Abstract
Background
Postherpetic neuralgia (PHN) is a refractory neuropathic pain syndrome after herpes zoster and is often complicated by medication intolerance in older adults. The dorsal root ganglion (DRG) is a biologically plausible target for interventional analgesia.
Methods
We conducted a retrospective, observational single-arm cohort study including consecutive adults with PHN, who underwent DRG-targeted transforaminal injection (DRG-TFI; local anesthetic with or without corticosteroid) between January 2022 and June 2024. Time zero was the first DRG-TFI during the study period. The primary endpoint was within-patient change in 0–10 numeric rating scale (NRS) pain from baseline to 4 weeks. We report observed pain outcomes and responder rates with confidence intervals (CIs). A linear mixed-effects model is presented to summarize repeated NRS assessments with incomplete follow-up. Exploratory summaries added sleep-disturbance NRS, absolute pain-reduction thresholds, analgesic co-medication trajectories, and calibration of the short-term responder model.
Results
Among 298 included patients (mean age 65.7 ± 10.9 years; 58.1% female), baseline NRS was 7.3 ± 1.1 and a 4-week assessment was available for 279 (93.6%). In the mixed-effects model, mean NRS decreased by − 3.1 points at 4 weeks (95% CI − 3.4 to − 2.8; P < 0.001), with an observed 4-week NRS of 4.1 ± 1.9. At 4 weeks, 191/279 (68.5%, 95% CI 62.7–73.9%) achieved ≥ 30% pain reduction and 118/279 (42.3%, 95% CI 36.4–48.3%) achieved ≥ 50% reduction. Procedure-related adverse events were documented in 26/298 (8.7%), with no documented persistent neurologic deficits or infections. At 4 weeks, 232/279 (83.2%, 95% CI 78.2–87.4%) achieved a ≥ 2-point NRS reduction and 145/279 (52.0%, 95% CI 45.9–58.0%) achieved a ≥ 4-point reduction; mean sleep-disturbance NRS decreased from 6.5 ± 1.8 at baseline to 4.2 ± 2.1 at 4 weeks.
Conclusions
In this single-center retrospective cohort, DRG-TFI was associated with clinically meaningful short-term pain reduction and no documented persistent neurologic deficit or procedure-related infection. However, chart-based adverse-event ascertainment and historical particulate-steroid exposure preclude definitive safety conclusions. Longer-term outcomes should be interpreted as observed outcomes under routine care because rescue interventions occurred during follow-up.
Supplementary Information
The online version contains supplementary material available at https://doi.org/10.1186/s12883-026-05135-6.
Keywords: Postherpetic neuralgia, Herpes zoster, Dorsal root ganglion, Transforaminal injection, Observational study, Pain outcomes
Introduction
Herpes zoster results from reactivation of latent varicella–zoster virus in cranial nerve or dorsal root ganglia, with viral spread along sensory nerves producing a painful, dermatomal rash and acute neuritis [1]. A substantial proportion of patients, particularly older adults, develop persistent zoster-associated pain that can last for months or years. Postherpetic neuralgia (PHN) is commonly defined as pain persisting more than 90 days after rash onset in the previously affected dermatome [2–4]. PHN is associated with severe pain intensity, sleep disturbance, reduced physical functioning, and frequent health-care utilization [2].
Because the principal risk factor for PHN is increasing age, the burden of disease is expected to rise in settings with rapidly aging populations [1, 2]. In China, synthesis of published epidemiologic studies suggests that herpes zoster remains common and that incidence increases substantially with age [5]. Recent nationwide survey data also indicate substantial pain-related morbidity across adult age strata [6].
Guideline-based management of PHN emphasizes antiviral therapy for acute herpes zoster and multimodal management of established neuropathic pain, typically including gabapentinoids, certain antidepressants, and topical therapies. Opioids may be considered for selected patients when first-line agents are ineffective or not tolerated [2, 3, 7, 8]. However, evidence syntheses consistently show that available medications provide incomplete pain relief for many patients and are limited by adverse effects, drug–drug interactions, and contraindications, particularly in older adults with comorbidities [7, 8]. These challenges motivate evaluation of interventional approaches that may reduce pain intensity and facilitate de-escalation of systemic analgesics. This medication framework is consistent with contemporary neuropathic-pain and PHN guidance emphasizing gabapentinoids, selected antidepressants, topical therapy, and cautious opioid use when needed [2, 3, 7, 8].
Interventional strategies for zoster-associated pain include epidural or paravertebral injections, peripheral nerve blocks, and neuromodulation techniques, with mixed evidence depending on disease stage and intervention type [9, 10]. In established PHN, the dorsal root ganglion (DRG) is a biologically plausible target because primary sensory neuron cell bodies reside within the DRG and may generate ectopic activity that sustains neuropathic pain [11]. DRG-adjacent transforaminal injection may provide short-term suppression of abnormal afferent firing via local anesthetic and may reduce perineural inflammation when corticosteroid is used. Although pharmacologic effects are time-limited, short-term interruption of afferent barrage may plausibly reduce pain amplification and enable functional recovery and medication adjustment in some patients. Prior epidural or paravertebral injection studies have mainly evaluated broader epidural/paravertebral approaches or prevention and early treatment of zoster-associated pain rather than DRG-adjacent transforaminal treatment for established PHN [9, 12]. Similarly, pulsed radiofrequency and DRG neuromodulation studies address different intervention intensity and durability questions than a single-injection DRG-TFI cohort [10, 13–15]. The DRG commonly occupies the neural foramen near the pedicle, with level-dependent anatomical variability. In this cohort, DRG-TFI denoted a target-defined modification of the transforaminal route: the intended endpoint was the posterior/inferior foraminal margin adjacent to the DRG, whereas conventional TFESI generally prioritizes periradicular/epidural spread along the exiting nerve root, often more anteriorly; the access corridor overlaps, but the intended target differs [11, 16, 17].
Despite increasing use of DRG-targeted transforaminal injection in routine care, large real-world cohorts with prespecified outcome windows, transparent procedural reporting, and systematic safety assessment remain limited. We conducted a retrospective single-center cohort study evaluating pain trajectories, responder rates, medication burden, rescue interventions, and documented adverse events after DRG-targeted transforaminal injection for PHN. Exploratory analyses assessed whether selected baseline or procedural factors were associated with short-term clinically meaningful response.
