Benzodiazepines in the Treatment of Alcoholism


This post comprises three sections that cover the main aspects of benzodiazepines and alcohol: (1) the basic pharmacology of benzodiazepines; (2) use of benzodiazepines in the treatment of withdrawal; and (3) the use of benzodiazepines in treating alcoholics. The basic studies suggest that a major site of action of alcohol may be the GABA/benzodiazepine receptor complex and that compensatory alterations in this complex may underly withdrawal. In the section on alcohol withdrawal, interactions between the GABA/benzodiazepine receptor complex, sympathetic nervous system, and hypothalamic-pituitary-adrenal axis are discussed. Use of benzodiazepines in the treatment of the alcohol withdrawal syndrome are reviewed, including the possibility that the benzodiazepines may prevent withdrawal-induced “kindling”. Lastly, we review indications for, and efficacy of, benzodiazepines in long-term treatment of patients with alcoholism. Benzodiazepines are not indicated for the treatment of alcoholism. Furthermore, they have very few indications in alcoholics and their dependency-producing potency has to be appreciated when they are used in patients with alcoholism.

The benzodiazepines () are a group of compounds that were first synthesized in the 1880s and shown to have tranquilizing actions in the 1950s. The earliest clinically used drugs were chlordiazepoxide and diazepam but in the last 30 years a large number of effective analogs have been developed. These all share the common property of binding to the brain-specific benzodiazepine receptor and, in general, differ only in affinity and efficacy at that receptor. This means that doses may vary between drugs but that their clinical and neurochemical effects are broadly equivalent. Indeed, one of the most remarkable aspects of the pharmacology of the benzodiazepines is the confidence with which one can predict clinical effects from their basic pharmacology.

Benzodiazepine Receptors

In the mid-1970s it became apparent that benzodiazepines acted by facilitating inhibitory neurotransmission in the CNS. In particular these drugs enhanced the action of endogenously released γ-aminobutyric acid (GABA). The molecular basis of this became clear when a high-affinity binding site for diazepam was discovered in the brain — the brain-specific benzodiazepine receptor (henceforth referred to as the benzodiazepine receptor). This receptor is concentrated in those regions of the brain that would, on physiological and anatomical grounds, be predicted to mediate the actions of the benzodiazepines, especially the anticonvulsant and anxiolytic ones. Furthermore, the evolution of this receptor shows a clear increase up the phylogenetic scale with the highest density being in human cortex. The close association of the benzodiazepine receptor with GABA function was discovered by binding experiments that revealed an enhancement of the binding of benzodiazepines by GABA and vice versa. Recently, the receptor has been cloned. It consists of a macromolecular complex of both GABA and the benzodiazepine-binding sites (GABA/benzodiazepine receptor complex). This complex also has binding sites for the barbiturate anticonvulsants and convulsants, such as picrotoxin and pentylenetetrazol ().

It has been realized for some time that the benzodiazepines do not have direct actions of their own in in vitro systems, which is surprising given their potent in vivo actions. The reasons for this have become clear recently as it has been established that benzodiazepines act solely to enhance the actions of GABA, and thus may be considered neuromodulators (as opposed to the directly acting neurotransmitters). GABA acts by increasing the permeability of a chloride ion channel and when benzodiazepine receptors are activated the likelihood of a GABA-receptor interaction causing channel opening is increased. Unlike barbiturates, benzodiazepines do not directly affect chloride flux through the ion channel. This probably explains the relative safety of benzodiazepines in overdose, as compared with the barbiturates.

