Clinical Therapeutics/Volume 32, Number 5, 2010

844 Volume 32 Number 5

The results of this study were presented as a poster at the Annual Meeting of the Canadian Pain Society, May 23–26, 2007, Ottawa, Ontario, Canada.

Accepted for publication March 22, 2010.

Express Track online publication April 21, 2010.doi: 10.1016/j.clinthera.2010.04.0180149-2918/$ - see front matter

© 2010 Excerpta Medica Inc. All rights reserved.

ABSTRACTBackground: Buprenorphine is a mixed-activity, partial

µ-opioid agonist. Its lipid solubility makes it well suited for transdermal administration.

Objective: This study assessed the efficacy and safety profile of a 7-day buprenorphine transdermal system (BTDS) in adult (age >18 years) patients with moderate to severe chronic low back pain previously treated with ≥1 tablet daily of an opioid analgesic.

Methods: This was a randomized, double-blind, placebo-controlled crossover study, followed by an open-label extension phase. After a 2- to 7-day washout of previous opioid therapy, eligible patients were random-ized to receive BTDS 10 µg/h or matching placebo patches. The dose was titrated weekly using 10- and 20-µg/h patches (maximum, 40 µg/h) based on efficacy and tolerability. After 4 weeks, patients crossed over to the alternative treatment for another 4 weeks. Patients who completed the double-blind study were eligible to enter the 6-month open-label phase. Rescue analgesia was provided as acetaminophen 325 mg to be taken as 1 or 2 tablets every 4 to 6 hours as needed. The primary outcome assessments were daily pain intensity, measured on a 100-mm visual analog scale (VAS), from no painto excruciating pain, and a 5-point ordinal scale, from

0 = none to 4 = excruciating. Secondary outcome as-sessments included the Pain and Sleep Questionnaire (100-mm VAS, from never to always), Pain Disability Index (ordinal scale, from 0 = no disability to 11 = total disability), Quebec Back Pain Disability Scale (categori-cal scale, from 0 = no difficulty to 5 = unable to do), and the 36-item Short Form Health Survey (SF-36). Patients and investigators assessed overall treatment effectiveness at the end of each phase; they assessed treatment preference at the end of double-blind treat-ment. After implementation of a precautionary amend-ment, the QTc interval was measured 3 to 4 days after randomization and after any dose adjustment. All as-sessments performed during the double-blind phase were also performed every 2 months during the open-label extension. Adverse events were collected by non-directed questioning throughout the study.

Buprenorphine Transdermal System in Adults With Chronic Low Back Pain: A Randomized, Double-Blind, Placebo-Controlled Crossover Study, Followed by an Open-Label Extension Phase

Allan Gordon, MD1; Denis Callaghan, MD2; Donald Spink, MD3; Christian Cloutier, MD4; Peter Dzongowski, MD5; William O’Mahony, MD6; Duncan Sinclair, MD7; Saifudin Rashiq, MD8; Norm Buckley, MD9; Geoffrey Cohen, MD10; James Kim, MD11; Aline Boulanger, MD12; Paula S. Piraino, PhD13; John Eisenhoffer, MD13; Zoltan Harsanyi, MBA13; Andrew C. Darke, PhD13; and Kenneth J. Michalko, PharmD13

1Wasser Pain Management Centre, Toronto, Ontario, Canada; 2private practice, Hamilton, Ontario, Canada; 3Brookdale Research, Peterborough, Ontario, Canada; 4CHUS–Hôpital Fleurimont, Sherbrooke, Quebec, Canada; 5London East Medical Centre, London, Ontario, Canada; 6Corunna Medical Services Ltd., Corunna, Ontario, Canada; 7private practice, Aylmer, Ontario, Canada; 8University of Alberta, Edmonton, Alberta, Canada; 9Department of Anesthesia, McMaster University, Hamilton, Ontario, Canada; 10Winston Churchill Medical Centre, Mississauga, Ontario, Canada; 11private practice, Brampton, Ontario, Canada; 12CHUM–Hotel Dieu de Montreal, Montreal, Quebec, Canada; and 13Purdue Pharma, Pickering, Ontario, Canada

May 2010 845

A. Gordon et al.

to severe low back pain in patients who had previously received opioids. The improvements in pain scores were sustained throughout the 6-month, open-label extension. (Current Controlled Trials identification number: ISRCTN 06013881) (Clin Ther. 2010;32:844–860) © 2010 Excerpta Medica Inc.

Key words: buprenorphine, low back pain, transdermal system, chronic noncancer pain.

INTRODUCTIONApproximately 4 in 5 Canadian adults experience at least one episode of back pain during their lifetime.1

Chronic back pain is generally defined as pain lasting longer than 3 months, although the pain may never fully resolve; patients may experience repeated exacer-bations, with lifetime recurrences in up to 85% of in-dividuals.2 Reviews of the societal costs of low back pain in several developed countries have noted the substantial economic consequences of treatment decisions regarding low back pain.3,4 These consequences involve both direct costs (eg, physician and hospital services, medical de-vices, medications, diagnostic testing) and indirect costs (eg, employment costs due to absenteeism and presentee-ism, loss of household productivity).

Management of low back pain is often multidisci-plinary. An evidence-based clinical practice guideline from the American College of Physicians and the Ameri-can Pain Society contains 7 distinct recommendations for the evaluation and treatment of low back pain.5

Among the treatment recommendations are discussion of self-care options, consideration of pharmacotherapy, and incorporation of nonpharmacologic therapy (eg, exercise, spinal manipulation, massage therapy) when patients do not improve with self-care.

The guideline recommends consideration of several pharmacotherapeutic options for patients with low back pain, including acetaminophen, NSAIDs, opioids, muscle relaxants, tricyclic antidepressants, benzodiazepines, and some herbal remedies.5 While acetaminophen and NSAIDs are generally first-line treatments for most patients with chronic low back pain, opioid analgesics are an option for patients with severe, disabling pain that is not, or is unlikely to be, controlled by first-line therapies. This is consistent with the World Health Organization’s analgesic ladder, which recommends that opioid analgesics be administered only after non-opioid options have failed, have become inadequate, or are deemed inappropriate.6

Results: Of 78 randomized patients, 52 (66.7%) completed at least 2 consecutive weeks of treatment in each study phase without major protocol violations (per-protocol [PP] population: 32 women, 20 men; mean [SD] age, 51.3 [11.4] years; mean weight, 85.5 [19.5] kg; 94% white, 4% black, 2% other). The mean (SD) dose of study medication during the last week of treatment was 29.8 (12.1) µg/h for BTDS and 32.9 (10.7) µg/h for placebo (P = NS). During the last week of treatment, BTDS was associated with significantly lower mean (SD) pain intensity scores compared with placebo on both the VAS (45.3 [21.3] vs 53.1 [24.3] mm, respec-tively; P = 0.022) and the 5-point ordinal scale (1.9 [0.7] vs 2.2 [0.8]; P = 0.044). The overall Pain and Sleep score was significantly lower with BTDS than with placebo (177.6 [125.5] vs 232.9 [131.9]; P = 0.027). There were no treatment differences on the Pain Disability Index, Quebec Back Pain Disability Scale, or SF-36; however, BTDS was associated with significant improvements compared with placebo on 2 individual Quebec Back Pain Disability Scale items (get out of bed: P = 0.042; sit in a chair for several hours: P = 0.022). Of the 48 patients/physicians in the PP popula-tion who rated the effectiveness of treatment, 64.6% of patients (n = 31) rated BTDS moderately or highly effective, as did 62.5% of investigators (n = 30). Among the 50 patients in the PP population who answered the preference question, 66.0% of patients (n = 33) pre-ferred the phase in which they received BTDS and 24.0% (n = 12) preferred the phase in which they re-ceived placebo (P = 0.001), with the remainder having no preference; among investigators, 60.0% (n = 30) and 28.0% (n = 14) preferred the BTDS and placebo phases, respectively (P = 0.008), with the remainder having no preference. The mean placebo-adjusted change from baseline in the QTc interval ranged from –0.8 to +3.8 milliseconds (P = NS). BTDS treatment was associated with a significantly higher frequency of nausea (P < 0.001), dizziness (P < 0.001), vomiting (P = 0.008), somnolence (P = 0.020), and dry mouth (P = 0.003), but not constipation. Of the 49 patients completing 8 weeks of double-blind treatment, 40 (81.6%) entered the 6-month, open-label extension study and 27 completed it. Improvements in pain scores achieved during the double-blind phase were maintained in these patients.

