213 213..220

BackgroundBackground The prevalenc e ofThe prevalenc e of

benz odi azep ine c ons u mp ti on in Eur opeanbenz odi azep ine c ons u mp ti on in European

count ries remains at 2^3% of the generalcount ries remains at 2^3% of the general

popula tion despite the we ll-docume ntedpopu l a tion desp it e the well-docume nted

disadvantages of long-term use.disadvantages of long-term use.

AimsAims To review systemat i cally theTo rev i ew systemat i cally the

success rates of different benzodia zepinesuccess rat es of differ ent benzodiazepi ne

discont inuat ion strat egies.discontinuation strategies.

MethodMethod Meta-analysis ofcomparableMeta-analysis ofcomparable

intervention studies.intervention studies.

ResultsResults Tw e n t y - n i n e a r t i c l e s m e tTw e n t y - n i n e a r t i c l e s m e t

inclusion criteria.Two groups ofinclusion criteria.Two groups of

i ntervent ions were ident ified; m i n i malinterventions were identif ied; minimal

intervention (e.g. giving simple advice inintervention (e.g. giving simple advice in

the form of a let ter or meeting to a largethe form of a lett er or meeti ng t o a large

group of people;group of people; nn¼3), and systematic3), and systematic

d i s c ontinua tion ( defined as tr eat m entd i s c onti n uat i on ( defi ned as tr eat m ent

programmesled by a physician orprogrammesled by a physician or

psychologist;psychologist; nn¼ 26). Both were found to26). Bothwere found to

be significantly more effective thanbe sig nificant ly more effecti ve than

treatment as usual: m in imal in terventionstreatment as usual: minimal interventions

(pooled OR(pooled OR¼2.8, 95% CI1.6 ^5.1);2.8,95% CI1.6^5.1);

sy stem a ti c d i s c onti n ua tion alone ( onesy stem a ti c d i sc o nti n ua tion alone ( one

st udy,ORst ud y,O R¼ 6.1,95% CI 2.0^18.6).6.1,95% CI 2.0^18.6).

Augmentation of systematicAugmentation of systematic

discontinuation with imipramine (twodiscontinuation with imipramine (two

st udies,ORst udies,OR¼ 3.1,95% CI1.1^9.4) or group3.1,95% CI1.1^9.4) or group

cognitive^ behaviouraltherapy forcognitive^ behavioural therapy for

patients with insomnia (two studies,patients withinsomnia (two studies,

OROR¼ 5.5,95% CI 2.3^14.2) was super ior to5.5,95% CI 2.3^1 4.2 ) was superior to

sy stem a ti c d i s c onti n ua tion alone.systemat ic d iscont inuation alone.

ConclusionsConclusions Ev idence was found forEv idence was found for

the efficacy of stepped care (minimalthe efficacy of stepped care (minimal

intervention followed by systematicintervention followed by systematic

discontinuation alone) in discontinuingdiscontinuation alone) in discontinuing

long-term benzodiazepine use.long-term benzodiazepine use.

Declaration of interestDeclaration of interest None .None .

Since the early 1960s benzodiazepines haveSince the early 1960s benzodiazepines have

become widely available, reaching prescrip-become widely available, reaching prescrip-

tion peaks in the 1970s (Lader, 1991).tion peaks in the 1970s (Lader, 1991).

Subsequently more and more data wereSubsequently more and more data were

reported indicating the disadvantages ofreported indicating the disadvantages of

long-term benzodiazepine use, such as thelong-term benzodiazepine use, such as the

risk of dependence, a higher risk of acci-risk of dependence, a higher risk of acci-

dents and falls, and cognitive disturbancesdents and falls, and cognitive disturbances

(Taylor(Taylor et alet al, 1998). In the past few years, 1998). In the past few years

the prevalence rate of benzodiazepine con-the prevalence rate of benzodiazepine con-

sumption in most European countries issumption in most European countries is

estimated to be stable or slightly decreasingestimated to be stable or slightly decreasing

(Stillwell & Fountain, 2002), but remains(Stillwell & Fountain, 2002), but remains

at levels varying between 2% and 3% ofat levels varying between 2% and 3% of

the general population (Zandstrathe general population (Zandstra et alet al,,

2002). Although long-term therapeutic use2002). Although long-term therapeutic use

of benzodiazepines is controversial, limitedof benzodiazepines is controversial, limited

evidence suggests long-term efficacy inevidence suggests long-term efficacy in

specific diagnostic groups such as panic dis-specific diagnostic groups such as panic dis-

order and social phobia (Schweizerorder and social phobia (Schweizer et alet al,,

1993; Otto1993; Otto et alet al, 2000). The prevalence of, 2000). The prevalence of

these disorders among people who arethese disorders among people who are

long-term benzodiazepine users, however,long-term benzodiazepine users, however,

is relatively low (Zandstrais relatively low (Zandstra et alet al, 2004)., 2004).

Problems experienced by patients stop-Problems experienced by patients stop-

ping long-term benzodiazepine use initiatedping long-term benzodiazepine use initiated

the development of treatment strategiesthe development of treatment strategies

for discontinuing these drugs. Russell &for discontinuing these drugs. Russell &

Lader (1993) proposed a stepped careLader (1993) proposed a stepped care

approach to address the problem of long-approach to address the problem of long-

term use. They advised starting with aterm use. They advised starting with a

minimal intervention and, if this failed,minimal intervention and, if this failed,

gradually intensifying treatment fromgradually intensifying treatment from

supervised gradual withdrawal after patientsupervised gradual withdrawal after patient

assessment to specialised care includingassessment to specialised care including

augmentation strategies. In order to sum-augmentation strategies. In order to sum-

marise the evidence for the individual stepsmarise the evidence for the individual steps

of such programmes, we carried out meta-of such programmes, we carried out meta-

analyses of the success rates of the differentanalyses of the success rates of the different

benzodiazepine discontinuation strategies.benzodiazepine discontinuation strategies.

METHODMETHOD

Identification of studiesI dentification of studi es

An initial search was made of the databasesAn initial search was made of the databases

PubMed and PsycINFO for the periodPubMed and PsycINFO for the period

1966 to September 2004 and the Cochrane1966 to September 2004 and the Cochrane

Library in December 2004, using theLibrary in December 2004, using the

keywords BENZODIAZEPINE(S) inkeywords BENZODIAZEPINE(S) in

combination with WITHDRAWAL,combination with WITHDRAWAL,

DETOXIFICATION, DEPENDENCE,DETOXIFICATION, DEPENDENCE,

DISCONTINUATIONDISCONTINUATION oror LONG-TERM.LONG-TERM.

This search was extended by a manualThis search was extended by a manual

search of the reference lists of all benzodia-search of the reference lists of all benzodia-

zepine discontinuation studies and benzo-zepine discontinuation studies and benzo-

diazepine discontinuation augmentationdiazepine discontinuation augmentation

studies (Fig. 1).studies (Fig. 1).

Inclusion criteriaInclusion criteria

Papers were included in the review if theyPapers were included in the review if they

met the following criteria:met the following criteria:

(a)(a) the study had a randomised controlledthe study had a randomised controlled

design;design;

(b)(b) the outcomes of discontinuation werethe outcomes of discontinuation were

presented separately for each treatmentpresented separately for each treatment

arm;arm;

(c)(c) long-term benzodiazepine use waslong-term benzodiazepine use was

defined as daily use for at least 3defined as daily use for at least 3

months.months.

Excluded were case series, review papers,Excluded were case series, review papers,

double publications, experimental researchdouble publications, experimental research

or clinical trials evaluating the efficacy ofor clinical trials evaluating the efficacy of

benzodiazepine treatment for a fixedbenzodiazepine treatment for a fixed

period, and animal research. Authorsperiod, and animal research. Authors

R.C.O.V. and J.E.C. independently checkedR.C.O.V. and J.E.C. independently checked

the inclusion and exclusion criteria of thethe inclusion and exclusion criteria of the

identified studies.identified studies.

Selection procedure, dataSelection procedure, data

extraction and quality assessmentextraction and quality assessment

Included studies were coded twice byIncluded studies were coded twice by

R.C.O.V. and J.E.C. Discrepancies in theR.C.O.V. and J.E.C. Discrepancies in the

two coding forms were resolved by consen-two coding forms were resolved by consen-

sus after discussion or by referring to thesus after discussion or by referring to the

data in the original article. This methoddata in the original article. This method

yielded one coding form per article. The in-yielded one coding form per article. The in-

tervention type was added to the codingtervention type was added to the coding

form by distinguishing between minimalform by distinguishing between minimal

interventions and systematic discontinua-interventions and systematic discontinua-

tion programmes. Minimal interventionstion programmes. Minimal interventions

were defined as simple interventionswere defined as simple interventions

applicable to large groups of people, for ex-applicable to large groups of people, for ex-

ample, an advisory letter or a meeting inample, an advisory letter or a meeting in

which patients who are long-term benzo-which patients who are long-term benzo-

diazepine users are advised to stop takingdiazepine users are advised to stop taking

the drug. Systematic discontinuation pro-the drug. Systematic discontinuation pro-

grammes were defined as treatmentgrammes were defined as treatment

programmes guided by a physician orprogrammes guided by a physician or

psychologist. We sub-categorised thesepsychologist. We sub-categorised these

treatment programmes into systematic dis-treatment programmes into systematic dis-

continuation alone or discontinuation withcontinuation alone or discontinuation with

either psychotherapy or pharmacotherapy.either psychotherapy or pharmacotherapy.

