Postherpetic neuralgia (PHN) is the most common complication of herpes zoster (HZ) and has a reported incidence of 10% to 70%, depending on its definition and the patient age.1–3 The incidence of PHN in patients with HZ has been rising over the past decades.4 The development of PHN is thought to be related to certain types of neuronal damage incited by the varicella zoster virus. Immunohistochemical investigation and biopsy of the affected dermatome in patients with PHN show loss of afferent nerve fibers in the skin.5,6 Neurophysiological examination indicates that both small and large sensory fibers are affected in these patients, that is, loss of small nociceptive fibers may cause constant pain and demyelination of large A-beta fibers, which may lead to paroxysmal pain.7
Because the effect of treatment is unsatisfactory once PHN has developed, physicians recommend identification of those patients who are most likely to develop PHN and treat them using preventive strategies.8–12 However, several clinical factors, such as older age and greater acute pain severity, are recognized as risk factors of PHN that can be used to develop multiple prediction models.13–16 Several biomarkers might serve as neoteric predictors for related nervous system diseases as these can objectively indicate characteristic pathogenic processes of the disease.17 Previous studies have shown different expression profiles of serum miRNAs between PHN patients and acute HZ patients, providing new evidence of circulating biomarkers for diagnosing PHN.18 Taking into account the pathogenic processes of HZ, we hypothesized that identification of certain type circulating biomarkers of neuronal damage might be valuable in predicting PHN.
In this study, we examined several easily measurable circulating biomarkers that reflect neuronal damage, including cell-free DNA (cfDNA), a nonspecific biomarker of injury19,20; myelin basic protein (MBP), a biomarker related to demyelination21–23; and soluble protein-100B (S100B), a biomarker related to glial reactivity after nerve injury.24 We compared plasma levels of these biomarkers among healthy controls and patients with HZ who did or did not develop PHN and evaluated whether these biomarkers could serve as predictors for PHN. Finally, we constructed and evaluated a prediction model for PHN based on potential markers of neuronal damage and clinical risk factors.
MATERIALS AND METHODS
This study was conducted at Xiangya Hospital of Central South University from November 7, 2016 to December 21, 2017. All procedures were approved by the Ethics Committee of Xiangya Hospital of Central South University (IRB no.: 201609599) in accordance with the Declaration of Helsinki and registered with the Chinese Clinical Trial Registry (Ref: ChiCTR-ROC-16009748). All participants provided their informed consent for inclusion before participating in the study.
We recruited patients with a specialist-confirmed diagnosis of HZ; only those who were within 90 days after rash onset were included. We excluded patients with cognitive disorders, psychosis, tumors, infectious diseases, nervous system demyelinating diseases, or previous vaccination against HZ, and those who had undergone trauma or surgery in the previous months. Volunteers without a history of HZ were recruited as controls.
This was a nonmatched, prospective, nested case-control study. Patients with HZ were asked to complete a survey questionnaire about their baseline characteristics and clinical features including age, sex, body mass index (BMI), comorbid diabetes, history of chickenpox/HZ, rash location, prodromal pain duration, rash duration, interval from rash onset to delayed antiviral prescriptions, and acute pain severity (maximum visual analog scale [VAS] score for pain during the first 30 d after rash onset). Peripheral blood samples were collected from patients with HZ and controls to test for biomarkers of neuronal damage on the first visit day they were included.
Patients’ pain severity was evaluated (using a 10-point VAS)25 at the first visit and at 30, 60, and 90 days after rash onset, through telephone follow-up calls. Pain relief was evaluated at 60 and 90 days, calculated as follows: (acute pain severity−VAS at evaluating timepoints)/acute pain severity, ranging from 1 (complete relief) to 0 (no relief), even negative scores (worse). PHN was defined as persistent clinically significant pain (VAS score ≥3) for >90 days after rash onset.26,27 All patients received pharmacologic treatment including antiviral drugs (valaciclovir; Sichuan Med-shine Pharmaceutical Co. Ltd, Chengdu, China), calcium channel regulators (gabapentin; Jiangsu Nhwa Pharmaceutical Co. Ltd, Xuzhou, China), analgesics (dezocine; Jiangsu Nhwa Pharmaceutical Co. Ltd), and cobalt compounds (cobamamide; Hebei Zhitong Pharmaceutical Group Co. Ltd, Shijiazhuang, China). According to the patients’ willingness and the evaluation of a specialist, a number of patients received neurointerventional treatment, including intrathecal injection and nerve block using ropivacaine (AstraZeneca, Wilmington, DE) combined with betamethasone (Merck, Kenilworth, NJ) and pulsed radiofrequency treatment.
