Cocaine, already a significant drug problem in North and South America, has become a more prominent part of the European drug scene. Cocaine dependence has major somatic, psychological, psychiatric, socio-economic, and legal implications. No specific effective pharmacological treatment exists for cocaine dependence. Recent advances in neurobiology have identified various neuronal mechanisms implicated in cocaine addiction and suggested several promising pharmacological approaches. Data were obtained from Medline, EMBASE, and PsycINFO searches of English-language articles published between 1985 and June 2007 using the key words: cocaine, addiction, cocaine dependence, clinical trials, pharmacotherapy(ies) singly and in combination. Large well-controlled studies with appropriate statistical methods were preferred. Pharmacological agents such as GABA agents (topiramate, tiagabine, baclofen and vigabatrin) and agonist replacement agents (modafinil, disulfiram, methylphenidate) seem to be the most promising in treatment of cocaine dependence. The results from trials of first- and second-generation neuroleptics are largely negative. Aripiprazole, a partial dopaminergic agonist that may modulate the serotonergic system, shows some promise. Preliminary results of human studies with anti-cocaine vaccine, N-acetylcysteine, and ondansetron, are promising, as are several compounds in preclinical development. While no medication has received regulatory approval for the treatment of cocaine dependence, several medications marketed for other indications have shown efficacy in clinical trials. An anti-cocaine vaccine and several compounds in preclinical development have also shown promise. Findings from early clinical trials must be confirmed in larger, less selective patient populations.
Cocaine dependence is a chronic, relapsing disorder characterized by compulsive drug-seeking and drug use despite negative consequences (Dackis and O'Brien, 2001; Goldstein and Volkow, 2002). Cocaine dependence is a significant worldwide public health problem with somatic, psychological, socio-economic, and legal complications. Associated public health risks include infectious diseases (viral hepatitis, HIV) (Friedman et al., 2006) and increased crime and violence (Hoaken and Stewart, 2003).
This psychostimulant has become a noticeable part of the European drug scene. The European Monitoring Centre for Drugs and Drug Abuse estimates that about 3.5 million Europeans (approximately 1% of the adult population) had used cocaine at least once during the past year, a substantial increase over the past decade. Only 13% of lifetime cocaine users had used cocaine in the previous month (EMCDDA, 2006), suggesting that many users do not become chronic users. Cocaine use varies widely by country, with the highest past year rates in Spain (2.7%) and the United Kingdom (2%), comparable to the 2.4% past year use rate in the USA in 2004 (Substance Abuse and Mental Health Services Administration, 2005). As with most illegal drugs, the highest rates of cocaine use are among young males aged 15–24 years. Cocaine use has become the third commonest reason (after opiate and cannabis use) for patients in the European Union to enter drug abuse treatment, accounting for approximately 8% of all treatment admissions in 2004 (EMCDDA, 2006).
Almost all psychoactive drugs causing addiction in humans activate the meso-cortico-limbic system by increasing dopamine release within the nucleus accumbens (Cami and Farre, 2003; Koob and Le Moal, 2001; Nestler, 2001; Wise, 2002). This dopamine system plays a critical role in motivational, emotional, contextual, and affective information processing of behaviour and drug reinforcement mechanisms. Cocaine directly increases synaptic dopamine levels in this meso-cortico-limbic system by blocking the transporter that pumps dopamine out of the synapse into the presynaptic nerve terminal. There is strong evidence from animal and human studies (including brain imaging by positron emission tomography; PET) that increased dopamine transmission contributes significantly to the reinforcing effects of cocaine (Volkow et al., 2006). However, the original hypothesis of a simple dysfunction of the meso-cortico-limbic dopaminergic system is no longer adequate to explain all aspects of cocaine addiction. Cocaine also blocks the serotonin and norepinephrine presynaptic transporters. The rewarding cocaine subjective experience (‘high’) is mediated by increases in both synaptic dopamine and norepinephrine levels (Filip et al., 2005; Wee et al., 2006). Cocaine withdrawal alters function of both serotonin and dopamine neurons. All three of these biogenic amine transporters are targets of interest for the development of therapeutic compounds.
Cocaine influences other neurotransmitter systems including glutamate, GABA, endocannabinoid, and corticotrophin-releasing hormone (Arnold, 2005; Backstrom and Hyytia, 2006; Lee et al., 2003; Lhuillier et al., 2007). These other neurotransmitter systems interact with and modulate the reward, motivation, and memory systems in the brain (Kalivas et al., 2005; Lingford-Hughes and Nutt, 2003). Adaptations of these neurotransmission systems are observed after repeated cocaine administration in rats (Kalivas, 2004).