Conclusions: In the 8-week, double-blind portion of this study, BTDS 10 to 40 µg/h was effective compared with placebo in the management of chronic, moderate

846 Volume 32 Number 5

Clinical Therapeutics

(5, 10, and 20 µg/h).22 The BTDS is designed to be worn for 7 days, and steady-state pharmacokinetic studies have reported maintenance of stable plasma buprenorphine concentrations throughout this period, avoiding the frequent peaks and troughs associated with immediate-release oral or sublingual analgesic formulations.22 Through avoidance of hepatic first-pass metabolism, BTDS may be of potential benefit in pa-tients with altered bioavailability due to concomitant gastrointestinal disease or those who are vomiting or have difficulty swallowing.

Three previous studies have reported the efficacy and tolerability of BTDS in patients with chronic non-cancer pain. In an open-label, parallel-group, noninfe-riority study, adult patients with a clinical and radiologic diagnosis of osteoarthritis who were experiencing mod-erate to severe pain (mean Box Scale 11 score ≥4 during treatment with acetaminophen 4 g/d) were randomized to receive either low-dose BTDS (5–20 µg/d) or twice-daily prolonged-release tramadol (150–400 mg/d) for 12 weeks.23 Doses of both medications were titrated as needed to achieve stable pain control. One hundred thirty-four patients were randomized to treatment (69 BTDS, 65 tramadol; 59.4% and 53.8% female, respectively; 98.6% and 100% white; mean [SD] age, 64.4 [11.1] and 64.2 [9.3] years). Both BTDS and tra-madol were associated with clinically meaningful least squares mean changes from baseline in pain intensity scores (–2.26 [95% CI, –2.76 to –1.76] and –2.09 [95% CI, –2.61 to –1.58], respectively). The incidence of adverse events was comparable between treatment groups.

A double-blind, parallel-group, maintenance-of-analgesia study compared BTDS and placebo in adult patients with a ≥2-month history of noncancer pain for which they had received oral opioid combination analgesics.24 BTDS was titrated to effect over a 7- to 21-day, open-label run-in period. Patients who achieved stable pain control and were able to tolerate BTDS were randomly assigned to continue BTDS at the dose achieved during the run-in period or to receive place-bo for up to 14 days. Patients completed the study on day 14 or when they met predefined criteria for inef-fective treatment (needing >1 g/d of acetaminophen, requiring a change in dose of study drug, having dif-ficulty keeping the patch affixed, or discontinuing due to ineffective treatment without meeting any of the first 3 criteria). Five hundred eighty-eight patients entered the open-label run-in phase, and 267 were

A number of randomized controlled trials have reported the efficacy of opioids in the treatment of chronic noncancer pain, including back pain, osteoarthritis, and neuropathic pain.7–14 Based on an extensive review of the evidence, the American Pain Society has recently published a com-prehensive guideline on the use of opioids for chronic noncancer pain.15 This guideline supports the efficacy of opioids in the treatment of selected patients who have failed a documented trial of first-line treatments, with close monitoring for adverse effects and aberrant drug-related behavior. The guideline also recommends establishment of a prospective treatment plan with the patient, includ-ing agreement on objectives for control of pain and res-toration of function, as well as discussion of the risks and benefits of opioid therapy.15 The American Pain Society guideline is consistent with previous recommendations concerning universal precautions in pain medicine.16

Buprenorphine is a derivative of thebaine; it is a partial agonist of the µ- and opioid receptor-like 1 opioid receptors, and has antagonist activity at both the κ- and δ-opioid receptors.17,18 Because of its high affinity for and slow dissociation from the µ-opioid receptor (Ki = 1.33 nM, vs 230 nM for morphine), buprenorphine has a relatively long duration of action.17 At low doses (0.6 mg/d sublingual), buprenorphine is considered an alternative to codeine, whereas at higher doses (up to 3 mg/d sublingual), it is considered equivalent to 180 mg/d of oral morphine.6 No clinically important accumulation of metabolites has been observed with buprenorphine. As such, no dose adjustment is recom-mended in patients with renal dysfunction.19 Therefore, a consensus statement from an international expert panel recommended buprenorphine as a top-line choice for opioid treatment of the elderly.19

Buprenorphine recently became available in North America as a sublingual formulation for use in combi-nation with naloxone to deter intravenous abuse in the management of opioid dependence. Although it has low oral bioavailability (14%–16%),20 buprenorphine is highly lipid soluble (logarithmic octanol/water coeffi-cient T log P = 3.9)21 and is thus well suited for trans-dermal administration. A transdermal formulation of buprenorphine for analgesic use has been available in Europe for 5 years. The buprenorphine transdermal delivery system (BTDS)* provides 3 nominal flux rates

*Trademark: BuTranstm (Purdue Pharma, Pickering, Ontario, Canada). Marketed as BuTranstm or Norspan® in some areas other than North America.

May 2010 847

A. Gordon et al.

PATIENTS AND METHODSInclusion and Exclusion Criteria

Eligible patients were men and nonpregnant women aged >18 years who had experienced low back pain of at least moderate intensity (rated ≥2 on a 5-point ordinal scale [0 = none, 1 = mild, 2 = moderate, 3 = severe, 4 = excruciating])26 for >3 months and currently required ≥1 tablet daily of an opioid analgesic.

Patients whose history indicated that their pain was refractory to opioids and those with an allergy to opioids or acetaminophen were excluded. Patients who were undergoing any procedure or treatment (eg, physio-therapy, surgery) that was likely to change their pain and those with a clinically significant alternative source of pain were also excluded, as were those requiring external application of heat. Patients with suspected dependence on narcotics or alcohol and those with a history of major psychiatric disorders were excluded. Other exclusion criteria included elevated liver enzymes, decreased serum potassium or magnesium, severe organ dysfunction, head injury or seizures, chronic obstructive pulmonary disease, asthma, respiratory depression, cor pulmonale, congestive heart failure, atrial fibrillation, myocardial ischemia, tachycardia (>100 beats/min at rest) or bradycardia (<45 beats/min at rest), and peptic ulcer disease or inflammation of the gastrointes-tinal tract.

Because of the evidence that some opioids may cause prolongation of the QT interval,27 a precautionary amendment to the protocol was implemented ~7 to 8 months after the first patient was recruited to define additional exclusion criteria. From that point forward, patients with a personal or family history of congenital long QT syndrome and those who were receiving treat-ment with class IA or III antiarrhythmics or drugs known to prolong the QT interval or induce torsades de pointes were excluded. In addition, patients with a mean QTc interval >450 milliseconds or a single QTc interval >500 milliseconds at any time during the crossover trial were excluded. Any patients enrolled before implemen-tation of the amendment who met any of the above exclusion criteria were immediately withdrawn from the trial.

Research ethics boards at the 13 participating Cana-dian centers approved the protocol and informed-consent form, and each patient gave written informed consent before participating in the study. After imple-mentation of the amendment, all patients who were currently in either the double-blind or open-label phase

randomized to double-blind treatment (129 BTDS, 138 placebo; 61.2% and 63.8% female, respectively; 99.2% and 98.6% white; mean [SD] age, 56.2 [13.3] and 59.2 [11.5] years). The proportion of patients with ineffective treatment (primary efficacy measure) was significantly lower with BTDS than with placebo (51.2% vs 65.0%; 95% CI, 1.09–2.95). In the second-ary analyses, the median time from the first dose of double-blind medication to ineffective treatment was significantly longer with BTDS than with placebo (10 vs 3 days; P = 0.011). Buprenorphine was associated with higher frequencies of certain adverse events compared with placebo: pruritus (9.3% vs 5.1%, respectively), headache (3.9% vs 2.2%), and somnolence (2.3% vs 0.7%).

A recent randomized, double-blind, placebo-controlled crossover study evaluated BTDS in adult patients with chronic low back pain of moderate or greater severity for ≥6 weeks who had not achieved adequate pain relief with nonopioid analgesics.25 A large proportion (58%) of enrolled patients had not received previous opioid treatment. Patients discontin-ued prestudy analgesics before being randomized to receive either BTDS 5 µg/h or placebo. The dose was titrated to 10 µg/h and then to a maximum of 20 µg/h based on pain relief and adverse effects. After 4 weeks of the assigned treatment, patients crossed over to the alternative treatment for another 4 weeks. Seventy-nine patients were enrolled in the study, and 53 (28 men, 25 women; mean [SD] age, 54.5 [12.7] years) completed ≥2 consecutive weeks in each phase of the crossover and were evaluable for efficacy. BTDS was associated with lower mean (SD) daily pain intensity scores (pri-mary efficacy measure) compared with placebo on both a visual analog scale (VAS) (37.6 [20.7] vs 43.6 [21.2] mm, respectively; P = 0.049) and an ordinal scale (1.7 [0.6] vs 2.0 [0.7]; P = 0.036). The number of patients reporting ≥1 adverse event was significantly greater during receipt of BTDS than during receipt of placebo (98.6% vs 89.2%; P = 0.014).