Alcohols, including ethanol, also enhance the behavioral, physiological, and pharmacological actions of GABA, and this may contribute to many of the actions of these substances. It is, however, unlikely that it explains all the effects of ethanol. In some test systems ethanol has direct actions that are depressant (barbituratelike). These effects tend to occur at higher concentrations than those needed to enhance GABA function. Ethanol may also increase benzodiazepine receptor binding, which could explain the synergistic interaction of these two classes of drug. There is evidence that on chronic administration of ethanol compensatory changes occur in the GABA/benzodiazepine receptor complex. Such alterations of receptor function could account for many of the changes seen in alcohol withdrawal, since when GABA function is downregulated, e.g., by administration of antagonists such as bicuculline, a behavioral profile similar to withdrawal is seen. A possible mechanism for the actions of ethanol on receptor function comes from work in our laboratory that showed a specific depletion of polyunsaturated phospholipid species from the brains of ethanol-treated rats. It seems that these phospholipid s (especially phosphatidylserine and phosphatidylethanolamine)are concentrated around the protein/ receptor moieties in plasma membranes and may be vital for their proper functioning. Their depletion by ethanol could significantly disrupt GABA/benzodiazepine receptor binding or function.

New Classes of Benzodiazepines: A Unique Receptor

The discovery of the benzodiazepine receptor obviously raised the question of the natural ligand for them. This problem is still unresolved () but the search for the endogenous ligand has lead to the discovery of new compounds that have actions very different from those of the classical benzodiazepines, and to the realization that the benzodiazepine receptor is unusual and possibly unique.

In 1980, Braestrup and collegues, extracted a β-carboline (b-CCE) from human urine with very high affinity for the benzodiazepine receptor. When its in vivo actions were examined, it was quite unexpectedly discovered to have an opposite pharmacology to that of the benzodiazepines. It was both proconvulsant and anxiogenic. The more metabolically stable homologue FG 7142 was given to volunteers and this produced severe, paniclike anxiety. These drugs with actions opposite to those of the benzodiazepines have become known as inverse agonists or contragonists. The fact that they act at the benzodiazepine receptor in vivo was established by their ability to reverse the actions of benzodiazepines and, later, by the discovery that the benzodiazepine antagonist Ro 15-1788 (flumazenil) was able to block the effects of both benzodiazepine agonists and inverse agonists ().

Ro 15-1788 is one of several high-affinity ligands for the benzodiazepine receptor that lack intrinsic activity in most behavioral tests but potently and completely antagonize all effects of benzodiazepines. They do not antagonize the actions of ethanol, which argues against ethanol having a direct interaction with the benzodiazepine receptor. These drugs have potential clinical utility in treating benzodiazepine overdose and can be used to determine whether there is activity of endogenous ligands for the benzodiazepine receptor in vivo ().

Inverse Agonists as Alcohol Antagonists (Ro 15-4513)

Although the first inverse agonists were j3-carbolines, it is now established that drugs with other structures have these properties. There has been a lot of interest in one, Ro 15-4513, as a possible alcohol antagonist. This is an imidazodiazepine, i.e., a benzodiazepine, that is a potent partial inverse agonist (see Table 1). Thus it is proconvulsant and anxiogenic. Other inverse agonists (e.g., FG 7142) also oppose many of the behavioral actions of ethanol. However, Ro 15-4513 is reported to be a highly potent antagonist of ethanol’s enhancement of GABA-mediated chloride flux. Since this property is not shared by β-carboline inverse agonists it may represent an important finding about the mechanism of action of ethanol. However, in almost all behavioral tests Ro 15-4513 seems to reverse the actions of ethanol because of its intrinsic proconvulsant or anxiogenic activity. Thus it acts in an opposite direction to ethanol but is not a true antagonist. The convulsant action of Ro 15-4513, especially in alcohol withdrawal, makes the clinical potential of this drug highly questionable.

Are There Endogenous Benzodiazepine Receptor Ligands and Do They Contribute to Alcoholism?

The discovery of the anxiogenic and proconvulsant activities of drugs acting at the benzodiazepine receptor raised possible etiological explanations for a variety of psychiatric and neurological conditions in which benzodiazepines are effective treatments. It may be postulated that since inverse agonists produce a syndrome similar to alcohol withdrawal, the release of an endogenous compound with these properties might occur during chronic alcohol ingestion. This might serve to oppose the sedative and depressant actions of ethanol, but would, when unopposed, lead to withdrawal symptoms. Evidence for this in patients was generated by studies of the urinary excretion of a substance, later called tribulin, which displaced benzodiazepines in binding assays. Increased excretion was observed in withdrawing alcoholics. However, the nature of tribulin has proved difficult to determine and a series of animal experiments have undermined this hypothesis to a large extent. Two groups have used the benzodiazepine antagonist Ro 15-1788 to test the hypothesis directly, since if an inverse agonist is contributing to withdrawal its effects would be fully antagonized. Neither group found any evidence of an effect of the antagonist and it appears that benzodiazepine inverse agonists do not contribute to the signs of alcohol withdrawal in rats. Clinical studies are clearly required to finally decide this issue, but it should be pointed out that for the benzodiazepines animal studies are remarkably predictive of the clinical actions of these drugs, so it is unlikely that the present conclusions will change.

