- Open Access
- Open Peer Review
Guidelines for the pharmacological acute treatment of major depression: conflicts with current evidence as demonstrated with the German S3-guidelines
BMC Psychiatry volume 19, Article number: 265 (2019)
Several international guidelines for the acute treatment of moderate to severe unipolar depression recommend a first-line treatment with antidepressants (AD). This is based on the assumption that AD obviously outperform placebo, at least in the case of severe depression. The efficacy of AD for severe depression can only be definitely clarified with individual patient data, but corresponding studies have only been available recently. In this paper, we point out discrepancies between the content of guidelines and the scientific evidence by taking a closer look at the German S3-guidelines for the treatment of depression. Based on recent studies and a systematic review of studies using individual patient data, it turns out that AD are marginally superior to placebo in both moderate and severe depression. The clinical significance of this small drug-placebo-difference is questionable, even in the most severe forms of depression. In addition, the modest efficacy is likely an overestimation of the true efficacy due to systematic method biases. There is no related discussion in the S3-guidelines, despite substantial empirical evidence confirming these biases. In light of recent data and with their underlying biases, the recommendations in the S3-guidelines are in contradiction with the current evidence. The risk-benefit ratio of AD for severe depression may be similar to the one estimated for mild depression and thus could be unfavorable. Downgrading of the related grade of recommendation would be a logical consequence.
Guidelines may be crucial for adequate treatment if they systematically and critically evaluate the evidence and infer treatment recommendations in a rational and transparent manner. This way, guidelines are an important interface between science and clinical practice. The obvious benefit of guidelines vanishes if the recommendations are misleading, for example because of biases in the synthesis of the evidence [1, 2], or simply because the evidence in the guidelines is outdated and conflicting with current evidence. Correcting the discrepancies between the content of the guidelines and current evidence is of utmost importance to avoid potentially harming patients. This seems to be the case for the acute pharmacological treatment of unipolar depression (synonymous to major depression), as we demonstrate in this article. We will mainly focus on the German S3-guidelines from 2015 (with updates until March 2017) . However, algorithms in other guidelines are largely comparable, for example in the guidelines of organizations such as RANZCP (Australia and New Zealand) or NICE (UK) [4, 5], thus our findings are relevant beyond Germany.
We reviewed the sections of the S3-guidelines about the acute pharmacological treatment of unipolar depression (sections 3.4.1. to 3.4.4) with two objectives. First, we investigated if the data about the efficacy of antidepressants (AD) is still in line with current meta-analytic evidence, and also if the clinical importance of the findings is discussed. Since main arguments of the treatment recommendations rely heavily on the efficacy of AD for different levels of depression severity, we included a simple systematic review of related efficacy studies based on individual patient data. We therefore systematically searched PubMed on November 21, 2018, using the following terms: (“individual participant” OR “individual patient” OR “participant level” OR “patient level” OR “individual level”) AND (“meta” OR “meta-analysis”) AND (depression OR SSRI OR SNRI OR antidepressants OR “mood disorder” OR “affective disorder”). This resulted in 185 hits. After screening the abstracts, 149 studies could be excluded because they obviously did not include relevant information. The remaining 36 studies were screened in detail and 10 studies included primary information of interest [6,7,8,9,10,11,12,13,14,15]. We also checked the references of these studies and could find one more relevant study . The 11 relevant studies are summarized in Table 2. The second objective was to review if empirically supported method-biases were adequately addressed as limitations in the judgment of the evidence .
Results and discussion
Efficacy of antidepressants
Comparing the evidence in the guideline with current evidence
In the S3-guidelines, the efficacy of antidepressants (AD) in the acute treatment of major depression is summarized as follows :
To prove a clinically relevant efficacy of acute antidepressant treatment in placebo-controlled trials, a minimum improvement of 50% on established scales (e.g., the Hamilton Rating Scale) is suggested […] In these kinds of clinical trials with a maximum duration of up to twelve weeks, the response rates mostly range between 50 and 60%, the placebo response rates about 25–35% (p. 67).Footnote 1
Thus, the difference in response rates between AD and placebo is reported to be around 25%. This conclusion is based on two outdated studies; a meta-analysis and a review [18, 19]. The 25%-difference contradicts the results from current meta-analyses which reported a difference of about 10% [20, 21], with response rates of approximately 50 and 40% for AD and placebo, respectively (Table 1). A common counter-argument is that response rates for placebo have increased over the years, leading to decreasing AD-placebo differences. This argument is often based on an outdated meta-analysis of Walsh et al. from 2002 . However, a recent meta-analysis found that the placebo-response rates did not increase from 1991 onwards . Therefore, the 25–35% placebo response rate and the approximately 25% difference in response rates between AD and placebo reported in the S3-guidelines substantially deviate from the current evidence.
