The Effect of Treatment With Selective Serotonin Reuptake Inhibitors in Comparison to Placebo in the Progression of Dementia

A Systematic Review and Meta-analysis

Helen E. Jones; Alka Joshi; Susan Shenkin; Gillian E. Mead


Age Ageing. 2016;45(4):448-456. 

In This Article


Background: selective serotonin reuptake inhibitors (SSRIs) may affect the neurodegenerative process of dementia, enhancing cognition. This systematic review aims to determine whether SSRIs influence cognitive performance, mood and function in people with any type of dementia.

Method: randomised placebo-controlled studies of SSRIs in people with dementia, which recorded cognitive outcomes, were identified in ALOIS (ALzheimer's and cOgnitive Improvement Studies register) in April 2013 and updated in January 2015. Data were extracted on cognition, agitation, mood, activities of daily living (ADLs) and adverse events. End of treatment statistics were calculated.

Results: twelve studies met inclusion criteria (1,174 participants), of which seven studies (710 participants) provided data for meta-analysis on cognition. There was no difference in MMSE score at end of treatment; mean difference (MD) was 0.28 (95% CI −0.83 to 1.39) (six studies, 470 participants). For change in MMSE scores, there was a small improvement; MD was 0.53 (95%CI −0.07 to 1.14) (three studies, 352 participants). The remaining studies showed no improvement in cognition. There was no statistically significant benefit of SSRIs on mood (four studies, 317 participants); standard mean difference (SMD) −0.10 (95% CI −0.39 to 0.2), agitation (three studies, 189 participants); SMD −0.01(95% CI −0.86 to 0.83), or ADLs at end of treatment (four studies, 336 participants); SMD −0.15(95% CI −0.45 to 0.15). There was no difference in mortality between the two groups. Study quality was mixed with concerns over incomplete data.

Conclusion: a small number of relatively low-powered studies showed no benefit or harm from SSRIs in terms of cognition, mood, agitation or ADLs. Large, methodologically robust studies are needed.