Hooked on Hype? A Critical Examination of Recent Anti-Antidepressant Reporting

COMMENTARY

Regular reports are published in the lay media regarding the alleged lack of efficacy of antidepressants. This time, it was Newsweek’s turn, with Adam Piore¹ highlighting, amongst other reports, a new analysis by Stone et al,² in which it was found that antidepressants work better than placebo only 15% of the time.

As usual in this type of reporting, “critical” voices are heard mostly unopposed, and little room is made for actual evidence-based interpretation of the data presented. In this article, I will review some of the claims made, provide evidence that the statistic 15% is most likely an underestimation of real-world antidepressant effects, and cast doubt on some other claims made about the (lack of) efficacy of antidepressants.

No Stone Unturned

In their analysis, Stone et al included all antidepressants and possible age groups. Thus, their estimate represents the average effects across all antidepressant trials submitted to the US Food and Drug Administration (FDA), but not necessarily one of the average antidepressants used. The 5 worst drugs, according to their analysis, are very rarely used for depression. Thus, the much more commonly used SSRIs outscore the least efficacious antidepressant, dothiepine, by a considerable margin. Further, antidepressants are probably less effective in treating depression in children and adolescents than in adults.³ However, they are also much less prescribed in the younger population, but the researchers did not account for that.

The main outcome measure in the Stone et al analysis is the sum score of the 17 items of the Hamilton Depression Rating Scale (HDRS-17-sum). The use of this measure has been widely criticized for many reasons: it lacks one-dimensionality,⁴ it might reflect factors other than depressive symptoms,⁴ and the majority of the measured symptoms are not present in many individuals with depression.⁵ Drug-placebo differences were found to be substantially larger when the sum score of 6 core depressive symptoms included in the unidimensional HDRS-6 scale were used.^5,6 It was also found that drug-placebo differences for this subscale are similar in those with severe and nonsevere depression,⁷ although the Newsweek article made a contrary claim against antidepressants.

HDRS items not included in the HDRS-6 are prone to pick up adverse effects from treatment, thus further diminishing apparent drug-placebo differences as measured using HDRS-17 sum score.⁸ Although adverse effects are certainly important to consider, including them in an efficacy measure is unreliable; for example, this measure is unable to distinguish between a treatment with low efficacy but no adverse effects and a highly effective treatment with more adverse effects.

Dosing and Other Issues

Dosing is also an important issue. Drug-placebo differences are about twice as large at doses above the common starting one for the most frequently prescribed SSRIs compared.⁹ Yet, in the Stone et al analysis, all possible doses were included, including those shown to not be significantly more effective than placebo. In a recent publication,⁶ for example, a 22.3% drug-placebo difference was found for optimal SSRI doses in adults when using response (50% score reduction from baseline) on the HDRS-6 sum score as outcome measure; however, this number is more likely an underestimation than an overestimation.

Other issues in clinical trials include difficulties in finding drug-placebo separation, noncompliance to randomized drugs,¹⁰ high drop-out rates,¹¹ patient recruitment difficulties,¹² and substantial placebo response rates.¹³ Further difficulties are presented by the diagnosis itself, major depressive disorder being a multi-faceted, highly heterogenous disorder with low diagnostic reliability.¹⁴ Thus, it is unreasonable to expect substantial response and remission rates for any treatment for depression. Additionally, using strictly regulated trial protocols disallowing many concurrent treatments also impacts response and remission rates.

Although only company-sponsored trials were analyzed in most papers cited here, as well as in the Stone et al paper, unsponsored trials were found to show about 50% higher drug-placebo differences.¹⁵ It is possible that the studied populations in the latter more accurately reflect patients receiving treatment in clinical practice. This is another issue not reflected in most critical opinions about antidepressants, where it is instead stressed that trials are funded by pharmaceutical companies, with the results thus being inflated and untrustworthy.

Psychotherapy

Another treatment highlighted in the Newsweek article is psychotherapy, which is heralded as safer and more effective than antidepressants. Since safety data are very rarely collected in psychotherapy trials, comparisons in this regard are thus impossible. Also, neither the assumption of better efficacy holds true. In those few trials in which psychotherapy is compared to pill placebo, there is a difference of about 0.25 standardized mean differences (SMD),¹⁶ effects similar to the average antidepressant. However, every one of those trials was unblinded, as the patient was certainly able to distinguish psychotherapy from pills. This fact is entirely overlooked by Irving Kirsch, PhD, associate director of the Program in Placebo Studies and lecturer in medicine at the Harvard Medical School and Beth Israel Deaconess Medical Center, and others when interviewed by Newsweek. (This is ironic, considering that he used the possibility of unblinding in antidepressant trials, which evidence shows is a nonissue,¹⁷ as an argument to dismiss the similar differences between antidepressants and placebo.¹⁸)