We also noted a discrepancy between the summary statement regarding the efficacy of AD (50–60% responders on AD as compared to 25–35% on placebo) and the two studies that were cited in support of this statement [18, 19]. One study  claimed that “there is a far-reaching agreement” that two-third of patients respond to AD, but this is not supported by the referenced evidence (Table 1). Furthermore, both cited studies reported differences in response rates between AD and placebo of only 20% and not 25%. In addition, it is surprising that the S3-guidelines did not include meta-analyses that were already available before the guidelines were updated and published [6, 7, 23,24,25,26,27,28] (see Table 1). These newer meta-analyses found substantially lower differences in response rates between AD and placebo than the reported 25%, and also much higher placebo response rates. Thus, even without the latest meta-analyses published after 2017, the overall assessment of efficacy should have been different.
The impression of an exaggerated presentation of the efficacy of AD also occurs in the discussion of the efficacy of different types of AD. For SSRIs, the following is claimed:
The group of selective serotonin-reuptake-inhibitors (SSRI) […] increases the central serotonergic neurotransmission by selectively inhibiting the reuptake of serotonin from the synaptic cleft. This explains the antidepressant effects as well as the side effects. The efficacy of selective serotonin reuptake inhibitors (SSRIs) in the treatment of acute depressive episodes has been demonstrated in many clinical studies versus placebo and in corresponding meta-analyses. (p. 69).
Some of the SSRI-trials cited in the S3-guidelines reported rather small effect-sizes and this should have raised doubts on the summary efficacy statement mentioned above. More importantly, the largest and most recent meta-analysis cited in the S3-guidelines  reported a high response rate for placebo (41–47%), which grossly deviates from the summary statement (25–35%).
One reason why recent meta-analyses reported smaller differences between AD and placebo lies in the fact that they were based on both published and unpublished studies, whereas earlier meta-analyses exclusively relied on studies published in scientific journals [20, 21, 30]. A related well known publication bias is that positive studies were almost always published in scientific journals (sometimes multiple times), but negative trials were rarely published [31, 32]. According to a comprehensive analysis of the trial-results available to the FDA, only 51% of studies were positive and 97% of these studies were published as positive studies in journals. In contrast, only 3% of negative studies were published as being negative in a journal. Furthermore, 21% of negative studies were published as being positive, for example by only reporting on a secondary outcome that was then falsely reported to be the primary outcome, or by only reporting the results of a subgroup. All other negative studies remained unpublished . Thus, despite that only about half of the AD-trials were positive, nearly all related published studies report positive findings . This important bias is briefly mentioned in the S3-guidelines, but the implications are not considered any further in the evaluation of the evidence from published AD trials.
One common explanation for the modest efficacy of AD in more recent studies is that there is a trend to only include less severely depressed patients or those without frequent prior depressive episodes  (p. 308). However, this does not seem to be the case, instead, it was the rate of drop-outs due to inefficacy in placebo-groups that has changed . The average drop-out rate in the year 1985 was 58% and of those who discontinued the studies early, 93% stated lack of efficacy as a reason. In the year 2009, only 20% of patients in the placebo-group dropped out, and only 15% attributed this to lack of efficacy . The massive reduction of placebo-dropouts due to lack of efficacy is crucial, because this can fully explain the reduced efficacy of AD in more recent studies. Moreover, this effect appears to be robust and consistent, as it is independent of the length of the study or sample-size. Thus, instead of the typical explanation that the placebo-response is miraculously greater in more recent studies, a more accurate interpretation is that patients on placebo do not immediately drop-out if they do not recognize some effect of the drug  (this also raises the question of successful blinding of patients and doctors in older trials). Since patients could be kept longer in more recent studies, it seems that substantially more patients in the placebo-group achieve spontaneous remission until the end of the trial, leading to a reduction of the difference between AD and placebo, even when they may not perceive a drug effect.
