sign in sign up
<<
Home , Benzatropine

BRITISH JOURNAL OF PSYCHIATRY (2006), 189, 433^440. doi: 10.1192/bjp.bp.105.019307

Comparative effectiveness of second-generation explained to them and who signed an informed consent statement were included. and in acute Institutional review board approval was obtained for this study. {{ Pregnant or lactating women and schizophrenia patients with a medical condition in which pharmacotherapy would prove a significant ROBERT E. MCCCUE, RUBINA WAHEED, LEONEL URCUYO, clinical risk were excluded. Patients who GERALDINE ORENDAIN, MICHEL D. JOSEPH, RICHARD CHARLES had a clear history of response or lack and SYED M. HASAN of response to a particular drug and who, in the judgement of the treating psychiatrist, would best be Background ThereThereislittleinformation is little information Second-generation antipsychotic drugs have treated accordingly, were not entered into ononthe the comparative effectiveness of been heralded as a significant advance in the the study. Patients with a diagnosis of treatment of patients with schizophrenia. bipolar disorder, major depressive disorder second-generation antipsychotic agents. However, except for , none has or substance-induced psychotic disorder Aims Todetermine if any of five second- been conclusively shown to be superior in were also excluded. resolving the symptoms of schizophrenia. generation antipsychotics or haloperidolis Head-to-head studies are lacking. There is more effective in treating acutely ill little rational basis for selecting one over Study design patients with schizophrenia, another other than a patient’s history of Patients were admitted to one of the six schizoaffective disorder or response, lack of response or side-effects. general adult in-patient psychiatric units schizophreniform disorder. The purpose of this study was to determine based on bed availability, and this deter- if any of five second-generation antipsycho- mined the treating psychiatrist. All units MethodMethod Asampleof327newlyA sample of 327 newly tics was more effective in treating acutely ill have the same number of patients and staff- hospitalised patients with schizophrenia, ing, and are indistinguishable with respect admitted patients were randomised to schizoaffective disorder or schizophreni- to diagnoses and acuity of patients. Newly open-labeltreatment with , form disorder, and whether any of these admitted patients with a diagnosis of haloperidol, , , drugs had an advantage over haloperidol. schizophrenia, schizoaffective disorder or or for a minimum Two important features of this study were schizophreniform disorder were given in- of 3 weeks.Measures of effectiveness that it was designed to reflect clinical prac- formation about the study and asked to tice as a pragmatic (March etet participate and provide informed consent. wereimprovementinwere improvementin mental status sothatso that alal, 2005) and that it was not supported by Consenting patients were randomly the patient no longer required acute in- pharmaceutical companies. assigned to treatment with one of six patientcare, and changesin Brief antipsychotics: aripiprazole, haloperidol, Psychiatric Rating Scale (BPRS) scores. olanzapine, quetiapine, risperidone and METHOD ziprasidone. A randomised medication ResultsResults By the first measure, assignment list was prepared before the SampleSample haloperidol (89%), olanzapine (92%) and study using the randomisation website The study examined patients 18 years and http://www.randomization.com. Hospital risperidone (88%) were significantlymoresignificantly more older of either gender, who were newly staff with no clinical responsibilities and effective than aripiprazole (64%), admitted to the hospital’s psychiatric in- no knowledge of the patients oversaw the quetiapine (64%) andziprasidoneand ziprasidone (64%). patient service between January 2004 and assignment procedure and assigned medica- Changes in BPRS ratings were not February 2005. The 135-bed psychiatric tions in sequential order, strictly following significant among treatments. in-patient service treats acutely ill adult the randomised list. The treating psy- patients and is part of a 413-bed general chiatrist did not have access to this list. Conclusions Haloperidol, olanzapine hospital which serves an impoverished Both the patient and the treating psy- and risperidone are superior to urban population. Approximately 70% of chiatrist were aware of the antipsychotic admissions are involuntary. being prescribed. The treating psychiatrists aripiprazole, quetiapine and ziprasidone All patients in the study were diagnosed followed standardised dosing guidelines for the acute treatmentof psychosisin with schizophrenia, schizoaffective disorder based on the manufacturers’ recommenda- hospitalised patients with schizophrenia, or schizophreniform disorder according to tions, with the objective of obtaining a schizoaffective disorder or DSM–IV criteria (American Psychiatric maximum recommended dosage within Association, 1994). Patients with a history 1–21–2weeks.weeks. Patients were given at least schizophreniform disorder. of substance misuse were included if the a 3-weeka3-weektrial of the antipsychotic to deter- Declaration of interest None.None. above diagnoses were present. Patients mine its effectiveness. As needed doses of were included regardless of whether they haloperidol, lorazepam and diphenhydra- had recently taken antipsychotics before mine for agitation were permitted. Following admission. Only patients who understood current practice at the facility, these medica- {{See editorial, pp. 391^392, thisissue. the nature of the study when it was fully tions are generally administered together and

433433

Downloaded from https://www.cambridge.org/core. 02 Oct 2021 at 15:09:14, subject to the Cambridge Core terms of use. McCUE ET AL

