Greenlab.npih.ucla.edu

Psychopharmacology (2006) 187:312–320DOI 10.1007/s00213-006-0428-x The neurocognitive effects of aripiprazole: an open-labelcomparison with olanzapine Robert S. Kern & Michael F. Green &Barbara A. Cornblatt & J. Randall Owen &Robert D. McQuade & William H. Carson & Mirza Ali &Ron Marcus Received: 26 November 2005 / Accepted: 3 May 2006 / Published online: 30 June 2006 Objectives The present paper is the first report, to our knowl- Rationale Cognitive deficits are a core feature of schizo- edge, on the neurocognitive effects of aripiprazole. Unlike phrenia. As a target of intervention, improvements in other second-generation antipsychotics, aripiprazole is a D2 cognition may lead to improvements in functional outcome.
and D3 receptor partial agonist. It is unknown what effects thisunusual pharmacological profile may yield on neurocognition.
Materials and methods The present open-label study Department of Psychiatry and Biobehavioral Sciences, included data on 169 patients with schizophrenia or schizoaffective disorder who were randomly treated with aripiprazole or olanzapine. Subjects received a neuro- cognitive battery at baseline, week 8, and 26.
VA Greater Los Angeles Healthcare Center (MIRECC 210A), Results The aripiprazole group had a significantly greater dropout rate than the olanzapine group. Neurocognitive data were reduced through a principal components analysis that yielded a three-factor solution. The factors were general cognitive functioning, executive functioning, and verbal learning. For general cognitive functioning, both groups Mental Illness Research, Education, and Clinical Center, improved from baseline and the effects were relatively stable over the 26-week protocol. There were no differential treatment effects. For executive functioning, neither group improved significantly from baseline. For verbal learning, the aripiprazole group improved significantly from baseline to the 8th and 26th week of assessment, and there was a between-group effect favoring aripiprazole over olanzapine North Shore-Long Island Jewish Health System, that was largely attributable to the differences in performance within the 8th week. Separate analyses were conducted for a measure of sustained attention (Continuous Performance Test–Identical Pairs). There were no differential treatment Conclusions The findings from this open-label study R. D. McQuade W. H. CarsonOtsuka America Pharmaceutical, Incorporated, suggest that the neurocognitive effects of aripiprazole are at least as good as those of olanzapine.
Keywords Schizophrenia . Aripiprazole . Olanzapine .
Otsuka Maryland Research Institute,Rockville, MD, USA Neurocognition . Learning . Memory . Dopamine . Agonists dopamine receptors, behaving as a functional antagonist inhyperdopaminergic states and as a functional agonist in hypodopaminergic states (Burris et al. Kikuchi et al.
Shapiro et al. ). Aripiprazole exhibits a high binding affinity for D2 and D3 receptors, a moderate affinity for D4 receptors, and a low affinity for D1 receptors.
Preclinical studies have also indicated that aripiprazole has a relatively high affinity for serotonin 5HT2A and 5HT1A receptors. It displays partial agonist activity at the 5HT1A receptor and antagonistic activity at the 5HT (Jordan et al. McQuade et al. ). It is not known what neurocognitive effects might be expected from an antipsychotic with this complex mechanism of action. This paper reports on the results of a multicenter study that assessed the neurocognitive effects of aripiprazole vsolanzapine using an open-label, randomized study design.
Cognitive deficits are now widely recognized as a corefeature of schizophrenia, largely independent of the psychot-ic symptoms of the illness (Gold The cognitive Subjects The study included 19 sites and 255 randomly domains affected include, among others, impairments in picked patients (aripiprazole=128 and olanzapine=127). All attention, working memory, verbal and visual learning, speed patients in the study were outpatients who met the of processing, language abilities, and reasoning and prob- Diagnostic and Statistical Manual of Mental Disorders–IV lem-solving (Aleman et al. Cornblatt et al. diagnostic criteria for schizophrenia or schizoaffective disorder, were between ages of 18 and 65, able to speak Nuechterlein and Dawson ; Nuechterlein et al. and understand English, were on a stable dose of an oral Saykin et al. ). It is important to note that there is a typical antipsychotic, risperidone, or quetiapine for at least growing awareness that these deficits are linked to impair- 1 month, and had not been hospitalized for psychiatric ments in community outcome and rehabilitation success treatment for at least 2 months before randomization.
