Articles

Sustained-release dexamfetamine in the treatment of chronic

cocaine-dependent patients on heroin-assisted treatment:
a randomised, double-blind, placebo-controlled trial
Mascha Nuijten, Peter Blanken, Ben van de Wetering, Bastiaan Nuijen, Wim van den Brink, Vincent M Hendriks

Summary
Background Heroin-assisted treatment is eective for methadone treatment-refractory heroin-dependent patients, but
continued comorbid cocaine dependence remains problematic. Sustained-release dexamfetamine is a promising
agonist pharmacotherapy for cocaine dependence and we aimed to assess its acceptance, ecacy, and safety.
Methods In this multicentre, randomised, double-blind, placebo-controlled trial, patients who were treatmentrefractory, as indicated by at least two earlier failed treatments aimed at reducing or abstaining from cocaine use, and
who regularly (8 days/month) used crack-cocaine were enrolled from four heroin-assisted treatment centres in the
Netherlands. Eligible patients were randomly assigned (1:1) to receive either 12 weeks of daily, supervised prescription
of 60 mg/day oral sustained-release dexamfetamine or placebo in addition to co-prescribed methadone and
diacetylmorphine. Randomisation was done by the collaborating pharmacist, using a computer-generated random
number sequence with stratication by treatment centre in blocks of four per stratum. Randomisation was masked to
patients, sta, and researchers throughout the study. The primary outcome was the number of self-reported days of
cocaine use during study treatment, assessed every 4 weeks. Primary and safety analyses were done in the intentionto-treat population. The study was registered with the European Union Drug Regulating Authorities Clinical Trials
(EUdraCT 2013-004024-11) and with The Netherlands Trial Register (NTR2576).
Findings Between Aug 8, 2014, and Feb 27, 2015, 111 patients were assessed for eligibility, of whom 73 were enrolled
and randomised; 38 patients were assigned to the sustained-release dexamfetamine group and 35 to the placebo group.
Sustained-release dexamfetamine treatment resulted in signicantly fewer days of cocaine use than placebo treatment
(mean 449 days [SD 294] vs 606 days [243], respectively [95% CI of dierence 31284]; p=0031; Cohens
standardised eect size d=058). One or more adverse events were reported by 28 (74%) patients in the dexamfetamine
group and by 16 (46%) patients in the placebo group. Most adverse events were transient and well-tolerated.
Interpretation Sustained-release dexamfetamine is a well accepted, eective, and safe agonist pharmacotherapy for
comorbid treatment-refractory cocaine dependence in heroin-dependent patients in heroin-assisted treatment. Future
research should aim to replicate these ndings in chronic cocaine-dependent and other stimulant-dependent patients
in more routine treatment settings, including strategies to optimise treatment adherence like medication management
interventions and contingency management.
Funding Netherlands Organisation for Health Research and Development.

Introduction
Heroin-assisted treatment is an eective treatment
for methadone treatment-refractory heroin-dependent
patients, resulting in reduced illicit heroin use and
improvements in mental status, physical health, and
social functioning, as has been shown in seven
randomised controlled trials1 and two cohort studies.2,3
However, many heroin-dependent patients are also
cocaine-dependent, which worsens the prognosis of
treatment,4 as is also shown among patients in heroinassisted treatment, who often show no or only slight
reductions in cocaine use.3 Agonist pharmacotherapy for
chronic cocaine dependence among patients in opioid
agonist treatment might be a viable strategy. However, a
recent randomised placebo-controlled trial with
immediate-release methylphenidate (30 mg twice daily)

in cocaine-dependent patients currently in heroinassisted treatment did not show benets in terms of
reduced cocaine use.5
Reviews of substitution treatments for cocaine
dependence, including psychostimulants and (other)
dopamine agonists,6,7 suggest that sustained-release
dexamfetamine is probably the most promising agonist
drug with respect to reductions in cocaine use and
craving, but previous studies were restricted by low
adherence, and cocaine-related outcomes often did not
reach statistical signicance.810
We aimed to assess the acceptance, ecacy, and safety
of a robust dose of 60 mg/day oral sustained-release
dexamfetamine in chronic crack-cocaine-dependent
patients with comorbid heroin dependence, currently on
heroin-assisted treatment.

www.thelancet.com Published online March 22, 2016 http://dx.doi.org/10.1016/S0140-6736(16)00205-1

Published Online
March 22, 2016
http://dx.doi.org/10.1016/
S0140-6736(16)00205-1
See Online/Comment
http://dx.doi.org/10.1016/
S0140-6736(16)00563-8
Parnassia Addiction Research
Centre (PARC, Brijder Addiction
Treatment), The Hague,
Netherlands (M Nuijten MSc,
P Blanken PhD,
Prof V Hendriks PhD); Antes
Group, Bouman, Rotterdam,
Netherlands
(B van de Wetering PhD);
Antoni van Leeuwenhoek/The
Netherlands Cancer Institute,
Medical Centre Slotervaart,
Amsterdam, Netherlands
(B Nuijen PhD); Amsterdam
Institute for Addiction
Research, Department of
Psychiatry, Academic Medical
Centre, University of
Amsterdam, Amsterdam,
Netherlands
(Prof W van den Brink PhD); and
Curium, Leiden University
Medical Centre, Department of
Child and Adolescent
Psychiatry, Leiden University,
Leiden, Netherlands
(Prof V Hendriks PhD)
Correspondence to:
Mascha Nuijten, Parnassia
Addiction Research Centre
(PARC, Brijder Addiction
Treatment), Monsterseweg 83,
2553 RJ The Hague, Netherlands
mascha.nuijten@brijder.nl

