Efficacy of High-Voltage Pulsed Radiofrequency in Zoster-Associated Pain: A Meta-Analysis and Systematic Review

Background Zoster-associated pain (ZAP) is often refractory to conventional treatments and can seriously affect patients' physical and mental health. High-voltage pulsed radio frequency (H-PRF) is a new method for treating ZAP with pulse voltages above 60 V or even up to 100 V. The purpose of this paper was to conduct a systematic review and meta-analysis to evaluate the efficacy of H-PRF in the management of ZAP. Methods PubMed, Embase, and the Cochrane library were searched from their inception to June 2022 to identify controlled trials which evaluated the effectiveness of H-PRF compared with standard PRF and sham operations. The primary outcome was pain scores at different treatment times. The secondary outcomes included SF-36 scores, rescue analgesic dose, and side effects. Results We reviewed 6 randomized controlled trials involving 428 patients. There was no significant difference between the H-PRF and standard PRF pain scores at 1 week after surgery and the sham operation group at 1 month. At 1, 3, and 6 months, the H-PRF group had better pain score than the standard PRF group, and at 3 months, the pain score was better than the sham operation group. The H-PRF group showed improvement in the SF-36 score, and there were no significant complications in the H-PRF group. Conclusions H-PRF is an effective and safe treatment method that has better effects in relieving pain and improving the quality of life and physical and mental health. Although H-PRF provides pain relief rates comparable to those of the control group in the early stages, it remains the preferred and alternative treatment for relieving herpes zoster-related pain.


Introduction
Zoster-associated pain (ZAP) is a neuropathic pain characterized by hyperalgesia and paresthesia, including herpetic pain and postherpetic neuralgia (PHN).It is an intractable, persistent, severe pain along the innervated skin area that occurs due to herpes zoster, which is troublesome to treat [1,2].In the United States, the overall incidence rate of PHN was 57.5 cases per 100,000 person-years, and the annual growth rate is 3.1% [3].Te prevalence of herpes zoster (HZ) and PHN in China is 7.7% and 2.3%, respectively, and about 29.8% of HZ patients will develop PHN, whose incidence increases with age, especially in elderly patients over 70 years [4].Te acute phase of herpes zoster usually occurs within one month after the eruption.Postherpetic neuralgia (PHN) is a kind of neuralgia that lasts one month after the explosion [5,6].ZAP is characterized by acupuncture, burning like an electric shock and severe pain, which leads to sleep disorders, anxiety, and depression and seriously afects patients' quality of life [7].
ZAP is a kind of refractory neuropathic pain.Te commonly used treatment methods include oral drug therapy, nerve injection therapy, continuous epidural analgesia, etc., but these treatment methods are not efective enough for ZAP [8][9][10][11].Pulsed radiofrequency (PRF) is an improvement to thermal radiofrequency therapy, in which a short (20 ms) burst of high-voltage current is followed by a silent phase (480 ms) to achieve heat transfer and maintain the target tissue temperature at 42 °C.PRF technology is widely used in treating various neuropathic pains, especially in ZAP [12,13].Trough image guidance, radiofrequency electrodes are punctured into the target nerve's dorsal root ganglion for treatment.Te standard PRF (S-PRF) output voltage is 45 V, and its painrelieving efect is controversial [14].Some studies have shown that its analgesic efect is signifcant [15], but some believe its efect is limited and the recurrence rate is high [16].Terefore, there is now a high-voltage pulsed radio frequency (H-PRF) modality that has been shown in some studies to provide better results [17].Teixeira and Sluijter frst reported that a highvoltage PRF ablation of 60 V used to treat patients with discogenic pain attained satisfactory efcacy that lasted over 3 months [18].In 2013, Luo et al. found that there was a signifcant negative correlation between postoperative Numerical Rating Scale (NRS) scores and the output voltage of PRF in the treatment of idiopathic trigeminal neuralgia [19].Subsequently, more and more studies have been conducted to treat ZAP by increasing the voltage value of pulsed radio frequency [20][21][22].A randomized controlled trial (RCT) conducted by Wan et al. showed that no signifcant diference in improvement in pain scores was observed 1 month after treatment, and at 3 and 6 months, the high voltage group scored signifcantly lower than the standard voltage group; however, the incidence of ecchymosis in the high voltage group (19.2%) was higher than in the standard voltage group (12.1%) [23].Terefore, further analysis is needed to determine whether H-PRF ablation is better than S-PRF ablation at diferent time points and whether H-PRF ablation is a safe treatment for ZAP.Te purpose of this study was to compare the efectiveness of H-PRF in controlling ZAP, in order to better guide clinical work and improve the treatment efect of herpes zoster.