Discussion of clinical significance
There is a controversy about the appropriateness of using response rates, because this can lead to an overestimation of the efficacy of a treatment  (also see footnote 2). This problem is briefly mentioned in the S3-guidelines:
Furthermore, the efficacy in comparison to placebo is mostly based on the higher response rate, whereas the difference in remission-rates or the reduction of summary-scores of depression rating-scales is often not significant (p. 67).
However, it is not discussed what “not significant” actually implies. In the meantime, it has been replicated many times that even though the AD-placebo difference is statistically significant, this effect may not be clinically significant [17, 21, 36]. This was already discussed in publications available at the time well before the S3-guidelines were published [35, 37, 38]. For example, Kirsch and colleagues demonstrated that most variance (> 75%) in the outcome in the SSRI groups can be attributed to placebo-responses, and the rest may result from enhanced placebo responses due to perceived side-effects of AD . According to the most recent meta-analysis of Cipriani and colleagues , the overlap between AD and placebo is even larger (88%) [17, 39].
Admittedly, there is no universal definition of “clinical significance” (see Footnote 2). However, AD do not meet any criterion for clinical significance, not even the most liberal [17, 39]. This is not surprising, because the average difference of AD compared to placebo is only about 2 points on the HAMD-17 depression rating scale that has a range from 0 to 52 points (most items are scored between 0 and 4). This is intuitively a very modest and unimportant effect, which is also confirmed when the 2 point difference is compared to clinical judgments made by mental health professionals. If the HAMD is compared to the clinical evaluation using the Clinical Global Impression Improvement Scale (CGI-I), then 0–3 points improvement on the HAMD correspond to “no improvement” on the CGI-I. It needs at least 7 points improvement on the HAMD scale to achieve a corresponding “minimal improvement” on the CGI-I. None of the AD come anywhere near this criterion .
Furthermore, the S3-guidelines seem to have a contradictory use of clinical significance, because it is questioned in one section and then taken for granted in other sections. When the efficacy of AD for mild depression is discussed (p. 68), the criterion of 3 HAMD-points for clinical significance is questioned with the argument that this criterion was removed from the current NICE guidelines. This is wrong, because the NICE guidelines from 2010 did include this criterion in an appendix .Footnote 2 Doubts on the criterion for clinical significance also appear when discussing a study which reported less than 3 HAMD-points difference between AD and placebo for both mild and more severe depression . Interestingly, this important study is then ignored in the following section (also p. 68) about the treatment of moderate to severe depression. Instead, it is stated that for severe depression, AD are clinically superior to placebo, based on the 3-point criterion for clinical significance.
Efficacy of AD in relation to depression severity – guidelines versus current evidence from a systematic review
The S3-guidelines report that, for mild depression, AD are not superior to placebo, resulting in an unfavorable negative risk-benefit ratio because of the side-effects of AD. The NICE guidelines include very similar arguments: “Do not use antidepressants routinely to treat persistent subthreshold depressive symptoms or mild depression because the risk-benefit ratio is poor (p. 327)” . Likewise, the RANZCP guidelines recommend that “patients with mild-moderate depression should be offered one of the evidence based psychotherapies as first line treatment” (p. 1108)  (the negative risk-benefit ratio is not explicitly stated but the logical argument behind this conclusion is given).
For moderate to severe depression, the S3-guidelines report that AD have a clinically significant effect:
For medium to severe depression, however, the difference in efficacy between antidepressants and placebo is more pronounced, since in the most severe forms up to 30% of treated patients benefit from antidepressants above the placebo rate. Thus, HDRS scores of > 24 are associated with the most consistent difference between the response to drug and placebo, whereby these differences in the direction of the active antidepressant are also clinically significant (p. 68).
This statement is based on a single citation, referring to a study by Khan et al. (2005), but this study is not related to depression at all and is most likely a citation error. We guess that the authors of the S3-guidelines wanted to refer either to another publication of Khan , or to the meta-analysis of Fournier et al.  that is frequently cited in this context.