Currently, there is no specific pharmacological therapy with established efficacy for the treatment of cocaine dependence, nor is any medication approved by regulatory authorities for such treatment. Recent advances in neurobiology have identified various neuronal mechanisms implicated in cocaine addiction (Goldstein and Volkow, 2002; Koob, 2000) and suggested several promising pharmacological approaches. Preliminary clinical trials, including the Cocaine Rapid Efficacy Screening Trial (CREST) programme conducted by the United States National Institute on Drug Abuse (NIDA) (Leiderman et al., 2005), point to medications affecting the GABAergic and glutamatergic systems as holding substantial promise. Recent reviews on pharmacotherapy for cocaine dependence have been published (Gorelick et al., 2004; Sofuoglu and Kosten, 2005; Vocci and Elkashef, 2005), but this is a quickly evolving area with new clinical trials and preclinical findings being reported frequently.
This review will focus on medications which have been clinically evaluated (Dackis, 2004; Vocci and Elkashef, 2005) and shown efficacy in published clinical trials, especially agents affecting the neurotransmitter systems mentioned above. While glutamatergic agents show promise in animal studies (Kalivas, 2004), we are not aware of any published clinical trials showing efficacy of purely glutamatergic medication (Berger et al., 2005; Ciraulo et al., 2005). We also discuss preliminary human studies results with anti-cocaine vaccine, N-acetylcysteine, and promising compounds based on preclinical data. We will not discuss other medications that have been tested as treatment for cocaine dependence in patients with comorbid psychiatric disorders. This review is based on Medline, EMBASE, and PsycINFO searches of English-language articles published between 1985 and June, 2007.
Clinical pharmacology and efficacy data of current medications
Topiramate is an anticonvulsant that is also used clinically to prevent migraine headaches. It has several neuropharmacological actions: blockade of voltage-dependent sodium channels, enhancement of GABA neurotransmission at GABAA receptors, blockade of glutamate receptors (AMPA/kainate subtype), and inhibition of carbonic anhydrase.
Topiramate is well absorbed after oral administration (80% bioavailability), with peak plasma concentrations achieved after 2 h. Its half-life is around 21 h, with plasma steady state reached after 4 d. Topiramate is weakly bound to plasma proteins (15–40%). Approximately 70% is excreted unchanged in the urine.
In animal studies, topiramate reduces cocaine self-administration and the dopaminergic response to cocaine administration and exposure to cocaine-associated cues (Johnson, 2005).
An open-label pilot study involved six cocaine-dependent outpatients (also alcohol users) who received increasing doses of topiramate (up to 300 mg/d) for 6 wk, along with weekly cognitive-behavioural therapy (CBT) (Johnson, 2005). Urine tests remained negative for cocaine and alcohol consumption was reduced throughout the study.
The first randomized, double-blind clinical trial involved 40 cocaine-dependent outpatients who received topiramate (up to 200 mg/d) or placebo for 13 wk, in conjunction with twice weekly CBT (Kampman et al., 2004).The topiramate group had significantly less cocaine use than the placebo group; 59% maintained continuous abstinence for at least 3 wk.
Baclofen is a GABAB receptor agonist which inhibits mono- and poly-synaptic reflexes at the spinal cord level, but does not affect neuromuscular transmission. It is primarily used to reduce muscle spasticity in neurological diseases such as multiple sclerosis.
Baclofen has good absorption after oral administration (75%), with peak serum concentrations achieved in 2–4 h. Its half-life is 3–4 h. Baclofen is weakly bound (30%) to plasma proteins. It is eliminated primarily via the kidneys, 85% as the unchanged parent compound.
Baclofen reduces ongoing cocaine self-administration (Roberts, 2005; Roberts and Brebner, 2000), reinstatement of cocaine self-administration after extinction (Campbell et al., 1999), and cocaine-seeking behaviour (Di Ciano and Everitt, 2003) in rats. These behavioural effects may be mediated by its antagonism of cocaine-induced dopamine release in the nucleus accumbens (Fadda et al., 2003). In preliminary clinical studies, baclofen (20–40 mg/d) significantly reduced cocaine craving in cocaine-dependent subjects (Brebner et al., 2002; Ling et al., 1998). The first randomized, double-blind clinical trial involved 70 cocaine-dependent outpatients who received 60 mg/d baclofen or placebo for 16 wk. While there was no significant overall difference between the two groups, baclofen significantly reduced cocaine use in the subgroup of patients who had heavier cocaine use (Shoptaw et al., 2003). A recent human laboratory study showed that 60 mg/d baclofen reduced cocaine self-administration in non-opioid-dependent, non-treatment-seeking cocaine addicts (Haney et al., 2006). Baclofen may have clinical utility beyond any anti-addiction effect. Cocaine-dependent patients have high rates of psychiatric comorbidity; there is preliminary evidence that baclofen has some anxiolytic (Breslow et al., 1989; Drake et al., 2003) and antidepressant (Krupitsky et al., 1993) effects.
