Alcohol & Alcoholism Vol. 33, No. 2, pp. 103-115, 1998
REVIEW
THE DRUG TREATMENT OF ALCOHOL WITHDRAWAL SYMPTOMS:A SYSTEMATIC REVIEW
DAVID WILLIAMS and ANDREW J. McBRIDE*1
Ty Bryn. St Cadoc's Hospital, Caerleon, Gwent NP6 1XQ and 'Community Addiction Unit, Cardiff Community HealthcareNHS Trust, Whitchurch Hospital, Cardiff CF4 7XB, UK
(Received 1 April 1997; in revised form 9 September 1997; accepted 11 November 1997)
Abstract — A computer-assisted and cross-reference literature search identified trials of therapy foralcohol withdrawal symptoms. Those with a randomized, double-blind placebo-controlled design weresystematically assessed for quality of methodology. Fourteen studies were identified investigating 12different drugs. The quality of methodological design, even among this highly selected group ofpublished studies, was often poor. Study populations were generally under-defined, most studiesexcluded severely ill patients, control groups were poorly matched, and the use of additional medicationmay have confounded results in some studies. Twelve different rating scales were used to assess severityof symptoms. All 12 compounds investigated were reported to be superior to placebo, but this has onlybeen replicated for benzodiazepines and chlormethiazole. Further research using better methods isrequired to allow comparison of different drugs in the treatment of alcohol withdrawal symptoms. On theevidence available, a long-acting benzodiazepine should be the drug of first choice.
INTRODUCTION
Contemporary views on alcohol withdrawal symp-toms have evolved since the work of Victor andAdams (1953) and Isbell et al. (1955). Victor andAdams (1953) described a series of alcohol-dependent patients admitted to a specialist unitin the USA. They identified the now well-recognized spectrum of symptoms, including:tremor, nausea, anxiety, tinnitus, muscle cramps,diaphoresis, seizures, hallucinations and deliriumtremens, which comprise the alcohol withdrawalsyndrome. Isbell et al. (1955) gave large quantitiesof alcohol to recently detoxified drug-dependentpatients and described the symptoms they sufferedwhen the alcohol was abruptly stopped. Alcoholwithdrawal symptoms were later incorporated intothe construct of the 'alcohol dependence syn-drome' (Edwards et al., 1981) and continue to becentral to diagnostic classifications (e.g. WorldHealth Organization, 1992).
•Author to whom correspondence should be addressed.
The onset of the alcohol withdrawal syndromeis most commonly gradual, but may be abrupt.Classically the milder symptoms appear withinhours of cessation or reduction of alcoholconsumption and subside over 2-7 days. Seizuresmay appear singly or as a series, usually within24-48 h of alcohol withdrawal. Seizures areusually, but not always, preceded by other with-drawal symptoms (Saunders, 1987). Thirty percent of patients who suffer fits may go on todevelop delirium tremens (Victor, 1966). Deliriumtremens most characteristically occurs after48-72 h of alcohol withdrawal, is sometimesprecipitated by additional stresses such as infec-tion or head injury, and resolves over ~ 7 days.The proportion of drinkers affected by deliriumtremens has been estimated to be as high as 5% inuntreated withdrawal (Editorial, 1981) and may beexperienced at some time by 12% of alcoholics(Schuckit, 1995).
Although the symptoms of the alcohol with-drawal syndrome are now well described, theunderlying biochemical causes are still uncertain.It is not clear, for example, whether the with-
Decreased: B h B2, B6, B12, folic acidElectrolytes and glucose
Low: H + , Mg r+, Ca2+, Zn2+, K+, PO42~, HCO3",
glucose
Adapted from Glue and Nutt (1990) and Lieber (1991).
drawal syndrome represents the various manifes-tations of a single process, or the commonpathways of several abnormalities which com-monly occur together. This uncertainty has beenreflected in the treatments employed. It is nowthought that chronic alcohol use causes alterationsin several neurophysiological parameters to com-pensate for chronic central nervous systemdepression. Following cessation there is an over-correction of these changes leading to withdrawalsymptoms (Sellers and Kalant, 1976). Some of thephysiological changes which have been describedare shown in Table 1. The overall effect is that ofrebound neuronal and biochemical overactivityfollowing the prolonged depression of activitycaused by alcohol. Disturbed electrolyte andvitamin levels may reflect poor dietary intakeduring the period of alcohol abuse.
TREATMENT OF ALCOHOL WITHDRAWAL
The treatment of alcohol withdrawal symptomshas tended to follow the prevailing theory as totheir aetiology. Inevitably the great majority ofalcohol withdrawal episodes occur with nomedical supervision and the majority of medicallysupervised alcohol detoxifications occur in non-specialist settings, such as general medical andgeneral psychiatry wards or, increasingly, at home(Shaw, 1995). Most episodes of alcohol with-drawal do not require sedative medication. Thoseindividuals experiencing objectively observablewithdrawal symptoms, and those with a past
history of serious withdrawal symptoms, such asseizures or delirium tremens, are clearly in need ofdrug treatment, but criteria for predicting thosecases which will benefit from medication have notbeen clearly defined, even among inpatients(Shaw, 1995).
