Efficacy and safety in treating refractory epilepsy with a combination of traditional Chinese and Western medicine: A meta-analysis of randomized controlled trials

Abstract

Background:

Refractory epilepsy (RE) remains a significant topic in neurology. With the advantages of multi-target mechanisms and higher compliance in traditional Chinese medicine (TCM), RE patients shift their focus to this therapy. However, contradictions exist between reported efficacy and safety. The purpose of this meta-analysis was to evaluate the efficacy and safety of integrated traditional Chinese and Western medicine in RE.

Methods:

Seven databases (CNKI, Wanfang, VIP, PubMed, Web of Science, Cochrane Library, Embase) were searched from inception to March 2023 in English or Chinese. The primary outcomes were epilepsy frequency, clinical efficacy (total effective rate). The secondary outcomes were quality of life score, TCM symptom score, electroencephalogram abnormality rates, epilepsy duration, and incidence of adverse reactions. We performed meta-analysis with fixed- or random-effects model based on heterogeneity (I² > 50% or P < .10). Heterogeneity was detected by subgroup analysis and sensitivity analysis. The risk of bias tool was used to quality assessment. Trial sequential analysis (TSA) was performed with primary outcome. We conducted a Grading of Recommendations Assessment, Development, and Evaluation assessment for all outcomes in this meta-analysis.

Results:

Twenty randomized controlled trials (RCTs) with 1521 RE patients were included, with 760 cases in the treatment group and 761 cases in the control group. Meta-analysis showed that the treatment group for RE reduced epilepsy frequency [MD = −2.37, 95% Cl = (−3.17, −1.57), P < .00001]; alleviated clinical symptoms [MD = −2.32, 95% Cl = (−4.13, −0.51), P = .01]; shortened the duration of epilepsy [MD = −1.20, 95% Cl = (−1.71, −0.68), P < .00001]; improved clinical outcome [RR = 1.27, 95% Cl = (1.19, 1.35), P < .00001]; and improved quality of life [MD = 4.45, 95% Cl = (1.60, 7.30), P = .002]. However, there was no statistically significant difference between the treatment and control groups in the improvement of electroencephalogram abnormalities [RR = 0.87, 95% Cl = (0.70, 1.10), P = .25] and the incidence of adverse effects [RR = 0.75, 95% Cl = (0.48, 1.16), P = .19]. TSA analysis confirmed the clinical efficacy of the primary outcome measure.

Conclusion:

The combination of traditional Chinese and Western medicine may be an effective treatment option for RE, but further studies are needed to be assessed.

1. Introduction

Refractory epilepsy (RE) was defined by the International League Against Epilepsy in 2010 as a situation in which a patient’s seizure-free duration does not reach 3 times the longest pretreatment seizure interval or 1 year after rational selection and proper use of at least 2 well-tolerated antiepileptic drugs (AEDs), alone or in combination, depending on seizure type.^[1] It is estimated that approximately one-third of epilepsy patients worldwide are affected by RE.^[2,3] These patients face significant challenges in their medical treatment journey, which not only severely impacts their quality of life and psychological well-being but also imposes a heavy economic burden on their families and society.^[4,5] In recent years, research on the combination of traditional Chinese and Western medicine in the treatment of RE has received increasing attention.^[6] Traditional Chinese medicine (TCM) treatment for epilepsy includes herbal medicine, acupuncture, and tui na, and has a rich theoretical system and practical experience.^[7] Compared with Western medicine, TCM treatment is associated with fewer adverse effects and higher patient compliance.^[8–10] Some studies have shown that the combination of traditional Chinese and Western medicine can improve the treatment efficiency of RE, reduce patients’ symptoms, and improve the quality of life.^[11] However, existing studies have limitations in terms of sample size, methodological design, and outcome reporting, which make it difficult to provide strong evidence to support clinical practice.

This study aimed to systematically evaluate the efficacy and safety of combining traditional Chinese and Western medicine in the treatment of RE utilizing meta-analysis to provide more objective, comprehensive, and reliable evidence for clinical practice. Compared with previous studies, all the studies included in this study used TCM specialty therapies, including oral TCM and topical medications, and the dosage forms included decoctions, pills, creams, tablets, granules, and implants. This is of guiding significance for the combination of Chinese and Western medicine in the treatment of RE on the basis of continuous innovation of Chinese medicine treatment methods. Through the combined analysis of multiple randomized controlled trials (RCTs), this study is expected to identify the advantages and limitations of combined traditional Chinese and Western medicine treatment and provide references and insights for future research and clinical decision-making.

2. Materials and methods

2.1. Protocol and registration

This study was prospectively registered with PROSPERO (CRD42023445413) and conducted in accordance with the PRISMA guidelines. Ethics approval was not required for this meta-analysis as it utilizes publicly available, previously published, and anonymized data, ensuring no direct involvement of human subjects.

