Effect of Buprenorphine Implants on Illicit Opioid Use Among Abstinent Adults With Opioid Dependence Treated With Sublingual Buprenorphine: A Randomized Clinical Trial

Importance  The effectiveness of buprenorphine treatment of opioid dependence is limited by suboptimal medication adherence, abuse, and diversion.

Objective  To determine whether 6-month buprenorphine implants are noninferior to daily sublingual buprenorphine as maintenance treatment for opioid-dependent patients with stable abstinence.

Design, Setting, and Participants  Outpatient, randomized, active-controlled, 24-week, double-blind, double-dummy study conducted at 21 US sites from June 26, 2014, through May 18, 2015. Outpatients were prescribed daily sublingual buprenorphine for 6 months or more, were abstinent while taking 8 mg/d or less of sublingual buprenorphine for 90 days or longer, and were determined to be clinically stable by their physician.

Interventions  Participants were randomized to receive sublingual buprenorphine plus 4 placebo implants or sublingual placebo plus four 80-mg buprenorphine hydrochloride implants (expected efficacy, 24 weeks).

Main Outcome Measure  The primary end point was between-group difference in proportion of responders (≥4 of 6 months without opioid-positive urine test result [monthly and 4 times randomly] and self-report). The noninferiority established for the lower bound of the 95% confidence interval was greater than −0.20 (P < .025). Secondary end points included cumulative percentage of negative opioid urine results, abstinence, and time to first illicit opioid use. Safety was assessed by adverse event reporting.

Results  Of 177 participants (mean age, 39 years; 40.9% female), 90 were randomized to sublingual buprenorphine with placebo implants and 87 to buprenorphine implants with sublingual placebo; 165 of 177 (93.2%) completed the trial. Eighty-one of 84 (96.4%) receiving buprenorphine implants and 78 of 89 (87.6%) receiving sublingual buprenorphine were responders, an 8.8% difference (1-sided 97.5% CI, 0.009 to ∞; P < .001 for noninferiority). Over 6 months, 72 of 84 (85.7%) receiving buprenorphine implants and 64 of 89 (71.9%) receiving sublingual buprenorphine maintained opioid abstinence (hazard ratio, 13.8; 95% CI, 0.018-0.258; P = .03). Non–implant-related and implant-related adverse events occurred in 48.3% and 23% of the buprenorphine implant group and in 52.8% and 13.5% of participants in the sublingual buprenorphine group, respectively.

Conclusions and Relevance  Among adults with opioid dependence maintaining abstinence with a stable dose of sublingual buprenorphine, the use of buprenorphine implants compared with continued sublingual buprenorphine did not result in an inferior likelihood of remaining a responder. However, the study population had an exceptionally high response rate in the control group, and further studies are needed in broader populations to assess the efficacy in other settings.

Trial Registration  clinicaltrials.gov Identifier: NCT02180659

Quiz Ref IDOpioid dependence is a growing public health concern in the United States, associated with spread of human immunodeficiency virus and hepatitis C and fatal overdose when left untreated.^1-3 Medical treatment of opioid dependence decreases illicit opioid use at a greater rate compared with psychosocial intervention or placebo alone^4-7 and reduces morbidity, mortality,⁸ and spread of infectious diseases.⁹ Because only 3 available medications—methadone, buprenorphine, and naltrexone—are approved for treatment of opioid dependence,^5,10 innovation is needed.

Quiz Ref IDA 6-month buprenorphine implant may improve treatment adherence, improve rates of opioid abstinence, and reduce susceptibility to increasing public health concerns of abuse, diversion, and unintentional pediatric exposure.¹¹ The Centers for Disease Control and Prevention reported that 9.5% of emergency hospitalizations for drug ingestion among children younger than 6 years were caused by buprenorphine/naloxone, despite buprenorphine accounting for 2.2% of retail opioid prescriptions and 0.16% of all retail prescriptions.¹²

Opioid dependence is a chronic, relapsing disorder. Little research has investigated tactics to support long-term remission; long-acting implants may sustain long-term remission. In double-blind, phase 3, placebo-controlled trials among opioid-dependent, treatment-seeking participants newly inducted to 12 to 16 mg/d of sublingual buprenorphine and subsequently randomized, buprenorphine implants were superior to placebo implants in reducing illicit opioid use over the initial 6-month treatment period.^13,14 The purpose of this trial was to determine if 6-month subdermal buprenorphine implants maintained low to no illicit opioid use relative to daily sublingual buprenorphine among currently stable opioid-dependent patients receiving buprenorphine maintenance treatment.

