Research Review: Alcohol Treatment Medications

The medical and psychosocial burden of alcohol dependence has motivated hundreds of studies to discover better medications. Clinical trials to find effective treatments began in the 1950s, but currently there are only three approved pharmacotherapies. These are disulfiram, acamprosate and naltrexone. However, while these medications are effective in subsets of patients, many still do not respond to these interventions, and recent research has investigated the causes for such variability.

Disulfiram

Disulfiram has been available to treat alcohol dependence since 1949, and its mechanism of action has been known since about that time. Alcohol is metabolized into acetaldehyde, which is further metabolized into acetic acid by the enzyme acetaldehyde dehydrogenase. Disulfiram inhibits this enzyme, leading to the accumulation of acetaldehyde which provokes flushing, sweating, headache, nausea and vomiting. But the efficacy of disulfiram is poor; patients just don’t take it because of its adverse effects. While treatment adherence is 61% in court mandated treatment, it drops to less than 20% in voluntary patients (See J. Subst Abuse Treat. and Alcohol Clin Exp Res.). Moreover, disulfiram does not address the underlying neurobiology of alcohol dependence (learn more). 

Acamprosate

Acamprosate has been in use for alcohol dependence since 1989 and was approved by the FDA to treat alcohol dependence in 2004. However, clinical trials using acamprosate in Europe have consistently shown a greater effect than those in the United States. One reason for this may be that the European trials tend to be performed in highly dependent subjects that are enrolled in public treatment programs, whereas the American studies usually include less severe patients recruited from advertisements (learn more). Nonetheless, the number needed to treat (NNT) for acamprosate is high, between 8 and 12, which means that on average 8 to 12 patients would have to be treated to prevent only one patient from returning to heavy drinking (See these articles: in JAMA; in CDSR; and in Alcohol Clin Exp Res).

Because acamprosate has some promise, experimental work has been performed to identify its mechanism of action in the brain, with the aim of improving its efficacy. Two major neurotransmitter systems in the brain are glutamate, which is excitatory, and GABA, which is inhibitory. Alcohol largely increases GABA transmission in the brain, leading to sedating effects. Alcohol dependence is associated with a hyper-glutamatergic state, which can eventually lead to withdrawal symptoms and even seizures. It is thought that acamprosate increases GABA signaling or inhibits glutamate. However, despite numerous experiments to identify the molecular mode of action, the mechanism remains unclear.

The weight of evidence now suggests that acamprosate affects the glutamate system, but the actual glutamate receptor targeted remains debated. In fact, a recent and provocative study performed in rats suggests that acamprosate does not have any direct action on glutamate receptors, but that the effect of acamprosate could originate from the calcium moiety attached to the molecule (learn more). Based on this, the authors propose that calcium supplementation alone, which may have effects on the brain, could have beneficial effects for alcohol dependence. Though interesting, this hypothesis will have to be confirmed by clinical studies.

Naltrexone

Clinical trials investigating naltrexone for alcohol dependence were first published in 1992 based on rodent studies showing that naltrexone decreases alcohol consumption (See Arch Gen Psychiatry). Naltrexone was approved by the FDA in 1994, and a series of clinical trials have investigated its efficacy since then, but with mixed results. The largest of these clinical trials was the COMBINE study, in which subjects received medical management with 16 weeks of naltrexone or acamprosate – both with and without behavioral intervention – or behavioral intervention without medication. The behavioral intervention integrated aspects of cognitive behavioral therapy, 12-step programs and motivational interviewing. The results showed that medical management was associated with better outcomes when patients also received either naltrexone or behavioral intervention (read ​JAMA).

Using a meta-analysis to derive an NNT for naltrexone shows that about 9 to 20 patients would need to be treated for one successful response (See ​JAMA and CDSR). Similar to acamprosate, the fact that naltrexone has a known effect on alcohol dependence, but low effectiveness in clinical trails, has led to research on the mechanism behind this.

Unlike acamprosate, the molecular action of naltrexone is well known. Naltrexone binds to and blocks the mu-opioid beta-receptor – which is important for the modulation of pleasure and pain - which makes it very effective in blocking the effects of opiates, such as morphine or heroin. Because alcohol is known to increase levels of an endogenous brain opioid (beta-endorphins), naltrexone is thought to exert its action by blocking this effect. However, it remains unclear why only some patients respond to naltrexone.

There is a variant of the gene that codes for the mu receptor called A118G Asn40Asp (asparagine-to-aspartate amino acid substitution at position 40). Studies have shown that people with this version of the gene for the mu receptor have a greater "high" in response to alcohol, in agreement with the subsequent observation that they have a higher dopamine response to alcohol in brain reward circuitry (Read Alcohol Clin Exp Res and Mol. Psychiatry).

Based on this data, research has been performed to investigate the effect of this genotype on treatment with naltrexone. Oslin et al. demonstrated that the subjects of the Asp40 genotype treated with naltrexone had lower relapse rates and a longer time to return to heavy drinking than those without this allele (homozygous for the Asn40 allele). Anton et al. re-examined the results of the COMBINE study and showed similar results: carrying one or two copies of the Asp40 allele was associated with a better response to naltrexone (87.1%) compared to subjects homozygous for the Asn40 genotype (54.8%, which was not different from the placebo rate). This finding has been replicated in other studies as well, but not all (See J Alcohol Drug Depend).

In order to better characterize the impact of the Asn40Asp variation on alcohol reward and naltrexone response,a recent study used two humanized  mouse lines, in which the mouse gene had been replaced with either the human Asn40 or Asp40 allele. The Asp40 animals displayed enhanced reward response to alcohol, as predicted by the phenotype in humans carrying this mutation. Interestingly, naltrexone was able to blunt the rewarding effects of alcohol in Asp40 mice and to more efficiently decrease alcohol self-administration in the animals carrying the mutation. This study therefore supports the hypothesis based on human data, that subjects carrying the Asn40Asp mutation are more likely to respond to naltrexone.

Conclusion

The fact that the number to treat is high for these medications is not a reason to avoid using them in practice. But they do indicate that more medications with other mechanisms of action are needed, as well as a better understanding of which patients will respond. For naltrexone, it is important to remember that the genotype is a moderating effect, and that not all subjects of either genotype will have a predictable response. Also the asp40 genotype is not common: it’s only seen in about 15% of individuals of European ancestry, 35% in Asians, and less than 5% in individuals of African descent.

Other medications are currently being tested for alcoholism. Medications that target the GABA receptor system have shown some promise (topiramate, baclofen, gabapentin), each of which is currently FDA-approved for other indications, but the results of clinical trails investigating these are mixed. Recent studies have also investigated the use of varenicline, which partially activates the nicotine receptor. Results of these studies are also mixed, but varenicline might reduce alcohol consumption in dependent subjects who smoke. Prazozin, which blocks the adrenergic receptor and reduces the excitability of neurons, may be helpful in subjects with co-morbid PTSD, but the data are limited. Ondansetron, which inhibits a subtype of serotonin receptor, has been shown to have an effect on alcohol dependence, and may be most promising for early onset dependence. However, recent work is indicating that response to odansetron and topriamate may also vary by genotype (See work by Keena GA et al. and by Kranzler HR et al.).