A randomised, double-blind, placebo-controlled,
efficacy study of nalmefene, as-needed use, in
patients with alcohol dependence
Antoni Guala,*, Yuan Heb, Lars Torupb, Wim van den Brinkc,1,
Karl Mannd,1, for the ESENSE 2 Study Group
aNeurosciences Institute, Hospital Clinic, Barcelona, Spain
bH. Lundbeck A/S, Copenhagen, Denmark
cAcademic Medical Center, Department of Psychiatry, University of Amsterdam, Amsterdam, The Netherlands
dCentral Institute of Mental Health, University of Heidelberg, Mannheim, Germany
Received 16 October 2012; received in revised form 17 January 2013; accepted 28 February 2013
This study evaluated the efficacy of as-needed use of the opioid system modulator nalmefene in
reducing alcohol consumption in patients with alcohol dependence. Seven hundred and eighteen
patients (placebo=360; nalmefene=358), ≥18 years of age, with a diagnosis of alcohol dependence,
≥6 heavy drinking days and an average alcohol consumption ≥WHO medium drinking risk level in the
4 weeks preceding screening, were randomised (1:1) to 24 weeks of as-needed placebo or nalmefene
The co-primary efficacy analyses showed a significantly superior effect of nalmefene compared to
placebo in the change from baseline to month 6 in heavy drinking days (group difference: −1.7 days/
month [95% CI −3.1; −0.4]; p=0.012) and a better but not significant effect in reducing total alcohol
consumption (group difference: −5.0 g/day last month [95% CI −10.6; 0.7]; p=0.088). A subgroup
analysis showed that patients who did not reduce their drinking prior to randomisation benefitted
more from nalmefene. Improvements in Clinical Global Impression and reductions in liver enzymes
were greater in the nalmefene group than in the placebo group. Adverse events were more common
with nalmefene; the incidence of adverse events leading to dropout was similar in both groups.
This study provides evidence for the efficacy of nalmefene, which constitutes a new pharma-
cological treatment paradigm in terms of treatment goal (reduced drinking) and dosing regimen
(as-needed), in alcohol dependent patients unable to reduce alcohol consumption on their own.
& 2013 Elsevier B.V. and ECNP . All rights reserved.
Europe has the highest overall consumption of alcohol (World
Health Organization, 2010), and in general, the European
Union can be characterised by a lower proportion of abstainers
0924-977X/$-see front matter & 2013 Elsevier B.V. and ECNP. All rights reserved.
⁎Correspondence to: Department of Psychiatry, Alcohol Unit,
Institute of Neurosciences, Hospital Clinic, 08036 Barcelona, Spain.
Tel.: +34 932271719; fax: +34 932275548.
E-mail addresses: firstname.lastname@example.org.
email@example.com (A. Gual).
1These authors contributed equally.
European Neuropsychopharmacology (2013) 23, 1432–1442
and a higher proportion of the population drinking more than
20 g of pure alcohol per day than the rest of the world (World
Health Organization, 2011). Worldwide, the European region
suffers the highest impact of alcohol with 6.5% of all deaths
and 11.6% of all disability-adjusted life years attributable to
alcohol (Rehm et al., 2009).
Most of the alcohol-attributable mortality is due to alcohol
dependence, presumably by means of heavy drinking (Rehm
et al., 2013). Approximately 15 million persons in the European
Union are alcohol-dependent (Wittchen et al., 2011). There is a
large treatment gap, with less than 10% of people in Europe with
a diagnosis of any alcohol disorder (including alcohol depen-
dence) actually receiving any treatment (Alonso et al., 2004).
Reduction of alcohol consumption is increasingly accepted as
a viable treatment goal (European Medicines Agency, 2010;
Luquiens et al., 2011). However, the three currently registered
pharmacological treatments for alcohol dependence are indi-
cated only for the maintenance of abstinence following
A large proportion of patients in abstinence-oriented
treatments experience relapses (Anton et al., 2006; Mann
et al., 2004; Merkx et al., 2011; Miller et al., 2001), and
abstinence-oriented treatments might not be desirable or
acceptable to many patients (Gastfriend et al., 2007;
Marlatt and Witkiewitz, 2002). Allowing patients to choose
between abstinence and reduced drinking as their treat-
ment goal may enhance engagement with the treatment,
ultimately leading to better treatment outcomes for the
population at large (Adamson et al., 2010; Heather et al.,
2010). Furthermore, research has shown that any reduction
in alcohol consumption for a person who consumes more
than 10 g of alcohol per day will reduce the annual and life-
time risk of alcohol-related death (Rehm et al., 2011).
Therefore, there is clearly a need for new pharmacolo-
gical treatments allowing for reduction of alcohol consump-
tion as a treatment goal.
Nalmefene is an opioid system modulator, which in several
studies in patients with alcohol use disorders has been asso-
ciated with a reduction of heavy drinking. Although Anton et al.
(2004) were unable to show a reduction in alcohol use
compared to placebo, other studies in patients with alcohol-
use disorders indicate that treatment with nalmefene causes a
reduction of heavy drinking (Karhuvaara et al., 2007; Mason
et al., 1994, 1999). A recently published large phase 3 study in
patients with alcohol dependence showed that nalmefene,
taken on an as-needed basis was superior to placebo in reducing
alcohol consumption (Mann et al., 2013).
Here, we present results from another recently completed
phase 3 study in patients with alcohol dependence that
assessed the efficacy and safety of as-needed use of nalmefene
in reducing alcohol consumption, measured as the monthly
changes from baseline in the number of heavy drinking days
(days in last month) and mean total alcohol consumption (g/day
in last month) during a treatment period of 24 weeks.
