The effect of COVID‐19 restrictions on sleep in individuals with cognitive impairment and their carers

Background
COVID-19 restrictions have been associated with apathy, anxiety and depression in individuals with cognitive impairment1-3 and increased stress in caregivers4. But despite links between these restrictions and sleep impairment in the general population5, effects on sleep in this cohort are not widely reported. Here we determine overall sleep quality in these populations during the first UK lockdown, how this changed over time and explanatory factors associated with longitudinal change.

Method
Cognitively impaired group (CIG), carer group (CG) and older healthy participants completed online self-reported validated questionnaires (PSQI, GAD-7, PHQ-8 and DBAS) alongside bespoke questions comprising the ‘SleepQuest’ study at two comparable stringent lockdown periods within the UK [t1 – (29/4/20-13/5/20) and t2 – (5/11/20-2/12/20)]. Analysis utilised paired t-tests with Bonferroni correction with 95% confidence intervals. Multivariable linear regression evaluated predictive factors.

Result
Within the CIG, baseline n = 79 [MCI, n = 33; AD, n = 19; Other, n = 27], mean age 68.7±9.7, with follow-up data for n = 37. Within the CG, baseline n = 183, mean age 59.4±11.1, follow-up n = 116. Overall sleep quality by PSQI Total was impaired at baseline (CIG = 7.8±4.3; CG = 7.9±4.0) compared to 6.6±3.6 for older healthy participants (n = 2354). There was a trend towards sleep quality deterioration from t1-t2 in both groups (CIG - Mean diff = 0.84, CI[-0.20,1.84], p = 0.111; CG - Mean diff = 0.51, CI[-0.01,1.02], p = 0.053). PSQI subcomponent analysis revealed increasing sleep medication use in both groups (CIG – Mean diff = 0.83, CI[0.36,1.32], pcorr = 0.007; CG – Mean diff = 0.99, CI[0.81,1.17], pcorr<0.001). However, levels of daytime dysfunction improved in both groups (CIG – Mean diff = -0.54, CI[-0.87,-0.21], pcorr = 0.014; CG – Mean diff = -0.46, CI[-0.61,-0.30], pcorr<0.001). Within the CIG, affective and behavioural factors at t1 did not predict sleep at t2. Within carers, increased anxiety (total GAD-7) at t1 predicted worse sleep at t2 (Est[normalised PSQI total] = 0.22, SE = 0.10, p = 0.041).

Conclusion
Overall sleep quality for individuals with cognitive impairment and their carers was poor at baseline and trended towards further deterioration, although daytime function improved. Within this cohort, potential compensatory strategies included increased use of sleep medication (associated with adverse outcomes). Given impaired sleep may accelerate progression of dementia neuropathology6,7, it is important to understand and ameliorate the impact of prolonged socio-political stressors on people with dementia and their carers.