Skip navigation
Link to Back issues listing | Back Issue Listing with content Index | Subject Index

Discontinuation rates with newer antidepressants

Systematic review
Results
Comment
Persistence etc

Insufficient attention is given to discontinuation rates in clinical trials. Yet it is a major, and important, outcome. If people cannot take the tablets because they have insufficient clinical effect, or because they suffer unacceptable adverse effects, this limits utility of a medicine or intervention.

Greater acceptability of medicines or interventions is likely to be a major determinant both of therapeutic strategy and overall cost. Moreover, when healthcare systems have limited capacity, use of therapies with lower acceptability places often unacceptable additional strains on already overburdened staff.

Discontinuation rates are sometimes collected in meta-analyses, but meta-analyses of discontinuation rates themselves are uncommon. So one examining discontinuation in a sometimes difficult area, like major depression, is welcome both for the insight it gives us on differences between treatments, and because it makes us think about what evidence is useful for, and the methods we use to obtain it.

Systematic review

The review [1] used five electronic databases to find studies, plus manual searching. Only English language studies were searched for, to early 2004. Studies had to be randomised and double blind, of at least six weeks duration, and with at least 40 patients. Outcomes used were overall discontinuation rates, discontinuation because of lack of efficacy, and discontinuation because of adverse effect.

Only those directly comparing a selective serotonin reuptake inhibitor (SSRI) with another second-generation antidepressant (venlafaxine, mirtazapine, buproprion) were used. Each of these three was considered to act through different pharmacological mechanisms, and comparison was with each separately. Studies with only a placebo comparator were not used.

Results

Twenty studies were found and used, with a few small studies in different populations or with methodological problems found but not used. Most studies used were in people younger than 60 years, used standard diagnostic criteria of major depression, and most had moderate to severe depression.

For seven of the 10 comparisons with venlafaxine (75-225 mg), fluoxetine (20-60 mg) was the SSRI used. For mirtazapine (15-45 mg), paroxetine (20-40 mg) was the SSRI most often used, in three of the four comparisons. For buproprion (150-450 mg, most often sustained release formulation), sertraline (50-200 mg) was the SSRI most often used, in three of six comparisons.

The main results are shown in Table 1. There was no statistical difference in discontinuation overall, or for particular cause, between SSRIs and venlafaxine, mirtazapine, or buproprion.



Table 1: Discontinuation rates in head-to-head comparisons of SSRIs with other second-generation antidepressants



Number of
Discontinuation rates (%)
Comparison
Trials
Patients
Overall
Adverse events
Lack of efficacy
SSRI
10
1154
24
9
5.3
Venlafaxine
1160
25
11
3.6
SSRI
4
596
31
13
2.2
Mirtazapine
608
29
14
2
SSRI
6
631
17
6.7
4.1
Bupropion
623
14
6.7
3.1


Comment

The finding was that there were no differences in discontinuation rates overall, or because of either lack of efficacy or adverse effects. The use of only studies where there were sufficient numbers of patients in direct comparisons between SSRIs and other second-generation antidepressants means that we can trust these results, especially because only randomised and double blind trials were accepted.

This is unusual, and often only placebo-controlled studies are available, so that only indirect comparisons are possible. There is no reason to think that there would be a major difference in using indirect comparisons (Bandolier 110), and it would have been instructive to know in this case how well direct and indirect comparisons would have fared.

The authors of the paper provide the bottom line. It says that clinicians can focus on other practical or clinical issues in making treatment choices, given similar efficacy for these drugs. Those might be cost, differences in particular adverse events, or onset of action.

Persistence etc

This is a topic Bandolier has examined on a number of occasions. Bandolier online has a new Pharmacy section, where other studies of compliance, concordance, persistence or adherence have been pulled together. Good studies on improving compliance are few, as a systematic review in Bandolier 127 pointed out. Bandolier would be grateful if pharmacists seeing good studies would let us know of them.

Reference:

  1. G Gartlehner et al. Discontinuation rates for selective serotonin reuptake inhibitors and other second-generation antidepressants in outpatients with major depressive disorder: a systematic review and meta-analysis. International Clinical Psychopharmacology 2005 20: 59-69.

previous or next story