OUP user menu

A Multicenter, Randomized, Double-Blind, Controlled Study of NGX-4010, a High-Concentration Capsaicin Patch, for the Treatment of Postherpetic Neuralgia

Gordon A. Irving MD, Miroslav M. Backonja MD, Edwin Dunteman MD, E. Richard Blonsky MD, Geertrui F. Vanhove MD, PhD, Shiao-Ping Lu MS, Jeffrey Tobias MD
DOI: http://dx.doi.org/10.1111/j.1526-4637.2010.01004.x 99-109 First published online: 1 January 2011


Objectives. To confirm the efficacy, tolerability, and safety of NGX-4010, an 8% capsaicin dermal patch (capsaicin 640 µg/cm2), in patients with postherpetic neuralgia (PHN). PHN is a chronic pain disorder that can be difficult to treat and for which current treatment options are often limited by poor tolerability.

Design. A total of 418 patients were randomized to receive a single 60-minute application of NGX-4010 or a 0.04% capsaicin control patch (3.2 µg/cm2) in a multicenter, double-blind, confirmatory, phase 3 study.

Patients. Patients were 18–90 years old with a diagnosis of PHN, pain for at least 6 months, and an average baseline Numeric Pain Rating Scale (NPRS) score of 3–9.

Outcome Measures. The primary efficacy end point was the percentage change in NPRS score from baseline to weeks 2–8.

Results. NGX-4010 recipients had a significantly greater mean reduction from baseline in pain during weeks 2–8 compared with the control group (32.0% vs 24.4%; P = 0.011). A ≥30% reduction in mean NPRS scores was achieved in 46% of NGX-4010 recipients compared with 34% of controls (P = 0.02). Pain was significantly lower in NGX-4010 recipients than controls by week 2, and greater pain reduction was maintained throughout the remaining 12-week study period. Most treatment-emergent adverse events were application site specific (notably erythema and pain), transient, and generally mild to moderate in severity.

Conclusions. In patients with PHN, a single 60-minute application of NGX-4010 produced significant reduction in pain that was maintained over a 12-week period.

  • Neuropathic Pain
  • Postherpetic Neuralgia
  • Capsaicin


Postherpetic neuralgia (PHN) is the most common complication of herpes zoster, which results from reactivation of latent varicella zoster virus that causes chicken pox. It is a chronic pain disorder that often leads to considerable disability and suffering. The definition of PHN varies, but it is typically defined as the presence of pain lasting more than 3 months from the onset of skin lesions [1–3].

PHN occurs in about 8–19% of patients with herpes zoster, although prevalence increases with age. It is estimated that between 25% and 50% of patients with herpes zoster older than 50 years and up to 75% of patients with herpes zoster over 70 years develop PHN [1,3,4].

PHN has a significant impact on quality of life as a result of severe and debilitating pain and interference with daily functioning [3,5,6]. Common treatment options include anticonvulsants such as pregabalin and gabapentin, topical lidocaine, opioids, tricyclic antidepressants, and selective serotonin and norepinephrine reuptake inhibitors. However, the use of many of these treatments is often limited by poor tolerability, the need for titration, and administration of multiple daily doses. In addition, many patients continue to experience significant pain while taking these treatments [7–10], stressing the need for more effective pain management.

Improved understanding of the pathophysiological mechanisms of neuropathic pain may enable a more specific pain mechanism-based approach in managing PHN. Because sensitization of peripheral nociceptors that express transient receptor potential vanilloid 1 receptor (TRPV1) is evident in many patients with PHN [11], targeting of TRPV1 would seem a logical approach to pain management. Capsaicin activates TRPV1 ligand-gated cation channels on nociceptive nerve fibers, causing depolarization, action potential initiation, and the transmission of pain signals to the spinal cord. Continued capsaicin exposure leads to reversible defunctionalization of and associated decrease in epidermal TRPV1-expressing sensory nerve endings, inhibiting the initiation of pain transmission [12–14].