This randomised, double-blind, placebo-controlled, parallel-group
study included patients from 57 sites in Belgium, the Czech Republic,
France, Italy, Poland, Portugal, and Spain. Patients were recruited
from in- and out-patient clinics, from the study site's patient pool,
and by spontaneous referrals to the study site. Advertisements were
used in the Czech Republic, France, Italy, and Spain. Eligible patients
were men and women aged ≥18 years with a primary diagnosis of
alcohol dependence according to Diagnostic and Statistical Manual of
Mental Disorders, 4th Edition, Text Revision (DSM-IV-TR™ [American
Psychiatric Association, 2000]) (assessed using the Mini-International
Neuropsychiatric Interview [MINI; Lecrubier et al., 1997]) and a blood
alcohol concentration o0.02% at the screening visit. Exclusion
criteria were (a) o6 heavy drinking days in the 4 weeks before
screening (European Medicines Agency, 2010; a day with alcohol
consumption ≥60 g for men and ≥40 g for women), (b) an average
alcohol consumption below medium drinking risk level according to
the World Health Organization (WHO) in the 4 weeks before screen-
ing (≤40 g alcohol/day for men and ≤20 g alcohol/day for women;
World Health Organization, 2000), (c) 414 consecutive abstinent
days in the 4 weeks preceding screening, (d) a score ≥10 on the
revised version of the Clinical Institute Withdrawal Assessment for
Alcohol (CIWA-Ar; Sullivan et al., 1989), indicating the need for
medication supported detoxification, (e) aspartate aminotransferase
or alanine aminotransferase (ALAT) values 43 times of upper normal
limit, (f) a current DSM-IV Axis I disorder other than alcohol
dependence (except nicotine dependence), (g) a DSM-IV Axis II
(h) recent (within 1 week prior to the screening visit) treatment
with opioid agonists or partial agonists. For the full list of selection
criteria, see Supplementary material.
This study was designed and conducted in accordance with the
principles of the Declaration of Helsinki and Good Clinical Practice,
and each site started patient inclusion only after ethics committee
approval. All patients gave written informed consent.
2.2.Randomisation and blinding
At baseline (week 0), eligible patients were assigned to 24 weeks of
treatment with as-needed use of either placebo or nalmefene
18 mg (base) in a 1:1 ratio, according to a computer generated
randomisation list (in blocks of 4), provided by the sponsor.
Randomisation for the run-out period was also done at baseline.
Patients, investigators, staff and the sponsor were blind to
treatment assignment. Two sets of sealed envelopes containing
study medication details for each patient were prepared. One set
was kept by the sponsor and one set by the investigator or
pharmacist. The randomisation code was only to be broken by the
investigator in case of an emergency situation. The randomisation
code was not broken for any patient during the study. Nalmefene
and placebo tablets were identical in appearance.
The study consisted of a 1 to 2-week screening period, a 24-week
double-blind main treatment period with nalmefene or placebo,
and a 4-week double-blind run-out period (to evaluate any treat-
ment discontinuation effects) during which nalmefene-treated
patients were randomised to placebo or nalmefene (1:1) and
placebo-treated patients continued with placebo. A safety follow-
up was scheduled 4 weeks after completion or dropout.
Patients were instructed to take one tablet on each day they
perceived a risk of drinking alcohol (as-needed dosing), preferably
1–2 h prior to the anticipated time of drinking. Tablets could be
taken up to once daily and were supplied in wallet cards with space
for the patient to record the date of study medication intake. The
Timeline Follow-back (Sobell and Sobell, 1992) was used to obtain
estimates of daily drinking as well as to record daily medication
intake. In addition, all patients took part in a motivational and
adherence-enhancing intervention (BRENDA [Volpicelli et al., 2001;
1433Nalmefene, as-needed use, in patients with alcohol dependence
Starosta et al., 2006]) to support them in changing their behaviour
and to enhance adherence to treatment, starting at randomisation
and subsequently at all scheduled visits. No treatment goal was
defined, i.e. both abstinence and reduction were accepted; no
information was collected on individual treatment goals. Measure-
ments needed for the assessment of efficacy and safety were
performed at screening (week-1 or week-2), baseline, weeks 1, 2,
and 4, followed by monthly assessments. For a full description of
timing of assessments, the reader is referred to the Supplementary
Monthly drinking variables were derived from the Timeline
Follow-back (Sobell and Sobell, 1992) that provided information
of daily number of standard drinks. To define standard drinks, a
conversion card was provided. The conversion of recorded standard
drinks to grams was performed by a statistical programmer using
the following country specific factors:: Belgium, Italy, Poland, and
Spain 10 g; France 12 g; Portugal 14 g; Czech Republic 16 g.
At screening, patients reported their daily drinking over the
previous month (=28 consecutive days). At subsequent visits, they
reported drinking since the previous visit.
The pre-defined co-primary outcome measures were change from
baseline in heavy drinking days and change from baseline in total
alcohol consumption (g/day) at month 6.
Key-secondary outcome measure was drinking risk level response
(from very high drinking risk level at baseline to medium drinking
risk level or below, or from high or medium drinking risk level at
baseline to low drinking risk level or below) at month 6.
Secondary outcome measures reported here are Clinical Global
Impression-Severity of Illness (CGI-S) and Global Improvement (CGI-
I) scores (Guy, 1976) and γ-glutamyltransferase (GGT) and ALAT
values at week 24. Other secondary variables will be reported
elsewhere. For the full list of outcome variables, the reader is
referred to the Supplementary material.
Clinical status was based on Clinical Global Impression-Severity
of Illness, the Alcohol Dependence Scale (Skinner and Horn, 1984),
and the Drinker Inventory of Consequences (Miller et al., 1995).
Safety assessments consisted of evaluation of adverse events
(including pre-treatment and treatment-emergent adverse events),
clinical safety laboratory tests, vital signs, weight, electrocardio-
grams, and Profile of Mood States. To capture any signal related to
psychiatric adverse events, a group of selected adverse events was
pre-defined (see Supplementary material). Adverse events poten-
tially related to suicide were identified using the sub-standardised
Medical Dictionary for Regulatory Activities query “suicide/self-
The sample size calculation was based on a standard deviation for
the change from baseline in number of heavy drinking days of 7 days
and the change from baseline in total alcohol consumption of
36.5 g/day and a correlation of 0.7 between heavy drinking days
and total alcohol consumption. With a significance level of 5%, 300
patients in each treatment group would provide a power of 90% for
detecting a difference between the treatment groups of three
heavy drinking days and 12 g/day in the total alcohol consumption,
accounting for an expected drop-out rate of 35% at month 6.
Three datasets were pre-specified in the study protocol:
The all-patients-randomised set, comprising all randomised
patients, was used to calculate the incidence of serious adverse
events, in order to account for any pre-treatment serious adverse
The all-patients-treated set, comprising all randomised patients
but excluding from the dataset those with no recorded study
medication intake and all study medication returned. This dataset
was used for all remaining safety analyses.