Circulating Biomarker Measurement
Blood samples were collected in vacuum tubes containing EDTA K2 anticoagulant. Samples were then centrifuged (800g at 4°C) for 10 minutes. The supernatant was then transferred to a new tube, and centrifuged (1600g at 4°C) for another 10 minutes.28 The plasma was then distributed into several cryogenic vials and kept frozen at −80°C for further testing.
Quantification of cfDNA
Plasma samples were unfrozen at 4°C and centrifuged (16,000g at 4°C) for 10 minutes. Then, cfDNA was extracted from 250 μL of supernatant by binding to a silica-based column and eluting with an elution buffer, using the EasyPure Micro Genomic DNA kit (TransGen Biotech, Beijing, China). Finally, cfDNA was eluted with 30 μL elution buffer (TransGen Biotech). Absolute quantification of cfDNA was done using real-time quantitative polymerase chain reaction. Three target genes were examined, including 1 nuclear-encoded gene: GAPDH (ID: 2597), and 2 mitochondrial-encoded genes: mtATP8 (ID: 4509) and mtND4 (ID: 4538).29,30 The primers of the target genes were designed using Primer3Plus (www.primer3plus.com) based on information from the database of the National Center for Biotechnology Information (NCBI) and verified using the NCBI primer-BLAST tool. (sequence information shown in Table S1, Supplemental Digital Content 1, http://links.lww.com/CJP/A742). The amount of cfDNA was calculated using the standard curve derived from serial 10-fold dilutions of a recombinant plasmid (pUC57) solution (Sangon Biotech Co. Ltd, Shanghai, China) of known concentration, which contained the ampliconic sequence of the corresponding primers (Table S1, Supplemental Digital Content 1, http://links.lww.com/CJP/A742). Real-time quantitative polymerase chain reaction was performed using the ViiA 7 QuantStudio Real-Time PCR System (Applied Biosystems, Branchburg, NJ) under the following conditions: a first denaturation for 10 minutes at 95°C, followed by 40 cycles of 15 seconds at 95°C, 29 seconds at 60°C, and 24 seconds at 72°C. The concentration of cfDNA in the plasma was equal to the concentration in the DNA sample times 25/3, that is, the volume ratio of plasma for cfDNA extraction to elution buffer.30
Quantification of Specific Neuronal Damage Biomarkers
Plasma levels of human MBP and S100B were determined using an enzyme-linked immunosorbent assay kit (Cusabio Biotech Co. Ltd, Wuhan, China), according to the manufacturer’s protocol. Briefly, 100 μL of each plasma sample and standard were added into a well of a microplate precoated with a specific antibody. This was then incubated with 100 μL antibiotin antibody, followed by incubation with 100 μL avidin-horseradish peroxidase. Then, 90 μL TMB (chromogenic substrate) was added and the mixture was incubated; this was followed by addition of 50 μL stop solution. Finally, the optical density of each well was measured using a microplate reader at 450 nm, and the concentration of target protein in each sample was calculated based on its optical density and the standard curve.22
All statistical analyses were conducted using SPSS 18.0 (SPSS Inc., Chicago, IL) and GraphPad Prism 7.0 (GraphPad Software Inc., San Diego, CA).
Parametric data were reported as mean±SD and nonparametric data were reported as number (%). The Shapiro-Wilk test was used for tests of normality. Regarding the comparison of variables that were normally distributed between 2 groups, an unpaired t test was used if their variances were equal (using Levene test for equality of variances); otherwise, a t test was used. As to the comparison of variables that were not normally distributed between 2 groups, the Mann-Whitney U test was used. For comparisons of variables that were normally distributed among 3 groups, a 1-way analysis of variance was used, and posthoc analyses were performed using Holm-Sidak multiple comparisons test. Regarding comparison of variables that were not normally distributed among 3 or more groups, we used a Kruskal-Wallis test and posthoc analyses were performed using Dunn multiple comparison test. To analyze the relationships between circulating biomarkers and clinical factors, bivariate correlation analyses were used. Since these variables were not normally distributed (tests of normality was attached in the Supplementary Materials, Supplemental Digital Content 2, http://links.lww.com/CJP/A743), Spearman rank correlation was used.31
For screening predictors of PHN, we used univariate analysis for preliminary screening of candidate predictors, followed by multiple logistic regression analysis for the prediction models. Prediction models were constructed based on clinical factors, with or without circulating biomarkers, using logistic regression analysis. The prognosis (PHN or non-PHN) was set as the dependent variable, and candidate predictors were set as independent variables. The assignment criterion is shown in Table S4 (Supplemental Digital Content 3, http://links.lww.com/CJP/A744). The models were tested for goodness-of-fit using the Cox-Snell R2 and Nagelkerke R2. Receiver operating characteristic (ROC) curve analyses were used to evaluate the performance of MBP and prediction models by the function value of each. Finally, performance was evaluated using the area under the ROC curve (AUC), as well as sensitivity and specificity at the cutoff point of the ROC curve.