Although low-concentration capsaicin creams (0.025% and 0.075%) have demonstrated efficacy in the treatment of PHN [4,15,16], they require continued multiple daily applications over several weeks and often cause a burning sensation at the application site, resulting in a lack of compliance. NGX-4010 is a high-concentration capsaicin dermal patch (capsaicin, 8%) developed to quickly deliver a therapeutic dose of capsaicin locally into the skin during a single application. A dose-finding study in patients with PHN investigated the efficacy of NGX-4010 applied for three different durations (30, 60, or 90 minutes) and identified the 60-minute duration as the dose to be evaluated in subsequent studies [17]. This dose was confirmed in a phase 3 study that demonstrated that a single 60-minute application of NGX-4010 significantly reduced pain in patients with PHN [18]. The current study sought to confirm the results of the previous multicenter, randomized, double-blind, controlled phase 3 study that compared the efficacy and safety of a single 60-minute application of NGX-4010 with that of a low-concentration capsaicin control patch (capsaicin, 0.04%) over a 12-week posttreatment period in patients with PHN [18].



Sixty-one outpatient sites in the United States and Canada including private practices, university hospitals, and community hospitals participated in this trial. Patients were recruited directly by the study centers from their existing patient database, through written advertising, and radio and television ads. Patients 18–90 years old with a diagnosis of PHN and an average Numeric Pain Rating Scale (NPRS) [19] score of 3–9 (inclusive) were eligible if at least 6 months had elapsed since shingles vesicle crusting. Patients taking chronic pain medications, such as anticonvulsants, non-SSRI antidepressants, opioids, nonsteroidal anti-inflammatory drugs (NSAIDS), salicylates, or acetaminophen, had to be on a stable dose of those medications for at least 21 days before the day of study patch application and remain on a stable dose throughout the study period. Women of childbearing age were required to have a negative pregnancy test and be willing to use an effective method of contraception for 30 days after exposure to study medication.

Exclusion criteria were as follows: use of any topically applied pain medication on the painful area within 21 days before the day of application of the study patch; current use of any investigational drug or class 1 antiarrhythmic drug; uncontrolled diabetes mellitus or uncontrolled hypertension; significant pain of an etiology other than PHN; painful PHN areas located only on the face, above the scalp hairline, or near mucous membranes; and hypersensitivity to capsaicin, local anesthetics, oxycodone hydrochloride, hydrocodone, or adhesives. As prior use of high-dose opioids could limit the responsiveness to the optional oral opioid analgesics for treatment-associated discomfort used during the treatment procedure, patients were excluded if they were using concomitant opioid medication that was not orally or transdermally administered, or exceeded a total dose of 60 mg/day morphine equivalent.

The study was approved by the Institutional Review Boards/Independent Ethics Committees at all participating sites and conducted in accordance with the ethical principles of the Declaration of Helsinki, Good Clinical Practice guidelines, and applicable regulatory requirements. Written informed consent was obtained from all participating patients before initiating study-related procedures.


The double-blind study included a ≥14-day baseline screening period followed by a treatment day (day 0) and a 12-week posttreatment assessment period with clinic visits at weeks 4, 8, and 12. Eligible patients were randomized 1:1 to receive NGX-4010 (capsaicin 640 µg/cm2, 8%; NeurogesX, Inc., San Mateo, CA) or a control patch that was identically formulated but contained a lower concentration of capsaicin (3.2 µg/cm2, 0.04%). Randomization was stratified by gender and cardiovascular risk status. The low-concentration capsaicin control patches were used in place of placebo patches to provide effective blinding in the study as topical capsaicin can produce local erythema and a burning sensation.

The actual treatment assigned to individual patients was determined by a randomization scheme prepared by Fisher Clinical Services (Allentown, PA). Numbers were assigned only once, and no subject was randomized into the study more than once. The NGX-4010 and control patches were identical in appearance, as were the blinded study drug kits. Each kit was designated by a unique kit number, which was printed on the investigational drug label affixed to the outer bag enclosure and on each individual patch envelope.

All patients were pretreated with a topical local anesthetic cream (LMX4®, lidocaine 4%; Ferndale Laboratories, Inc., Ferndale, MI) for 60 minutes before the application of the study or control patches, which were applied directly to the painful area(s) (up to 1,120 cm2) for 60 minutes. After patch removal, the area was cleansed with a proprietary cleansing gel formulated to remove residual capsaicin. Patients were monitored for 2 hours after patch removal. Local cooling as well as oxycodone hydrochloride oral solution (1 mg/mL) or equivalent could be administered at the onset of treatment-associated discomfort and as needed. Patients could take short-acting opioid medication (hydrocodone bitartrate/acetaminophen 5 mg/500 mg) for up to 5 days after patch application for treatment-associated discomfort as needed. Topical pain medications were not permitted during the 12-week study period. Patients were allowed to take acetaminophen up to 3 g/day as needed for aches and pains.

