Executive function predicts antidepressant treatment noncompletion in late-life depression

Objective: To examine whether executive function (EF) is associated with nonremission and noncompletion of antidepressant pharmacotherapy in older adults with depression. Design: In this prospective study (July 2009 to May 2014), older adults (aged ≥ 60 years; n = 468) with a DSM-IV-defined major depressive episode diagnosed via structured interview received 12 weeks of venlafaxine extended release with the goal of achieving remission. A hypothesis was made that worse baseline EF would predict both nonremission and noncompletion (primary outcomes). Treatment-related factors, including side effects and nonadherence, were also studied. Methods: Baseline EF, including response inhibition and set-shifting, was assessed with subtests of the Delis-Kaplan Executive Function System and the semantic fluency subtest of the Repeatable Battery for the Assessment of Neuropsychological Status (RBANS). Attention, immediate memory, delayed memory, visuospatial ability, and global cognition were also assessed with the RBANS. Results: Of 468 participants, 96 (21%) failed to complete the treatment trial, 191 (41%) completed and remitted, and 181 (39%) completed and did not remit. Univariate analyses indicated that some EFs (setshifting and semantic fluency) and other cognitive variables (attention, immediate memory, visuospatial ability, and global cognition) predicted treatment noncompletion, whereas no cognitive variables predicted nonremission. In a multivariate logistic regression model, semantic fluency (P = .003), comorbid medical burden (P < .001), and early nonadherence (P < .001) were significant predictors of treatment noncompletion. Conclusions: Poorer EF predicted treatment noncompletion. These findings suggest that EFs of initiation and set maintenance (examined by the semantic fluency task) may allow depressed elderly individuals to engage and stay in treatment. Identification of those at risk for noncompletion may help implementation strategies for personalized care.