Transdermal Buprenorphine Relieves Neuropathic Pain: A Randomized, Double-Blind, Parallel-Group, Placebo-Controlled Trial in Diabetic Peripheral Neuropathic Pain
OBJECTIVE To evaluate the efficacy and safety of transdermal buprenorphine in patients with diabetic peripheral neuropathic pain (DPNP).
RESEARCH DESIGN AND METHODS This multicenter, randomized, double-blind, placebo-controlled, parallel-group trial enrolled patients with type 1 or type 2 diabetes and stable glycemic control who had been experiencing moderate to severe DPNP for at least 6 months on maximal tolerated conventional therapy. Patients were randomly assigned to receive buprenorphine (5 μg/h) or placebo patches. The dose was titrated to effect to a maximum of 40 μg/h. Paracetamol was available as rescue analgesia. The severity of pain and other symptoms of DPNP were assessed daily in a patient diary and at clinic visits.
RESULTS One hundred eight-six patients were enrolled, with 93 randomized to either buprenorphine or placebo. A high proportion of patients did not complete the study (buprenorphine 37 of 93, placebo 24 of 93). The main reason for premature withdrawal in the buprenorphine group was adverse events commonly due to untreated nausea and/or vomiting. Among the per-protocol population, more patients in the buprenorphine group (86.3%) experienced a 30% reduction in average versus baseline pain at week 12 than those in the placebo group (56.6%, P < 0.001). A nonsignificant trend favored the buprenorphine group within the intention-to-treat analysis of the same end point (51.7% vs. 41.3%, P = 0.175).
CONCLUSIONS Transdermal buprenorphine, when tolerated, is an effective therapy for DPNP and provides another option to manage this challenging painful condition. Nausea and constipation need to be managed proactively to optimize treatment outcomes.
Peripheral neuropathy is a common complication of both type 1 and type 2 diabetes, occurring in up to 47% of patients (1). The most common painful neuropathy observed with diabetes is diabetic peripheral neuropathic pain (DPNP), a burning, tingling, or aching sensation that is worse at night, often disturbs sleep, and subsequently impairs daytime function (2–4). Patients with DPNP often experience allodynia, leading to acute distress on contact with normally nonpainful stimuli such as clothing or bed sheets (3).
The symptomatic management of DPNP is challenging because painful symptoms can be refractory to treatment (4), and often, the best achievable clinical outcome is a 30–50% reduction in pain severity (2). Antidepressants and antiepileptics are the first-line medications for the relief of DPNP, but generally, well-designed prospective clinical trials establishing their effectiveness are limited (5). Currently, pregabalin and duloxetine are the only medications specifically approved for DPNP (2). Their clinical effect is modest, with the number needed to treat (NNT) for one extra patient to achieve 50% reduction in pain over placebo being 6.3 (95% CI 4.6, 10) for 600 mg/day pregabalin, 11 (95% CI 6.1, 54) for 300 mg/day pregabalin (6), and 5 (95% CI 4, 7) for duloxetine (7). Although neuropathic pain appears to be relatively unresponsive to opioid analgesics, some evidence demonstrates efficacy of strong opioids (oxycodone and morphine) in neuropathic pain states, including DPNP (8,9).
Buprenorphine is a synthetic opioid analgesic with a well-established clinical profile in the management of moderate to severe chronic pain. It is pharmacologically different from other opioid analgesics, and these differences may enhance its potential in the management of neuropathic pain states (10–12). Buprenorphine is a partial μ-opioid receptor agonist. At clinical doses, it exhibits full-dose–dependent analgesia but a plateau in respiratory depression (11,13,14). Neuropathic pain is associated with loss of pertussis toxin-sensitive G-protein activity (15) on which the analgesic action of pure μ-opioids, such as morphine, depends. Buprenorphine, however, is the only opioid where its antinociception is not blocked by the pertussis toxin (11). Abnormal pain sensitivity (hyperalgesia and allodynia) caused by neuropathy often is nonresponsive to opioid therapy (11). Buprenorphine, unlike the pure μ-opioid agonist fentanyl, has demonstrated a significant long-lasting antihyperalgesic action in a human pain model where hyperalgesia was induced by central sensitization (16). A small, but growing body of clinical evidence supports the potential of buprenorphine in neuropathic pain (10), including case series (17,18), a retrospective observational study in patients with refractory neuropathic pain (19), and a postmarketing surveillance study (20). To date, no placebo-controlled trials have assessed transdermal buprenorphine in neuropathic pain or specifically DPNP. The aim of this randomized, double-blind, placebo-controlled trial was to evaluate the efficacy and safety of transdermal buprenorphine in patients with DPNP.
