This study demonstrates that stress related to the COVID-19 lockdown differentially affects the anxiety levels among CI and CU participants of the TRIAD cohort. Before the COVID-19 pandemic, anxiety levels are higher among the CI participants and are positively associated with tau deposition in the brain. However, during the COVID-19 lockdown, the anxiety levels increased in the CU and remained high in the CI participants. We further found that COVID-19 related stress is associated with higher anxiety levels in the CU group, while among the CI group, only those with lower tau load in the brain had higher anxiety levels. Our findings suggest distinct aetiologies explaining high anxiety levels observed among participants of the TRIAD cohort during the first COVID-19 lockdown. Apart from the persistent endogenous anxiety associated AD pathophysiology, CI individuals exhibited anxiety related to COVID-19 related stress.
Neuropsychiatric symptoms (NPS) are commonly reported in patients with cognitive impairment. In a population-based study, 43% of MCI and 75% of dementia participants exhibited NPS, with anxiety being a commonly reported symptom (9.9% in MCI and 21.5% in dementia)18. In a clinical cohort, anxiety was reported in 68–71% of the AD patients19. Patients with anxiety are also associated with earlier conversion to AD19. Emerging evidence suggests that anxiety may be driven by AD-related neuropathological changes. In cognitively normal older individuals, higher amyloid load in the brain is associated with increasing anxious-depressive symptoms over mean 3.8 years20. More recently, our group showed that NPS, measured using the neuropsychiatric inventory (NPI-Q) correlated with higher tau load in the parietal association area, superior frontal, temporal, and medial occipital lobes of participants from the TRIAD cohort21. Therefore, our findings of anxiety being associated with higher tau uptake in the brain provides further evidence that NPS such as anxiety represent an endogenous presentation of the AD biological process.
Since the beginning of the COVID-19 pandemic, studies have shown that the restrictions imposed by the lockdown has had a significant impact on the mental health status of the population22,23, such as an increased level of anxiety24–26. In a population study of 1498 participants in Iran, 12.4% reported anxiety higher than moderate while 8% have severe or extremely severe stress27.
On the other hand, we found that in the CI group the anxiety levels remained high, suggesting that CI participants lockdown measures did not impose additional stress in patients. Biologically, this finding may be a result of the impairment of cognitive processes such as inattention, lack of insight, executive dysfunction, impaired social cognition and memory impairment that the participants are already experiencing prior to the pandemic. This could make them less able to be affected by the increased anxiety and stress from the pandemic. Indeed, we found an association between lower MMSE scores during lockdown and lower knowledge on COVID-19 related matters, which suggests a lack of understanding of the present situation. In this regard, Banovic et al. (2018) have previously shown that dementia is a root cause for communication difficulties, with persons with dementia displaying lower results in the area of understanding and verbal expression, repetition, reading and writing28. Nonetheless, anxiety remains common in the CI population with a prevalence from 8–71%, which results in poor outcome and quality of life even after controlling for depression29. Beyond quality of life and functional outcome, anxiety symptoms in dementia are also associated with environmental and social variables, preserved insight and greater prevalence of other neuropsychiatric and behavioral symptoms30,31. Therefore, the anxiety levels in CI are more likely to find its root cause in the AD pathology32 as described above, rather than being induced specifically by the lockdown.
When considering the individuals’ knowledge on COVID-19, it was revealed that CI participants performed significantly worse than the CU group. Nonetheless, this decreased knowledge on COVID-19 in CI participants did not seem to impact their change in anxiety over time. Perhaps, the temporary increase in time spent with caregivers helped to mitigate the increase in anxiety levels among the CI participants. A recent systematic review of 81 studies with 3347 participants revealed insights into the holistic experience of caring for a person with dementia, which encompass positive aspects of strengthening relationships and personal accomplishment, and negative aspects of emotions33.
In addition, the COVID-19-induced stress score was not significantly associated with the anxiety scores during lockdown in people with cognitive impairment, which further suggests that anxiety levels in these individuals might be biased by the load of disease pathology. This was in contrast with the CU individuals, where there was a strong association between anxiety and COVID-19-induced stress. Although increased anxiety is not a desired outcome of the measures that were put in place, decreased levels of anxiety could be indicative for people not taking necessary precautions. A study investigating the relationship between anxiety and risk taking behaviour has previously highlighted that higher levels of anxiety predict less risk taking in high-ambiguity situations34.
Interestingly, we found an inverse relationship between baseline tau load in the brain and change in anxiety levels due to COVID-19 pandemic. This indicates that CI participants with less tauopathy are more vulnerable to increased levels of anxiety during the lockdown. With recent evidence describing a link between the presence of tau and a more rapidly declining cognition, as measured by the MMSE score35, this could possibly explain the better situational understanding and the associated increase in anxiety levels in those with less tau burden.
The strength of this study is that this cohort consists out of participants with well-characterised cognitive and behavioral assessments enriched with tau PET scans. The prospective design of this study enables the evaluation of anxiety levels specifically due to the COVID-19 lockdown when compared to the baseline anxiety levels. Finally, the fact that our patients were not exposed to COVID supports that these findings were not confounded by its long-term neurological effects. There are several limitations to this study. Firstly, the study is observational and not mechanistic in design. Therefore, while we observed the anxiety trends in the CI and CU groups, we can only draw associations with other clinical markers obtained through the questionnaires.
In conclusion, stress related to COVID-19 pandemic differentially affects anxiety levels among CI and CU participants of the TRIAD study. This finding may have implications in management of CI patients in the clinic, and the interpretation of clinical trials of CI and CU participants in which anxiety constitutes an outcome measure.