Short-term treatment with risperidone or haloperidol in first-episode schizophrenia: 8-week results of a randomized controlled trial within the german research network on schizophrenia

International Journal of Neuropsychopharmacology (2008), 11, 985–997. Copyright f 2008 CINP haloperidol in first-episode schizophrenia:8-week results of a randomized controlled trialwithin the German Research Networkon Schizophrenia Hans-Ju¨rgen Mo¨ller1, Michael Riedel1, Markus Ja¨ger1, Florian Wickelmaier1,Wolfgang Maier2, Kai-Uwe Ku¨hn2, Gerhard Buchkremer3, Isabella Heuser4,Joachim Klosterko¨tter 5, Markus Gastpar6, Dieter F. Braus7, Ralf Schlo¨sser8,Frank Schneider9, Christian Ohmann10, Mathias Riesbeck11, Wolfgang Gaebel11,for the German Study Group on First-Episode Schizophrenia 1 Department of Psychiatry and Psychotherapy, Ludwig-Maximilians-University, Munich, Germany ; 2 Department of Psychiatryand Psychotherapy, University of Bonn, Bonn, Germany ; 3 Department of Psychiatry and Psychotherapy, University of Tuebingen,Tuebingen, Germany ; 4 Department of Psychiatry and Psychotherapy, Charite´-Campus Benjamin Franklin, Berlin, Germany ;5 Department of Psychiatry and Psychotherapy, University of Cologne, Cologne, Germany ; 6 Department of General Psychiatry,Rhein Hospital Essen, Essen, Germany ; 7 Central Institute of Mental Health, Mannheim, Germany ; 8 Department of Psychiatryand Psychotherapy, University of Jena, Germany ; 9 Department of Psychiatry and Psychotherapy, RWTH Aachen University,Germany ; 10 Coordinating Centre for Clinical Trials, Heinrich-Heine-University, Duesseldorf, Germany ; 11 Departmentof Psychiatry and Psychotherapy, Heinrich-Heine-University, Duesseldorf, Germany Patients with first-episode schizophrenia appear to respond to lower doses of neuroleptics, and to be moresensitive to developing extrapyramidal side-effects. The authors therefore compared in such patients theefficacy and extrapyramidal tolerability of comparatively low dosages of the atypical neuroleptic risperi-done and of the conventional neuroleptic haloperidol. Risperidone was hypothesized to have better extra-pyramidal tolerability and efficacy in treating negative symptoms. Patients were randomly assigned underdouble-blind conditions to receive risperidone (n=143) or haloperidol (n=146) for 8 wk. The primaryefficacy criterion was the estimated difference in the mean change in the Positive and Negative SymptomScale (PANSS) negative score between treatment groups ; secondary efficacy criteria were changes on thePANSS total score and other PANSS subscores, and several other measures of psychopathology and generalfunctioning. The primary tolerability criterion was the difference in baseline-adjusted occurrence rates ofextrapyramidal side-effects measured with the Simpson–Angus Scale (SAS) compared between treatmentgroups. The main hypothesis was that risperidone would be superior in terms of improving negativesymptoms and lowering the risk of extrapyramidal symptoms. Secondary tolerability criteria were the otherextrapyramidal symptoms, measured with the Hillside Akathisia Scale (HAS) and the Abnormal Invol-untary Movement Scale (AIMS). The average mean daily doses were 3.8 mg (S.D.=1.5) for risperidone and3.7 mg (S.D.=1.5) for haloperidol. There were similar, significant improvements in both treatment groups inthe primary and secondary efficacy criteria. At week 8 nearly all scores of extrapyramidal side-effectsindicated a significantly higher prevalence of extrapyramidal side-effects with haloperidol than withrisperidone [SAS : risperidone 36.5 % of patients ; haloperidol 51.5 % of patients ; likelihood ratio test,x2(1)=7.8, p=0.005]. There were significantly fewer drop-outs [risperidone n=55, drop-out rate=38.5 % ;haloperidol n=79, drop-out rate=54.1 %, x2(1)=7.1, p=0.009] and a longer non-discontinuation time[risperidone : average of 50.8 d to drop-out ; haloperidol : average of 44.0 d to drop-out ; log rank test,x2(1)=6.4, p=0.011] in the risperidone group. Risperidone and haloperidol appear to be equally effective intreating negative and other symptoms of first-episode schizophrenia. Risperidone has better extra-pyramidal tolerability and treatment retention rate than the equivalent dose of haloperidol in these patients.
Received 2 August 2007 ; Reviewed 16 September 2007 ; Revised 22 February 2008 ; Accepted 9 March 2008 ;First published online 9 May 2008 Key words : First-episode schizophrenia, haloperidol, risperidone.
Address for correspondence : Prof. H.-J. Mo¨ller, Department of Psychiatry and Psychotherapy, Ludwig-Maximilians-University, Nussbaumstr. 7,80336 Munich, Germany.