Articles

Research in context
Evidence before this study
Our reference point was the Cochrane review (Castells et al
[2010]), based on 16 randomised parallel group
placebo-controlled clinical trials (RCTs) on the ecacy and safety
of stimulant medications (bupropion, dexamfetamine,
methylphenidate, modanil, mazindol, methamfetamine, and
selegiline) for the treatment of cocaine use disorders until July
24, 2008. As a group, these stimulants did not reduce cocaine
use. When type of medication was included in the analysis, the
proportion of patients achieving sustained cocaine abstinence
was higher with bupropion (three RCTs) and dexamfetamine
(three RCTs) than with placebo. The authors concluded that the
evidence for stimulants in the treatment of cocaine dependence
was inconclusive, but also that promising results existed for
dexamfetamine and bupropion.
We searched MEDLINE, Embase, PsycINFO, and the Cochrane
Central Register of Controlled Trials for clinical trialspublished
between July 25, 2008 and Nov 1, 2013on the ecacy of
dexamfetamine in the treatment of cocaine dependence, using
the same search terms as Castells et al in 2010. Restricting our
results to RCTs on the ecacy of dexamfetamine among
treatment seeking cocaine-dependent patients in terms of
clinical (cocaine use) outcomes, we retrieved two potentially
relevant articles. One study (n=81) tested a combination of
mixed amphetamine salts and topiramate (Mariani et al [2012]),
making it impossible to know the contribution of

Methods
Study design and participants

For the study protocol see

https://www.brijder.nl/study_
protocol_sr_dexamphetamine

This multicentre, randomised, double-blind, placebocontrolled trial was part of a larger project testing three
pharmacological drugs (topiramate, modanil, and
sustained-release dexamfetamine) in separate studies in
crack-cocaine-dependent patients in the Netherlands.11
Study participants were recruited from the population
of patients currently receiving oral methadone plus
inhalable or injectable diacetylmorphine for their
concurrent heroin dependence in supervised heroinassisted treatment programmes in two treatment centres
in Amsterdam, one in Rotterdam, and one in The Hague.
Eligible patients: (1) met inclusion criteria for heroinassisted treatment, including minimum age of 25 years,
methadone treatment-refractory heroin dependence,
(nearly) daily heroin use, and poor physical, mental or
social functioning (for full details, see van den Brink and
colleagues);12 (2) met cocaine dependence criteria
according to the Diagnostic and Statistical Manual of
Mental Disorders IV edition13 in the past year and
previous 5 years; (3) used cocaine on at least 8 days in the
previous month; (4) administered cocaine primarily by
means of basing (also known as freebasing and means
smoking crack-cocaine); (5) had at least two earlier failed
treatments aimed to reduce or abstain from cocaine use

dexamfetamine to the eect. The second study (n=73) compared

the eects of dexamfetamine, modanil, and the combination of
dexamfetamine plus modanil with placebo (Schmitz et al
[2012]). Modanil and the combination of modanil plus
dexamfetamine were associated with increased cocaine use and
dexamfetamine alone did not clearly separate from placebo in
terms of cocaine use.
Added value of this study
Previous studies on the eect of dexamfetamine in
cocaine-dependent patients were promising, but often restricted
by small sample size, high treatment dropout and, consequently,
cocaine use-related outcomes did not reach statistical signicance.
Our study on sustained-release dexamfetamine in comorbid
cocaine- and heroin-dependent patients, participating in
heroin-assisted treatment, oered a context in which medication
adherence could be optimised, allowing us to assess the real
potential of sustained-release dexamfetamine in the treatment of
cocaine dependence.
Implications of all the available evidence
Sustained-release dexamfetamine is an eective and safe
agonist medication for the treatment of patients with cocaine
dependence when medication adherence can be established.
Replication of these ndings in treatment-refractory
cocaine-dependent and other stimulant-dependent patients in
routine, optimised treatment settings, with clinical measures to
enhance medication adherence, is warranted.

(treatment-refractory); (6) were able and willing to

participate in the 12-week study; and (7) provided written
informed consent.
Patients were excluded in case of (1) severe medical
problems (eg, electrocardiography or blood abnormalities) or severe psychiatric problems (eg, acute
psychosis or suicidality); (2) pregnancy or breastfeeding;
(3) pharmacotherapy with a potentially eective drug for
cocaine dependence (ie, disulram, acamprosate,
methylphenidate, modanil, topiramate, immediaterelease dexamfetamine, or baclofen); (4) insucient
command of the Dutch language; and (5) current
participation in another addiction treatment trial.
The study was approved by the medical ethics
committee of the Academic Medical Centre of the
University of Amsterdam. The study protocol is available
online.

Randomisation and masking

Following screening and baseline assessment, eligibility
was determined by the treatment physician, and eligible
patients were randomly assigned (1:1) to receive either
12 weeks oral sustained-release dexamfetamine or
identical placebo along with continued heroin-assisted
treatment. Randomisation was conducted by the
collaborating pharmacist, using a computer-generated

www.thelancet.com Published online March 22, 2016 http://dx.doi.org/10.1016/S0140-6736(16)00205-1

Articles

random number sequence with stratication by treatment

centre (four centres) in blocks of four per stratum.
Treatment packs with sucient study medication or
placebo for the 12-week period were numbered
sequentially and dispensed by the pharmacist to eligible
patients in order of study entry. Randomisation was
concealed for patients, sta, and researchers throughout
the study.