The coding form consisted of the followingThe coding form consisted of the following

items:items:

213213

BRITISH JOURNAL OF PSYCHIATRYBRITISH JOURNAL OF PSYCHIATRY (2 0 0 6), 1 89, 213 ^ 2 20 . doi : 10 .119 2/ bjp.18 9. 3. 213(2006), 189, 213^220. doi: 10.1192/bjp.189.3.213 REVIEW ARTICLEREVIEW ARTICLE

Strategies for discontinuing long-termStrategies for discontinuing long-term

benzodiazepine usebenzodiazepine use

Meta-analysisMeta-analysis

RICHARD C . OUDE VO SHAAR, J AAP E . COUVEE,RICHARD C.OUDE VOSHAAR,JAAPE.COUVE

ANTONJ.L.M.VANBALKOM,PAULG.H.MULDERandFRANSG.ZITMANANTONJ.L.M.VANBALKOM,PAULG.H.MULDERandFRANSG.ZITMAN

https://doi.org/10.1192/bjp.189.3.213 Published online by Cambridge University Press

OUD E VO S HA A R E T A LOUD E VO S HA A R E T A L

(a)(a) inclusion criteria (minimum durationinclusion criteria (minimum duration

of benzodiazepine use 3.6 or 12of benzodiazepine use 3.6 or 12

months) and diagnosed benzodiazepinemonths) and diagnosed benzodiazepine

dependence (yes/no);dependence (yes/no);

(b)(b) results at post-treatment outcome;results at post-treatment outcome;

(c)(c) year of publication;year of publication;

(d)(d) domain of use (i.e. psychiatric diagnosisdomain of use (i.e. psychiatric diagnosis

or symptoms of included patients);or symptoms of included patients);

(e)(e) steps of taper (abrupt, fixed orsteps of taper (abrupt, fixed or

symptom-guided);symptom-guided);

(f)(f) tapered withdrawal after transfer to atapered withdrawal after transfer to a

long-acting benzodiazepine (yes/no);long-acting benzodiazepine (yes/no);

(g)(g) history of benzodiazepine use (dosage,history of benzodiazepine use (dosage,

type, duration of use);type, duration of use);

(h)(h) in-patient treatment (yes/no);in-patient treatment (yes/no);

(i)(i) setting (primary care, psychiatric clinicsetting (primary care, psychiatric clinic

or addiction clinic).or addiction clinic).

Mean equivalent benzodiazepine do-Mean equivalent benzodiazepine do-

sages were obtained from the articles orsages were obtained from the articles or

calculated in diazepam equivalents (Zitmancalculated in diazepam equivalents (Zitman

& Couvee, 2001). If no information was& Couve

e, 2001). If no information was

available to calculate the dosage in diaze-available to calculate the dosage in diaze-

pam equivalents, we categorised thepam equivalents, we categorised the

dosages as low (within the therapeuticdosages as low (within the therapeutic

range, or less than 15 mg), high (above therange, or less than 15 mg), high (above the

therapeutic range, or more than 30 mg) ortherapeutic range, or more than 30 mg) or

medium (patients using benzodiazepinesmedium (patients using benzodiazepines

within and above the therapeutic range, orwithin and above the therapeutic range, or

15–30 mg).15–30 mg).

The quality of the included articles wasThe quality of the included articles was

assessed twice by R.C.O.V., J.E.C. and/assessed twice by R.C.O.V., J.E.C. and/

or A.J.L.M.v.B. using the Amsterdam–or A.J.L.M.v.B. using the Amsterdam–

Maastricht consensus list, which coversMaastricht consensus list, which covers

the Chalmers criteria usually applied inthe Chalmers criteria usually applied in

the assessment of study quality (Van Tulderthe assessment of study quality (Van Tulder

et alet al, 1997; Van Boeijen, 1997; Van Boeijen et alet al, 2005)., 2005).

Statistical analysisStatistical analysis

Since we were interested in the success ratesSince we were interested in the success rates

of benzodiazepine discontinuation (binaryof benzodiazepine discontinuation (binary

outcome) and because in some studies dataoutcome) and because in some studies data

were sparse, we used stratified exact (con-were sparse, we used stratified exact (con-

ditional) methods with odds ratios as fixed-ditional) methods with odds ratios as fixed-

effects association measures. Exacteffects association measures. Exact PP-values-values

for testing significance and homogeneity offor testing significance and homogeneity of

odds ratios across studies were calculated,odds ratios across studies were calculated,

and exact 95% confidence intervals wereand exact 95% confidence intervals were

estimated. In cases in which homogeneityestimated. In cases in which homogeneity

had to be rejected (had to be rejected (PP550.05) we introduced0.05) we introduced

a random effect in order to account fora random effect in order to account for

between-study variability of the oddsbetween-study variability of the odds

ratios. In such cases the asymptotic directratios. In such cases the asymptotic direct

pooling method was used for calculatingpooling method was used for calculating

significance levels and confidence limits.significance levels and confidence limits.

RESULTSRESULTS

The initial search yielded 5264 referenceThe initial search yielded 5264 reference

titles in PubMed, 1260 in PsychINFO andtitles in PubMed, 1260 in PsychINFO and

666 in the Cochrane Library. Of these,666 in the Cochrane Library. Of these,

275 titles were identified by R.C.O.V. and275 titles were identified by R.C.O.V. and

J.E.C. as having possible relevance to dis-J.E.C. as having possible relevance to dis-

continuation of long-term benzodiazepinecontinuation of long-term benzodiazepine

use. (The full reference list is presented inuse. (The full reference list is presented in

data supplement 1 to the online version ofdata supplement 1 to the online version of

this paper.) After screening of the abstractsthis paper.) After screening of the abstracts

and if necessary the full text, 246 papersand if necessary the full text, 246 papers

were excluded (Fig. 1) and 29 paperswere excluded (Fig. 1) and 29 papers

met the inclusion criteria (Tyrermet the inclusion criteria (Tyrer et alet al,,

1981; Lader & Olajide, 1987; Ashton1981; Lader & Olajide, 1987; Ashton etet

alal, 1990; Cantopher, 1990; Cantopher et alet al, 1990; Jones,, 1990; Jones,

1990; Udelman & Udelman, 1990;1990; Udelman & Udelman, 1990;

Garcia-BorregueroGarcia-Borreguero et alet al, 1991; Schweizer, 1991; Schweizer

et alet al, 1991; Di Costanzo & Rovea, 1992;, 1991; Di Costanzo & Rovea, 1992;

LaderLader et alet al, 1993; Otto, 1993; Otto et alet al, 1993; Bashir, 1993; Bashir

et alet al, 1994; Cormack, 1994; Cormack et alet al, 1994; Schweizer, 1994; Schweizer

et alet al, 1995; Tyrer, 1995; Tyrer et alet al, 1996; Lemoine, 1996; Lemoine etet

alal, 1997; Hantouche, 1997; Hantouche et alet al, 1998; Garfinkel, 1998; Garfinkel

et alet al, 1999; Petrovic, 1999; Petrovic et alet al, 1999; Rickels, 1999; Rickels etet

alal, 1999, 2000; Cialdella, 1999, 2000; Cialdella et alet al, 2001;, 2001;

Zitman & Couvee, 2001; GerraZitman & Couve

e, 2001; Gerra et alet al,,

2002; Vorma2002; Vorma et alet al, 2002; Baillargeon, 2002; Baillargeon etet

alal, 2003; Oude Voshaar, 2003; Oude Voshaar et alet al, 2003, 2003aa; Rynn; Rynn

et alet al, 2003; Morin, 2003; Morin et alet al, 2004)., 2004).

