Comparison of Adjuvant Hypertonic Saline and Normal Saline for Epidural Block in Patients with Postherpetic Neuralgia: A Double-Blind, Randomized Trial

Background In patients with postherpetic neuralgia (PHN), the effectiveness of epidural block and the benefits of adjuvant hypertonic saline (HS) have not been fully determined. Therefore, we investigated these issues in this study. Methods At a tertiary medical center's single pain clinic in Seoul, Republic of Korea, patients complaining of PHN even after 4 months of herpes zoster onset were enrolled and randomly assigned to either the HS or normal saline (NS) group. After epidural block with adjuvant HS or NS administration according to each protocol, outcomes were assessed at baseline and one and three months after the intervention. The primary outcome was pain intensity on the numerical rating scale (NRS). The secondary outcomes were the insomnia severity index (ISI), the medication quantification scale (MQS), and the global perceived effect of satisfaction (GPES). Results Thirty-six patients (NS: 17, HS: 19) were included in the intention-to-treat analysis. The estimated pain intensity decreased in both groups at one and three months after the procedure (P < 0.001), without a significant group difference. The estimated ISI and MQS were not significantly different at 1 month compared with baseline but significantly decreased at 3 months in each group (P < 0.001 and P < 0.001, respectively), without group differences. In addition, there was no difference between the groups on the GPES scale at one and three months after the procedure. Conclusions Epidural steroid injection may have the advantages of short-term pain relief, improved sleep quality, and decreased medication usage in patients with PHN. In addition, adjuvant HS administration with epidural steroid injection did not show beneficial effects in patients with PHN. Further studies are needed to clarify the potential effectiveness of HS in treating neuropathic pain such as PHN. This trial is registered with KCT0002845.


Introduction
As a sequelae of acute herpes zoster (AHZ) in sensory neurons, postherpetic neuralgia (PHN) often lasts from several years to a lifetime. Te incidence of PHN is 7%-50% at 3 months and 5%-32% at 6 months after AHZ [1]. Tese patients often sufer from severely agonizing pain that leads to sleep disturbance, impaired social activities, and decreased quality of life [1][2][3].
Although several pharmacologic treatment options for PHN, such as anticonvulsants, antidepressants, opioids, topical lidocaine, and topical capsaicin, can be used, they are sometimes insufcient to control the pain. In this case, interventional treatments such as epidural or intrathecal block with or without steroid injection, spinal cord stimulation, and sympathetic block can be considered [2,4,5]. Te most popular interventional method for treating PHN was epidural blocks. However, this procedure may not be helpful for long-term pain relief in patients with PHN [2,3]. In addition, the efectiveness of epidural blocks for PHN has not been fully determined [6].
In general, the analgesic efect of epidural injection with local anesthetics and steroids is mediated by inhibiting the synthesis of infammatory mediators and decreasing neural excitability. Previously, it has been shown that the administration of adjuvant hypertonic saline (HS) for epidural block has the benefts of decreasing pain intensity and extending the duration of pain relief in patients with other types of chronic pain, such as pain due to spinal stenosis [7][8][9]. Although the mechanism of action of HS for epidural block is not yet fully clarifed, high concentrations of sodium and chloride ions may play a role in blocking noxious stimuli with a neuromodulatory efect as well as in reducing nerve edema [7,8]. Tese efects of HS may be reasonably applicable to patients with PHN.
Terefore, the present study aimed to investigate the efectiveness of adjuvant HS administration for epidural block in patients with PHN.

Participants and Study Design.
Tis study was a randomized, double-blind, active-controlled clinical trial. Te Institutional Review Board of our institution gave its consent for this study to be conducted (IRB approval number, 2017-0433). Each patient who took part in this trial provided their written, informed permission. Participants' confdentiality and privacy were upheld in every way. In addition, this study was performed in compliance with the Helsinki Declaration and fled with the Clinical Research Information Service (https://cris.nih.go.kr/KCT0002845) [10]. In reporting this study, we adhered to the CONSORT recommendations.
We evaluated the eligibility of PHN patients who visited our center's pain management clinic between May 2018 and December 2020. Te following were the criteria for inclusion: (1) patients who complained of PHN even after 4 months of herpes zoster onset [11]; (2) patients (aged ≥18 years) who could understand and cooperate with the research purpose; (3) patients whose location of PHN was a dermatome below C2; (4) patients with a pain intensity ≥6 (out of 10) on the numerical rating scale (NRS); (5) patients who failed to control pain even when an efective dose of analgesics was administered; and (6) patients who voluntarily agreed in writing to participate in the present trial.
Exclusion criteria were as follows: (1) patients who refused to participate in this clinical trial; (2) patients whose PHN was located in the head or face; (3) patients who underwent any interventional procedure such as an epidural steroid injection, paravertebral block, pulsed radiofrequency, or spinal cord stimulation within 3 months; (4) patients who were pregnant or lactating; (5) patients who had hypersensitivity to local anesthetics, steroids, contrast dyes, or HS; (6) patients who had a coagulopathy; (7) patients with red fag signs (infection, cancer, fracture, and neurological defcits or cauda equine syndrome); and (8) patients with uncontrolled medical or psychiatric diseases, including opioid disuse.

