NL | RO | FR | BE

An alcohol & drug rehab clinic

Methadone Increases Craving for Heroin

 A study of chronic opiate users receiving methadone substitution treatment.

By H. Valerie Curran1, Judi Bolton1, Shamil Wanigaratne2 & Ciaron Smyth3

1Department of Psychology, University College London, 2Department of hospital Psychology, 3Institute of Psychiatry & Substance Misuse Services, Camden and Islington NHS Trust, London, UK


Aims: To assess the acute-on-chronic effects of methadone on drug craving, mood and cognitive and psychomotor functioning in patients on long-term methadone substitution treatment.

Design and participants: A double-blind, cross-over design was used to compare the effects of a 33% increase in patient's daily dosage of methadone with a matched placebo linctus. Eighteen patients completed the study; all were assessed pre- and post-drug on two separate testing days. Findings. Methadone significantly increased both positive craving (expected positive effects) and negative craving (expected relief of withdrawal discomfort) for heroin. Patients were unable to distinguish between methadone and placebo treatments. No differences between treatments emerged in cognitive or psychomotor effects. In terms of mood, patients were more alert and more contented following placebo than following methadone. Conclusions. Additional methadone may "prime" cravings for heroin in methadone substitution patients.


Methadone is the most widely used pharmacological treatment for people dependent on opioids, with methadone substitution programmes remaining popular in many Countries since the advent of HIV and MDS (Gossop & Grant, 1991; Ball & Ross, 1991). Despite its widespread use, relatively little is known of the acute on chronic effects of methadone on drug craving, cognitive function and mood. Methadone is an opiate agonist acting at µ opioid receptors (Koob, 1992). It is well absorbed with oral bioavailability of 90% (Sawe, 1986) and peak plasma concentration being reached 2-4 hours after oral administration (Koob, 1992). Its plasma elimination half-life is 16-24 hours in opioid-naive people, but in chronic users extends to 24-48 hours) showing considerable individual variation (Tennant, 1987). This variation in how individuals metabolize thc drug may contribute to differences in withdrawal symptoms reported by patients on methadone substitution following acute daily dosage (Dyer & White, 1997).Tolerance to the various effects of most centrally acting drugs builds up over repeated use. However, tolerance may develop at different rates to the drug's differing effects. For example, with benzodiazepines, tolerance to sedative effects develops faster than tolerance to anxiolytic or cognitive effects (Curran, 1991).

Tolerance to the euphoric and analgesic effects of opioids such as methadone is thought to develop over repeated use (Gossop, 1987) but it is not clear whether tolerance also builds up to methadone's other effects. The prescribed dose is seldom escalated within treatment settings, but patients may obtain extra methadone as well as other opioids outside the treatment setting.

Studies of the effects of methadone on mood have focused predominantly on mood-related opiate withdrawal symptoms and craving. Musselman & Kell (1995) report improvements over a period of 8 months in anxiety and dysthymia in 71 patients enrolled in a methadone maintenance programme. However, it was not clear from this study whether patients' mood was stabilizing over time on methadone or whether patients were also using illicit drugs. Hiltunen et al (1995) found plasma levels of methadone over a 24-hour period correlated with self-rated levels of anxiety and irritability, although lack of placebo control hinders interpretation of the self report data. Dawe & Gray (1995) compared methadone to clonidine in 16 detoxifying opiate addicts. No significant difference in the effects of these two drugs was found in terms of people’s desire to use heroin either for its pleasant effect or for alleviating unpleasant mood or feelings.

Drug craving is a central concern for most addicts and for many theoretical conceptualizations of drug dependence. Recently) Tiffany (1992) 1997) has produced a critique of research on drug craving at both a conceptual and an empirical level. At a conceptual level, he has proposed a cognitive processing model of drug craving and drug use. At an empirical level, he has developed several multi-item questionnaires for thc reliable assessment of craving in different drug using groups. The most recent scale is thc 'heroin craving questionnaire" (HCQ) (Tiffany et al, 1999). Thc HCQ was initially validated in a sample of 230 heroin users who were not trying to stop using heroin. It has also been used in research with opiate users in the United Kingdom (Weinstein et al, 1998). The HCQ assesses five components of craving for heroin: (i) desire to use (ii) intention to use (iii) anticipation of positive outcome (iv) relief from withdrawal and dysphoria and (v) lack of control over use. Factors (iii) and (iv) correspond, respectively, to positive and negative craving.

