British Journal of Psychiatry (1992), 161, 84—89

Disulfiram Treatment of Alcoholism JONATHAN CHICK, KEVIN GOUGH, WOJCIECH FALKOWSKI, PETER KERSHAW, BRIAN HORE, BRIJ MEHTA, BRUCE RITSON, RICHARD ROPNER and DENIS TORLEY

To assessthe efficacy of superviseddisulfiramas an adjunct to out-patient treatment of alcoholics,a randomised,partiallyblind,six-monthfollow-up study was conductedin which 126 patients received 200 mg disulfiram or 100 mg vitamin C under the supervision of a nominated informant. In the opinion of the (blinded) independent assessor, patients on

disulfiramincreasedaveragetotal abstinentdays by 100 and patientson vitamin C by 69, thus enhancing by one-third this measure of treatment outcome. Mean weekly alcohol consumption was reduced by 162 units with disulfiram, compared with 105 units with vitamin

C. and the disulfirampatients reducedtheir total six-monthalcoholconsumptionby 2572 units compared with an average reduction of 1448 units in the vitamin C group. Serum gamma-GTshoweda meanfall of 21 lU/I in patientsondisulfirambut roseby a meanof 13 lU/l with vitamin C. Unwanted effects in the disulfiram group led to a dose reduction in seven patients and to treatment withdrawal in four (and in one vitamin C patient). Two-thirds of

the disulfiramgroup asked to continuethe treatment at the end of the study. There were no medicallyseriousadversereactions.

Disulfiram (Antabuse) is an agent which inhibits metabolism of alcohol, resulting in the unpleasant symptoms

(flushing,

headache,

nausea,

dizziness,

tachycardia) of the disulfiram—alcoholreaction. Although it has been available for many years as an adjunct to counselling in the treatment of chronic alcoholism, and despite a dearth of therapies for this

condition (Vaillant, 1983),disuffiram is not commonly prescribed in the UK owing partly to concern that the agent may cause hepatic damage (Peachey & Naranjo,

1983; Peachey, 1988). In addition, the early

literature provided poor evidence of the efficacy of disulfiram, but these studies often lacked adequate controls (Bourne eta!, 1966; Edwards & Dill, 1974;

Bigelow et a!, 1976) or supervision of patient compliance (Fuller & Roth, 1979). Compliance with the disulfiram regime, found to be as low as 20%

already relapsed after previous

therapy

or other support

wereinvitedto participate,since we felt the memoryof previousfailurewouldaidtheircompliancewiththestudy treatment. Pregnant women were excluded, as were subjects with cardiac disease, psychosis, or habitual drug abuse,

and those showing abnormally high levels of serum bilirubin, aspartate aminotransferase (AST) or alanine aminotransferase (ALT).

The protocol was approvedby local hospital ethical reviewcommittees.Allpatientsgavetheirwritten,informed consent to receive one tablet a day of disulfiram (200 mg, dispersed in water) or vitamin C (100 mg) for six months

under supervision,and an informant was nominated, usually the spouse (occasionally another relative, colleague,

ora memberof theclinicstaff,withwhomtheyhadcontact at least once a week). Treatments were randomly placed against lists of numbers, supplied to the pharmacist at the various centres, who then allocated the numbers sequentially to patients entering the trial.

in a study in the USA (Fuller et a!, 1986), can be improved with supervision by the spouse or clinic (Oerrein et a!, 1972; Robichaud et a!, 1979; Azrin

allocation

et a!, 1982;Serenyet a!, 1986). We report here the first UK controlled study of supervised disulfiram as an adjunct to out

given to the patients and families concerned.

patient treatment of alcoholics, in which safety and acceptability were assessed in addition to the effect

of the treatment on alcohol consumption and related problems.

Forethicalreasonsthetreatmentcodeswerebrokenafter so that a thorough

explanation

of the use of

disulfiram and the associated risks of drinking, to include written information and a pocket warning card, could be If left blind,

patients might have been tempted to test whether or not drinking

could trigger a reaction.

The vitaminC group was includedto control for the effects

of receiving

supervised

medication

and out-patient

counselling,andpatientsweretoldthis;if theyaskedfurther theyweretold that vitaminC waschosenfor the control medication because alcoholics may have vitamin deficiencies, of which this is one.

