OUP user menu


DOI: http://dx.doi.org/10.1093/alcalc/agh093 542-547 First published online: 29 September 2004


Aims: Relapse prevention treatment with both acamprosate and naltrexone has been shown to be efficacious in the treatment of alcoholism. Whereas both compounds act pharmacologically differently, there is up to now only limited evidence as to whether combined treatment is efficacious and pharmacologically safe. It remains to be answered whether data justify the combination of both drugs in clinical practice. Methods: Review of the three pre-clinical and four clinical studies that have been published to date on either combined tolerability or efficacy. Results: Data available up to now show no occurrence of severe adverse events during combined treatment. Diarrhoea and nausea were shown to be the most significant side-effects. Whereas pre-clinical studies regarding efficacy of combined treatment are not yet conclusive, clinical data show the superiority of combined treatment compared with both placebo and acamprosate monotherapy. The synergistic effect of combined treatment remained after 12 weeks of drug-free follow-up. Conclusions: The combination of acamprosate with naltrexone in a clinical sample seems to be efficacious and safe. Numerous alcohol dependent patients could benefit, particularly those that responded insufficiently on monotherapeutic treatment with either acamprosate or naltrexone.

(Received 9 June 2004; first review notified 28 July 2004; in revised form 8 August 2004; accepted 8 August 2004)


The last decade has seen the accumulation of an extensive body of clinical trial data indicating that both acamprosate and naltrexone are of use in the treatment of alcohol dependence. The vast majority of randomized placebo-controlled clinical trials with acamprosate have demonstrated that this drug significantly increases the proportion of patients remaining abstinent after acute detoxification (Mason, 2001; Mann et al., 2004). The most reproducible finding with naltrexone has been that it reduces relapse into heavy drinking (usually defined as five or more drinks a day; Kranzler and Van Kirk, 2001; Streeton and Wheelan, 2001). Certain studies have shown that naltrexone reduces the desire to drink in social drinkers and in non-abstinent and abstinent alcohol dependent subjects. However, neither naltrexone nor acamprosate allow every single detoxified alcohol dependent patient to achieve durable abstinence from alcohol. For this reason, it has been suggested that a combination of the two pharmacotherapeutic approaches may provide greater benefit than either alone. During the last 3 years, three preclinical and four clinical studies on pharmacodynamic and pharmacokinetic aspects of combined treatment have been published (Table 1). These studies provide evidence for beneficial effects of the co-administration of acamprosate and naltrexone in the treatment of alcohol addiction. This review will clarify the background and results of these studies in order to discuss clinical implications.

View this table:
Table 1.

Trials on combined treatment with acamprosate (ACP) and naltrexone (NTX)

Type of trialReferenceSubjectsOutcome
Pre-clinicalStromberg et al. (2001)24 male Wistar rats following ethanol habituation (during limited access)Reduction of ethanol intake: NTX > ACP; NTX + ACP > ACP; NTX + ACP = NTX (both dosages)
Heyser et al. (2003)117 male Wistar rats following ethanol habituation (after 5 days of ethanol deprivation)Reduction of ethanol intake: NTX > PLA; NTX + ACP > NTX (day 1 and dependent on dosages used)
Kim et al. (2004)42 male C57BL/6 mice following ethanol habituation (during limited access)Reduction of ethanol intake: NTX > PLA; NTX + ACP > PLA (dependent on dosages used)
ClinicalMason et al. (2002)24 healthy volunteers; drug interaction studyHigh tolerability of NTX + ACP; no negative interactions on measures of safety and cognitive function
Johnson et al. (2003)23 non-treatment-seeking alcohol dependent subjects; drug interaction studyInfrequent moderate adverse events following NTX + ACP: anger, depression, somnolence, nervousness, diarrhoea, headache
COMBINE study group (2005)108 alcohol dependent subjects; feasibility studyHigh tolerability of NTX + ACP; no negative interactions on measures of toxicity and physical complaints
Kiefer et al. (2003a)160 alcohol dependent subjects; treatment trialAbstinence duration: NTX > PLA; ACP > PLA; NTX + ACP > PLA; NTX + ACP > ACP; adverse events following NTX + ACP: diarrhoea, nausea
  • PLA = placebo.