2.2. Search strategy

Seven databases were searched by computer as the main databases for this study: CNKI, Wanfang, VIP, PubMed, Web of Science, Cochrane Library, and Embase. All searches published from the inception of the databases to March 2023 were included according to predefined criteria. Search strategies were based on subject terms combined with free-text words.

2.3. Inclusion criteria

Literature meeting the criteria was included according to the PICOS principles: P (population): The study subjects were patients^[12] who met the diagnostic criteria for RE and were not limited to gender, age, disease duration, or geography; I (intervention): The treatment group received integrated traditional Chinese and Western medicine therapy, which included TCM treatments such as decoctions, acupuncture, or catgut implantation, while control group were the same conventional Western medicine treatments as the treatment group; C (control measures): The control group applied conventional administration of AEDs, and chose different drugs according to Western medical epilepsy treatment guidelines and individualized differences; O (outcome index): The primary outcomes were epilepsy frequency, clinical efficacy (total effective rate). The secondary outcomes were quality of life score, TCM symptom score, electroencephalogram (EEG) abnormality rates, epilepsy duration, and incidence of adverse reactions; S (study design): The study types were all RCTs of traditional Chinese and Western medicine in the treatment of RE. The language of the literature was Chinese and English.

2.4. Exclusion criteria

Studies with duplicate publications or duplicate data; irrelevant papers, reviews, systematic evaluations, meta-analysis, animal experiments, editorials, commentaries, opinions, short reports, etc. that are not relevant to this study; studies with non-RCTs or treatment group interventions that are not a combination of traditional Chinese and Western medicine; studies with missing data or too few cases to provide sufficient data; low quality of studies which cannot assess were excluded; studies without AEDs in the control group.

2.5. Data extraction

With reference to the inclusion and exclusion criteria set in advance, 2 independent researchers (JL and ZW) were arranged to screen the literature separately, using Endnote X9 software to de-weight the literature, and then by browsing through the title, abstract, and full text layer by layer, and the whole process was completed independently without communication between the 2 researchers (JL and ZW). After checking, if there were differences in the final results, a third researcher (YZ) was arranged to confer together to reach a consensus and make a decision. Excel software was applied to extract the data from the final included literature, and if there was disagreement, a third researcher was arranged to confer together to reach a consensus and make a decision. The main extracts included: first author and year of publication, random allocation method, sample size of subjects, age, combined traditional Chinese and Western medicine treatment regimen, treatment period, outcome index, incidence of adverse effects, and disease duration.

2.6. Risk of bias and GRADE assessment

The risk of bias evaluation tool (risk of bias) recommended by the Cochrane Collaboration was used to evaluate the quality of included studies in 7 areas, including random sequence generation, group concealment, blinding of study subjects and investigators, blinding of outcome measures, data integrity, selective reporting, and other biases. The Grading of Recommendations Assessment, Development, and Evaluation (GRADE) profiler version 3.6.1 was used to assess the quality of the evidence for each outcome. The certainty of evidence was assessed as high, moderate, low, or very low depending on the risk of bias, consistency, directness, precision, publication bias, large effect, and plausible confounding.

2.7. Statistical analysis

Meta-analysis was performed by RevMan 5.4.1 software. Relative risk was used as an effect indicator for dichotomous variables, and mean difference was used as an effect indicator for continuous variables, and 95% confidence interval (Cl) was set for both. The heterogeneity of the included studies was assessed based on the P value and I² statistic, and the appropriate model was selected based on the heterogeneity test results. If P > .10 and I² < 50%, the heterogeneity was small and the fixed-effects model was used for the combined analysis; if P < .10 and I² > 50%, the heterogeneity was large and the random-effects model was used for the combined analysis. If significant heterogeneity existed, sensitivity analysis or subgroup analysis was performed to account for the stability of the combined results and to search for possible sources of heterogeneity. Publication bias analysis was performed for outcome indicators for no fewer than 10 included studies. To account for type-I and type-II errors and predetermined number of patients was reached, trial sequential analysis (TSA) was performed via version V.0.0.5.10 Beta. The TSA was set with a power level of 90% and a tow-tailed α level of 0.05.

3. Results

3.1. Literature search results

The computer initially searched 547 literature based on the search formula, including CNKI (n = 31), Wanfang (n = 43), VIP (n = 11), PubMed (n = 418), Web of Science (n = 16), Cochrane Library (n = 0), and Embase (n = 28). We removed 62 duplicates; browsed the titles and abstracts for primary screening, and found 387 irrelevant papers, reviews, systematic evaluations, meta-analyses, animal experiments, editorials, reviews, opinions, short reports, etc; browsed the full text for re-screening, and found 77 non-RCT studies, treatment groups not combining traditional Chinese and Western medicine, non-AEDs for control groups, etc; and excluded 1 literature with missing data or too few cases, and finally included 20 papers that met the study conditions. The specific literature screening flow chart is shown in Figure 1.