The present study was designed to further assess the therapeutic potential of BTDS. The clinical efficacy and safety profile of a higher initial dose (10 µg/h) and higher maximum dose (40 µg/h) of BTDS than in the previous study25 were assessed in an 8-week, randomized, double-blind, placebo-controlled crossover study in patients with chronic low back pain who were already receiving opioids. This was followed by a 6-month, open-label extension phase.

848 Volume 32 Number 5

Clinical Therapeutics

out in the same manner in both phases of the crossover. Patients who successfully completed both phases of the double-blind study were eligible to continue receiving BTDS in the 6-month, open-label extension, with the investigator’s concurrence.

Study AssessmentsThroughout the washout period and the double-blind

treatment phase, patients recorded their pain intensity in a diary twice daily (8:00 am and 8:00 pm) using an unmarked 100-mm VAS (from no pain to excruciating pain) and the 5-point ordinal scale (from 0 = none to 4 = excruciating). At each clinic visit during the double-blind phase and every 2 months during the open-label phase, the 100-mm VAS and 5-point ordinal scale were also used to rate pain intensity over the previous 24 hours and over the previous week.

The impact of pain on sleep since the last evaluation was assessed using an 8-item Pain and Sleep Question-naire at baseline, crossover, and the end of double-blind treatment, and every 2 months during the open-label extension.13,28–30 This instrument draws from 2 validated measures of sleep disturbance—the Chronic Pain Sleep Inventory and the Verran and Snyder-Halpern Sleep Scale.31–33 Six items were rated on a 100-mm VAS from never to always; quality of sleep was rated on a 100-mm VAS from very poor to excellent, and the final item involved the number of hours of sleep per night. Items 1 through 5 (trouble falling asleep, needed pain medica-tion to sleep, needed sleep medication to sleep, awakened by pain at night, and awakened by pain in the morning) were summed to derive an overall score.13,28,30

Patients rated their pain-related disability at baseline, crossover, and the end of double-blind treatment, and every 2 months during the open-label phase using the Pain Disability Index (PDI).34,35 The PDI consists of 7 subscales, each rated on an ordinal scale from 0 = no disability to 11 = total disability: family/home responsibilities, recreation, social activity, occupation, sexual behavior, self-care, and life-support activity. An overall disability score was determined by summing the individual ratings.7,12,13,28,29

Functional disability was assessed at baseline, cross-over, and the end of double-blind treatment, and every 2 months during the open-label phase using the Quebec Back Pain Disability Scale (QBPDS).36 The scale consists of the following 20 items: get out of bed; sleep through the night; turn over in bed; ride in a car; stand up 20 to 30 minutes; sit in a chair for several hours; climb one

of the study were presented with the new safety infor-mation, and a second written informed consent was obtained.

Study DesignThis was a randomized, double-blind, placebo-

controlled crossover study, followed by an open-label extension phase. Screening included a medical history and physical examination, including assessment of the nociceptive and neuropathic components of pain. Stan-dard clinical laboratory tests (hematology, serum chem-istry, and urinalysis) were performed. Three 12-lead ECGs were obtained 10 minutes apart for determination of the mean QTc interval.

Eligible patients underwent a 2- to 7-day washout of opioid analgesics before receiving a sequential patient number and being randomized to receive BTDS 10 µg/h or matching placebo patches. A block-randomization procedure was used to generate the treatment alloca-tions: for every 4 successive patients, 2 received BTDS in the first phase and 2 received BTDS in the second phase. The randomization code was generated using PROC PLAN in SAS version 6.12 (SAS Institute Inc., Cary, North Carolina). Study monitors, investigators, coordinators, pharmacists, patients, and sponsor clinical research personnel remained blinded to treatment al-location throughout the conduct of the study.

All patches were to be worn for 6 to 8 days. The initial dose was titrated weekly to 20 µg/h and a maxi-mum of 40 µg/h using 10- and 20-µg/h patches based on pain relief and adverse events. An intermediate 30-µg/h dose was permitted only when the investigator considered 20 µg/h too low (due to lack of pain control) or 40 µg/h too high (due to adverse events). Patients in the placebo arm underwent titration with corresponding placebo patch strengths.

Both treatment arms were provided with acetamino-phen tablets (325 mg) to be taken as 1 or 2 tablets every 4 to 6 hours as needed for unmanageable pain. Patients were permitted to continue nonopioid analgesics at doses that had been stable for 2 weeks before enroll-ment, and antidepressants or anticonvulsants at doses that had been stable for 8 weeks before enrollment.

Patients returned to the clinic for weekly visits. After 4 weeks of the assigned treatment, they crossed over to the alternative treatment for an additional 4 weeks. After implementation of the amendment, ECGs were obtained 3 to 4 days after randomization to study drug and after any dose change. Dose titration was carried

May 2010 849

A. Gordon et al.

done to subjects already experiencing withdrawal symp-toms) that are expected to alter the severity of opioid withdrawal symptoms.40,41 A modified version of the SOWS was used as a structured probe for symptoms potentially related to opioid withdrawal during the double-blind phase of the study. It consisted of 22 symptoms, 15 taken from the original SOWS (item 16, I feel like shooting up today, is omitted as not being relevant to the target population) and an additional 7 that have been reported in association with opioid withdrawal in pain patients: my sleeping habits have changed; I had to sneeze; I had diarrhea; I felt weak; I felt my heart beating faster; I had an unexplained fever; and I did not feel like eating. The intensity of each of the 22 symptoms was rated by patients on a 5-point scale (0 = not at all, 1 = a little, 2 = moderately, 3 = quite a bit, 4 = extremely). The mean of all 22 ratings was then calculated.

Additional adverse events were captured by nondi-rected, open-ended questioning. Adverse events, whether spontaneously reported by patients or observed by investigators, were recorded (including their severity [1 = mild, 2 = moderate, or 3 = severe] and potential causality) at each clinic visit during the double-blind and open-label phases. Patients used a 100-mm VAS to retrospectively assess nausea (from no nausea to severe nausea) twice daily and drowsiness (from no drowsinessto extreme drowsiness) once daily, and recorded their assessments in the patient diary.

Outcome Measures and Statistical AnalysisDuring protocol development, the difference in VAS

pain scores with BTDS was estimated to be ~10 mm. Using estimates of SD and intrapatient correlation for VAS pain scores from 2 previous crossover studies of opioids in patients with chronic low back pain,25,39 it was estimated that 44 patients completing the study would provide 80% power (β = 0.20) at the 5% signifi-cance level (α = 0.05) to detect a 10-mm difference in VAS pain intensity in the primary analysis.

The primary efficacy measures were the VAS and 5-point ordinal scale ratings of pain intensity from the patient diaries. Mean pain intensity scores were calcu-lated by day (across the 2 daily assessments), by week, and by phase. Because the buprenorphine dose was titrated in each phase, mean pain intensity values were calculated for the last week of treatment in each phase for the primary analysis. All means are presented with the associated SDs to represent the variability of the

flight of stairs; walk a few blocks; walk several miles; reach up to high shelves; throw a ball; run one block; take food from the fridge; make the bed; put on socks; clean the bathtub; move a chair; pull or push heavy doors; carry 2 bags of groceries; and lift and carry a heavy suitcase. Each item is rated on a 6-point categori-cal scale (0 = no difficulty, 1 = minimally difficult, 2 = somewhat difficult, 3 = fairly difficult, 4 = very difficult, 5 = unable to do). The individual item scores are then summed to yield a total score.

At baseline, crossover, and the end of double-blind treatment, and every 2 months during the open-label phase, the 36-item Short Form Health Survey (SF-36), a general health status outcome measure, was admin-istered.37,38 The SF-36 includes one multi-item scale measuring 8 health concepts: physical functioning, role limitations due to physical health problems, bodily pain, general health, vitality (energy/fatigue), social functioning, role limitations due to emotional problems, and mental health (psychological distress and psychological well-being). The SF-36 Physical Component Summary Scale is derived from the physical functioning, role limitations–physical, bodily pain, and general health items, and the SF-36 Mental Component Summary Scale is derived from the vitality, social functioning, role limitations–emotional, and mental health items. For all SF-36 items, higher scores indicate better functioning.

The effectiveness of treatment was assessed by pa-tients and investigators at screening, baseline, crossover, and the end of double-blind treatment, and every 2 months during open-label treatment using a 4-point categorical scale (0 = not effective, 1 = slightly effective, 2 = mod-erately effective, 3 = highly effective).12,13,39 Overall treatment-phase preference was assessed by patients and investigators at the end of the study, before unblind-ing of treatment allocation, by responding to the ques-tion, “Which treatment period (period 1, period 2, or no preference) did you prefer in the management of your/your patient’s pain?” Clinical benefit was assessed by investigators at the end of the double-blind phase and the end of the open-label phase using a 4-point ordinal scale (1 = great deal of benefit, 2 = moderate benefit, 3 = slight benefit, 4 = no benefit).