Randomization and
Blinding. Participated patients were randomly assigned to one of two groups: the HS group (administered an epidural steroid injection with HS) or the normal saline (NS) group (administered an epidural steroid injection with NS). Based on a computer-generated randomization method, an impartial data manager divided the patients among the groups. Each patient's randomization number was concealed from both the trial participants and the outcome assessor, an independent pain specialist. From the time of the procedure to the follow-up period, the interventional pain physicians remained unaware of which group the patients belonged to.

Intervention: Epidural Steroid Injection with HS or NS.
Two pain physicians with more than ten years of experience conducted every procedure in this study. No premedication or sedatives were used in all participants. In patients selected according to the inclusion criteria, the target site of injection was the epidural space of the nerve afected by herpes zoster, through the parasagittal approach. All epidural steroid injection procedures were performed under fuoroscopic guidance. All epidural blocks were performed under a fuoroscopy system (OEC 9800, GE Healthcare, Little Chalfont, UK). During the procedure, each patient was placed in the prone position, and vital signs (pulse oximetry, blood pressure, and electrocardiography) were monitored. Sterile preparation and draping of the injection area were performed. Regarding the parasagittal interlaminar approach, the anteroposterior (AP) radiographic view was obtained to confrm the targeted vertebral level. Ten, the image intensifer was cephaladly or caudally tilted to make the targeted interlaminar space more visible. After deciding the appropriate needle insertion point, local infltration with 2 mL of 1% lidocaine was performed on the skin and soft tissues. A 20-gauge Tuohy needle was gently advanced under guidance by fuoroscopy. After placing the needle tip between the laminae, being ipsilateral to the symptomatic side of the patient on AP view, the needle was gently advanced using the lateral or contralateral oblique view while confrming the depth by visualizing the spinolaminar line or ventral interlaminar line, respectively [12][13][14]. Upon feeling the ligamentous resistance, the stylet was removed. Ten, the needle was further advanced in millimeter increments using the loss of resistance technique to reach the epidural space. In a case of cervical epidural block above the C6 level, an 18gauge Tuohy needle was inserted at the C6-7 or C7-T1 level for an epidural catheter insertion to reach the target cervical level. After an aspiration test to exclude cerebrospinal fuid or blood, 0.5-2.0 mL of contrast dye (Omnipaque 300, GE Healthcare AS, Oslo, Norway) was injected under a real-time fuoroscopy guide to ensure sufcient dispersion to the epidural space and to avoid additional potential intrathecal or intravascular injection. Following verifcation of the radiographic imaging's proper needle placement and sufcient contrast distribution, 6 mL of 1% lidocaine was administered. Before administering the drugs, a nurse who was not related to this study opened the envelope containing the paper with the assigned group and flled the syringe with the drug according to the assignment in a place where the physician could not see it. Ten, the syringe that was not discernible by color or volume was delivered to the physician. After 3 min of lidocaine injection, 3 mL of 10% HS containing 5 mg dexamethasone in the HS group or 3 mL of NS containing 5 mg dexamethasone in the NS group was slowly administered over 3 min. After the procedure, the administration of neuropathic medication and/or analgesics was allowed.

