Journal reference: Eyding D, Lelgemann M, Grouven U, et al. Reboxetine for acute treatment of major depression: systematic review and meta-analysis of published and unpublished placebo and selective serotonin reuptake inhibitor controlled trials. BMJ 2010;341:c4737 
|Reboxetine (Edronax) is no more effective than placebo for major depressive disorder.
Reboxetine should be avoided as a choice of antidepressant.
to assess the benefits and harms of reboxetine versus placebo or selective serotonin reuptake inhibitors (SSRIs) in the acute treatment of depression, and to measure the impact of publication bias in trials of reboxetine
Trial eligibility criteria:
- double blind, randomised control design
- adult patients with major depressive disorder as primary diagnosis (either ICD, DSM or Research Diagnostic Criteria)
- acute treatment: at least 6 weeks duration
- comparison of reboxetine with placebo or any antidepressant
- language: English, German or French (or any other language if English title/abstract seemed relevant)
- availability of full text document
- searched various databases including Medline, Embase, PsycINFO, BIOSIS and the Cochrane Library
- reference lists were scrutinised to identify further trials
- clinical trial registries and trial results databases were screened
- websites of the European Medicines Agency and US Food and Drug Administration (FDA) were screened
- manufacturer of reboxetine (Pfizer) asked to supply unpublished trials and unpublished data from published trials.
- Benefit outcomes:
- Hamilton depression rating scale (HDRS), response and remission rates.
- response: reduction in HDRS of ≥ 50% from baseline to end of study
- remission: HDRS ≤ 10 at the end of study (except in one trial where the threshold was ≤ 8 )
- Harm outcomes:
- rates of patients with at least one adverse event
- rates of withdrawals owing to adverse events
- rates of serious adverse events
- data extraction and assessment of bias was always conducted by one person and checked by another
- risk of bias at the study level and outcome level were formally assessed
- data was pooled by meta-analysis
- heterogeneity of studies was measured
- Studies (after exhaustive search and collection) that met the inclusion criteria = 13. All were included in the qualitative synthesis and 12 were included in the quantitative synthesis.
- 3 were placebo controlled
- 5 were active controlled
- 5 had both placebo and active controlled arms
- In four fluoxetine controlled trials and one citalopram controlled trial, the SSRIs were potentially underdosed compared to reboxetine
- No statistical significant difference between reboxetine and placebo for remission rate (OR 1.17, 95% CI 0.91-1.51; P=0.216)
- There was substantial heterogeneity (I2=67.3%, p=0.003) in the meta-analysis for response rate so no composite estimate was calculated
- there was a very substantial difference in effect between inpatient and outpatient responses
- one impatient trial was a clear statistical outlier with an OR of 11.43 (95% CI 3.10-42.12)
- No statistical significant difference between reboxetine and placebo for response rate in outpatient setting (OR 1.05, 95% CI 0.73-1.50; P=0.796)
- Similarly, no statistical significant difference was seen between reboxetine and placebo for response rate for both inpatient and outpatient settings when the aforementioned outlier was excluded.
- Reboxetine was inferior to SSRIs for remission rates (OR 0.80, 95% CI 0.67-0.96; P=0.015)
- Reboxetine was inferior to SSRIs for response rates (OR 0.80, 95% CI 0.67-0.95; P=0.010)
- reboxetine vs placebo:
- at least 1 adverse effect: OR 2.14, 95% CI 1.59-2.88; P<0.001
- withdrawal due to adverse effect: OR 2.21, 95% CI 1.45-3.37; P<0.001
- reboxetine vs SSRI:
- no significant difference for at least 1 adverse effect
- however, inferior to fluoxetine for withdrawal from study due to adverse effect: OR 1.79, 95% CI 1.06-3.05; P=0.031
- Substantial proportion of patient data (74%) had not previously been published
- Comparing the published dataset to the full dataset demonstrated that the published data overestimated the beneficial effect of reboxetine compared to placebo by 99-115% and of reboxetine compared to SSRIs by 19-23%.
Reboxetine is, overall, an ineffective and potentially harmful antidepressant. Published evidence is affected by publication bias, underlining the urgent need for mandatory publication of trial data.
- A total of 4098 patients analysed
- 2256 in the reboxetine vs placebo comparisons
- 2641 in the reboxetine vs SSRI comparisons
- Baseline demographic data given for the included trials but not in composite form
- mean age of most studies around 40 years with SD of 12 years
- more women than men
- Access only to aggregate data; a patient level meta-analysis would provide more information especially when considering the degree of heterogeneity found.
- Acute treatment trials for major depression only.
- No analysis possible on subtypes of depression or use of reboxetine for other clinical indications.
- Inclusion of a large quantity of previously unpublished data with extensive and rigorous effort in identifying the unpublished data.
- Most studies included were at low risk of bias.
- Relative under-dosing of the SSRI control arm of some of the studies suggests that the demonstrated inferiority of reboxetine to SSRI for depression remission and response rate is likely to be on the conservative end of the spectrum.
Biases and conflicts of interests
- no support from any company for this article
- apart from declared potential conflicts of interests, none seem obvious
- Eyding D (primary author) was employed by H Lundbeck A/S, Copenhagen, between January 2006 and April 2007;
- Härter M (fourth author) received remuneration from Boehringer Ingelheim and Lilly Pharma for three talks on depression guidelines in 2008;
- and Grouven U (third author), Kromp M (fifth author), Kaiser T (sixth author), Kerekes MF (seventh author), and Wieseler B (ninth author) are employees of IQWiG (The German Institute for Quality and Efficiency in Health Care; Institut für Qualität und Wirtschaftlichkeit im Gesundheitswesen)
Clinical relevance to primary health care
Publication bias of clinical trials involving antidepressants have been demonstrated to be problematic. Even the clinical significance and effectiveness of SSRIs have been called into question when unpublished data is included in meta-analysis. Kirsch and colleagues demonstrated substantial publication bias for SSRIs in their meta-analysis that included unpublished data obtained from the US FDA under the Freedom of Information Act; clinical benefit from SSRIs seemed to be limited to those with the most severe depression . This was recently confirmed in a patient-level meta-analysis by Fournier and colleagues; there is little if any benefit of SSRIs in patients with major depression with mild to moderate symptoms .
Reboxetine (marketed as Edronax) is an selective noradrenaline reuptake inhibitor, a newer and less commonly used antidepressant in Australia. This study convincingly demonstrates that there has been marked publication bias on reboxetine that over-estimates its effectiveness. It appears that for the acute treatment of major depression:
- reboxetine is ineffective; no more effective than placebo for remission and response rate
- reboxetine is inferior to SSRIs for remission and response rate
- reboxetine is inferior to fluoxetine for medication discontinuation due to adverse effects
Arguably, reboxetine should be avoided for the treatment of unipolar major depressive disorder in the primary health care setting. Consideration should be given to transfer patients suffering severe major depressive disorder with poor response from reboxetine onto more efficacious therapy.
- Eyding D, Lelgemann M, Grouven U, et al. Reboxetine for acute treatment of major depression: systematic review and meta-analysis of published and unpublished placebo and selective serotonin reuptake inhibitor controlled trials. BMJ 2010;341:c4737
- Kirsch I, Deacon BJ, Huendo-Medina TB, et al. Initial severity and antidepressant benefits: A meta-analysis of the data submitted to the Food and Drug Administration. PLoS Med. 2008;5(2): e45
- Fournier JC, DeRubeis RJ, Hollon SD, et al. Antidepressant drug effects and depression severity: a patient-level meta-analysis. JAMA. 2010;303(1):47-53