The impact of COVID‐19 lockdown on cognitive decline over time

Abstract

AbstractBackgroundDuring the COVID‐19 pandemic, memory clinic patients reported worries for faster cognitive decline as a consequence of lockdown measures, e.g. loss of structure and feelings of loneliness and depression. We aimed to investigate the impact of the COVID‐19 lockdown on cognitive decline over time in a mixed memory clinic population, compared to historical controls.MethodWe included patients who visited Alzheimer Center Amsterdam six months to a week before the first COVID‐19 lockdown in the Netherlands (mid‐March 2020), and a second visit one year later, after this lockdown period (n = 114; 67±7 years old; 30% female; Mini‐Mental State Examination (MMSE): 25±3; diagnosis: n = 55 dementia, n = 32 mild cognitive impairment, n = 18 subjective cognitive decline, n = 9 postponed diagnosis). We selected historical controls with a visit in 2016 or 2017 and a second visit one year later (n = 641). Historical controls were matched 1:1 to patients during COVID‐19 lockdown by optimal Mahalanobis distance matching based on age, sex, diagnosis, MMSE, type of visit (baseline/follow‐up) and time between the first and second visit. Matching was successful and led to balanced groups (both n = 114) on all matching variables (table 1). We used linear mixed models with terms for time, group (i.e. lockdown patients vs. historical controls) and the interaction between time and group. Outcome measures were MMSE, Trail Making Test part A and B, Rey‐Auditory Verbal Learning Test (RAVLT) immediate and delayed recall and category fluency.ResultAs expected, there was no effect of group on baseline test performance for any of the tests. We found significant interactions between time and group for RAVLT immediate recall [B(SE) = ‐2.74(1.07), p = 0.011] and delayed recall [B(SE) = ‐1.07(0.34), p = 0.002], indicating steeper cognitive decline during lockdown. In non‐matched historical controls, we found similar results; significant interactions between group and time for RAVLT immediate recall [B(SE) = ‐3.34(0.80), p<0.001] and delayed recall [B(SE) = ‐0.94(0.23), p<0.001], and additionally steeper decline on category fluency [B(SE) = ‐1.42(0.51), p = 0.005].ConclusionMemory clinic patients during COVID‐19 lockdown show steeper decline in memory than matched historical controls, providing evidence that lockdown regulations contributed to faster cognitive decline.