Outcome Assessments and Follow-Up.
All study participants' baseline characteristics were gathered. Assessments of the procedure's outcome were made at the baseline, 1 month, and 3 months later. Before the procedure, all study participants were instructed on how to use the 11-point NRS (0 = no pain and 10 = worst possible pain) to assess the pain intensity of their PHN, along with the insomnia severity index (ISI), a 7-item self-report questionnaire with a total score ranging from 0 to 28, to rate their quality of sleep [15][16][17]. Furthermore, patient satisfaction and improvement were evaluated using the 7-point Likert scaled global perceived efect of satisfaction (GPES) [18]. In addition, analgesic alterations were accessed using the medicine quantifcation scale (MQS) III [19]. Adverse events that occurred throughout therapy and follow-up were documented.
Te primary outcomes were the mean diference of NRS scores in pain intensity from baseline at one month and three months. Te secondary outcomes included changes in ISI, MQS, and GPES and adverse events in each group over the course of the follow-up. At each follow-up evaluation, the diferences from baseline in pain intensity, ISI, and MQS were assessed. If procedural complications occurred, they were recorded. Also, any adverse events were further assessed during the follow-up period.
All patients were informed of this guideline before participating in the trial. According to the patients' residual pain intensity at each follow-up, the prescription doses of each analgesic were either decreased or increased.

Sample
Size. Te study sample size was determined on the basis of a previous publication [8], in which the average degree of pain reduction 3 months after the procedure was estimated at 2.93 for the HS group, 1.52 for the NS group, and 1.6 for the standard deviation. Assuming a type 1 error of 0.05 (two-tailed) and a power of 0.80, a minimum number of 34 patients was required for between-group comparisons with the repeated analysis of variance test. By setting the dropout rate at 15%, we determined a total of 40 patients as the sample size.

Statistical Analysis.
Absolute numbers and percentages are used to represent categorical variables. Continuous variables are displayed as medians with interquartile range or means with standard deviation or 95% confdence intervals. Te categorical variables were evaluated by the χ 2 test or Fisher's exact test appropriately. Te Student's t-test or the Mann-Whitney U test was used to evaluate continuous variables, as applicable, to compare data between groups. All observed data were analyzed by intention-to-treat (ITT) analysis, despite loss to follow-up or dropout from the study. Given that data loss resulting from loss to follow-up was anticipated, a linear mixed-efect model was used to analyze and compare changes in continuous variables (NRS, ISI, MQS, and GPES) at baseline and one month and three months after the procedure within and between groups. SPSS version 21 (IBM Corp, Armonk, NY) and Stata version 13.1 (StataCorp LP, College Station, TX) were used for all data handling and statistical analysis. A statistically significant diference was determined to exist when the two-tailed P value was less than 0.05.

Results
Between May 2018 and December 2020, 88 consecutive patients diagnosed with PHN were screened for eligibility to be included in this study. A total of 40 patients who fulflled both the exclusion and inclusion criteria agreed to participate in this study. Seven patients did not meet the inclusion criteria, 20 patients met the exclusion criteria, and 21 patients declined to participate. After randomization, 20 patients were assigned to the NS group and 20 patients were assigned to the HS group. Among the 20 eligible NS group patients, 1 patient did not receive the allocated intervention because of improved symptoms and 2 patients refused to participate after randomization. Among the 20 eligible HS group patients, 1 patient did not receive the allocated intervention because he refused to participate in the study. Finally, a total of 36 patients (17 in the NS group and 19 in the HS group) were included in the ITT analysis. All 36 patients were assessedat 1 at one and three months after the procedure, two NS group patients dropped out, whereas no HS group patient dropped out. At 3 months after the procedure, one NS group patient and three HS group patients dropped out ( Figure 1). Te reasons for dropout were as follows: fve patients did not return for the follow-up visit, and one patient underwent another intervention before completing the study.
As shown in Table 1, the baseline demographic characteristics were not signifcantly diferent between the two groups. Te most common site of lesions was the thoracic spine, without a group diference. As the primary outcome, the estimated mean changes in pain intensity (NRS) over the    Tables 3 and 4, respectively. Although the estimated ISI and MQS were not diferent at 1 month compared with baseline, the ISI and MQS in both groups were improved at 3 months compared with baseline after each intervention (P < 0.001 and P < 0.001, respectively), without group diferences (P � 0.123 and P � 0.899). No signifcant interactions in ISI and MQS were observed between groups and time (P � 0.981 and P � 0.465, respectively). In addition, the GPES scales at one and three months after the procedure were not signifcantly diferent between groups (Table 5). We did not observe any serious complications during the injections. Several patients reported temporary injection pain during needling and injection, which was manageable and did not need any additional care. Tere were no other complications, such as intravenous injection, dural puncture, infection, or persistent neurologic impairment. Postprocedure complications were also not reported during the follow-up.