The reasons for attempting combination therapy with acamprosate and naltrexone are theoretical and pragmatic. From a theoretical point of view, the two drugs have different pharmacological mechanisms of action. Naltrexone is an antagonist at opiate receptors in the central nervous system that mediate the rewarding effects of opioid peptides. These receptors are associated with the dopaminergic neurones in the ventral tegmental area and their projection areas in the limbic system which are believed to be critical for the manifestation of reward-directed behaviours. Naltrexone is thus thought to attenuate the rewarding nature of addictive drugs, including alcohol. Although less well-defined, the mechanism of action of acamprosate appears to involve indirect inhibition of N-methyl-d-aspartate (NMDA) receptor-mediated glutamatergic neurotransmission. These receptors become up-regulated following chronic alcohol intake, and contribute to central nervous system hyperexcitability when alcohol is no longer present in the brain. This may lead to unpleasant feelings of tension, anxiety and dysphoria, which may be attenuated by acamprosate.

One of the drivers for relapse in abstinent alcohol dependent patients is craving to drink. Different aspects of craving have been postulated (Anton et al., 1995; Verheul et al., 1999), which reflect different aspects of the neurobiology of alcohol dependence. Widely accepted are the notions of reward (or positive) craving which arises as a conditioned cue to drink in order to experience the hedonic effects of alcohol, and relief (or negative) craving, a conditioned cue to avoid the unpleasant effects associated with the absence of alcohol (Fig. 1).

Figure 1.

Dynamics of alcohol craving and consumption, indicating proposed neurophysiological mechanisms and possible points of impact of naltrexone and acamprosate.

The first of these is believed to involve activation of dopaminergic and opioidergic pathways in the midbrain, and the second modulation of the balance between excitatory glutamatergic activity and inhibitory gamma-aminobutyric acid (GABAergic) activity in the cortex. In line with this, it might be expected that naltrexone would specifically attenuate reward craving while acamprosate would diminish relief craving. The combination of the two drugs would then act simultaneously on two different aspects of craving and have a more incisive effect on the risk of relapse than either treatment alone.

In addition, acamprosate and naltrexone might act on different aspects of drinking behaviour. The former seems to increase the probability that currently abstinent subjects remain abstinent, without necessarily affecting the craving for alcohol, whereas the latter might reduce the quantity of alcohol ingested, by attenuating the priming effect of initial alcohol intake or alcohol-associated cues.

From a pragmatic point of view, both drugs are on the whole well tolerated and have no propensity for potentially dangerous drug–drug interactions. Meta-analyses revealed effect sizes between 0.25 and 0.30 for both compounds (Berglund et al., 2003; Mann et al., 2004). Hence, neither treatment shows an effect size that could not be ameliorated.


The utility of combining acamprosate and naltrexone has been assessed in various animal models of alcoholism. Stromberg et al. (2001) used a limited access paradigm in rat strains outbred to drink ethanol solutions spontaneously, and found robust decreases in alcohol consumption after administration of naltrexone. The effects of acamprosate were more modest and only observed at high doses (200 mg/kg), and no evidence for additive effects of the two treatments were seen. However, this experimental paradigm is more a model of consumption in active alcohol dependence rather than a model of consumption following withdrawal. The latter phenomenon has been studied by Heyser et al. (2003), using a model of ethanol deprivation wherein rats previously rendered alcohol dependent undergo an imposed period of abstinence and are then re-exposed to alcohol (Heyser et al., 1997; Spanagel and Holter, 1999). Under these conditions, rats 'over-consume' during the re-exposure period compared to pre-abstinent conditions. In this model, both acamprosate and naltrexone attenuate rebound drinking, and the association of acamprosate (25 mg/kg as well as 100 mg/kg) with a high dose of naltrexone (0.125 mg/kg) produces a significantly greater reduction in rebound drinking at post deprivation day 1 than does the same dose of naltrexone given alone. These data are in support of the notion that naltrexone and acamprosate may be more effective in combination. The third study, published by Kim et al. (2004), also applied a limited access alcohol model to C57BL/6 mice. Whereas naltrexone monotherapy reduced alcohol intake at sub-maximal (1.0 mg/kg) but not at low dosages (0.025 mg/kg), co-administration of acamprosate (both 50 mg/kg and 200 mg/kg) resulted in reduced alcohol intake across both dosages of naltrexone. However, acamprosate was not effective immediately, but only after 8 days of pre-treatment.