To estimate the study sample size, we used the computational formula of group design for measurement data, as follows:
In the formula, n refers to the sample size of each group; δ is the difference between the means of 2 populations, and σ is the SD of the population (assuming their SDs are equal). We set α=0.05 (both sides), β=0.20, and estimated that δ/σ=0.9, thereby determining that the sample size required for univariate analysis in each group was about 20. For logistic regression analysis, the total sample size was suggested to be 10 times the number of variables. According to previous studies and the characteristics of this study, the number of variables was estimated to be 5 to 8, considering a possible 5% dropout rate, the total required sample size was 52 to 85.
A total of 37 volunteer controls and 92 patients with HZ were included in the study. Three patients were lost to follow-up because they did not answer their telephone. Among the 89 patients with HZ who were finally enrolled, 32 developed PHN and 57 did not (Fig. 1); the incidence of PHN was 35.96% in these patients. There was no difference in age (64.32±9.86 vs. 64.37±14.84 y, P=0.499), BMI (23.26±2.81 vs. 22.53±3.45 kg/m2, P=0.256), and sex distribution (P=0.721) between controls and patients with HZ.
Blood samples were collected from all 37 controls and from 71 patients with HZ; the remaining 21 patients refused to provide blood samples. The 71 patients who consented to blood sample collection showed no significant difference with the total 89 patients who had HZ, with respect to age (P=0.513), sex distribution (P=0.490), and BMI (P=0.905). Among the 71 patients who consented to blood sample collection, 25 developed PHN and 46 did not. The incidence of PHN was not different between these 71 patients and the total 89 patients (P=0.922).
Increased Levels of Circulating cfDNA and MBP in Patients With HZ
We divided the blood samples of patients with HZ into 2 groups: 1 group was from patients presenting with rash (n=30) and the other was from patients after their rash had healed (n=41). As shown in Table 1, the levels of cfDNA (including GAPDH, mtATP8, and mtND4) were higher in both groups of patients with HZ than the levels in controls (P<0.001). However, these cfDNA levels were not different between patients with HZ who had rash and those whose rash had healed. Similarly, compared with controls, MBP levels were increased in patients with HZ who had rash and those whose rash had healed (P<0.01 and <0.001); however, MBP levels were not different between these 2 HZ groups. These results indicated that the activation of HZ virus in the nervous system causes elevation of circulating damage-related biomarkers, and skin rash may not contribute to the elevation of these biomarkers.
TABLE 1 -
Plasma Concentrations of cfDNA, MBP, and S100B in Controls and Patients With HZ Presenting Rash and After Rash Healing
||Patients Presenting Rash
||Patients After Rash Healing
|Plasma concentration of cfDNA (fmol/L)
| Plasma concentration of MBP (μg/L)
| Plasma concentration of S100B (ng/L)
Data are expressed as mean±SD.
cfDNA indicates cell-free DNA; HZ, herpes zoster; MBP, myelin basic protein; S100B, soluble protein-100B.
***P<0.001 versus healthy controls, using Dunn multiple comparison test.
Compared with controls, the levels of S100B were lower in patients with HZ who had rash and those whose rash had healed (P<0.001 and <0.01). It has been shown that myelin repair after peripheral nerve injury may deplete S100B, leading to a decrease in its concentration.32 Thus, the decrease of circulating S100B may reflect its neuroprotective role33 but not neuronal damage in patients with HZ.
Prediction of PHN Based on Circulating MBP Level
Levels of circulating cfDNA, MBP, and S100B were further compared between patients with HZ who developed PHN (n=25) and those who did not (n=46). We found no difference in levels of all 3 cfDNAs and S100B between the 2 groups of patients with HZ (Table S2, Supplemental Digital Content 4, http://links.lww.com/CJP/A745). In contrast, MBP levels were significantly higher in patients who developed PHN than in those who did not (14.98±7.89 vs. 9.90±4.99 μg/L, P=0.002; Fig. 2). This result indicated that demyelination injury may contribute to PHN development. In addition, an increased MBP level could be a candidate predictor for PHN. Prediction of PHN based on circulating MBP level was made using ROC curve analysis, with an AUC of 0.720 (95% confidence interval [CI]: 0.591 to 0.848, P=0.002). At the cutoff (10.43 μg/L), the sensitivity was 76.0% and the specificity was 63.0%.