The full-analysis set, comprising all patients in the all-patients-
treated set with at least one valid post-baseline assessment of
alcohol consumption, was used for all efficacy analyses.
Baseline for drinking variables in the main treatment period was
defined as the month preceding the screening visit. For all other
variables, baseline was defined as the assessment at the screening visit.
The co-primary outcome measures were analysed using mixed
model repeated measures, using observed cases, with the baseline
score as covariate, and site, sex, time (months 1–6), and treatment
as fixed effects; baseline score-by-time interaction and treatment-
by-time interaction were also included in the model.
Sensitivity analyses were performed using analysis of covariance
by month with the same covariates and fixed effects as in the mixed
model repeated measures analysis, using (a) observed cases, (b) last
observation carried forward and baseline observation carried for-
ward imputation, (c) placebo mean imputation (imputing the month
one estimation in the placebo group to all time points with missing
data) and (d) multiple imputation assuming that the future
behaviour of withdrawn patients is the same as those in the placebo
group with a similar past (Little and Yau, 1996).
Post-hoc analyses of the co-primary variables were performed to
estimate the effect of nalmefene versus placebo in the subgroup of
patients who, reduced or did not reduce their drinking to less than
6 heavy drinking days per month or below medium drinking risk
level already in the period between screening and randomisation.
All patients were classified as having (yes/no) at least a medium
drinking risk level and at least 6 heavy drinking days in the period
between screening and randomisation (extrapolated to 4 weeks).
These analyses were performed using the primary mixed model
repeated measures model including alcohol consumption at rando-
misation (yes/no)-by-time-by-treatment interaction with randomi-
sation score as response assuming no systematic difference between
the treatment groups.
The null hypothesis of no difference in treatment effect on heavy
drinking days and total alcohol consumption had to be rejected in
order to proceed with formal testing of the key-secondary outcome
measure, which was analysed by logistic regression by month, with
country, sex, baseline drinking risk level, and treatment as fixed
effects, imputing missing values with response based on mixed
model repeated measures—predicted total alcohol consumption
values from the primary analysis.
The odds ratio of nalmefene compared to placebo with 95%
confidence interval and corresponding p-value based on the likelihood
ratio test was estimated from the model. Sensitivity analyses were
performed for the same logistic regression model using observed
cases, last observation carried forward, non-response or sustained
response imputation for missing values, where sustained response was
defined as response at the current month and response at the previous
month with LOCF imputation for missing values.
The secondary outcome measures (CGI-S, CGI-I, log-transformed
GGTand ALAT values) were analysed with similar models as used for
the co-primary variables. The CGI-S baseline score was included as a
covariate in the model for CGI-I.
Adverse events were coded using the lowest level term according
to Medical Dictionary for Regulatory Activities, Version 13.0.
The principal statistical software used was SAS®, Version 9.2.
From March, 2009 to July, 2010, 941 patients were screened, of
whom 718 were randomised (Figure 1). There were no clinically
relevant differences in baseline demographic or clinical char-
acteristics between the groups (Table 1). All but eight patients
were Caucasian, approximately 70% were men, and the mean
A. Gual et al.1434
age was 45 years. Mean age of onset of alcohol problems was
In the month before screening, patients had on average
19 heavy drinking days and drank on average 90 g of alcohol
per day. The mean CGI-S score of 4, the Drinker Inventory of
Consequences score of 46, and the Alcohol Dependence
Scale score of approximately 14 confirmed that these were
moderately ill patients with significant adverse conse-
quences of drinking. Mean values of liver parameters were
close to or slightly above the reference ranges. The
majority of patients had not previously been treated for
either alcohol dependence (429 [60%] of 718) or alcohol
withdrawal symptoms (593 [83%] of 718).
The all-patients-treated set comprised 678 patients and the
full-analysis set comprised 655 patients. Importantly, 218 of
these patients (33%) reduced their drinking to o6 heavy
drinking days/month or below medium drinking risk level
already in the period between screening and randomisation,
i.e., prior to taking any study medication (Table 1). Baseline
characteristics for the patients reducing their consumption
prior to treatment initiation were similar to those that did not
do so (Supplementary material). However, the early reducers
had somewhat lower alcohol consumption as shown by the
parameters total alcohol consumption and heavy drinking
days, and more often fell within the medium drinking risk
level category. During the main treatment period, 127 (38%) of
the placebo-treated patients and 140 (41%) of the nalmefene-
treated patients dropped out from the study; the most
frequent primary reason was withdrawal of consent in both
the placebo and nalmefene groups (Figure 1). Two hundred
and five patients (61%) in the placebo and 194 (57%) in the
nalmefene group completed the entire study.
On average, patients on placebo took study medication
on 65% of the days in the main treatment period, whereas
patients on nalmefene took study medication on 57% of the
days (Table 2).
The mean number of heavy drinking days decreased from 20 to
7 days/month and the mean total alcohol consumption
Main treatment period
comprising all randomised patients but excluding from the dataset those with no recorded study medication intake and all study
medication returned. ‡=Full-analysis set, comprising all patients in the all-patients-treated set with at least one valid post-baseline
assessment of alcohol consumption. Eleven (11) and twelve (12) patients in the placebo and nalmefene groups, respectively were
not included in the full-analysis set.
Trial profile. *Adverse events were not set to primary reason for dropout by default. †=All-patients-treated set,
1435Nalmefene, as-needed use, in patients with alcohol dependence
decreased from 93 to 30 g/day in the nalmefene group at
month 6 (Table 3). In the placebo group, the mean number of
heavy drinking days decreased from 18 to 8 days/month and
the mean total alcohol consumption decreased from 89 to
33 g/day at month 6. A statistically significant reduction in the
number of heavy drinking days and total alcohol consumption
in favour of nalmefene was observed already at month 1
(Figure 2). The co-primary efficacy analyses showed a statis-
tically significantly superior effect of nalmefene compared to
placebo in the change from baseline to month 6 in heavy
drinking days (group difference: −1.7 days/month [95% CI
−3.1; −0.4]; p=0.012) and a better but not statistically
Demographics and baseline clinical characteristics.