Exclusion criteria included current suicidality, neurological A number of reports have examined the effects of second- disorder (e.g., epilepsy), acute or unstable medical condition, generation antipsychotic (SGA) medications on cognition in a clinically significant laboratory test value, gastrointestinal schizophrenia. Most report advantages when compared with resection or stapling that may interfere with study medica- conventional antipsychotic agents (Bilder et al. see tion absorption, and alcohol- or substance-dependence Keefe et al. ; Meltzer and McGurk for reviews).
within the past 3 months. Patients were also excluded if However, there are relatively few replicated findings of they had received aripiprazole in a prior clinical study, had specific treatment effects at the domain level (e.g., Green et taken a selective serotonin reuptake inhibitor within 2 weeks al. Haggar et al. ; Hoff et al. Kern et al.
before screening, or if they had taken an investigational drug Lee et al. ; Rossi et al. ). In general, the within 4 weeks before randomization. Patients were included effect sizes for SGAs compared with standard doses of in the analyses if they had a neurocognitive assessment at conventional agents were small to medium (Woodward et al.
baseline and at least one follow-up assessment (aripipra- zole=76 and olanzapine=93, total n=169). Table presents This is the first report, to our knowledge, of the neuro- the demographic and symptom ratings for the sample, cognitive effects of aripiprazole. Aripiprazole was shown to divided according to treatment group. Of the 255 patients be effective, safe, and well-tolerated for the treatment of enrolled in the study, 109 (43%) completed the entire 26- positive and negative symptoms in persons with schizophre- week protocol [60 subjects (47%) from the olanzapine group nia and schizoaffective disorder (Potkin et al. and to and 49 subjects (38%) from the aripiprazole group]. Of the have a low liability for extrapyramidal side-effects and a low 146 patients who discontinued the study, 21 were lost to incidence of clinically significant weight gain (Marder et al.
follow-up, 23 withdrew consent for personal reasons, 46 McQuade et al. It has a unique pharmacology discontinued due to adverse events, 39 discontinued due to among SGAs in that it is a partial agonist at D2 and D3 lack of clinical response or worsening of clinical symptoms of schizophrenia, 13 were excluded for noncompliance, and At baseline, subjects were administered the North 4 were excluded for violation of protocol. The overall American Adult Reading Test (NAART; Blair and Spreen difference in dropout between aripiprazole and olanzapine ), a widely used measure of premorbid intelligence. In was attributable to differences in the number of subjects who addition, subjects were administered the Vocabulary, Block discontinued the study due to personal reasons (14 vs 9), Design, and Information subtests from the Wechsler Adult adverse events (25 vs 21), lack of clinical response (22 vs Intelligence Scale (WAIS III; Wechsler to measure 17), and medication noncompliance (9 vs 4); not because of lost to follow-up (8 vs 13) or protocol violations (1 vs 3).
After complete description of the study to the subjects, written informed consent was obtained.
The study was designed to assess short- and longer-term The neurocognitive tests used in the study are as follows: neurocognitive treatment effects and included a baselineand two postbaseline assessments scheduled at 8 and 1. California Verbal Learning Test (CVLT; Delis et al.
26 weeks after randomization (1:1 ratio) to either 30 mg ). The CVLT is a test of learning and memory. A of oral aripiprazole or 15 mg of oral olanzapine (olanza- 16-item list is presented over five learning trials with pine-treated patients received 10 mg for the first 7 days and recall assessed after each trial. After the fifth trial, a new 15 mg from day 8 onwards). The selection of the list of 16 items is presented and recall assessed. Tests of olanzapine dosing regimen was based on published clinical cued and free recall for the first list are administered efficacy data (Beasley et al. and clinical practice at after short and long (20 min) delays. After the long delay the time of the initiation of the study. Randomization was recall, a 44-item recognition test is administered. The stratified according to prior antipsychotic therapy (conven- same form was used across assessment points. The tional vs atypical). Administration of nonstudy antipsychot- dependent measures selected for the current study were: ic medications was prohibited during the course of the (1) list A trials 1–5 total recall (a measure of learning), study. However, prior antipsychotic medications could be (2) semantic clustering ratio (a measure of learning tapered during the first 2 weeks of the study protocol to strategy), and (3) the discriminability index (a measure prevent destabilization. Patients on mood stabilizing med- ications before the onset of the study were maintained on 2. Benton Visual Retention Test–Revised (BVRT-R; these medications during the study. Anxiolytic medications Benton ). The BVMT-R is a test of visual memory.
(e.g., lorazepam) were permitted for use during the study The subject is shown ten cards that contain abstract for treatment of anxiety or insomnia. Twenty-one percent of geometrical designs. After presentation of each card, aripiprazole patients compared with 10% of olanzapine the subject is asked to draw it from memory. The test patients received anxiolytic medication as a concomitant includes different administration conditions. For the treatment. Anticholinergics (e.g., benztropine) were tapered current study, Administration A was used. This admin- and discontinued by the end of the second week of study istration involves a 10-s exposure with no delay. The treatment. At baseline, 12% of olanzapine patients and 10% test includes three alternate forms that were counter- of aripiprazole patients were receiving anticholinergic balanced across assessment points. The dependent medications for treatment of extrapyramidal symptoms (EPS). During the study, EPS were treated with medications 3. Wisconsin Card Sorting test (WCST; Heaton et al.
other than anticholinergics (e.g., propranolol). Administra- ). The WCST is a test of concept formation and the tion of antidepressant agents was prohibited after random- ability to maintain and shift cognitive set. The comput- ization. At baseline, 12% of both olanzapine and erized version used in the present study measures the aripiprazole patients were receiving antidepressant medica- ability of subjects to match a deck of stimulus cards to tion. Patients who elected to discontinue the study before one of four key cards. The cards can be matched completion of the 26-week trial were administered the according to color, shape, or number of figures. Subjects neurocognitive battery at the time of departure.
are provided little instruction about how to match thecards; only whether their attempt(s) were correct or not.