The differences between the results of these studies, especially regarding the efficacy of acamprosate, might be attributable to procedural differences among the experiments. Whereas the study of Heyser et al. (2003) administered acamprosate twice a day for 5 days during ethanol deprivation (25, 100 or 200 mg/kg) before giving access to alcohol using an operant lever press procedure, Stromberg et al. (2001) exposed rats to 4 days acamprosate once a day (50 and 200 mg) without pre-testing alcohol-free days. Kim et al. (2004) administed acamprosate for 10 days (50 and 200 mg) during which half of the daily dosage was administered 12 h before the start of the limited access session and the other half 30 min before the start of the limited access session. Thus it seems possible that procedural differences (duration of pre-treatment, deprivation phase, application procedure) might be responsible for the reduced efficacy of acamprosate in some studies and, as a consequence, a lack of clearly detectable synergistic effect in combination with naltrexone.

However, taken together, pre-clinical data do provide some evidence that the co-administration of acamprosate and naltrexone might reduce ethanol intake more effectively than either on its own.


Three human studies have assessed the safety of combining acamprosate and naltrexone. The first of these was a combined pharmacodynamic and pharmacokinetic drug interaction study performed in 24 healthy adult volunteers (Mason et al., 2002). The study used a randomized, double-blind, cross-over design in which standard doses of acamprosate (2 g/day) and naltrexone (50 mg/day) and a combination of both were compared within the same subjects. Each treatment was given for 7 days, separated by a 7-day wash-out period.

It was observed that co-administration of the two drugs significantly increased the rate and extent of absorption of acamprosate and thus the extent of exposure. The area under the curve of acamprosate plasma concentration was increased by 25%. This observation is presumably explained by a decreased rate of gastric emptying produced by naltrexone. On the other hand, there was no change in the plasma concentrations of either naltrexone or its principal metabolite 6-β-naltrexol compared to naltrexone given alone. All treatments were tolerated and there was no impact of combined treatment on a panel of tests of cognitive function.

The second study was a randomized, placebo-controlled, double-blind, cross-over study performed in 23 non-treatment-seeking, alcohol dependent patients (Johnson et al., 2003). The study examined kinetic and dynamic factors to determine the pharmacological and behavioural safety and tolerability of low versus high doses of naltrexone (50 mg/day vs 100 mg/day) and acamprosate (2 g/day vs 3 g/day), both independently and combined. A placebo washout preceded treatment with low- or high-dose naltrexone or acamprosate. Thereafter, the alternative medication type at its lower and higher doses, respectively, was administered with continuation of the first medication. Predetermined behavioural, performance and pharmacological criteria determined significant pathological change from baseline. Significant increases in symptoms from baseline with monotherapy included nervousness and fatigue with 3 g acamprosate and somnolence and headache with 50 mg and 100 mg naltrexone, respectively. Combined treatment at various doses evinced anger, depression, somnolence, nervousness, diarrhoea and headache. For all but one subject who dropped out, increased symptoms did not produce any remarkable clinical deterioration. Naltrexone administration significantly again increased plasma acamprosate levels. Naltrexone and acamprosate, both alone and in combination at the tested doses, were behaviourally and pharmacologically safe. Adverse events were infrequent, of moderate intensity, and resolved with reassurance and symptomatic treatment. More side effects were noted with the combination of medications than with either medication alone.