The merits of any drug treatment need to beconsidered in the light of evidence for theeffectiveness of non-pharmacological approaches.Whitfield et al. (1978) successfully treated 1024of 1114 (91.8%) alcoholic inpatients withoutdrugs. Staff, all college graduates, were 'givenbasic medical instruction' and trained to 'reassureand relate to disturbed alcoholics . . . restorefamiliar surroundings . . . be authoritative (toenable patients take control of their treatment)'within a specialist alcohol unit. Ninety patients(8.2%) were considered to be sufficiently'seriously ill' to require medication and weresent on to a hospital. There was one case ofdelirium tremens and 12 seizures among theremaining 1024 patients. The mean inpatient staywas 8 days. On average, non-pharmacologicaltreatment was required for only 2 days of the stay.The authors were unable to identify characteristicswhich distinguished the group requiring moreintensive treatment before the development ofseizures or delirium tremens. The major metho-dological weaknesses of the study were failure to:(1) use recognized diagnostic criteria; (2) definethe study group adequately; (3) use a rating scalefor severity of withdrawal symptoms; (4) define'seriously ill'. The study is noteworthy for the
Table 2. Clinical trial quality rating scale
(a) Degree to which randomization was truly blind(b) Inclusion of data from subjects who subsequently
withdrew from the study(c) Degree to which assessors were blind to the treatment
allocation(d) Whether subjects were assessed to determine if they
had accurately guessed their treatment status(e) Statement of criteria for improvement(f) Use of multiple informants for assessment of outcome(g) Method of determining dose of drug(h) Whether concurrent treatment was held constant(i) Length of baseline observation(j) Control for previous treatment(k) Control for co-morbidity
Adapted from Chalmers et al. (1981).
DRUG TREATMENT OF ALCOHOL WITHDRAWAL 105
number of patients studied and the radicalapproach taken. The results suggest that placebocontrol groups are required in any evaluation ofdrug treatment efficacy if high-risk individuals areexcluded.
The drug with the longest history of use andmost widely used by drinkers for alleviation ofwithdrawal symptoms is alcohol itself. It has beenestimated that > 150 different drugs have beenused to treat alcohol withdrawal symptoms duringthe past 30 years (Litten and Allen, 1991). Perhapsthe most significant achievement of modernmanagement of alcohol withdrawal has been todecrease the mortality from 15% in the 1960s(Victor, 1966) to <2% (Guthrie, 1989). However,factors other than advances in drug treatment, suchas improvement in the overall health and nutri-tional status of Western populations, are likely tohave been of equal or even greater significance.
This review of treatment research was con-ducted to update that carried out by Moscowitz etal. (1983), to determine whether or not morerecent clinical research had learned from themethodological mistakes of the past and to see ifclearer guidance could now be given on the mostappropriate treatment of the alcohol withdrawalsyndrome.
METHODS
A computer-assisted literature search (Medline)using key words (alcohol withdrawal, detoxifica-tion and treatment), plus manual cross-referencesearch of articles, review articles and contempor-ary text books, identified 51 trials of pharmaco-logical treatment for alcohol withdrawalsymptoms. Fourteen trials of double-blind pla-cebo-controlled design were identified and scoredfor quality of design using a modified version of ascale developed by Chalmers et al. (1981). Thescale comprises 11 aspects of study design andquality and is shown in Table 2. Each item isscored on a 4-point scale, 0 indicating that thepaper failed to mention anything to enable thatcriterion to be rated, a score of three indicating a'textbook' description of that criterion beingsatisfied. Only the source paper was scored, sothat a study may score 0 on some criteria, becauseit referred only to a previous description ofmethod with no further elaboration.
Twenty-two further comparison trials were
identified that included at least one of the drugsthat had been subjected to randomized, double-blind placebo-controlled studies. Although notanalysed in depth, they are presented in theResults section. The remaining 15 trials wereeither open trials or comparison studies in whichnone of the compounds had been subjected todouble-blind placebo-controlled studies. These arenot described.
RESULTS
Table 3 describes briefly the 14 trials that wereof a randomized, double-blind placebo-controlleddesign. Table 4 describes, more briefly, the 22randomized, double-blind controlled trials whichdid not include a placebo group but whichincluded one or more of the drugs which hadalso been investigated using a placebo-controlleddesign. The results show mainly efficacy data, asfew trials published safety data to allow detailedcomparison. The trials are grouped according tothe drug being studied. Overall quality scoresranged from 8/33 to 22/33 (mean = 13.8). Studieswith the highest quality designs generally had thesmallest sample sizes.
Benzodiazepines
Benzodiazepine drugs are used as anxiolytics,hypnotics, premedication for surgery, musclerelaxants and anticonvulsants. Benzodiazepinesact by potentiating GABA acivity and it is thoughtthat their action in the withdrawal period is relatedto this effect.
Almost as soon as the prototypes were launchedin the late 1950s and early 1960s, trials wereperformed to assess the efficacy of benzodi-azepines in the treatment of alcohol withdrawalsymptoms. One of the first large trials ofchlordiazepoxide was published in the BritishMedical Journal in 1965 (Sereny and Kalant,1965). On the basis of these early trials chlordia-zepoxide and diazepam have become the favouredtreatment in the USA and vie for the top positionwith chlormethiazole in Europe.