Given current concerns about "drug driving" and road traffic accidents, it is also important to know whether methadone impairs the psychomotor skills and speed of reaction time which are central to driving skills. Patients taking other prescribed psychotropics such as benzodiazepines and certain antidepressants are warned about deleterious effects of these drugs on driving and on operating machinery. In reviewing the literature on the cognitive effects of a wide range of opioids, Zaeny (1995) points out that most studies of methadone have methodological problems such as lacking appropriate controls, using small numbers of participants, not allowing for patients' polydrug use and often relying on one or two arbitrarily selected cognitive tests (especially DSST). One exception is a study by Rothenberg et al (1977) which assessed the effects of two doses of methadone (5 mg, 10 mg) and placebo in 12 patients on methadone substitute treatment and in 12 healthy volunteers. The drug impaired the performance of healthy volunteers on both simple and choice reaction-time tasks, but did not affect the performance of patients, implying that tolerance to the psychomotor or sedative effects of methadone builds up during maintenance treatment.

If maintenance patients become differentially tolerant over long-term use to the effects of methadone on drug craving, mood, cognitive and psychomotor function) then an acute increase in methadone dose should affect measures of these four aspects differentially. Therefore the present study was designed to assess the acute-on-chronic effects of methadone on these measures in maintenance patients. The increased dose was selected as a percentage (33%) increase of each individual's total daily dose. The 33% increase was considered to be the highest dosage useable whilst minimizing any risk of respiratory depression in the outpatient population assessed. The increase was additional to the patient's normal, once-daily dose of methadone. A double-blind, balanced cross-over design was chosen to compare additional methadone with a carefully matched placebo. Of central concern was whether patients' craving, mood, cognitive and psychomotor function changed following additional methadone compared with additional placebo. Our hypothesis was that the increased dose would have global sedative effects which could affect performance across a range of tasks, rather than specific effects on particular cognitive processes. It was also predicted that sedative effects would be apparent following methadone on the subjective mood rating scale employed. In terms of drug craving, the literature does not support a unidirectional hypothesis. Pharmacologically, one would expect a reduction in craving for heroin following additional methadone. However, an opposite effect (increased craving) would be predicted from studies where a drug "primes" craving for more of that drug (e.g. Hodgson, Rankin & Stockwell, 1979).

To test during peak methadone effect, assessments were administered before and 3-4 hours following the additional treatments (Koob, 1992). A range of measures were employed to assess the effects of this increase on craving, cognitive and psychomotor function and mood.



Patients over 18 years of age were included who had been taking prescribed methadone for a minimum of 6 months and who had been on the same dose in the 20-100 mg per day range for the previous 4 weeks. In addition, participants needed to have basic literary skills and bc able and willing to give informed consent. Patients were excluded if they had any past history of severe head injury or organic cognitive dysfunction. Eighteen participants fulfilling our inclusion criteria were recruited via their key workers and completed all assessments on both testing days.

The project took place at the Camden and Islington Substance Misuse Services, an out-patient service with over 2000 registered patients, most of whom are classified as opiate-dependent and polydrug users. Ethical approval was given by the ethics committees of University College London and of Camden and Islington Community Health Services NHS Trust. All participants gave written, informed consent.


A double-blind, placebo-controlled, cross-over design was employed with participants being assigned randomly to treatment order (methadone or placebo on the first testing day). Testing occurred pre-drug and then began again 3 hours post-drug on each test day. Versions of tests were counterbalanced across subjects and design.

Participants attended two separate sessions with a 1-week "washout" period between. They were informed that a urine sample would required and breath alcohol levels would be measured on each day. Participants were assured that results of these tests (and all other information obtained in the study) were confidential and would not be reported to their key workers.


