Journal of Acupuncture Research 2025; 42:21-31
Published online January 23, 2025
https://doi.org/10.13045/jar.24.0033
© Korean Acupuncture & Moxibustion Medicine Society
Correspondence to : Min-seok Go
Department of Korean Acupuncture and Moxibustion Medicine, Gwangju Jaseng Hospital of Korean Medicine, 287 Uncheon-ro, Seo-gu, Gwangju 61946, Korea
E-mail: rhalstjr86@naver.com
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
Bell’s palsy, characterized by sudden-onset facial paralysis of unknown cause, is the most prevalent form of its kind, and electroacupuncture (EA) is one of the popular treatments in Korea. This review aimed to evaluate the effectiveness of EA in treating Bell’s palsy. Among the randomized controlled trials (RCTs) targeting patients with Bell’s palsy, studies comparing EA treatment with other therapies in PubMed, OVID, Cochrane RCTs, and China National Knowledge Infrastructure were searched up to April 2024. Data extraction and risk-of-bias assessment were performed independently. For the primary outcome, the House-Brackmann facial nerve grading system was used, and a significant difference was found in the EA group, although the results were inconsistent across all three studies. Secondary outcomes, including the total effective rate and recovery duration, generally favored EA, with some studies reporting significant differences compared with the control groups. The quality of evidence, evaluated by the grading of recommendations assessment and the development and evaluation approach, was moderate to low. This systematic review proposes that EA may be partially effective on Bell’s palsy and shortening the recovery time. However, further high-quality RCTs are needed to confirm these findings and address the limitations identified in the current evidence.
Keywords Bell palsy; Electroacupuncture; Systematic review
Bell’s palsy, characterized by sudden-onset facial paralysis of unknown cause, is the most prevalent form of its kind [1]. It primarily affects individuals aged 20–40 years, with an estimated annual prevalence of 20–25 cases per 100,000 people [2]. In Korea, traditional Korean medicine (TKM) remains a popular choice for treating Bell’s palsy, despite its recommendation level being categorized as D by the Acute Facial Nerve Paralysis Guideline Center of the Korean Otological Society, indicating limited empirical support compared with conventional medical approaches. Moreover, 54.4% of patients with Bell’s palsy in Korea opt for TKM treatments, with 41% choosing electroacupuncture (EA), making it the fourth most utilized TKM treatment after basic acupuncture, joint puncture, and infrared therapy. In contrast, 23.3% of patients seek conventional medical care. These figures underscore the ongoing preference for TKM therapies, including EA, among patients with Bell’s palsy in Korea, despite varying levels of clinical endorsement in established guidelines [3].
This review aimed to assess the clinical efficacy of EA in treating Bell’s palsy by synthesizing available evidence from randomized controlled trials (RCTs). Given the rising popularity of EA in TKM, despite limited empirical support, this study seeks to determine whether EA can significantly improve recovery outcomes in patients with Bell’s palsy compared with conventional treatments such as medication therapy, acupuncture, and physical therapy. By analyzing the results of selected RCTs, this review will provide insights into the overall effectiveness, recovery duration, and clinical relevance of EA. In addition, we will highlight gaps in the current evidence, suggest areas for further investigation, and propose recommendations for future clinical trials on EA for Bell’s palsy.
This systematic review of the effectiveness of EA in Bell’s palsy was conducted in compliance with the guidelines of the Preferred Reporting Items for Systematic Reviews and Meta-Analysis 2020. The study protocol was registered on The International Prospective Register of Systematic Reviews (registration no. CRD42024573728).
The search included both full texts and abstracts of articles registered or published in databases up to April 2024. The search was conducted in MEDLINE (via PubMed), Embase (via OVID), Cochrane Register of Controlled Trials, and China National Knowledge Infrastructure (CNKI). Within each search engine, three researchers independently screened records for inclusion, and researchers were blinded to each other’s decision. Disagreements were resolved by majority votes using an Excel spreadsheet (Microsoft). Search strategies for an online search are listed in Table 1.
Table 1 . Search strategies for an online search
PubMed | #1. Search: (Bell palsy [Mesh Terms] OR facial palsy [All Fields] OR facial paralysis [All Fields] OR Bell’s palsy) |
#2. Search: (Electroacupuncture [Mesh terms] OR Electric acupuncture [All fields] OR Electrical acupuncture) | |
#3. Search: (Randomized controlled trial [All Fields] OR Randomized clinical trial [All Fields]) | |
#4. #1 AND #2 AND #3 | |
OVID | (AllFields:(‘Bell’s palsy’ OR ‘facial palsy’ OR ‘facial paralysis’) AND (AllFields:(‘electroacupuncture’ OR ‘electric acupuncture’ OR ‘electrical acupuncture’)) AND (AllFields:(‘randomized controlled trial’ OR ‘randomized clinical trial’))) |
Cochrane Library | #1. ((Bell’s palsy OR facial palsy OR facial paralysis OR Bell palsy)):ti,ab,kw |
#2. ((electroacupuncture OR electric acupuncture OR electrical acupuncture)):ti,ab,kw | |
#3. ((randomized controlled trial OR randomized clinical trial)):ti,ab,kw | |
#4. #1 AND #2 AND #3 | |
CNKI | Used professional search and checked cross-language search |
TKA = (‘Bell palsy’ + ‘facial palsy’ + ‘facial paralysis’) AND TKA = (electroacupuncture + electric acupuncture + electrical acupuncture) AND TKA = (randomized controlled trial + randomized clinical trial) |
CNKI, China National Knowledge Infrastructure.
Inclusion criteria: 1) Study design: RCTs, 2) Population: patients with Bell’s palsy (including idiopathic facial paralysis and facial paralysis), 3) Intervention: including EA, 4) Comparison: all treatments, including oral medication, acupuncture, or physical therapy. Exclusion criteria: Excluding studies with EA in both groups.
A researcher independently extracted the data, and the other two researchers checked the extracted data. Data were extracted based on the target disease, intervention, duration and number of interventions, assessment scale, control group setting, number of participants, treatment efficacy, and year of publication. Data were recorded in an Excel spreadsheet, and disagreements were resolved by majority votes by an Excel spreadsheet.
The risk-of-bias (RoB) was assessed using the RoB tool for randomized trials (RoB 2) [4] developed by the Cochrane group. Studies were assessed as “low,” “some concerns,” or “high” based on six domains: biases arising from the randomization process, deviations from intended interventions, missing outcome data, measurement of the outcomes, and selection of the reported result and overall bias. Each domain has signaling questions, and the questions are answered “yes,” “probably yes,” “probably no,” “no,” and “no information.” Depending on the response, the RoB judgment is primarily evaluated through the algorithm results. The authors followed the algorithm results.
The primary outcome measure was the House-Brackmann facial nerve grading system (HBGS). Secondary outcome measures were the total facial nerve index (TFNI) score, Sunnybrook facial grading system (SFG) score, total effective rate, recovery duration, effective rate, cured rate, World Health Organization Quality of Life-BREF (WHOQOL-BREF), and facial disability index (FDI) including social functioning and well-being (FDIs) and physical function (FDIp).
The grading of recommendations assessment and the development and evaluation (GRADE) method in the GRADE Pro tool was used to rate the quality of evidence. The quality of evidence from the studies began at a high level; however, the scores gradually decreased to “moderate,” “low,” or “very low” based on the limitations in design and implementation (RoB), inconsistency (unexplained heterogeneity), indirectness, imprecision (spare data), and publication bias.
From the 141 articles searched, 30 were selected after excluding 17 duplicates and articles that did not target EA or Bell’s palsy based on the title and abstract. Then, through a secondary full-text review, 11 articles [5-15] were selected after excluding protocol, case reports, and articles that used EA in both groups. The above contents are presented in Fig. 1. The 11 studies were analyzed and summarized in Table 2.
