I REVIEWS

Disulfiram Treatment of Alcoholism CURTIS WRIGHT, M.D., M.P.H., RICHARD

D. MOORE,

M.D., M.H.Sc.,

PURPOSE: F o r 40 years, disulfiram has been the alcohol-aversive drug used most frequently by American physici~n~ in the treatment of alcohol dependency disorders. We reviewed the clinical lite r a t u r e regarding the risks, benefits, indications, and efficacy of this controversial drug and summarized current knowledge of this therapy. CONCLUSIONS: Disulfiram will produce an aversive reaction with ethanol, usually at a dose between 250 mg/day and 500 mg/day, although some patients may not have an aversive reaction at this level. Cardiac, hepatic, and neurologic toxicity can also o c c u r within this dosage range. If disulfiram is to be used, the patient must clearly understand the risks of drinking while t~ldng the drug, and the physician and patient must agree about the need for continued clinical supervision and monitoring for efficacy and side effects. T h e physician must also recognize that disulfiram is only an adjunctive therapy and that continued support, supervision, and other therapeutic measures are required. Disulfiram is probably effective in reducing the frequency of alcohol consumption in the compliant patient over the short term (e.g., 6 months). Certain subgroups of patients, such as those who are older, those who are more socially stable, and those who are well-motivated, may experience a beneficial effect for longer periods. The drug may be most effective in reducing short-term alcohol consumption when the compliance of the patient is supervised, although consideration of this kind of therapy includes the practical problems of supervising the patient and concerns that the supervising person may be placed in a difficult position. Prescription of disulfiram without accompanyin~ education, counseling, and concomitant alcoholism therapy is not beneficial. Disulfiram has no proven effect on the long-term outcome of alcoholism.

From the Division of Internal Medicine, Department of Medicine, and the Division of Behavioral Biology, Department of Psychiatry, the Johns Hopkins University School of Medicine, Baltimore, Maryland. Support for Dr. Wright was provided by United States Public Health Service Grant USPHS/ NIH 5-32 DAO 7209. Dr. Moore is a Burroughs Wellcome Scholar in Pharmacoepidemiology. Requests for reprints should be addressed to Richard D. Moore, M.D., M.H.S., Johns Hopkins University School of Medicine, 1830 East Monument Street, Room 8059, Baltimore, Maryland 21205. Manuscript submitted September 20, 1989, and accepted in revised form December 28, ]989.

Baltimore, Maryland

lcohol dependence is a common substance use disorder in the United States, if not in the world. In A caring for the alcoholic patient, the clinician is confronted with the need to provide treatment that will eliminate abusive drinking. To provide this treatment, clinicians must learn to identify and confront the active alcoholic in their practice, manage the patient's detoxification from alcohol, and provide basic information regarding further treatment alternatives. In addition to these basic interventions, alcohol-aversive drugs are also available to treat alcoholism, and surveys have suggested that about half of all practitioners who treat alcoholism use these drugs [1,2]. This article reviews the data on the pharmacology of chemical aversive agents used in the management of alcoholism, as well as the behavioral and clinical issues involved in their use. Indications for such therapy, its risks, and the potential benefits are also covered. To ensure that all relevant articles were included in this review, we searched under disulfiram, calcium carbamide, and metronidazole and alcoholism treatment in MEDLINE, the National Institute on Alcohol Abuse and Alcoholism (NIAAA) database on alcoholism, and the NIAAA monograph series on treatment effectiveness in alcoholism and on the effectiveness of disulfiram. The references of the primary and review articles thus found were then manually searched in an iterative fashion until no new references were found. Articles were included in the review if they were the primary reports in the literature of the results of clinical or basic research on aversive drug therapy. Articles were excluded if it was not possible either to define the treatment or to quantify treatment outcomes. Preliminary analysis showed that disulfiram has been used in very different ways since it was introduced, and the papers were grouped accordingly.

THE HISTORYOF ALCOHOL-AVERSIVEDRUGS AversionTherapy Aversion therapy is a behavioral treatment technique in which a controlled drinking situation is linked to an electrical, chemical, or behavioral aversive stimulus administered by the therapist. It is an old technique that has been reviewed by Nathan [3], who describes the use of emetine, electric shock, succinylcholine-induced paralysis, aversive images, and a host of other aversive events in a comprehensive review of the treatment of alcoholism by aversive therapy. Aversive conditioning can often be difficult to administer, and is relatively uncommon in this country. Although classical aversive conditioning is restricted to a few centers that report a consistently moderate success rate [4-6], another form of aversive therapy using alcohol-aversive drugs has become quite common.

AlcohoI-AversiveDrugs The two major alcohol-aversive drugs, disulfiram and carbamide, are both industrial chemicals that June 1990 The American Journal of Medicine Volume 88

647

DISULFIRAMTREATMENTOF ALCOHOLISM/ WRIGHTAND MOORE

were noted to produce distinct occupational illness in workers who handled these substances and then drank alcohol. Disulfiram was first reported to be an alcoholaversive agent by an occupational medicine physician in 1937 [7], and carbamide followed in 1956 [8]. Reports of adverse reactions to alcohol in subjects taking the antitrichomonal drug metronidazole led Taylor [9] to suggest its use in alcoholism treatment. The appeal of these drugs was that the aversive reaction was due to the drug's effect, not to the action of the therapist, and would occur in any drinking situation. Following the publication of the first reports of the clinical use of disulfiram in 1949, drugs of this type began to be used enthusiastically and widely in the treatment of alcoholism.

(resulting in very low plasma levels), and is excreted as diethyldithiocarbamate and its conjugates, diethylamine and carbon disulfide. Disulfiram is a relatively non-specific irreversible inhibitor of sulfhydryl-containing enzymes. The target enzyme for the pharmacologic effect of the drug is aldehyde dehydrogenase, an NAD-requiring enzyme, which converts acetaldehyde to acetate in the metabolism of ethanol. Aldehyde dehydrogenase is not the only enzyme the drug affects; dopamine ~-hydroxylase, superoxide dismutase, and some of the enzymes of intermediary metabolism are reduced in activity as well. Ingestion of a single dose of oral disulfiram will begin to affect ethanol metabolism within 1 to 2 hours, peak at 12 hours, and recede over 12 to 72 hours, depending on the rate of new enzyme synthesis [26].

