REVIEW OF: Sommerlad A et al, Int J Geriatr Psychiatry 2021;37(1):10

STUDY TYPE: Retrospective cohort study

Trazodone is often used to address insomnia and agitation in patients with dementia. Small randomized controlled trials have not found trazodone to be helpful in improving cognition or reducing the incidence of dementia. Nonetheless, recent mouse model studies have suggested trazodone might reverse aggregation of pathologic proteins found in dementia. Could this be a mechanism to protect against neurodegeneration? And could trazodone have a similar disease-modifying effect in ­humans?

To test out this theory, researchers conducted an observational cohort study using three UK mental health service registers. This was a retrospective study in which records of patients with dementia were reviewed to see if trazodone use correlated with improved dementia outcomes, as measured by Folstein Mini Mental Status Exam (MMSE) scores. 

The study period varied by site between 2006 and 2016. Individuals were included if they had dementia and were prescribed one of three antidepressants (trazodone, citalopram, or mirtazapine) for at least six weeks. All 2,199 study participants (average age of 79.3 + 3.0 years old; 65.3% female) had ICD-10 diagnostic codes for dementia. The predominant dementia subtype (78.8%) was neurodegenerative, non-vascular dementia. Average baseline MMSE score was 18.8 + 2.6. The average daily dose of trazodone was 101.8 mg/day (range 50–300 mg/day). 

Participants were followed until the last date of study medication use, date of death, or predetermined end of the study period. Participants taking both trazodone and either citalopram or mirtazapine were included in the trazodone group. When trazodone was switched to either citalopram or mirtazapine during the study, only data during trazodone use were analyzed. Individuals taking both mirtazapine and citalopram were excluded from the study, although the authors did not explain their rationale for this decision. 

The primary outcome was a change in MMSE scores over time. Disappointingly, no statistically significant cognitive benefit could be attributed to trazodone compared to citalopram or mirtazapine. After adjustment for dementia subtype, dementia severity at baseline, and severity of neuropsychiatric symptoms, these results remained nonsignificant. 

This was an observational, unblinded, nonrandomized study with baseline clinical differences between groups (including a much smaller trazodone group with significantly lower MMSE scores). As the researchers used a secondary mental health care database, we also don’t know about medication indications, concurrent medications, comorbid medical/psychiatric conditions, or the participants’ level of functioning. 

Carlat Take 

This large observational study revealed no signal that trazodone might protect against neurodegeneration. However, the study had a number of severe limitations, including opaque methodology and unequal groups at baseline, which may have influenced the results. The only way to know for sure whether trazodone is ineffective for dementia is to do a controlled clinical trial, but it’s unlikely that such a study will be conducted given these unpromising findings. However, trazodone remains useful as an off-label, adjunctive treatment for behavioral and psychological symptoms of dementia, and it remains a relatively safe medication to help with sleep.