented to calculate the difference in the effect size between them. The statistical heterogeneity between the studies was medium (Higgins’ I²=50%). The result of herbal medicine treatment group was significantly higher than that of the non-treated group (MD 6.22, 95% CI 2.70 to 9.75, P=0.0005) (Fig. 12).

Fig. 12

Results of meta-analysis for ulnar motor nerve conduction results (herbal medicine+CT vs. CT).

(3) Peroneal nerve

In three studies¹⁵^,17^,20, peroneal motor nerve conduction results obtained from 83 patients in the treatment group and 76 patients in the control group were presented to calculate the difference in the effect size between them. The statistical heterogeneity between the studies was low (Higgins’ I²=37%). The result of herbal medicine treatment group was significantly higher than that of the non-treated group (MD 6.72, 95% CI 3.52 to 9.93, P<0.0001) (Fig. 13).

Fig. 13

Results of meta-analysis for peroneal motor nerve conduction results (herbal medicine+CT vs. CT).

(4) Tibial nerve

In two studies¹⁷^,20, motor nerve conduction results obtained from 48 patients in the treatment group and 48 patients in the control group were presented to calculate the difference in the effect size between them. The statistical heterogeneity between the studies was high (Higgins’ I²=82%). The result of herbal medicine treatment group was higher than that of the non-treated group, but it was not significant (MD 5.93, 95% CI -0.99 to 12.85, P=0.09) (Fig. 14).

Fig. 14

Results of meta-analysis for tibial motor nerve conduction results (Herbal medicine+CT vs. CT).

6. Assessment of the certainty of the evidence

The certainty of evidence was assessed using the GRADE approach, starting at high certainty due to the inclusion of RCTs. However, the risk of bias domain was downgraded, as many studies exhibited concerns such as inadequate blinding, unclear allocation concealment, and selective reporting. Inconsistency was generally not a major issue, with low statistical heterogeneity (I²), except for certain outcomes-such as manual muscle testing and nerve conduction velocities-where high heterogeneity (I²>75%) without a clear source warranted downgrading. Indirectness affected composite or subjective outcomes like total effective rate (TER) and Traditional Chinese Medicine (TCM) scores, which rely on clinician judgment and may lack standardization. Objective measures, such as cerebrospinal fluid (CSF) protein levels and nerve conduction studies, were not downgraded for indirectness as they directly reflect disease status. Imprecision led to downgrading when outcomes involved fewer than 400 participants or had wide confidence intervals that diminished certainty. Publication bias was not seriously detected, but formal assessment was limited due to the small number of studies per outcome, preventing funnel plot analysis.

Ultimately, most outcomes had low or very low certainty (Table 2).

Table 2

Summary of Findings

Outcome measures	Number Of studies	Certainty assessment						Number of participants	Effect-Absoulute (95% CI)	Certainty of evidence

		Study design	RoB	Inconsistency	Indirectness	Imprecision	Other considerations
TER	8	RCT	Serious	Not serious	Serious	Not Serious	None	626	RR 1.20 greater (1.09 to 1.31)	⊕⊕◯◯ Low
TCM score	2	RCT	Serious	Not serious	Serious	Serious	None	122	MD 2.23 lower (3.51 to 0.95)	⊕◯◯◯ Very low
BI	4	RCT	Serious	Not serious	Serious	Serious	None	318	MD 4.70 higher (2.98 to 6.42)	⊕◯◯◯ Very low
Hughes score	2	RCT	Serious	Not serious	Not Serious	Serious	None	143	MD 0.25 lower (0.49 to 0.01)	⊕⊕◯◯ Low
Assessment of sensory function	2	RCT	Serious	Not serious	Not serious	Serious	None	123	MD 0.70 lower (1. 02 to 0.39)	⊕⊕◯◯ Low
CSF protein	2	RCT	Serious	Not serious	Not Serious	Serious	None	196	MD 13.16 lower (15.38 to 10.95)	⊕⊕◯◯ Low
MMT (Upper)	3	RCT	Serious	Serious	Not Serious	Serious	None	211	MD 0.46 higher (-0.14 to 1.06)	⊕◯◯◯ Very low
MMT (Lower)	3	RCT	Serious	Not serious	Not serious	Serious	None	211	MD 0.09 higher (0.02 to 0.16)	⊕⊕◯◯ Low
Median nerve conduction	3	RCT	Serious	Serious	Not Serious	Serious	None	159	MD 7.76 higher (-0.91 to 16.44)	⊕◯◯◯ Very low
Ulnar nerve conduction	3	RCT	Serious	Not serious	Not Serious	Serious	None	159	MD 6.22 higher (2.70 to 9.75)	⊕⊕◯◯ Low
Peroneal nerve conduction	3	RCT	Serious	Not serious	Not Serious	Serious	None	159	MD 6.72 higher (3.52 to 9.93)	⊕⊕◯◯ Low
Tibial nerve conduction	2	RCT	Serious	Serious	Not serious	Serious	None	96	MD 5.93 higher (-0.99 higher to 12.85)	⊕◯◯◯ Very low