To clarify if AD are more efficacious for severely depressed patients, individual-level data from patients are needed, because using group means leads to substantial biases (referred to as ecological fallacy) . It is surprising that this argument is completely lacking in the S3-guidelines, even more so, as two such studies with individual patient data were cited in the S3-guidelines, and these studies addressed the problems resulting from group-level data [6, 9]. In addition, one of these studies did not find AD to be clinically effective for severe depression , but this study was not discussed appropriately, as we already noted above.
Our simple systematic review of studies with individual patient-level data could locate 11 relevant studies that are summarized in Table 2. It can be concluded that most patient-level meta-analyses, especially the more recent and larger ones, reported that AD are not clinically significantly superior to placebo, even for severe depression (< 3 HAMD-points difference between AD and placebo). One exception is a study in older patients, where one subgroup (severely and chronically depressed patients) responded much better to AD than to placebo . However, this could be a false positive finding because of multiple testing of many different subgroups. Also, according to the meta-analysis of Fournier et al. , AD were substantially more efficacious than placebo in patients with a baseline score of ≥23 on the HAMD, but this was refuted in recent and larger meta-analyses. One very recent study reported that placebo is slightly more effective than AD for the most severely depressed patients . Finally, it was also found that AD were not more efficacious for the melancholic subtype of depression – which is associated with higher depression-scores and seen as the most severe form of depression by many experts .
Discussion of method biases
The S3-guidelines did not include a discussion of important biases, except for the publication bias:
In the perception of the (specialist) public, the efficacy of antidepressants is rather overestimated, since studies in which the antidepressant performed better than placebo are published much more frequently in scientific journals than those in which the antidepressant was not superior to placebo (p. 67).
So the publication bias is briefly mentioned, but it was not considered elsewhere. This is problematic in sections where treatments were compared with each other, based on single or very few published studies. Due to the publication and sponsorship bias, where negative results are rarely published, these comparisons are likely biased . Moreover, throughout the guidelines, the efficacy of different treatment approaches is often based on statistical significance alone. It is known that statistical significance is not informative about the size of a difference or about clinical significance .
There are many more biases that may lead to an overestimation of the efficacy of AD, but they were not discussed in the S3-guidelines. Such biases include unblinding due to specific side-effects of AD, exclusion of patients who improve in the placebo lead-in phase, withdrawal effects in the placebo group due to abrupt discontinuation of pre-trial AD prescriptions, inadequate handling of missing data with last observations-carried forward, and other biases [44,45,46]. Some of these biases, for example the breaking of the double-blinding due to correct guessing of placebo or drug, have been replicated in various empirical studies and are known for a long time [47, 48]. There is also sound evidence that unblinded physicians judge the drug as being more effective than blinded physicians [49, 50]. Just recently, it was found that trials with a placebo lead-in phase produce significantly larger efficacy estimates than the minority of trials without such a lead-in phase (d = 0.31 vs. d = 0.22) . This was long expected by various experts, because patients who improve during the placebo lead-in phase are excluded from the trial, biasing the results in favor of AD. Thus, it can be concluded with a high degree of certainty, that the efficacy of AD is overestimated in typical clinical trials. In contrast, we are not aware of empirical studies confirming postulated biases leading to an underestimation of the efficacy of AD [52, 53]. On the contrary, some of these biases were refuted in the meantime. For example, it is often claimed that AD work much better in real-world patients. However, AD are no more effective in patients treated in the real-world routine practice compared to those selected for clinical trials, as clearly demonstrated in the STAR*D study [54, 55] or in a meta-analysis of real-world primary care patients . Some other assumed biases do not seem very plausible, for example the argument that patients lie about their depression to be included in studies in order to obtain treatment for free or to receive some money. Even if this is so, there is no plausible explanation as to why this should lead to biased drug-placebo differences, since these malingerers would be randomly assigned to treatment arms. In any case, there is no empirical evidence that would support such an assumption, and as such it is no more than an untested hypothesis. Another popular argument is that some trials allow additional treatment with benzodiazepines and other tranquilizers, but this would affect both the AD and the placebo groups similarly, so this is no systematic bias and both direction and size of the bias are still unknown.