Methods
Study design and setting
This observational, non-interventional, retrospective, single-arm cohort study used routinely collected electronic medical record (EMR) data and procedure documentation from the Affiliated Hospital of North Sichuan Medical College. Index procedures were performed between January 2022 and June 2024. Eligibility criteria, follow-up windows, and the primary endpoint were specified before data extraction to reduce analytic flexibility. Because treatment was not assigned and there is no concurrent comparator, analyses are descriptive and hypothesis-generating [18].
This study was conducted in accordance with the Declaration of Helsinki. The protocol was reviewed and approved by the Ethics Committee of the Affiliated Hospital of North Sichuan Medical College, which waived the requirement for informed consent because this was a retrospective analysis of de-identified routinely collected data and posed no more than minimal risk to participants. All data were anonymized prior to analysis.
Participants eligibility
Eligible participants were consecutive adults (≥ 18 years) with PHN defined as pain in a dermatomal distribution consistent with prior herpes zoster that persisted > 90 days after documented rash onset [3, 4]. Herpes zoster history was determined from clinician-documented diagnosis with a compatible unilateral dermatomal eruption. Pain intensity was assessed using a standard published 11-point Numeric Rating Scale (NRS; 0 = no pain, 10 = worst imaginable pain) [19]. Baseline pain intensity was defined as the last recorded NRS within 7 days prior to the index procedure; baseline NRS ≥ 3 was required. PHN duration (months) was calculated as the time from documented rash onset date to the index procedure date. When rash onset was not explicitly recorded, the earliest clinician-documented patient-reported onset date at presentation was used.
Exclusion criteria were alternative dominant pain diagnoses at the treated segment documented by the treating clinician (e.g., compressive radiculopathy judged to be the primary driver of symptoms), missing baseline NRS, incomplete index procedure documentation, or no recorded follow-up encounter within the hospital system after the index procedure (precluding outcome ascertainment).
Time zero was defined as the date of the index DRG-TFI (the first DRG-targeted transforaminal injection during the study period).
DRG-targeted transforaminal injection procedure
The intervention was a DRG-targeted transforaminal injection (DRG-TFI) performed under computed tomography (CT) or fluoroscopic guidance according to operator preference and anatomical considerations. Target level(s) were selected based on dermatomal pain distribution and clinical documentation. Ropivacaine 0.25% was injected adjacent to the targeted DRG within the neural foramen; corticosteroid (dexamethasone or triamcinolone) could be added at the discretion of the treating physician. Procedural elements abstracted from the record included dermatomal region, treated level(s), number of levels treated, imaging modality, corticosteroid use and type, injectate volume per treated level, and use of periprocedural sedation. Sedation agents and depth were not consistently documented in routine records and were therefore abstracted only as use versus no use.
Procedural notes indicated use of a 22-gauge spinal needle (Quincke-type). For fluoroscopy-guided cases, an oblique transforaminal view was obtained and the needle was advanced through a posterior/infraneural trajectory; for CT-guided cases, intermittent axial CT images were used to advance the needle to the posterior aspect of the target neural foramen. The intended needle tip was adjacent to the affected DRG within the posterior/inferior foramen and outside the central canal. After negative aspiration for blood or cerebrospinal fluid, 0.3–0.5 mL nonionic contrast was injected to confirm perineural/epidural spread and exclude intravascular uptake; fluoroscopy-guided contrast injection was observed in real time. Digital subtraction angiography was not routinely documented in the EMR. Injectate was administered slowly and incrementally after repeat negative aspiration. When steroid was used, the usual per-level dose was dexamethasone 5 mg or triamcinolone 10 mg; triamcinolone was used only at thoracic/lumbar levels, and cervical steroid injections used nonparticulate dexamethasone only. This posterior/inferior ganglion-adjacent endpoint contrasts with conventional TFESI techniques that primarily seek epidural/periradicular spread around the exiting nerve root. Steroid selection reflected historical routine care; the study was not designed to compare steroid formulations or establish formulation-specific safety or efficacy.
Outcome definitions and measurement windows
The primary endpoint was within-patient change in NRS pain intensity from baseline to the 4-week assessment window, prespecified as Day 21–45 after time zero to reflect routine follow-up scheduling while limiting measurement heterogeneity. Secondary pain endpoints included NRS change at 1 week (Day 5–14), 3 months (Day 75–105), and 6 months (Day 165–210). For each window, when multiple NRS values were available, the value closest to the target day was selected.
Responder endpoints were prespecified as achieving ≥ 30% and ≥ 50% reduction from baseline NRS at each time window, consistent with IMMPACT guidance for clinically meaningful change [20]. Responder proportions are reported with exact (Clopper–Pearson) binomial 95% confidence intervals (CIs).
Sleep disturbance was assessed using a 0–10 NRS when recorded in routine pain-clinic notes and was analyzed descriptively using the same follow-up windows. Standardized physical-function instruments and patient global impression of change were not routinely captured in the EMR and therefore could not be analyzed.
Medication outcomes included changes in opioid use expressed as morphine-equivalent daily dose (MEDD, mg/day) among baseline opioid users and changes in gabapentinoid dose expressed as gabapentin-equivalent daily dose (GED, mg/day) among baseline gabapentinoid users. Baseline medication doses were abstracted from the active medication list and/or prescriptions documented within 7 days prior to the index procedure. Follow-up doses were abstracted from prescriptions or active medication lists documented within the relevant follow-up window, selecting the entry closest to the target day when multiple entries were available. GED was calculated using the commonly used approximation pregabalin mg/day ×6 = gabapentin-equivalent mg/day; GED is used as a descriptive standardization metric. Antidepressant (TCA/SNRI) and topical lidocaine patch use were abstracted as binary co-medications at baseline and follow-up because standardized dose-equivalent conversions were not available. Medication tapering or discontinuation after DRG-TFI was not protocol-mandated; changes were clinician-directed during routine follow-up based on pain relief, adverse effects, comorbidities, and patient preference. Patients generally received follow-up recommendations at approximately 1 and 4 weeks, with later visits individualized according to symptoms and rescue-treatment needs.
Rescue interventions were counted from zero through the 6-month follow-up interval. The analysis dataset retained cumulative patient-level occurrence but did not contain uniformly usable dates for each rescue procedure. Therefore, interval-specific counts and time-to-rescue estimates were not calculated. Follow-up was routine clinical care rather than a study-mandated structured safety visit. A structured chart abstraction reviewed procedure notes, immediate post-procedure/recovery documentation when available, and subsequent EMR follow-up entries; no prospective standardized adverse-event checklist or study-specific patient contact was used. Because rescue interventions may influence subsequent pain scores, outcomes beyond 4 weeks are interpreted as observed outcomes under routine care.
