Get Alert
Please Wait... Processing your request... Please Wait.
You must sign in to sign-up for alerts.

Please confirm that your email address is correct, so you can successfully receive this alert.

Articles   |    
Meta-Regression Analysis of Placebo Response in Antipsychotic Trials, 1970–2010
Ofer Agid, M.D.; Cynthia O. Siu, Ph.D.; Steven G. Potkin, M.D.; Shitij Kapur, M.B.B.S., Ph.D.; Eric Watsky, M.D.; Douglas Vanderburg, M.D., M.P.H.; Robert B. Zipursky, M.D.; Gary Remington, M.D., Ph.D.
Am J Psychiatry 2013;170:1335-1344. doi:10.1176/appi.ajp.2013.12030315
View Author and Article Information

Dr. Agid was a principal investigator for a Pfizer-sponsored study of ziprasidone and risk factors for metabolic syndrome; he has also received grant support or funding from or has served as a consultant for Janssen-Ortho (Johnson & Johnson), Eli Lilly U.S., Eli Lilly Canada, Novartis, Sepracor U.S., and Sunovion U.S. Dr. Siu has served as a consultant for Pfizer, Sunovion U.S. (Dainippon Sumitomo Pharma/Sepracor), Memory Pharmaceutical (Roche Laboratories), and Wyeth (Pfizer). Dr. Potkin has received grant support, funding, or honoraria or served as a consultant for the American Psychiatric Association, the Alzheimer’s Association, Baxter, Baylor College of Medicine, Bioline, Bristol-Myers Squibb, Cephalon, Cortex, Eisai, Eli Lilly, Forest Laboratories, Genentech, Janssen Pharmaceutical, Lundbeck, Merck, National Institute on Alcohol Abuse and Alcoholism, National Institute of Biomedical Imaging and Bioengineering, NIH National Center for Research Resources, Novartis, Organon, Otsuka, Pfizer Wyeth, Shire Development, Solvay, Sunovion, Roche, Takeda Pharmaceuticals International, Takeda Global Research and Development, University of Southern California, University of California San Francisco, and University of California San Diego. Dr. Kapur has received grant support from or served as consultant or speaker for AstraZeneca, Bioline, Bristol-Myers Squibb, Eli Lilly, GlaxoSmithKline, Janssen (Johnson & Johnson), Lundbeck, NeuroSearch, Otsuka, Pfizer, Roche, Servier, and Solvay Wyeth. Dr. Watsky and Dr. Vanderburg are full-time employees of Pfizer, Inc. Dr. Zipursky has served as a consultant to Roche, served on advisory panels for Amgen, Roche, and Sunovion, and received research contract support from Roche. Dr. Remington has received research support from the Canadian Diabetes Association, the Canadian Institutes of Health Research, Medicure, Neurocrine Biosciences, Novartis Canada, Research Hospital Fund–Canada Foundation for Innovation, and the Schizophrenia Society of Ontario and has served as a consultant or speaker for Novartis, Laboratorios Farmacéuticos Rovi, and Roche.

Editorial support was provided by Cynthia O. Siu, Ph.D., and Cecilia Rosenblatt, B.Pharm., of Data Power, Inc., and was funded by Pfizer, Inc.

Pfizer was not involved in the design and conduct of the meta-analysis, or in the systematic review, collection, management, and analysis of the data.

From the Centre for Addiction and Mental Health, Toronto; the Department of Quantitative Methodology, Data Power, Inc., Ringoes, N.J.; UCI Medical Center–UCI Brain Imaging Center, Orange, Calif.; Institute of Psychiatry, London; Pfizer Neuroscience, New York; and the Department of Psychiatry, McMaster University, Hamilton, Ontario.

Presented at the 48th annual meeting of the American College of Neuropsychopharmacology, Hollywood, Fla., Dec. 6–10, 2009; and at the 163rd annual meeting of the American Psychiatric Association, New Orleans, May 22–26, 2010.

Address correspondence to Dr. Agid (ofer_agid@camh.net).

Copyright © 2013 by the American Psychiatric Association

Received March 08, 2012; Revised October 10, 2012; Revised January 21, 2013; Accepted March 11, 2013.


Objective  Large placebo response presents a major challenge for psychopharmacologic drug development and contributes to the increasing failure of psychiatric trials. The objective of this meta-regression analysis was to identify potential contributors to placebo response in randomized controlled trials of antipsychotic treatment in schizophrenia.