Measures and Data Analysis


Efficacy was evaluated with daily NPRS scores captured at 9 pm every evening in a paper diary throughout the 12-week study period. The NPRS is an 11-point scale (0–10) with 0 indicating no pain and 10 indicating the worst possible pain [19]. These NPRS assessments were for “average pain for the past 24 hours.” Patient Global Impression of Change (PGIC; patients reported how they felt after treatment as compared with before treatment on a scale of −3 indicating “very much worse” to +3 indicating “very much improved,” with 0 being “no change”) and investigator-rated Clinical Global Impression of Change (CGIC) [20] were evaluated at weeks 4, 8, and 12. The modified Brief Pain Inventory (BPI) [21] and Short-Form McGill Pain Questionnaire (SFMPQ) [22] were collected at screening, and weeks 8 and 12, and the SF-36v2 questionnaire [23,24] was collected at screening and week 8.

The primary efficacy end point was the percentage change in NPRS scores from baseline to weeks 2 through 8. Baseline NPRS scores were recorded from day −14 to day −1. To avoid the potential confounding effect of allowed opioid medications for treatment-associated discomfort during days 0–5, week 1 NPRS scores were not included in the primary analysis. Other efficacy measures included: percentage change in NPRS scores from baseline to weeks 2–12; mean absolute changes in NPRS scores from baseline to weeks 2–8 and to weeks 2–12; the weekly percentage change from baseline in average pain; the percentage of patients who were responders (mean percentage decrease in NPRS score of ≥30%) and percentage of patients with a ≥50% or a ≥2-point reduction in NPRS score from baseline to weeks 2–8 and weeks 2–12; the percentage of patients considered improved (slightly, much, or very much) on the PGIC and CGIC at weeks 4, 8, and 12; and changes from screening in the BPI, SFMPQ, and SF-36v2 questionnaire.

All data were collected and analyzed by NeurogesX, Inc., and reviewed by an independent company and the academic author (GI). Efficacy analyses were based on the intent-to-treat population that consisted of all patients who received any study treatment and had at least 3 days of available NPRS scores during the baseline period. Mean percentage changes and mean numeric changes in NPRS scores from baseline to weeks 2 through 8 and 2 through 12 were analyzed using a gender-stratified analysis of covariance (ancova) model with baseline pain score as the only covariate. For each subject, the mean NPRS score for weeks 2 to n, where n is any number greater than 2, was computed as the average of the NPRS scores from day 8 to day 7 × n. Missing posttreatment NPRS scores were imputed using a modified last-observation-carried-forward approach. If the NPRS score was missing on any of days 0–8 or missing on day 8 and one or more consecutive days, then the baseline score was imputed for those days. If the NPRS score was missing for any day past day 8, then the missing score was imputed by the latest available non-missing score collected before that day. If all posttreatment NPRS scores were missing (including day 0), all scores were imputed using the baseline score. The percentage of responders (≥30% reduction) and the percentage of patients achieving a ≥50% or a ≥2-point reduction in NPRS score were compared between groups using logistic regression analyses, with baseline pain score and gender as covariates. A mixed-effect repeated measures ancova was used to analyze weekly percentage change from baseline in average pain. For the analysis of the weekly percentage change from baseline, no imputation was performed for missing pain scores posttreatment, as mixed-effects models allow the use of existing data without imputation. As a first step in the implementation of the repeated measures ancova, a full ancova that included gender, treatment, and week as main effects, two-way interactions and three-way interactions with baseline pain score as the covariate was performed to assess the effect of each treatment. Only the main effects and interactions that were statistically significant (P < 0.05) were kept in the final model. The main treatment effect, two-way interaction of gender with treatment, and treatment with week were significant and kept in the final model. Subject was included in the model as a random effect. The method took into consideration the week-to-week correlation of the pain score within the same subject assuming the autoregressive covariance structure. Such covariance structure was estimated from the data, which showed higher correlation in earlier weeks than that in later weeks up to week 12.

The percentage of patients reporting PGIC/CGIC score improvement was compared between treatment groups using the Fisher exact test. Changes from screening in the BPI, SFMPQ, and SF-36v2 questionnaire were tested between the active and the control group using a gender-stratified ancova model with screening score as the covariate.

