Efficacy and safety of therapies related to acupuncture for acute herpes zoster: A PRISMA systematic review and network meta-analysis

Abstract

Background:

Herpes zoster (HZ) is mainly characterized by intense pain and severe skin lesions, particularly during the acute phase, which seriously affects the patient’s quality of life. Acupuncture is a widely used and effective treatment for HZ. However, there are many types of acupuncture, which have different curative efficacy. This study employed a network meta-analysis (NMA) to assess and rank the clinical efficacy of different acupuncture therapies.

Methods:

The database of Cochrane Library, Web of Science, PubMed, MEDLINE, Embase, China National Knowledge Infrastructure (CNKI), Chinese BioMedical Database, VIP Database, and Wanfang Database were searched from inception to December 31, 2022 to identify eligible randomized controlled trials (RCTs) of acupuncture related therapies in the treatment of acute HZ. The outcome indicators measured were visual analogue scale (VAS), date of cessation of herpes increase (DCHI), effective rate (ER), postherpetic neuralgia (PHN), and adverse events (AEs). Bayesian network meta-analyses were performed using the GeMTC package (version 1.0-1) and R software (version 4.2.3).

Results:

A total of 59 RCTs with 3930 patients were included. The results of this NMA were as follows: compared with pharmacotherapy, electroacupuncture (EA) + pricking and cupping (PC) shown the best efficacy to improve VAS score and reduce DCHI. In terms of ER, EA + fire needle (FN) had the highest results of probability ranking. PC was more effective in reducing the incidence of PHN. Furthermore, this study shown that the incidence of AEs associated with acupuncture-related therapies was acceptable.

Conclusions:

This study indicated that therapies related to acupuncture were both effective and safe in treating acute HZ, and could significantly reduce patients’ symptoms such as pain and skin lesions with fewer adverse events. Clinically, the selection of the appropriate therapy should be based on practical considerations. However, due to the limitations of this study, more high-quality trials are required to evaluate the efficacy and safety of acupuncture-related therapy for the treatment of acute HZ.

1. Introduction

Herpes zoster (HZ) is an acute viral infectious disease caused by the reactivation of varicella-zoster virus (VZV) that persists within the human body.^[1] Herpes zoster has three clinical stages: prodromal phase, acute phase, and cure phase,^[2] the acute phase of HZ is characterized by severe pain, and unilateral maculopapular rash, which then evolves into vesicles.^[3] Moreover, HZ may cause complications including postherpetic neuralgia (PHN),^[4] which is defined as pain in a dermatomal distribution, lasting for at least 90 days after acute HZ,^[5] and has a significant impact on the patient’s quality of life. Global epidemiology investigation shows that the median incidence of HZ is about 0.4% to 0.45%^[6] with a higher incidence rate among females.^[7] Furthermore, with the increase of age, the incidence and hazard of HZ further increase.^[8]

Current treatment for HZ mainly focuses on controlling pain and using antiviral medication.^[2] During the acute phase of HZ, the primary goal of treatment is using appropriate analgesics to control pain.^[9] Antiviral medication, such as acyclovir, valacyclovir, and famciclovir, is often used in treating HZ to help reduce the severity and duration of symptoms,^[10] A study shown that antiviral therapy could decrease the duration of lesions and severity of pain, even reduce the incidence of PHN.^[11] Although antiviral therapy has demonstrated effectiveness in some cases, there are also negative results associated with antiviral therapy for HZ. A review revealed that antiviral treatment did not significantly reduce the incidence of PHN.^[12] Another study found that combing antiviral therapy with corticosteroids might increase the risk of adverse events.^[13]

Acupuncture-related therapies have been shown to be effective in treating the acute phase of HZ. A clinical trial illustrated that acupuncture was effective in reducing severe acute pain associated with HZ.^[14] And a meta-analysis indicated that acupuncture combined with moxibustion was effective in reducing both pain and complications associated with HZ.^[15] In addition, other related therapies, like cotton sheet moxibustion, were also believed to be effective in HZ.^[16]

However, despite the various acupuncture-related therapies, such as electroacupuncture (EA), fire needle (FN), pricking, and cupping (PC), no systematic comparison of their actual efficacy and safety in treating the acute phase of HZ has been conducted. Therefore, this study aimed to perform the network meta-analysis (NMA) method to compare the efficacy and safety of different therapies in the treatment of acute HZ. This will provide valuable evidence for clinical treatment and identify preferable interventions.