Calcium Carbamide Disulfiram is the only alcohol-aversive agent marketed for use in the United States, but in order to consider disulfiram in its proper context, it is helpful to also discuss calcium carbamide and metronidazole. In 1953, concern over the severity of the ethanol-disulfiram reaction and a desire for a less aversive drug led the Alcoholism Research Foundation of Ontario to fund basic and clinical research into finding alternative aversive drugs for the treatment of alcoholism. Carbamide, another industrial compound with known alcohol sensitizing effects, was suggested for use in alcoholism treatment by Ferguson [8] in 1956. The citric acid salt (citrated calcium carbamide, Temposil, Abstem) was introduced into Canadian and European practice following a series of reports [10-16] demonstrating that the usual daily dose of 50 mg of this drug produced a milder aversive reaction than disulfiram, did not inhibit dopamine #-hydroxylase, produced less severe cardiovascular effects, and was clinically equivalent in trials in volunteers and in short-term trials in alcoholic patients. Continued work at the Addition Research Foundation [17] has shown that the rapid onset and short duration of action of calcium carbamide may be important advantages of this drug in some clinical situations. Calcium carbamide is prescribed about half as frequently as disulfiram in Sweden [18], where both drugs are available, but it is not currently available to the American physician. Metronidazole In 1964, Taylor [9] used the antitrichomonal drug metronidazole as an alcohol aversive and reported that the drug was effective not only in producing an aversive response, but also in reducing the craving for alcohol. Several authors [19-23] attempted to replicate these results and were not able to demonstrate either an anti-craving or a reliable aversive effect. Rothstein and Clancy [24] initiated and then abandoned a trial using a combination of disulfiram and metronidazole due to a greatly increased incidence of toxic psychosis (20%) in patients taking both drugs. The Food and Drug Administration has never approved the marketing of metronidazole as an alcohol-aversive drug.

DISULFIRAM (ANTABUSE) Pharmacology Faiman [25] has extensively reviewed the pharmacology of disulfiram (Antabuse, Antabus) in an excellent article. The drug is absorbed rapidly following an oral dose, attaches to visceral tissues within an hour 648

June 1990 The American Journal of Medicine Volume 88

The EthanoI-DisulfiramReaction The ethanol-disulfiram reaction is presumed to be due to the combination of the effects of circulating acetaldehyde with an intrinsic host susceptibility; it consists of a highly aversive episode of flushing, weakness, and nausea, as well as occasional hypotension. The drug is usually given in an oral form as a liquid or tablet, and a wide range of doses have been used in clinical practice. The medically supervised ethanoldisulfiram reaction was commonly performed in the years between 1949 and 1955. This method [27-34] consisted of detoxifying the patient from alcohol, loading him or her with several days of disulfiram therapy, and then having the patient drink alcohol under medical supervision. Following the reaction (which might be repeated several times), the dose of disulfiram was adjusted to the minimum individual dose that would produce an aversive response. By 1950, myocardial ischemia during such treatment was reported [35], and in 1952 the first report of a fatal disulfiram reaction was published in J A M A [36]. This was followed by Jacobsen's [37] report of 26 fatalities in an experimental study by Markham and Hoff [38], which suggested that more than 90% of patients drinking while taking disulfiram experienced transient electrocardiographic changes suggestive of ischemia. These reports led most clinicians to abandon the provoked ethanol-disulfiram reaction as having too great a risk for the claimed benefits and to begin to use the drug to prevent drinking rather than as an aversive stimulus.

Dosage Although it is known that very large doses of disulfiram will produce an ethanol-disulfiram reaction, the minimum dose which must be given to produce a reaction is uncertain. Experimental studies in volunteers have shown doses of 100 mg/day [39] to 250 mg/70 kg [40] to not reliably cause a reaction. Brewer [41] studied alcoholic patients who either had had a supervised ethanol-disulfiram challenge or who drank while taking the drug. In that study, 30 of 63 alcoholic patients failed to show a response when challenged with alcohol following a dose of 200 mg/70 kg of disulfiram daily. Table I lists the loading doses and the range of maintenance doses required to produce objective reactions in the early studies of induced ethanol-disulfiram reactions [27-34]. As can be seen from the table, the maintenance dose of disulfiram required to elicit a supervised ethanol-disulfiram reaction is between

DISULFIRAM TREATMENT OF ALCOHOLISM / WRIGHT AND MOORE

0.125 g/day and 1.00 g/day, with the median effective dose being about 300 rag/day to 500 rag/day. Since it is now possible to measure aldehyde dehydrogenase activity and to determine the degree of inhibition for an individual patient, such monitoring may allow more precise individualization of the dose in future research. Toxicity

Many patients complain of a variety of subjective side effects from disulfiram, but a well-designed, randomized, double-blind clinical trial of 241 patients conducted by Christensen et al [42] failed to show any more subjective or objective side effects in subjects taking 200 mg/70 kg of disulfiram daily than with a placebo. Disulfiram does alter the metabolism of many drugs dependent on hepatic metabolism [43-46] and may alter the metabolism of endogenous amines. Inhibition of dopamine ~-hydroxylase and other monoamine enzyme systems has been implicated in the development of hypertension [47], altered serotonin metabolism [48], and altered sympathetic tone [49] in patients taking the drug. Disulfiram has been associated with serious neurologic side effects. Reports of a toxic psychosis from disulfiram first appeared in 1950 [50] in European journals and in 1951 in English [51]. The clinical presentation of this condition became progressively clearer as more cases were accumulated [52-55]. Patients developing this syndrome complain of progressive fatigue, forgetfulness, and confusion after beginning disulfiram. This reaction may occur in as many as 5% to 25% of patients taking doses in excess of 500 mg/day, and can progress to affective changes, ataxia, stupor, and a frank toxic psychosis or encephalopathy if the drug is continued. Although the exact mechanism of such an encephalopathy is unknown, one author [56] has suggested that the mechanism is metabolically produced carbon disulfide. As with most drugs toxic to the central nervous system (CNS), peripheral neuropathy with disulfiram has been reported but is claimed by one author [57] to be infrequent in doses under 500 mg. In addition to the reports of CNS toxicity, Keefe and Smith [58] reported two cases of acute hepatitis following disulfiram administration. Many reports of similar cases followed [59-61], with histologic examination [62] finding atypical inclusion bodies in the hepatocytes of alcoholics treated with the drug. Iber and co-workers [63] reviewed the routine serum chemistry values obtained as part of the Veterans Administration Cooperative Disulfiram Study and were not able to demonstrate a significant association between disulfiram administration and subclinical hepatitis, but there was a low overall compliance with taking the drug in this study. EFFECTIVENESS OF DISULFIRAM T H E R A P Y Non-Randomized Controlled Studies

Disulfiram was introduced into clinical practice on the basis of anecdotal reports, case series, and uncontrolled trials. Despite the poor methodology of these early studies, they constitute the majority of the disulfiram literature. The first report of large-scale use of disulfiram without a supervised drinking session was reported by Hoff [64] in 1955. He reported a case series of 1,020 alcoholics who chose to take disulfiram and

