The Effect of Apathy on Cognitive Decline in Late-Life Depression

Introduction Apathy – defined as a reduction in voluntary goal-directed behavior – is a persistent and debilitating symptom often found in older adults with late-life depression (LLD). The presence of apathy often correlates with worse cognitive functioning in LLD, and apathy is commonly conceptualized as a harbinger of cognitive decline in predementia states. For example, the presence of both apathy and depression together in older adults increases the likelihood of transitioning from mild cognitive impairment to dementia when compared to depression alone. Considering the emphasis on failure to initiate and maintain goal-directed behavior, apathy has often been linked to the neuropsychological construct of executive functioning. However, the longitudinal association between apathy and executive functioning and other cognitive domains in older adults with LLD remains unclear. The current analyses addressed the following two questions: 1) do patients who generally score higher on apathy also experience lower levels of cognitive functioning? and 2) when a patient exhibits an increase in apathy above their personal average, does s/he also experience an associated decrease in cognition? We hypothesized that greater apathy would be associated with worse executive functioning in a sample of LLD followed over five years. Methods Data originated from a naturalistic observation of the clinical and cognitive trajectories of LLD over time. Upon entering the study, all participants were age 60+, were without a clinical diagnosis of dementia, scored >= 25 on the Mini-Mental State Exam (MMSE), and met DSM-IV criteria for a current episode of Major Depressive Disorder as confirmed by clinical interview with a geriatric psychiatrist. Upon entry, most depressed participants were offered open label treatment with sertraline but the antidepressant regimen differed across participants based on clinical history and need. These analyses used data from 115 LLD patients who were followed upwards of five years. Annual data are used in the analyses. Measures The geriatric psychiatrist administered Montgomery Asberg Depression Rating Scale (MADRS) and Apathy Evaluation Scale (AES) were used to record depression and apathy severity, respectively. Medical co-morbidities were recorded with the Cumulative Illness Rating Scale – Geriatric Version (CIRS-G). There was a total of eight cognitive outcomes that were derived from neuropsychological tests of processing speed, executive functioning (e.g., Trail Making, Letter Fluency), memory (e.g., Logical Memory), and global cognitive functioning (i.e., the CERAD total composite score). Statistical Analyses We first used spaghetti plots in order to visualize the longitudinal MADRS and AES scores. We also computed the correlation between MADRS and AES over time. We next used two-level linear mixed-effects models with within-person centering, which allows the separation of time-varying within-person effects from time-invariant between-person effects of apathy and depression upon cognitive function. Initial unconditional mixed-effects models examined a within-person centered (WP) apathy variable and a person mean (PM) apathy variable on cognitive outcomes. The WP variable was entered as a time-varying (level-1) variable. The PM variable was centered on its mean, and entered as a time-invariant (level-2) variable in the equation for the random intercept (the PM main effect). Subsequent models controlled for demographics, medical comorbidities, and WP and PM depression severity. Results Depression severity significantly decreased over the course of the five year-study while apathy severity experienced a more modest decline. Apathy and depression severity were moderately correlated over the five-year period (r >= .5). In the unconditional model, increases in apathy between patients (PM) was associated with corresponding declines across multiple cognitive domains. After controlling for demographics (i.e., sex, age, education) and medical comorbidities, increases in apathy between patients (PM) was associated with corresponding declines in processing speed measured with Trail Making Part A. When WP/PM depression was added to the model, apathy was no longer associated with processing speed or any other cognitive outcomes. Conclusions We found that increases in apathy between patients was mirrored by corresponding declines in processing speed, even when accounting for demographics and medical co-morbidities. However, this finding was no longer statistically significant when depression was added to the model. We did not find a within-person association between apathy and cognition – presumably due to the fact that apathy changed little over the course of this study in adults treated with antidepressants. Future research may further seek to disentangle apathy, depression, and cognitive decline in LLD. This research was funded by NIMH R01MH108578