Table 1 lists the scores for method-Table 1 lists the scores for method-

ological quality of the included studiesological quality of the included studies

measured with the Amsterdam–Maastrichtmeasured with the Amsterdam–Maastricht

consensus list. The sumscore (range 0–18)consensus list. The sumscore (range 0–18)

can be considered to be a proxy of studycan be considered to be a proxy of study

quality. For studies evaluating psychother-quality. For studies evaluating psychother-

apy augmentation strategies, however, theapy augmentation strategies, however, the

maximum score is 17. The quality of themaximum score is 17. The quality of the

included studies ranged from 8 to 17, corre-included studies ranged from 8 to 17, corre-

sponding with a moderate to excellentsponding with a moderate to excellent

study quality. Recency of the study cor-study quality. Recency of the study cor-

related moderately with better qualityrelated moderately with better quality

(Spearman’s rank correlation coefficient(Spearman’s rank correlation coefficient

0.44,0.44, PP¼0.02). Patient numbers and demo-0.02). Patient numbers and demo-

graphic characteristics of the samples ingraphic characteristics of the samples in

the included papers are summarised inthe included papers are summarised in

Table 2. The numbers of patients leavingTable 2. The numbers of patients leaving

the studies were relatively low, which canthe studies were relatively low, which can

be explained by the fact that patientbe explained by the fact that patient

withdrawal was classified as discontinua-withdrawal was classified as discontinua-

tion failure in the 14 studies reportingtion failure in the 14 studies reporting

intention-to-treat analyses (Table 1). Nointention-to-treat analyses (Table 1). No

difference was found in withdrawal ratesdifference was found in withdrawal rates

between studies of different treatmentbetween studies of different treatment

modalities. Compared with those usingmodalities. Compared with those using

benzodiazepine in the general population,benzodiazepine in the general population,

minimal intervention studies included aminimal intervention studies included a

higher proportion of women and thehigher proportion of women and the

participants had a relatively higher ageparticipants had a relatively higher age

(Zandstra(Zandstra et alet al, 2002). Age and gender dis-, 2002). Age and gender dis-

tribution of patients recruited in the onlytribution of patients recruited in the only

controlled study of systematic discontinua-controlled study of systematic discontinua-

tion alone was comparable with that oftion alone was comparable with that of

long-term benzodiazepine users in thelong-term benzodiazepine users in the

population, as found by Zandstrapopulation, as found by Zandstra et alet al

(2002). Systematic discontinuation studies(2002). Systematic discontinuation studies

with augmentation strategies, on thewith augmentation strategies, on the

contrary, included a lower proportion ofcontrary, included a lower proportion of

women and a relatively lower agewomen and a relatively lower age

compared with the ‘average’ person usingcompared with the ‘average’ person using

benzodiazepines in the population. Thebenzodiazepines in the population. The

characteristics of the included studiescharacteristics of the included studies

214214

Fig. 1Fig. 1 Search strategy. Note: the study by Oude VoshaarSearch strategy. Note: the study by OudeVoshaar et alet al (20 03(2003aa ) was included twice owing to its) was included twice owing to its

three-condition randomised controlled design.three-condition randomised controlled design.

https://doi.org/10.1192/bjp.189.3.213 Published online by Cambridge University Press

DISCONTINUING LONG-TERM BENZODIAZEPINESD I S C ON T INUIN G LONG -T E R M B E NZO DI A Z E P IN E S

215215

Ta b l e 1Ta b l e 1 Validity scores of included studies assessed with the Amsterdam ^Maastricht consensus listValidity scores of included studies assessed with the Amsterdam^Maastricht consensus list

JonesJones

(1990)(1990)

Corm ackCormack

et alet al (1994)(1994)

BashirBashir

et alet al (1994)(1994)

Oude V oshaarOude Voshaar

et alet al (2003(2003aa))

OttoOtto

et alet al (1993)(1993)

VormaVorma

et alet al (2002)(2002)

BaillargeonBaillargeon

et alet al (2003)(2003)

MorinMorin

et alet al (2004)(2004)

TyrerTyrer

et alet al (1981)(1 981)

Lader &Lade r &

Olajide (1987)Olajide (1987)

Validity criteriaV al idity criteria

Adeq uate randomisation procedureAdequate randomisati on procedure 00++++++++++++++00++

Concealed random allocation of treatme ntsCon cealed random allocation of treatments 77 ++ 77 ++ 77 ++ 77 00 7777

Baseline similarity testedBaseline similarity tested ++ 7777 ++++++++++ 7777

Control for co-intervent ions in desig nControl for co-intervent ions in desig n 77777777 ++++++++0000

Check for adhe re nce to interve nt i onsCheck for adhe ren ce to intervent i ons ++ 7777 ++ 77 ++++++ 77 ++

V al i d outcome measuresValid outcome measures 77 ++ 77 ++++++++++ 77 ++

Rele vant outcome measuresRele vant outcome measures ++++++++++++++++++++

Outcome assessor maskedOutcome assessor masked 77 0000++0000000000++

Care provider maskedCare provid e r masked 7777777777777777 ++++

Patient maskedPatient masked 7777777777777777 ++++

Withdrawals (proportion; inequalit y between groups; reasons forWithd rawals (proportion; inequality between groups; reasons for

withd rawal reported)withd rawal reported)

++++++++++++++++++++

Ident ical timin g of outcome assessment for all intervent i on groupsIde nt ical timi ng of outcome assessment for all intervent i on groups ++++++++++++++++ 77 ++

Intent ion-to-treat analysisIntent ion-to-treat analysis 00 7777 ++++++++++++++

Descript iv e criteriaDescriptiv e criteria

Specification of eligibil ity criteriaSpecification of eligibility criteria ++++++++++++++++++++

Descript ion of the inte rvent i onsDesc ript ion of the interve nt ions ++++++++++++++++++++

Follow-upFollow-up 7777 ++ 77 ++00++++++ 77

Adverse effectsAdverse effects 00 77 ++++++ 77 ++++++++

Statist ical crite riaStat i stical criteria

Sample size presented at randomisati on and outcomeSample size present ed at randomisation and outcome 00 77 ++++++++++++++00

Presentati on of point estimates and distri but ion measuresPresentati on of point estimates and distri but ion measures ++00++++++++++++0000

Total score (range 0^19)Total score (range 0^19) 88991010 1515 1414 1414 1515 1515 1010 1 313

+, Pr esent;+, Present; 77,absent;0,notreported.,absent;0,notreported.

https://doi.org/10.1192/bjp.189.3.213 Published online by Cambridge University Press

OUD E VO S HA A R E T A LOUD E VO S HA A R E T A L

21621 6

Table 1 (continued)Table 1 (conti nued) Validity scores of included studies assessed with the Amsterdam ^Maastricht consensus listValidity scores of included studies assessed with the Amsterdam^Maastricht consensus list

AshtonAshton

et alet al

(1990)(1990)

CantopherCantopher

et alet al

(1990)(1990)

UdelmanUdelman

&Udelman&Udelman

(1990)(1990)

Garcia-Garcia-

BorregueroBorreguero

et alet al (1991)(1991)

Schw ei ze rSchwe ize r

et alet al

(1991)(1991)

Di CostanzoDi Costanzo

&Rovea&Rovea

(1992)(1992)

LaderLader

et alet al

(1993)(1993)

Schwe ize rSchwe i zer

et alet al

(1995)(1995)

TyrerTyrer

et alet al

(1996)(1996)

LemoineLemoine

et alet al

(1997)(1997)

Validity criteriaValidity criteria

Adequate randomisation procedureAdeq uate randomisation procedu re ++++++00++++++++++++

Concealed random allocation of treatmentsConcealed random allocation of treatments 7777 00 7777 ++ 77777777

Baseline similarity testedBaseline similarity tested ++++++ 77 ++++ 777777 ++

Control for co-intervent i ons in desi g nControl for co-intervent i ons in desi g n ++++++++ 77 0000 7777 00

Check for adhe ren ce to intervent i onsCheck for adhe re nce to interve nt ions ++++++ 77 ++++++++ 77 ++

V al id outcome measuresVali d outcome measures ++++++++++++++++00++

Rele vant outcome measuresRelevant outcome measures ++++++++++++++++++++

Outcome assessor maskedOutcome assessor masked ++0000 77 00++00000000

Care provide r maskedCare provid er masked ++++++ 77 ++00++++++++

Patient maskedPat ient masked ++++++ 77 ++++++++++++

Withdra wals (proportion; inequal it y between groups; reason forWithdrawals (proportion; inequality between groups; reason for

withd rawal reported)withdrawal reported)

++++++++++++++++++++

Ident i cal timin g of outcome assessment for all interve nt ion grou psIde nt ical timi ng of outcome assessment for all intervent i on groups ++++0000++++++++++++

Intent i on-to-treat analysisIntent ion-to-treat analys is 00 77 ++++ 77 ++000000++

Descript iv e criteriaDescript iv e criteria

Specification of eli gi bil ity criteriaSpecification of eligi bility criteria ++++++++++++++++++++

Descripti on of the intervent i onsDescri pt ion of the inte rvent i ons ++++++++++++++++++++

Follow-upFollow-up ++00 7777777777 ++ 77 ++

Adverse effectsAdverse effects ++++++++++00++++++++

Statistical criteriaStatist i ca l criteria

Sample size presented at randomisation and outcomeSam ple size presented at randomisati on and outcome ++++0000++ 7777 ++ 7777

Presentati on of point estimates and dist ri but ion measuresPresentat ion of point esti mates and distribut i on measures 00++00++++++++++++++

Total score (range 0^19)Total score (range 0^19) 1616 1 515 1313 991414 1414 1212 1414 1010 1 515

+, Present;+, Present; 77,absent;0,notreported., absent; 0, not reported.

https://doi.org/10.1192/bjp.189.3.213 Published online by Cambridge University Press

DISCONTINUING LONG-TERM BENZODIAZEPINESD I S C ON T INUIN G LONG -T E R M B E NZO DI A Z E P IN E S

217217

Table 1 (conti nued)Table 1 (conti nued) Validity scores of included studies assessed with the Amsterdam^Maastricht consensus listValidity scores of included studies assessed with the Amsterdam ^Maastricht consensus list

GarfinkelGarfinkel

et alet al

(1999)(1999)

HantoucheHantouche

et alet al

(1998)(1998)

PetrovicPetrovic

et alet al

(1999)(1999)

RickelsRickels

et alet al

(1999)(1999)