intramuscularly for aggressive and threa- minimal involvement with managed-care variables were also examined with analyses tening behaviour. Oral doses of diphenhy- health insurance plans; as a result, decisions of variance. Categorical variables were ana- dramine were also administered, at the about discharge were made solely on lysed using a ww22 test. Logistic regression was patient’s request, for sleep. Benzatropine clinical grounds and not influenced by used to explore the effect of other indepen- could also be prescribed for extrapyramidal insurance arrangements. dent variables on the categorical outcome side-effects; it was the treating psychia- variable. All initial analyses used a two- trist’s decision whether to prescribe this Data collection tailedtailed aa level of 0.05. prophylactically or after side-effects devel- The two main measures of effectiveness oped. After the second week of treatment, used were the ability to discharge the RESULTSRESULTS an , mood stabiliser or anxio- patient from acute in-patient care and the lytic could be added at the psychiatrist’s total score on the Brief Psychiatric Rating From January 2004 to February 2005 a discretion for significant mood symptoms Scale (BPRS; Overall & Gorham, 1988). total of 584 admissions to the psychiatric or impulsivity. These medications are often Ratings were made at baseline, weekly up in-patient service with the diagnoses of considered essential in the acute treatment to 3 weeks, and at end-point. The end-point schizophrenia, schizoaffective disorder or of schizophrenia (McCue et aletal, 2003).,2003). was when the antipsychotic was determined schizophreniform disorder were screened If the treating psychiatrist assessed the to be effective or ineffective. for entry into the study; 368 were ran- patient to be improving on the medication, A clinician masked to the patient’s anti- domised. This included some patients who it was continued until the patient was psychotic regimen administered the BPRS. had previously participated in the study well enough to be discharged. On the Before the study began, this clinician had and who were rehospitalised during its other hand, if the patient showed no sig- 6 h of training per week for 2 months with course and were randomised a second time nificant improvement after at least 3 weeks the study’s senior authors (R.E.M. and L.U.) if they consented. For the purpose of this of treatment with the randomly assigned in using the BPRS. At the end of the training study, only the first randomised entry of antipsychotic, the treating psychiatrist period there was a sufficiently high correla- those entered more than once (nn¼41) was41)was could discontinue the medication and the tion of BPRS ratings. At the study’s mid- used for data analysis. Of the 327 patients patient would be withdrawn from the point, a revalidation of the clinician’s randomised, 8 were withdrawn from the study. A period of 3 weeks was chosen BPRS ratings was performed with the study for reasons unrelated to antipsychotic because treatment guidelines (American study’s senior authors (R.E.M. and L.U.). treatment and were not included in the data Psychiatric Association, 2004) have recom- Side-effects were recorded concurrently analysis. A total of 319 patients were mended waiting 2–4 weeks before changing with BPRS ratings by a clinician masked included in the analysis: of these, 301 had antipsychotic pharmacotherapy, although to the patient’s antipsychotic regimen. at least a 3-week trial of the antipsychotic there is evidence that the lack of improve- Side-effect data were elicited by spon- and in 18 cases participation was dis- ment in the first week or so of treatment taneous report and clinical evaluation. A continued because of side-effects or clinical predicts non-response (Correll et aletal,, clinician masked to the patient’s treatment deterioration (Fig. 1). 2003). At any time, if the treating psy- assessed Parkinsonian side-effects with the chiatrist believed that continuing treatment Simpson–Angus Scale (Simpson & Angus, with the selected antipsychotic would not Patient characteristics 1970) and with the Barnes be in the patient’s best interest (e.g. signifi- Akathisia Rating Scale (Barnes, 1989). Table 1 shows the baseline characteristics cant side-effects, medical instability and of the 319 patients whose data were used clinical deterioration), the medication was for analysis. No significant difference was Data analyses discontinued. found among the six groups in BPRS total AnAn a prioriapriori power analysis was performed score, gender, diagnosis, length of illness using G*POWER (Erdfelder et aletal, 1996).,1996). or comorbid substance misuse. There was Classification of outcome For six experimental groups, an aa of 0.05of0.05 a significant difference in the age of partici- The antipsychotic was classified as effective and a postulated modest effect size of pants among the six treatment groups: postpost if the patient’s mental status improved 0.25, the study needed a total sample size hochoc analyses using Fisher’s protected least sufficiently to no longer necessitate acute of 324 to have a power (17b7b) of 0.95.0.95.)of significant difference (PLSD) test showed in-patient care. Such patients were either Using these assumptions, the goal was to that patients in the olanzapine group were discharged to the community or moved to have each treatment cell contain approxi- significantly younger than patients in an alternative form of care. The anti- mately 54 patients. The software StatView the aripiprazole (PP¼0.004), risperidone psychotic was classified as ineffective if, in version 5.0 (SAS Institute, Cary, North ((PP¼0.03) and quetiapine (PP¼0.03) groups. the treating psychiatrist’s assessment, the Carolina, USA) was used for all other ana- In addition, patients given haloperidol were patient had made no significant improve- lyses. The primary hypothesis was that the significantly younger than those given aripi- ment after at least 3 weeks of treatment, six treatments would be differentially effec- prazole (prazole(PP¼0.03). As a result, age was and the drug was discontinued. If the medi- tive in treating acutely ill patients with included in analyses as a covariable. cation was discontinued before the end of a schizophrenia, schizoaffective disorder or 3-week trial owing to side-effects or signif- schizophreniform disorder. The effect of icant deterioration in the patient’s mental the antipsychotic on the main continuous Treatment characteristics state, it was also classified as ineffective. outcome variable (BPRS score) was ana- The maximum daily dosage of anti- The study site was a public hospital, with lysed with analysis of variance evaluating psychotic used in each treatment group the psychiatric in-patient service having change from baseline. Other continuous was as follows: aripiprazole, mean

434

Downloaded from https://www.cambridge.org/core. 02 Oct 2021 at 15:09:14, subject to the Cambridge Core terms of use. COMPARATIVE EFFECTIVENESS OF ANTIPSYCHOTICS

Fig. 11Fig. Progress of participants through the trial.