The Subjective Opioid Withdrawal Scale (SOWS) is a validated, 16-item rating scale used to assess the signs and symptoms of withdrawal in patients with an estab-lished opiate addiction.40 It has been found to be sensitive to interventions (eg, administration of opioid antagonists to subjects not in withdrawal, administration of metha-

850 Volume 32 Number 5

Clinical Therapeutics

a given ECG trace and the RR interval was taken from the rhythm strip over ≥5 beats. The resulting corrected values were compared with baseline values adjusted for the effect of placebo using the paired t test.43

Adverse events were coded using Coding Symbols for a Thesaurus of Adverse Reaction Terms IV preferred terms. The McNemar test, which can be used only when patients are exposed to both treatments, was applied to detect differences in the overall frequency of adverse events between treatments.43

Patients’ demographic characteristics were summa-rized using descriptive statistics. All data are reported as mean (SD). Statistical significance was defined as P < 0.05 for a 2-tailed hypothesis. Analyses were per-formed using SAS version 6.12 (SAS Institute Inc.).43

RESULTSPatient Disposition and Baseline Characteristics

Of the 99 patients screened for eligibility, 78 were ran-domized to treatment (47 women, 31 men; mean [SD] age, 50.7 [11.9] years). Among all randomized patients, the mean duration of low back pain was 12.9 (10.6) years. Forty-seven patients had neuropathic pain (mean dura-tion, 11.4 [8.4] years), 60 had nociceptive pain (mean duration, 12.0 [9.2] years), and some had both types of pain. Before the study, 88.5% of patients (n = 69) were taking combination opioid/nonopioid preparations, including codeine combinations (n = 51) and oxycodone combinations (n = 18); 38.5% of patients (n = 30) were taking single-entity opioids before enrollment, including morphine (n = 7), hydromorphone (n = 6), oxycodone (n = 5), buprenorphine (n = 4), codeine (n = 2), pethidine (n = 2), methadone (n = 2), pentazocine (n = 1), and tramadol (n = 1).

Twenty-nine patients withdrew after randomization (Figure). The reasons for withdrawal were adverse events (19), lack of efficacy (5), loss to follow-up (2), with-drawal of consent (1), protocol violation (1), and early termination due to an upcoming planned surgery (1). Of the 19 patients who discontinued due to adverse events, 16 were receiving BTDS and 3 were receiving placebo at the time of withdrawal. Fifty-five patients completed ≥2 consecutive weeks of treatment in each phase of the study, although 3 of them were excluded from the PP analysis because of protocol violations (1 for noncompliance with dosing of study medication and 2 for altering the dosing of concomitant pain medi-cations). Therefore, 52 patients were included in the PP population (32 women, 20 men; mean [SD] age,

mean estimates. The standard pain-intensity measure is a scale from 0 to 10, for which means and SDs or SEs (rather than medians) are the generally reported mea-sures of central tendency.

Secondary end points included the Pain and Sleep Questionnaire, PDI, QBPDS, SF-36, assessments of treat-ment effectiveness and preference, SOWS, and VAS ratings of nausea and drowsiness from the patient diaries.

The primary efficacy analysis included all patients who completed ≥2 consecutive weeks of treatment in each study phase with no major protocol violations (per-protocol [PP] population). All patients who received study medication were evaluated for safety. The full-analysis set (intent-to-treat [ITT] population) was used to support the results of the primary efficacy analysis and the assessment of treatment preference. Because there were 2 primary end points (VAS and ordinal scale) and the primary comparison between BTDS and placebo would not be considered statistically significant unless both components were significant, no adjustment for multiple testing was considered necessary; similarly, because prespecified secondary end points are not con-firmatory, no adjustment for multiple testing was con-sidered necessary.42

For all primary and secondary variables, 3-way ANOVA was used to test for the effect of treatment, phase, and sequence (carryover) using mean scores by treatment.43 The patient-within-sequence variance was used as the error term for testing the sequence effect. Analysis of the SF-36 was based on recommendations from its developer, the Medical Outcomes Trust. Rates of treatment preference were tested using the Prescott test,43 which accounts for responses of no preference. Use of rescue medication was compared by treatment based on mean daily consumption, summarized by week.

The change from baseline in the primary and second-ary end points was calculated for each treatment during the double-blind phase. The change from baseline dur-ing the open-label phase was calculated using the last week of treatment during the double-blind phase as baseline. In both cases, the paired t test was used to test the null hypothesis of no change from baseline.43

In keeping with Health Canada and International Conference on Harmonisation44 norms, correction of the QT interval for heart rate was made using both the Frid-ericia cube-root correction (QTcF = QT/RR[0.33]) and the Bazett square-root correction (QTcB = QT/RR[0.5]), where the QT interval was taken as the mean of ≥3 beats on

May 2010 851

A. Gordon et al.

acetaminophen (3.6 [3.4] vs 3.9 [3.2] tablets/d, respectively).

Pain IntensityAt baseline, the mean (SD) daily pain intensity ratings

were 60.9 (15.4) mm on the VAS scale and 2.6 (0.5) on the 5-point ordinal scale. During the last week of treat-ment, the mean VAS score was significantly lower in those receiving BTDS compared with those receiving placebo (45.3 [21.3] vs 53.1 [24.3] mm, respectively; P = 0.022). BTDS treatment was associated with a 25.6% decrease from baseline in VAS pain intensity (P < 0.001), compared with a 12.9% decrease with placebo (P = 0.018). During the last week of treatment, 5-point ordinal scale ratings were also significantly lower with BTDS compared with placebo (1.9 [0.7] vs

51.3 [11.4] years; mean weight, 85.5 [19.5] kg; 94% white, 4% black, 2% other).

Use of Study Medication and Rescue MedicationDuring the last week of double-blind treatment, the

majority of patients were using the 40-µg/h patch strength of BTDS and placebo (55% and 67%, respec-tively). In the PP population, the mean (SD) dose of study medication during the last week of treatment was 29.8 (12.1) µg/h for BTDS and 32.9 (10.7) µg/h for placebo (P = NS). In the ITT population, the mean dose during the last week of treatment was significantly lower for BTDS than for placebo (26.5 [12.9] vs 32.6 [10.9] µg/h, respectively; P = 0.006).

There was no significant difference between BTDS and placebo with respect to consumption of rescue

Randomized(N = 78)

BTDS(n = 39)

Withdrawn (n = 10)AE (9)Lost to follow-up (1)

Placebo(n = 29)

Withdrawn (n = 6)Lack of efficacy (3)AE (2)Consent withdrawn (1)

BTDS(n = 35)

Withdrawn (n = 9)AE (7)Lack of efficacy (1)Other (1)

Placebo(n = 39)

Withdrawn (n = 4)AE (1)Lack of efficacy (1)Lost to follow-up (1)Protocol violation (1)

ITT/safety analysis (n = 78)Completed ≥2 consecutive weeks in each phase (n = 55)

Excluded from efficacy analysis (n = 3)PP population (n = 52)

Completed 8 weeks of treatment (n = 49)

Phase 1(4 wk)

Phase 2(4 wk)

Figure. Patient disposition in the randomized, double-blind portion of the study. BTDS = buprenorphine transdermal system; AE = adverse event; ITT = intent-to-treat; PP = per-protocol.

852 Volume 32 Number 5

Clinical Therapeutics

improvements from baseline on 7 of 8 items (P range, <0.001–0.034), compared with significant improvements on 4 items with placebo (P range, <0.001–0.014).

Pain and Disability IndexBoth BTDS and placebo were associated with sig-

nificant improvements from baseline in the total PDI score (16.8% [P = 0.003] and 11.9% [P = 0.033], re-spectively). BTDS was associated with significant im-provements from baseline on 4 of 7 PDI subscales (P range, <0.001–0.020), compared with improvements on 3 subscales with placebo (P range, 0.003–0.043). During the final week of treatment, there were no sig-nificant differences between BTDS and placebo on any of the 7 subscales or the total PDI score.

Quebec Back Pain Disability ScaleBoth BTDS and placebo were associated with sig-

nificant improvements from baseline in the total QBPDS score (19.3% [P < 0.001] and 14.0% [P = 0.001], re-spectively), with no significant difference between treat-ments. There were significant improvements from base-line on 17 of 20 QBPDS items with BTDS (P range, <0.001–0.041), and on 11 items with placebo (P range, <0.001–0.026). BTDS was associated with significantly better scores on 2 QBPDS items compared with placebo:

2.2 [0.8]; P = 0.044). BTDS was associated with a 23.9% decrease from baseline in ordinal scale pain intensity (P < 0.001), compared with a 15.1% decrease with placebo (P = 0.001).