Tel. : 0049 89 5160 5501 Fax : 0049 89 5160 5522 E-mail : [email protected] in treating negative symptoms (Mo¨ller, 2000b, 2003),special emphasis should be placed on this psycho- Neuroleptics are more effective in first-episode schizo- pathological domain in the treatment of FES pa- phrenia (FES) than in patients with multiple episodes (Ja¨ger et al., 2006 ; Lieberman et al., 1996 ; Sanger et al., Efficacy results from studies on FES are inconsistent 1999). Lower daily doses are required for adequate for negative symptoms ; they have been found to improvement, and patients are more sensitive to de- improve similarly and significantly from baseline veloping extrapyramidal side-effects (EPS) (Merlo with risperidone and haloperidol (Emsley and the et al., 2002 ; Sanger et al., 1999 ; Zhang-Wong et al., Risperidone Working Group, 1999), and significantly more with olanzapine than haloperidol (Sanger et al., Only rather limited evidence is available for the 1999). One study found no difference between olan- efficacy and tolerability of second-generation anti- zapine and haloperidol in the last observation carried psychotics (SGAs) in FES, especially in comparison to forward (LOCF) analysis (Lieberman et al., 2003), but first-generation antipsychotics (FGAs) (Rummel et al., significantly greater improvement with olanzapine in 2003). Earlier short-term trials on acute treatment of FES patients investigated risperidone alone (Kopala The study was planned in 1999 as part of the et al., 1998 ; Merlo et al., 2002), risperidone vs. halo- German Research Network on Schizophrenia (a peridol (Emsley and the Risperidone Working Group, nationwide research network funded by the German 1999), and olanzapine vs. haloperidol (Keefe et al., Ministry of Education and Research ; BMBF) (Gaebel 2004 ; Lieberman et al., 2003 ; Sanger et al., 1999). Apart et al., 2004 ; Wolwer et al., 2003). The entire FES study from one study with a dose range of 5–20 mg/d halo- programme consisted of an 8-wk acute study and a peridol (mean dose 10.8 mg/d) (Sanger et al., 1999), the subsequent 2-yr, long-term treatment phase, the 1-yr haloperidol dose was kept in the range 2–8 mg/d outcome data of which were recently published (Emsley and the Risperidone Working Group, 1999 ; (Gaebel et al., 2007). The aim was to evaluate whether Lieberman et al., 2003). Besides the short-term studies risperidone shows better efficacy, especially in terms on FES, long-term studies comparing SGAs with halo- of negative symptoms, and fewer EPS than halo- peridol or comparing several SGAs with each other peridol when administered at equivalent dosages have recently been published (Green et al., 2006 ; Keefe in a design aimed at keeping the dose as low as poss- et al., 2007 ; McEvoy et al., 2007 ; Schooler et al., 2005).
In some of these long-term studies the average dose ofstudy drug was lower than in the short-term studies on acute patients. For example, in the risperidone–haloperidol study by Schooler et al. (2005) the mean modal dose was 3.3 mg for risperidone and 2.9 mg for Subjects were selected from patients admitted to the in-patient departments of the participating centres.
All of the studies with haloperidol found the SGA to Inclusion criteria were : (1) acute manifestation of FES be significantly superior in terms of extrapyramidal according to ICD-10 F20 criteria ; (2) age 18–60 yr ; tolerability. The efficacy results are not so consistent.
(3) adequate proficiency in German ; (4) no involun- In some studies superior efficacy was found for SGAs tary in-patient treatment (at the date of inclusion) ; either in terms of global outcome, remission rates, re- (5) written informed consent. Exclusion criteria lapse rates or scores/subscores of symptom rating were : (1) pregnancy ; (2) insufficient response to pre- scales. The non-discontinuation rate, used as a cri- treatment with risperidone or haloperidol ; (3) other terion in recent effectiveness studies such as the contraindications for risperidone or haloperidol ; Clinical Antipsychotic Trials of Intervention Effec- (4) mental retardation ; (5) organic brain disease ; tiveness (CATIE ; Lieberman et al., 2005), was also (6) substance abuse ; (7) history of suicidal behav- found in some studies to be more advantageous for iour ; (8) severe physical disease ; (9) participation in Because negative symptoms in general are of great relevance for social functioning, prognosis and course characteristics (Mo¨ller et al., 1988, 2000, 2002, Mo¨ller,2004, 2007), as was also described for FES (Haas and This was an 8-wk, multicentre, double-blind, parallel- Sweeney, 1992 ; Siegel et al., 2006), and because SGAs group, randomized, controlled study in in-patients have been described to show better efficacy than FGAs suffering from acute FES. It was conducted in 13 Treatment of first-episode schizophrenia German psychiatric university hospitals according efficacy criteria were : PANSS total, positive and gen- to the principles of good clinical practice and the eral subscale, Scale for the Assessment of Negative Declaration of Helsinki. Approval was obtained from Symptoms (SANS ; Andreasen, 1982), Hamilton Rating the ethics committees of the coordinating centre and Scale for Depression (HAMD ; Hamilton, 1960), Calgary Depression Scale for Schizophrenia (CDSS ;Addington et al., 1990 ; Muller et al., 1999), CGI, Young Mania Rating Scale (YMRS ; Young et al., 1978).
Compliance was monitored at every visit with a Patients were assessed for suitability at screening and questionnaire. Prognosis was assessed at baseline baseline. Patients pretreated with psychotropic drugs with the Strauss–Carpenter Prognosis Scale (SCPS ; underwent a washout period of 4–7 d, if clinically Strauss and Carpenter, 1978). At both baseline and justified. After random assignment to treatment week 8, functioning was measured with the Global groups in a 1 : 1 ratio, patients received 2 mg/d of the Assessment of Functioning Scale (GAF ; Frances et al., study medication (once daily). Investigators, hospital 1994) and the Social and Occupational Functioning staff and patients were blinded to treatment assign- Assessment Scale (SOFAS ; APA, 1994). Several rater ment. Patients at each centre were consecutively trainings took place. Inter-rater reliability yielded a allocated a labelled container of medication, supplied satisfactory to good concordance which fitted to by the Coordinating Centre. The identification con- values in other publications (e.g. intra-class corre- sisted of a letter representing the centre, a three-digit lation coefficient of the PANSS positive scale=0.74, patient number, a three-letter abbreviation of the p<0.001 ; Shrout and Fleiss, 1979).
study name and a number representing the investi- Adverse events were assessed at every visit and EPS recorded using the Simpson–Angus Scale (SAS ; Dose could be increased by 1–2 mg/d between day 3 Simpson and Angus, 1970), Hillside Akathisia and week 1, if required, and then at each weekly as- Scale (HAS ; Fleischhacker et al., 1989) and Abnormal sessment up to a maximum of 8 mg/d, whereby the Involuntary Movement Scale (AIMS ; Guy, 1976a).
total dose should not exceed 4 mg/d by week 2. The A broad range of safety laboratory assessments criterion for a dose increase (apart from between day 3 [differential blood count, clinical chemistry (sodium, and the week 1) was the non-achievement of symptom potassium, calcium, glucose, total bilirubin, aspartate improvement of at least one level on the Clinical aminotransferase, alanine aminotransferase, alkaline Global Impression (CGI ; Guy, 1976b). If EPS ap- phosphatase, gamma-glutamyltransferase, creatinine, peared, dosage reductions were permitted in steps of urea, uric acid, creatinine kinase, lactate dehydro- 1–2 mg/d at the weekly evaluations.
genase, total protein), thyroid-stimulating hormone, Co-medication with psychotropic substances was blood sedimentation rate, urine status] were per- generally not allowed, with the exception of the fol- formed at selection and repeated at baseline and at weeks 2, 4, 6, and 8. An ECG was conducted at selec-tion and week 8.