Tiagabine is an anticonvulsant that increases GABA neurotransmission by blocking the presynaptic reuptake of GABA.
Tiagabine is rapidly and almost completely (>90%) absorbed after oral administration, with peak plasma concentration achieved after 45 min. The elimination half-life is 7–9 h, which shortens to 2–5 h with chronic dosing because of induction of drug-metabolizing enzymes. Tiagabine is 96% bound to plasma proteins. Only 2% is excreted as parent compound in urine and faeces.
Tiagabine (20 mg/d) showed a trend towards reduced cocaine use in a 10-wk screening trial (Winhusen et al., 2005). Two randomized, placebo-controlled clinical trials in cocaine-dependent outpatients maintained on methadone to treat concurrent opiate dependence confirmed the efficacy of tiagabine. The first trial involved 45 patients who received 12 mg/d, 24 mg/d, or placebo for 10 wk, along with weekly CBT (Gonzalez et al., 2003). Tiagabine dose-dependently decreased cocaine use compared with placebo. The second trial involved 50 patients who received 24 mg/d or placebo for 10 wk, along with weekly individual CBT (Gonzalez et al., 2006). Tiagabine again significantly decreased cocaine use. Tiagabine may have clinical utility beyond its anti-addiction effect. There is clinical evidence that it may be an anxiolytic agent (Schwartz and Nihalani, 2006).
Vigabatrin (γ-vinyl-GABA) is an anticonvulsant that increases GABA neurotransmission by inhibiting GABA transaminase, an enzyme that breaks down GABA. In animals, it reduces cocaine self-administration and cocaine-induced dopamine release in the nucleus accumbens (Gerasimov et al., 2001; Kushner et al., 1999). In three open-label studies involving 78 cocaine- and/or methamphetamine-dependent outpatients receiving medication for up to 9 wk, vigabatrin (1.5–3 g/d) was well tolerated and associated with substantial drug abstinence, albeit with about a 50% dropout rate (Brodie et al., 2003, 2005; Fechtner et al., 2006). Vigabatrin is not marketed in some countries (e.g. USA) because of concern over ophthalmological side-effects, but none were observed during these short-term studies (Fechtner et al., 2006).
Bupropion: a dopamine reuptake inhibitor
Bupropion is an antidepressant also approved as treatment for nicotine (tobacco) dependence. It has stimulant-like effects in animals and is a weak inhibitor of the presynaptic dopamine transporter, but the mechanism of its therapeutic action is unclear. Bupropion (300 mg/d), along with weekly or bi-weekly individual counselling, was not effective in reducing cocaine use in a 12-wk, multi-site, controlled clinical trial in outpatients also maintained on methadone for the treatment of concurrent opiate dependence (Margolin et al., 1995). A post-hoc analysis of this study showed a significant beneficial effect in the patients with comorbid depression.
In a recent controlled clinical trial also conducted in outpatients maintained on methadone for the treatment of concurrent opiate dependence, bupropion (300 mg/d) potentiated the effect of contingency management (i.e. reward for producing a cocaine-free urine specimen) in reducing cocaine use, while having no effect in patients receiving non-contingent reward (Poling et al., 2006).
The combination of levodopa (l-dopa) and carbidopa is approved for the treatment of Parkinson's disease.
l-dopa, the precursor of dopamine, readily crosses the blood–brain barrier into the central nervous system (CNS), where it is metabolized to dopamine by aromatic-l-amino-acid decarbxylase. This conversion also occurs in peripheral tissues, causing adverse effects and decreasing the dopamine available to the CNS.
Carbidopa inhibits decarboxylation of l-dopa but does not cross the blood–brain barrier. Hence, co-administration of carbidopa makes more l-dopa available for transport into the brain and its conversion there to dopamine.
l-dopa–carbidopa is a dopamine replenishment treatment strategy. In three randomized, double-blind, placebo-controlled trials in cocaine-dependent outpatients, l-dopa–carbidopa (300/75 mg/d, 400/100 mg/d, or 800/200 mg/d) did not significantly reduce cocaine use or craving (Mooney et al., 2007; Shoptaw et al., 2005).
Dopamine antagonists and partial agonists
Cocaine acts directly on the dopaminergic system, so blockade of dopamine receptors is a plausible therapeutic approach for cocaine dependence. Conventional (first-generation) neuroleptics, which act chiefly as dopamine D2 receptor antagonists, are not effective in reducing cocaine use in patients with major psychiatric comorbidity (e.g. schizophrenia) (Green, 2005). Newer second-generation neuroleptics, which also act on serotonin receptors, are currently being evaluated.