21.8 mg, s.d.¼8.1, range 10–45; haloperi- range for each medication (American of was significantly dif- dol, mean 16.0 mg, s.d.¼7.6, range 4–30; Psychiatric Association, 2004). ferent among the six groups, and there olanzapine, mean 19.1 mg, s.d.¼7.1, range7.1,range The use of additional medication was a significant medication66age interac- 5–40; quetiapine, mean 652.5 mg, s.d. throughout the study is shown in Table 2. tion effect (FF¼2.63, d.f.2.63,d.f.¼5,307,5,307, PP¼0.02).0.02). ¼280.8, range 50–1200; risperidone, mean There was no significant overall difference UsingUsing post hocposthoc analyses with Fisher’s PLSD 5.2 mg, s.d.¼1.8, range 2–9; ziprasidone, among the six treatment groups in the need test, patients treated with aripiprazole re- mean 151.2 mg, s.d.¼32.4, range 40–240. for haloperidol and lorazepam for quired significantly more diphenhydramine These fell within the recommended dosage aggressive or agitated behaviour. The use than patients treated with olanzapine

Ta b l e 11Tab Baseline characteristics of participants receiving randomised treatment with one of six antipsychotics

Antipsychotic treatment group TestAnalysis

Aripiprazole HaloperidolHaloperidol Olanzapine Quetiapine RisperidoneRisperidone ZiprasidoneZiprasidone d.f. PP nn¼5353 nn¼5757 nn¼5252 nn¼5050 nn¼57 nn¼5050

Age, years: mean (s.d.) 40.5 (12.6)(12.6)40.5 35.7 (10.8)33.8 (10.1)39.0 (11.0) 38.6 (12.9)(12.9)38.6 38.3 (11.9)(11.9)38.3 FF¼2.302.305,313 0.04 BPRS total score: mean (s.d.)41.3 (10.2)42.0 (11.3)41.1 (11.0)43.6 (10.4) 42.3 (9.0)(9.0)42.3 43.4 (11.0)(11.0)43.4 FF¼0.490.495,313 0.78 Length of illness, years: mean (s.d.)14.9 (11.4)12.2 (10.3) 11.7 (8.6)(8.6)11.7 14.5 (9.4)(9.4)14.5 13.1 (10.7)(10.7)13.1 12.9 (9.5)(9.5)12.9 FF¼0.815,313 0.54 Gender, nn (%)(%) ww22¼10.2550.07 5 0.07 Male 27 (51)42 (74)37 (71) 32 (64)(64)32 34 (60)(60)34 26 (52)26(52) FemaleFemale 26 (49)26(49) 15 (26)(26)15 15 (29)(29)15 18 (36)23 (40)24 (48) Diagnosis, nn (%) ww22¼11.4510 0.32 Schizophrenia 41 (77)43 (75)39 (75) 36 (72)(72)36 45 (79)45(79) 38 (76)38(76) Schizoaffective 12 (23)9(16) 9 (16) 9(17) 12 12(24) (24)8(14) 8 (14) 12 12(24) (24) Schizophreniform 0 (0) 5 (9)5(9) 4 (8)4(8)2 (4)4 (7) 0 (0)0(0) Substance misuse, nn (%) ww22¼6.226.22 550.29 Yes 20 (38) 22 (39)22(39) 25 (48)(48)25 16 (32)(32)16 17 (30)(30)17 14 (28)(28)14 No 33 (62)(62)33 35 (61)(61)35 27 (52)(52)27 34 (68)(68)34 40 (70)36 (72)

BPRS, Brief Psychiatric Rating Scale.

435435

Downloaded from https://www.cambridge.org/core. 02 Oct 2021 at 15:09:14, subject to the Cambridge Core terms of use. McCUE ET AL

Ta b l e 2 Psychotropic and medication used in addition to the randomised antipsychotic

Additional medication Antipsychotic treatment group TestAnalysis

Aripiprazole HaloperidolHaloperidol Olanzapine Quetiapine RisperidoneRisperidone ZiprasidoneZiprasidone d.f. PP nn¼5353 nn¼57 nn¼5252 nn¼5050 nn¼57 nn¼5050

Dosage: mean (s.d.) HaloperidolHaloperidol11 22.1 (39.9)(39.9)22.1 11.1 (18.5)12.5 (19.0) 20.7 (34.2)(34.2)20.7 9.0 (16.8)(16.8)9.0 17.3 (28.4)(28.4)17.3 FF¼1.581.5822 5,307 0.160.16 Lorazepam11 7.7 (13.7)4.7 (8.4)4.6 (7.7) 7.9 (11.9)(11.9)7.9 2.8 (5.3)7.3 (12.0) FF¼1.7322 5,307 0.130.13 Diphenhydramine11 104.2 (225.2) 51.2 (95.6)(95.6)51.2 35.6 (66.1)76.7 (181.1)72.8 (152.4)65.0 (154.3) FF¼3.5722 5,3070.004 Benzatropine33 0 (0)0(0) 1.9 (0.6)(0.6)1.9 0 (0)0(0)2.2 (1.1) 1.7 (0.5)(0.5)1.7 2.8 (1.1) FF¼0.910.9122 3,463,460.44 Patients receiving additional medication, nn (%) Mood stabiliser44 7(13)4 (7) 1 (2)1(2) 4 (8)4(8) 5 (9)5(9) 1 (2)1(2) ww22¼7.5750.18 5 0.18 Antidepressant44 0 (0)0(0)3(5) 3 (5) 2 2(4) (4) 1 (2)1(2) 0 (0)0(0) 1 (2)1(2) ww22¼5.6550.34 5 0.34 AnxiolyticAnxiolytic44 3(6)4 4(7) (7) 5 (10)5(10) 4 (8)4(8) 3 (5)3(5)5(10) 5 (10) ww22¼1.4950.91 5 0.91 Anticholinergic55 0 (0)0(0) 27 (47)27(47) 0 (0)0(0)5(10) 5 (10) 17 17(30) (30)5(10) 5 (10) ww22¼69.1169.115 550.0001

1. Total amount of medication in milligrams used as required for agitated or aggressive behaviour throughout the study period. 2. Age used as a covariate. 3. Daily dosage in milligrams. 4. Psychotic medication added after the second week of antipsychotic treatment for significant mood symptoms or impulsivity. 5. Benzatropine used on an ongoing basis for extrapyramidal side-effects.