2. Methods

The systematic review and NMA has been registered (registration number: CRD42022367196) with the International Prospective Register of Systematic Reviews (PROSPERO). And this research was reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement.^[17] The details of PRISMA are shown in Supplementary Material S1, http://links.lww.com/MD/M415.

2.1. Search strategy

Cochrane Library, Web of Science, PubMed, MEDLINE, Embase, China National Knowledge Infrastructure (CNKI), Chinese BioMedical Database (SinoMed), VIP Database, and Wanfang Database will be searched from inception to December 31, 2022, without language, country, or article type restrictions. The search term will use the combination of intervention (acupuncture, moxibustion, etc.), disease (herpes zoster), and study design (RCT, comparative studies, etc.). When searching different databases, search terms will be adjusted accordingly. The PubMed database search strategy was shown in Table 1.

Table 1.

PubMed database search strategy.

Number	Search terms
#1	Acupuncture Therapy [MeSH Terms]
#2	Acupuncture [Title/Abstract]
#3	Electroacupuncture [Title/Abstract]
#4	Fire Needle [Title/Abstract]
#5	Warm Needle [Title/Abstract]
#6	Blood-letting Therapy [Title/Abstract]
#7	Moxibustion [Title/Abstract]
#8	Auricular Needle [Title/Abstract]
#9	Acupoint Catgut Embedding [Title/Abstract]
#10	#1 OR #2 OR #3 OR #4 OR #5 OR #6 OR #7 OR #8 OR #9
#11	Herpes Zoster [MeSH Terms]
#12	Zona [Title/Abstract]
#13	Zoster [Title/Abstract]
#14	Shingles [Title/Abstract]
#15	#11 OR #12 OR #13 OR #14
#16	Randomized Controlled Trial [Publication Type]
#17	Controlled Clinical Trial [Publication Type]
#18	Randomized [Title/Abstract]
#19	Randomly [Title/Abstract]
#20	Trial [Title/Abstract]
#21	#16 OR #17 OR #18 OR #19 OR #20
#22	#10 AND #15 AND #21

2.2. Trial selection criteria

Eligible trials included randomized clinical trials (RCTs) with 2-arm or more arms, and compared therapies related to acupuncture with medication, placebo, sham acupuncture, or other acupuncture-related therapies. Studies had to report at least one targeted outcome measurement: visual analogue scale (VAS), the date of cessation of herpes increase (DCHI), therapeutic effective rate (ER), postherpetic neuralgia (PHN) incidence at 1 month, and adverse events. Studies were excluded if the RCT included patients who experienced acute HZ onset for more than 7 days or if the included patients belonged to special populations, such as pregnant women. Additionally, studies were excluded if participants had special types of herpes zoster, such as Hunter syndrome or genital shingles.

2.3. Trial identification

ZOTERO software (version 6.0.26) was used to manage the retrieved literature records, and automatically excluded the duplicate documents. Two researchers (X.Y.L. and X.C.) independently performed the preliminary screening by reading the title and abstract, after that, retrieve the full text of the remaining records, and screened the retrieved articles using the predetermined inclusion and exclusion criteria. Any disagreement between reviewers was solved through discussion with a third viewer (X.M.L.).

2.4. Data extraction

The following data were independently extracted by two researchers (X.Y.L. and X.C.) using EpiData 4.6 software after read the full literature: Basic information: title, first author, year of publication, language, country, and contact information; Study design: sample size, number of groups, randomization, allocation hiding, blinding; According to the STRICTA standard, the type of intervention, frequency, duration, and other information of each group; and Outcome indicators.

If there is a disagreement between the data results, the two researchers should discuss and decide, and if there is still disagreement, a third researcher should intervene in the discussion and make a decision.