Covariates and exploratory predictors
Covariates for longitudinal pain summaries were prespecified to improve precision and support missing-at-random assumptions for follow-up availability. These included age, sex, PHN duration (months; also analyzed as < 6 months vs. ≥ 6 months), dermatomal region (thoracic vs. non-thoracic), baseline opioid use (yes/no), baseline gabapentinoid use (yes/no), imaging modality (CT vs. fluoroscopy), corticosteroid use (yes/no), and number of treated levels (1 vs. ≥ 2). Baseline NRS was not included as a separate covariate in the longitudinal repeated-measures model because baseline NRS is incorporated as an outcome observation. A sensitivity model additionally adjusted for baseline antidepressant (TCA/SNRI) use and baseline topical lidocaine patch use.
For the exploratory 4-week ≥ 50% responder analysis, we used a parsimonious prespecified predictor set aligned with the presented model output: age (per 10 years), sex, PHN duration < 6 months, baseline NRS, thoracic involvement, corticosteroid use, CT guidance (vs. fluoroscopy), and baseline opioid use.
Statistical analysis
Continuous variables are summarized as mean ± standard deviation (SD) or median [interquartile range, IQR] as appropriate, and categorical variables as n (%). Longitudinal NRS outcomes were summarized using observed means and responder proportions within prespecified windows. To additionally summarize trajectories using all available repeated NRS observations, we fit a linear mixed-effects model with a patient-specific random intercept and categorical time window fixed effects (baseline, 1 week, 4 weeks, 3 months, 6 months) and the prespecified covariates above. Time was modeled categorically because follow-up occurs at irregular clinical visit times and we did not want to impose a linear (or other parametric) time trend. Two-sided P values for the primary baseline-to-4-week contrast were obtained from Wald tests and are unadjusted. We emphasize inference for the prespecified baseline-to-4-week contrast, while secondary contrasts are presented descriptively without multiplicity correction. The mixed model used a patient-specific random intercept, which induces an exchangeable/compound-symmetry within-patient covariance structure; an unstructured covariance matrix was not used because follow-up was incomplete across fixed clinical windows. Model parameters were estimated by restricted maximum likelihood (REML).
Responder outcomes are reported with exact binomial 95% CIs. For the exploratory 4-week ≥ 50% responder model, we fit a ridge-penalized multivariable logistic regression (glmnet) with penalty selection by cross-validation and report odds ratios (ORs) with 95% CIs. Model discrimination was summarized using bootstrap optimism-corrected AUC in accordance with TRIPOD reporting guidance [21]. Ridge shrinkage was used to reduce overfitting in this exploratory prediction model. Calibration was assessed with a decile calibration plot, calibration-in-the-large, calibration slope, and Brier score; discrimination and calibration were internally assessed using bootstrap optimism correction.
Medication changes among baseline users were assessed using paired comparisons (Wilcoxon signed-rank tests for MEDD and paired t-tests for GED) and are interpreted descriptively; 6-month medication values are reported as observed, recognizing incomplete follow-up and potential administrative censoring. All analyses were performed using R (Version 4.4.2, R Foundation for Statistical Computing, Vienna, Austria), including lme4/lmerTest for mixed models, glmnet for penalized logistic regression, and pROC/rms for discrimination and calibration summaries.
Missing data
Patients missing baseline NRS or key eligibility variables were excluded because the primary endpoint requires a defined baseline. For follow-up outcomes, analyses are reported using the number of patients with an NRS recorded within each prespecified window; we did not impute missing pain scores. Reasons for missing longer-term follow-up were not uniformly recorded; therefore, no inverse-probability weighting was performed, and attrition is addressed as a potential source of bias.
Sample size and precision
Sample size was determined by the number of eligible patients during the accrual period. Because this is a single-arm retrospective cohort, we emphasize clinically meaningful effect sizes with uncertainty (CIs) rather than relying on statistical significance across multiple endpoints.
Results
Study cohort and follow-up availability
Among 324 screened patients, 26 were excluded (zoster-associated pain duration < 90 days after rash onset, n = 14; missing baseline NRS, n = 6; incomplete index documentation, n = 4; and alternative dominant pain diagnosis, n = 2), yielding 298 included patients (Fig. 1). Mean age was 65.7 ± 10.9 years, 173/298 (58.1%) were female, median PHN duration was 6.2 months [4.1, 9.8], and thoracic involvement predominated (216/298, 72.5%) (Table 1). Baseline mean NRS was 7.3 ± 1.1, with 214/298 (71.8%) using gabapentinoids and 82/298 (27.5%) using opioids (Table 1). NRS follow-up was available for 270/298 (90.6%) at 1 week, 279/298 (93.6%) at 4 weeks, 220/298 (73.8%) at 3 months, and 191/298 (64.1%) at 6 months (Fig. 1).
Fig. 1.
Cohort flow diagram. Flow diagram showing identification of eligible patients and completeness of follow-up assessments. Numbers indicate patients with an available NRS measurement within each prespecified window
Table 1.
Baseline characteristics of the study cohort
| Characteristic | Value |
|---|---|
| Age, years | 65.7 ± 10.9 |
| Female sex | 173 (58.1%) |
| Body mass index, kg/m2 | 24.1 ± 3.6 |
| PHN duration, months | 6.2 [4.1, 9.8] |
| Dermatomal region involved | |
| Cervical | 34 (11.4%) |
| Thoracic | 216 (72.5%) |
| Lumbar | 48 (16.1%) |
| Baseline NRS pain (0-10) | 7.3 ± 1.1 |
| Baseline sleep disturbance NRS (0-10) | 6.5 ± 1.8 |
| Baseline gabapentinoid use | 214 (71.8%) |
| Baseline opioid use | 82 (27.5%) |
| Baseline antidepressant (TCA/SNRI) use | 94 (31.5%) |
| Baseline topical lidocaine patch use | 57 (19.1%) |
| Hypertension | 124 (41.6%) |
| Diabetes mellitus | 68 (22.8%) |
| Cardiovascular disease | 49 (16.4%) |
| Chronic kidney disease | 21 (7.0%) |
| Immunosuppression* | 19 (6.4%) |
| Current smoker | 52 (17.4%) |
| Prior herpes zoster vaccination | 18 (6.0%) |
Values are mean ± SD, median [IQR], or n(%). *Includes systemic corticosteroid or other immunosuppressive therapy within the prior 3 months
Primary and secondary pain outcomes
Observed mean NRS decreased from 7.3 ± 1.1 at baseline to 4.1 ± 1.9 at 4 weeks. In the mixed-effects model, the estimated mean change from baseline to 4 weeks was − 3.1 points (95% CI − 3.4 to − 2.8; P < 0.001), meeting the prespecified primary endpoint (Table 2). Improvements were also observed at 1 week (− 2.5; 95% CI − 2.8 to − 2.2), 3 months (− 2.9; 95% CI − 3.3 to − 2.5), and 6 months (− 2.6; 95% CI − 3.0 to − 2.2) (Table 2 and Fig. 2). A sensitivity mixed model additionally adjusted for baseline TCA/SNRI and lidocaine patch use yielded a similar 4-week NRS estimate (− 3.0 points; 95% CI − 3.3 to − 2.7).
