Psychopharmacology (2002) 160:344–352 DOI 10.1007/s00213-001-0975-0
O R I G I N A L I N V E S T I G AT I O N
Mark K. Greenwald · Kory J. Schuh John A. Hopper · Charles R. Schuster Chris-Ellyn Johanson
Effects of buprenorphine sublingual tablet maintenance on opioid drug-seeking behavior by humans Received: 15 May 2001 / Accepted: 6 November 2001 / Published online: 31 January 2002 © Springer-Verlag 2002
Abstract Rationale: Buprenorphine can decrease opioid self-administration by humans and animals, but its ability to decrease drug-seeking behavior and craving (i.e. motivational measures) among outpatient volunteers using clinically relevant dosing schedules has not been extensively studied. Objectives: We investigated whether daily versus alternating-day administration of high versus low buprenorphine doses influenced choice of, and operant responding for, hydromorphone versus money. Methods: Fourteen heroin-dependent outpatients were maintained under four buprenorphine sublingual tablet (double blind) dose conditions using a within-subject, randomized crossover design. All participants received, for 2 weeks each, buprenorphine doses of 2 mg daily, 4 mg/placebo on alternating days, 16 mg daily, and 32 mg/placebo on alternating days. In each laboratory test session, participants chose between money ($2/choice) and drug (1/8 of total hydromorphone, 4 or 24 mg IM in different sessions) alternatives using an eight-trial non-independent progressive ratio schedule (FR 100, 200,...12,800). The drug dose and money amount earned was delivered after the end of the 2.5-h work period. Results: Hydromorphone 24 mg was more reinforcing than 4 mg. Higher versus lower average buprenorphine doses (regardless of daily versus alternateday schedule) significantly decreased hydromorphone 24 mg choice and increased money choice. Baseline heroin craving questionnaire scores predicted drug choice, and craving scores were significantly decreased by highdose buprenorphine. Conclusions: High-dose buprenorphine attenuated opioid drug-seeking behavior, heroin M.K. Greenwald (✉) · J.A. Hopper · C.R. Schuster C.-E. Johanson Addiction Research Institute, Wayne State University School of Medicine, 2761 E. Jefferson, Detroit, MI 48207, USA e-mail:
[email protected] Tel.: +1-313-9933965, Fax: +1-313-9931372 K.J. Schuh Eli Lilly and Company, Lilly Corporate Center, Indianapolis, IN 46285, USA
craving self-reports and increased sensitivity to alternative reinforcement. These beneficial effects were retained when high-dose buprenorphine was administered on alternate days. Keywords Buprenorphine · Heroin · Progressive ratio · Self-administration · Drug-seeking · Opioid dependence
Introduction Buprenorphine (BUP) is a mu-opioid partial agonist being developed as a treatment for opioid dependence (Lewis et al. 1983; Bickel and Amass 1995). Due to its lower intrinsic activity relative to full mu agonists, BUP produces limited respiratory depression even after administration of high doses (Walsh et al. 1994; Kishioka et al. 2000) that occupy virtually all brain mu-opioid receptors (Zubieta et al. 2000). On the other hand, because buprenorphine has high affinity for opioid receptors (Hambrook and Rance 1976; Sadee et al. 1982; Lewis et al. 1983), moderately high BUP doses can block subjective and physiological effects of mu-opioids for more than 1 day (Jasinski et al. 1978; Rosen et al. 1994; Schuh et al. 1999). Consistent with this human laboratory evidence, BUP has been shown to retain its efficacy using less-than-daily dosing schedules in several clinical trials (Fudala et al. 1990; Amass et al. 1994, 1998, 2000; Johnson et al. 1995b; Bickel et al. 1999; Petry et al. 1999). Buprenorphine has been shown to decrease opioid self-administration by primates (Mello et al. 1983; Winger and Woods 1996; Mello and Negus 1998; Briscoe et al. 2000), and human inpatient (Mello and Mendelson 1980; Mello et al. 1982), and outpatient volunteers (Johnson et al. 1995a). Comer et al. (2001) showed that BUP 16 versus 8 mg daily sublingual tablet doses decreased progressive ratio breakpoint values for intravenous heroin, but this BUP dose effect was only significant for one of three heroin doses (12.5 mg but not 6.25 mg or 25 mg). The inability of BUP 16 mg consis-
345
tently to block heroin reinforcement in that study could be due, in part, to the modest (two-fold) range of maintenance doses. There remain several gaps in our knowledge of the ability of BUP to attenuate the reinforcing effects of opioids. First, it is important to understand the ability of medications such as BUP to decrease opioid drug-seeking behavior and craving questionnaire scores, i.e. measures of drug motivation, because these antecedent behaviors could serve as useful targets for intervention. Drug-seeking behavior can be defined as the choice of (and operant responding for) a drug that precedes its actual consumption. Craving can be defined as a verbal behavioral (e.g. questionnaire) measure of drug-use intent by drug-abusing individuals, which is influenced by environmental, pharmacological and individual difference factors (Greenwald 2002). Second, although clinical trials have shown that BUP administered on a less-thandaily basis can decrease heroin use and block the subjective and physiological effects of opioids, human laboratory studies have not evaluated whether BUP alternateday dosing can decrease opioid reinforcement. This is important in order to determine the predictive validity of the human laboratory model. One pre-clinical study found that BUP 1 mg/kg (acute