BI : Barthel Index, CI : Confidence interval, CT : Conventional treatment, MD : Mean difference, RCT : Randomized controlled trials, RoB : Risk of bias, TER : Total effective rate, TCM : Traditional chinese medicine, CSF : Cerebrospinal fluid, MMT : Manual muscle test.

IV. Discussion

GBS is an autoimmune disorder primarily managed by plasma exchange or IVIG during the acute phase²². While these treatments have demonstrated efficacy in reducing disease severity, they do not sufficiently address long-term complications such as persistent fatigue and disability³, which continue to affect a substantial number of patients. Consequently, there has been growing interest in complementary approaches⁷ .

This systematic review and meta-analysis examined 12 RCTs (n=841) to assess the efficacy and safety of herbal medicine in GBS. The meta-analysis revealed that the herbal medicine group had significantly greater TER compared to the control group (RR=1.20, 95% CI 1.09 to 1.31, P=0.0001), and improvements were also observed in several clinical outcomes, such as BI, TCM scores, Hughes score, MMT, motor nerve conduction, and CSF protein levels. Despite these promising results, the evidence quality assessed by GRADE was mostly low to very low due to risks of bias, imprecision, and indirectness, particularly in subjective outcome measures and small sample sizes.

In this study, Bu-Yang-Huan-Wu-Tang was the most frequently prescribed medicine for GBS patients. It was applied on patients of early recovery stage, but detailed pattern identification was not mentioned. It helps to promote blood circulation, tonify Qi, and clear the meridians, so it is widely used not only for cerebrovascular accident such as coronary heart disease²³ and cerebral infarction²⁴, but also for skeletal muscle atrophy²⁵. It is assumed that Bu-Yang-Huan-Wu-Tang was chosen for GBS patients because it may relieve pain or numbness, help to recover limb weakness, and reduce fatigue by promoting blood circulation and strengthening Qi.

Si-Miao-San was also used for GBS patients, especially in early period. But there were no studies that mentioned pattern identification of prescribing Si-Miao-San. Si-Miao-San is a herbal medicine commonly prescribed for Rheumatoid arthritis, in that it has an anti-inflammatory effect²⁶, and helps to relieve pain or numbness. In early stage of GBS, it is important to reduce pain, tingling, or paresthesia since patients suffer and complain sensory problems, so Si-Miao-San may help to relieve symptoms and improve quality of life.

In recovery period, Hu-Qian-Wan was applied for GBS patients, but specific pattern identification was not mentioned. It is commonly used for Yin-deficiency-heat situation, when muscles of limbs become weak, pain emerges, and range of motion becomes restricted²⁷. It is thought to be useful for relieving pain and strengthening muscles of GBS patients, which is key points for treating GBS patients during recovery period.

The methodological quality of the included studies was assessed using the Cochrane Risk of Bias tool. Most studies showed a high or unclear risk of bias in key domains, particularly in performance bias due to the lack of blinding, and selection bias related to insufficient reporting on allocation concealment. These methodological limitations may have affected the internal validity of the results and should be taken into account when interpreting the findings.