RickelsRickels

et alet al

(2000)(2000)

CialdellaCialdella

et alet al

(200(2001)1)

ZitmanZitman

&Couvee&Couve

(200(2001)1)

GerraGerra

et alet al

(2002)(2002)

RynnRynn

et alet al

(2003)(2003)

V alid ity criteriaV a l i d i ty criteria

Adequate randomisation procedur eAdeq uate randomisation procedure ++++00++++++++++++

Concealed random allocati on of treatmentsConcealed random allocation of treatme nts 77 00 77777777 ++ 7777

Baseline similarity testedBaseline similarity tested ++++ 77 ++++++++++++

Control for co-intervent ions in desig nControl for co-intervent i ons in desi g n 77 ++ 777777 ++ 7777 00

Check for adhe ren ce to intervent i onsCheck for ad here nce to interv ent i ons ++++ 77 ++++++++++++

V al id outcome measuresV al id outcome measures ++++++++++++++++++

Rele vant outcome measuresRelevant outcome measures ++++++++++++++++++

Out c om e assessor maske dOutcome assessor masked ++++00++++++00 77 ++

Care provid e r maskedCare provi d er masked ++00 77 00++++++ 77 ++

Patient maskedPat ient masked ++++++++++++++++++

Withd rawals (proportion; inequality between groups; reasons for withd rawalWithdraw als (proportion; inequalit y between groups; reasons for withd rawal

reported)reported)

++++++++++++++ 77 ++

Ident i cal timin g of outcome assessment for all interve nt ion grou psIde nt ical timin g of outcome assessment for all interve nt ion groups ++++++++++++++++++

Intent i on-to-treat analysisIntent i on-to-treat analysis ++00++00 77 00++00 77

Descript iv e criteriaDescript iv e criteria

Specification of eligibility criteriaSpecification of eli g ibi l ity criteria 00++++++++++++++++

Descript ion of the interve nt ionsDescription of the inte rvent i ons ++++++++++++++++++

Follow-upFollow-up ++ 777777 ++ 77 ++++++

Adverse effectsAdverse effects ++++++++++++++++++

Stat i stical criteriaStatistical criteria

Sample size presented at randomisation and outcomeSam ple size presented at randomisation and outcome ++++++++ 77 ++++00 77

Presentati on of point estimates and distri but ion measuresPresentati on of point estimates and dist ri but ion measures ++++++++00++++00++

Total score (range 0^19)Total score (range 0^19) 1616 1515 1 111 1414 1414 1616 1717 1010 1 515

+, Present;+, Present; 77,absent;0,notreported., absent; 0, not repor ted.

https://doi.org/10.1192/bjp.189.3.213 Published online by Cambridge University Press

OUD E VO S HA A R E T A LOUD E VO S HA A R E T A L

according to the main items of the codingaccording to the main items of the coding

form are given in data supplement 2 toform are given in data supplement 2 to

the online version of this paper.the online version of this paper.

Findings of the meta-analysisFindings of the meta-analysis

The three minimal intervention studiesThe three minimal intervention studies

including 298 patients were homogeneousincluding 298 patients were homogeneous

((PP¼ 0.76). The pooled odds ratio was 2.80.76). The pooled odds ratio was 2.8

(95% CI 1.6–5.1). We found only one(95% CI 1.6–5.1). We found only one

study that evaluated systematic discon-study that evaluated systematic discon-

tinuation alone using a randomisedtinuation alone using a randomised

controlled design (Oude Voshaarcontrolled design (Oude Voshaar et alet al,,

20032003aa) which showed an odds ratio of 6.1) which showed an odds ratio of 6.1

(95% CI 2.0–18.6). (Further information(95% CI 2.0–18.6). (Further information

is presented in data supplement 2 to theis presented in data supplement 2 to the

online version of this paper.)online version of this paper.)

All psychotherapy augmentationAll psychotherapy augmentation

strategies evaluated the effect ofstrategies evaluated the effect of

cognitive–behavioural therapy. Thesecognitive–behavioural therapy. These

studies appeared to be heterogeneous instudies appeared to be heterogeneous in

outcome values (outcome values (PP550.001), which could0.001), which could

be explained by the cofactors setting,be explained by the cofactors setting,

benzodiazepine dosage, groupbenzodiazepine dosage, group vv. individual. individual

therapy and diagnosis (see Table 2). How-therapy and diagnosis (see Table 2). How-

ever, the studies of Baillargeonever, the studies of Baillargeon et alet al (2003)(2003)

and Morinand Morin et alet al (2004) appeared to be com-(2004) appeared to be com-

parable with respect to all variables evalu-parable with respect to all variables evalu-

ated with the coding form. Both studiesated with the coding form. Both studies

evaluated group cognitive–behaviouralevaluated group cognitive–behavioural

therapy as an augmentation to systematictherapy as an augmentation to systematic

discontinuation alone using a fixed taperdiscontinuation alone using a fixed taper

programme in a psychiatric out-patientprogramme in a psychiatric out-patient

setting among patients using low-dosesetting among patients using low-dose

benzodiazepines for insomnia. Abenzodiazepines for insomnia. A post hocpost hoc

heterogeneity analysis confirmed this find-heterogeneity analysis confirmed this find-

ing (ing (PP¼1.00) and a pooled odds ratio of1.00) and a pooled odds ratio of

5.5 (95% CI 2.3–14.2) was found.5.5 (95% CI 2.3–14.2) was found.

We found five pharmacological aug-We found five pharmacological aug-

mentation strategies with the compoundsmentation strategies with the compounds

propranolol, buspirone, carbamazepine,propranolol, buspirone, carbamazepine,

trazodone and imipramine which weretrazodone and imipramine which were

each evaluated at least twice. Statisticaleach evaluated at least twice. Statistical

homogeneity was found for the studieshomogeneity was found for the studies

evaluating carbamazepine (evaluating carbamazepine (PP¼0.22), trazo-0.22), trazo-

done (done (PP¼ 0.35) and imipramine0.35) and imipramine

((PP¼0.051).0.051). The pooled analysis of studiesThe pooled analysis of studies

evaluating the addition of imipramineevaluating the addition of imipramine

found a significantly higher discontinuationfound a significantly higher discontinuation

success rate (success rate (PP¼0.03); augmentation with0.03); augmentation with

carbamazepine resulted in a higher successcarbamazepine resulted in a higher success

rate of borderline significance (rate of borderline significance (PP¼ 0.06);0.06);

whereas no significant effect was foundwhereas no significant effect was found

for the addition of trazodone (for the addition of trazodone (PP¼ 0.12).0.12).

The studies evaluating augmentation withThe studies evaluating augmentation with

propranolol and buspirone were heteroge-propranolol and buspirone were heteroge-

neous in odds ratios (neous in odds ratios (PP¼ 0.02 and0.02 and

PP¼0.004 respectively). The heterogeneity0.004 respectively). The heterogeneity

in odds ratios of the two studies evaluatingin odds ratios of the two studies evaluating

propranolol was explained by differences inpropranolol was explained by differences in

the steps of the tapering procedure, transferthe steps of the tapering procedure, transfer

to a long-acting benzodiazepine before do-to a long-acting benzodiazepine before do-

sage tapering, baseline benzodiazepine do-sage tapering, baseline benzodiazepine do-

sage, type of benzodiazepine and finallysage, type of benzodiazepine and finally

the diagnosis of included patients. The het-the diagnosis of included patients. The het-

erogeneity in odds ratios of the five studieserogeneity in odds ratios of the five studies

evaluating buspirone was explained by theevaluating buspirone was explained by the

transfer to a long-acting agent, hospitalisa-transfer to a long-acting agent, hospitalisa-

tion, baseline benzodiazepine dosage, typetion, baseline benzodiazepine dosage, type

of benzodiazepine used before tapering,of benzodiazepine used before tapering,

and diagnosis of included patients. Closerand diagnosis of included patients. Closer

inspection did not reveal combinations ofinspection did not reveal combinations of

studies evaluating the addition of buspironestudies evaluating the addition of buspirone

that might be homogeneous. Using athat might be homogeneous. Using a

random-effects model we also did notrandom-effects model we also did not

find significant effects of the additionfind significant effects of the addition

of propranolol (of propranolol (PP¼0.77) and buspirone0.77) and buspirone

((PP¼0.59).0.59).

DIS CUSSIONDIS CUSSION

The main finding of our meta-analysis wasThe main finding of our meta-analysis was

that minimal interventions are effectivethat minimal interventions are effective

strategies for reducing benzodiazepine con-strategies for reducing benzodiazepine con-

sumption, yielding an odds ratio of 2.8 insumption, yielding an odds ratio of 2.8 in

comparison with patients receiving usualcomparison with patients receiving usual

care. More-intensive treatment in the formcare. More-intensive treatment in the form

of systematic discontinuation with or with-of systematic discontinuation with or with-

out therapeutic augmentation was onlyout therapeutic augmentation was only

once compared with usual care (Oudeonce compared with usual care (Oude

VoshaarVoshaar et alet al, 2003, 2003aa), with the finding of), with the finding of

an odds ratio for patients receivingan odds ratio for patients receiving

systematic discontinuation alone of 6.1.systematic discontinuation alone of 6.1.