Figure 1.

Flow chart of literature screening.

3.2. General characteristics

Of the 20 studies included,^[13–32] involving 1521 patients with RE, 760 in the treatment group and 761 in the control group. 10 studies^{[13,14,16,17,23,24,26,29–31]} mentioned the random data form method. Fourteen studies^{[13–16,19–21,24,26–29,31,32]} mentioned the course of the disease. AEDs were used in all treatment groups and all treatment groups were control groups with TCM treatments. All studies mentioned the treatment period. The general characteristics of the included studies are shown in Table 1.

Table 1.

General characteristics of included studies.

First author and year	Random allocation method	Number of cases		Age (yr)		Intervention		Treatment period (mo)	Outcome	Adverse reaction rate (%)		Course of disease (yr)
		T	C	T	C	T	C			T	C	T	C
Chen CP,[13] 2018	Random data form	30	30	42.92 ± 13.61	41.32 ± 12.58	Chaihu Shugan Decoction (200 mL, bid) + AEDs	AEDs (Lamotrigine 25 mg, qd; gradually increase 50 mg, bid + Sodium Valproate 200 mg, tid)	3	(1) (3) (6) (7) (9)	13.33	20	5.83 ± 2.78	4.85 ± 1.92
Dai GJ,[14] 2017	Random data form	45	45	32.35 ± 3.11	32.26 ± 3.07	Chaibei Zhixian Decoction (bid) + AEDs	AEDs (Sodium Valproate 30 mg/kg/d + Carbamazepine 20 mg/kg/d)	6	(1) (2) (4)	17.78	11.11	26.26 ± 3.38 (mo)	26.19 ± 3.28 (mo)
He QC,[15] 2016	N/A	60	60	42.12 ± 4.54	41.02 ± 6.92	Chaihu Shugan Decoction (bid) + AEDs	AEDs	2	(1)	N/A		5.30 ± 2.47	4.73 ± 1.35
Liu WY,[16] 2021	Random data form	30	30	49.22 ± 14.61	49.03 ± 14.43	Chaibei Zhixian Decoction (200 mL, bid) + AEDs	AEDs	3	(1) (3) (6) (7) (9)	13.33	16.67	5.63 ± 2.51	5.66 ± 2.77
Luo HM,[17] 2015	Random data form	66	66	42.5 ± 11.4	43.2 ± 12.2	Annao Pill (6 g, bid) + AEDs	AEDs (Carbamazepine + Sodium Valproate + Lamotrigine)	1	(1) (2) (11)	6.06	21.21	N/A
Nie LY,[18] 2015	N/A	28	29	40.14 ± 18.79		Chaibei Zhixian Decoction (bid) + AEDs	AEDs	3	(1) (4) (12)	0	0	N/A
Pan LK,[19] 2009	N/A	32	32	24 ± 5	23.5 ± 5	Shenpu Decoction (4 g/kg, bid or tid) + AEDs	AEDs (1–2 tablets, bid or tid)	3	(2)	N/A		10.0 ± 0.10	19.0 ± 0.10
Pan YY,[20] 2012	N/A	20	20	22.45 ± 7.83	20.80 ± 6.92	Kangdian Cream + AEDs	AEDs	6	(2) (4) (8)	N/A		11.24 ± 7.35	9.255 ± 5.97
Shang XZ,[21] 2019	N/A	50	50	34.93 ± 5.22	34.72 ± 5.71	Zhebeimu Shugan Zhixian Decoction (bid) + AEDs	AEDs (Carbamazepine 0.2 g, tid + Sodium Valproate 0.2 g, bid + Levetiracetam 0.5 g, tid)	4	(1) (2) (3) (5)	0	0	4.38 ± 1.34	4.57 ± 1.51
Shu J,[22] 2012	N/A	46	46	36.33 ± 8.4	34.06 ± 6.3	Chaibei Zhixian Decoction (bid) + AEDs	AEDs (Phenytoin Sodium + Sodium Valproate + Carbamazepine + Diazepam)	6	(1) (2) (12)	0	0	N/A
Sun YD,[23] 2017	Random data form	32	31	19–60		Shaoxing Zhixian tablet (1.6 g, bid) + AEDs	AEDs	3	(3) (10) (13)	9.38	16.13	N/A
Tian LJ,[24] 2011	Random data form	31	31	30.68 ± 1.94	29.71 ± 1.66	Yuxianling Granule (8 g, bid) + AEDs	AEDs (Sodium Valproate 30 mg/kg/d + Carbamazepine 20 mg/kg/d + Clinnarizine 25 mg, tid+γ-aminobutyric acid 0.25 g, tid)	3	(1) (2) (3) (5) (14) (19)	0	0	10.81 ± 1.28	10.91 ± 0.99
Wang B,[25] 2012	N/A	40	41	68.45		Xuefu Zhuyu Decoction (bid) + AEDs	AEDs (Carbamazepine + Lamotrigine)	1	(2)	N/A		N/A
Wang JC,[26] 2015	Random data form	40	40	64.2 ± 4.4	63.9 ± 4.6	Xuefu Zhuyu Decoction (250 mL, bid) + AEDs	AEDs (Sodium Valproate 500 mg, bid)	3	(1) (2)	N/A		2.4 ± 0.4 (mo)	2.7 ± 0.3 (mo)
Wang XH,[27] 2019	N/A	28	32	16–80	16–61	Chaibei Zhixian Granule (a bag, bid) + AEDs	AEDs	3	(1) (3) (13)	3.57	6.25	4–50	3–51
Wu CH,[28] 2018	N/A	30	30	46.1 ± 10.3	45.8 ± 10.4	Chaibei Zhixian Decoction (150–200 mL, bid) + AEDs	AEDs (Sodium Valproate 0.5 g, bid + Phenytoin Sodium 0.1 g, tid + Carbamazepine 0.1 g, bid)	3	(2) (4) (15)	0	0	4.5 ± 1.1	4.2 ± 1.3
Wu WJ,[29] 2017	Random data form	35	37	35.43 ± 10.858	35.97 ± 9.158	Kangdian Cream + AEDs	AEDs	10	(1) (2) (4) (5) (8)	0	0	8.209 ± 4.172	8.638 ± 4.232
Xiao HL,[30] 2022	Random data form	43	40	47.77 ± 3.67	45.93 ± 5.23	Dingxain Pill (6 g, bid) + AEDs	AEDs	3	(1) (5) (6) (8) (16) (17)	9.3	10	N/A
Xie HT,[31] 2019	Random data form	47	47	25.50 ± 6.00	26.00 ± 3.00	Qingxin Wendan Decoction (250 mL, bid) + AEDs	AEDs	6	(1) (3) (10) (11)	8.51	4.25	6.45 ± 4.00	6.75 ± 4.00
Zhu H,[32] 2019	N/A	27	24	26.07 ± 11.45	29.25 ± 10.84	Catgut Implantation + AEDs	AEDs	6	(1) (3) (5)	0	0	10.31 ± 6.66	15.18 ± 11.25

