On bitopertin and sharing data from clinical trials

While modern antipsychotics are doing an OK job improving the positive symptoms of schizophrenia – such as auditory hallucinations or fixed delusional ideas – there is a lot left to be desired in term of how these medications – or any psychotropics for that matter – work against negative symptoms. colorful-pillsThe negative symptoms of schizophrenia, including deficits in a variety of mental functions such as will power, ability to enjoy things, express emotions, or associate with others, have been an elusive target from a treatment perspective. In other words, available medications for schizophrenia might help a patient deal with “the voices,” but, because they do not also improve that patient’s negative symptoms, his ability to return to a meaningful and productive life is not substantially helped.

A promising new medication for negative symptoms — almost

Not surprisingly, any intervention that promises to improve negative symptoms makes the news. This was exactly what happened with bitopertin, a new Roche medication aimed at negative symptoms, which generated enough enthusiasm during its preliminary testing (Phase I and II clinical trials) to move to Phase III trials.

So, one wonders, why didn’t a brand new bitopertin study (ClinicalTrials.gov identifier NCT00616798), with a solid design and impressive action on negative symptoms get any attention – from the media, psychiatrist gurus, patient’s advocates, anyone? The study, which just came out in the flagship psychiatric journal (JAMA Psychiatry) generated no press/media coverage that I am aware off.

Actually, this rather surprising silent treatment is not an accident.

It turns out that this study has been preempted by a press announcement that Roche made this last January stating that  bitopertin did not hold up to its promise following its Phase III testing.  Meaning that… the just published JAMA Psychiatry study — the very study which generated enough enthusiasm to justify the seemingly well-deserved promotion bitopertin received further down the pipeline — is old, outdated news.

To summarize: on one hand, great positive results from a just published study, which, as it turns out, are actually quite old. On the other hand, negative results in Phase III testing, according to a news conference, but no publications to date.

The result: bitopertin is at the center of what is essentially a skewed and likely misleading state of affairs – at least from the point of view of official scientific validation via publication in peer reviewed journals.

How did we get here?

It seems that this bitopertin situation is the predictable result of a series of unfortunate events including:

  1. delays in reporting NCT00616798 study results : the original study, according to the official clinicaltrials.gov data, was completed in February 2010, more than 3 1/2 years prior to submission for  publication (original submission date August 30, 2013)
  2. an extremely lengthy peer review process: only a bit shy of 1/2 year: the original submission date was August 30, 2013; final revision received January 24, 2014; accepted January 24, 2014.
  3. No shared public data for any of the relevant studies that have been completed.

There is a fairly straightforward solution to prevent such complications: opening up the data from clinical trials to the public.alltrials_basic_logo2 Europe is already  pursuing an initiative to make reporting of all clinical trials data mandatory. This is an important step forward. But real progress will not occur unless and until patient level data is made available to the general public as close in time as possible to the date of data collection completion. Adrian Preda, M.D.

5 thoughts on “On bitopertin and sharing data from clinical trials”

  1. Thanks for the interesting story Adrian. It’s really important to hear how barriers to clinical trial transparency are affecting researchers.

    I did want to clarify a couple points though about the AllTrials campaign.

    First, while AllTrials started in the UK, it is a global campaign and has been joined by hundreds of organisations around the world.

    Second, we are not calling for the full public release of patient level data. We do want to see all trials past and present registered, with their full methodology and summary results posted in publicly accessible database. The new European law will ensure this happens for drug clinical trials in Europe after the law is enacted (probably by 2016). We still need to work toward a worldwide solution that includes all past trials though.

    Thanks again for speaking out on this issue. I encourage everyone to learn more and sign the petition at http://www.alltrials.net.


    Ian Bushfield
    Campaigns Support Officer
    Sense About Science (an AllTrials co-founding organisation)


    1. Dear Ian,

      All Trials is a very important initiative that has my full support. It is a great beginning – however as a researcher I believe that original data is always best and clearly better than any descriptive or inferential stats.
      Why? As final stats on anything rest on two things really: original data plus statistician expertise.

      Wouldn’t be great not to not worry about the latter?


  2. As an interested observer, in favour of greater access to scientific / medical information, it’s not clear to me why you chose to cite this as an example of a ‘barrier to clinical trial transparency’ that is ‘affecting researchers’ (in Ian’s words). In this case, the risk of moving the compound to phase III was taken by the company that ran the phase II trial. How could your solution of ‘opening up the data to the public’ have made any difference to that decision? Do you mean that their analysis of the phase II data was flawed? If so, the company appear to be the main losers (aside from the disappointment that may be felt by physicians and families). Which other ‘researchers’ have been affected by the availability (or otherwise) of the data? Many drugs fail at phase III, despite robust phase II data, and for various reasons. In this case, the risks were clearly substantial because the development of a specific and effective treatment for negative symptoms is somewhat uncharted territory, requiring exploration of new study designs and concepts, and development and validation of new outcome measures focused on negative symptoms.
    Also to call this a ‘skewed and likely misleading state of affairs’ seems overly dramatic. The ph II results were actually presented at ACNP in 2010 (see link) and elsewhere, generating interest and enthusiasm at the time, and as a result the data have been mentioned (in brief) in recently published reviews of drugs in development for schizophrenia. It does seem to have taken an excessively long time for the data to be published in full, but in pointing to the apparent delay of 3 1/2 years you ignore the distinct possibility that the article may previously have been submitted to (and rejected by) another journal or journals. Even high quality research is not guaranteed of acceptance by ‘top-tier’ journals at first asking, and each submission could take up to a year. As to the phase III trials, these were long-term (one-year) trials (compared with 8 weeks for the phase II trial) so it doesn’t seem particularly surprising that it has taken until now for the trials to reach completion. In fact, in a press release this month (see link) the company noted that three more phase III trials were still ongoing, but that two of the three had been discontinued on the basis of futility analyses. It therefore seems unrealistic to expect the phase III data should have been published in a peer-reviewed journal by now, given that the original press release was in January.


    1. @David: Thanks for your thoughtful comment. There is nothing wring with the way Roche worked with bitopertin.

      You are right: many Phase II ‘graduates” fail when moved to Phase III – it is an inherent risk when one moves from exploratory to confirmatory hypotheses.

      My point is about an entirely different issue: If one is to do a systematic review/metanalysis of a glycine transporter-1 inhibitors in schizophrenia or more specifically bitopertin, what would they find?

      A JAMA paper – with positive findings, then grey literature (conference posters) with negative findings and finally news reports about further negative findings. To get a sense of the significance of the negative findings one would need to contact Roche and ask for the data. Which then can be put together with the positive data and bingo.

      So, what this comes down do, is that to get the full picture, in one way or another, researchers depend on Roche decision about releasing the full data.

      This would be a non-issue if research sponsors would be mandated to release data (positive and negative) at the completion of a trial.


  3. Five months from initial submission to acceptance doesn’t seem all that terrible. In my experience at the Archives, I typically get about 60 days to first revision and 5-6 months to first notice of acceptance.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s