Procedures
All patients were oered pharmaceutical-grade diacetylmorphine (maximum single dose 400 mg; maximum daily
dose 1000 mg) 3 times per day and 7 days per week in
designated treatment centres, along with once daily oral
methadone (maximum dose 150 mg). Methadone was coprescribed to achieve a stable base of opioid plasma
concentrations and to prevent withdrawal symptoms in
case patients missed a visit at the heroin-assisted treatment
centre for supervised use of diacetylmorphine.
The study treatment consisted of either ongoing heroinassisted treatment along with 12 weeks of treatment with
sustained-release dexamfetamine, prescribed in a robust,
single oral dose of 60 mg/day (2 tablets of 30 mg) in the
experimental group or ongoing heroin-assisted treatment
along with 12 weeks of identical placebo (2 tablets of
30 mg) in the placebo group. Study medication was
dispensed once daily during the patients morning visit at
the heroin-assisted treatment centre, and had to be taken
under supervision to allow intensive safety monitoring.
Study assessments were done at baseline, and at weeks
4, 8, and 12. At baseline, the Composite International
Diagnostic Interview Substance Abuse Module (cocaine
and alcohol dependence)14 and the Mini-International
Neuropsychiatric Interview on suicide risk15 were
undertaken. At all assessments we administered the
substance use section of the Addiction Severity Index,
supplemented with questions about illegal activities;12,16
the Time Line Follow-Back on self-reported cocaine use;17
the Obsessive Compulsive Drug Use Scale on past week
cocaine craving;18 the Maudsley Addiction Prole Health
Symptoms Scale (MAP-HSS) on physical health;19 and the
Brief Symptom Inventory (BSI) on mental health.20 In the
nal 4 study weeks, urine samples were collected (nonsupervised) twice weekly, on Mondays and Thursdays.
Samples were analysed for the presence of the cocainemetabolite benzoylecgonine (>300 ng/mL), using
qualitative rapid tests (nal von minden GmbH, Moers,
Germany). The urine tests had a sensitivity of 95% and a
specicity of 90%, and had no cross-reactivity with
dexamfetamine sulphate. Additional assessments
included blood sampling and electrocardiography
(screening and week 12 assessment); weekly medical
monitoring of heart rate, blood pressure, and
bodyweight; weekly standardised registration of (serious)
adverse events and co-medication; monthly pregnancy
testing; daily registration of supervised medication
adherence; and at week 12 the Client Satisfaction

Questionnaire, supplemented with a question to rate the

study medication on a scale ranging from 0 (very bad) to
10 (excellent).21 Participants received a maximum
remuneration of 85 for participating in the study
assessments.

Outcomes
The primary outcome was the number of
self-reported days of cocaine use during the 12-week
study (range 084 days) and was centrally assessed.
Secondary cocaine use-related outcomes were number of
cocaine-negative urine samples in the last 4 study weeks,
and the following TimeLine FollowBack-based outcomes:
longest period of consecutive cocaine abstinence;
percentage of patients with at least 21 consecutive days of
cocaine abstinence; days of cocaine abstinence during
the last 4 study weeks; and changes in so-called cocaine
hits (ie, cocaine self-administrations on days patients
used cocaine) and changes in days of cocaine abstinence
comparing the 4 weeks preceding the baseline and week
12 assessment.
Other secondary outcomes were changes in cocaine
craving, (self-reported) use of other substances, physical
and mental health, criminality, as well as medication
adherence, and safety (ie, [serious] adverse events) during
the 12-week study. Safety was assessed in terms of the
number of patients that reported at least one (serious)
adverse event, the number of (serious) adverse events,
and by electrocardiography and monitoring of heart rate,
blood pressure, and bodyweight.

Statistical analysis
For the power analysis, the mean dierence between
the sustained-release dexamfetamine group and the
placebo group in number of days of cocaine use during
the 12-week treatment period was estimated to be 10,
with a pooled SD of 17 days (d=059; ie, moderate eect
size). For this proof-of-principle study, a lenient alpha
of 010 was chosen to minimise the risk of a false
negative outcome (type 2 error).11 With a two-sided
alpha of 010 and power of 080, 36 patients were
required per study group.
An intention-to-treat approach, including all
randomised patients, was used to test group dierences
in all primary, secondary, and safety analyses. This
denition is more strict than the one in the study
protocol, which additionally required that patients took at
least one dose of the study drug.
The primary outcomeie, number of self-reported days
of cocaine use during the 12-week studywas analysed
with negative binomial regression analyses with
treatment group as the only independent variable and
the interaction of treatment group with treatment centre
as the only eect modier. To t the negative binomial
regression model, a reection transformation was done
on the negatively skewed data of the primary outcome
(ie, 84 days minus cocaine use days).

www.thelancet.com Published online March 22, 2016 http://dx.doi.org/10.1016/S0140-6736(16)00205-1

Articles

Concerning the secondary cocaine use-related outcomes,

dierent statistical analysis strategies were used based on
the nature and distribution of the outcome. Negative
binomial regression analyses with treatment group as the
only independent variable were used for the longest
period of consecutive cocaine abstinence and the mean
number of cocaine metabolite-free urine samples in the
4 weeks preceding the week 12 assessment. Achievement

111 patients assessed for eligibility

38 ineligible
73 enrolled

73 randomised

38 assigned to sustained-release
dexamfetamine

35 assigned to placebo

4 discontinued treatment
2 imprisonment
2 adverse events

4 discontinued treatment
1 imprisonment
2 adverse events
1 no experienced treatment eect

34 completed treatment

31 completed treatment

38 included in intention-to-treat analysis

35 included in intention-to-treat analysis

Figure: Trial prole

Sustained-release
dexamfetamine
group (n=38)

Placebo group
(n=35)

Demographic background
Age (years)

484 (66)

Men

35 (92%)

490 (53)
31 (89%)

European descent

26 (68%)

23 (66%)

Substance use
Lifetime regular crack-cocaine use (years)

191 (77)

199 (71)

Cocaine-positive baseline urine

38 (100%)

34 (97%)

Cocaine use days (past month)

235 (76)

237 (76)

Lifetime regular heroin use (years)

211 (84)

230 (85)

Heavy (5 units per day) alcohol use (1 day, past month)

13 (34%)

12 (34%)

Cannabis use (1 day, past month)

21 (55%)

13 (37%)

Treatment status and treatment history

Time in heroin-assisted treatment (months)
Medical heroin dose (mg)
Methadone dose (mg)
Previous addiction treatments
Data are mean (SD) or n (%).