Methadone formulated in linctus and placebo (the same linctus without methadone) was purchased from Martindale Pharmaceuticals. Eight drops of peppermint essence were added to each dose to give the same taste to methadone and placebo syrups. A hospital pharmacist dispensed both methadone (33% of each participant's daily prescribed dose) and a matched quantity of placebo linctus.


A battery of tests was administered at four time-points (pre- and post-drug on each test day). Two additional scales (heroin craving questionnaire and "guess on treatment") were given only at the post-drug testings. Participants arrived at the same time on each test day (either 9 or 10 a.m.) and had taken their daily dose of methadone approximately 40-80 minutes beforehand. The waiting time between test sessions was filled partly by chatting and giving participants time to relax. In addition, on the first test day, this time was also used for administering the National Adult Reading Test (NART) and a drug use questionnaire. One the second test day, participants also watched a video.

A range of tests were used to determine the effect of methadone on mood, cognitive and psychomotor functioning, and craving for heroin. Order of test administration is given in Table 1 .

National Adult Reading Test (NART). This test was administered once to provide an estimate of pre-morbid intellectual function ( Nelson, 1982).

Table 1. Order of test administration


Screening tests each dayUrine sample

Tests given pre- and post-drugImmediate prose recall

Mood rating scale

Digit symbol substitution test (DSST)

Simple reaction time test

Finger-tapping test

Digit cancellation test

Delayed prose recall
Tests given post-drugWord-stem completion

Heroin craving questionnaire

Guess on treatment
Tests given once onlyNART

Drug use questionnaire

Mood rating scale. This 16 item visual analogue mood rating scale was used to assess subjective feelings at each test time. Principal components analysis of this scale yields three mood factors. alertness, contentedness and calmness (Bond & Lader, 1974). A 17th scale (anchored by "very stoned"-"not at all stoned") was added as a single item.

Prose recall. Participants listened to an audio-taped prose passage which they then recalled aloud immediately and then again after a delay filled by other tests. This task taps explicit memory. The four parallel versions used were from the Rivermead Behavioural Memory Test (Wilson, Cockburn & Baddlele, 1985) and scoring was standard (for each of 2] "idea units", 1 point is given for correct recall or exact synonym; half a point for partial recall or partial synonym).

Word-stem completion task. This task taps implicit memory and was administered post-drug on each test day. Participants were shown a series of 20 words presented one at a time and asked to rate these semantically (how much they liked each word on a five-point scale). They were then given a list of 32 word-stems to complete with "the first English word that came into their minds". Half of these stems could begin words previously presented; the other half were used to obtain base rates of completion. Two sets of words were used on each test occasion: half the subjects were presented one set and half thc other set (cf. Curran & Gorcnstcin, 1993).

Digit Symbol Substitution Test (DSST). This sub-test from the Wechsler Adult Intelligence Scale (Wechsler, 1955) requires speeded recording of digits and symbols for 90 seconds.

Digit cancellation. This paper-and-pencil test requires visual selectivity at speed and taps focused attention. The participant is asked to delete a target number in rows of random numbers as carefully and quickly as possible. Forty instances of the target number are randomly interspersed with 400 numbers. Time to completion and errors of omission or commission are scored.

Tapping rate. This was used as a measure of motor sedation (Frith, 1967). Participants are required to press the space bar of a computer keyboard with two fingers of their preferred hand as quickly as possible for 60 seconds.

Simple reaction time test. Participants press the space bar of a computer as soon as they see a stimulus appear on the screen. Twenty-four trials with random inter-trial intervals were given and responses and reaction times were recorded automatically.

Heroin craving. The Tiffany Heroin Craving Questionnaire was administered post-drug on each test day. The participant has to rate each of 45 items on a seven-point scale from "strongly disagree to strongly agree". Scoring was standard to yield five components of heroin craving: (1) desire to use (2) intention and plan to use (3) anticipation of positive outcome (4) relief from withdrawal and dysphoria and (5) lack of control over use.

Guess on treatment. At the end of each test day, participants were asked to indicate on a form whether they thought they had received (i) an extra dose of methadone (ii) an inert placebo or (iii) were unsure what they had received.