T = treatment group, C = control group, N/A = not mentioned, (1) = epilepsy frequency, (2) = clinical efficacy (total effective rate), (3) = quality of life score, (4) = TCM symptom score, (5) = EEG abnormality rates, (6) = epilepsy duration, (7) = SOD/MDA/P-gp/IGF-1 levels, (8) = serum cytokine IL-1β/IL-2/IL-6/TNF-α levels, (9) = video EEG, (10) = epilepsy severity, (11) = TCM classification efficacy, (12) = EEG improvement rate, (13) =TCM syndrome score, (14) = drug retention rate, (15) = Montreal Cognitive Assessment (MoCA) score, (16) = self-rating anxiety scale (SAS)/self-rating depression scale (SDS)/Pittsburgh sleep quality index (PSQI) score, (17) = epileptogenic index, (18) = epilepsy score total score.

AEDs = antiepileptic drugs, EEG = electroencephalogram, TCM = traditional Chinese medicine.

3.3. Results of risk of bias evaluation

Of the 20 included studies, all used randomized sequence generation methods, but only 10 studies^{[13,14,16,17,23,24,26,29–31]} specified the use of the random data form method, evaluating it as low risk. Four studies^[24,27–29] mentioned the allocation concealment of the randomization scheme but did not describe it in detail, evaluating it unclear risk. Due to the disease-specific nature of RE, study participants were often concerned about their treatment and could not readily be given a placebo, and the treatment was relatively clinically limited, making it inconvenient to blind, but the observational outcome indicators were relatively objective, and all studies were evaluated as having unclear risk. One study^[30] blinded the evaluator of the outcome indicator and evaluated it as low risk. Seven studies^{[15,16,19,23,25,31,32]} had unknown general information or cases falling off without reported reasons, had poor data completeness, and were evaluated as unclear risk. Three studies^[19,20,25] did not report safety, had the potential for selective reporting of positive results, and were evaluated as high risk. Eight studies^{[14–16,19,21,23,25,26]} made no mention of “there is no conflict of interest in this paper” or did not mention follow-up or there was a large difference in the baseline comparison of general information between the treatment and control groups or inconsistency in the pre and post data for the same study population, and the other risk of bias was evaluated as the risk of not being clear. The results of risk of bias evaluation of the included studies are shown in Figures 2 and 3.

