First-generation antipsychotic long-acting injections v. oral antipsychotics in schizophrenia: systematic review of randomised controlled trials and observational studies

First-generation antipsychotic long-acting injections v. oral antipsychotics in schizophrenia: systematic review of randomised controlled trials and observational studies
of 10
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  10.1192/bjp.195.52.s20Access the most recent version at DOI: 2009, 195:S20-S28. BJP Peter M. Haddad, Mark Taylor and Omair S. Niaz randomised controlled trials and observational studiesantipsychotics in schizophrenia: systematic review of . oral v  First-generation antipsychotic long-acting injections References article cites 48 articles, 10 of which you can access for free at: permissionsReprints/ write to To obtain reprints or permission to reproduce material from this paper, please to this article atYou can respond;195/52/S20 from Downloaded The Royal College of PsychiatristsPublished by on November 9, 2014  go to: The British Journal of Psych iatry  To subscribe to  First-generation antipsychotic long-acting injections (FGA–LAIs)were introduced in the 1960s, 1 and continue to be widely usedtoday in both the USA 2 and the UK 3,4 for the maintenancetreatment of schizophrenia. In meta-analyses antipsychotics aresuperior to placebo in reducing relapse in schizophrenia, 5 andrandomised studies have shown that continuous maintenancemedication is associated with lower relapse rates than intermittenttargeted medication given only when there are early warning signsof a possible relapse. 6,7 In practice the effectiveness of maintenanceantipsychotic treatment is often undermined by poor adherence,with Cramer & Rosenheck estimating a medication adherence ratein schizophrenia of 58%. 8 Stopping antipsychotic medication is acommon cause of relapse, 9 and even a 10-day period of missedmedication has been associated with an increased risk of readmission due to relapse. 10 Partial adherence may lead to poorsymptom control irrespective of an increased risk of relapse, so by improving medication adherence LAIs may reduce relapse andimprove symptom control. The regular contact with nursing staff that accompanies LAI treatment may have further benefits. Animportant proviso is that this argument assumes that those whoadhere poorly to a regimen of tablets will accept an injection.Some patients will not, and so LAIs are not a panacea foradherence problems nor are they the only strategy by which toimprove adherence.In summary, there are intuitive reasons why LAIs may improve clinical outcomes but the key issue is whether evidencesupports this. In this article we systematically review studies thatcompare the effectiveness of FGA–LAIs with both first- andsecond-generation antipsychotic oral medication in schizophrenia.First-generation antipsychotic LAIs are used in other disorders,including bipolar disorder, 11 but this is less frequent and outsidethe remit of this review. We have reviewed both randomisedcontrolled studies and observational studies as they assess differentoutcomes. Randomised controlled trials (RCTs) assess efficacy, i.e.does a drug lead to benefit in ideal circumstances. Observationalstudies assess effectiveness, i.e. does a drug have benefit in the realworld where dose, patient characteristics and follow-up may bemore variable than in an RCT. MethodSearch strategy A systematic search of the MEDLINE, EMBASE and PsycINFOdatabases was conducted in September 2008 using the termsantipsychotic, depot and long-acting injection, and mapping toMeSH terms, with the limits humans, adults, clinical trial, meta-analysis, randomised controlled trial, comparative study, English.There were no limitations by date. Abstracts of articles werereviewed against our inclusion and exclusion criteria. Thereferences cited by included studies were reviewed for additionalrelevant cited articles, and the citation search facility wasemployed to identify further potentially relevant srcinal studies. Inclusion and exclusion criteria For inclusion, studies were required to:(a) include a group of patients treated with an FGA–LAI; s20 First-generation antipsychotic long-actinginjections  v.  oral antipsychotics in schizophrenia:systematic review of randomised controlled trialsand observational studies Peter M. Haddad, Mark Taylor and Omair S. Niaz Background Antipsychotic long-acting injections (LAIs) are often used inan attempt to improve medication adherence in people withschizophrenia. Aims To compare first-generation antipsychotic long-actinginjections (FGA–LAIs) with first- and second-generation oralantipsychotics in terms of clinical outcome. Method Systematic literature review. Results A meta-analysis of randomised controlled trials (RCTs)showed no difference in relapse or tolerability between oralantipsychotics and FGA–LAIs but global improvement wastwice as likely with FGA–LAIs. Four prospective observationalstudies were identified; two studies reported lowerdiscontinuation rates for FGA–LAIs compared with oralmedication and two found that outcome was either nodifferent or better with oral antipsychotics. Mirror-imagestudies consistently showed reduced in-patient days andadmissions following a switch from oral antipsychotics toFGA–LAIs. Conclusions The results are variable and inconclusive. Some evidencesuggests that FGA–LAIs may improve outcome comparedwith oral antipsychotics. Methodological issues may partlyexplain the variable results. Selective recruitment in RCTsand lack of randomisation in observational studies are biasesagainst LAIs, whereas regression to the mean in mirror-image studies favours LAIs. In terms of future research, along-term pragmatic RCT of an FGA–LAI against an oralantipsychotic, in patients with problematic adherence, wouldbe of value. Declaration of interest P.M.H. and M.T. have received fees for lecturing andconsultancy from the manufacturers of variousantipsychotics, including AstraZeneca, Bristol-Myers Squibb,Eli Lilly and Janssen-Cilag. The British Journal of Psychiatry  (2009)195, s20–s28. doi: 10.1192/bjp.195.52.s20 Review article  (b) include an oral antipsychotic comparator (first- or second-generation);(c) provide srcinal quantitative data on efficacy or effectiveness;(d) (for RCTs and prospective observational studies) be restrictedto those that recruited patients with schizophrenia, schizo-affective disorder or schizophreniform disorders.The last inclusion criterion was not applied to retrospectivestudies as these frequently reported on the outcome of a cohortof patients treated with LAIs irrespective of diagnosis. No specificquality threshold was set for inclusion of studies. Studies wereexcluded if there were fewer than 20 patients in the LAI arm,if no srcinal patient data were reported (e.g. ‘modelling’ studies)or if the comparator group was given a placebo, another FGA–LAIor risperidone LAI. Statistical analysis Included studies were divided into four groups: RCTs, prospectiveobservational studies, mirror-image studies and other retro-spective observational studies. Quantitative data were extracted.In some mirror-image studies admission and in-patient datawere presented only in graph form in the srcinal articles and/or  P- values were not given. Where possible we have extrapolatedthe missing data and calculated  P  -values using data from theoriginal publications. The summary table for mirror-imagestudies (see Table 2) indicates where secondary calculations havebeen made. No further statistical analysis was applied. ResultsSearch strategy The initial search strategy revealed 249 potentially relevant study abstracts, which were individually scrutinised against the inclusioncriteria. Seven further possible studies were identified throughcitation search. After inclusion and exclusion criteria were applied,the remaining studies were categorised as RCTs (1 meta-analysisthat considered FGA–LAIs as a total group and 1 RCT);prospective observational studies (4 studies); mirror-image studies(11 studies); other retrospective observational studies (2 studies).The one meta-analysis of FGA–LAIs  v  . oral medication that weidentified was part of a comprehensive systematic meta-review of LAIs by Adams  et al  . 12 This review was based on a synthesis of data from eight Cochrane reviews of individual FGA–LAIs inpatients with schizophrenia or schizophrenia-like illnesses. Sincethe Adams review was published, five of the Cochrane FGA–LAIreviews on which it was based have been updated. 13–17 Theseupdates either contain no data comparing LAIs with oralmedication or show no significant difference in efficacy betweenoral and LAI. Consequently the updated Cochrane reviews giveno reason to doubt a key result of the meta-analysis by Adams et al  , namely that relapse rates do not differ between LAI and oralmedication. 