pre-treatment) decreased alfentanil self-administration by primates for at least 48 h but less than 72 h (Briscoe et al. 2000). Third, with the exception of one report (Greenwald et al. 1999a), there is little understanding of the external validity of medication effects in humans, i.e. whether BUP influences laboratory and naturalistic measures of opioid reinforcement concurrently. Greenwald et al. (1999a) found that, in outpatient volunteers maintained on low to moderate BUP doses (2–8 mg daily sublingual liquid), there were significant individual differences in hydromorphone choice (laboratory setting) that related to outpatient heroin use. The present study investigated opioid drug-seeking behavior by heroin-dependent outpatient volunteers who were maintained on a wide range of buprenorphine sublingual tablet doses: 2 mg daily, 4 mg alternating with placebo, 16 mg daily, and 32 mg alternating with placebo. During each BUP dose regimen, each participant chose 8 times between money and hydromorphone (4 or 24 mg) and responded on a progressive ratio schedule to earn units of the reinforcer(s). This study also examined whether drug seeking differed when BUP pretreatment (i.e. prior to each test session) was double the average daily dose (4 mg and 32 mg), the daily dose (2 mg and 16 mg), or placebo (alternating with 4 and 32 mg). It was predicted that: (1) the higher hydromorphone dose would be more reinforcing, (2) the higher average daily buprenorphine dose would decrease hydromorphone reinforcement, (3) the ability of buprenorphine to decrease drug seeking would not significantly diminish with alternateday dosing (i.e. double the daily dose versus placebo pretreatment), and (4) buprenorphine would produce similar decreases in hydromorphone choice in the laboratory and outpatient heroin use.
Materials and methods Participants The Wayne State University Institutional Review Board approved this protocol. Eligible volunteers (males and females 18–50 years old) were heroin dependent based on self-reported daily use, opioid-positive urine samples, and Structured Clinical Interview for DSM-IV (First et al. 1996). Participants provided a medical history, blood and urine samples for laboratory and drug testing, and received a tuberculin skin test, physical examination, and electrocardiogram. Those selected had no current axis I psychiatric disorders except opioid and nicotine dependence, no chronic health problems, and were not taking prescribed medications. They were not seeking treatment but were willing to be maintained on buprenorphine and undergo short-term detoxification. Volunteers provided informed consent before the study. Participants were paid $15 per session, a $15 per session completion bonus, and could receive up to $16 per session depending on the number of money choices. Twenty-eight individuals enrolled, but 14 stopped attending or were terminated for using non-opioid drugs (mostly cocaine). Fourteen individuals (11 males and three females) completed the study. Their ages ranged from 30 to 47 years (mean=41.4). Age of first heroin use ranged from 13 to 30 years (mean=19.3) and chronicity of opioid abuse ranged from 4 to 29 years (mean=22.1). Participants reported spending an average of $67 per day (range: $30 to $100) on heroin. Seven participants reported using heroin intravenously and seven reported using intranasally. All participants reported daily cigarette use. Study design The study was designed to determine the effects of different buprenorphine dose conditions on hydromorphone versus money choice. Each participant was exposed to four BUP maintenance conditions in a varied order, using a within-subject crossover design. All participants were first maintained on low average daily doses of BUP (2 mg daily and 4 mg/placebo alternating-day doses). Then all participants were maintained on high average daily doses of BUP (16 mg daily and 32 mg/placebo alternating-day doses). The order of constant daily versus active/placebo alternating-day dosing schedules within each average BUP dose level was randomized and counterbalanced across participants. These dose conditions represent two repeated measures factors. The first factor was Low versus High average dose (BUP 2 versus 16 mg). The second factor was Session pretreatment, with three levels: BUP placebo (inactive dose on the alternating-day schedule), average dose (constant active dose, either 2 or 16 mg, on the every-day dosing schedule), and double dose (active dose, either 4 or 32 mg, on the alternating-day schedule). The third factor was the hydromorphone dose available to each participant on different test session days (within each BUP dose condition), either 4 or 24 mg. Drugs Buprenorphine HCl (manufactured by Reckitt and Colman, London, UK and supplied by Research Triangle Institute, Research Triangle Park, N.C., USA) was provided as 2 and 8 mg sublingual tablets, with corresponding placebo tablets. Participants received four sublingual tablets daily. Two tablets were held under the tongue for 10 min then two more tablets for another 10 min. Different combinations of the active and placebo tablets were used to deliver all experimental doses (0, 2, 4, 16 and 32 mg) under double blind conditions. Buprenorphine administrations were observed and timed by staff. Hydromorphone (Dilaudid-HP, 10 mg/ml; Knoll Laboratories, Mount Olive, N.J., USA) was administered in 4 mg or 24 mg IM doses (constant injection volumes of 2.4 ml into the deltoid muscle). For the 4 mg dose, 2 ml of 0.9% bacteriostatic saline solution was added.