Figure 2.

Quality assessment of included studies.

Figure 3.

Individual quality assessment of included studies.

3.4. Meta-analysis results

3.4.1. Epilepsy frequency

Fifteen studies^{[13–18,21,22,24,26,27,29–32]} mentioned epilepsy frequency and 10 studies^{[13–17,21,26,29–31]} recorded data in detail. Meta-analysis revealed high heterogeneity (P < .00001, I² = 98%), so a random-effects model was employed to combine the analyses. The results showed that the treatment group significantly reduced epilepsy frequency compared to the control group, with a statistically significant difference [MD = −2.37, 95% Cl = (−3.17, −1.57), P < .00001] (Fig. 4). The research results of He QC^[15] in the forest plot significantly deviated from the combined results, but none of the heterogeneity was found to be significantly decreased by the case-by-case exclusion study. Subgroup analysis was also performed separately according to the epilepsy frequency recording unit as well as the treatment period, and the results suggested that only the group with an epilepsy frequency recording unit of 3 times a month had a greater decrease in heterogeneity, while the rest of the groups still had higher heterogeneity (Figs. 5 and 6).

Figure 4.

Forest plot of epilepsy frequency.

Figure 5.

Subgroup analysis forest plot of epilepsy frequency recording units.

Figure 6.

Subgroup analysis forest plot of treatment cycles.

3.4.2. Clinical efficacy (total effective rate)

Clinical efficacy was mentioned in 11 studies.^{[14,17,19–22,24–26,28,29]} Meta-analysis suggested low heterogeneity (P = .20, I² = 26%), so a fixed-effects model was used to combine the analyses. The results showed that the enhanced clinical efficacy of the treatment group was better than that of the control group, with a statistically significant difference [RR = 1.27, 95% Cl = (1.19, 1.35), P < .00001] (Fig. 7).

Figure 7.

Forest plot of total efficiency rate.

3.4.3. Quality of life score

Eight studies^{[13,16,21,23,24,27,31,32]} referred to quality of life and 5 studies^{[13,16,21,24,31]} were based on improvement in scores after treatment. Meta-analysis suggested high heterogeneity (P < .00001, I² = 95%), so a random-effects model was used to combine the analyses. The results showed that the treatment group improved quality of life better than the control group, with a statistically significant difference [MD = 4.45, 95% Cl = (1.60, 7.30), P = .002] (Fig. 8). After a case-by-case exclusion could not significantly reduce heterogeneity, and because the pretreatment quality of life scores of the studies were already very different, it was difficult to find suitable possible influencing factors for subgroup analysis, and it was observed that the treatment courses of the studies were different, and some quality of life scores were even proportional to the length of treatment, and the longer the treatment time, the higher the scores, taking into account that the pretreatment quality of life scores as well as the treatment duration of treatment may be the factors contributing to the high heterogeneity.

Figure 8.

Forest plot of quality of life scores.

3.4.4. TCM symptom score

Five studies^{[14,18,20,28,29]} mentioned TCM clinical symptoms. Four studies^{[18,20,28,29]} recorded total TCM symptom scores, and meta-analysis suggested high heterogeneity (P < .00001, I² = 89%), so a random-effects model was used to combine the analyses. The results showed that the treatment group relieved clinical symptoms of epilepsy better than the control group, with a statistically significant difference [MD = −2.32, 95% Cl = (−4.13, −0.51), P = .01] (Fig. 9). Sensitivity analysis was conducted using a one by one exclusion method, and after removing 2 studies of Pan ^[20] and Wu,^[29] it was found that heterogeneity was significantly reduced (P = .8, I² = 0%). The treatment duration in these 2 studies was 6 months and 9 months, respectively, while the other 2 studies were both 3 months. Therefore, the treatment duration may be a source of heterogeneity.

Figure 9.

Forest plot of TCM symptom total integral. TCM = traditional Chinese medicine.