Table 2 . Summary of randomized controlled trials of EA in Bell’s palsy
Study | Sample size | Treatment group (sample size) | Control group (sample size) | Retaining time × no./duration | Outcome measurement | Results |
---|---|---|---|---|---|---|
Li et al. [5] | 60 | Treatment group (30): EA | Control group (30): TES | 30 min × 72/24 wk | 1. TFNI score 2. SFG score | 1. T > C (p < 0.05) 2. T > C (p < 0.05) (without 6 week) |
Li et al. [6] | 900 | Group C (180): EA (after 7 days) Staged acupuncture | - Group A (180): Staged acupuncture - Group B (180): Staged acupuncture + moxibustion - Group D (180): Staged acupuncture + line puncture on the muscle - Group E (180): acupuncture | 30 min × 20/ 4 wk | 1. Total effective rate | 1. T = C (p > 0.05) |
Shen et al. [7] | 279 | - Group A (74): Acupuncture (3 days after onset) EA (after 5 days) Prednisone (based on the date of onset), 30 mg (days 3–5), 15 mg(days 6–8), 10 mg (days 9–13), 5 mg (days 14–18) - Group B (70): Acupuncture (7 days after onset) EA (after 5 days) + control group - Group C (74): Acupuncture (10 days after onset) EA (after 5 days) Prednisone (based on the date of onset), 30 mg (days 3–5), 15 mg (days 6–8), 10 mg (days 9–13), 5 mg (days 14–18) | Control group (61): Prednisone (based on the date of onset), 30 mg (days 3–5), 15 mg (days 6–8), 10 mg (days 9–13), 5 mg (days 14–18) | 20 min × 25/25 d | 1. Total effective rate 2. Recovery duration | 1. T > C (p < 0.05) 2. T < C (p < 0.05) |
Liu et al. [8] | 86 | Treatment group (45): Staged acupuncture EA Chinese medicine modified Qianzheng powder (twice a day/50 g) | Control group (41): Staged acupuncture Chinese medicine modified Qianzheng powder (twice a day/50 g) | 20–30 min × 42/60 d | 1. Effective rate | 1. T > C (p < 0.05) |
Hong et al. [9] | 288 | Group C (96): EA (after 7 days) Staged acupuncture | - Group A (96): Staged acupuncture - Group B (96): Staged acupuncture, moxibustion | (not reported) × 20/4 wk | 1. FDI 2. WHOQOL-BREF and facial nerve paralysis score | 1. T = C (p > 0.05) 2. T = C (p > 0.05) |
An et al. [10] | 80 | Treatment group (40): EA Prednisone (based on the date of register), 30 mg (day 1), 15 mg (days 2–7), 5 mg (days 8–10), neutrophis-vitamin B1 | Control group (40): Prednisone (based on the date of register), 30 mg (day 1), 15 mg (days 2–7), 5 mg (days 8–10), neutrophis-vitamin B1 | 30 min × 10/10 d | 1. Total effective rate 2. Recovery duration | 1. T = C (p > 0.05) 2. T < C (p < 0.05) |
Yang et al. [11] | 197 | Control group (72): EA | Observed group (125): Acupoint injection (prednisolone 50 mg, vitamin B12 0.5 mg/mixed 3 mL inj. to four acupoint) Acupuncture Warming needle therapy | - Treatment group: 20 min × 30/ 6 wk - Control group: 30 min × 30/ 6 wk | 1. Cured rate 2. Total effective rate | 1. T < C (p < 0.01; only first evaluation) 2. T < C (p < 0.05) |
Chen et al. [12] | 304 | Group C (76): Staged acupuncture EA (after 7 days) | - Group A (76): Staged acupuncture - Group B (76): Staged acupuncture + moxibustion - Group D (76): Muscle-region alignment needling - Group E (76): Nonstaged acupuncture | 20–30 min × 70/70 d | 1. HBGS 2. FDI 3. WHOQOL-BREF 4. Total effective rate | 1. T > C (p < 0.01) 2. T = C (p > 0.05) 3. T = C (p > 0.05) 4. T = C (p > 0.05) |
Chen et al. [13] | 60 | Treatment group (30): Staged acupuncture Moxibustion EA (after 8 days) | Control group (30): Prednisone acetate tablets (until day 28) Methylcobalamin tablets (until day 28) | 30 min × 28/ 4 wk | 1. HBGS 2. SFG score 3-1. FDIs 3-2. FDIp 4. Total effective rate | 1. T = C (p > 0.05) 2. T > C (p < 0.05) 3-1. T > C (p < 0.05) 3-2. T < C (p < 0.05) 4. T > C (p < 0.05) |
Liu et al. [14] | 131 | EA group (45): EA Staged acupuncture | - Acupuncture group (44): Staged acupuncture - Medication and acupuncture group (42): Staged acupuncture Prednisone (oral, based on the date of onset 20 mg until day 3, 5 mg until day 7), acyclover (0.6 mg daily, oral), vitamin B1 (IM, 100 mg, daily after 7 days) and vitamin B12 (IM, 100 μg, daily, after 7 days) | 30 min × 20/ 4 wk | 1. Cured rate 2. Total effective rate | 1. T > C (p < 0.01) 2. T = C (p > 0.05) |
Xu [15] | 130 | - Group A (26): EA (sparse-dense wave > 2 Hz) + acupuncture - Group B (26): EA (sparse-dense wave < 2 Hz) + acupuncture - Group C (26): EA (continuous wave > 2 Hz) + acupuncture - Group D (26): EA (continuous wave < 2 Hz) + acupuncture | Control group (26): Acupuncture | 30 min × 24/ 4 wk | 1. HBGS 2. Total effective rate 3. Recovery duration | 1. T = C (p > 0.05) 2. T = C (p > 0.05) 3. T < C (p < 0.05) |
EA, electroacupuncture; TES, transcutaneous electrical stimulation; TFNI, total facial nerve index; SFG, Sunnybrook facial grading system; FDI, facial disability index; WHOQOL-BREF, World Health Organization Quality of Life-BREF; inj., injection; HBGS, House-Brackmann facial nerve grading system; FDIs, FDI including social functioning and well-being; FDIp, FID including physical function.
In seven studies [5,7,8,10,13-15], the treatment group showed beneficial results before and after evaluation compared with the control group. Six studies [6,8,9,12, 14,15] used the same acupuncture treatment for the control and treatment groups, and five of the studies [5,7,10,11,13] varied the acupuncture treatment by treatment period. Six studies [7,8,10,11,13,14] used medications for the control group inwhich four [7,10,13,14] used prednisone. The treatment duration was 4 weeks in five studies [6,9,13-15], with 72 sessions over 24 weeks in the longest study [5]. In all but one study, which did not specify the needle retention time, the needles were retained for 20–30 minutes. Excluding one study that did not specify the acupoints, Taiyang (extra points-head and neck 5) was the most-used acupoint for EA in eight studies, and gallbladder meridian 14 was used for EA in seven studies, followed by stomach meridia (ST) 4, ST7, etc. (Table 3). All studies were published in China and were written in Chinese, except for two studies [5,6] written in English. Six studies [6,8,9,12-14] used staged acupuncture, and acupuncture treatments varied by treatment period: acute phase, 1–7 days; resting phase, 8–20 days; and recovery period, 21–70 days [16]. Only one study [13] had different periods: acute phase, 1–7 days; subacute phase, 8–14 days; and recovery period, 15–28 days.
Table 3 . Frequency of acupoints used in the studies
Frequency (Hz) | Acupoints |
---|---|
8 | EX-HN5 |
7 | GB14 |
6 | ST4 |
4 | ST7 |
Ex-HN, extra points-head and neck; GB, gallbladder meridian; ST, stomach meridian.
Three studies [12,13,15] used HBGS as the outcome measure. Chen et al. [12] did not find a significant difference between the control and treatment groups (p > 0.05). However, Chen et al. [13] stated that before treatment, HBGS was not significant in both groups (p > 0.05); after treatment, a significant difference was found in HBGS in the treatment group compared with the control group (p < 0.01). In the study by Xu [15], all five groups had significantly decreased HBGS before treatment and at weeks 2 and 4 (p = 0.000). Only the comparison between groups A (EA with sparse-dense wave > 2 Hz) and E (acupuncture; p = 0.031) showed a significant difference. Comparisons between groups B (EA with sparse-dense wave < 2 Hz) and E (p = 0.827), groups C (EA with continuous wave > 2 Hz) and E (p = 1.000), and groups D (EA with continuous wave < 2 Hz) and E (p = 1.000) did not show significant differences. The GRADE analysis indicated that the evidence for HBGS was of low quality. The RoB and inconsistency of the studies downgraded the quality of evidence.