((PP¼0.02). To examine the interaction ef- 1 patient, 1 and ziprasidone, but not significantly better fect, patients were divided into two groups patient, 1 patient) or anxiolytic than each other. In addition, aripiprazole, by the median age (38 years). For older pa- (clonazepam 11 patients, lorazepam 5 quetiapine and ziprasidone did not differ tients, there was no significant difference in patients, 3 patients, buspirone significantly from one another. There was diphenhydramine use among treatments 2 patients, diphenhydramine 2 patients, no significant difference among treatments ((FF¼1.28, d.f.1.28,d.f.¼5,155,5,155, PP¼0.27); however, alprazolam 1 patient) after the second week in the number of days until a patient’s treat- there was a significant difference for of treatment. ment was classified as effective. younger patients (FF¼3.53, d.f.3.53,d.f.¼5,152,5,152, Improvement in the BPRS total score PP¼0.005). Using the Fisher’s PLSD test, from baseline to study end-point did not younger patients taking aripiprazole Clinical outcome differ significantly among the six treat- required significantly more diphenhydramine Of 319 patients, 301 (94.4%) received ments. However, as a group, patients (mean 234.5 mg, s.d.¼316.6) than patients at least a 3-week trial of the randomised taking haloperidol, olanzapine or risperi- taking haloperidol (mean 70.0 mg, s.d.¼ antipsychotic. The antipsychotic was done tended to have a greater decrease in 120.1, PP¼0.002), olanzapine (mean 28.7 mg, prematurely discontinued in 18 patients BPRS total score (mean 15.6, s.d.¼11.1)11.1) s.d.s.d.¼66.3,66.3, PP550.0001), quetiapine (mean (5.6%) – in 14 (4.4%) as a result of side- than the group of patients who took 99.3 mg, s.d.¼227.3,227.3, PP¼0.02), risperidone effects and in 4 (1.2%) because of a aripiprazole, quetiapine or ziprasidone (mean 65.4 mg, s.d.¼149.5,149.5, PP¼0.002) and0.002)and worsening of the patient’s mental state. (mean 13.8, s.d.¼12.5;12.5; tt¼1.38, d.f.1.38,d.f.¼317,317, ziprasidone (mean 89.6 mg, s.d.¼193.9,193.9, Table 3 shows the outcome of each PP¼0.08, one-tailed). There was signifi- PP¼0.009).0.009). medication group. cantly greater improvement (tt¼8.55, d.f.8.55,d.f. There was a significant difference in the There was an overall significant dif- ¼317,317, PP550.0001) in the BPRS total scores use of benzatropine for extrapyramidal ference in effectiveness among the six of patients whose treatment was classified side-effects (Table 2); significantly more antipsychotics, with haloperidol, olanza- as effective (mean 17.5, s.d.¼10.5) com-10.5)com- patients treated with haloperidol or ris- pine and risperidone being the most pared with those with ineffective treatment peridone were prescribed benzatropine, effective. To examine the influence of age (mean 5.3, s.d.¼11.2).11.2). whereas no patient treated with aripipra- on the effectiveness of the antipsychotics, Changes in BPRS factors (Guy, 1976) zole or olanzapine was. For those patients age was included with medication in a from baseline to end-point were also taking this anticholinergic medication there logistic regression of clinical outcome. examined. Differences among the six was no significant difference in the mean Results of the logistic likelihood ratio test medications were not statistically signifi- daily dosage of benzatropine among the indicate that antipsychotic treatment cant for thought disturbance (FF¼0.70,0.70, treatments.treatments. ((ww22¼31.89, d.f.31.89,d.f.¼5,5, PP550.0001) had a signif- d.f.d.f.¼5,307,5,307, PP¼0.62; age as covariable), The six treatment groups did not differ icant effect on clinical improvement, but negativism (FF¼0.85, d.f.0.85,d.f.¼5,307,5,307, PP¼0.51;0.51; significantly in the addition of a mood sta- age (age(ww22¼0.20, d.f.0.20,d.f.¼1,1, PP¼0.65) did not. age as covariable), anxiety/depression biliser (divalproex 12 patients, gabapentin Pairwise comparisons by logistic regression ((FF¼0.98, d.f.0.98,d.f.¼5,307,5,307, PP¼0.43; age as co- 5 patients, lithium 2 patients, lamotrigine of each antipsychotic’s effectiveness are variable), hostility (FF¼0.76, d.f.0.76,d.f.¼5,307,5,307, 2 patients, oxcarbazepine 2 patients, given in Table 4. Again, haloperidol, olan- PP¼0.58; age as covariable) and activation carbamazepine 1 patient), antidepressant zapine and risperidone were significantly ((FF¼0.65, d.f.0.65,d.f.¼5,307,5,307, PP¼0.66; age as ( 3 patients, 1 patient, more effective than aripiprazole, quetiapine covariable).

436

Downloaded from https://www.cambridge.org/core. 02 Oct 2021 at 15:09:14, subject to the Cambridge Core terms of use. COMPARATIVE EFFECTIVENESS OF ANTIPSYCHOTICS

Ta b l e 33Tab Clinical outcome of participants analysed according to antipsychotic treatment group

Antipsychotic treatment group TestAnalysis

Aripiprazole HaloperidolHaloperidol Olanzapine Quetiapine RisperidoneRisperidone ZiprasidoneZiprasidone d.f. PP nn¼53 nn¼5757 nn¼5252 nn¼5050 nn¼5757 nn¼5050