Procedure Neurocognitive and symptom assessments were Unbeknownst to the subject, the correct sorting rule conducted at baseline, week 8, and week 26. The neuro- changes after ten correct matches and the subject must cognitive battery assessed areas commonly impaired in adjust their sorting strategy to the new matching rule.
schizophrenia and included measures of attention, memory, The present protocol used an abbreviated 64-card executive functioning, and manual dexterity. The battery version. The primary dependent measures were (1) required approximately 90 min of administration time and number of categories, (2) percent conceptual level responses, and (3) percent perseverative errors.
4. Trail Making A and B (Army Individual Test Battery each of the two conditions. The primary dependent ). This test measures fine motor speed, visual measure for each condition was d-prime, which search, and the ability to alternate cognitive set. In Part indicates the response sensitivity for discrimination of A, subjects are asked to connect as quickly as possible a series of circled numbers that are haphazardly located ona sheet of paper. The numbered circles are to be connected in ascending order. Part B is similar to Aexcept that some circles contain numbers and others Severity of psychiatric symptoms was measured using the letters. The subject is instructed to connect the circles by Positive and Negative Symptom Scale (PANSS; Kay et al.
alternating between numbers and letters (e.g., 1-A-2-B- ). The PANSS includes three scales: positive scale, 3-C and so on). The dependent measures for Parts A and negative scale, and general psychopathology scale. The positive scale items include delusions, conceptual disorgani- 5. Verbal fluency (letter and category; Spreen and Benton zation, hallucinations, excitement, grandiosity, suspicious- Tests of verbal fluency assess verbal productiv- ness, and hostility. The negative scale items include blunted ity under selected search conditions. For the current affect, emotional withdrawal, poor rapport, passive pathetic study, verbal fluency was assessed under two condi- withdrawal, difficulty in abstract thinking, lack of spontaneity tions. In the letter (phonological) fluency condition, and flow of conversation, and stereotyped thinking. The subjects were asked to generate as many words as general psychopathology scale includes items for somatic possible that begin with a certain letter of the alphabet concern, anxiety, guilt, tension, mannerisms and posturing, (e.g., F, A, and S). In the category (semantic) fluency depression, motor retardation, uncooperativeness, unusual condition, subjects were asked to generate as many thought content, disorientation, poor attention, lack of exemplars of a particular category (e.g., animals, fruits, judgment, disturbance of volition, poor impulse control, and vegetables). All trials were 1 min in length. The preoccupation, and active social avoidance. The dependent dependent measure was the total number of correct measure was change from baseline for the PANSS total score.
words generated across the two conditions.
6. Letter–Number Sequencing subtest from the WAIS-III (Wechsler ). This test of auditory working memoryincludes two conditions. In the first condition (without reordering), subjects are presented strings of letters andnumbers of increasing length and asked to repeat them Statistical analyses were performed using Statistical Analysis back in the same order. The numbers/letters are System (SAS Data were initially examined for presented at the rate of approximately 1 s−1. In the normality of distribution and where violations occurred second condition (with reordering), strings of numbers transformations were performed. The neurocognitive data and letters are again presented, but this time subjects were reduced by means of a principal components analysis are asked to reorder them, first by saying aloud the (PCA) with the resultant factor scores used as the primary numbers in their ascending order and then the letters in outcome variables in the analyses. Data from the CPT-IP had alphabetical order. The dependent measures were a relatively large amount of invalid and missing data due to number of correct items/responses for each condition.
technical and administrative problems at a subset of the sites.
7. Grooved Pegboard test (Matthews and Klove ).
Because PCA requires complete cases for meaningful This test measures manual dexterity. The subject is results, these data for the CPT-IP were analyzed separately.
instructed to insert a series of grooved metal pegs into a We conducted two initial analyses to address the cognitive metal template with corresponding grooved slots as effects of aripiprazole vs olanzapine, and then two specific quickly as possible. Administration includes trials for follow-up analyses to consider the effects of clinical state right and left hand performance. The dependent and dropout on the study results. The primary analytic measure was time of completion for the dominant hand.
approach was the last observation carried forward (LOCF) 8. Continuous Performance Test–Identical Pairs version procedure, which was the original analytic plan and the way (CPT-IP; Cornblatt et al. ). The CPT-IP is a the data were presented at conferences. The data were then computerized measure of sustained attention. Subjects reanalyzed using a mixed model procedure for repeated are asked to respond to target stimuli that are flashed measures analysis of variance (Gueorguieva and Krystal briefly on a computer screen by lifting up their finger to confirm the LOCF results. Baseline factor scores from a mouse pad whenever a number repeats itself.
were entered as covariates. Efficacy was assessed by For the current study, the two- and four-digit conditions examining within- and between-group contrasts on the were used. Three hundred trials were administered for resultant factor scores at the weeks 8 and 26 assessments.