Diazepam was launched in 1961 and is the mostwidely used drug in Europe whereas chlordiazep-oxide has been considered the drug of choice foralcohol withdrawal in the USA. The majordifferences are the superior anticonvulsant effectof diazepam and the claimed greater safety of
o
Table 3. Summary of randomized, double-blind and placebo-controlled trials of drugs in the treatment of the alcohol withdrawal syndrome
Atenolol decreased craving and ANSsymptoms37% failure rate on AtenololPlacebo group more previous DTs andfits, longer Hx of use
64% of both groups successfulCBZ faster relief; 18 drop-outs in both;11 patients with side-effect.s in CPZgroup
r>on
n03
9ra
Seller* el al (1976) 19 Lilhium
Borg and Wcinholt (1982) 8 Bromocnpline
Gallimbem rial. (1989) 13 GHBA
Sampliner and Iber (1974) 15 Phenytoin
900 mg
7 5 mg
50 mg/kg
300 mg
60
23
157
Male, > 160 g ethanol EEG, tremometerdaily Motor tracking tabkNo liver disease, medical 34-nem scale + 17-Uemillness, contraindication to questionlithium or psychiatric Bloodsillness
Nil Lithium better than placebo insymptom relief subjectively
Male 24-71 years, IP,'gamma alcoholic'No exclusion specified
DSM-III WDNo drug abuse, fits, DT,medical illness, epileptictreatment
History of fitsConsecutive IP, 4 weekconstant intakeNo medical treatment
6-item 6-point scaleFSCL, global assessmentBlood prolactin
6-item 4-point scaleWord fluency
QuestionnairePresence of fitsBloods
BZD, CBZDixyrazme
Nil
CDP
Bromocripline relieved all symptoms8 drop-outsLower symptom score of bromoenptinegroup at start
GHBA good but high rale of side-effecK, especially dizziness, 7 h trial
No fits with phenytoinFits early with placebo11 in 11 patients
Key and abbreviations: =: equally effective; ANS = autonomic nervous system; Bloods: liver function tests, urea and electrolytes, and full blood count;BP = blood pressure; BZD = benzodiazepines; CBZ = carbamazepine; CDP = chlordiazepoxide; C1WA = Clinical Institute Withdrawal Assessment Scale;CMZ = chlormethiazole; CPZ = chlorpromazine; Dose: is the daily dosage of the drug unless otherwise specified; DT = delirium tremens; DZ = diazepam;EEG = electroencephalogram; FSCL = Fischer Symptom Check List; GHBA = gamma-hydroxybutyric acid; HI = head injury; HR = heart rate; Hx = pasthistory; IP = inpatient; NS = no severe withdrawal at onset included; OD = overdose; OP = outpatient; PP = propranolol; PZ = promazine; prn = as re-quired; SSA = selected severity scale; Sx = symptom; TSA = total severity assessment; VAS = visual analogue scale; WD = withdrawal.
108 D. WILLIAMS and A. J. McBRIDE
Table 4. Studies in which one or more of the drugs has been investigated using a placebo control group
Study Drugs Measurements Results
Chambers and Schulte (1965) DiazepamPromazine
Goldbert et al. (1967)
Kaim et al. (1969)
PromazineChlordiazepoxideParaldehydeAlcohol
ChlorpromazineChlordiazepoxideHydroxycineThiamine
4-point scale (not shown) Promazine = diazepam for DT andseizuresPromazine better for mild symptoms
Own doctor's assessment Paraldehyde most effective(now shown) Promazine least effective
BPRS5-point scale (not shown)Doctor's and nursingobservations
TSA, global ratings
7-point semantic scaleshown, Beck scale
Doctor's and patients'ratings (not shown)
Doctor's and patients'global rating (not shown)HARS
Baumgartner and Rowan (1987) ClonidineChlordiazepoxide
Madden et al. (1969)
Manhem et al. (1985)
Robinson et al. (1989)
Poutanen (1979)
Ritola and Malinen (1981)
ChlormethiazoleTrifluoperazine
ClonidineChlormethiazole
ClonidineChlormethiazole
Carbamazepine
ChlormethiazoleCarbamazepine
Own alcohol withdrawalscale (not shown)
Own 2-point scale (notshown)
Borg 4-point scale
Own scale (not shown)
Own scale (shown)
Own scale (shown)VAS
Chlordiazepoxide bestChlorpromazine worstMost drop-outs with thiamineSeizures worst with chlorpromazineMild symptoms all equally controlledChlorpromazine least effect on majorsymptoms
Chlormethiazole betterFewer drop-outs and fewer DT
Diazepam: quickerParaldehyde: more untoward eventsHaloperidol: better
Oral barbital better for frank DTcompared to i.m. diazepam
Lorazepam = diazepamOne adverse event with lorazepam11 drop-outs not evaluated
Alprazolam = chlordiazepoxideMore seizures in alprazolam group3 DT in each group
Alprazolam = chlormethiazole1 seizure with alprazolamChlormethiazole 10% drop-out groupnot evaluated
Alprazolam = diazepam
Clonidine more effective forautonomic symptoms and as effectivefor others
Chlormethiazole = trifluoperazineMore anxiety in chlormethiazole group
Clonidine = chlormethiazoleDT in both groups
Chlormethiazole more effectiveHigher drop-out rate for clonidine
Carbamazepine effective for allsymptoms. No seizures
Carbamazepine = chlormethiazole forsymptomsMore drop-outs with chlormethiazoleMore side-effects with carbamazepine
DRUG TREATMENT OF ALCOHOL WITHDRAWAL 109
Table 4. (continued)
Study Drugs Measurements Results
Agricola et al. (1982)
Flygenring et al. (1984)
Malcolm et al. (1989)
Stuppaeck et al. (1992)
Borg and Weinholt (1980)
CarbamazepineTiapride
CarbamazepineBarbital
OxazepamCarbamazepine
OxazepamCarbamazepine
ApomorphineBromocriptine
3-point scale (not shown)VAS
Own 5-point scale(shown)
CIWA, BDI, global scale,anxiety inventory
CIWA-A, global scale,self-rating scale
6-point scale (shown)
Carbamazepine betterQuicker for fear and hallucinationsCarbamazepine = barbital
Carbamazepine = oxazepam,including drop-out rate
Carbamazepine = oxazepam forsymptom relief days 1-5Carbamazepine better days 6-7
Bromocriptine better after day 5especially for tremorNo side-effects
Key and abbreviations: =: equally effective; BDI = Beck Depressive Inventory, BPRS = brief psychiatric rating scale;CIWA = Clinical Institute Withdrawal Assessment Scale; DT = delirium tremens; HARS = Hamilton Anxiety RatingScale; i.m. = intramuscular; TSA = total severity assessment; VAS = visual analogue scale.
chlordiazepoxide in overdose with alcohol(Serfaty and Masterton, 1993). Other differencesinclude a longer half-life of diazepam (14-70 hcompared to 4—29 h for chlordiazepoxide) andchlordiazepoxide causing less initial euphorialeading to a lower abuse potential. Other benzo-diazepines which have been used for alcoholdetoxification include oxazepam, lorazepam,alprazolam and flurazepam (Guthrie, 1989).
