Efficacy comparison of different acupuncture methods for herpes zoster: a systematic review and Bayesian analysis

Hua-Chong Xu,Ya-Wen Jiang,Yu-Cong Shi,Pei Liu,Li Deng*

1College of Traditional Chinese Medicine,Jinan University,Guangzhou 510632,China.

#Hua-Chong Xu,Ya-Wen Jiang and Yu-Cong Shi contributed equally to this work.

Abstract Objective: Acupuncture methods (including moxibustion) are used frequently in the treatment of herpes zoster. However, the choice is usually made only based on personal experience among different acupuncture methods. This study compared the effectiveness of different acupuncture methods for herpes zoster. Methods: All randomized controlled trials(RCTs) of acupuncture methods for herpes zoster were searched in seven databases including Cochrane Library, Embase, PubMed, Web of Science, Wan‐fang,CNKI, and CQVIP database. After screening process, effectiveness rate was extracted from all the included RCTs as primary outcomes. The Bayesian network meta‐analysis was conducted by GeMTC 0.14.3, Stata13.0 and Review Man 5.3. Results: 39 studies were included, which contained 3,042 participants among 11 interventions. Based on the results of network meta‐analysis and ranking probability, fire‐acupuncture plus electro‐acupuncture is considered to be the most effective method, followed by body‐acupuncture plus moxibustion, fire‐acupuncture,surround‐acupuncture plus moxibustion, moxibustion, surround‐acupuncture plus western medicine, surround‐acupuncture plus electro‐acupuncture, body‐acupuncture plus western medicine, surround‐acupuncture, western medicine, body‐acupuncture. Global and local inconsistency test suggested no significant difference between the results of direct and indirect comparisons. Conclusion: Acupuncture methods might be an effective alternative treatment for herpes zoster and fire‐acupuncture plus electro‐acupuncture might be considered the best option among the included treatments. However, the results of this study need to be interpreted with caution, because there may still be a problem of small sample size of some studies and interventions. Future research, with a standard methodology and design, requires large‐scale trials to validate the effect identified in this meta‐analysis.

Keywords:acupuncture; moxibustion; herpes zoster; Bayesian analysis

Introduction

Herpes zoster (HZ) is a neuro‐cutaneous disease, usually resulting from the reactivation of varicella‐zoster virus which remains dormant in the sensory ganglia following primary infection(chickenpox)[1,2].Herpes zoster causes the characteristic vesicular skin rash localized in the sensory region of the affected ganglia, and is often accompanied by severe pain, pustular skin lesions, or itching [3]. Approximately 25% of the world’s population have the risk of developing herpes zoster during their lifetime, and epidemiological data from North America, Europe and Asia Pacific indicate a lifetime incidence of HZ of 3.0‐5.0/1000 person‐years, with a similar prevalence in China,while the incidence of herpes zoster has been shown to be strongly related to age [4‐6]. Two‐thirds of all shingles patients are over the age of 50 due to decreasing varicella zoster virus specific cell‐mediated immunity with advancing age [7]. As a common disease, herpes zoster has a great impact on the quality of life especially due to the pain during the acute phase and post‐herpetic neuralgia (PHN) [8‐10]. Most of the patients suffer a great psychological and economic burden [10].

The aim of herpes zoster treatment is to relieve pain in the acute phase, to limit the spread and duration of skin lesions and to prevent or alleviate PHN and other acute and chronic complications.Treatments for herpes zoster include local therapy and systemic therapy. Systemic therapy is more commonly used with antiviral therapy, corticosteroid therapy, neuralgia therapy and neurotrophic therapy because local therapy have limited effect [3]. Currently,medications for antiviral therapy,including acyclovir,valaciclovir and famciclovir, can alleviate the symptoms and reduce healing time [3,11]. But Long‐term excessive use of antiviral drugs often cause some adverse reactions such as hemolytic anemia, thrombocytopenia, or renal insufficiency. Additionally, it is hard to receive a satisfactory effect for the elderly with reduced renal function or renal insufficiency patients due to the limited dosage and cumulative toxicity.