Recently it has become clear that certain naturally occurring steroid metabolites are potently active at several sites on the GABA/benzodiazepine receptor complex, although they do not directly alter benzodiazepine binding. Some may bind to the GAB A receptor and others may act at the picrotoxin site. Since increased levels of corticosteroids are found in alcohol withdrawal (see later), it is possible that this contributes to increased central excitability by reducing GABA function. At the present time the possibility cannot be ruled out that these steroid derivatives are enhancing central inhibition that is down regulated for other reasons.

Another class of putative endogenous ligands are peptides exemplified by diazepam-binding inhibitor (DBI). This peptide has been reported in brain and other tissues, e.g., liver. It displaces benzodiazepines from their binding sites and has been claimed to be anxiogenic in animal tests, and so it is probably an inverse agonist. Whether it is truly the natural ligand for the benzodiazepine receptor is still unclear, especially as its distribution in the body does not mirror that of the receptor. More recently the same group suggested that DBI may be a precursor of a more active ligand: octadecaneuropeptide (ODN). This might be a better candidate for the endogenous ligand but further studies are needed. We have measured DBI concentrations in the CSF of abstinent alcoholics and found no change compared with an age- and sex-matched control population. Similar measurements in alcoholics during withdrawal would be of interest, although, as previously detailed, the lack of effect of benzodiazepine antagonists in ethanol withdrawal argues against activity of endogenous compounds in withdrawal.

Chronic Ethanol and Benzodiazepine Receptors

If endogenous inverse agonists of the benzodiazepine receptor do not explain ethanol withdrawal, is GABA/benzodiazepine receptor complex function altered by chronic ethanol administration in such a way that central inhibition is reduced? A number of studies have addressed this question and results are conflicting. In vivo functional subsensitivity of GABA has been shown in behavioral and physiological experiments, although this may not occur in all brain regions. More recently, it has been suggested that this probably occurs as a result of a decrease in GABA binding, GABA-mediated chloride flux, or even GABAergic neuron number, rather than by a change in benzodiazepine receptor number of affinity (). Furthermore, several investigators have demonstrated that chronic ethanol administration does not alter the binding characteristics of the picrotoxin-binding site, although the coupling between it and the benzodiazepine site may be disrupted.

The downregulation of GABA receptor function after chronic ethanol administration probably explains the well-recognized cross-tolerance between alcohol and the benzodiazepines. It seems that part of the mechanism of benzodiazepine tolerance is a reduction in GABA function that is similar to that found following ethanol administration.

Benzodiazepine Receptors and Alcohol Sensitivity

A number of mouse and rat strains have been selectively bred for ethanol preference or sensitivity. The pharmacological basis of these functional differences are not well understood but research is focusing on the GABA/benzodiazepine receptor complex. It has recently been reported that mice that sleep for a short time after ethanol administration show a greater GABA enhancement of benzodiazepine binding in some brain regions (cortex and cerebellum) than mice that sleep for a long time. Recently, new mouse strains have been bred in which benzodiazepine function is markedly altered and it will be very interesting to know whether they show differential sensitivity to ethanol.

Peripheral Benzodiazepine Receptors

There is now good evidence that some benzodiazepines will also bind to receptors in peripheral tissue — the peripheral benzodiazepine receptor (though it should be noted that in some species these receptors are also found in some brain regions). The structure activity requirements of this binding site are quite different from those of the central-type receptor and it is generally thought that there is no functional relationship between the two. However, there is a single report that chronic ethanol administration increases the binding of Ro 5-4864, a ligand for this receptor, in brain, especially cerebellum, and that this change persists for up to a week.

