Cerebral Hemispheres 2


This was written as an addendum to my 2-Minute Neuroscience video on SSRIs.

If you were surprised, dubious, or even indignant in response to the last couple lines of the video, then please take the time to read this post. The last two lines of the video state: "Additionally, evidence has now emerged that questions the effectiveness of SSRIs. Some research has found they do not work much better than placebos, and may only benefit those who are most severely depressed, which is a minority of patients who actually take the drugs."

First, let me explain what the statement above does not mean: it does not mean that antidepressants will not or did not work for you. There is a great deal of individual variation in responses to drugs; what does work for one patient may not work for another and vice versa. Additionally, it's important to keep in mind that when we are looking at scientific studies of drug effectiveness, we are dealing with the average response among a group of people. Even if a study were to find that a drug isn't highly effective on average, it may still provide a benefit for an individual. And, as the statement indicates, some research has questioned the effectiveness of SSRIs, but there are many studies that support their effectiveness as well. 

The statement reflects a controversy that has been broiling in the area of antidepressant research since the late 2000s. Arguably, the study that ignited the controversy was conducted by Kirsch et al. (2008). Kirsch et al. asserted that estimates of antidepressant efficacy up to that point were biased because they focused primarily on published studies that were used to support the effectiveness of a drug, ignoring unpublished studies that might have been more likely to find negative results (i.e., the drugs didn't display greater effectiveness than placebo). This concern is not an uncommon one when it comes to looking at drug efficacy; there are often studies that find a drug is not superior to a placebo, but they are less likely to be submitted to academic journals or accepted for publication. Thus, the negative results may remain unknown, and positive studies that are published may skew the perception of just how effective a drug is.

Kirsch et al. thus conducted an overall analysis of the data from all controlled trials---both published and unpublished---for four antidepressants: fluoxetine (aka Prozac), venlafaxine (aka Effexor), nefazodone (aka Serzone---this drug has been largely discontinued), and paroxetine (aka Paxil). When Kirsch et al. combined the published and unpublished data, they found that the antidepressants did not meet standards set for clinical significance, an indication of whether the drugs were likely (on average) to make a meaningful difference in the symptoms of patients. Kirsch et al. also grouped patients according to the severity of their depression, and found that for patients with moderate depression there was no difference in efficacy between antidepressants and placebos. For patients with severe depression, antidepressants led to a small improvement in symptoms---although it still was not deemed to be clinically significant. In patients with the most severe depression, however, antidepressants did have clinically significant effects.

The study by Kirsch et al. had a major impact on the scientific community---and on the public perception of antidepressant effectiveness. Since the time of its publication, several other studies have been published that support the findings of Kirsch et al., both in terms of overall antidepressant efficacy (for example, see Jakobsen et al. 2017) and in terms of the finding that antidepressants are less effective in patients with milder or moderate symptoms of depression (for example, see Fournier et al. 2013).

On the other hand, a number of researchers have criticized the Kirsch et al. study for various reasons, including the interpretation of the results, the definition of clinical significance, and the determination of depression severity. And, a number of other studies have found conflicting results. The most notable example of these conflicting studies is the largest meta-analysis of antidepressants for depression to date, published in 2018 by Cipriani et al. This analysis of 21 antidepressants (including the most commonly prescribed drugs such as fluoxetine/Prozac, paroxetine/Paxil, escitalopram/Lexapro, and so on) included data from over 116,000 participants. It found that all antidepressants evaluated were statistically more effective than placebo.

As you would expect with any controversy, however, there are critics of the Cipriani et al. study as well. Some have argued, for example, that the results found by Cipriani et al. are actually quite similar to those seen by Kirsch et al. (2008), but the difference is that Kirsch et al. considered clinical significance rather than just statistical significance. The latter shows that there was a statistical difference in the effectiveness of the drug vs. the placebo, but again the former is an indication of if the change is big enough to make a meaningful difference in symptoms of depression. Other critiques about the methodology of the Cipriani et al. study have made some suggest the results are difficult to interpret. 

So, do antidepressants work or not? Well, it's complicated. The scientific community can't completely agree on how effective they are, but it's important to note that even Kirsch et al. and those who fall on that side of the controversy are not arguing that antidepressants have no effect. Instead, they are saying that---for some patients---antidepressants may not have a clinical benefit above and beyond what placebos can offer. But, the placebo effect can be quite powerful in certain conditions, and it can certainly help some people to feel better. 

At the same time, there are many in the scientific community who believe strongly that there is a clear benefit of antidepressants above and beyond the placebo effect. Thus, right now there are differing opinions among scientists as to just how effective antidepressants are compared to placebos. What does seem to be clear, however, is that there are many people who find antidepressants to be beneficial. The research does not deny those individual experiences, and many doctors continue to prescribe antidepressants because their patients often feel the drugs have led to a major improvement. Thus, while this debate will probably continue in scientific circles for some time, it does not mean antidepressants have no benefits to offer. And it should not make you hesitant to try an antidepressant if you are suffering from depression.