$ Short-acting benzodiazepines for insomnia.
$ Lorazepam was permitted at the lowest possible dose for the shortest possible time to arrest agitation, psychotic anxiety, etc. ; after week 4, the dose wasnot allowed to exceed 4 mg/d over 4 d/wk. Con- A sample size of 176 patients/group was estimated to tinuous administration of benzodiazepines was not test the hypothesis of an advantage for risperidone (two-sided test) at week 8 in the primary efficacy par-ameter PANSS negative score and the primary toler- $ If dose reduction of the study drug did not achieve the desired effect, the anticholinergic biperiden (up ability criterion SAS on the basis of the following to 6 mg/d) was prescribed to treat EPS, and the beta- assumptions : a=0.05 ; power 1xb=0.8, expected blocker propranolol (up to 80 mg/d) to treat aka- group difference of d=0.3 standard deviations.
thisia, usually for a maximum of 14 d at a time.
The intent-to-treat (ITT) sample comprised all ran- domized patients except those whose initial diagnosishad been revised, and the per protocol sample of all patients who completed the 8-wk trial as in-patients or The negative scale of the Positive and Negative who were discharged prior to week 8. Drop-outs were Syndrome Scale (PANSS ; Kay et al., 1987) was as- patients who discontinued for any reason. Differences sessed as the primary efficacy criterion. Secondary in time to drop-out between treatment groups were evaluated using Kaplan–Meier analysis. Continuous patients having a total score >0. LOCF analyses were outcome measures are presented as means and performed using logistic regression including terms standard deviations, categorical variables as absolute for baseline value and treatment. Observed cases frequencies (n) and percentages. Treatment group (weekly prevalence rates) were evaluated by general- differences at baseline were evaluated using two- ized estimation equation (GEE) models (Agresti, 2002 ; sample t tests and Pearson’s x2 tests. Two alternative Liang and Zeger, 1986) including a common intercept, analytical strategies were applied to the ITT data : (1) treatment-specific linear and quadratic slopes as ef- last observation (under regular treatment conditions) fects for treatment and week. The correlation in the carried forward (LOCF analysis) ; (2) the observed responses over time was accounted for by specifying cases were analysed using methods for longitudinal an autoregressive working correlation structure. This model was determined by starting from a preliminary Based on the special importance of negative symp- model allowing for treatment-specific intercepts, lin- toms and some positive findings for SGAs in this re- ear and quadratic slopes, and subsequently eliminat- spect, the estimated difference in the mean change in ing non-significant terms according to Wald tests. In the PANSS negative score between the two treatment addition, incidence of EPS was estimated as the pro- groups [analysis of covariance (ANCOVA) and portion of patients with a change in side-effects total mixed-model analysis] was chosen as the primary score from 0 at baseline to >0 at endpoint. Incidence efficacy criterion. The LOCF was analysed using rates were compared using Pearson’s x2 tests.
ANCOVA including terms for baseline value and Supplementary analyses were performed on the per treatment in the model (Vickers and Altman, 2001).
protocol sample in order to confirm the results ob- Baseline-to-endpoint differences were tested for with tained by the ITT analyses (results are not reported).
paired t tests. Observed cases (scores at each week) Statistical hypotheses were tested on a two-sided were analysed using mixed-effects models (Hedeker and Gibbons, 2006 ; Singer, 1998), including a commonintercept, treatment-specific linear slopes, and a com- mon quadratic slope as fixed effects for treatment and week. Subject-specific intercepts and slopes were in-cluded as random effects. A first-order autoregressive Between November 2000 and May 2004, 1372 patients covariance structure was chosen in order to account were screened. The unexpectedly slow recruitment for observations within a subject to be autocorrelated rate meant that recruitment had to be discontinued for over time. This model was determined by starting from pragmatic reasons before reaching the planned sample a preliminary model allowing for treatment-specific size. There were 1070 patients (78.0 %) unsuitable for intercepts, linear and quadratic slopes, and sub- the study : 755/1070 (70.6 %) patients did not fulfil in- sequently eliminating non-significant terms according clusion or fulfilled exclusion criteria, 298 (27.9 %) to likelihood ratio tests. Patients were classified as patients refused to participate, and 17 were not in- treatment responders if they had (1) a rating f3 in cluded for other reasons. Of the remaining 302 PANSS items 1–3, 5, 6, (2) a o30 % reduction from patients included (total sample), 105 completed the baseline in PANSS total score, and (3) a CGI severity 8-wk trial as in-patients, 50 were discharged before score f4 (cf. Lieberman et al., 2003). Time to response completion (per protocol sample 155 ; risperidone 88, was estimated using Kaplan–Meier analysis.
