Advertisement
Research Article

Publication Bias in Antipsychotic Trials: An Analysis of Efficacy Comparing the Published Literature to the US Food and Drug Administration Database

  • Erick H. Turner mail,

    turnere@ohsu.edu

    Affiliations: Department of Psychiatry, Oregon Health & Science University, Portland, Oregon, United States of America, Department of Pharmacology, Oregon Health & Science University, Portland, Oregon, United States of America, Center for Ethics in Health Care, Oregon Health & Science University, Portland, Oregon, United States of America, Behavioral Health and Neurosciences Division, Portland Veterans Affairs Medical Center, Portland, Oregon, United States of America

    X
  • Daniel Knoepflmacher,

    Affiliation: School of Medicine, Oregon Health & Science University, Portland, Oregon, United States of America

    X
  • Lee Shapley

    Affiliation: School of Medicine, Oregon Health & Science University, Portland, Oregon, United States of America

    X
  • Published: March 20, 2012
  • DOI: 10.1371/journal.pmed.1001189

Reader Comments (2)

Post a new comment on this article

Please see correction

Posted by turnere on 27 Mar 2012 at 19:07 GMT

However, the FDA also reported that iloperidone was significantly inferior to the active comparator (ziprasidone in this trial) in most of the comparisons. (These p-values were not reported in the FDA review, so they do not appear in Table 3.) As with the above-mentioned trials, iloperidone's statistical inferiority to the active comparator was not reported in the corresponding journal article [34].
http://plosmedicine.org/article/info:doi/10.1371/journal.pmed.1001189#article1.body1.sec3.sec5.sec5.p1

Trial 3101 was in fact the one trial (of four) in which iloperidone was not statistically inferior to the active comparator.

Following a query regarding our study, we would like to correct an error in the article. In the section on outcome reporting bias for the drug iloperidone, the paragraph on Trial 3101 states: "However, the FDA also reported that iloperidone was significantly inferior to the active comparator (ziprasidone in this trial) in most of the comparisons. (These p-values were not reported in the FDA review, so they do not appear in Table 3.)"

In an earlier version of the manuscript, the phrase "most of the comparisons" was accurately used to describe only the first three iloperidone trials (3000, 3004, and 3005). However, during the manuscript revision process, this phrase became erroneously associated specifically with Trial 3101. An error was then introduced in Table 3 such that it incorrectly states, in reference to Trial 3101, "FDA reported as statistically inferior without P values".

Here we report the correct information from the FDA Drug Approval Package. (The page numbers referenced pertain to the FDA documents made available in the OHSU Digital Resources Library—see Methods section.) According to a March 27, 2009 memorandum by Division Director Tom Laughren (page 12 of 21): "The sponsor points out that iloperidone was shown to be equivalent to ziprasidone in study 3101, a placebo-controlled short-term study. Comment: I agree that this was the case. However, this was the only one of the 4 short-term placebo controlled trials in which this was the case."

Subsequently, in a May 6, 2009 memorandum by Office Director Robert Temple, M.D. summarized this and related issues as follows (page 2 of 7): " A not approvable letter was issued on 7/25/08 based on the applicant's failure to have submitted two adequate and well-controlled studies showing effectiveness and a strong tendency toward substantial inferiority to the comparator drugs used in the trials. Dr. Laughren's July 11, 2008 review and my July 21, 2008 review detailed the reasons. Both of us agreed that study 3101, a 4 week 3-arm comparison of iloperidone 24 mg (12 mg bid), ziprasidone 160 mg and placebo, the only study sponsored by the applicant, Vanda, clearly showed an effect on the PANSS (the PANSS and the BPRS are standard measures of severity of schizophrenia, regularly used in clinical trials) that was significantly greater than placebo and approximately equal to the effect of ziprasidone."

We clarify that other parts of the paper were not affected by this error. Figure 1 shows Trial 3101 as FDA-positive, which is consistent with the following statement from Dr Laughren's summary review (page 8 of 21): "We still considered study 3101 the only unambiguously positive study." Also, the header "Iloperidone trials (n=3)" still indicates the correct number of iloperidone trials showing evidence of outcome reporting bias (Trials 3000, 3004, and 3005). Finally, iloperidone's effect size was unaffected because the meta-analytic data underwent double data extraction and entry and excluded data from active comparators.

Competing interests declared: Please see competing interests listed in main article.