Cocaine abuse is a serious health problem in many areas of the world, yet there are no proven effective medications for the treatment of cocaine dependence. Preclinical studies suggest that the reinforcing effect of cocaine that promotes its abuse is mediated by blockade of the presynaptic dopamine transporter. This results in increased dopamine activity in the mesolimbic or meso-accumbens dopamine reward system of brain. Development of new medications to treat cocaine dependence has focused on manipulation of this dopamine system, either by direct action on dopamine binding sites (transporter or receptors) or indirectly by affecting other neurotransmitter systems that modulate the dopamine system. In principle, a medication could act via one of three mechanisms: (i) as a substitute for cocaine by producing similar dopamine effects; (ii) as a cocaine antagonist by blocking the binding of cocaine to the dopamine transporter; or (iii) as a modulator of cocaine effects by acting at other than the cocaine binding site. The US National Institute on Drug Abuse has a Clinical Research Efficacy Screening Trial (CREST) programme to rapidly screen existing medications. CREST identified four medications warranting phase II controlled clinical trials: cabergoline, reserpine, sertraline and tiagabine. In addition, disulfiram and selegiline (deprenyl) have been effective and well tolerated in phase II trials. However, selegiline was found ineffective in a recent phase III trial. Promising existing medications probably act via the first or third aforementioned mechanisms. Sustained-release formulations of stimulants such as methylphenidate and amfetamine (amphetamine) have shown promise in a stimulant substitution approach. Disulfiram and selegiline increase brain dopamine concentrations by inhibition of dopamine-catabolising enzymes (dopamine-beta-hydroxylase and monoamine oxidase B, respectively). Cabergoline is a direct dopamine receptor agonist, while reserpine depletes presynaptic stores of dopamine (as well as norepinephrine and serotonin). Sertraline, baclofen and vigabatrin indirectly reduce dopamine activity by increasing activity of neurotransmitters (serotonin and GABA) that inhibit dopamine activity. Promising new medications act via the second or third aforementioned mechanisms. Vanoxerine is a long-acting inhibitor of the dopamine transporter which blocks cocaine binding and reduces cocaine self-administration in animals. Two dopamine receptor ligands that reduce cocaine self-administration in animals are also undergoing phase I human safety trials. Adrogolide is a selective dopamine D(1) receptor agonist; BP 897 is a D(3) receptor partial agonist.A pharmacokinetic approach to treatment would block the entry of cocaine into the brain or enhance its catabolism so that less cocaine reached its site of action. This is being explored in animals using the natural cocaine-metabolising enzyme butyrylcholinesterase (or recombinant versions with enhanced capabilities), catalytic antibodies, and passive or active immunisation to produce anti-cocaine binding antibodies. A recent phase I trial of a "cocaine vaccine" found it to be well tolerated and producing detectable levels of anti-cocaine antibodies for up to 9 months after immunisation.