Discussing the 15%

Understandably, the first author of the Stone et al paper—Marc B. Stone, MD, of the Center for Drug Evaluation and Research, US Food and Drug Administration—declined to comment in the Newsweek article, perhaps because media discussion lies beyond the obligations of FDA employees. My personal primary take-away point from this analysis is not the 15% number itself, but the finding of a subgroup of patients with depression responding uniquely to antidepressants but not to placebo. Previously, other authors including Kirsch attempted to dismiss the average drug-placebo difference as being smaller than a “minimum clinically important difference,” assuming that the average difference is what is being experienced by all patients who are treated with antidepressants. It was, however, pointed out that their suggested “minimal difference” instead corresponds to the maximum achievable effect in this type of trial, and that the average difference does not reflect a general effect, but a higher likelihood of lower depression scores in the drug-treatment group.¹³

In all, it might be that only 15% of those with depression experience significantly positive effects from antidepressant treatment compared to placebo, if the material studied was indeed representative of the average treatment, setting, and patient group. In real-world practice, the actual number might instead substantially differ depending on the drug used, the drug dose, the patient group treated, and the efficacy measure used. Effect estimates in clinical trials are more likely to underestimate than overestimate treatment effects.

The 15% number is then contrasted with 56%—the percentage of patients who apparently experience withdrawal effects. The author leaves us with a take-away that more people have withdrawal effects than antidepressant effects. Unfortunately, the 56% estimate was taken from a flawed¹⁹ review based mostly on survey studies²⁰ that were uncontrolled and thus disregarded potential nocebo effects. Tapering was often abrupt—which is against guidelines—and again no difference was made between different drugs, even though drugs with short half-life (eg, paroxetine, venlafaxine) were driving the results. Ignoring these issues and contrasting estimates from uncontrolled surveys to drug-placebo differences from randomized controlled trials is nothing short of a woeful misuse of statistics. A more honest view could have been obtained by contrasting the withdrawal estimate to an efficacy estimate from the largest survey study, which showed that 83% of those taking antidepressants believed the drugs had reduced their depression.²¹

Contrary Claims

Unfortunately, even more factual inaccuracies were presented in the Newsweek article. It was stated that antidepressants do not aid loss of energy, loss of interest, nor motivation, all being common symptoms in depression. In item-based analyses of SSRIs and the SNRI duloxetine, the opposite was revealed to be true, with small but replicable effects on all these symptoms.^22,23 To the best of my knowledge, no better evidence exists on the contrary, thus rendering these statements by Joanna Moncrieff, MD, in the Newsweek article untrue.

Another statement by the same author claims that the superiority of antidepressants can be attributed not to actual antidepressant effects, but to emotional numbing. A recent article in Psychiatric Times™ by Dawson and Pies rebutted this claim,²⁴ and even more evidence was found in a recent analysis of SSRI trials using the Montgomery-Åsberg Depression Rating Scale (MADRS) as outcome measure.²⁵ Here, an opposite effect to what was stated by the Newsweek interviewees was found, with a significant and positive drug effect against emotional numbing, as measured on the corresponding MADRS item. Thus, although emotional numbing may also be an adverse effect of SSRI treatment, the average effect instead corresponds to the individual with depression experiencing an improved ability to feel emotions.

Concluding Thoughts

The Newsweek article is just another unfortunate example of one-sided media reporting, ignoring evidence contradicting claims of those critical to drug treatment, and casting unnecessary doubt on drugs that, while not perfect, are more than useful in alleviating symptoms of major depression, as well as many other disorders (eg, anxiety syndromes, obsessive compulsive disorder, and premenstrual dysphoric disorder). Unfortunately, I expect this tradition of antidepressant bashing will continue. Lay media seem disinterested in correcting their claims (neither the journalist in question nor Newsweek themselves responded when I shared my concerns) and instead will likely continue to seek sensationalist headlines and short-term profit.

The opinions expressed in the interviews are those of the participants and do not necessarily reflect the opinions of Psychiatric Times™.

Dr Lisinski is in the Department of pharmacology at the Institute of Neuroscience and Physiology, Sahlgrenska Academy, at the University of Gothenburg, Sweden.

References

1. Piore A. Antidepressants work better than sugar pills only 15 percent of the time. Newsweek. September 9, 2022. Accessed November 29, 2022. https://www.newsweek.com/2022/09/30/antidepressants-work-better-sugar-pills-only-15-percent-time-1744656.html

2. Stone MB, Yaseen ZS, Miller BJ, et al. Response to acute monotherapy for major depressive disorder in randomized, placebo controlled trials submitted to the US Food and Drug Administration: individual participant data analysis. BMJ. 2022;378:e067606.