Benzodiazepines and the Ethanol Withdrawal Syndrome

As presently defined by the DSM-III, alcohol withdrawal occurs upon “the cessation of or reduction in heavy prolonged (several days or longer) ingestion of alcohol, followed within several hours by coarse tremor of the hands, tongue, and eyelids and at least one of the following: (1) nausea and vomiting, (2) malaise or weakness, (3) autonomic hyperactivity, e.g., tachycardia, sweating, elevated blood pressure, (4) anxiety, (5) depressed mood or irritability, or (6) orthostatic hypotension.” Additional symptoms include nightmares, disturbed sleep, itching, hyperacusis, and transitory illusions, hallucinations, and other perceptual disturbances.

Benzodiazepines are the drugs most commonly used for the treatment of the ethanol withdrawal syndrome. The following discussion will first review the pathophysiology of the ethanol withdrawal syndrome and the biochemical rationale behind the benzodiazepines effectiveness in its treatment. The appropriate treatment of the ethanol withdrawal syndrome with benzodiazepines will then be described. This will be followed by a discussion of the implications of the treatment of ethanol withdrawal with benzodiazepines, with suggestions for future research.


As discussed in the previous section, ethanol appears to act, at least in part, at the benzodiazepine/GABA-chloride receptor complex. It is thought that, upon withdrawal, the receptor complex becomes acutely devoid of ethanol’s enhancing effect on chloride flux, so there is a subsequent decrease in GABA functioning. As GABA is the major inhibitory neurotransmitter in the CNS, this results in disinhibition. The consequences may be ethanol withdrawal seizures and other symptoms of the ethanol withdrawal syndrome. Seizures, for example, have been associated with a reduction in GABA function in both animals and man and are also indicative of severe ethanol withdrawal. Benzodiazepines, as well as the barbiturates, may ameliorate the signs and symptoms of withdrawal by substituting for the GABA-enhancing effects of ethanol. Benzodiazepines have been demonstrated to decrease the signs of ethanol withdrawal through the GABA/ benzodiazepine receptor, in that diazepam’s ameliorating effects on withdrawal-associated tremor in the rat is antagonized by Ro 15-1788.

Signs of sympathetic nervous system overactivity, such as diaphoresis, tachycardia, hypertension, and tremor, observed during alcohol withdrawal indicate increased activity of the noradrenergic system. This increase has been substantiated by reports of elevated levels of cerebrospinal fluid (CSF) norepinephrine (NE) and 3-methoxy-4-hydroxyphenylglycol (MHPG), the central metabolite of norepinephrine, in alcoholics during ethanol withdrawal. As NE is highly polar, and does not readily cross the blood-brain barrier, the increase in CSF NE indicates a central elevation in noradrenergic activity. Signs and symptoms of noradrenergic overactivity have also been shown to be positively correlated with CSF MHPG during alcohol withdrawal, and the levels of NE and MHPG in the CSF decrease in parallel with resolution of symptoms. In addition, clonidine, an alpha2-adrenoceptor agonist, which decreases noradrenergic activity, has been demonstrated to reduce many of the signs and symptoms of ethanol withdrawal.

Several lines of evidence suggest that the benzodiazepines may interact with noradrenergic neuronal functioning. Various studies have demonstrated that stress-induced alterations in NE and MHPG are blocked by benzodiazepines and relatively low doses of diazepam and chlordiazepoxide have been reported to decrease spontaneous single-unit activity in the locus coeruleus which contains approximately 50% of the noradrenergic cell bodies in the CNS. Clinically, acute diazepam pretreatment antagonized anxiety produced by yohimbine, an α2-adrenoceptor antagonist. The benzodiazepines may therefore ameliorate symptoms of ethanol withdrawal such as diaphoresis, tachycardia, anxiety, and tremor through an interaction with noradrenergic neuronal functioning, resulting in a decrease in noradrenergic activity.