Several human laboratory studies found that the second-generation neuroleptics risperidone or olanzapine reduced cocaine-induced euphoria or cue-induced cocaine craving (Newton et al., 2001; Smelson et al., 1997, 2002, 2004, 2006). One case- series of 21 outpatients on methadone maintenance for opiate dependence found that olanzapine (5–10 mg/d) substantially reduced cocaine use (Bano et al., 2001). However, several outpatient clinical trials found that neither olanzapine (10 mg/d for 8 wk or 12 wk) (Kampma, et al., 2003; Reid et al., 2005) nor risperidone (2–8 mg/d for 12 wk or 2–4 mg/d for 26 wk) (Grabowski et al., 2000, 2004a), always in combination with weekly or twice-weekly CBT, significantly reduced cocaine use in patients without psychiatric comorbidity. Aripiprazole is a second-generation neuroleptic which acts as a partial antagonist at both the dopamine D2 and 5-HT1A receptors (which regulate dopamine release) (El-Sayeh et al., 2006). It blocked reinstatement of cocaine self-administration after extinction, an animal model of relapse (Feltenstein et al., 2006). Aripiprazole decreased cocaine craving in an 8-wk open-label pilot study involving 10 cocaine-dependent patients with comorbid schizophrenia (Beresford et al., 2005). Clinical trials in cocaine-dependent patients are currently underway to evaluate these promising findings.
Agonist replacement therapy
Agonist replacement therapy uses a drug from the same pharmacological family as the abused drug to suppress withdrawal and drug craving (Grabowski et al., 2004b). A clinical example is methadone treatment of heroin dependence. The abused drug itself can be used in treatment, especially in a slower onset formulation which has less abuse liability (Gorelick, 1998). An example of such use is nicotine gum or skin patch to treat tobacco dependence.
The World Health Organization (WHO) is exploring the use of agonist replacement medication; several countries are developing programmes to pursue this approach. Potential agonist medications include methylphenidate, modafinil, disulfiram, d-amphetamine, and oral cocaine.
Methylphenidate is approved for the treatment of attention deficit hyperactivity disorder (ADHD).
Methylphenidate binds to both the dopamine and norepinephrine presynaptic transporters, but not to the serotonin transporter. The functional effect is to block catecholamine reuptake from and increase catecholamine release into the synapse.
Methylphenidate is rapidly and extensively absorbed following oral administration. Owing to extensive first-pass metabolism, oral bioavailability is around 30%. Peak plasma concentrations of 7.8 ng/ml were observed 2 h after administration of 0.30 mg/kg as the standard tablet. The extended-release formulation has similar bioavailability with a broader peak plasma concentration lasting 2–6 h after administration. Methylphenidate has a mean plasma half-life of around 3.5 h. Following oral administration, 78–97% of the dose is excreted in urine and 1–3% in faeces within 48–96 h.
The commonest adverse effects are nervousness, insomnia, and decreased appetite.
Methylphenidate (5 mg+20 mg sustained release) was no better than placebo in an 11-wk, double-blind, placebo-controlled trial in 24 cocaine-dependent outpatients without psychiatric comorbidity (Grabowski et al., 1997). However, ADHD is a common psychiatric comorbidity among cocaine-dependent persons, occurring in up to 30% in some studies (Schubiner, 2005). Because methylphenidate is an effective treatment for ADHD, it was hypothesized that it might have a beneficial effect in cocaine-using patients with this comorbid disorder.
Controlled clinical trials of methylphenidate in this population have shown mixed results. In a 12-wk, double-blind, placebo-controlled trial of methylphenidate (90 mg/d immediate-release formulation) in 48 cocaine-dependent adults with comorbid ADHD, methylphenidate was no better than placebo for cocaine use and craving, but did improve subjective reports of ADHD symptoms (Schubiner et al., 2002). A recent 14-wk, double-blind, placebo-controlled trial of sustained-release methylphenidate (60 mg/d) in 106 cocaine-dependent outpatients with comorbid ADHD found a significant decrease in cocaine use compared with placebo, with most of the benefit occurring in patients who had improvement in ADHD symptoms (Levin et al., 2007). This pattern of findings suggests that sustained-release methylphenidate may be more effective than immediate-release, and that the beneficial effect may be mediated, in part, by a reduction in ADHD symptoms.