haloperidol 67), and 134 discontinued prematurely The primary tolerability criterion in terms of EPS (risperidone 55, haloperidol 79). Besides fulfilling the was the SAS score. AIMS and HAS were secondary criteria for schizophrenia according to ICD-10 F20, tolerability criteria. Scores of these tolerability criteria 289/302 (95.70 %) patients also fulfilled the respective were dichotomized (0, >0) because >50 % of the sample showed total scores of zero throughout the The ITT sample consisted of 289 patients (risperi- study (presumably due to the low dosages of medi- done 143, haloperidol 146) as 13 patients were not cation applied). Therefore, since the continuous total eligible for inclusion in the analysis. Details of the scores markedly deviated from normal, methods for numbers selected and the number of drop-outs are categorical data analysis were applied to the dichot- omized values ; average continuous total scores of the With the actual sample size of n=143 (risperidone) tolerability criteria are not reported because they are and n=146 (haloperidol) only a slightly larger effect strongly influenced by outlying values. Instead, size of d=0.33 standard deviations was detectable at prevalence of EPS was estimated by the proportion of Treatment of first-episode schizophrenia Assessed for eligibility: n=1372
(% of base)
Not included (in acute study)
(78.0/1372)
Not meeting inclusion/
meeting exclusion criteria
(70.6/1070)
i.e. (multiple answers possible):- Diagnostic criteria of F20 not fulfilled - Participation in other (incompatible) trial - Contraindication of neuroleptic treatment Refused participation
(27.9/1070)
Other reasons
(0.6/1070)
Not documented
(1.0/1070)
Included in acute study: 302 (22% of 1372)
Not randomized: n=6
Figure 2. Prevalence of movement disorders over the 8-wk study, defined by the proportion of patients havingSimpson–Angus Scale (SAS) total score >0. Differences Randomized: n =296
between risperidone ($) and haloperidol (#) are significantbased on last observation carried forward (LOCF) analysis Risperidone: n=148
Haloperidol: n=148
[logistic regression, likelihood ratio test, x2(1)=7.8, p=0.005] (50% of 296)
(50% of 296)
and generalized estimation equation analysis [Wald test, Excluded: n=5
Excluded: n=2
dropped out due to EPS with haloperidol [8.2 % (12/ Risperidone: n =143
Haloperidol: n=146
46)] than with risperidone [4/143 (2.8 %), Fisher’s exacttest, p=0.069]. Further reasons for drop-out were non- Drop-out n=55 (38.5%)
Drop-out n=79 (54.1%)
compliance [risperidone 15/143 (10.5 %), haloperidol 11/146 (7.5 %)], withdrawal of informed consent [ris- peridone 8/143 (5.6 %), haloperidol 16/146 (11 %)], insufficient response [risperidone 7/143 (4.9 %), halo- peridol 13/146 (8.9 %)] and others [risperidone 15/143 (10.5 %), haloperidol 15/146 (10.3 %)] ; the frequency ofdrop-outs in each of these categories did not differsignificantly between groups.
Completer sample
Completer sample
(completed study or
(completed study or
discharged):
discharged):
Risperidone: n=88
Haloperidol: n=67
Figure 1. Flow of subjects through the study.
The ITT sample consisted of 117/289 (40.5 %) womenand 172/289 (59.5 %) men, mean age 30.1¡9.8 yr.
The mean PANSS total score at baseline was 79.1, There were significantly fewer drop-outs with risperi- indicating a severity level typical for an acute episode done (n=55, drop-out rate=38.5 %) than haloperidol [n=79, drop-out rate=54.1 %, x2(1)=7.1, p=0.009] The only significant difference in baseline charac- (Figure 1). The Kaplan–Meier analysis of the discon- teristics between the two groups was the proportion tinuation rates mirrors these results [log rank test, of patients with an AIMS total score >0 [x2(1)=4.2, x2(1)=6.4, p=0.011] (Figure 2) : average of 50.8 d to p=0.041, see Table 1]. The baseline SANS total score drop-out for a risperidone patient, and 44.0 d for a showed a numerical difference of about 4 points haloperidol patient. Side-effects were the most fre- between the two treatment groups (risperidone quent reason for drop-out in both groups, but more 36.1¡28.5, haloperidol 40.3¡25.7 ; n.s.), but this was often the reason in the haloperidol group [risperidone not reflected in the PANSS negative score (risperidone 10/143 (7 %), haloperidol 24/146 (16.4 %) ; Fisher’s 19.0¡8.3, haloperidol 19.6¡8.1 ; n.s.). Values of the exact test, p=0.017]. A higher proportion of patients mean PANSS total score, an indicator of the severity of Table 1. Sample characteristics and drug-group differences at entry in the study (ITT sample ; two-sample t test forcontinuous variables, Pearson’s x2 test for categorical variables) ITT, Intent to treat ; PANSS, Positive and Negative Syndrome Scale ; SANS, Scale for the Assessment of Negative Symptoms ;HAMD, Hamilton Depression Rating Scale ; CDSS, Calgary Depression Scale for Schizophrenia ; CGI, Clinical GlobalImpression ; YMRS, Young Mania Rating Scale ; SCPS, Strauss–Carpenter Prognosis Scale ; GAF, Global Assessment ofFunctioning ; SOFAS, Social and Occupational Functioning Assessment Scale ; SAS, Simpson–Angus Scale ; AIMS, AbnormalInvoluntary Movement Scale ; HAS, Hillside Akathisia Scale.
a Reduced n in single scales due to missing values.
the psychopathological symptoms, as well as mean p=0.197] or propranolol [risperidone 4.0 % of patients, SCPS score, an indicator of the global prognosis, were haloperidol 6.5 % of patients ; x2(1)=0.73, p=0.393].
The prescription rate (risperidone 38.7 % of patients,haloperidol 34.7 % of patients ; n.s.) and average dose (risperidone 4.1¡1.6 mg/d, haloperidol 4.0¡ The aim to keep the average neuroleptic dose as low 1.6 mg/d ; n.s.) of biperiden was similar in both as possible was achieved and was y4 mg/d for both drugs (risperidone 3.8¡1.5 mg/d, haloperidol 3.7¡1.5 mg/d).
The results of the ITT sample are presented.
Both treatment groups showed marked, statistically There was no significant difference between groups in significant improvements from baseline to week 8 in the prescription rates of lorazepam [risperidone 54.8 % the primary efficacy criterion (PANSS negative score ; of patients, haloperidol 62.9 % of patients ; x2(1)=1.67, paired t test, p<0.001) and in the secondary efficacy Treatment of first-episode schizophrenia criteria (paried t test, p<0.001 for all parameters ex- endpoint (note that Fisher’s exact test was used be- cept CDSS, for which p<0.01) (Table 2), indicating ef- cause of the small number of cases per cell). The re- ficacy in various symptom domains. Efficacy did not sults were similar, and the results for SAS and HAS differ significantly in either the LOCF or mixed-model also reached statistical significance (SAS p=0.008, analysis (Table 2) ; the same was true for global func- AIMS p=0.017, HAS p=0.001) (Table 3c).