Patients randomised (APRS) Placebo Nalmefene
Body mass index (kg/m2)
Age at the onset of drinking problems
Drinking risk levela
Total monthly heavy drinking days (days)a
Total alcohol consumption (g alcohol/day)a
Clinical global impression—severity of illness
Alanine aminotransferase (IU/L)b
Mean corpuscular volume (fL)b
Percentage carbohydrate-deficient transferrin (%)
Drinker inventory of consequences total score
Alcohol dependence scale total score
Previously treated for alcohol dependence
Previously treated for alcohol withdrawal
Family history of alcohol problems
o6 heavy drinking days or drinking risk level omedium at randomisationd
Data are mean (SD) or number of participants (%).
SD=Standard deviation. APRS=All-patients-randomised set.
aBased on Timeline Follow-back data from the month preceding the screening visit.
cSituation at screening. Percentages based on the full-analysis set.
dPatients having o6 heavy drinking days or a drinking risk level below medium in the period between screening and randomisation,
extrapolated to 4 weeks; percentages based on the full-analysis set.
Distribution of percentage of days with study medication intake in the main treatment period.
PatientsSummary statistics % of days with
Only patients in the all-patients-treated set with Timeline Follow-back study medication records are included.
aDistribution of the individual patient percentages of days with study medication intake.
A. Gual et al.1436
significant effect in reducing the total alcohol consumption
(group difference: −5.0 g/day last month [95% CI −10.6; 0.7];
p=0.088) (Table 4). All sensitivity analyses were numerically
in favour of nalmefene.
At baseline, the mean number of non-drinking days (stan-
dard error) was 5.4±0.3 in the placebo group and 5.0±0.3 in
the nalmefene group. The mean number of non-drinking days
increased to 14.7±0.7 in the placebo group and to 14.7±0.8 in
the nalmefene group at month 6.
The patients (33% [218 of the 655 patients in the full-
analysis set]) who in the period between screening and
randomisation had reduced the number of heavy drinking
days to o6 days/month or had a drinking risk level below
medium had reduced their alcohol consumption by 13.9
heavy drinking days and 62.6 g/day at the time of rando-
misation (Figure 3). In this group, the low level of alcohol
consumption was maintained throughout the study with no
difference between the treatment groups during the treat-
ment period. However, in the patients (67% [437 of the 655
patients in the full-analysis set]) who in the period between
screening and randomisation had ≥6 heavy drinking days/
month and at least a medium drinking risk level, nalmefene
treatment resulted in a significant reduction compared to
placebo in the mean number of heavy drinking days (−2.0
days/month [95% CI −3.6; −0.4]; p=0.012) as well as in the
mean total alcohol consumption (−7.0 g/day last month
[95% CI −13.6; −0.4]; p=0.037) at month 6 (Figure 3).
The key secondary outcome measure, analysis of drinking risk
level response was numerically in favour of nalmefene (odds
ratio=1.28; [0.89; 1.83]; p=0.1833 [Table 5]). Similar results
were obtained in the sensitivity analyses of the key secondary
outcome measure, except for the non-response imputation.
A decrease in the CGI-S score from baseline to week 24
was observed in both treatment groups (Figure 4), with
greater mean improvement (group difference: −0.2 [−0.44;
−0.02]; p=0.029) in the nalmefene group than in the
placebo group. The difference in the CGI-I score at week
24 was also in favour of the nalmefene group (group
difference: −0.2 [−0.38; 0.04]; p=0.111).
For GGT and ALAT, the analysis showed improvements
from baseline in both treatment groups; there was a greater
reduction from baseline to week 24 in ALAT in the nalme-
fene group than in the placebo group (p=0.049 [Table 6]).
There was no difference between nalmefene and placebo in
GGT at week 24.
3.3.Safety and tolerability
During the main treatment period, 199 (59%) of the patients in
the placebo group and 232 (68%) of the patients in the
nalmefene group had treatment-emergent adverse events
(Table 7). The majority of these events were mild or moderate.
Approximately half of the patients with treatment-emergent
adverse events had treatment-emergent adverse events with an
onset within 1 day after the first dose of study medication. Of
the most common treatment-emergent adverse events (inci-
dence ≥5%) nausea, dizziness, and insomnia had an incidence
two times higher in the nalmefene group than in the placebo
group. Fourty three patients dropped out due to treatment-
emergent adverse events during the main treatment period: 20
(5.9%) in the placebo group and 23 (6.7%) in the nalmefene
group (Table 7). Treatment-emergent adverse events with an
incidence ≥0.5%, leading to dropout comprised dizziness,
nausea, vomiting, anxiety, and insomnia in the nalmefene
Baseline and month 6 efficacy variables.
Monthly number of heavy drinking days (FAS, OC)
Monthly total alcohol consumption (g/day) (FAS, OC)
deviation; FAS=full-analysis set;
Monthly HDDs adjusted Mean Change
Monthly TAC (g/day) adjusted Mean
Change from Baseline
change from baseline in monthly heavy drinking days (HDDs).
(B) Adjusted mean change from baseline in monthly total
alcohol consumption (TAC; g/day). Baseline data for HDDs and
TAC were derived from the Timeline Follow-back for the month
preceding the screening visit. Patient numbers contributing to
each monthly period are shown below the x-axis for each
treatment group. *po0.05 (difference to placebo). B=base-
line. Bars indicate standard errors.
Change in alcohol consumption. (A) Adjusted mean
1437 Nalmefene, as-needed use, in patients with alcohol dependence
group, and anxiety, depression, and major depression in the
Serious adverse events were reported for 25 patients
(including four patients with pre-treatment serious adverse
events): 17 patients in the placebo group and 8 patients in the
nalmefene group. No serious adverse event was reported in
more than one patient in either treatment group, except for
intentional overdose (two patients in the placebo group).
All the serious adverse events in the nalmefene group, and
the majority of serious adverse events in the placebo group
were considered not related to study medication by the
investigator. There was no pattern in the distribution of serious
adverse events across system organ classes, and no indication of
specific serious adverse events occurring in the nalmefene
Two patients died during the main treatment period: a 50-
year-old woman in the placebo group (due to liver disorder as
a result of hepatocellular carcinoma) and a 61-year-old man in
the nalmefene group (sudden death; unknown cause; assessed
by the investigator as being not related to study medication).