Methylphenidate, like cocaine, is a cardiovascular stimulant with the potential to cause cardiac adverse events. A placebo-controlled, cross-over oral methylphenidate (placebo, 60 mg, 90 mg immediate-release formulation)–intravenous cocaine (saline, 20 mg, 40 mg) interaction study in seven cocaine-dependent adults found no significant effect of methylphenidate on the cardiovascular response to cocaine or cocaine pharmacokinetics (Winhusen et al., 2006). A similar placebo-controlled, cross-over interaction study in seven adult cocaine abusers with comorbid ADHD found that sustained-release oral methylphenidate (40 mg or 60 mg) increased the cardiovascular response to intravenous cocaine (16 mg or 48 mg/70 kg every 14 min for four doses), but not to a degree that required termination of study sessions (Collins et al., 2006). In both studies, methylphenidate pretreatment reduced the positive subjective effects of cocaine. These findings suggest that methylphenidate may be safe to use in outpatient cocaine abusers.
Oral methylphenidate in the immediate-release formulation has potential for abuse (Arria and Wish, 2006; White et al., 2006), while the sustained-release formulation has much less abuse potential (Greenhill, 2006).
Modafinil, a functional stimulant, is used to treat disorders of excessive sleepiness, such as narcolepsy or idiopathic hypersomnia (Bastuji and Jouvet, 1988; Billiard et al., 1994; Laffont et al., 1994).
Neurochemical mechanisms underlying modafinil's therapeutic actions remain unresolved. This drug has been demonstrated to occupy both the dopamine and norepinephrine transporter at clinically relevant doses (Madras et al., 2006; Mignot et al., 1994), consistent with stimulant-like action In addition, modafinil appears to increase release of glutamate, an excitatory neurotransmitter, and decrease release of GABA, an inhibitory neurotransmitter (Ballon and Feifel, 2006).
Modafinil has good absorption following oral administration, with peak plasma concentrations attained after 2–4 h. Modafinil is moderately bound to plasma proteins (60%), with little evidence of displacement of other medications. Modafinil acid, the primary metabolite, is pharmacologically inactive. Modafinil and its metabolites are primarily excreted by the kidneys; only a small proportion (<10%) is excreted as the unchanged parent compound. Modafinil has a half-life of about 15 h, with steady state reached after 2–4 d of dosing. Thus, once- or twice-daily dosing is adequate. In clinical trials for sleep disorders, up to 3% of patients experienced cardiovascular side-effects such as hypertension, tachycardia, and palpitations. Patients with mitral valve prolapse or left ventricular hypertrophy may be at increased risk of chest pain or ischaemic ECG changes.
Modafinil's stimulant-like activity may diminish the symptoms of cocaine withdrawal, including hypersomnia, lethargy, dysphoric mood, cognitive impairment, and increased appetite (Dackis et al., 2003), thereby reducing the desire to use cocaine. It has weak cocaine-like reinforcement effects in animals (Deroche-Gamonet et al., 2002) and stimulant-like subjective effects in humans (Rush et al., 2002a,b). In this light, modafinil may be envisaged as a ‘substitution treatment’ for cocaine dependence, analogous to methadone or buprenorphine treatment of heroin dependence.
The first randomized, double-blind clinical trial involved 62 cocaine-dependent outpatients who received either a single 400-mg dose of modafinil (n=32) or of placebo (n=30) daily for 8 wk in conjunction with CBT (Dackis et al., 2005). Patients taking modafinil had significantly less cocaine use (measured by urine drug testing) than did patients treated with placebo. No significant adverse effects were noted. A recently completed multi-site, controlled clinical trial involved 210 cocaine-dependent outpatients who received either modafinil (200 mg/d or 400 mg/d) or placebo (Elkashef and Vocci, 2007). Modafinil significantly reduced cocaine use only in the subgroup of patients without alcohol dependence. Additional clinical trials are needed to evaluate the efficacy and clinical role of modafinil treatment.
Modafinil does not appear to produce euphoria or evoke cocaine craving (Ballon and Feifel, 2006; O'Brien et al., 2006), suggesting that it has low abuse potential (Jasinski, 2000; Jasinski and Kovacevic-Ristanovic, 2000). Chronic cocaine users are capable of discriminating between cocaine and modafinil effects and report no euphoric effects from the latter (Rush et al., 2002b). Human laboratory studies report no clinically significant adverse interactions between modafinil and cocaine (Dackis et al., 2003; Donovan et al., 2005; Hart et al., 2007; Malcolm et al., 2006). Post-marketing surveillance studies have reported no evidence of significant abuse liability (Myrick et al., 2004).
Disulfiram: a dopamine metabolism inhibitor
Disulfiram has been used for more than half a century in the treatment of alcoholism (Suh et al., 2006). It inhibits aldehyde dehydrogenase, the enzyme that transforms acetaldehyde into acetate during alcohol metabolism. When a person drinks alcohol while taking disulfiram, the resulting acetylaldehyde accumulation causes an aversive reaction (flushing, sweating, headache, nausea, tachycardia, palpitations, arterial hypotension, hyperventilation) which discourages further drinking.
Disulfiram also inhibits dopamine β-hydroxylase, the enzyme that transforms dopamine into norepinephrine. Such inhibition would increase dopamine levels and decrease norepinephrine levels in the nervous system. This effect is considered to be therapeutic for cocaine dependence. In addition, a disulfiram metabolite may block glutamatergic receptors (Ningaraj et al., 2001).
Disulfiram is readily absorbed after oral administration, but then rapidly reduced to diethyldithiocarbamate, which in turn is broken down into several metabolites, some of which may be pharmacologically active. The elimination half-life is about 7 h, but enzyme inhibition may persist for at least a week after discontinuation of chronic dosing. Disulfiram is excreted in urine almost entirely as metabolites.
The initial impetus for the use of disulfiram to treat cocaine dependence was the high rate of comorbidity between cocaine abuse or dependence and alcohol abuse or dependence, up to 85% in some studies (Gossop and Carroll, 2006). It was hoped that reduction in alcohol use would lead to secondary reduction in cocaine use. Abstinence from alcohol would also prevent formation of cocaethylene, a metabolite formed when alcohol and cocaine are present together. Cocaethylene has pharmacological actions similar to cocaine, but may be longer acting (Hart et al., 2000). A recent 12-wk open study in outpatients abusing both cocaine and alcohol found that the four patients receiving disulfiram (400 mg/d) along with CBT had fewer urine samples positive for cocaine and cocaethylene than did the four patients receiving CBT alone (Grassi et al., 2007).
Several short-term clinical trials in outpatients using both cocaine and alcohol showed that disulfiram (250–500 mg/d), along with CBT or a 12-step self-help group, significantly reduced cocaine and alcohol use (Carroll et al., 1998; Higgins et al., 1993). In one study, the reduction in cocaine use was still present one year after treatment (Carroll et al., 2000).
Three published randomized, placebo-controlled clinical trials have found that disulfiram has a direct effect in reducing cocaine use, i.e. in outpatients who are not also abusing alcohol. Two of these studies involved outpatients who were also opiate-dependent and receiving opiate agonist maintenance treatment (either methadone or buprenorphine) (George et al., 2000; Petrakis et al., 2000). A larger trial in outpatients without concurrent opiate dependence found that disulfiram (250 mg/d), along with CBT or interpersonal therapy, significantly reduced cocaine use over the 12-wk study (Carroll et al., 2004).
Caution has been raised regarding the clinical use of disulfiram. In human laboratory studies, disulfiram inhibits cocaine metabolism, increasing cocaine plasma levels when the two are administered together (Baker et al., 2006). In some studies, this has been associated with enhanced cardiovascular response to cocaine. Both cocaine and the disulfiram–alcohol interaction can produce severe cardiovascular effects. A patient who relapsed to cocaine and/or alcohol use while taking disulfiram might be at risk for serious, perhaps life-threatening, adverse events. This might limit the use of disulfiram to patients who are highly motivated for abstinence, have an active social support network for early detection of relapse, and are in good cardiovascular health.
Dextroamphetamine binds to presynaptic dopamine and norepinephrine transporters and promotes release of these neurotransmitters. d-amphetamine decreases cocaine self-administration by rhesus monkeys (Negus and Mello, 2003). Three double-blind, placebo-controlled studies using d-amphetamine (15–60 mg/d sustained-release formulation) in cocaine-dependent or in cocaine- and heroin-dependent patients showed decreased cocaine use at the higher doses (30–60 mg/d) (Grabowski et al., 2001, 2004a; Shearer et al., 2003).
Oral formulations of cocaine
Coca leaf chewing is common among indigenous inhabitants of the Andean region (Bolivia, Peru, Colombia). Oral formulations of cocaine, such as coca tea (infusions of the leaf) and tablets, are also used in this region.
Cocaine is a crystalline tropane alkaloid found in leaves of the coca plant. About a third of an oral dose is systemically absorbed, with detectable blood levels appearing after about 30 min. Physiological and psychotropic effects typically appear approximately 1 h after ingestion. Cocaine is rapidly and extensively metabolized, with only about 1% excreted unchanged in the urine. The major metabolic process is hydrolysis, either by butyrylcholinesterase in plasma, brain, and lung (resulting in ecgonine methylester) or by carboxyesterases in the liver (resulting in benzoylecgonine) (Warner and Norman, 2000).
A human laboratory study found that pretreatment with oral cocaine (400 mg/d in capsules) decreased the subjective and physiological responses to an intravenous cocaine challenge (25 mg or 50 mg) (Walsh et al., 2000). Anecdotal reports suggest that coca tea may have anti-craving effects in cocaine users (Siegel et al., 1986; Weil, 1978). A case- series of 23 coca-paste smokers in Lima, Peru found that coca tea (20–60 mg/d cocaine) plus counselling reduced cocaine craving and relapse (Llosa, 1994).
Future studies including controlled clinical trials are needed to evaluate the efficacy of this treatment approach.
Promising medications for the future
Several other promising compounds are undergoing clinical trials or are in preclinical development.
Promising compounds in clinical trials
Vaccine pharmacotherapy uses anti-cocaine antibodies to sequester cocaine molecules in the peripheral circulation. The cocaine-antibody complexes are too large to cross the blood–brain barrier, thus keeping cocaine from its site of action in the CNS (Kosten and Biegel, 2002). The cocaine molecule by itself is too small to be antigenic (i.e. to evoke an antibody response), so it (or a stable congener) must be coupled to a larger antigenic molecule, e.g. cholera B toxin (Heading, 2002). Cocaine vaccines significantly reduce the behavioural effects of cocaine in animals (Fox et al., 1996; Kantak et al., 2000). An anti-cocaine vaccine might have two advantages over conventional medication: (1) no direct psychoactive effects and, therefore, no abuse liability and (2) therapeutic effects persisting for months, improving patient adherence to treatment (Kosten and Owens, 2005). A disadvantage might be a lag time of up to several months before therapeutic antibody levels were achieved.
A phase I, randomized, double-blind placebo-controlled trial of vaccine in 34 cocaine abusers over 12 months found that cocaine-specific IgG cocaine antibodies were induced in a time- and dose-dependent manner. The vaccine was well tolerated with no serious adverse effects (Kosten et al., 2002). More recently, the cocaine vaccine was tested in 18 cocaine-dependent subjects in an open label, 14-wk, dose-escalation study (Martell et al., 2005). Ten subjects received four 100 µg injections over 8 wk; eight subjects received five 400 µg vaccinations over 12 wk. The vaccine was well tolerated. There was a significantly higher mean antibody titre response in the 2000-µg group than in the 400-µg group, with detectable antibody titres still present after 6 months. Despite relapse in both groups, most subjects reported attenuation of cocaine-induced euphoria (‘high’) (Martell et al., 2005).
N-acetyl cysteine is approved for the treatment of pulmonary complications of cystic fibrosis and paracetamol (acetaminophen) overdose.
As the N-acetyl derivative of the amino acid l-cysteine, it is a major precursor to the antioxidant glutathione. N-acetyl cysteine is rapidly absorbed from the gastrointestinal tract, but has low bioavailability due to first-pass metabolism. Peak plasma concentrations are observed 0.5–1 h following oral administration of 200–600 mg. Its terminal half-life is approximately 6 h.
Chronic cocaine use decreases basal levels of glutamate within the nucleus accumbens in rats (Baker et al., 2003a). N-acetyl cysteine can exchange extracellular cysteine for intracellular glutamate, resulting in increased levels of glutamate. N-acetyl cysteine treatment reduces cocaine-induced reinstatement of cocaine self-administration in rats (an animal model of relapse) (Baker et al., 2003b).
A double-blind, placebo-controlled, cross-over in-patient study in 13 non-treatment-seeking cocaine-dependent subjects found N-acetyl cysteine well tolerated, with some evidence of decreased cocaine craving and withdrawal symptoms (LaRowe et al., 2006).
A 4-wk, open-label pilot study of three doses of N-acetyl cysteine (1200, 2400, or 3600 mg/d) in 23 treatment-seeking cocaine-dependent outpatients found all doses safe and well tolerated (Mardikian et al., 2007). The majority of subjects reduced their cocaine use; treatment retention was longer for the two higher-dose groups. Controlled clinical trials are currently underway to evaluate the promise of this compound.
Ondansetron, a serotonin 5-HT3 receptor antagonist, is approved as an anti-emetic agent. 5-HT3 receptor activation increases dopamine activity in the nucleus accumbens (Dremencov et al., 2006), making blockade of these receptors a potential treatment approach. A recent 10-wk, double-blind, placebo-controlled trial in 63 cocaine-dependent outpatients receiving ondansetron (0.25 mg, 1.0 mg, or 4.0 mg twice daily) or placebo plus weekly CBT found that ondansetron was well tolerated. The 4.0-mg ondansetron group had a lower dropout rate and higher percentage of participants with a cocaine-free week than the other three groups (Johnson et al., 2006).