2.5. Evaluation of the risk of bias

Two reviewers assessed the risk of bias independently, used the second version of the Cochrane Risk of Bias (RoB 2.0).^[18] The assessment focused on five domains of the studies: randomization process, deviations from intended interventions, missing outcome data, measurement of the outcome, and selection of the reported result. Risk of bias was identified by selecting the appropriate response options to the different signaling questions, and classified as “low risk of bias,” “some concerns,” and “high risk of bias.” Finally, the overall risk of bias judgment will be provided. If any disagreement exists, discuss it with a third viewer.

2.6. Statistical analysis

A frequency-based paired meta-analysis was conducted to analyze all direct comparisons. Risk ratio (RR) and its 95% confidence interval (CI) were used to evaluate the efficacy of dichotomous data, while standard mean deviation (SMD) and its 95% CI were used to evaluate the efficacy of continuous data.

A NMA was performed using GeMTC package (version 1.0-1) and R software (version 4.2.3) within a Bayesian framework. Supplementary Material S2, http://links.lww.com/MD/M416 provides the running code for the NMA. For the network plot, studies with three or more arms were split into comparable two-arm trials to identify comparative relationships among interventions. The Markov Chain Monte Carlo (MCMC) method was adopted to perform network meta-analysis on the fixed effects model with vague priors. Four MCMC chains were run simultaneously, with the number of iterations set as 30,000 times (the first 5000 times for annealing, and the last 25,000 times for sampling). And performed the Brooks-Gelman-Rubin method^[19] which evaluated the Potential Scale Reduction Factor (PSRF) to assess the convergence of MCMC. If the PSRF value was close to 1 and remains stable, it represents that the MCMC model converges well.^[20] If there was no convergence, the number of iterations was increased until the convergence was achieved.

For dichotomous data, RR with 95% credibility intervals (CrI)^[21] were used as statistics to evaluate the efficacy. Continuous data will be analyzed using mean difference (MD) with 95% CrI. However, NMA comparisons are difficult to conduct when certain types of acupuncture-related therapies are lacking, thus, this study combined them into the category of “other acupuncture-related treatments” (OAT) for analysis. If necessary, a narrative description of the results would be provided. To rank the effectiveness of each treatment and identify the most effective option, study illustrated the probability ranking results as a rankogram and calculated the surface under the cumulative ranking curve (SUCRA).^[22]

And the evaluation of local and global inconsistency were performed, which indicates the credibility of the NMA result.^[23] Among the evaluation of the global inconsistency of the NMA, the Bayesian unrelated mean effects (UME) model was performed,^[24] and its deviance information criterion (DIC) was compared with the consistency model.^[25] If the UME model with a smaller DIC value of more than 5 units,^[26] this was evidence of possible inconsistency in the network.^[27] The node-splitting model was used to evaluate the local inconsistency if closed loops existed in the NMA diagram.^[28] The summary of direct and indirect evidence for valid treatment comparisons was provided,^[29] and any comparison with a Bayesian P value of less than 0.05 was considered to have significant inconsistency between the direct and indirect evidence. Further details regarding the inconsistency analysis can be found in Supplementary Material S3, http://links.lww.com/MD/M417. A comparison-adjusted funnel plot was conducted to assess the potential publication bias and small sample effect for all included studies. According to the Cochrane Handbook,^[30] this study performed a sensitivity analysis on the VAS scores, to assess the robustness and reliability of the NMA results.

3. Results

3.1. Literature screen results

A total of 2184 literatures were retrieved, after the preliminary screening by title and abstract, and rescreened the literature by reading the full text, 59 RCTs with 3930 patients were finally included.^[31–89] The literature screening process is shown in Figure 1.

Figure 1.

Literature screening process.

3.2. Characteristics of the Included Studies

Of the final 59 studies included, 50 studies investigated 3358 individuals with an average age of 46.13 (SD: 13.44) years, and 1710 (50.92%) were female, there is no difference in gender composition (χ² = 1.1447 P > .05). In relation to outcome measures reported, 35 trials (59.32%) assessed pain intensity by VAS, 36 trials (61.02%) assessed the date of cessation of herpes increase, 46 trials (77.97%) assessed effective rate of interventions, 32 trials (54.24%) assessed postherpetic neuralgia, 23 trials (38.98%) assessed adverse events of interventions. The characteristics of the included studies are illustrated in (Supplementary Material S4.1, http://links.lww.com/MD/M418).