Research Design and Methods
Patients with type 1 or type 2 diabetes for the previous 6 months with stable glycemic control for the past 3 months and having experienced DPNP for a minimum of 6 months (on stable conventional nonopioid analgesic therapy) were eligible for this study. Patients had DPNP attributable to diabetic sensorimotor polyneuropathy based on clinical grounds. This included experiencing daily neuropathic pain in the lower limbs for at least 6 months, pain initially starting in the feet with relatively symmetrical onset, and a score >3 on the physical assessment section of the Michigan Neuropathy Screening Instrument. Eligible patients had pain of at least moderate intensity (numeric rating scale [NRS] score ≥4 [out of 10]) at the screening visit and had a mean pain intensity NRS score ≥4 during the screening week.
Patients were allowed to continue stable doses of antidepressants, antiepileptics, or other medications with activity against neuropathic pain. However, the use of weak opioid analgesics (e.g., codeine containing, propoxyphene containing, tramadol), nonsteroidal anti-inflammatory drugs, and any topical DPNP therapies or other nondrug therapies for DPNP were discontinued at the screening visit. Patients who had currently or previously treated their DPNP with a strong opioid (e.g., oxycodone, morphine) or were currently taking a strong opioid for any condition were excluded from the study. Patients with eczema, cutaneous atrophy, or a dermatological disorder that may preclude correct use of the patch; those with a known hypersensitivity to opioid analgesics, transdermal delivery systems, or patch adhesive; and those who needed treatment with direct external heat sources, which may interfere with the patch delivery system, were also excluded from the study.
This multicenter, randomized, double-blind, placebo-controlled, parallel-group trial evaluated the efficacy and safety of buprenorphine patches in patients with DPNP. The study consisted of three phases: a 1-week screening phase, a 12-week assessment phase, and a 3-week blinded downward-titration completion phase. At the screening visit, patients were assessed for suitability for entry into the study. At the randomization visit, participants were randomly allocated by an interactive Web response system to active or placebo treatment in equal proportions. The assessment phase consisted of six weekly visits followed by three biweekly visits. A flexible dosage titration regimen was allowed during this 12-week period. Any patient on a dose >10 μg/h at the end of week 12 or at study withdrawal underwent a blinded downward titration off the study drug.
The participants recorded pain intensity and pain relief over the previous 24 h daily in their patient diary before going to bed. Pain intensity was measured on an 11-point NRS (0 = no pain, 10 = worst possible pain). Patients assessed their average, least, worst, and night pain intensity by using separate NRSs. Pain relief was assessed by using a categorical rating scale (0 = complete relief, 1 = a lot of relief, 2 = moderate relief, 3 = slight relief, 4 = no relief, and 5 = pain worse). The frequency of use of rescue medication and how much the participant’s pain interfered with sleep (on the basis of the Daily Sleep Interference Scale [DSIS]) was also recorded in the patient dairy.
At clinic visits, investigators assessed the participants’ general health by reviewing vital signs, recorded changes in concomitant medications, and adverse events experienced. At specific visits, blood was taken for laboratory testing. At each clinic visit, participants assessed the effect of the study medication on a range of spontaneous and evoked neuropathic pain manifestations by using the Neuropathic Pain Symptom Inventory (NPSI). In addition, pain and its interference with daily activities was assessed at baseline and weeks 6 and 12 by using the short-form McGill Pain Questionnaire (SF-MPQ) and the Brief Pain Inventory interference scales, respectively.