Discussion
Te present study found that the intensity of pain was decreased by 3 months after epidural block in patients with PHN. However, the diference in pain intensity decrease between the HS and NS groups was not signifcant. Te estimated ISI and MQS were also decreased3 at months without group diferences.
Epidural nerve blocks are often performed in patients with PHN that are not well controlled by conservative treatment with neuropathic pain medication and/or analgesics. According to the report by Nahm et al., epidural nerve blocks were performed in 52.1% (interlaminar epidural injection in 32.7% and transforaminal epidural injection in 19.4%) of patients with PHN in the spinal area in Korea [6]. However, there are few data on the efectiveness or efcacy of epidural blocks in PHN, whereas the quality of evidence about relieving pain in AHZ and preventing PHN after AHZ is moderate [20,21]. Te speculation that nerve blocks, including epidural injections, might have some benefts for PHN was only based on some case series [20,22]. In another previous report, satisfactory analgesia with more than 50% of global pain relief was achieved only in 20% of the 12 cases of epidural steroid injection at 1 week and 24 weeks after the procedures. In this report, the mean extent of pain relief was not presented as numbers like the NRS, but just as graphs [23]. A recent retrospective study reported that 45.2% of the patients with PHN had moderate-to-good pain relief at 3 months after epidural block [24]. Until now, the efectiveness of epidural injection in PHN has not been completely determined [2]. In the same context, an evidence-based guideline suggested that epidural steroid or anesthetic injection for the treatment of PHN is not supported by enough research [5]. A recent systematic review did not even mention epidural injection as an interventional treatment for PHN [25]. Nonetheless, in the present study, 30% pain relief was observed in the short-term period (3 months) after epidural injection in PHN, regardless of whether adjuvant HS was administered or not. Interestingly, administration of adjuvant HS helped in improving the sleep quality and decreasing the medication requirement of patients. To our knowledge, this study is unique in that it reported NRS diferences, showing the extent of pain reduction in a randomized controlled study.
Te pain relief achieved in both groups implies that epidural block with local anesthetics and steroids would be efective for PHN regardless of adjuvant HS administration. Te sensory nerve afected by a neuropathic condition such as PHN has altered ion channel functioning and is usually Data are expressed as a number (%), mean ± standard deviation, or median (interquartile range). a Te individualized baseline drugs for managing pain were anticonvulsants (gabapentinoids and carbamazepine), antidepressants (tricyclic antidepressants and serotonin norepinephrine reuptake inhibitors), topical agents (transdermal lidocaine patches and capsaicin cream), and analgesics (opioids and NSAIDs). ISI, insomnia severity index; MQS, medication quantifcation scale; NRS, numerical rating scale; NSAIDs, nonsteroidal antiinfammatory drugs.  sensitized with a lower threshold for action potentials. Tis produces disproportionate responses to stimuli and even spontaneous discharges. Local anesthetics block the generation and propagation of action potentials by inhibiting voltage-gated sodium channels. Generally, the afnity for ion channels in depolarized membranes is higher in neuropathic conditions [26]. Te results of this study suggest that this mechanism may be applicable to PHN, similar to other neuropathic conditions. Meanwhile, corticosteroids are known to have roles in interrupting the infammatory cascade, relieving edema of the afected nerves, and promoting nerve restoration [27]. However, epidural steroid injections may have more important roles in AHZ than in PHN, although the present results showed short-term effectiveness in managing PHN [28]. Epidural injection of HS was reported as an efective treatment in other neuropathic pain syndromes. In previous studies, epidural adhesiolysis with HS has shown favorable outcomes for short-term and long-term pain relief in patients with spinal stenosis, postspine surgery syndrome, and radiculopathy of any etiology, with moderate evidence [28][29][30]. Furthermore, a randomized controlled study by Shin et al. demonstrated the superiority of the addition of 10% HS to transforaminal epidural steroid injections (TFESI) for patients with chronic lumbosacral radicular pain caused by lumbar lateral canal stenosis [8]. Te same research group retrospectively analyzed 246 patients with chronic low back pain associated with radicular pain and also showed that adding 10% HS to conventional TFESI provided a superior response over a 6-month period compared with conventional TFESI [7]. In light of the positive efects of HS reported earlier in other neuropathic pain syndromes, we hypothesized that epidural block with adjuvant HS may have merits for PHN, which is another specifc group of neuropathic pain syndromes. According to the present results, adjuvant injection of 10% HS did not reduce pain intensity, improve sleep quality, or decrease drug utilization to a greater extent than conventional epidural block in patients with PHN.
As a conceivable reason, the chronicity of the disease might be a contributing factor. Te mean total duration of pain was 26.6 (9.0-38.0) months in total participants,   Te insomnia severity index (ISI) was used to assess insomnia. A mixed linear model was used in the statistical analysis. Te data are presented as the estimated mean with a 95% confdence interval. Te group diference was not signifcant (P � 0.123). Te time efect was signifcant (P � 0.014). Te P values of interaction between the group and time were not signifcant (P � 0.981). a Estimated diference in values between groups. b P < 0.001 compared with the baseline in each group. 6 Pain Research and Management without substantial diference between the groups. According to the previous literature, nerve block in the acute stage of herpes zoster helps prevent PHN [28,31]. Generally, it is accepted that the longer the duration of neuropathic pain, the higher the chance of central sensitization. Central sensitization is known as a crucial factor for poor response to any treatment [32][33][34]. In addition, the irreversible damage to the nervous system might be more extensive, involving the spinal cord and peripheral nerve endings from the beginning, which could also contribute to peripheral sensitization and hyperalgesia in PHN, unlike in spinal degenerative lesions investigated in prior studies. Considering that PHN does not originate from a mechanical spine problem but from damage to the nervous system itself, PHN caused by remnant nerve damage may be difcult to reverse or improve compared with neuropathic pain related to mechanical spine problems. Nevertheless, it is difcult to rule out that adjuvant HS administration with epidural steroid injection might relieve pain in patients with AHZ or early stage of PHN.
Tis pilot study has several limitations. First, the 16.7% follow-up loss rate with a small sample size may be criticized. Because this study was conducted in one of the largest hospitals in Korea, a considerable percentage of the patients came from various cities. Te extent of follow-up loss may have been infuenced by the high percentage of patients from remote areas. Although an ITT-based analysis and a linear mixed model were utilized to adjust for missing variables, the 3-month follow-up period in the present study may have been insufcient to evaluate the long-term clinical efectiveness of the epidural block. Terefore, it is necessary to further evaluate the long term treatment efect beyond 3 months. Second, the volume of injectate in the parasagittal interlaminar epidural injection in the present study was decided at 6 mL arbitrarily. Although the suggested volume of injectate for epidural injection varied in a few literature reports, the optimal volume is still to be determined [35][36][37]. However, a 6 mL volume in the parasagittal approach was considered sufcient to cover the targeted spinal nerve regardless of the spinal level [38]. Tird, the protocol was confned to only a single bolus injection of HS. Since a single adjuvant injection is not necessarily the answer, this protocol can be modifed to another, such as a several-day protocol with catheter insertion, which might afect the results in diferent ways [39]. Finally, the patients' initial emotional states could not be assessed. Te perception of pain intensity and functional outcomes following the procedure may have been impacted by subjective emotions. However, considering that individuals with psychiatric or neurological disorders were not allowed to enroll in the trial, we think that emotional factors had a negligible impact on our fndings.