There are 3 three-arm studies and 56 two-arm studies. In terms of intervention categories, all but 7 studies were established pharmacotherapy groups, the rest of the treatment groups received single or combined acupuncture-related therapies. Furthermore, the average number of treatment courses was 11.24 days (SD: 3.25). The characteristics of intervention measures are shown in (Supplementary Material S4.2, http://links.lww.com/MD/M418).

3.3. Results of the risk of bias

Overall, 8 articles had a low risk of bias (RoB), 49 were judged to have some concerns regarding the RoB, and 2 were rated as having a high RoB. In terms of each RoB domain, only 8 studies had low RoB in the randomization process, while the others had an unclear RoB due to unspecified allocation concealment. Since acupuncture therapies have specific characteristics, there is no blinding possible for participants or personnel. Most studies had low RoB of intended interventions, but 5 studies were identified as having either high or unclear RoB due to reported deviations from the intended intervention. Two studies reported the loss to follow-up and withdrawal, but the missing data were assumed to be unrelated to its true value, thus all studies rated as having low RoB. One study measured the outcomes of the different groups at different periods and was judged as having a high risk of selective reporting. The details of the RoB assessments in each study and the summary of the risk of bias evaluation are shown in (Fig. 2).

Figure 2.

Results of the risk of bias evaluation (A for each study, B for summary).

3.4. Results of the paired meta-analysis

The paired meta-analysis involved 59 studies comprising a total of 3930 patients. The results of this paired meta-analysis provided effect estimates and their corresponding 95% CI for each study, as well as summary data for each intervention group. The detailed results of the paired meta-analysis can be found in (Supplementary Material S5, http://links.lww.com/MD/M419).

3.5. Evidence network diagram

35 studies reported VAS, involved 11 treatment categories, which included pharmacotherapy, MA, EA, FN, PC, MA + FN, MA + PC, EA + FN, EA + PC, FN + PC, and OAT. 36 studies reported DCHI, involved 10 treatment categories, which included pharmacotherapy, MA, EA, FN, MA + FN, MA + PC, EA + FN, EA + PC, FN + PC, and OAT. 45 studies reported ER, involved 11 treatment categories, which included pharmacotherapy, MA, EA, FN, PC, MA + FN, MA + PC, EA + FN, EA + PC, FN + PC, and OAT. 35 studies reported PHN, involved 10 treatment categories, which included pharmacotherapy, MA, EA, FN, PC, MA + FN, MA + PC, EA + FN, FN + PC, and OAT.

Figure 3 shows the network graph of eligible comparisons. The size of nodes represented the number of participants in each intervention, while the thickness of lines between interventions corresponded to the number of studies in each comparison. Except for PC, MA + FN, the remaining nodes formed distinct closed loops, which indicated that there was no direct comparison between PC and MA + FN.

Figure 3.

Network graph of eligible comparisons (A for VAS, B for DCHI, C for ER, D for PHN). EA = Electroacupuncture, FN = Fire Needle, MA = Manual Acupuncture, OAT = Other Acupuncture-related Treatments, PC = Pricking and Cupping.

3.6. Network meta-analysis results

3.6.1. Network meta-analysis results of VAS scores

35 studies with 2303 participants, contributed to this analysis. Compared with pharmacotherapy, all categories of treatments were found to be significantly more effective, as presented by the forest plot (Fig. 4A). Supplementary Material S6.1, http://links.lww.com/MD/M420 illustrates the pairwise comparisons matrix, including the mean difference with 95% CrI between the different interventions, the result shown that PC was more effective than other interventions and with a significant statistical difference. Figure 5A displays the probability ranking results of VAS. According to the SUCRA, the complete probability ranking results from highest to lowest were: PC > EA > EA + PC > FN + PC > EA + FN > FN > OAT > MA + PC > MA + FN > MA > Pharmacotherapy, as shown in (Supplementary Material S7.1, http://links.lww.com/MD/M421).