The incidence of EPS was also calculated in relation Both drugs were effective with respect to treatment to average dose per day, with a cut-off point at 4 mg response defined after Lieberman et al. (2003) [i.e. (1) a (<4 mg vs. o4 mg). The differences between the two rating f3 in PANSS items 1–3, 5, 6, (2) a o30 % re- treatment groups were again in favour of risperidone, duction from baseline in PANSS total score, and (3) a and for SAS and HAS were more pronounced between CGI severity score f4]. At week 8, 66/134 (49.3 %) of patients showed a response with risperidone and 63/127 (49.6 %) with haloperidol. The Kaplan–Meier esti- mated average time to response was 41.0 d with ris-peridone and 38.6 d with haloperidol [log rank test, Only one of the 146 patients receiving haloperidol showed clinically significant laboratory test abnor-malities. There were no abnormal laboratory values inthe risperidone group in the safety parameters men- Both compounds were safe, i.e. no serious adverse The dose of both neuroleptics was kept as lowas possible in this study (average dose y4 mg/d).
The dose equivalence for risperidone and haloperidol After 8 wk, all three EPS-related scores indicated a has not been finally clarified. In this fixed flexible higher prevalence of EPS, defined by the number of dosing study, a relationship of 1 : 1 was observed, patients having a total parameter score >0 with halo- which is similar to some other studies (Marder et al., peridol than with risperidone. LOCF analysis found a 2003 ; Schooler et al., 2005), while others, for ex- significant difference in baseline-adjusted prevalence ample, reported a 1 : 2.5 ratio (Csernansky et al., rates after 8 wk in favour of risperidone for both the primary tolerability criterion, the SAS [risperidone The results in the outcome parameters show that 36.5 % of patients, haloperidol 51.5 % of patients ; like- risperidone and haloperidol are similarly efficacious lihood ratio test, x2(1)=7.8, p=0.005] (Figure 3), and in treating FES. Contrary to the hypothesis, risperi- the AIMS [risperidone 8.8 % of patients, haloperidol done was not superior in treating negative symptoms.
21.7 % of patients ; likelihood ratio test, x2(1)=6.4, Other short-term FES studies that compared an SGA p=0.011] (Table 3a). The resulting odds ratios suggest with haloperidol found inconsistent results in negative that the risk of EPS is more than doubled in the halo- symptoms. Emsley and the Risperidone Working peridol group at endpoint. The prevalence rates of the Group (1999) found that negative symptoms sig- HAS showed no significant difference. GEE analyses nificantly improved with both risperidone and confirmed the results of LOCF analyses, indicating a haloperidol. Lieberman et al. (2003) found similar re- higher weekly increase in risk of EPS for haloperidol ; ductions in symptom severity with olanzapine and this effect was again significant for SAS and AIMS, but haloperidol in the LOCF analysis, but olanzapine had significantly greater decreases in the PANSS negative Another categorical analysis (Lieberman et al., scale, among others, in a mixed-model analysis.
2003), evaluated the incidence of EPS defined by a However, the second part of the hypothesis, the change in total parameter score from 0 at baseline to lower risk of risperidone in terms of EPS, could be >0 at endpoint; only the AIMS score showed a sig- substantiated. Both the LOCF and the longitudinal nificant difference between the two treatment groups data analysis confirmed the superiority of risperidone [risperidone 5.9 % of patients, haloperidol 14.7 % of in terms of extrapyramidal tolerability, especially at patients ; x2(1)=5.7, p=0.017] (Table 3b). In a sup- dose levels >4 mg. EPS were more prevalent in the portive analysis, incidence was re-defined by a change haloperidol group, as shown by a significantly greater in EPS total score from <1 at baseline to o1 at proportion of scores >0 in the SAS and AIMS scales, Table 2. Treatment response : psychopathological parameters at baseline and end of study (ITT sample), LOCF (ANCOVA to adjust for possible imbalance at baseline) and Difference in baseline-to-endpoint improvement Mean scores¡S.D. at baseline and end of study ITT, Intent to treat ; LOCF, last observation carried forward ; PANSS, Positive and Negative Syndrome Scale ; SANS, Scale for the Assessment of Negative Symptoms ; HAMD,Hamilton Depression Rating Scale ; CDSS, Calgary Depression Scale for Schizophrenia ; CGI, Clinical Global Impression ; GAF, Global Assessment of Functioning ; SOFAS, Socialand Occupational Functioning Assessment Scale.
a Reduced n in single scales due to missing values.
Treatment of first-episode schizophrenia Table 3a. Prevalence of movement disorder side effects estimated by the proportion of patients having a total parameterscore >0. Drug-group differences at end of study (ITT sample), LOCF (logistic regression to adjust for possible imbalance atbaseline) and GEE analysis ITT, Intent to treat ; LOCF, last observation carried forward ; GEE, generalized estimation equation ; SAS, Simpson–Angus Scale ;AIMS, Abnormal Involuntary Movement Scale ; HAS, Hillside Akathisia Scale.
a Reduced n in single scales due to missing values.
Table 3b. Incidence of movement disorder side effects defined (after Lieberman et al., 2003) by a change in total parameterscore from 0 at baseline to >0 at endpoint (week 8, LOCF, Pearson’s x2 test) LOCF, Last observation carried forward ; SAS, Simpson–Angus Scale ; AIMS, Abnormal Involuntary Movement Scale ; HAS,Hillside Akathisia Scale.
a Reduced n in single scales due to missing values.
and a numerically greater frequency of scores >0 in the HAS scale. Surprisingly, this difference was notmirrored by the prescription rates of the anti- cholinergic biperiden. In general risperidone, like otherSGAs, has better extrapyramidal tolerability than haloperidol (Leucht et al., 1999 ; Mo¨ller, 2000a), an ad-vantage which was found consistently in FES studies (Emsley and the Risperidone Working Group, 1999 ;Lieberman et al., 2003 ; Sanger et al., 1999 ; Schooler et al., 2005). The differences favouring risperidone interms of EPS were especially convincing in the sub- analysis of the FUTURIS study (Kopala et al., 2003), Cumulative probability of remaining in study which compared in an ex-post analysis patients with equivalent doses of risperidone and haloperidol inFES, i.e. 1 mg vs. 1 mg, 2 mg vs. 2 mg, etc. On the other Figure 3. Kaplan–Meier analysis of discontinuation rates.
side, based on a double-blind RCT in a small sample The difference in time to discontinuation between treatment groups is significant [log rank test, x2(1)=6.4, =40) of FES patients comparing 2 mg and 8 mg of haloperidol, it was suggested that even a dose around Table 3c. Incidence of movement disorder side effects defined (after Lieberman et al., 2003) by a change in total parameterscore from <1 at baseline to o1 at endpoint (week 8, LOCF, Fisher’s exact test) LOCF, Last observation carried forward ; SAS, Simpson–Angus Scale ; AIMS, Abnormal Involuntary Movement Scale ; HAS,Hillside Akathisia Scale.
a Reduced n in single scales due to missing values.
