Antidepressants and the Tangle of Treatment-Related Suicidality

A Q&A with Martin Plöderl

Martin Plöderl, PhD is a clinical psychologist and psychotherapist at a public psychiatric hospital in Salzburg, Austria. In addition, he gives lectures and trainings in suicide prevention and is a member of the expert panel of the national suicide prevention program SUPRA Austria. His research focus is on suicide prevention, with a recent focus on the efficacy of antidepressants. The views presented here are his own.

---

Aftab: I suspect we have our differences in views when it comes to the association between antidepressants and suicidality, but you are someone I respect a lot, and you are quite familiar with the literature, so I am hoping to use you as a guide to introduce readers to the complexity of the topic. It’s unfortunate that despite more than decades of research, opinions in the psychiatric community are still divided on how to understand this relationship.

Plöderl: Thank you for giving me the opportunity to respond to your important questions. The divided opinions in the field are not surprising. Suicide is a rare event and consequently, it is notoriously hard to prove if a treatment affects suicide rates. Even after decades of research, there is much uncertainty and it is just natural that subjective beliefs – more or less supported by the evidence – enter the discussion. In addition, antidepressants have been used for decades with the intent to improve depression and to reduce suicide risk. Inconvenient scientific findings cause substantial cognitive dissonance which are often resolved with “but I see these meds work in clinical practice.” However, it is of utmost importance to stick to the principles of evidence-based medicine. Awais, I am not sure if we really disagree on our conclusions about the suicide risk associated with antidepressants, but perhaps more on what this means for the use and overall harm/benefit ratio for antidepressants.

Aftab: How did you get interested in the topic of antidepressants and suicidality?

Plöderl: I have worked with suicidal patients for nearly 20 years now and because I try to regularly screen the current literature about suicide prevention. I, of course, became aware of the controversy surrounding the role of antidepressants. For example in 2005, when two meta-analyses of clinical trials reported significantly increased rates of suicide attempts among adults on SSRI’s (Fergusson et al., 2005; Gunnell et al., 2005). Then there were at least three issues which motivated me to do some research myself. First, I became aware of studies that shattered my trust in evidence-based medicine, such as the re-analysis of Study 329 (Le Noury et al., 2015) or the study by Sharma et al. (2016). These studies provided evidence that suicidal events were underreported or misclassified in clinical studies to make the results more convenient for antidepressants. Second, I became aware of the gap between the evidence about the efficacy of antidepressant and prescription practices. Third, at a plenary session in a major international suicide conference with a very large audience, a systematic review authored by many prominent suicidologists was presented. Antidepressants were declared as clearly suicide preventive and none of the problems we discuss later were mentioned.

Children and Adolescents

Aftab: What do we know about the risk of suicidality (ideation, attempts, suicide deaths) with antidepressants in children and adolescents?

Plöderl: I focus on suicidal behavior for simplicity and with “antidepressants,” I refer to second generation antidepressants (SSRIs, SNRIs, atypical antidepressants) but not to newer drugs such as esketamine (we have published here, too, e.g., Plöderl et al., 2022). It has been known for a long time now from clinical trials that antidepressants, relative to placebo, can increase the risk for suicidal ideation and behavior in children and adolescents, leading the FDA to issue a black box warning. Since then, systematic reviews continued to report an increased risk for suicidality (suicidal ideation or worse) (Hetrick et al., 2012) or suicidal behavior with antidepressants for children and adolescents (Hetrick et al., 2021) and also young adults (Stone et al., 2009). No suicides occurred in the clinical trials, which is not surprising, because suicides among children and adolescents during trials are rare, especially when suicidality is an exclusion criterion, and much larger samples would be needed. However, observational studies confirmed the increased risk for suicidal behavior (Barbui et al., 2009), and these studies were deemed as high-quality and convincing even in a review published in JAMA Psychiatry (Dragioti et al., 2019).

Aftab: What are the sources of disagreement on this issue?

Plöderl: I think that there is actually quite some consensus that antidepressants can increase the risk for suicidal behavior in children and adolescents. However, it is interesting to look at reviews on the topic where relevant evidence is sometimes not mentioned or downplayed, or it is only said that suicide ideation can be increased but not suicidal behavior, or that the fact that there were no suicides in the trials. What you also find as argument in defense of antidepressant even by prominent suicidologists is that untreated depression causes more harm, compared to the iatrogenic risk for suicide for a few. To give you a quote from such an important systematic review in Lancet authored by many prominent suicidologists (Zalsman et al., 2016):

“In children and adolescents, increased risk of suicidal thoughts has to be taken into account when starting pharmacotherapy for depression. However, given the increased risk of suicide in untreated depression and the absence of an increased risk of suicide associated with pharmacotherapy, currently available evidence does not support the avoidance of initiation and continuation of pharmacotherapy for depression in children and adolescents. Therefore, the ongoing discussion about possible induction of suicidality in minors should not prevent physicians from prescribing SSRIs.”

This argument is at odds with the best evidence, even more so when considering recent evidence where efficacy in children and adolescents for common antidepressants, at least when summarized as group (SSRIs, SNRIs), was found to be well below clinical significance (Hetrick et al., 2021). Even very recent papers and reviews fail to acknowledge the fact that efficacy was also poor for fluoxetine according to the most recent Cochrane meta-analysis (Hetrick et al., 2021). Fluoxetine has been the drug of choice because it was found to be effective in earlier reviews. As a side note, this is a demonstration of the so-called novelty bias, that is, the efficacy of drugs decreases over time because established drugs, first tested by the producer, later become comparator drugs, tested by competing pharmaceutical companies.