Of the eight studies [6,7,10-15] that used the total effective rate as an endpoint, only two reported significantly higher results in the treatment group than in the control group. Li et al. [6] set the criteria of the total effective rate as HBGS, FDI scale, and degree of facial nerve paralysis scale. Li et al. [6] also stated that in the comparisons of the total effective rates of the five groups, no significant differences were noted: group A (staged acupuncture), 87.0; group B (staged acupuncture with moxibustion), 87.6; group C (staged acupuncture with EA), 86.0; group D (staged acupuncture with line puncture on muscles), 85.4; group E (acupuncture), 84.4 (all comparisons p > 0.05). Shen et al. [7] set the criteria of the total effective rate as HBGS, eyes closed, bilateral forehead symmetry, bilateral nasolabial folds, cheek symmetry, and exposed teeth. The treatment groups showed significantly higher results than the control group: group A (EA and drug with acupuncture 3 days after onset), 98.6; group B (acupuncture and EA and drug with acupuncture 7 days after onset), 95.7; group C (acupuncture and EA and drug with acupuncture 10 days after onset), 94.6; and group D (drug group), 72.1 (groups A, B, and C compared with group D, p < 0.05). An et al. [10] set the criteria of the total effective rate based on the HBGS. The difference between the treatment and control groups was not significant (treatment group, 100.00; control group, 100.00; p = 1.00). Yang et al. [11] set the criteria as facial asymmetry and movement. In the study by Yang et al. [11], the difference between the control and treatment groups was significantly high (control group, 100.0; treatment group, 86.1; p < 0.01). Chen et al. [12] set the criteria of the total effective rate based on the HBGS and FDI. In their study, the difference between the treatment and control groups was not significant: group A (staged acupuncture), 92.3; group B (staged acupuncture with moxibustion), 85.9; group C (staged acupuncture with EA), 88.0; group D (muscle-region alignment needling), 83.1; group E (nonstaged acupuncture group), 84.3 (all comparisons, p > 0.05). Chen et al. [13] set the criteria of the total effective rate as the HBGS. Chen et al. [13] stated that the total effective rate of the treatment group was significantly higher than that of the control group (treatment group, 93.3; control group, 83.3; p < 0.05). Liu et al. [14] set the criteria of the total effective rate as the HBGS. Liu et al. [14] stated that the EA group had significantly higher total effective rates than the other groups (EA group, 93.4; acupuncture group, 76.6; medication and acupuncture group, 78.6; comparison of EA group, p < 0.05). Xu [15] set the criteria of the total effective rate as the HBGS. Xu [15] evaluated the total effective rate by weeks 2 (w2) and 4 (w4) of treatment. In this study, all group comparisons were not significant (w2: group A, 88.0; group B, 76.9; group C, 69.2; group D, 61.9; group E, 68.0; p = 0.123; w4: group A, 100.0; group B, 96.1; group C, 96.1; group D, 100.0; group E, 92.0; p = 0.307). The GRADE analysis indicated that the evidence for the total effective rate was of low quality. The RoB and inconsistency of the studies downgraded the quality of evidence.
All three studies [7,10,15] that used recovery duration as an endpoint reported that recovery took significantly less time in the treatment group than in the control group. In the study by Shen et al. [7], the recovery duration significantly increased gradually in the order of groups A, B, and C and control group (group A, 31.90 ± 8.15; group B, 35.58 ± 9.10; group C, 39.91 ± 11.49; control group, 47.21 ± 12.04; each comparison p < 0.05). In the study by An et al. [10], the treatment group had a significantly longer recovery duration than the control group (treatment group, 28.16 ± 8.62; control group, 29.35 ± 6.26; p = 0.031). In the study Xu [15], group A had a significantly shorter recovery duration than group E. However, the comparisons between groups B and E, groups C and E, and groups D and E did not reveal significant differences (group A, 19.00 ± 4.75; group B, 21.00 ± 0.19; group C, 23.33 ± 1.41; group D, 23.33 ± 1.41; group E, 23.30 ± 1.49; groups A vs. E, p = 0.010; groups B vs. E, p = 0.164; groups C vs. E, p = 0.663; groups D vs. E, p = 0.985). The GRADE analysis indicated that the evidence for the recovery duration was moderate. The RoB of the studies downgraded the quality of evidence.
Two studies [5,13] that used the SFG score as an endpoint reported significant results in the treatment group compared with the control group. Li et al. [5] evaluated the SFG score by weeks 6 (6w), 12 (12w), and 24 (24w) of treatment. At week 6, no significant difference was found between the treatment and control groups (treatment group, 63.20; control group, 59.57; p = 0.74). At weeks 12 and 24, a significant increase was noted in the treatment group compared with the control group (12w: treatment group, 61.10; control group, 54.30; p = 0.005; 24w: treatment group, 60.50; control group, 52.30; p = 0.001). Chen et al. [13] reported that the treatment group showed significantly higher SFG scores than the control group (treatment group, 87.50; control group, 80.00; p < 0.05). The GRADE analysis indicated that the evidence for the SFG score was of moderate quality. The RoB of the studies downgraded the quality of evidence.
Of the three studies [9,12,13] that used FDI as an endpoint, two [9,12] found no significant difference between the treatment and control groups, whereas one [13] reported significant differences in FDIs and FDIp between the treatment and control groups. Hong et al. [9] stated no significant difference among groups A (staged acupuncture), B (staged acupuncture with moxibustion), and C (EA with staged acupuncture). The FDIp values were as follows: group A, 23.14 ± 2.58; group B, 23.02 ± 2.68; group C, 22.83 ± 2.71 and FDIs values were: group A, 9.01 ± 3.64; group B, 9.42 ± 3.47; group C, 9.28 ± 3.38, with all comparisons showing p > 0.05. Chen et al. [12] did not find a significant difference among the five groups (FDIp: group A, 23.08 ± 2.65; group B, 22.55 ± 2.73; group C, 22.36 ± 3.05; group D, 21.99 ± 3.54; group E, 22.88 ± 2.74; FDIs: group A, 8.97 ± 3.70; group B, 9.50 ± 3.53; group C, 9.20 ± 3.58; group D, 9.21 ± 3.69; group E, 9.04 ± 3.46; all comparisons of each group, p > 0.05). Chen et al. [13] reported significantly higher FDIp in the treatment group; however, the treatment group showed significantly lower FDIp than the control group (FDIp: treatment group, 25.00; control group, 24.00; FDIs: treatment group, 10.00; control group, 11.00; both FDIp and FDIs compared with the control group, p < 0.05). The GRADE analysis indicated that the evidence for FDI was of moderate quality. The RoB of the studies downgraded the quality of evidence.
Two studies [11,14] used the cured rate as an endpoint, and Liu et al. [14] reported significantly higher results in the treatment group than in the control group when there was an abnormal lesion in the upper part of the chorda tympany (EA group, 85.2; acupuncture group, 54.2; medication and acupuncture group, 48.0; p < 0.01). However, no significant difference was found between the treatment and control groups when there was an abnormal lesion on the lower part of the chorda tympany (EA group, 66.7; acupuncture group, 50.0; medication and acupuncture group, 70.6; p > 0.05). Yang et al. [11] reported significantly higher results in the control group than in the treatment group in the first evaluation (three evaluations were conducted over 2 weeks; first, second, and third evaluations of the treatment group, 22.9, 42.8, and 34.3, respectively; first, second, and third evaluations in the control group, 42.1, 44.2, and 13.7, respectively; only first evaluation, p < 0.05; other p-values were not reported). The GRADE analysis indicated that the evidence for the cured rate was of low quality. The RoB and inconsistency of the studies downgraded the quality of evidence. Two studies [9,12] used the WHOQOL-BREF, and they did not report significant differences. One study [8] used the effective rate as an outcome measurement, which was significant.
Cochrane RoB 2 was used to assess the 11 studies, and the results are summarized in the graphs shown in Fig. 2.
In D1, all studies [5-15] were assessed as “low.” Randomization was performed in all studies.
Five studies [5,7,10,13,15] were assessed as “low,” and the remaining six studies [6,8,9,11,12,14] were assessed to have “some concerns” because they did not clearly define the intentional intervention.
In D3, all studies [5-15] were assessed as “low.” In all studies, the baseline characteristics of the participants were randomized to each study group. The authors cannot find any evidence about the result being biased by missing outcome data.
In D4, eight studies [6,8,10-15] were assessed as “high,” and three studies [5,7,9] were assessed as “low.” The studies assessed as “low” clearly separated the intervention providers and the data assessors through blinding. The papers assessed as “high” did not clearly separate the intervention providers from the data assessors.
In D5, all studies [5-15] were assessed as having “some concerns.” Researchers cannot find a selective response about an incomplete outcome and cannot find a protocol or statistical analysis plan.
This systematic review of 11 RCTs (selected from 119 articles obtained from PubMed, OVID, Cochrane Register of Controlled Trials, and CNKI) compared the effectiveness of EA with those of traditional acupuncture and medications. Although several systematic reviews and meta-analyses of acupuncture for Bell’s palsy have been published, no systematic reviews focused on EA. Thus, this study may provide evidence of whether EA leads to a shorter time to recovery and is more effective than acupuncture or medications.
Studies utilizing the SFG and TFNI generally reported favorable outcomes for the treatment group, suggesting enhanced facial nerve recovery. However, studies using the HBGS and the total effective rate showed mixed results, with some reporting no significant differences between the treatment and control groups. This disparity may be attributed to the differences in assessment tools and their sensitivity to changes in facial nerve function.
The treatment duration and specific protocols varied widely among studies, from 4 to 24 weeks. This variation likely contributed to the mixed results. Studies with longer treatment durations tended to report more significant improvements, highlighting the potential need for prolonged EA for optimal outcomes.
In seven studies, the treatment group showed a significant improvement compared with the control group. However, HBGS, which was set as the primary outcome, was not significant in two out of three studies.
One study [6] mentioned one adverse event of fainting during acupuncture treatment with symptoms of dizziness, vomiting, and sweating and noted that the symptoms were relieved after the acupuncture needles were immediately withdrawn, followed by 10 minutes of relaxation with warm water. The remaining studies either stated that no adverse events occurred or did not mention adverse events.
A meta-analysis could not be performed in this study because of the following reasons. First, lack of consistency between studies: The results of the three studies using the HBGS are inconsistent, making it difficult to meet the assumption of homogeneity required for the meta-analysis. Second, differences in sample sizes and study design: The studies had varied sample sizes and treatment protocols, making it challenging to draw a unified conclusion in a meta-analysis. Third, quality issues in the results: Many studies were assessed to have a high RoB, which reduces the reliability of the outcomes.
This study has a few limitations. First, all the 11 studies selected were published in China, which was a somewhat significant deviation from the initial intention to include studies from different East Asian countries, including Korea, China, and Japan. This possibly limits the generalizability of the findings to other populations, and the lack of studies from other regions or unpublished data could skew the overall findings. Second, the number of studies was small, making it difficult to conduct a meta-analysis. Some outcomes showed significant differences, whereas others did not, and the studies had varied sample sizes, with some having relatively small populations. This could affect the precision of the estimates, particularly when significant results were found in smaller studies. Third, many studies lacked an experimental design, did not mention adverse events, and had no rigorous randomization, blinding, or allocation concealment. Eight studies had high risk because of outcome measurement issues, which might lower the quality of evidence. Fourth, the overall RoB in studies is quite high.
Future research should focus on large-scale, multicenter RCTs with standardized protocols to better evaluate the efficacy of EA. Studies should also explore the underlying mechanisms of EA in facial nerve recovery and identify subgroups of patients who may benefit the most from this treatment.
Several conclusions were drawn from these 11 RCTs of EA for Bell’s palsy. First, whether EA is a more effective treatment of Bell’s palsy than other treatments is not yet established; however, some studies have shown that it is more effective than other treatments. Second, the EX-HN5 acupoint was the most used for EA for Bell’s palsy. Third, further research is required to confirm the effectiveness of EA for Bell’s palsy.
The quality of evidence regarding the efficacy of EA for treating Bell’s palsy is moderate to low. Although some evidence suggests that EA might offer benefits in terms of recovery time and effectiveness compared with conventional treatments, significant limitations, such as RoB, inconsistency in results, and potential publication bias, prevent a strong recommendation. Future well-designed RCTs with more diverse populations and robust methodologies are necessary to confirm these findings and determine the optimal treatment protocols for Bell’s palsy.
This review highlights the need for a cautious interpretation of the current evidence and suggests that while EA may be effective, it should be considered part of a broader treatment strategy pending further research.
Conceptualization: MG. Data curation: MG, TK, SL. Formal analysis: MG, SB, HS. Investigation: MG. Methodology: MG, HS, JO, CL. Project administration: MG. Supervision: SY. Writing – original draft: MG.
The authors have no conflicts of interest to declare.
None.
This research did not involve any human or animal experiments. The study protocol was registered on The International Prospective Register of Systematic Reviews (registration no. CRD42024573728).
Journal of Acupuncture Research 2025; 42(): 21-31
Published online January 23, 2025 https://doi.org/10.13045/jar.24.0033
Copyright © Korean Acupuncture & Moxibustion Medicine Society.
Min-seok Go1 , Seong-hyun Lee1
, Ji-eun Oh1
, Seung-chul Yeom1
, Tae-ha Kwon2
, Sae-young Bong3
, Cha-young Lee4
, Hye-jin Shin5
1Department of Korean Acupuncture and Moxibustion Medicine, Gwangju Jaseng Hospital of Korean Medicine, Gwangju, Korea
2Department of Korean Internal Medicine, Gwangju Jaseng Hospital of Korean Medicine, Gwangju, Korea
3Department of Korean Rehabilitation Medicine, Gwangju Jaseng Hospital of Korean Medicine, Gwangju, Korea
4Department of Korean Medicine Neuropsychiatry, Gwangju Jaseng Hospital of Korean Medicine, Gwangju, Korea
5Department of Korean Acupuncture and Moxibustion Medicine, Teunteun Hospital of Korean Medicine, Hwaseong, Korea
Correspondence to:Min-seok Go
Department of Korean Acupuncture and Moxibustion Medicine, Gwangju Jaseng Hospital of Korean Medicine, 287 Uncheon-ro, Seo-gu, Gwangju 61946, Korea
E-mail: rhalstjr86@naver.com
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
Bell’s palsy, characterized by sudden-onset facial paralysis of unknown cause, is the most prevalent form of its kind, and electroacupuncture (EA) is one of the popular treatments in Korea. This review aimed to evaluate the effectiveness of EA in treating Bell’s palsy. Among the randomized controlled trials (RCTs) targeting patients with Bell’s palsy, studies comparing EA treatment with other therapies in PubMed, OVID, Cochrane RCTs, and China National Knowledge Infrastructure were searched up to April 2024. Data extraction and risk-of-bias assessment were performed independently. For the primary outcome, the House-Brackmann facial nerve grading system was used, and a significant difference was found in the EA group, although the results were inconsistent across all three studies. Secondary outcomes, including the total effective rate and recovery duration, generally favored EA, with some studies reporting significant differences compared with the control groups. The quality of evidence, evaluated by the grading of recommendations assessment and the development and evaluation approach, was moderate to low. This systematic review proposes that EA may be partially effective on Bell’s palsy and shortening the recovery time. However, further high-quality RCTs are needed to confirm these findings and address the limitations identified in the current evidence.
Keywords: Bell palsy, Electroacupuncture, Systematic review
Bell’s palsy, characterized by sudden-onset facial paralysis of unknown cause, is the most prevalent form of its kind [1]. It primarily affects individuals aged 20–40 years, with an estimated annual prevalence of 20–25 cases per 100,000 people [2]. In Korea, traditional Korean medicine (TKM) remains a popular choice for treating Bell’s palsy, despite its recommendation level being categorized as D by the Acute Facial Nerve Paralysis Guideline Center of the Korean Otological Society, indicating limited empirical support compared with conventional medical approaches. Moreover, 54.4% of patients with Bell’s palsy in Korea opt for TKM treatments, with 41% choosing electroacupuncture (EA), making it the fourth most utilized TKM treatment after basic acupuncture, joint puncture, and infrared therapy. In contrast, 23.3% of patients seek conventional medical care. These figures underscore the ongoing preference for TKM therapies, including EA, among patients with Bell’s palsy in Korea, despite varying levels of clinical endorsement in established guidelines [3].
This review aimed to assess the clinical efficacy of EA in treating Bell’s palsy by synthesizing available evidence from randomized controlled trials (RCTs). Given the rising popularity of EA in TKM, despite limited empirical support, this study seeks to determine whether EA can significantly improve recovery outcomes in patients with Bell’s palsy compared with conventional treatments such as medication therapy, acupuncture, and physical therapy. By analyzing the results of selected RCTs, this review will provide insights into the overall effectiveness, recovery duration, and clinical relevance of EA. In addition, we will highlight gaps in the current evidence, suggest areas for further investigation, and propose recommendations for future clinical trials on EA for Bell’s palsy.
This systematic review of the effectiveness of EA in Bell’s palsy was conducted in compliance with the guidelines of the Preferred Reporting Items for Systematic Reviews and Meta-Analysis 2020. The study protocol was registered on The International Prospective Register of Systematic Reviews (registration no. CRD42024573728).
The search included both full texts and abstracts of articles registered or published in databases up to April 2024. The search was conducted in MEDLINE (via PubMed), Embase (via OVID), Cochrane Register of Controlled Trials, and China National Knowledge Infrastructure (CNKI). Within each search engine, three researchers independently screened records for inclusion, and researchers were blinded to each other’s decision. Disagreements were resolved by majority votes using an Excel spreadsheet (Microsoft). Search strategies for an online search are listed in Table 1.
Table 1 . Search strategies for an online search.