Table 3d. Incidence of movement disorder side effects (cf. Lieberman et al., 2003) by average dose per day, and difference inincidence between treatment groups 21 (29.6) 10 (20.4) 25 (32.9) 21 (47.7) x3.3 12 (16.9) 11 (22.4) 18 (23.4) 17 (38.6) x6.5 SAS, Simpson–Angus Scale ; AIMS, Abnormal Involuntary Movement Scale ; HAS, Hillside Akathisia Scale.
4 mg might be too high in FES and that 2 mg might be drop-out rates of 20 % and 31 %, respectively (Emsley a more suitable dosage (Oosthuizen et al., 2004).
and the Risperidone Working Group, 1999), and the However, the results of the study by Oosthuizen et al.
12-wk comparison of olanzapine vs. haloperidol of (2004) seem to be somewhat in contrast to the results of 32 % and 46 %, respectively (Lieberman et al., 2003).
the 7-arm sertindole–haloperidol study by Zimbroff The relevance of the lower drop-out rate becomes ap- et al. (1997) which compared in a double-blind, parent when considering that the non-discontinuation randomized design three dosages of sertindole, three estimation by a survival analysis approach was dosages of haloperidol and placebo in a sample of recently used as the primary outcome measure in 497 patients. This non-FES study found no significant so-called ‘effectiveness ’ studies (Lieberman et al., difference between 4, 8 and 16 mg haloperidol.
One of the central, clinically relevant findings of Although the sample size was smaller than orig- this study is that the overall rate of discontinuation is inally planned, the difference in power between the significantly lower in the risperidone group (38.5 % vs.
planned and actual sample was very small (0.08), i.e.
54.1 % with haloperidol ; p<0.01). Other FES studies there was no relevant effect on the power of the study.
also found higher drop-out rates with haloperidol Thus, the study would not have reached different than the SGA comparator, although these did not conclusions if a larger number of patients had been always reach statistical significance. For example, included. Although a proportion of patients was not the 6-wk study of risperidone vs. haloperidol found drug naive, in contrast to some of the other FES studies Treatment of first-episode schizophrenia (Emsley and the Risperidone Working Group, 1999 ; bureau for AstraZeneca GmbH, Bristol–Myers Squibb Kopala et al., 1998), it is presumed that pre-treatment GmbH & Co. KG, GlaxoSmithKline, Janssen-Cilag with neuroleptics did not confound the main results of GmbH, Lilly Deutschland GmbH, Lundbeck GmbH, the study. Another limitation is that the inclusion of Novartis Pharma GmbH, Sanofi-Synthelabo GmbH/ in-patients only means that the results can only be generalized to outpatients with caution.
The efficacy and tolerability results of this study, together with the findings in the literature in first- and multi-episode patients, seem to indicate that it is much Addington D, Addington J, Schissel B (1990). A depression easier to demonstrate the advantage of SGAs in the rating scale for schizophrenics. Schizophrenia Research 3, dimension of EPS, where the indications are quite ro- bust and consistent across the studies. In contrast, Agresti A (2002). Categorical Data Analysis. New York : Wiley.
differences concerning the hypothesized broader effi- Andreasen NC (1982). Negative symptoms in schizophrenia.
cacy spectrum, especially in terms of negative or de- Definition and reliability. Archives of General Psychiatry 39, pressive symptoms, are not so pronounced and the results are quite varied (Mo¨ller, 2003, 2005). Although APA (1994). Diagnostic and Statistical Manual of Mental Disorders (4th edn). Washington DC : American Psychiatric FES patients show negative symptoms, they might not be as good a target population to show differences Coldham EL, Addington J, Addington D (2002). Medication between SGAs and FGAs as multi-episode patients, adherence of individuals with a first episode of psychosis.
and especially patients with enduring and predomi- Acta Psychiatrica Scandinavica 106, 286–290.
nant negative symptoms (Mo¨ller et al., 1994).
Csernansky JG, Mahmoud R, Brenner R (2002). A comparison of risperidone and haloperidol for theprevention of relapse in patients with schizophrenia.
New England Journal of Medicine 346, 16–22.
This study was conducted within the framework of the Emsley RA, the Risperidone Working Group (1999).
German Research Network on Schizophrenia, which is Risperidone in the treatment of first-episode psychoticpatients : a double-blind multicenter study. Risperidone funded by the German Federal Ministry for Education Working Group. Schizophrenia Bulletin 25, 721–729.
and Research BMBF (grants 01 GI 9932 and 01 GI Fleischhacker WW, Bergmann KJ, Perovich R, Pestreich LK, 0232). Risperidone and haloperidol and the blind- Borenstein M, Lieberman JA, Kane JM (1989). The ing and randomization procedure were provided by Hillside Akathisia Scale : a new rating instrument for Janssen-Cilag, Germany. The authors thank all those neuroleptic-induced akathisia. Psychopharmacology Bulletin who were involved in the study, especially the fol- lowing people and their colleagues : Fritz Henn, M.D., Frances A, Pincus HA, First MB (1994). The Global Paul Hoff, M.D., Ph.D., Andreas Marneros, M.D., Assessment of Functioning Scale (GAF). Diagnostic and Eckart Ru¨ther, M.D., Lutz Schmidt, M.D.
Statistical Manual of Mental Disorders – IV. Washington DC :American Psychiatric Association.
