Two reviews by John Ioannidis assess the strength of the evidence supporting the use of antidepressant drugs as treatment for depression.
In his 2008 editorial, “Effectiveness of antidepressants: An evidence myth constructed from a thousand randomized trials?” Ioannidis argues the full body of evidence supporting antidepressant use is weaker than the published evidence would suggest. Citing Erick Turner’s 2008 meta-analysis, he notes that while nearly all published antidepressant trials show statistically significant benefits, only half of all evidence presented to the FDA did the same, with negative trials either remaining unpublished or becoming distorted in publication. In a second meta-analysis, antidepressants were only found to be more effective than placebo in cases of less severe depression, which indicates that any perceived benefits within this population may be due to the placebo effect (even in more severe depression, antidepressants were no more effective, and significant benefits resulted from a lack of response to placebo).
Ioannidis concludes that multiple factors contributed to the widespread belief in antidepressants’ effectiveness, despite a lack of evidentiary support:
- Antidepressant trials have been interpreted as “positive” when they indicate any statistically significant benefits associated with the drugs, regardless of the clinical significance or magnitude of those benefits.
- Restrictive participant inclusion criteria may select a group of subjects who are more likely to respond to treatment but whose results may not generalize to the broader population to whom antidepressants have been subsequently prescribed and marketed.
- Most trials test the drugs over a short period of time (rarely exceeding eight weeks) and therefore provide little support for the long-term impact of antidepressants or the effect of these drugs on hard outcomes of depression (such as depression resolution).
- Demonstrably selective reporting and publication mean any meta-analysis that looks only at published literature is likely to overstate the drugs’ true benefits.
- Biological arguments, such as marketing SSRIs as drugs to correct an imbalance of serotonin in the brain, distract from the weakness of the evidence supporting their use and are not aligned with our current understanding of these conditions.
Ioannidis concludes:
Based on the above considerations, antidepressants are probably indicated only in select patients with major depression, probably preferentially in those who have severe symptoms and have not responded to anything else. For most patients with some depressive symptoms who are currently taking antidepressants, using these drugs would not have been the preferred option, placebo would be practically as good, if not better, and would save toxicities and cost.
“This may sound depressing for the current state of clinical trials research on antidepressants and beyond,” he adds. “Nevertheless, even if one feels a bit depressed by this state of affairs, there is no reason to take antidepressants, they probably won’t work.”
In 2018, Ioannidis et al. published a meta-analysis of 522 antidepressant trials specific to patients with major depression. These trials covered 21 drugs and 116,477 participants in total. All antidepressants were more effective than placebo with modest effect size, which contrasts with similar analyses in children and adolescents that have found only fluoxetine to be effective. Nearly all trials were industry-funded and short, and excluded patients with treatment-resistant depression. This evidence, therefore, may represent a biased data set that cannot be extrapolated to all forms of treatment.
Taken in the context of other pieces featured on CrossFit.com, Ioannidis’ analysis is consistent with the interpretation that antidepressants play a role in the short-term treatment of major depression and that the evidence fails to justify more widespread, long-term use. Support for the latter use relies upon selective interpretation of a biased evidence base, and a more complete assessment of the research suggests antidepressants’ benefits are overstated.
Comments on John Ioannidis on the effectiveness of antidepressants
Placebo is just as good. The power of the mind. Power of feeling that someone understands us and that will help us to get out of depression is probably the most important point here. The best medicine will be a support team that puts you on track with proper food and exercise.
Great write up on an important topic. There are several meta-level issues lurking in the background here, namely, the emphasis on publishing positive findings (the flipside being negative, or lack of effect, findings are ignored) and the standard approach to care these days which is chronically medicating. Ionnadis saying, "Based on the above considerations, antidepressants are probably indicated only in select patients with major depression, probably preferentially in those who have severe symptoms and have not responded to anything else," is a much more conservative and calculated approach as opposed to the widespread use of antidepressants in less severe cases. Given the side effects, this approach is wise.
Furthermore, I would like to see more research spent in determining the causes of depression in modern society and then strategies that can be put in place to *prevent* depression in the first place. Prevention, as a rule of thumb, is generally preferable to treatment. Depression and other mental disorders are clearly not going anywhere on their own.
The problem Ioannidis draws attention to here may have effects that stretch far beyond the antidepressant literature, and speak to the core of the concept of evidence-based medicine.
When a clinician or an academic argues they are following the evidence in their recommendations, that implies they believe the evidence available to them is an accurate reflection of the evidence that exists regarding a particular condition, drug, therapy, etc. What Ioannidis describes is an area in which the published literature fails to even address the majority of the evidence available - that is, most of the evidence has never actually been published. Simultaneously, the evidence that HAS been published is systematically and predictably biased in ways that would lead to better mean results in clinical trials than in subsequent clinical practice.
Before Turner 2008 and other papers on antidepressants, it would have been difficult for any clinician or researcher to assess whether these forms of bias existed simply by looking at the published literature. Surely the same problems exist elsewhere and we are just not aware of them, or their extent.
John Ioannidis on the effectiveness of antidepressants
3