We further evaluated the correlation between MBP level and pain relief at 60 and 90 days after rash onset. We found that among patients who were included within 30 days after rash onset, those with higher MBP levels (>10.43 μg/L) had less pain relief than patients with normal MBP levels at 60 days (0.60±0.42 vs. 0.84±0.29, P=0.0093) and 90 days (0.66±0.42 vs. 0.88±0.24, P=0.0284). In patients who were included between 30 and 60 days after rash onset, MBP levels showed no significant correlation with pain relief at 90 days (0.51±0.40 vs. 0.70±0.28, P=0.2467). These results suggested that MBP levels within 30 days after rash onset were correlated with short-term and long-term pain relief during the disease course.
Prediction of PHN Based on Clinical Factors
Table 2 shows the results of univariate analysis of potential clinical predictors of PHN in the 89 patients with HZ. Age, acute pain severity, and response to treatment drugs showed significant differences between the PHN group and non-PHN group. The remaining clinical factors were excluded in the analysis, including sex, BMI, comorbid diabetes, history of chickenpox/HZ, rash location, prodromal pain duration, rash duration, delayed antiviral prescriptions, and neurointerventional treatment.
TABLE 2 -
Association between Each Candidate Clinical Factor and Prognosis, in Univariate Analysis
|Age, mean±SD (y)
|Sex, n (%)
|BMI, mean±SD (kg/m2)
|Comorbid diabetes, n (%)
|History of chickenpox/HZ, n (%)
|Rash location, n (%)
| Trigeminal nerve region
| Cervical nerve region
| Thoracic nerve region
| Lumbar nerve region
| Sacral nerves region
|Prodromal pain duration, mean±SD (d)
|Rash duration, mean±SD (d)
|Acute pain severity (VAS), mean±SD
|Delayed antiviral prescriptions, mean±SD (d)
|Response to treatment drugs, n (%)
| No relief
| Pain relief <50%
| Pain relief ≥50%
|Neurointerventional treatment, n (%)
BMI indicates body mass index; HZ, herpes zoster; PHN, postherpetic neuralgia; VAS, visual analog scale.
Multiple logistic regression analysis was performed based on the 3 included clinical factors. Age, acute pain severity, and response to treatment drugs were identified as independent clinical predictors for PHN. A prediction model (Logit Mc) based on multiple logistic regression analysis of these independent clinical predictors showed an AUC of 0.823 (95% CI: 0.728 to 0.917, P<0.001). At the cutoff point (−1.618), the sensitivity and specificity of Logit Mc was 96.0% and 58.7%, respectively.
Prediction of PHN by Combining Circulating MBP Level and Clinical Factors
Prediction of PHN was further investigated using multiple logistic regression analysis of the combination of circulating MBP level and clinical factors (Logit Mc+MBP), which showed better predictive validity than circulating MBP level and the clinical prediction model (Fig. 3). The AUC of Logit Mc+MBP was 0.853 (95% CI: 0.764 to 0.943, P<0.001). At the cutoff point (−0.44), the sensitivity was 92.0% and the specificity was 69.6%. In the prediction model Logit Mc+MBP, circulating MBP level was an independent predictor for PHN. Table 3 illustrates the contribution of each predictor in the prediction model of PHN. Among these predictors, circulating MBP level showed the largest odds ratio (OR)=5.032. The second largest OR was for acute pain severity (OR=1.212), followed by age (OR=1.092) and response to treatment drugs (OR=0.195). It should be noted that the partial regression coefficient of response to treatment drugs was a negative value, indicating that response to treatment drugs should act as a protective factor. Together, the above results indicated that circulating MBP level is an independent predictor of PHN. The combination of circulating MBP level with clinical factors achieved better prediction of PHN than each of these alone.
TABLE 3 -
Independent Predictors of PHN Based on Logistic Regression Analysis
||95% CI for OR
|Acute pain severity
|Response to treatment drugs
B indicates partial regression coefficient; CI, confidence interval; MBP, myelin basic protein; OR, odds ratio; PHN, postherpetic neuralgia.
