Clinical Trial Registration: Australian New Zealand Clinical Trials Registry—ACTRN12606000272594
Nicole Lee PhD, Associate Professor, Amy Pennay BA (Hons), Research Fellow, Robert Hester PhD, Senior Research Fellow, Rebecca McKetin PhD, Senior Research Fellow, Suzi Nielsen PhD, Senior Research Fellow, Jason Ferris MBioStats, Senior Statistician. Correspondence to Miss Amy Pennay, Turning Point Alcohol and Drug Centre, 54-62 Gertrude Street, Fitzroy, Vic. 3065, Australia. Tel: +61 3 8413 8413; Fax: +61 3 9416 3420; E-mail: email@example.com
Introduction and Aims
There are no medications approved for the treatment of methamphetamine withdrawal. Wake-promoting agent modafinil has recently been proposed as a viable option. This paper reports on the results of a pilot study that tested the feasibility of modafinil in an inpatient withdrawal setting during acute methamphetamine withdrawal.
Design and Methods
In a double-blind, randomised, placebo-controlled study, 19 methamphetamine dependent participants received modafinil (n = 9) or placebo (n = 10) daily for 7 days (200 mg for the first 5 days and 100 mg on days 6 and 7). Primary outcomes were retention in treatment and severity of withdrawal symptoms. Secondary outcomes were methamphetamine craving, sleep and physiological outcomes.
There were no significant differences between groups on retention in treatment, withdrawal severity, craving, sleep or physiological outcomes. There were no adverse events or side-effects reported.
Modafinil was found to be tolerable and well accepted by methamphetamine users and feasible for short-term inpatient withdrawal, but the sample was too small to detect treatment effects. Larger trials are needed to establish efficacy.
The United Nations Office on Drugs and Crime  estimates that there are up to 52 million people globally who use amphetamine-type stimulants at least once a year, a prevalence rate of 1.2%; more than twice that of opiates. In Australia, the rate is 2.3% . An estimated 56% of regular methamphetamine users are dependent on the drug, with dependence being more common with more frequent use, longer duration of use, use of high purity crystalline methamphetamine and among people who smoke or inject .
Despite large numbers of people who experience problems with methamphetamine, there is little in the way of effective treatments, and treatment completion for methamphetamine is one of the lowest of all drugs at around 50%, with another 50% likely to relapse following treatment completion [4, 5].
Engaging and retaining methamphetamine users in treatment can be difficult because of the nature of withdrawal and relapse . Symptoms of withdrawal such as dysphoria, insomnia or hypersomnia, increased appetite, psychomotor impairment and agitation  can reduce treatment compliance and increase the risk of relapse. Cravings are a critical factor in precipitating relapse, and are often intolerable, leading to cessation of treatment and re-uptake of stimulant use [8, 9]. Poor concentration, disturbed circadian rhythms and agitation that occur through the first week of withdrawal can reduce engagement in comprehensive treatment programs, such as behavioural treatments, which have shown some effectiveness for treating methamphetamine dependence . Medications that aim to address some of these symptoms may improve retention in withdrawal treatment.
The lack of an effective withdrawal medication has been identified by methamphetamine users as a substantial barrier to treatment [11, 12] and no medications have been approved for this purpose . Only two medications have been trialled for methamphetamine withdrawal in randomised controlled trial (RCT) conditions.
Cruickshank et al.  concluded that 12 days of mirtazapine, a tricyclic antidepressant, in an outpatient setting, was no more effective than placebo in reducing methamphetamine withdrawal symptoms; while over a decade ago, treatment with amineptine, an atypical tricyclic antidepressant, showed some improvements in mood and energy over 14 days of treatment, compared with placebo [15, 16]. Amineptine is no longer publicly available because of concerns about abuse liability . The only other study of withdrawal medication for methamphetamine was a non-randomised trial, which showed that up to 10 days of treatment with either modafinil or mirtazapine reduced the severity of withdrawal symptoms compared with treatment as usual (antipsychotic pericyazine), with modafinil being associated with the least withdrawal symptoms .
Two RCTs of cocaine dependence have shown that modafinil-treated patients are more likely to abstain from cocaine compared with those receiving placebo [19, 20]. Two recent trials showed no benefit of modafinil for methamphetamine dependence post withdrawal [21, 22], but no studies have examined its use during withdrawal using an RCT design. Previous research has shown that modafinil is efficacious in promoting wakefulness, decreasing food intake and does not significantly increase heart rate [23, 24]. As hypersomnolence and fatigue are prominent features of methamphetamine withdrawal, the wake-promoting properties of modafinil make it a strong candidate for withdrawal pharmacotherapy .