Chlordiazepoxide. There have been threedouble-blind, randomized placebo-controlledtrials. All scored highly on the quality scale, butsample sizes were small, ranging from six to 20.Burroughs et al. (1985) found chlordiazepoxide tobe equivalent in efficacy to chlormethiazole andsuperior to bromocriptine or placebo, for treatmentof mild and severe symptoms. In a study ofvolunteers, chlordiazepoxide was as effective aspromazine and more effective than placebo, inrelieving mild withdrawal symptoms, but was lesslikely to be associated with seizures (Sereny andKalant, 1965). Subjects developing delirium tre-mens prior to day 7 of the study were not includedin the results. Sellers et al. (1977) foundchlordiazepoxide to be inferior to propranololbut superior to placebo. The study groupscomprised only six patients.
Diazepam. One randomized, double-blind pla-cebo-controlled study has been made of diazepam.Sellers et al. (1983) treated 25 patients inmoderate to severe alcohol withdrawal with 20
mg diazepam repeated 2-hourly until symptomsabated. All patients with a previous history ofseizures were given additional anticonvulsants.Ninety per cent of the sample responded withinnine doses of diazepam and required no furthermedication, the other 10% required 3-5 days oftreatment. The apparent success of this short-termapproach was thought to be due to the very longhalf-lives of diazepam and its metabolites. Theplacebo response rate was 56%.
Chlormethiazole
Chlormethiazole acts as a central nervoussystem depressant, causing sedation and sleep(Lundqvist, 1966). Chlormethiazole was first usedin the treatment of delirium tremens during 1957and introduced for milder symptoms in 1965(Glatt et al., 1965). Chlormethiazole is now awidespread treatment for inpatient withdrawal inEurope, but has never been licensed for use in theUSA. Two randomized, double-blind placebo-controlled studies have shown chlormethiazole tobe equivalent in efficacy to chlordiazepoxide(Burroughs et al., 1985) and superior to placebo(Glatt et al., 1965). Group sizes were 49 and 20respectively. The patients in the Glatt et al. (1965)study were all given anticonvulsants. The placebogroup had a lower initial severity of withdrawalsymptoms and there was a 40% placebo responserate.
110 D. WILLIAMS and A. J. McBRIDE
Clonidine
Clonidine is an alpha-2-adrenoceptor partialagonist and was first produced as a nasaldecongestant but was subsequently found to havehypotensive properties. It was first studied inalcohol withdrawal because of the drug's knownefficacy in opioid withdrawal states (Bjorkvist,1975). In the Bjorkvist (1975) study, clonidine-treated patients' subjective ratings of withdrawalsymptoms (other than sleep disturbance) resolvedmore quickly than the control group. Objectivemeasures failed to show a significant difference.All patients were given additional sedation andanticonvulsants. The placebo group included morepatients with a history of previous withdrawalseizures. A 6 h cross-over study, on a sample of 11patients, found clonidine to control those symp-toms thought to be due to sympathetic overactivity(tachycardia, hypertension, tremor, diaphoresis)more quickly than placebo following a single dose(Wilkins et al, 1983).
Beta-blockers
Beta-blockers were first synthesized in 1958.Propranolol, the prototype drug, is non-selective,whereas atenolol is a specific beta-1-receptorblocker. Atenolol is thought to be safer indiabetics and unlike propranolol does not crossthe blood-brain barrier. Beta-blockers are used incardiovascular medicine and for controlling symp-toms in phaeochromocytomas, hyperthyroidismand panic attacks. The use of beta-blockers inalcohol withdrawal was first proposed in theLancet (Editorial, 1973) to treat symptoms dueto autonomic nervous system dysfunction.
Propranolol. One randomized, double-blind andplacebo-controlled trial has shown propranolol tobe superior to chlordiazepoxide or a combinationof both drugs in the treatment of mild withdrawalsymptoms (Sellers et al., 1977). Each samplegroup contained only six patients. Propranolol wassubsequently reported to cause hallucinations, aknown rare side-effect, more commonly whenused in alcohol withdrawal treatment (Guthrie,1989).
Atenolol
Two randomized, double-blind and placebo-controlled studies, using patient samples of > 100,have been published by the same centre in the
USA (Kraus et al., 1985; Horwitz et al., 1989).Atenolol was used as an adjunctive treatment tobenzodiazepines in mild withdrawal syndromes.The once-daily dosages were dependent on patientheart rate; no drug was given if the pulse was <50beats per minute (bpm), 50 mg if rates were50-79 bpm and 100 mg if the pulse was >80 bpm.Results suggested that the addition of atenololshortened the length of inpatient stay, hastenedimprovement of symptoms and decreased therequirement for benzodiazepines (Kraus et al.,1985). When used on an outpatient sample,patients on additional atenolol were less likely todrop-out of treatment and experienced less sub-jective craving for alcohol (Horwitz et al., 1989).The apparent anticraving effect appeared to belimited to the withdrawal period. Whether theeffects are similar for severe withdrawal phenom-ena has not been studied (Litten and Allen, 1991).The major criticism of both the American studiesis the poor matching of the control groups for age,sex and alcohol consumption history. There was ahigh overall treatment failure rate in the Kraus etal. study. Seventy-two patients were excludedfrom analysis as they developed seizures, or'combative or uncooperative states' during thewithdrawal phase.
AnticonvulsantsCarbamazepine. The principal uses of carba-
mazepine are as an anticonvulsant, in chronic painsyndromes especially trigeminal neuralgia, and inthe treatment of manic depressive psychosis.Carbamazepine acts on several neurotransmittersystems including GABA and noradrenaline, butits mechanism of action in the withdrawal state isuncertain. Carbamazepine was introduced in the1960s and has been extensively used in Scandi-navia for treatment of alcohol withdrawal states.Carbamazepine was first used in the treatment ofdelirium tremens in combination with sedatives inthe early 1970s (Brune and Busch, 1971).