The S3-guidelines and other international guidelines do not recommend AD as first-line treatment for mild depression, because:
Due to the unfavorable risk-benefit ratio, antidepressants are not generally useful in the initial treatment of mild depressive episodes, since antidepressant medication is hardly superior to a placebo condition (p. 74, citations removed).
As we have shown in this paper and discussed elsewhere [17, 39], AD are indeed hardly superior to placebo in mild depression, but the same holds for moderate and severe depression (i.e., less than three points on the HAMD scale or approximately 10% difference in response rates). This already modest efficacy is most likely an overestimation of the true effect size due to various systematic method biases inherent in clinical trials. Therefore, the degree of recommendation for the pharmacological acute treatment of moderate and severe depression with AD should be downgraded on the basis of the guidelines’ own logic. We are not alone with such conclusions. Munkholm et al.  recently re-analyzed the trial data for moderate to severe depression collected by Cipriani et al. , and based on the poor efficacy estimates and the many systematic biases in these trials, they concluded that “the evidence does not support definitive conclusions regarding the efficacy of antidepressants for depression in adults, including whether they are more efficacious than placebo” (p. 8). Consequently, this impacts the risk-benefit ratio of AD in the acute treatment of major depression, as well as comparisons of AD with alternative treatments. Therefore, treatment recommendations should be critically discussed in light of the current evidence. This clearly goes beyond the scope of this paper, but good examples are available . We hope that our review can inform clinicians until the guideline will be updated accordingly.
Availability of data and materials
The list of studies of the systematic review and additional information can be found here: https://osf.io/4kh2a/
All quotations from the S3-guidelines were translated into English by the authors.
Unfortunately, the NICE guidelines did not justify the different definitions of clinical significance. Three different criteria for clinical significance were defined: First, a ≥ 3 points difference between AD and placebo on the HAMD scale (or the BDI scale). Second, an effect-size of d ≥ 0.5 (equivalent to approximately 3.8 points difference on the HAMD scale ). Third, a risk-ratio (RR) of RR ≤ 0.8 for response rates. Of note, these criteria are an absolute minimum, corresponding to a “no improvement” clinical judgment, but this is not mentioned. Furthermore, the three criteria are not equivalent, leading to contradictory conclusions. For example, the average effect-size in a recent meta-analysis  was d = 0.3 (clearly below the required d = 0.5), corresponding to a 2.4 HAMD points difference (below the required 3 points), but to a risk ratio of RR = 0.8 (only just fulfilling the criterion).
Beck Depression Inventory
Clinical Global Impression Scale
Food and Drug Administration
- HAMD or HDRS:
Hamilton Depression Rating Scale
Hamilton Depression Rating Scale, version with 17 items
Hamilton Rating Scale points difference between AD and Placebo
National Institute for Health and Care Excellence
Royal Australian and New Zealand College of Psychiatrists
Randomized controlled trial
Serotonin-noradrenaline reuptake inhibitor
Selective serotonin reuptake inhibitor
Woolf SH, Grol R, Hutchinson A, Eccles M, Grimshaw J. Clinical guidelines: potential benefits, limitations, and harms of clinical guidelines. BMJ. 1999;318:527–30.
Ioannidis JPA. Professional societies should abstain from authorship of guidelines and disease definition statements. Circ Cardiovasc Qual Outcomes. 2018;11. https://doi.org/10.1161/CIRCOUTCOMES.118.004889.
DGPPN. S3-Leitlinie/Nationale Versorgungsleitlinie Unipolare Depression - Langfassung, 2. Auflage, 5. Version; 2015. https://doi.org/10.6101/AZQ/000364.
Malhi GS, Bassett D, Boyce P, Bryant R, Fitzgerald PB, Fritz K, et al. Royal Australian and new Zealand College of Psychiatrists clinical practice guidelines for mood disorders. Aust N Z J Psychiatry. 2015;49:1087–206.
National Institute for Health and Clinical Excellence. Depression: the NICE guideline on the treatment and management of depression in adults. Updated edition 2018. Leicester: British Psychological Society; 2010. https://www.ncbi.nlm.nih.gov/pubmed/22132433.
