The association between measures of sleepiness and subjective cognitive decline symptoms in a diverse population of cognitively normal older adults

Abstract

Subjective cognitive decline (SCD) is associated with preclinical Alzheimer's disease (AD). Suboptimal sleep is also a risk factor for cognitive decline, but with unclear relationship to SCD. We conducted a retrospective cross-sectional study in a biracial research cohort of 148 cognitively normal older adults who underwent quantification of SCD (Cognitive Change Index; CCI), sleepiness (Epworth Sleepiness Scale; ESS), depression (Geriatric Depression Scale; GDS), and amyloid/tau PET. ESS score was associated with total, amnestic, and non-amnestic CCI scores, after adjustment for GDS, amyloid/tau burden, and race. This supports future longitudinal work on how sleepiness impacts SCD outcomes.

Keywords

Alzheimer's disease amnestic cognitive aging cognitive dysfunction depression excessive sleepiness imaging biomarkers non-amnestic subjective cognitive decline

Introduction

Subjective cognitive decline (SCD) is a self-experienced deterioration of memory and other cognitive capacity despite normal objective cognitive performance.¹ SCD is associated with the preclinical stage of Alzheimer's disease (AD), with correlations to cerebral amyloid and tau levels and subsequent objective cognitive decline. As such, SCD forecasts progression to mild cognitive impairment (MCI) and AD dementia.² Therefore, recognizing, monitoring, and perhaps even treating SCD directly may be important elements of early intervention to prevent or slow down cognitive decline.

Optimal sleep health is also a significant factor in an individual's overall health, rejuvenation, and recovery process, and plays a role in regulating mood, and enhancing well-being and quality of life^3,4 and brain health.^5,6 Sleep is frequently affected in early AD stages, suggesting that there may be a bidirectional relationship between sleep disorders, cognitive decline, and progression to MCI and dementia.^7,8 Moreover, there is an association between sleep disorders and AD biomarkers in cognitively normal individuals.⁹ However, suboptimal sleep may also impact brain health through other pathways as well.

SCD and sleep are also linked, as individuals with sleep problems tend to report SCD, as well as associated mood symptoms.⁷ Yet, it is unclear if and how sleep dysfunction relates to SCD apart from AD biomarkers, or how symptoms related to SCD and sleep dysfunction, such as depression, might play a role. Furthermore, studies of SCD and sleep dysfunction have not distinguished between amnestic and non-amnestic domains, and have also primarily focused on non-Hispanic White (NHW) populations despite Black/African Americans (B/AA) being more susceptible to poor sleep.³ Therefore, to fill these gaps, we investigated the cross-sectional relationship between excessive sleepiness in cognitively normal older adults and SCD symptoms linked to cognitive domains in a cohort of cognitively normal NHWs and B/AA.

Methods

Study participants

This retrospective cross-sectional study included a de-identified dataset of 148 cognitively normal older participants evaluated at the NYU Alzheimer's Disease Research Center (ADRC) between January 2019 and January 2023. The NYU ADRC study is approved by the NYU Langone Grossman School of Medicine Institutional Review Board. The NYU ADRC recruits community-dwelling older adults, with inclusion criteria of age >60, the presence of a study partner, fluency of both participant and study partner in either English or Spanish, and consent to neuroimaging (MRI or amyloid and/or tau PET MRI). Exclusion criteria include major neurological disease outside of AD and related disorders, major psychiatric illness, organ failure/transplant, and alcohol/drug abuse. Participants underwent annual evaluation that included neuropsychological and other assessments using the National Alzheimer's Coordinating Center Uniform Data Set 3.0 and additional NYU-specific measures. Normal cognition was confirmed via psychometric testing and consensus diagnosis. Inclusion criteria for this study included NHW or B/AA race and completion of relevant clinical measures (see subsequent section).