One patient in the placebo–placebo group in the run-out
period had an intentional overdose of study medication and
suicidal behaviour (both serious adverse events). In addition,
nine patients were identified with potentially suicide-related
adverse events (five on placebo and two on nalmefene in the
main treatment period; one in the nalmefene–placebo group
in the main treatment and run-out period, and one in the
safety follow-up period, after having received placebo); for
the majority of these patients, the events were identified as
intentional overdoses of the study medication in order to
obtain increased efficacy. Fourteen patients (four on placebo,
10 on nalmefene) had one of the selected psychiatric adverse
events; none was serious and all the patients fully recovered.
In addition one patient had an ongoing selected psychiatric
adverse event at baseline. There were no apparent trends in
the incidence of potentially clinically significant clinical safety
Co-primary efficacy analysis at month 6.
Efficacy variableAdjusted change from baseline to month 6 Difference to placebo
Number of heavy drinking days (MMRM; OC)
Number of heavy drinking days—sensitivity analyses
Total alcohol consumption (g/day) (MMRM; OC) 229
Total alcohol consumption (g/day)—sensitivity analyses
−1.7 [−3.1; −0.4]0.012
−1.9 [−3.4; −0.4]
−1.8 [−3.0; −0.6]
−0.7 [−2.1; 0.7]
−1.1 [−2.22; −0.04] 0.042
−1.0 [−2.4; 0.3]
−4.9 [−10.6; 0.7]
−5.6 [−11.7; 0.5]
−5.9 [−11.1; −0.7]
−3.1 [−9.7; 3.5]
−5.2 [−10.2; −0.2]
−2.9 [−8.5; 2.8]
SE=Standard error; MMRM=mixed model repeated measures; OC=observed cases; ANCOVA=analysis of covariance; LOCF=last
observation carried forward; BOCF=baseline observation carried forward; PMI=placebo mean imputation; MI=multiple imputation.
Monthly TAC (g/day) adjusted mean
change from baseline
Monthly HDDs adjusted mean change
Placebo –non-early reducers
Nalmefene –non-early reducers
Placebo -early reducers
Nalmefene -early reducers
Placebo –non-early reducers
Nalmefene –non-early reducers
Placebo -early reducers
Nalmefene -early reducers
gorised according to alcohol consumption at randomisation.
(A) Adjusted mean change from baseline in monthly heavy
drinking days (HDDs). (B) Adjusted mean change from baseline
in monthly total alcohol consumption (TAC; g/day). Early redu-
cers=patients having less than 6 HDDs or below medium drinking
risk level at randomisation. *po0.05 (difference to placebo),
B=baseline, R=randomisation. Bars indicate standard errors.
Change in alcohol consumption for patients cate-
A. Gual et al. 1438
laboratory values and no clinically relevant changes over time
or differences between the treatment groups were seen in the
vital signs, weight, electrocardiogram parameters, or Profile
of Mood States total scores.
This study is the second study in the recently completed
clinical phase 3 programme using nalmefene as-needed as a
means of reducing alcohol consumption in patients with
alcohol dependence. Patients were predominantly middle-
aged men, with the majority having a high or very high
drinking risk level; the average baseline consumption was
90 g alcohol per day. In line with EMA recommendations,
patients with significant withdrawal symptoms were not
eligible for participation and thus some of the most severe
alcohol dependent patients had to be excluded. The study
population is comparable to patients that are likely to
present in primary care (Willenbring et al., 2009).
Despite their alcohol problems having started more than
10 years ago, the majority of the patients had never
received any treatment. The nalmefene treatment para-
digm thus addresses an unmet medical need as it obviously
has the potential to engage alcohol dependent patients in
treatment who may otherwise not have sought help.
Nalmefene does not require the patients to achieve
and maintain complete abstinence and the as-needed
dosing regimen engages patients with alcohol dependence
in active and responsible management of their illness,
which may improve patient adherence and persistence to
The idea of administrating an opioid antagonist (naltrex-
one) to patients that are still currently drinking, and only
when drinking is anticipated (as-needed/targeted use) in
order to reduce alcohol consumption has previously been
proposed (for review see Sinclair, 2001). As-needed use has
been studied with a variety of medications in patients with
an alcohol use disorder: nalmefene (Karhuvaara et al.,
2007), acamprosate (Laaksonen et al., 2008), disulfiram
(Laaksonen et al., 2008), and naltrexone (Heinälä et al.,
2001; Kranzler et al., 1997, 2003, 2009; Laaksonen et al.,
2008). However, the current study and the recently pub-
lished study by Mann et al. (2013) are the largest rando-
mised controlled studies of the as-needed use approach
with reduction of alcohol consumption as a primary outcome
measure in alcohol dependent patients to date.
Compared to baseline, there was a substantial reduction
in alcohol consumption in both treatment conditions on
both co-primary outcome measures: number of heavy
drinking days and total alcohol consumption decreased by
approximately65% in thenalmefenegroupand by
Key secondary analysis (drinking risk level response) at month 6.
Method for handling missing data PlaceboNalmefene Odds ratio 95% CI
326 206 (63%)329 221 (67%)1.28 [0.89; 1.83]0.1833
OR=odds ratio; CI=confidence interval; drinking risk level response=response defined for patients at very high risk at baseline: as a
downward shift to medium risk or below, and for patients at high or medium risk at baseline: as a downward shift to low risk or below;
OC=observed cases; LOCF=last observation carried forward; TAC=total alcohol consumption; MMRM=mixed model repeated
aMissing values were imputed by response evaluation based on individual patient-predicted values of TAC at each month derived
from the MMRM model used in the primary analysis.
B1 2481216 2024
Adjusted change from baseline
in CGI-S scores
1 248 1216 20 24
Adjusted CGI-I scores
change from baseline in Clinical Global Impression-Severity of
Illness (CGI-S) scores. (B) Adjusted Clinical Global Impression-
Global Improvement (CGI-I) scores. *po0.05 (difference to
placebo), B=baseline. Bars indicate standard errors.