The ITT analysis supported the results of the PP analysis. Pain intensity scores were significantly lower with BTDS compared with placebo on both the VAS (44.6 [21.4] vs 52.4 [24.0] mm, respectively; P = 0.005) and the ordinal scale (2.0 [0.7] vs 2.2 [0.8]; P = 0.016). The decreases from baseline in VAS pain intensity were 25.4% for BTDS (P < 0.001) and 12.0% for placebo (P = 0.011); the corresponding decreases in ordinal scale pain intensity were 22.5% (P < 0.001) and 13.3% (P = 0.001).

Pain and Sleep QuestionnaireThe overall score on the Pain and Sleep Questionnaire

was significantly lower with BTDS compared with placebo (P = 0.027), and scores on 2 of the 8 individual questionnaire items were significantly lower with BTDS compared with placebo: trouble falling asleep (P = 0.012) and awakened by pain in the morning (P = 0.036) (Table I). The overall Pain and Sleep Questionnaire score was significantly improved from baseline with both BTDS (37.3%; P < 0.001) and placebo (18.9%; P = 0.004). BTDS was also associated with significant

Table I. Pain and Sleep Questionnaire scores* during the last week of treatment, per-protocol population. Values are mean (SD).

Questionnaire ItemScore at Baseline

Score in Last Week of Treatment P, BTDS Versus

PlaceboBTDS Placebo

Trouble falling asleep 64.7 (26.5) 36.4 (29.1)† 51.2 (29.5)‡ 0.012Needed pain medication to sleep 57.1 (32.2) 34.9 (30.8)† 38.2 (33.6)‡ 0.603Needed sleeping medication to sleep 33.4 (39.7) 24.0 (33.1) 33.0 (37.6) 0.066Awakened by pain at night 62.5 (27.1) 38.8 (29.7)† 52.3 (31.5)‡ 0.070Awakened by pain in the morning 64.4 (29.6) 43.5 (34.0)† 57.1 (29.8) 0.036Partner awakened 48.1 (35.3) 34.0 (31.2)† 37.4 (33.4) 0.543No. of hours of sleep 5.5 (1.6) 6.1 (1.5)† 5.7 (1.2) 0.102Quality of sleep 21.6 (19.5) 39.6 (24.1)† 36.2 (25.7)‡ 0.732 Overall score 283.4 (123.6) 177.6 (125.5)† 232.9 (131.9)‡ 0.027

BTDS = buprenorphine transdermal system. * The f irst 6 items were rated on a 100-mm visual analog scale (VAS) (from never to always). Quality of sleep was rated on

a 100-mm VAS (from very poor to excellent). The overall score was the sum of items 1 through 5. † P range, <0.001 to 0.034 versus baseline. ‡ P range, <0.001 to 0.014 versus baseline.

May 2010 853

A. Gordon et al.

Adverse EventsOverall, patients reported significantly more adverse

events when they received BTDS than when they received placebo (P = 0.003). Sixty-five patients receiving BTDS (89.0% of the 73 patients who received BTDS and provided data) reported 349 adverse events (mean maximum severity, 1.8), and 44 patients receiving pla-cebo (64.7% of the 68 who received placebo and provided data) reported 179 adverse events (mean maximum severity, 1.6).

The most frequently reported adverse events are listed in Table II. Compared with placebo, BTDS was associ-ated with a significantly higher frequency of nausea (P < 0.001), dizziness (P < 0.001), vomiting (P = 0.008), somnolence (P = 0.020), and dry mouth (P = 0.003). The frequency of constipation did not differ significantly between BTDS and placebo.

Skin reactions, specifically at the patch-application site, were reported in 20 and 18 patients receiving BTDS and placebo, respectively. The skin reaction was rated as severe in only 1 patient each. No serious adverse events were reported during the double-blind phase of the study.

Based on VAS scores recorded in the patient diary, BTDS was associated with significantly more nausea compared with placebo (mean [SD], 19.7 [22.5] vs 9.7 [18.9] mm, respectively; P < 0.001). BTDS was also associated with significantly more drowsiness compared with placebo (29.6 [26.7] vs 23.4 [28.1] mm; P = 0.032).

Modified SOWSThe mean modified SOWS score was significantly

higher at the baseline assessment (before randomization) than at the screening visit (before washout of prestudy opioids) (1.2 [0.8] vs 0.8 [0.6], respectively; P < 0.001). When analyzed independently of treatment sequence, the mean modified SOWS score did not differ signifi-cantly between the first week of the BTDS phase and the first week of the placebo phase (0.8 [0.6] vs 0.8 [0.7], respectively). Similarly, there was no significant differ-ence in modified SOWS score in the first week of treat-ment after the baseline evaluation between those who received BTDS first and those who received placebo first (0.8 [0.6] and 0.8 [0.6], respectively).

In the subgroup of patients who crossed over to placebo after 4 weeks of BTDS treatment, modified SOWS scores in the last week of BTDS and the first week of placebo were 0.8 (0. 7) and 0.9 (0.7) (P = NS), respectively, compared with an increase from 0.7 (0.6)

get out of bed (P = 0.042) and sit in a chair for several hours (P = 0.022).

SF-36Both BTDS and placebo were associated with signifi-

cant improvements from baseline on 4 of the 8 SF-36 categories—role limitations due to physical health problems, bodily pain, vitality, and social function-ing (BTDS: P range, <0.001–0.017; placebo: P range, <0.001–0.025). BTDS and placebo also were associated with significant improvements from baseline on the physical summary score (18.2% [P = 0.002] and 14.3% [P = 0.011], respectively). There were no significant differences between treatments on any of the 8 catego-ries, nor on the physical or mental health summary scores.

Treatment Effectiveness and PreferenceThe effectiveness of BTDS was rated significantly

higher than that of placebo by both patients (1.8 [1.1] vs 1.0 [1.1], respectively; P = 0.016) and investigators (1.8 [1.0] vs 1.0 [1.1]; P = 0.013). In the 48 patients in the PP population who rated the effectiveness of treat-ment, 64.6% (n = 31) rated BTDS moderately or high-ly effective, compared with 37.5% (n = 18) rating pla-cebo moderately or highly effective. The corresponding proportions of investigators rating treatment moderate-ly or highly effective were 62.5% (n = 30) and 35.4% (n = 17).

Among the 50 patients in the PP population who answered the preference question, 66.0% (n = 33) preferred the phase in which they received BTDS, 24.0% (n = 12) preferred the phase in which they received placebo, and 10.0% (n = 5) had no preference (P = 0.001, BTDS vs placebo). Similarly, among investi-gators, 60.0% (n = 30) and 28.0% (n = 14) preferred the BTDS and placebo phases, respectively (P = 0.008, BTDS vs placebo), whereas 12.0% (n = 6) had no preference.

These results were supported by the ITT analysis, in which 57.4% (35/61) of patients preferred the phase in which they received BTDS, 27.9% (17/61) preferred the phase in which they received placebo, and 14.8% (9/61) had no preference (P = 0.004, BTDS vs placebo). Similarly, investigators preferred the BTDS phase for 52.5% (32/61) of patients, preferred the placebo phase for 29.5% of patients (18/61), and had no preference for 18.0% of patients (11/61) (P = 0.019, BTDS vs placebo).

854 Volume 32 Number 5

Clinical Therapeutics

and 1.5 milliseconds); in both cases, the QTcF interval was <450 milliseconds.

Open-Label ExtensionOf 49 eligible patients completing the double-blind

study, 40 (81.6%) chose to continue BTDS in the open-label phase (22 women, 18 men; mean age, 51.4 [12.1] years; mean weight, 85.5 [19.9] kg). Thirteen patients withdrew from the open-label phase due to adverse events that included nausea, dizziness, somnolence, asthenia, vomit-ing, sweating, constipation, and rash (7 patients, some with >1 adverse event); loss to follow-up (2); insufficient therapeutic effect (1); withdrawal of consent (1); with-drawal by the investigator (1); and remission of the original indication for treatment (1). Thus, 27 patients (67.5%) completed the 6-month open-label extension. The mean (SD) duration of exposure to BTDS in all patients who entered the open-label phase was 136.1 (65.5) days; when this was combined with data from the double-blind phase, the mean duration of exposure to BTDS was 164.0 (65.6) days. The mean final doses of BTDS in the open-label phase were 22.6 (10.1) µg/h for all evaluable patients and 24.3 (10.6) µg/h for those who completed

at screening to 1.2 (1.0) after withdrawal of prestudy opioids (P = 0.004). Investigators reported no symptoms of opioid withdrawal at any time during the study.