Activation of the hypothalamic-pituitary-adrenal (HPA) axis has been reported during ethanol withdrawal in both animal and human studies. Elevations in corticosteroid levels have been associated with central nervous system (CNS) structural and physiological changes such as neuronal loss, disrupted neurotransmission and disturbances in the detection and interpretation of sensory stimuli. Alterations in mood are observed in patients with Cushing’s syndrome and corticosteroid therapy. Furthermore, cortisol hypersecretion in depression has been correlated with cognitive impairment. In addition, the severity of withdrawal symptoms, such as seizures and depression, have been correlated with HPA axis activation.

Signs and symptoms of ethanol withdrawal such as fatigue, weakness, hypertension, mental confusion, and depression may therefore be partially related to or consequences of the excessive glucocorticoid levels observed during withdrawal. The benzodiazepines have been demonstrated to decrease stress-induced elevations in ACTH and corticosteroids, and this effect of the benzodiazepines is antagonized by the benzodiazepine antagonist Ro 15-1788. It is suggested that the benzodiazepines may therefore ameliorate the severity of ethanol withdrawal by blocking the associated HPA axis activation.

Benzodiazepine Treatment of the Ethanol Withdrawal Syndrome

Other Complications of Ethanol Withdrawal

Ethanol withdrawal is probably the most common cause of adult onset seizure disorder. Seizures are estimated to occur in 5-33% of alcoholics. The prophylactic treatment for ethanol withdrawal seizures should be initiated only in patients with a history of seizures. These patients should be treated upon presentation with diazepam 20 mg or chlordiazepoxide 50 mg p.o. repeated hourly for the next 2 hr. Patients with significant liver disease may instead be treated with benzodiazepines such as oxazepam or lorazepam in doses similar to those used for ethanol withdrawal. Lorazepam has also been reported to maintain seizure control longer than diazepam. An advantage of the benzodiazepines over phenytoin in the prophylactic treatment of withdrawal seizures is their efficacy in the treatment of other accompanying withdrawal symptoms. When a seizure occurs in ethanol withdrawal it should be treated with the same regimen described above for prophylactic treatment.

Symptoms of alcohol withdrawal delirium, or delirium tremens (DTs), include confusion, disorientation, motor agitation with gross tremor, and autonomic hyperactivity (tachycardia, diaphoresis, and elevation of temperature). DTs usually occurs 3-5 days following the cessation or marked reduction of drinking. For reasons similar to those noted in the treatment of ethanol withdrawal, benzodiazepines should be used to treat DTs. Diazepam 10 mg should be given every 30-60 min until symptoms subside or the patient becomes markedly sedated. Diazepam should be administered intravenously due to the difficulty of administering medications p.o. and the risk of aspiration. Large doses of diazepam may be required; initial doses of 15-215 mg i.v. diazepam have been reported for symptom control or up to 1355 mg over a 4-day period. Prior to each diazepam infusion, the patient should be evaluated for response and symptoms of benzodiazepines toxicity. High-potency neuroleptics may be required for extremely agitated and confused patients, and beta blockers or clonidine may be administered to treat tremors and hypertension that do not respond to the benzodiazepines. It should be noted that the timely and appropriate treatment of the ethanol withdrawal syndrome and coexisting medical illnesses should prevent progression to DTs.

Both ethanol withdrawal seizures and DTs are diagnoses of exclusion; other etiologies such as CNS infection, subdural bleeding, metabolic disturbances, and drug withdrawal as well as other complicating medical problems should be considered.

Cardiovascular symptoms of the ethanol withdrawal syndrome such as arrhythmias and high blood pressure usually require interventions other than the benzodiazepines. Arrhythmias, for example, are most frequently secondary to electrolyte imbalance, and generally respond to appropriate potassium and magnesium supplementation.

Benzodiazepines in the Treatment of Alcoholism: Future Directions

Long-Term Use of Benzodiazepines in Treatment of Patients with Alcoholism

Benzodiazepines are efficacious anxiolytics but they also possess low, yet definite, potential for producing dependence. The desirable and untoward effects of benzodiazepines have to be weighed carefully when considering whether to medicate an alcoholic patient with a benzodiazepine.