It was estimated that to achieve 90% power at the 0.05 significance level with a standard deviation of 31%, 203 patients per treatment arm (total of 406) were required to detect a difference of 10% in change from baseline in NPRS scores between the NGX-4010 and control groups.


Safety and tolerability assessments during the 12-week study included adverse events, dermal assessments (0- to 7-point severity score) [25], pain assessment during and after the procedure using NPRS scores for “pain now,” the proportion of subjects with early patch removal, use of concomitant medications (name, strength, indication, dose, units, frequency, route, start date, stop date), 12-lead centrally read electrocardiograms (ECGs) (heart rate, RR interval, PR interval, QRS duration, QT interval, QTcB, and QTcF), clinical laboratory tests (hematocrit, white blood cell count, platelet count, aspartate aminotransferase, alanine aminotransferase, alkaline phosphatase, chloride, sodium, potassium, blood urea nitrogen, and creatinine), vital signs, targeted neurologic sensory examinations (light brush, pinprick, warmth, vibration, allodynia), physical examination, and medication use for treatment-associated discomfort.

Adverse events were coded using the Medical Dictionary for Regulatory Activities (version 9.0). Medication use for treatment-associated discomfort from days 0–5, the number of subjects completing the intended patch duration, demographics, and baseline clinical characteristics were compared between groups with Fisher exact tests or t-tests, as appropriate. The proportions of subjects reporting each level of dermal response at each of the clinical assessment time points were compared between the active group and the control group using the Cochran–Mantel–Haenszel test for trend. The results of the neurologic sensory exams were compared at each visit using the Wilcoxon rank sum test and the Cochran–Mantel–Haenszel test for trend. For the total surface area of allodynia, differences between the active and control groups in the percentage change from screening to weeks 8 and 12 were compared using gender-stratified ancova model with screening allodynia area as the covariate.



A total of 418 patients were enrolled in the study; 416 received study medication and are included in all safety and efficacy analyses. Of those subjects receiving treatment, 212 were treated with NGX-4010 and 204 received the control low-dose capsaicin patch (Table 1). Two patients randomized to the NGX-4010 group were not treated: one had an adverse reaction to the pretreatment topical anesthetic, and the other was found not to meet the inclusion criteria (Figure 1). Thirty-eight patients prematurely withdrew from the study, 20 in the NGX-4010 group (including the two patients who did not receive treatment) and 18 in the control group.

View this table:
Table 1

Patient demographics and clinical characteristics

NGX-4010 (N = 212)Control (N = 204)
Age, mean, years (SD)70.2 (12.25)70.4 (12.87)
Male, N (%)93 (44)97 (48)
Race, N (%)
  White197 (93)191 (94)
  African American6 (3)8 (4)
  Asian3 (1)2 (1)
  Other6 (3)3 (1)
Clinical characteristics
Mean duration of postherpetic neuralgia, years (SD)3.1 (3.55)3.3 (3.72)
Mean baseline pain level (SD)*5.7 (1.59)5.8 (1.63)
Any baseline pain treatment, N (%)150 (71)153 (75)
Baseline concomitant pain medication use, N (%)117 (55)118 (58)
  Opioids only13 (6)16 (8)
  Anticonvulsants only42 (20)55 (27)
  Non-SSRI antidepressants only14 (7)17 (8)
  Opioids and anticonvulsants21 (10)16 (8)
  Opioids and non-SSRI antidepressants5 (2)0
  Anticonvulsants and non-SSRI antidepressants15 (7)9 (4)
  All three7 (3)5 (2)
Mean size of painful area at screening, cm2 (SD)329.8 (225.87)325.7 (215.03)
  • * All available Numeric Pain Rating Scale scores from day −14 to day −1.

  • Patients were considered to be using baseline concomitant pain medication if using anticonvulsants, non-SSRI antidepressants, or opioids on day −1 and for at least 7 consecutive days.

  • SD = standard deviation; SSRI = selective seretonin reuptake inhibitor.

Figure 1

Method of randomization for treatment. aThe number of patients withdrawn includes two patients who were randomized to the NGX-4010 group but never received treatment. One subject had an adverse reaction to the topical lidocaine pretreatment (adverse event), and one subject did not meet entry criteria (other).

Demographics, baseline pain scores, duration of PHN, the size of the painful area, the size of the treated area, and use of baseline concomitant neuropathic pain medications (opioids, antidepressants, and anticonvulsants) were similar between treatment groups (Table 1).