Although the clinical relevance was limitedAlthough the clinical relevance was limited

by the fact that systematic discontinuationby the fact that systematic discontinuation

alone was evaluated in one study only, thealone was evaluated in one study only, the

62% success rate of systematic discontinua-62% success rate of systematic discontinua-

tion alone in this study was comparabletion alone in this study was comparable

with the median success rate of 58% (rangewith the median success rate of 58% (range

25–100) in the control groups of studies25–100) in the control groups of studies

evaluating systematic discontinuation aug-evaluating systematic discontinuation aug-

mentation strategies which consisted ofmentation strategies which consisted of

systematic discontinuation alone or sys-systematic discontinuation alone or sys-

tematic discontinuation with placebo.tematic discontinuation with placebo.

Moreover, two large and well-designedMoreover, two large and well-designed

(but uncontrolled) studies of benzodiaze-(but uncontrolled) studies of benzodiaze-

pine discontinuation also found disconti-pine discontinuation also found disconti-

nuation success rates of 62% (Rickelsnuation success rates of 62% (Rickels etet

alal, 1990; Schweizer, 1990; Schweizer et alet al, 1990). The three, 1990). The three

minimal intervention studies, as well asminimal intervention studies, as well as

the study by Oude Voshaarthe study by Oude Voshaar et alet al (2003(2003aa),),

were conducted in general practice. There-were conducted in general practice. There-

fore, evidence for treatment of patientsfore, evidence for treatment of patients

referred for help with benzodiazepinereferred for help with benzodiazepine

discontinuation is scarce.discontinuation is scarce.

A total of 17 different augmentationA total of 17 different augmentation

strategies were evaluated. Although thesestrategies were evaluated. Although these

studies were conducted in a variety ofstudies were conducted in a variety of

settings, the age and gender distributionsettings, the age and gender distribution

of patients in the samples suggests selec-of patients in the samples suggests selec-

tive recruitment towards younger, maletive recruitment towards younger, male

patients. Six augmentation strategies werepatients. Six augmentation strategies were

evaluated in at least two studies each; forevaluated in at least two studies each; for

imipramine, carbamazepine and trazodoneimipramine, carbamazepine and trazodone

augmentation the studies were homoge-augmentation the studies were homoge-

neous. Of these three agents, only forneous. Of these three agents, only for

imipramine was a significantly superiorimipramine was a significantly superior

effect on benzodiazepine discontinuationeffect on benzodiazepine discontinuation

success rate found (success rate found (PP ¼ 0.03); the effect of0.03); the effect of

carbamazepine did not reach significancecarbamazepine did not reach significance

((PP¼ 0.06). A0.06). A post hocpost hoc analysis showed thatanalysis showed that

21821 8

Ta b l e 2Ta b l e 2 Demographic characteristics of the population in the selected studiesDemographic characteristics of the population in the selected studies

Interve nt i onInt erve nt ion Stud iesStudies

ParticipantsParticipants Gender ratioGender ratio

M:FM:F

Age, meanAge, mean

(years)(years)

TotalTotal

Withd rewWithd rew

CompletedCompleted

Mi nimal intervent i onMi ni mal inte rvent i on 33601601 7575 526526 1 : 51:5 7171

Sy stemat ic discont in uation aloneS ystem atic discont in uati on alone

11107107 2323 8484 1 : 2.61:2.6 6262

Sy stemati c disconti nuati on with psychotherapySystemati c discontin uati on with psychotherap y

55 357357 4040 317317 1 : 1.41:1.4 5656

Sy stemati c discont inuat ion with pharmacotherapyS ystem atic discont in uati on with pharmacotherapy 2121 13331333 130130 1 1881188 1 : 1.31:1.3 5252

1. The study by OudeVoshaar1. The study by OudeVoshaar et alet al (2003(2003aa ) is included twice because it was a three-condition, controlled study.) is included twice because it was a three-condition, controlled study.

https://doi.org/10.1192/bjp.189.3.213 Published online by Cambridge University Press

DISCONTINUING LONG-TERM BENZODIAZEPINESD I S C ON T INUIN G LONG -T E R M B E NZO DI A Z E P IN E S

group cognitive–behavioural therapy hadgroup cognitive–behavioural therapy had

additive value for patients using low-doseadditive value for patients using low-dose

benzodiazepines (benzodiazepines (5515 mg diazepam15 mg diazepam

equivalent) for insomnia. Finally, theequivalent) for insomnia. Finally, the

following strategies showed significantlyfollowing strategies showed significantly

higher benzodiazepine discontinuation suc-higher benzodiazepine discontinuation suc-

cess rates in single studies: group cognitive–cess rates in single studies: group cognitive–

behavioural therapy for patients with panicbehavioural therapy for patients with panic

disorder, melatonin therapy for patientsdisorder, melatonin therapy for patients

with insomnia, and for long-term benzodia-with insomnia, and for long-term benzodia-

zepine use generally also sodium valproatezepine use generally also sodium valproate

or flumazenil (Ottoor flumazenil (Otto et alet al, 1993; Garfinkel, 1993; Garfinkel

et alet al, 1999; Rickels, 1999; Rickels et alet al, 1999; Gerra, 1999; Gerra

et alet al, 2002)., 2002).

LimitationsLim itations

Large generalisations from our meta-Large generalisations from our meta-

analysis are limited owing to heterogeneityanalysis are limited owing to heterogeneity

of the included studies. We strove toof the included studies. We strove to

explain heterogeneity with variables thatexplain heterogeneity with variables that

have previously been suggested to behave previously been suggested to be

associated with discontinuation outcomeassociated with discontinuation outcome

(Ashton(Ashton et alet al, 1990; Rickels, 1990; Rickels et alet al, 1990,, 1990,

2000; Schweizer2000; Schweizer et alet al, 1990, 1998; Murphy, 1990, 1998; Murphy

& Tyrer, 1991; Oude Voshaar& Tyrer, 1991; Oude Voshaar et alet al,,

20032003bb). However, the current state of). However, the current state of

knowledge precludes any firm conclusionknowledge precludes any firm conclusion

as to the effects of these variables. Inas to the effects of these variables. In

addition, more important variables mightaddition, more important variables might

not have been identified or measured innot have been identified or measured in

the included studies, such as a clear DSM–the included studies, such as a clear DSM–

IV Axis I diagnosis (American PsychiatricIV Axis I diagnosis (American Psychiatric

Assocation, 1994) or personality character-Assocation, 1994) or personality character-

istics. For example, in a relatively large,istics. For example, in a relatively large,

uncontrolled study (uncontrolled study (nn¼ 165) personality165) personality

factors were found to explain 24% offactors were found to explain 24% of

the variance in discontinuation outcomethe variance in discontinuation outcome

(Schweizer(Schweizer et alet al, 1998)., 1998).

Clinical implicationsClinical implications

Although establishing the efficacy ofAlthough establishing the efficacy of

individual treatment strategies is clinicallyindividual treatment strategies is clinically

relevant, stepped care approaches are evenrelevant, stepped care approaches are even

more important for treatment planning inmore important for treatment planning in

the case of treatment-resistant benzodiaze-the case of treatment-resistant benzodiaze-

pine dependence. This meta-analysis waspine dependence. This meta-analysis was

conducted in order to establish the clinicalconducted in order to establish the clinical

evidence for the individual steps in aevidence for the individual steps in a

stepped care approach in order to dis-stepped care approach in order to dis-

continue long-term benzodiazepine use.continue long-term benzodiazepine use.

Following the stepped care approachFollowing the stepped care approach

proposed by Russell & Lader (1993), weproposed by Russell & Lader (1993), we

now know that use of the first twonow know that use of the first two

steps – namely starting with a minimalsteps – namely starting with a minimal

intervention strategy, followed by systema-intervention strategy, followed by systema-

tic discontinuation alone for cases resistanttic discontinuation alone for cases resistant

to treatment in primary care – is supportedto treatment in primary care – is supported

by the results of randomised controlledby the results of randomised controlled

trials. With respect to this statement, ittrials. With respect to this statement, it

has to be mentioned that the single studyhas to be mentioned that the single study

evaluating systematic discontinuation aloneevaluating systematic discontinuation alone

was conducted among people with long-was conducted among people with long-

term benzodiazepine use who did notterm benzodiazepine use who did not

respond to a minimal intervention strategyrespond to a minimal intervention strategy

(Oude Voshaar(Oude Voshaar et alet al, 2003, 2003aa; Gorgels; Gorgels et alet al,,

2005). However, much research has still2005). However, much research has still

to be conducted in this field; for example,to be conducted in this field; for example,

we do not know which variables and treat-we do not know which variables and treat-

ment characteristics are associated with ament characteristics are associated with a

favourable outcome. The taper schedulesfavourable outcome. The taper schedules

described in published studies vary fromdescribed in published studies vary from

abrupt discontinuation (Rickelsabrupt discontinuation (Rickels et alet al,,

1990), to 25% weekly reduction of dosage1990), to 25% weekly reduction of dosage

(Schweizer(Schweizer et alet al, 1990; Oude Voshaar, 1990; Oude Voshaar et alet al,,