Method  The authors extracted trial design and clinical variables from eligible randomized controlled trials (N=50) identified through searches of MEDLINE (1960–2010) and other sources. Standardized mean change (SMC) was used as the effect size measure for placebo response, based on change scores on the Brief Psychiatric Rating Scale or the Positive and Negative Syndrome Scale from baseline to endpoint (2 to 12 weeks).

Results  The results suggest significant heterogeneities (Q=387.83, df=49) in the magnitude of placebo response (mean SMC, −0.33, range −1.4 to 0.9) and in study quality. Both placebo SMC and study quality increased over time. Younger age, shorter duration of illness, greater baseline symptom severity, and shorter trial duration were significantly associated with greater placebo response, while country (United States compared with other countries) was not. More study sites, fewer university or Veterans Affairs treatment settings, and a lower percentage of patients assigned to receive placebo were associated with a greater placebo response, but these were not independent of publication year. Study quality affected the variability but not mean levels of placebo response.

Conclusions  This study identified important patient characteristics and trial design factors affecting the level of placebo response and hence the likelihood of detecting efficacy signals in randomized controlled trials. Future studies should test whether controlling these factors improves the detection of an antipsychotic effect.

Abstract Teaser
Figures in this Article

Your Session has timed out. Please sign back in to continue.
Sign In Your Session has timed out. Please sign back in to continue.
Sign In to Access Full Content
Sign in via Athens (What is this?)
Athens is a service for single sign-on which enables access to all of an institution's subscriptions on- or off-site.
Not a subscriber?

Subscribe Now/Learn More

PsychiatryOnline subscription options offer access to the DSM-5 library, books, journals, CME, and patient resources. This all-in-one virtual library provides psychiatrists and mental health professionals with key resources for diagnosis, treatment, research, and professional development.

Need more help? PsychiatryOnline Customer Service may be reached by emailing PsychiatryOnline@psych.org or by calling 800-368-5777 (in the U.S.) or 703-907-7322 (outside the U.S.).

FIGURE 1. Flow Diagram for the Stages of Data Extraction for a Meta-Regression Analysis of Randomized Controlled Trials of Antipsychoticsa

a PANSS=Positive and Negative Syndrome Scale; BPRS=Brief Psychiatric Rating Scale.

b Studies were excluded because they focused on add-on treatments (N=256), augmentation treatments (N=83), adjunctive medication (N=75), or long-acting injectables (N=138); had children or adolescents as study subjects (N=37); provided no information about the placebo group (N=929); analyzed previously published results (N=12); used rating scales other than the PANSS or BPRS (N=25); did not provide baseline rating scale scores (N=4); or had missing values for the rating scales used. Five long-term studies had available data on week 6 to week 12, which were extracted and included in the analysis.

FIGURE 2. Impact of Placebo Response Size on Drug-Placebo Differencea

a Mean improvement is defined as −1 × standardized mean change (SMC) from baseline in symptom score, so a positive value indicates improvement and a negative value indicates worsening in this figure. A significant positive association was observed between placebo response and active-drug response (t=9.24, df=58, p<0.001). Line A shows that drug improvement is 1.30 for trials with larger mean placebo improvement (0.75), compared with a lower drug improvement, 0.19, for trials showing worsening in the placebo arm (−0.75). Line B represents the identity line with the slope equal to 1, where the size of drug response (vertical axis) is equal to the placebo response (horizontal axis); vertical deviation from this identity line shows the drug-placebo difference. For fixed multiple-dose studies involving second-generation antipsychotics, we selected the treatment arm based on the optimum dosages as reported in the literature (aripiprazole, 10–30 mg/day; olanzapine, 10–20 mg/day; quetiapine, >250 mg/day; risperidone, 4–6 mg/day; sertindole, 16–24 mg/day; and ziprasidone, 120–160 mg/day) (11). If several arms of a drug fell within its optimum dosage ranges as reported in the literature (e.g., an aripiprazole study with 10 mg/day and 30 mg/day arms), the treatment arm representing the highest dosage was selected. This applies to iloperidone, 12–24 mg/day; sertindole, 20 mg/day; and bifeprunox, 20 mg/day. Eligible studies of lurasidone (80 mg/day compared with placebo), paliperidone (12 mg/day compared with placebo), and asenapine (10 mg/day) were single, fixed-dosage designs.