Mean NPRS scores at baseline were similar between the NGX-4010 and the control groups (5.7 ± 0.11 and 5.8 ± 0.11; Table 2). Patients treated with a single 60-minute application of NGX-4010 reported a mean of 32.0 ± 2.07% reduction in pain during weeks 2–8 compared with 24.4 ± 2.11% in the control group (P = 0.011; Table 2). Forty-six percent of NGX-4010 patients were considered to have responded to treatment (i.e., experienced a ≥30% mean decrease from baseline in pain) during weeks 2–8 compared with 34% of controls (P = 0.02). The proportion of patients who achieved a ≥50% decrease in pain scores from baseline to weeks 2–8 was significantly greater for the NGX-4010 group (29% vs 20%; P = 0.04). In addition, more NGX-4010 than control patients experienced a ≥2-point reduction in pain from baseline to weeks 2–8 (42% vs 26%; P = 0.002). Similar results were observed during weeks 2–12 (Table 2).

View this table:
Table 2

Clinical efficacy of NGX-4010 in postherpetic neuralgia

NGX-4010 (N = 212)Control (N = 204)P Value
Numeric Pain Rating Scale scores
  Baseline, mean (SE)5.7 (0.11)5.8 (0.11)NS
  95% CI5.46 to 5.895.54 to 5.99
Change, mean (SE)
  Baseline to weeks 2–8−1.7 (0.12)−1.3 (0.12)0.034
  95% CI−1.93 to −1.47−1.58 to −1.11
  Baseline to weeks 2–12−1.7 (0.12)−1.4 (0.12)0.058
  95% CI−1.94 to −1.47−1.62 to −1.14
Percentage change, mean (SE)
  Baseline to weeks 2–8−32.0 (2.07)−24.4 (2.11)0.011
  95% CI−36.06 to −27.92−28.57 to −20.27
  Baseline to weeks 2–12−32.3 (2.13)−25.0 (2.17)0.017
  95% CI−36.49 to −28.12−29.29 to −20.76
Patients with ≥30% reduction, %
  Baseline to weeks 2–846340.020
  Baseline to weeks 2–1247350.021
Patients with ≥50% reduction, %
  Baseline to weeks 2–829200.042
  Baseline to weeks 2–1230210.035
Patients with ≥2-point reduction, %
  Baseline to weeks 2–842260.002
  Baseline to weeks 2–1243290.003
Patients taking concomitant pain medication, N (%)117 (55)118 (58)
Percentage change, mean (SE)
  Baseline to weeks 2–8−26.9 (2.37)−18.4 (2.36)0.012
  95% CI−31.55 to −22.21−23.09 to −13.79
  Baseline to weeks 2–12−26.8 (2.45)−19.3 (2.44)0.032
  95% CI−31.61 to −21.94−24.10 to −14.47
Patients not taking concomitant pain medication95 (45)86 (42)
  Baseline to weeks 2–8−38.4 (3.55)−32.4 (3.73)NS
  95% CI−45.42 to −31.42−39.81 to −25.09
  Baseline to weeks 2–12−39.3 (3.63)−32.7 (3.82)NS
  95% CI−46.44 to −32.10−40.24 to −25.16
Patient Global Impression of Change
Much improved (very much, much), N (%)
  Week 871 (37)49 (26)0.020
  Week 1283 (41)50 (26)0.001
Improved (very much, much, slightly), N (%)
  Week 8119 (62)99 (51)0.030
  Week 12123 (61)91 (47)0.005
  • CI = confidence interval; SE = standard error; NS = not significant.

Patients treated with NGX-4010 had significantly greater mean percentage reductions in NPRS scores compared with controls as early as week 2. Pain reduction remained significantly greater in the NGX-4010 group compared with the control group from week 2 to week 12 (P < 0.05; Figure 2).

Figure 2

Numeric Pain Rating Scale scores for 12 weeks after patch application in treated and control patients. Values shown are mean and standard error of the mean. NGX-4010 showed greater improvement in mean pain scores at all assessment times from week 2 to week 12 (*P < 0.05; P < 0.005 vs control).

Greater reductions in NPRS scores from baseline to weeks 2–8 were observed in patients receiving NGX-4010 compared with control, both in patients receiving concomitant pain medication (26.9% vs 18.4%; P = 0.01) and in those with no concomitant pain medication (38.4% vs 32.4%); however, the difference was not significant between NGX-4010 and control in the latter group (P > 0.05). Similar findings were seen during weeks 2–12 (Table 2).

Analysis of PGIC demonstrated that significantly more NGX-4010 recipients considered themselves to have improved (slightly, much, or very much) compared with the control group at both week 8 (62% vs 51%; P = 0.030) and week 12 (61% vs 47%; P = 0.005). Additionally, more NGX-4010 patients reported being much or very much improved compared with control patients both at week 8 (37% vs 26%; P = 0.020) and at week 12 (41% vs 26%; P = 0.001) (Table 2).

The results of the CGIC were comparable to the results of the PGIC. More NGX-4010 patients were judged by the study investigators to have been improved (very much, much, or slightly) compared with control patients at week 8 (63% vs 52%; P = 0.030) and week 12 (63% vs 48%; P = 0.003). Study investigators also rated more NGX-4010 patients as much or very much improved compared with control patients at week 8 (36% vs 26%; P = 0.047) and at week 12/termination (41% vs 27%; P = 0.003). Other subject-rated questionnaires (BPI and SFMPQ) frequently demonstrated larger improvements in the NGX-4010 group compared with the control group; however, the differences were not significant (P > 0.05). Similar mean changes on the SF-36v2 subject-rated questionnaire were seen for both study groups.


NGX-4010 was well tolerated in the majority of patients. Ninety-eight percent of NGX-4010 patients and 100% of control patients completed at least 90% of the intended patch application duration.

During days 0–5, medication for the relief of treatment-associated discomfort was used by more patients in the NGX-4010 group than in the control group (53% vs 21%; P < 0.0001). The majority of this medication use occurred on the day of treatment (day 0). On day 1, 19% of NGX-4010 patients and 8% of control patients used medication for treatment-associated discomfort; by day 5, only 8% of NGX-4010 patients and 3% of control patients were using medication for treatment-associated discomfort.

The overall incidence of adverse events was higher among NGX-4010-treated patients compared with controls. The proportion of patients with at least one treatment-emergent adverse event was 98% in the NGX-4010 group and 87% in the control group (Table 3). The higher incidence of treatment-emergent adverse events in the NGX-4010 group was primarily due to expected, capsaicin-related application site events reported by 204 (96%) NGX-4010 and 160 (78%) control patients. Common application site events included application site erythema, application site pain, application site papules, and application site edema. Most events were mild or moderate; the incidence of severe application site events was higher in the NGX-4010 group (8%) compared with the control group (1%). The incidence of non-application site-related adverse events was similar in the two treatment groups.

View this table:
Table 3

Treatment-emergent adverse events that occurred in ≥3% of patients in either treatment group by system organ class and preferred term (safety population)

System Organ Class and Preferred TermNGX-4010, N (%) (N = 212)Control, N (%) (N = 204)
Any adverse event208 (98)177 (87)
Gastrointestinal disorders21 (10)22 (11)
  Nausea11 (5)5 (2)
  Vomiting6 (3)0 (0)
General disorders and administration site conditions204 (96)160 (78)
  Application site edema13 (6)0 (0)
  Application site erythema194 (92)141 (69)
  Application site pain134 (63)57 (28)
  Application site papules15 (7)5 (2)
  Application site pruritus6 (3)3 (1)
Infections37 (17)37 (18)
  Sinusitis6 (3)1 (1)
  Upper respiratory tract infection8 (4)6 (3)
Musculoskeletal and connective tissue disorders14 (7)16 (8)
Nervous system disorders29 (14)29 (14)
  Dizziness3 (1)6 (3)
  Headache4 (2)10 (5)
  Postherpetic neuralgia15 (7)12 (6)
Respiratory, thoracic, and mediastinal disorders7 (3)14 (7)
Skin and subcutaneous tissue disorders31 (15)22 (11)
  Erythema13 (6)16 (8)
  Pruritus7 (3)2 (1)

Serious adverse events occurred in 5% and 4% of NGX-4010 and control patients, respectively. No serious adverse event was considered related to treatment. Five patients withdrew from the study due to adverse events: two (1%) NGX-4010 and three (1%) control patients. All events leading to study withdrawal were serious adverse events and not considered related to treatment. One patient, an 81-year-old female who was treated with NGX-4010, died on day 31 due to diverticulitis; this event was considered not related to study medication.

