ReviewEfficacy of treatment in older depressed patients: A systematic review and meta-analysis of double-blind randomized controlled trials with antidepressants
Introduction
Many randomized controlled trials (RCTs) in depressed older patients have been published and the vast majority of these reported significantly greater efficacy of antidepressants than placebo (Taylor and Doraiswamy, 2004). Combining the results of these individual studies in a systematic review and meta-analysis has many advantages. Summarizing the results of individual studies can provide more precise estimates of the effect of treatment and overcomes the difficulty that many studies are underpowered. Moreover, subgroup analysis may answer important questions such as possible differences in efficacy between antidepressant classes in older patients. In adult patients, subgroup analysis has found a greater efficacy of tricyclic antidepressants (TCAs) compared with selective serotonin reuptake inhibitors (SSRIs) among inpatients but no difference in efficacy between more severely and less severely depressed patients (Anderson, 2000). The question whether TCAs are more effective than SSRIs in psychiatric inpatients or in severely depressed patients has not been addressed in systematic reviews among older patients.
In one of the first reviews of the efficacy of antidepressants in depression in the older patients, published in 1988, only 12 double-blind studies were found (Gerson et al., 1988). The conclusions were that antidepressants are clearly superior to placebo, and that older patients are just as likely as younger patients to go into remission given appropriate treatment. All reviews of placebo-controlled studies in older depressed patients published since then, also concluded that antidepressants are more effective than placebo (McCusker et al., 1998, Mittmann et al., 1997, Nelson et al., 2008, Taylor and Doraiswamy, 2004, Williams et al., 2000a, Wilson et al., 2001). These reviews also suggested that efficacy of antidepressants in older patients is comparable with efficacy in adult patients. Most of the reviews did not find a difference between classes of antidepressants in older patients, including a more recent review specifically aimed at comparing single- versus dual-action antidepressants (Mukai and Tampi, 2009). One review found a smaller Numbers-Needed-to-Treat (NNT) in the placebo controlled RCTs with TCAs (NNT 3.97, 95% confidence interval (CI) = 3.88–4.05) than with SSRIs (8.45; 95% CI = 8.38–8.53) (Wilson et al., 2001). However, in studies directly comparing TCAs with SSRIs, no differences in efficacy are found.
Reviews may also have disadvantages. The number of included RCTs in the reviews mentioned above ranges from 6 to 32, suggesting that there are not many studies in the older patients. However, a closer look at the inclusion and exclusion criteria of all these reviews reveals that only a minority of published studies was included in these reviews. Although this may increase the internal validity of the review, this usually limits the external validity (generalizability) of the results. Another example is the need to have comparable, quantitative data. Both Cochrane reviews in non-demented elderly excluded many studies because they did not publish extractable data on number of patients with remission or response, necessary for a meta-analysis (Mottram et al., 2006, Wilson et al., 2001). However, this exclusion criterion is somewhat arbitrary and moreover, these studies may contain important additional information concerning the differential efficacy of antidepressants. Finally, some reviews included studies that were not double-blind or did not have a blinded outcome assessor, which are both important qualitative factors to prevent bias.
Our aim was to provide a systematic review of all acute phase double-blind RCTs of antidepressants in older depressed patients without dementia. We included studies using outcome criteria that differ from the primary or secondary outcome criteria, in order to compare these results with studies that could be pooled into the meta-analysis.
Our questions were: (1) do the various classes of antidepressants (TCAs, SSRIs and other antidepressants) have different efficacy rates when compared to placebo; (2) are there differences in efficacy between the various classes of antidepressants when compared with each other; (3) are there differences between the various classes among subgroups of older depressed patients (e.g. in inpatients, severely depressed patients).
In addition to proving again that the various classes of antidepressants (TCAs, SSRIs as well as other antidepressants) are effective in non-demented older depressed patients and that efficacy would not be different between these classes when compared to each other, we hypothesized a priori that TCAs would be more effective than SSRIs among psychiatric inpatients and possibly also among the more severely depressed patients.
Section snippets
Methods
This systematic review aimed to include all published double-blind RCTs with antidepressants in the acute phase treatment of unipolar depression in patients with a minimum age of at least 55 years, or described as elderly, senile, geriatric or older adults. Because there are important differences between countries in registering drugs as antidepressants, we have used the Anatomical Therapeutic Chemical (ATC) classification system of the World Health Organization as an internationally accepted
Results
A total of 155 potentially relevant, double-blind RCTs were identified and screened for retrieval. The flow chart of selection of RCTs is presented in Fig. 1.