Modafinil stimulates central alpha-1 adrenergic receptors. There is some evidence to suggest that it regulates dopamine, adrenaline and noradrenaline . Modafinil may also enhance extracellular monoamine, glutamate and histamine levels, and decrease gamma-aminobutyric acid levels . Given that multiple neurotransmitter system are affected by stimulant use , pharmacotherapy targeting multiple neurotransmitter systems are likely to be a viable strategy for managing withdrawal.
The current pilot study tested the feasibility of modafinil to manage acute methamphetamine withdrawal in an inpatient setting over 7 days using a double-blind, randomised, placebo-controlled study design. We hypothesised that withdrawal symptoms would be reduced and retention rates increased in the modafinil group compared with placebo.
Twenty participants were recruited through the admissions of two inpatient clinical sites, advertising and word of mouth. The two study sites both were short-term (7-day stay) metropolitan residential withdrawal units.
Volunteers were eligible if they were between 18–65 years of age, met Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) criteria for methamphetamine dependence and had used methamphetamine in the 48 h before inpatient admission (confirmed by a saliva test: iscreenOFD). Participants were excluded if they were dependent on drugs other than methamphetamine, cannabis and nicotine, or had an active psychosis or major depression, significant medical conditions (e.g. liver cirrhosis, pneumonia, cardiac disease), a previous diagnosis of attention deficit hyperactivity disorder or narcolepsy, previous adverse reaction to modafinil, had been treated with modafinil or dexamphetamine in the past month, or were pregnant or breast-feeding.
Ethics approval for the project was granted by the Human Research Ethics Committee of Western Health, Victoria, and all procedures where in accordance with the standards of this committee.
Retention in treatment
Retention in treatment was calculated using number of full days as an inpatient.
Severity of withdrawal
Severity of withdrawal during the previous 24 h was measured each day (days 1–7) using the participant-administered Amphetamine Withdrawal Questionnaire (AWQ)  and Amphetamine Cessation Symptoms Assessment (ACSA) , and the clinician-administered Amphetamine Selective Severity Assessment (ASSA) [29, 30].
Craving for methamphetamine
Craving for methamphetamine during the previous 24 h was rated on a visual analogue scale. Participants placed a line between 1 and 100 in response to the questions: ‘Rate the highest intensity of the desire for methamphetamine you have felt in the last 24 h’ and ‘Identify how often you have felt the urge to use methamphetamine in the last 24 h’.
Sleep was measured using the St Mary's Hospital Sleep Questionnaire . Quality of sleep was measured using two items: ‘How well did you sleep last night?’ (a six-item Likert scale from 1—very bad to 6—very well) and ‘How satisfied were you with last night's sleep? (a five-item Likert scale from 1—very unsatisfied to 5—completely satisfied).
Physiological measures of blood pressure, heart and respiratory rate were taken daily.
Twenty dependent methamphetamine users were randomly assigned to either modafinil or placebo for 7 days (200 mg of modafinil or matching placebo capsules for days 1–5, and 100 mg for days 6 and 7, to titrate the dose before discharge). Modafinil (Modavigil®) and placebo matched for weight and appearance was prepared, randomised and dispensed by the study pharmacist. All investigators, doctors, nursing staff, researchers and participants were blind to medication assignment. Randomisations were created in blocks of four using a computerised randomisation program.
Potential participants were screened by a researcher to ensure they met the study eligibility criteria. Final eligibility for inclusion in the trial, including dependence on methamphetamine and absence of dependence on other substances, was assessed by the study doctors at the inpatient sites. A saliva test confirming methamphetamine consumption in the previous 48 h (and absence of other drugs except cannabis) was undertaken on day 1. Following the tests, participants were given their first dose of study medication.
Medication was administered each morning to reduce potential sleep disturbance. Daily measures of withdrawal and physiology were completed every afternoon.
It was considered that 20 participants would be sufficient to determine trial feasibility and the trial was deemed complete upon commencement of the 20th participant. One participant was later found to be ineligible and was removed from the study with that participants’ data excluded from analysis, leaving nine participants in the modafinil group and 10 in the placebo group.
Analysis was undertaken using stata version 10.0 (StataCorp LP, College Station, TX, USA). Retention was compared at the end of treatment using a Pearson's χ2 test, and retention across the 7-day period using a log–rank equality of survivor test. Data from the ASSA, ASCA and AWQ were highly skewed, and therefore medians and interquartile ranges are presented. Group comparisons were made using the Kruskal–Wallis test, and trends over time were examined using a non-parametric trend test . Missing data as the result of loss to follow up are assumed to be missing-at-random  and disregarded in the analysis.