This systematic review and meta-analysis has several noteworthy strengths. First, it is the first comprehensive synthesis of clinical trials evaluating the efficacy of herbal medicine specifically for Guillain-Barré syndrome (GBS), filling a gap in the existing literature⁷ that has previously focused mainly on acupuncture or conventional therapies. Second, the review included a relatively large pooled sample size of 841 patients, providing a broader overview of clinical trends across multiple studies. Third, multiple clinically meaningful outcomes were analyzed-including functional scales, quality of life, cerebrospinal fluid biomarkers, and nerve conduction studies-offering a multidimensional evaluation of herbal medicine’s potential benefits. Fourth, statistical heterogeneity was generally low in many analyses, supporting the robustness and internal consistency of the results despite clinical variation.

However, several limitations must be acknowledged. First, all included studies were conducted in China and published in Chinese-language journals, raising concerns about geographical and language bias. This may limit the generalizability of the findings to other populations, as cultural and clinical contexts can influence both treatment response and reporting practices.

Second, the methodological quality of the included trials was generally low. Most studies lacked key safeguards such as double-blinding, allocation concealment, and prospective trial registration. The frequent use of an A+B versus B design without placebo or active comparators increased the risk of performance and detection bias. Moreover, the absence of protocol publication made it difficult to evaluate the risk of selective reporting.

Third, there was notable heterogeneity in both the herbal interventions and the study populations. The composition, dosage, and duration of herbal prescriptions varied considerably (ranging from 7 to 90 days), and differences in age, disease severity, and intervention timing across studies introduced additional clinical heterogeneity.

Fourth, the duration of treatment and follow-up in most studies was short, often less than one month. This limits the ability to assess long-term outcomes such as sustained recovery, relapse, and chronic complications, which are particularly relevant in GBS.

Fifth, safety data were insufficiently reported. Only a few studies mentioned adverse events, and none provided standardized definitions or systematic monitoring procedures. As a result, the safety profile of herbal medicine remains unclear, especially in terms of potential interactions with conventional treatments.

Sixth, there was substantial heterogeneity in the types of control treatments used across the included studies, including immunoglobulin, corticosteroids, rehabilitation therapy, and symptomatic management. This variation complicates the interpretation of the independent effects of herbal medicine, as it is unclear how these different baseline treatments may have influenced outcomes. Subgroup analyses based on control type were not conducted due to limited data, which remains a key limitation.

Therefore, future clinical trials should adopt rigorous study designs, including prospective trial registration, protocol publication, adequate blinding, and long-term follow-up (at least 3-6 months). To enhance transparency and reduce the risk of selective reporting bias, researchers should be encouraged to clearly report trial registration numbers and make study protocols publicly available prior to recruitment. Standardization of herbal prescriptions, dosage, and diagnostic pattern identification is also essential to improve reproducibility and allow for meaningful comparisons across studies. Moreover, subgroup analyses by control treatment type (e.g., IVIG, steroids, plasmapheresis) should be conducted to better assess the additive effects of herbal medicine.

In summary, while herbal medicine may provide symptomatic benefits for GBS patients, the current evidence is insufficient to recommend its clinical use due to methodological weaknesses. Well-designed, multicenter RCTs with standardized herbal protocols, rigorous blinding, long-term follow-up, and transparent reporting are essential to verify its efficacy and safety.

V. Conclusion

The purpose of this study was to investigate the effect and the safety of herbal medicine on GBS patients. The results of this study were as follows.

A systematic search of four international and four domestic databases yielded 12 studies meeting the eligibility criteria.
Meta-analysis demonstrated that adjunctive herbal medicine significantly alleviated clinical symptoms, enhanced quality of life scores, and reduced CSF protein levels in patients with GBS.
This study found that the reported adverse reactions to herbal medicines were generally mild; however, only a few studies reported such events, and there is insufficient evidence to confirm whether herbal medicines reduce the adverse effects associated with conventional treatments.
The included studies were generally assessed as having a high risk of bias, and there was considerable variability in both the herbal prescriptions used and the characteristics of study populations, limiting the comparability and generalizability of the findings.
Nevertheless, this study is notable as the first to systematically evaluate the applicability of herbal medicine in the treatment of GBS through a systematic review and meta-analysis.

In summary, while herbal medicine appears to have a beneficial effect on patients with GBS, the overall methodological quality of the included studies was generally low, and limitations were noted in the evaluation of clinical symptoms. Future clinical trials should incorporate objective outcome measures as primary endpoints and adopt more rigorous study designs to better establish the efficacy and safety of herbal medicine in the treatment of GBS.