PubMed | #1. Search: (Bell palsy [Mesh Terms] OR facial palsy [All Fields] OR facial paralysis [All Fields] OR Bell’s palsy) |
#2. Search: (Electroacupuncture [Mesh terms] OR Electric acupuncture [All fields] OR Electrical acupuncture) | |
#3. Search: (Randomized controlled trial [All Fields] OR Randomized clinical trial [All Fields]) | |
#4. #1 AND #2 AND #3 | |
OVID | (AllFields:(‘Bell’s palsy’ OR ‘facial palsy’ OR ‘facial paralysis’) AND (AllFields:(‘electroacupuncture’ OR ‘electric acupuncture’ OR ‘electrical acupuncture’)) AND (AllFields:(‘randomized controlled trial’ OR ‘randomized clinical trial’))) |
Cochrane Library | #1. ((Bell’s palsy OR facial palsy OR facial paralysis OR Bell palsy)):ti,ab,kw |
#2. ((electroacupuncture OR electric acupuncture OR electrical acupuncture)):ti,ab,kw | |
#3. ((randomized controlled trial OR randomized clinical trial)):ti,ab,kw | |
#4. #1 AND #2 AND #3 | |
CNKI | Used professional search and checked cross-language search |
TKA = (‘Bell palsy’ + ‘facial palsy’ + ‘facial paralysis’) AND TKA = (electroacupuncture + electric acupuncture + electrical acupuncture) AND TKA = (randomized controlled trial + randomized clinical trial) |
CNKI, China National Knowledge Infrastructure..
Inclusion criteria: 1) Study design: RCTs, 2) Population: patients with Bell’s palsy (including idiopathic facial paralysis and facial paralysis), 3) Intervention: including EA, 4) Comparison: all treatments, including oral medication, acupuncture, or physical therapy. Exclusion criteria: Excluding studies with EA in both groups.
A researcher independently extracted the data, and the other two researchers checked the extracted data. Data were extracted based on the target disease, intervention, duration and number of interventions, assessment scale, control group setting, number of participants, treatment efficacy, and year of publication. Data were recorded in an Excel spreadsheet, and disagreements were resolved by majority votes by an Excel spreadsheet.
The risk-of-bias (RoB) was assessed using the RoB tool for randomized trials (RoB 2) [4] developed by the Cochrane group. Studies were assessed as “low,” “some concerns,” or “high” based on six domains: biases arising from the randomization process, deviations from intended interventions, missing outcome data, measurement of the outcomes, and selection of the reported result and overall bias. Each domain has signaling questions, and the questions are answered “yes,” “probably yes,” “probably no,” “no,” and “no information.” Depending on the response, the RoB judgment is primarily evaluated through the algorithm results. The authors followed the algorithm results.
The primary outcome measure was the House-Brackmann facial nerve grading system (HBGS). Secondary outcome measures were the total facial nerve index (TFNI) score, Sunnybrook facial grading system (SFG) score, total effective rate, recovery duration, effective rate, cured rate, World Health Organization Quality of Life-BREF (WHOQOL-BREF), and facial disability index (FDI) including social functioning and well-being (FDIs) and physical function (FDIp).
The grading of recommendations assessment and the development and evaluation (GRADE) method in the GRADE Pro tool was used to rate the quality of evidence. The quality of evidence from the studies began at a high level; however, the scores gradually decreased to “moderate,” “low,” or “very low” based on the limitations in design and implementation (RoB), inconsistency (unexplained heterogeneity), indirectness, imprecision (spare data), and publication bias.
From the 141 articles searched, 30 were selected after excluding 17 duplicates and articles that did not target EA or Bell’s palsy based on the title and abstract. Then, through a secondary full-text review, 11 articles [5-15] were selected after excluding protocol, case reports, and articles that used EA in both groups. The above contents are presented in Fig. 1. The 11 studies were analyzed and summarized in Table 2.
Table 2 . Summary of randomized controlled trials of EA in Bell’s palsy.
Study | Sample size | Treatment group (sample size) | Control group (sample size) | Retaining time × no./duration | Outcome measurement | Results |
---|---|---|---|---|---|---|
Li et al. [5] | 60 | Treatment group (30): EA | Control group (30): TES | 30 min × 72/24 wk | 1. TFNI score 2. SFG score | 1. T > C (p < 0.05) 2. T > C (p < 0.05) (without 6 week) |
Li et al. [6] | 900 | Group C (180): EA (after 7 days) Staged acupuncture | - Group A (180): Staged acupuncture - Group B (180): Staged acupuncture + moxibustion - Group D (180): Staged acupuncture + line puncture on the muscle - Group E (180): acupuncture | 30 min × 20/ 4 wk | 1. Total effective rate | 1. T = C (p > 0.05) |
Shen et al. [7] | 279 | - Group A (74): Acupuncture (3 days after onset) EA (after 5 days) Prednisone (based on the date of onset), 30 mg (days 3–5), 15 mg(days 6–8), 10 mg (days 9–13), 5 mg (days 14–18) - Group B (70): Acupuncture (7 days after onset) EA (after 5 days) + control group - Group C (74): Acupuncture (10 days after onset) EA (after 5 days) Prednisone (based on the date of onset), 30 mg (days 3–5), 15 mg (days 6–8), 10 mg (days 9–13), 5 mg (days 14–18) | Control group (61): Prednisone (based on the date of onset), 30 mg (days 3–5), 15 mg (days 6–8), 10 mg (days 9–13), 5 mg (days 14–18) | 20 min × 25/25 d | 1. Total effective rate 2. Recovery duration | 1. T > C (p < 0.05) 2. T < C (p < 0.05) |
Liu et al. [8] | 86 | Treatment group (45): Staged acupuncture EA Chinese medicine modified Qianzheng powder (twice a day/50 g) | Control group (41): Staged acupuncture Chinese medicine modified Qianzheng powder (twice a day/50 g) | 20–30 min × 42/60 d | 1. Effective rate | 1. T > C (p < 0.05) |
Hong et al. [9] | 288 | Group C (96): EA (after 7 days) Staged acupuncture | - Group A (96): Staged acupuncture - Group B (96): Staged acupuncture, moxibustion | (not reported) × 20/4 wk | 1. FDI 2. WHOQOL-BREF and facial nerve paralysis score | 1. T = C (p > 0.05) 2. T = C (p > 0.05) |
An et al. [10] | 80 | Treatment group (40): EA Prednisone (based on the date of register), 30 mg (day 1), 15 mg (days 2–7), 5 mg (days 8–10), neutrophis-vitamin B1 | Control group (40): Prednisone (based on the date of register), 30 mg (day 1), 15 mg (days 2–7), 5 mg (days 8–10), neutrophis-vitamin B1 | 30 min × 10/10 d | 1. Total effective rate 2. Recovery duration | 1. T = C (p > 0.05) 2. T < C (p < 0.05) |
Yang et al. [11] | 197 | Control group (72): EA | Observed group (125): Acupoint injection (prednisolone 50 mg, vitamin B12 0.5 mg/mixed 3 mL inj. to four acupoint) Acupuncture Warming needle therapy | - Treatment group: 20 min × 30/ 6 wk - Control group: 30 min × 30/ 6 wk | 1. Cured rate 2. Total effective rate | 1. T < C (p < 0.01; only first evaluation) 2. T < C (p < 0.05) |
Chen et al. [12] | 304 | Group C (76): Staged acupuncture EA (after 7 days) | - Group A (76): Staged acupuncture - Group B (76): Staged acupuncture + moxibustion - Group D (76): Muscle-region alignment needling - Group E (76): Nonstaged acupuncture | 20–30 min × 70/70 d | 1. HBGS 2. FDI 3. WHOQOL-BREF 4. Total effective rate | 1. T > C (p < 0.01) 2. T = C (p > 0.05) 3. T = C (p > 0.05) 4. T = C (p > 0.05) |
Chen et al. [13] | 60 | Treatment group (30): Staged acupuncture Moxibustion EA (after 8 days) | Control group (30): Prednisone acetate tablets (until day 28) Methylcobalamin tablets (until day 28) | 30 min × 28/ 4 wk | 1. HBGS 2. SFG score 3-1. FDIs 3-2. FDIp 4. Total effective rate | 1. T = C (p > 0.05) 2. T > C (p < 0.05) 3-1. T > C (p < 0.05) 3-2. T < C (p < 0.05) 4. T > C (p < 0.05) |
Liu et al. [14] | 131 | EA group (45): EA Staged acupuncture | - Acupuncture group (44): Staged acupuncture - Medication and acupuncture group (42): Staged acupuncture Prednisone (oral, based on the date of onset 20 mg until day 3, 5 mg until day 7), acyclover (0.6 mg daily, oral), vitamin B1 (IM, 100 mg, daily after 7 days) and vitamin B12 (IM, 100 μg, daily, after 7 days) | 30 min × 20/ 4 wk | 1. Cured rate 2. Total effective rate | 1. T > C (p < 0.01) 2. T = C (p > 0.05) |
Xu [15] | 130 | - Group A (26): EA (sparse-dense wave > 2 Hz) + acupuncture - Group B (26): EA (sparse-dense wave < 2 Hz) + acupuncture - Group C (26): EA (continuous wave > 2 Hz) + acupuncture - Group D (26): EA (continuous wave < 2 Hz) + acupuncture | Control group (26): Acupuncture | 30 min × 24/ 4 wk | 1. HBGS 2. Total effective rate 3. Recovery duration | 1. T = C (p > 0.05) 2. T = C (p > 0.05) 3. T < C (p < 0.05) |
EA, electroacupuncture; TES, transcutaneous electrical stimulation; TFNI, total facial nerve index; SFG, Sunnybrook facial grading system; FDI, facial disability index; WHOQOL-BREF, World Health Organization Quality of Life-BREF; inj., injection; HBGS, House-Brackmann facial nerve grading system; FDIs, FDI including social functioning and well-being; FDIp, FID including physical function..