Data Sources and Search Strategies.
Two researchers independently conducted comprehensive searches of PubMed, Embase, and the Cochrane library from their inception to June 2022.We used the following medical subject terms and keywords for our search: "postherpetic neuralgia," "PHN," "herpes zoster," "herpetic neuralgia," "zoster-associated neuralgia," "pulsed radiofrequency," "PRF," "High-Voltage," and "Voltage."We use the "related articles" function in PubMed to expand the search scope and summarize all retrieved abstracts, studies, and citations.Tere were no language restrictions in this search.Te search strategy is in supplementary material (available here).We preregistered the protocol of the systematic review on PROSPERO (CRD42022348310).Abbreviations can be found in Table 1.

Inclusion and Exclusion Criteria.
Studies that met the following selection criteria were included in the metaanalysis: (1) Te study was a randomized controlled trial (RCT).(2) Te study group included an H-PRF and a control group, while the control group included standard PRF and sham operation.(3) Te study subjects were adults with acute herpes, subacut herpes, and PHN, collectively referred to as ZAP.( 4) Tey reported at least one outcome indicator of the visual analog scale (VAS), Numerical Rating Scale (NRS), 36-Item Short-Form Health Survey (SF-36), rescue analgesic dose, and side efects.
Te exclusion criteria were as follows: (1) patients with sequelae of cerebral thrombosis, severe cardiopulmonary disease, serious liver and kidney dysfunction, and other serious systemic diseases; (2) reviews, animal studies, and case reports; (3) the data from the study were not suitable for statistical analysis; and (4) patients with other neuropathies or neurological diseases (e.g., diabetes mellitus neuropathy, intercostal neuralgia, and primary trigeminal neuralgia).

Data Extraction and Quality Assessment.
Two reviewers independently extracted baseline data and result data.We extracted the following data from the included studies: the name of the frst or corresponding author, the year of publication, the type of study, the sample size, the last follow-up time, the age and sex of the study population, the location of herpes, RF parameters, and outcome indicators.Te quality of RCTs was assessed using the bias risk assessment tool of Cochrane Collaboration.Te judgments of bias were reported as "low risk," "high risk," or "unclear risk" according to their methodological section.

Statistical Analysis.
Tis study used RevMan5.3(Nordic Cochrane Collaborative Centre Cochrane, Copenhagen, Denmark) for meta-analysis.We used weighted mean diferences and their 95% confdence intervals for analysis for continuous variables.A P value <0.05 was assigned statistical signifcance, and a P value <0.01 was given signifcant statistical signifcance [24,25].I 2 test was used to test the heterogeneity between studies.Te random efect model is used in this study.Subgroup analysis was used to assess heterogeneity.If there were only graphs in the literature, two researchers independently used the GetData Graph Digitizer 2.25 to extract the data and calculate the mean and standard deviation [26,27].

Study Selection.
Figure 1 is a fowchart depicting the screening and selection of studies.A search of this database yielded 428 articles.After reviewing the titles and abstracts, 416 articles were excluded.Finally, six full-text articles that met our requirements were retrieved.[20,22,28] included patients who had sufered ZAP within 1-3 months.Two trials [14,23] included patients who had sufered ZAP within 1 month, while one trial [29] involved patients who had sufered ZAP for around 3 months.Te herpes sites in three trials [20,22,28] were the trigeminal nerve, two in the thoracic back [23,29], and one in the neck and upper extremities [14].Before and after treatment, all included trials [14,20,22,23,28,29] reported pain scores to assess pain intensity.Four trials [20,22,28,29] reported SF-36 while others [14,23] reported PSQI or SQS.Four trials reported changes in drug dosage [20,22,23,28], and fve trials reported adverse events after treatment [14,20,22,28,29].Follow-up lasted for 2 days to 12 months in the included studies.Te characteristics of this study are shown in Table 2.