20032003aa), discontinuation in steps of about), discontinuation in steps of about

one-eighth of the daily dose every 2 weeksone-eighth of the daily dose every 2 weeks

(Russell & Lader, 1993) to, finally,(Russell & Lader, 1993) to, finally,

symptom-guided withdrawal with the timesymptom-guided withdrawal with the time

needed for withdrawal varying from aboutneeded for withdrawal varying from about

4 weeks to a year or more (Ashton,4 weeks to a year or more (Ashton,

1987). However, different taper schedules1987). However, different taper schedules

have never been directly compared in ahave never been directly compared in a

randomised controlled study. We also dorandomised controlled study. We also do

not know which strategy should benot know which strategy should be

followed if the first two steps fail. Althoughfollowed if the first two steps fail. Although

augmentation was not evaluated amongaugmentation was not evaluated among

patients who failed to discontinue theirpatients who failed to discontinue their

benzodiazepine use by systematic disconti-benzodiazepine use by systematic disconti-

nuation alone, our meta-analysis found anuation alone, our meta-analysis found a

higher discontinuation success ratehigher discontinuation success rate

after the addition of imipramine andafter the addition of imipramine and

carbamazepine in general, or groupcarbamazepine in general, or group

cognitive–behavioural therapy for patientscognitive–behavioural therapy for patients

with insomnia. Moreover, adding sodiumwith insomnia. Moreover, adding sodium

valproate or flumazenil and addingvalproate or flumazenil and adding

melatonin or group cognitive–melatonin or group cognitive–

behavioural therapy in specific patientbehavioural therapy in specific patient

groups (e.g. those with panic disorder) cangroups (e.g. those with panic disorder) can

be an option. It should be noted that thesebe an option. It should be noted that these

suggestions are based on small, singlesuggestions are based on small, single

studies (patient numbersstudies (patient numbers nn¼27 to27 to

nn¼40). Future research should40). Future research should

evaluate more rigorously stepped careevaluate more rigorously stepped care

programmes and promising augmentationprogrammes and promising augmentation

strategies.strategies.

REFE RENCE SREF ERENCE S

American Psychiatric Assocation (1994)American Psychiatric Assocation (199 4) DiagnosticDi agnostic

and Statistical Manual of Mental Disordersand Statistical Manual of Mental Disorders (DSM^ IV).(DSM^IV).

Washington, DC: APA.Washington, DC: APA.

Ashton, H . (1987)Ashton, H. ( 1987) Benzodiazepine withdrawal:Benzodiazepine wi thdrawal:

outcome in 50 patients.outcome in 50 patients. British Journal of AddictionBritish Journal of Addiction,, 8282,,

665^671.665^671.

Ashton, C. H., Rawl ins, M. D. & Tyrer, S. P. (199 0)Ashton, C. H., Rawlins, M. D. & Tyrer, S. P. (1990) AA

double-blind placebo-controlled study of buspirone indoubl e-blind placebo-contro lled stud y of buspirone in

diazepam wi thdrawal in chronic benzodiazepine users.diazepam withdrawal in chronic benzodiazepine users.

British Journal of PsychiatryBritish Journal of Psychiatry,, 157157,232^238., 232^238.

Baillargeon, L., Landreville, P., Verreault, R.,Baillargeon, L., Landreville, P., Verreault, R., et alet al

(2003)(2003)

D i sc ont inuation of benzodi azep i ne among olde rD iscontinua tion of benzodi azep ine among older

insomniac adults treated with cognitive ^ behaviouralinsomniac adults treated with cognitive ^ behavioural

the rap y comb ined w ith gradual tapering: a random i zedtherap y comb ined with gradua l taperi ng: a random i z ed

trial.trial. Canadian Medical Association JournalCanadian Medical Association Journal,, 169169,,

101 5^1020.1015^1020.

Bashir, K., King, M. & Ashworth, M. (1994)Bashir, K., King, M. & Ashworth, M. (1994)

Contro l l ed evaluation of br ief in tervention by generalCont r o lled evaluation of bri ef int e rvention by general

practi tioners to reduce chronic use of benzodiazepines.practi tione rs t o reduc e chronic use of benzodi azep ines.

British Journal of General PracticeBritish Journal of General Practice,, 4444, 408^412., 408^412.

Cantopher, T., Olivieri, S., Cleave, N.,Cantopher, T., Olivieri, S., Cleave, N., et alet al (19 9 0)(19 9 0)

Chr oni c benzodiazepine dependence. A comparati veChroni c benzodiazepine dependence. A comparative

study of abrupt withdrawal under propranolol coverst u d y of abrupt withd ra wa l under propranol ol cove r

versu s gradual withd ra wa l.versus gradua l withd ra wal . British Journal of PsychiatryBritish Journal of Psychiatry,,

156156, 406^411., 4 0 6^411.

Cialdella,P.,Boissel,J.P.&Belon,P.(2001)Cialdella, P., Boissel, J. P. & Belon, P. (2001)

Homeopathic specialties as substitutes forHomeopathic specialties as substitutes for

benzodiazepines: double -blindbenzodiazepines: double-blind vv.placebostudy..placebostudy.

Therapi eThe

rapie,, 5656, 397^402.,397^402.

Cormack, M. A., Sweeney, K. G., Hughes-Jones, H.,Cormack, M. A., Sweeney, K. G., Hughes-Jones, H.,

et alet al (1994)(19 94)

Evaluation of an easy, cost-effective strategyEvaluation of an easy, cost-effective strategy

for cutting benzodiazepine use in general practice.for cutting benzodiazepine use in general practice. Br it ishBri ti sh

Journal of General PracticeJournal of General Practice,, 4444, 5^8.,5^8.

Di Costanzo, E. & Rovea, A. (1992)Di Costanzo, E. & Rovea, A. (1992) [The prophy lax i s[The prophy l ax is

of benzodiazepine withdrawal syndrome in the elderly:of benzodiazepine withdrawal syndrome in the elderly:

the effectiveness of carbamazepine. Double-blind studythe effectiveness of carbamazepine. Double-blind study

vv.placebo](inItalian)..placebo](inItalian).Minerva PsychiatricaMinerva Psychiatrica,, 3333,301^304.,301^304.

Garcia-Borreguero, D., Bronisch, T., Apelt, S. ,Garcia-Borreguero, D., Bronisch, T ., Apelt, S., et alet al

(19 91)(19 91)

Treatment of benzodiazepine withdrawalTreatment of benzodia zepine withdrawal

symptoms with carbamazepine.symptoms with carbamazepine. Eur opean Ar chi v es ofEuropean Archives of

Psychiatry and Clinical NeurosciencePsychiatry and Clinical Neuroscience,, 241241, 145^ 15 0.,14 5^15 0.

Garfinkel, D., Zisapel, N., Wainstein, J.,Garfinkel, D., Zisapel, N., Wainstein, J., et alet al (19 9 9 )(1999)

Fac ilitat i on of benzodi azep ine dis c ontinua tion byFac ili ta tion of benzodiazepine di s c ontinua tion by

mela t onin: a new cl i n i cal approach.mela t on i n: a new clinical approach. Archives of InternalArchives of Internal

MedicineMedicine,, 159159 , 2456^2460., 2456^24 60.

Gerra , G., Zaimovic, A., Giusti, F .,Gerra, G., Zaimovi c, A., Giusti, F ., et alet al (2002)(2002)

Intravenous flumazenil versus oxazepam tapering in theIntravenousflumazenilversusoxazepamtaperinginthe

tr eat m ent of benzodiazep ine withd raw al: a randomized,treatmen t of benzodiazepine wi t hdrawa l: a randomi zed,

placebo -control l ed study.placebo -controlled study. Addi ction BiologyAddi ction Biology,, 77, 385^395., 385^395.

Gorgels, W. J. M. J., Oude Voshaar, R. C., Mol, A. J. J.,Gorgels,W.J.M.J.,OudeVoshaar,R.C.,Mol,A.J.J.,

et alet al (2 005)(2005)

D i s c ont inuation of long-t e rmD i s c ontinua tion of long-t e r m

219219

R.C.OUDE VOSHAAR, MD, PhD, Department of Psychiatry, Radboud University Nijmegen Medical Center,R. C. OU D E VOS HAAR, MD, PhD , Department of Psychiatry, Radboud U ni versi ty Nijmegen Medical Cent er,

Nijmegen; J. E.COUVEE, PhD,GlaxoSmithKline, Zeist; A. J. L. M.VAN BALKOM, MD, PhD, Department ofNijmegen; J. E. COUVE

E, Ph D , GlaxoSmithKline, Zeist; A. J. L. M. VAN BALKOM, MD, PhD, Depar tment of

Psychiatry and Institute for Research and Extramural Medicine, Free University Medical Centre, Amsterdam;Psychiatry and Institute for Research and Extramural Medicine, Free University Medical Centre, Amsterdam;

P.G. H. MULDER, PhD, Depar tment of Epidemiology and Biostatistics, Erasmus Medical Center, Rotterdam;P.G. H. MULDER, PhD, Depar tment of Epidemiology and Biostatistics, Erasmus Medical Center, Rotterdam;

F.G. ZITMAN, MD, PhD, Department of Psychiatry, Leiden University Medical Centre,Leiden,The NetherlandsF.G. ZITMAN, MD, PhD, Department of Psychiatry, Leiden University Medical Centre, Leiden,The Netherlands

Correspondence: Dr J. E.Couvee, Head Clinical Development CNS, Anti-Infectives and Oncology,Correspondence : Dr J. E.Couve

e, Head Clinical Development CNS, Anti-Infectives and Oncology,

GlaxoSmith Kline, Medical Department, Huis ter Heideweg 62, 3705 LZ Zeist,The Netherlands.GlaxoSmithKline, Medical Department, Huis ter Heideweg 62, 3705 LZ Zeist,The Netherlands.

