Effectiveness of antidepressants
BMJ 2018; 360 doi: https://doi.org/10.1136/bmj.k1073 (Published 09 March 2018) Cite this as: BMJ 2018;360:k1073
All rapid responses
Rapid responses are electronic comments to the editor. They enable our users to debate issues raised in articles published on bmj.com. A rapid response is first posted online. If you need the URL (web address) of an individual response, simply click on the response headline and copy the URL from the browser window. A proportion of responses will, after editing, be published online and in the print journal as letters, which are indexed in PubMed. Rapid responses are not indexed in PubMed and they are not journal articles. The BMJ reserves the right to remove responses which are being wilfully misrepresented as published articles or when it is brought to our attention that a response spreads misinformation.
From March 2022, the word limit for rapid responses will be 600 words not including references and author details. We will no longer post responses that exceed this limit.
The word limit for letters selected from posted responses remains 300 words.
I endorse Dr Hopcraft's reflections on the length of time doctors keep patients on anti-depressants, and on the assumption in NICE guidelines that the average course will last many months. As I discussed in my recent Personal View in the BMJ, I see no scientific basis for this. I teach psychiatric trainees, firstly, not to prescribe lightly, and, secondly, if they do decide to prescribe they should review at no more than 4-6 weeks. If there is no clear-cut difference in the patient's state at that point the anti-depressants should be stopped. We have a clinical duty not to persist with ineffective treatments! Very commonly one encounters patients who have been kept on an SSRI anti-depressant for years, yet remain 'depressed' and seeking another modality of treatment. It is ironic that many patients find that their doctor will prescribe 2 years of anti-depressants more readily than 2 weeks sleeping tablets!
Competing interests: No competing interests
McCormack and Korownyk's interesting editorial analysis suggests that, if 10 patients with moderate to severe depression taken an antidepressant for two months, five will report being 'better' but in four the response will not be because of the drug.
Yet NICE guidance in this area suggests that a person at remission from an episode of depression should continue medication for at least six months.
The implication is that the four patients who simply 'happened' to get better on antidepressants will subsequently represent at least 24 months, in total, of unnecessary treatment, which has a significant cost (in all senses) to them and the NHS.
Should we therefore be rethinking the current mantra of long courses of antidepressants? A trial without treatment after two months of apparently succesful therapy would prevent an unnecessarily long course of medication in four patients, the risk being that the patient who genuinely has benefitted might relapse. Is this a trade-off we should be considering, or which we should at least be discussing with our patients?
Competing interests: No competing interests
One of the problems with "Evidence Based Medicine" (EBM) is that it can be misinterpreted as "Medicine Supported by the Most Evidence" (MSME). Innovation is stifled by this error, since MSME simply looks at whether what we are doing works compared with not doing it or doing something else.
Basing the medicine we practise on available evidence can permit innovation but the protagonists of MSME may not trust new interpretations and applications of old evidence.
I declare interest having used anatomical evidence (the juxtaposition of the 3rd & 4th Cranial nerve nuclei with the Edingher-Westphal nucleus) and observational evidence (noticing the typical gaze of neonates) as a basis to describe and use a simple technique, 'BabyGaze' for treating anxiety (not depression) in adults (1).
Though many colleagues welcome a simple non-pharmacological approach in anxiety, others have accused me of putting patients in danger by using a treatment based on evidence rather than that which is supported by the sort of complex and expensive studies that attract pharmaceutical funding to give validation to drugs through MSME.
While there is MSME evidence that Antidepressants are effective for some people in certain circumstances, there are other treatments such as psychological therapies that are also of benefit and that are based on evidence, even if the investment in randomised controlled trials to prove their MSME worth (rather than as a basis for understanding their effectiveness) is more sparse.
Basing Medicine, including innovation, on evidence is a good thing but we should beware of the "never mind the quality feel the width' argument of MSME when it is passed off as EBM.