Demographic information. This included a full drug history and details of current use of psychotropics.

Statistical analyses. Repeated measures analysis of variance was employed with drug (methadone versus placebo) and test time (pre versus post-drug) as within subjects factors and treatment order (methadone or placebo first) as a between subjects factor. Pearson correlation’s were use to correlate scores on the five HCQ craving factors and to assess relationships between daily dose of methadone and variables showing significant treatment differences.

Table 2. Means (and standard deviations) on mood rating scale, pre- and post-treatments


Alert-Drowsy39.17 (16.64)42.30(14.85)40.90 (14.99)34.73 (11.68)
Contented-Discontented40.27 (17.91)37.59 (14.44)43.46 (14.65)31.80 (13.96)
Calm-anxious36.11 (23.33)29.47 (16.28)37.17 (19.99)29.64 (18.22)
Stoned78.78 (16.85)71.72 (18.27)72.12 (16.21)73.62 (23.80)

Table 3. Means (and standard deviations) of performance pre-and post-drug on cognitive and psychomotor test


Simple reaction time (msecs)317.2 (45.5)311.8(44.4)305.7 (60.3)315.7 (77.1)
Tapping speed (N taps)355.9 (36.5)344.4 (51.7)356.1(41.6)351.9 (70.7)
DSST (N substitutions)51.8 (10.4)54.1(10.4)52.1(11.9)51.3 (10.6)
Digit cancellation (secs)73.8 (17.9)71.3 (21.6)68.1(27.3)66.9 (27.7)
Prose recall immediate10.5 (3.0)7.0 (2.8)7.9 (2.1)7.2 (3.2)
Prose recall delayed6.7 (3.0)6.3 (2.6)6.7 (2.8)6.3 (2.9)


Demographic and drug use information

The 18 participants had a mean age of 37.5 (± 7.9) years, 14 were male, most (12) were unemployed, although five had unskilled jobs and one professional employment. In this, the sample was similar to most methadone patients (Department of Health, 1996). They were of slightly higher educational achievement than most patients (five had degrees, three had GCEs, two vocational qualifications and nine no qualifications). NART scores for the group gave an estimated IQ mean of 108 (± 11.95).

The average daily prescribed methadone dose was 43.5 (± 16.2) ml, range 20-80 ml, and therefore the mean 33% increase in dose administered was 14.5 ml (range 6.7-26.6 ml). This fairly low methadone dose is in keeping with the policy of the service in which the study was based where the goal is abstinence from illicit opiate use and eventual total detoxification. All subjects had been using methadone for a minimum of 6 months, as stipulated in the inclusion criteria, and had been on the same dose for at least 4 weeks before participating in the study.

Participants' reported illicit drug use was fairly typical of this treatment population; 72% of participants reported using at least one illicit drug in the previous month; 67% of participants reported using heroin in the week prior to testing. All but two participants reported frequent use of cannabis; amphetamine use was less frequent than cannabis or heroin. History of opiate use ranged from 2 to 5 years. One of the sample reported that he did not always take his full prescribed methadone dose as he was trying to cut down independently. Four participants reported using extra methadone on top of their prescribed dose. Five participants were taking diazepam prescribed for their benzodiazepine dependence.

Urine screening and breathalyser

Urinalysis indicated that 50% of participants had taken illicit opiates, three were positive for amphetamines, two for cocaine and three for cannabis. Five participants who had not reported illicit opiate use were positive for opiates, and two who had reported using were negative. Twenty-seven per cent of participants were negative for illicit drugs at both test screens, 17% were positive for the same drugs at each screen and 56% were positive for different drugs at each screen. Only one subject had detectable levels of breath alcohol and these were very low on both test days.

