Comparative Efficacy and Acceptability of 12 New-Generation Antidepressants

The featured paper is ‘Comparative efficacy and acceptability of 12 new-generation antidepressants: a multiple-treatments meta-analysis’ by Cipriani and colleagues. This is a bold paper which throws down the gauntlet and which has potentially significant clinical implications. In summary, this is a meta-analysis of 12 new-generation antidepressants which have been directly compared with each other in randomised controlled trials using efficacy and acceptability as end-points. The authors conclude that Reboxetine is the least effective and least acceptable and that

‘Reboxetine should not be used as a routine first-line acute treatment for major depression’

They also conclude that

‘Sertraline might be the best choice when starting treatment for moderate to severe major depression in adults because it has the most favourable balance between benefits, acceptability and acquisition cost’

In order to arrive at these conclusions the authors have firstly identified relevant papers. They wanted to find randomised controlled trials which compared one antidepressant against another and consulted the Cochrane collaboration depression, anxiety and neurosis review group controlled trials registers up to the end of November 2007. The authors also approached the regulatory authorities as well as the pharmaceutical companies for other study data. Apart from a comment on how many papers were identified from the pharmaceutical companies websites, it wasn’t clear how successful the authors were in using this approach. One obvious difficulty is that there are a number of negative findings that wouldn’t have been published and may not have been on the controlled trials registers and in the conclusions the authors comment on publication bias.

Once the papers were identified, the authors sorted through them and excluded a number of papers. There were a number of interesting points here. For instance, most of the trials took place in Europe and North America and it is not clear why other global regions are not represented so highly. There might be a number of explanations including only english language articles being sought (although this was not clear) or this might have closely matched the profile of drug trial activity. What was also interesting was that 68 of the 345 initial articles were duplicates! The reasons for this are unclear. The authors were unable to extract ‘any data’ from 18 of the articles! They also identified 15 unpublished studies from pharmaceutical websites.

A clear difficulty with the paper is that it includes 8-week trials. This is not a very long time. Indeed it is just enough to cover the often-quoted 6-week period in which antidepressant effect becomes apparent. This does cause obvious difficulties in interpretation particularly as this doesn’t give sufficient time for an adequate titration of dosage if necessary. The authors also made a number of assumptions before performing the analysis. Firstly it was assumed that all of the participants could be grouped together for the analysis (whether directly or indirectly). The problem with this is that drug trials often have a highly selected population which excludes complex concurrent pathology which might confound the results but this will vary between trials. The selection criteria may preclude the results from being generalised to populations seen in routine clinical practice in primary or secondary care. Thus comparing groups with different selection criteria may cause problems in the interpretation. The authors have also compared dosages of antidepressants. For antipsychotics, there are the well recognised chlorpromazine or percentage equivalents but here the authors use lower and upper dosing quartiles. The difficulties with this become apparent when considering the receptor-binding profiles of different antidepressants which are often complex and where a clear response between dosage and binding is influenced by pharmacokinetics. However, a strategy with these objections is probably better than no strategy at all. A number of responses on different measures of depression were grouped together to produce a standard response rate and again this approach can be questioned although the scales will have been validated and usually compared with others in common usage. Acceptability was measured using drop out rates although the authors point out that this was used for convenience and they hadn’t looked at issues such as side-effects because of the impracticality of the analysis in these circumstances.

The authors then constructed a neat diagram showing the inter-relationships between the antidepressants in the trials. They used a Markov analysis which allowed them to compare the antidepressants both directly and indirectly and determine probabilities that each would have different rankings in acceptability and efficacy. The authors had assumed that the constructed network was coherent and tested this by looking at the agreement between the direct and indirect evidence. Direct evidence for instance would be comparing citalopram and escitalopram in RCT’s which directly compared them whereas indirect evidence would involve a third antidepressant which is also compared against both. Again there is a neat diagram which compares the efficacy and acceptability of the antidepressants against each other using a rectangular grid which makes a visual inspection of the data straightforward.

The authors found that mirtazapine, escitalopram, venlafaxine and sertraline were more efficacious than a number of the other antidepressants and that escitalopram, sertraline, citalopram and bupropion were better tolerated than a number of other antidepressants. The authors draw conclusions including those outlined above although cautioning that an economic analysis is also needed as costing implications are complex. While the study wasn’t funded, a number of the authors declare conflicts of interests including funding by GSK. However good faith is assumed and it is not clear in any case how GSK would benefit from the conclusions reached here.

This is an interesting study which produces very specific recommendations. However the duration of the studies as well as the possible sampling methods and publication bias should be considered when interpreting the results.

Steps to Treatment 2 (Further analysis of selected antidepressants using additional criteria, incorporation into policy if appropriate)


Cipriani A et al. Comparative efficacy and acceptability of 12 new-generation antidepressants: a multiple-treatments meta-analysis. Lancet. 2009. 373. 746-58.

Steps To Treatment (STT)

STT = Steps To Treatment. An estimate of the number of steps between the results and translation into practice i.e. treatment. This is an opinion.


If you have any comments, you can leave them below or alternatively e-mail


The comments made here represent the opinions of the author and do not represent the profession or any body/organisation. The comments made here are not meant as a source of medical advice and those seeking medical advice are advised to consult with their own doctor. The author is not responsible for the contents of any external sites that are linked to in this blog.


  1. What I don’t understand is the author’s recommendation that sertraline be used instead of placebo as the new benchmark for new anti-depressant drugs. Escitalopram was clearly the superior drug overall ranking 2nd for efficacy and 1st for acceptability while sertraline ranked 4th for efficacy and 2nd for acceptability. I know the cost factor was mentioned but to my undestanding the purpose of testing against placebo is to highlight efficacy/tolerability benefits and not cost. Therefore, why choose the 2nd best overall dug and not the best overall drug as the new ‘benchmark’?


    • Dear Michael,
      Thank you for your comment. I agree – your argument is flawless. Looking at the conclusions all I could find was a statement that sertraline ‘could be used as a standard comparator’ and in the last sentence ‘the need of new treatments to show either greater efficacy or acceptability than an existing standard therapy’. The first statement gives some implicit scope for considering the alternative antidepressants in the analysis although as you say it is curious that they have opted for Sertraline as a comparator in Phase III trials. The only other point is that they also refer to effectiveness trials.
      The difficulty I have with such conclusions is that the trials are reasonably short – it would be interesting to see the results of longer trials particularly if they are suggesting use as comparators on the basis of their analysis. The authors also note that a number of other factors have been omitted from their analysis and the ranking of the antidepressants on the basis of 2 outcomes is perhaps a little artificial. However at least it has started a debate which will hopefully contribute to the decision-making process for prescribing!



  2. I found very informative. The article is professionally written and I feel like the author knows the subject very well. keep it that way.


  3. Hi there, just became alert to your blog through Google, and found that it is really informative. I am gonna watch out for brussels. I’ll appreciate if you continue this in future. Lots of people will be benefited from your writing. Cheers!


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s