Tel: +31 30 6938145; fax: +31 30 6938459; email: jaap.e.couveeTel: +3130 6938145; fax: +3130 69 38459; email: jaap.e.couvee@@gsk.comgsk.com

(First received 14 July 20 03, f inal revision 4 March 20 05, accepted 28 July 20 05)(First received 1 4 July 20 03, f inal revision 4 March 20 05, accepted 28 July 20 05)

https://doi.org/10.1192/bjp.189.3.213 Published online by Cambridge University Press

OUD E VO S HA A R E T A LOUD E VO S HA A R E T A L

benzodiazepine use by sending a letter to users in familybenzodiazepine use by sending a letter to users in family

practice: a prospectiv e cont ro lled in t erven t ion study.practice: a prospective controlled intervention study.

Drug and Alcohol DependenceDr u g and Al cohol Dependence,, 7878,49^56., 49^56.

Hantouche, E. G., Guelf i, J. D. & Comet, D. (1998)Hantouche, E. G., Guelfi, J. D. & Comet, D. (1998)

[Alp h a - b et a[Alph a - b et a LL-aspartat e magnesium in treatmen t of-aspartate magnesium in treatment of

chronic benzodiazepine misuse: controlled and double -chronic benzodiazepine misuse: controlled and double-

bl ind study versus placebo ] (in French).blind study versus placebo] (in French). EncephaleEnce

phale,, 2424,,

469^479.469^479.

Jones, D. (1990)Jones, D. (1990) Weaning elderly patients offWeaning elderly patients off

psy cho trop i c drugs in gene ra l practice: a randomi sedpsy cho trop i c drugs i n gene ral practice: a randomised

control led trial .controlled trial . HealthTrendsHealthTrends,, 2222, 164^166.,164^166.

Lader, M. (1 991)Lader, M. (19 91) History of benzo diazepineHistory of benzodiazepine

dependenc e .dependence . Journal of Substance Abuse and TreatmentJournal of Substance A buse and Treatment,, 88,,

53^59.53^59.

Lader, M. & Olaji de, D. ( 1987)Lader, M. & Olajide, D. (1987) AcomparisonofAcomparisonof

buspirone and placebo in relieving benzodiazepinebuspirone and placebo in reliev ing benzodiazepine

withdrawal symptoms.withdrawal symptoms. Journal of ClinicalJournal of Clinical

PsychopharmacologyPsychopharmacology,, 77, 11^15.,11^15.

Lader, M., Farr, I. & Morton, S. (1993)Lader, M., Farr, I. & Morton, S. (1993) AcomparisonAcomparison

of alpi dem and placebo in relievi ng benzodiazepineof al pidem and placebo in re l i evi ng benzodiazepine

withdrawal symptoms.withdrawal symptoms. International ClinicalInternational Clinical

PsychopharmacologyPsychopharmacology,, 88, 31^36.,31^36.

Lemoine, P., Touchon, J. & Billardon, M. (1997)Lemoine, P., Touchon, J. & Billardon, M. (1997)

[ C ompari son of 6 di f fe r e nt methods for lorazepam[Comparison of 6 different methods for lorazepam

withdrawal. A control led study, hydro xyzine versusw ithdrawal. A controlled study, hydroxyzine versus

placebo] (in French).placebo] (in French). EncephaleEnce

phale,, 2323, 290^29 9., 29 0^299.

Morin,C.M.,Bastien,C.,Guay,B.,Morin,C.M.,Bastien,C.,Guay,B.,et alet al (2004)(2004)

Randomized cli nical trial of supervised tapering andRandomized clinical trial of supervised tapering and

cognitive behavior therapy to facilitate benzodiazepinecognitive behavior therapy to facilitate benzodiazepine

discontinuation in older adults with chronic insomnia.d i s c ont inuation in olde r adults with chr oni c insomn i a.

American Journal of PsychiatryAmerican Journal of Psychiatry,, 161161, 332^342., 332^342.

Murphy, S. M. & Tyrer , P. (1991)Murphy, S. M. & Tyrer, P. ( 1991 ) A double-blindAdouble-blind

comparison of the effects of gradual withdrawal ofcomparison of the effects of gradual withdrawal of

lorazepam, diazepam and br omazepam inlorazepam, diazepam and bromazepam in

benzod i azepine dependence .benzodi azepine dependence. British Journal of PsychiatryBr itish Journal of Psychiatry,,

158158, 511 ^516.,511^516.

Otto, M.W., Pollack, M. H., Sachs, G. S.,Otto, M.W., Pollack, M. H., Sachs, G. S., et alet al (19 9 3)(19 93 )

Discont inuation of benzodiazepine treatment: eff icacyDiscontinuation of benzodiazepine treatment: eff icacy

of cogn i ti ve ^ behav i oral therapy for pati en ts with pani cof cognitive ^ behavioral therapy for patients with panic

disorder .disorder . American Journal of PsychiatryAmerican Journal of Psychiatry,, 150150, 14 85^149 0.,1485^1490.

Otto, M.W., Pollack, M. H., Gould, R. A.,Otto, M. W., Pollack, M. H., Gould, R. A., et alet al (2000)(2000)

A comparison of the efficacy of clonazepam andA comparison of the efficacy of clonazepam and

cogni ti ve ^ behavioral group therapy for the treatmentcognitive ^ behavioral group therapy for the treatment

of socia l phobia.of socia l phobi a. Journal of Anxiety DisordersJournal of Anxiety Disorders,, 1414, 345^358., 345^358.

OudeVoshaar,R.C.,Gorgels,W.J.M.J.,Mol,A.J.J.,OudeVoshaar,R.C.,Gorgels,W.J.M.J.,Mol,A.J.J.,

et alet al (2003(2003aa))

Tapering off long-term benzodiazepine useTaperi ng off long-te rm benzodi azepine use

with or without group cognit ive ^ behavioural therapy:with or without group cognitive^ behavioural therapy:

th r ee-condition , randomised contr olled tr i a l.th ree-condition, randomised control led trial. Br itishBr it ish

Jour nal of Psychi atryJournal of Psychiatry,, 182182, 498^5 0 4., 498^50 4.

Oude Voshaar, R. C., Mol, A. J. J., Gorgels, W. J. M. J.,Oude Voshaar, R. C., Mol, A. J. J., Gorgels, W. J. M. J.,

et alet al (2003(2003bb))

Cross-validation, time course andCr oss-valida tion , time cou rse and

predi cti ve val id i ty of the Benzodiazepine Dependencepredi cti ve val id ity of the Benzodiazepi ne Dependence

Self-Report Questionnaire in a benzodiazepineSelf-Repor t Questionnaire in a benzodiazepine

discontinuation trial.discontinuation trial . Comprehensive PsychiatryComprehensive Psychiatry,, 4444,,

247^255.247^255.

Petrovic, M., Pevernagic, D.,Van den Noortgate, N.,Petrovic, M., Pevernagic, D., V an den N oortgate, N.,

et alet al (1999)(19 9 9)

A programme for short-term withdrawalA program me for short -term w ithd ra wa l

fr om benzodiazep ines in geriatr i c hospital inpat i ents:from benzodiazepines in ge r iat r ic hospi ta l i n patients:

success rate and effect on subjective sleep quality.success rate and effect on subjective sleep quality.

International Journal of Geriatric PsychiatryInternational Journal of Geriatric Psychiatry,, 1414, 754^76 0., 754^76 0.

Rickels, K., Schweizer, E., Case,W. G.,Rickels, K., Schweizer, E., Case,W. G., et alet al (19 9 0 )(19 9 0)

Long-term therapeutic use of benzodiazepines. I. EffectsLong-term therapeutic use of benzodiazepines. I. Effects

of abrupt discontinuation.of abrupt discontinuation. Ar chives of Genera l Psychi atryArchives of General Psychiatry,,

4747,899^907., 89 9^9 07.

Rickels, K., Schweizer, E., Garcia, E. F.,Rickels, K., Schweizer, E., Garcia, E. F., et alet al (19 9 9)(1999)

Trazodone and valproate in patients discontinuing long-Trazodone and va lproat e in pat i e nts dis c ont inu ing long-

term benzodiazepine therapy: effects on withdrawalterm benzodiazepine therapy: effects on withdrawal

sympt oms and taper outcome.symptoms and taper outcome. PsychopharmacologyPsychopharmacology

(Berlin)(Berlin),, 141141 ,1^5., 1^5.

