WILSHIRE TREATMENT CENTER, INC.

Hayley Pinto, (M.B.B.S., B.Sc., M.R.C.Psych.)a,, Vivienne Maskrey, (R.M.N., R.G.N.)b,

Louise Swift, (Ph.D.)b, Daphne Rumball, (M.B.Ch.B., D.R.C.O.G., F.R.C.Psych.)a,

Ajay Wagle, (M.B.B.S., M.D., M.R.C.Psych.)a,

Richard Holland, (B.M.B.Ch., D.A., D.P.H., M.F.P.H.)b

aNorfolk & Waveney Mental Health NHS Foundation Trust (NWMHFT), UK

This prospective patient-preference study examined the effectiveness in practice of methadone versus buprenorphine maintenance treatment

and the beliefs of subjects regarding these drugs. A total of 361 opiate-dependent individuals (89% of those eligible, presenting for treatment

over 2 years at a drug service in England) received rapid titration then flexible dosing with methadone or buprenorphine; 227 patients chose

methadone (63%) and 134 buprenorphine (37%). Participants choosing methadone had more severe substance abuse and psychiatric and

physical problems but were more likely to remain in treatment. Survival analysis indicated those prescribed methadone were over twice as likely

to be retained (hazard ratio for retention was 2.08 and 95% confidence interval [CI] = 1.49–2.94 for methadone vs. buprenorphine), However,

those retained on buprenorphine were more likely to suppress illicit opiate use (odds ratio = 2.136, 95% CI = 1.509–3.027, p b .001) and achieve

detoxification. Buprenorphine may also recruit more individuals to treatment because 28% of those choosing buprenorphine (10% of the total

sample) stated they would not have accessed treatment with methadone. © 2010 Elsevier Inc. All rights reserved.

Keywords: Buprenorphine; Methadone; Maintenance; Treatment retention; Opiate dependence

Opiate maintenance treatment with both methadone (MMT) and buprenorphine (BMT) has been shown to reduce the negative consequences of opiate dependence: reducing mortality, illicit drug use, criminal behavior, and spread of blood borne viruses and improving physical and mental health and social functioning (National Institute forClinical Excellence [NICE], 2007).  NICE guidance (NICE, 2007) in the UK recommends the availability of both and states that where both are “equally appropriate,” methadone should be first-line, based on its reduced cost and evidence from existing trials (summarized in the recently updated Cochrane review by Mattick, Kimber, Breen, & Davoli, 2008) that retention with methadone is better; whereas suppression of illicit opiate

use is equivalent with either treatment. However, many of the trials on which the Cochrane review was based (Fischer et al., 1999; Johnson et al., 2000; Lintzeris et al., 2004; Mattick et al., 2000; Neri et al., 2001; Petitjean et al., 2001; Strain, Stitzer, Liebson, & Bigelow, 1994) used relatively low doses of buprenorphine with slow inflexible induction, and this may have skewed the retention

rates in favor of methadone. In addition, there is virtually no research identifying features that predict a better outcome with either drug or directly examining other outcomes, such as social functioning or psychological health, to guide theclinician in deciding whether indeed both drugs are equally appropriate. The decision as to which drug to prescribe is therefore currently based on predicted differential effects based on the known pharmacology of the drugs and clinicians' opinions and patient choice, about which very little is known.  Methadone is a long-acting full agonist at the mu opiate receptor. Consequently it can offer a degree of persisting intoxication (including a degree of sedation), which may be attractive to some and limiting to others. Continued use of illicit heroin is common, limiting the reduction in harm associated with treatment. It is capable of producing fatal respiratory depression and prolongs the QTc interval (Fanoe,Hvidt, Ege, & Jensen, 2007), contributing significantly to opiate-related deaths (Auriacombe, Franques, & Tignol,2001). Withdrawal from methadone is believed by manyopiate users to be worse than heroin withdrawal. This, inaddition to the requirement for daily dosing, leads some toview methadone simply as an “alternative addiction” that prolongs dependency. These beliefs form part of a negative “lore” among some heroin users regarding methadone, which has been shown to inhibit some individuals from engaging with treatment (Kayman, Goldstein, Deren, & Rosenblum, 2006; Schwartz et al., 2008). Buprenorphine's receptor profile as a partial agonist with high affinity at the opiate receptors should in theory predict superiority over methadone in terms of safety (due to a ceiling effect on respiratory depression; Walsh, Preston, Stitzer, Cone, & Bigelow, 1994), suppression of illicit opiate use (due to increased receptor blockade), improvements in social functioning (due to reduced sedation), and progression to detoxification (due to greater ease of withdrawal).  Translation of theory into practice has however been patchy. To date, studies have shown no consistent benefit interms of suppression of illicit opiate use or social functioning. Progression to detoxification has not been examined.  Buprenorphine does appear to be safer than methadone for those individuals who remain in treatment (Auriacombe et al.,2001; Bell, Butler, Lawrence, Batey, & Salmelainen, 2009), but the impact of reduced retention rates cannot be ignored in this context, as mortality rates of injecting drug users are considerably higher when not in treatment (Gronbladh, Ohlund, & Gunne, 1990) as are levels of criminal activity and chaotic social functioning that impact on their immediate families (including their children) and society in general. Negative aspects of buprenorphine include the potential to precipitate withdrawal symptoms at induction. In addition, the sublingual tablet tastes unpleasant and dissolves slowly, reducing its acceptability and making supervision of its use more difficult. The practice of crushing and injecting the tablets also leads to significant harm. Whether the availability of an alternative has attracted into treatment individuals averse to methadone has not been examined; however, since the introduction of BMT to the UK, the number of prescriptions for both drugs has increased (NTA, 2007), implying that more individuals are being treated rather than that buprenorphine has replaced methadone in existing patients.  There are no existing trials comparing BMT and MMT based in the UK, which differs in important respects to many other countries, partly due to the free availability of both treatments in most areas and a comprehensive benefit system. These factors contribute to difficulty recruiting to randomized controlled trials to the extent that their generalizability can be severely compromised. A local pilot study found that the number of eligible subjects prepared to accept randomization constituted less than 2% (Pinto, Rumball, Maskrey, & Holland, 2008). Randomization and effective blinding also (in theory) divorce the results from patients' expectations about the effectiveness, for them as individuals, of the two alternative treatments. This could be predicted to have a powerful effect on their outcome in routine clinical practice where patients do know which treatment they are receiving.  The authors were therefore interested to explore the effectiveness of BMT versus MMT in UK clinical practice where treatment is freely available, and patient choice is now supported (NICE, 2007), outside the narrower confines of a randomized control trial. Effectiveness was measured in terms of retention, achievement of opiate detoxification, and suppression of illicit opiate use. Secondary outcomes included measurement of social functioning and psychological health. Uniquely, comprehensive baseline data were gathered regarding participants' history, detail on reasons for choice, and beliefs about both treatments to explore the basis on which patients chose between these drugs and establish whether any preexisting features or expectations predicted outcomes in treatment.