Box Section Ref ID

This randomized, double-blind, double-dummy, active-controlled, 26-week, multisite study in adult clinically stable outpatients receiving 8 mg/d or less of sublingual buprenorphine evaluated the safety and efficacy of 4 simultaneously placed, 26 × 2.5-mm buprenorphine implants (buprenorphine hydrochloride, 80 mg each) compared with daily sublingual buprenorphine (Figure 1). The study was conducted at 21 US office-based buprenorphine outpatient treatment sites. Each site approved the protocol (Supplement) through a centralized or local institutional review board (IRB) and obtained written informed consent from each volunteer prior to participation. Participants received an IRB-approved stipend for attending study visits (average per visit = $40; total stipend range, $350-$725).

Eligible participants had a primary diagnosis of opioid dependence,¹⁵ were aged 18 to 65 years, received sublingual buprenorphine for at least 24 weeks as an outpatient at a stable dosage of 8 mg/d or less, and showed no evidence of opioid withdrawal or illicit opioid-positive urine samples for at least 90 days prior to study entry. Female participants of childbearing potential agreed to use contraception during the study.

Exclusion criteria included pregnancy, lactation, or planning pregnancy; lack of appropriate implant sites (recent scars, history of keloids); coagulopathy within 90 days; screening serum aspartate and alanine aminotransferase levels 3-fold higher than upper limits of normal; total bilirubin or creatinine levels 1.5-fold higher than upper limits of normal; clinically significant thrombocytopenia; use of strong cytochrome P450 3A4 inhibitors (azole antifungals, macrolide antibiotics, or protease inhibitors) or an anticoagulant; chronic pain requiring opioids; AIDS; significant medical problems potentially affecting volunteer safety if enrolled; primary diagnosis of substance dependence other than opioids or nicotine; or pending legal action or other factors/conditions that could adversely affect participant safety and adequate adherence.

Quiz Ref IDUsing a centralized computer system for blinded allocation, participants were randomized 1:1 within a block size of 4 without stratification to daily sublingual buprenorphine tablets (dosage same as prerandomization; Amneal Pharmaceuticals) with 4 placebo subdermal implants or daily sublingual placebo tablets with 4 active buprenorphine implants. Four implants were expected to yield plasma buprenorphine concentrations (0.5-1.0 ng/mL) at a range comparable with 8 mg/d or less of sublingual buprenorphine. Based on independent analysis of 1-year insurance data for a cohort exclusive of patients taking multiple opioid medications, approximately 25% of persons receiving buprenorphine maintenance therapy for at least 1 year receive sublingual buprenorphine, 8 mg/d or less (Suboxone equivalents). Placebo and sublingual buprenorphine tablets matched in appearance.

On randomization day, implant insertion was performed during an approximate 10- to 15-minute in-office procedure. Following local anesthetic injection, 4 subdermal implants were inserted one at a time through a single 2.5- to 3-mm incision in the inner upper arm using an applicator designed for the implants. The incision was closed with steri-strips and a pressure bandage was applied for 24 hours. To maintain blinding, staff involved in implant insertion and removal at 6 months did not participate in study evaluation as active and placebo implants varied slightly in appearance. Unblinded supplemental sublingual buprenorphine use (2-mg tablets; Actavis Elizabeth LLC) was allowed as needed during the study.

