Association of antidepressant medication use and social engagement on the level of depressive symptoms in community-dwelling, older persons with dementia

Objective: Depression is a common co-morbid condition at the time of dementia diagnosis. Current clinical evidence for pharmacological and non-pharmacological interventions for management of depression in dementia has shown mixed results. We wanted to explore the feasibility of designing a clinical trial to understand the efficacy of pharmacological therapy, non-pharmacological therapy, or the combination for depression management in dementia. Therefore, we examined how one pharmacological measure (antidepressant medication use) and two non-pharmacological measures of social engagement (social network size and; perceived social isolation) are associated individually and in combination with the level of depressive symptoms at the time of initially meeting criteria for dementia.;Methods: Participants were older, community-dwelling participants who developed dementia in a combined analytical cohort from two longitudinal studies on aging (the Rush Memory and Aging Project, and the Minority Aging Research Study). At the first visit where criteria for dementia were met, eligible participants had documentation of current medications to determine antidepressant use, and self-reported measures of social network size, perceived social isolation, and depressive symptoms. In this cross-sectional analysis, proportional odds models were constructed with total depressive symptom score from the Center for Epidemiologic Studies Depression Scale as the outcome of interest and antidepressant medication use, social network size and perceived social isolation as independent variables. Subsequent models examined the interaction of the pharmacologic and non-pharmacologic measures with depressive symptoms.;Results: Dementia occurred in 402 participants (16.3% of the total number of participants). The mean age was 86 years (SD = 7), mean education was 14 years (SD = 3) and mean depressive symptom score was 2 (SD = 2). Women accounted for 70% of the cohort, with 16% of the total cohort being African American and 25% using at least one antidepressant medication. The mean social network size was 6 persons (SD = 5) and mean perceived social isolation score was 2.6 units (SD = 0.6). The proportional odds model adjusted for demographics showed no association between antidepressant medication use and total depressive symptom score (estimate= -0.15, SE = 0.21, p = 0.5). In contrast, each additional person in the social network was associated with a lower depressive symptom score (estimate = 0.07, SE = 0.02, p < 0.001) and each unit increase in perceived social isolation was associated with a higher depressive symptom score (estimate = -1.42, SE = 0.18, p < 0.001). No interaction of antidepressant medication use with either social engagement measure was observed.;Conclusion: At the time of initial dementia diagnosis, high social network size and low perceived social isolation (but not antidepressant medication use) were associated with lower depressive symptoms. Also, the association of social engagement measures with depressive symptoms was not modified by antidepressant use. Based on these results, a clinical trial for reducing depressive symptoms at the time of initial diagnosis of dementia probably should focus on comparing interventions aimed to increase social network size and/or reduce perceived social isolation.;Keywords: Depression, Antidepressant therapy, Social Network, Social Isolation, Dementia.