As a safe and natural therapy, acupuncture and moxibustion treatments are considered as an effective alternative treatment and recommended in Chinese medicine clinical practice guidelines for treating the acute symptoms of herpes zoster in China [12]. Some clinical studies [13, 14] and systematic analysis [15‐18] have shown that acupuncture treatments,alone or combined with drugs,may have a better therapeutic effect than simple drug therapy in pain reduction and skin conditions of herpes zoster. There are various acupuncture treatments used for herpes zoster in the clinic, including fire‐acupuncture, surround‐acupuncture, moxibustion, body acupuncture, electro‐acupuncture, combination treatments and so on.However, the previous studies [15‐18] only explored the effectiveness between single acupuncture treatment and a control intervention or only studied its effectiveness considering all the acupuncture treatments as a whole. As lacking of authoritative evidence and guidelines, the choice of different acupuncture therapies is usually made basing on the experience of clinical doctors, for which may lead to an unsatisfactory effect and longer course of treatment.Overall, the comparison of the efficacy among various acupuncture therapies is urgently need a detailed and in‐depth exploration.

Compared to previous meta‐analysis, recent researches have been incorporated into a network meta‐analysis to perform a critical evaluation for the commonly used methods of acupuncture and moxibustion in herpes zoster. Our review aimed to compare the effectiveness of different acupuncture treatments for herpes zoster and find a better selection guideline for clinicians and patients.

Methods

We followed the methods of a published articles [19] and our published protocol of this study[20].PROSPERO registration number:CRD42020175189.

Search strategy

To identify eligible RCTs for this network meta‐analysis, we searched seven electronic databases systematically from inception to January 30, 2020, including Cochrane Library, Embase, PubMed, Web of Science, Wan‐fang database, CNKI (China National Knowledge Infrastructure database), and CQVIP (VIP Chinese Science and Technique Journals database). The following search terms were conducted: (“Acupuncture” OR “Acupuncture Therapy” OR“Moxibustion”) AND (“Herpes Zoster” OR “Shingles” OR “Herpes zoster Virus” OR “Varicella Zoster Virus Infection”) AND(“Randomized Controlled Trial” OR “Controlled Clinical Trial” OR“Randomized” OR “Controlled”). The search strategy for Embase is shown in Supplementary Materials Appendix 1 and each database was adopted to its own unique characteristics. The languages were restricted to English and Chinese.

Inclusion criteria

Types of trials.Clinical randomized controlled trials were selected,which were published in the journal with English and Chinese.

Participants.Participants should have a primary diagnosis of herpes zoster according to diagnostic criteria or clinical diagnosis, regardless of age, gender and disease duration. The main diagnostic criteria for herpes zoster include the following [21]: Standards for Diagnosis and Curative Effect of Chinese Medical Symptom; Dermatovenerology;Modern dermatology; Cecil’s medicine; Internal medicine; Chinese clinical dermatology; Diagnosis of conventional.

The detailed diagnostic standards are as follows [21]: (1) Skin lesions are mostly blisters the size of mung beans, which are clustered with relatively tense blisters with red basal base and usually unilateral distribution and arranged into ribbons.In severe cases,the lesion may be hemorrhagic or gangrenous; (2) The rash is usually preceded by a tingling or burning sensation of the skin, which may be accompanied by mild discomfort and fever throughout the body;(3) Conscious pain is obvious, but there may be unbearable severe pain or pain left after the rash subsides.

Interventions and comparison.The interventions for the test group must be one of acupuncture methods (include acupuncture and moxibustion), or a combination treatment of any two methods, or a combination treatment of one acupuncture method and western medicine.

The interventions for the control group were western medicine or a treatment above differ with the test group. Additionally, western medicine treatment must be antiviral therapy in guideline [3], alone or in combination with other treatments such as corticosteroid therapy, neuralgia therapy and neurotrophic therapy. Drugs for antiviral therapy include acyclovir, valaciclovir and famciclovir.Studies with three or more intervention methods will not be included in order to eliminate the interference of multiple factors.

