Association of Posttraumatic Stress Disorder With Accelerated Cognitive Decline in Middle-aged Women

IMPORTANCE Posttraumatic stress disorder (PTSD) has been hypothesized to lead to impaired cognitive function. However, no large-scale studies have assessed whether PTSD is prospectively associated with cognitive decline in middle-aged adults. OBJECTIVE To assess the association between PTSD and decline in cognitive function over time. DESIGN, SETTING, AND PARTICIPANTS This cohort study included participants from the Nurses' Health Study II, an ongoing longitudinal cohort study involving community-dwelling middle-aged female nurses residing in the US who had at least a 2-year nursing degree at the time of enrollment in 1989. The present study included 12 270 trauma-exposed women who were enrolled in the PTSD substudy of the Nurses' Health Study II and completed 1 to 5 cognitive assessments. Data were collected from March 1, 2008, to July 30, 2019. EXPOSURES Lifetime PTSD symptoms, assessed using a validated questionnaire between March 1. 2008, and February 28, 2010. MAIN OUTCOMES AND MEASURES The main outcome was evaluated using the Cogstate Brief Battery, a self-administered online cognitive battery. Cognitive function was measured by a psychomotor speed and attention composite score and a learning and working memory composite score. Women completed the Cogstate Brief Battery every 6 or 12 months (up to 24 months) from October 3, 2014, to July 30, 2019. Linear mixed-effects models were used to evaluate the association of PTSD symptoms with the rate of change in cognition over follow-up, considering a broad range of relevant covariates, including the presence of depression symptoms and history of clinician-diagnosed depression. The rate of cognitive change was adjusted for potential practice effects (ie, potential changes in test results that occur when a test is taken more than once) by including indicators for the number of previous tests taken. RESULTS Among 12 270 women, the mean (SD) age at the baseline cognitive assessment was 61.1 (4.6) years; 125 women (1.0%) were Asian, 75 (0.6%) were Black, 156 (1.3%) were Hispanic, 11767 (95.9%) were non-Hispanic White, and 147 (1.2%) were of other race and/or ethnicity. A higher number of PTSD symptoms was associated with worse cognitive trajectories. Compared with women with no PTSD symptoms, women with the highest symptom level (6.7 symptoms) had a significantly worse rate of change in both learning and working memory (beta = -0.08 SD/y; 95% CI, -0.11 to -0.04 SD/y; P < .001) and psychomotor speed and attention (beta = -0.05 SD/y; 95% CI, -0.09 to -0.01 SD/y; P = .02), adjusted for demographic characteristics. Associations were unchanged when additionally adjusted for behavioral factors (eg, 6-7 symptoms in the analysis of learning and working memory: beta = -0.08 SD/y; 95% CI, -0.11 to -0.04 SD/y; P < .001) and health conditions (eg, 6-7 symptoms in the analysis of learning and working memory: beta = -0.08 SD/y; 95% CI. -0.11 to -0.04 SD/y; P < .001) and were partially attenuated but still evident when further adjusted for practice effects (eg, 6-7 symptoms in the analysis of learning and working memory: beta = -0.07 SD/y; 95% CI, -0.10 to -0.03 SD/y; P < .001) and comorbid depression (eg, 6-7 symptoms in the analysis of learning and working memory: beta = -0.07 SD/y; 95% CI, -0.11 to -0.03 SD/y; P < .001). CONCLUSIONS AND RELEVANCE In this large-scale prospective cohort study, PTSD was associated with accelerated cognitive decline in middle-aged women, suggesting that earlier cognitive screening among women with PTSD may be warranted. Given that cognitive decline is strongly associated with subsequent Alzheimer disease and related dementias, better understanding of this association may be important to promote healthy aging.