ECG AnalysisBecause the amendment requiring calculation of the

QTc interval was introduced after enrollment had al-ready commenced, baseline ECGs were available for only 47 of 78 patients randomized to treatment in the double-blind phase. During randomized treatment, ECGs were available for 42, 30, and 17 patients receiv-ing BTDS doses of 10, 20, and 40 µg/h, respective-ly. Mean placebo-adjusted changes from baseline in the QTcF and QTcB intervals ranged from –0.8 to +3.8 milliseconds (P = NS). No patients had a QTc interval >500 milliseconds or an increase from baseline of >100 milliseconds. During the double-blind phase, 5 patients had a >60-millisecond increase in QTcB interval to >450 milliseconds while receiving BTDS (1 patient also had an increase while receiving placebo), although in all cases, the increase in QTcF was <60 mil-liseconds. During open-label treatment, 2 patients had a QTcB interval that exceeded 450 milliseconds (by 1.8

Table II. Incidence of most common adverse events (AEs).

BTDS (n = 73)

Placebo (n = 68)

AENo. (%) of Patients

Mean Maximum Severity*

No. (%) of Patients

Mean Maximum Severity* P

Overall 65 (89.0) 1.8 44 (64.7) 1.6 0.003

Individual AEs Nausea 39 (53.4) 1.5 12 (17.6) 1.1 <0.001 Dizziness 24 (32.9) 1.4 3 (4.4) 1.3 <0.001 Pruritus 17 (23.3) 1.2 14 (20.6) 1.5 0.467 Vomiting 16 (21.9) 1.9 3 (4.4) 1.3 0.008 Somnolence 16 (21.9) 1.7 5 (7.4) 1.4 0.020 Rash 13 (17.8) 1.2 12 (17.6) 1.3 1.000 Dry mouth 13 (17.8) 1.2 0 0 0.003 Constipation 12 (16.4) 1.5 4 (5.9) 1.0 0.058 Sweating 10 (13.7) 1.8 2 (2.9) 1.5 0.059 Asthenia 9 (12.3) 1.9 5 (7.4) 1.6 0.103 Headache 9 (12.3) 1.2 3 (4.4) 1.3 0.132

BTDS = buprenorphine transdermal system.*Rated on a 3-point scale (1 = mild, 2 = moderate, 3 = severe).

May 2010 855

A. Gordon et al.

which would be expected to reduce the magnitude of difference in pain intensity between treatment phases, did not differ significantly between treatment BTDS and placebo. However, as has been reported previously, the use of an active rescue analgesic does not preclude demonstration of significant differences between active treatment and placebo, whether the rescue analgesic is an opioid7,25,39 or a nonopioid medication.12,13,45 A sig-nificant majority of patients and investigators expressed a blinded preference for the BTDS treatment phase, suggesting that the difference in pain scores between BTDS and placebo was clinically meaningful to patients and investigators. In addition, more patients and inves-tigators rated BTDS as more effective than placebo, and approximately two thirds of patients rated BTDS mod-erately or highly effective. BTDS was associated with an improvement in the overall Pain and Sleep Question-naire score compared with placebo, as well as improve-ments from baseline on 7 of 8 questionnaire items. Sleep disturbance is a common complaint among patients with chronic pain, and improvement in sleep quality is an important goal of pain management.46 All assessments were scheduled for day 7 (range, 6–8 days) to evaluate efficacy at the end of the dosing period.

The results of the 6-month, open-label extension phase suggest that BTDS was effective in the long-term management of chronic low back pain. Although this was an enriched population, >80% of eligible patients chose to continue receiving BTDS in the open-label phase. The reductions in pain intensity and improvements in sleep achieved during the double-blind phase were sus-tained for up to 6 months in the open-label phase. The mean final dose of BTDS in the open-label phase was numerically lower than the mean dose during the double-blind phase, suggesting that pain control was maintained without development of analgesic tolerance. Furthermore, 84.7% of patients reported receiving at least moderate benefit from BTDS during the open-label phase.

Although there were no serious adverse events during the double-blind phase of the study, patients receiving BTDS had significantly more adverse events compared with those receiving placebo (P = 0.003). Adverse events such as nausea, dizziness, vomiting, somnolence, and dry mouth, which occurred more frequently with BTDS than with placebo, are well-documented adverse effects of opioids. The most frequent adverse events with BTDS were nausea (53.4%) and dizziness (32.9%), both of which decreased substantially (37.5% and 10.0%, re-spectively) during long-term, open-label treatment.

6 months of open-label treatment; the mean dose of BTDS during the last week of double-blind treatment was 28.5 (12.1) µg/h.

The significant improvements in measures of pain intensity (VAS and ordinal scale), functionality (PDI, QBPDS), sleep (Pain and Sleep Questionnaire), and quality of life (SF-36) achieved during the double-blind phase were maintained throughout the open-label phase (Table III). At the end of the open-label phase, the mean clinical benefit was 1.8 (0.8), compared with 2.2 (1.0) at the end of the double-blind phase (P = NS). At the last open-label visit, 18 of 39 patients (46.2%) reported receiving moderate benefit from BTDS and 15 of 39 (38.5%) reported receiving a great deal of benefit from BTDS.

During the open-label extension, 40 patients reported 204 adverse events (mean maximum severity, 1.9). The most frequently reported adverse event was nausea in 15 patients (37.5%), with a mean maximum severity of 1.6. Pruritus and rash were each reported by 12 pa-tients (30.0%), with a mean maximum severity of 1.5 and 1.3, respectively. Somnolence was reported by 8 pa-tients (20.0%), with a mean maximum severity of 1.0. Constipation and asthenia were each reported by 5 pa-tients (12.5%), with a mean maximum severity of 2.2 and 1.2, respectively. Headache and dizziness were each reported by 4 patients (10.0%), with a mean maximum severity of 2.0 and 1.3, respectively. Four patients ex-perienced serious adverse events during or shortly after the open-label phase: 1 case of myocardial infarction and 1 case of pneumonia were judged unrelated to study drug, and 2 cases of meningioma (one of them a recur-rence) were judged unlikely to be related to study drug.

DISCUSSIONIn this randomized, double-blind crossover study of BTDS initiated at a dose of 10 µg/h and titrated to effect (maximum dose, 40 µg/h), BTDS was more effective than placebo in adults with chronic low back pain who had previously received opioids. This study extends the results of a previous study in which BTDS was initiated at 5 µg/h and titrated to a maximum of 20 µg/h.25

The BTDS dose in the present study ranged from 10 to 30 µg/h in 45% of patients, with the remainder titrated to the upper limit allowed by the protocol.

In the PP analysis, BTDS was associated with lower scores compared with placebo on the primary measures of pain intensity; these results were supported by the ITT analysis. Consumption of rescue acetaminophen,

856 Volume 32 Number 5

Clinical Therapeutics

withdrawal from BTDS due to adverse events were 20.5% (16/78) during the double-blind phase and 17.5% (7/40) during the open-label phase. Similar rates of withdrawal due to adverse events have been reported in previous studies of opioids, including a study of controlled-release tramadol in chronic noncancer pain (19.7%)28 and a study of controlled-release oxycodone in chronic low back pain (15.4%).39 Adverse events as-sociated with the transdermal delivery system (eg, pruritus, rash) did not differ significantly between BTDS and placebo. Only 3 patients (2 BTDS, 1 placebo) dis-continued the study due to a skin reaction during the double-blind phase, and only 1 discontinued due to a skin reaction during the open-label phase (data not shown). Most skin reactions during receipt of BTDS were mild to moderate in severity; only 1 skin reaction during receipt of BTDS was reported as severe. The number of patients with skin reactions at the previous patch site was lower than the number of patients with

Although patients receiving BTDS had a numerical-ly higher frequency of constipation than did those re-ceiving placebo (16.4% and 5.9%, respectively), the difference was not statistically significant. In a previous crossover study of BTDS in chronic low back pain, patients in both study arms were provided with acetaminophen/codeine for rescue analgesia.25 In that study, BTDS did not augment the rate of constipation elicited by the opioid rescue medication compared with placebo. Combined with the results of the present study, this suggests a favorable adverse-effect profile for BTDS with respect to constipation, a common opioid-related adverse effect to which patients do not normally develop tolerance.22

The overall rate of withdrawals in the double-blind portion of the present study was 37.2%, with more patients withdrawing while receiving BTDS compared with placebo (19 vs 10, respectively). The most common reason for withdrawal was adverse events. Rates of

Table III. Clinical effects of buprenorphine transdermal system during the 6-month, open-label extension phase. Values are mean (SD).