Some controversy arose from seemingly increasing rates of suicides after the FDA issued a black box warning (Gibbons et al., 2007; Lu et al., 2018), but the research supporting this argument is problematic for several reasons (Stone, 2014, 2018). We discuss this in our papers, too (e.g., Plöderl & Hengartner, 2023), and we will come back to this controversy later. Nonetheless, you will regularly stumble upon arguments based on such problematic ecological data where rates of prescriptions were compared with rates of suicides. Perhaps you remember the 90s when suicide rates decreased with increasing rates of prescriptions of SSRIs. This was seen as success story for the suicide preventive effect of antidepressants in some psychiatric publications. However, the association now reversed, that is, ever increasing prescriptions are now associated with increasing suicide rates. This is hardly mentioned in psychiatric publications. 

Adults

Aftab: What do we know about the risk of suicidality with antidepressants in adults?

Plöderl: Let’s start with randomized clinical trials because these provide the highest level of evidence. According to the analysis by the FDA (Stone et al., 2009), the data is inconclusive for suicides in all age-groups. For suicide attempts, there is an age-dependent effect: for antidepressants (compared to placebo) the risk is significantly increased for young adults, significantly decreased for older adults (65+ years), and results are inconclusive for adults of other ages. When we re-analyzed trial data from the FDA for adults we found a significantly increased risk for suicide attempts for antidepressants (Hengartner & Plöderl, 2019a, 2019b). The findings for suicides were comparable but statistical significance depended on the choice of method — it is a challenge to combine data with many zero events. Our re-analysis was initially only a brief letter to the editor, pointing out that the original paper used a problematic method. We received quite some criticism for our analysis and even a full-blown paper was published in response (Kaminski & Bschor, 2020). A problem in our analysis was that the data also included trials with open label extension phases without a comparable placebo phase. Whereas suicidal behavior occurs primarily in the beginning of trials, we don’t know how much our analysis led to an overestimation of the problem. On the other hand, we should keep in mind that suicidal behavior is underreported or misclassified in a way that makes the findings more favorable for antidepressants (Healy & Aldred, 2005; Högberg et al., 2015; Le Noury et al., 2015; Sharma et al., 2016). We discussed all this in a short piece that we published together with psychiatrists who initially criticized us (Plöderl et al., 2020); a nice example about handling disagreements in a scientific way.

Our findings are supported by results from observational studies. We recently updated a previous systematic review of observational studies and found a statistically significant increase of suicide attempts among patients on antidepressants compared to those who did not use antidepressants (Hengartner et al., 2021). The findings for suicides were also in the same direction but not always statistically significant. We also found strong evidence for biases. Studies where lead-authors had financial ties to the pharmaceutical industry did not show an increased risk of suicidal behavior with antidepressant, whereas studies without such conflicting interests reported that antidepressant were associated with a significantly increased risk for suicidal behavior (Plöderl, Amendola, et al., 2023). We also found strong evidence for reporting bias, meaning that studies which found an increased risk for antidepressants were not published. Of course, observational studies are limited by analytic flexibility (e.g., by choice of confounders, time-frame, statistical method) and also selection bias, that is, those who are on antidepressant may have more severe depression and thus a higher risk for suicide. Some studies take care to control for confounders such as severity of depression or use sophisticated statistical methods and still report problematic findings for antidepressants. Observational studies always remain limited and cannot replace randomized controlled trials. On the other hand, randomized clinical trials are not designed to detect rare adverse events. Furthermore, they usually exclude suicidal patients and they are too short, raising doubt if findings can be generalized to people we see in clinical practice and for longer use of antidepressants. Of note, there are two systematic reviews of longer-term randomized clinical trials which also found that, compared to placebo, the use of antidepressants was significantly increased for suicide attempts and borderline significant for suicides (Baldessarini et al., 2016; Braun et al., 2016). Overall, my take is that the evidence for adults remains somewhat inconclusive. However, what can be said with quite some certainty is that antidepressants overall do not decrease suicidal behavior. Put differently, the claim that “antidepressants are live-saving” is not supported by the evidence, at least not for the average patient. This is striking, given the widespread assumption that antidepressants effectively treat depression, one of the most important risk factors for suicide. A reduction of suicidal behavior should of course be expected, but we don’t see this in the studies. 

Plöderl: My take is that the evidence for adults remains somewhat inconclusive. However, what can be said with quite some certainty is that antidepressants overall do not decrease suicidal behavior… This is striking, given the widespread assumption that antidepressants effectively treat depression, one of the most important risk factors for suicide. A reduction of suicidal behavior should of course be expected, but we don’t see this in the studies.

Aftab: Some reviews have suggested that while antidepressants are associated with an increase in suicidality in children, they are associated with a reduction in suicidality in older adults. Is that your reading of the data as well?

Plöderl: It depends on the level of evidence and what you mean with “suicidality.” Yes, a reduced rate of suicide attempts was found among older adults (65+ years) on antidepressants compared to placebo in randomized controlled trials (Stone et al., 2009). However, data for suicides remain inconclusive so that effects in both directions cannot be ruled out. Furthermore, well-controlled observational studies reported an increased risk of suicide with antidepressant use (Coupland et al., 2011). Consequently, I would not consider the issue as settled.

