Table 1.
Studies of subjects ≥65 years included in the meta-analysis
| Author Year | Drug & Dose | Time (weeks) | N | Mean Age (years) | Qualitya | Scale | Primary outcome | Statistical benefit |
|---|---|---|---|---|---|---|---|---|
| Evans 1997 [14]b | FLU 20 mg | 8 | 62 | 82 | Lowd | HDRS-17 | Response | No |
| Fraguas 2009 [15]bc | CIT 20-40 mg | 8 | 37 | 74 | Low | HDRS-17 | Response | No |
| Hewett 2010 [16] | BUP 105–300 mg | 10 | 418 | 71 | Lowd | MADRS | Change from baseline | Inconsistente |
| Kasper 2005 [17] |
ESC 10 mg FLU 20 mg |
8 | 517 | 75 | Moderated | MADRS | Change from baseline | No |
| Katona 2012 [18] | DUL 60 mg | 8 | 452 | 71 | Highd | HDRS-24 | Change from baseline | Yes |
| Raskin 2008 [19] b | DUL 60 mg | 8 | 311 | 73 | Lowd | HDRS-17 |
Response/ remission |
Yes |
| Robinson 2014 [20]b | DUL 60 mg | 12 + 24 extension | 370 | 73 | Low-d | HDRS-17 Maier subscale | Change from baseline | No |
| Roose 2004 [21] | CIT 10–20 mg | 8 | 174 | 80 | Highd | HDRS |
Response/ remission |
No |
| Schatzberg 2006 [22]b |
FLU 40–60 mg VEN 150–225 mg |
8 | 300 | 71 | Highd | HDRS-21 |
Response/ remission |
No |
a Quality based on Cochrane risk of bias criteria; b Enrolled subjects with dementia/did not explicitly exclude dementia subjects; c Subjects had heart failure; d Trial had industry funding/industry employees as authors; e The pre-specified ANCOVA analysis was not statistically significant (p = 0.09). Post-hoc rank-based ANCOVA analysis was statistically significant (p = 0.03). Response was statistically significant (p = 0.01), but remission was not statistically significant (p = 0.17)
FLU Fluoxetine, CIT Citalopram, BUP Buspirone, ESC Escitalopram, DUL Duloxetine, VEN Venlafaxine, HDRS Hamilton Depression Rating Scale, MADRS Montgomery-Åsberg Depression Rating Scale