2. Materials and methods

2.1. Subjects and setting

Treatment occurred according to usual clinic practice and was not influenced by participation in the trial. Patients chose either methadone or buprenorphine in collaboration with the treating clinician. Induction in line with national guidance (Orange book) on to the drug of choice occurred over a titration period of 3 days during which the patient attended daily for monitoring. Doses were given under H. Pinto et al. / Journal of Substance Abuse Treatment 39 (2010) 340–352 341 supervision, and multiple doses could be given over the course of a day. Following titration, most patients continued

to receive their medication under supervision on most days either in the clinic or a community pharmacy. In line with normal UK practice, take-home doses were introduced on an individual basis when it was felt appropriate and safe to do so. Alteration to doses either up or down could be negotiated with the treating clinician at any point. Where general practitioners were willing, prescriptions for stable subjects were transferred to primary care. Patients were discharged from the trial if they failed to take their prescription for more than 7 days or were discharged

from the service. Criteria for discharge were production of a tampered urine sample; serious aggressive behavior to staff or on the premises; illicit drug use on the premises; and failure to engage constructively with therapy after appropriate attempts by workers to facilitate engagement. In addition to pharmacological treatment, subjects were seen individually by key workers, initially weekly then at a frequency negotiated with the subject according to their needs. Key workers act as care coordinators, providing individual support sessions that are not structured therapy.

This includes signposting and assistance with a range of needs, for example, accessing housing support, education and employment advice, and finance and benefits advice.

They act as advocates for the patients and monitor physical and mental health, making referrals to additional services as

appropriate. Medical review occurs in the service six monthly or on request of the key worker.

2.3. Measures

Detailed baseline data were collected on a very wide variety of demographic and background variables (N50). In addition, at entry, participants completed a detailed questionnaire regarding their knowledge and beliefs about the two alternative drugs and their reasons for their treatment choice. The following beliefs or knowledge were explored: taste of drug, likelihood of experiencing withdrawals when starting on the drug, effect on craving, blocking effect on heroin use, overdose risk, level of intoxication (sedation, clear headedness, emotional numbing), stigma, ease of

detoxification, and degree to which the drug was perceived as an alternative addiction.  The primary outcome was retention in treatment at 6 months or successful detoxification (i.e., a “positive” outcome). Successful detoxification was defined as patients who were recorded to be opiate-free at the point of discharge from the service during the trial. Subjects were considered as “not retained” in the trial if they failed to take their maintenance drug for seven consecutive days. A secondary outcome was suppression of illicit drug use evidenced by urine toxicology. The frequency of urine sample testing in the service is flexible according to the level of engagement and suspected or reported illicit drug use. This could have led to a wide variation in the numbers of available samples for patients in the trial. To eliminate this discrepancy and prevent possible bias, it was decided at the outset to obtain copies of the first urine toxicology report each month only from the clinical notes. Christo inventories (a measure of social functioning, addictive and risk behavior, and treatment engagement; Christo, Spurrell, & Alcorn, 2000) were completed by key workers at baseline, 3 months, and 6 months. Additional selfreport data on alcohol use (brief Alcohol Use Disorder Identification Test, brief AUDIT), illicit drug use (adapted Office for National Statistics questionnaire), and psychological health (Brief Symptom Inventory [BSI]) were collected at baseline, 3 months, and 6 months. 2.4. Sample size A pilot study (n = 42) demonstrated approximately equal numbers chose each drug and suggested that methadone retained about 15% more subjects at 6 months (68% vs. 55%)13. A total sample size of 326 was calculated to be sufficient, with 80% power, to detect a difference of 55% versus 70% using the chi-squared test at a 5% significance level. After the trial began (at n = 109), recruitment demonstrated that methadone was the more popular choice (2:1 ratio). We therefore recalculated the sample size at that point taking account of this imbalance, which raised the sample size required to 363 patients.

2.5. Analysis

Because this was a nonrandomised study, baseline data were compared to investigate differences between groups.  Retention was compared between groups using Fisher's exact test. To control for possible confounders, all variables measured at baseline that exhibited both (a) a p value less than .2 when tested for bivariate association with the primary outcome and (b) a p value less than .2 when tested between groups were included as independent variables in a logistic regression to obtain an adjusted odds ratio (OR). In addition, the Cox proportional hazards model was used to compare time retained between groups, using the same potential confounders as in the logistic regression to provide adjusted estimates. Analyses were conducted on an

intention-to-treat basis, classifying participants according to the drug they first chose, then repeated on a per-protocol basis analyzing participants according to the drug they completed the study on. Finally, a sensitivity analysis was conducted, broadening our definition of retention to include those who were transferred on a prescription to another drug

treatment service. Urine results were analyzed as repeated measures to allow for correlation between results on each patient. In addition, sustained abstinence from illicit use was investigated by comparing groups in terms of the proportion providing six negative urines, or five negative and one missed urine. Changes in Christo, BSI, and drug and alcohol use scores for those retained at follow-up were 342 H. Pinto et al. / Journal of Substance Abuse Treatment 39 (2010) 340–352 compared using a linear regression model to adjust for the corresponding baseline measure.  Within each drug category, we also investigated the effect on outcome of participants' baseline knowledge, beliefs, and reason for choice of drug. Initially, this was by a univariate analysis for each drug to explore whether those with apositive outcome had different beliefs to those with a negative outcome. We then planned to undertake a multivariate analysis if a number of beliefs appeared potentially predictive of success.

