NPC Archive Item: Newer antidepressants produce clinically significant benefits over placebo only in the severely depressed

NOTE – This is an archive post from the NPC and has not been updated since first publication. Therefore, some hyperlinks may no longer be working.
MeReC Rapid Review

NPC Logo

Kirsch I, et al. Initial severity and antidepressant benefits: A meta-analysis of data submitted to the Food and Drug Administration. PLoS Med 2008 5(2): 45. doi:10.1371/journal.pmed.0050045

What is the background to this?
Antidepressants have undoubtedly helped many people with depression.  However, they are not recommended in NICE guidance on management of depression for the initial treatment of mild depression, because the risk–benefit ratio is poor.  The authors of this meta-analysis wanted to analyse all the information, including both published and unpublished studies, on several newer drugs (fluoxetine, paroxetine, venlafaxine and nefazadone). Through the US Freedom of Information Act, they sought information that had been submitted to the US Food and Drug Administration (FDA) by manufacturers applying for marketing authorisation for their antidepressants.  They matched this with published versions of the FDA trials, identified using a PubMed search. Relevant studies were also identified through citations quoted in retrieved studies and review articles, and from a partly overlapping list of published versions of trials submitted to the Swedish drug regulatory authority.

What did the study find?
In line with earlier research, fluoxetine, nefazadone, paroxetine and venlafaxine were found to produce statistically significant improvements over placebo of a mean of 1.8 points in the Hamilton Rating Scale for Depression (HRSD).  However, this difference did not reach the 3-point drug-placebo criterion of clinical significance adopted by NICE.  When analysis was confined to those trials which included only subjects who were severely depressed (scoring 28 or more on the HRSD), the difference between drug and placebo exceeded NICE’s criterion for clinical significance. Further analyses indicated that drug type did not moderate this effect, when baseline severity of trial subjects’ depression was taken into account.

So what?
This paper has received widespread media coverage.  It is important to note that the meta-analysis found that the effect of antidepressants was statistically significantly different from placebo. However this effect was, on average, not clinically significant for all but the most severely depressed people.  It is also important to note that, due to the limitations of the data available, the results refer to averages.  We do not know the proportions of people who showed clinically significant improvement during antidepressant treatment, no improvement or worsened depression.

Less widely reported, but in our view at least as troubling, was that mean improvement scores were not available, despite exhaustive searching, for four sertraline trials and one citalopram trial which were reported as not showing a statistically significant drug effect.  These omissions represent 38% of patients in sertraline trials and 23% of patients in citalopram trials. Inclusion in the meta-analysis of only those sertraline and citalopram trials for which means were reported to the FDA would have constituted a form of reporting bias similar to publication bias and would have led to overestimation of drug–placebo differences for these drugs.  As we blogged recently, several studies have found publication bias in trials of antidepressants by the pharmaceutical industry.  Selective reporting biases the evidence base in favour of antidepressant drugs and as a result of these perceived benefits, inappropriate management decisions may be made by the most astute of clinicians and well informed patients. Selective reporting also deceives secondary researchers who rely on published information, and to those who volunteer to participate in these trials. The authors call for a change in the regulations for licensing new drugs which would require manufacturers to make publicly available all data submitted to regulatory agencies.

Prescribers are reminded of NICE guidance for the management of depression (see MeReC Briefing, 2005 and the depression section of NPC). For mild depression, watchful-waiting/ active monitoring and non-drug interventions, such as brief psychological therapy or exercise, may be all that is needed and drug treatment is not recommended initially. Antidepressants are recommended by NICE for people with moderate to severe depression; however, they may not be suitable for, or preferred by everyone.

Prescribers need to be aware of the limitations in the evidence for the effectiveness of antidepressants, and be mindful that published evidence may be biased in favour of their effectiveness. Rather than relying on results of evidence from individual articles, clinicians would be well advised to base their prescribing decisions on evidence-based advice from independent trusted sources such as NICE, Cochrane, Clinical Evidence, and the National Prescribing Centre. However, as these sources largely rely on published evidence, unless public access to all clinical trials becomes mandatory, even the most trusted sources of evidence-based guidance will have their limitations.

Please comment on this blog in the NPC discussion rooms, or using our feedback form