In the one double-blind, placebo-controlled trial(Bjorkvist et al., 1976), a reducing dose ofcarbamazepine was given to 105 male outpatients.A hypnotic was also given. Results showed asignificantly faster decrease in withdrawal symp-toms in the carbamazepine group when comparedwith the placebo group. Despite one of theinclusion criteria being 'considered co-operative',there was a high drop-out rate, 18 patients from
DRUG TREATMENT OF ALCOHOL WITHDRAWAL 111
each group. Eleven of the carbamazepine groupsuffered side-effects such as dizziness.
Phenytoin. In a randomized, double-blindplacebo-controlled trial, Sampliner and Iber(1974) gave phenytoin 100 mg three times daily,in addition to unspecified chlordiazepoxideregimes, in the prevention of withdrawal seizuresin 150 patients with a past history of fits. Noseizures were seen in the 78 patients givenphenytoin, whereas 11 of 77 patients givenplacebo suffered one seizure each.
It has been suggested that any anticonvulsantwould exert a similar protective effect for thegroup of patients at high risk of fits (Saunders,1987) and that in most cases a sufficiently highdose of a benzodiazepine or chlormethiazole isadequate to prevent seizures (Edwards, 1987).
Other drugs
Lithium. Sellers et al. (1976) postulated thatlithium would lessen the activity of the sodium/potassium ATPase pump, which is increased inalcohol withdrawal. In a double-blind placebo-controlled trial on 18 patients with mild symp-toms, nine patients who commenced lithium0.3 mg three times daily prior to withdrawal hadsubjectively decreased symptoms. There was nochange in objective measurements of tremor orvital signs in these patients. The effect was onlypresent if treatment was commenced prior toabstaining from alcohol. No other treatment wasgiven.
Bromocriptine. Alcohol is known to activatecentral dopamine, therefore possibly creating aperiod of dopamine receptor subsensitivity in thewithdrawal period. Bromocriptine is a dopaminereceptor agonist, primarily used in reproductivedisorders, Parkinson's disease and acromegaly.
In a randomized, double-blind placebo-con-trolled trial, bromocriptine was found to have asignificant effect on anxiety, restlessness, depres-sion, tremor, nausea and sweating in withdrawalstates, when compared with placebo (Borg andWeinholt, 1982). Interpretation of results iscomplicated by all patients having been prescribedbenzodiazepines and carbamazepine, in varyingdoses, in addition to the study medication.
The therapeutic effect of bromocriptine wasfurther questioned by Burroughs et al. (1985) whofound it to be significantly inferior to chlormethia-zole and chlordiazepoxide.
Gamma-hydroxybutyric acid. The most recentdrug to be investigated in the treatment of alcoholwithdrawal states is gamma-hydroxybutyric acid(GHBA). This is a normal constituent of mammalbrain found especially in the hypothalamus andbasal ganglia. It is thought to be a neurotransmit-ter, having its own specific receptor sites. It isused in the treatment of narcolepsy, as it decreasesrapid eye movement (REM) sleep. It has beenproposed that GHBA may be of use in withdrawalstates because REM sleep is known to beincreased. In a randomized, double-blind pla-cebo-controlled study of 23 patients, GHBA wasshown to decrease tremor, sweating, nausea,depression, anxiety and restlessness occurringduring alcohol withdrawal in comparison withplacebo (Gallimberti et al., 1989). GHBA causedprominent side-effects, particularly 'dizziness',not further defined, which affected seven of the11 GHBA-treated patients in the first 2 days of thetrial.
DISCUSSION
Comparison of the randomized, double-blindplacebo-controlled trials described in this paper isdifficult due to methodological problems. Defini-tion of alcohol dependence or abuse ranged fromDSM-IIIR to Jellinek gamma type, whilst eight ofthe 14 studies simply used the term 'alcoholic'with no further elaboration. In addition to thepossible diversity in sample selection that thisimplies, comparison is further hampered by thefailure to use a single withdrawal symptom ratingscale. Of the 14 scales used, only five had beenpreviously published and each of these was used inonly one of the 14 studies.
If the studies are difficult to compare, themethodological quality of most studies adds to thedifficulties of the reviewer. Major failings in thestudies were commonplace. Few papers recordeddetails of inclusion and exclusion criteria orrevealed the proportion of the proposed studygroup thereby excluded. If 'severe' problems wereexcluded, severity was not clearly defined. Samplenumbers were usually small. There was a generalfailure to consider or monitor treatment compli-ance, and to control for previous treatment or co-morbidity. Complex drug regimes, in addition tothe trial drug, including the use of drugs known tobe effective in the treatment of alcohol withdrawal
112 D. WILLIAMS and A. J. McBRIDE
symptoms, were common.The clinical relevance of the studies is limited
because seven of the 14 trials excluded 'severelyill' patients. It is possible that these authorsconsidered any placebo treatment unethical forthe severely ill, but the work of Whitfield et al.(1978), discussed earlier in this paper, the highresponse rates in placebo groups and the lowincidence of delirium tremens and seizures in allstudies suggest that there are means by whichthese reservations can and should be overcome.
The aims of drug treatment for the alcoholwithdrawal syndrome are to enable the patient tostop drinking without psychological and physicalmorbidity or mortality. The effectiveness of anytreatment may be measured against the extent towhich it achieves these aims. The ideal drugtreatment for the alcohol withdrawal syndromeshould: (a) be effective against all withdrawalsymptoms; (b) be effective as an anticonvulsant;(c) be effective in the prevention and treatment ofdelirium tremens and hallucinosis; (d) have a rapidonset of action; (e) have an easily adjusted dosage;(f) be available in oral and injectable forms; (g)prevent craving and minimize short-term relapse;(h) have a wide safety range; (i) have no side-effects; (j) have no interactions with alcohol; (k)not be liable to abuse.