FIGURE 3. Placebo Response Compared With Patient Characteristics and Trial Design Factorsa

a In panel A, greater placebo response was observed in studies with shorter illness duration (χ2=14.66, p<0.001). In panel B, greater placebo response was observed in studies with higher baseline Clinical Global Impressions Scale (CGI) severity scores (χ2=10.53, p=0.001). In panel C, greater placebo response was observed in studies with shorter trial duration (slope=0.081; χ2=8.07, p<0.005).

FIGURE 4. Placebo Response and Omnibus Rating of Study Quality: Variability Comparisonsa

a The group of lower-quality studies had significantly greater variations and hence less consistency in standardized mean change (SMC) in score on the Positive and Negative Syndrome Scale or the Brief Psychiatric Rating Scale (from baseline) compared with the group with higher-quality studies (F=9.2, df=1, 59, p=0.004; Brown and Forsythe’s test for equal SMC variance as assessed by absolute deviation from the group median). All studies with exceptionally poor to moderately poor ratings were published in the period 1966–1970.

Anchor for Jump
TABLE 1.Findings for Meta-Regression Analysisa
Table Footer Note

a Coefficient (slope) indicates association between standardized mean change (SMC) in symptom score (Positive and Negative Syndrome Scale or Brief Psychiatric Rating Scale) and explanatory variables. Reduced symptom score from baseline would result in a negative SMC value. LR=likelihood ratio; CGI=Clinical Global Impressions Scale. Study year refers to year of publication.

Table Footer Note

b In this column, period refers to before versus after 1993. Intercept was allowed to differ according to study period.

Table Footer Note

c Each of these four variables was significant at p<0.05, based on likelihood ratio chi-square (SAS generalized linear model).

Table Footer Note

d Illness duration was derived from time since onset of psychotic symptoms based on clinical interview.

Table Footer Note

e Trial duration was based on nominal duration stated in the short-term (<12 weeks) protocol.

Table Footer Note

f Washout period was defined as the number of days between termination of previous antipsychotic treatment and the date that study treatment was initiated; it was categorized as ≤2 days, 3–7 days, or 8–84 days.

Table Footer Note

g Treatment setting was coded as follows: 1=community hospital (N=3), 2=university hospital (N=16), 3=state/provincial hospital (N=0), 4=mixed patients (N=31), 5=not mentioned (N=8), and 6=Veterans Affairs (VA) hospital (N=3). In the analysis, treatment setting was categorized as university or Veterans Affairs; mixed; or community or not mentioned). Missing data were imputed with median values from available data in the other studies.

Table Footer Note

h Number of treatment arms and percent in the placebo group were analyzed by study period using 1997 as cutoff, excluding the four studies that used a crossover design.