Observational Studies

Aftab: In observational studies, suicidality seems to improve after initiation of antidepressants. The question in the observational context is whether that is simply the effect of time and placebo – natural history, regression to the mean, etc. However, it does fit with the clinical observation that most patients with suicidality who are prescribed antidepressants experience a reduction in suicidality, whether this is because of antidepressant or not.

So, in this observational context, if we are looking at increase in suicidality with antidepressants, would this manifest as a lesser reduction in suicidality with antidepressants (on average) vs natural history/placebo? A related issue is that of different trajectories of suicidality with antidepressant treatment. One would expect that some people experience improvement in suicidality, some people experience no change, and some people experience worsening or new onset of suicidality, and these differences would be obscured by average rates.

Plöderl: Yes to your question. When you only look at the group of people who use antidepressants, suicidality peaks before the initiation of the treatment and then reduces over the course of therapy. This would most likely be very similar for treatment with placebo, psychotherapy etc. As you point out, the observed course of symptoms is a mix of selection processes and regression to the mean, because people enter treatment around the peak of symptoms and, on average, get better due to the natural course, placebo effects, and actual effects of therapies. This can explain why clinicians easily get the impression that their treatment caused the improvement and why scientific studies with control groups are essential. We discussed that in response to a paper which only looked at those who initiated antidepressant treatment but used no control groups and overstated the results (Plöderl & Hengartner, 2022). A different picture emerges in observational studies where those who used antidepressants were compared to people who did not use these drugs, or where persons served as their own control by using different time-frames. Then it turns out that, relative to the control group/period, risk for suicidal behavior peaks at the initiation of antidepressants. There are similar patterns in randomized controlled trials.

Plöderl: When you only look at the group of people who use antidepressants, suicidality peaks before the initiation of the treatment and then reduces over the course of therapy. A different picture emerges in observational studies where those who used antidepressants were compared to people who did not use these drugs, or where persons served as their own control by using different time-frames. Then it turns out that, relative to the control group/period, risk for suicidal behavior peaks at the initiation of antidepressants.

You also refer to treatment effect heterogeneity, that is, the average treatment effect may not tell the full story because some patients may benefit more or less than the observed average effect. Indeed, this is the position of the FDA: antidepressants can trigger suicidality in some patients (in all age groups, by the way) and reduce suicidality in others because of the effective treatment of depression. Whereas this argument sounds plausible, expert biostatisticians remain skeptical about assuming treatment effect heterogeneity. Furthermore, clinically meaningful predictors of treatment success were not detected yet, despite massive research efforts. Pointing out to treatment effect heterogeneity typically arises when efficacy is poor to justify the use of certain treatments. My position is that this should be explored and perhaps can be explored with patient level data of all available RCTs.

Aftab: There is the issue of treatment effect heterogeneity, yes, but I am also thinking of a more basic heterogeneity with regards to the natural course of suicidality. Take untreated depression and suicidality. Some depressed individuals experience worsening of suicidality, progressing to attempted or completed suicide. Other depressed and suicidal individuals may remain chronically suicidal for years without worsening or improving substantially. And other depressed and suicidal individuals experience improvement or resolution of suicidality with time. So, there is heterogeneity right there in natural course. This heterogeneity wouldn’t be eliminated by an intervention, but the proportion of people in different trajectories would shift.

Placebo

Aftab: You mentioned that a reduction of suicidal behavior should be expected with an intervention that effectively improves depression, but we don’t see this in the studies. What about placebo? I’m using “placebo” broadly here to encompass all non-specific forms of care. We know that depression is responsive to placebo in clinical trials. We also know this from clinical experience that sometimes the mere act of carefully listening, evaluating, empathizing with a patient, or providing them a safe therapeutic environment, can bring improvement. So, if “placebo” reduces depression above and beyond natural history of depression, does “placebo” improve suicidal behavior? This is important because if “placebo” does improve suicidality, active treatments are competing with placebo to show their effect. And if “placebo” doesn’t improve suicidality, maybe we need to rethink the relationship between improving depression and improving suicidality.

Aftab: If “placebo” reduces depression above and beyond natural history of depression, does “placebo” improve suicidal behavior? This is important because if “placebo” does improve suicidality, active treatments are competing with placebo to show their effect. And if “placebo” doesn’t improve suicidality, maybe we need to rethink the relationship between improving depression and improving suicidality.

Plöderl: Let me clarify that, when I said we don’t see an improvement of suicide risk in clinical trials, I referred to the drug-placebo difference. I am not aware of studies on suicidal patients where treatment with antidepressants (or placebo) was compared with no treatment. I guess no such study will make it through review boards for ethical reasons. However, there are very interesting data from the esketamine phase 3 trials on depressed and acutely suicidal patients. Here, participating patients were randomized to receive either esketamine or placebo in addition to standard antidepressant treatment. It turned out that throughout the trials, the esketamin-placebo difference for suicidal ideation was negligible but both groups substantially improved in suicide ideation. Thus, it seems that nonspecific factors such as regression to the mean, extra care, being in a scientific trial, regular contact with clinicians, expectations etc. are more important to improve than medication itself. I doubt that similar results would be seen in a control group without any treatment. There is also evidence from clinical trials that increasing the frequency of clinical encounters and a good relationship with the clinician is associated with more symptom improvement in an extend that is comparable to the drug-placebo difference. So, what does it mean in the discussion when someone says that we cannot leave depressed people untreated? Does it mean no treatment at all? (Here I would agree that we shouldn’t leave people untreated.) Or does “untreated” means no treatment with medication? And what does this mean for GPs and psychiatrists: can they offer clinical care without medication and how would this fit to patients expectations?