3.4.5. EEG abnormality rates

Five studies^{[21,22,24,29,30]} mentioned EEG abnormality rates. Meta-analysis suggested high heterogeneity (P < .00001, I² = 94%), so a random-effects model was used to combine the analyses. The results showed no statistically significant difference in the improvement of EEG abnormalities in the treatment group compared with the control group [RR = 0.87, 95% Cl = (0.70, 1.10), P = .25] (Fig. 10). Sensitivity analysis was conducted using a one by one exclusion method, and it was found that 2 studies of Shang ^[21] and Tian ^[24] were the main reasons affecting the results. After exclusion, heterogeneity was significantly reduced (P = .65, I² = 0%). It was further found that both studies had normal EEG before treatment, which was not the case in the other 3 studies, and that neither the treatment nor the control group converted the abnormal EEG back to normal after treatment, but both confirmed that the treatment group reduced the number of cases of severe EEG abnormalities to moderate or mild in a higher percentage of cases than the control group. Thus, the degree of EEG abnormality before treatment in patients with RE may be a source of heterogeneity.

Figure 10.

Forest plot of EEG abnormality rates. EEG = electroencephalogram.

3.4.6. Epilepsy duration

Three studies^[13,16,30] mentioned epilepsy duration. Meta-analysis suggested high heterogeneity (P < .00001, I² = 92%), so a random-effects model was used to combine the analyses. The results showed that the treatment group shortened the duration of epilepsy better than the control group, and the difference was statistically significant [MD = −1.20, 95% Cl = (−1.71, −0.68), P < .00001] (Fig. 11). It was observed that the 3 studies had different TCM syndrome of epilepsy, which may be one of the sources of heterogeneity.

Figure 11.

Forest plot of epilepsy duration.

3.4.7. Incidence of adverse reactions

Fifteen studies^{[13,14,16–18,21–24,27–32]} mentioned safety issues and 7 studies^{[13,14,16,17,23,30,31]} specified the number of cases of adverse reactions. Meta-analysis suggested low heterogeneity (P = .33, I² = 13%), so fixed-effects model analysis was used. The results showed no statistically significant difference in the incidence of adverse reactions between the treatment and control groups [RR = 0.75, 95% Cl = (0.48, 1.16), P = .19] (Fig. 12). The main adverse events in both groups were gastrointestinal reactions in 28 cases (T-12/C-16), abnormal liver function in 17 cases (T-5/C-12), dizziness and headache in 13 cases (T-5/C-8), rash in 7 cases (T-4/C-3), drowsiness in 4 cases (T-2/C-3), and irritability in 3 cases (T-3) (Table 2).

Figure 12.

Forest plot of incidence of adverse events.

Table 2.

Adverse envents.

Groups	Gastrointestinal reactions (N)	Abnornal liver function (N)	Dizziness and headache (N)	Rash (N)	Drowsiness (N)	Irritability (N)
Treatment groups	12	5	5	4	2	3
Control groups	16	12	8	3	3	0

N = number of cases.

3.4.8. Grade analyses for all outcomes

We conducted a GRADE assessment for all outcomes in this meta-analysis, with overall evidence quality ranging from moderate to very low. Among these, the outcomes of total clinical efficacy, TCM symptom score, and the number of adverse events were rated as moderate quality. Epilepsy frequency was rated as low quality. However, quality of life scores, EEG abnormalities, and epilepsy duration were defined as very low quality due to excessive heterogeneity and small sample sizes (Fig. 13).

Figure 13.

Grade analyses for all outcomes.

3.4.9. Publication bias

The analysis of epilepsy frequency and total effective rate in RE by combining traditional Chinese and Western medicine and drawing funnel plots revealed that the left and right funnel plots of epilepsy frequency were asymmetrical, indicating that there might be some publication bias (Fig. 14A); the funnel plot of total effective rate was basically symmetrical and the RCT study focused on the narrow part above the funnel, suggesting that there was no publication bias (Fig. 14B).

Figure 14.

Funnel plot of epilepsy frequency (A) and total efficiency rate (B).

3.4.10. TSA for total effective rate

We conducted a TSA of the primary outcome of “total effective rate,” and analyzed the cumulative data from the 11 included RCTs. As shown in Figure 15, the cumulative Z-curve exceeded the traditional significance boundary (Z = ±1.96), indicating that the effect size of this indicator was statistically significant. The cumulative evidence proved the effectiveness of integrated traditional Chinese and Western medicine in the treatment of RE and significantly reduced the random error caused by repeated measurements.

Figure 15.

Trial sequence analysis for total effective rate.

4. Discussion

RE is a difficult disease to cure and requires a combination of treatment options to achieve optimal results. In recent years, the efficacy of combining traditional Chinese and Western medicine in the treatment of RE has received widespread attention. Combination of traditional Chinese and Western medicine is gradually gaining attention and application by clinicians as an alternative treatment option.