2.6. Ethical approval

The two groups appeared to be well matched (Table 1). Most were White, unemployed, male polydrug users involved in criminal activity who had been dependent on heroin (which most injected) for an average of 10 years. Most described a history of childhood adversity, with more than one third reporting some form of abuse and two thirds reporting a parent with an alcohol problem. Most participants also had poor physical and mental health. Some differences were noted. Those choosing MMT were more likely to be female (28% vs. 19%), to have resident children (23% vs. 14%), to have a drug-using partner (62.6% vs. 49.4%), to express a higher level of psychological distress (BSI's Global Symptom Index 53.6 vs. 50.3), to have a physical diagnosis (51% vs. 41%), and to be taking some medication for that condition. They had slightly poorer social functioning reflected in higher average Christo scores (11 vs. 9.7), were more likely to be using heroin intravenously (68% vs. 55%) at a higher reported dose (0.8g vs. 0.5g), and to have positive urine toxicology for cocaine.

3.3. Doses

Mean dose of buprenorphine at the end of Day 1 of titration was 6.9 mg (range = 2–12 mg), 9.9 mg (range = 2– 20 mg) at the end of Day 2, and 11.3 mg (range = 4–20) at the end the Day 3. This compared to a mean of 50.7 mg Day 1 (range = 0–135 mg), 63.8 mg Day 2 (range = 5–160 mg), and 69.6 mg at the end of Day 3 for methadone (range = 5– 170 mg). Mean maximum doses throughout the trial were 73.3 mg of methadone (range = 10–170 mg) and 11.7 mg of buprenorphine (range = 4–20 mg). More patients on buprenorphine chose to reduce their dose during the trial

rather than to increase it.

3.4. Positive outcome (i.e., retention or detoxification) At 6 months, significantly fewer of those selecting BMT (50%), than those selecting MMT (70%), achieved a positive outcome (p b .001; Fisher's exact test). This equates to a relative risk of a positive outcome for BMT of 0.71 (i.e., a reduced probability of a successful outcome) or an OR of 0.43 (95% confidence interval [CI] = 0.20–0.67, p b .001). This OR decreased (i.e., worsened) to 0.34 (95% CI = 0.20–0.59, p b .001) after adjusting for potential confounders. Almost identical results were found in a per-protocol

analysis adjusting for seven participants who swapped from buprenorphine to methadone soon after entering the study. The difference in positive outcome observed, favoring methadone, was wholly due to the difference in retention rates as detoxification was achieved by more of those receiving buprenorphine (10/134, 7.5%) than those receiving methadone (1/361, 0.3%).

3.5. Survival analysis

it was clear that they believed these drugs to be very different (see Table 4). Most marked differences (N1 point difference, on each 1- to 5-point Likert scale) were taste (buprenorphine

worst), blocking effect on heroin use (buprenorphine more effective), risk of overdose (buprenorphine safer), ease of stopping (buprenorphine easier), and time to stop (buprenorphine

less). Because this was a pragmatic study, we did not have direct control over the information clinicians discussed with subjects prior to treatment initiation.  However, subjects were asked about their sources of information and to identify that which they considered to be the main source in making their treatment choice. Eightyone percent identified their own experience as their primary

source of information (vs. 1% who identified drug workers as their primary source). In addition, subjects were asked separately whether they felt free choose their treatment; 92% responded that they did. We then investigated whether those choosing buprenorphine as a group differed in their beliefs about the two drugs to those choosing methadone. This analysis demonstrated

that those choosing buprenorphine appeared to have a consistently more negative view of methadone than those choosing methadone, whereas participants were reasonably

consistent in their view of buprenorphine irrespective of their drug choice. In particular, those choosing buprenorphine, as a group, believed that if they used methadone, then they

would be more likely to crave heroin, be less clear headed, be more drowsy, and experience more emotional numbing. They felt methadone would be harder to stop, that using methadone was more like swapping one addiction for another, and that they would be viewed more negatively by others. In contrast, the only areas where there was a difference between the two groups in views about buprenorphine were that those choosing methadone believed titration on to buprenorphine would be more uncomfortable than titration on to methadone and that craving would be less

reduced on buprenorphine (see Table 5).  Next, the relationship between beliefs and retention was investigated. A small number of these were potentially significant on a univariate analysis (p b .1). However, when these were entered together into a multivariate analysis, only one variable (belief as to whether buprenorphine blocked the effect of heroin) remained significant after adjusting for

group, age, previous custodial sentences, parental drug user, and cocaine use. The adjusted odds of retention were 5.21

Fig. 2. Survival analysis showing retention in treatment for methadone versus buprenorphine.

Table 2

Reasons for participants not completing study

Reasons

Methadone

noncompleters

(n = 69)

n (%)

Buprenorphine

non-completers

(n = 77)

n (%)

Planned detox' attempt 1 (1.5) 10 (13.0)

Transferred to other service 3 (4.3) 4 (5.2)

Client dropped out 33 (47.8) 40 (51.9)

Discharged by service

Aggression 2 (2.9) 0 (0)

Tampered urine 3 (4.3) 3 (3.9)

Persistent on-top use 4 (5.8) 4 (5.2)

Other reason 3 (4.3) 3 (3.9)

Detained (police/prison) 18 (26.1) 13 (16.9)

Death 2 (2.9) 0 (0)

346 H. Pinto et al. / Journal of Substance Abuse Treatment 39 (2010) 340–352

(95% CI = 1.71–15.85, p = .004), with those believing that

buprenorphine acted as an effective blocker being more

likely to be retained than those who did not. However, it

should be noted that more than 70% of participants

considered it to be an effective blocker (retention = 63%),

whereas only 6% thought it was not (37% retention).