Of the 20 participants recruited to the study, 13 were male, their average age was 34.3 (range 21–48), most were born in Australia (n = 16), unmarried (n = 16) and living in rental property or boarding houses (n = 13). Only one participant had higher than secondary school education and six were employed. Participants had been using methamphetamine for an average of 6.5 years (SD 6.04). The sample had high rates of current polydrug use, including benzodiazepines (n = 14), cannabis (n = 12), alcohol (n = 12), ecstasy (n = 6) and heroin (n = 2). This profile is typical of methamphetamine users accessing treatment in Australia .
Participants in each treatment arm were similar aside from gender and drug use (Table 1). The modafinil group had been using methamphetamine for an average of 3 years longer than the placebo group, and had also used more cannabis and alcohol in the month before the trial.
Table 1. Descriptive statistics of demographics and drug use by treatment arm
Placebo (n = 10)
Modafinil (n = 9)
aIncludes: married, divorced, separated and widowed.
The screening to randomisation ratio was 2.7:1 (see Figure 1). Five participants were ineligible because of polydrug dependence or medical reasons. Of the 25 people that were screened but were not enrolled in the study, 11 could not be contacted after screening, four had become abstinent from methamphetamine before the trial, three declined inpatient withdrawal (due to work or caring for children), three were seeking long-term daily maintenance treatment, two were seeking long-term inpatient rehabilitation treatment, one was seeking medicated assisted counselling and one wanted non-medicated withdrawal treatment.
Retention in treatment
Eleven of the 19 participants remained in treatment for the full 7 days. Sixty-seven per cent (n = 6) of the modafinil group and 50% (n = 5) of the placebo group completed treatment, although the difference was not significant (χ2(d.f.=1) = 0.5, P = 0.463). There was no difference between the groups in retention over the seven-day period (log–rank test for equality of survivor function, χ2(d.f.=1) = 0.35, P = 0.553).
Of the three people who discontinued treatment before day 7 in the modafinil group, two were due to family issues and one to an aggressive incident with staff. In the placebo group, three left to use methamphetamine, one left feeling better and one left to pursue counselling.
Men were more likely to complete treatment than women (with 9 of 13 men completing treatment but only 2 of 6 women); however, age did not differ between completers (mean age 34 years) and non-completers (mean age 33 years), nor did polysubstance use at baseline [at baseline 74% of participants had used benzodiazepines and 63% had used alcohol and cannabis in the past month, and of completers, 8 of 11 (73%) participants indicated using these three substances at baseline]. Years of methamphetamine use did not predict treatment completion; however, those who completed treatment in the modafinil arm had a slightly longer history of methamphetamine use (8.5 years) than those in the modafinil arm who did not complete treatment (7.9 years). A reverse pattern was seen for the placebo arm (5 years at baseline and 3.2 years at treatment completion).
The reduction in withdrawal symptoms over the 7-day period was statistically significant for both treatment conditions across all three measures (P <0.01). For each day in treatment there was a drop in each of the withdrawal measures: for AWQ the slope (or coefficient) was −1.41, for ASSA it was −4.70 and for ACSA it was −2.41 (Figure 2). However, there were no significant differences between the two groups on any of the withdrawal measures (P > 0.05).
A significant decreasing trend in cravings for both intensity and frequency for each treatment arm was observed (P <0.01); yet, the difference between the two treatment arms for both craving measures was not significant (P > 0.05) (Figure 3).
At day 1, eight of the 19 participants indicated that they slept ‘very badly’ the previous night, and only six indicated sleeping ‘fairly well’ or better. At day 3, only three of the remaining 15 participants indicated that their sleep was ‘fairly bad’ or worse. However, the non-parametric test for trends was not significant (P > 0.05); nor was the differences between the two treatment arms (P > 0.05). At day 1, nine of the 19 participants were ‘very unsatisfied’ with the previous night's sleep and 14 indicated that their sleep was ‘slightly unsatisfying’ or worse. At day 3, seven of the remaining 15 participants indicated that their sleep was ‘slightly unsatisfying’ or worse (Figure 4). Both the test for trends and the comparison of responses between treatment arms were not significant.