In this study, the levels of circulating cfDNA and MBP were higher in patients with HZ than in controls, indicating damage to the nervous system. The level of MBP was higher in patients who developed PHN than in those who did not, suggesting that demyelination injury may contribute to the development of PHN. The MBP level was an independent predictor for PHN. The combination of MBP level and clinical factors enhanced the predictive power for PHN, compared with prediction based solely on clinical factors.
Increased circulating cfDNA levels indicate enhanced cell death in the tissues of patients with acute HZ. Skin lesions could be a source of dead cells that release cfDNA when the patient has a rash. However, we found that the level of cfDNA remained high after rash healing; this suggests ongoing cell death, taking into account the short half-life of cfDNA in blood circulation.34,35 We consider that the increase of cfDNA may reflect activation of virus-induced neuronal damage, including neuritis or even substantial nerve degeneration, such as nerve fiber loss and demyelination injury, which may exist with or without skin lesions. Progression from neuritis to substantial nerve degeneration has been thought to contribute to the development of most cases of PHN.36 However, the level of cfDNA showed no difference between patients with HZ who developed PHN and those who did not, indicating that increased cfDNA levels may not reflect the exact type of neuronal damage. Thus, cfDNA in circulation may be considered a biomarker reflecting neuronal damage in patients with HZ, but this is not an appropriate predictor of PHN.
Plasma concentrations of MBP were higher in patients with HZ in comparison with controls. Unlike circulating cfDNA, MBP is a characteristic biomarker of the nervous system and accounts for 30% of all myelin proteins.37,38 Elevated circulating MBP levels indicate nerve demyelination.21,23,39 Except for that induced by HZ invasion, preexisting nerve demyelination may be present in some patients, especially older patients.17,28 However, increased MBP levels in our study likely reflect HZ invasion rather than age-induced nerve demyelination, as the bivariate correlation analyses showed that MBP level had no correlation with age (P=0.135) or with other clinical factors (Table S3, Supplemental Digital Content 5, http://links.lww.com/CJP/A746). Moreover, it has been shown that the expression of MBP in the nervous system declines with age,40 which may not cause upregulation of MBP levels in the circulation in elderly adults. Apart from the 92 included patients, 14 patients with HZ >90 days after rash onset were recruited to determine the characteristics of PHN, especially regarding changes in MBP. We outlined the time distribution characteristics of plasma MBP levels, finding that the MBP level had no correlation with time (Fig. 4). Regarding patients with HZ for >90 days, their MBP levels were not very high, and some were even lower than those of controls.
We found that MBP levels were higher in patients with HZ who developed PHN than in those who did not, suggesting that demyelination injury may contribute to the development of PHN. The pain of PHN may result from damage to afferent large-diameter myelinated axons41 and consequent loss of their predominantly inhibitory action on higher-order nociceptive transmission neurons in the dorsal horn of the spinal cord.42 The damage to primary afferents may also induce ectopic impulse, eventually leading to peripheral and central sensitization, an important mechanism underlying the allodynia in PHN.42,43 However, nerve demyelination may not be the only pathology leading to PHN. It is thought that PHN likely represents a disorder in which several mechanisms operate alone or at the same time.7 Our results showed that some patients with HZ who had normal MBP levels also developed PHN. Similarly, PHN has been shown to be the result of damage to or loss of small unmyelinated fibers.6,44 In some cases, PHN even develops without obvious neuronal damage.36,45 Moreover, a recently published paper showed that the circulating plasminogen, thrombin, and vitronectin were significantly downregulated in HZ patients, suggesting that multiple cellular and molecular processes may be involved in the pathology of HZ.46 Thus, consistent with previous reports,36,41,42 nerve demyelination represents an important but not the only mechanism in PHN development.
We analyzed several clinical features in screening potential predictors, finding that older age, severe acute pain, and poor response to treatment drugs were correlated with PHN and might act as predictors of PHN. Older age is an accepted risk factor of PHN.13,14 Severe acute pain might indicate initiation of central sensitization and excitotoxic damage in the dorsal horn, leading to PHN.47 In fact, some clinical factors that are associated with rash and neuronal damage, including rash severity and other neuropathic characteristics, have been suggested to be risk factors of PHN in previous studies.13,48 However, because these are highly correlated with acute pain severity and are difficult to quantify, it is inappropriate to include these factors in the prediction model of PHN. Poor response to treatment drugs might be owing to several reasons; however, these patients did not obtain satisfactory pain relief, leading to a higher risk of PHN.