12 In view of this the individual updated Cochranereviews are not detailed further in this paper. Randomised controlled trials The meta-analysis by Adams  et al   of FGA–LAIs  v  . oral anti-psychotics provided data on several outcomes, including relapse(Fig. 1). 12 The relapse data are based on a total sample of 848patients randomised to an FGA–LAI (fluphenazine decanoate,fluspirilene decanoate, pipotiazine palmitate) or an FGA–oralmedication (including chlorpromazine, haloperidol, penfluridoland trifluoperazine) (Fig. 1). The duration of included trialsvaried (4 weeks to 2 years), but most patients took part in trialsof at least a year in duration. The risk of relapse did not differbetween the two groups (RR=0.96, 95% CI 0.8–1.1). In an analysisof 127 patients treated with three FGA–LAIs (fluphenazinedecanoate, fluphenazine enanthate and haloperidol decanoate),global improvement (assessed using the Clinical GlobalImpressions scale) was more likely with FGA–LAI than withFGA–oral medication, with a number needed to treat (NNT) of 4 (95% CI 2–9). The FGA–LAI and FGA–oral groups were similarin terms of study attrition, the need for adjunctive anticholinergicmedication and incidence of tardive dyskinesia (Fig. 1). Anti-cholinergic medication, a proxy marker for the presence of extrapyramidal symptoms, was prescribed to 69% of the FGA–LAI cohort and 65% of the FGA–oral cohort. The prevalence of tardive dyskinesia in the FGA–LAI cohort was 9.0% and in theFGA–oral cohort it was 14.1%.We identified one RCT not included in the srcinal or updatedCochrane reviews of FGA–LAIs, namely that by Arango  et al  . 18 This small RCT compared oral zuclopenthixol ( n =20) withzuclopenthixol decanoate ( n =26) over 1 year in patients withschizophrenia and a history of violence. A lower frequency of violent acts was seen in the LAI group but end-point scores onthe Positive and Negative Syndrome Scale (PANSS) did not differ. Prospective observational studies We identified four prospective observational studies thatcompared an FGA–LAI with one or more oral antipsychoticcohorts (Table 1). 19–23 These studies had various pragmaticoutcome measures, including risk of readmission and time toall-cause discontinuation of medication. Results were mixed.Two studies found a better outcome for FGA–LAI compared withan FGA–oral. 19,20 The Schizophrenia Outpatient HealthOutcomes (SOHO) study found poorer outcomes for FGA–LAIthan oral olanzapine, 21,22 and a fourth study found oral anti-psychotics to be superior to haloperidol decanoate but equivalentto fluphenazine decanoate. 23 Tiihonen  et al   assessed the outcome of patients after their firstadmission with schizophrenia or schizoaffective disorder. 19 Theother studies in Table 1 had samples wholly or largely comprisingpatients who had had schizophrenia for several years. TheTiihonen study assessed a nationwide cohort, all first admissionsin Finland occurring over a 5½-year period, and had a meanfollow-up period of 3.6 years. Analysis was performed on theten most commonly used antipsychotics, which included oneinjectable formulation: perphenazine LAI. Multivariate modelsand propensity score methods were used to adjust estimates of effectiveness, and comparisons were made with oral haloperidol.Initial use of perphenazine LAI was associated with a significantly lower adjusted risk of all-cause discontinuation than that forhaloperidol and the second lowest discontinuation rate of theten drugs studied. In an analysis of rehospitalisation rates,calculated according to the ongoing antipsychotic, perphenazineLAI had the lowest risk of rehospitalisation (68% reduction infully adjusted relative risk compared with haloperidol) (Fig. 2).Oral perphenazine showed no difference from oral haloperidolin terms of adjusted risk of discontinuation and rehospitalisation,suggesting that it was the mode of administration rather than thedrug  per se  that was responsible for the improved outcome withperphenazine LAI. 24 Zhu  et al   used data from the US Schizophrenia Care andAssessment Program (US–SCAP) study to assess the time to all-cause medication discontinuation in the first year after initiationof an FGA–LAI or oral antipsychotic. 