Participant assessments occurred at week 1, then at 4-week intervals. Urine samples were obtained for drug testing at scheduled monthly visits and 4 times randomly for a total of 10 urine samples. Study follow-up was conducted by telephone 1 week posttreatment; a follow-up visit was scheduled for 2 weeks posttreatment. All participants received mandatory manual-guided individual drug counseling during each monthly study visit.¹⁶ Participant-requested or physician-directed supplemental visits, telephone calls, or additional counseling; supplemental sublingual buprenorphine or other pharmacological interventions; and reasons for each were recorded. Medication adherence was captured through visual inspection and palpation of each subdermal implant and via pill counts at study visits.

Efficacy outcome analyses were based on the predefined intention-to-treat population consisting of all randomized participants who received study medication and provided any efficacy data. The primary efficacy end point was the difference in proportion of responders, defined as participants with at least 4 of 6 months without evidence of illicit opioid use (based on urine test and self-report composites) by treatment group.

Illicit drug use was measured by urine testing and self-report. Urine samples, adjusted for creatinine level, were quantitatively analyzed for opioids (codeine, morphine, hydrocodone, oxymorphone, hydromorphone, oxycodone, methadone, and fentanyl) using liquid chromatography–tandem mass spectrometry and qualitatively tested for other drugs of abuse (cocaine, barbiturates, amphetamines, phencyclidine, and tetrahydrocannabinol). Self-report of illicit opioid use and other drugs of abuse was recorded using a Timeline Followback interview¹⁷ on an illicit drug use self-report form. When both urine samples and self-report were negative for opioids within the reporting month, the assessment was considered opioid negative for evaluation of the primary efficacy end point; either an opioid-positive urine test result or self-reported opioid use qualified as an opioid-positive assessment.

Secondary efficacy analyses included treatment retention, time to first illicit opioid use, percentage of illicit opioid use by month, and cumulative percentage of negative illicit opioid urine results at 6 months. Opioid craving, withdrawal, and supplemental use of sublingual buprenorphine were also measured. Safety—based on adverse event reporting, including implant site reactions—was assessed in all participants receiving study medication.

Opioid craving was determined through self-reported “desire to use” and “need to use,” measured with unipolar 100-mm visual analog scales ranging from 0 (no desire or need to use) to 100 (strongest possible desire or need to use).¹⁸ Opioid withdrawal measures included the Subjective Opioid Withdrawal Scale (SOWS) and Clinical Opioid Withdrawal Scale (COWS).^19,20 Outcomes were completed at screening, day 1, and monthly.

For demographic data collection, participants designated their race/ethnicity based on investigator-defined options.

Analyses followed the statistical analysis plan (Supplement) developed in collaboration with and approved by the US Food and Drug Administration (FDA). The study tested noninferiority of buprenorphine implants relative to sublingual buprenorphine on the primary efficacy outcome, with noninferiority established for a lower bound of the 95% confidence interval greater than −0.20 within the intention-to-treat population (1-sided testing with a significance threshold of P < .025). Per FDA statistical guidance for noninferiority trials, once noninferiority criteria were met, a nonprespecified sequential χ² test was performed for superiority.²¹ All superiority analyses were 2-sided with a significance threshold of P < .05. A sample size of 90 participants per group (180 participants overall) was selected to achieve 87.3% power assuming a 75% responder rate for both treatment groups. Missing urine test data were imputed by randomly generated binary outcome (positive or negative for opioid use) using a 20% relative penalty against the buprenorphine implant group based on urine test results from each group. Sensitivity analyses were conducted, including analysis of the effect of imputing all missing urine samples as opioid positive with response defined as 6-month abstinence from illicit opioids and adjudication of randomized participants who did not provide any efficacy data as nonresponders.

For secondary efficacy analyses, variables based on percentage were analyzed by χ²; time to first evidence of illicit opioid use was analyzed by log-rank test with treatment effects. Changes from baseline in craving scale, SOWS, and COWS scores were derived by subtracting baseline values from postbaseline values; negative changes indicated improvement. Changes from baseline were analyzed using analysis of covariance with treatment and baseline effects and via mixed-model repeated measures as a sensitivity analysis. Adverse events were tabulated by treatment group among all participants in the safety population. Commercially available software (SAS version 9.2; SAS Institute Inc) was used for all statistical analyses.