346 Procedures Participants were first maintained on 2 mg/day buprenorphine for at least 7 days. During this initial week, they received non-contingent injections of 4 mg and 24 mg hydromorphone in separate sessions. The order in which HYD doses were sampled was counterbalanced across participants. The dose administered during the first session was labeled “drug A” and the dose administered during the second session was “drug B.” Participants were instructed to attend to the effects produced by each drug because, later in the protocol, they would be able to choose these drugs. After sampling sessions, each participant was then maintained on the four different BUP dose schedules that lasted 2 weeks each. Four laboratory sessions were completed during the second week of each BUP dose condition to assure stable BUP blood levels (i.e. four half-lives), even with constant daily versus alternate-day dosing (Kuhlman et al. 1998). During two of these sessions, 4 mg (i.e. drug A or B) was the total available HYD dose and during the other two sessions 24 mg (i.e. drug B or A) was the total available dose. During conditions when active BUP doses alternated with placebo, 4 mg and 24 mg HYD were each available on a day when active (2× maintenance dose) BUP was administered, and on a day when placebo BUP was administered. Participants arrived at the laboratory at 0900 hours for test sessions. Breath alcohol and urine drug testing were performed. Subjective and physiological measures were completed. At 0945 hours, four sublingual tablets were administered. At 1030 hours, subjective and physiological measures were again completed. Lunch was provided at 1100 hours. Participants could only smoke cigarettes during the lunch break. At 1130 hours, participants began a computer choice program that lasted up to 2.5 h. At 1400 hours, subjective and physiological measures were again completed. At 1430 hours, hydromorphone was injected in a single bolus. The HYD dose administered was determined by the number of drug choice ratios that the participant completed (0–8) and the total dose available for that session (4 or 24 mg). If no drug choice ratios were completed the participant received a saline injection (i.e. to control for choosing money simply to avoid injections). Following injection, vital signs were monitored for safety. At 1630 hours, participants underwent a brief neurological test and, if unimpaired, they were paid and discharged. After the last BUP experimental dose condition, participants were detoxified over 5 weeks by administering BUP 8, 4, 2, 1, and 0 mg/day for 1 week at each dose. Urine testing Urine samples were analyzed using an Abbott ADx system (Abbott Laboratories, Abbott Park, Ill., USA) to determine participants’ use of opioids, methadone, benzodiazepines and barbiturates (cutoff values=200 ng/ml) and cocaine (cutoff value for cocaine metabolites=250 ng/ml). During non-session weeks, urine samples were collected on Monday, Wednesday and Friday. During test session weeks, urine samples were collected and analyzed before each session. If the urine test indicated any drug use other than non-methadone opioids, that session was rescheduled. The ability of buprenorphine maintenance to reduce outpatient drug use was assessed by determining the proportion of subjects with (opioid and other) drug-free urine samples on the morning of the first test session in each BUP dose condition. This strategy allowed maximal time for the maintenance dose to take effect (minimum of 10 days) and washout from residual heroin use, while avoiding the confounding effects of experimental hydromorphone administrations on urine samples during session weeks. Drug versus money choice task The progressive ratio choice task was designed to measure hydromorphone preference (completed choices) and drug-seeking responses (successively increasing operant performance requirements) in the presence of buprenorphine and alternative money reinforce-