Table 2 Factor structure for neurocognitive battery Letter–no. sequencing (without reordering), total correct Letter–no. sequencing (with reordering), total correct Verbal Fluency (total score for phonologic+semantic conditions) Highest factor loadings for each variable appear in bold face type.
Treatment group changes in psychiatric symptoms and differential rates of attrition are sources of variance thatcould confound interpretation of study results. To address Examination of scores for normality of distribution showed the contribution of symptom changes, the LOCF analyses the Trail Making A and B scores to be negatively skewed.
were reconducted with the PANSS total score entered as a These scores were subsequently transformed using a log time-varying covariate. To address the effects of attrition, transformation. The neurocognitive battery included 13 we first examined group differences in dropout (i.e., selected variables. To condense the data and reduce Type I subjects with baseline plus at least one postbaseline error, we performed a PCA (Nunnally ). Factors assessment) using chi-square analyses. In response to the yielding eigenvalues greater than 1.00 were retained (i.e., differential dropout rates, the data were then reanalyzed they extracted at least as much as the equivalent of one using propensity score weighting (Hirano et al. original variable; Gorsuch Kaiser for varimax (orthogonal) rotation. The results of the PCA yielded a three- The propensity-weighting method is based on the factor solution. The resultant factors were labeled general assumption that subjects remaining in a study share, to cognitive functioning (factor 1), executive functioning differing degrees, characteristics with those who drop out, (factor 2), and verbal learning (factor 3; see Table ). A and these characteristics or predictor variables can be used number of neurocognitive variables including measures of to weigh scores to assess possible biasing effects of visual memory, manual dexterity, verbal recognition memo- dropout. Propensity weighting involves a two-step process.
ry, working memory, verbal fluency, and psychomotor speed First, logistic regression is used to identify variables that loaded on the general cognitive functioning factor. This predict dropout and generate predicted probabilities for factor accounted for approximately 40% of the overall each case remaining in the study. Second, the analyses are variance. The remaining two factors loaded primarily on done using the inverse of these probabilities as case two measures, the WCST and CVLT, respectively. These weights. Cases who remained in the study, but whose factors best represent individual measures and the method propensity scores indicate they were “likely to drop out” are variance tied to those measures. Consistent with the data thus similar to those who actually did drop out. Such cases reduction purpose of these analyses, the factor labels were are therefore weighted more heavily in the analyses. In the used to facilitate presentation of results and not to identify current study, demographic and baseline neurocognitive constructs. Factors 2 and 3 accounted for an additional 13 factor scores were examined in the first step to determine and 9% of the overall variance, respectively.
their relationship to dropout. Variables yielding a p value Initially, the data from the two treatment groups were <0.10 were retained in the logistic regression. The logistic examined for baseline comparability on demographics, regression model generated predicted probabilities for symptoms, measures of current and premorbid intellectual remaining cases (propensities). The inverse of those functioning, and baseline scores from the neurocognitive propensities, rescaled so that the sum of weights in each battery (Table ). The two groups were comparable in their treatment group equaled one, were then used as case demographic characteristics, premorbid and current intel- weights in the propensity-weighted data analyses.
lectual functioning, and symptom severity ratings.
Ariprazole
Olanzapine
Change from baseline
General Cognitive
Executive
Verbal Learning
Functioning
Functioning
Significant Improvement from baseline
Significant difference between groups
The results from the LOCF analyses are illustrated in df=1,166, and p=0.027). Follow-up of between-group Fig. An increase in factor score value indicated contrasts at the assessment points revealed the group improvement in neurocognitive functioning. For factor 1 difference to be due largely to a significant difference (general cognitive functioning), both aripiprazole and favoring aripiprazole at week 8 (F=5.58, df=1,166, and olanzapine showed significant improvement from baseline p=0.019) that was smaller and nonsignificant at week 26. In at week 8 (p=0.023 and 0.015, respectively) that fell to a sum, both the LOCF and mixed model analyses showed trend at week 26 (p=0.055 and 0.087, respectively). There within group improvement from baseline for aripiprazole at were no significant between-group differences at either the weeks 8 and 26 assessment points and both analyses week 8 or 26 comparisons. Results from the mixed model showed a significant group difference favoring aripiprazole.
analyses were similar to the LOCF analyses. Both groups The LOCF analyses indicated a more stable pattern over showed improvement from baseline to week 8 (aripipra- the 26-week period, whereas the more conservative zole: t=2.29, df=166, and p=0.024; olanzapine: t=2.68, mixed model analyses indicated group differences only at df=166, and p=0.008), and there were no overall group differences (F=0.00, df=1,166, and p=0.99). The primarydifference in the results for the mixed model analyses was that the group trends for improvement at week 26 seen in data from the two conditions of the CPT-IP were analyzed the LOCF analyses were no longer evident.
