Antidepressants have a good track record in old age depression. This new meta analysis is trying to provide a more precise estimate of the effect of treatment and to see whether any class of medications hold advantage over others.
The first review of this subject found only 12 double blind studies.(Gerson et al., 1988). Since then various reviews have shown that antidepressants are better than placebo but no difference in efficacy was found between different classes of antidepressants.
Kok et al from Netherlands reviewed all acute phase double-blind RCTs of antidepressants in older depressed patients (above 55 yrs!) without dementia.Trials that include only patients suffering from concomitant physical illness were excluded. Studies included major depressive disorder, dysthymia, minor depression, sub clinical or sub threshold depression or non-specified depression.The primary outcome of the meta-analyses was remission, dichotomized according to a final score lower than a cut-off on scale used.ie score 7 or 10 on Hamilton Depression Rating Scale (HAM-D dependent on the version) and score 12 on MADRS. Response was defined as a 50% decrease on the HAM-D, MADRS or other depression rating scale or as a Clinical Global Impressions-Improvement Scale (CGI-I) score of 1 or 2 (very much or much improved).
A total of 155 potentially relevant, double-blind RCTs were identified. After applying exclusion criteria, 51 of the remaining 92 RCTs were included in at least one of the meta-analyses.
Efficacy against Placebo
If all antidepressants are combined, the difference in achieving remission is statistically significantly different from placebo (OR 1.36; 95% CI 1.07– 1.73) with a NNT of 14.4 (95% CI 8.3–50). The difference in achieving a response is also statistically significant different from placebo (OR 1.78 with 95% CI 1.42–2.24) with a NNT of 6.7 (95% CI 4.8–10). Limiting these analyses to studies that only included patients with major depression resulted in non-significant changes in ORs.
Response is achieved by 48 % of patients on antidepressants compared with 38.6% in patients using a placebo. Remission is achieved by 33.7% of patients treated with antidepressants compared with 27.2% in patients using placebo.
Different classes against Placebo: TCAs were compared with placebo in 12 RCTs of which only 3 compared remission rates with no statistically significant difference with placebo. However Response rates when pooled, antidepressants were much better. ((OR (fixed) 2.63, 95% CI 1.52–4.55) and the NNT for response is 4.2 (95% CI = 2.8–7.7).
SSRIs were compared with placebo in 11 RCTs.The difference was statistically not significant (OR 1.29; 95% CI = 0.93–1.80) in remission. However, again, the response rate was better for SSRI, statistically significant (OR 1.61, 95% CI 1.18–2.20) and the NNT for response is 10 (95% CI=6.7–20).
Other antidepressants (venlafaxine,trazodone,moclebemide, and bupropion): Of the 10 RCTs that compared ‘Other antidepressant’ with placebo, only 4 compared remission rates and no significant difference was found with placebo. As above, response rate was significantly better for these Antideps ( Six studies, (OR 1.83, 95% CI 1.21–2.78) and a NNT of 3.5 (95% CI = 3.5–14.3)).
The absence of difference in remission against placebo, (against different classes ) is in part be explained by a type 2 error.(ie small number of trails , not having enough power to show the difference).
Any class better than other classes ?
16 RCTs that used the primary outcome criterion of remission, no significant difference could be demonstrated between treatment with a TCA compared neither with SSRIs nor with other antidepressants nor between SSRIs compared with other antidepressants.
In the 27 RCTs that used response criteria , no significant difference could be demonstrated between TCAs and SSRIs nor between TCAs and other antidepressants nor between SSRIs and other antidepressants
Inpatients: No differences were found between TCAs and SSRIs or other antidepressants in remission( 3 RCTs) or response (5 studies)
Severe depression:No difference could be demonstrated between TCAs and SSRIs in more severely depressed patients, neither in achieving remission (2 studies) nor in achieving response (5 studies)
Strengths: More studies included than previous reviews. Authors claim that this is the first meta-analysis using remission defined as a score below a predefined cut-off on a depression rating scale as primary outcome.
Limitations: Significant heterogeneity ( variation between studies) make pooling questionable even though random effects model was used. No assessment of quality of the included studies. The group of ‘other antidepressants’ is very heterogeneous. Though the authors say that they have repeated the most important analyses with studies limited to major depression, this was not clear in the analysis provided.
This review show that in old age depression antidepressants are effective. They do not differ in efficacy. The choice might be much based on side effects. Currently we have very limited number of studies to inform choice based on efficacy.
Efficacy of treatment in older depressed patients: A systematic review and meta-analysis of double-blind randomized controlled trials with antidepressants. Kok RM, Nolen WA, Heeren TJ.J Affect Disord. 2012 Dec 10;141(2-3):103-15.