Diagnostic Considerations

The trend in psychopharmacotherapy has been away from symptomatic treatment. Rather, diagnostically uniform groups of patients are being treated with medications thought to ameliorate the pathophysiology of a specific mental disorder. Using this model, the following paragraphs examine diagnostic considerations relevant for the choice of pharmacotherapy in patients with alcoholism.

Adoption studies by Cloninger et al. defined two genetically distinct forms of alcoholism, called type I and type II by the authors. Type I is the more common form of the illness. It affected 75% of men alcoholics and all women alcoholics in the sample investigated by Cloninger et al For the genetic predisposition to be expressed, a man who is at risk needs to be exposed to an adverse environment. Women may express the gene effect without environmental provocation. Type II alcoholism afflicts only men. Approximately 25% of alcoholic men have this disorder. It is inherited from fathers to sons, but not daughters, and the gene effect is expressed without environmental provocation. Excessive consumption of alcohol starts at an early age and antisocial personality disorder and criminality coexist with alcoholism in the families. Women with this genetic background often exhibit symptoms of somatization. Type II male alcoholics may have a childhood history of hyperactivity.

Patients with type II alcoholism are unlikely to benefit from benzodiazepines. Indeed, these drugs are relatively contraindicated in such patients, particularly if they exhibit impulsive behaviors, which may be aggravated by benzodiazepines. The biochemical basis of alcoholism and impulsivity in type II alcoholics may be that these patients have deficient CNS serotonin function. Benzodiazepines may further reduce serotonin turnover. If psychotropics are indicated in type II alcoholics to treat symptoms such as poor impulse control and dysthymia, medications such as carbamazepine and lithium are probably preferable. However, no controlled studies are available on pharmacological treatment of type II alcoholism with or without other mental disorders.

Type I alcoholics free of other mental disorders probably will not benefit from any pharmacotherapy. Patients with type II alcoholism who exhibit symptoms or have a history of other mental disorders should be divided into primary and secondary alcoholics depending on the relative times of onset of alcoholism and other mental disorders. This classification does not apply to type II alcoholics with a history of hyperactivity because the hyperactivity may represent an age-appropriate expression of the gene effect. Patients with primary type I alcoholism and secondary mental disorder other than alcoholism need to maintain abstinence and the secondary mental disorder needs to be treated vigorously. Successful treatment of alcoholism as such may reduce symptoms of the secondary mental disorder.

Patients with secondary alcoholism should be treated for the primary mental disorder. The treatment of the primary mental disorder is expected to reduce the risk of relapse to drinking, particularly if drinking is the result of an attempt to self-medicate with alcohol. The issue of alcoholics being at a high risk to be addicted to medications needs to be considered in choosing appropriate pharmacotherapy. Furthermore, measures to encourage abstinence need to be actively promoted because concomitant use of psychotropics and alcohol may, over and beyond being dangerous, block therapeutically necessary pharmacological effects of psychotropics.

Alcoholism and Other Mental Disorders

Anxiety, eating, and affective disorders are more common among alcoholics and their families than the rest of the population. The relationship between anxiety and alcoholism is particularly complex because alcohol can both alleviate and exacerbate anxiety. The approach to pharmacotherapy of anxiety in alcoholics depends on a careful differential diagnosis of anxiety disorders.

Alcohol and Panic Attacks

Alcoholism and Social Phobias

Social phobias can be successfully treated with beta blockers administered p.r.n. A behavioral treatment program involving exposure may be effective with or without adjunctive pharmacotherapy. In patients with alcoholism, p.r.n. benzodiazepines should be avoided due to a risk of producing dependence. Beta blockers, on the other hand, do not produce dependence. If beta blockers are used in an alcoholic patient who experiences difficulty in maintaining sobriety, a drug such as atenolol, whose elimination is primarily dependent on kidney rather than liver function, is probably preferable.

Alcoholism and Simple Phobias

Behavioral therapy, in the form of desensitization, is the current treatment of choice for simple phobias.