Antidepressants vs CBT

Aftab: How does the risk of suicidality with antidepressants compare to CBT (or other forms of psychotherapy)?

Plöderl: The data-base for psychotherapy is much smaller than compared to antidepressants. However, psychotherapy studies often include patients after suicide attempts or those who were referred to emergency departments, a population known to be at high risk for suicide. This is a crucial difference to drug trials where suicidal patients are nearly always excluded. In meta-analysis, CBT significantly reduced suicide attempts and the results for suicides are promising but inconclusive (Witt et al., 2020). However, there are methodological biases raising doubt on the results. Blinding is hardly possible and effects decrease with methodological rigor (Witt et al., 2020). My take is that even when considering the method biases, the results look better for CBT than for antidepressants. More data is needed to see if the promising results for CBT with respect to suicide attempts will also be found for suicides and in studies with higher quality. We also do not have sufficient data for other forms of psychotherapy than CBT. Interestingly, there are some promising results for brief interventions (e.g., follow-up after hospitalization, safety planning). But again, efficacy must be put to test in high-quality replication studies independent of the original study groups as allegiance bias typically leads to better outcomes, as for example, seen in the results for DBT.

Aftab: There are RCTs that have directly compared antidepressants and CBT in the treatment of depression (i.e. both arms have the same inclusion/exclusion criteria and similar baseline characteristics). Has anyone done a review or meta-analysis of such head-to-head RCTs to compare the occurrence of suicidal behaviors? (We may also wonder about combined treatment with antidepressants and psychotherapy!)

Plöderl: Fortunately, just as I am responding, a study appeared which can help answer your question. Zainal (2023) meta-analyzed clinical trials where antidepressant medication was compared to psychotherapy (mostly CBT or interpersonal psychotherapy) or to a combination of psychotherapy and antidepressants. The outcome was the prevention of suicidal behavior and serious adverse events (it was mostly suicidal behavior, though). It turned out that among youth, psychotherapy alone was statistically significantly superior to antidepressants alone and also superior to the combination of psychotherapy/antidepressants. Among adults, results were generally favoring psychotherapy but only the difference between the combination of antidepressants/psychotherapy over antidepressants was statistically significant.

Plöderl: Among youth, psychotherapy alone was statistically significantly superior to antidepressants alone and also superior to the combination of psychotherapy/antidepressants in the prevention of suicidal behavior.

Aftab: In the Zainal (2023) meta-analysis, there wasn’t a statistically significant difference between psychotherapy alone vs antidepressants alone with regards to suicide attempts in adults. Would you say that this suggests that in adults, the risk of treatment-emergent suicidality with psychotherapy may be comparable to the risk of treatment-emergent suicidality with antidepressants? Or that, at the very least, we can’t say much with confidence until better data is available?

Plöderl: The 95% confidence intervals of the odds ratios were 0.22–0.71 (p < 0.05) for youth and 0.25–1.18 (p > 0.05) for adults. A non-significant finding does not imply that there is a zero difference between the groups but that the evidence remains inconclusive. The results are concerning but more data is needed.

Two Scenarios

Aftab: The discussion about increased risk of suicidality with antidepressants seems to be referring to, at least, two different sorts of scenarios. In the first case we have suicidality as a complex, multicausal phenomenon, with multiple interacting causes, and antidepressants become one factor in this causal mix with a small causal contribution, and we need larger samples to detect that signal while other controlling for other causal factors. In the other case, suicidality is driven predominantly by initiation (or discontinuation, in the context of withdrawal) of treatment with antidepressant. These people weren’t suicidal prior to treatment and don’t have a lot of other active risk factors, but they started the antidepressant, and now they are actively suicidal in a way that they weren’t before. This latter scenario often seems to be driven by iatrogenic effects such as akathisia or mood lability. Any thoughts on thinking about the antidepressant-suicidality relationship in this manner?

Plöderl: These are very good points and based on your arguments, it becomes clear when antidepressants may indeed be life-saving and what it means to use antidepressant in a rational way. There is consensus that the suicidal process is complex, that is, as you point out, there are many different causal factors which are interacting over time. This explains why it is impossible to predict suicidal behavior on the longer-term (we did some work on this issue, by the way). Ideally, clinicians try to come up with a valid case-formulation to explain when and why a person’s suffering became so unbearable that suicide is seen as only solution. And here I believe that antidepressants’ psychotropic effects can be beneficial for certain people at certain times (e.g., by taking the edge of intense feelings of anxiety). In the framework of complex systems theory, this may help to desynchronize a pathological pattern (e.g., a suicidal crisis or the suicidal mode). Awais, I imagine you will reply that antidepressants’ mechanisms go beyond psychotropic effects, such as enhanced neurogenesis and neuroplasticity. From a complex systems viewpoint, this could be considered as a change of control-conditions which can destabilize a system, making it easier to change pathological patterns and establish more healthy one’s. We should see this as better longer-term outcome in the combination of psychotherapy and medication. However, the data is not very convincing (Furukawa et al., 2021) and there are arguments that antidepressants can infer negatively with adaptive processes (Hollon, 2020), so I remain respectfully skeptical. Anyway, compare this way of thinking – prescribing antidepressants only as long as necessary and depending if the desired psychotropic effect is appropriate to counteract the suicidal process – and the current treatment algorithms, which appear rather “blind” to me. I am sure we both agree that there is room for improvement in research and practice and I second your statement that “time and energy could instead be spent examining the appropriate place of antidepressants in the clinical treatment of depression, how to personalize treatment, and how to utilize and improve access to non-pharmacological interventions” (Aftab, 2022).