TABLE I Doses of Disulfiram Used in Medically Supervised EthanoI.Disul. firam Reactions

Reference

Day I (g)

Day 2 (g)

Day 3 (g)

[27] [28] [29] [30] [31] [33] [34]

2.0 1-2 2.0 2.0 2.0 1.0 2.0

1.5 1.0 1.5 1.5 1.5 1.0 1.5

1.0 0.75 1.0 1.0 1.0 0.750 1,0

Thereafter (g) 0.125-1.000 0.250-0.750 0.250-0.500 0.125-0.500 0.250-1.000 0.250-1.000 0.250

TABLE II Effect of Disulfiram on the Percentage of Clients in Remission for Each Facility Type*

6 Months Disulfiram No disulfiram 12 Months Disulfiram No disulfiram

Facility Type 0 HO

I0

Total

82 80

80 63

75 56

73 67

77 64

68 56

66 64

74 74

62 67

68 66

H

I

73 67 76 69

Facility H = hospitalprogram;Facility I = intermediateprogram;Facility0 = outpatient program; FacilityHO = hospitaloutpatient;FacilityIO = intermediateoutpatient. * Randreport disulfiramdata [65] (adiustedfor severity, duration, socialstability, age, and race).

484 who refused to take the drug, and described a significant improvement in outcome among those choosing drug therapy. Of greater significance than these results was the description by the author of how the experience of an ethanol-disulfiram challenge reaction was omitted in cases where the patient was allowed to observe the reaction in another patient. When the increasing reports of danger to the patient during such reactions led to the abandonment of their use, physicians began to give the drug to patients and then warn them not to drink. The drug was no longer being used as an aversive agent, but was now being used as an agent to prevent drinking. This presents the paradox that the desired outcome of drug therapy with disulfiram is that the patient will never experience the pharmacologic action of the drug. With the replacement of an experienced ethanoldisulfiram reaction by a physician's warning to avoid this reaction, the drug began to be used as an adjunct to the outpatient management of alcoholism. In one of the largest surveys of the effectiveness of such therapy, Armor et al [65] published the outcomes of patients selecting disulfiram in a number of treatment settings in the Rand report. These data are summarized in Table II. There was a better short-term (6 months) outcome among those taking disulfiram, but the effect was lost at 12 months, and there is no way to determine if the improvement was due to a selection bias or to the disulfiram. Other case series have also suggested that patients who accept and comply with disulfiram may have a better outcome than those who do not [66,67]. In contrast, Schuckit [68] investigated this question in a large (550 patient) study comparing veterans who elected to take disulfiram with those who June 1990 The American Journal of Medicine Volume 88

649

DISULFIRAMTREATMENT OF ALCOHOLISM/ WRIGHTAND MOORE

refused to do so. In this 1-year follow-up study, the intake interviews of the disulfiram acceptors differed from those of the disulfiram refusers with regard to a number of measures of pre-morbid severity, although no effect of disulfiram was noted in terms of any therapeutic outcome variable (drinking, medical complications, or behavioral impairment) as measured by a series of follow-up interviews with the patient and verified by a community resource person who documented the patient's self-report. Other studies on the possible bias introduced by self-selection of disulfiram treatment [69-71] suggested that patients who accepted disulfiram did differ from those who refused the drug, but analyses are complicated by reports that even disulfiram acceptors comply poorly with the drug. From these reports, we do not know if an improved outcome with disulfiram is due to the drug or to other factors such as the patient's motivation or some other type of selection bias. Randomized Clinical Trials

There have only been two randomized clinical trials of disulfiram of adequate power that have addressed the pharmacologic and psychologic effects of disulfiram as an adjunct to detoxification and counseling. Fuller and Roth [72] approached the problem of disentangling the pharmacologic and the psychologic effects of disulfiram in the first such study in 1979. These authors reasoned that there must be not one, but two, control groups in any trial of disulfiram in addition to the group receiving the active drug. They suggested that the pharmacologic action of the drug and the psychologic action of the drug were, in fact, separate variables, and must be so addressed by their research design. Ethics forbid the giving of disulfiram to any patient without his or her knowledge, and the physicians could not deceive the patients by telling them they were receiving medication when this was not true. They solved this dilemma by designing a double-blind placebo-controlled trial in which the original patient cohort was divided into not two but three groups. Group 1 received 250 mg of disulfiram and Group 2 received 1 mg (a pharmacologically ineffective dose); both groups were told that they were receiving disulfiram and warned most strongly not to drink by physicians who were unaware of the dose. Patients in Group 3 received a placebo in the form of a multivitamin tablet and were told they were not receiving disulfiram. Thus, in this trial, there was a usual care group, a group that received a disulfiram "warning" but a placebo dose, and a group that received a pharmacologic dose and a "warning." In addition, all medications used in this trial were marked with a riboflavin marker t h a t could be used to measure compliance. After 6 months of follow-up, there was a small but statistically significant increase in abstinence from alcohol between patients taking disulfiram and placebo, but there was no difference between the group taking the 1-mg "placebo dose" and the group taking the full 250-mg pharmacologic dose. This study suggested that the major therapeutic intervention effect was due to the patient's fear of an ethanol-disulfiram reaction and not to a subjective experience of the pharmacologic action of the drug. Measures of compliance by urinary riboflavin assay were measured across all three groups and were not significantly different. 650

June1990 The American Journalof Medicine Volume88

Fuller and others [73] used this methodology again in a 605-patient multicenter trial reported in 1986. That study was drawn from a group of 6,629 patients being seen at nine Veterans Administration hospitals. Selection criteria for the study were that the patient had no major illness other than alcoholism, lived with an individual who could verify the patient's history, was younger than age 69, had no history of drug abuse or destructive behavior, lived less than 80 km from the hospital, and met the National Council on Alcoholism's diagnostic criteria for alcoholism. Of the 6,629 patients presenting, 5,018 were ineligible by these criteria and 1,006 refused to participate. The 605 patients who remained were randomly assigned to one of three conditions (250 mg/1 mg/placebo), and the groups were checked for homogeneity on numerous severity measures. Of the study group, 28 dropped out, and 577 were followed for the full follow-up period. Neither the Veterans Administration alcoholic population nor the subpopulation eligible for this study can be expected to be representative of the entire alcoholic population, since both have been selected by a number of criteria that affect prognosis. However, the study population was from nine hospitals from both coasts and the Midwest, located in rural and urban areas, and results may well be generalizable to male, blue-collar, high school graduates. After randomization, the patients and their respondents were surveyed at 2-month intervals, and biologic samples were collected at these times. Outcome measures were abstinence, time to first drink, drinking days, employment status, and place of residence. No differences were found between the groups with regard to length of abstinence, time to first drink, employment, and social stability. However, there was a significant and graded relationship between the number of drinking days in 12 months and the dose of disulfiram. For those receiving 250 mg of disulfiram, there were 49.0 =E8.4 (mean + SD) drinking days; for those receiving 1 mg, there were 75.4 + 11.9 drinking days, and for those receiving placebo, there were 86.5 + 13.6 drinking days. It was also found that older, more socially stable men were most likely to have a beneficial effect lasting as long as 1 year. Compliance with Disulfiram Therapy

Jacobsen and Martinsen-Larsen [74], in one of the earliest reports in English regarding disulfiram, made the cogent remark that "as long as the patients take the tablets, they will not drink." In the case of many patients, this means that they quickly stop taking the tablets in order to drink. Paulson et al [75] developed a breath test for carbon disulfide, which showed that only 25 of 52 outpatients claiming to be taking the drug were actually doing so. Similar results were found when Gordis and Peterson [76] evaluated a urine test for disulfiram metabolites. Fuller and Neiderhiser [77] reported the compliance with therapy in a group of 21 outpatients taking disulfiram for I year as part of a clinical trial of disulfiram efficacy. In that series of samples from 10 men randomly assigned disulfiram, 76 of 149 samples (51%) tested positive for the drug's metabolites. In the first of Fuller et al's [78] clinical trials, 22% of the patients were rated as compliant as defined by having 15 or more positive urine specimens for riboflavin. In Fuller and co-worker's [73] second study, the overall compli-

DISULFIRAM TREATMENT OF ALCOHOLISM / WRIGHT AND MOORE

ance rate as measured by riboflavin marker was again 20% and was found to be constant across all conditions. Fuller et al reported a significant relationship between disulfiram compliance and treatment outcome. Of those patients who were found to be compliant with drug or placebo, 43% were abstinent in contrast with an 8% abstinence rate among noncompliant patients. This relationship was stable across all treatment groups regardless of the drug condition (Table III). Fuller's work with riboflavin markers has not yet been fully replicated, and it is not known how accurate this method is in measuring compliance. In a similar study of lithium use in alcoholic populations (where compliance could be very accurately determined by serum lithium levels), medication compliance was a more powerful predictor of outcome than the drug effect being studied [79]. Although compliance with drug therapy seems to be strongly associated with a positive treatment outcome, these results are not definitive, as neither of the trials stratified patients on compliance before randomization to treatment conditions. Compliance is a confounder of the assessment of disulfiram efficacy, and various strategies have been proposed to deal with this compliance, one of which was surgically implanted disulfiram tablets.

Disulfiram Implants Clinicians were well aware that many alcoholics would accept disulfiram but might not take the medication for any extended period of time. As an alternative to oral therapy, Kellan and Wesolkowski [80] described a technique of disulfiram implantation in a 1968 case report in L a n c e t . Two reports followed that suggested that the implants resulted in a significant improvement in abstention [81,82] and in marriage, job, and social functioning [81]. These were promising results, but further studies indicated that patients were returning to drinking after implant without an ethanol-disulfiram reaction [83-85]. These findings led Wilson et al [86] to study the 2year post-implantation sobriety of 10 patients who had disulfiram implanted and 10 randomized control subjects who had sham operations. These investigators reported that there was no significant difference between the two groups in the first 6 months of followup, but that there was a small effect at 1 to 2 years. These authors [87] later added more cases to their series by randomizing additional patients until they had a total of 40 patients who had accepted disulfiram and who received an implant, 40 patients who had accepted disulfiram and had received a placebo implant, 10 patients who had accepted disulfiram and had the operation delayed, and 10 disulfiram refusers. The major outcome in this study was days of abstinence from alcohol. The patients who had implants averaged 307 to 331 days of abstinence in 365 days of follow-up, whereas the patients who had not had implants averaged only 24 to 31 days. There was no significant difference between the patients who had dis u l f i r a m i m p l a n t s and t h o s e who had p l a c e b o implants, whereas those who had no implant did more poorly regardless of apparent motivation. The conclusion that the primary effect of disulfiram implantation was behavioral and not pharmacologic is further supported by two European studies. Morland

TABLE III Relationshipof Drug Complianceto AbstinenceCompared AcrossDrug Treatment Classes*

Treatment Class

Percentage Abstinent Compliant Non-Compliant (20% of total) (80% of total)

250 mg Disulfiram 1 mg Disulfiram Placebo

38 50 43

10 9 6

Across all classes

43

8

From[73].

et al [88] carefully studied the reactions in normal volunteers and alcoholic patients with implanted disulfiram both by measuring metabolites of disulfiram in the blood and by giving alcohol challenges. These authors found that the amounts of disulfiram absorbed were very low, and they showed the implants to be without significant pharmacologic effects. Borg et al [89] reported on a placebo implant controlled trial of 6 months' duration that randomized 206 patients into two groups that received either disulfiram or placebo implants. All patients were followed every month for 6 months by physical examination, physician interview, and symptom and alcohol-use questionnnaires by physicians unaware of the type of implant. These authors were able to follow 171 of 209 patients (83%) and determined that 56 of the 209 patients completed the study without relapse. They compared the treatment and control groups with respect to time to relapse, social conditions, medical care utilization, fidelity of f o l l o w - u p , d r i n k i n g days, and r e p o r t e d ethanol-implant reactions. In no case was a significant difference found between the placebo and the disulfiram implant groups. These results were confirmed in still another more recent randomized double-blind clinical trial [90]. The evidence from these studies argues persuasively against the effectiveness of the disulfiram implant, and such implants are not licensed for use in the United States at this time.

