Antidepressant Effectiveness – A Meta-Analysis Says They Work Well for a Minority

adminGPS Leave a Comment

How can we assess antidepressant effectiveness? A series of articles have come to seemingly incompatible conclusions about how well they work.

An article in the British Medical Journal published in 2022 suggests that the answer is more complicated than a simple yes or no.

This article is another attempt to examine existing data from the thousands of people who have participated in a clinical trial of an antidepressant medication in the last 40 years.

Another effort to look at this data was discussed on the antidepressant comparison page. It used the technique of Network Meta-Analysis to try to tease out differences between antidepressants.

The authors of this meta-analysis examined 232 randomized, double blind, placebo controlled trials of drug monotherapy for major depressive disorder submitted by drug developers to the FDA between 1979 and 2016. These studies included 73,388 adults and children. They analyzed the raw participant level data, unlike many previous meta-analyses.

They concluded that the model that best fit the data had three patterns of response:

  • Minimal response – This group had a very small change in depression scores (Hamilton Depression Rating Scale average change of -1.68 points). Significantly fewer individuals treated with active medication (12%) had this response compared with placebo treated individuals (21%).
  • Non-specific response – This group had a moderate change of depression scores (HAMD average change of -8.94 points). This was the most common type of response and roughly the same number of individuals receiving active medication (63%) had this response as did those treated with placebo (69%).
  • Large response – This group had an average 16 point reduction of HAMD scores (equivalent to going from moderate to severe depression to not being depressed). 24% of subjects treated with active drug had this response, compared to only 10% of those treated with placebo. A large placebo response was even less common in those who were older than age 25.
Magnitude of improvement in depression scores broken down by age, sex and severity of depression.

A large response to an antidepressant was more likely in men and was quite a bit more likely in those with more severe depression.

The authors also confirmed earlier findings that some antidepressants were more likely to have a large response.

Size of antidepressant response (drug vs placebo difference) broken down by antidepressant.

Venlafaxine, amitriptyline and clomipramine had the greatest drug vs placebo differences. Trazodone and fluvoxamine had the smallest differences.

In conclusion, this article suggests that the answer to the question, “do antidepressants work” is more complicated than a “yes” or “no” answer.

Antidepressants work better than placebo (which often works quite well) for a significant minority of individuals in clinical research trials (overall about 15% of all individuals and about 20% of those older than 25).

This group cannot be clearly identified prior to treatment, but they have a large improvement on standard measures of depression after they receive treatment.

They tend to be more depressed and older than 25 (teenagers and young adults have a larger positive response to placebo and so more of the improvement seen in this age group is “non-specific” – meaning not different in active and placebo treated individuals).

One issue that likely has contributed to the fact that a large response to an antidepressant occurs relatively infrequently is the heterogeneity of the condition we call “major depression.” In fact, major depression is one of the diagnoses in the Diagnostic and Statistical Manual (DSM5) that has the largest variability in how it presents ( a person with depression may be more or less hungry, may get more or less sleep, may feel agitated or slowed down, etcetera, etcetera). As a result, it has the lowest concordance across clinicians (two psychiatrists interviewing the same patient have a greater chance of disagreeing about whether that patient has a major depression than they have of disagreeing about the diagnosis of almost any of the common psychiatric conditions).


Stone M B, Yaseen Z S, Miller B J, Richardville K, Kalaria S N, Kirsch I 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 doi:10.1136/bmj-2021-067606

Arroll B, Elley CR, Fishman T, et al. Antidepressants versus placebo for depression in primary care. Cochrane Database Syst Rev2009;3):CD007954. doi:10.1002/14651858.CD007954 pmid:19588448

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 Psychopharmacol2002;22:40-5. doi:10.1097/00004714-200202000-00007 pmid:11799341

Undurraga J, Baldessarini RJ. Randomized, placebo-controlled trials of antidepressants for acute major depression: thirty-year meta-analytic review. Neuropsychopharmacology2012;37:851-64. doi:10.1038/npp.2011.306 pmid:22169941

Watanabe N, Omori IM, Nakagawa A, et al. Mirtazapine versus other antidepressive agents for depression. Cochrane Database Syst Rev2011; doi:10.1002/14651858.CD006528.pub2 pmid:22161405

Cipriani A, Koesters M, Furukawa TA, et al. Duloxetine versus other anti-depressive agents for depression. Cochrane Database Syst Rev2012; doi:10.1002/14651858.CD006533.pub2 pmid:23076926

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. Lancet2018;391:1357-66. doi:10.1016/S0140-6736(17)32802-7. pmid:29477251

Jakobsen JC, Katakam KK, Schou A, 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 Psychiatry2017;17:58. doi:10.1186/s12888-016-1173-2. pmid:28178949

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 Med2008;5:e45. doi:10.1371/journal.pmed.0050045. pmid:18303940

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 Psychiatry2012;69:572-9. doi:10.1001/archgenpsychiatry.2011.2044 pmid:22393205

Fournier JC, DeRubeis RJ, Hollon SD, et al. Antidepressant drug effects and depression severity: a patient-level meta-analysis. JAMA2010;303:47-53. doi:10.1001/jama.2009.1943 pmid:20051569

Moncrieff J, Kirsch I. Empirically derived criteria cast doubt on the clinical significance of antidepressant-placebo differences. Contemp Clin Trials2015;43:60-2. doi:10.1016/j.cct.2015.05.005. pmid:25979317

Hengartner MP, Plöderl M. Estimates of the minimal important difference to evaluate the clinical significance of antidepressants in the acute treatment of moderate-to-severe depression. BMJ Evid Based Med2022;27:69-73. doi:10.1136/bmjebm-2020-111600. pmid:33593736