Gaebel W, Moller HJ, Buchkremer G, Ohmann C, Riesbeck M, Wolwer W, Von Wilmsdorff M, Bottlender R,Klingberg S (2004). Pharmacological long-term H.-J. Mo¨ller has received/is receiving research grants/ treatment strategies in first episode schizophrenia – study support from, serves as a consultant or is on the ad- design and preliminary results of an ongoing RCT visory board for, or is a member of the speaker bureau within the German Research Network on Schizophrenia.
for AstraZeneca, Bristol–Myers Squibb, Eli Lilly, Eisai, European Archives of Psychiatry and Clinical Neuroscience GlaxoSmithKline, Janssen-Cilag, Lundbeck, Merck, Novartis, Organon, Pfizer, Sanofi Aventis, Sepracor, Gaebel W, Riesbeck M, Wo¨lwer W, Klimke A, Eickhoff M, Servier, Wyeth. M. Gastpar is a consultant or member Von Wilmsdorff M, Jockers-Scheru¨bl MC, Ku¨hn K, of the advisory board for AstraZeneca, Lundbeck, Lemke M, Bechdolf A, et al. (2007). Maintenance Pfizer, Servier, Wyeth. F. Schneider has received/ treatment with risperidone or low-dose haloperidol infirst-episode schizophrenia : one-year results of a is receiving research grants/support from, serves randomized controlled trial within the German Research as a consultant or is on the advisory board for Network on Schizophrenia. Journal of Clinical Psychiatry 68, AstraZeneca, Janssen-Cilag, Lundbeck Neuroscience, Wyeth. W. Gaebel has received/is receiving research Green AI, Lieberman JA, Hamer RM, Glick ID, Gur RE, grants/support from, serves as a consultant or is on the Kahn RS, McEvoy JP, Perkins DO, Rothschild AJ, advisory board for, or is a member of the speakers’ Sharma T, et al. (2006). Olanzapine and haloperidol in first episode psychosis : two-year data. Schizophrenia Research Lieberman JA, Koreen AR, Chakos M, Sheitman B, Woerner M, Alvir JM, Bilder R (1996). Factors influencing Guy W (1976a). Abnormal Involuntary Movement Scale treatment response and outcome of first-episode schizophrenia : implications for understanding the Psychopharmacology (revised edn). Washington DC : US pathophysiology of schizophrenia. Journal of Clinical Department of Health, Education and Welfare.
Guy W (1976b). Clinical Global Impression (CGI). In : ECDEU Lieberman JA, Stroup TS, McEvoy JP, Swartz MS, Assessment Manual for Psychopharmacology (revised edn).
Rosenheck RA, Perkins DO, Keefe RS, Davis SM, Davis Washington DC : US Department of Health, Education and CE, Lebowitz BD, Severe J, Hsiao JK (2005). Effectiveness of antipsychotic drugs in patients with chronic Haas GL, Sweeney JA (1992). Premorbid and onset features schizophrenia. New England Journal of Medicine 353, of first-episode schizophrenia. Schizophrenia Bulletin 18, Lieberman JA, Tollefson G, Tohen M, Green AI, Gur RE, Hamilton M (1960). A rating scale for depression. Journal of Kahn R, McEvoy J, Perkins D, Sharma T, Zipursky R, Neurology, Neurosurgery and Psychiatry 23, 56–62.
Wei H, Hamer RM (2003). Comparative efficacy and safety Hedeker DR, Gibbons RD (2006). Longitudinal Data Analysis.
of atypical and conventional antipsychotic drugs in first- episode psychosis : a randomized, double-blind trial of Ja¨ger M, Riedel M, Messer T, Pfeiffer H, Laux G, Naber D, olanzapine versus haloperidol. American Journal of Gaebel W, Huff W, Schmidt LG, Heuser I, et al. (2006).
Psychopathological characteristics and treatment outcome Marder SR, Glynn SM, Wirshing WC, Wirshing DA, Ross of first episode compared with multiple episode D, Widmark C, Mintz J, Liberman RP, Blair KE (2003).
schizophrenic disorders. European Archives of Psychiatry and Maintenance treatment of schizophrenia with risperidone Clinical Neuroscience Published online : 10 October 2006.
or haloperidol : 2-year outcomes. American Journal of Kay SR, Fiszbein A, Opler LA (1987). The positive and McEvoy JP, Lieberman JA, Perkins DO, Hamer RM, Gu H, negative syndrome scale (PANSS) for schizophrenia.
Lazarus A, Sweitzer D, Olexy C, Weiden P, Strakowski Schizophrenia Bulletin 13, 261–276.
SD (2007). Efficacy and tolerability of olanzapine, Keefe RS, Seidman LJ, Christensen BK, Hamer RM, Sharma quetiapine, and risperidone in the treatment of early T, Sitskoorn MM, Lewine RR, Yurgelun-Todd DA, Gur psychosis : a randomized, double-blind 52-week RC, Tohen M, et al. (2004). Comparative effect of atypical comparison. American Journal of Psychiatry 164, 1050–1060.
and conventional antipsychotic drugs on neurocognition Merlo MC, Hofer H, Gekle W, Berger G, Ventura J, in first-episode psychosis : a randomized, double-blind Panhuber I, Latour G, Marder SR (2002). Risperidone, trial of olanzapine versus low doses of haloperidol.
2 mg/day vs. 4 mg/day, in first-episode, acutely psychotic American Journal of Psychiatry 161, 985–995.
patients : treatment efficacy and effects on fine motor Keefe RS, Sweeney JA, Gu H, Hamer RM, Perkins DO, functioning. Journal of Clinical Psychiatry 63, 885–891.
McEvoy JP, Lieberman JA (2007). Effects of olanzapine, Mo¨ller HJ (2000a). Definition, psychopharmacological basis quetiapine, and risperidone on neurocognitive function and clinical evaluation of novel/atypical neuroleptics : in early psychosis : a randomized, double-blind methodological issues and clinical consequences. World 52-week comparison. American Journal of Psychiatry Journal of Biological Psychiatry 1, 75–91.
