Over the last 20 years, the role of adjuvant pharmacotherapy in optimising outcome in rehabilitation programmes for alcohol-dependent patients has become increasingly evident. New avenues for rational drug treatment have arisen from better understanding of the neurobiological substrates of alcohol dependence, including adaptive changes in amino acid neurotransmitter systems, stimulation of dopamine and opioid peptide systems, and, possibly, changes in serotonergic activity.Disulfiram, naltrexone and acamprosate are currently the only treatments approved for the management of alcohol dependence. However, there is still no unequivocal evidence from randomised controlled clinical trials that disulfiram improves abstinence rates over the long term. Aversive therapy with disulfiram is not without risk for certain patients, and should be closely supervised. Both naltrexone and acamprosate improve outcome in rehabilitation of alcohol-dependent patients, but seem to act on different aspects of drinking pathology. Naltrexone is thought to decrease relapse to heavy drinking by attenuating the rewarding effects of alcohol. However, data from the naltrexone clinical trial programme are somewhat inconsistent, with several large studies being negative. Acamprosate is believed to maintain abstinence by blocking the negative craving that alcohol-dependent patients experience in the absence of alcohol. The clinical development programme has involved a large number of patients and studies, of which the vast majority have shown a beneficial effect of acamprosate on increasing abstinence rates. Both drugs are generally well tolerated; nausea is reported by around 10% of patients treated with naltrexone, while the most frequent adverse effect reported with acamprosate is diarrhoea.Another opioid receptor antagonist, nalmefene, has shown promising activity in pilot studies, and may have a similar profile to naltrexone. Data from studies of SSRIs in alcohol dependence are somewhat heterogeneous, but it appears that these drugs may indirectly improve outcome by treating underlying depression rather than affecting drinking behaviourper se. Similarly, the anxiolytic buspirone may act by ameliorating underlying psychiatric pathology. Dopaminergic neuroleptics, benzodiazepines and antimanic drugs have not yet demonstrated evidence of activity in large controlled clinical trials. Trials with drugs acting at serotonin receptors have yielded disappointing results, with the possible exception of ondansetron.Because the biological basis of alcohol dependence appears to be multifactorial, the future of management of alcoholism may be combination therapy, using drugs acting on different neuronal pathways, such as acamprosate and naltrexone. Pharmacotherapy should be used in association with appropriate psychosocial support and specific treatment provided for any underlying psychiatric comorbidities.