separately. For the two-digit condition, the results of the For factor 2 (executive functioning), the results from the LOCF analyses on the smaller CPT-IP data set indicated a LOCF analyses failed to show significant improvement significant improvement from baseline for aripiprazole at from baseline to week 8 or 26 for either group (all p>0.20) week 8 (p=0.034) and week 26 (p=0.027); but no and there were no between-group differences. The results significant changes for olanzapine. There were no differen- from the mixed model analyses yielded the same findings.
tial treatment effects at either week 8 or 26. For the four- For factor 3 (verbal learning), the results from the digit condition, there were no significant within- or LOCF analyses revealed that aripiprazole showed a between-group effects for either aripiprazole or olanzapine significant improvement from baseline at both week 8 (all p>0.30). For the mixed model analyses of the two-digit (p<0.0001) and week 26 (p<0.0001); olanzapine did not.
condition, aripiprazole showed a significant improvement at Examination of between-group differences at these assess- week 8 [t(109)=2.74 and p=0.007] that was weaker and ment points revealed a significant difference in favor of the nonsignificant at week 26 (p=0.16). There were no aripiprazole group compared to the olanzapine group at both week 8 (p=0.020) and week 26 (p=0.040). For themixed model analyses, aripiprazole showed a significant Analyses with PANSS as time-varying covariate When the improvement from baseline at both week 8 (t=4.94, df=166, LOCF analyses of the factor scores and CPT data were and p<0.0001) and week 26 (t=3.02, df=166, and p=0.003); conducted with the PANSS total score entered as a time- olanzapine showed a trend at week 8 (t=1.96, df=166, varying covariate for weeks 8 and 26 assessment points, the and p=0.052) but no significant improvement at week 26 results remained essentially the same. That is, the signifi- (p>0.25). There was a significant overall effect of cant results remained significant and nonsignificant results group favoring aripiprazole over olanzapine (F=4.97, Propensity-weighted analyses A chi-square analyses of schizophrenia and is frequently used in studies of neuro- treatment group×dropout indicated a significant overall cognitive predictors of functional outcome (Green ; effect (χ2=4.08 and p=.043), indicating greater dropout Green et al. Performance gains associated with for the aripiprazole group relative to the olanzapine group.
olanzapine treatment on this list learning measure were Examination of predictor variables indicated that older age, more modest than that noted in other studies (Harvey et al.
male gender, and lower scores on the baseline neuro- ; Purdon et al. ; Stip et al. The perfor- cognitive executive functioning factor were the strongest mance differences may be linked to differences in the predictors of dropout. To assess whether the differential measures, characteristics of the samples, dosing levels, or treatment effect favoring aripiprazole on the verbal learning length of treatment, and may have contributed to the factor was influenced by dropout, propensity weighted magnitude of the effect size favoring aripiprazole in the scores were calculated for this factor and these scores were current study. It should also be noted that the group entered into a mixed model analyses. The overall group difference favoring aripiprazole over olanzapine in the effect favoring aripiprazole vs olanzapine remained signif- mixed model analyses was largely due to differences at icant (F=4.80, df=1,166, and p=0.030). Like the mixed week 8 that were smaller and nonsignificant at week 26.
model analyses without the propensity weightings, the The waning strength of effects may be due to declining overall group effect was largely due to group differences sample size or changes in the composition of the sample.
at week 8 (F=5.81, df=1,166, and p=0.017) that were The difference between the LOCF and mixed model smaller and nonsignificant at week 26.
analyses may be due to larger performance differencesbeing carried forward in the LOCF analyses compared withthe parameter estimates used for missing cases in the mixed Higher levels of performance on list learning measures The primary aim of this open-label study was to compare such as the CVLT are linked to better community outcome, the neurocognitive effects of aripiprazole vs olanzapine in better social problem-solving ability, and rehabilitation schizophrenia and schizoaffective disorder outpatients over success (Green et al. ). There is optimism in the field a 26-week protocol. In general, the findings revealed that improvements in cognition such as verbal learning will aripiprazole and olanzapine to be comparable in their lead to improvements in real world functioning. However, it effects on neurocognition. On a factor that represented is unknown whether changes in cognition will lead directly general cognitive functioning, and accounted for most of to improvements in functional outcome in patients with the variance in the overall battery, both the aripiprazole schizophrenia. It may be more likely that changes in group and the olanzapine group showed small but signif- cognition will enable patients to acquire component skills icant improvements from baseline that were relatively (e.g., coping strategies and communication skills) necessary stable over the 26-week period. The two other factors, to succeed in the workplace or social environments, and the executive functioning and verbal learning, that loaded incorporation of these skills in their daily lives will lead to primarily on the WCST and CVLT, respectively, accounted improvements in work and social functioning.
for much less of the variance in the neurocognitive battery.