Clinical measures

SCD characteristics were measured by the Cognitive Change Index (CCI), a validated, 20-question measure of SCD using a Likert scale.¹⁰ The CCI was further divided into subscores of amnestic domains (i.e., memory, questions 1–12) and non-amnestic (i.e., executive, questions 13–17) domains. The Geriatric Depression Scale (GDS) was measured to quantify depressive symptoms in older adults.¹¹ Finally, the widely used Epworth Sleepiness Scale (ESS) was used to quantify the severity of sleep-related problems, including the likelihood of dozing off in various everyday situations.¹²

Neuroimaging

PET acquisition: A subset of participants underwent brain amyloid assessment using second generation 18F-florbetaben tracer and smaller subset also had brain tau assessment using either ¹⁸F-PI-2620 or ¹⁸F-MK-6240 tracers. PET imaging was compliant with the acquisition and image reconstruction protocol prescribed by NIH/NIA Standardized Centralized Alzheimer's & Related Dementias Neuroimaging (SCAN) initiative (http://www.scan.naccdata.org). Amyloid PET MRI was completed by 129 participants (87.2% of the total). Tau PET MRI was completed by 70 participants (47.3% of the total). The number of study participants who completed both amyloid and tau PET exams was 57 (38.5% of the total). There were 6 participants who did not complete amyloid or tau PET and consented to MRI alone.

PET image processing: Standardized uptake value ratio (SUVR) for ¹⁸F-florbetaben was determined following the recommended ADNI method [Landau 2021]. A composite neocortical region of interest (ROI) comprising the frontal, lateral parietal, and lateral temporal cortex was used as the target region and the whole cerebellum served as the reference region. The neocortical SUVR was computed based on the average, volume-weighted activity within. The neocortical ROI and whole cerebellum were defined based on FreeSurfer-derived Desikan-Killiany atlas regions. SUVR positivity cutoff was 1.08.¹³ Tau PET SUVR (for both ¹⁸F-PI-2620 and ¹⁸F-MK-6240) was determined with cerebellar gray matter as the reference region and the target regions that included the fusiform, parahippocampal, hippocampus, posterior cingulate, and temporal cortex.¹⁴ The tau positivity threshold was 1.25.¹⁵

Statistical analysis

Descriptive statistics were used to summarize demographics characteristics of the sample. CCI total, amnestic, and non-amnestic scores were divided by the number of questions contributing to each score. Multivariable linear regression models tested the association of ESS with the three CCI outcome measures (total score, non-amnestic domain, amnestic domain), adjusting for covariates (GDS, amyloid SUVR, tau SUVR, race). Given the exploratory nature of this analysis, we did not adjust for multiple testing of the association of ESS with the three CCI measures and report 95% confidence intervals and not p-values in the text. The confidence intervals should be interpreted in this context. All statistical analyses were performed using R, version 4.2.1 (R Core Team, Vienna, Austria).

Results

In this study, there were 148 cognitively normal subjects from the NYU ADRC who fit inclusion criteria and who were included in this analysis. Their characteristics are summarized in Table 1, showing robust representation of women (nearly 69%) and B/AA participants (∼41%). ESS total scores ranged from 0 to 17. GDS scores ranged from 0 to 10. Of those who completed amyloid PET MRI, 24.8% (32/129) were positive. Of those who completed tau PET MRI, 15.7% (11/70) were positive.

Table 1.

Participant demographics and biomarker characteristics of study population. Values are expressed as mean ± standard deviation, or number of participants (percentage) %).

Characteristic	N (%) or mean +/–or– SD
Age, y (SD)	72.93 +/– 6.50
Education, y (SD)	16.66 +/–2.26
Women	102 (68.9%)
Black/ African American	61 (41.2%)
Non-Hispanic White	87 (58.8%)

We first assessed the association between sleepiness (ESS) and SCD when including all domains of the CCI (CCI Total). These results, as shown in Table 2, indicate a positive association between increased sleepiness and SCD symptomatology; a one unit increase in ESS total is associated with a 0.03 (95% CI: 0.008, 0.052) unit increase in CCI total. Of note, this was found via multivariable linear regression to adjust for amyloid and tau SUVR, depression (GDS), and race. Furthermore, as expected, there was an independent association between CCI Total and the extent of depressive symptoms as measured by the GDS. Surprisingly, there was no statistically significant association found between CCI Total and amyloid or tau burden in this cohort after adjustment. Lastly, race (NHW versus B/AA) did not correlate with the magnitude of SCD symptoms.

Table 2.

Multivariable linear regression to assess correlation of Cognitive Change Index (CCI) total to sleepiness and covariates.