Zipursky  RB;  Darby  P:  Placebo-controlled studies in schizophrenia: ethical and scientific perspectives: an overview of conference proceedings.  Schizophr Res 1999; 35:189–200
[CrossRef] | [PubMed]
Temple  RJ:  When are clinical trials of a given agent vs placebo no longer appropriate or feasible? Control Clin Trials 1997; 18:613–620
[CrossRef] | [PubMed]
Laughren  TP:  The scientific and ethical basis for placebo-controlled trials in depression and schizophrenia: an FDA perspective.  Eur Psychiatry 2001; 16:418–423
[CrossRef] | [PubMed]
Kemp  AS;  Schooler  NR;  Kalali  AH;  Alphs  L;  Anand  R;  Awad  G;  Davidson  M;  Dubé  S;  Ereshefsky  L;  Gharabawi  G;  Leon  AC;  Lepine  JP;  Potkin  SG;  Vermeulen  A:  What is causing the reduced drug-placebo difference in recent schizophrenia clinical trials and what can be done about it? Schizophr Bull 2010; 36:504–509
[CrossRef] | [PubMed]
Chen  YF;  Wang  SJ;  Khin  NA;  Hung  HMJ;  Laughren  TP:  Trial design issues and treatment effect modeling in multi-regional schizophrenia trials.  Pharm Stat 2010; 9:217–229
[CrossRef] | [PubMed]
Kinon  BJ;  Potts  AJ;  Watson  SB:  Placebo response in clinical trials with schizophrenia patients.  Curr Opin Psychiatry 2011; 24:107–113
Leucht  S;  Arbter  D;  Engel  RR;  Kissling  W;  Davis  JM:  How effective are second-generation antipsychotic drugs? A meta-analysis of placebo-controlled trials.  Mol Psychiatry 2009; 14:429–447
[CrossRef] | [PubMed]
Laporte  JR;  Figueras  A:  Placebo effects in psychiatry.  Lancet 1994; 344:1206–1209
[CrossRef] | [PubMed]
Cressey  D:  Traditional drug-discovery model ripe for reform.  Nature 2011; 471:17–18
[CrossRef] | [PubMed]
Nutt  D;  Goodwin  G:  ECNP Summit on the future of CNS drug research in Europe 2011: report prepared for ECNP by David Nutt and Guy Goodwin.  Eur Neuropsychopharmacol 2011; 21:495–499
[CrossRef] | [PubMed]
Leucht  S;  Corves  C;  Arbter  D;  Engel  RR;  Li  C;  Davis  JM:  Second-generation versus first-generation antipsychotic drugs for schizophrenia: a meta-analysis.  Lancet 2009; 373:31–41
[CrossRef] | [PubMed]
Pilla Reddy  V;  Kozielska  M;  Johnson  M;  Vermeulen  A;  de Greef  R;  Liu  J;  Groothuis  GMM;  Danhof  M;  Proost  JH:  Structural models describing placebo treatment effects in schizophrenia and other neuropsychiatric disorders.  Clin Pharmacokinet 2011; 50:429–450
[CrossRef] | [PubMed]
Welge  JA;  Keck  PE  Jr:  Moderators of placebo response to antipsychotic treatment in patients with schizophrenia: a meta-regression.  Psychopharmacology (Berl) 2003; 166:1–10
Mallinckrodt  CH;  Zhang  L;  Prucka  WR;  Millen  BA:  Signal detection and placebo response in schizophrenia: parallels with depression.  Psychopharmacol Bull 2010; 43:53–72
Leucht  S;  Wahlbeck  K;  Hamann  J;  Kissling  W:  New generation antipsychotics versus low-potency conventional antipsychotics: a systematic review and meta-analysis.  Lancet 2003; 361:1581–1589
[CrossRef] | [PubMed]
McMahon  RP;  Kelly  DL;  Boggs  DL;  Li  L;  Hu  Q;  Davis  JM;  Carpenter  WT  Jr:  Feasibility of reducing the duration of placebo-controlled trials in schizophrenia research.  Schizophr Bull 2008; 34:292–301
[CrossRef] | [PubMed]
Kane  JM;  Leucht  S:  Unanswered questions in schizophrenia clinical trials.  Schizophr Bull 2008; 34:302–309
[CrossRef] | [PubMed]
Dunlop  BW;  Thase  ME;  Wun  CC;  Fayyad  R;  Guico-Pabia  CJ;  Musgnung  J;  Ninan  PT:  A meta-analysis of factors impacting detection of antidepressant efficacy in clinical trials: the importance of academic sites.  Neuropsychopharmacology 2012; 37:2830–2836
[CrossRef] | [PubMed]
Leucht  S;  Heres  S;  Hamann  J;  Kane  JM:  Methodological issues in current antipsychotic drug trials.  Schizophr Bull 2008; 34:275–285
[CrossRef] | [PubMed]
Potkin  S;  Agid  O;  Siu  C;  Watsky  E;  Vanderburg  D;  Remington  G:  Placebo response trajectories in short-term and long-term antipsychotic trials in schizophrenia.  Schizophr Res 2011; 132:108–113
[CrossRef] | [PubMed]
Merlo-Pich  E;  Alexander  RC;  Fava  M;  Gomeni  R:  A new population-enrichment strategy to improve efficiency of placebo-controlled clinical trials of antidepressant drugs.  Clin Pharmacol Ther 2010; 88:634–642