Promising compounds in preclinical development
Selective dopamine reuptake inhibitor
Dopamine reuptake inhibitors with a slower onset of effect and longer duration of action than cocaine might act as functional cocaine antagonists. One such compound, the 3-phenyltropane analogue RTI-336, reduced cocaine self-administration in rat and rhesus monkey models (Carroll et al., 2006).
D3 receptor ligands
D3 selective antagonists may influence the ability of drug-associated cues to induce drug-seeking behaviour (Cervo et al., 2007). They inhibit cocaine-induced drug seeking and decrease cocaine self-administration in rodents (Cervo et al., 2007), but not in rhesus monkeys (Martelle et al., 2007). The D3 partial agonist CJB090 blocks the discriminative and reinforcing stimulus effects of cocaine in rhesus monkeys, without producing cocaine-like effects (Martelle et al., 2007).
Dual dopamine-serotonin releasers
Medications that release both dopamine and serotonin are plausible candidates for the treatment of cocaine addiction, based on several lines of evidence (Rothman et al., 2006). Withdrawal from chronic cocaine use produces a dual deficit of both neurotransmitters in the brain; drug-seeking behaviour is reduced by the administration of dopamine and serotonin releasing agents separately or together; increased levels of extracellular serotonin can antagonize psychomotor stimulant actions of dopamine releasers, and dual dopamine-serotonin releasers have low abuse ability. One such candidate medication, PAL287, is minimally reinforcing itself and suppresses cocaine self-administration in rhesus monkeys (Rothman et al., 2005).
Cannabinoid CB1 receptor antagonists
Cannabinoid CB1 receptors are expressed in the brain reward circuit, modulate the dopamine-releasing effects of drugs of abuse, and are involved in relapse to drug seeking for many addictive drugs (Maldonado et al., 2006). Blockade of cannabinoid CB1 receptors inhibits cocaine- or cocaine cue-induced reinstatement of cocaine seeking in animals (Xi et al., 2006). CB1 receptor antagonists are not themselves self-administered by animals, suggesting little or no abuse liability (Beardsley et al., 2002). These preclinical findings suggest that CB1 receptor antagonists may be a useful target for medication development.
Corticotropin-releasing factor (CRF) receptor antagonists
CRF is a neuropeptide that evokes release of ACTH in response to physiological or behavioural stress. Cocaine self-administration by rhesus monkeys stimulates ACTH release and activates the hypothalamic–pituitary–adrenal (HPA) axis (Broadbear et al., 2004), while administration of CRF promotes reinstatement of cocaine self-administration after extinction in rats (Erb and Brown, 2006). Stress-induced activation of the HPA axis during in-patient treatment was negatively correlated with time to relapse to cocaine use during a 90-d outpatient follow-up period in a recent study of 49 cocaine-dependent patients (Sinha et al., 2006). These findings suggest the HPA axis as a useful target for medication development.
The selective CRF1 receptor antagonist CP154,526 decreased cocaine-induced reinstatement of cocaine self-administration in rats, while not affecting ongoing cocaine self-administration or cocaine discrimination (Przegalinski et al., 2005). Another selective CRF1 receptor antagonist, antalarmin, had no effect on cocaine discrimination or self-administration in rhesus monkeys (reinstatement was not studied) (Mello et al., 2006).
Given that cocaine has become a more prominent part of the European drug scene, pharmacological treatment has become important for combating this increasingly popular drug of abuse. No pharmacological treatment has yet proven broadly effective for cocaine addiction. The studies reviewed in this paper reveal some promising findings, although most studies were of short duration and in selected patients.
Recent controlled clinical trials have highlighted the promise of several medications, especially GABAergic agents, agonist replacement therapy (amphetamine or perhaps modafinil) which modulate the cortico-meso-limbic dopaminergic brain circuits on which cocaine acts. Partial agonists at the dopamine D2 receptor, the 5-HT3 receptor antagonist ondansetron and anti-cocaine vaccines are other promising treatment approaches.
Recent studies suggest that the targets of interest in developing new treatments for cocaine dependence should include all three biogenic amine neurotransmitters (not only dopamine), as well as the GABAergic, glutamatergic, and endocannabinoid systems.
Finally, several issues beyond the scope of this review, such as integrating pharmacological with psychosocial treatment, psychiatric comorbidity, and concurrent dependence (on alcohol or other substances), should be taken into consideration when implementing pharmacological treatment for cocaine dependence.
Preparation of this review was supported in part by the Intramural Research Program, US National Institutes of Health, National Institute on Drug Abuse. Dr Weinstein (Israel) and Dr Karila (France) are supported by the Fondation Rashi (Israel).
Statement of Interest