Re your point about non-suicidal people becoming suicidal with antidepressants: the evidence is convincing that antidepressants can lead to treatment emergent suicidality. For this reason, the FDA released a black box warning. You also mentioned suicidality caused by withdrawal effects. Indeed, an increase of suicidality was found in a large observational study (Coupland et al., 2015) or in the tapering phase of Study 329, which was only published due to the RIAT initiative (Le Noury et al., 2016). Let me add an anecdote: some time ago I met someone in my neighborhood who did not know anything about me and my research and thinking about antidepressants. He told me about his panic disorder and that he used sertraline for many years (!) now. I asked about how well he tolerated the medication and as usual, he only talked about sexual dysfunctions after I explicitly asked him. I also wanted to know if he ever had considered tapering. He said that he tried to do so once by halving the dose, as recommended by his GP. The following night, for the first time in his life, he became so suicidal that he would have committed suicide and only thinking of his kids prevented him to do so. The unbearable state was likely a mixture of anxiety and akathisia. Since then, he does not dare to reduce his medication. So many people are on antidepressants for years or even decades, and the problems with tapering these drugs, including suicidality, are not adequately considered in research and practice, in my opinion. Related to that, there are still common clinical practices which may put patients unnecessarily at risk for suicide. For example, patients are not always closely monitored after initiating an antidepressant. After nonresponse, it is still common to increase the dose or to switch to a different antidepressant, but these practices are not effective according to the available evidence and may be associated with an increase of suicide risk. I also still come across too many patients who do not know that stopping antidepressants is risky.

Strawn, et al. 2023

Aftab: You recently commented on an article by Strawn et al. 2023, a RCT of escitalopram for generalized anxiety disorder in children and adolescents. 9.5% of subjects in the escitalopram arm had suicidal ideation vs only 1.5% in the placebo arm, a striking and alarming difference. The suicidality wasn’t really taken into account in the conclusions and escitalopram was considered by the authors to be effective and “well tolerated.” I was also struck by the lack of curiosity about this treatment-related suicidality. There was practically no discussion of why this surprising increase may have happened, or what the subjects reported about their experience of suicidality, or how this should impact clinical decision-making. This sort of lack of curiosity seems to be pervasive in the research literature.

Plöderl: I only stumbled upon this RCT because of a weekly PubMed alert and saw that suicidality was not discussed at all. A simple statistical test shows that the risk for suicide ideation is significantly increased for escitalopram. Only the main outcome, the reduction of anxiety symptoms, was just above the threshold of statistical significance, but the effect size is small. The five secondary outcomes were all null-effects. Of note, this trial made the FDA approve escitalopram for generalized anxiety disorders in children and youth. I think it is another example of how low the bar for approval has got and that peer-review has obvious deficits. I wonder if and how the authors will respond to our commentary (Plöderl, Horowitz, et al., 2023). 

Aftab: The Strawn et al. RCT makes me wonder if the presence and magnitude of antidepressant-suicidality associations also differs based on the diagnosis. Is it different, say, for major depression, vs bipolar depression, vs anxiety disorders?

Plöderl: Indeed, the discussion above is nearly exclusively about evidence for the treatment of depression. The data-base for other mental health problems is sure much smaller, and as far as I know, there is also uncertainty about the association of antidepressants and suicide risk. Admittedly, I am not very familiar with the evidence here, so I cannot reply to your question adequately. However, this leads to the question if antidepressants reduce suicidality among severely depressed people or suicidal depressed people. This is crucial because it seems that it is a no-go to not prescribe antidepressants here. We do know that suicide ideation is a predictor of poorer outcome for antidepressants, but it is not known how much this also applies to the placebo arms of trials. And we don’t know if antidepressants are more suicide preventive among severely depressed. At worst, antidepressants’ adverse effects may add additional burden to severely depressed suicidal people with hardly any meaningful clinical benefit (at least for the majority). We clearly need more studies here.

FDA Black Box Warning

Aftab: There is a pervasive myth that FDA adding a black box warning to antidepressants led to an increase in suicide rates in children in the US. You’ve written about this before. Can you say why this claim is mistaken?