3. Zhou X, Teng T, Zhang Y, et al. Comparative efficacy and acceptability of antidepressants, psychotherapies, and their combination for acute treatment of children and adolescents with depressive disorder: a systematic review and network meta-analysis. Lancet Psychiatry. 2020;7(7):581-601.

4. Fried EI, van Borkulo CD, Epskamp S, et al. Measuring depression over time . . . Or not? Lack of unidimensionality and longitudinal measurement invariance in four common rating scales of depression. Psychol Assess. 2016;28(11):1354-1367.

5. Bech P, Gram LF, Dein E, et al. Quantitative rating of depressive states. Acta Psychiatr Scand. 1975;51(3):161-170.

6. Lisinski A, Hieronymus F, Nilsson S, Eriksson E. Impact of chosen cutoff on response rate differences between selective serotonin reuptake inhibitors and placebo. Transl Psychiatry. 2022;12(1):160.

7. Hieronymus F, Lisinski A, Nilsson S, Eriksson E. Influence of baseline severity on the effects of SSRIs in depression: an item-based, patient-level post-hoc analysis. Lancet Psychiatry. 2019;6(9):745-752.

8. Hieronymus F, Lisinski A, Eriksson E, Ostergaard SD. Do side effects of antidepressants impact efficacy estimates based on the Hamilton Depression Rating Scale? A pooled patient-level analysis. Transl Psychiatry. 2021;11(1):249.

9. Hieronymus F, Nilsson S, Eriksson E. A mega-analysis of fixed-dose trials reveals dose-dependency and a rapid onset of action for the antidepressant effect of three selective serotonin reuptake inhibitors. Transl Psychiatry. 2016;6(6):e834.

10. Reis M, Aberg-Wistedt A, Agren H, et al. Compliance with SSRI medication during 6 months of treatment for major depression: an evaluation by determination of repeated serum drug concentrations. J Affect Disord. 2004;82(3):443-446.

11. Cipriani A, Furukawa TA, Salanti G, 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. Focus (Am Psychiatr Publ). 2018;16(4):420-429.

12. Brody B, Leon AC, Kocsis JH. Antidepressant clinical trials and subject recruitment: just who are symptomatic volunteers? Am J Psychiatry. 2011;168(12):1245-1247.

13. Hieronymus F, Lisinski A, Hieronymus M, et al. Determining maximal achievable effect sizes of antidepressant therapies in placebo-controlled trials. Acta Psychiatr Scand. 2021;144(3):300-309.

14. Fried EI, Nesse RM. Depression is not a consistent syndrome: an investigation of unique symptom patterns in the STAR*D study. J Affect Disord. 2015;172:96-102.

15. Munkholm K, Paludan-Muller 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(6):e024886.

16. Cuijpers P, Turner EH, Mohr DC, et al. Comparison of psychotherapies for adult depression to pill placebo control groups: a meta-analysis. Psychol Med. 2014;44(4):685-695.

17. Hieronymus F, Lisinski A, Nilsson S, Eriksson E. Efficacy of selective serotonin reuptake inhibitors in the absence of side effects: a mega-analysis of citalopram and paroxetine in adult depression. Mol Psychiatry. 2018;23(8):1731-1736.

18. Kirsch I. The emperor's new drugs: medication and placebo in the treatment of depression. Handb Exp Pharmacol. 2014;225:291-303.

19. Jauhar S, Hayes J. The war on antidepressants: what we can, and can't conclude, from the systematic review of antidepressant withdrawal effects by Davies and Read. Addict Behav. 2019;97:122-125.

20. Davies J, Read J. A systematic review into the incidence, severity and duration of antidepressant withdrawal effects: are guidelines evidence-based? Addict Behav. 2019;97:111-121.

21. Read J, Cartwright C, Gibson K. Adverse emotional and interpersonal effects reported by 1829 New Zealanders while taking antidepressants. Psychiatry Res. 2014;216(1):67-73.

22. Lisinski A, Hieronymus F, Naslund J, et al. Item-based analysis of the effects of duloxetine in depression: a patient-level post hoc study. Neuropsychopharmacology. 2020;45(3):553-560.

23. Hieronymus F, Emilsson JF, Nilsson S, Eriksson E. Consistent superiority of selective serotonin reuptake inhibitors over placebo in reducing depressed mood in patients with major depression. Mol Psychiatry. 2016;21(4):523-530.

24. Dawson G, Pies RW. Antidepressants do not work by numbing emotions. Psychiatric Times. September 26, 2022. https://www.psychiatrictimes.com/view/antidepressants-do-not-work-by-numbing-emotions

25. Hieronymus F, Lisinski A, Ostergaard SD, Eriksson E. The response pattern to SSRIs as assessed by the Montgomery-Asberg Depression Rating Scale: a patient-level meta-analysis. World Psychiatry. 2022;21(3):472-473.