Table 2.
Clinical outcome. Pain outcomes over time
| Timepoint (window) | N assessed | NRS mean ± SD | Observed mean change | Adjusted mean change (95% CI)† | P value |
|---|---|---|---|---|---|
| Baseline (≤7 days pre-procedure) | 298 | 7.3 ± 1.1 | — | — | — |
| 1 week (Day 5-14) | 270 | 4.8 ± 1.8 | -2.5 | -2.5 (-2.8 to -2.2) | <0.001 |
| 4 weeks (Day 21-45) [Primary] | 279 | 4.1 ± 1.9 | -3.2 | -3.1 (-3.4 to -2.8) | <0.001 |
| 3 months (Day 75-105) | 220 | 4.4 ± 2.0 | -2.9 | -2.9 (-3.3 to -2.5) | <0.001 |
| 6 months (Day 165-210) | 191 | 4.7 ± 2.1 | -2.6 | -2.6 (-3.0 to -2.2) | <0.001 |
† From a covariate-adjusted linear mixed-effects model with patient-specific random intercept and categorical time window fixed effects. Baseline NRS is modeled as a repeated outcome observation and was not included as a separate covariate
Fig. 2.
Mean NRS pain trajectory over time. Line plot of observed mean NRS pain (0–10) with 95% CIs across prespecified follow-up windows after the index DRG-TFI. Windows correspond to baseline (≤7 days pre-procedure) and follow-up assessments at 1 week (Day 5–14), 4 weeks (Day 21–45), 3 months (Day 75–105), and 6 months (Day 165–210). Ns at each time window were: baseline n=298; 1 week n=270; 4 weeks n=279; 3 months n=220; and 6 months n=191
Clinically meaningful responder rates
At 4 weeks, 191/279 (68.5%, 95% CI 62.7–73.9%) achieved ≥ 30% NRS reduction and 118/279 (42.3%, 95% CI 36.4–48.3%) achieved ≥ 50% reduction (Table 3). Early improvement was common at 1 week (≥ 30%: 198/270, 73.3%; ≥50%: 129/270, 47.8%). Among those with observed longer-term data, responder rates were 145/220 (65.9%) and 86/220 (39.1%) at 3 months, and 109/191 (57.1%) and 71/191 (37.2%) at 6 months for ≥ 30% and ≥ 50% thresholds, respectively (Table 3).
Table 3.
Clinical outcome. Responder rates
| Timepoint | N assessed | ≥30% NRS reduction | Exact 95% CI | ≥50% NRS reduction | Exact 95% CI |
|---|---|---|---|---|---|
| 1 week (Day 5-14) | 270 | 198/270 (73.3%) | 67.6–78.5% | 129/270 (47.8%) | 41.7–53.9% |
| 4 weeks (Day 21-45) | 279 | 191/279 (68.5%) | 62.7–73.9% | 118/279 (42.3%) | 36.4–48.3% |
| 3 months (Day 75-105) | 220 | 145/220 (65.9%) | 59.2–72.1% | 86/220 (39.1%) | 32.6–45.9% |
| 6 months (Day 165-210) | 191 | 109/191 (57.1%) | 49.7–64.2% | 71/191 (37.2%) | 30.3–44.4% |
Using absolute pain-reduction thresholds, ≥ 2-point NRS reductions were observed in 214/270 (79.3%) at 1 week, 232/279 (83.2%) at 4 weeks, 176/220 (80.0%) at 3 months, and 143/191 (74.9%) at 6 months; corresponding ≥ 4-point reductions were 113/270 (41.9%), 145/279 (52.0%), 100/220 (45.5%), and 79/191 (41.4%), respectively (Table 4).
Table 4.
Clinical outcome. Absolute NRS reduction thresholds
| Timepoint | N assessed | ≥2-point NRS reduction | Exact 95% CI | ≥4-point NRS reduction | Exact 95% CI |
|---|---|---|---|---|---|
| 1 week (Day 5-14) | 270 | 214/270 (79.3%) | 73.9–83.9% | 113/270 (41.9%) | 35.9–48.0% |
| 4 weeks (Day 21-45) | 279 | 232/279 (83.2%) | 78.2–87.4% | 145/279 (52.0%) | 45.9–58.0% |
| 3 months (Day 75-105) | 220 | 176/220 (80.0%) | 74.1–85.1% | 100/220 (45.5%) | 38.7–52.3% |
| 6 months (Day 165-210) | 191 | 143/191 (74.9%) | 68.1–80.9% | 79/191 (41.4%) | 34.3–48.7% |
Among patients with recorded sleep scores, mean sleep-disturbance NRS decreased from 6.5 ± 1.8 at baseline to 4.2 ± 2.1 at 4 weeks; the adjusted mean change was − 2.2 (95% CI − 2.5 to − 1.9; P < 0.001) (Table 5).
Table 5.
Clinical outcome. Sleep-disturbance outcomes
| Timepoint (window) | N assessed | Sleep-disturbance NRS mean ± SD | Observed mean change | Adjusted mean change (95% CI)† | P value |
|---|---|---|---|---|---|
| Baseline (≤7 days pre-procedure) | 298 | 6.5 ± 1.8 | — | — | — |
| 1 week (Day 5-14) | 248 | 4.7 ± 2.1 | -1.8 | -1.8 (-2.1 to -1.5) | <0.001 |
| 4 weeks (Day 21-45) | 252 | 4.2 ± 2.1 | -2.3 | -2.2 (-2.5 to -1.9) | <0.001 |
| 3 months (Day 75-105) | 193 | 4.5 ± 2.2 | -2.0 | -1.9 (-2.3 to -1.5) | <0.001 |
| 6 months (Day 165-210) | 166 | 5.0 ± 2.3 | -1.5 | -1.5 (-1.9 to -1.1) | <0.001 |
Analgesic burden and rescue interventions
Among baseline opioid users (n = 82), median MEDD decreased from 22.0 mg/day [15.0, 35.0] at baseline to 12.5 mg/day [0.0, 25.0] at 4 weeks (P < 0.001), with 27/82 (32.9%) discontinuing opioids by 4 weeks (Table 6). At 6 months, median MEDD among baseline opioid users was 14.0 mg/day [0.0, 30.0] and 30/82 (36.6%) had discontinued opioids (Table 6). These 6-month values are descriptive given incomplete follow-up. Among baseline gabapentinoid users (n = 214), mean GED decreased from 915 ± 420 mg/day at baseline to 754 ± 410 mg/day at 4 weeks (− 17.6%; P < 0.001) and was 808 ± 430 mg/day at 6 months (− 11.7%) (Table 6). Because TCA/SNRI and lidocaine patch doses were not standardized in the EMR, these co-medications are presented as use trajectories rather than dose reductions: among baseline TCA/SNRI users, 86/94 (91.5%) continued use at 4 weeks and 76/94 (80.9%) at 6 months; among baseline lidocaine patch users (n = 57), 49/57 (86.0%) and 41/57 (71.9%) continued use at 4 weeks and 6 months, respectively (Table 6).
Table 6.
Analgesic burden and rescue interventions
| Outcome | Baseline | 4 weeks | 6 months | P value (baseline vs 4 weeks)† |
|---|---|---|---|---|
| Baseline opioid users, n | 82 | |||
| MEDD (mg/day), median [IQR] | 22.0 [15.0, 35.0] | 12.5 [0.0, 25.0] | 14.0 [0.0, 30.0] | <0.001 |
| Opioid stopped, n (%) | — | 27/82 (32.9%) | 30/82 (36.6%) | — |
| Baseline gabapentinoid users, n | 214 | |||
| GED (mg/day), mean ± SD | 915 ± 420 | 754 ± 410 | 808 ± 430 | <0.001 |
| GED percent change vs baseline | — | -17.6% | -11.7% | — |
| Baseline TCA/SNRI users, n | 94 | |||
| TCA/SNRI continued, n (%) | — | 86/94 (91.5%) | 76/94 (80.9%) | — |
| Baseline topical lidocaine patch users, n | 57 | |||
| Lidocaine patch continued, n (%) | — | 49/57 (86.0%) | 41/57 (71.9%) | — |
| Rescue interventions during6-month follow up from time zero* | ||||
| Repeat DRG-TFI | 52 (17.4%) | |||
| Pulsed radiofrequency | 23 (7.7%) | |||
| Spinal cord stimulation | 6 (2.0%) | |||
| Other peripheral nerve blocks | 12 (4.0%) |
MEDD morphine-equivalent daily dose, GED gabapentin-equivalent daily dose
† Within-user paired comparisons among baseline users; P values correspond to baseline versus 4-week comparisons for MEDD and GED. Six-month values are descriptive
* Rescue interventions are cumulative patient-level occurrences from time zero through the 6-month follow-up; exact event dates were not uniformly available in the analytic extract, so interval-specific timing and time-to-event statistics are not reported
Within 6 months, 52/298 (17.4%) underwent repeat DRG-TFI and 23/298 (7.7%) underwent pulsed radiofrequency as rescue interventions, while spinal cord stimulation was uncommon (6/298, 2.0%) (Table 6). Because rescue interventions can influence later pain scores, 3- and 6-month pain outcomes should be interpreted as observed outcomes under usual-care management after the index procedure rather than effects attributable solely to the index DRG-TFI.
Procedural characteristics and documented adverse safety
CT guidance was used in 194/298 (65.1%) cases and corticosteroid was included in 227/298 (76.2%); most procedures treated one level (201/298, 67.4%), with 78/298 (26.2%) treating two levels and 19/298 (6.4%) treating ≥ 3 levels (Table 7). Any documented procedure-related adverse event occurred in 26/298 (8.7%; 95% CI 6.0% to 12.5%), most commonly transient dizziness (3.0%), transient sensory disturbance (2.3%), and vasovagal reactions (1.7%) (Table 7). Less frequent events included transient motor weakness (0.7%) and isolated reports of headache, local hematoma, and transient blood pressure elevation (each 0.3%), with no documented persistent neurologic deficit or procedure-related infection (Table 7). Three procedures (1.0%) documented intravascular contrast uptake during test injection; in each case, the needle was repositioned and therapeutic injection was not delivered until repeat contrast showed nonvascular spread.
Table 7.
Procedural characteristics and documented adverse events
| Characteristic | Value |
|---|---|
| Imaging guidance | |
| CT-guided | 194 (65.1%) |
| Fluoroscopy-guided | 104 (34.9%) |
| Injectate corticosteroid | |
| Yes | 227 (76.2%) |
| No | 71 (23.8%) |
| Steroid type among steroid group (n=227) | |
| Dexamethasone | 163 (71.8%) |
| Triamcinolone | 64 (28.2%) |
| Local anesthetic | Ropivacaine 0.25% (100%) |
| Injectate volume per treated level, mL | 1.7 ± 0.4 |
| Number of levels treated | |
| 1 level | 201 (67.4%) |
| 2 levels | 78 (26.2%) |
| >=3 levels | 19 (6.4%) |
| Periprocedural sedation used | 45 (15.1%) |
| Any documented procedure-related AE |
26 (8.7%) 95% CI 6.0% to 12.5% |
| Transient dizziness | 9 (3.0%) |
| Transient sensory disturbance | 7 (2.3%) |
| Vasovagal reaction | 5 (1.7%) |
| Transient motor weakness | 2 (0.7%) |
| Headache | 1 (0.3%) |
| Local hematoma | 1 (0.3%) |
| Transient blood pressure elevation | 1 (0.3%) |
| Persistent neurologic deficit | 0 (0.0%) |
| Procedure-related infection | 0 (0.0%) |
Values are n (%) unless otherwise stated. Adverse events were retrospectively identified from routine documentation without a standardized prospective checklist; counts are documented events, not complete incidence estimates. Regional counts by steroid formulation were not available in the analytic extract
Exploratory prognostic model for 4-week clinically meaningful response
In an exploratory ridge-penalized logistic regression model for 4-week ≥ 50% response (events = 118/279), shorter PHN duration (< 6 months) and higher baseline NRS were associated with higher odds of response (PHN duration < 6 months: OR 1.65, 95% CI 1.05 to 2.58; baseline NRS: OR 1.25 per 1-point increase, 95% CI 1.05 to 1.48) (Fig. 3A). Model discrimination was moderate (bootstrap-corrected AUC 0.72) (Fig. 3B). Calibration assessment showed calibration-in-the-large of 0.03, bootstrap-corrected calibration slope of 0.91, and Brier score of 0.20 (Fig. 3C).
Fig. 3.
Exploratory predictors, discrimination, and calibration of 4-week clinically meaningful response: A forest plot of penalized odds ratios (ORs) with 95% CIs from a multivariable ridge-penalized logistic regression model for 4-week clinically meaningful response (≥50% NRS reduction). B ROC curve for the same model as described in (A). C Decile calibration plot comparing mean predicted probability with observed response proportion
Discussion
In this retrospective single-center cohort of patients with PHN treated with DRG-targeted transforaminal injection, we observed clinically meaningful short-term pain reduction and no documented persistent neurologic deficit or infection,. The primary endpoint was met with an estimated mean NRS decrease of − 3.1 points at 4 weeks, and 42.3% of patients with 4-week follow-up achieved ≥ 50% pain reduction. Analgesic burden decreased among baseline opioid and gabapentinoid users, and rescue interventions occurred in a subset of patients during follow-up. The absolute-threshold analyses and sleep-disturbance outcomes provide a broader description of clinical change beyond mean pain scores alone.
The biological plausibility of targeting the DRG in PHN is supported by viral latency in sensory ganglia and the potential for ganglionitis and neuronal injury to sustain neuropathic pain beyond rash resolution [1, 2, 11]. Operationally, the study target was the posterior/inferior foramen adjacent to the DRG, whereas conventional TFESI generally seeks periradicular/epidural spread along the exiting nerve root; the approaches share a transforaminal corridor but differ in intended endpoint (Supplementary Figure S1). A DRG-adjacent transforaminal injection may transiently suppress abnormal afferent firing and, when corticosteroid is used, may reduce perineural inflammation. While anesthetic effects are short-lived, short-term interruption of afferent barrage may plausibly reduce pain amplification and support functional recovery and medication adjustment in some patients. These findings should not be interpreted as evidence that DRG-TFI is superior to epidural/paravertebral injection, pulsed radiofrequency, or neuromodulation. Epidural steroid literature for zoster-related pain often addresses acute-phase prevention or early pain using broader epidural spread, whereas pulsed radiofrequency and DRG/SCS neuromodulation involve different mechanisms, invasiveness, and durability expectations [9, 10, 12–15]. The main contribution of the present cohort is therefore not comparative efficacy, but a larger routine-care description of DRG-adjacent injection outcomes, medication/rescue trajectories, sleep disturbance, and safety documentation in established PHN.
Contemporary guidance places interventional neuropathic-pain care after optimized pharmacologic and topical management, while acknowledging that the supporting evidence is heterogeneous and generally limited. NeuPSIG made only weak recommendations for selected invasive neuropathic-pain procedures, and PHN-specific systematic reviews and recent meta-analyses suggest possible benefit from pulsed radiofrequency or short-term spinal cord stimulation in selected refractory patients but emphasize uncertainty [7, 22–26]. DRG-TFI should therefore be viewed as a lower-intensity, hypothesis-generating adjunct rather than a guideline-established substitute for standard therapy.
Our longer-term pain outcomes must be interpreted cautiously. Rescue interventions (repeat DRG-TFI, pulsed radiofrequency, and other procedures) occurred during follow-up and may contribute to 3- and 6-month pain scores. Therefore, later outcomes represent observed trajectories under routine care rather than effects attributable solely to the index procedure. In addition, follow-up completeness at 6 months may reflect both loss to follow-up and administrative censoring depending on the EMR extraction date. Medication reductions are clinically salient because pharmacologic management of PHN is frequently limited by incomplete efficacy and adverse effects [2, 3, 7, 8]. Medication tapering was not protocolized and may be influenced by clinician practice and patient preference; nonetheless, the observed reductions are consistent with meaningful symptom improvement in some patients. Safety is central when offering transforaminal interventions to older adults. Although no documented persistent neurologic deficits or infections were identified, rare serious neurologic complications have been reported after epidural corticosteroid injections, underscoring the importance of standardized technique reporting, appropriate steroid selection, avoidance of deep sedation, and transparent adverse-event ascertainment in future prospective studies [16, 27]. Importantly, this cohort does not establish the safety of thoracic triamcinolone. Safety recommendations favor nonparticulate dexamethasone for transforaminal injections above L3, and a recent meta-analysis found no meaningful efficacy advantage of particulate steroid for lumbar radicular pain [16, 17, 27, 28]. We therefore do not endorse particulate corticosteroid for cervical or thoracic DRG-TFI; where steroid is used at these levels, a nonparticulate formulation should be preferred. The largest mean NRS reduction at 4 weeks may reflect the time course of local anesthetic plus corticosteroid effects, clinician-directed medication adjustment, regression to the mean or natural fluctuation, and selective follow-up. The higher 1-week responder proportions despite a smaller mean change likely reflect threshold behavior: many patients crossed percentage reduction thresholds early, whereas larger reductions in a subset by 4 weeks increased the mean change. At 6 months, missingness could plausibly operate in either direction: patients with good relief might not return, whereas nonresponders might seek outside care or receive rescue interventions, producing survivorship bias.
Our exploratory prognostic analysis suggested that shorter PHN duration and higher baseline pain intensity were associated with higher odds of short-term response. These patterns are clinically plausible, but percent-reduction endpoints are mathematically coupled to baseline NRS, and the model was internally validated only. External validation and prospective impact assessment are required before any clinical use [21].