The third study corresponds to a preliminary safety analysis from the COMBINE study. This is a large randomized, placebo-controlled, double-blind, study in alcohol dependent patients in North America designed to assess the efficacy of different combinations of naltrexone, acamprosate and behavioural interventions in promoting abstinence. The ongoing study is planned to include 1375 patients. Analyses of the safety data from an initial sample of 108 patients have been published (COMBINE Study Group, 2003) including retention, adherence to study parameters and medication, physical complaints and physiologic toxicity. In addition to pharmacological treatment with naltrexone and/or acamprosate, individuals received Medical Management (MM) provided by a health care practitioner alone or in combination with an enhanced behavioural intervention, Combined Behavioural Intervention (CBI) delivered by a trained therapist. A final group received CBI alone without pills. All participants were treated and assessed for a maximum of 16 weeks.

The attendance at therapy and research visits, and medication adherence and tolerability were good with no statistical differences between the medication and behavioural intervention groups. Over 75% of participants completed the week 16/end-of-study assessment and the average medication adherence (percent of total pills taken) was ∼65%. The level and types of physical complaints were similar among the groups as was the incidence of patient-reported adverse events. An observation that during combined treatment more subjects experienced diarrhoea (75% compared with 59% in the placebo group) was statistically insignificant. Plasma concentrations of liver enzymes (AST and ALT), bilirubin and creatinine did not differ significantly between the pharmacological combination and monotherapy groups. No unexpected adverse events were reported in patients with combined acamprosate and naltrexone.


Kiefer et al. (2003a) performed a placebo-controlled, double-blind randomized clinical trial comparing treatment with acamprosate, naltrexone and the combination of both. A double dummy treatment assignment, whereby every patient took seven capsules a day, allowed an effective double-blind to be maintained. This study included 160 individuals with alcohol dependence as defined in DSM-IV admitted for acute detoxification as an inpatient. Patients with present or past abuse or dependence on substances other than nicotine or alcohol, those currently taking any psychotropic medication and those with a history of psychosis were excluded. Patients were treated with the study medication for 3 months. They were also provided with 90 min of behavioural group psychotherapy on a weekly basis. Data on drinking behaviour were obtained weekly using drink diaries filled in by the patient, which were checked for their validity by reference to laboratory parameters, breath alcohol control, and reports of relatives and general practitioners (GPs). The main outcome measure was time to relapse into heavy drinking (defined as ≥5 drinks per day for males or ≥4 drinks for females), as well as occurrence of 5 drinking days within 1 week (1 drink = 13 g ethanol), time to first drink, and cumulative abstinence duration. At the end of the 3 month study period, patients discontinued medication and entered a further 3-month open-label follow-up period, at the end of which the drinking status of all participants was re-assessed.

It was observed that mean time to first relapse into heavy drinking was significantly longer (P < 0.05) in all three active treatment groups compared to placebo. There was no significant difference between the naltrexone alone and acamprosate alone treatment groups. However, time to first relapse was longer in the combination treatment group than in the acamprosate alone group. Similar differences emerged when the time to first drink was analysed, as well as in the proportion of patients who had relapsed at the study end (placebo: 75%; acamprosate: 50%; naltrexone: 35%; combined treatment: 27%). Even though further relapse occurred during the follow-up period, the relative treatment benefits between the three active treatment groups and placebo were maintained at the end of the 3-month open label phase (placebo: 80%; acamprosate: 54%; naltrexone: 53%; combined treatment: 34%; Kiefer et al., 2003b). At this time, there was no statistically significant difference between the three active treatment groups (naltrexone alone, acamprosate alone, and combination). With respect to the emergence of adverse events, combination therapy was generally well tolerated, with no unexpected novel side-effects noted. However, the incidence of diarrhoea (13.8% per visit) and nausea (5.6% per visit) was significantly greater in the combination group than with monotherapy.