Assessment

End of Double-Blind

Treatment

Last Visit in Open-Label

Phase P

Pain intensity in past wk VAS* 43.4 (23.3) 42.1 (25.9) 0.639 Ordinal scale† 1.9 (0.7) 1.9 (0.9) 1.000

Pain intensity in past 24 h VAS* 41.5 (21.7) 40.6 (27.4) 0.765 Ordinal scale† 1.8 (0.8) 1.8 (1.0) 0.895

Overall Pain and Sleep Questionnaire score‡ 169.2 (121.0) 184.4 (129.5) 0.424

Overall Pain and Disability Index§ 30.1 (13.8) 29.5 (16.1) 0.838

Total Quebec Back Pain Disability Scale∥ 47.4 (20.9) 48.9 (21.4) 0.656

SF-36 Physical Component Summary score¶ 30.5 (9.4) 29.9 (10.0) 0.742

SF-36 Mental Health Component Summary score¶ 45.5 (14.4) 43.4 (16.1) 0.182

VAS = visual analog scale; SF-36 = 36-item Short Form Health Survey. * 100-mm VAS (from no pain to excruciating pain). † 5-Point ordinal scale (0 = none, 1 = mild, 2 = moderate, 3 = severe, 4 = excruciating). ‡ Sum of 5 items, each rated on a 100-mm VAS (from never to always). § Sum of scores on 7 subscales, each rated on an 11-point ordinal scale (from no disability to total disability). ∥ Sum of 20 items rated on a 6-point categorical scale (0 = no diff iculty, 1 = minimally diff icult, 2 = somewhat diff icult,

3 = fairly diff icult, 4 = very diff icult, 5 = unable to do). ¶ The SF-36 Physical Component Summary Scale is derived from the physical functioning, role limitations–physical,

bodily pain, and general health items, and the SF-36 Mental Component Summary Scale is derived from the vitality, social functioning, role limitations–emotional, and mental health items. Higher scores indicate better functioning.

May 2010 857

A. Gordon et al.

with opioids.25 In the present study, BTDS was suitable for titration to higher doses in patients already receiving opioids.

This study was adequately powered to detect a sig-nificant difference between BTDS and placebo. Although the study was limited by a small number of patients relative to parallel studies, the sample size was consistent with power calculations for crossover study designs.43

In terms of the duration of exposure to double-blind treatment, the randomized portion of the study had a duration of 4 weeks, while the 6-month extension phase provided a longer period of open-label treatment in which to monitor changes in dose, tolerability, and pain scores over time. The open-label population consisted of patients who chose to continue treatment after a satisfactory experience with BTDS; therefore, they were, by definition, an enriched subgroup. Nonetheless, the outcomes during the open-label phase were consistent with those during the double-blind phase. Although adverse events were generally spontaneously reported or observed, 2 common opioid-related adverse events (nausea and drowsiness) were assessed at predetermined intervals to strengthen the comparison between active treatment and placebo. Finally, an effort was made to keep the enrollment criteria as broad as possible while remaining within clinical practice norms for the use of opioids in suitable patients with chronic low back pain. Enrolled subjects met these criteria, although a skewed distribution of subjects by white race emerged, repre-senting the catchment populations of the investigation-al sites.

CONCLUSIONSIn the randomized, double-blind crossover phase of this study, pain control was significantly better during receipt of BTDS compared with placebo in these adult patients with moderate to severe chronic low back pain who had previously been treated with opioids. Patients’ and investigators’ ratings of treatment effectiveness and preference suggested a meaningful therapeutic benefit with BTDS. Although more adverse events were reported with BTDS than with placebo, 81.6% of eligible patients chose to continue into the open-label phase. The improve-ments achieved during the double-blind phase were maintained for up to 6 months in the open-label phase, with no evidence of analgesic tolerance. Thus, BTDS provided effective analgesia with an acceptable tolera-bility profile when initiated at 10 µg/h and titrated upward to a maximum of 40 µg/h.

skin reactions at the current patch site (9 vs 15, respec-tively), suggesting that skin reactions after patch removal were of short duration. Varying the site of administra-tion would be expected to minimize skin irritation in any particular area.

Buprenorphine is unique in having mixed κ-opioid–receptor antagonist and partial µ-opioid–receptor ago-nist activity; therefore, an assessment for possible opioid withdrawal effects was included in the study. There were no differences between BTDS and placebo with respect to modified SOWS scores, either during the treatment phases or on discontinuation of buprenorphine (before crossover to placebo in the double-blind phase). In addition, there were no investigator reports of opioid withdrawal at any time during the study. This is con-sistent with the findings of another clinical trial of BTDS, in which discontinuation of BTDS was not associated with symptoms related to withdrawal.24

In view of evidence suggesting a potential for opioids to prolong the QT interval,27 the present study exam-ined the cardiac safety profile of BTDS in a subset of patients receiving doses of 10, 20 and 40 µg/h. QTc interval changes >5 milliseconds, with an absolute QTc interval of >500 milliseconds, have been associated with an increased risk of cardiac arrhythmias.44 The present analysis provided no evidence of a clinically meaningful prolongation of the QTc interval with BTDS doses of up to 40 µg/h. In 2 studies in opioid-dependent patients, methadone was associated with a dose-dependent prolongation of the QT interval, where-as buprenorphine was not,47,48 despite administration of doses several-fold higher than those of BTDS in the present study.

Fentanyl is the only other opioid currently available in a transdermal formulation, although this formulation requires dosing every 3 days.49 Both BTDS and trans-dermal fentanyl are indicated for use in patients who require continuous opioid analgesia for an extended period of time. However, transdermal fentanyl is indi-cated for the treatment of persistent moderate to severe chronic pain that is not adequately managed with short-acting or weak opioids or combination products, and it is indicated only in patients who are already receiv-ing opioid therapy at a total daily dose ≥60 mg of morphine (or analgesic equivalent).49 In contrast, BTDS is indicated for the management of persistent moderate pain, with no restriction as to previous opioid use.22

An earlier study found BTDS to be effective and well tolerated in patients who had not previously been treated

858 Volume 32 Number 5

Clinical Therapeutics

9. Caldwell JR, Hale ME, Boyd RE, et al. Treatment of os-teoarthritis pain with controlled release oxycodone or fixed combination oxycodone plus acetaminophen added to nonsteroidal antiinflammatory drugs: A double blind, randomized, multicenter, placebo controlled trial. J Rheu-matol. 1999;26:862–869.

10. Peloso PM, Bellamy N, Benson W, et al. Double blind randomized placebo control trial of controlled release co-deine in the treatment of osteoarthritis of the hip or knee. J Rheumatol. 2000;27:764–771.

11. Roth SH, Fleischmann RM, Burch FX, et al. Around- the-clock, controlled-release oxycodone therapy for osteoarthritis-related pain: Placebo-controlled trial and long-term evaluation. Arch Intern Med. 2000;160:853–860.

12. Watson CP, Babul N. Efficacy of oxycodone in neuro-pathic pain: A randomized trial in postherpetic neuralgia. Neurology. 1998;50:1837–1841.

13. Watson CP, Moulin DE, Watt-Watson J, et al. Controlled release oxycodone relieves neuropathic pain: A randomized controlled trial in painful diabetic neuropathy. Pain. 2003; 105:71–78.

14. Gimbel JS, Richards P, Portenoy RK. Controlled-release oxycodone for pain in diabetic neuropathy: A randomized controlled trial. Neurology. 2003;60:927–934.

15. Chou R, Fanciullo GJ, Fine PG, et al, for the American Pain Society–American Academy of Pain Medicine Opioids Guidelines Panel. Clinical guidelines for the use of chronic opioid therapy in chronic noncancer pain. J Pain. 2009;10: 113–130.

16. Gourlay DL, Heit HA, Almahrezi A. Universal precautions in pain medicine: A rational approach to the treatment of chronic pain. Pain Med. 2005;6:107–112.

17. Lewis JW. Clinical pharmacology of buprenorphine in rela-tion to its use as an analgesic. In: Cowan A, Lewis JW, eds. Buprenorphine: Combating Drug Abuse With a Unique Opioid. New York, NY: Wiley-Liss Inc; 1995:151–163.

18. Russo MA, Wasiak J. A clinical snapshot of transdermal buprenorphine in pain management. Eur J Pain Suppl. 2007;1:74–77.

19. Pergolizzi J, Böger RH, Budd K, et al. Opioids and the management of chronic severe pain in the elderly: Con-sensus statement of an International Expert Panel with focus on the six clinically most often used World Health Organization Step III opioids (buprenorphine, fentanyl, hydromorphone, methadone, morphine, oxycodone). Pain Pract. 2008;8:287–313.

20. Cowan A, Friderichs E, Strassburger W, Raffa RB. Basic pharmacology of buprenorphine. In: Budd K, Raffa RB, eds. Buprenorphine—the Unique Opioid Analgesic. Pharmacol-ogy and Clinical Application. New York, NY: Thieme Medical Publishing; 2005:3–21.