Risk of Bias Assessment.
Five trials [14,20,22,28,29] were assessed as "low risk" for sequence generation because participants were grouped by a numerical randomisation method.One trial [23] was classifed as "high risk" due to the lack of reporting of appropriate random sequence generation methods and the lack of assignment concealment.
Because of the lack of information, three trials [14,28,29] were deemed to be concealing the assigned "undefned risk."Two trials [20,22] were rated as "low risk" for the project because their allocation procedures depended on computergenerated allocation sequences.Tree trials [14,23,29] showed performance bias due to a lack of blind information about people and outcome evaluation.Tree trials [20,22,28] reported that the trial was classifed as "low risk" of performance bias due to the carefully designed S-PRF group and sham surgery group, with staf and participants unaware of the grouping.Due to the lack of information, detection bias was assessed as an "undefned risk" in all trials.Five patients in two trials [20,22] lost follow-up without explanation, which may have resulted in incomplete bias in the outcome data.Other trials were considered "low risk" because no incomplete outcome data was detected.None of the included trials showed selective reporting bias, but four of them [14,20,22,23] did not report conficts of interest and funding sources, which could have contributed to other biases.As no more than 10 studies were included, publication bias was not assessed.Te quality evaluation of randomized controlled trials is shown in Figure 2.  3.

H-PRF and Sham.
Two studies [14,20] compared the pain scores at diferent times after H-PRF and sham treatment, both of which used VAS scores, and were grouped and analyzed according to 1 month after treatment and three months after treatment.Two studies [14,20] showed no statistically signifcant in the H-PRF group and the sham treatment group at 1 month (P � 0.06; WMD, −2.08; 95% CI, −4.21 to −0.06;I 2 � 97%).Two studies [14,20] showed that the H-PRF group had lower pain scores at 3 months after treatment than the sham treatment group (P � 0.03; WMD, −2.22; 95% CI, −4.18 to −0.26;I 2 � 97%), as shown in Figure 4.  [14,20,22,23] reported the efect of pain control by recording the dose of analgesics.Due to the diferences in the category of analgesic drugs, data format, data collection time, and unit in the included randomized controlled trials, we did not have this result in the metaanalysis but only made a descriptive analysis.Wan et al. [22] and Wan et al. [20] reported pregabalin dosage in the H-PRF group compared with the control group.Wang et al. [23] and Lin et al. [14] reported the dosage of gabapentin and tramadol.Due to the diference between the time and unit of data collection, these data cannot be summarized.All four randomized controlled trials [14,20,22,23] showed that the analgesic dose of the H-PRF group was signifcantly lower than that of the standard PRF group and sham group.Anesthesiology Research and Practice 3.7.Side Efects.All RCTs did not report serious adverse efects after H-PRF, such as intraspinal/paravertebral hematoma, hemorrhage, infection, hoarseness/aphonia, pneumothorax hypoesthesia of face or weakness of masseter muscle, and intracranial hemorrhage.

Sensitivity Analysis.
All included studies were excluded one by one, and the results showed no signifcant changes in heterogeneity between the studies, while the meta-analysis results remained consistent.