Figure 4.

Forest plots of the network meta-analyses (A for VAS, B for DCHI, C for ER, D for PHN). EA = Electroacupuncture, FN = Fire Needle, MA = Manual Acupuncture, OAT = Other Acupuncture-related Treatments, PC = Pricking and Cupping.

Figure 5.

Rankograms for eligible treatments (A for VAS, B for DCHI, C for ER, D for PHN). EA = Electroacupuncture, FN = Fire Needle, MA = Manual Acupuncture, OAT = Other Acupuncture-related Treatments, PC = Pricking and Cupping.

The evaluation of consistency was performed, and the DIC value indicated that there is no significant global inconsistency (consistency model DIC = 139.8, UME model DIC = 140.6). As local inconsistency results shown that, except for Pharmacotherapy versus EA, the effects estimated from direct and indirect comparisons were consistent in VAS (P > .05).

3.6.2. Network meta-analysis results of date of cessation of herpes increase

36 studies with 2323 participants, contributed to this analysis. Compared with pharmacotherapy, most of the included interventions shown a better treatment efficacy but MA and MA + PC were without statistical significance, the forest plot is presented in Figure 4B. Supplementary Material S6.2, http://links.lww.com/MD/M420 illustrates the pairwise comparisons matrix, including the mean difference with 95% CrI between the different interventions. Figure 5B shows the probability ranking results of DCHI, and according to the SUCRA, the complete probability ranking results were as follows: EA + PC > MA + FN > EA + FN > FN + PC > FN > OAT > MA + PC > MA > EA > Pharmacotherapy, as shown in (Supplementary Material S7.2, http://links.lww.com/MD/M421).

The evaluation of consistency was performed, and the DIC value indicated that there is no significant global inconsistency (consistency model DIC = 145.6, UME model DIC = 147.2).

Significant local inconsistency was found in a closed loop, namely, Pharmacotherapy and MA and MA + PC. especially in MA versus MA + PC, the comparison between direct and indirect have opposite results.

3.6.3. Network meta-analysis results of effective rate

45 studies with 3019 participants, contributed to this analysis. Compared with pharmacotherapy, all categories of treatments but MA + FN shown a larger treatment efficacy with a statistical significance, the forest plot is presented in Figure 4C. Supplementary Material S6.3, http://links.lww.com/MD/M420 illustrates the pairwise comparisons matrix, including the risk ratio with 95% CrI between the different interventions, it lists comparisons between all interventions. Figure 5C shows the probability ranking results of ER, and according to the SUCRA, the complete probability ranking results were as follows: EA + FN > PC > EA + PC > EA > FN > MA + PC > FN + PC > MA > OAT > MA + FN > Pharmacotherapy, as shown in (Supplementary Material S7.3, http://links.lww.com/MD/M421).

The evaluation of consistency was performed, and the DIC value indicated that there is no significant global inconsistency (consistency model DIC = 137.6, UME model DIC = 142.3). As local inconsistency result shown that, except for MA versus MA + PC, the effects estimated from direct and indirect comparisons were consistent (P > .05).

3.6.4. Network meta-analysis results of postherpetic neuralgia

32 studies with 2603 participants, contributed to this analysis. Compared with pharmacotherapy, all categories of interventions shown a larger treatment efficacy, however, MA and MA + FN with no statistical significance, the forest plot is presented in Figure 4D. Supplementary Material S6.4, http://links.lww.com/MD/M420 illustrates the pairwise comparisons matrix, including the risk ratio with 95% CrI between the different interventions. Except for pharmacotherapy, no differences between groups were noted among the treatments. Figure 5D shows the probability ranking results of PHN, and according to the SUCRA, the complete probability ranking results were as follows: PC > FN > MA + PC > EA + FN > OAT > EA > MA + FN > FN + PC > MA > Pharmacotherapy, as shown (Supplementary Material S7.4, http://links.lww.com/MD/M421).

The evaluation of consistency was performed, and the DIC value indicated that there is no 0significant global inconsistency (consistency model DIC = 106.5, UME model DIC = 108.5). And the local inconsistency result shown that there was no significant local inconsistency (P > .05).