1. Ashworth AJ, Dutton PV BabyGaze: A rapid neurobiological Intervention for Anxiety, Panic and Anger. Int J Psychiatry Vol 2 Issue 2 2017
Competing interests: I have published on alternatives to psychotropic medication
McCormack and Korownyk provide a very helpful analysis [1]. They point out that the average odds ratio for remission within 8 weeks was 1.56 so that if the probability of remission within 8 weeks was 0.4 on placebo, then on treatment it would be 1/(1+1/(0.4/0.6*1.56)) = 0.51, the NNT being 1/(0.51 – 0.4) = 9.
If the probability of remission in a patient with mild depression were 0.95, then by applying the same assumptions and calculation, the probability of remission on treatment would increase only to 1/(1+1/(0.95/0.05*1.56))= 0.967, the NNT being 1/(0.967-0.95) = 58. Similarly if the probability of a patient with severe depression going into remission were 0.05, the probability of doing so on treatment would be 1/(1+1/(0.05/0.95*1.56))= 0.076, the NNT being 1/(0.076-0.05) = 38. Further advice therefore is to estimate the probability of a patient going into remission without treatment and to estimate the probability of remission with treatment using the odds ratio as shown above.
This illustrates the ‘triage principle’ of treatment where patients with mild and severe conditions may benefit little from treatments. It is failure to understand this when interpreting the average results of RCTs that is the cause of over-diagnosis and over-treatment. It should be understood that it is the odds ratio that is constant for all absolute risk reductions. The relative risk reduction is NOT constant as widely asserted but only provides reasonable approximations for low baseline probabilities.
Reference
1. BMJ 2018;360:k1073
Competing interests: No competing interests
Dear Editors,
Recent evidence reveals that administered antidepressants increase suicide risks by 2-5 times. [2][3][4][5][6][9]
Cognitive behavioural therapy halves the risk of repeated suicide attempts. [12]
A recent meta-analysis, level I evidence, clearly demonstrated that SSRIs double the risk of suicide and violence in adults. [4]
Furthermore, antidepressants increase all cause mortality by 33%! [10][11]
Another meta-analysis published in the British Journal of Psychiatry has found that even patients with the most severe depression can expect to get as much benefit from cognitive behavioural therapy (CBT) as those with less severe symptoms. [7]
Even Behavioural Activation effectively decreases depressive symptoms. [8]
References
[2] http://journals.sagepub.com/doi/pdf/10.1177/0141076816666805
[3] http://www.bmj.com/content/348/bmj.g3510
[4] http://www.bmj.com/content/352/bmj.i65
[5] http://nordic.cochrane.org/sites/nordic.cochrane.org/files/public/upload...
[6] http://www.bmj.com/content/355/bmj.i6103
[7] http://bjp.rcpsych.org/content/210/3/190.long
[8] http://www.bmj.com/content/356/bmj.j914
[9] http://www.bmj.com/content/358/bmj.j3697/rr-4
[10] http://dailynews.mcmaster.ca/article/antidepressants-associated-with-sig...
[11] https://www.ncbi.nlm.nih.gov/pubmed/28903117
[12] https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5650127/
Competing interests: No competing interests
Re: Effectiveness of antidepressants
There are several major problems with research on antidepressants which are evident in the network meta-analysis by Cipriani et al.
The first is the short duration of trials usually of 8 weeks. This is at odds with clinical observation that improvement continues after 8 weeks.
Secondly there are many sources of bias especially direct or indirect bias due to pharmaceutical industry sponsorship as noted in your editorial. There are many ways for this influence to infiltrate complex studies. Declarations of conflict of interest are not adequate to offset this influence. We need to undertake analyses of trials where the authors have not had any links with industry. It would be helpful to remove all studies with pharma links (the vast majority) in the Cipriani study and confine analysis to the remainder. I would hypothesize that vortioxetine and agomelatine would not rank but fluoxetine would!
Competing interests: No competing interests