Change in Clinical Global Impression. (A) Adjusted
1439Nalmefene, as-needed use, in patients with alcohol dependence
approximately 60% in the placebo group. Nalmefene was
statistically significantly superior to placebo in reducing the
number of heavy drinking days at month 6, with a group
difference of 1.7 days/month, and also showed a numeri-
cally better effect on total alcohol consumption, although
non-significantly. The sensitivity analyses were consistently
in favour of nalmefene, despite the fact that approaches
like placebo mean and multiple imputation minimises the
differences between the treatment groups after dropout.
There is no clear-cut answer to what constitutes a
clinically relevant magnitude of reduction of heavy drink-
ing. However, the European Medicines Agency's guideline on
the development of medicinal products for the treatment of
alcohol dependence (European Medicines Agency, 2010)
states that efficacy should also be evaluated in terms of
the difference in the percentage of treatment responders,
e.g. the difference in the percentage of patients with a
two-category downshift in the WHO drinking risk levels.
The result of the responder analysis based on a two-
category downshift in drinking risk level was consistently
numerically in favour of nalmefene, with the exception of
the sensitivity analysis that imputed missing values as non-
response. However, the non-response imputation can be
considered very conservative, with all patients dropping out
being considered non-responders, irrespective of the value
of total alcohol consumption at the time of dropout; an
assumption that is not supported by published data (Project
MATCH Research Group, 1998).
The difference between nalmefene and placebo in the
number of heavy drinking days per month at month 6 trans-
lates into a reduction of about 3 weeks of heavy drinking
days per year. From a public health perspective, this
difference is relevant, since evidence from epidemiological
data have shown that every heavy drinking day carries an
increased risk of accidents, aggression, suicide, and cardiac
arrest (Rehm et al., 2010). Furthermore, reduction of total
alcohol consumption is associated with reduced risk of
morbidity and mortality: any reduction in alcohol consump-
tion for a person who consumes more than 10 g of alcohol
per day will reduce the annual and life-time risk of alcohol-
related death (Rehm et al., 2011).
Efficacy variables independent of the Timeline Follow-
back data also provided evidence of the effect of nalme-
fene. The reduced alcohol consumption was associated with
reductions in the liver enzymes GGTand ALAT; the reduction
in ALAT was greater in the nalmefene group than in the
placebo group. Furthermore, the improvement in the CGI-S
scale score, which reflects the global clinical judgement of
the severity of illness by an expert clinician, was greater in
the nalmefene group than in the placebo group.
The adverse event profile was as expected from published
data (Anton et al., 2004; Karhuvaara et al., 2007; Mason
et al., 1994, 1999) and reflects the pharmacological profile of
nalmefene. There were more patients with serious adverse
events in the placebo group compared to the nalmefene
group; the incidence of treatment-emergent adverse events
leading to dropout was comparable between the groups.
Overall, as-needed use of nalmefene was safe and well
tolerated and no safety issues were raised in this study.
Secondary efficacy variables: γ-glutamyltransferase and alanine aminotransferase at week 24.
Efficacy variable PlaceboNalmefeneRatio to placebo
Mean Ratio 95% CI
γ-glutamyl transferase (IU/L)
Baseline (geometric mean)
Adjusted geometric mean at week 24
Alanine aminotransferase (IU/L)
Baseline (geometric mean)
Adjusted geometric mean at week 24
43.4 0.96[0.86; 1.08] 0.529
25.0 0.92 [0.84; 1.00]0.049
SD=Standard deviation; CI=confidence interval.
TEAEs leading to
TEAEs leading to dropout (≥0.5%)a
199 (59.1) 232 (68.0)
Data are numbers of patients (%).
TEAE=treatment-emergent adverse event.
aIn the main treatment period.
bIn the entire study period; percentages based on the all-
A. Gual et al. 1440
There are also limitations of this study. The main limita-
tion was the large non-specific treatment response, as
evident from the high proportion of patients (approximately
33%) that reduced their drinking prior to start of treatment,
before any intervention (medication and BRENDA). At
randomisation, these patients consumed such a small
amount of alcohol that there was little room for further
improvement, irrespective of treatment. This is a phenom-
enon that has been observed in other alcohol treatment
studies, including the recently published study by Mann
et al. (2012), and can indeed have an impact on study
outcome (Epstein et al., 2005; Litten et al., 2012). No doubt
motivational factors (readiness to change), expectancy and
natural course could explain why some patients self-
initiated a reduction in alcohol consumption immediately
after they had been informed about the study and con-
sented to participate and before they started on any
treatment intervention. When taking this non-specific treat-
ment response into account, and performing a post-hoc
analysis of the patients who did not reduce their drinking
prior to treatment, the effect of nalmefene was shown to
be statistically significant for both co-primary outcome
measures, thereby confirming the pharmacological effects
Secondly, the dropout rate in the current study was high,
but not very much higher than in another published 6-month
study in alcohol dependence (Garbutt et al., 2005) or in
studies of a shorter duration (Johnson et al., 2003; Anton
et al., 2006). As treatment adherence is a well-known
prerequisite for treatment success, nalmefene should only
be prescribed in conjunction with continuous psychosocial
support, focusing on motivation and treatment adherence.
Thirdly, the results from this study should be interpreted
in view of the fact that the study population was limited by
the selection criteria, e.g. patients with significant axis I co-
morbidity were excluded. However, this is directly in line
with the European Medicines Agency guideline (European
Medicines Agency, 2010).
In conclusion, meeting one of the two predefined co-
primary outcome measures in the total population, this
study supports the concept of reduction of alcohol con-
sumption with an as-needed use dosing regimen of nalme-
fene in patients with alcohol dependence that are unable to
reduce alcohol intake on their own.
Role of the funding source
The sponsor was involved in the study design, data collec-
tion, data analysis, and interpretation of the data, but not
in the decision to submit the report for publication. An
employee of the sponsor provided medical writing assis-
tance in the preparation of the report. The corresponding
author had full access to all study data and had final
responsibility for the decision to submit for publication.
Antoni Gual was the signatory investigator for the study. All authors
were involved in the design of the study, data analysis and
interpretation. Antoni Gual, Yuan He and Lars Torup wrote the
manuscript in collaboration with a medical writer. All authors
reviewed and approved the manuscript before submission.
Conflict of interest
Antoni Gual has received honoraria and travel grants from Lund-
beck, Janssen, D&A Pharma and Servier.