	Coefficient	Standard Error	95% CI	p
ESS Total	0.02979	0.01120	[0.008, 0.052]	0.01042
Amyloid SUVR	–0.13906	0.21186	[–0.55, 0.28]	0.51455
Tau SUVR	0.29999	0.39890	[–0.48, 1.08]	0.45548
GDS Total	0.07118	0.02170	[0.029, 0.11]	0.00187
Race	–0.09853	0.09618	[–0.29, 0.09]	0.31046

Unadjusted p-values <0.05 are listed in boldface. CI: confidence interval (of coefficient); ESS: Epworth sleepiness scale; SUVR: standardized uptake value ratio; GDS: Geriatric Depression Scale

We then assessed whether these associations were domain-specific using two additional multivariable linear regression models. Analysis focusing on amnestic SCD symptoms is reported in the top of Table 3. The multivariable linear regression model included the same covariates. These results revealed the same pattern as with CCI Total; a one unit increase in ESS total is associated with a 0.035 (95% CI: 0.013, 0.057) unit increase in CCI amnestic. It is also positively associated with depression. We again found no adjusted correlations with amyloid and tau burden, nor with race. A third model with the same covariates examined only non-amnestic, executive SCD symptoms on the CCI (Table 3, bottom). This analysis also demonstrated a positive correlation with ESS; a one unit increase in ESS total is associated with a 0.034 (95% CI: 0.0099, 0.058) unit increase in CCI non-amnestic. It does not appear to be associated with amyloid and tau burden and race. Different from the amnestic subtotal, it does not appear to be associated with depression.

Table 3.

Multivariable linear regression to assess correlation of Cognitive Change Index (CCI) subdomains to sleepiness and covariates.

Cognitive Change Index, Amnestic
	Coefficient	Standard Error	95% CI	p
ESS Total	0.03491	0.01521	[0.005, 0.065]	0.02583
Amyloid SUVR	−0.32335	0.28772	[−0.89, 0.24]	0.26636
Tau SUVR	0.73663	0.54173	[−0.33, 1.80]	0.17989
GDS Total	0.08719	0.02947	[0.029, 0.14]	0.00468
Race	−0.20535	0.13062	[−0.46, 0.051]	0.12210

Cognitive Change Index, Non−Amnestic
	Coefficient	Standard Error	95% CI	p
ESS Total	0.033764	0.012169	[0.0099, 0.058]	0.00771
Amyloid SUVR	−0.023986	0.230239	[−0.69, 0.21]	0.91744
Tau SUVR	−0.001616	0.433501	[−0.85, 0.85]	0.99704
GDS Total	0.040775	0.023583	[−0.005, 0.087]	0.08986
Race	0.149781	0.104522	[−0.055, 0.35]	0.15796

Unadjusted p-values <0.05 are listed in boldface. CI: confidence interval (of coefficient); ESS: Epworth sleepiness scale; SUVR: standardized uptake value ratio; GDS: Geriatric Depression Scale

Discussion

Here we find that sleepiness measured via ESS correlates with SCD magnitude on the CCI, across both amnestic and non-amnestic domains. These effects were independent of cerebral amyloid, tau deposition, and depression, with only depression being independently associated with CCI.

The implications of our findings depend on if sleepiness has a causal impact on SCD outcomes. If sleepiness influences SCD magnitude without negatively influencing cognitive outcomes, sleepiness would be interpreted as confounding the establishment of SCD as a preclinical AD stage. This would necessitate that clinicians screen and treat suboptimal sleep first, to determine whether SCD symptoms persist. Alternatively, the influence of sleepiness on SCD magnitude may indeed associate with negative cognitive outcomes. As such, the identification of SCD, no matter the underlying causes - whether they be sleepiness, tau, amyloid, depression and other associated risks, or some combination, should warrant some level of concern as a preclinical predictor of cognitive decline. Indeed, individuals with poor sleep quality and excessive sleepiness have a greater risk of cognitive decline compared with those with good sleep quality.¹⁶ In this case, sleepiness or the drivers of impaired sleep (e.g., obstructive sleep apnea¹⁷) may directly impact brain health apart from amyloid and tau accumulation. In other words, SCD in some individuals may be a preclinical stage for non-AD etiologies that are influenced by poor sleep. For example, studies have revealed that cognitively normal individuals who have experienced sleep disturbances have a higher risk of cerebrovascular disease, specifically white matter hyperintensities (WMH),¹⁸ which in turn can shape SCD.¹⁹ Other studies have suggested that there is a mechanistic link between sleepiness and neuroinflammation.^20–22 In our study, the involvement of both amnestic and non-amnestic (executive) symptoms of SCD suggest that both temporal and frontal lobes may be impacted by suboptimal sleep, as suggested by FDG PET.²³