3.9. Reasons for choice of treatment

Subjects choosing BMT were asked at recruitment if they

would have come into treatment if methadone were the only

treatment available, 38 (28% of the BMT group—10% of the

total sample) answered “no” to this question.

3.11. Urine results

Those prescribed buprenorphine showed evidence of

using less illicit opiate than those prescribed methadone.

The odds of having opiate-negative urine samples were

significantly higher for the buprenorphine group whether

considering all patients (OR = 2.14, 95% CI = 1.51–3.03)

or only those retained at 6 months (OR = 2.74, 95% CI =

1.77–4.22; Table 6). Evidence of sustained abstinence from

heroin was also significantly more likely in the BMT group.

Two proxy measures for this were used. First, a complete

set of six opiate-negative urine results, and then less

stringent criteria including those with five negative and one

missing urine result. On either of these measures, the

buprenorphine group was more likely to achieve sustained

abstinence. This difference was most marked in those

retained at 6 months (OR = 6.08, 95% CI = 2.15–17.16;

Table 6). No significant difference was found in urine levels

of other illicit drugs, which appeared to be unaffected by

treatment (Table 6).

Fig. 3. Comparing retention at 6 months by maximum dose of drug prescribed.

Table 3

Baseline variables predicting retention at 6 months on multivariate analysis

Variables

OR for

retention at

6 months 95% CI p value

Buprenorphine preferred 0.29 0.17–0.48 b.001

Age 1.05 1.01–1.08 .005

No. previous custodial sentences 0.92 0.88–0.97 .001

Parent was drug user 0.51 0.27–0.97 .04

No. of medications, M (Mdn) 1.25 1.06–1.47 .008

Baseline cocaine positive 0.56 0.34–0.93 .03

Baseline benzodiazepine positive 0.53 0.30–0.93 .03

H. Pinto et al. / Journal of Substance Abuse Treatment 39 (2010) 340–352 347

Christo data were available for almost all retained clients

at baseline, 3 months, and 6 months (94%–100% of retained

subjects). Although both groups showed an improvement in

Christo and BSI scores over follow-up, the changes adjusted

for baseline values were not significantly different between

groups. Regarding alcohol consumption, at 6 months the

BMT group showed an improvement (decrease in AUDIT

scores), whereas the MMT group demonstrated deterioration

(increased scores; p = .052 between groups; see Table 7).

3.13 Side effects

M (Mdn)

Methadone,

M (Mdn)

No. for matched

pairs analysis

(% subjects)

p value

(Wilcoxon matched pairs)

Taste 1.7 (1) 2.9 (3) 290 (80) b.001

Withdrawal at start (5 = bad withdrawal) 3.5 (3.5) 3.0 (3) 252 (70) b.001

Hold (5 = no withdrawals) 4.2 (5) 4.4 (5) 246 (68) .05

Craving for heroin (5 = craving stopped) 3.6 (4) 3.0 (3) 248 (69) b.001

Change effect of heroin (5 = total block) 4.3 (5) 1.7 (1) 285 (79) b.001

Risk of overdose (5 = low) 3.3 (3) 1.6 (1) 173 (48) b.001

Clearheaded (5 = not at all) 1.7 (1) 2.6 (3) 258 (71) b.001

Drowsiness (5 = very sleepy) 1.8 (2) 2.8 (3) 252 (70) b.001

Emotional numbing (5 = emotional numbing) 1.9 (1) 2.8 (3) 214 (59) b.001

Others' views of addiction (5 = not addict) 2.2 (1) 1.3 (1) 308 (85) b.001

Ease of stopping drug (5 = very hard) 3.0 (3) 4.3 (5) 245 (68) b.001

Time to stop (% days/weeks vs. months) 49.3% 12.8% 219 (61) b.001 a

Swapping one addiction for another (5 = not true) 3.3 (4) 2.4 (2) 286 (79) b.001

a McNemar test for matched pairs.

Table 5

Comparison of participants' views by group about buprenorphine and methadone treatment

Views

Buprenorphine group,

M (SD)

Methadone group,

M (SD)

p value

(Mann–Whitney test)

Buprenorphine taste (5 = great) 1.8 (0.9) 1.7 (0.8) .05

Methadone taste (5 = great) 3.1 (1.1) 2.8 (1.1) .006

Importance of taste (5 = not at all) 4.2 (1.3) 3.8 (1.3) .01

Withdrawals experienced when starting buprenorphine (5 = yes, badly) 3.2 (1.1) 3.7 (1.3) b.001

Withdrawals experienced when starting methadone (5 = yes, badly) 2.5 (1.3) 3.2 (1.2) b.001

Buprenorphine stop withdrawals (hold; 5 = no withdrawals) 4.2 (1.0) 4.1 (1.1) .55

Methadone stop withdrawals (holds; 5 = no withdrawals) 4.2 (1.0) 4.4 (0.9) .19

Heroin craving when on buprenorphine (5 = expect cravings to stop) 3.9 (1.1) 3.4 (1.5) .001

Heroin craving when on methadone (5 = expect cravings to stop) 2.6 (1.3) 3.2 (1.3) b.001

Change effect of heroin when on buprenorphine (5 = totally blocked) 4.5 (0.9) 4.3 (1.1) .15

Change effect of heroin when on methadone (5 = totally blocked) 1.5 (0.9) 1.8 (1.1) .04