Yuan He and Lars Torup are Lundbeck employees.
Wim van den Brink has received honoraria from Lundbeck, Merck
Serono, Schering-Plough, Reckitt Benckiser, Pfizer, and Eli Lilly,
speaker fees from Lundbeck, investigator initiated industry grants
from Alkermes, Neurotech, and Eli Lilly, is a consultant to Lund-
beck, Merck Serono, Schering-Plough, and Teva, and has performed
paid expert testimony for Schering-Plough.
Karl Mann has received research grants from Schering-Plough,
Alkermes, Lundbeck, McNeil, and Merck. He has been a paid
consultant to Alkermes and Desitin, is a consultant to Lundbeck
and Pfizer, and has received speaker fees from Lundbeck.
We thank all patients for their participation in the ESENSE 2 Study,
all research staff and the ESENSE 2 Study Group (see Supplementary
material) for their contributions. We also thank Johan Hellsten, an
employee of Lundbeck for providing medical writing assistance in
the preparation, revision, and editing of the manuscript.
Appendix A.Supporting information
Supplementary data associated with this article can be
found in the online version at http://dx.doi.org/10.1016/
Adamson, S.J., Heather, N., Morton, V., Raistrick, D., UKATT
Research Team, 2010. Initial preference for drinking goal in
the treatment of alcohol problems: II. Treatment outcomes.
Alcohol Alcohol. 45, 136–142.
Alonso, J., Angermeyer, M.C., Bernert, S., Bruffaerts, R., Brugha, T.
S., Bryson, H., et al., 2004.Use of mental health services in
Europe: results from the European Study of the Epidemiology of
Mental Disorders (ESEMeD) project. Acta. Psychiatr. Scand.
Suppl. 109 (suppl 420), 47–54.
American Psychiatric Association (APA), 2000. Diagnostic and Sta-
tistical Manual of Mental Disorders, Fourth Ed., Text Revision
(DSM-IV-TR). American Psychiatric Association, Washington DC.
Anton, R.F., Pettinati, H., Zweben, A., Kranzler, H.R., Johnson, B.,
Bohn, M.J., et al., 2004.A multi-site dose ranging study of
nalmefene in the treatment of alcohol dependence. J. Clin.
Psychopharmacol. 24, 421–428.
Anton, R.F., O'Malley, S.S., Ciraulo, D.A., Cisler, R.A., Couper, D.,
Donovan, D.M., et al., 2006.Combined pharmacotherapies and
behavioral interventions for alcohol dependence: the COMBINE
study: a randomized controlled trial. J. Am. Med. Assoc. 295,
Epstein, E.E., Drapkin, M.L., Yusko, D.A., Cook, S.M., McCrady, B.
S., Jensen, N.K., 2005. Is alcohol assessment therapeutic?
Pretreatment change in drinking among alcohol-dependent
women. J. Stud. Alcohol 66, 369–378.
European Medicines Agency (EMEA), 2010. Guideline on the devel-
opment of medicinal products for the treatment of alcohol
1441Nalmefene, as-needed use, in patients with alcohol dependence
Garbutt, J.C., Kranzler, H.R., O'Malley, S.S., Gastfriend, D.R.,
Pettinati, H.M., Silverman, B.L, for the Vivitrex Study Group,
2005. Efficacy and tolerability of long-acting injectable naltrex-
one for alcohol dependence a randomized controlled trial. J.
Am. Med. Assoc. 293, 1617–1625.
Gastfriend, D.R., Garbutt, J.C., Pettinati, H.M., Forman, R.F.,
2007. Reduction in heavy drinking as a treatment outcome in
alcohol dependence. J. Subst. Abuse Treat. 33, 71–80.
Guy, W. (Ed.), National Institute of Mental Health, Rockville, MD
Publication no. 76-338.
Heather, N., Adamson, S.J., Raistrick, D., Slegg, G.P., UKATT
Research Team, 2010. Initial preference for drinking goal in
the treatment of alcohol problems: I. Baseline differences
between abstinence and non-abstinence groups. Alcohol Alco-
hol. 45, 128–135.
Heinälä, P., Alho, H., Kiianmaa, K., Lönnqvist, J., Kuoppasalmi, K.,
Sinclair, J.D., 2001. Targeted use of naltrexone without prior
detoxification in the treatment of alcohol dependence: a
factorial double-blind, placebo-controlled trial. J. Clin. Psycho-
pharmacol. 21, 287–292.
Johnson, B.A., Ait-Daoud, N., Bowden, C.L., DiClemente, C.C.,
Roache, J.D., Lawson, K., et al., 2003.Oral topiramate for
treatment of alcohol dependence: a randomised controlled
trial. Lancet 361, 1677–1685.
Karhuvaara, S., Simojoki, K., Virta, A., Rosberg, M., Löyttyniemi,
E., Nurminen, T., et al., 2007.Targeted nalmefene with simple
medical management in the treatment of heavy drinkers: a
randomized double-blind placebo-controlled multicenter study.
Alcohol. Clin. Exp. Res. 31, 1179–1187.
Kranzler, H.R., Tennen, H., Penta, C., Bohn, M.J., 1997. Targeted
naltrexone treatment of early problem drinkers. Addict. Behav.
Kranzler, H.R., Armeli, S., Tennen, H., Blomqvist, O., Oncken, C.,
Petry, N., et al., 2003.Targeted naltrexone for early problem
drinkers. J. Clin. Psychopharmacol. 23, 294–304.
Kranzler, H.R., Tennen, H., Armeli, S., Chan, G., Covault, J., Arias,
A., et al., 2009.Targeted naltrexone for problem drinkers. J.
Clin. Psychopharmacol. 29, 350–357.
Laaksonen, E., Koski-Jännes, A., Salaspuro, M., Ahtinen, H., Alho,
H., 2008. A randomized, multicentre, open-label, comparative
trial of disulfiram, naltrexone and acamprosate in the treatment
of alcohol dependence. Alcohol Alcohol. 43, 53–61.
Lecrubier, Y., Sheehan, D.V., Weiller, E., Amorim, P., Bonora, I.,
Harnett Sheehan, K., et al., 1997.The Mini International Neu-
ropsychiatric Interview (MINI). A short diagnostic structured
interview: reliability and validity according to the CIDI. Eur.