The drugs described in this review have notbeen equally or systematically tested against mostof these criteria. What follows is a summary ofwhat can reasonably be extracted from thepublished literature.
Benzodiazepines are superior to placebo in therelief of alcohol withdrawal symptoms apart fromhallucinosis, have a cleaner side-effect profilecompared with all other drugs tested (Moscowitzet al., 1983) and are safe in the high doses oftenrequired in delirium tremens (Woo and Greenblatt,1979). Diazepam is available in oral, rectal andintravenous forms. The main disadvantage of thebenzodiazepines is the risk of subsequent depen-dency, although this should be avoided if use isconfined to the withdrawal period. Side-effectsand interactions with other central nervous systemdepressants, particularly alcohol, which mayrarely cause apnoea are more likely in the elderly.Care should also be taken when commencingtreatment in those with hepatic impairmentespecially if using long-acting compounds. Seiz-ures are more likely to occur with short-acting
compounds (Hill and Williams, 1993).Chlormethiazole is effective and is available in
intravenous form for rapid sedation of acutelydisturbed patients. A major disadvantage ofchlormethiazole is its potentially lethal interactionwith alcohol, causing respiratory depression andarrest (Mclnnes, 1987). This is particularlyimportant to bear in mind for outpatients whomay be at greater risk of drinking whilst onmedication. Other potential problems includedepression of the gag reflex (predisposing thepatient to aspiration pneumonia) and confusion.As with benzodiazepines, dependency can beavoided by confining prescribing to the with-drawal period. However, concerns regarding thesafety of chlormethiazole have led to the manu-facturer advising outpatient use only in excep-tional circumstances and the refusal to issue asafety licence for it in the USA by the Food andDrug Administration of that country.
Clonidine and atenolol are ineffective in pre-venting major withdrawal effects such as deliriumtremens and have no anticonvulsant properties;they can only be considered as possible adjunctivetreatments for the suppression of sympatheticnervous system overactivity (Brewer, 1995) andpossibly in the reduction of craving during with-drawal.
Of the other drugs, carbamazepine may yet bedemonstrated to have the most important role toplay. The advantages of carbamazepine are that itis effective in severe alcohol withdrawal syn-drome, including delirium tremens, and is welltolerated. Carbamazepine does not interact withalcohol, is not contraindicated in cirrhosis andmay have an effect on the kindling process,thereby protecting against further withdrawalepisodes (Ballenger and Post, 1984). Whetherthis last action is common to other anticonvulsantsis not known. The disadvantages of carbamaze-pine are the potentially serious side-effects,including the small risk of potentially fatalhaematological complications, and the higherrelative cost compared to benzodiazepines. Allother anticonvulsants share the risk of side-effectsand it has been suggested that these agents mayactually increase the incidence of seizures duringwithdrawal (Hillbom and Hjelm-Jager, 1984).None of the other new agents has proved to besuperior to the older drugs. None is of proven usein severe withdrawal states, and all are much more
DRUG TREATMENT OF ALCOHOL WITHDRAWAL 113
expensive.With advances in neurophysiological and phar-
macological knowledge, there are many newcompounds that may prove to be useful in thetreatment of alcohol withdrawal symptoms. How-ever, future research will add substantially tocurrent knowledge only if such research isproperly conducted and described using: (1)recognized diagnostic criteria; (2) stated inclusionand exclusion criteria; (3) recognized, validatedand reliable rating scales for withdrawal symp-toms; (4) simple, clearly described, prescribingregimes; (5) sufficiently large sample sizes. Onlywith such studies will it be possible for newmethods and existing methods to be properlycompared, and integrated into clinical practice. Inaddition, further investigation into the biochemicaland neurophysiological changes, that occur duringalcohol dependence and withdrawal, is required togive a better understanding of drug action and toenable appropriate biochemical and neurophysio-logical measures to be used to monitor treatmentof the withdrawal phase.
In their review of studies up to 1981, Moscowitzet al. (1983) concluded that the only consistent,statistically significant finding for mild to moder-ately ill patients was the superiority of benzodia-zepines over placebo. The published evidencesince then suggests that several drugs are probablyeffective against some alcohol withdrawal symp-toms in some samples of drinkers. The most robustevidence remains for the use of adequate doses of along-acting benzodiazepine, with due considera-tion given to the small additional risks in theelderly and those with liver disease.
REFERENCES
Agricola, R. Malzarin, Q. M. and Surani, R. (1982)Treatment of alcohol withdrawal syndrome withcarbamazepine: A double blind comparison withtiapride. Journal of International Medical Research10, 160-165.
Ballenger, J. C. and Post, R. M. (1984) Carbamazepinein alcohol withdrawal syndromes and schizophrenicpsychosis. Psychopharmacology Bulletin 20,572-584.
Baumgartner, G. R. and Rowan. R. C. (1987) Clonidineversus chlordiazepoxide in the management ofalcohol withdrawal syndrome. Archives of InternalMedicine 147, 1223-1226.
Bjorkvist, S. E. (1975) Clonidine in alcohol withdrawal.Ada Psychiatrica Scandinavica 52, 256-263.
Bjorkvist, S. E., Isohanni, M , Makela, R. and Mazinen,L. (1976) Ambulant treatment of alcohol withdrawalsymptoms with carbamazepine. Acta PsychiatricaScandinavica 53, 333-342.
Borg, V. and Weinholt, T. (1980) A preliminary doubleblind study of 2 dopaminergic drugs in the treatmentof alcohol withdrawal syndrome. Current Therapeu-tic Research 27, 170-177.
Borg, V. and Weinholt, T. (1982) Bromocriptine in thetreatment of alcohol withdrawal syndrome. ActaPsychiatrica Scandinavica 65, 101-111.