Conclusions
Epidural steroid injection may have advantages of shortterm pain relief, improved sleep quality, and decreased medication usage in patients with PHN. In addition, adjuvant HS administration with epidural steroid injection did not show benefcial efects in patients with PHN. Further studies are needed to clarify the potential efectiveness of HS in treating neuropathic pain such as PHN.

Data Availability
Te Institutional Review Board at Asan Medical Centre, Seoul, Korea, has put restrictions on the data used to support the study's conclusions in order to preserve patient privacy. For researchers who meet the requirements for access to secret data, the corresponding author will provide the data.

Disclosure
Te funder played no part in the design and conduct of the study; data collection, management, analysis, and interpretation; preparation, approval, or review of the manuscript; or making the choice to submit the manuscript for publication.

Conflicts of Interest
Te authors declare that they have no conficts of interest.

Authors' Contributions
Data curation, methodology, formal analyses, visualization, reporting, and writing-original draft were done by HJK; investigation, methodology, and validation by DHK; formal analyses, software,and validation by SSiC; data curation, visualization, and validation by BJ; conceptualization, data curation, formal analyses, and writing-review and editing by MHK; conceptualization, methodology, investigation, formal analyses, funding acquisition, and supervision by SSoC. Myong-Hwan Karm and Seong-Soo Choi contributed equally to this project.