There are several possible explanations for the superior efficacy of combination treatment. Firstly, it may be that subgroups of patients exist who respond selectively to acamprosate or naltrexone. In this hypothesis, the added benefit of combination therapy would be explained by the recruitment of additional responder patients. It is unlikely that all patients respond to only one of these drugs, since otherwise a clear additive effect of the treatments would be expected, which is not observed. However, even if a minority of patients respond preferentially to naltrexone or to acamprosate, this could explain our findings. Such patient subgroups may be characterized, for example, by dominance of reward craving in the urge to drink in the hypothetical ‘naltrexone responders’ and of relief craving in the hypothetical ‘acamprosate responders’. However, in our study, exploratory analyses of the subgroups ‘reward cravers’ vs ‘relief cravers’ (based on an open interview at baseline on motivational factors of alcohol intake) showed no differential treatment interaction. A new and more promising attempt to differentiate positive (reward) craving from negative (relief-) craving is based on the measurement of the affect-modulated eyeblink startle reflex during confrontation with alcohol-associated visual stimuli (Heinz et al., 2003). Additional factors that have been shown to be possibly useful to predict efficacy of pharmacological anti-craving treatment are: (i) for naltrexone: high levels of baseline craving, somatic distress, and anxiety (Volpicelli et al., 1995) high baseline depression and female gender (Kiefer et al., 2004); and (ii) for acamprosate: according to the typological differentiation of chronic alcoholism (Lesch et al., 2001), affiliation with type I (Whitworth et al., 1996; Kiefer et al., 2004) and type II (Whitworth et al., 1996); and a low scoring for items of somatic distress (Kiefer et al., 2004). Johnson et al. (2000) suggested that, in general, pharmacological anti-craving treatment might be expected to be most efficacious in early-onset patients. However, larger prospective studies are needed to evaluate the possibly predictive capacity of these factors in single or combined use. One has to bear in mind that for the majority of other complex bio-psycho-social disorders (e.g. depressive or psychotic disorders) a matching of distinct pharmacotherapeutic strategies with biological or psychosocial variables has had little success so far.

A second hypothesis for the synergistic effects of acamprosate–naltrexone co-administration would be that the combination produces a more incisive anti-craving effect in each individual patient. Such a synergy could result from the two drugs interfering with distinct biological aspects of the craving process (reward and relief craving). If this hypothesis is true, we would not expect to find distinct populations who respond preferentially to either one or other of the two drugs.

Thirdly, pharmacokinetic interaction might underly the observed treatment benefit, whereby the bioavailability of one or both drugs might be enhanced by co-administration of the other drug. The observation that plasma levels of acamprosate are increased by ∼33% by co-administration of naltrexone (Mason et al., 2002; Johnson et al., 2003) favours this hypothesis.


Studies to date suggest higher efficacy of combined treatment with acamprosate and naltrexone compared to monotherapy, regarding prevention of relapse into heavy drinking and maintenance of abstinence. This added benefit could be explained by subgroups of drug-specific responders, by synergistic effects within each patient, or by a pharmacokinetic interaction. The explorative assessment of potentially discriminating features (nature and intensity of craving, psychopathology, and typology) gives hints on possibly useful differentiations. However, prospective studies with larger sample size are warranted to evaluate the value of matching these subgroups with specific pharmacological interventions.

Combination treatment appears to be well tolerated, with no severe adverse events reported, although they imply an increased incidence of diarrhoea and nausea, perhaps due to a pharmacokinetic interaction. It can be concluded that there is accumulating evidence that the combination of acamprosate and naltrexone is both efficacious and safe. Given the large proportion of alcohol dependent subjects responding insufficiently to monotherapy with either acamprosate or naltrexone with the consequence of early relapse after detoxification, many patients might benefit from enhancing the efficacy of relapse-prevention treatment by combining acamprosate and naltrexone.


View Abstract