In seven studies [5,7,8,10,13-15], the treatment group showed beneficial results before and after evaluation compared with the control group. Six studies [6,8,9,12, 14,15] used the same acupuncture treatment for the control and treatment groups, and five of the studies [5,7,10,11,13] varied the acupuncture treatment by treatment period. Six studies [7,8,10,11,13,14] used medications for the control group inwhich four [7,10,13,14] used prednisone. The treatment duration was 4 weeks in five studies [6,9,13-15], with 72 sessions over 24 weeks in the longest study [5]. In all but one study, which did not specify the needle retention time, the needles were retained for 20–30 minutes. Excluding one study that did not specify the acupoints, Taiyang (extra points-head and neck 5) was the most-used acupoint for EA in eight studies, and gallbladder meridian 14 was used for EA in seven studies, followed by stomach meridia (ST) 4, ST7, etc. (Table 3). All studies were published in China and were written in Chinese, except for two studies [5,6] written in English. Six studies [6,8,9,12-14] used staged acupuncture, and acupuncture treatments varied by treatment period: acute phase, 1–7 days; resting phase, 8–20 days; and recovery period, 21–70 days [16]. Only one study [13] had different periods: acute phase, 1–7 days; subacute phase, 8–14 days; and recovery period, 15–28 days.
Table 3 . Frequency of acupoints used in the studies.
Frequency (Hz) | Acupoints |
---|---|
8 | EX-HN5 |
7 | GB14 |
6 | ST4 |
4 | ST7 |
Ex-HN, extra points-head and neck; GB, gallbladder meridian; ST, stomach meridian..
Three studies [12,13,15] used HBGS as the outcome measure. Chen et al. [12] did not find a significant difference between the control and treatment groups (p > 0.05). However, Chen et al. [13] stated that before treatment, HBGS was not significant in both groups (p > 0.05); after treatment, a significant difference was found in HBGS in the treatment group compared with the control group (p < 0.01). In the study by Xu [15], all five groups had significantly decreased HBGS before treatment and at weeks 2 and 4 (p = 0.000). Only the comparison between groups A (EA with sparse-dense wave > 2 Hz) and E (acupuncture; p = 0.031) showed a significant difference. Comparisons between groups B (EA with sparse-dense wave < 2 Hz) and E (p = 0.827), groups C (EA with continuous wave > 2 Hz) and E (p = 1.000), and groups D (EA with continuous wave < 2 Hz) and E (p = 1.000) did not show significant differences. The GRADE analysis indicated that the evidence for HBGS was of low quality. The RoB and inconsistency of the studies downgraded the quality of evidence.
Of the eight studies [6,7,10-15] that used the total effective rate as an endpoint, only two reported significantly higher results in the treatment group than in the control group. Li et al. [6] set the criteria of the total effective rate as HBGS, FDI scale, and degree of facial nerve paralysis scale. Li et al. [6] also stated that in the comparisons of the total effective rates of the five groups, no significant differences were noted: group A (staged acupuncture), 87.0; group B (staged acupuncture with moxibustion), 87.6; group C (staged acupuncture with EA), 86.0; group D (staged acupuncture with line puncture on muscles), 85.4; group E (acupuncture), 84.4 (all comparisons p > 0.05). Shen et al. [7] set the criteria of the total effective rate as HBGS, eyes closed, bilateral forehead symmetry, bilateral nasolabial folds, cheek symmetry, and exposed teeth. The treatment groups showed significantly higher results than the control group: group A (EA and drug with acupuncture 3 days after onset), 98.6; group B (acupuncture and EA and drug with acupuncture 7 days after onset), 95.7; group C (acupuncture and EA and drug with acupuncture 10 days after onset), 94.6; and group D (drug group), 72.1 (groups A, B, and C compared with group D, p < 0.05). An et al. [10] set the criteria of the total effective rate based on the HBGS. The difference between the treatment and control groups was not significant (treatment group, 100.00; control group, 100.00; p = 1.00). Yang et al. [11] set the criteria as facial asymmetry and movement. In the study by Yang et al. [11], the difference between the control and treatment groups was significantly high (control group, 100.0; treatment group, 86.1; p < 0.01). Chen et al. [12] set the criteria of the total effective rate based on the HBGS and FDI. In their study, the difference between the treatment and control groups was not significant: group A (staged acupuncture), 92.3; group B (staged acupuncture with moxibustion), 85.9; group C (staged acupuncture with EA), 88.0; group D (muscle-region alignment needling), 83.1; group E (nonstaged acupuncture group), 84.3 (all comparisons, p > 0.05). Chen et al. [13] set the criteria of the total effective rate as the HBGS. Chen et al. [13] stated that the total effective rate of the treatment group was significantly higher than that of the control group (treatment group, 93.3; control group, 83.3; p < 0.05). Liu et al. [14] set the criteria of the total effective rate as the HBGS. Liu et al. [14] stated that the EA group had significantly higher total effective rates than the other groups (EA group, 93.4; acupuncture group, 76.6; medication and acupuncture group, 78.6; comparison of EA group, p < 0.05). Xu [15] set the criteria of the total effective rate as the HBGS. Xu [15] evaluated the total effective rate by weeks 2 (w2) and 4 (w4) of treatment. In this study, all group comparisons were not significant (w2: group A, 88.0; group B, 76.9; group C, 69.2; group D, 61.9; group E, 68.0; p = 0.123; w4: group A, 100.0; group B, 96.1; group C, 96.1; group D, 100.0; group E, 92.0; p = 0.307). The GRADE analysis indicated that the evidence for the total effective rate was of low quality. The RoB and inconsistency of the studies downgraded the quality of evidence.
All three studies [7,10,15] that used recovery duration as an endpoint reported that recovery took significantly less time in the treatment group than in the control group. In the study by Shen et al. [7], the recovery duration significantly increased gradually in the order of groups A, B, and C and control group (group A, 31.90 ± 8.15; group B, 35.58 ± 9.10; group C, 39.91 ± 11.49; control group, 47.21 ± 12.04; each comparison p < 0.05). In the study by An et al. [10], the treatment group had a significantly longer recovery duration than the control group (treatment group, 28.16 ± 8.62; control group, 29.35 ± 6.26; p = 0.031). In the study Xu [15], group A had a significantly shorter recovery duration than group E. However, the comparisons between groups B and E, groups C and E, and groups D and E did not reveal significant differences (group A, 19.00 ± 4.75; group B, 21.00 ± 0.19; group C, 23.33 ± 1.41; group D, 23.33 ± 1.41; group E, 23.30 ± 1.49; groups A vs. E, p = 0.010; groups B vs. E, p = 0.164; groups C vs. E, p = 0.663; groups D vs. E, p = 0.985). The GRADE analysis indicated that the evidence for the recovery duration was moderate. The RoB of the studies downgraded the quality of evidence.
Two studies [5,13] that used the SFG score as an endpoint reported significant results in the treatment group compared with the control group. Li et al. [5] evaluated the SFG score by weeks 6 (6w), 12 (12w), and 24 (24w) of treatment. At week 6, no significant difference was found between the treatment and control groups (treatment group, 63.20; control group, 59.57; p = 0.74). At weeks 12 and 24, a significant increase was noted in the treatment group compared with the control group (12w: treatment group, 61.10; control group, 54.30; p = 0.005; 24w: treatment group, 60.50; control group, 52.30; p = 0.001). Chen et al. [13] reported that the treatment group showed significantly higher SFG scores than the control group (treatment group, 87.50; control group, 80.00; p < 0.05). The GRADE analysis indicated that the evidence for the SFG score was of moderate quality. The RoB of the studies downgraded the quality of evidence.