Supervised Disulfiram Supervised administration of oral disulfiram is an alternative method of ensuring compliance with the drug. Bourne et al [91] reported an uncontrolled experimental program in Atlanta that offered disulfiram treatment as an alternative to punishment in indigent, alcohol-related recurrent judicial offenders. In this study, daily taking of disulfiram was supervised by a probation officer, and a 3-month follow-up reported a 50% abstinence rate in 64 patients so treated. Haynes [92] reported an uncontrolled study on the use of supervised disulfiram given on alternate days in 141 alcoholic offenders with multiple arrests on parole and demonstrated a response to treatment over 1 year in 46% of the cases at follow-up. These findings were supported by further work by Brewer and Smith [93] in a case series of 16 habitual criminal offenders given up to 375 mg of disulfiram daily as a condition of parole. T h a t series showed that, over 1 year, supervised disulfiram use was effective in preventing drinking in 12 of the 16 criminal offenders offered disulfiram as an alternative to incarceration. Sereny et al [94] used thrice weekly mandatory (a June 1990 The American Journal of Medicine Volume 88

651

DISULFIRAM TREATMENT OF ALCOHOLISM / WRIGHT AND MOORE TABLE IV Contraindications to Disulfiram Use* Absolute contraindication

Prior psychiatric, neurologic, hepatic, or cardiovascular toxic reaction to disulfiram

Strong relativecontraindications Ischemic heart disease Cardiomyopathy with failure Chronic cardiac arrhythmias Severe pulmonary insufficiency Hepatic insufficiency Chronic renal failure Severeorganic mental disorder Severe peripheral neuropathy Psychosis Severe personality disorders Diabetes mellitus Pituitary and adrenal disorders Pregnancy Anti-convulsant and anti-coagulant therapy

From [103].

condition for remaining in the program) supervised disulfiram in a case series of 68 patients in an alcoholism program who were returning for visits but had repeatedly been unable to remain abstinent for longer than a few weeks at a time. Over 6 months, 58% of these patients remained abstinent. In contrast, Gallant and co-workers [95] were unable to replicate these results in so-called "skid row" alcoholic judicial offenders, with only a 10% rate of improvement in drinking behavior. These judicial and other mandatory use studies indicate that there is an effect of disulfiram in about half of such cases. Although the experimental design in each of these reports was a case series using each subject as his or her own control, the habitual alcoholic behavior at baseline made a substantial reduction in arrests, hospitalizations, and parole violations appear clinically important. However, the effect of disulfiram is confounded by concomitant social and legal sanctions and cannot be generalized to other less coercive treatment settings. Supervised voluntary use of disulfiram is a less restrictive alternative to compulsory use. Azrin [96] studied such a use of disulfiram as part of a community-based comprehensive program of alcoholism services in 1976. In an intensive experimental study of 20 subjects, he linked disulfiram use to daily habits and adverse life events, monitored compliance by having a spouse or family member observe ingestion, and had the client take disulfiram at each appointment. He was able to show good program efficacy over non-randomized controls who did not take disulfiram in regard to drinking day, employment, absences from home, and need for institutionalization, but disulfiram was studied as a part of a complex mixture of interventions and the isolated effect of the drug was not determined. Robichaud et al [97] studied disulfiram in a cross-over study of a workplace alcoholism program and found a reversible fivefold reduction in absenteeism during the supervised administration of disulfiram to 21 alcholic hospital employees. The behavioral aspects of supervised disulfiram administration were studied in a randomized trial by Bigelow et al [98,99], who showed, in a randomized study of disulfiram compliance, that contingency contracting for a small monetary reward or for other reinforcers such as methadone resulted in a significant decrease in days drinking and a significantly decreased arrest rate compared with control subjects. Gerrein et al [100] studied a group of 49 alcoholics 652

June 1990 The American Journal of Medicine Volume 88

who volunteered for disulfiram use in a supervised setting. In this randomized trial of disulfiram acceptors, one group was given disulfiram under observation three times a week, a second group prescribed the drug but administration was not supervised, and a third group was a no-drug control. Supervised disulfiram was statistically more effective than any other condition in promoting compliance with treatment and abstinence throughout the 8-week program, and there was no difference between the unsupervised administration of disulfiram and no-drug controls. Azrin et al [101] extended their earlier work by studying the effect of three treatment conditions on drinking, employment, and health care utilization. In this randomized study, a traditional educational program with voluntary disulfiram use was contrasted with supervised disulfiram use or a combination of supervised disulfiram use and comprehensive behavioral therapy. The authors found that the amount of disulfiram actually taken was proportional to the amount of concurrent behavioral therapy, and that drinking, absenteeism, and illness were significantly associated not only with disulfiram but also with the amount of behavioral theraPy received. These studies of non-judicial supervised disulfiram use suggest that this type of use does result in shortterm benefits. It should be noted, however, that the volunteers for these studies are likely to have been particularly well-motivated and this may limit the generalizability to similarly well-motivated patients. In support of this, it has been shown that disulfiram acceptors are also more likely to accept and comply with other t r e a t m e n t recommendations compared with disulfiram refusers [102]. COMMENTS

Disulfiram is likely to be the only alcohol aversive approved for use in the United States for at least the next decade. It clearly has no place in the intoxicated individual, in the patient undergoing alcohol withdrawal, alcohol-induced delirium, or dementia, or in the actively psychotic or suicidally depressed patient. The role of the drug is limited to individuals who are free of alcohol at the time of initiation of therapy and either have a desire (or in the case of judicial programs, an externally imposed need) to abstain from or control their consumption of alcohol. Disulfiram can produce an aversive reaction in both alcoholic and non-alcoholic volunteers when taken at sufficient dosage. Clinical trials as well as pharmacologic experiments have shown that the drug may have a narrow therapeutic window. Doses less than 250 mg have failed to produce aversive reactions in a significant number of individuals, whereas doses in excess of 250 mg may increase the risk of several forms of doserelated toxic side effects, with hypertension, neuropathy, hepatitis, and psychosis becoming unacceptably frequent at doses above 500 mg/70 kg/day. Clinicians should be aware that in some patients the dose that will produce toxicity is lower than the dose that will produce an aversive reaction. Disulfiram has produced pharmacologic effects when given every other day in judicial programs, but has had to be given in doses of at least 500 mg/70 kg every other day to do so. There is evidence from placebo-controlled trials of disulfiram implants that the only effective route of administration of the drug cur-

DISULFIRAM TREATMENT OF ALCOHOLISM / WRIGHT AND MOORE TABLE V Recommended Monitoring*

Initially Initial medical history and examination Interim medical history and examination Evaluation of drug and alcohol use Establish or review the goals of treatment Psychosocial assessment Evaluation of treatment compliance Comprehensive diagnostic evaluation as directed by history and physical examination, including hepatic, renal, cardiovascular, and endocrine evaluation Screening for hepatocellular injury Periodic laboratory or diagnostic evaluation as indicated by the patient's medical condition and directed by history or physical findings *

Monthly

Every 3 Months

Every 6 Months

X X X X X

X X X X X

X X X X X

X X X X X

First month

X X

From [i03].