Limitations of our study include the single-arm observational design, procedural heterogeneity, incomplete follow-up, reliance on EMR documentation for medication and safety outcomes, and lack of standardized neuropathic pain questionnaires and quality-of-life measures in routine records [29]. Without a concurrent comparator, observed improvement could reflect the procedure, concurrent medication changes, regression to the mean, or the underlying course of PHN. Adverse-event ascertainment was based on routine chart documentation rather than a standardized prospective checklist. Transient numbness, headache, injection-site discomfort, or other self-limited symptoms may therefore have been missed, and the 8.7% documentation rate should not be interpreted as a complete incidence estimate. Exact dates of rescue procedures were not uniformly available in the analytic extract, preventing time-to-rescue analysis. Referral bias and intervention-candidacy bias are possible because patients offered DRG-TFI may differ from medication-managed PHN patients. Survivorship bias is also possible because later outcomes include only patients with recorded follow-up. Although sleep-disturbance scores were available, standardized physical-function instruments and patient global impression of change were not collected consistently.
In conclusion, DRG-targeted transforaminal injection was associated with clinically meaningful short-term pain reduction, reductions in analgesic burden among baseline users, and no documented persistent neurologic deficit or infection. Prospective comparative studies are needed to clarify effectiveness, optimize patient selection, and define the role of DRG-TFI within guideline-based PHN care.
Supplementary Information
Acknowledgements
Not applicable.
Authors’ contributions
The authors confirm contribution to the paper as follows: study conception and design: L.X.; data collection: H.T., Y.P., L.H.; analysis and interpretation of results: H.T., Y.P., L.H.; draft manuscript preparation: H.T., Y.P., L.H., L.X. All authors reviewed the results and approved the final version of the manuscript.
Funding
No funding.
Data availability
Data sets generated during the current study are available from the corresponding author on reasonable request.
Declarations
Ethics approval and consent to participate
This study was conducted in accordance with the Declaration of Helsinki. The protocol was reviewed and approved by the Ethics Committee of the Affiliated Hospital of North Sichuan Medical College, which waived the requirement for informed consent because this was a retrospective analysis of de-identified routinely collected data and posed no more than minimal risk to participants. All data were anonymized prior to analysis.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Cohen JI. Clinical practice: Herpes zoster. N Engl J Med. 2013;369(3):255–63. 10.1056/NEJMcp1302674. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Johnson RW, Rice ASC. Clinical practice: Postherpetic neuralgia. N Engl J Med. 2014;371(16):1526–33. 10.1056/NEJMcp1403062. [DOI] [PubMed] [Google Scholar]
- 3.Gross GE, Eisert L, Doerr HW, Fickenscher H, Knuf M, Maier P, Maschke M, Müller R, Pleyer U, Schäfer M, Sunderkötter C, Werner RN, Wutzler P, Nast A. S2k guidelines for the diagnosis and treatment of herpes zoster and postherpetic neuralgia. J Dtsch Dermatol Ges. 2020;18(1):55–78. 10.1111/ddg.14013. [DOI] [PubMed] [Google Scholar]
- 4.Centers for Disease Control and Prevention. (2024, May 10). Clinical features of shingles (herpes zoster). Retrieved February 14, 2026.
- 5.Zhang Z, Liu X, Suo L, Zhao D, Pan J, Lu L. The incidence of herpes zoster in China: A meta-analysis and evidence quality assessment. Hum Vaccines Immunotherapeutics. 2023;19(2):2228169. 10.1080/21645515.2023.2228169. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Xia Y, Ye X, Zhu W, Ai J, Shen Y, Shi Z, Cui C, Yuan Y, Zhu D, He P. Epidemiology of herpes zoster and post-herpetic neuralgia in China: A nationwide population-based survey. Int J Infect Dis. 2025;159:108005. 10.1016/j.ijid.2025.108005. [DOI] [PubMed] [Google Scholar]
- 7.Soliman N, Moisset X, Ferraro MC, Ciampi de Andrade D, Baron R, Belton J, Bennett DLH, Calvo M, Dougherty P, Gilron I, Hietaharju AJ, Hosomi K, Kamerman PR, Kemp H, Enax-Krumova EK, McNicol E, Price TJ, Raja SN, Rice ASC, Smith BH, Talkington F, Truini A, Vollert J, Attal N, Finnerup NB, Haroutounian S, NeuPSIG Review Update Study Group. Pharmacotherapy and non-invasive neuromodulation for neuropathic pain: A systematic review and meta-analysis (updated NeuPSIG recommendations). Lancet Neurol. 2025;24(5):413–28. 10.1016/S1474-4422(25)00068-7. [DOI] [PubMed] [Google Scholar]
- 8.Guo Z, Xia Y, Zhang Z. Efficacy and safety of different medications compared for the treatment of postherpetic neuralgia: A network meta-analysis. Front Pharmacol. 2025;16:1614587. 10.3389/fphar.2025.1614587. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.van Wijck AJM, Opstelten W, Moons KGM, van Essen GA, Stolker RJ, Kalkman CJ, Verheij TJM. The PINE study of epidural steroids and local anaesthetics to prevent postherpetic neuralgia: A randomised controlled trial. Lancet. 2006;367(9506):219–24. 10.1016/S0140-6736(06)68032-X. [DOI] [PubMed] [Google Scholar]
- 10.Rui M, Ni H, Xie K, Xu L, Yao M. Progress in radiofrequency therapy for zoster-associated pain about parameters, modes, targets, and combined therapy: A narrative review. Pain Therapy. 2024;13(1):23–32. 10.1007/s40122-023-00561-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Devor M. Rethinking the causes of pain in herpes zoster and postherpetic neuralgia: The ectopic pacemaker hypothesis. Pain Rep. 2018;3(6):e702. 10.1097/PR9.0000000000000702. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Nahm FS, Choi E, Han WK, Lee HJ, Gil HY, Kim JH, Lee PB, Ju H. Transforaminal epidural steroid injection for zoster-related pain: The golden period for the best outcome. Pain Physician. 2021;24(5):E669–76. [PubMed] [Google Scholar]