Gibbons RD, Hur K, Brown CH, Davis JM, Mann JJ. Benefits from antidepressants: synthesis of 6-week patient-level outcomes from double-blind placebo-controlled randomized trials of fluoxetine and venlafaxine. Arch Gen Psychiatry. 2012;69:572–9.
Nelson JC, Delucchi KL, Schneider LS. Moderators of outcome in late-life depression: a patient-level meta-analysis. Am J Psychiatry. 2013;170:651–9.
Thase ME, Pritchett YL, Ossanna MJ, Swindle RW, Xu J, Detke MJ. Efficacy of duloxetine and selective serotonin reuptake inhibitors: comparisons as assessed by remission rates in patients with major depressive disorder. J Clin Psychopharmacol. 2007;27:672–6.
Fournier JC, DeRubeis RJ, Hollon SD, Dimidjian S, Amsterdam JD, Shelton RC, et al. Antidepressant drug effects and depression severity: a patient-level meta-analysis. JAMA. 2010;303:47–53.
Khan A, Bhat A, Faucett J, Kolts R, Brown WA. Antidepressant-placebo differences in 16 clinical trials over 10 years at a single site: role of baseline severity. Psychopharmacology. 2011;214:961–5.
Harada E, Schacht A, Koyama T, Marangell L, Tsuji T, Escobar R. Efficacy comparison of duloxetine and SSRIs at doses approved in Japan. Neuropsychiatr Dis Treat. 2015;11:115–23.
Cuijpers P, Weitz E, Lamers F, Penninx BW, Twisk J, DeRubeis RJ, et al. Melancholic and atypical depression as predictor and moderator of outcome in cognitive behavior therapy and pharmacotherapy for adult depression. Depress Anxiety. 2017;34:246–56.
Debray TP, Schuit E, Efthimiou O, Reitsma JB, Ioannidis JP, Salanti G, et al. An overview of methods for network meta-analysis using individual participant data: when do benefits arise? Stat Methods Med Res. 2018;27:1351–64.
Furukawa TA, Maruo K, Noma H, Tanaka S, Imai H, Shinohara K, et al. Initial severity of major depression and efficacy of new generation antidepressants: individual participant data meta-analysis. Acta Psychiatr Scand. 2018;137:450–8.
Nakabayashi T, Hara A, Minami H. Impact of demographic factors on the antidepressant effect: a patient-level data analysis from depression trials submitted to the pharmaceuticals and medical devices Agency in Japan. J Psychiatr Res. 2018;98:116–23.
Rabinowitz J, Werbeloff N, Mandel FS, Menard F, Marangell L, Kapur S. Initial depression severity and response to antidepressants v. placebo: patient-level data analysis from 34 randomised controlled trials. Br J Psychiatry. 2016;209:427–8.
Hengartner MP, Plöderl M. Statistically significant antidepressant-placebo differences on subjective symptom-rating scales do not prove that antidepressants work: effect size and method bias matter! Front Psychiatry. 2018;9. https://doi.org/10.3389/fpsyt.2018.00517.
Oeljeschläger B, Müller-Oerlinghausen B. Wege zur Optimierung der individuellen antidepressiven Therapie. Dtsch Ärztebl. 2004;19:A 1337–40.
Walsh BT, Seidman SN, Sysko R, Gould M. Placebo response in studies of major depression: variable, substantial, and growing. JAMA. 2002;287:1840–7.
Cipriani A, Furukawa TA, Salanti G, Chaimani A, Atkinson LZ, Ogawa Y, et al. Comparative efficacy and acceptability of 21 antidepressant drugs for the acute treatment of adults with major depressive disorder: a systematic review and network meta-analysis. Lancet. 2018;391:1357–66.
Jakobsen JC, Katakam KK, Schou A, Hellmuth SG, Stallknecht SE, Leth-Møller K, et al. Selective serotonin reuptake inhibitors versus placebo in patients with major depressive disorder. A systematic review with meta-analysis and Trial sequential analysis. BMC Psychiatry. 2017;17. https://doi.org/10.1186/s12888-016-1173-2.