Surprisingly, unlike sleepiness, we found that amyloid and tau did not have statistically significant independent correlations with SCD magnitude. This may be due to the heterogenous cohort, as AD biomarkers have been less predictive of elevated dementia risk in B/AA persons.^24,25 Alternatively, it has been shown that relationship to AD biomarkers increases with additional variables related to memory domain predominance, worry about memory, informant concordance, and APOE ε4 carrier positivity.¹ It is possible that these factors were less represented in this sample. We did find a correlation between depression and CCI and its amnestic domain, as expected from prior studies.² This suggests that affect plays a role in driving SCD in our cohort, but that it exerts an independent and phenotypically different effect since it does not correlate with non-amnestic features like sleepiness.

Strengths of our study include a racially heterogeneous cohort, quantification of SCD magnitude, the breakdown of SCD across domains, and the incorporation of AD imaging biomarkers. This permitted the simultaneous assessment of sleepiness, SCD, tau, amyloid, and depression within a single study. This study has some limitations as well. This is a retrospective cross-sectional study that does not allow us to understand whether the associations revealed are causal. The implementation of longitudinal study design will help in the understanding of the directionality of these relationships and associations of ESS and SCD. Another limitation is the use of a self-reported scale as opposed to an objective measure of sleepiness that may offer greater accuracy. One reason this might be problematic is that older adults may potentially underestimate their own sleepiness symptoms.

In summary, our findings suggest that SCD is influenced by sleepiness apart from AD biomarkers and depression. Future work should examine the impact of factors related to sleepiness, including overall quality of sleep, sleep habits, and sleep disorders. Moreover, to draw firmer conclusions on any causal links between sleepiness and SCD, future studies should ascertain in a longitudinal fashion whether SCD outcomes are modulated by sleepiness level, or if interventions to improve sleep impact SCD outcomes over time.

Footnotes

Acknowledgments

This work was supported by NIH P30AG066512, the Louis J. And June E. Kay Foundation, and Alzheimer's Association AARF-D-24-1243696.

ORCID iDs

Anthony Q Briggs

Carolina Boza-Calvo

Mark A Bernard

Henry Rusinek

Rebecca A Betensky

Arjun V Masurkar

Author contributions

Anthony Q Briggs (Conceptualization; Writing – original draft; Writing – review & editing); Carolina Boza-Calvo (Writing – review & editing); Mark A Bernard (Writing – review & editing); Henry Rusinek (Data curation; Formal analysis; Methodology); Rebecca A Betensky (Data curation; Formal analysis; Methodology); Arjun V Masurkar (Conceptualization).

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Declaration of conflicting interests

Dr Masurkar is a council member of the Alzheimer's Association International Research Grant Program, is a steering committee member of the Alzheimer's Disease Cooperative Study, and serves on the editorial boards of Journal of Neuro-ophthalmology and Alzheimer's and Dementia: Translational Research and Clinical Interventions.

References

Jessen

Amariglio

Van Boxtel

, et al. A conceptual framework for research on subjective cognitive decline in preclinical Alzheimer's disease. Alzheimers Dement 2014; 10: 844–852.

Parfenov

Zakharov

Kabaeva

, et al. Subjective cognitive decline as a predictor of future cognitive decline: a systematic review. Dement Neuropsychol 2020; 14: 248–257.

Baron

Gilyard

Williams

, et al. Sleep-related attitudes, beliefs, and practices among an urban-dwelling African American community: a qualitative study. Sleep Health 2019; 5: 418–425.

Jean-Louis

Kripke

Cole

, et al. Sleep detection with an accelerometer actigraph: comparisons with polysomnography. Physiol Behav 2001; 72: 21–28.