rently available is the oral route. The drug is known to inhibit the metabolism of other drugs to an extent sufficient to cause toxicity, and there have been reports of severe adverse reactions when it has been given in combination with other alcohol-aversive drugs. The contraindications to disulfiram therapy are listed in Table IV. The only absolute contraindications to the use of disulfiram are previous psychiatric, neurologic, hepatic, or cardiovascular toxic reactions to the drug. It may be expected that ethanol-disulfiram reactions will frequently produce significant hypotension and probable myocardial ischemia, and such reactions should never be a planned part of therapy with the drug. Prescription of this agent obligates the prescribing physician to take special care in informing the patient not only of the hazard of an ethanol-disulfiram reaction, but also of the need for continued medical supervision while taking the drug, suggestions for which are outlined in Table V. Patients who have demonstrated hypertensive reactions, elevation of serum enzymes indicating hepatocellular injury, or altered metabolism of essential drugs probably should discontinue the use of disulfiram promptly. The drug should not be given to a patient who is known to be pregnant, a patient who has any condition that would make a hypotensive episode unusually hazardous, or a patient who cannot comprehend the risk of drinking while taking the drug. There is evidence from two placebo-controlled randomized clinical trials by Fuller et al [72,73] that simply prescribing oral disulfiram to alcoholic patients in outpatient settings may be expected to reduce the total number of days the patient drinks but should not be expected to sustain abstinence or ameliorate the behavioral aspects of the disorder. Supervised use of disulfiram, whereby the patient's ingestion of the drug is monitored by someone other than the patient, does appear to be associated with an improved outcome. There is evidence that even small incentives increase patient compliance in taking the medication. Studies of disulfiram programs as an alternative to incarceration have been shown to result in improvement in about half of the patients, but the

confounding due to social and legal sanctions in such cases makes it difficult to determine the individual contribution of disulfiram to these outcomes. Supervised administration of a drug is only one of a large number of compliance monitoring and enhancement techniques and is ethically repugnant to many physicians. There is a current need for well-designed clinical trials of alternative methods of improving compliance in disulfiram studies, and an absolute need to measure compliance as a part of any future studies of the drug. Disulfiram has been in clinical use in America for 40 years, and its efficacy in the treatment of alcoholism has been a subject of controversy for most of that time. We have reviewed the current evidence and find that disulfiram is effective in the short-term reduction in alcohol consumption but only when it is used in conjunction with a comprehensive treatment program and the compliance of the patient in taking the drug is monitored. There is no evidence from the available literature that disulfiram alters the long-term outcome of alcoholism, that it is effective in the absence of concomitant addictions treatment, or that unmonitored self-administration of the drug has any beneficial effect. REFERENCES 1. Jones RW, Helrich AR: Treatment of alcoholism by physicians in private practice, Q J Stud Alcohol 1973; 33: 117-131. 2. Sjolin-Fosberg G, Sandberg M, Osterlund L, Bergmann U: Patterns of drug utilization in the treatment of alcoholism. In: Workshop on pharmacological treatment of alcoholism, National Board of Health and Welfare. Stockholm: Drug Information Committee of Sweden, 1985; 99-112. 3. Nathan PE: Aversion therapy in the treatment of alcoholism: success and failure. Ann NY Acad Sci 1985; 443: 357-364. 4. Jackson TR, Smith JW: A comparison of two aversion treatment methods for alcoholism. J Stud Alcohol 1978; 39: 187-191. 5. Hedberg AG, Campbell L: A comparison of four behavioral treatments of alcoholism. J Behav Ther Exp Psychiatry 1974; 5: 251-256. 6. Neuberger OW, Matarazzo JD, Schmitz RE, Pratt HH: One year follow-up of total abstinence in chronic alcoholic patients following emetine counterconditioning. Alcoholism 1980; 4: 306-312. 7. Williams EE: The effect of alcohol on workers with carbon disulfide. JAMA 1937; 109: 1472-1473. 8. Ferguson JKW: A new drug for the treatment of alcoholism. Can Med Assoc J 1956; 74: 793-799. 9. Taylor JA: Metronidazole: a new agent for combined somatic and psychic therapy of alcoholism. Bull Los Angeles Neurol Soc 1964; 29: 158-162.