Alcoholism and Generalized Anxiety Disorder

Until the early 1980s benzodiazepines were thought to be the treatment of choice for generalized anxiety disorder with or without symptoms of depression. Two relatively large-scale studies, one American and the other English, challenged this impression. They found that imipramine and amitriptyline were more effective in the long-term treatment of generalized anxiety than chlordiazepoxide or diazepam, respectively. This finding was true regardless of the severity of depression in the patients. Thus, tricyclic antidepressants may be preferred to benzodiazepines in the treatment of generalized anxiety in alcoholics.

Buspirone, a somewhat specific 5-HTiA receptor agonist, is anxiolytic in humans. It is different from benzodiazepines because it does not have adverse cognitive or psychomotor interactions with alcohol. Furthermore, it has little or no abuse potential in humans including alcoholics, and it shares the advantage of the benzodiazepines of being relatively nontoxic when overdosed. Thus, buspirone appears to be close to an optimal anxiolytic in alcoholics if it proves to be efficacious in treating generalized anxiety in these patients. Theoretically, similar to serotonin reuptake inhibitors, buspirone may reduce drinking as well.

Alcoholism and Depression

Depression is especially common in women with alcoholism. The only benzodiazepine reported to be equiefficacious to imipramine in the treatment of depression, based on a controlled large-scale study, is alprazolam. Relatively high doses of it are needed for the treatment of depression, and administering it to alcoholics may represent an unacceptable risk of producing dependence. The current treatment of choice for depression in patients with alcoholism is probably secondary amine tricyclic antidepressants such as desipramine or nortriptyline. This may, however, change in the near future in favor of relatively selective serotonin reuptake inhibitors. Strict maintenance of abstinence may be necessary for antidepressants to manifest their therapeutic effects.

Alcoholism and Symptoms of Other Mental Disorders Not Fulfilling DSM-III Diagnostic Criteria: Insomnia

Especially after withdrawal, most alcoholics experience disrupted sleep. Whether most patients with type I alcoholism have abnormal sleep patterns even during long-term abstinence is not clear at the present time.

In abstaining type II alcoholics, insomnia may be an effect of the postulated serotonin deficit producing a functional disturbance of the suprachiasmatic nucleus. Benzodiazepines are not recommended for these patients.

After other sleep disorders have been carefully ruled out, benzodiazepines and other hypnotics are indicated only in the treatment of transient insomnia. The similarities in the mechanisms of action of alcohol and benzodiazepines have been described above. Furthermore, alcoholics may be at an increased risk of developing dependence. Thus, hypnotics not affecting the GABA-benzodiazepine-chloride ionophore complex, such as sedative antihistamines, are probably preferable in patients with alcoholism. Again, behavioral and sleep hygiene improving interventions should be exhausted prior to resorting to medication.

Anxiety and Depression

Severity of symptoms of anxiety and depression correlates positively with the severity of alcoholism. Moreover, anxiety is reduced over time during abstinence. When not a component of a diagnosable mental disorder, symptoms of distress may be direct consequences of heavy alcohol consumption and therefore improve during abstinence even without pharmacological treatment.

Are Benzodiazepines Indicated in Alcoholism?

As the above review reveals, benzodiazepines are seldom needed for long-term treatment of mental disorders in patients with alcoholism. Furthermore, they should not be used to treat alcoholism, other than withdrawal per se, because no studies have demonstrated efficacy of benzodiazepines in this condition. One of the few indications for benzodiazepine in alcoholics is adjustment disorder with anxiety since other medications have not been demonstrated to be effective in this disorder. If they are to be administered to alcoholics certain differences between the various benzodiazepines have to be considered when choosing a particular compound. These differences fall into three partially overlapping categories: (1) abuse potential, (2) facilitation of aggressive outbursts, and (3) pharmacokinetics.

Relative Abuse Potential of Various Benzodiazepines

In the 1980s, a series of human and animal studies were conducted to investigate abuse potential of benzodiazepines. Compared with other psychotropics, diazepam has been observed to have less abuse potential than pentobarbital and more than chlorpromazine. Among the benzodiazepines, diazepam and lorazepam seem to have a higher abuse potential than chlordiazepoxide and, especially, oxazepam.