The effectiveness of combined traditional Chinese and Western medicine in the treatment of RE is mainly reflected in adjusting the entire internal environment of the patient’s body, correcting the dysfunction of the organism, and enhancing the body’s self-healing and disease-resistance abilities. At the same time, treatment modalities such as AEDs and surgery^[33] used in Western medicine can reduce the severity and frequency of epilepsy through targeted effects, thus effectively improving the therapeutic effect. It is worth mentioning that Chinese medical treatments such as acupuncture, massage, and herbal decoction can also help improve the patient’s mental state, such as depression and anxiety, and promote physical health. In conclusion, combined traditional Chinese and Western medicine treatment has the potential of proving highly effective.

We found some contradictory findings in the RCT studies. There are several possible reasons for this. There are different types and characteristic clinical manifestations of RE. These multiple types with otherwise unclear and complex mechanisms also allow for random errors in treatment outcomes. The small number of participants in the study may have led to insufficient statistical significance of the results to draw consistent conclusions. Participants in a study have different characteristics, baseline status, or individual differences, which may cause them to respond differently to the treatment, which in turn affects the consistency of the results. There are differences in treatment protocols across studies, such as drug dosage, duration of treatment, and mode of follow-up, and these differences may lead to conflicting results. The results of published studies are usually positive or significant, while insignificant or negative results may not be published, which may lead to contradictory results. To address this issue, more studies are needed to validate and replicate the results with more comprehensive and specialized data analysis and comprehensive assessment.

In this study, based on the results of meta-analysis, we found that the combination of traditional Chinese and Western medicine can significantly reduce the frequency of epilepsy, relieve clinical symptoms, shorten the duration of epilepsy, enhance clinical efficacy, and improve the quality of life compared with the treatment of RE by Western medicine alone. However, most of the above meta results were highly heterogeneous among the studies, and in this meta-analysis, heterogeneity analyses were performed to explore the sources of heterogeneity. We also need more high-quality studies to confirm our conclusions. Thankfully, however, despite these differences in heterogeneous factors, they do not have a decisive biasing effect on the subject of our study. Whether it is the difference in age or the duration of the disease, etc, the final RCT study confirms that the combination of Chinese and Western medicine plays a significant role in the treatment of RE, and surely has a clinical significance in guiding the innovative treatment. The source of heterogeneity in the forest plot of epilepsy frequency is unclear, and the grouping of single factors explored cannot be reasonably explained, therefore, it may be the result of a combination of multiple factors such as quality of literature, recording unit, and treatment period, and the source of heterogeneity should be debated based on multiple considerations, but it is certain that the effect of combining traditional Chinese and Western medicine to reduce epilepsy frequency is definite. There was no significant difference in the incidence of adverse events between the treatment group and the control group, which may mean that the combination of traditional Chinese and Western medicine did not reduce the occurrence of adverse events to some extent, but of course, it may also be related to the insufficient number of RCT studies or the small number of cases, etc, which cannot accurately reflect the results. As to whether there are also differences in the specific adverse events occurring in the 2 groups and what are the influencing factors associated with these differences, it is not known, and therefore more clinical RCT studies and observations are needed to objectively evaluate safety. Based on comparisons of adverse events, it seems that the combination of Chinese and Western medicine has more side effects in terms of psychological irritability, but further future research is needed to determine whether this is clinically significant. Likewise, we should focus on the potential side effects of integrated traditional Chinese and Western medicine therapy. In addition, each secondary outcome indicator was mentioned in only 1 or 2 papers, and therefore no forest plot analysis was performed. Two studies^[13,16] indicated that integrated traditional Chinese and Western medicine treatment can effectively reduce the expression of P-glycoprotein, which may enhance the efficacy of AEDs by improving blood-brain barrier permeability. Additionally, epilepsy induced by traumatic brain injury raised curiosity in clinical neurology researchers. This meta-analysis included 2 studies^[25,26] focusing on this point, suggesting that traumatic brain injury may influence treatment efficacy. Further, patients often have a higher rate of EEG abnormalities detected during epilepsy, and there was no significant difference in the rate of post-treatment EEG abnormalities between the treatment and control groups in 5 studies,^{[21,24,29,30,32]} but this does not indicate that the combination of traditional Chinese and Western medicine does not better improve the EEG situation, as it is closely related to the duration of the disease, the disease situation, and the duration of treatment. Long-range video EEG plays an important role in the diagnosis of RE, and the results of 2 studies^[13,16] showed that the combination of traditional Chinese and Western medicine is effective in improving epileptiform discharges exactly. Because of the complex mechanism of EEG formation, the visual determination of EEG and video EEG requires in-depth research in principle and time. China has made progress in this field, the new version of the EEG reporting system of the China Association Against Epilepsy can provide a large amount of researchable EEG data information, which serves as a clinical guideline and is an important tool for future data sharing and big data research.^[34] Among the various forms of TCM interventions, the Chaibei Zhixian Decoction was the most frequently used. Although its efficacy in improving clinical symptoms and reducing epilepsy frequency has been confirmed, the mechanism of this decoction on EEGs remains unclear. In case of the extreme complexity of EEG changes, further in-depth research is needed to elucidate the effects of the Chaibei Zhixian Decoction on brain neurophysiology. There are also many new advances in the treatment of Western medicine, such as magnetic resonance-guided stereotactic laser ablation, pirampanel, cannabidiol, and soticlestat,^[35–39] but the combination with TCM has not been studied enough. In conclusion, the combination of traditional Chinese and Western medicine is effective in the treatment of RE, and no new serious adverse effects have been observed, which is worthy of clinical promotion and application. Of course, TCM combined with AEDs therapy requires adherence to the course of treatment, and the drugs should not be withdrawn at will or taken irregularly.

