Depression is very common among institutionalized elders. Because of the increased risk of cognitive impairment/dementia, and mortality we want to describe the evolution of depression and analyze predictive factors.

In the Aging Trajectories Study (Instituto Superior Miguel Torga - Coimbra), we followed up a sample of 83 nondemented persons (M±SD baseline age = 79.51±6.58; men: 17; women: 66). In a 2-year prospective cohort analysis (2010-2011, and 2013), we assessed depression using the Geriatric Depressive Scale/GDS as screening tool and the Mini International Neuropsychiatric Interview to diagnose depression. We also used the UCLA Loneliness Scale, the Geriatric Anxiety Inventory/GAI, the Positive And Negative Affect Scale/PANAS. Sociodemographics, and health were control variables. We performed a multinomial logistic regression to identify predicitive factors.

Fifty participants had depression at baseline, nine developed, 49 maintained, nine remitted, and 16 maintained without depression.

Having depression was associated with worse scores in UCLA, GAI, and PANAS. Not having depression was correlated with higher positive affect.

Baseline higher GAI and UCLA, and lower positive affect and satisfaction predicted recurrent depression.

Improvement in GDS, GAI, and positive affect predicted depression remission.

Results show that depression is a concern issue for professionals working with institutionalized elderly. Anxiety, loneliness, low positive affect and satisfaction constitute a risk factor for maintaing depression in institutionalized elderly and low anxiety and depressive symptoms are a protective factors for depression. These results could be used in depression prevention programs.