Table 1: Baseline characteristics

462 (343)

575 (351)

5822 (2008)

6350 (1883)

676 (281)

701 (243)

62 (32)

84 (72)

of cocaine abstinence for at least 21 consecutive days was

analysed by logistic regression analysis, using treatment
group as the only independent variable. Group dierences
in changes in number of days of cocaine abstinence and
cocaine hits in the 4 weeks preceding baseline and week
12 were analysed by multilevel analyses (generalised linear
mixed models) with a random intercept, and with the two
assessments and treatment group as xed eects.
Multilevel analyses were used, instead of repeated
measures analyses of variance mentioned in the study
protocol, to t the non-normal distribution of the data.
Cohens d eect sizes were calculated for continuous
outcomes and numbers needed to treat (NNT) for
dichotomous outcomes.
The other secondary outcomesie, changes in craving,
use of other substances, health status, and criminality
were analysed by generalised estimating equation models
with treatment group, assessment (baseline, weeks 4, 8,
and 12) and the interaction between treatment group and
assessment as independent variables, and using an
unstructured correlation matrix. Except craving, all
secondary outcomes were non-normally distributed and,
therefore, dichotomised based on the presence or
absence of past month illicit heroin use, or cannabis use,
or heavy (5 units per day) alcohol use (all 1 day); poor
physical health (MAP-HSS total score 8); poor mental
health (BSI total score 056 for men and 071 for
women); and past month criminality (1 day).3
Because the medication was dispensed daily and intake
was supervised, adherence with the study medication
was registered on a daily basis. Dierences between the
study groups and treatment centres were described and
analysed by means of negative binomial regression
analyses in terms of the number of days of medication
intake during the 12-week study, the number of
consecutive weeks in which patients were fully compliant,
and days of medication intake in the nal 4 weeks.
Data for one patient in the dexamfetamine group was
missing from week 4 onward due to imprisonment, and,
following the most conservative strategy, all missing
TimeLine FollowBack-days were considered as cocaine
use days. Furthermore, 516 (88%) of the 584 scheduled
urine samples were submitted, and the remaining
68 missing urine samples were considered cocainepositive. Agreement between self-reported crack-cocaine
use in the 3 days before the last urine sample (week 12)
and a cocaine-positive urine was 892% with a Kappavalue of 064 in both study groups; almost 50% of the
patients with no self-reported crack-cocaine use (n=15)
did have a cocaine-positive urine.
Data monitoring was conducted by the investigators
and the independent supervisory pharmacist (Amsterdam
Academic Medical Centre); there was no independent
data monitoring committee.
Data were analysed with SSPS (version 23).
The study was registered with the European Union
Drug Regulating Authorities Clinical Trials (EUdraCT

www.thelancet.com Published online March 22, 2016 http://dx.doi.org/10.1016/S0140-6736(16)00205-1

Articles

2013-004024-11) and with The Netherlands Trial Register

(NTR2576).

Role of the funding source

The funder of the study had no role in the study design,
data collection, data analysis, data interpretation, or
writing of the report. The corresponding author had full
access to all study data and had nal responsibility for
the decision to submit for publication.

Results
Between Aug 8, 2014, and Feb 27, 2015, 111 patients were
assessed for eligibility, of whom 73 were enrolled;
38 patients were randomly assigned to sustained-release
dexamfetamine and 35 to placebo (gure). Patient
recruitment was terminated when the aimed number of
patients according to the power calculation was achieved.
Patients were mainly men from European descent, on
average 49 years old (SD 6), with a long history of regular
illicit heroin and cocaine use, who had multiple previous
treatments, and who had used cocaine on an average of
24 days (SD 8) in the past month (table 1). Patients
participated in heroin-assisted treatment for on average
4 years (SD 3). One patient injected cocaine; all others
smoked crack-cocaine. Baseline characteristics were
balanced between the two treatment groups.
Analysis of the primary outcome showed that the mean
number of self-reported days of cocaine use in the 84 days
treatment period was signicantly lower in the
dexamfetamine group than in the placebo group
(449 days [SD 294]) vs 606 days [243], respectively
[95% CI of dierence 31284 days]; Wald =466, df=1;
p=0031) (table 2). There was no signicant interaction
between treatment centre and treatment group
(Wald =202, df=3, p=0569).
With regards to secondary cocaine use-related outcomes,
the longest consecutive period of self-reported cocaine
abstinence was signicantly higher in the dexamfetamine
group than in the placebo group (Wald =1617, df=1,
p<00001; table 2). Similarly, patients in the
dexamfetamine group were more often abstinent from

cocaine for at least 3 consecutive weeks than those in the

placebo group (Wald =552, df=1, p=0019), and reported
more days of cocaine abstinence in the nal 4 weeks of the
study (Wald =845, df=1, p=0004; table 2). Eight patients
(21%) in the dexamfetamine group had at least one
cocaine-negative urine in the last 4 weeks compared with
two patients (6%) in the placebo group, with a signicantly
higher proportion of cocaine-negative urine samples in
the dexamfetamine group (Wald =511, df=1, p=0024).
Additionally, the average number of days of cocaine
abstinence in the 4 weeks preceding baseline compared
with the 4 weeks preceding week 12 increased signicantly
more in the dexamfetamine than in the placebo group
(65 days [SD 69] to 152 days [108] days vs 54 days [72]
to 75 [91], respectively [treatment by time interaction:
F=470; df=1; p=0032; d=094]). Moreover, patients in the
dexamfetamine group showed higher reductions in
the mean number of cocaine hits than did patients in the
placebo group on days they used cocaine (89 cocaine hits
per day [SD 59] to 51 [44] vs 83 cocaine hits per day
[44] to 77 [59], respectively [treatment by time
interaction: F=745; df=1; p=0007; d=059]).
With respect to the other secondary outcomescocaine
craving, use of other substances, health, and criminality
we noted signicant changes from baseline to week 12 for
cocaine craving, heavy alcohol use, and physical health
problems, but no signicant group dierences over time
on any of these variables (all p0098; table 3). Finally,
patients in the dexamfetamine group rated the study
medication at week 12 on average more positively than
patients in the placebo group (76 [SD 14] vs 57 [23],
respectively; t=427, df=551, p<00001).
At the week-12 assessment, the study blind was tested.
In the dexamfetamine group, 54% of the patients correctly
identied their group allocation compared with 60% in
the placebo group (Kappa=014), indicating that blinding
was successful until the end of the study and that patients
were not able to discern beyond chance what they had
been prescribed. Study medication was taken on a mean
of 77 (SD 152) of the 84 study days (92%), with no
dierence between the study groups (75 days [SD 169] in

Sustained-release
dexamfetamine
group (n=38)

Placebo group Exp(B) (95% CI)