This study points the way to clinical work and future research. Since TCM still adopts a fixed diagnosis and treatment model and lacks the awareness and knowledge of evidence-based medicine, it has led to insufficient and less credible clinical research literature on TCM. In order to solve this problem, the recognized diagnostic and therapeutic evaluation standards of Western medicine can be borrowed to establish objective and reliable diagnostic standards for TCM symptoms and a safety evaluation system for TCMs and compound formulas, so as to reflect the combination of evidence-based medical thinking and TCM evidence-based treatment. Clinical information on RE is still insufficient or lacks veracity. Therefore, future efforts need to focus on conducting population-based incidence surveys within each country, including the associated risk factors and morbidity characteristics of RE. This will not only help to strengthen the theoretical innovation of combined Chinese and Western medicine treatments by further confirming the epidemiological profile of the disease, but also facilitate the government’s encouragement of epilepsy treatment facilities. According to the consultation needs of different RE patients, the future research direction can vigorously develop the unique advantages of external treatment methods of Chinese medicine. Using a variety of traditional treatment modalities such as acupuncture, moxibustion, buried threads, cupping, tuina, plasters, medicinal baths, etc., in combination with Western medicine, we can develop new treatments and products that are effective and easy to be accepted by RE patients.

However, the present meta-analysis also has some limitations. First, there were differences in the treatment regimens involved in the study, which prevented us from assessing whether the effects of the various treatment regimens were the same. These sources of heterogeneity may be the size of the sample size, different age stratification, differences in TCM treatment protocols, length of treatment cycles and duration of illness, and even differences in etiology. Second, the number of studies was small, and further large-scale studies are needed to confirm the effectiveness and safety of combining traditional Chinese and Western medicine in the treatment of RE that is reproducible. Third, the quality of the RCT literature included in this study was also low overall and needs further improvement. For example, the study by He QC et al,^[15] Liu WY,^[16] Nie LY et al^[18] did not provide details on the dosing of AEDs. Meanwhile, the study by Shang XZ,^[21] Shu J,^[22] Wang B^[25] did not provide the amount of oral medication per dose. In addition, for Wu CH’s^[28] study, Chen CP’s^[13] study, He QC’s^[15] study, etc, the doses of AEDs provided in these studies may not be clinically effective in the treatment of easily controllable epilepsy, let alone in the treatment of RE. Therefore, the authenticity and reliability of the data have yet to be thoroughly examined and explored. Finally, special attention should be paid to the fact that there are also some safety issues with TCM, such as acupuncture point application and acupuncture point burial techniques that require physicians to strictly master and avoid misuse in order to reduce the occurrence of adverse reactions.

5. Conclusion

The combination of traditional Chinese and Western medicine has significant efficacy in the treatment of RE, but further improvements and studies are needed to verify the safety of long-term use. A multicenter, large-scale study to confirm its advantages and to enhance the comparison and analysis of various treatment options will provide a more useful reference for clinical TCM and Western medicine treatment. It is recommended that the combination of traditional Chinese and Western medicine be more widely used in the treatment of RE in clinical practice and that the optimization of treatment protocols be enhanced.

Acknowledgments

We thank professor HC for his guidance on the methodology and discussion of this meta-analysis.

Author contributions

Data curation: Junjie Luo, Zhao Wang, Yichen Zhao.

Formal analysis: Junjie Luo, Zhao Wang, Yichen Zhao, Yunze Shi, Yongkang Lian, Fuxiong Zhang.

Methodology: Junjie Luo, Zhao Wang, Yunze Shi, Hong Chen.

Software: Junjie Luo, Zhao Wang, Yichen Zhao, Yunze Shi, Yongkang Lian.

Supervision: Hong Chen.

Validation: Hong Chen.

Visualization: Junjie Luo, Zhao Wang, Yichen Zhao, Yunze Shi, Yongkang Lian, Fuxiong Zhang.

Writing – original draft: Junjie Luo.

Writing – review & editing: Hong Chen.