Twenty-one RCTs were excluded because they did not use antidepressants but acetyl-l-carnitine (Bella et al., 1990, Gecele et al., 1991), ACTH 4–9 analog (Frederiksen et al., 1985), adinazolam (Feighner et al., 1990), brofaromine (Moller and Volz, 1993), diclofensine (Gentili et al., 1984, Jansen et al., 1982), flupenthixol (Hostmaelingen
Discussion
Our systematic review included 92 double-blind RCTs in depressed older patients, of which 51 were pooled into one of the meta-analyses. Other reviews limited to double-blind RCTs have included only 10–18 RCTs (Katona and Livingston, 2002, Mittmann et al., 1997, Mukai and Tampi, 2009, Nelson et al., 2008, Sneed et al., 2008). The median number of RCTs included in reviews using lesser strict criteria is only 18 (McCusker et al., 1998, Mottram et al., 2006, Pinquart et al., 2006, Rajji et al., 2008
Limitations
An important limitation is that we could only find 4 controlled trials that have not been published, although it was not clear which of these studies were randomized and double-blind (Cassano et al., 1998, Giakis et al., 1993, Elly Lilly, 1993, Tourigny-Rivard et al., 1996). Two additional randomized studies were published with only an abstract available, and it is not stated if these studies were double-blind (Dong-Ming et al., 2006, Mingjun, 2007). If the results of the 2 studies of which the
Summary
This extensive systematic review confirmed the efficacy of antidepressants. The results of the meta-analyses were in line with the majority of RCTs that did not produce extractable results to enter the meta-analysis. All classes of antidepressants were equally effective, also in more severely depressed older patients. As other reviews suggested that patients receiving SSRIs are less likely to be withdrawn due to side effects compared with TCAs, SSRIs may be the best treatment option in older
Conflict of interest
Rob M. Kok has received a research grants from Wyeth and Lundbeck and has received speaker's honoraria from GlaxoSmithKline, Lundbeck, Pfizer and Wyeth.
Thea J. Heeren has received speaker's honoraria from Eli Lilly and Lundbeck.
Willem A. Nolen has received research grants from Astra Zeneca, GlaxoSmithKline and Wyeth; has served as consultant for Astra Zeneca, Eli Lilly, GlaxoSmithKline, Johnson & Johnson and Pfizer and has received speaker's honoraria from Astra Zeneca, Eli Lilly,
Role of funding source
This study was not sponsored by any pharmaceutical company or other organization with a potential conflict of interests.
Acknowledgments
The authors would like to thank Carlijn van Baarsen from the Dutch Knowledge Centre of Psychiatry in the Elderly for assisting with data input and analysis.
References (168)
Selective serotonin reuptake inhibitors versus tricyclic antidepressants: a meta-analysis of efficacy and tolerability
Journal of Affective Disorders
(2000)- et al.
Escitalopram in the treatment of depressed patients aged 60 years or older
The American Journal of Geriatric Psychiatry
(2008) - et al.
Fluoxetine and amitriptyline in elderly depressed patients. A 10-week, double-blind study on course of neurocognitive adverse events and depressive symptoms
Archives of Gerontology and Geriatrics. Supplement
(1998) - et al.
Randomized, double-blind, placebo-controlled trial of fluoxetine treatment for elderly patients with dysthymic disorder
The American Journal of Geriatric Psychiatry
(2005) - et al.
Comparative efficacy of sertraline vs. fluoxetine in patients age 70 or over with major depression
The American Journal of Geriatric Psychiatry
(1999) - et al.
Comparative efficacy and safety of MAOIs versus TCAs in treating depression in the elderly
Biological Psychiatry
(1986) Tolerance and side-effects of paroxetine in elderly depressed patients
Archives of Gerontology and Geriatrics
(1997)- et al.
Escitalopram prevents relapse in older patients with major depressive disorder
The American Journal of Geriatric Psychiatry
(2007) - et al.
Escitalopram in the treatment of depressed elderly patients
The American Journal of Geriatric Psychiatry
(2005) - et al.
How well do antidepressants work in older people? A systematic review of number needed to treat
Journal of Affective Disorders
(2002)