Several clinical factors were excluded because they showed no correlation with PHN in univariate analysis during preliminary screening, including sex, BMI, comorbid diabetes, history of chickenpox/HZ, rash location, prodromal pain duration, rash duration, delayed antiviral prescriptions, and neurointerventional treatment. Several studies have reported the association between sex and PHN; however, these have drawn the opposite conclusion, and whether women are at higher risk of PHN than men requires further study.13 One large-scale study suggested that diabetes is a risk factor for PHN49; the occurrence of peripheral neuropathy in patients with poorly controlled diabetes may increase the risk of PHN. However, blood glucose control status and peripheral neurological complications in diabetics were not evaluated in that study. In our study, no patients with diabetes had any symptoms of peripheral neuropathy. This may be the reason why diabetes was excluded in the multivariate analysis. Rash location, prodromal pain duration, and rash duration showed no correlation with PHN in this study; whether these factors are risk factors of PHN remains controversial in related studies.13,14,48
We constructed a multivariable regression model that included MBP level and clinical factors. The results showed that increased MBP level was an independent predictor of PHN. Importantly, the OR of MBP level was higher than those of age, acute pain severity, and response to treatment drugs, which have most frequently been identified as predictors in previous studies.13,14 Prediction of PHN was made based on plasma MBP level, clinical factors, and both of these combined. Prediction of MBP level showed an AUC of 0.720, and prediction based on clinical factors an AUC of 0.823, which were comparable with previous findings.14,27,50 The maximum predictive power was achieved in the model incorporating both MBP level and clinical factors, which showed an AUC of 0.853. Taken together, our results indicated that increased MBP level is an independent predictor of PHN, and better prediction could be obtained by a combination of MBP level and clinical factors than by either of these alone.
A limitation of this study is we only tested these circulating neuronal damage biomarkers once for each patient with HZ. Because the duration after onset of HZ usually varies among patients who present to the hospital, the levels of these neuronal damage biomarkers obtained for each patient may also be different, especially S100B, the level of which is increased for only a short period after injury,39,51 MBP may be better, as it has been found that serum MBP concentrations generally remain elevated for up to 2 weeks after injury.52 The prolonged presence of MBP in the circulation may make it more applicable for detecting neuronal damage than other biomarkers. However, as we did not take continuous measurements, we do not know exactly how long increased MBP levels may persist in patients with HZ. Although in this study, MBP levels were tested before patients received treatment, preventing a potential effect on MBP owing to treatment, it is also unclear whether treatment for HZ may affect subsequent variation in MBP levels. Further studies should be conducted to address the above issues.
In conclusion, this study showed that patients with HZ had higher levels of circulating cfDNA and MBP than healthy people. Increased MBP level, which may reflect nerve demyelination, was an independent predictor for PHN. The combination of clinical factors and MBP level enhanced the predictive accuracy for PHN, in comparison with prediction made using clinical factors alone.
The authors acknowledge Fei Ren, MD, Jianqin Yan, MD, Zhigang Cheng, MD, Nianyue Bai, MD, Shenghui Yang, MD and Jiangang Luo, MD in the Pain Medicine Unit of Xiangya Hospital Central South University (Changsha, China) for their support.
1. Kost RG, Straus SE. Postherpetic neuralgia–pathogenesis, treatment, and prevention. N Engl J Med. 1996;335:32–42.
2. Gauthier A, Breuer J, Carrington D, et al. Epidemiology and cost of herpes zoster and post-herpetic neuralgia in the United Kingdom. Epidemiol Infect. 2009;137:38–47.
3. Yawn BP, Saddier P, Wollan PC, et al. A population-based study of the incidence and complication rates of herpes zoster before zoster vaccine introduction. Mayo Clin Proc. 2007;82:1341–1349.
4. Klompas M, Kulldorff M, Vilk Y, et al. Herpes zoster and postherpetic neuralgia surveillance using structured electronic data. Mayo Clin Proc. 2011;86:1146–1153.
5. Oaklander AL, Romans K, Horasek S, et al. Unilateral postherpetic neuralgia is associated with bilateral sensory neuron damage. Ann Neurol. 1998;44:789–795.
6. Xu M, Cheng Z, Ding Z, et al. Resveratrol enhances IL-4 receptor-mediated anti-inflammatory effects in spinal cord and attenuates neuropathic pain following sciatic nerve injury. Mol Pain. 2018;14:1744806918767549.
7. Truini A, Galeotti F, Haanpaa M, et al. Pathophysiology of pain in postherpetic neuralgia: a clinical and neurophysiological study. Pain. 2008;140:405–410.