Mo¨ller HJ (2000b). State of the art of drug treatment of Kopala L, Rabinowitz J, Davidson M (2003). Extra- schizophrenia and the future position of the novel/atypical pyramidal signs and symptoms (EPS) in recent onset antipsychotics. World Journal of Biological Psychiatry 1, schizophrenia : a comparison of risperidone and haloperidol. European Neuropsychopharmacology 13 (Suppl.
Mo¨ller HJ (2003). Management of the negative symptoms of schizophrenia. New treatment options. CNS Drugs 17, Kopala LC, Good KP, Fredrikson D, Whitehorn D, Lazier L, Honer WG (1998). Risperidone in first-episode Mo¨ller HJ (2004). Course and long-term treatment of schizophrenia : improvement in symptoms and pre- schizophrenic psychoses. Pharmacopsychiatry 37 (Suppl. 2), existing extrapyramidal signs. International Journal of Psychiatry in Clinical Practice 2, 19–25.
Mo¨ller HJ (2005). Antidepressive effects of traditional and Leucht S, Pitschel-Walz G, Abraham D, Kissling W (1999).
second generation antipsychotics : a review of the clinical Efficacy and extrapyramidal side-effects of the new data. European Archives of Psychiatry and Clinical antipsychotics olanzapine, quetiapine, risperidone, and sertindole compared to conventional antipsychotics and Mo¨ller HJ (2007). Clinical evaluation of negative placebo. A meta-analysis of randomized controlled trials.
symptoms in schizophrenia. European Psychiatry 22, Liang K-Y, Zeger SL (1986). Longitudinal data analysis using Mo¨ller HJ, Bottlender R, Groß A, Hoff P, Wittmann J, generalized linear models. Biometrika 73, 13–22.
Wegner U, Strauss A (2002). The Kraepelinian dichotomy : Treatment of first-episode schizophrenia preliminary results of a 15-year follow-up study on Swyzen W, De Smedt G (2005). Risperidone and functional psychoses : focus on negative symptoms.
haloperidol in first-episode psychosis : a long-term randomized trial. American Journal of Psychiatry 162, Mo¨ller HJ, Bottlender R, Wegner U, Wittmann J, Strauß A (2000). Long-term course of schizophrenic, affective Shrout PE, Fleiss JL (1979). Intraclass correlations : uses and schizoaffective psychosis : focus on negative in assessing rater reliability. Psychological Bulletin 86, symptoms and their impact on global indicators of outcome. Acta Psychiatrica Scandinavica 102 (Suppl. 407), Siegel SJ, Irani F, Brensinger CM, Kohler CG, Bilker WB, Ragland JD, Kanes SJ, Gur RC, Gur RE (2006). Prognostic Mo¨ller HJ, Schmid Bode W, Cording-To¨mmel C, Wittchen variables at intake and long-term level of function in HU, Zaudig M, von Zerssen D (1988). Psychopathological schizophrenia. American Journal of Psychiatry 163, 433–441.
and social outcome in schizophrenia versus affective/ Simpson GN, Angus JWS (1970). A rating scale for schizoaffective psychoses and prediction of poor outcome extrapyramidal side effects (EPS). Acta Psychiatrica in schizophrenia. Results from a 5–8 year follow-up. Acta Scandinavica 212 (Suppl. 44), 11–19.
Psychiatrica Scandinavica 77, 379–389.
Singer JD (1998). Using SAS PROC MIXED to fit multilevel Mo¨ller HJ, van Praag HM, Aufdembrinke B, Bailey P, models, hierarchical models, and individual growth Barnes TR, Beck J, Bentsen H, Eich FX, Farrow L, models. Journal of Educational and Behavioral Statistics 24, Fleischhacker WW, et al. (1994). Negative symptoms in schizophrenia : considerations for clinical trials. Working Strauss JS, Carpenter Jr. WT (1978). The prognosis of group on negative symptoms in schizophrenia.
schizophrenia : rationale for a multidimensional concept.
Psychopharmacology (Berlin) 115, 221–228.
Muller MJ, Marx-Dannigkeit P, Schlosser R, Wetzel H, Vickers AJ, Altman DG (2001). Analysing controlled trials Addington D, Benkert O (1999). The Calgary Depression with baseline and follow up measurements. British Medical Rating Scale for Schizophrenia : development and interrater reliability of a German version (CDSS-G). Journal Wolwer W, Buchkremer G, Hafner H, Klosterkotter J, Maier of Psychiatric Research 33, 433–443.
W, Moller HJ, Gaebel W (2003). German research network Oosthuizen P, Emsley R, Jadri TH, Keyter N (2004). A on schizophrenia-bridging the gap between research and randomized, controlled comparison of the efficacy and care. European Archives of Psychiatry and Clinical tolerability of low and high doses of haloperidol in the treatment of first-episode psychosis. International Journal of Young RC, Biggs JT, Ziegler VE, Meyer DA (1978). A rating Neuropsychopharmacology 7, 125–131.
scale for mania : reliability, validity and sensitivity. British Rummel C, Hamann J, Kissling W, Leucht S (2003). New Journal of Psychiatry 133, 429–435.
generation antipsychotics for first episode schizophrenia.
Zhang-Wong J, Zipursky RB, Beiser M, Bean G (1999).
Cochrane Database of Systematic Reviews. Issue 4. Art. no. : Optimal haloperidol dosage in first-episode psychosis.
Canadian Journal of Psychiatry 44, 164–167.
Sanger TM, Lieberman JA, Tohen M, Grundy S, Beasley Jr.
Zimbroff DL, Kane JM, Tamminga CA, Daniel DG, Mack C, Tollefson GD (1999). Olanzapine versus haloperidol RJ, Wozniak PJ, Sebree TB, Wallin BA, Kashkin KB, and treatment in first-episode psychosis. American Journal of the Sertindole Study Group (1997). Controlled, dose- response study of sertindole and haloperidol in the Schooler N, Rabinowitz J, Davidson M, Emsley R, Harvey treatment of schizophrenia. American Journal of Psychiatry PD, Kopala L, McGorry PD, Van HI, Eerdekens M,

Source: http://homepages.uni-tuebingen.de/florian.wickelmaier/pubs/MoellerRiedel2008IJNPP.pdf