An increase in mean NPRS scores was evident during the NGX-4010 patch application; after completion of the treatment, scores declined, returning to near pre-procedure levels (+0.4) at 85 minutes after patch removal. By the evening of the treatment day, mean NPRS scores were decreased below baseline in both the NGX-4010 and control groups. Dermal irritation was generally mild and transient. Two hours following treatment, 3% of NGX-4010 recipients had dermal assessment scores >2 (definite erythema or minimal edema or minimal papular response) compared with 0.5% of control subjects. Four weeks following treatment, no patients in either the NGX-4010 or control groups had dermal assessment scores >2.

There were no significant differences between the treatment groups in light brush, pinprick, vibration, and warmth sensations at screening, week 8, or week 12 (study termination). Changes from screening to week 8 and week 12 were similar between treatment groups for all sensory modalities evaluated. No trends consistent with a decrease in sensory function were observed in either group over the 12-week study period. The area of allodynia decreased similarly in both treatment groups from screening to week 12.

Transient changes in blood pressure were noted during and shortly after the treatment procedure. Blood pressure decreased following topical anesthetic application and increased after patch application in both treatment groups. Increases in blood pressure seen during and after patch application were greater in the NGX-4010 group than those observed in the control group. Mean increases in systolic blood pressure were ≤9.9 mm Hg for the NGX-4010 and ≤3.5 mm Hg for the control group; mean increases in diastolic blood pressure were ≤4.0 mm Hg and ≤1.6 mm Hg, respectively. Blood pressure began returning toward baseline values within 60 minutes after patch removal.

There was no evidence of an effect of NGX-4010 on any laboratory parameter evaluated. Hematologic and serum chemistry laboratory values showed no trends for any parameter during the 12-week study in either treatment group. There were no ECG changes related to treatment with NGX-4010. No other safety concerns were identified.


This study demonstrates that, in patients with PHN, a single 60-minute application of NGX-4010 can provide pain relief that is maintained for up to 12 weeks following treatment. Pain scores began declining as early as day 1 after treatment (data not shown) and were significantly lower than control by week 2 after patch application. This reduction continued over the 12-week trial period. Compared with control patients, significantly more patients treated with the NGX-4010 patch had a clinically meaningful reduction in pain (≥30% reduction in NPRS score [19]). NGX-4010 provided additional pain relief when administered with other concomitant pain medications. This finding suggests that this peripherally targeted capsaicin-based therapy can provide an additive benefit to the effects of other systemically administered medications and is important as treatment of peripheral neuropathic pain syndromes such as PHN commonly requires the use of multiple medications [26]. Similar to previous reports, pain reductions reported by patients using concomitant pain medications were smaller than pain reductions observed in patients not using concomitant pain medications in both treatment groups [18,27]. This might suggest that patients already taking concomitant pain medications have more treatment-resistant pain or that each additional pain medication has an additive but smaller effect than the preceding medication. In contrast to the previous phase 3 study [18], the treatment difference between NGX-4010 and control patients was smaller in patients not using concomitant pain medications compared with patients using concomitant pain medications. This was mostly due to a larger pain reduction in control patients in this subgroup compared with control patients in the previous phase 3 study, while pain reduction in patients treated with NGX-4010 was comparable. Taken together, these data support the conclusion that NGX-4010 is efficacious regardless of concomitant pain medication use.

Although the pungency of capsaicin is well recognized, treatment with NGX-4010 was well tolerated in the majority of patients, with nearly all patients completing the full duration of treatment. Despite an initial increase in mean NPRS scores that was evident during NGX-4010 patch application, scores returned to near pre-procedure levels (+0.4) within 85 minutes after patch removal. By the evening of the day of treatment, patients had less pain compared with baseline in both the NGX-4010 and control groups. Capsaicin-related local application site reactions were the most common adverse events and were transient, mostly mild to moderate, and self-limited. Application site pain could be adequately managed by local cooling or, if needed, by short-acting oral opioid analgesics. Transient increases in systolic and diastolic blood pressure seen during or shortly after treatment were associated with and likely due to treatment-related discomfort. A direct effect of capsaicin on blood pressure is unlikely as systemic capsaicin levels following treatment with NGX-4010 have been shown to be low (<5 ng/mL) and transient [28].

The current results confirm and extend the findings of previous phase 2 and 3 clinical trials, demonstrating the efficacy and tolerability of a 60-minute application of NGX-4010 over a 12-week posttreatment period in patients with PHN [17,18,29]. Results from a previous 1-year extension study suggest that pain relief can be maintained in patients with PHN given up to four NGX-4010 applications [29].