Enrollment began in June 2014 and the study ended May 2015; 177 patients were randomized. One participant randomized to sublingual buprenorphine did not receive study medication; 3 participants randomized to buprenorphine implants provided no study outcome data. The predefined intention-to-treat population included 173 of 177 participants (n = 89 sublingual buprenorphine; n = 84 buprenorphine implants); 176 of 177 participants were included in the safety population (Figure 1). In total, 84 of 90 participants receiving sublingual buprenorphine and 81 of 87 with buprenorphine implants completed the study. Table 1 shows participant demographics and clinical characteristics.

In the buprenorphine implant and sublingual buprenorphine groups, 81 of 84 participants (96.4%) and 78 of 89 participants (87.6%), respectively, were responders. The difference was 8.8% (1-sided 97.5% CI, 0.009 to ∞; P < .001 for noninferiority; P = .03 for superiority) on the primary outcome measure, with a calculated number needed to treat of 11.36 vs sublingual buprenorphine (Table 2).

In a sensitivity analysis for all randomized participants, with all missing urine samples imputed as positive for opioids and no illicit opioid use for all 6 months, 70 of 87 participants (80.5%) receiving buprenorphine implants and 60 of 90 (66.7%) receiving sublingual buprenorphine were abstinent, resulting in a proportion difference of 13.8% (1-sided 97.5% CI, 0.010 to ∞; P < .001 for noninferiority). Additional sensitivity analyses were consistent with primary efficacy results (Table 2).

Relative to sublingual buprenorphine, a larger proportion of participants receiving buprenorphine implants demonstrated no evidence of illicit opioid use throughout 6 months of treatment (Figure 2A). At 6 months, cumulative abstinence was 72 of 84 (85.7%) for buprenorphine implants vs 64 of 89 (71.9%) for sublingual buprenorphine (hazard ratio, 13.8; 95% CI, 0.018-0.258; P = .03), with a number needed to treat of 7.25. This statistically significant difference became apparent starting at month 3 and was sustained throughout month 6. Time to first evidence of illicit opioid use was significantly longer for buprenorphine implants relative to sublingual buprenorphine (hazard ratio, 0.49; 95% CI, 0.25-.97; P = .04) (Figure 2B).

There was no significant difference in desire-to-use scores for participants receiving buprenorphine implants or sublingual buprenorphine at baseline (Table 1) or at end of treatment (P = .83) (Table 2). The need-to-use scores were similar at baseline (Table 1) between buprenorphine implant and sublingual buprenorphine groups and at end of treatment (P = .51) (Table 2).

Baseline COWS scores were low for participants receiving buprenorphine implants and sublingual buprenorphine (Table 1), with no statistically significant changes from baseline to end of treatment (P = .92) (Table 2). Baseline SOWS scores were low for participants receiving buprenorphine implants and sublingual buprenorphine (Table 1), with no changes from baseline to end of treatment (P = .43) (Table 2).

Supplemental sublingual buprenorphine was permitted as needed throughout the study as determined by the treating physician. Fifteen participants with buprenorphine implants (17.9%) and 13 given sublingual buprenorphine (14.6%) received supplemental medication (P > .05) (Table 2). In both groups, the majority of participants receiving supplemental medication were prescribed small dosages (2 mg/d) on 4 or fewer occasions within the 6-month period.

Serious adverse events occurred in 5 participants, 3 in the sublingual buprenorphine group (biliary colic, chronic cholecystitis, bronchitis) and 2 in the buprenorphine implant group (convulsions, worsening bipolar I disorder). One buprenorphine implant participant discontinued because of an adverse event (muscle spasms).

Medication-related and implant-related adverse events were consistent with the known safety profile of buprenorphine²² and the previously published implantation procedure¹⁴ (Table 3). Buprenorphine implants were associated with higher incidences of local site adverse events vs sublingual buprenorphine. In addition to implant-related or non–implant-related adverse events, 1 accidental pediatric tablet exposure occurred in the sublingual buprenorphine group, resulting in overnight hospitalization and notification of local child protective services. Two participants receiving sublingual buprenorphine were admitted to inpatient rehabilitation. In each group, 1 participant was jailed for non-drug-related charges and 1 medication theft was reported.