Table 4. Means (and standard deviations) of scores of the five heroin craving factors following treatments with methadone and placebo


HCQ factorPost-methadone  Post-placebo
Anticipation of positive outcome39.06 (8.79)32.78 (11.96)
Lack of control36.80 (6.56)33.67 (8.79)
Desire to use23.67 (6.75)18.50 (10.59)
Intention to use30.72 (7.28)27.06 (7.28)
Relief from withdrawal36.56 (6.89)30.56 (11.40)
Total craving166.81(22.34)142.57 (43.25)

Table 5. Correlation’s (and significance of correlation’s) between the five craving factors on the HCQ


 Anticipation of pos. outcomeLack controlDesireIntent
Lack control0.50 (.002)   
Desire0.53 (.001)0.45 (0.006)  
Intent0.50 (.002)0.49 (0.003)0.63 (0.001) 
Relief0.86 (.0001)0.48 (0.003)0.51(0.002)0.49 (0.003)

Mood rating scale. Principal components analysis yielded three mood factors. There was a main effect of drug treatment on alertness (mood factor 1) and contentedness (mood factor 2) (Table 2). Participants were more alert following placebo and slightly more drowsy following additional methadone (F1,16 = 10.0l, p<0.01). They were also more contented following placebo than following methadone (F1,16 = 8.81, p <0.01). No treatment differences emerged on mood factor 3 (calm-anxious) or on the single item "stoned".

Prose recall. There were no significant treatment effects (Table 3). Participants allocated to treatment order of methadone first had higher immediate recall scores pre-drug than those allocated to treatment order of placebo first (F1,15 = 4.7, p<0.05). Word-stem completion showed no treatment or order effects.

Simple reaction time tapping rate, DSST; digit cancellation. No significant treatment or treatment order effects emerged on any of these tasks.

Craving for heroin. Scores on the craving questionnaire were analysed as total score and as five factors (Table 4). Total craving was increased following methadone as compared with placebo (P1,16 = 5.64, p < 0.03). Significant main effects of drug were found on two craving factors: anticipation of positive outcome (F1,16 = 5.35, p <0.03) and relief from withdrawal or dysphoria (F1,16 = 5.0l,p<0.04). Scores on both were higher after methadone than after placebo, suggesting that an increase in methadone dosage increases craving for heroin. There was no significant difference found in either "lack of control" or "intention to use" factors. Desire to use heroin showed a marginally significant main effect of drug (p <0.06) with ratings again being higher post-methadone than post-placebo. Scores on desire to use showed a significant interaction of drug with treatment order (F1,16 = 6.2, p<0.02). This reflected higher scores on the first compared with the second test day for participants given methadone first, whereas those given placebo first showed little change. For the methadone first group, "desire to use" scores were 24.8 ± 7.6 after methadone and 13.3 ± 6.1 after placebo. Corresponding figures for the placebo first group were 22.6 ± 6.0 and 23.7 ± 11.9. It can also be seen from Table 4 that the variance in craving scores was lower following methadone than following placebo. This is the case with each of the five craving factors, though levene's test for equality of variance had shown that ANOVA assumptions had not been violated for any factor except "lack of control". Correlations between the five craving factors are shown in Table 5. There were significant correlations between all five craving factors. "Anticipation of positive outcome" and "relief from withdrawal" correlated most highly of all the craving factors (r =0.86), sharing 74% of the variance. "Intention" and "desire to use" also correlated fairly highly (r = 0.63), whereas all other factors inter-correlated at around 0.50.

Relationship between dose of methadone and treatment effects. Correlations were carried out between daily dose of methadone and change in mood factor scores (methadone (post-pre-drug)) minus placebo post-pre-rug). These were positive but small and non-significant (for alertness changes, Pearson's r = 0.28; for contentedness changes, r = 0.17). Correlations were also carried out between daily dose and craving factor scores (methadone ,minus placebo). These correlations were again small and non-significant (anticipation of positive outcome, r = 0.18; relief from withdrawal, r = 0.04; total HCQ score, r = 0.02).

Guess on treatment. There were a total of 36 treatment presentations (18 subjects attending two sessions). Participants' guess on treatment was at chance levels (correct in 39% of cases, incorrect in 36% of cases and in the remaining 25% of cases participants stated that they were unsure which treatment they had received). Participants therefore could not differentiate between methadone and placebo treatments. Thus double-blind procedures used, including the matching of placebo linctus with drug linctus, had been effective.