(n=35)

Wald
(df=1)

p value

Eect size

0031

d=058

Primary outcome
Days of cocaine use during 12-week study

449 (294)

606 (243)

167 (105267)

466

Longest period of consecutive cocaine abstinence (days)

179 (249)

Consecutive cocaine abstinence for 21 days

11 (29%)

67 (117)

269 (166436)

1617

<00001

2 (6%)

672 (1373297)

552

0019

Secondary cocaine use-related outcomes

d=058
NNT=43

Days of cocaine abstinence in nal 4 weeks

152 (108)

75 (91)

204 (126331)

845

0004

d=077

Proportion cocaine-negative urine samples in nal 4 weeks

106 (251)

39 (179)

260 (114594)

511

0024

d=031

Data are mean (SD) or n (%), unless otherwise specied. Exp(B)=exponentiated value of regression coecient B; odds ratio. df=degrees of freedom. d=Cohens d, which is a
standardised eect size. NNT=number needed to treat.

Table 2: Primary and secondary cocaine use-related outcomes

www.thelancet.com Published online March 22, 2016 http://dx.doi.org/10.1016/S0140-6736(16)00205-1

Articles

Estimated marginal means*

Baseline

GEE parameters (baselineweek 12)*

Week 4

Week 8

Week 12

Time

Group

Group x time

Wald=5236;
p<0001

Wald=652;
p=0011

Wald=458;
p=0205

Cocaine craving (range 020)

Sustained-release dexamfetamine group
(n=38)

874

600

580

511

Placebo group (n=35)

980

813

706

729

..

..

..

Illicit (non-prescribed) heroin use (1 day past month)

Sustained-release dexamfetamine group
(n=38)

021

024

019

024

Placebo group (n=35)

034

034

026

031

Wald=276;
p=0431
..

Wald=144;
p=0230
..

Wald=022;
p=0975
..

Heavy (5 units per day) alcohol use (1 day past month)

Sustained-release dexamfetamine group
(n=38)

034

024

024

034

Placebo group (n=35)

034

034

040

040

Sustained-release dexamfetamine group

(n=38)

055

050

047

049

Placebo group (n=35)

037

046

049

057

Sustained-release dexamfetamine group

(n=38)

076

062

060

052

Placebo group (n=35)

057

054

057

046

Sustained-release dexamfetamine group

(n=38)

037

033

030

028

Placebo group (n=35)

043

031

034

031

Sustained-release dexamfetamine group

(n=38)

013

019

022

013

Placebo group (n=35)

020

029

033

029

Wald=858;
p=0035
..

Wald=068;
p=0411
..

Wald=592;
p=0115
..

Cannabis use (1 day past month)

Wald=270;
p=0440
..

Wald=010;
p=0758
..

Wald=630;
p=0098
..

Physical health problems

Wald=1590;
p=0001
..

Wald=091;
p=0340
..

Wald=388;
p=0275
..

Mental health problems

Wald=514;
p=0162
..

Wald=009;
p=0764
..

Wald=057;
p=0904
..

Illegal activities (1 day past month)

Wald=448;
p=0214
..

Wald=201;
p=0157
..

Wald=069;
p=0875
..

GEE=generalised estimating equation. *Estimated marginal means were based on generalised estimating equation models, using an unstructured correlation matrix,
and assuming missing data (seven of 292 [four 73] assessments; 2%) were missing completely at random. Maudsley Addiction Prole 8. Brief Symptom Inventory
(071 [women] or 056 [men]).

Table 3: Longitudinal changes in secondary outcomescocaine craving, substance use, health problems, and criminality (intention-to-treat sample, n=73)

Sustained-release dexamfetamine group

(n=38)
Sleeping problems

Placebo group (n=35)

13 (34%)

3 (9%)

6 (16%)

2 (6%)

Physical arousal

5 (13%)

2 (6%)

Gastrointestinal problems

5 (13%)

3 (9%)

Changes in appetite

6 (16%)

2 (6%)

Changes in weight

5 (13%)

2 (6%)

Inuenza

3 (8%)

3 (9%)

Dizziness

3 (8%)

0 (0%)

Respiratory complaints

0 (0%)

2 (6%)*

Craving

0 (0%)

2 (6%)

Headache

1 (3%)

1 (3%)

Agitation/irritability

Data are n of patients (%). *Including one patient with a serious adverse event (admission to hospital).

Table 4: Adverse events reported by at least two patients

the dexamfetamine group vs 79 days [129] in the placebo

group; Wald =005, df=1, p=0828). Similarly, no group
dierences were noted between the number of
6

consecutive weeks with full medication adherence (mean

87 weeks [SD 37] in the dexamfetamine group vs
93 weeks [34] in the placebo group; Wald =009, df=1,
p=0767) and medication acceptance in the nal 4 weeks
(247 days [SD 81] in the dexamfetamine group vs
251 days [70] in the placebo group; Wald =001, df=1,
p=0943). Additionally, medication adherence did not
dier between the four treatment centres (all p>091).
Four patients in the dexamfetamine group and four in the
placebo group discontinued their medication intake
prematurely: three were imprisoned, four due to adverse
events, and one had limited treatment eects (gure).
One or more adverse events were reported by 28 (74%)
patients in the dexamfetamine group and by 16 (46%)
patients in the placebo group (OR 333 [95% CI 125887];
p=0016). Together, 95 adverse events were registered and
adverse events that were reported by at least two patients
are summarised in table 4. Patients in the dexamfetamine
group reported 69 adverse events, of which 58 (84%) events
were possibly, probably, or certainly related to the study
medication. Most of these adverse events (51 events; 74%)

www.thelancet.com Published online March 22, 2016 http://dx.doi.org/10.1016/S0140-6736(16)00205-1

Articles

Sustained-release dexamfetamine
group (n=36)

Placebo group (n=35)

Baseline

Baseline

Heart rate (beats per min)

Week 12

Group time

Week 12

682 (119)

761 (116)

693 (100)

687 (126)

F=958, df=1, p=0003

1281 (157)

1274 (147)

1265 (158)

1249 (148)

F=006, df=1, p=0809

Diastolic blood pressure (mm Hg)

793 (93)

812 (92)

805 (96)

793 (97)

F=234, df=1, p=0130

Bodyweight (kg)

769 (187)

772 (184)

740 (182)

739 (179)

F=021, df=1, p=0645

Systolic blood pressure (mm Hg)

Data are mean (SD), unless otherwise specied.