Zhao

Veasey

. Neural consequences of chronic short sleep: reversible or lasting? Front Neurol 2017; 8: 235.

Borges

Poyares

Studart-Neto

, et al. Amyloid profile is associated with sleep quality in preclinical and prodomal Alzheimer’s disease older adults. Sleep Medicine 2024; 121: 359–364.

Exalto

Hendriksen

Barkhof

, et al. Subjective cognitive decline and self-reported sleep problems: the SCIENCe project. Alzheimers Dement (Amst) 2022; 14: e12287.

Zuelsdorff

Okonkwo

Norton

, et al. Stressful life events and racial disparities in cognition among middle-aged and older adults. J Alzheimers Dis 2020; 73: 671–682.

Winer

Deters

Kennedy

, et al. Association of short and long sleep duration with amyloid-β burden and cognition in aging. JAMA Neurol 2021; 78: 1187–1196.

10.

Rattanabannakit

Risacher

Gao

, et al. The cognitive change index as a measure of self and informant perception of cognitive decline: relation to neuropsychological tests. J Alzheimers Dis 2016; 51: 1145–1155.

11.

Yesavage

Brink

Rose

, et al. Development and validation of a geriatric depression screening scale: a preliminary report. J Psychiatr Res 1982; 17: 37–49.

12.

Johns

. A new method for measuring daytime sleepiness: the epworth sleepiness scale. Sleep 1991; 14: 540–545.

13.

Royse

Minhas

Lopresti

, et al. Validation of amyloid PET positivity thresholds in centiloids: a multisite PET study approach. Alzheimers Res Ther 2021; 13: 99.

14.

Pascoal

Therriault

Benedet

, et al. 18F-MK-6240 PET for early and late detection of neurofibrillary tangles. Brain 2020; 143: 2818–2830.

15.

Bullich

Mueller

De Santi

, et al. Evaluation of tau deposition using (18)F-PI-2620 PET in MCI and early AD subjects-A MissionAD tau sub-study. Alzheimers Res Ther 2022; 14: 105.

16.

Leng

Yin

Zhang

, et al. Sleep quality and health-related quality of life in older people with subjective cognitive decline, mild cognitive impairment, and Alzheimer disease. J Nerv Ment Dis 2020; 208: 387–396.

17.

Barletta

Abreu

Ramos

, et al. Role of obstructive sleep apnea in cognitive impairment. Int J Head Neck Surg 2019; 10: 57.

18.

Zhao

Kong

Wang

, et al. Association of excessive daytime sleepiness with cerebral small vessel disease in community-dwelling older adults. Nat Sci Sleep 2022; 14: 765–773.

19.

Rothstein

Zhang

Briggs

, et al. Impact of white matter hyperintensities on subjective cognitive decline phenotype in a diverse cohort of cognitively normal older adults. Int J Geriatr Psychiatry 2023; 38: e5948.

20.

Hall

Smagula

Boudreau

, et al. Association between sleep duration and mortality is mediated by markers of inflammation and health in older adults: the health, aging and body composition study. Sleep 2015; 38: 189–195.

21.

Xing

Zhou

, et al. Sleep disturbance induces depressive behaviors and neuroinflammation by altering the circadian oscillations of clock genes in rats. Neurosci Res 2021; 171: 124–132.

22.

Duncan

Guerriero

Kohler

, et al. Chronic fragmentation of the daily sleep-wake rhythm increases amyloid-beta levels and neuroinflammation in the 3xTg-AD mouse model of Alzheimer’s disease. Neuroscience 2022; 481: 111–122.

23.

André

Tomadesso

de Flores

, et al. Brain and cognitive correlates of sleep fragmentation in elderly subjects with and without cognitive deficits. Alzheimers Dement (Amst) 2019; 11: 142–150.

24.

Deters

Napolioni

Sperling

, et al. Amyloid PET imaging in self-identified non-hispanic black participants of the anti-amyloid in asymptomatic Alzheimer's disease (A4) study. Neurology 2021; 96: e1491–e1500.

25.

Morris

Schindler

McCue

, et al. Assessment of racial disparities in biomarkers for Alzheimer disease. JAMA Neurol 2019; 76: 264–273.