Outcomes.The primary outcome was effectiveness rate. Clinically,the effectiveness rate of acupuncture and moxibustion (including the objective ratio of skin lesion area reduction and subjective pain scale)is a commonly used and widely recognized outcome indicator for the efficacy of acupuncture and moxibustion in treating herpes zoster.Effectiveness rate was mainly reported by measuring symptom improvement according to Standards for Diagnosis and Curative Effect of Chinese Medical Symptom. The effectiveness rate was calculated as the ratio of the number of patients treated effectively to the total number of patients.Effective treatment includes those who have cured or shown improvement.

According to Standards for Diagnosis and Curative Effect of Chinese Medical Symptoms, evaluation criteria are defined as three levels[21], (1) Cured: the rash disappeared, clinical symptoms disappeared,no pain sequela; (2) Improvement: the rash subsided by more than 30%, and the pain was significantly reduced; (3) Inefficacious: the rash subsided by less than 30% and pain was not relieved.

Exclusion criteria

Studies with the following conditions will be excluded [19]: (1)Republicated studies; (2) Studies without outcomes for effectiveness rate; (3) Patients with serious complications; (4) Trials with interventions of cupping or bleeding; (5) Trials with interventions of Chinese herbal medicine or proprietary Chinese medicine.

Study selection and data extraction

Endnote X9 and Excel 2018 were used for study selection and data extraction. Firstly, two independent investigators screened titles and abstracts after removing duplicate studies in Endnote. Secondly, they read the full‐text of relevant studies after titles‐abstracts screen,according to inclusion and exclusion criteria. Any discrepancies were resolved by the third investigators. Finally, included studies were coded and extracted the relevant information [19]: study characteristics (author, publication time); participant characteristics(diagnose criteria, age, disease course, cases); intervention information (intervention detail, treatment duration, follow‐up,adverse events) and treatment outcome.

Study quality evaluation

According to Cochrane risk of bias assessment tool[19,22],the risk of bias is assessed by two investigators independently using Review Man 5.3. The following aspects were evaluated [22]: (1) Random sequence generation; (2) Allocation concealment; (3) Blinding of participants and personnel; (4) Blinding of outcome assessment; (5) Incomplete outcome data; (6) Selective reporting; (7) other bias. Any discrepancies would be decided by other investigators within the review team.

Statistical analysis

Firstly, inconsistency test and network graphs were conducted using Stata 13.0 with the network and network graphs packages(Description details were shown in Supplementary Materials Appendix 2.Inconsistency test was performed at two levels:global inconsistency and local inconsistency. Global inconsistency test reflects the fit and parsimony of consistency and inconsistency models for this study.Local inconsistency test calculates the difference between direct and indirect estimates in all closed loops in the network.

Secondly, to increase the stability of network meta‐analysis, we chose a random‐effects model using a Bayesian framework in GeMTC 0.14.3 (Generate Mixed Treatment Comparisons) with MCMC(Markov Chain Monte Carlo) and further analysis with Stata 13.0[19]. It estimated the posterior probability according to the prior probability[19].When the MCMC reached a stable convergence state,estimations and inferences would be conducted with parameters as follows [19, 23, 24]: the initial value is set to 2.5; the number of simulation iterations is set to 50,000, and 20,000 adjustment iterations are performed first to eliminate the influence of the initial value; the step size (sparse interval) is set to 10 when the number of chains is 4. Brooks‐Gelman‐Rubin diagnosis plot and the potential scale reduced factor (PSRF) reflect the convergence of the model [19,24]. The calculation method of PSRF is to analyze the results of the last iteration of all MCMC chains, and to pair variables in the chain with those between the chains. The PSRF value tends to 1, indicating that the model converges satisfactorily. Brooks and Gelman pointed out that a PSRF value below 1.2 is acceptable [24].

Finally, the figure of ranking probability was made for all interventions and the node‐splitting method was used to evaluate local inconsistency [19], which separated evidence on a particular comparison into direct and indirect evidence.The odds ratio(OR)was calculated for dichotomous outcomes (effectiveness rate) with 95%confidence interval (CI). ThePvalues was evaluated for inconsistency test, andP< 0.05 was regarded as statistically significant [19]. We detect publication bias using the funnel plot and Egger’s test.