Risk of overdose on buprenorphine (5 = low) 3.4 (1.5) 3.2 (1.5) .46

Risk of overdose on methadone (5 = low) 1.7 (1.0) 1.6 (1.0) .40

Clearheaded on buprenorphine (5 = not at all) 1.7 (0.8) 1.7 (1.0) .84

Clearheaded on methadone (5 = not at all) 3.1 (1.2) 2.3 (1.0) b.001

Drowsiness on buprenorphine (5 = very sleepy) 1.9 (0.9) 1.7 (1.0) .14

Drowsiness on methadone (5 = very sleepy) 3.4 (1.2) 2.5 (1.1) b.001

Emotional numbing effect of buprenorphine (5 = disconnected from emotions) 1.9 (1.1) 1.8 (1.1) .19

Emotional numbing effect of methadone (5 = disconnected from emotions) 3.4 (1.3) 2.5 (1.2) b.001

View of other people about buprenorphine (5 = not a drug addict) 2.2 (1.4 2.2 (1.5) .63

View of other people about methadone (5 = not a drug addict) 1.2 (0.7) 1.4 (0.9) .04

Ease of stopping buprenorphine (5 = very hard) 2.8 (1.2) 3.1 (1.3) .09

Ease of stopping methadone (5 = very hard) 4.6 (0.8) 4.2 (0.9) b.001

How long to stop buprenorphine (% days or weeks vs. months) 50.5% 48.1% .71

How long to stop methadone (% days or weeks as opposed to months) 11.2% 17.3% .16

Buprenorphine is swapping one addiction for another (5 = not true) 3.4 (1.4) 3.2 (1.5) .37

Methadone is swapping one addiction for another (5 = not true) 1.9 (1.2) 2.7 (1.5) b.001

Importance of receiving a drug experienced before 3.0 (1.6) 2.7 (1.6) .11

348 H. Pinto et al. / Journal of Substance Abuse Treatment 39 (2010) 340–352

(55% vs. 44%, p = .01). The most common side effects in the

MMT group were sweating (19%), sedation (12%), and

constipation (10%). Side effects most frequently reported in

the BMT group were sedation (9%) and constipation (8%).

3.14. Adverse events

Two patients died during the trial period, both in the

methadone group. One death was due to a heroin overdose

shortly after the start of treatment related to adverse social

circumstances. The second was recorded as methadone

toxicity in combination with two other sedative drugs in a

patient prescribed the same dose for some time.

4. Discussion

This is the first large study to compare the effectiveness

of buprenorphine maintenance with methadone maintenance

in ordinary UK clinical practice. The results demonstrate

that those treated with buprenorphine are more than 50%

less likely to remain in treatment for 6 months than those

receiving methadone (adjusted hazard ratio = 0.47, 95% CI

= 0.32–0.69, p b .001). This is considerably greater than the

effect size pooled from randomized controlled trials

(relative risk = 0.85) by Mattick et al. (2008) and similar

to that Burns et al. (2009) found in a large (nonrandomized)

retrospective cohort study. However, this poorer retention

needs to be balanced against a finding of superiority in

suppression of illicit opiate use by buprenorphine in those

who were retained. This was demonstrated in several ways

including an overall reduction in the number of urine

samples positive for opiates, increased rates of sustained

abstinence from illicit use, and increased rates of progression

to detoxification. These findings are in contrast to

previous randomized trials but similar to the results that

Vigezzi et al. (2006) found in another nonrandomized,

nonblinded field trial. In addition, our results suggest that

BMT may be associated with reduced alcohol use but did

not demonstrate any difference in social functioning or

psychological health.

The hypothesis that poor retention is due to slow titration

or inadequate dosing is not supported by the results of this

trial. Rapid titration was achieved, and although loss during

titration was more common in the BMT group and made a

small contribution to the overall outcome, an increased rate

of loss was seen throughout the trial. The mean maximum

doses of buprenorphine achieved (11.7 mg) were slightly

below the current UK guidelines (12–16 mg for most

patients; DOH, 2007) but in the upper range of the doses

used in the flexible dosing trials Mattick et al. (2008)

included in the Cochrane review, almost identical to

estimated average doses found in Australian clinical practice

by Burns et al. (2009) but above the average dose (8.9 mg)

found in a recent national audit of actual prescribing 2004–

2005 by the NTA (2007). Dosing was flexible, and subjects

did not appear to seek higher doses. In fact, more of those on

buprenorphine chose to reduce their dose than to increase it.

Furthermore, our results suggest that above 8 mg increasing

doses for buprenorphine did not improve retention. No

studies to date have specifically compared retention for doses

greater than 8 mg, but four studies, which examined dose of

buprenorphine as a predictor of retention in a secondary

analysis, failed to find a relationship (Gerra et al., 2004,

2006; Soyka, Zingg, Koller, & Kuefner, 2008; Vigezzi et al.,

2006). Kakko et al. (2007) reported 78% retention using a

very high mean dose (29 mg), but this was an intention-totreat

analysis where only 35% of the BMT group actually

remained on buprenorphine at 6 months.

Examination of predictors of retention in the sample as a

whole revealed treatment of choice as by far the strongest

factor. Additional predictors (older age, fewer custodial

sentences, and absence of drug-using parents), although less

powerful, made intuitive sense. Additional use of benzodiazepines

and cocaine/crack is known to negatively effect

prognosis (Marsden et al., 2009; Rooney, Kely, Bamford,

Sloan, & O'Connor, 1999). The use of more prescribed

medications was surprising but could imply more engage-

Table 6

Urine test results: repeated measures analysis and analysis of opiate abstinence

Variables

All subjects Subjects retained at 6 months

No. of tests

OR of a negative urine

sample for the BMT vs.

MMT group (95% CI) p value No. of tests

OR of a negative urine

sample for the BMT vs.