Patient outcome, nn (%)(%) EffectiveEffective11 34 (64)(64)34 51 (89)(89)51 48 (92)(92)48 32 (64)(64)32 50 (88)50(88)32 (64) ww22¼30.4430.44 55 550.00010.0001 Ineffective 19 (36) 6 (11)6(11) 4 (8)4(8) 18 (36)(36)18 7(12) 7 (12) 18 18(36) (36) Lack of clinical response22 15 (28)(28)15 1 (2)1(2) 2 (4)2(4)18 (36)5 (9) 13 (26)(26)13 Side-effectsSide-effects33 3(6)5 5(9) (9) 0 (0)0(0) 0 (0)0(0) 2 (4)2(4) 4 (8)4(8) Deterioration44 1 (2)1(2) 0 (0)0(0) 2 (4)2(4) 0 (0)0(0) 0 (0)0(0) 1 (2)1(2) Change in BPRS total score: mean (s.d.)55 12.9 (12.3)16.4 (11.4)14.9 (11.3)14.2 (12.5) 15.4 (10.6)(10.6)15.4 14.2 (12.9)(12.9)14.2 FF¼1.1366 5,3070.34 Time to ‘Effective’, days: mean (s.d.)77 17.6 (10.5)18.6 (10.6)19.5 (13.1) 16.8 (8.0)(8.0)16.8 20.4 (13.5)(13.5)20.4 19.5 (8.5) FF¼0.2466 5,2355,2350.94

BPRS, Brief Psychiatric Rating Scale. 1. No longer needing acute in-patient care. 2. Minimal or no improvement after at least a 3-week trial. 3. Unable to complete a 3-week trial because of side-effects. 4. Unable to complete a 3-week trial because of worsening of mental state. 5. Change in BPRS total score from baseline to end-point. 6. Age used as a covariate. 7. Number of days of treatment until a patient’s medication was classified as effective.

Side-effects significantly larger proportion of patients DISCUSSION treated with either haloperidol (55%) or The following side-effects caused 14 ziprasidone (58%) reported side-effects, This study demonstrates differences in patients to leave the trial: nausea, dizziness whereas patients treated with aripiprazole effectiveness among six antipsychotics in and akathisia (aripiprazole); , reported significantly fewer (31%). Through- treating acutely ill hospitalised patients and akathisia (haloperidol); out the remaining 2 weeks of the study, with schizophrenia, schizoaffective disorder anxiety and (risperidone); and including at end-point, there was no signif- or schizophreniform disorder. Haloperidol, rash, akathisia, and derealisation icant difference among the six treatments olanzapine and risperidone were more (ziprasidone). The haloperidol and ziprasi- in the proportion of patients reporting effective than aripiprazole, quetiapine and done groups had the most withdrawals side-effects (week 2: ww22¼8.24, d.f.d.f.8.24, ¼5,5, ziprasidone. These results were obtained because of side-effects whereas the olanza- PP¼0.14; week 3: ww22¼2.89, d.f.2.89,d.f.¼5,5, PP¼0.72;0.72; with minimum bias, using a randomised pine and quetiapine groups had none. The end-point:end-point: ww22¼4.43, d.f.d.f.4.43, ¼5,5, PP¼0.49).0.49). design, without support from the phar- difference among the six treatments in rate There was no significant difference maceutical industry. The latter point is of withdrawals because of side-effects was among treatment groups in change in important as a study’s findings must be not statistically significant (ww22¼9.15,9.15, Simpson–Angus Scale ratings from baseline interpreted in light of the source of funding d.f.d.f.¼5,5, PP¼0.10).0.10). to end-point (FF¼0.61, d.f.0.61,d.f.¼5,307,5,307, PP¼0.69;0.69; (Als-Nielsen et aletal, 2003). The definition of The proportion of patients reporting age as covariable). In addition, there was effectiveness was a pragmatic one that mir- side-effects throughout the first 3 weeks no significant difference among treatment rored clinical practice: an ill patient is of the trial and at the end-point was groups in the change in score on the Barnes admitted, treated and, when sufficiently examined. After a week of treatment there Akathisia Rating Scale from baseline to improved, is discharged. In this study, an was a significant difference among treat- end-point (FF¼1.45, d.f.1.45,d.f.¼5,307,5,307, PP¼0.20;0.20; effective antipsychotic improved a patient’s ments (ments(ww22¼12.42, d.f.¼5,5, PP¼0.03). A0.03).A age as covariable). psychosis enough so that he or she could be

Ta b l e 4 Comparisons of the relative effectiveness of the six antipsychotics used (logistic regressions with age included as an independent variable)

Reference antipsychotic Comparison antipsychotic

Aripiprazole Haloperidol Olanzapine Quetiapine Risperidone OR (95% CI) PP OR (95% CI) PP OR (95% CI) PP OR (95% CI) PP OR (95% CI) PP

Haloperidol 0.20 (0.07^0.57)0.002 Olanzapine 0.14 (0.04^0.47)0.001 0.70 (0.19^2.64)0.60 Quetiapine 1.00 (0.45^2.24)1.00 4.87 (1.74^13.62) 0.0020.0026.95 (2.14^22.60)0.001 Risperidone 0.25 (0.09^0.65)0.005 1.21 (0.38^3.85) 0.750.751.72 (0.47^6.29)0.41 0.24 (0.09^0.66)0.005 Ziprasidone 0.99 (0.44^2.23) 0.990.994.85 (1.74^13.56)0.003 6.92 (2.13^22.48) 0.0010.0011.00 (0.44^2.26)0.99 4.02 (1.51^10.70)0.005

