Addiction as a Brain Disease

American psychiatry has made remarkable progress in recategorizing the addictive disorders from moral failures to brain diseases, but the need for community education continues. The concept of moral failure is by no means gone from the discussion of addictive disorders, as evidenced by our country's investment in criminal justice rather than treatment, including the denial of health insurance parity for addictive disorders and the court ruling that alcoholism among military personnel was “willful misconduct,” not a disease. Ironically, this political denial comes at a time when our new understanding of the neurobiology and genetics of addictions offers innovative treatments (1, 2).

Three recent Journal articles, including two in this issue, reflect our progress in understanding the genetics of alcoholism. In one study (3), genotypic variation associated with midbrain serotonin transporter binding sites and mRNA levels distinguished comparison subjects from alcohol-dependent patients. A similar study of allelic variation related to the dopamine transporter showed an association with paranoia in cocaine dependence (4). These candidate gene association studies for complex disorders such as addictions require identification of appropriate candidate genes, but the more difficult task is often defining an appropriate phenotype with a neurobiological foundation (5, 6).

How should we define the phenotype of substance abusers who will have the strongest genetic determinants for their disorder, and can we identify a neurobiological correlate of this phenotype that will suggest candidate genes? These questions are at the core of the large Collaborative Study on the Genetics of Alcoholism. In this issue of the Journal, two articles separately address the nosological issue of defining a phenotype and the neurobiological issue of identifying a neurobiological foundation for severe alcohol dependence. It is most critical to have a neurobiological foundation for these genetic studies, because genes fundamentally code for or regulate the production of proteins that ultimately have an effect on behavior. Both studies focus on the behavior of alcohol withdrawal, and experiencing alcohol withdrawal might identify a clinically relevant phenotype that has genetic neurobiological determinants.

Schuckit et al. have made an important contribution by addressing nosological issues with data from a study that is primarily designed for genetic analyses, and this example needs to be followed by other investigators. The Collaborative Study on the Genetics of Alcoholism is quite relevant to nosology, and it showed an important difference from a previous study of cocaine abusers (7), in which withdrawal did not appear very important for distinguishing subgroups of patients. This difference from cocaine may reflect the generally modest severity of a cocaine withdrawal syndrome compared to alcohol withdrawal. Even in the previous study, withdrawal symptom rates and factor loadings were higher for the cocaine abusers who had concurrent alcohol dependence. Thus, a phenotype that is defined by the severity of withdrawal symptoms may be relevant, as suggested by a recent study of the serotonin transporter in alcoholism (8), and future analyses will benefit from having subjects with the full range of abused drugs, such as in the DSM-IV field trial (9).

If alcoholics who experience withdrawal symptoms are expressing a phenotype with strong genetic determination, then its neurobiological correlates might be specific for alcoholism, and genetic risk factors might differ across various drugs of abuse. This issue of specificity of inheritance for different abused drugs has a long history of study in the substance abuse field. The self-medication hypothesis (10) has suggested that substance abusers are usually specific in their selection of substances to abuse and that this selection may reflect psychological deficits that are ameliorated by the abused substance. This hypothesis has run into neurobiological difficulties, because fundamental neurobiological properties of reinforcement may follow similar rather than different final common pathways. For example, the dopamine-rich connection between the ventral tegmental region and the nucleus accumbens has been identified repeatedly as the site for reinforcement with many abused substances (11). On a psychosocial level this specificity has also run into difficulty, because many substance abusers abuse multiple substances that cross widely divergent drug classes, such as alcohol or opioids and cocaine, and the dependence syndrome, which was derived initially from the psychological correlates of alcohol dependence, has been productively applied to most other abused substances (12). As an alternative to this focus on acute withdrawal syndromes, which are often quite different across drugs, protracted withdrawal may be more similar across abused substances (13). Since protracted withdrawal obviously requires first experiencing acute withdrawal, acute withdrawal may be a marker for this more subtle heritable phenomenon.

Protracted withdrawal follows the acute state and may last for weeks. Interestingly, as reported in this issue of the Journal, Tsai et al. assessed glutamatergic neurotransmission and oxidative stress after acute withdrawal and 4 weeks later and found persistent abnormalities in cerebrospinal fluid (CSF). As a trait-dependent characteristic, these alcoholics might have a degenerative brain disorder like Alzheimer's disease that has alcohol as a cofactor in disease progression (14). These patients are then vulnerable to this degenerative process during protracted withdrawal because of their enhanced excitatory neurotransmission. Future studies of at-risk populations and alcoholics with long-term abstinence will be able to address this issue of the trait versus state dependence of this finding.

These studies of CSF also encourage use of other measures of excitatory amino acid (e.g., N-methyl-d-aspartic acid [NMDA]) neurotransmission in the brains of living humans. Two specific approaches are receptor neuroimaging using positron emission tomography (PET) or single photon emission computed tomography (SPECT) and magnetic resonance spectroscopy (MRS). Glutamatergic metabolites can be measured by using MRS, but in both MRS and CSF studies the major source of glutamate is from metabolic activity, not neurotransmission. Because of this complication in measuring glutamate directly, other metabolites, such as N-acetylaspartylglutamate (NAAG), and metabolite ratios have been examined, as in the study by Tsai et al. More specificity for neuronal changes could be obtained with receptor neuroimaging using a ligand for the NMDA receptor. This neuroimaging could measure localized changes in receptor density relative to normal controls or neurotransmitter turnover by examining displacement of the ligand during activation procedures, as recently confirmed for the dopamine system in cocaine abusers (15, 16). Thus, study of markers of excitatory neurotransmission might profit from convergent neuroimaging studies examining abnormalities in alcohol dependence and withdrawal.

Candidate gene studies similar to those with the serotonin transporter in alcoholism and the dopamine transporter in cocaine dependence may be warranted for genetic variants of the excitatory amino acid system (3, 4, 8, 17). Will these demonstrations that addictive disorders are genetically influenced brain diseases persuade our leaders and fellow citizens that these patients deserve the same level of compassion and treatment as is provided to other medical patients? Not without our help in educating them.

Address reprint requests to Dr. Kosten, Department of Psychiatry 116A, VA Connecticut Healthcare System, 950 Campbell Ave., West Haven, CT 06516; [email protected] (e-mail). Supported by grants DA-04060 and DA-09250 from the National Institute on Drug Abuse (to the Yale Medications Development Research Center).