Benzodiazepine Dependence

If emergence of withdrawal symptoms upon stopping medication is taken as a sign of dependence, then two forms of benzodiazepine dependence can be distinguished: (1) dependence on regular therapeutic doses and (2) dependence on high doses, usually escalated by the patient.

A multitude of symptoms have been described to appear in some patients upon discontinuation of long-term treatment with low doses of benzodiazepines. These occur more commonly in patients who have received the drugs for more than four months. When giving a history, many of these patients maintain that their symptoms are newly emergent. The symptoms are thought compatible with an exacerbation of anxiety as well. The lack of detailed longitudinal studies describing patients’ symptoms prior to starting treatment hampers full characterization of the syndrome. Nevertheless, stopping treatment with benzodiazepines is best accomplished by gradual tapering of the dose and long-term treatment should be avoided whenever possible.

Only a very small percentage of patients escalate the dose on their own and become abusers of benzodiazepines. Whether alcoholics are more likely to escalate their dose than the rest of the patients receiving benzodiazepines has not been established. It is, however, only prudent to assume this to be the case. To avoid severe withdrawal symptoms, such as seizures, benzodiazepines have to be tapered under close supervision in abusers.

Facilitation of Aggressive Outbursts by Benzodiazepines

Use of benzodiazepines is occasionally associated with unexpected outbursts of aggressive or violent behavior. Studies on healthy volunteers have found that in situations designed to provoke hostility, chlordiazepoxide facilitates its expression more than oxazepam. The clinical relevance of these findings is unknown, but administration of benzodiazepines to type II alcoholics should generally be avoided.

Relevance of Pharmacokinetics for Choice of Benzodiazepines

If a clear indication to use a benzodiazepine for treatment other than withdrawal exists in a patient with alcoholism, certain pharmacokinetic differences may be of relevance.

Alcoholics commonly have gastrointestinal symptoms, which are often treated with histamine 2 receptor antagonists or antacids. Cimetidine inhibits hydroxylation of benzodiazepines such as diazepam, chlordiazepoxide, and clorazepate, flurazepam, and alprazolam. Ranitidine, another histamine 2 receptor antagonist, does not inhibit diazepam hydroxylation. The clearance of benzodiazepines is also reduced in cirrhosis of the liver and by antituberculosis drugs. The pharmacodynamic consequences of these metabolic alterations are probably relatively minor, but theoretically lorazepam and oxazepam offer advantages due to their unaltered pharmacokinetics under these circumstances.

Some physicians still prescribe disulfiram to alcoholics despite lack of proven efficacy and potential neurotoxicity. Disulfiram reduces the clearance of diazepam, but the clinical significance of this interaction is unknown.

Benzodiazepines in the Treatment of Alcoholism: Conclusions

(1) Benzodiazepines work by enhancing the affinity of GABA to the GABA A receptor. The consequence of this action is increased chloride conductance across the cell membrance and in most neurons hyperpolarization of the cell. This inhibits neuronal firing. Behaviorally, these pharmacological effects are expressed as anxiolysis, sedation, and anticonvulsant activity.

(2) One of alcohol’s many pharmacological effects is to increase chloride flux. This may be the mechanism of the anxiolytic effect of alcohol as well as the source of cross-tolerance between benzodiazepines and alcohol. Whether this mechanism contributes to the abuse potential of benzodiazepines in alcoholics is unknown.

(3) Benzodiazepines are the treatment of choice for alcohol withdrawal.

(4) Benzodiazepines are not indicated in the treatment of alcoholism per se.

(5) There are few indications for long-term treatment of alcoholics with benzodiazepines. The likelihood of abuse should always be considered when prescribing benzodiazepines to an alcoholic.

(6) Oxazepam probably has less abuse potential than diazepam and lorazepam.


Selections from the book: “Recent Developments in Alcoholism. Volume 7. Treatment Research: Alcoholics Anonymous. The Family. Serotonin and Preference. Clinical Pharmacology.” Edited by Marc Galanter. An Official Publication of the American Medical Society on Alcoholism, the Research Society on Alcoholism, and the National Council on Alcoholism. 1989.