The study yielded a differential dropout for the two For the executive functioning factor, neither treatment drugs with more patients dropping from the aripiprazole group showed significant improvement from baseline or than the olanzapine group. The differences in dropout may any differential treatment effects. For the verbal learning be due to differences in dose-related side effects (e.g., factor, aripiprazole showed a differential treatment effect restlessness) or other random factors, but may be due to the compared to olanzapine that was primarily accounted for by open-label nature of the study and the possibility that differences at week 8. For the separate analyses conducted clinicians had a lower threshold for removing subjects from on the measure of sustained attention, no differential the arm with an investigational drug (aripiprazole was not an FDA-approved antipsychotic medication at the time of Regarding the differential treatment effect on verbal the study). The 30-mg/day of aripiprazole dose may have led learning, the magnitude of the effect size for aripiprazole vs to some of the excess dropout in this group. Premarketing olanzapine was modest, falling in the small to medium studies do not suggest any additional clinical benefit from range (d=0.36; Cohen ). The two CVLT variables that doses above 15 mg/day, and the higher dose may have been loaded highest on this factor included the ability to acquire associated with increased side effects. It is not possible to new verbal information (total recall trials 1–5) and the use know for sure what the results would have yielded had of efficient organizing strategies to facilitate recall (seman- dropout for the two drugs been comparable. However, when tic clustering ratio). The CVLT total recall measure is a the effects of dropout were estimated in the statistical analyses, commonly used measure to assess verbal learning in the primary findings on cognition remained the same.
The primary finding from this open label study is that design. Given the suggestion for a treatment-related aripiprazole is at least as good as olanzapine on general difference in verbal learning, future studies may wish to cognitive functioning, executive functioning, and attention, target this neurocognitive domain using a more specialized and may be better at improving verbal learning. Interpretation battery aimed at assessing a wide range of memory-related of the results on verbal learning requires consideration of a processes. If replicated, these findings would suggest that number of potential confounds. First, the open-label nature of the D2 partial agonism of aripiprazole might have specific the study should be considered. Findings from open-label cognitive benefits compared with the full antagonism of studies deserve a more conservative interpretation than double- blind studies. However, it is not obvious how a potential biasor difference in expectation would lead to a differential The authors would like to thank the patients and staff of the participating hospitals and clinics who made this study treatment effect in one cognitive domain, but not the others.
possible. Dr. Kern, Dr. Green, and Dr. Cornblatt have served as Second, the differential treatment effect could be related to consultants for Otsuka America Pharmaceutical and Dr. Green and differential effects of the two drugs on psychiatric symptoms.
Dr. Cornblatt have also served as consultants for Bristol-Myers Squibb However, the results were unchanged when the PANSS total Company. Dr. Kern, Dr. Green, and Dr. Cornblatt received no funds orother compensation for preparation of this manuscript. The data score was entered as a time-varying covariate in the analyses.
analyses were conducted by Jim Mintz, Ph.D., UCLA Department of Third, group differences at baseline could potentially bias the Psychiatry and Biobehavioral Sciences and Department of Veterans results in favor of one group over the other. Here, randomi- Affairs VISN 22 MIRECC. Funding for this research study was zation procedures resulted in the two groups being comparable provided by Otsuka America Pharmaceutical.
across demographic, clinical, and neurocognitive variables atbaseline. In addition, the contribution of baseline neuro-cognitive functioning was addressed in the analyses byentering baseline neurocognitive factor scores as covariates.
Fourth, the results cannot be due to differential coadministra-tion of adjunctive anticholinergic medications that can produce Aleman AA, Hijman R, de Haan EHF, Kahn RS (1999) Memory adverse effects on memory functioning because the study impairment in schizophrenia: a meta-analysis. Am J Psychiatr protocol required discontinuation of anticholinergics 2 weeks Army Individual Test Battery (1944) Manual of directions and scoring.
after randomization to drug. Fifth, dosing is a general War Department, Adjutant General’s Office, Washington, DC consideration in studies of this type. The level of olanzapine Beasley CM, Tollefson G, Tran P, Satterlee W, Sanger T, Hamilton S used in the present study falls within the therapeutic range, was et al (1996) Olanzapine versus placebo and haloperidol-acute comparable to that used in a multisite trial of the neuro- phase results of the North American double-blind olanzapinetrial. Neuropsychopharmacology 14:111–123 cognitive effects of olanzapine vs haloperidol and risperidone Benton AL (1974) Revised visual retention test, 4th edn. The (Purdon et al. ), and was held stable over the course of Psychological Corporation, San Antonio, Texas the trial. Sixth, interpretation of the factors derived from the Bilder RM, Goldman RS, Volavka J, Czobor P, Hoptman M, Sheitman PCA is limited in that we had different indices but not B, Lindenmayer JP, Citrome L, McEvoy J, Kunz M, Chakos M,Cooper TB, Horowitz TL, Lieberman JA (2002) Neurocognitive different measures for the executive functioning and verbal effects of clozapine, olanzapine, risperidone, and haloperidol in learning factors. Hence, these factors may represent method patients with chronic schizophrenia or schizoaffective disorder.