Affective state over the previous 2 weeks was assessed at baseline and weeks 6 and 12 by using the 21-item Beck Depression Inventory-II and the Profile of Mood States. In addition, health-related quality of life (HRQOL) was assessed with the Medical Outcomes Study 36-Item Short Form survey. Global ratings of improvement and satisfaction with treatment were assessed by both patients and investigators at baseline and weeks 6 and 12 by using the Participant Global Impression of Change (PGIC) scale and the Clinician Global Impression of Change (CGIC) scale.
The primary efficacy end point of the study was the proportion of patients achieving a 30% reduction in average pain intensity at the end of week 12 compared with baseline by using the daily NRS recorded in the patient diary. Secondary end points included the proportion of patients achieving a 50% reduction in average pain intensity at the end of week 12 compared with baseline; change from baseline in the mean NRS score at week 12 for the average pain intensity; change from baseline in the mean NRS score at each week for least pain, worst pain, and night pain; mean categorical scale score for pain relief at each week; change from baseline in NPSI scores for pain intensity as measured at clinic visits; median average daily dose of paracetamol rescue medication used; differences in total medication usage between active and placebo groups at weeks 6 and 12; change in mean DSIS score for each week of the study; change in SF-MPQ, including total score, sensory and affective pain dimension subscores, the visual analog scale rating of pain over the previous 7 days, and present pain intensity score at weeks 6 and 12; change from baseline versus weeks 6 and 12 for pain interference with daily activities, affective state, and HRQOL; and PGIC and CGIC at weeks 6 and 12.
The sample size was calculated on the assumption that 30% of patients receiving buprenorphine patches and 10% of patients receiving placebo patches would achieve the primary end point. Ninety-three patients were required in each group to provide 90% power to detect a difference in response rates.
Buprenorphine (NORSPAN transdermal drug delivery system; Mundipharma) or placebo patch was applied weekly during the assessment phase of the trial. Therapy commenced with the 5 μg/h buprenorphine patch or matching placebo, and the dose was titrated to effect at study visits, weekly during the first 6 weeks and every second week during weeks 7 to 12 of the assessment phase to a maximum dose of 40 μg/h. The permitted doses were 5, 10, 20, 30, and 40 μg/h. Paracetamol (500–1,000 mg) up to four times per day was available for use as a rescue analgesic. If participants experienced severe uncontrolled DPNP despite maximal usage of rescue paracetamol, uptitration could occur earlier at 5-day intervals. Downtitration was allowed at any stage if the patient experienced intolerable adverse events. Common opioid-related adverse effects were treated with antiemetics and aperients.
For the primary efficacy analysis, the proportion of responders in each group was compared using logistic regression. The dependent variable was response or nonresponse as defined by a 30% reduction in pain at week 12. Patients who had dropped out of the study or had missing values at week 12 were considered nonresponders. Primary efficacy analysis was based on the intention-to-treat (ITT) population by using a significance level (α) of 0.05. The same analysis was conducted in the per-protocol (PP) population.
The ITT population included all participants randomized to receive at least one dose of study medication and had at least one valid after-treatment efficacy measurement. The PP population included all participants in the ITT population who completed the study, were compliant with study medication, and did not have any major protocol violations.
All secondary efficacy end points were compared between treatment groups by using the ITT population. Generalized linear mixed models were used to examine the treatment effect over the course of the study in the secondary efficacy models. Missing data were considered missing at random. Differences between treatment groups in the number of daily rescue tablets taken were compared by Wilcoxon two-sample test. All statistical analyses were carried out using two-sided tests at the 5% level of significance.
The affect of antidepressant or antiepileptic use on the effectiveness of buprenorphine patches was investigated by adding interaction terms to the logistic model of the primary outcome. All safety end points were summarized by treatment group. The safety population included all participants randomized to receive at least one dose of study medication and who had at least one safety assessment during the assessment phase of the study.
A total of 186 patients were enrolled in the study, with 93 randomized to receive either buprenorphine or placebo patches. A high number of patients did not complete the study. Of those participating, 37 (39.8%) in the buprenorphine group and 24 (25.8%) in the placebo group withdrew from the study. The main reasons for discontinuation were adverse events (n = 28) in the buprenorphine group and inadequate pain control (n = 9) in the placebo group. Figure 1 summarizes the patient flow (randomized, withdrawn, and analyzed).
Participant demographic characteristics at baseline were comparable between the two groups (Table 1). The mean average pain at baseline on an 11-point NRS was 5.7 (SD 1.1) for the buprenorphine group and 5.9 (1.3) for the placebo group. All but 1 patient (in the buprenorphine group) were taking concomitant medications during the study; 57 (61.3%) participants in the buprenorphine group and 64 (68.8%) in the placebo group were taking antiepileptics or antidepressants.
The primary end point was the proportion of patients from the ITT population that achieved at least a 30% reduction in NRS average pain intensity at week 12. A nonsignificant trend was found for more patients in the buprenorphine group (51.7% [46 of 89]) who achieved this end point compared with placebo (41.3% [38 of 92]). The odds ratio (OR) from logistic regression was 1.56 (95% CI 0.82, 2.97; P = 0.175). This outcome was highly significant, favoring buprenorphine in the PP population (buprenorphine 86.3% [44 of 51], placebo 55.6% [35 of 63], OR 6.88 [95% CI 2.20, 21.47], P < 0.001). According to the PP analysis on the efficacy of buprenorphine patches among the participants who tolerated this therapy, the main reason for study withdrawal in the buprenorphine group was adverse events, whereas lack of efficacy was the main reason for study withdrawal in the placebo group.
Nausea and vomiting were the most common adverse events that led to study discontinuation in buprenorphine group (37.8% [14 of 37]); however, only one patient in the buprenorphine group received antiemetic treatment despite this being allowed within the protocol. In contrast, constipation (the next most commonly reported adverse event) was actively managed in 14.0% of participants in the buprenorphine group and resulted in only two participants withdrawing from the study, one of whom also reported nausea/vomiting as reason for withdrawal.
The effect of concomitant antidepressants and antiepileptics on the primary end point was investigated. The OR for the estimated effect of antidepressants was 0.77 (95% CI 0.39, 1.52) and 1.48 (0.58, 3.81) for antiepileptics. The wide CIs indicate that the concomitant use of adjuvants did not have a meaningful effect on the efficacy of transdermal buprenorphine.
The mixed-model analysis of the change in NRS score for pain intensity from baseline at each study visit (ITT population) showed a significant difference between buprenorphine and placebo from week 1, which was maintained throughout the 12-week study period (Fig. 2). The mean treatment effect was −1.20 (95% CI −1.83, −0.57; P < 0.001) at week 12. Similar results were observed in the PP analysis (data not shown), with the magnitude of the difference favoring buprenorphine being greater than in the ITT analysis.
The observation of a significant improvement in average pain intensity from week 1 with buprenorphine patches was also observed for patient ratings of least pain, worst pain, and night pain (Fig. 3). Improvements in pain intensity reached peak effect by week 6, and this level of pain control was maintained until week 12.
Table 2 summarizes the remaining secondary efficacy end points at week 12. The proportion of patients experiencing at least a 50% reduction in NRS pain intensity at week 12 was significantly higher in the buprenorphine group (34.8% [31 of 89]) than in the placebo group (20.7% [19 of 92]). The OR for the treatment effect was 2.19 (95% CI 1.10, 4.38; P = 0.026), and the NNT was 7.05.
The mean daily use of rescue medication did not differ between the two groups (Table 2). The cumulative use of paracetamol was lower for the buprenorphine group than in the placebo group, and this was significantly lower in the second 6 weeks of the study (P < 0.05).
Buprenorphine patch use was associated with significant improvements in pain relief, total pain intensity (from the NPSI), and sleep. In this study, buprenorphine use was not associated with statistically significant improvements in physical function, mood, or HRQOL, with the exception of the dimension bodily pain (P = 0.024). PGIC was significantly better for buprenorphine than for placebo (P < 0.05). CGIC was similar to that of patients; however, this result was not statistically significant.
The extent of exposure to study medication was lower for buprenorphine than for placebo. This is due to fewer participants in the buprenorphine group being titrated to the maximum dose of 40 μg/h (11.8% vs. 58.1% for placebo) (Table 3) and a lower mean duration of exposure in the buprenorphine group (66.9 ± 34.3 vs. 84.3 ± 31.3 days for placebo) due to more participants discontinuing therapy in the buprenorphine group.
The majority of study participants reported at least one adverse event (93.6% [87 of 93] in the buprenorphine group vs. 81.7% [76 of 93] in the placebo group). The most frequently reported treatment-emergent adverse events were nausea (43.0%) and constipation (31.2%) in the buprenorphine group and upper respiratory tract infection (12.9%) and headache (9.7%) in the placebo group. Most adverse events were mild or moderate in severity. A similar number of participants reported severe adverse events (10 and 9 in the buprenorphine and placebo groups, respectively). Seventeen serious adverse events were reported by seven participants in the buprenorphine group, and five serious adverse events were reported by four participants in the placebo group. All serious events were unrelated (per investigator assessment) to study medication except for one case of vomiting (buprenorphine) and one case of respiratory failure (placebo). In addition, one case each of supraventricular tachycardia, depression, and diarrhea was considered possibly related to buprenorphine use. There were no deaths and few clinically important changes in laboratory assessments, vital signs, and physical examination findings in both treatment groups.
The management of DPNP is challenging, and although opioids are not a first-line therapy, they are commonly used with adjuvants to manage neuropathic pain (21,22). To our knowledge, this is the first high-quality, randomized, double-blind, placebo-controlled trial to assess the effectiveness of buprenorphine patches for the symptomatic relief of DPNP. In the ITT population, there was a trend for more participants in the buprenorphine group to experience at least a 30% reduction in average pain at week 12 than in the placebo group (P = 0.175). The high withdrawal rate, most commonly due to untreated nausea and vomiting, in the buprenorphine group appears to have contributed to the primary end point not being achieved. Any participant who discontinued therapy before study end was treated as a treatment failure for the primary end point. Because more participants in the buprenorphine group than in the placebo group dropped out of the study, more treatment failures were recorded for this reason. This is supported by the PP analysis, which demonstrated a statistically significant difference, favoring buprenorphine patches (P < 0.001). Because the rate of discontinuations differed between groups, the PP analysis provides a better indication of the effectiveness of buprenorphine patches to relieve DPNP. The magnitude of the mean reduction in pain over placebo was 1.2 points on an 11-point NRS. Although this reduction is modest, it was highly significant and likely to be clinically meaningful for patients experiencing moderate to severe DPNP.
The high incidence of adverse effects experienced by participants in the buprenorphine group is similar to that reported in other trials of opioid analgesics, wherein 80% experienced any adverse event, with constipation (41%) and nausea (32%) being the most common (23). Despite the protocol allowing for the use of antiemetics and laxatives to manage opioid-related nausea and constipation, the use of antiemetics was limited. Almost 40% of premature study withdrawals in the buprenorphine group were due to intolerable nausea/vomiting, yet only one patient received treatment with antiemetics. When buprenorphine patches are initiated, patients should be informed that the drug can cause nausea, that this adverse effect is typically short lived, and that if nausea occurs, it can be effectively managed with antiemetics (24,25). To assist patients in persisting with buprenorphine therapy, antiemetics and/or aperients should be coprescribed to proactively manage opioid-induced nausea/vomiting and constipation (25).
The relative impact of difference in discontinuations may have been amplified by the fact that over time, a proportion of patients will experience a reduction in pain severity as part of the normal clinical course and thus become responders. Because a greater proportion of the placebo group completed the study, relatively more participants in this group potentially became responders by the end of the study because of this phenomenon than in the buprenorphine group. Longitudinal multivariate analysis accounts for the difference in withdrawal rates, and this modeling demonstrated a greater reduction in pain intensity in the buprenorphine group than in the placebo group, with this treatment effect being statistically significant at all time points.