Furukawa TA, Cipriani A, Leucht S, Atkinson LZ, Ogawa Y, Takeshima N, et al. Is placebo response in antidepressant trials rising or not? A reanalysis of datasets to conclude this long-lasting controversy. Evid Based Ment Health. 2018;21:1–3.
Furukawa TA, Cipriani A, Atkinson LZ, Leucht S, Ogawa Y, Takeshima N, et al. Placebo response rates in antidepressant trials: a systematic review of published and unpublished double-blind randomised controlled studies. Lancet Psychiatry. 2016;3:1059–66.
Weitz ES, Hollon SD, Twisk J, van Straten A, Huibers MJH, David D, et al. Baseline depression severity as moderator of depression outcomes between cognitive behavioral therapy vs pharmacotherapy: an individual patient data meta-analysis. JAMA Psychiatry. 2015;72:1102.
Undurraga J, Baldessarini RJ. Randomized, placebo-controlled trials of antidepressants for acute major depression: thirty-year meta-analytic review. Neuropsychopharmacology. 2012;37:851–64.
Melander H, Salmonson T, Abadie E, van Zwieten-Boot B. A regulatory apologia — a review of placebo-controlled studies in regulatory submissions of new-generation antidepressants. Eur Neuropsychopharmacol. 2008;18:623–7.
Arroll B, Macgillivray S, Ogston S, Reid I, Sullivan F, Williams B, et al. Efficacy and tolerability of tricyclic antidepressants and SSRIs compared with placebo for treatment of depression in primary care: a meta-analysis. Ann Fam Med. 2005;3:449–56.
Storosum JG, Elferink AJA, Van Zwieten BJ, Van den Brink W, Huyser J. Natural course and placebo response in short-term, placebo-controlled studies in major depression: a meta-analysis of published and non-published studies. Pharmacopsychiatry. 2004;38:32–6.
McCormack J, Korownyk C. Effectiveness of antidepressants. BMJ. 2018;:k1073.
Monden R, Roest AM, van Ravenzwaaij D, Wagenmakers E-J, Morey R, Wardenaar KJ, et al. The comparative evidence basis for the efficacy of second-generation antidepressants in the treatment of depression in the US: a Bayesian meta-analysis of Food and Drug Administration reviews. J Affect Disord. 2018;235:393–8.
Melander H, Ahlqvist-Rastad J, Meijer G, Beermann B. Evidence b(i)ased medicine—selective reporting from studies sponsored by pharmaceutical industry: review of studies in new drug applications. BMJ. 2003;326:1171–3.
Turner EH, Matthews AM, Linardatos E, Tell RA, Rosenthal R. Selective publication of antidepressant trials and its influence on apparent efficacy. N Engl J Med. 2008;358:252–60.
de Vries YA, Roest AM, de Jonge P, Cuijpers P, Munafò MR, Bastiaansen JA. The cumulative effect of reporting and citation biases on the apparent efficacy of treatments: the case of depression. Psychol Med. 2018;48:2453–5.
Schalkwijk S, Undurraga J, Tondo L, Baldessarini RJ. Declining efficacy in controlled trials of antidepressants: effects of placebo dropout. Int J Neuropsychopharmacol. 2014;17:1343–52.
Kirsch I, Moncrieff J. Clinical trials and the response rate illusion. Contemp Clin Trials. 2007;28:348–51.
Kirsch I, Deacon BJ, Huedo-Medina TB, Scoboria A, Moore TJ, Johnson BT. Initial severity and antidepressant benefits: a meta-analysis of data submitted to the Food and Drug Administration. PLoS Med. 2008;5:e45.
Kirsch I, Moore TJ, Scoboria A, Nicholls SS. The emperor’s new drugs: an analysis of antidepressant medication data submitted to the US Food and Drug Administration. Prev Treat. 2002;5:23a.
Kirsch I, Sapirstein G. Listening to Prozac but hearing placebo: A meta-analysis of antidepressant medication. Prev Treat. 1998;1:2a.
Hengartner MP. What is the threshold for a clinical minimally important drug effect? BMJ Evid-Based Med. 2018;23:225–7.