Results

Literature search results

Overall, 1,676 citations were identified through database search. 321 potentially eligible studies were retrieved in full text after duplicates removed and titles‐abstracts screened. Based on full‐text screening,282 studies were excluded with the following reasons: Not meet the inclusion of interventions (96); No relevant outcome (182);Republication (4). Finally, 39 RCTs were selected for this study include 1 RCT [25] in English and 38 RCTs in Chinese. The selection process is shown in Figure 1.

Figure 1 selection process for literatures

Study characteristics

39 included RCTs were published between 2003 and 2019 containing a total of 3,042 participants, of which 1,557 patients were in the test group and 1,485 patients in the control group. The mean sample size of studies was 78 participants,ranging between 24 and 97.There were 9 interventions in the test group as follows: FA, Fire‐acupuncture (6 trials); SA, Surround‐acupuncture (2 trials); MOX, Moxibustion (7 trials); BA+WM, Body acupuncture + Western medicine (3 trials);BA+MOX, Body acupuncture + Moxibustion (5 trials); FA+EA,Fire‐acupuncture + Electro‐acupuncture (4 trials); SA+EA,Surround‐acupuncture + Electro‐acupuncture (3 trials); SA+MOX,Surround‐acupuncture + Moxibustion (5 trials); SA+WM,Surround‐acupuncture + Western medicine (4 trials). In addition, the control group of 26 trials were treated with Western drugs, and the other 13 trials were treated with Body acupuncture (5 trials),Surround‐acupuncture + Electro‐acupuncture (5 trials) and Surround‐acupuncture (3 trials). The treatment duration ranged from 5 days to 30 days, but most are 10 days. Drug therapy of 26 trials all included antiviral drugs, but only one used drugs at doses recommended in clinical guidelines [26]. 3 trials used local antiviral therapy, which is not in clinical guidelines [27‐29]. Vitamin B1 and B12(or mecobalamin)are commonly used for nerve repair in China as 17 trials used. Neuralgia therapy often used non‐steroidal analgesics[30‐32] (ibuprofen and indomethacin) and additional low‐potency opioid analgesics [33] (carbamazepine). Corticosteroid therapy of prednisolone only used in one trail [33].

All included trials have clear outcome criteria and diagnostic criteria except one trail [34]. Characteristics of included studies are shown in Table 1.

Table 1 Characteristics of included studies

Table 1 Characteristics of included studies(Continued)

Adverse events and follow-up

6 trials observed adverse events[28,30,40,45,52,63].3 trials found no adverse reactions when the other 3 reported several adverse events. One trial detailed the adverse events that the test group and the control group had 5 and 12 patients with high fever and gastrointestinal discomfort, respectively. Some patients in both two groups experienced mild gastrointestinal discomfort and increased blood pressure, but there was no significant difference. The other trial reported 2 patients in the control group experienced discomfort but lacked a detailed description. All adverse events mentioned above were resolved by acupuncture adjustment,drug adjustment and verbal communication. No patient was reported to withdraw due to adverse events.

Follow‐ups were adopted in only 5 trials [30, 41, 47, 55, 57] for lasting effect and the incidence of postherpetic neuralgia ranging between 1 month to 3 months. one study suggested that the recurrence rate of the control group was higher[41].The remaining 4 studies showed the incidence of postherpetic neuralgia in the control group was higher than that in the test group, and the difference was statistically significant.

Risk of bias for research quality evaluation

The assessment results showed a medium quality for included studies.38 studies reported randomization, but only 12 studies described detailed and reliable random grouping methods, 11 of which used random number tables and one [45] used coin random method. Both concealment of allocation and blinding of outcome were not mentioned in all studies except one [40]. Blinding was not used in all studies because of the special nature of acupuncture treatment.Hence,a high risk of performance bias probably existed in literature quality assessment. All studies had complete outcome measures, and no studies had a high risk of reporting bias or other bias. The quality evaluation details are shown in the Figure 2 and Figure 3.