Method One hundredand twenty-sixsubjectsenteredthe trial from among patients of either sex, aged 18—67 years, attending

sevenalcoholismtreatmentcentres.Onlypatientswhohad

Medicationwasusuallysuperviseddailyby the informant; wheretheinformantwasnot thespousethedoseon a day whenthe informantand patientdid not meetwas either giventhedaybeforeorgivento takeunsupervised athome. The informantwas encouragedto telephonethe clinicif

84

85

DISULFIRAM TREATMENT OF ALCOHOLISM the patient refused the medication or lost touch, so that advice could be offered. No written contract, however, was involved, and no sanctions were invoked if the patient ceased taking the medication. Patients were either already in, or were offered, a range of out-patient and community counselling and support, which varied between centres. A few patients were offered day-patient places. Marital therapy, relaxation therapy, attendance at Alcoholics Anonymous (AA), vitamin B supplements, and supportivegroup therapywere also used by some patients. At intake, the clinicanconducted a physicalexamination of the patient, which included blood tests, and took a

medical and psychiatric history. During treatment the clinicianmonitoredcompliance(checkingwiththe informant) and drug safety at each visit, recording any unusual symptoms reported by the patient, and reviewed patient progress at the end of the six-month trial. Blood tests (haematology and biochemistry, including liver function tests and blood

alcohol,

plus serum

gamma-glutamyl

transferase ((IT) and mean red cell volume (MCV) as markers of regular alcohol consumption

(Chick eta!, 1981))

wererepeatedafter one, threeand six months of treatment. Each centre appointed as an independent assessor to obtain follow-up data someone with previous experience with alcoholics: medical practitioners, nurses, or trained research interviewers. They were to stay blind to the medication received. Patients and informants were reminded

at each contact not to give any information which could reveal the medication. The assessor saw patient and informant at intake and again, separately,at weeks 2 and 4, and thereaftermonthly until the final interviewat six months. Interviewquestions concerned alcohol consumption, alcohol dependence (the Severity of Alcohol Dependence Questionnaire (SADQ) (Stockwell eta!, 1983)) and 13 alcohol-related health and social problems (Chick et a!, 1988), all with reference to the period since the last visit (or the previous six months for the first and final visits). ‘¿Typical week's consumption' was according to a retro spective diary of a typical week during which the patient drank. In addition, at each interviewthe assessor obtained an estimateof total consumptionin the previousfour weeks. These were aggregated in the analysis to give ‘¿total units consumed in past six months'. At intake the assessor had obtainedan estimateof the priorsix months' consumption, with anchor dates to aid memory. Informants'information was included, and was evaluated by the assessor in making a judgement, because patients are known to report less consumption than informants (Fuller et a!, 1988). Statistical

analysis

All data were used, on an ‘¿intention-to-treat' basis, irrespective of patient compliance, attempts having been madeto follow up all patients.Categoricaldatawereanalysed using Fisher's exact test for 2 x 2 tables and Pearson's x2 test for larger tables. Otherwise, the Mack-Skilling's test (Mack & Skillings, 1980), taking account of the weighting at different centres, was used to test for significant treatment differences. Laboratory blood data were analysed by fitting

anadditivelinearmodelof centreandtreatmenteffectsand using a (-test to compare treatments (Searle, 1971). Where

possible,

differences

from pre-treatment

were

analysed.All tests weretwo-tailedwith a significancelevel of 5%.

Results Thetwo groupsof patientscommencingtreatment(64 on disulfiram, 62 on vitamin C) had similar demographic and social backgrounds. The overall mean age was 43 years (range 18-67); 84% were male, 65% were unemployed, and 46% lived with a spouse or other cohabitee. The commonest

illnesssufferedwasgastrointestinaldisease(21%of patients),

of which85%wasalcohol-related. Two-thirds of thedisulfiram groupand halfthe vitaminC grouphad had in-patienttreat ment for alcoholism. Informantsweremainlyspouses(4l%) or members of the clinic staff (33%).

Fifty-sevenpatients(28on disulfiram,29 on vitaminC) did not adhere to their allocated treatment, 45 through failure to keep appointments or by withdrawing consent. Follow-upinterviewswerenot obtainedin 20% (15disulfiram patients, 14vitaminC patients).Both initialand final blood samples were available in only 57%, because at follow-up some patients were interviewedby telephone, and at intake and follow-up some samples were not analysable because of delay or damage. Four patients on disulfiram and one on vitamin C were withdrawnwith adversereactions:two of the formerowing to allergicskin rash,one withsuspectedneuropathy,and one with dizziness and nausea, while the patient on vitamin C was admitted suffering left hemiparesis. A further two withdrawals from the vitamin C group were due to increasing

problemswith drinking.(Four of the patientson vitaminC who withdrewtheir consent did so because they wanted to take disulfiram; in addition, three initial recruits had withdrawnas soon as they heardthey had been assigned to vitamin C and were thus excluded from the trial population.) Unusual

symptoms

were

reported

equally

in the two

groups (e.g. depression:4 disulfiram, 5 vitamin C patients; nausea:

2 disulfiram,

2 vitamin

C patients)

except for

headache (11 disulfiram, 5 vitamin C patients), fatigue (12 disulfiram, 6 vitamin C patients), and skin rash (4 disulfiram

patients only). Seven patients on disulfiram had their dose reducedbecauseof side.effects.Disulfiram-alcoholreactions were reportedon 29 occasions, but none led to a reduction of dose. Five patients had their disulfiram dose increased because the alcohol reaction was mild or absent, and one such patientrefusedto have the dose increased.There were no abnormalities of liver function during treatment. Treatment effects are summarised in Tables 1-4. Both treatment groups achieved a reduction in alcohol con sumption which by most estimates was greater with disulfiram, the treatment difference reaching statistical significance for values at 6 months (Table 1). However, at the final assessmentthe numberof days since the last drink and alcohol consumption in the last month of the study revealed no significant treatment difference (Table 3). The mean (s.d.) SADQ score at intake was similarin the

two groups, and fellequally(disulfiram:intake 31.6(13.8)

@ @ @

@q.@

E

86

.@

8

CHICKET AL

@ @

C-)

00 (0

0)00

00

00(0

IC)

10

lC)'@

10

‘¿*

6R 10(0

@

E.@

@

>

‘¿@

(O@

COO)

CD(S)

LOIS) ON

C').-

0010

+

00

@00

‘¿2 CD o

I

IC)

c'co

‘¿@‘* 00

I

100)

@0)

2

CD

0000

‘¿*

NO

N

0

N

+

I

0@ N

10N

cod,

000)

.@

0

ION

-5 C

@N

100

•¿â€” 0

C')O

I

100

N

+

•¿

N

N N C')N +

I

•¿

bC')

C

c

c

,@

@C')

ION

(DC')

00'CON

NIt)

+

@C')

N

I

‘¿0

@

E

0 .@-

C

o

10(0

—¿5

U) •¿

1010

@N

100)

CD

ION N(S)

m

E o

C')0)

‘¿-CD

(ON N(O

0)00

NN

U

+

‘¿2

NCD IO@ CD'lOS') C')N

.-N ‘¿-0

I

LOS')

@o C')C4)

2

NN

*IO ,-00 N.-

+

0

@C')

tom NIO N.-

‘¿

@

@

I

10.— 0)00 C')N

C')N

0

@

C')O)

1015)

I

.-N ON o@

2

@ @ @

C')C')

N

N N

@N N(O N.— I

I

(‘)O) C')N

+

I

I

.@

o @

c

C

ION

.—.@.

E.@

Nb

@ .@

@

.C

@ @

@

NN

ON

N

+

@

@

NO)

.-‘-

N@

•¿

2

0

o •¿ .2@ O

0

0

E

R@

10(0

5

C')N

100)

N

+

2

c@

.-N

NC')

0

00

(@

00

@R

‘¿*N

.-.-

0000

0

C')IO

+

2@

N

CV@O

ci

E

—¿

NIt)

@N

I

ION

NN

It)

OCO

(0'* o

N

lOIS) NN

C

o

‘¿*‘* I

‘¿*0)

o

1010

C') N

.-‘* C')N

NIt)

@L0

‘¿*00N0 R@

(010

0) (0

C')CD

N.-

oo0@ o@ N0

+

I

C

0

E@

.2 —¿

F

‘¿I-

•¿

.f—¿ ! @

@

o

.2

.2

C

.@

8

.@

@.

.c

a

110

a

•¿

.@ .n

a

•¿

-a

-a 2

.@

c

.@

.c

I@ !h! !II

J IIU1!IIiI1!!IiI1!!!?IJ!@Iif!1! :ii

DISULFIRAM Table

TREATMENT

Table4

2

Blood test markers: comparison of changes from intake after treatment

87

OF ALCOHOLISM

Final opinions of ability to control drinking

(means (s.d.))

MarkerDisulfiram (n)MCV:

(n) VitaminC

(fI)intake96.6 (52)change

OpinionDegree valuePatientworse of controlFrequency Disulfiram

Vitamin CP

(5.8)(57) 97.8 (5.8)

after treatment—3.0 (33)Treatment (5.1) (31) —¿2.6 (4.5) mean(95% difference: 2.1)P confidenceinterval)—0.4 (—2.8to value0.78Serum lU/lintake49 GT: (57)change (63) (55) 51(72) (38)Treatment after treatment—21 (65) (32) + 13 (83) mean(95% difference: —¿7)P (—74to confidenceinterval)—34 value0.02

nochange

3

20

moderate full0 13