There were no cases of migration of implants beyond the local insertion site. Ultrasound was used to locate nonpalpable implants for removal in 9 cases, 5 of which were placebo.

Quiz Ref IDBecause there are limited data on patients stable on sublingual buprenorphine,²³ it is important to study maintenance and improvement of stability in patients who achieve good clinical response to initial buprenorphine treatment. In this study, buprenorphine implants were noninferior to sublingual buprenorphine for the primary end point of at least 4 of 6 months with no illicit opioid use by drug testing and self-report among opioid-dependent adults maintaining clinical stability with sublingual buprenorphine. Additionally, relative to participants using sublingual buprenorphine, participants receiving buprenorphine implants were more likely to sustain abstinence from illicit opioids from month 3 through month 6. At a 6-month cumulative end point, 85.7% of the buprenorphine implant group was abstinent relative to 71.9% of the sublingual buprenorphine group, providing a number needed to treat of 7.25. To our knowledge, this was the first comparative trial to evaluate efficacy and safety of 6-month buprenorphine implants relative to sublingual buprenorphine in this understudied population of patients clinically stable taking sublingual buprenorphine.

Buprenorphine is an effective treatment for opioid dependence^4,7,8,10; however, adherence to daily dosing for management of chronic disorders is challenging.²⁴ An implantable buprenorphine delivery system reduces adherence issues and may improve efficacy. Furthermore, buprenorphine implants may reduce the need for sublingual buprenorphine, decreasing its availability for diversion, misuse,²³ and harms such as accidental pediatric exposure.²⁵ This study demonstrated the efficacy of buprenorphine implants in treatment of patients maintaining clinical stability with sublingual buprenorphine in comparison with an active control; preliminary evidence of efficacy relative to placebo was previously demonstrated by reducing opioid use in patients initiating buprenorphine treatment in double-blind, phase 3, placebo-controlled trials.^13,14

Although very few participants experienced relapses to opioid use during the 24 weeks of the study, statistically significant differences in time to first illicit opioid use between the buprenorphine implant and sublingual buprenorphine groups demonstrated sustained abstinence (Figure 2A). Despite a conservative method applying a 20% relative penalty for imputing missing urine values to the buprenorphine implant group, a 14% difference in illicit opioid-free urine samples was present at 6 months. These results were maintained in further sensitivity analyses (Table 2). These data support long-term maintenance of clinical stability provided by buprenorphine implants relative to the current standard, sublingual buprenorphine.

Transition from sublingual buprenorphine to the implant was clinically nondisruptive, obviating concerns regarding increases in opioid withdrawal, craving, need, or desire to use. The majority of study participants did not require supplemental sublingual buprenorphine, although more participants with buprenorphine implants received supplemental medication relative to the sublingual buprenorphine group. In prior studies, supplemental buprenorphine use was higher among patients new to treatment and receiving buprenorphine implants.¹⁴ This is likely due to differences in the sample enrolled and is also reflected in retention rates; the current study retained 93.1% of participants compared with a lower retention in buprenorphine treatment (64%-66%) in studies enrolling participants new to treatment.^13,14

A limitation of this study was that it was not powered to detect differences in adverse outcomes. Additionally, generalizability of the study findings may be limited because the majority of the participants enrolled were white, employed, with at least a high school education, and dependent on prescription opioids. The population enrolled herein may demonstrate a demographic shift in opioid use disorder and differs from the traditional population in clinical studies of participants new to treatment, such as those recruited in prior studies of buprenorphine implants.^13,14 This study included clinically stable participants who maintained abstinence for 90 days prior to randomization and were in buprenorphine treatment for an average of 3.5 years. These factors may account for the higher than anticipated response rate in the control group (88% vs 75% in power analysis).

Quiz Ref IDA question for future study includes determining the rate and predictors of relapse after implant discontinuation. Many patients wish to discontinue sublingual buprenorphine following a few months of maintenance, perhaps because of unrealistic expectations,²⁶ subjecting them to significant risk of relapse.²⁷ Maintenance dosages of sublingual buprenorphine appear to be inversely associated with risk of opioid relapse after cessation.²⁸ Thus, it is important to study the feasibility of longer-term patient treatment with 6-month implant replacement. Additionally, larger trials could address infrequent but clinically and societally important public health adverse events, including accidental pediatric exposure, and more common events associated with buprenorphine, such as diversion.