Figure 2 Risk of bias graph:review authors’judgements about each risk of bias item presented as percentages across all included studies

Figure 3 Risk of bias of each included study

Network plot for interventions

Network plot for 11 interventions was conducted by Stata 13.0(Figure 4). The width of the lines is proportional to the number of trials comparing every pair of treatments, and the size of every circle is proportional to the number of participants of each intervention. The result shows that the sample of Western medicine, Moxibustion and Surround‐acupuncture + Electro‐acupuncture ranked in the top three interventions in the study. Except for Fire‐acupuncture +Electro‐acupuncture and Surround‐acupuncture + Western medicine,other interventions have two or more direct comparisons. Network meta‐analysis was performed to combine direct comparison with indirect comparison.

Figure 4 Network plot of included interventions for efficacy The width of the lines is proportional to the number of trials comparing every pair of treatments, and the size of every circle is proportional to the number of participants of each intervention.

Results of network meta-analysis

Both global inconsistency test(P=0.578>0.05,details were shown in Supplementary Materials Appendix 3 and local inconsistency suggested that data showed no significant inconsistency. Local inconsistency test was shown in loop inconsistency map: all the 95%CI include zero and all the IF are close to zero (Figure 5), which represents no significant inconsistency between direct and indirect effects.

Figure 5 The result of loops inconsistency test The 95% CI include zero and IF is close to zero, represent no significant inconsistency between direct and indirect effects.(A:Western medicine;B:Body acupuncture;C:Fire‐acupuncture;D:Surround‐acupuncture;E:Moxibustion;F:Body‐acupuncture+Western medicine;G:Body‐acupuncture + Moxibustion; H: Fire‐acupuncture + Electro‐acupuncture; I: Surround‐acupuncture + Electro‐acupuncture; J: Surround‐acupuncture +Moxibustion;K: Surround‐acupuncture +Western medicine.

A random‐effects Bayesian framework was selected for the network meta‐analysis,and the results was shown in Table 2.Firstly,compared with western medicine, the following interventions can significantly improve the effectiveness for herpes zoster treatment: SA+MOX[OR=9.32,95%CI(4.14,30.26),P<0.05],SA+WM[OR=7.24,95%CI(2.88,19.83),P<0.05], FA[OR =13.39, 95%CI(4.17, 61.95),P< 0.05], MOX [OR = 7.43, 95% CI (3.46, 17.88),P< 0.05], BA +WM[OR=4.11,95%CI(1.59,26.01),P<0.05],BA+MOX[OR=20.17,95%CI(4.18,367.21),P<0.05],FA+EA[OR=43.61,95%CI (14.46, 155.02),P< 0.05] and SA + EA [OR = 5.72, 95% CI(2.48,13.57),P<0.05].Secondly,compared with body acupuncture,the following methods showed significantly improvement: SA +MOX[OR = 27.27, 95% CI (4.98, 100.29),P< 0.05], SA + WM [OR =17.85, 95% CI (2.70, 83.47),P< 0.05], FA [OR = 41.02, 95% CI(3.88,223.81),P<0.05],SA[OR=5.77,95% CI(1.36,29.93),P<0.05], MOX [OR = 19.46, 95% CI (3.42, 80.41),P< 0.05], BA +WM [OR = 12.58, 95% CI (2.76, 53.07),P< 0.05], BA + MOX [OR=64.91, 95% CI(5.31, 898.10),P< 0.05], FA+EA [OR= 117.61,95% CI (15.37, 624.34),P< 0.05] and SA + EA [OR = 14.34, 95%CI(2.27, 60.33),P< 0.05]. Thirdly, the following interventions were significantly less effective than Fire‐acupuncture +Electro‐acupuncture:SA+WM[OR=0.16,95%CI(0.04,0.59),P<0.05], SA [OR = 0.05, 95% CI (0.01, 0.30,P< 0.05], MOX [OR =0.17,95%CI(0.04,0.61),P<0.05]and BA+WM[OR=0.09,95%CI (0.02, 0.81),P< 0.05]. Fourth, compared with Surround‐acupuncture + Moxibustion, Surround‐acupuncture is less effective with OR = 0.21 and 95% CI (0.07, 0.91). The remaining paired comparisons showed no significant differences in effectiveness rate.