differences between these and other measures in the battery and not necessarily independent cognitive constructs. Sev- Blair JR, Spreen O (1989) Predicting premorbid IQ: a revision of the National Adult Reading Test. Clin Neuropsychol 3:129–136 enth, it is possible that sampling methods may have affected Burris KD, Molski TF, Xu C, Ryan E, Tottori K, Kikuchi T, Yocca the study results. Subjects who had received aripiprazole in a FD, Molinoff PB (2002) Aripiprazole, a novel antipsychotic, is a previous clinical trial were not included, but subjects who had high affinity partial agonist at human dopamine D2 receptors. J previously been maintained on olanzapine were. Therefore, it Cornblatt BA, Lenzenweger MF, Dworkin RH, Erlenmeyer-Kimling L is possible that the study included a number of subjects who (1985) Positive and negative schizophrenic symptoms, attention, had a poor previous clinical response to olanzapine and were and information processing. Schizophr Bull 11:397–408 seeking an opportunity to try a novel drug. Eighth, given that Cornblatt BA, Risch NJ, Faris G, Friedman D, Erlenmeyer-Kimling L the study did not use alternate forms for the CVLT, it is (1988) The continuous performance test, identical pairs version(CPT-IP). 1. New findings about sustained attention in normal- possible that the findings on verbal learning reflect enhanced practice effects and not direct benefits on learning per se.
Cohen J (1988) Statistical power analysis for the behavioral sciences.
However, it should also be noted that schizophrenia patients generally show minimal practice effects on list learning Delis DC, Kramer JH, Kaplan E, Ober BA (1987) The California Verbal Learning Test (manual). The Psychological Corporation, This study used an open-label study and it will be Gold JM (2004) Cognitive deficits as treatment targets in schizophrenia.
important to replicate these findings in a double-blind Goldberg TE, Weinberger DR, Berman KF, Pliskin NH, Podd MH Lee MA, Thompson P, Meltzer HY (1994) Effects of clozapine on (1987) Further evidence for dementia of the prefrontal type in cognitive function in schizophrenia. J Clin Psychiatry 55(Suppl B): schizophrenia? A controlled study of teaching the Wisconsin Card Sorting Test. Arch Gen Psychiatry 44:1008–1014 Marder SR, McQuade RD, Stock E, Kaplita S, Marcus R, Safferman Gorsuch RL (1983) Factor analysis, 2nd edn. Erlbaum, Hillsdale, NJ AZ, Saha A, Ali M, Iwamoto T (2003) Aripiprazole in the Green MF (1996) What are the functional consequences of neuro- treatment of schizophrenia: safety and tolerability in short-term, cognitive deficits in schizophrenia? Am J Psychiatr 153:321–330 placebo-controlled trials. Schizophr Res 61:123–126 Green MF, Marshall BD Jr, Wirshing WC, Ames D, Marder SR, Matthews CG, Klove H (1964) Instruction manual for the adult McGurk S, Kern RS, Mintz J (1997) Does risperidone improve neuropsychology test battery. University of Wisconsin Medical verbal working memory in treatment-resistant schizophrenia? Am McQuade RD, Burris KD, Jordan S, Tottori K, Kurahashi N, Kikuchi Green MF, Kern RS, Braff DL, Mintz J (2000) Neurocognitive deficits T (2002) Aripiprazole: a dopamine-serotonin system stabilizer and functional outcome in schizophrenia: are we measuring the (abstract). Int J Neuropsychopharmacol 5(Suppl 1):S176 McQuade RD, Stock E, Marcus R, Jody D, Gharbia NA, Vanveggel S, Green MF, Kern RS, Heaton RK (2004) Longitudinal studies of Archibald D, Carson WH (2004) A comparison of weight change cognition and functional outcome in schizophrenia: implications during treatment with olanzapine or aripiprazole: results from a randomized, double-blind study. J Clin Psychiatry 65(Suppl 18): Gueorguieva R, Krystal JH (2004) Move over ANOVA—progress in analyzing repeated-measures data and its reflection in papers Meltzer HY, McGurk SR (1999) The effects of clozapine, risperidone, published in the archives of general psychiatry. Arch Gen and olanzapine on cognitive function in schizophrenia. Schizophr Haggar C, Buckley P, Kenny JT, Freidman L, Ubogy D, Meltzer HY Nuechterlein KH, Dawson ME (1984) Information processing and (1993) Improvement in cognitive function and psychiatric symp- attentional functioning in the developmental course of schizo- toms in treatment-refractory schizophrenic patients receiving phrenia disorders. Schizophr Bull 10:160–203 Nuechterlein KH, Barch DM, Gold JM, Goldberg TE, Green MF, Harvey PD, Keefe RS (2001) Studies of cognitive change in patients Heaton RK (2004) Identification of separable cognitive factors in with schizophrenia following novel antipsychotic treatment. Am Nunnally JC (1967) Psychometric theory. McGraw-Hill, New York Harvey PD, Bowie CR, Loebel A (2006) Neuropsychological Potkin SG, Saha AR, Kujawa MJ, Carson WH, Mirza A, Stock E, normalization with long-term atypical antipsychotic treatment: Stringfellow J, Ingenito G, Marder SR (2003) Aripiprazole, an results of a six-month randomized, double-blind comparison of antipsychotic with a novel mechanism of action, and risperidone ziprasidone vs. olanzapine. J Neuropsychiatry Clin Neurosci 18: vs placebo in patients with schizophrenia and schizoaffective disorder. Arch Gen Psychiatry 60:681–690 Heaton RK, Chelune GJ, Talley JL, Kay GG, Curtiss G (1993) Purdon SE, Jones BDW, Stip E, Labelle A, Addington D, David SR, Wisconsin Card Sorting Test manual revised and expanded.