Plöderl: I mentioned this issue already, but let’s dig a bit deeper. Some say that after the black box warning, rates of prescriptions for children/adolescents decreased whereas suicide rates increased. For example, there was this widely cited paper by Gibbons et al. (2007) who reported a reduction of prescriptions by 22% from 2003 to 2005, which was accompanied by in increase of suicides by 14%. However, when you take a closer look at this paper, you will see that the increase of suicides was only in the first year (2003 to 2004) where there was hardly any change of prescription rates. From 2004 to 2005, antidepressant prescriptions indeed decreased, but so did suicide rates, and this was not reported in the paper. Furthermore, a look at the longer-term trends of suicides and prescriptions leads to a contradiction. And where are critical voices that the ever increasing prescription rates, with an all-time record in recent years, could be seen as example that the black box warning is not working? Nonetheless, the study by Gibbons et al. is still used in discussions, even among experts in suicidology (I made some interesting experiences here). Another group of researchers used interrupted regression models to prove that the black box warning caused suicides. This statistical method models the trend of suicides before the black box warning to estimate the expected trend after the warning and thus allows a comparison with the actual observed trend of suicides. However, the results of interrupted regression models are very sensitive to the choice of the time-frame before and after the black box warning. In another such study, authors used overdoses as outcome (with and without intent to die) and reported detrimental effects of the black box warnings. But when intentional self-harm was used as more appropriate outcome, an opposite finding was observed (Stone, 2014). There is also the problem that the data is all too noisy and putative changes of trends after the warning could not be conformed with statistical methods (Plöderl & Hengartner, 2023). I refer the interested readers to the in-depth discussion between proponents and opponents of the black box warning (Lu et al., 2018; Stone, 2018).

Misconceptions

Aftab: Are there other common misconceptions about antidepressants and suicidality?

Plöderl: While I am critical about antidepressant and believe that the harm-benefit ratio is problematic for the majority (a point where we probably disagree), suicidality caused by antidepressants is rare in a statistical sense. For example, there were 0.12% suicides in the drug arms versus 0.04% in the placebo arms of clinical trials for depression (Hengartner & Plöderl, 2019a). Of course, even small differences are a problem for outcomes such as suicides, but antidepressants are no killer-drugs as sometimes depicted on social media. It is also not appropriate to blame antidepressants for mass-shootings simply because many of the perpetrators were on antidepressants. To establish causality, it takes more than knowing that someone was on an antidepressant. For an example, see the cases discussed by Teicher et al. (1990) or our recent case-report where we discuss why discontinuation of antidepressants likely caused suicidality in a patient, based on a challenge-dechallenge-rechallence paradigm (Kostic et al., in press, The Journal of Clinical Psychiatry).

Plöderl: While I am critical about antidepressant and believe that the harm-benefit ratio is problematic for the majority (a point where we probably disagree), suicidality caused by antidepressants is rare in a statistical sense... antidepressants are no killer-drugs as sometimes depicted on social media.

I would also like to point out what I consider as misconception in research and clinical practice: GPs and psychiatrists too often equate “no prescription” with “untreated” and watchful waiting is not often considered as viable option. There is also the belief that randomized controlled studies cannot resolve the questions we discussed here. The pandemic showed that large, pragmatic randomized trials can be done in short time. We now even know if aspirin or Vitamin D is working to prevent severe courses of COVID. In psychiatric research, the evidence for antidepressants is almost exclusively provided by the pharmaceutical industry, with all the known biases. We still don’t know if and how strong antidepressants or lithium affect suicidal behavior and for whom, especially in the longer-term. This is in striking contrast to the widespread use of these drugs. I hope that pragmatic trials will become common in our field.

Thanks again for the opportunity to provide my thoughts. I hope you keep on bringing work of critics and skeptics into mainstream psychiatry (I don’t like these terms but I don’t know better ones) in a way that will help to correct false beliefs and improve treatment for people with mental health problems.

Aftab: Thank you!

Comments are open for all subscribers.

---

---

References

• Aftab, A. (2022, November 20). The Case for Antidepressants in 2022 [Substack].

  The Case for Antidepressants in 2022

• Baldessarini, R. J., Lau, W. K., Sim, J., Sum, M. Y., & Sim, K. (2016). Suicidal risks in reports of long-term controlled trials of antidepressants for major depressive disorder II. International Journal of Neuropsychopharmacology, 20, 281–284. https://doi.org/10.1093/ijnp/pyw092

• Barbui, C., Esposito, E., & Cipriani, A. (2009). Selective serotonin reuptake inhibitors and risk of suicide: A systematic review of observational studies. Canadian Medical Association Journal, 180(3), 291–297. https://doi.org/10.1503/cmaj.081514

• Braun, C., Bschor, T., Franklin, J., & Baethge, C. (2016). Suicides and suicide attempts during long-term treatment with antidepressants: A meta-analysis of 29 placebo-controlled studies Including 6,934 patients with major depressive disorder. Psychotherapy and Psychosomatics, 85(3), 171–179. https://doi.org/10.1159/000442293

• Coupland, C., Dhiman, P., Morriss, R., Arthur, A., Barton, G., & Hippisley-Cox, J. (2011). Antidepressant use and risk of adverse outcomes in older people: Population based cohort study. BMJ (Clinical Research Ed.), 343, d4551. https://doi.org/10.1136/bmj.d4551

• Coupland, C., Hill, T., Morriss, R., Arthur, A., Moore, M., & Hippisley-Cox, J. (2015). Antidepressant use and risk of suicide and attempted suicide or self harm in people aged 20 to 64: Cohort study using a primary care database. BMJ, 350(feb18 32), h517–h517. https://doi.org/10.1136/bmj.h517

• Dragioti, E., Solmi, M., Favaro, A., Fusar-Poli, P., Dazzan, P., Thompson, T., Stubbs, B., Firth, J., Fornaro, M., Tsartsalis, D., Carvalho, A. F., Vieta, E., McGuire, P., Young, A. H., Shin, J. I., Correll, C. U., & Evangelou, E. (2019). Association of Antidepressant Use With Adverse Health Outcomes: A Systematic Umbrella Review. JAMA Psychiatry, 76(12), 1241. https://doi.org/10.1001/jamapsychiatry.2019.2859

• Fergusson, D., Doucette, S., Glass, K. C., Shapiro, S., Healy, D., Hebert, P., & Hutton, B. (2005). Association between suicide attempts and selective serotonin reuptake inhibitors: Systematic review of randomised controlled trials. British Medical Journal, 330, 396–403.

• Furukawa, T. A., Shinohara, K., Sahker, E., Karyotaki, E., Miguel, C., Ciharova, M., Bockting, C. L. H., Breedvelt, J. J. F., Tajika, A., Imai, H., Ostinelli, E. G., Sakata, M., Toyomoto, R., Kishimoto, S., Ito, M., Furukawa, Y., Cipriani, A., Hollon, S. D., & Cuijpers, P. (2021). Initial treatment choices to achieve sustained response in major depression: A systematic review and network meta-analysis. World Psychiatry: Official Journal of the World Psychiatric Association (WPA), 20(3), 387–396. https://doi.org/10.1002/wps.20906

• Gibbons, R. D., Brown, C. H., Hur, K., Marcus, S. M., Bhaumik, D. K., Erkens, J. A., Herings, R. M. C., & Mann, J. J. (2007). Early evidence on the effects of regulators’ suicidality warnings on SSRI prescriptions and suicide in children and adolescents. Am J Psychiatry, 164(9), 1356–1363. https://doi.org/10.1176/appi.ajp.2007.07030454

• Gunnell, D., Saperia, J., & Ashby, D. (2005). Selective serotonin reuptake inhibitors (SSRIs) and suicide in adults: Meta-analysis of drug company data from placebo controlled, randomised controlled trials submitted to the MHRA’s safety review. BMJ, 330(7488), 385. https://doi.org/10.1136/bmj.330.7488.385

• Healy, D., & Aldred, G. (2005). Antidepressant drug use & the risk of suicide. International Review of Psychiatry, 17(3), 163–172. https://doi.org/10.1080/09540260500071624

• Hengartner, M. P., Amendola, S., Kaminski, J. A., Kindler, S., Bschor, T., & Plöderl, M. (2021). Suicide risk with selective serotonin reuptake inhibitors and other new-generation antidepressants in adults: A systematic review and meta-analysis of observational studies. Journal of Epidemiology and Community Health, 75, 523–530. https://doi.org/10.1136/jech-2020-214611

• Hengartner, M. P., & Plöderl, M. (2019a). Newer-Generation Antidepressants and Suicide Risk in Randomized Controlled Trials: A Re-Analysis of the FDA Database. Psychotherapy and Psychosomatics, 88(4), 247–248. https://doi.org/10.1159/000501215

• Hengartner, M. P., & Plöderl, M. (2019b). Reply to the Letter to the Editor: „Newer-Generation Antidepressants and Suicide Risk: Thoughts on Hengartner and Plöderl’s Re-Analysis“. Psychotherapy and Psychosomatics, 88(6), 373–374. https://doi.org/10.1159/000502485

• Hetrick, S. E., McKenzie, J. E., Bailey, A. P., Sharma, V., Moller, C. I., Badcock, P. B., Cox, G. R., Merry, S. N., & Meader, N. (2021). New generation antidepressants for depression in children and adolescents: A network meta-analysis. Cochrane Database of Systematic Reviews, 2021(5). https://doi.org/10.1002/14651858.CD013674.pub2

• Hetrick, S. E., McKenzie, J. E., Cox, G. R., Simmons, M. B., & Merry, S. N. (2012). Newer generation antidepressants for depressive disorders in children and adolescents. The Cochrane Database of Systematic Reviews, 11, CD004851. https://doi.org/10.1002/14651858.CD004851.pub3

• Högberg, G. N., Antonuccio, D. O., & Healy, D. (2015). Suicidal risk from TADS study was higher than it first appeared. The International Journal of Risk & Safety in Medicine, 27(2), 85–91. https://doi.org/10.3233/JRS-150645

• Hollon, S. D. (2020). Is cognitive therapy enduring or antidepressant medications iatrogenic? Depression as an evolved adaptation. American Psychologist, 75(9), 1207–1218. https://doi.org/10.1037/amp0000728

• Kaminski, J. A., & Bschor, T. (2020). Antidepressants and suicidality: A re-analysis of the re-analysis. Journal of Affective Disorders, 266, 95–99. https://doi.org/10.1016/j.jad.2020.01.107

• Le Noury, J., Nardo, J. M., Healy, D., Jureidini, J., Raven, M., Tufanaru, C., & Abi-Jaoude, E. (2015). Restoring Study 329: Efficacy and harms of paroxetine and imipramine in treatment of major depression in adolescence. BMJ, h4320. https://doi.org/10.1136/bmj.h4320