June 1990 The American Journal of Medicine Volume 88

653

DISULFIRAM TREATMENT OF ALCOHOLISM / WRIGHT AND MOORE 10. Smith JA, Wolford JA, Weber M, McLean D: Use of citrated calcium carbamide (Temposil) in treatment of chronic alcoholism, lAMA 1957; 165: 2181-2184. 11. Glatt MM: Disulfiram and citrated carbamide in the treatment of alcoholism. J Ment Sci 1959; 105: 476-481. 12. MarconiJ, Solari G, GaeteS, PiazzaL: Comparativeclinical study of the effects of disulfiram and calcium carbamide. Q J Stud Alcohol 1960; 21: 642-654. 13. Marconi J, Solari G, Gaete S: Comparative clinical study of the effects of disulfiram and calcium carbamide. Q J Stud Alcohol 1961; 22: 46-51. 14. Levy MS, Livingstone BL, Collins DM: A clinical comparison of disulfiram and calcium carbamide. Am J Psychiatry 1967; 123: 1018-1022. 15. Lader MH: Alcohol reactions after single and multiple doses of calcium carbamide. Q J Stud Alcohol 1967; 28: 468-475. 16. Mellor CS, Sims P: Citrated calcium carbamide/alcohol reaction: its severity and effectiveness as a deterrent. Br J Addict 1971; 66: 123-128. 17. PeachyJE, Annis H: New strategies for using the alcohol-sensitizingdrugs. In: Naranjo CA, Sellars EM, eds. Research advances in new psychopharmacological treatments for alcoholism. New York: Elsevier Science Publishers, 1985; 199216. 18. Linton PH, I/lain JD: Metronidazole in the treatment of alcoholism. Q J Stud Alcohol 1966; 28: 544-547. 19. Goodwin DW: Metronidazolein the treatment of alcoholism: a negative report. Am J Psychiatry 1967; 123: 1276-1278. 20. Egan WP, Goetz R: Effect of metronidazole on drinking alcoholics. Q J Stud Alcohol 1968; 29: 899-902. 21. Gallant DM, Bishop MP, Tisdale C: A six-month controlled evaluationof metronidazole in chronic alcoholics. Curr Ther Res 1968; 10: 82-87. 22. Merry J, Whitehead A: Metronidazole and alcoholism. Br J Psychiatry 1968; 114: 859-861. 23. Swinson RP: Long-term trial of metronidazolein male alcoholics. Br J Psychiatry 1971; 119: 85-89. 24. Rothstein E, Clancy DD: Toxicity of disulfiram combined with metronidazole. N Eng J Med 1969; 280: 1006-1007. 25. Faiman MD: Biochemical pharmacology of disulfiram. In: MajcrowiczE, Nobel P, eds. Biochemistry and pharmacologyof ethanol, volume 2. New York: Plenum Press, 1979; 325-347. 26. Detrich RA, Hellerman L: Diphosphopyridine nucleotide linked aldehyde dehydrogenase inhibitors. J Biol Chem 1963; 238: 1683-1689. 27. Glud E: The treatment of alcoholicpatients in Denmarkwith Antabuse. Q J Stud Alcohol 1949; 10: 185-197. 28. Jacobsen E, Martinsen-Larsen O: Treatment of alcoholism with tetraethylthiuram disulfide, lAMA 1949; 139: 918-922. 29.Gelbman F, Epstein NB: Initial clinical experience with Antabuse. Can Med Assoc J 1949; 60: 549-552. 30. Barrera SE, Osinski WA, DavidoffE: The use of Antabuse in chronic alcoholics. Am J Psychiatry 1950; 107: 8-13. 31. Dale PW, Ebaugh FG: Antabuse therapy in chronic alcoholism. Am J Med Sci 1950; 107: 774-780. 32. Child GP, Osinski W, Bennet RE, Davidoff E: Therapeutic results and clinical manifestations following the use of tetraethylthiuram disulfide. Am J Psychiatry 1951; 107: 774-780. 33. Bowman KM, Simon A, Hine CH, et al: A clinical evaluationof tetraethylthiuramidsulfide in the treatment of problem drinkers. Am J Psychiatry 1951; 107: 832838. 34. Winship GM: Antabusetreatment. In: Wallenstein RS, ed. Hospitaltreatment of alcoholism. New York: Basic Books, 1957; 23-51. 35. Hine CH, Anderson HH, Macklin EA, etal: Some observationson the effects of small dosesof alcohol in patients receivingAntabuse.J PharmacolExpTher 1950; 98: 13. 36. Becker MC, Sugarman G: Death following test drink of alcohol in patients receiving Antabuse. lAMA 1952; 149: 568-569. 37. Jacobsen E: Deaths of alcoholic patients treated with disulfiram. Q J Stud Ncohol 1952; 13: 16-26. 38. Markham JD, Hoff EC: Toxic manifestationsof the An]abusealcohol reaction. lAMA 1953; 152: 1597-1601. 39. Sauter AM, Boss D, Von Wartburg JP: Reevaluationof the disulfiram alcohol reaction in man. J Stud Alcohol 1977; 38: 1680-1695. 40. Peachy JE, Zlim DH, Robinson GM, Jacob M, Cappell H: A placebo controlled double blind comparative clinical study of the disulfiram and calcium carbamide acetaldehyde mediated ethanol reactions in social drinkers. Alcoholism (NY) 1983; 7: 180-187. 41, Brewer C: How effective is the standard dose of disulfiram? Br J Psychiatry 1984; 144: 200-202. 42, Christensen JK, Ronsted P, Vaag UH: Side effects after disulfiram. Acta Psychiatr Scand 1984; 69: 265-273. 43. Ciraulo DA, Barnhill J, Boxenbaum H: Pharmacokineticinteraction of calcium carbamide and antidepressants. Am J Psychiatry 1985; 142: 1373-1374. 44, Kioboe E: Phenytoin intoxication during treatment with Antabuse. Epilepsia 1966; 7: 246-249. 45, O'Reilly RA: Interaction of sodium warfarin and disulfiram in man. Ann Intern Med 1973; 78: 73-76. 46, Vesall ES, PassanatiniGT, Lee CH: Impairment of drug metabolism by disulfiram in man. Clin Pharmacol Ther 1971; 12: 785-792. 47, Vilicer L, Nelson K: Developmentof reversible hypertension during disulfiram

654

June 1990

The American Journal of Medicine

Volume 88

therapy. Arch Intern Med 1984; 144: 1294-1297. 48. Borg OB, Holmstedt B, Kvand H, Skroder R: Concentration of serotonin in CSF from alcoholics before and during disulfiram therapy. Acta Pharmacol Toxicol (Copenh) 1980; 47: 305-307. 49. Lake CR, Major LF, Zeigler MG, Kopin IJ: Increasedsympathetic nervous system activity in alcoholic patients treated with disulfiram. Am J Psychiatry 1977; 134: 1411-1414. 50. Wolff K: Nachteilige nebenwirkungen bei behandlung nit dem alkohol-vergallungsmittel antabus. Schweitz Med Wochenschr 1950; 80:1151-1152. 51. Martinsen-Larsen O: Psychotic phenomena provoked by tetraethylthiuram disulfide. Q J Stud Alcohol 1951; 12: 206-216. 52. KneeST, RazaniJ: Acute organic brain syndrome: a complication of disulfiram therapy. Am J Psychiatry 1974; 131: 1281-1282. 53. Liddon SC, Satran R: Disulfiram (Antabuse) psychosis. Am J Psychiatry 1967; 123: 1284-1289. 54. Hotson JR, Langston JW: Disulfiram induced encephalopathy. Arch Neurol 976; 33: 141-142. 55. Guild J, Epstein NB: Psychosis during the treatment of alcoholism with tetraethylthiuram disulfide. Q J Stud Alcohol 1951; 12: 360-365. 56. RaineyJM: Disulfiram toxicity and carbon disulfide poisoning. Am J Psychiatry 1977; 134: 371-377. 57. Bradley WG, Hewer RL: Peripheral neuropathy due to disulfiram. Br Med J 1966; 2: 449-450. 58. Keefe EB, Smith FW: Disulfiram hypersensitivity hepatitis, lAMA 1974; 230: 435-438. 59. Ranek L, Andreasen PB: Disulfiram hepatotoxicity. Br Med J 1977; 2: 94-96. 60. Eisen JH, Ginsberg AL: Disulfiram hepatotoxicity. Ann Intern Med 1981; 83: 673-674. 61. Bartle WR, Fisher MM, Kerenyi NO: Disulfiram induced hepatitis. Dig Dis Sci 1985; 30: 834--837. 62. VazquezJJ, Pardo-MindanJ: Liver cell injury in alcoholics treated with disulfiram. Histopathology 1979; 3: 377-384. 63. Iber FL, Kelvin L, Lacoursiere R, Fuller R: Liver toxicity encountered in the Veteran's Administration trial of disulfiram in alcoholics. Alcoholism 1987; 11: 301-305. 64. Hoff EC: The use of disulfiram. Conn Med 1955; 19: 793-799. 65.Armor DJ, Polich J, Stambul HB: Specific therapies: In: Armor DJ, Polich J, Stambul HB, eds. Alcoholism and treatment. New York: Wiley, 1978: 146-147. 66. Baekeland F, Lundwall L, Kissin B, Shanahan T: Correlates of outcome in disulfiram treatment of alcoholism. J Nerv Ment Dis 1971; 153: 1-9. 67. Bromet E, Moos R, Bliss F, Wuthmann C: Post-treatment functioning of alcoholic patients. J Consult Clin Psychol 1977; 45: 829-842. 68. Schuckit MA: A one-year follow-up of male alcoholics given disulfiram. J Stud Alcohol 1985; 46: 191-195. 69. Obitz FW: Control orientation and disulfiram. J Stud Alcohol 1978; 39: 12971298. 70. O'Neil PM, Roitzch JC, Giancinto JP, Miller WC: Disulfiram acceptors and refusers: do they differ? Addict Behav 1982; 7: 207-209. 71. Rush BR, MallaA: A comparison of disulfiram acceptors and refuserson selected demographic characteristics. Drug Alcohol Depend 1984; 14: 75-85. 72. Fuller RK, Roth HP: Disulfiram for the treatment of alcoholism: an evaluationof 128 men. Ann Intern Med 1979; 90: 901-904. 73. Fuller RK, Branchley L, Brightwell DR, et ak Disulfiram in the treatment of alcoholism: a VeteransAdministration CooperativeStudy. lAMA 1986; 256: 14491455. 74, Jacobsen E, Martinsen-Larsen O: Treatment of alcoholism with tetraethylthiuram disulfide, lAMA 1949; 139: 918-922. 75. Paulson SM, Krause S, Iber FL: Development and evaluation of a compliance test for patients taking disulfiram. Johns Hopkins Med J 1977; 141: 119-125. 76. Gordis E, Peterson K: Disulfiram therapy in alcoholism: patient compliance studied with a urine detection procedure. Alcoholism 1977; 1: 213-216. 77. Fuller RK, NeiderhiserDH: Evaluationand applicationof a urinary diethylamine method to measurecompliance with disulfiram therapy. J Stud Alcohol 1981; 42: 202-207. 78. Fuller R, Roth H, LongS: Compliancewith disulfiram treatment of alcoholism. J Chronic Dis 1983; 36: 161-170. 79. FawcettJ, Clark CA, AagesenCA: A double blind placebocontrol trial of lithium carbonate for alcoholism. Arch Gen Psychiatry 1986; 4: 248-258. 80. Kellan AMP, Wesolkowski JM: Disulfiram implantation for alcoholism. Lancet 1968; 1: 925-926. 81. Malcolm MT, MaddenJS: The use of disulfiram implantation in alcoholism. Br J Psychiatry 1973; 123: 41-45. 82. Whyte CR, O'Brian PMJ: Disulfiram implant: a controlled trial. Br J Psychiatry 1974; 124: 42-44. 83. ObholzerAM: A follow-up study of nineteen alcoholic patients treated by means of tetraethylthiuram disulfide implants. Br J Addict 1974; 69: 19-23. 84. Malcolm MT, MaddenJS, WilliamsAE: Disulfiram implantation critically evaluated. Br J Psychiatry 1974; 125: 485-489. 85. Lewis MJ, Bland RC, Baile WB: Disulfiram implantation for alcoholism. Can Psychiatry Assoc J 1975; 20: 283-286. 66. Wilson A, Davidson WJ, Blanchard R, White J: Disulfiram implantation: a placebo controlled trial with two year follow-up. J Stud Alcohol 1978; 39: 809819.