21. Sittl R. Transdermal buprenorphine in clinical practice. In: Budd K, Raffa RB, eds. Buprenorphine—the Unique Opioid

ACkNOWLEDGMENTSThis study was funded by Purdue Pharma, Pickering, Ontario, Canada. The study design was finalized with input from Drs. Gordon, Callaghan, Spink, Cloutier, Dzongowski, O’Mahony, Sinclair, Rashiq, Buckley, Cohen, Kim, and Boulanger, each of whom received compensation for their participation. Drs. Piraino, Eisenhoffer, Darke, and Michalko, and Mr. Harsanyi are employees of Purdue Pharma or were employed by the company at the time the study was conducted. All authors were involved in critical review and revision of the manuscript, and agreed to its content and conclu-sions. The authors have indicated that they have no other conflicts of interest with regard to the content of this article.

The authors acknowledge the late John T. Sibley, MD, of the Royal University Hospital, Saskatoon, Sas-katchewan, Canada, for his intellectual contributions and for recruitment of patients for this study.

REFERENCES1. Statistics Canada. Musculoskeletal diseases: Back pain.

http://www.statcan.gc.ca/pub/82-619-m/2006003/ 4053542-eng.htm. Accessed December 9, 2008.

2. Woolf AD, Pfleger B. Burden of major musculoskeletal conditions. Bull World Health Organ. 2003;81:646–656.

3. Asche CV, Kirkness CS, McAdam-Marx C, Fritz JM. The societal costs of low back pain: Data published between 2001 and 2007. J Pain Palliat Care Pharmacother. 2007;21: 25–33.

4. Dagenais S, Caro J, Haldeman DC. A systematic review of low back pain cost of illness studies in the United States and internationally. Spine J. 2008;8:8–20.

5. Chou R, Qaseem A, Snow V, et al, for the Clinical Efficacy Assessment Subcommittee of the American College of Phy-sicians and the American College of Physicians/American Pain Society Low Back Pain Guidelines Panel. Diagnosis and treatment of low back pain: A joint clinical practice guideline from the American College of Physicians and the American Pain Society [published correction appears in Ann Intern Med. 2008;148:247–248]. Ann Intern Med. 2007; 147:478–491.

6. World Health Organization. Cancer Pain Relief, With a Guide to Opioid Availability. 2nd ed. Geneva, Switzerland: World Health Organization; 1996.

7. Arkinstall W, Sandler A, Goughnour B, et al. Efficacy of controlled-release codeine in chronic non-malignant pain: A randomized, placebo-controlled clinical trial. Pain. 1995; 62:169–178.

8. Moulin DE, Iezzi A, Amireh R. Randomised trial of oral morphine for chronic non-cancer pain. Lancet. 1996;347: 143–147.

May 2010 859

A. Gordon et al.

41. Elkader AK, Brands B, Callaghan R, Sproule BA. Exploring the relation-ship between perceived inter-dose opioid withdrawal and patient char-acteristics in methadone mainte-nance treatment. Drug Alcohol De-pend. 2009;105:209–214.

42. European Agency for the Evaluation of Medicinal Products, Committee for Proprietary Medicinal Products. Points to consider on multiplicity issues in clinical trials. http://www.emea.europa.eu/pdfs/human/ewp/ 090899en.pdf. Accessed January 8, 2010.

43. Jones B, Kenward MG. Design and Analysis of Cross-Over Trials. 2nd ed. Boca Raton, Fla: Chapman & Hall/CRC; 2003.

44. International Conference on Harmon-isation of Technical Requirements for Registration of Pharmaceuticals for Human Use. ICH Harmonised Tripar-tate Guideline. The clinical evalua- tion of QT/QTc interval prolongation and proarrhythmic potential for non- antiarrhythmic drugs: E14. http://www.ich.org/cache/compo/475-272-1.html#E14. Accessed March 23, 2010

45. Thorne C, Beaulieu AD, Callaghan DJ, et al. A randomized, double-blind, crossover comparison of the efficacy and safety of oral controlled-release tramadol and placebo in pa-tients with painful osteoarthritis. Pain Res Manag. 2008:13:93–102.

46. Morin CM, Gibson D, Wade J. Self-reported sleep and mood distur-bance in chronic pain patients. Clin J Pain. 1998;14:311–314.

47. Fanoe S, Hvidt C, Ege P, Jensen GB. Syncope and QT prolongation among patients treated with metha- done for heroin dependence in the city of Copenhagen. Heart. 2007;93: 1051–1055.

48. Wedam EF, Bigelow GE, Johnson RE, et al. QT-interval effects of metha-done, levomethadyl, and buprenor-phine in a randomized trial. Arch In-tern Med. 2007;167:2469–2475.

30. Hagen NA, Thirlwell M, Eisenhoffer J, et al. Efficacy, safety, and steady-state pharmacokinetics of once-a-day controlled-release morphine (MS Contin XL) in cancer pain. J Pain Symp-tom Manage. 2005;29:80–90.

31. Snyder-Halpern R, Verran JA. In-strumentation to describe subjec-tive sleep characteristics in healthy subjects. Res Nurs Health. 1987;10: 155–163.

32. Kosinski M, Janagap C, Gajria K, et al. Pain relief and pain-related sleep disturbance with extended-release tramadol in patients with osteoar-thritis. Curr Med Res Opin. 2007;23: 1615–1626.

33. Miaskowski C. Pharmacologic man-agement of sleep disturbances in noncancer-related pain. Pain Manag Nurs. 2009;10:3–13.

34. Pollard CA. Preliminary validity study of the pain disability index. Percept Mot Skills. 1984;59:974.

35. Tait RC, Chibnall JT, Krause S. The Pain Disability Index: Psychometric properties. Pain. 1990;40:171–182.

36. Kopec JA, Esdaile JM, Abrahamo- wicz M, et al. The Quebec Back Pain Disability Scale. Measurement prop-erties. Spine (Phila Pa 1976). 1995; 20:341–352.

37. Brazier JE, Harper R, Jones NM, et al. Validating the SF-36 health survey questionnaire: New outcome mea-sure for primary care. BMJ. 1992; 305:160–164.

38. Bronfort G, Bouter LM. Respon-siveness of general health status in chronic low back pain: A compari-son of the COOP Charts and the SF-36. Pain. 1999;83:201–209.

39. Sibley J, Kelly A, Rashiq S, et al. Controlled release oxycodone and acetaminophen plus codeine in chronic low back pain. Presented at: 10th World Congress on Pain; August 17–22, 2002; San Diego, Calif. Abstract 429.

40. Handelsman L, Cochrane KJ, Aron-son MJ, et al. Two new rating scales for opiate withdrawal. Am J Drug Alcohol Abuse. 1987;13:293–308.

Analgesic. Pharmacology and Clinical Ap-plication. New York, NY: Thieme Medi-cal Publishing; 2005:92–101.

22. BuTrans (buprenorphine transdermal system) product monograph. Picker-ing, Ont: Purdue Pharma; 2010.

23. Karlsson M, Berggren AC. Efficacy and safety of low-dose transdermal buprenorphine patches (5, 10, and 20 microg/h) versus prolonged- release tramadol tablets (75, 100, 150, and 200 mg) in patients with chronic osteoarthritis pain: A 12-week, randomized, open-label, controlled, parallel-group noninferiority study. Clin Ther. 2009;31:503–513.

24. Landau CJ, Carr WD, Razzetti AJ, et al. Buprenorphine transdermal de-livery system in adults with persistent noncancer-related pain syndromes who require opioid therapy: A multi- center, 5-week run-in and random-ized, double-blind maintenance-of-analgesia study [published correction appears in Clin Ther. 2009;31:677]. Clin Ther. 2007;29:2179–2193.

25. Gordon A, Rashiq S, Moulin D, et al. Buprenorphine transdermal system for opioid therapy in patients with chronic low back pain. Pain Res Manag. In press.

26. Dworkin RH, Turk DC, Farrar JT, et al, for IMMPACT. Core outcome measures for chronic pain clinical trials: IMMPACT recommendations. Pain. 2005;113:9–19.

27. Katchman AN, McGroary KA, Kil-born MJ, et al. Influence of opioid agonists on cardiac human ether-a-go-go related gene K(+) currents. J Pharmacol Exp Ther. 2002;303:688– 694.

28. Beaulieu AD, Peloso P, Bensen W, et al. A randomized, double-blind, 8-week crossover study of once-daily controlled-release tramadol versus immediate-release tramadol taken as needed for chronic noncancer pain. Clin Ther. 2007;29:49–60.

29. Russell A, Watson CP, Clark AJ, et al. Evaluation of dosing guidelines for use of controlled-release codeine in chronic noncancer pain. Pain Res Manag. 2003;8:143–148.

860 Volume 32 Number 5

Clinical Therapeutics

49. Duragesic (fentanyl transdermal system) product monograph. To-ronto, Ont: Janssen-Ortho Inc; 2009.

Address correspondence to: John Eisenhoffer, MD, Purdue Pharma, 575 Granite Court, Pickering, ON, Canada, L1W 3W8. E-mail: John.Eisenhoffer@purdue.ca