8. Ding Z, Cao J, Shen Y, et al. Resveratrol promotes nerve regeneration via activation of p300 acetyltransferase-mediated VEGF signaling in a rat model of sciatic nerve crush injury. Front Neurosci. 2018;12:341.
9. Rauck RL, Irving GA, Wallace MS, et al. Once-daily gastroretentive gabapentin for postherpetic neuralgia: integrated efficacy, time to onset of pain relief and safety analyses of data from two phase 3, multicenter, randomized, double-blind, placebo-controlled studies. J Pain Symptom Manage. 2013;46:219–228.
10. Sommer C, Cruccu G. Topical treatment of peripheral neuropathic pain: applying the evidence. J Pain Symptom Manage. 2017;53:614–629.
11. Park J, Baek SJ, Baek SH, et al. Response to transforaminal epidural block as a useful predictive factor of postherpetic neuralgia. J Clin Med. 2019;8:3.
12. Wilhelm IR, Tzabazis A, Likar R, et al. Long-term treatment of neuropathic pain with a 5% lidocaine medicated plaster. Eur J Anaesthesiol. 2010;27:169–173.
13. Forbes HJ, Thomas SL, Smeeth L, et al. A systematic review and meta-analysis of risk factors
for postherpetic neuralgia. Pain. 2016;157:30–54.
14. Drolet M, Brisson M, Schmader K, et al. Predictors of postherpetic neuralgia among patients with herpes zoster: a prospective study. J Pain. 2010;11:1211–1221.
15. Balfour HH Jr. Varicella zoster virus infections in immunocompromised hosts. A review of the natural history and management. Am J Med. 1988;85:68–73.
16. Volpi A, Gatti A, Pica F, et al. Clinical and psychosocial correlates of post-herpetic neuralgia. J Med Virol. 2008;80:1646–1652.
17. He R, Yan X, Guo J, et al. Recent advances in biomarkers for Parkinson’s disease. Front Aging Neurosci. 2018;10:305.
18. Huang Y, Li X, Tao G, et al. Comparing serum microRNA levels of acute herpes zoster patients with those of postherpetic neuralgia patients. Medicine. 2017;96:e5997.
19. Aarthy R, Mani S, Velusami S, et al. Role of circulating cell-free DNA in cancers. Mol Diagn Ther. 2015;19:339–350.
20. Varhaug KN, Vedeler CA, Myhr KM, et al. Increased levels of cell-free mitochondrial DNA in the cerebrospinal fluid of patients with multiple sclerosis. Mitochondrion. 2017;34:32–35.
21. Whitaker JN, Lisak RP, Bashir RM, et al. Immunoreactive myelin basic protein in the cerebrospinal fluid in neurological disorders. Ann Neurol. 1980;7:58–64.
22. Derkus B, Acar Bozkurt P, Tulu M, et al. Simultaneous quantification of Myelin Basic Protein and Tau proteins in cerebrospinal fluid and serum of Multiple Sclerosis patients using nanoimmunosensor. Biosens Bioelectron. 2017;89(pt 2):781–788.
23. Zou Y, Xu F, Tang Z, et al. Distinct calcitonin gene-related peptide expression pattern in primary afferents contribute to different neuropathic symptoms following chronic constriction or crush injuries to the rat sciatic nerve. Mol Pain. 2016;12:1744806916681566.
24. Gonzalez-Martinez T, Perez-Pinera P, Diaz-Esnal B, et al. S-100 proteins in the human peripheral nervous system. Microsc Res Tech. 2003;60:633–638.
25. Zhang YL, Ou P, Lu XH, et al. Effect of intraoperative high-dose remifentanil on postoperative pain: a prospective, double blind, randomized clinical trial. PLoS One. 2014;9:e91454.
26. Johnson RW, Rice AS. Clinical practice. Postherpetic neuralgia. N Engl J Med. 2014;371:1526–1533.
27. Opstelten W, Zuithoff NP, van Essen GA, et al. Predicting postherpetic neuralgia in elderly primary care patients with herpes zoster: prospective prognostic study. Pain. 2007;132 (suppl 1):S52–S59.
28. Wen Y, Wang M, Yang J, et al. A comparison of fentanyl and flurbiprofen axetil on serum VEGF-C, TNF-alpha, and IL-1ss concentrations in women undergoing surgery for breast cancer. Pain Pract. 2015;15:530–537.