MMT group (95% CI) p value

Opiate 1,400 2.136 (1.509–3.027) .000 1,122 2.735 (1.771–4.224) .000

Cocaine 1,400 1.514 (0.989–2.318) .056 1,122 1.331 (0.766–2.313) .311

Amphetamine 1,395 0.809 (0.364–1.794) .601 1,121 0.771 (0.260–2.290) .640

Benzodiazepines 1,395 1.262 (0.745–2.137) .388 1,120 1.350 (0.670–2.718) .401

Cannabis 1,199 1.166 (0.722–1.884) .529 969 1.432 (0.782–2.624) .245

OR BMT vs. MMT

groups (95% CI)

OR BMT vs. MMT

groups (95% CI)

Opiate abstinence (6 negative urines) 2.693 (1.071–6.767) .04 6.079 (2.153–17.163) .001

Opiate abstinence (6 negative urines

or 5 negative + 1 missing)

2.391 (1.122–5.096) .02 4.748 (2.081–10.831) .03

H. Pinto et al. / Journal of Substance Abuse Treatment 39 (2010) 340–352 349

ment with treatment systems or experience of more negative

consequences of drug use leading to greater motivation to

engage. The numbers retained were too small to confidently

examine baseline predictors of retention in the buprenorphine

group separately. Previous studies, which have

examined baseline factors, identified comorbid depression

and chronicity of dependence, but these have not been

successfully replicated (Gerra et al., 2004, 2006; Poirier

et al., 2004; Schottenfield, Pakes, & Kosten, 1998; Soyka

et al., 2008).

Rather than any simple demographic variable, an

alternative hypothesis is that the disparity in retention

between methadone and buprenorphine is more closely

related to the match between individuals' needs, treatment

aspirations, and expectations.

Firstly, with regard to aspiration, maintenance treatment

is usually considered as a route to abstinence from illicit drug

use. However, for some injecting drug users, it may be seen

as a means to reestablish control over their use, allowing

them to comfortably forgo alternative opiates (usually

heroin) but retain the option to use occasionally, either

recreationally or as a coping strategy. Although possible, this

is less easy for those using buprenorphine, a drug that blocks

opiate receptors. It is possible, therefore, that more of those

choosing methadone did so with these more achievable

aspirations. The finding that retention was prolonged in

those who expressed stronger baseline beliefs about

buprenorphine as a “blocker” with less intoxicating effects,

and chose not to take it, supports such a hypothesis.

Furthermore, those choosing methadone commonly cited

fear of withdrawal and issues around desire for expected

opioid effects as reasons for choosing it.

In contrast, apart from a dislike of methadone, those

choosing buprenorphine commonly cited its ability to block

heroin use, ease of detoxification, and lack of intoxicating

effects as reasons for their choice. This may reflect higher

aspirations around elimination of illicit opiate use and

detoxification. A previous study (Pinto, Rumball, &

Holland, 2008) found that buprenorphine is often viewed

as a drug for those who are “serious” about treatment.

Although this would explain the better performance in this

group regarding illicit opiate use, it would also leave those

individuals who failed to achieve these more challenging

goals vulnerable to disillusionment and disengagement.

Secondly buprenorphine is intrinsically less rewarding than

methadone and provides less sedative effect. This might have

been expected to be of particular value to those expressing

more psychological distress, but BSI scores did not predict

retention in this study. Finally, because of its reduced

potency, a prescription for buprenorphine may simply be

easier for subjects to disengage from than high-dose

methadone.

The numbers lost to treatment on BMT are of concern,

and the authors suggest that further research should focus on

the relationship between treatment aspirations, treatment

choice, and retention in treatment.

Table 7

Comparison of Christo, AUDIT, and BSI scores over follow-up

Measures Baseline, M (SD) p value a 3 months, M (SD) p value a 6 months, M (SD) p value a

Mean change

0–3 months (SD) p value b

Mean change

0–6 months (SD) p value b

Christo Bup 9.7 (3.7), n = 133 .044 5.9 (3.8), n = 77 .014 5.5 (3.5), n = 57 .121 3.9 (3.6) .098 4.4 (3.8) .205

Meth 11.0 (3.6), n = 227 7.1 (4.0), n = 187 6.5 (3.8), n = 156 3.6 (4.3) 4.2 (4.2)

Audit Bup 3.5 (3.3), n = 134 .843 3.9 (3.5), n = 46 .546 3.0 (3.3), n = 31 .193 0.0 (3.6) .863 0.8 (3.9) .052

Meth 3.6 (3.5), n = 227 3.5 (3.0), n = 101 3.9 (3.3), n = 84 0.5 (2.9) −0.7 (3.0)

GSIT Bup 50.3 (13.7), n = 133 .066 50.2 (12.1), n = 46 .649 50.2 (13.4), n = 31 .779 2.3 (11.9) .995 2.0 (12.5) .852

Meth 53.6 (11.8), n = 227 51.5 (13.3), n = 101 51.7 (14.1), n = 85 2.8 (10.1) 2.0 (11.7)

PSDIT Bup 48.7 (12.8), n = 133 .071 47.9 (11.0), n = 46 .759 45.2 (11.6), n = 31 .197 2.3 (11.1) .970 5.2 (12.0) .362

Meth 51.7 (11.0), n = 227 49.3 (12.5), n = 101 49.0 (13.6), n = 85 3.0 (9.9) 3.9 (10.7)

PSTT Bup 51.7(14.7), n = 133 .122 52.4 (13.0), n = 46 .818 52.5 (13.4), n = 31 .808 1.1 (12.1) .468 0.5 (14.1) .878

Meth 54.6 (12.5), n = 227 52.1 (12.6), n = 101 53.5 (13.95), n = 85 3.0 (10.7) 0.6 (13.0)

Note. Those retained at time of measurement only. Sample sizes for changes correspond to those of corresponding follow-up period. Changes for all measures are baseline minus follow-up, positive is improvement

as in all cases higher scores = worse health state. GSIT = Global Severity Index T score; PSDIT = Positive Symptom distress index T score; PSTT = Positive Symptom Total T score.

a Wilcoxon Mann–Whitney.

b Linear regression including baseline measurement as independent variable.