Brewer, C. (1995) Second line and 'alternative'treatments for alcohol withdrawal. Alcohol andAlcoholism 30, 799-803.
Brune, F. and Busch, H. (1971) Anticonvulsant-sedativetreatment of delirium alcoholism. Quarterly Journalof Studies on Alcohol 32, 334-342.
Burroughs, A. K., Morgan, M. Y. and Sherlock, S.(1985) Double blind controlled trial of bromocrip-tine, chlordiazepoxide and chlormethiazole for al-cohol withdrawal symptoms. Alcohol andAlcoholism 20,263-271.
Chalmers, T. C , Smith, H., Blackburn, B., Silverman,B., Shroeder, B., Reitman, D. and Ambroz, A. (1981)A method for assessing the quality of a randomisedcontrolled trial. Controlled Clinical Trials 2, 31-49.
Chambers, J. and Shultz, J. D. (1965) Double blindstudy of three drugs in the treatment of acutealcoholic states. Quarterly Journal of Studies onAlcohol 26, 10-17.
Editorial (1973) Beta Blockers for Withdrawal Symp-toms? Lancet ii, 650-651.
Editorial (1981) Management of alcohol withdrawaltreatment. British Medical Journal 282, 502.
Edwards, G. A. (1987) Treatment of Drinking Problems,2nd edn. Blackwell, London.
Edwards, G. A., Arif, A. and Hodgson, R. (1981)Nomenclature and classification of drug and alcoholproblems. Bulletin of the World Health Organization59, 225-242.
Flygenring, J., Hasen, J., Hoist, B., Etersen, E. andSorensen. A. (1984) Treatment of alcohol with-drawal symptoms in hospitalised patients. ActaPsychiatrica Scandinavica 69, 398-408.
Gallimberti, L., Canton, G., Gentile, N., Ferri, M.,Cibin, M., Ferrara, S. D., Fadda, F. and Gessa, G. L.(1989) Gamma-hydroxy-butyric acid for treatment ofalcohol withdrawal syndrome. Lancet ii, 787-789.
Glatt, M., George, H. R. and Frisch, E. P. (1965)Controlled study of chlormethiazole in treatment ofthe alcohol withdrawal phase. British MedicalJournal 2, 401-404.
Glue, P. and Nutt, D. (1990) Overexcitement anddisinhibition: dynamic neurotransmitter interactionsin alcohol withdrawal. British Journal of Psychiatry157, 491^99.
Goldbert, T. M., Sanz, C. J., Rose, H. D. and Leitschuh,T. H. (1967) Comparative evaluation of treatment ofthe alcohol withdrawal syndrome. Journal of theAmerican Medical Association 201, 99-102.
Guthrie, S. (1989) The treatment of alcohol withdrawal.Pharmacolherapy 9, 131-143.
114 D. WILLIAMS and A. J. McBRIDE
Hill, A. and Williams, D. (1993) Hazards associatedwith use of benzodiazepines in alcohol detoxifica-tion. Journal of Substance Abuse Treatment 10,449^51.
Hillbom, M. and Hjelm-Jager, M. (1984) Should alcoholwithdrawal seizures be treated with anti-epilepticdrugs. Ada Neurologica Scandinavica 69, 39^4-2.
Horwitz, R., Gottlieb, L. D. and Kraus, M. L. (1989)The efficacy of atenolol in the outpatient manage-ment of the alcohol withdrawal syndrome. Archivesof Internal Medicine 149, 1089-1093.
Isbell, H., Fraser, H. R, Wilder, A., Beleville, R. E. andEisenman, A. J. (1955) An experimental study of theetiology of 'rum fits' and delirium tremens. Quar-terly Journal of Studies on Alcohol 16, 1-33.
Kaim, S. C , Klett, C. J. and Rothfield, B. (1969)Treatment of alcohol withdrawal symptoms. Acomparison of four drugs. American Journal ofPsychiatry 125, 54-59.
Kolin, I. S. and Linnet, D. I. (1981) Double blindcomparison of alprazolam and diazepam for sub-chronic withdrawal from alcohol. Journal of ClinicalPsychiatry 42, 169-173.
Kramp, P. and Rafaelsen, O. J. (1979) Deliriumtremens. A double blind comparison of diazepamand barbital treatment. Ada PsychiatricaScandinavica 58, 174-190.
Kraus, M. L., Gottlieb, L., Horwitz, R. and Anscher, M.(1985) Randomised clinical trial of atenolol inpatients with alcohol withdrawal. New EnglandJournal of Medicine 313, 905-909.
Lieber, C. S. (1991) Hepatic, metabolic and toxic effectsof ethanol. Alcoholism: Clinical and ExperimentalResearch 15, 573-592.
Litten, R. Z. and Allen, R. (1991) Pharmacotherapies foralcoholism: promising agents and clinical issues.Alcoholism: Clinical and Experimental Research 15,620-633.
Lundqvist, G. The clinical use of chlormethiazole. AdaPsychiatrica Scandinavica 1996, 42 (Suppl. 92),113-114.
Madden, J. S., Jones, D. and Frisch, E. P. (1969)Chlormethiazole and trifluoperazine in alcoholwithdrawal. British Journal of Psychiatry 115,1191-1192.
Malcolm, R., Ballenger, J., Sturgis, E. and Anton, R.(1989) Double blind comparison trial comparingcarbamazepine to oxazepam treatment of alcoholwithdrawal. American Journal of Psychiatry 146,617-621.
Manhem, P., Nilsson, L. H., Moberg, A., Wadstein, J.and Hokfelt, B. (1985) Alcohol withdrawal: effectsof clonidine treatment on sympathetic activity.Alcoholism: Clinical and Experimental Research 9,238-244.
McGrath, S. D. (1975) A controlled study of chlor-methiazole and chlordiazepoxide in treatment of theacute withdrawal phase of alcoholism. BritishJournal of Addiction 70 (Suppl.), 81-90.