29. Phillips NR, Sprouse ML, Roby RK. Simultaneous quantification of mitochondrial DNA copy number and deletion ratio: a multiplex real-time PCR assay. Sci Rep. 2014;4:3887.
30. Kohler C, Radpour R, Barekati Z, et al. Levels of plasma circulating cell free nuclear and mitochondrial DNA as potential biomarkers for breast tumors. Mol Cancer. 2009;8:105.
31. Zhou H, Martinez E, Lin HH, et al. Inhibition of the prefrontal projection to the nucleus accumbens enhances pain sensitivity and affect. Front Cell Neurosci. 2018;12:240.
32. Celikbilek A, Akyol L, Sabah S, et al. S100B as a glial cell marker in diabetic peripheral neuropathy. Neurosci Lett. 2014;558:53–57.
33. Sorci G, Riuzzi F, Arcuri C, et al. S100B protein in tissue development, repair and regeneration. World J Biol Chem. 2013;4:1–12.
34. Diehl F, Schmidt K, Choti MA, et al. Circulating mutant DNA to assess tumor dynamics. Nat Med. 2008;14:985–990.
35. Lo YM, Zhang J, Leung TN, et al. Rapid clearance of fetal DNA from maternal plasma. Am J Hum Genet. 1999;64:218–224.
36. Delaney A, Colvin LA, Fallon MT, et al. Postherpetic neuralgia: from preclinical models to the clinic. Neurotherapeutics. 2009;6:630–637.
37. Whitaker JN. Myelin basic protein in cerebrospinal fluid and other body fluids. Mult Scler. 1998;4:16–21.
38. Zou Y, Tang Z, Ahn W, et al. Increase in regional oxygen saturation in the thigh indicates successful block of the femoral nerve after lumbar plexus block. Br J Anaesth. 2018;121:973–974.
39. Zetterberg H, Smith DH, Blennow K. Biomarkers of mild traumatic brain injury in cerebrospinal fluid and blood. Nat Rev Neurol. 2013;9:201–210.
40. Hou X, Weng Y, Wang T, et al. Suppression of HDAC2 in spinal cord alleviates mechanical hyperalgesia and restores KCC2 expression in a rat model of bone cancer pain. Neuroscience. 2018;377:138–149.
41. Baron R, Haendler G, Schulte H. Afferent large fiber polyneuropathy predicts the development of postherpetic neuralgia. Pain. 1997;73:231–238.
42. Rowbotham MC, Fields HL. Post-herpetic neuralgia: the relation of pain complaint, sensory disturbance, and skin temperature. Pain. 1989;39:129–144.
43. Lim ET, Grant D, Pashenkov M, et al. Cerebrospinal fluid levels of brain specific proteins in optic neuritis. Mult Scler. 2004;10:261–265.
44. Baron R, Saguer M. Mechanical allodynia in postherpetic neuralgia: evidence for central mechanisms depending on nociceptive C-fiber degeneration. Neurology. 1995;45(suppl 8):S63–S65.
45. Rowbotham MC, Fields HL. The relationship of pain, allodynia and thermal sensation in post-herpetic neuralgia. Brain. 1996;119(pt 2):347–354.
46. Wang T, Shen H, Deng H, et al. Quantitative proteomic analysis of human plasma using tandem mass tags to identify novel biomarkers for herpes zoster. J Proteomics. 2020;225:103879.
47. Bennett GJ. Hypotheses on the pathogenesis of herpes zoster-associated pain. Ann Neurol. 1994;35(suppl):S38–S41.
48. Jung BF, Johnson RW, Griffin DR, et al. Risk factors
for postherpetic neuralgia in patients with herpes zoster. Neurology. 2004;62:1545–1551.
49. Jih JS, Chen YJ, Lin MW, et al. Epidemiological features and costs of herpes zoster in Taiwan: A national study 2000 to 2006. Acta Derm Venereol. 2009;89:612–616.
50. Cho SI, Lee CH, Park GH, et al. Use of S-LANSS, a tool for screening neuropathic pain, for predicting postherpetic neuralgia in patients after acute herpes zoster events: a single-center, 12-month, prospective cohort study. J Pain. 2014;15:149–156.
51. Thelin EP, Zeiler FA, Ercole A, et al. Serial sampling of serum protein biomarkers for monitoring human traumatic brain injury dynamics: a systematic review. Front Neurol. 2017;8:300.
52. Thomas DG, Palfreyman JW, Ratcliffe JG. Serum-myelin-basic-protein assay in diagnosis and prognosis of patients with head injury. Lancet. 1978;1:113–115.