350 H. Pinto et al. / Journal of Substance Abuse Treatment 39 (2010) 340–352

Exploration of subjects' views about the drugs revealed

that they are generally very well informed, with the largest

perceived differences between the drugs being in taste,

ability to block heroin use, and ease of coming off the drug.

The stated reasons for choice, in addition to aspirations,

revealed more about what is perceived to be important. Past

experience was commonly cited by both groups, but safety

was rarely mentioned, highlighting the mismatch in priorities

between patients and their clinicians.

The pragmatic, open-label design of this study has

strengths and weaknesses. A very high recruitment rate

was achieved (89% of the eligible population), and subjects

received the treatment available in ordinary National Health

Service (NHS) community drug services. This makes the

results highly generalizable. However, the limitations must

be acknowledged. The absence of randomization is one

limitation. Although the groups appeared similar on most

baseline variables, there were some differences between

groups. These differences suggested that the methadone

group had more risk factors for poor retention. Despite this,

adjustment for baseline differences only increased the gap in

retention rates between BMT and MMT. Nevertheless, we

cannot be certain that unmeasured differences did not favor

the methadone group. No measures were taken to reduce the

possible impact of the views and expectations of treating

staff and subjects because these are inevitably present and

contribute to effectiveness in real clinical practice. In

addition, we deliberately included an exploration of subjects'

beliefs and expectations as a part of our study. Furthermore,

because of the pragmatic nature of the trial, we were unable

to dictate the frequency of urine sampling. As a result, our

data are restricted to one per month, which gives only a

limited picture of possible drug use and therefore weakens

the findings regarding suppression of illicit use.

The marked difference in retention rates raises important

issues regarding the use of buprenorphine for maintenance

in the UK, where commissioners have imposed targets of

80% retention at 3 months, linked to funding (BMT falls

below this at 61.2%, whereas MMT reaches 84.1%). In

addition, buprenorphine is more expensive than methadone.

However, the extra cost may be offset for services (treatment

is free for NHS patients) if BMT increases progression to

abstinence and detoxification allowing shorter episodes of

treatment. Our results also provide evidence that the

availability of buprenorphine recruits additional individuals

into treatment. Not only did 10% of the sample state that

they would not have entered treatment where methadone is

the only available option, but all of those accepting MMT

because they thought they could not be titrated on to BMT

dropped out of treatment. In addition, the group choosing

BMT expressed a more negative view of methadone, and

more than half cited dislike of methadone as the reason they

chose buprenorphine. This would suggest that although

buprenorphine should, as per NICE Guidance, be an

available option for all, services should closely monitor

those on this drug for the first 3 months and maintain a low

threshold for conversion to methadone for those not

achieving illicit abstinence.

In conclusion, this study has shown that in a UK drug

treatment context, those choosing methadone maintenance

are considerably more likely to be retained in treatment at

6 months than those choosing buprenorphine. This

difference appears unrelated to dose, titration procedure,

or subjects' beliefs and expectations about the drugs.

Those retained on buprenorphine appear more able to

suppress illicit opiate use and progress to detoxification. In

addition, provision of buprenorphine has merits in

increasing recruitment to treatment.

5. Addresses of sites

Trust Alcohol and Drug Service, NWMHPT, Norfolk,

England:

• 7 Unthank Road , Norwich NR2 2PA,

• 22-24 Colgate, Norwich NR3 1BQ

• The Weavers Centre, Hellesdon Hospital, Drayton

High Road, Norwich and

• The Willows Centre, Northgate Hospital, Northgate,

Great Yarmouth.

Acknowledgments

The authors would like to thank the East Norfolk and

Waveney Research Consortium for funding for this project,

Laura Vincent and Lisa Regan for their invaluable assistance

with data entry, the staff of TADS, and all patients who took

part in this study.

References

Auriacombe, M., Franques, P., & Tignol, J. (2001). Deaths attributable to

methadone vs. buprenorphine in France. Journal of the American

Medical Association, 285, 45.

Bell, J. R., Butler, B., Lawrence, A., Batey, R., & Salmelainen, P.

(2009). Comparing overdose mortality associated with methadone

and buprenorphine treatment. Drug & Alcohol Dependence, 104,

73−77.

Burns, H., Randall, D., Hall, H., Law, M., Butler, T., Bell, A., et al. (2009).

Opioid agonist pharmacotherapy in New South Wales from 1985 to

2006: Patient characteristics and patterns and predictors of treatment

retention. Addiction, 104, 1363−1372.

Christo, G., Spurrell, S., & Alcorn, R. (2000). Validation of the Christo

Inventory for Substance-misuse Services (CISS): A simple outcome

evaluation tool. Drug & Alcohol Dependence, 59, 189−197.

DOH. (2007). Drug misuse and dependence: UK Guidelines on Clinical

Management. London: Stationary Office.

Fanoe, S., Hvidt, C., Ege, P., & Jensen, G. B. (2007). Syncope and QT

prolongation among patients treated with methadone for heroin

dependence in the city of Copenhagen. Heart, 93, 1051−1055.

Fischer, G., Gombas, W., Eder, H., Jagsch, R., Peternell, A., Stuhlinger, G.,

et al. (1999). Buprenorphine versus methadone maintenance for the

treatment of opioid dependence. Addiction, 94, 1337−1347.

Gerra, G., Borella, F., Zaimovic, A., Moi, G., Bussandri, M., Bubici, C.,

et al. (2004). Buprenorphine versus methadone for opioid dependence:

H. Pinto et al. / Journal of Substance Abuse Treatment 39 (2010) 340–352 351

Predictor variables for treatment outcome. Drug & Alcohol Dependence,

75, 37−45.

Gerra, G., Leonardi, C., D'Amore, A., Strepparola, G., Fagetti, R., Assi, C.,

et al. (2006). Buprenorphine treatment outcome in dually diagnosed

heroin dependent patients: A retrospective study. Progress in Neuro-

Psychopharmacology & Biological Psychiatry, 30, 265−272.