3.7. Publication bias and small sample effect estimation

The comparison-adjusted funnel plot was employed to evaluate the primary outcome indicator, VAS, and to detect possible publication bias or small sample effect, as shown in (Fig. 6). The results demonstrated an asymmetrical funnel plot, particularly with regards to the highly positive outcome results observed in the studies conducted by Zhang and Xin^[85] and Li,^[52] which could potentially represent the main source of publication bias or small sample effect.

Figure 6.

Comparison-adjusted funnel plot of VAS. EA = Electroacupuncture, FN = Fire Needle, MA = Manual Acupuncture, OAT = Other Acupuncture-related Treatments, PC = Pricking and Cupping.

3.8. Sensitivity analysis

Studies conducted by Zhang Y, Li X, and colleagues were excluded due to their highly positive effect results. Then the analysis was re-performed on the pain intensity assessed by the VAS score. The results revealed a deviation in the ranking probability from the main NMA analysis, resulting in decreased ranks for some interventions. And this result is potentially more robust. Detailed results of sensitivity analyses are presented in (Supplementary Material S8, http://links.lww.com/MD/M422).

3.9. Adverse events

A total of 23 studies reported adverse events, as shown in (Supplementary Material S9, http://links.lww.com/MD/M423). It can be seen that the major adverse reactions of acupuncture-related therapies included fainting and hematoma formation, while the main adverse reactions of pharmacotherapy included dizziness, nausea, and abdominal discomfort. However, due to the limitation of included literature data, this review abstained from performing a quantitative analysis of adverse events.

4. Discussion

HZ has become a significant burden, with millions of people affected and exhibits a rising incidence rate.^[90] Research have shown that early effective treatment can reduce the duration and severity of the acute HZ,^[10] also potentially reduce the incidence of PHN, compared to placebo or delayed treatment.^[91] As a safe and reliable alternative therapy, acupuncture plays an important role in HZ treatment.^[15] However, there are various types of acupuncture-related therapies, which have different clinical effects. Therefore, this systematic review and network meta-analysis evaluated the effects of different therapies and aimed to identify the optimal acupuncture therapy for patients with acute HZ.

This study assessed the effects of different acupuncture-related therapies on VAS, DCHI, ER, and PHN in patients with acute HZ. VAS score measured HZ patients’ perception of pain, which is the primary symptom during the acute phase of HZ. This study found that PC is more effective than other interventions, according to the probability ranking results. However, study conducted a sensitivity analysis in which Zhang’s and Li’s studies were excluded, the probability ranking results of re-analysis shown that EA + PC was the most effective intervention. This change was due to the removal of larger positive results, which strengthened the robustness of the study’s conclusions. DCHI indicates the effectiveness of herpes skin lesions treatment. A study have shown that skin lesions of HZ are the most influential factors affecting the duration of acute herpetic pain.^[92] This NMA results indicated that EA + PC significantly outperformed the other interventions in the probability ranking results. However, no significant statistical difference was found between MA and pharmacotherapy, or between MA + PC and pharmacotherapy. ER comprehensively assesses the effect of various interventions with information about pain, skin lesions, and other factors.^[93] The probability ranking results shown that EA + FN was the most effective intervention, while MA + FN was not statistically superior to pharmacotherapy according to the NMA results. PHN is the most common complication of herpes zoster, even 80% of cases occur in patients 50 years or older.^[6] According to the probability ranking results, PC had the best effect in preventing PHN. Additionally, the NMA results shown no statistical improvements with the use of MA or MA + FN compared to using pharmacotherapy alone in the prevention of PHN. Moreover, 23 trials reported adverse events, among them, acupuncture often resulted in fainting and hematoma, while pharmacotherapy usually leaded to dizziness, nausea, and abdominal discomfort. On the whole, fewer adverse events were associated with acupuncture therapy compared to pharmacotherapy, and no serious adverse events were reported. In clinical practice, those adverse reactions produced by acupuncture-related therapies should be acceptable.