ment. During this computerized task (modified from Heishman et al. 2000), participants were instructed that they had eight opportunities to choose between drug (HYD), money or neither during each session. Responding on a second-order progressive ratio schedule was required to obtain unit amounts of the drug or money reinforcers. Participants were required to press in sequence on the computer keyboard “d” then “r” to respond for drug and “m” then “o” to respond for money. The words “Drug” and “Money” were always visible on the computer monitor. The current number of completed ratios for each option was displayed above these words. At the start of each ratio, once the participant emitted a single response, he/she was committed to that choice (i.e. irreversible option) and a box surrounded the word on the monitor. Every tenth response produced a flash of the word, “Drug” or “Money” and an audible “beep”. A different “beep” sounded to signal completion of a ratio and there was a 15-s pause, during which a counter decreased from 15 to 0 s, the border of the computer screen changed from green to red, and responses had no effect. The first ratio required 100 responses to obtain the first reinforcer. Each successive ratio, regardless of the previous choice, required the participant to perform twice the number of responses as the previous ratio (100, 200, 400, 800, 1600, 3200, 6400, and 12800 responses for the eight ratios). Because responding for one reinforcer affected (i.e. continued to double) the response requirement for the other reinforcer, this procedure is called a non-independent progressive ratio schedule. Participants earned $2 for each completed money choice ratio ($16 maximum) and one-eighth of the total hydromorphone dose available that session (4 or 24 mg) for each completed drug choice ratio. Participants were informed at the beginning of each session whether drug “A” or drug “B”was available. The computer program stopped 2.5 h after it began (1400 hours) unless the participant had already completed all eight ratios (at which time the program stopped).
Subjective effects Computerized questionnaires were used to measure subjective drug effects. Opioid agonist and withdrawal symptoms were assessed using a 32-item questionnaire (Schuster et al. 1995). Sixteen items described agonist symptoms and 16 items described withdrawal symptoms. Each item was scored on a scale from 0 (not at all) to 4 (extremely), yielding two scales that each ranged from 0 to 64. Visual analog scales required the participants to place a mark on a 100-mm line (0=not at all, 100=extremely) that appeared on the computer screen. The scales were “any drug effect,” “high,” “good drug effect,” “bad drug effect,” “stimulated,” and “sedated.” The Addiction Research Center Inventory (ARCI; Martin et al. 1971) was used. It has 49 true/false items which are grouped into five scales: MBG (euphoria); PCAG (sedation), LSD (somatic and dysphoric changes); and BG and A scales (stimulant effects). Drug craving was assessed using the Heroin Craving Questionnaire developed by Tiffany et al. (personal communication of unpublished data; see Appendix in Schuster et al. 1995). This is a 45-item Likert-type questionnaire with each item scored from 1 (strongly disagree) to 7 (strongly agree). There are five a priori (i.e. theory-derived) scales, each with nine items. These scales are Desire to Use Drug (e.g. “I want heroin so bad now I can almost taste it”), Intent to Use (e.g. “I am going to use heroin as soon as possible”), Anticipation of Positive Outcome (e.g. “I would feel so good and happy if I used heroin now”), Withdrawal Relief (e.g. “I would feel less sick now if I used heroin”), and Lack of Control (e.g. “I could not stop myself from using heroin if I had some here now”). Factor analysis of responses by opioid-abusing participants (Tiffany et al., personal communication) yielded four independent factors and a score comprising the 34 items that load on these four factors (Schuster et al. 1995). The 34-item score was the primary measure of heroin craving in this study because it is more reliable than the nine-item scales (Greenwald 2002), and it is sensitive to opioid agonist dose (Greenwald and Roehrs 2001; Greenwald 2002) and individual differences in heroin abstinence (Greenwald et al. 1999a).
347 Physiological signs Pupil diameter was measured using an Olympus D600L digital camera. Images were transferred to a Macintosh computer, digitally filtered to enhance the contrast between the pupil and iris, and then scored for vertical and horizontal diameter on a Macintosh color monitor using the application PhotoFlash version 2.02 (Greenwald et al. 1999b). Blood pressure, heart rate and oxygen saturation measures were taken using a Criticare 507S noninvasive monitor.
Urinalysis results (outpatient drug use) just prior to HYD testing in each phase were analyzed with a two-way repeated measures ANOVA using Average daily buprenorphine dose (2 versus 16 mg) and Schedule (daily versus alternating) as factors.