Rickels, K., DeMartinis, N., Garcia-Espana, F.,Rickels, K., DeMartinis, N., Garcia-Espana, F., et alet al

(2000)(2000)

Imipramine and buspirone in treatment ofImipramine and buspirone in treatment of

patients with generalized anxiety disorder who arepatien ts wi th general ized anxiety di sorder who are

d i sc ont inui ng long-ter m benzodi azepine therapy.disc ontinui ng long-ter m benzodiazepine therapy.

American Journal of PsychiatryAmerican Journal of Psychiatry,, 157157 ,1973^1979., 1973^1979.

Russel,V. J. & Lader, M. H. (eds) (1993)Russel,V. J. & Lader, M. H. (eds) (1993) Gui delines f orGuidelines for

the Prevention and Treatment of Benzodiazepinethe Prevention and Treatment of Benzodiazepine

DependenceDependence. London: Mental Health Foundation .. L ondon: Mental Healt h Founda tion .

Rynn, M., Garcia-Espana, F., Greenblatt, D. J.,Rynn, M., Garcia-Espana, F., Greenblatt, D. J., et alet al

(2003)(2003)

Imipramine and buspirone in patients with panicImipramine and buspirone in patients with panic

disorder who are discontinuing long-termdisorder who are discontinuing long-term

benzodiazepine therapy.benzodiazepine therapy. Journal of ClinicalJournal of Clinical

PsychopharmacologyPsychopharmacology,, 2323, 505^508., 505^508.

Schweizer, E., Rickels, K., Case, W. G.,Schweizer, E., Rickels, K., Case, W. G., et alet al (19 9 0)(19 9 0 )

Long-term therapeutic use of benzodiazepines. II.Long-term therapeutic use of benzodiazepines. II.

Effects of gradual taper.Effects of gradual taper. Archives of General PsychiatryArchives of General Psychiatry,,

4747,908^915., 90 8^915.

Schweizer, E., Rickels, K., Case,W. G.,Schweizer, E., Rickels, K., Case, W. G., et alet al (19 91)(19 91)

Carbamazepine treatment in patients discontinuing long-Carbamazepine treatment in patients discontinuing long-

term benzodiazepine therapy. Effects on withdrawalterm benzodiazepine therapy. Effects on withdrawal

severity and outcome.severity and out come. Archives of General PsychiatryArchives of General Psychiatry,, 4848,,

4 4 8^452.44 8^452.

Schweizer, E., Rickels, K., Weiss, S.,Schweizer, E., Rickels, K., Weiss, S., et alet al (19 9 3)(19 9 3)

Maintenance drug treatment of panic disorder. I. ResultsMai ntenance drug treatmen t of pan i c disor der. I . Resu l ts

of a prospective, placebo -control l ed comparison ofof a prospective, placebo - con trolled comparison of

alprazolam and imipramine.alprazolam and imipramine. Archives of GeneralArchives of General

PsychiatryPsychiatry,, 5050, 51^60.,51^60.

Schweizer, E., Case,W. G., Garcia-Espana, F.,Schweizer, E., Case,W. G., Garcia-Espana, F., et alet al

(19 95)(19 9 5)

Progesterone co -admin istration in patientsProgesterone co-administration in patients

discontinuing long-term b enzodiazepine therapy: effectsdiscontinuing long-term benzodiazepine therapy: effects

on wi thdrawa l severi ty and taper out c ome.on withdrawal severity and taper outcome.

Psychopharmacology (Berlin)Psychopharmacology (Berlin),, 117117, 424^429., 424^429.

Schweizer, E., Rickels, K., De Martinis, N .,Schweizer, E., Rickels, K., De Martinis, N., et alet al

(19 9 8)(19 9 8)

The effect of personality on withdrawal severityThe effect of personality on withdrawal severity

and taper out c om e in benzodiazep ine dependentand taper outcome in benzodiazep ine dependen t

patients.patients. Psychologica l MedicinePsychologi cal Medici ne,, 2828, 713^720., 713^720.

Stillwell, G. & Fountain, J. (2002)Stillwell, G. & Fountain, J. (2002) BenzodiazepineBenzodiazepine

Use ^ A Report of a Survey of BenzodiazepineUse ^ A Report of a Survey of Benzodiazepine

Consumption in the Member Countries of the PompidouConsumption in the Member Countries of the Pompidou

GroupGroup. Geneva: World Healt h Org an i zati on.. Geneva: Wor ld Health Organ ization.

Taylor, S., McCracken, C. F.,Wilson, K. C.,Taylor, S., McCracken,C. F.,Wilson, K. C., et alet al (19 9 8)(19 9 8)

Exten t and appropriat eness of benzodiazepine use.Extent and appropriateness of benzodiazepine use.

Results from an elderly urban community.Results from an elderly urban community. Br i ti sh Journa lBri ti sh Journal

of Psychiatryof Psychiatry,, 173173, 433^438., 433^438.

Tyrer, P., Rutherford, D. & Huggett, T.Tyrer, P., Rutherford, D. & Huggett, T. (1981)(19 81)

Benzodiazepine withdrawal sympt oms and propranolol.Benzodiazepine withdrawal symptoms and propranolol.

LancetLanc e t,, ii, 520^52 2., 520^522.

Tyrer, P., Ferguson, B., Hallstrom, C.,Tyrer, P., Ferguson, B., Hallstrom, C., et alet al (19 9 6)(19 9 6) AA

control led trial of dothiepin and placebo in treatingcontrolled trial of dothiepin and placebo in treating

benzodiazepine withdrawal symptoms.benzodiazepine withdrawal symptoms. Bri t i sh Journa l ofBr it ish Journal of

PsychiatryPsychiatry,, 168168, 457^4 61., 457^4 61.

Udelman, H. D. & Udelman, D. L. (1990)Udelman, H. D. & Udelman, D. L. (19 9 0) Concurren tConcurrent

use of buspirone in anxious patients during withdrawaluse of buspirone in anxious patients during withdrawal

from alprazolam therapy.from alprazolam therapy. Journal of Clinical PsychiatryJournal of Clinical Psychiatry,,

5151 (suppl.), 46^5 0.(suppl .), 46^50.

Van Boeijen, C. A.,Van Balkom, A. J. L. M.,VanVan Boeijen, C. A.,Van Balkom, A. J. L. M.,Van

Oppen, P .,Opp en, P ., et alet al (2005)(2005)

Efficacy of self-help manuals forEfficacy of self-help manuals for

anxiety disorders in primary care, a review.anxiety disorders in primary care, a review. Fami lyFami ly

PracticePr actice,, 2222, 192^196.,192^196.

VanTulder, M.W., Assendelft,W. J. J., Koes, B.W.,VanTulder, M.W., Assendelft,W. J. J., Koes, B.W., et alet al

(19 9 7)(19 9 7)

Methodologic guidelines for systematic reviews inMethodologic guidel ines for systematic reviews in

the Cochrane Collaboration Back Review Group forthe Cochrane Collaboration Back Review Group for

Spi nal D isor ders.Spinal D isor ders. Spi neSp i ne,, 2222, 2323^2330., 2323^2330.

Vorma, H., Naukkarinen, H., Sarna, S.,Vorma, H., Naukkarinen, H., Sarna, S ., et alet al (2002)(2002)

Treatment of out-patients with complicatedTr e a t m e n t o f o u t - p a t i e n t s w i t h c o m p l i c a t e d

benzod i azepine dependence: compar i son of twobenzodiazepine dependence: compar i son of two

approaches.approaches. AddictionAddiction,, 9797, 851^859.,851^859.

Zandstra, S. M., Fuhrer, J.W.,Van de Lisdonk, E. H.,Zandstra, S. M., Fu

hrer, J.W., Van de Lisdonk, E. H.,

et alet al (2002)(2002)

Different study crit eria affect the prevalenceDi fferent study crit eri a affect the prevalence

of benzodiazepine use.of benzodiazepine use. Social Psychiatry and PsychiatricSocial Psychiatry and Psychiatric

EpidemiologyEpidemio logy,, 3737, 139^144., 139^144.

Zandstra, S. M.,Van Rijswijk, E., Rijnders,C. A.,Zandstra, S. M.,Van Rijswijk, E., Rijnders, C. A., et alet al

(2004)(2004)

Long-term benzodiazepine users in familyLong-term benzodiazepine users in fami ly

practice: differences from short-te rm users in mentalpracti ce: diffe renc es from short-term users in mental

healt h, coping behaviou r and psy cho l og i ca lhealth, coping behaviour and psychological

characteristics.characteristics. Famil y PracticeFamil y Practice,, 2121 , 266^269., 266^2 69.

Zitman, F. G. & Couvee , J. E. (2001)Zitman, F. G. & Couve

e, J. E. (2001) Chroni cChronic

benzodi azepi ne use in general practi ce pat ien ts w i thbenzodiazepine use in general practice patients with

depression: an evaluation of contr olled treat me nt anddepress i on: an eva luation of contr o lled t r eat me nt and

taper-off: report on behalf of the Dutch Chronictaper-off: report on behalf of the Dutch Chronic

Benzodiazepine W orki ng G roup.Benzodiazepine Working Group. Bri ti sh Journal ofBri t ish Journa l of

PsychiatryPsychiatry,, 178178, 317 ^324., 317 ^324.

220220

https://doi.org/10.1192/bjp.189.3.213 Published online by Cambridge University Press