According to the results of this NMA, the top rankings of therapies include EA, PC, and FN. The precise mechanism by which acupuncture relieves the symptoms of herpes zoster is not fully understood. Modern scientific research has found that acupuncture could recruit immune cells through the activation of the NF-κB pathway.^[94] Additionally, acupuncture could stimulate the release of endogenous opioids and other neurotransmitters such as serotonin and dopamine,^[95] which have analgesic and anti-inflammatory effects. EA, PC, and FN are further developments and additions to acupuncture therapy. Electroacupuncture could activate a variety of bioactive chemicals and desensitized peripheral nociceptors,^[96] thereby blocking pain caused by acute HZ.^[97] PC has been found to be effective in reducing serum substance P (SP) levels and relieving symptoms in patients with HZ at the acute stage.^[98] FN is capable of alleviating the pain of acute HZ probably by inhibiting the release of inflammatory cytokines and reducing the level of pain mediators such as SP.^[99]

This study found that the acupoints most commonly selected were Ashi points and Jiaji (EX-B2), which are often located close to the regions of skin lesions. Recent scientific research has suggested that electroacupuncture at Ashi points could attenuate local neuropeptide levels, leading to a reduction in the inflammatory response.^[100] Furthermore, stimulating EX-B2 could alleviate hyperalgesia by reducing the local secretion of prostaglandin E2 in rats.^[101]

This study has several limitations. First, the search language was restricted to Chinese and English, which might cause some bias. Second, some lower-quality studies could be included in this NMA research, which might impact the accuracy and reliability of the conclusions. Third, in the pharmacotherapy category, this NMA did not perform a separate analysis of the type and dose of medications, and issues might be the focus of future studies. Fourth, this study evaluated the efficacy of different acupuncture-related therapies in the treatment of HZ, however, due to insufficient reporting on certain therapies, such as thread moxibustion and cotton-sheet moxibustion, it was not possible to compare the efficacy of all therapies. Consequently, these therapies were included in the OAT category for analysis. Fifth, inconsistency investigations indicated no significant global inconsistency, but considerable local inconsistencies were presented. These inconsistencies might arise from factors such as study populations, interventions, or study designs. Sixth, there had some publication bias and small sample effects in this NMA study, which could result in unreliable conclusions. A sensitive analysis was conducted, the results shown some differences from the main NMA analysis in probability ranking results, which are mainly manifested as a decline in the rankings of excluded therapies, and leaded to more robust results.

5. Conclusions

In conclusion, it can be speculated from the NMA analysis results, EA + PC might be the best therapy for alleviating pain symptoms and skin lesions in patients with acute HZ. The results of ER indicated that EA + FN might be the most efficacious therapy. And PC had some advantages in reducing the incidence of PHN. In clinical practice, the selection of the appropriate therapy should consider practical circumstances. However, this NMA study could be affected by bias due to the limited number and quality of available literature. Therefore, more multi-center, large-sample, prospective RCTs studies are needed to validate conclusions.

Acknowledgments

All authors of this article would like to express their sincere gratitude to the Chengdu University of Traditional Chinese Medicine.

Author contributions

Conceptualization: XingYu Liang, Xi Chen.

Data curation: Sha Yang, Hong Zhang.

Formal analysis: SiYing Wang, MingMing Guo.

Investigation: SiYing Wang, DanDan Ma.

Methodology: XingYu Liang, XueMei Li.

Software: XingYu Liang, Xi Chen.

Supervision: Sha Yang, Hong Zhang.

Writing – original draft: XingYu Liang, Xi Chen.

Writing – review & editing: XingYu Liang, XueMei Li.

Supplementary Material

medi-103-e38006-s001.docx ^{(32.3KB, docx)}

medi-103-e38006-s002.docx ^{(22.9KB, docx)}

medi-103-e38006-s003.docx ^{(22.9KB, docx)}

medi-103-e38006-s004.docx ^{(41.2KB, docx)}

medi-103-e38006-s005.docx ^{(1.6MB, docx)}

medi-103-e38006-s006.docx ^{(29.5KB, docx)}

medi-103-e38006-s007.docx ^{(670.6KB, docx)}

medi-103-e38006-s008.docx ^{(296.4KB, docx)}

medi-103-e38006-s009.docx ^{(17.4KB, docx)}