Results Hydromorphone effects from sampling sessions
Data analysis To determine whether the two hydromorphone doses produced different subjective and physiological effects during sampling sessions, analyses of variance (ANOVAs) were conducted using HYD Dose (4 versus 24 mg) and Session Time as repeated measures factors. We also assessed whether the four buprenorphine dosing schedules influenced subjective effects. Data from the 45 min post-BUP time point (prior to the choice task) were analyzed by two-way ANOVA with Average daily BUP dose (2 versus 16 mg) and Session pretreatment (placebo on alternating days versus average daily dose versus double-dose on alternating days) as repeated measures factors. The primary data analyses determined whether the number of completed HYD choices (range: 0–8) and cumulative HYD responding (range: 0–25,500) varied across BUP dose conditions. A three-way repeated measures ANOVA was performed using Hydromorphone dose (4 versus 24 mg), Average daily buprenorphine dose (2 versus 16 mg) and Session pretreatment (placebo on alternating days versus constant daily dose versus double-dose on alternating days) as factors. Huynh-Feldt adjustments were used to correct for violations of sphericity when appropriate (Session time and Pretreatment factors). To assess whether subjective effects during sampling predicted HYD choice during BUP maintenance, two-tailed Pearson correlations were calculated between subjective effects during the sampling phase and HYD (24 mg dose only) choice during BUP 2 and 16 mg maintenance.
Fig. 1 Effects of sampling 4 and 24 mg intramuscular hydromorphone during initial buprenorphine sublingual tablet maintenance (2 mg/day) on mean (+SEM) scores for subjective “high” (upper left), heroin craving scores (upper right), agonist symptoms (lower left), and Addiction Research Center Inventory MBG scale (lower right)
Figure 1 illustrates hydromorphone dose and time-related effects for selected subjective effects during sampling sessions when participants were maintained on buprenorphine 2 mg daily during week 1. Table 1 summarizes statistical results for all subjective effects that were significant. Relative to HYD 4 mg, HYD 24 mg injection sigTable 1 Significant hydromorphone dose effects during sampling sessions HYD dose
HYD dose×Time
Measures with five time points Agonist symptoms Heroin craving ARCI-MBG
F(1,13), (P) 12.62 (0.005) 4.08 (0.07) 8.60 (0.02)
F(4,52), (P) NS 2.72 (0.05) NS
Measures with 6 time points High Any effect Good effect Stimulated Sedated
F(1,13), (P) 23.51 (0.001) 19.52 (0.001) 22.92 (0.0005) 11.90 (0.005) 5.95 (0.03)
F(5,65), (P) 2.67 (0.10) 5.12 (0.005) 8.01 (0.0001) 4.69 (0.002) 3.13 (0.04)
348 Fig. 2 Effects of average daily buprenorphine dose (2 versus 16 mg sublingual tablet) and total available hydromorphone dose (4 versus 24 mg intramuscular) on mean (+SEM) number of completed drug choices (upper left) and money choices (upper right), and on progressive ratio responding for drug (lower left) and money (lower right)
Table 2 Mean (±SEM) number of hydromorphone (HYD) 24 mg and alternative money choices (top portion), and HYD 4 mg and alternative money choices (bottom portion) as a function of buprenorphine (BUP) treatment conditions BUP average dose
2 mg/day
16 mg/day
Session pretreatment
0 mg
2 mg
4 mg
0 mg
16 mg
32 mg
HYD (3.0 mg/choice) Money ($2/choice) HYD (0.5 mg/choice) Money ($2/choice)
4.9 (0.6) 2.5 (0.6) 0.4 (0.2) 7.0 (0.4)
4.0 (0.7) 3.3 (0.6) 1.3 (0.7) 6.0 (0.5)
3.9 (0.8) 3.6 (0.8) 1.0 (0.6) 6.3 (0.6)
2.8 (0.8) 4.6 (0.8) 0.4 (0.2) 7.1 (0.3)
2.5 (0.9) 4.8 (0.8) 0.1 (0.1) 6.7 (0.3)
2.6 (0.7) 4.6 (0.7) 0.3 (0.2) 6.9 (0.4)
nificantly increased scores for several visual analog scale ratings (any drug effect, high, good drug effect, stimulated and sedated), the ARCI MBG scale and opioid agonist symptoms, and decreased heroin craving scores. Measures that did not significantly change following HYD 24 versus 4 mg were physiological signs (although there were mean decreases in pupil diameter, oxygen saturation and blood pressure), visual analog scale ratings of bad drug effect, the opioid withdrawal symptom scale, and all other ARCI scales. Buprenorphine dose-effects on symptom reports Relative to low average-dose (2 mg) maintenance, high average-dose (16 mg) buprenorphine maintenance decreased heroin craving scores [F(1,13)=20.18, P