Among adults with opioid dependence maintaining abstinence with a stable dose of sublingual buprenorphine, the use of buprenorphine implants compared with continued sublingual buprenorphine did not result in an inferior likelihood of remaining a responder. However, the study population had an exceptionally high response rate in the control group, and further studies are needed in broader populations to assess the efficacy of buprenorphine implants vs sublingual buprenorphine in other settings.

Corresponding Author: Richard N. Rosenthal, MD, Mount Sinai West, 1000 10th Ave, 8C-02, New York, NY 10019 (rrosenthal@chpnet.org).

Correction: This article was corrected on August 11, 2016, for errors in Figure 2.

Author Contributions: Dr Rosenthal had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Beebe.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Rosenthal, Lofwall, Kim, Beebe.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Chen.

Administrative, technical, or material support: Beebe.

Study supervision: Kim.

Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Rosenthal reports grants and nonfinancial support from Braeburn Pharmaceuticals during the conduct of the study. Dr Lofwall reports research contract funding from Braeburn Pharmaceuticals during the conduct of the study; consulting fees from Braeburn, CVS Caremark, and Orexo Pharmaceuticals; and honoraria for developing and giving continuing medical education talks for PCM Scientific outside the submitted work. Dr Kim is an employee of Braeburn Pharmaceuticals. Dr Chen reports personal fees from Braeburn Pharmaceuticals during the conduct of the study and outside the submitted work. Dr Beebe is an employee of Titan Pharmaceuticals. Dr Vocci reports personal fees and other from Braeburn Pharmaceuticals during the conduct of the study; personal fees and other from Pinney Associates Inc; personal fees and other from Indivior; and personal fees from Demerx outside the submitted work. No other disclosures were reported.

Group Information: The PRO-814 Study Group investigators include Valentin Isacescu, MD (Oceanside, California), Paul W. Schkolnik, PhD (Columbus, Ohio), Richard N. Rosenthal, MD (New York, New York), Mohammad Munir, MD (New Bedford, Massachusetts), John Bernard, MD (Belvidere, New Jersey), Azfar Malik, MD (St Louis, Missouri), Amit Vijapura, MD (Jacksonville, Florida), Edwardo Cifuentes, MD (Charleston, South Carolina), Kyle Kampman, MD (Philadelphia, Pennsylvania), James Sullivan, MD (Birmingham, Alabama), Scott Segal, MD (Miami, Florida), Walter Ling, MD (Los Angeles, California), Stacey Sigmon, PhD (Burlington, Vermont), Rishi Kakar, MD (Lauderhill, Florida), Genie Bailey, MD (Fall River, Massachusetts), Jelena Kunovac, MD (Las Vegas, Nevada), Michael Frost, MD (Conshohocken, Pennsylvania), Kent Hoffman, DO (Maitland, Florida), Boyde Harrison, MD (Haleyville, Alabama), Michelle Lofwall, MD (Lexington, Kentucky), Brent Boyett, DO (Hamilton, Alabama).

Funding/Support: This study was funded by Braeburn Pharmaceuticals.

Role of the Funders/Sponsors: Braeburn Pharmaceuticals was responsible for the design and conduct of the study. Medical monitoring was conducted by the sponsor and PPD Inc; clinical monitoring, serious adverse event management, and data management were conducted by PPD. Urine toxicology was measured by NMS Laboratories. Braeburn Pharmaceuticals personnel were involved in the collection, management, analysis, and interpretation of the data and the preparation and review of the manuscript. Final approval of the content of the manuscript and the decision to submit the manuscript for publication was solely the responsibility of the corresponding author (R.N.R.) and the other authors who have provided certification as such.

Additional Contributions: Editorial support was provided by Terri Schochet, PhD, AlphaBioCom LLC, and funded by Braeburn Pharmaceuticals.