DISULFIRAM TREATMENT OF ALCOHOLISM / WRIGHT AND MOORE 87. Wilson A, Davidson WJ, Blanchard R, White J: Disulfiram implantation: a trial using placebo implants and two types of controls. J Stud Alcohol 1980; 41: 429436. 88. Morland J, JohnsonJ, Bache-WiligJE, RipelA, StensrudT, StowellA: Implanted disulfiram: pharmacologicand clinical studies in man. In: Workshop on pharmacological treatment of alcoholism, National Board of Health and Welfare. Stockholm: Drug Information Committee of Sweden, 1985; 89-95. 89. Borg S, Halldin J, Kyhlhorn Mannerfelt M, StrandbegK: Implantation of disulfiram: resultsfrom a placebocontrolled multicenter study. In: Workshop on pharmacological treatment of alcoholism, National Board of Health and Welfare. Stockholm: Drug Information Committee of Sweden, 1985; 65-88. 90. Johnsen J, Stowell A, Bache-Wilig JE, et ak A double-blind placebo controlled study of male alcoholicsgiven a subcutaneousimplantation. Br J Addict 1987; 82: 607-613. 91. Bourne PG, AIford JA, Bowcock JZ: Treatment of skid row alcoholics with disulfiram. Q J Stud Alcohol 1963; 27: 42-48. 92. HaynesSN: Contingency managementin a municipally administeredAntabuse program for alcoholics. J Behav Ther Exp Psychiatry 1973; 4: 31-32. 93. Brewer C, Smith J: Probation linked supervised disulfiram in the treatment of habitual drunken offenders. Br Med J 1983; 287: 1282-1283. 94. SerenyG, SharmaV, Holt J, GordisE: Mandatory supervisedAntabusetherapy in an outpatient alcoholism program: a pilot study. Alcoholism: clinical and experi-

mental research 1986; 10: 290-292.

95. Gallant DM, Faulkner MA, Simpson L, et ak A comparative evaluation of compulsory and voluntary treatment of the chronic alcoholic municipal court offender. Psychosomatics 1968; 9: 306-310. 96. Azrin NH: Improvements in the community reinforcement approachto alcoholism. Behav Res Ther 1976; 14: 339-348. 97. Robichaud C, Strickler D, Bigelow G, Liebson I: Disulfiram maintenance employee alcoholism treatment. Behav Res Ther 1979; 17: 618-621. 98. Bigelow G, Strickler D, Liebson I, Griffiths R: Maintaining disulfiram ingestion among outpatient alcoholics: a security deposit contingency contracting procedure. Behav Res Ther 1976; 14: 378-379. 99. Liebson IA, Tommasello A, Bigelow G: A behavioral treatment of alcoholic methadone patients. Ann Intern Med 1978; 89: 342-344. 100. GerreinJR, RosenbergCM, ManoharV: Disulfiram maintenancein outpatient treatment of alcoholism. Arch Gen Psychiatry I973; 28: 798-802. 101. Azrin NH, Sisson RW, Meyers R, Godley M: Alcoholism treatment by disulfiram and community reinforcement therapy. J BehavTher Exp 1982; 13:105-112. 102. Rush BR, Malla A: A comparison of disulfiram accepters and refusers on selected demographic and clinical characteristics. Alcohol Drug Dependence 1984; 14: 75-85. 103. SellarsEM, NaranjoCA, PeacheyJE: Drugsto decreasealcohol consumption. N Engl J Med I981; 305: 1255-1261.

June 1990 The American Journal of Medicine

Volume 88

655