437437

Downloaded from https://www.cambridge.org/core. 02 Oct 2021 at 15:09:14, subject to the Cambridge Core terms of use. McCUE ET AL

discharged. This outcome is meaningful to for extrapyramidal side-effects. The meta- whereas no one taking olanzapine or quetia- both clinicians and their patients. analysis by Davis et aletal (2003) found pine did so. Patients in all six medication risperidone and olanzapine to be more groups reported having side-effects about Comparisons among second- efficacious than first-generation antipsy- one-third or more of the time. Patients tak- generation antipsychotics chotics, including haloperidol. Geddes et aletal ing haloperidol and ziprasidone had more Although treatment guidelines for schizo- (2000) found no advantage of the second- complaints at the beginning, but at end- phrenia (McEvoy et aletal, 1999; National generation antipsychotics over haloperidol point the distribution of side-effects was Institute for Clinical Excellence, 2002; for either efficacy or side-effects when an fairly even among the six treatments. American Psychiatric Association, 2004) optimal dosage of haloperidol of 6–12 mg Except for those elicited by rating scales, recommend starting with a second- per day was used. The mean daily dosage side-effects were obtained from the generation antipsychotic because of the of 16 mg in our study was higher than this. patient’s report. The validity of these improved side-effect profile, there is little Perhaps if lower dosages had been used in reported side-effects is open to question, as to guide clinicians in choosing among them. conjunction with prophylactic anticholiner- patients were often taking other medica- Studies that have compared risperidone gic medication, side-effects would have tions or had physical symptoms possibly and olanzapine have not been definitive. been less of a problem. The use of halo- unrelated to antipsychotic treatment. How- One study (Tran et aletal, 1997) compared peridol as an effective and inexpensive ever, these reported side-effects are relevant: olanzapine and risperidone in a double- treatment, even compared with olanzapine the patient’s perception that they were blind prospective trial and found some and risperidone, has had additional support caused by the antipsychotic would certainly advantage with olanzapine, whereas (Hunter(Hunter et aletal, 2003; Rosenheck et aletal, 2003;,2003; affect the individual’s present comfort and Conley & Mahmoud (2001) also compared KeefeKeefe et aletal, 2004; Kilian et aletal, 2004).,2004). future adherence to the drug regime. these two medications and found that Patients given aripiprazole or olanzapine risperidone was more efficacious. Both of Concomitant psychotropic required no concomitant anticholinergic these studies were supported by pharma- medication medication, whereas a small percentage of ceutical companies. A third study (Ho etet The use of as needed medication, including patients on quetiapine or ziprasidone and alal, 1999), without such support, found haloperidol, during the study period was an a significant minority of patients on halo- risperidone and olanzapine to be equally unavoidable complicating factor. For safety peridol or risperidone did need it. These re- effective in the acute treatment of schizo- reasons it was necessary for the staff to sults are consistent with each drug’s phrenia. Of these three studies, the first have at their disposal the conventional reported propensity to cause extrapyrami- two dealt with efficacy (how a drug treatments used for emergency situations. dal side-effects. No significant change was performs in controlled trials) and the third Although not to a degree of statistical found among treatments in ratings of studied effectiveness (how a drug works in significance, patients treated with aripipra- parkinsonism and akathisia using the real-world populations). Effectiveness studies zole, quetiapine and ziprasidone required Simpson–Angus Scale and the Barnes such as the one reported here may provide more haloperidol and lorazepam than Akathisia Rating Scale. An interpretation clinically useful information about pharma- patients in the other three medication of this result is that extrapyramidal side- cotherapy that is not obtainable from groups. However, this extra use of effects were not a problem for the majority studies of efficacy (Summerfelt & Meltzer, haloperidol, one of the more effective anti- of patients in this study and were resolved 1998).1998). psychotics in this trial, would probably have with anticholinergic medication if present. had a positive effect on the clinical outcome An exception is a small number of patients Comparison with haloperidol of patients treated with it. The as needed use taking haloperidol who had significant pro- We chose haloperidol as a comparator of haloperidolhaloperidol might also have obscured a blems with these side-effects. As fewer than because of its proven efficacy in treating differencein its effectiveness as the primary half of the patients given haloperidol were schizophrenia. Although there were more antipsychotic and the other two more also given anticholinergic medication, a withdrawals because of side-effects with effective drugs, olanzapine and risperidone. more consistent use of it prophylactically this drug, those who were able to tolerate Younger patients prescribed aripipra- might have prevented extrapyramidal side- it had a response rate of 98%. Trials zole required significantly more diphen- effects. Owing to the relatively short that have examined efficacy of the second- hydramine compared with younger patients treatment period of this study, the generation antipsychotics used in this study taking other medications. An interpretation important side-effects of weight gain, (Marder & Meibach, 1994; Beasley et aletal,, is that aripiprazole was much more activat- hyperglycaemia, lipid abnormalities and 1996; Arvanitis & Miller, 1997; Carnahan ing in younger patients. However, diphen- were not evaluated. et aletal, 2001; Kane et aletal, 2002) reported that hydramine is usually administered with these drugs were equal to first-generation haloperidol and lorazepam when as needed antipsychotics such as haloperidol. medication is used at the facility. It is also Study limitations Subsequent meta-analyses that have com- possible that younger patients taking ari- Qualifying any conclusion about effective- pared efficacy between second-generation piprazole required diphenhydramine more ness is the lack of differentiation among antipsychotics and haloperidol have been often for sleep. At this point, firm conclu- the antipsychotics with respect to the BPRS inconclusive. Leucht et aletal (1999) found sions cannot be drawn from this finding. total score. As there was a significant differ- a slight advantage of risperidone and ence in this variable between effectively and olanzapine over haloperidol for efficacy, Side-effects ineffectively treated patients, the BPRS and a larger advantage of risperidone, More patients taking haloperidol and zipra- total score did have validity as an indicator olanzapine and quetiapine over haloperidol sidone left the study because of side-effects, of clinical improvement. As a group, the