Psychiatry 12, 224–231.
Litten, R.Z., Fertig, J.B., Falk, D.E., Ryan, M.L., Mattson, M.E.,
Collins, J.F., et al., 2012.A double-blind, placebo controlled trial
to assess the efficacy of Quetiapine Fumarate XR in very heavy-
drinking alcohol dependent patients. Alcohol. Clin. Exp. Res. 36,
Little, R, Yau, L., 1996. Intent-to-treat analysis for longitudinal
studies with drop-outs. Biometrics 52, 1324–1333.
Luquiens, A., Reynaud, M., Aubin, H.J., 2011. Is controlled drinking
an acceptable goal in the treatment of alcohol dependence? A
survey of French alcohol specialists. Alcohol Alcohol. 46,
Mann, K., Lehert, P ., Morgan, M.Y., 2004. The efficacy of acamprosate in
the maintenance of abstinence in alcohol-dependent individuals:
results of a meta-analysis. Alcohol. Clin. Exp. Res. 28, 51–63.
Mann, K., Bladström, A., Torup, L., Gual, A., van den Brink, W.,
2013. Extending the treatment options in alcohol dependence: a
randomized controlled study of as-needed nalmefene. Biol.
Psychiatry 73, 706–713.
Marlatt, G.A., Witkiewitz, K., 2002. Harm reduction approaches to
alcohol use: health promotion, prevention, and treatment.
Addict. Behav. 27, 867–886.
Mason, B.J., Ritvo, E.C., Morgan, R.O., Salvato, F.R., Goldberg, G.,
Welch, B., et al., 1994.A double-blind, placebo-controlled pilot
study to evaluate the efficacy and safety of oral nalmefene HCl
for alcohol dependence. Alcohol. Clin. Exp. Res. 18, 1162–1167.
Mason, B.J., Salvato, F.R., Williams, L.D., Ritvo, E.C., Cutler, R.B.,
1999. A double-blind, placebo-controlled study of oral nalme-
fene for alcohol dependence. Arch. Gen. Psychiatry 56, 719–724.
Merkx, M.J., Schippers, G.M., Koeter, M.W., Vuijk, P.J., Oudejans,
S.C., Stam, R.K., et al., 2011.Guidelines for allocating out-
patient alcohol abusers to levels of care: predictive validity.
Addict. Behav. 36, 570–575.
Miller, W.R., Tonigan, J.S., Longabaugh, R., 1995. The Drinker
Inventory of Consequences (DrInC): An Instrument for Assessing
Adverse Consequences of Alcohol Abuse. Project MATCH Mono-
graph Series. National Institute on Alcohol Abuse and Alcohol-
ism, Rockville, MD.
Miller, W.R., Walters, S.T., Bennett, M.E., 2001. How effective is
alcoholism treatment in the United States? J. Stud. Alcohol 62,
Project MATCH Research Group, 1998. Matching alcoholism treat-
ments to client heterogeneity: project MATCH three-year drink-
ing outcomes. Alcohol. Clin. Exp. Res. 22, 1300–1311.
Rehm, J., Mathers, C., Popova, S., Thavorncharoensap, M., Teer-
awattananon, Y., Patra, J., 2009. Global burden of disease and
injury and economic cost attributable to alcohol use and
alcohol-use disorders. Lancet 373, 2223–2233.
Rehm, J., Baliunas, D., Borges, G.L., Graham, K., Irving, H., Kehoe,
T., et al., 2010.The relation between different dimensions of
alcohol consumption and burden of disease: an overview.
Addiction 105, 817–843.
Rehm, J., Zatonksi, W., Taylor, B., Anderson, P., 2011. Epidemiology
and alcohol policy in Europe. Addiction 106 (Suppl 1), 11–19.
Rehm, J., Shield, K.D., Rehm, M.X., Gmel, G., Frick, U., 2013.
Modelling the impact of alcohol dependence on mortality burden
and the effect of available treatment interventions in the
European Union. Eur. Neuropsychopharmacol. 23, 89–97.
Sinclair, J.D., 2001. Evidence about the use of naltrexone and for
different ways of using it in the treatment of alcoholism. Alcohol
Alcohol. 36, 2–10.
Skinner, H.A., Horn, J.L., 1984. Alcohol Dependence Scale: Users
Guide. Addiction Research Foundation, Toronto.
Sobell, L.C., Sobell, M.B., 1992. Timeline Follow-back: a technique
for assessing self-reported ethanol consumption. In: Litten, R.Z.,
Allen, J.P . (Eds.), Measuring Alcohol Consumption: Psychosocial
and Biological Method. Humana Press, Totowa, NJ, pp 41–72.
Starosta, A.N., Leeman, R.F ., Volpicelli, J.R., 2006. The BRENDA model:
integrating psychosocial treatment and pharmacotherapy for the
treatment of alcohol use disorders. J. Psychiatr. Pract. 12, 80–89.
Sullivan, J.T., Sykora, K., Schneiderman, J., Naranjo, C.A., Sellers,
E.M., 1989. Assessment of alcohol withdrawal: the revised
clinical institute withdrawal assessment for alcohol scale
(CIWA-Ar). Br. J. Addict. 84, 1353–1357.
Volpicelli, J.R., Pettinati, H.M., McLellan, A.T., O'Brien, C.P., 2001.
Combining Medication and Psychosocial Treatments for Addic-
tions; the BRENDA Approach. The Guilford Press, New York, NY.
Willenbring, M.L., Massey, S.H., Gardner, M.B., 2009. Helping
patients who drink too much: an evidence-based guide for
primary care clinicians. Am. Fam. Physician 80, 44–50.
Wittchen, H.U., Jacobi, F., Rehm, J., Gustavsson, A., Svensson, M.,
Jönsson, B., et al., 2011.The size and burden of mental
disorders and other disorders of the brain in Europe 2010. Eur.
Neuropsychopharmacol. 21, 655–679.
ing Alcohol Consumption and Related Harm. WHO/MSD/MSB/00.4.
World Health Organisation (WHO), 2010. Global Strategy to Reduce
the Harmful use of Alcohol.
World Health Organisation (WHO), 2011. Global Status Report on
Alcohol and Health.
A. Gual et al.1442