Brier A, Tollefson GD (2000) Neuropsychological change in early Psychological Assessment Resources, Odessa, Florida phase schizophrenia during 12 months of treatment with olanza- Heinrichs RW, Zakzanis KK (1998) Neurocognitive deficit in schizo- pine, risperidone, or haloperidol. Arch Gen Psychiatry 57:249–258 phrenia: a quantitative review of the evidence. Neuropsychology Rosenbaum PR, Rubin DB (1983) The central role of the propensity score in observational studies for causal effects. Biometrika 70:41–55 Hirano K, Imbens GW, Ridder G (2003) Efficient estimation of Rossi A, Mancini F, Stratta P, Mattei P, Gismondi R, Pozzi F et al average treatment effects using the estimated propensity score.
(1997) Risperidone, negative symptoms, and cognitive deficit in schizophrenia: an open study. Acta Psychiatr Scand 95:40–43 Hoff AL, Faustman WO, Wieneke M, Expinoza S, Costa M, SAS (1990) SAS Version 6.12. SAS Institute, Cary, NC Wolkowitz O, Csernansky JG (1996) The effects of clozapine Saykin AJ, Gur RC, Gur RE, Mozley PD, Mozley LH, Resnick SM, on symptom reduction, neurocognitive function, and clinical Kester DB, Stafiniak P (1991) Neuropsychological function in management in treatment-refractory state hospital schizophrenic schizophrenia: selective impairment in memory and learning.
inpatients. Neuropsychopharmacology 15:361–369 Jordan S, Koprivica V, Chen R, Tottori K, Kikuchi T, Altar CA (2002) Shapiro DA, Renock S, Arrington E, Chiodo LA, Liu L-X, Sibley DR, The antipsychotic aripiprazole is a potent, partial agonist at the Roth BL, Mailman R (2003) Aripiprazole, a novel atypical human 5-HT1A receptor. Eur J Pharmacol 441:137–140 antipsychotic drug with a unique and robust pharmacology.
Kaiser HF (1960) Directional statistical decisions. Psychol Rev 67:16–167 Kay SR, Opler LA, Fiszbein A (1992) Positive and negative syndrome Spreen O, Benton AL (1977) Neurosensory center comprehensive scale (PANSS) manual. Multi-Health Systems, North Tonawanda, NY examination for aphasia (NCCEA) (revised edition). University of Keefe RSE, Silva SG, Perkins DO, Lieberman JA (1999) The effects Victoria, Neuropsychology Laboratory, Victoria, British Columbia of atypical antipsychotic drugs on neurocognitive impairment in Stip E, Remington GJ, Dursun SM, Reiss JP, Rotstein E, MacEwan schizophrenia: a review and meta-analysis. Schizophr Bull 25: GW, Chokka PR, Jones B, Dickson RA, Canadian Switch Study Group (2003) A Canadian multicenter trial assessing memory Kern RS, Green MF, Marshall BD Jr, Wirshing WC, Wirshing D, and executive functions in patients with schizophrenia spectrum McGurk S, Marder SR, Mintz J (1999) Risperidone vs.
disorders treated with olanzapine. J Clin Psychopharmacol 23: haloperidol on secondary memory: can newer medications aid Wechsler D (1997) Wechsler Adult Intelligence Scale, 3rd edn. The Kikuchi T, Tottori K, Uwahodo Y, Hirose T, Miwa T, Oshiro Y, Morita Psychological Corporation, San Antonio, TX S (1995) 7-{4-[(2,3-Dichlorophenyl)-1-piperazinyl]butyoxy}- Woodward ND, Purdon SE, Meltzer HY, Zald DH (2005) A meta- 3,4-dihydro-2 (1H)-quinolinone (OPC-14597), a new putative analysis of neuropsychological change to clozapine, olanzapine, antipsychotic drug with both presynaptic and postsynaptic D2 quetiapine, and risperidone in schizophrenia. Int J Neuropsycho- receptor antagonist activity. J Pharmacol Exp Ther 274:329–336

Source: http://greenlab.npih.ucla.edu/Publications/Kern%20-%20Psychopharmacology%20-%202006.pdf