438

Downloaded from https://www.cambridge.org/core. 02 Oct 2021 at 15:09:14, subject to the Cambridge Core terms of use. COMPARATIVE EFFECTIVENESS OF ANTIPSYCHOTICS

more effective antipsychotics were asso- ROBERT E. McCUE, MD, RUBINA WAHEED, MD, LEONEL URCUYO, MD,GERALDINE ORENDAIN, MD, ciated with a greater mean change in BPRS MICHEL D. JOSEPH, MD,RICHARDMD, RICHARD CHARLES, DO, SYED M. HASAN, MD, Department of Psychiatry,Woodhull total score than the less effective ones, Medical and Mental Health Center,Center,Brooklyn, Brooklyn, New York,York,USA USA although not to a statistically significant degree. A likely possibility is that our study Correspondence:DrCorrespondence: Dr Robert E.McCue,Department of Psychiatry,Woodhull Medical and Mental Health might not have had sufficient power to Center, 760 Broadway,Brooklyn, NewYork11206,USA.New York 11206,USA. Email: mccuer@@nychhc.org detect differences among the six treatments. (First received 11November 2005, final revision 9 May 2006, accepted 2 June 2006) AA post hocposthoc power analysis of this compari- son showed a power of 0.39. There might also have been aspects of the patient’s clin- ical condition relating to discharge that were not reflected in the BPRS total score; for the majority of the patients by the the other three antipsychotics tested, which for example, haloperidol, olanzapine and criteria used in this study, the marginal ben- may account for their superior effectiveness risperidone might have been more success- efit of a longer trial would probably be (Kapur(Kapur et aletal, 2000). Olanzapine and risper- ful at controlling disturbed behaviour and minimal.minimal. idone were better tolerated in the short as a result patients treated with these would A psychiatrist who was not masked to term than haloperidol; however, greater have been more readily discharged. How- the antipsychotic being used made the deci- use of anticholinergic medication with ever, if sedation alone accounted for the sion that a patient no longer needed acute haloperidol would probably have improved results then quetiapine – one of the most in-patient care, a major outcome variable. its tolerability. This study did not address sedating of the six antipsychotics – would However, this decision was not made by long-term effectiveness and side-effects. have had an advantage. In addition, no dif- the treating psychiatrist in isolation and The number of patients with schizophrenia, ference was found among the medications was the product of input from the patient, schizoaffective disorder and schizophreni- in changes in the BPRS factors, including the patient’s family and other members of form disorder who require acute treatment hostility and activation. Although the defi- the treatment team. During the study peri- is substantial and more studies with nition of effectiveness used in this study od there was no significant difference in minimal bias are greatly needed to assist may be a reflection of improvement in only the length of stay of patients of the 14 psy- clinicians in making thoughtful treatment some of the clinical manifestations of chiatrists who participated in the study decisions.decisions. schizophrenia, improving the condition of ((FF¼1.50, d.f.1.50,d.f.¼13,164,13,164, PP¼0.12). There is patients so that they can be discharged also the possibility that, as a result of bias, ACKNOWLEDGEMENTS sooner remains a clinically important the psychiatrists waited longer with some objective.objective. of the drugs before classifying them as in- The authors thank Gabrielle Marcelin-Capelli, MSW, Adriana Mateoc, MD,Babatunde Asemota, MD, and The presence of a statistically sig- effective, thereby increasing the chance of Lisardo Augustin, MD, for assistance in data collec- nificant – although not clearly clinically a favourable outcome. However, in addi- tion to there being no difference in the time tion.Wetion. We are also grateful to Jeannette Torres-Baillie significant – difference in age among for her invaluable organisational and clerical efforts. the treatment groups may indicate that needed for a drug to be effective, there was no significant difference in the number of there was unsuccessful randomisation. The REFERENCES patients were assigned treatment from a list days until a treatment was classified as inef- fective (fective(FF¼0.82, d.f.0.82,d.f.¼5,48,5,48, PP¼0.54). An-0.54).An- prepared before the study began and by Als-Nielsen, B., Chen,W., Gluud, C., et aletal (2003)(2003) someone who had no knowledge of the other limitation of the study is that Association of funding and conclusions in randomized patients, including their age, so it is unlikely although standard recommended dosages drug trials. JAMAJAMA,, 290,,921^928. 921^928. that this represented an intentional bias. were used, optimal therapeutic dosing for American Psychiatric Association (1994) Diagnostic the newer second-generation antipsychotics and Statistical Manual of Mental Disorders (4th edn) Although the differences in age cannot be (DSM^IV).Washington, DC: APA. explained, age was not a significant factor is still uncertain. As aripiprazole, quetia- American Psychiatric Association (2004) Practice in determining effectiveness. pine and ziprasidone are further studied, perhaps the recommended therapeutic do- guideline for the treatment of patients with A major weakness of this study is its schizophrenia, second edition. American Journal of questionable generalisability. The results, sages of these drugs will be revised and, PsychiatryPsychiatry,, 161161,1^56. hence, their effectiveness. although robust, may reflect idiosyncrasies Arvanitis, L. A. & Miller, B. G. (1997) Multiple of clinical practice by the psychiatric in- fixed doses of ‘Seroquel’ (quetiapine) in patients with acute exacerbation of schizophrenia: a comparison patient service at our facility. Also, the with haloperidol and placebo. Biological Psychiatry,, 42,, definition of effectiveness was relevant to Clinical implications 233^246. hospitalised patients. The effectiveness of Based on these findings, haloperidol, risper- Barnes,Barnes,T. T. R. (1989) A rating scale for drug-induced these medications in out-patients might be idone and olanzapine are more effective akathisia. British Journal of Psychiatry,, 154,672^676., 672^676. different. By American guidelines, a 3-week antipsychotics for the acute treatment of Beasley,C.Beasley, C. M. J., Tollefson, G., Tran,Tran,P., P., et aletal (19 9 6) minimum trial would be sufficient to hospitalised patients with schizophrenia, Olanzapine versus placebo and haloperidol: acute phase determine an antipsychotic’s effectiveness; schizoaffective disorder or schizophreni- results of the North American double-blind olanzapine trial.trial. Neuropsychopharmacology,, 1414,111^123. however, this might be considered too short form disorder. These drugs are reasonable Carnahan, R. M., Lund, B. C. & Perry, P.J.P. J. (2001) for European psychiatric practice, where a first choices unless the patient’s history Ziprasidone, a new drug. minimum of 6 weeks is needed (National suggests otherwise. Haloperidol, risperi- Pharmacotherapy,, 2121, 717^730.,717^730.