Table 5: Baseline to week 12 changes in heart rate, blood pressure, and bodyweight

were resolved before the end of the study treatment.

Sleeping problems was the adverse event reported by most
patients (n=13; 34%). In the placebo group, 26 adverse
events were reported, of which 18 (69%) were possibly,
probably, or certainly related to the study medication.
One serious adverse event occurred: a patient in the
placebo group was admitted to hospital during the study
period due to an exacerbation of chronic obstructive
pulmonary disease, which was not related to the study
drug. After admission, this patient resumed treatment.
In six other patients, adverse events resulted in
(temporary) discontinuation of study treatment: two in
the placebo group and four in the dexamfetamine group.
Of the latter four patients, two resumed treatment with a
dose of 30 mg/day sustained-release dexamfetamine, one
patient discontinued medication intake due to psychotic
symptoms, and one patient due to concurrent adverse
events of mild to moderate severity.
Heart rate signicantly increased from baseline to week
12 among patients in the dexamfetamine group compared
with those in the placebo group (table 5). No signicant
group by time interaction eects were noted for blood
pressure or bodyweight. Week 12 ECG data were available
for 67 patients (dexamfetamine n=34; placebo n=33) with
only one abnormality in terms of a repolarisation
disturbance in a patient in the placebo group.

Discussion
This multicentre, randomised, double-blind, placebocontrolled trial shows the acceptance, ecacy, and safety
of 60 mg/day oral sustained-release dexamfetamine as a
substitution drug in the treatment of chronic
crack-cocaine dependence in heroin-dependent patients,
currently in heroin-assisted treatment. Sustained-release
dexamfetamine was superior to placebo in terms of the
primary cocaine-related outcome (d=058), and all selfreported and urine-based secondary cocaine use-related
outcomes (d=058094 and d=031, respectively).
Sustained-release
dexamfetamine
was
generally
well-accepted, with high medication adherence. No
serious adverse events occurred in the dexamfetaminetreated patients. There were no unexpected adverse
events and most adverse events were transient and
well-tolerated.

Our ndings are an important contribution to the

search for eective pharmacotherapies for cocaine
dependence: it is the rst study that shows the benets of
a robust dose of sustained-release dexamfetamine as a
valuable agonist medication in the treatment of cocaine
dependence. This is by contrast with previous studies in
which strong inferences could not be made because of
high rates of premature treatment discontinuation and
promising, but often non-signicant, indications of
cocaine use reductions. In two randomised controlled
trials by Grabowski and colleagues,8,9 reductions in
cocaine use were larger in the 60 mg/day sustainedrelease dexamfetamine group than in the 30 mg/day
sustained-release dexamfetamine and placebo groups,
but both studies had treatment discontinuation rates of
up to 60% and the reported eects were only signicant
in subgroup analyses. In a small randomised controlled
trial of 30 cocaine injectors, signicant reductions in
cocaine use and cocaine-related improvements were
noted in the dexamfetamine group (60 mg/day), but not
in the between-group comparison.10 In a randomised
placebo-controlled trial (n=73) on the ecacy of 60 mg/day
sustained-release dexamfetamine, 400 mg/day modanil,
and the combination of both compounds, 60% of the
patients had discontinued treatment at 12 weeks and no
benets of sustained-release dexamfetamine over placebo
were noted.22 Finally, in a recent pilot randomised placebocontrolled trial, 70 mg/day of the prodrug lisdexamfetamine (containing approximately 30 mg/day
dexamfetamine) resulted in reduced craving but not in an
increase of cocaine abstinence.23 Thus, sustained-release
dexamfetamine has repeatedly shown to be a promising
treatment for cocaine dependence, but no studies so far
have shown a convincing benet in terms of signicant
and substantial reductions in cocaine use, most likely due
to small samples and low treatment adherence. By
contrast, our study shows very good medication adherence
and superiority of dexamfetamine in reducing selfreported and urine-based cocaine use.
We believe that our high medication adherence resulted
from daily supervised intake that enabled the treatment
sta to motivate patients and intensively monitor potential
side eects of the study medication, which are important
strategies to optimise adherence.24 Additionally, increased

www.thelancet.com Published online March 22, 2016 http://dx.doi.org/10.1016/S0140-6736(16)00205-1

Articles

doses of sustained-release dexamfetamine, such as

60 mg/day, are likely to result in more robust ndings.9,23
The observed eect sizes in our study were fair (urinebased cocaine use) to moderate (self-reported cocaine use)
and at least comparable to eect sizes in studies on other
chronic disorders, including alcohol dependence25 and
many other psychiatric and general medical conditions.26
Ecacy of sustained-release dexamfetamine was not
shown for our secondary, health-related outcome
measures. This could be due to the fact that our study
population already participated in heroin-assisted
treatment for on average 4 years, whereas much improvement in physical and mental health and reduction in
criminality already occurs at an early phase of heroinassisted treatment, as was shown by Blanken and
colleagues.3 Hence, in this ageing population with a long
history of cocaine and heroin dependency in heroinassisted treatment, there might be little room for further
improvements in these areas.
The study has several limitations. First, the sample size
was limited, but in view of the a priori expected eect
size, the study was adequately powered and a larger
sample would not be approved by a medical ethics
committee for this the proof-of-principle study. Moreover,
the study was undertaken in four treatment centres and
the eects were not driven by only one specic centre.
Second, a study with a duration of 12 weeks can not give a
conclusive answer about the best treatment for a chronic
relapsing disorder such as cocaine dependence. However,
at 12 weeks, 89% of the patients were still in treatment
and there are no reasons to expect that drop-out was
imminent or that the eect of the treatment was waning.
We therefore believe that this study provides a strong
indication for the potential long-term eectiveness of
sustained-release dexamfetamine as an agonist
pharmacotherapy for cocaine dependence. Third, the
ecacy of sustained-release dexamfetamine on selfreported cocaine abstinence might be somewhat overestimated in our study in view of the substantially smaller
eect size in urine-based cocaine use and the modest
agreement between self-reported and urine-based cocaine
use in the nal 4 weeks of the study. Therefore, we did a
post-hoc sensitivity analysis with days of cocaine
abstinence in the last 4 study weeks as dependent variable,
in which self-reported days of cocaine abstinence in the
2 days preceding the urinalysis were converted into nonabstinent days in case of a cocaine-positive or missing
urine sample (data not shown). This analysis showed that
patients in the dexamfetamine group on average still had
signicantly more cocaine-free days (89 days) than did
those in the placebo group (39 days), with a comparable
eect size (adjusted d=071 vs original d=077). However,
we also have to consider the possibility that our urinebased ecacy of sustained-release dexamfetamine on
cocaine abstinence is an underestimation of the
true eect, because qualitative urine tests only
provide dichotomous outcomes (cocaine-positive or
8