• Le Noury, J., Nardo, J. M., Healy, D., Jureidini, J., Raven, M., Tufanaru, C., & Abi-Jaoude, E. (2016). Study 329 continuation phase: Safety and efficacy of paroxetine and imipramine in extended treatment of adolescent major depression. International Journal of Risk & Safety in Medicine, 28(3), 143–161. https://doi.org/10.3233/JRS-160728

• Lu, C. Y., Penfold, R. B., Toh, S., Sturtevant, J. L., Madden, J. M., Simon, G., Ahmedani, B. K., Clarke, G., Coleman, K. J., Copeland, L. A., Daida, Y. G., Davis, R. L., Hunkeler, E. M., Owen-Smith, A., Raebel, M. A., Rossom, R., Soumerai, S. B., & Kulldorff, M. (2018). Near Real-time Surveillance for Consequences of Health Policies Using Sequential Analysis: Medical Care, 56(5), 365–372. https://doi.org/10.1097/MLR.0000000000000893

• Plöderl, M., Amendola, S., & Hengartner, M. P. (2023). Observational studies of antidepressant use and suicide risk are selectively published in psychiatric journals. Journal of Clinical Epidemiology, 162, 10–18. https://doi.org/10.1016/j.jclinepi.2023.07.015

• Plöderl, M., & Hengartner, M. P. (2022). Learning about the course of suicidal behavior but not about the effects of SSRIs. Neuropsychopharmacology, 47(4), 803–803. https://doi.org/10.1038/s41386-021-01224-x

• Plöderl, M., & Hengartner, M. P. (2023). Effect of the FDA Black Box Suicidality Warnings for Antidepressants on Suicide Rates in the USA: Signal or Noise? Crisis, 44(2), 128–134. https://doi.org/10.1027/0227-5910/a000843

• Plöderl, M., Hengartner, M. P., Bschor, T., & Kaminski, J. A. (2020). Commentary to „antidepressants and suicidality: A re-analysis of the re-analysis“. Journal of Affective Disorders, 273, 252–253. https://doi.org/10.1016/j.jad.2020.04.025

• Plöderl, M., Hengartner, M. P., & Volkmann, C. (2022). Suizidpräventive Effekte von Ketamin und Esketamin: Suicide-preventive effects of ketamine and esketamine. Nervenheilkunde, 41(04), 240–245. https://doi.org/10.1055/a-1761-3192

• Plöderl, M., Horowitz, M. A., & Hengartner, M. P. (2023). Re: “A Multicenter Double-Blind, Placebo-Controlled Trial of Escitalopram in Children and Adolescents with Generalized Anxiety Disorder” by Strawn et al.—Concerning Harm–Benefit Ratio in a Recent Trial About Escitalopram for Generalized Anxiety Disorder. Journal of Child and Adolescent Psychopharmacology, 33(7), 295–296. https://doi.org/10.1089/cap.2023.0029

• Sharma, T., Guski, L. S., Freund, N., & Gøtzsche, P. C. (2016). Suicidality and aggression during antidepressant treatment: Systematic review and meta-analyses based on clinical study reports. BMJ. https://doi.org/10.1136/bmj.i65

• Stone, M. B. (2014). The FDA Warning on Antidepressants and Suicidality—Why the Controversy? New England Journal of Medicine, 371(18), 1668–1671. https://doi.org/10.1056/NEJMp1411138

• Stone, M. B. (2018). In Search of a Pony: Sources, Methods, Outcomes, and Motivated Reasoning. Medical Care, 56(5), 375–381. https://doi.org/10.1097/MLR.0000000000000895

• Stone, M. B., Laughren, T. P., Jones, M. L., Levenson, M., Holland, P. C., Hughes, A., Hammad, T. A., Temple, R., & Rochester, G. (2009). Risk of suicidality in clinical trials of antidepressants in adults: Analysis of proprietary data submitted to US Food and Drug Administration. BMJ, 339, 1–10.

• Teicher, M. H., Glod, C., & Cole, J. O. (1990). Emergence of intense suicidal preoccupation during fluoxetine treatment. The American journal of psychiatry, 147(2), 207–210.

• Witt, K., Townsend, E., Arensman, E., Gunnell, D., Hazell, P., Taylor Salisbury, T., Van Heeringen, K., & Hawton, K. (2020). Psychosocial Interventions for People Who Self-Harm: Methodological Issues Involved in Trials to Evaluate Effectiveness. Archives of Suicide Research, 24(sup2), S32–S93. https://doi.org/10.1080/13811118.2019.1592043

• Zainal, N. H. (2023). Is combined antidepressant medication (ADM) and psychotherapy better than either monotherapy at preventing suicide attempts and other psychiatric serious adverse events for depressed patients? A rare events meta-analysis. Psychological Medicine, 1–16. https://doi.org/10.1017/S0033291723003306

• Zalsman, G., Hawton, K., Wasserman, D., van Heeringen, K., Arensman, E., Sarchiapone, M., Carli, V., Höschl, C., Barzilay, R., Balazs, J., Purebl, G., Kahn, J. P., Sáiz, P. A., Lipsicas, C. B., Bobes, J., Cozman, D., Hegerl, U., & Zohar, J. (2016). Suicide prevention strategies revisited: 10-year systematic review. The Lancet Psychiatry, 3(7), 646–659. https://doi.org/10.1016/S2215-0366(16)30030-X