The proportion of participants who discontinued therapy was higher in the buprenorphine group, which might have affected the blinding of the trial. This can have such consequences as influencing the use of adjunctive therapies, dose titration, participant response to the intervention, and ultimately, study outcomes (26).
Another factor that contributed to the primary end point was the higher-than-expected response to placebo. Power calculations were based on 10% of patients in the placebo group experiencing a 30% reduction in pain, whereas the actual result was 41%. This finding may be due to a stronger belief in the effectiveness of the patches, especially because more patients in the placebo group were taking the highest strength patch (58.1%) than in the buprenorphine group (11.8%). In addition, the stronger-than-expected placebo effect may be due to the concept of regression toward the mean, where extreme values on the first measurement tend toward the mean value on subsequent assessments (27,28). However, another explanation is that the assumptions about the placebo response were underestimated and contributed to inadequate study power, as other trials have reported higher response rates to placebo as well (29).
Pain at night that disrupts sleep is one of the main clinical challenges caused by DPNP. In the current study, transdermal buprenorphine demonstrated effective relief of night pain. There was a significant reduction in night pain from the first week, with the peak effect being achieved by week 6 and maintained for the duration of the study. Buprenorphine patch use was also associated with improvements in quality of sleep.
The assessment phase was limited to 12 weeks, although DPNP is a chronic condition. However, most placebo-controlled assessments of neuropathic pain are of similar duration due to ethical reasons. The main strength of this study was the robust clinical study design that evaluated the clinical effectiveness of buprenorphine in a typical patient population by using a realistic titration schedule.
No one therapy is effective for all patients with DPNP, and there is no way to predict which patient will respond to which therapy (5,30). Additional analgesics with robust evidence of clinical effectiveness are needed. The current study has demonstrated that buprenorphine patches are an effective therapy for DPNP when tolerated. Twice as many patients in the buprenorphine group experienced at least a 50% reduction in pain than in the placebo group. This equated to an NNT of 7.05, which is comparable to existing first-line therapies such as high-dose pregabalin (NNT 6.3) (6) and duloxetine (NNT 5) (7). Because opioids are normally a second-line therapy for the management of DPNP, they are often used in combination with first-line therapies (30,31). It is reassuring that the current study demonstrated that buprenorphine patches are effective in patients concomitantly taking antidepressants or antiepileptics.
This study indicates that when tolerated, buprenorphine patches are an effective therapy for DPNP and provide another option to manage this challenging chronic condition. Common opioid-related adverse effects, such as nausea and constipation, need to be anticipated and proactively managed with both patient education and coprescription of antiemetics and aperients to optimize tolerability and to facilitate an adequate trial of buprenorphine patch therapy.
Acknowledgments. The authors thank the coinvestigators in the multicenter study. Lead investigators for each study site are listed in the Supplementary Data. The authors also thank George Krassas from Scius Healthcare Solutions Pty Ltd. (Northbridge, NSW, Australia) for assisting with the preparation of the manuscript.
Duality of Interest. This clinical trial and George Krassas’s contribution were financially supported by Mundipharma, Sydney, NSW, Australia. NORSPAN is a registered trademark. No other potential conflicts of interest relevant to this article were reported.
Author Contributions. R.W.S. contributed as a study investigator and to the outline, review, editing, and final approval of the manuscript. J.H.W. contributed to the clinical trial design, statistical analysis, and outline, review, editing, and final approval of the manuscript. R.W.S. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Prior Presentation. Parts of this study were presented in abstract form at the Annual Scientific Meeting of the Australian Diabetes Society and Australian Diabetes Educators Association, Southwest Victoria, VIC, Australia, 27–29 August 2014.
Clinical trial reg. no. ACTRN12609000647235, www.anzctr.org.au.
This article contains Supplementary Data online at http://care.diabetesjournals.org/lookup/suppl/doi:10.2337/dc16-0123/-/DC1.
- Received January 19, 2016.
- Accepted May 25, 2016.
- © 2016 by the American Diabetes Association.
Readers may use this article as long as the work is properly cited, the use is educational and not for profit, and the work is not altered. More information is available at http://diabetesjournals.org/site/license.