cocaine-negative) and can not detect reductions in the

amount of cocaine that was used, such as those noted in
our study. In view of the rationale for agonist substitution
treatment with sustained-release dexamfetamine, cocaine
use reductions and stabilisation rather than cocaine
abstinence are also valid treatment goals. Fourth, the
study was undertaken in a quite specic treatment
setting, and it is an important question whether the
demonstrated ecacy of sustained-release dexamfetamine
can be generalised to cocaine-dependent patients outside
heroin-assisted treatment. We believe that the results are
generalisable to chronic cocaine-dependent patients with
comorbid heroin dependence in methadone maintenance
treatment with daily visits and supervised methadone
intake, because these patients have very similar patterns
of cocaine use, largely similar clinical characteristics,
and daily supervised intake of sustained-release
dexamfetamine can be established in this context.
Generalisability is less clear when it comes to cocaine and
comorbid heroin-dependent patients in methadone
maintenance treatment without daily supervised
methadone intake and regular take home doses
of methadone. In this context, sustained-release
dexamfetamine prescription should be made conditional
on regularly supervised intake of the medication and on
other measures to improve compliance, such as
compliance enhancement therapy24 or contingency
management directed at treatment attendance and
cocaine-free urines.27,28 Similar measures are needed in
cocaine-dependent patients without comorbid heroin
dependence, and thus not in substitution treatment. In
view that in previous studies of dexamfetamine
twice weekly visits to the treatment centre were related to
high treatment discontinuation rates,8,9,22 future studies
on the ecacy of sustained-release dexamfetamine
among more general populations of cocaine-dependent
patients in routine addiction treatment services should
incorporate medication management interventions,
including frequent monitoring of medication adherence
and adverse events, frequent dose evaluations and
motivational enhancement,24 combined with providing
relevant incentives for both treatment attendance and
cocaine abstinence29 to improve treatment outcomes.
Finally, our ndings might not only be relevant for the
treatment of patients with cocaine dependence, but
possibly also for the many patients with other stimulant
addictions, although the ecacy of sustained-release
dexamfetamine in these populations still has to be
established.30 Our study has shown that agonist pharmacotherapy with a robust dose of sustained-release
dexamfetamine is possible and safe, and shows at least
one way to improve medication adherence and treatment
outcomes in chronic patients with a stimulant dependence,
and our approach and the discussed strategies can be
examples for future studies in the eld of addiction.
We conclude that sustained-release dexamfetamine is a
well-accepted, eective, and safe agonist drug for the

www.thelancet.com Published online March 22, 2016 http://dx.doi.org/10.1016/S0140-6736(16)00205-1

Articles

treatment of cocaine-dependent patients currently in

heroin-assisted treatment. Replication of these ndings in
treatment-refractory cocaine-dependent and other
stimulant-dependent patients in less specic treatment
settings is warranted, using multiple strategies to optimise
treatment adherence, such as medication management
interventions and contingency management.
Contributors
VH, PB, WB, and BW designed the study and wrote the protocol.
BN was involved in the development and manufacturing of the
investigational medicinal product, the blinding procedure and
randomisation. MN and PB undertook the statistical analysis, and
MN wrote the rst draft of the manuscript with support of all co-authors.
All authors contributed to and have approved the nal manuscript.
Declaration of interests
VH, WB, PB, and MN report a grant from the Netherlands Organisation
for Health Research and Development (ZonMw, project number
31160012) to conduct this study. WB also received grants from Alkermes
and personal fees from Lundbeck, Novartis, Indivior, D&A Pharma,
Bioproject, Pzer, and Mundipharma, but there were no competing
interests regarding this study. BN developed and manufactured the
dexamfetamine investigational medicinal product and placebo tablets as
employee of the pharmacy of the Antoni van Leeuwenhoek hospital
(Amsterdam, Netherlands).
Acknowledgements
We thank the Netherlands Organization for Health Research and
Development (ZonMw) for their nancial support. We also thank our
research assistant for the data collection. All participating patients,
committed medical sta, pharmacists, and other collaborators of
MSC Bouman Rotterdam, GGD Amsterdam, and MSC Brijder The Hague
are cordially thanked for their contributions to the execution of the study.
References
1
Ferri M, Davoli M, Perucci CA. Heroin maintenance for chronic
heroin-dependent individuals. Cochrane Database Syst Rev 2011;
12: CD003410.
2
Rehm J, Gschwend P, Steen T, Gutzwiller F, Dobler-Mikola A,
Uchtenhagen A. Feasibility, safety, and ecacy of injectable heroin
prescription for refractory opioid addicts: a follow-up study. Lancet
2001; 358: 141723.
3
Blanken P, van den Brink W, Hendriks VM, et al. Heroin-assisted
treatment in the Netherlands: History, ndings, and international
context. Eur Neuropsychopharmacol 2010; 20: S10558.
4
Marsden J, Eastwood B, Bradbury C, et al. Eectiveness of
community treatments for heroin and crack cocaine addiction in
England: a prospective, in-treatment cohort study. Lancet 2009;
374: 126270.
5
Dursteler-MacFarland KM, Farronato NS, Strasser J, et al.
A randomized, controlled, pilot trial of methylphenidate and
cognitive-behavioral group therapy for cocaine dependence in
heroin prescription. J Clin Psychopharmacol 2013; 33: 10408.
6
Castells X, Casas M, Perez-Mana C, Roncero C, Vidal X, Capella D.
Ecacy of psychostimulant drugs for cocaine dependence.
Cochrane Database Syst Rev 2010; 2: CD007380.
7
Herin DV, Rush CR, Grabowski J. Agonist-like pharmacotherapy for
stimulant dependence: preclinical, human laboratory, and clinical
studies. Ann N Y Acad Sci 2010; 1187: 76100.
8
Grabowski J, Rhoades H, Schmitz J, et al. Dextroamphetamine for
cocaine-dependence treatment: a double-blind randomized clinical
trial. J Clin Psychopharmacol 2001; 21: 52226.
9
Grabowski J, Rhoades H, Stotts A, et al. Agonist-like or
antagonist-like treatment for cocaine dependence with methadone
for heroin dependence: two double-blind randomized clinical trials.
Neuropsychopharmacology 2004; 29: 96981.
10 Shearer J, Wodak A, van Beek I, Mattick RP, Lewis J.
Pilot randomized double blind placebo-controlled study of
dexamphetamine for cocaine dependence. Addiction 2003; 98: 113741.

11

12

13

14

15

16

17

18

19

20
21

22

23

24
25

26

27

28

29

30

Nuijten M, Blanken P, van den Brink W, Hendriks V.

Cocaine addiction treatments to improve control and reduce harm
(CATCH): new pharmacological treatment options for crack-cocaine
dependence in the Netherlands. BMC Psychiatry 2011; 11: 135.
van den Brink W, Hendriks VM, Blanken P, Koeter MW,
van Zwieten BJ, van Ree JM. Medical prescription of heroin to
treatment resistant heroin addicts: two randomised controlled trials.
BMJ 2003; 327: 310.
American Psychiatric Association. Diagnostic and Statistical
Manual of Mental Disorders (4th ed). Washington DC:
American Psychiatric Publishing; 1994.
Cottler LB. Composite International Diagnostic
Interview- Substance Abuse Module (SAM), version 4.1. St Louis,
MO: Department of Psychiatry, Washington University School of
Medicine; 2000.
Sheehan DV, Lecrubier Y, Sheehan KH, et al. The Mini-International
Neuropsychiatric Interview (M.I.N.I.): the development and
validation of a structured diagnostic psychiatric interview for
DSM-IV and ICD-10. J Clin Psychiatry 1998; 59: 2233.
Hendriks VM, Kaplan CD, van Limbeek J, Geerlings P.
The Addiction Severity Index: reliability and validity in a Dutch
addict population. J Subst Abuse Treat 1989; 6: 13341.
Sobell LC, Sobell MB. Timeline Follow-back: a technique for
assessing self-reported alcohol consumption. In: Litten RZ, Allen J,
eds. Measuring alcohol consumption: Psychosocial and biological
methods. Towota, NJ: Humana Press; 1992: pp 4172.
De Wildt WA, Lehert P, Schippers GM, Nakovics H, Mann K,
van den Brink W. Investigating the structure of craving using
structural equation modeling in analysis of the
obsessive-compulsive drinking scale: a multinational study.
Alcohol Clin Exp Res 2005; 29: 50916.
Marsden J, Gossop M, Stewart D, et al. The Maudsley Addiction
Prole (MAP): a brief instrument for assessing treatment outcome.
Addiction 1998; 93: 185767.
Derogatis LR, Melisaratos N. The Brief Symptom Inventory:
an introductory report. Psychol Med 1983; 13: 595605.
De Wilde EF, Hendriks VM. The Client Satisfaction Questionnaire:
psychometric properties in a Dutch addict population.
Eur Addict Res 2005; 11: 15762.
Schmitz JM, Rathnayaka N, Green CE, Moeller FG, Dougherty AE,
Grabowski J. Combination of modanil and d-amphetamine for the
treatment of cocaine dependence: a preliminary investigation.
Front Psychiatry 2012; 3: 77.
Mooney ME, Herin DV, Specker S, Babb D, Levin FR, Grabowski J.
Pilot study of the eects of lisdexamfetamine on cocaine use:
a randomized, double-blind, placebo-controlled trial.
Drug Alcohol Depend 2015; 153: 94103.
Weiss RD. Adherence to pharmacotherapy in patients with alcohol
and opioid dependence. Addiction 2004; 99: 138292.
Maisel NC, Blodgett JC, Wilbourne PL, Humphreys K, Finney JW.
Meta-analysis of naltrexone and acamprosate for treating alcohol
use disorders: when are these medications most helpful? Addiction
2013; 108: 27593.
Leucht S, Hierl S, Kissling W, Dold M, Davis JM. Putting the
ecacy of psychiatric and general medicine medication into
perspective: review of meta-analyses. Br J Psychiatry 2012;
200: 97106.
Prendergast M, Podus D, Finney J, Greenwell L, Roll J.
Contingency management for treatment of substance use disorders:
a meta-analysis. Addiction 2006; 101: 154660.
Petry NM, Rash CJ, Byrne S, Ashraf S, White WB. Financial
reinforcers for improving medication adherence: ndings from
a meta-analysis. Am J Med 2012; 125: 88896.
Schierenberg A, van Amsterdam J, van den Brink W, Goudriaan AE.
Ecacy of contingency management for cocaine dependence
treatment: a review of the evidence. Curr Drug Abuse Rev
2012; 5: 32031.
Galloway GP, Buscemi R, Coyle JR, et al. A randomized,
placebo-controlled trial of sustained-release dextroamphetamine for
treatment of methamphetamine addiction. Clin Pharmacol Ther
2011; 89: 27682.

www.thelancet.com Published online March 22, 2016 http://dx.doi.org/10.1016/S0140-6736(16)00205-1