- 13.Sun CL, Li XL, Li CW, He N, Zhang J, Xue FS. High-voltage, long-duration pulsed radiofrequency to the dorsal root ganglion provides improved pain relief for herpes zoster neuralgia in the subacute stage. Pain Physician. 2023;26(3):E155–62. [PubMed] [Google Scholar]
- 14.Kim JH, Apigo A, Fontaine C. Dorsal root ganglion stimulation for refractory post-herpetic neuralgia. Pain Pract. 2021;21(7):794–8. 10.1111/papr.13017. [DOI] [PubMed] [Google Scholar]
- 15.Verma A, Francois E, Maiti T, Cassidy L, Tolba R. Dorsal root ganglion stimulator-a targeted therapy for post-herpetic neuralgia: The Middle East experience. Pain Pract. 2024;24(3):567–72. 10.1111/papr.13325. [DOI] [PubMed] [Google Scholar]
- 16.Rathmell JP, Benzon HT, Dreyfuss P, Huntoon M, Wallace M, Baker R, Riew KD, Rosenquist RW, Aprill C, Rost NS, Buvanendran A, Kreiner DS, Bogduk N, Fourney DR, Fraifeld E, Horn S, Stone J, Vorenkamp K, Lawler G, Summers J, Kloth D, O’Brien D, Jr., Tutton S. Safeguards to prevent neurologic complications after epidural steroid injections: Consensus opinions from a multidisciplinary working group and national organizations. Anesthesiology. 2015;122(5):974–84. 10.1097/ALN.0000000000000614. [DOI] [PubMed] [Google Scholar]
- 17.Van Boxem K, Rijsdijk M, Hans G, de Jong J, Kallewaard JW, Vissers K, van Kleef M, Rathmell JP, Van Zundert J. Safe use of epidural corticosteroid injections: Recommendations of the WIP Benelux Work Group. Pain Pract. 2019;19(1):61–92. 10.1111/papr.12709. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.von Elm E, Altman DG, Egger M, Pocock SJ, Gøtzsche PC, Vandenbroucke JP, Initiative STROBE. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: Guidelines for reporting observational studies. Lancet. 2007;370(9596):1453–7. 10.1016/S0140-6736(07)61602-X. [DOI] [PubMed] [Google Scholar]
- 19.Hartrick CT, Kovan JP, Shapiro S. The numeric rating scale for clinical pain measurement: a ratio measure? Pain Pract. 2003;3(4):310–6. 10.1111/j.1530-7085.2003.03034.x. [DOI] [PubMed] [Google Scholar]
- 20.Dworkin RH, Turk DC, Wyrwich KW, Beaton D, Cleeland CS, Farrar JT, Haythornthwaite JA, Jensen MP, Kerns RD, Ader DN, Brandenburg N, Burke LB, Cella D, Chandler J, Cowan P, Dimitrova R, Dionne R, Hertz S, Jadad AR, Katz NP, Kehlet H, Kramer LD, Manning DC, McCormick C, McDermott MP, McQuay HJ, Patel S, Porter L, Quessy S, Rappaport BA, Rauschkolb C, Revicki DA, Rothman M, Schmader KE, Stacey BR, Stauffer JW, von Stein T, White RE, Witter J, Zavisic S. Interpreting the clinical importance of treatment outcomes in chronic pain clinical trials: IMMPACT recommendations. J Pain. 2008;9(2):105–21. 10.1016/j.jpain.2007.09.005. [DOI] [PubMed] [Google Scholar]
- 21.Collins GS, Moons KGM, Dhiman P, Riley RD, Beam AL, Van Calster B, Ghassemi M, Liu X, Reitsma JB, van Smeden M, Boulesteix A-L, Camaradou JC, Celi LA, Denaxas S, Denniston AK, Glocker B, Golub RM, Harvey H, Heinze G, Hoffman MM, Kengne AP, Lam E, Lee N, Loder EW, Maier-Hein L, Mateen BA, McCradden MD, Oakden-Rayner L, Ordish J, Parnell R, Rose S, Singh K, Wynants L, Logullo P. TRIPOD + AI statement: Updated guidance for reporting clinical prediction models that use regression or machine learning methods. BMJ. 2024;385:e078378. 10.1136/bmj-2023-078378. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Dworkin RH, O’Connor AB, Kent J, Mackey SC, Raja SN, Stacey BR, Levy RM, Backonja M, Baron R, Harke H, Loeser JD, Treede RD, Turk DC, Wells CD. Interventional management of neuropathic pain: NeuPSIG recommendations. Pain. 2013;154(11):2249–61. 10.1016/j.pain.2013.06.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Lin CS, Lin YC, Lao HC, Chen CC. Interventional treatments for postherpetic neuralgia: A systematic review. Pain Physician. 2019;22(3):209–28. 10.36076/ppj/2019.22.209. [DOI] [PubMed] [Google Scholar]
- 24.Xue S, Yang WJ, Cao ZX, Sun T. Comparing the efficacy and safety of short-term spinal cord stimulation and pulsed radiofrequency for zoster-related pain: A systematic review and meta-analysis. Medicine. 2022;101(11):e29073. 10.1097/MD.0000000000029073. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Liu Z, Weng Y, Liu F, Jiang D, Wu C, Chen Y, Duan X, Zhong Q. Efficacy and safety of short-term spinal cord stimulation and pulsed radiofrequency in the treatment of postherpetic neuralgia: A meta-analysis. Front Neurol. 2025;16:1586995. 10.3389/fneur.2025.1586995. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Serrano-Afonso A, Gálvez R, Paramés E, Navarro A, Ochoa D, Pérez-Hernández C. Update on interventional management of neuropathic pain: A Delphi consensus of the Spanish Pain Society Neuropathic Pain Task Force. Medicina. 2022;58(5):627. 10.3390/medicina58050627. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Food US, Administration D. (2014, April 23). FDA drug safety communication: FDA requires label changes to warn of rare but serious neurologic problems after epidural corticosteroid injections for pain.
- 28.Kim SJ, Park JM, Kim YW, Yoon SY, Lee SC. Comparison of particulate steroid injection vs nonparticulate steroid injection for lumbar radicular pain: A systematic review and meta-analysis. Arch Phys Med Rehabil. 2024;105(9):1756–69. 10.1016/j.apmr.2024.01.002. [DOI] [PubMed] [Google Scholar]
- 29.Truini A, Aleksovska K, Anderson CC, Attal N, Baron R, Bennett DL, Bouhassira D, Cruccu G, Eisenberg E, Enax-Krumova E, Davis KD, Di Stefano G, Finnerup NB, Garcia-Larrea L, Hanafi I, Haroutounian S, Karlsson P, Rakusa M, Rice ASC, Sachau J, Smith BH, Sommer C, Tölle T, Valls-Solé J, Veluchamy A. Joint European Academy of Neurology–European Pain Federation–Neuropathic Pain Special Interest Group of the International Association for the Study of Pain guidelines on neuropathic pain assessment. Eur J Neurol. 2023;30(8):2177–96. 10.1111/ene.15831. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
Data sets generated during the current study are available from the corresponding author on reasonable request.