The study had limitations that should be considered when interpreting the results. In order to address the difficulty in blinding the use of topical high-concentration capsaicin, a low-concentration capsaicin control was used instead of an inert placebo. The low-concentration patch delivered an amount of capsaicin that, like NGX-4010, was capable of producing local application site reactions. Erythema, pain, and use of medication for treatment-associated discomfort occurred with both capsaicin formulations, suggesting that the blind could not be broken based on an individual subject's initial reaction to treatment. The response to the low-concentration control patch (−24% to −25%) observed in this study was greater than the median placebo response of −15% to −16% observed for several PHN studies reviewed in a recent analysis by Quessy and Rowbotham [30]. Although the control patch delivered a lower dose of capsaicin compared with NGX-4010, an analgesic effect of the low-concentration capsaicin control patch cannot be excluded and may have led to an underestimation of the efficacy of NGX-4010. In addition, it is possible that the placebo response associated with the control patch was enhanced due to the mechanical manipulation related to the treatment procedure and the application site reactions that resulted from the capsaicin in the control.

The mechanism of action of capsaicin involves primarily a reversible defunctionalization of cutaneous nociceptors that have been shown to be hyperactive in painful peripheral neuropathies. This hyperactivity has been shown to be associated with decreased epidermal nerve fiber density [2,12,31–33] and a reduction of TRPV1 immunoreactivity [34,35]. Moreover, data suggest a positive correlation between the extent of epidermal nerve fiber loss and the severity of pain [31–33]. Although capsaicin most likely exerts its effect through TRPV1 receptors expressed on cutaneous nociceptors, TRPV1 receptors have also been reported to be expressed in keratinocytes and other epithelial and mesenchymal cell types [36–40]. It is possible that despite a loss of TRPV1-expressing epidermal nerve fibers in peripheral neuropathies, TRPV1 expression in keratinocytes and other epidermal and dermal cells persists [41], although the physiological relevance of TRPV1 in these cells remains unclear [42].

In summary, treatment with a single 60-minute topical application of NGX-4010 provided a significant reduction in pain that was maintained for up to 3 months in patients with PHN and was generally well tolerated.


The authors would like to acknowledge and remember E. Richard Blonsky, MD, for his important contributions to this study as well as to the field of pain medicine. The authors would like to thank Biao Lu, PhD, for his statistical assistance in the preparation of this article. Funding for this study was provided by NeurogesX, Inc. Gordon Irving and Misha Backonja are consultants for NeurogesX. Jeffrey K. Tobias and Geertrui F. Vanhove are NeurogesX employees and own NeurogesX stock. Shiao-Ping Lu is a former NeurogesX employee and owns NeurogesX stock. The remaining authors have no conflicts of interest.


  • NGX-4010 C117 Study Group members: C. D. Anderson, R. Aung-Din, R. J. Bargar, J. C. Barlow, S. Behiels, C. S. Bell, H. Bhamjee, E. S. A. Blue, M. A. Boileau, D. M. Brandon, H. V. Cabasares, J. S. Cook, B. J. Cutler, M. C. DeSantis, F. Diaz-Mitoma, M. Y. Dubois, W. T. Ellison, D. N. Fiske, S. K. Gazda, G. L. Gibson, J. E. Greenwald, W. P. Gulliver, S. Gupta, D. Headapohl, W. C. Howland, R. T. Jermyn, S. R. Kaster, A. J. Kelly, T. W. Khan, S. S. Kirzinger, A. M. Klymiuk, J. Kooy, S. Lam, T. L. Lee, T. S. Levy, M. Luterman, M. E. Lynch, L. J. Mate, R. B. Matthews, C. M. McCarthy, B. J. McGuire, J. E. McNeff, S. Mraz, W. H. Noland, M. Nunez, M. F. J. O'Mahony, A. E. Palencia, J. F. Peppin, H. M. Prupas, R. L. Rauck, Q. Rizvi, J. R. Roberts, G. Schacter, R. L. Schwartz, T. Shamim, W. M. Smith, D. R. Spink, R. J. Struble, M. D. Tark, R. Thakur, S. Venkataraman, D. B. Vine, S. A. Vota, D. Walk, and R. M. Webb.

  • This study is registered in http://clinicaltrials.gov as number: NCT00300222.


View Abstract