Of the three studies [9,12,13] that used FDI as an endpoint, two [9,12] found no significant difference between the treatment and control groups, whereas one [13] reported significant differences in FDIs and FDIp between the treatment and control groups. Hong et al. [9] stated no significant difference among groups A (staged acupuncture), B (staged acupuncture with moxibustion), and C (EA with staged acupuncture). The FDIp values were as follows: group A, 23.14 ± 2.58; group B, 23.02 ± 2.68; group C, 22.83 ± 2.71 and FDIs values were: group A, 9.01 ± 3.64; group B, 9.42 ± 3.47; group C, 9.28 ± 3.38, with all comparisons showing p > 0.05. Chen et al. [12] did not find a significant difference among the five groups (FDIp: group A, 23.08 ± 2.65; group B, 22.55 ± 2.73; group C, 22.36 ± 3.05; group D, 21.99 ± 3.54; group E, 22.88 ± 2.74; FDIs: group A, 8.97 ± 3.70; group B, 9.50 ± 3.53; group C, 9.20 ± 3.58; group D, 9.21 ± 3.69; group E, 9.04 ± 3.46; all comparisons of each group, p > 0.05). Chen et al. [13] reported significantly higher FDIp in the treatment group; however, the treatment group showed significantly lower FDIp than the control group (FDIp: treatment group, 25.00; control group, 24.00; FDIs: treatment group, 10.00; control group, 11.00; both FDIp and FDIs compared with the control group, p < 0.05). The GRADE analysis indicated that the evidence for FDI was of moderate quality. The RoB of the studies downgraded the quality of evidence.
Two studies [11,14] used the cured rate as an endpoint, and Liu et al. [14] reported significantly higher results in the treatment group than in the control group when there was an abnormal lesion in the upper part of the chorda tympany (EA group, 85.2; acupuncture group, 54.2; medication and acupuncture group, 48.0; p < 0.01). However, no significant difference was found between the treatment and control groups when there was an abnormal lesion on the lower part of the chorda tympany (EA group, 66.7; acupuncture group, 50.0; medication and acupuncture group, 70.6; p > 0.05). Yang et al. [11] reported significantly higher results in the control group than in the treatment group in the first evaluation (three evaluations were conducted over 2 weeks; first, second, and third evaluations of the treatment group, 22.9, 42.8, and 34.3, respectively; first, second, and third evaluations in the control group, 42.1, 44.2, and 13.7, respectively; only first evaluation, p < 0.05; other p-values were not reported). The GRADE analysis indicated that the evidence for the cured rate was of low quality. The RoB and inconsistency of the studies downgraded the quality of evidence. Two studies [9,12] used the WHOQOL-BREF, and they did not report significant differences. One study [8] used the effective rate as an outcome measurement, which was significant.
Cochrane RoB 2 was used to assess the 11 studies, and the results are summarized in the graphs shown in Fig. 2.
In D1, all studies [5-15] were assessed as “low.” Randomization was performed in all studies.
Five studies [5,7,10,13,15] were assessed as “low,” and the remaining six studies [6,8,9,11,12,14] were assessed to have “some concerns” because they did not clearly define the intentional intervention.
In D3, all studies [5-15] were assessed as “low.” In all studies, the baseline characteristics of the participants were randomized to each study group. The authors cannot find any evidence about the result being biased by missing outcome data.
In D4, eight studies [6,8,10-15] were assessed as “high,” and three studies [5,7,9] were assessed as “low.” The studies assessed as “low” clearly separated the intervention providers and the data assessors through blinding. The papers assessed as “high” did not clearly separate the intervention providers from the data assessors.
In D5, all studies [5-15] were assessed as having “some concerns.” Researchers cannot find a selective response about an incomplete outcome and cannot find a protocol or statistical analysis plan.
This systematic review of 11 RCTs (selected from 119 articles obtained from PubMed, OVID, Cochrane Register of Controlled Trials, and CNKI) compared the effectiveness of EA with those of traditional acupuncture and medications. Although several systematic reviews and meta-analyses of acupuncture for Bell’s palsy have been published, no systematic reviews focused on EA. Thus, this study may provide evidence of whether EA leads to a shorter time to recovery and is more effective than acupuncture or medications.
Studies utilizing the SFG and TFNI generally reported favorable outcomes for the treatment group, suggesting enhanced facial nerve recovery. However, studies using the HBGS and the total effective rate showed mixed results, with some reporting no significant differences between the treatment and control groups. This disparity may be attributed to the differences in assessment tools and their sensitivity to changes in facial nerve function.
The treatment duration and specific protocols varied widely among studies, from 4 to 24 weeks. This variation likely contributed to the mixed results. Studies with longer treatment durations tended to report more significant improvements, highlighting the potential need for prolonged EA for optimal outcomes.
In seven studies, the treatment group showed a significant improvement compared with the control group. However, HBGS, which was set as the primary outcome, was not significant in two out of three studies.
One study [6] mentioned one adverse event of fainting during acupuncture treatment with symptoms of dizziness, vomiting, and sweating and noted that the symptoms were relieved after the acupuncture needles were immediately withdrawn, followed by 10 minutes of relaxation with warm water. The remaining studies either stated that no adverse events occurred or did not mention adverse events.
A meta-analysis could not be performed in this study because of the following reasons. First, lack of consistency between studies: The results of the three studies using the HBGS are inconsistent, making it difficult to meet the assumption of homogeneity required for the meta-analysis. Second, differences in sample sizes and study design: The studies had varied sample sizes and treatment protocols, making it challenging to draw a unified conclusion in a meta-analysis. Third, quality issues in the results: Many studies were assessed to have a high RoB, which reduces the reliability of the outcomes.
This study has a few limitations. First, all the 11 studies selected were published in China, which was a somewhat significant deviation from the initial intention to include studies from different East Asian countries, including Korea, China, and Japan. This possibly limits the generalizability of the findings to other populations, and the lack of studies from other regions or unpublished data could skew the overall findings. Second, the number of studies was small, making it difficult to conduct a meta-analysis. Some outcomes showed significant differences, whereas others did not, and the studies had varied sample sizes, with some having relatively small populations. This could affect the precision of the estimates, particularly when significant results were found in smaller studies. Third, many studies lacked an experimental design, did not mention adverse events, and had no rigorous randomization, blinding, or allocation concealment. Eight studies had high risk because of outcome measurement issues, which might lower the quality of evidence. Fourth, the overall RoB in studies is quite high.
Future research should focus on large-scale, multicenter RCTs with standardized protocols to better evaluate the efficacy of EA. Studies should also explore the underlying mechanisms of EA in facial nerve recovery and identify subgroups of patients who may benefit the most from this treatment.
Several conclusions were drawn from these 11 RCTs of EA for Bell’s palsy. First, whether EA is a more effective treatment of Bell’s palsy than other treatments is not yet established; however, some studies have shown that it is more effective than other treatments. Second, the EX-HN5 acupoint was the most used for EA for Bell’s palsy. Third, further research is required to confirm the effectiveness of EA for Bell’s palsy.
The quality of evidence regarding the efficacy of EA for treating Bell’s palsy is moderate to low. Although some evidence suggests that EA might offer benefits in terms of recovery time and effectiveness compared with conventional treatments, significant limitations, such as RoB, inconsistency in results, and potential publication bias, prevent a strong recommendation. Future well-designed RCTs with more diverse populations and robust methodologies are necessary to confirm these findings and determine the optimal treatment protocols for Bell’s palsy.
This review highlights the need for a cautious interpretation of the current evidence and suggests that while EA may be effective, it should be considered part of a broader treatment strategy pending further research.
Conceptualization: MG. Data curation: MG, TK, SL. Formal analysis: MG, SB, HS. Investigation: MG. Methodology: MG, HS, JO, CL. Project administration: MG. Supervision: SY. Writing – original draft: MG.
The authors have no conflicts of interest to declare.
None.
This research did not involve any human or animal experiments. The study protocol was registered on The International Prospective Register of Systematic Reviews (registration no. CRD42024573728).
Table 1 . Search strategies for an online search.