[CrossRef] | [PubMed]
Kocsis  JH;  Gerber  AJ;  Milrod  B;  Roose  SP;  Barber  J;  Thase  ME;  Perkins  P;  Leon  AC:  A new scale for assessing the quality of randomized clinical trials of psychotherapy.  Compr Psychiatry 2010; 51:319–324
[CrossRef] | [PubMed]
Becker  BJ:  Synthesizing standardized mean-change measures.  Br J Math Stat Psychol 1988; 41:257–278
Dunlap  WP;  Cortina  JM;  Vaslow  JB;  Burke  MJ:  Meta-analysis of experiments with matched groups or repeated measures designs.  Psychol Methods 1996; 1:170–177
Morris  SB:  Distribution of the standardized mean change effect size for meta-analysis on repeated measures.  Br J Math Stat Psychol 2000; 53:17–29
[CrossRef] | [PubMed]
Glass  GV:  Integrating findings: the meta-analysis of research, in  Review of Research in Education , vol 5. Edited by Schulman  LS.  Itasca, Ill,  FE Peacock, 1977
Cohen  J:  Statistical Power Analysis for the Behavioral Sciences , 2nd ed.  Hillsdale, NJ,  Lawrence Erlbaum Associates, 1988
DerSimonian  R;  Laird  N:  Meta-analysis in clinical trials.  Control Clin Trials 1986; 7:177–188
[CrossRef] | [PubMed]
Higgins  JPT;  Green  S:  Cochrane Handbook for Systematic Reviews of Interventions 4.2.6, in  The Cochrane Library .  Chichester, UK,  Wiley & Sons, 2006
Cleveland  WS;  Devlin  SJ:  Locally-weighted regression: an approach to regression analysis by local fitting.  J Am Stat Assoc 1988; 83:596–610
West S, King V, Carey TS, Lohr KN, McKoy N, Sutton SF, Lux L: Systems to Rate the Strength of Scientific Evidence (Evidence Report/Technology Assessment No 47, AHRQ Publication No 02-E016). Rockville, Md, Agency for Healthcare Research and Quality, April 2002
Allison  DB;  Loebel  AD;  Lombardo  I;  Romano  SJ;  Siu  CO:  Understanding the relationship between baseline BMI and subsequent weight change in antipsychotic trials: effect modification or regression to the mean? Psychiatry Res 2009; 170:172–176
[CrossRef] | [PubMed]
Yudkin  PL;  Stratton  IM:  How to deal with regression to the mean in intervention studies.  Lancet 1996; 347:241–243
[CrossRef] | [PubMed]
Altman  DG;  Andersen  PK:  Bootstrap investigation of the stability of a Cox regression model.  Stat Med 1989; 8:771–783
[CrossRef] | [PubMed]
Derksen  S;  Keselman  HJ:  Backward, forward, and stepwise automated subset selection algorithms: frequency of obtaining authentic and noise variables.  Br J Math Stat Psychol 1992; 45:265–282
Patel  NC;  Crismon  ML;  Miller  AL;  Johnsrud  MT:  Drug adherence: effects of decreased visit frequency on adherence to clozapine therapy.  Pharmacotherapy 2005; 25:1242–1247
[CrossRef] | [PubMed]
Taiminen  T;  Syvalahti  E;  Saarijarvi  S;  Niemi  H;  Lehto  H;  Ahola  V;  Salokangas  RKR:  Prediction of positive placebo response among chronic schizophrenic outpatients.  J Nerv Ment Dis 1996; 184:109–113
[CrossRef] | [PubMed]
Fournier  JC;  DeRubeis  RJ;  Hollon  SD;  Dimidjian  S;  Amsterdam  JD;  Shelton  RC;  Fawcett  J:  Antidepressant drug effects and depression severity: a patient-level meta-analysis.  JAMA 2010; 303:47–53
[CrossRef] | [PubMed]
References Container

CME Activity

There is currently no quiz available for this resource. Please click here to go to the CME page to find another.
Submit a Comments
Please read the other comments before you post yours. Contributors must reveal any conflict of interest.
Comments are moderated and will appear on the site at the discertion of APA editorial staff.

* = Required Field
(if multiple authors, separate names by comma)
Example: John Doe

Related Content
The American Psychiatric Publishing Textbook of Psychopharmacology, 4th Edition > Chapter 46.  >
The American Psychiatric Publishing Textbook of Psychopharmacology, 4th Edition > Chapter 44.  >
Dulcan's Textbook of Child and Adolescent Psychiatry > Chapter 26.  >
The American Psychiatric Publishing Textbook of Psychopharmacology, 4th Edition > Chapter 46.  >
The American Psychiatric Publishing Textbook of Geriatric Psychiatry, 4th Edition > Chapter 17.  >
Topic Collections
Psychiatric News
Read more at Psychiatric News >>
APA Guidelines