Objective: The authors conducted a prospective cohort study to estimate the risk of incident mild cognitive impairment in cognitively normal elderly (aged $70 years) individuals with or without neuropsychiatric symptoms at baseline. The research was conducted in the setting of the populationbased Mayo Clinic Study of Aging. Method: Ac lassification of normal cognitive aging, mild cognitive impairment, and dementia was adjudicated by an expert consensus panelbased onpublished criteria. Hazard ratios and 95% confidence intervals were computed usingCoxproportionalhazardsmodel,withage as a time scale. Baseline Neuropsychiatric Inventory Questionnaire data were available for 1,587 cognitively normal persons who underwent at least one follow-up visit. Results: The cohort was followedto incident mild cognitive impairment (N=365) or censoring variables (N=179) for a median of 5 years. Agitation (hazard ratio=3.06, 95% CI=1.89– 4.93),apathy(hazard ratio=2.26,95%CI=1.49– 3.41), anxiety (hazard ratio=1.87, 95% CI=1.28–2.73), irritability (hazard ratio=1.84, 95% CI=1.31–2.58), and depression (hazard ratio=1.63, 95% CI=1.23–2.16), observed initially, increased risk for later mild cognitive impairment. Delusion and hallucination did not. A secondary analysis, limited in significance by the small number of study participants, showed that euphoria, disinhibition, and nighttime behaviors were significant predictors of nonamnestic mild cognitive impairment but not amnestic mild cognitive impairment. By contrast, depression predicted amnestic mild cognitive impairment (hazard ratio=1.74, 95% CI=1.22–2.47) but notnonamnesticmildcognitiveimpairment. Conclusions: An increased incidence of mild cognitive impairment was observed in community-dwelling elderly adults who had nonpsychotic psychiatric symptoms at baseline. These baseline psychiatric symptoms were of similar or greater magnitude as biomarkers (genetic and structural MRI) in increasing the risk of incident mild cognitive impairment. (Am J Psychiatry 2014; 171:572–581)
Read full abstract