PubMed | #1. Search: (Bell palsy [Mesh Terms] OR facial palsy [All Fields] OR facial paralysis [All Fields] OR Bell’s palsy) |
#2. Search: (Electroacupuncture [Mesh terms] OR Electric acupuncture [All fields] OR Electrical acupuncture) | |
#3. Search: (Randomized controlled trial [All Fields] OR Randomized clinical trial [All Fields]) | |
#4. #1 AND #2 AND #3 | |
OVID | (AllFields:(‘Bell’s palsy’ OR ‘facial palsy’ OR ‘facial paralysis’) AND (AllFields:(‘electroacupuncture’ OR ‘electric acupuncture’ OR ‘electrical acupuncture’)) AND (AllFields:(‘randomized controlled trial’ OR ‘randomized clinical trial’))) |
Cochrane Library | #1. ((Bell’s palsy OR facial palsy OR facial paralysis OR Bell palsy)):ti,ab,kw |
#2. ((electroacupuncture OR electric acupuncture OR electrical acupuncture)):ti,ab,kw | |
#3. ((randomized controlled trial OR randomized clinical trial)):ti,ab,kw | |
#4. #1 AND #2 AND #3 | |
CNKI | Used professional search and checked cross-language search |
TKA = (‘Bell palsy’ + ‘facial palsy’ + ‘facial paralysis’) AND TKA = (electroacupuncture + electric acupuncture + electrical acupuncture) AND TKA = (randomized controlled trial + randomized clinical trial) |
CNKI, China National Knowledge Infrastructure..
Table 2 . Summary of randomized controlled trials of EA in Bell’s palsy.
Study | Sample size | Treatment group (sample size) | Control group (sample size) | Retaining time × no./duration | Outcome measurement | Results |
---|---|---|---|---|---|---|
Li et al. [5] | 60 | Treatment group (30): EA | Control group (30): TES | 30 min × 72/24 wk | 1. TFNI score 2. SFG score | 1. T > C (p < 0.05) 2. T > C (p < 0.05) (without 6 week) |
Li et al. [6] | 900 | Group C (180): EA (after 7 days) Staged acupuncture | - Group A (180): Staged acupuncture - Group B (180): Staged acupuncture + moxibustion - Group D (180): Staged acupuncture + line puncture on the muscle - Group E (180): acupuncture | 30 min × 20/ 4 wk | 1. Total effective rate | 1. T = C (p > 0.05) |
Shen et al. [7] | 279 | - Group A (74): Acupuncture (3 days after onset) EA (after 5 days) Prednisone (based on the date of onset), 30 mg (days 3–5), 15 mg(days 6–8), 10 mg (days 9–13), 5 mg (days 14–18) - Group B (70): Acupuncture (7 days after onset) EA (after 5 days) + control group - Group C (74): Acupuncture (10 days after onset) EA (after 5 days) Prednisone (based on the date of onset), 30 mg (days 3–5), 15 mg (days 6–8), 10 mg (days 9–13), 5 mg (days 14–18) | Control group (61): Prednisone (based on the date of onset), 30 mg (days 3–5), 15 mg (days 6–8), 10 mg (days 9–13), 5 mg (days 14–18) | 20 min × 25/25 d | 1. Total effective rate 2. Recovery duration | 1. T > C (p < 0.05) 2. T < C (p < 0.05) |
Liu et al. [8] | 86 | Treatment group (45): Staged acupuncture EA Chinese medicine modified Qianzheng powder (twice a day/50 g) | Control group (41): Staged acupuncture Chinese medicine modified Qianzheng powder (twice a day/50 g) | 20–30 min × 42/60 d | 1. Effective rate | 1. T > C (p < 0.05) |
Hong et al. [9] | 288 | Group C (96): EA (after 7 days) Staged acupuncture | - Group A (96): Staged acupuncture - Group B (96): Staged acupuncture, moxibustion | (not reported) × 20/4 wk | 1. FDI 2. WHOQOL-BREF and facial nerve paralysis score | 1. T = C (p > 0.05) 2. T = C (p > 0.05) |
An et al. [10] | 80 | Treatment group (40): EA Prednisone (based on the date of register), 30 mg (day 1), 15 mg (days 2–7), 5 mg (days 8–10), neutrophis-vitamin B1 | Control group (40): Prednisone (based on the date of register), 30 mg (day 1), 15 mg (days 2–7), 5 mg (days 8–10), neutrophis-vitamin B1 | 30 min × 10/10 d | 1. Total effective rate 2. Recovery duration | 1. T = C (p > 0.05) 2. T < C (p < 0.05) |
Yang et al. [11] | 197 | Control group (72): EA | Observed group (125): Acupoint injection (prednisolone 50 mg, vitamin B12 0.5 mg/mixed 3 mL inj. to four acupoint) Acupuncture Warming needle therapy | - Treatment group: 20 min × 30/ 6 wk - Control group: 30 min × 30/ 6 wk | 1. Cured rate 2. Total effective rate | 1. T < C (p < 0.01; only first evaluation) 2. T < C (p < 0.05) |
Chen et al. [12] | 304 | Group C (76): Staged acupuncture EA (after 7 days) | - Group A (76): Staged acupuncture - Group B (76): Staged acupuncture + moxibustion - Group D (76): Muscle-region alignment needling - Group E (76): Nonstaged acupuncture | 20–30 min × 70/70 d | 1. HBGS 2. FDI 3. WHOQOL-BREF 4. Total effective rate | 1. T > C (p < 0.01) 2. T = C (p > 0.05) 3. T = C (p > 0.05) 4. T = C (p > 0.05) |
Chen et al. [13] | 60 | Treatment group (30): Staged acupuncture Moxibustion EA (after 8 days) | Control group (30): Prednisone acetate tablets (until day 28) Methylcobalamin tablets (until day 28) | 30 min × 28/ 4 wk | 1. HBGS 2. SFG score 3-1. FDIs 3-2. FDIp 4. Total effective rate | 1. T = C (p > 0.05) 2. T > C (p < 0.05) 3-1. T > C (p < 0.05) 3-2. T < C (p < 0.05) 4. T > C (p < 0.05) |
Liu et al. [14] | 131 | EA group (45): EA Staged acupuncture | - Acupuncture group (44): Staged acupuncture - Medication and acupuncture group (42): Staged acupuncture Prednisone (oral, based on the date of onset 20 mg until day 3, 5 mg until day 7), acyclover (0.6 mg daily, oral), vitamin B1 (IM, 100 mg, daily after 7 days) and vitamin B12 (IM, 100 μg, daily, after 7 days) | 30 min × 20/ 4 wk | 1. Cured rate 2. Total effective rate | 1. T > C (p < 0.01) 2. T = C (p > 0.05) |
Xu [15] | 130 | - Group A (26): EA (sparse-dense wave > 2 Hz) + acupuncture - Group B (26): EA (sparse-dense wave < 2 Hz) + acupuncture - Group C (26): EA (continuous wave > 2 Hz) + acupuncture - Group D (26): EA (continuous wave < 2 Hz) + acupuncture | Control group (26): Acupuncture | 30 min × 24/ 4 wk | 1. HBGS 2. Total effective rate 3. Recovery duration | 1. T = C (p > 0.05) 2. T = C (p > 0.05) 3. T < C (p < 0.05) |
EA, electroacupuncture; TES, transcutaneous electrical stimulation; TFNI, total facial nerve index; SFG, Sunnybrook facial grading system; FDI, facial disability index; WHOQOL-BREF, World Health Organization Quality of Life-BREF; inj., injection; HBGS, House-Brackmann facial nerve grading system; FDIs, FDI including social functioning and well-being; FDIp, FID including physical function..
Table 3 . Frequency of acupoints used in the studies.
Frequency (Hz) | Acupoints |
---|---|
8 | EX-HN5 |
7 | GB14 |
6 | ST4 |
4 | ST7 |
Ex-HN, extra points-head and neck; GB, gallbladder meridian; ST, stomach meridian..
Jin-Sol Yoon, Joonyong Noh, Seonju Ahn, Yun-Gwon Seon, Hong-Wook Choi, Sun-Joong Kim, Jae-Uk Sul, Jae-Hong Kim
Journal of Acupuncture Research 2023; 40(3): 214-237Seong-Kyeong Choi, Jeong-Hyun Moon, Woo-Seok Jang, Jung-Eun Jang, Si-Hyun Park, Won-Suk Sung, Chan-Yung Jung, Byung-Kwan Seo, Seung-Deok Lee, Kyung-Ho Kim, Eun-Jung Kim,
Journal of Acupuncture Research 2023; 40(1): 16-34Junyeop Oh, Jihoon Kong, Jeonghoon Lyu, Tae Han Yook, Beom Yong Song, Yoomin Choi, Jong Uk Kim
Journal of Acupuncture Research 2025; 42(): 59-72