Discussion
Tere are three main fndings in our meta-analysis.First, H-PRF signifcantly alleviated ZAP when compared with the control group, but the advantages are not so obvious in the early stages.Second, H-PRF seemed to better improve the quality of life and physical and mental health of patients.Tird, no serious adverse events were observed, and the reported complications were comparable between the two groups.
ZAP is refractory neuralgia, and no standard treatment algorithm is devised for all ZAP patients [30].According to the search results, there is no meta-analysis of the efcacy of H-PRF compared with other treatment methods, so this study is the frst.In this meta-analysis, we investigated the efectiveness of H-PRF in reducing pain, improving quality of life, and enhancing physical and emotional functioning in patients with HZ.According to the search, there was no diference in pain relief between H-PRF and S-PRF at 1 week.Compared with the sham operation group, there was no diference in pain relief at 1 month.However, the improvement of pain scores at 1, 3, and 6 months after operation, the H-PRF outperformed the S-PRF.Te H-PRF group was better than the sham group in pain scores at 3 months after operation.In terms of pain improvement in the medium to long term, the H-PRF has more advantages.Te results show that the analgesic efects of PRF appear to develop slowly and take longer to achieve optimal efects.At the same time, this fnding is consistent with other literature suggesting that the neuroregulatory efects of PRF occur gradually and reach their maximum at 3 months after the intervention [31].Te meta-analysis by Wang et al. [32] also showed that the PRF group did not begin to outperform the control group in terms of pain reduction until after 1 week.SF-36 is a practical and common tool for assessing health status and is often used to evaluate the quality of life of ZAP patients [21,33,34].Tis meta-analysis showed that patients in the H-PRF group had higher mental health, physical function, and social function scores than in the S-PRF group.Compared with the sham group, the scores of all items in  Anesthesiology Research and Practice SF-36 in the H-PRF group were higher.Te H-PRF group improved quality of life, which was associated with what we believe to be better Analgesia.H-PRF can provide ZAP patients with the advantage of better mental health and faster return to life and society.Although we cannot summarize and analyze the dosage of analgesic drugs, from a single study perspective, H-PRF can better reduce the dosage of drugs, which means that patients can reduce the impact of drug side efects.PRF will not destroy the anatomical basis of pain pulse transmission nor cause nerve injury and protein coagulation.Its main role is regulating nerve function [12,31,35,36].Tis study found that high-voltage pulse radiofrequency surgery will not cause numbness, nerve injury, or other symptoms, so the safety is as high as the pulse frequency.Compared to traditional thermocoagulation radiofrequency therapy, pulsed radiofrequency current is uninterrupted.Tis energy transfer does not disrupt the anatomical basis of pain pulse transmission, nor does it    Anesthesiology Research and Practice cause nerve damage and protein coagulation [29].Te pulse radio frequency has a silent stage and sufcient time to eliminate heat.Terefore, even if the voltage is increased, it still does not cause damage to the nerves.
In the acute phase of herpes zoster, the VZV-induced ganglions cause a strong local sympathetic response, leading to vasoconstriction, ischemic nerve damage, and pain [37].Pathological changes in the acute phase of HZ can cause peripheral and central sensitization in time, which will downregulate the central pain inhibition pathway, change the gene expression encoding neuropeptides, and expand the receptive feld [38].Sensitization results in the hyperexcitability of dorsal horn neurons.Tese damaged sensory neurons produce abnormal electrical impulses that are transmitted to the spinal cord, inducing spontaneous pain, which is characterized by ectopic pain, hyperalgesia, burning, and electric shock-like symptoms [37][38][39].Current studies have shown that PRF generates radiofrequency electric feld efects at target tissues, modulates abnormally active synaptic conduction in chronic pain, and at the same time, changes nerve fber structure [40], and afects peripheral nerve cell ion channels to improve ZAP peripheral sensitization.PRF can increase the expression of transcription activator-3 in C and Aδ fbers of dorsal root ganglion pain conduction, thereby activating the descending inhibitory system of the brainstem, resulting in an analgesic efect [41].
PRF is a new therapeutic technique.Currently, how to maximize the therapeutic efect of PRF has always been a signifcant problem faced by clinicians and scientists.Nowadays, many basic experiments and clinical studies have been carried out on parameters such as target, time, waveform, temperature, voltage, etc.Currently, no international guidelines recommend an ideal PRF setting for treating neuropathic pain [42][43][44].Te standard RF mode has a frequency of 2 Hz, output voltage of 45 V, pulse width of 20 ms, and upper temperature limit of 42 °C, but many clinical patients cannot achieve satisfactory curative efect [8][9][10][11].Terefore, some scholars began to change the voltage parameters and found that good clinical efcacy could be achieved.In 2006, Teixeira et al. used a high voltage of 60 V for the frst time in patients with lumbar discogenic pain and achieved satisfactory results [18].It was believed that the therapeutic efect of PRF was due to the infuence of the electric feld efect generated by the electrode on the nerve rather than the thermal efect.In the study we included, there are two modes of high voltage: one is to increase from 40 V to the patient's maximum tolerance gradually, usually 60−100 v [14,20,22,23], and the other is to increase progressively from 40 V to 65 V [28,29], but keep the other parameters the same as the standard RF.
Our research has several limitations.First, the number of RCTs included is small, and the sample size is small.Secondly, considering the limited number of randomized controlled studies on high voltage, we are unable to classify herpes zoster-related neuralgia into herpetic neuralgia and postherpetic neuralgia for subgroup analysis.Tird, the follow-up time included in the study is short, no more than half a year, so it is impossible to compare the long-term efcacy.Finally, the data extraction method (GetData graph digitizer) may limit the statistical power and accuracy.

Disclosure
Yinghao Song and Ziheng Yu contributed equally to this work and are the co-frst authors.

Figure 2 :
Figure2: Risk of bias assessment using the cochrane risk-of-bias tool for randomized controlled trials included in a meta-analysis.

Figure 4 :
Figure4: Forest plot of pain scores for the H-PRF and sham group at one month and three months after operation in a meta-analysis.

Figure 5 :
Figure 5: Forest plot of SF-36 score for H-PRF and S-PRF at the last follow-up time after operation in a meta-analysis.

Figure 6 :
Figure 6: Forest plot of SF-36 score for the H-PRF and sham group at the last follow-up time after operation in a meta-analysis.

Table 1 :
Abbreviations used in the text.

Table 2 :
Te characteristics of the studies included in this meta-analysis.
Figure 3: Forest plot of pain scores for H-PRF and S-PRF at one week, one month, three months, and six months after operation in a meta-analysis.
Anesthesiology Research and Practice