Khan A, Leventhal RM, Khan SR, Brown WA. Severity of depression and response to antidepressants and placebo: an analysis of the Food and Drug Administration database. J Clin Psychopharmacol. 2002;22:40–5.
Lau J, Ioannidis JP, Schmid CH. Summing up evidence: one answer is not always enough. Lancet. 1998;351:123–7.
Moncrieff J, Kirsch I. Efficacy of antidepressants in adults. BMJ. 2005;331:155–7.
Flacco ME, Manzoli L, Boccia S, Capasso L, Aleksovska K, Rosso A, et al. Head-to-head randomized trials are mostly industry sponsored and almost always favor the industry sponsor. J Clin Epidemiol. 2015;68:811–20.
Gøtzsche PC. Deadly psychiatry and organised denial. Kopenhagen: People’s Press; 2015.
Hengartner MP. Methodological flaws, conflicts of interest, and scientific fallacies: implications for the evaluation of antidepressants’ efficacy and harm. Front Psychiatry. 2017;8.
Wang S-M, Han C, Lee S-J, Jun T-Y, Patkar AA, Masand PS, et al. Efficacy of antidepressants: bias in randomized clinical trials and related issues. Expert Rev Clin Pharmacol. 2018;11:15–25.
Fisher S, Greenberg RP. How sound is the double-blind design for evaluating psychotropic drugs? J Nerv Ment Dis. 1993;181:345–50.
Even C, Siobud-Dorocant E, Dardennes RM. Critical approach to antidepressant trials. Br J Psychiatry. 2000;177:47–51.
Hrobjartsson A, Thomsen ASS, Emanuelsson F, Tendal B, Hilden J, Boutron I, et al. Observer bias in randomised clinical trials with binary outcomes: systematic review of trials with both blinded and non-blinded outcome assessors. BMJ. 2012;344:e1119.
Hrobjartsson A, Thomsen ASS, Emanuelsson F, Tendal B, Hilden J, Boutron I, et al. Observer bias in randomized clinical trials with measurement scale outcomes: a systematic review of trials with both blinded and nonblinded assessors. Can Med Assoc J. 2013;185:E201–11.
Munkholm K, Paludan-Müller AS, Boesen K. Considering the methodological limitations in the evidence base of antidepressants for depression: a reanalysis of a network meta-analysis. BMJ Open. 2019;9:e024886.
Möller HJ. Isn’t the efficacy of antidepressants clinically relevant? A critical comment on the results of the metaanalysis by Kirsch et al. 2008. Eur Arch Psychiatry Clin Neurosci. 2008;258:451–5.
Hegerl U, Mergl R. The clinical significance of antidepressant treatment effects cannot be derived from placebo-verum response differences. J Psychopharmacol (Oxf). 2010;24:445–8.
Pigott HE. The STAR*D trial: it is time to reexamine the clinical beliefs that guide the treatment of major depression. Can J Psychiatr. 2015;60:9–13.
Kirsch I, Huedo-Medina TB, Pigott HE, Johnson BT. Do outcomes of clinical trials resemble those “real world” patients? A reanalysis of the STAR* D antidepressant data set. Psychol Conscious Theory Res Pract. 2018;5:339–45.
Arroll B, Elley CR, Fishman T, Goodyear-Smith FA, Kenealy T, Blashki G, et al. Antidepressants versus placebo for depression in primary care. Cochrane Database Syst Rev. 2009. https://doi.org/10.1002/14651858.CD007954.
Gartlehner G, Gaynes BN, Amick HR, Asher GN, Morgan LC, Coker-Schwimmer E, et al. Comparative benefits and harms of antidepressant, psychological, complementary, and exercise treatments for major depression: an evidence report for a clinical practice guideline from the American College of Physicians. Ann Intern Med. 2016;164:331–42.
Ethics approval and consent to participate
No human subject was involved in this study.
Consent for publication
The authors declare that they have no competing interest. MP works as a clinical psychologist in a public psychiatric hospital, where the pharmacological treatment of psychiatric disorders is a central part. The conclusions from this paper could potentially lead to conflicts. Therefore, MP decided to prepare this manuscript in his leisure time and also wants to express that the content of this paper is not related to his clinical psychological practice with patients.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.