Antidepressants in the dock

Antidepressant drugs have been caught in the eye of two separate media storms over the past month.  

Both involve claims that manufacturers failed to make publicly available important data on the safety and efficacy of their products. Crucially, both also hinge on key studies that were carried out in the 1990s - long before the pharmaceutical sector introduced its current policy requiring greater openness about the results of clinical research.  

In the one case, the UK’s Medicines and Healthcare products Regulatory Agency (MHRA) completed its investigation into claims that GlaxoSmithKline (GSK) withheld information on the dangers to young people of using Seroxat, one of the selective serotonin reuptake inhibitor (SSRI) class of drugs. The regulator concluded that the company could have informed it much earlier of data showing an increased risk of attempted suicide in the under 18s, but that there were no grounds to proceed with a prosecution. GSK firmly denied that it had behaved improperly, saying it informed the MHRA as soon as data from clinical trials had become available and pointing out that the drug was never intended for use in this group. UK health minister Dawn Primarolo subsequently promised to tighten legislation on clinical trial disclosure. 

The second case focused on a meta-analysis carried out by Irving Kirsch of the University of Hull, and colleagues,¹ on the results of 47 clinical trials involving four different antidepressant products.  

Like previous reports, this study found that treatment was little more effective than placebo in reducing the symptoms of mild to moderate depression. In patients with very severe depression the difference between the responses of the treatment and placebo groups was more marked, although this was partly due to a reduction in the size of the placebo effect (see In the pipeline  for further analysis). 

Kirsch was able to gather information on many of the trials only after submitting an application for the US Food and Drug Administration (FDA) to release files under the Freedom of Information Act. He told Chemistry World  that he didn’t know if it would be possible to retrieve similar details from the European authorities, but noted that some unpublished trial data was withheld from the UK National Institute for Clinical Excellence during a review of the efficacy of antidepressant medicines. 

Registration rules 

However, data from any clinical trials undertaken since 2005 will be available, according to Richard Ley, head of communications at the UK Association of the British Pharmaceutical Industry. ’We have a policy that companies should not only register forthcoming trials but should also publish the results, whether positive or negative, within a year of them taking place.’ 

The information could appear on a company website or one of a number of international databases established by bodies such as the World Health Organization. Because of the multitude of possible publication routes, tracking down all trials within a particular clinical discipline might take some effort but the international trade association, the International Federation of Pharmaceutical Manufacturers (IFPMA), has established a central portal to make things easier, he said. 

Until now, registration of clinical trails has been entirely voluntary, although major journals such as the New England Journal of Medicine  refuse to publish studies that have not appeared in databases such as ClinicalTrials.gov, the largest single registry of clinical trials, which is managed by the US National Library of Medicine. 

Deborah Zarin, director of ClinicalTrials.gov, recently noted in Science²  that ’over 30 per cent of trials of 12 antidepressants submitted to the FDA for review, primarily those with negative results, have not been published 3 One effect of such "positive publication bias" is a boost in the apparent efficacy of these interventions’.

Open to openness 

But the registration rules are about to change. The FDA Amendments Act, enacted by Congress on 27 September 2007, mandates that all new or ongoing trials (other than Phase I) must be registered with ClinicalTrials.gov. And from September this year, all trials must publicly disclose their results through the website (while not necessarily revealing the underlying data) within one year of the last patient being examined. 

Ken Johnson, vice president of the US industry trade association PhRMA, says that member companies support the goals of the FDA Amendments Act. ’We are committed to making available clinical trial information to ensure that patients and their physicians have access to all relevant data from ongoing clinical testing. The process - which includes clinical trial registries and databases - should be transparent and accessible.’ 

Irving Kirsch agrees that the US system is potentially very useful for those carrying out independent research on the results of clinical trials. ’This is a great step forward - but there are still some gaps. One of these is the failure to mandate disclosure of the methodological details that are often important in evaluating the results of a study.’ 
John Bonner