How reviews covered the unfolding scientific story of gabapentin for bipolar disorder ☆ John W. Williams Jr., M.D., M.H.S. a, ⁎ , Leah Ranney, Ph.D. b , Laura C. Morgan, M.A. b , Lynn Whitener, Dr.P.H., M.S.L.S. b a Center for Health Services Research in Primary Care, Durham VAMC and Center for Clinical Health Policy Research, Duke University, Durham, NC 27705, USA b The Cecil G. Sheps Center for Health Services Research, The University of North Carolina at Chapel Hill, Chapel Hill, NC 27599, USA Received 12 November 2008; accepted 20 February 2009 Abstract Background: Despite the lack of randomized controlled trials (RCTs), gabapentin use increased rapidly in the 1990s for mental health conditions. Subsequent RCTs did not demonstrate efficacy for bipolar disorder (BD). We examined the characteristics of review articles to determine their potential role in the growth of gabapentin for BD. Methods: We searched MEDLINE, the International Pharmaceutical Abstracts and LexisNexis for review articles or commentaries examining the role of gabapentin for BD. Electronic searches were supplemented by manual searches of reference lists. Articles were abstracted for the types of evidence cited, source of evidence, the proportion of available RCTs cited and narrative blurbs discussing the role of gabapentin for BD. Review articles were classified as narrative versus systematic and positive, neutral or negative regarding the role of gabapentin in BD. Results: We included 27 review articles published between 1998 and 2008, but no commentaries met eligibility criteria. Most did not describe potential conflicts of interest or a funding source, and the 3 systematic reviews were of low quality. The 11 reviews published prior to the first RCT of gabapentin for BD cited uncontrolled trials or case series (n=9), basic science (n=6), chart reviews (n=3) or unpublished RCTs (n=2). Six recommended gabapentin, 3 were neutral and 2 were negative. The 16 articles published after the first gabapentin RCT continued to cite uncontrolled trials and basic science; only 5 cited all the available RCTs. However, more of these reviews (n=10) reached negative conclusions about the role of gabapentin for BD. Conclusions: Narrative and low-quality systematic reviews, principally those published prior to RCTs, may have contributed to the growth of gabapentin use for BD. High-quality systematic reviews are needed to inform clinicians and policymakers about the effectiveness of new treatments. Published by Elsevier Inc. Keywords: Gabapentin; Bipolar disorder; Systematic review; Evidence-based medicine 1. Introduction Gabapentin (Neurontin) is an antiepileptic drug (AED) licensed in June 1993 [1]. The medication underwent a very rapid growth in sales in the late 1990s in spite of generally narrow FDA indications as treatment for epilepsy and certain types of neuropathic pain such as post-herpetic neuralgia. Yet, by 1999, the second most common use of gabapentin was for mental health indications [2]. This occurred despite an absence of randomized controlled trials (RCTs) demon- strating the efficacy of gabapentin for mental health indications or FDA approval for mental health conditions. During the 1990s, multiple articles including case reports and reviews were published, examining the use of gabapentin for bipolar disorder (BD). This wave of publications coincided with the rapid growth of off-label use. Our group was funded by the Neurontin Executive Committee (a consortium of state attorneys general) to disseminate evidence-based information on the roles of Available online at www.sciencedirect.com General Hospital Psychiatry 31 (2009) 279 – 287 ☆ This supplement is supported by a consortium of state attorneys general, independent of any industry funding, which placed no restrictions on the content of this publication. Funding for this project was administered by the Vermont Office of the Attorney General. The authors report no competing interests. ⁎ Corresponding author. E-mail address: jw.williams@duke.edu (J.W. Williams). 0163-8343/$ – see front matter. Published by Elsevier Inc. doi:10.1016/j.genhosppsych.2009.02.006