Publication Bias and the Pharmaceutical Industry: The Case of Lamotrigine in Bipolar Disorder

S. Nassir Ghaemi, MD, MPH; Arshia A. Shirzadi, DO; Megan Filkowski, BA


Medscape J Med. 2008;10(9):211 

In This Article


Given the above analysis, it still appears that negative outcomes, instead of being seen as important results in their own right, are treated by the pharmaceutical industry and many academic authors as if they represent signets, putting a seal on all further discussion, thus needing to be suppressed or explained away. Even though GSK has now made the studies available, the information has not been widely appreciated by the medical community.

The clinical relevance of the specific results with lamotrigine that are available through the GSK Web site is not insignificant. Two non-industry-funded studies found some benefit with lamotrigine in acute bipolar depression (1 published[28] and another recent one not yet published).[29] Yet the unpublished negative studies described here are needed to put such positive studies in context. Many clinicians and academics widely view lamotrigine as an effective treatment for, colloquially, "bipolar depression,"[30] even though the entire literature is less consistent on this point than many clinicians and academics appear to assume. This impression has been furthered by the selective publishing of one of the RCTs[25] with a positive secondary outcome on Montgomery Asberg Depression Rating Scale without emphasizing its negative outcome on the study's primary outcome on Hamilton Depression Rating Scale. Further, numerous review papers of lamotrigine in bipolar disorder cite the single published study as evidence of benefit in acute bipolar depression, but make no mention of the unpublished studies.[30,31,32] This problem takes nothing away from the evidence of lamotrigine's benefit in prophylaxis of bipolar disorder, particularly in prevention of depressive episodes, although it should be noted that the benefits seen in those prophylaxis studies are limited to patients who initially tolerated and appeared to respond to lamotrigine for acute mood episodes. But preventive benefit does not necessarily translate to acute efficacy, or vice versa.[33] Clinicians must learn to distinguish treatment of acute bipolar depression from its prevention. The lamotrigine data, positive and negative, could have been an important means of providing empirical evidence of the importance of this distinction.

The importance of academic access to such unpublished data can hardly be underestimated. For instance, the published study of lamotrigine in rapid cycling bipolar disorder is conservatively reported: the negative result in the primary outcome is acknowledged in the abstract, and the benefit in a secondary outcome in type II bipolar disorder may be clinically plausible and important. However, despite this acceptable presentation of this study, the nonpublication of the second RCT, where there was no indication at all of any benefit in type II rapid cycling bipolar disorder, leaves a mistaken impression in the published literature that lamotrigine is likely effective in that population. One out of 1 positive RCTs provides a different context than 1/2 positive RCTs. Instead, post-hoc exploratory analyses of the unpublished data are published, which, after extensive data analysis, find some possible benefit (somewhat more euthymia in lamotrigine-treated patients vs placebo utilizing a secondary outcome measure, the Life Chart Methodology).[34] While these exploratory analyses might be relevant, and even correct, the nonpublication of the straightforward findings of the original RCT leads to a distorted context whereby the psychiatric literature overestimates the benefit of the drug.

Further, it is well known that systematic reviews, especially meta-analyses, are vulnerable to major biases if negative or failed trials showing statistical nonseparation of a drug vs placebo or an active comparator are not available for analysis.[4,35] Further, narrative reviews of agents for bipolar disorder will provide the false impression that certain agents, such as lamotrigine, have been consistently effective whenever they were studied, since positive studies are evidently more likely to be reported at Web sites or in published articles. The practice of burying negative data in infrequently read review papers seems to be one approach to "publishing" such data, as exemplified by the single GSK-sponsored review paper which briefly summarized 5 negative studies with limited detail[26]; yet this approach is not likely to provide adequate information for clinicians and researchers. Even when the negative studies are made available in more detail, as with a recent meta-analysis of 5 negative acute bipolar depression studies with lamotrigine,[27] that kind of analysis should not replace earlier availability of each specific study. Further, since meta-analysis is an observational data analytic exercise, it too is not definitive, and is open to differing interpretations.[36] Too often, it can take the form of "statistical alchemy,"[37] whereby huge samples provide statistical significance to clinically meaningless differences. There is some danger of this kind of interpretation of the recent lamotrigine meta-analysis.[27]

Attention to publishing negative RCTs in a straightforward manner might be less necessary if clinicians were more critical readers of the research literature. However, many clinicians appear unable to exercise methodological critique in their evaluation of papers, thus not sufficiently realizing, for instance, that the 2 main published studies of lamotrigine in acute bipolar depression needed to be interpreted with much caution (one used a crossover design,[28] and the other emphasized a positive secondary outcome[25]). Thus, publishing negative data, though an improvement on the status quo, will not be a panacea; clinicians will still need more training in critically interpreting research.

The clinical relevance of the lamotrigine studies is notable: Taking the negative outcomes into account, as of now, one might say that this agent is reasonably effective in maintenance treatment of bipolar disorder, particularly in prevention of depression, among patients who initially tolerate and may benefit from acute lamotrigine treatment. It is proven ineffective in acute mania, rapid cycling, and acute bipolar depression. Exploratory analyses suggest that a small effect size of benefit might occur in some patients with bipolar depression,[27] perhaps among those more severely ill,[27] and the hypothesis of possible benefit in type II rapid cycling bipolar disorder was raised by one study but disconfirmed by another. This context of where the drug is effective, and where it is not, is vital for scientifically valid and ethically honest clinical practice and research.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.