Table 2 Network meta-analysis comparisons

Comparison of the effectiveness of different interventions

Figure 6 shows the ranking probability basing on the MCMC theory for probability evaluation. In the 11 treatment interventions for herpes zoster in this study, Fire‐acupuncture + Electro‐acupuncture is considered to be the most effective method, followed by Body‐acupuncture + Moxibustion, Fire‐acupuncture,Surround‐acupuncture + Moxibustion, Moxibustion,Surround‐acupuncture+Western medicine,Surround‐acupuncture+Electro‐acupuncture, Body‐acupuncture + Western medicine,Surround‐acupuncture, Western medicine, Body‐acupuncture.

Figure 6 The rank probability of efficacy for included interventions 1: Western medicine; 2: Body acupuncture; 3: Fire‐acupuncture; 4:Surround‐acupuncture; 5: Moxibustion; 6: Body‐acupuncture + Western medicine; 7: Body‐acupuncture + Moxibustion; 8: Fire‐acupuncture + Electro‐acupuncture; 9:Surround‐acupuncture +Electro‐acupuncture;10: Surround‐acupuncture +Moxibustion; 11: Surround‐acupuncture +Western medicine.

According to the Brooks‐Gelman‐Rubin diagnosis plot, the median value of the reduction factor and 97.5% tend to be stable after 25,000 iterations, and then Bayesian model was calculated to 25,000 iterations. The PSRF value tends to 1, indicating that the model converges satisfactorily (Table 3). Additionally, the local inconsistency was also performed using the node‐splitting method.Except for the comparison between Western medicine and moxibustion, the P values of the other comparison groups were greater than 0.05, indicating that the direct comparisons were consistent with the indirect comparisons(Table 4).

Table 3 The PSRF value

Table 4 Node-splitting test result

Publication bias

According to the funnel plot (Figure 7), we found no obvious publication bias. Furthermore, we conducted verification by Egger’s test (Figure 8) and found no publication bias (P =0.206 > 0.05).More details of publication bias were shown in Supplementary Materials Appendix 4.

Figure 7 Comparison-adjusted funnel plot for the network meta-analysis

Figure 8 Egger's publication bias plot, (P= 0.206 >0.05)

Discussion

This Bayesian network meta‐analysis represents the most comprehensive synthesis of data for currently acupuncture and moxibustion treatments for herpes zoster. It incorporated the direct comparison and indirect comparison of various acupuncture therapies.The review found that acupuncture and moxibustion treatments are considered to have benefit on pain relief, rash healing and reduction of the incidence of PHN. Compared to other interventions,fire‐acupuncture + electro‐acupuncture seems to be more effective.Except for body acupuncture and surround‐acupuncture, other acupuncture and combined therapies are significantly better than drug therapy. The results of network meta‐analysis and the ranking probability showed the specific information above (Figure 6).However, these findings should be interpreted with caution clinically,as most of the included studies were of low to medium quality. Low quality methodology, literature bias risk and small sample size of some intervention nodes are important factors affecting the reliability of clinical interpretation of the results of this study.