Mclnnes, G. T. (1987) Chlormethiazole and alcohol: alethal cocktail. British Medical Journal 294, 592.
Moscowitz, G., Chalmers, T. L., Sacks, U. S.,
Fagerstrom, R. M. and Smith, H. (1983) Deficienciesof clinical trials of alcohol withdrawal. Alcoholism:Clinical and Experimental Research 7, 42-46.
O'Brien, J. E., Meyer, R. E. and Thomas, D. C. (1983)Double blind comparison of lorazepam and diaze-pam in the treatment of acute alcohol abstinencesyndrome. Current Therapeutic Research 34,825-831.
Palestine, M. L. and Alatorre, E. (1976) Control of acutealcohol withdrawal symptoms. Current TherapeuticResearch 20, 289-299.
Poutanen, P. (1979) Experience with carbamazepine intreatment of withdrawal symptoms in alcoholabusers. British Journal of Addiction 74, 201-204.
Ritola, E. and Malinen, L. (1981) A double blindcomparison of carbamazepine and chlormethiazolein treatment of alcohol withdrawal syndrome. AdaPsychiatrica Scandinavica 64, 254-259.
Robinson, B. J., Robinson, G. M., Maling, T. J. B. andJohnson, R. H. (1989) Is clonidine useful intreatment of alcohol withdrawal? Alcoholism: Clin-ical and Experimental Research 13, 95-98.
Sampliner, R. and Iber, F. L. (1974) Diphenhydantoincontrol of withdrawal seizures. Journal of theAmerican Medical Association 230, 1430-1432.
Saunders, J. B. (1987) Drug treatment in alcoholism. InAntimanics, Anticonvulsants and Other Drugs inPsychiatry, Burrow, G. D. and Norman, T. R. eds,Vol. 4, pp. 343-369. Elsevier, Amsterdam.
Schuckit, M. A. (1995) Alcohol and depression, aclinical perspective. Ada Psychiatrica Scandinavica377 (Suppl.), 28-32.
Serfaty, M. and Masterton, G. (1993) Fatal poisoningsattributed to benzodiazepines in Britain during the1980's. British Journal of Psychiatry 163, 386-393.
Sellers, E. M. and Kalant, H. (1976) Alcohol intoxica-tion and withdrawal. New England Journal ofMedicine 294, 757-762.
Sellers, E. M., Cooper, S. D., Zilm, D. H. and Shanks, C.(1976) Lithium treatment during alcohol withdrawal.Clinical Pharmacology and Therapy 20, 199-206.
Sellers, E. M., Zilm, D. H. and Degani, N. C. (1977)Comparative efficacy of propranolol and chlordia-zepoxide in alcohol withdrawal. Quarterly Journalof Studies on Alcohol 38, 2096-2108.
Sellers, E. M., Naranjo, C. A., Harrison, M., Devenyi,P., Roach, C. and Sykora, K. (1983) Diazepamloading. Simplified treatment of alcohol withdrawal.Clinical Pharmacology and Therapy 34, 822-826.
Sereny, G. and Kalant, H. (1965) Comparative clinicalevaluation of chlordiazepoxide and promazine intreatment of alcohol withdrawal syndrome. BritishMedical Journal 1, 92-97.
Shaw, G. K. (1995) Detoxification, the use ofbenzodiazepines. Alcohol and Alcoholism 30,765-770.
Stuppaeck, C. H., Miller, E., Pycha, R., Whitworth, A.B., Oberbauer, H. and Fleischhallker, W. W. (1992)Carbamazepine versus oxazepam in the treatment ofalcohol withdrawal. Alcohol and Alcoholism 27,153-158.
Thompson, W. L., Johnson, A. D. and Maddrey, W. L.
DRUG TREATMENT OF ALCOHOL WITHDRAWAL 115
(1975) Diazepam and paraldehyde for treatment ofsevere delirium tremens. Annals of Internal Medicine82, 175-180.
Turbridy, P. (1988) Alprazolam versus chlormethiazolein acute alcohol withdrawal. British Journal ofAddiction 83, 581-585.
Victor, M. (1966) Treatment of alcoholic intoxicationand the withdrawal syndrome. PsychosomaticMedicine 28, 636-650.
Victor, M. and Adams, R. (1953) The effect of alcoholon the nervous system. Research Publication of theAssociation for Nervous and Mental Disorders 32,526-573.
Whitfield, C. L., Thompson, G., Lamb, A., Spencer, U.,Pfiefer, M. and Browning-Ferrado, M. (1978)Detoxification of 1024 alcoholic patients without
psychoactive drugs. Journal of the American Medi-cal Association 239, 1409-1410.
Wilkins, A. J., Jenkins, W. J. and Steiner, J. A. (1983)Efficacy of clonidine in treatment of alcohol with-drawal state. Psychopharmacology 81, 78-81.
Wilson, A. and Vufcano, B. A. (1985) Double blind trialof alprazolam and chlordiazepoxide in managementof acute ethanol withdrawal syndrome. Alcoholism:Clinical and Experimental Research 9, 23-27.
Woo, E. and Greenblatt, D. (1979) Massive benzodi-azepine requirements in acute alcohol withdrawal.American Journal of Psychiatry 136, 821-823.
World Health Organization (1992) The ICD 10Classification of Mental and Behavioural Disorders:Clinical Descriptions and Diagnostic Guidelines.WHO, Geneva.
![Introduction WHO ECDD Questionnaire: delta-9 ... · Nomenclature Regulations 1967, etc.), the [Ministry of Health][other responsible governmental ... Alcohol withdrawal symptoms Opioid](https://static.documents.pub/doc/80x56/5fa3fd5ab0979e312e45de1e/introduction-who-ecdd-questionnaire-delta-9-nomenclature-regulations-1967.jpg)