Gronbladh, L., Ohlund, L. S., & Gunne, L. M. (1990). Mortality in heroin

addiction: Impact of methadone treatment. Acta Psychiatrica

Scandinavica, 82, 223−227.

Johnson, R. E., Chutuape, M. A., Strain, E. C., Walsh, S. L., Stitzer, M. L.,

& Bigelow, G. E. (2000). A comparison of levomethadyl acetate,

buprenorphine, and methadone for opioid dependence. New England

Journal of Medicine, 343, 1290−1297.

Kakko, J., Gronbladh, L., Svanborg, K. D., von Wachenfeldt, J., Ruck, C.,

Rawlings, B., et al. (2007). A stepped care strategy using buprenorphine

and methadone versus conventional methadone maintenance in heroin

dependence: A randomized controlled trial. American Journal of

Psychiatry, 164, 797−803.

Kayman, D. J., Goldstein, M. F., Deren, S., & Rosenblum, A. (2006).

Predicting treatment retention with a brief “opinions about methadone"

scale. Journal of Psychoactive Drugs, 38, 93−100.

Lintzeris, N., Ritter, A., Panjari, M., Clark, N., Kutin, J., & Bammer, G.

(2004). Implementing buprenorphine treatment in community settings in

Australia: Experiences from the Buprenorphine Implementation Trial.

American Journal on Addictions, 13(Suppl 1), S29−S41.

Marsden, J., Eastwood, B., Bradbury, C., Dale-Perera, A., Farrell, M.,

Hammond, P., et al. (2009). Effectiveness of community treatment for

heroin and crack cocaine addiction in England: A prospective, intreatment

cohort study. Lancet, 374, 1262−1270.

Mattick, R. P., Ali, R., White, J. M., O'Brien, S., Wolk, S., & Danz, C.

(2000). Buprenorphine versus methadone maintenance therapy: A

randomised double-blind trial with 405 opioid-dependent patients.

Addiction, 98, 441−452.

Mattick, R. P., Kimber, J., Breen, C., & Davoli, M. (2008). Buprenorphine

maintenance versus placebo or methadone maintenance for opioid

dependence. Cochrane Database of Systematic Reviews, 2, CD002207.

Neri, S., Bruno, C. M., Pulvirenti, D., Malaguarnera, M., Italiano, C.,

Mauceri, B., et al. (2001). Randomised clinical trial to compare the

effects of methadone and buprenorphine on the immune system in drug

abusers. Journal.

NICE technology appraisal guidance 114, (2007) Methadone and

buprenorphine for the management of opioid dependence. Available

at: www.nice.org.uk/TA114.

NTA. (2007). National Prescribing Audit: An assessment of prescribing

practices for opioid substitution treatment in England, 2004–2005.

Available at www.nta.nhs.uk.

Petitjean, S., Stohler, R., Deglon, J., Livoti, S., Waldvogel, D., Uehlinger,

C., et al. (2001). Double-blind randomised trial of buprenorphine and

methadone in opiate dependence. Drug and Alcohol Dependence, 62,

97−104.

Pinto, H., Rumball, D., & Holland, R. (2008). Attitudes and knowledge of

substance mis-users regarding buprenorphine and methadone maintenance

therapy. Journal of Substance Use, 13, 143−153.

Pinto, H., Rumball, D., Maskrey, V., & Holland, R. (2008). A pilot study for

a randomized controlled and patient preference trial of buprenorphine

versus methadone maintenance treatment in the management of opiate

dependent patients. Journal of Substance Use, 13, 73−82.

Poirier, M. F., Laqueille, X., Jalfre, V., Willard, D., Bourdel, M. C.,

Fermanian, J., et al. (2004). Clinical profile of responders to

buprenorphine as a substitution treatment in heroin addicts: Results of

a multi-center study of 73 patients. Progress in Neuro-Psychopharmacology

& Biological Psychiatry, 28, 267−272.

Rooney, S., Kely, G., Bamford, L., Sloan, D., & O'Connor, J. J. (1999).

Co-abuse of opiates and benzodiazepines. Irish Journal of Medical

Science, 168, 36−41.

Schottenfield, R. S., Pakes, J. R., & Kosten (1998). Prognostic factors in

buprenorphine vs. methadone maintained persons. Journal of Nervous

and Mental Diseases, 186, 35−43.

Schwartz, R. P., Kelly, S. M., O'Grady, K. E., Mitchell, S. G., Peterson,

J. A., Reisinger, H. S, et al. (2008). Attitudes toward buprenorphine

and methadone among opioid-dependent individuals. American

Journal on Addiction, 17, 396−401.

Soyka, M., Zingg, C., Koller, G., & Kuefner, H. (2008). Retention rate and

substance use in methadone and buprenorphine maintenance therapy and

predictors of outcome: Results from a randomized study. International

Journal of Neuropsychopharmacol, 11, 641−653.

Strain, E. C., Stitzer, M. L., Liebson, I. A., & Bigelow, G. E. (1994).

Comparison of buprenorphine and methadone in the treatment of opioid

dependence. American Journal of Psychiatry, 151, 1025−1030.

Vigezzi, P., Guglielmino, L., Marzorati, P., Silenzio, R., De Chiara, M.,

Corrado, F., et al. (2006). Multimodal drug addiction treatment: A field

comparison of methadone and buprenorphine among heroin- and cocainedependent

patients. Journal of Substance Abuse Treatment, 31, 3−7.

Walsh, S. L., Preston, K. L., Stitzer, M. L., Cone, E. J., & Bigelow, G. E.

(1994). Clinical pharmacology of buprenorphine: Ceiling effects at high

doses. Clinical Pharmacology and Therapeutics, 55, 569−580.

352 H. Pinto et al. / Journal of Substance Abuse Treatment 39 (2010) 340–352