In traditional Chinese medicine, shingles is called “snake string ulcer”, which is caused by accumulation of pathogenic toxin,emotional paralysis, and poor circulation of qi and blood in the viscera. Acupuncture and moxibustion therapies are often considered to have the function of clearing local qi and blood,activating channels and collaterals, and enhancing the ability to resist pathogenic factors,so as to prevent the spread of the virus and alleviate neuralgia.However, the modern biological mechanism of acupuncture and moxibustion treatment of herpes zoster is not fully revealed. Studies[2, 12] have shown that acupuncture and moxibustion treatments are important analgesic and anti‐inflammatory therapies in the acute stage of herpes zoster. On the one hand, the neurochemical basis of acupuncture analgesia has been substantially investigated by basic research. Electro‐acupuncture has agonist effect at all types of opioid receptor both central and spinal level, with stimulating release of all types of opioid peptide such as methionine enkephalin, the dynorphins, β‐endorphin, and the endorphins [64‐66]. Animal experiments indicated acupuncture can accelerate the synthesis and release of both serotonin and noradrenaline which have analgesic action via the descending inhibitory spinal pathway of the dorsolateral funiculus [67]. Fire acupuncture can protect central and peripheral nerves [68], significantly improve the expression of apoptosis‐related proteins in rats with spinal cord injury, and reduce the occurrence of apoptosis[69].Fire acupuncture relieves pain better than other acupuncture treatments because its intensity of stimulation in the cerebral cortex is much higher than that of other acupuncture treatments [70]. On the other hand, the antiviral and anti‐inflammatory effects of acupuncture methods had been researched but it remains uncertain. Fire acupuncture can reduce the expression of inflammatory cytokines interleukin 1 (IL‐1) and cyclooxygenase‐2 (COX‐2) in the spinal cord of sciatic nerve injury model rats [71] and electroacupuncture may alleviate inflammation.Acupuncture treatments can accelerate local blood circulation,lymphatic circulation, promote the absorption of inflammatory exudate, relax spasmodic muscles, and be beneficial to tissue repair[72]. It is consistent with existing basic research which the combination of fire and electric acupuncture is considered to be more effective than other therapies. Generally, the treatments associated with fire acupuncture or moxibustion seem to work well for herpes zoster.

In order to improve the quality of this study, we have made strict design in literature selection, software and method selection. In the inclusion criteria, drugs of control group were designated as standard western medicine to improve the homogeneity of the research results,which did not include Chinese herbal medicine with unclear mechanism of action. What’s more, studies of three or more intervention methods will not be included in order to eliminate the interference of multiple factors. A combination of more than three interventions is sometimes considered to be effective enough for further research, although its mechanism is complex and difficult to analyze. To obtained a more credible result, we prefer GeMTC software for the final simulation iteration with more strictive and stable iterative conditions. Actually, both the PSRF (Table 3) and the results of local inconsistency test (Figure 5 and Table 4) had proved that an appropriate model has been successfully established for our network meta‐analysis.

Our study has some limitations as follows. Firstly, Vitamin B1 and B12 (or mecobalamin) are commonly used for nerve repair in china,but it is unclear whether the addition of neurotrophic drugs affects the efficacy of valacyclovir. Secondly, acupoint selection between different therapies may be inconsistent and empirical due to the preference and experience of acupuncturists, which may increase the heterogeneity of the study. Most of the studies chose Ashi point and Jiaji point (EX‐B2) as the main points (Table 1), while other points were selected dialectically. What’s more,most included studies lack of standard follow‐up and adverse events report, which still not enough to indicate the long‐term effect of acupuncture. Only 6 trials observed adverse events while follow‐up were adopted to only 5 trials. In particular, the incidence of postherpetic neuralgia was considered to be different between the treatment group and the control group in 5 studies [30, 41, 47, 55, 57], but only one was followed for three months [41], consistent with the diagnosis of postherpetic neuralgia.Finally, there may still be a problem of small sample size of some studies and interventions although the funnel plot and Egger’s test results indicate that there is no significant publication bias among all studies.

Overall, although the results of this study need to be interpreted with caution, acupuncture methods would be effective alternative treatments, especially for the elderly with reduced renal function or renal insufficiency patients due to the limited dosage and cumulative toxicity. Future research, with a standard methodology and design，requires large‐scale trials to validate the effect identified in this meta‐analysis.

Conclusions

Notwithstanding these caveats, acupuncture methods might be an effective alternative treatment for herpes zoster and fire‐acupuncture+electro‐acupuncture might be considered the best option among the included treatments when acupuncture treatments is necessary.However,the results of this study need to be interpreted with caution,because there may still be a problem of small sample size of some studies and interventions. Future research, with a standard methodology and design，requires large‐scale trials to validate the effect identified in this meta‐analysis.

Data Availability

The data used to support the findings of this study are included within the supplementary appendix and more detailed information are available from the corresponding author upon request:Appendix 1.The search strategy for Embase;Appendix 2.Description for Stata 13.0;Appendix 3.Assessment of Inconsistency results;Appendix 4.Tests for Publication Bias;Appendix 5.Others Information.