Over the study period the average systolic pressure ranged between 116 and 124 mmHg (average SD 15 mmHg) and the average diastolic pressure ranged between 76 and 83 mmHg (average SD 16 mmHg) with no discernible variation over the 7 days, nor between treatment arms. The average heart rate (beats per minute) ranged from 79 to 87 across the 7-day period (average SD 10), this being slightly elevated during the first 3 days (mean 84; SD 10) compared with the remaining 4 days (mean 80; SD 8); however, this difference was not significant, nor was there a significant difference across time (Z = −1.34; P = 0.182) or between treatment groups (Z = 1.06; P = 0.288). The average respiratory rate (breaths per minute) over the 7-day period ranged from 15 to 18 (average SD 3), with no significant trend across time (Z = −1.33; P = 0.184) or between groups (Z = 0.55; P = 0.580). There were no adverse events and no side-effects reported other than both lethargy and disturbed sleep, which were reported by both groups.
This is the first RCT to examine the feasibility of modafinil during acute methamphetamine withdrawal. Modafinil was found to be well accepted by methamphetamine users and feasible for short-term inpatient withdrawal. There were no adverse effects on physiological measures such as blood pressure, heart or respiratory rate, supporting the safety and tolerability of modafinil during acute methamphetamine withdrawal. In addition, our ratio from screening to randomisation (2.7:1) indicates that recruitment to a larger study is feasible, and retention was higher in our study than has previously been found among methamphetamine users [4, 5], further supporting the feasibility of a larger study.
In terms of feasibility, it is important to note that detoxification is not suitable for all methamphetamine users; indeed, of the 25 people who were screened and declined involvement in the study, six were seeking longer term treatment options and three could not manage time off work or time off caring for children. In particular, the short-term nature of the medication (as opposed to daily maintenance) was a concern for some potential participants. While we suggest the current study be replicated in larger samples as a first step, future research might consider investigating whether daily maintenance treatment following medicated withdrawal treatment has any benefit for methamphetamine users.
It is important to note that subjective and objective ratings of withdrawal declined significantly among both groups (P <0.01), suggesting that withdrawal treatment, regardless of assisted pharmacotherapy, is likely to benefit methamphetamine users. In particular, in this study, withdrawal symptoms were resolved more quickly than in the previous withdrawal study of McGregor et al. . However, this finding may also reflect the lack of specificity of detoxification treatment. It has been proposed that drug use responds to treatment irrespective of what type of treatment is provided because treatment models do not always operate via their postulated theoretical underpinnings . This improvement may also be due to natural remission from methamphetamine withdrawal symptoms and/or selective attrition of participants who do not respond well to treatment.
In this pilot study we did not control for the range of mediating factors that may have influenced the results. For example, the participants’ drug-use histories showed that at baseline, the modafinil group had a greater duration (8 vs. 5 years) of methamphetamine use, as well as higher recent levels of cannabis and alcohol. These factors might indicate a higher severity of illness in the modafinil treated group, which has the potential to obscure any medication effect and bias outcomes towards a null result. If further research is conducted in larger samples, it will be important to control for such factors by stratifying the randomisation.
This study has a number of limitations. First the sample size was too small to detect significance. Second, despite dependence on drugs other than cannabis being an exclusion criterion for the study, there was a high frequency of cannabis, prescribed benzodiazepine and alcohol use in the sample. This may have potentially resulted in greater illness severity and it is possible that some of these subjects experienced withdrawal symptoms from other substances. A key strength of this study design is that medication compliance was assured and it is unlikely that illicit drugs were used during the withdrawal phase because of the inpatient nature of treatment.
Modafinil at 200 mg was feasible and tolerable for use within an inpatient withdrawal setting for dependent methamphetamine users. Given the promising results found in previous research  and the feasibility determined in this study, future research is warranted in larger samples to measure the clinical efficacy of modafinil as a medication for acute methamphetamine withdrawal. Lessons that can be taken from this pilot study include the need to: (i) ensure that groups are matched on demographics and drug use history; (ii) investigate whether modafinil reduces some withdrawal symptoms and not others; and (iii) consider the potentially confounding impact of withdrawal from other substances. Finally, there is a need to establish the effect of different doses of modafinil on methamphetamine withdrawal and whether this is affected by client characteristics, such as severity of methamphetamine dependence.
This research was supported by the Illicit Drugs Section, Drugs Strategy Branch, Australian Government Department of Health and Ageing. We gratefully acknowledge the clinical staff at Moreland Hall and the Drug and Alcohol Services of the Western Hospital who assisted with data collection. We thank Dr Adrian Dunlop for his contribution to the medical aspects of the trial in the project's developmental stage and Peter Muhleisen for his advice and assistance with the medication preparation.