Sleep Scales Diverge in Alzheimer’s and FTD Syndromes

Key Takeaways

• Estimated overnight sleep duration differed across groups, with significantly longer sleep than in cognitively healthy older volunteers in Alzheimer disease, logopenic variant primary progressive aphasia, and right temporal variant frontotemporal dementia.
• Daytime somnolence was higher in Alzheimer disease and semantic variant primary progressive aphasia while Pittsburgh Sleep Quality Index total scores did not differ significantly overall, even though the Alzheimer disease group had lower PSQI scores than controls and nonfluent variant primary progressive aphasia.
• Across the dementia cohort, greater sleepiness tracked with more behavioral symptoms and lower empathy, and the authors said standard sleep scales may need careful interpretation in dementia because questionnaire profiles may not mirror objective sleep disturbance.

Longer overnight sleep and higher daytime somnolence did not map neatly onto perceived nighttime sleep quality across dementia syndromes in a study of ESS and PSQI across dementia syndromes. Daytime sleepiness was higher in Alzheimer disease versus cognitively healthy volunteers (OR 3.50, p=0.006), even though reported sleep-quality scores in Alzheimer disease moved in the opposite direction. The cohort spanned Alzheimer disease and frontotemporal dementia-related syndromes, with caregivers completing questionnaires for patients while cognitively healthy older volunteers self-reported. Lower PSQI scores in Alzheimer disease therefore did not mirror lower daytime sleepiness.

A comparative observational cohort in Jiang and colleagues’ dementia sleep questionnaire analysis paired same-day sleep, cognitive, and behavioral assessments in 58 participants with primary progressive aphasia or right temporal variant frontotemporal dementia, 32 with Alzheimer disease, and 36 cognitively healthy older volunteers; the syndromic subgroups included 16 with logopenic variant PPA, 18 with nonfluent variant PPA, 15 with semantic variant PPA, and 9 with right temporal variant frontotemporal dementia. Primary caregivers completed the Epworth Sleepiness Scale for patients to rate daytime sleepiness over the previous week and the Pittsburgh Sleep Quality Index to rate nighttime sleep quality over the previous month, whereas controls completed their own questionnaires, with scores above 10 on ESS and above 5 on PSQI treated as abnormal. Age-adjusted ordinal logistic regression was used for questionnaire comparisons, and correlation analyses tested links with cognitive and behavioral measures, providing a common frame for syndrome-level comparisons.

Longer estimated overnight sleep was reported across syndromic groups, although the detailed results showed significantly longer sleep versus controls in Alzheimer disease, logopenic variant primary progressive aphasia, and right temporal variant frontotemporal dementia; elevated ESS scores were concentrated in Alzheimer disease and semantic variant PPA rather than spread evenly across diagnoses. PSQI totals, by contrast, were not significantly different overall even though the Alzheimer disease group showed lower scores than controls and the nonfluent variant subgroup; average rising time was also later in Alzheimer disease (F=10.44, p=0.030), and the Alzheimer disease group showed a significant ESS-versus-PSQI z-score difference (V=8, p<0.001). The two questionnaires therefore yielded divergent syndrome-level sleep profiles.

Across the combined patient cohort, higher ESS scores tracked with worse Cambridge Behavioural Inventory-Revised scores (R=0.38, p<0.001) and lower modified Interpersonal Reactivity Index empathy subscale scores (R=-0.24, p=0.036). Within Alzheimer disease, poorer PSQI scores also tracked with worse Cambridge Behavioural Inventory-Revised scores (R=0.47, p=0.018), lower modified Interpersonal Reactivity Index empathy subscale scores (R=-0.56, p=0.005), and lower Revised Self-Monitoring Scale totals (R=-0.45, p=0.027). The authors said standard sleep scales may miss pathologically increased overnight sleep time, reduced sleep latency, or sleep attacks in dementia, so lower PSQI scores may not necessarily indicate objectively better sleep.

Clinician Questions

How were the Epworth Sleepiness Scale and Pittsburgh Sleep Quality Index completed in Alzheimer’s disease and frontotemporal dementia participants versus controls?

Primary caregivers completed Epworth Sleepiness Scale and Pittsburgh Sleep Quality Index questionnaires for patients with Alzheimer disease, logopenic variant primary progressive aphasia, nonfluent variant primary progressive aphasia, semantic variant primary progressive aphasia, and right temporal variant frontotemporal dementia, while cognitively healthy older volunteers completed their own questionnaires. The ESS captured daytime sleepiness over the past week, the PSQI assessed nighttime sleep quality over the past month, and sleep, cognitive, and behavioral assessments were completed on the same day.

Which behavioral measures were associated with higher daytime somnolence across the dementia cohort?

Across the combined patient cohort, higher ESS scores correlated with worse Cambridge Behavioural Inventory-Revised scores (R=0.38, p<0.001) and lower modified Interpersonal Reactivity Index empathy subscale scores (R=-0.24, p=0.036). Within Alzheimer disease, higher ESS scores also correlated with worse Cambridge Behavioural Inventory-Revised scores, lower modified Interpersonal Reactivity Index total scores, and lower empathy scores.

Why did the authors caution against reading lower Pittsburgh Sleep Quality Index scores in Alzheimer’s disease as better sleep?

The authors said the Pittsburgh Sleep Quality Index may not capture pathologically increased overnight sleep time, reduced sleep latency, or sleep attacks that can be relevant in dementia, and they noted a possible mismatch between subjective and objective sleep disturbance. Lower Pittsburgh Sleep Quality Index scores in the Alzheimer disease group were reported alongside higher daytime somnolence, so that pattern should not be translated into objectively better sleep.