Institute for Clinical Excellence, 2002). done and olanzapine are also more potent Conley, R. R. & Mahmoud, R. (2001) A randomized Since all of the antipsychotics were effective antagonists of -2 receptors than double-blind study of risperidone and olanzapine in the

439439

Downloaded from https://www.cambridge.org/core. 02 Oct 2021 at 15:09:14, subject to the Cambridge Core terms of use. McCUE ET AL

treatment of schizophrenia or schizoaffective disorder. placeboinpatientswithschizophreniaandplacebo in patients with schizophrenia and McCue, R. E.,Waheed, R. & Urcuyo, L. (2003) American Journal of Psychiatry,, 158158, 765^774.,765^774. schizoaffective disorder. Journal of Clinical Psychiatry,, 63,, Polypharmacy in patients with schizophrenia. Journal ofofJournal 763^771. Clinical Psychiatry,, 6464, 984^989.,984^989. Correll, C. U., Malhotra, A. K., Kaushik, S., et aletal (2003)(2003) Early prediction of antipsychotic response in Kapur, S., Zipursky, R., Jones, C., et aletal (2000)(2000) McEvoy, J. P., Scheifler, P.P.L. L. & Frances, A. (1999) schizophrenia. American Journal of Psychiatry,, 160160,, Relationship between dopamine D22 occupancy, clinical Expert consensus panels for schizophrenia: treatment of 2063^2065.2063^2065. response, and side effects: a double-blind PET study of schizophrenia. Journal of Clinical Psychiatry,, 6060,1^80. Davis, J. M., Chen, N. & Glick, I. D. (2003) AA first-episode schizophrenia. American Journal of meta-analysis of the efficacy of second-generation PsychiatryPsychiatry,, 157157,514^520. National Institute for Clinical Excellence (2002) Core Interventions in the Treatment and Management of antipsychotics. Archives of General Psychiatry,, 60,, Keefe, R. S. E., Seidman, L. J., Christensen, B. K., etet 553^564.553^564. Schizophrenia in Primary and Secondary Care. London:.London: alal (2004)(2004) Comparative effect of atypical and NICE. Erdfelder, E., Faul, F. & Buchner, A. (1996) conventional antipsychotic drugs on neurocognition in G*POWER: a general power analysis program. Behavior first-episode psychosis: a randomized, double-blind trial Overall, J. E. & Gorham, D. R. (1988) The BriefTheBrief Research Methods, Instruments, and Computers,, 28,,1^11. 1 ^ 11. of olanzapine versus low doses of haloperidol. AmericanAmerican Psychiatric Rating Scale (BPRS): recent developments in Journal of Psychiatry,, 161161, 985^995. ascertaining and scaling. Psychopharmacology Bulletin,, 24,, Geddes, J., Freemantle, N., Harrison, P., et aletal (2000) 97^99.97^99. Atypical antipsychotics in the treatment of Kilian, R., Dietrich, S., Toumi, M., et aletal (2004)(2004) schizophrenia: systematic overview and meta-regression Quality of life in persons with schizophrenia in out- Rosenheck, R., Perlick, D., Bingham, S., et aletal (2003) analysis. BMJBMJ,, 321,1371^1376. patient treatment with first- or second-generation Effectiveness and cost of olanzapine and haloperidol in antipsychotics. Acta Psychiatrica Scandinavica,, 110,, Guy,Guy,W. W. (1976) ECDEU Assessment Manual for the treatment of schizophrenia. JAMAJAMA,, 290, 2693^2702. 108^118.108^118. Psychopharmacology.RevisedDHEWPub.(ADM).. Revised DHEW Pub. (ADM). Simpson, G. M. & Angus, J.W. S. (1970) Aratingscale Rockville, MD: National Institute of Mental Health. Leucht, S., Pitschel-Walz, G., Abraham, D., et aletal for extrapyramidal side effects. Acta Psychiatrica (1999)(19 9 9) Efficacy and extrapyramidal side-effects of the Ho,B.-C.,Miller,D.,Nopoulos,P.,Ho, B.-C., Miller, D., Nopoulos, P., et aletal (1999)(1999) AA Scandinavica,, 212212,11^19. comparative effectiveness study of risperidone and new antipsychotics olanzapine, quetiapine, risperidone, and sertindole compared to conventional antipsychotics olanzapine in the treatment of schizophrenia. Journal ofofJournal Summerfelt,W.T. & Meltzer,Meltzer,H.Y. H.Y. (1998) Efficacy vs.vs.Efficacy and placebo. Schizophrenia Research,, 3535, 51^68.,51^68. Clinical Psychiatry,, 60, 658^663.,658^663. effectiveness in psychiatric research. Psychiatric Services,, Hunter, R. H., Joy, C. B., Kennedy, E., et aletal (2003) March, J. S., Silva, S. G., Compton, S., et aletal (2005)(2005) 49, 834^835.,834^835. Risperidone versus medication for The case for practical clinical trials in psychiatry. Tran,P.V.,Tran, P.V.,Hamilton, Hamilton, S. H., Kuntz, A. J., et aletal (19 9 7) schizophrenia. Cochrane Library,issue2.Oxford:Update, issue 2.Oxford: Update American Journal of Psychiatry,, 162162, 836^846.,836^846. Double-blind comparison of olanzapine versus Software.Software. Marder, S. R. & Meibach, R. C. (1994) Risperidone in risperidone in the treatment of schizophrenia and other Kane, J. M., Carson,W.H., Saha, A. R., et aletal (2002)(2002) the treatment of schizophrenia. American Journal of psychotic disorders. Journal of Clinical Efficacy and safety of aripiprazole and haloperidol versus PsychiatryPsychiatry,, 151151, 825^835. Psychopharmacology,, 1717,407^418., 407^418.

440440

Downloaded from https://www.cambridge.org/core. 02 Oct 2021 at 15:09:14, subject to the Cambridge Core terms of use.