20 The study assessed thesame two antipsychotics – haloperidol and fluphenazine – in oralor LAI form. Compared with those treated with oral medication, s21 Review of FGA–LAI v. oral antipsychotic studies  Haddad et al those treated with LAI had a significantly longer mean time to all-cause medication discontinuation and were twice as likely tocontinue taking the medication (Fig. 3).The SOHO study was a pan-European observational study funded by Eli Lilly that recruited over 10000 patients with schizo-phrenia when they began a new antipsychotic medication regimenon an out-patient basis. 21 Patients were assessed at regularintervals for up to 3 years or until discontinuation of the baselineantipsychotic. The study included various SGA–oral cohorts plusa mixed cohort prescribed various FGA–LAIs and another mixedcohort taking various FGA–orals. Statistical comparisons weremade relative to oral olanzapine. The likelihood of not achievingremission, the risk of relapse and the all-cause discontinuation rateof medication were all higher for those treated with FGA–LAIcompared with oral olanzapine. 21,22 The proportion of individualswho had stopped medication by 3 years was 36.4% for thosetaking olanzapine, 50.2% for those who began FGA–LAI treatmentand 53.1% for those taking an oral FGA. 22 The hazard ratio (risk)for discontinuation relative to olanzapine for FGA–orals was1.70 (95% CI 1.46–1.97) and for FGA–LAIs it was 1.43 (95% CI1.19–1.70). 22 Conley   et al   assessed the risk of readmission in patientsdischarged from several in-patient psychiatric units in the Stateof Maryland, USA. 23 Cohorts discharged on fluphenazinedecanoate and haloperidol decanoate were compared with cohortsdischarged on one of three SGA–orals. The 1-year readmissionrisk (with adjustment for baseline variables) for each of the threeSGA–oral groups was lower than for the haloperidol decanoategroup but similar to that seen with fluphenazine decanoate.The only study in Table 1 that presented tolerability data wasthe SOHO study, albeit descriptive data without statisticalanalysis. 22 The presence of extrapyramidal symptoms and tardivedyskinesia was based on clinical judgement rather than ratingscales. The period prevalence for extrapyramidal symptoms s22 StudyDepot( n/N  )Oral( n/N  )RR(95% CI random)RR(95% CI random)Weight(%)01 DeathFluphenazine decanoate 2/78 1/78Subtotal (95% CI) 2/78 1/78Test for overall effect  z  =0.57,  P  =0.602 Global functioning: No important global changeFluphenazine decanoate 22/38 34/36Fluphenazine enanthate 5/16 7/15Haloperidol decanoate 8/11 9/11Subtotal (95% CI) 35/65 50/62Test for heterogeneity w 2 =1.85, d.f.=2,  P  =0.40Test for overall effect  z  = 7 3.27,  P  =0.00103 Mental state: General – relapseFluphenazine decanoate 129/339 112/345Fluspirilene enanthate 2/20 2/20Pipotiazine palmitate 15/61 10/63Subtotal (95% CI) 146/420 154/428Test for heterogeneity w 2 =1.88, d.f.=2,  P  =0.39Test for overall effect  z  = 7 0.50,  P  =0.0604 Leaving the study earlyFlupentixol decanoate 3/30 1/30Fluphenazine decanoate 85/298 77/310Fluspirilene decanoate 2/20 2/20Pipotiazine palmitate 16/85 15/81Subtotal (95% CI) 106/433 95/441Test for heterogeneity w 2 =0.89, d.f.=3,  P  =0.83Test for overall effect  z  =1.06,  P  =0.0305 Side-effects: I. Movement disorders – general – needing anticholinergic medicationFlupentixol decanoate 19/30 16/30Fluphenazine decanoate 54/75 54/80Fluspirilene decanoate 19/20 14/20Haloperidol decanoate 3/11 1/11Pipotiazine palmitate 42/61 49/63Subtotal (95% CI) 137/197 134/204Test for heterogeneity w 2 =6.46, d.f.=4,  P  =0.17Test for overall effect  z  =0.87,  P  =0.03406 Side-effects: 2. Movement disorders – tardive dyskinesiaFluphenazine decanoate 9/72 16/76Pipotiazine palmitate 3/61 3/63Subtotal (95% CI) 12/133 19/139Test for heterogeneity w 2 =0.39, d.f.=1,  P  =0.53Test for overall effect  z  = 7 1.21,  P  =0.02100.0 2.00 (0.19–21.61)100.0 2.00 (0.19–21.61)67.6 0.61 (0.46–0.81)6.6 0.67 (0.27–1.66)25.0 0.89 (0.56–1.40)100.0 0.68 (0.54–0.86)92.9 0.92 (0.77–1.11)0.9 1.00 (0.16–6.42)6.2 1.55 (0.76–3.18)100.0 0.96 (0.80–1.14)1.2 3.00 (0.33–27.24)82.8 1.15 (0.88–1.50)1.7 1.00 (0.16–6.42)14.3 1.02 (0.54–1.92)100.0 1.14 (0.90–1.45)13.4 1.19 (0.77–1.83)32.5 1.07 (0.87–1.31)21.8 1.36 (1.00–1.84)0.7 3.00 (0.37–24.58)31.6 0.89 (0.71–1.10)100.0 1.08 (0.90–1.30)81.2 0.59 (0.28–1.26)18.8 1.03 (0.22–4.92)100.0 0.66 (0.33–1.30)0.1 0.2 1 5 10Favours depot Favours oral 7777 Fig. 1  Outcomes for antipsychotic treatment: long-acting injection (LAI)  v.  oral. From Adams  et al  , 12 reproduced with permission.
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks