Sleep duration, genetic susceptibility, and Alzheimer's disease: a longitudinal UK Biobank-based study

Combined effects of genetic predisposition and sleep quality on acceleration of biological ageing: Findings from the UK biobank cohort

OBJECTIVE

Genetic risks can accelerate ageing, yet better quality sleep may slow down it. We thus examined the interaction and combined effects of genetic predisposition and sleep quality on the risk of accelerate aging.

METHODS

This study included 407,027 participants from the UK Biobank. Sleep index of each participant was retrieved from the following seven sleep behaviors: snoring, chronotype, daytime sleepiness, sleep duration, insomnia, nap and difficulties in getting up. The biological age (PhenoAge) were estimated by corresponding algorithms based on clinical traits, and their residual discrepancies with chronological age were defined as the age accelerations (PhenoAgeaccel). We explored the interaction and combined effects of genetic risk and sleep quality on accelerated ageing by constructing a linear model.

RESULTS

Compared with participants in low sleep quality group, those in medium and high sleep quality group decreased 0.727 (95%CI, 0.653 to 0.801) and 1.056 (95%CI, 0.982 to 1.130) years of PhenoAgeaccel, respectively. Compared with participants in low genetic risk group, those in medium and high genetic risk group increased 0.833 (95%CI, 0.792 to 0.874) and 1.543 (95%CI, 1.494 to 1.592) years of PhenoAgeaccel, respectively. There was interaction between the genetic risk and sleep quality (P-interaction<0.001). For combined effect, compared to the group with high sleep quality and lower genetic risk, people with low sleep quality and high genetic risk had 2.747 (95%CI, 2.602 to 2.892) years higher PhenoAgeaccel.

CONCLUSION

Our findings elucidate that better sleep quality could lessen accelerated biological ageing especially among population with high genetic risk.

Negative health outcomes in long sleepers: The societal sleep restriction hypothesis

Society imposes work and school schedules, as well as social expectations, that militate against consistently obtaining more than 7-9 h of sleep every 24 h. For most but not all adults this sleep duration is adequate. But among those who consistently obtain more than 9 h of sleep per day ("long sleepers"), there likely exists a subpopulation of individuals who are nevertheless failing to obtain enough sleep to satisfy their physiological sleep needs - a consequence of "restricting" their daily sleep durations to whatever extent they can tolerate so as to conform as closely as possible to society's norms and expectations. It is hypothesized that the 'long sleep arm' of the seemingly paradoxical U-shaped relationship between sleep duration and negative health outcomes can be explained, at least in part, by the existence of a subpopulation of such 'sleep-restricted long sleepers.'

Prolonged sleep duration as a predictor of cognitive decline: A meta-analysis encompassing 49 cohort studies

Despite numerous studies have explored the association between sleep duration and cognition, the link between sleep duration trajectories and cognition remains underexplored. This systematic review aims to elucidate this correlation. We analyzed 55 studies from 14 countries, comprising 36 studies focusing on sleep duration, 20 on insomnia, and 13 on hypersomnia. A total of 10,767,085 participants were included in 49 cohort studies with a mean follow-up duration of 9.1 years. A non-linear association between sleep duration and cognitive decline was identified. Both long (risk ratio (RR):1.35, 95 % confidence intervals (CIs):1.23-1.48) and short sleep durations (RR: 1.12, 95 % CIs:1.03-1.22) were associated with an elevated risk of cognitive decline compared to moderate sleep duration. Additionally, hypersomnia (RR:1.26, 95 % CIs: 1.15-1.39) and insomnia (RR: 1.16, 95 % CIs: 1.002-1.34) were also linked to an increased risk. Moreover, prolonged sleep duration posed a higher risk of cognitive decline than stable sleep duration (RR:1.42, 95 % CIs:1.27-1.59). Importantly, transitioning from short or moderate to long sleep duration, as well as persistent long sleep duration, exhibited higher RRs for cognitive decline (RRs: 1.94, 1.40, and 1.28, respectively) compared to persistent moderate sleep duration. Our findings underscore the significance of prolonged sleep duration, alongside short and long sleep durations, with an elevated risk of cognitive decline. The association is tied to the degree of sleep duration changes. Our study highlights the importance of considering changes in sleep patterns over time, not just static sleep durations.

Causal relationship between sleep disorders and the risk of Alzheimer's disease: A Mendelian randomization study

BACKGROUND AND OBJECTIVE

Epidemiological studies have shown that sleep disorders are risk factors for Alzheimer's disease (AD), but the causal relationship between sleep disorders and AD risk is unknown. We aim to assess the potential genetic causal association between sleep characteristics and AD, which may contribute to early identification and prediction of risk factors for AD.

METHODS

Seven sleep-related traits and the outcome phenotype AD were selected from published genome-wide association studies (GWASs). These sleep-related characteristics and instrumental variables (IVs) for AD were extracted. Two-sample and multivariate Mendelian randomization (MR) analyses were performed to assess the causal relationships between sleep characteristics and AD. The inverse variance weighted (IVW), weighted median (WME), weighted mode (WM), MR-Egger regression (MR-Egger) and simple mode (SM) models were used to evaluate causality. The existence of pleiotropy was detected and corrected by MR-Egger regression, MR pleiotropy residuals and outliers.

RESULTS

A two-sample MR study revealed a positive causal association between sleep duration and the onset of AD (OR = 1.002, 95 % CI: 1.000-1.004), and the risk of AD increased with increasing sleep duration. The MR-Egger regression method and MR-PRESSO were used to identify and correct pleiotropy, indicating that there was no horizontal pleiotropy. Heterogeneity was evaluated by Cochran's Q, which indicated no heterogeneity. In a multivariate MR study with seven sleep characteristics corrected for each other, we found that sleep duration remained causally associated with AD (OR = 1.004, 95 % CI: 1.000-1.007). Moreover, we found that after mutual correction, daytime napping had a causal relationship with the onset of AD, and daytime napping may reduce the risk of AD (OR = 0.995, 95 % CI: 0.991-1.000).

CONCLUSION

This study is helpful for the early identification and prediction of risk factors for AD, long sleep durations are a risk factor for AD, and daytime napping can reduce the risk of AD.

Sleep characteristics and risk of Alzheimer's disease: a systematic review and meta-analysis of longitudinal studies

BACKGROUND

Alzheimer's disease (AD) is on the rise in our aging society, making it crucial to identify additional risk factors to mitigate its increasing incidence. This systematic review and meta-analysis aimed to provide updated evidence regarding the association between sleep and AD.

METHODS

We conducted a comprehensive search of MEDLINE, EMBASE, and Web of Science databases from inception to July 2023 to identify longitudinal studies. Adjusted relative risks were pooled for each sleep characteristic, and a dose-response analysis was performed specifically for sleep duration.

RESULTS

A total of 15,278 records were initially retrieved, and after screening, 35 records were ultimately included in the final analysis. The results showed that insomnia (RR, 1.43; 95%CI, 1.17-1.74), sleep-disordered breathing (RR, 1.22; 95%CI, 1.07-1.39), as well as other sleep problems, including sleep fragmentation and sleep-related movement disorders, were associated with a higher risk of developing AD, while daytime napping or excessive daytime sleepiness (RR, 1.18; 95%CI, 1.00-1.40) only exhibited a trend toward a higher risk of AD development. Furthermore, our analysis revealed a significant association between self-reported sleep problems (RR, 1.34; 95%CI, 1.26-1.42) and the incidence of AD, whereas this association was not observed with sleep problems detected by objective measurements (RR, 1.14; 95%CI, 0.99-1.31). Moreover, both quite short sleep duration (< 4 h) and long duration (> 8 h) were identified as potential risk factors for AD.

CONCLUSIONS

Our study found the association between various types of sleep problems and an increased risk of AD development. However, these findings should be further validated through additional objective device-based assessments. Additional investigation is required to establish a definitive causal connection between sleep problems and AD.

Association between Sleep Duration, Social Jetlag, and the Metabolic Syndrome by Shift Works

Lifestyle factors, including sleep characteristics, have been implicated in the development of metabolic syndrome, particularly among shift workers. This study aimed to explore the relationship between shift work, sleep duration, social jetlag, and the risk of metabolic syndrome among U.S. workers and the moderating effect of sleep duration and social jetlag on this relationship. Data from the National Health and Nutrition Examination Survey (NHANES) in 2017-2020 March were analyzed. Poisson regression models were employed to examine associations. Among 4136 U.S. workers, 53.3% had metabolic syndrome, with a higher proportion of shift workers (63.8% vs. 56.7%, p = 0.001) and those sleeping less than 6 h or more than 9 h per week (22.3% vs. 19.1%, p = 0.044) in the affected group. Shift workers were initially found to have an increased risk of metabolic syndrome (Coef. = 0.03, 95% CI: 0.02, 0.16); however, this association was mitigated when accounting for the interaction with social jetlag. Specifically, 1 to <2 h of social jetlag interacted significantly, increasing metabolic risk (Coef. = 0.15, 95% CI: 0.09, 0.22), whereas 1 to <2 h alone showed a protective effect (Coef. = -0.11, 95% CI: -0.17, -0.06). These findings suggest that optimizing sleep schedules and addressing social jetlag may be crucial in mitigating metabolic syndrome risks among shift workers.

The association of sleep duration and leukocyte telomere length in middle-aged and young-old adults: A cross-sectional study of UK Biobank

BACKGROUND

The relationships between sleep duration and aging-associated diseases are intricate. Leukocyte telomere length (LTL) is a biomarker of aging, while the association of sleep duration and LTL is unclear.

METHODS

The 310,091 study participants from UK Biobank were enrolled in this cross-sectional study. Restricted cubic splines (RCS) analysis was firstly performed to assess the nonlinear relationship between sleep duration and LTL. Sleep duration was then categorized into three groups: <7 h (short sleep duration), 7-8 h (reference group), and >8 h (long sleep duration) and multiple linear regression was applied to analyze the association of short sleep and long sleep duration with LTL. We further performed subgroup analyses stratified by sex, age, chronotype and snoring.

RESULTS

RCS showed an inverted J-shaped relationship between sleep duration and LTL. Compared with the reference group, the inverse association of long sleep duration and LTL was statistically significant in fully-adjusted model (P = 0.001). Subgroup analyses showed that this association was more apparent in people over 50 years (51-60 y: P = 0.002; >60 y: P = 0.005), in men (P = 0.022), and in people preferred evening chronotype (P = 0.001).

CONCLUSION

Compared with participants sleeping 7-8 h, those sleep longer than 8 h had shorter LTL in middle-aged and young-old adults. The negative association between long sleep duration and LTL was more apparent in older people, in men, and in people preferred evening chronotype.

Feelings of tense and risk of incident dementia: A prospective study of 482,360 individuals

BACKGROUND

The relationship between feelings of tense, as a significant emotional distress, and dementia remains unclear. This study aimed to evaluate the association between feelings of tense and dementia.

METHODS

In UK Biobank, feelings of tense were measured with a standard item. The primary outcome was all cause of dementia (ACD) and its subtypes (Alzheimer's disease (AD), vascular dementia (VD), and other dementia). Cox regression models analyzed the association between feelings of tense and dementia risk, while linear regression examined the correlation with neuroimaging outcomes. The potential association and joint effects of AD and tenseness were evaluated based on the established genetic risk score (GRS).

RESULTS

During a median follow-up of 12.7 years among 482,360 participants, 7331 dementia cases were identified. Individuals with feelings of tense had a significantly increased risk of ACD (HR, 1.194; 95 % CI: 1.115-1.278), VD (HR, 1.164; 95 % CI: 1.007-1.346), and other dementia (HR, 1.181; 95 % CI: 1.081-1.289), but not AD in multi-adjusted models. This association persisted across various sensitivity analyses and exhibited some heterogeneity in subgroup analyses. Furthermore, feelings of tense are associated with total brain volume shrinkage, higher white matter hyperintensities, and decreased partial subcortical volume, particularly in the hippocampus. No interaction between tenseness and AD genetic susceptibility was observed (P for interaction =0.346).

LIMITATIONS

Our study only considered feelings of tense measured at a one-time point.

CONCLUSIONS

Our findings demonstrate a significant association between feeling of tense and elevated dementia risk, indicating that tenseness could serve as a modifiable psychological determinant for dementia.

Sleep features and long-term incident neurodegeneration: a polysomnographic study

STUDY OBJECTIVES

Sleep is altered early in neurodegenerative diseases (NDDs) and may contribute to neurodegeneration. Long-term, large sample-size studies assessing NDDs association with objective sleep measures are scant. We aimed to investigate whether video-polysomnography (v-PSG)-based sleep features are associated with long-term NDDs incidence.

METHODS

Retrospective cohort study of patients referred 2004-2007 to the Sleep Disorders Unit, Neurology, Medical University Innsbruck, Austria. All patients ≥ 18 years undergoing v-PSG and without NDDs at baseline or within 5 years were included. Main outcome was NDDs diagnosis ≥5 years after v-PSG.

RESULTS

Of 1454 patients assessed for eligibility, 999 (68.7%) met inclusion criteria (68.3% men; median age 54.9 (IQR 33.9-62.7) years). Seventy-five patients (7.5%) developed NDDs and 924 (92.5%) remained disease-free after a median of 12.8 (IQR 9.9-14.6) years. After adjusting for demographic, sleep, and clinical covariates, a one-percentage decrease in sleep efficiency, N3-, or rapid-eye-movement (REM)-sleep was associated with 1.9%, 6.5%, or 5.2% increased risk of incident NDDs (HR 1.019, 1.065, and 1.052). One-percentage decrease in wake within sleep period time represented a 2.2% reduced risk of incident NDDs (HR 0.978). Random-forest analysis identified wake, followed by N3 and REM-sleep percentages, as the most important feature associated with NDDs diagnosis. Additionally, multiple sleep features combination improved discrimination of incident NDDs compared to individual sleep stages (concordance-index 0.72).

CONCLUSIONS

These findings support contribution of sleep changes to NDDs pathogenesis and provide insights into the temporal window during which these differences are detectable, pointing to sleep as early NDDs marker and potential target of neuroprotective strategies.

Impact of sleep duration and sleep disturbances on the incidence of dementia and Alzheimer's disease: A 10-year follow-up study

The nature of the relationship between sleep problems and dementia remains unclear. This study investigated the relationship between sleep measures and dementia in older adults (≥ 65) using data from the English Longitudinal Study of Ageing (ELSA) and further investigated the causal association in Mendelian randomization (MR) analysis. In total of 7,223 individuals, 5.7 % developed dementia (1.7 % Alzheimer's disease (AD)) within an average of 8 (± 2.9) years. Cox regression models and MR were employed. Long sleep duration (>8 h) was associated with 64 % increased risk of incident dementia and 2-fold high risk of AD compared to ideal sleep duration (7-8 h). This association was particularly evident in older-older adults (≥70 years) and those who consumed alcohol. Short sleep duration (<7 h) was associated with lower risk of incident dementia among older-older but higher risk among younger-older adults. Sleep disturbances and perceived sleep quality were not associated with dementia or AD. The MR study did not reveal causal associations between sleep duration and dementia. These findings suggest that self-reported short sleep in younger-older and long sleep in older-older adults and those with frequent alcohol consumption are associated with dementia. Early detection of these sleep patterns may help identify individuals at higher dementia risk.

Recent advances in understanding of sleep disorders and disturbances for dementia risk and prevention

PURPOSE OF REVIEW

To synthesise the recent work examining the relationship between sleep disturbances and dementia, emphasising studies involving individuals with mild cognitive impairment (MCI) or Alzheimer's disease (AD) and/or those investigating AD biomarkers. Additionally, we provide an update on recent interventions targeting sleep-related issues in older adults with MCI or AD.

RECENT FINDINGS

Various studies have examined obstructive sleep apnoea, sleep duration, and circadian alterations in relation to Alzheimer's pathology and dementia risk, with an emerging body of evidence suggesting that cardiovascular disease, hypertension, glymphatic function, and inflammation might serve as plausible pathophysiological mechanisms contributing to dementia during critical brain periods. Conversely, recent studies investigating insomnia have produced disparate results. Regarding intervention studies, the scarcity of prospective randomised control trials poses a challenge in establishing the benefits of addressing sleep disorders and disturbances.

SUMMARY

Recent work examining the pathophysiological links between sleep and dementia is strongest for obstructive sleep apnoea and sleep duration, while findings in insomnia studies exhibit inconsistency, possibly due to varied associations with dementia among different insomnia subtypes. It is apparent that more longitudinal studies examining the underlying pathophysiological mechanisms are necessary, alongside more rigorous clinical trials. Although some trials are underway in this field, there is still scarcity in trials examining interventions for circadian disturbances.

Association of the Sleep Regularity Index With Incident Dementia and Brain Volume

BACKGROUND AND OBJECTIVES

Irregular sleep may increase the risk of cardiometabolic conditions, but its association with incident dementia is unclear. The aim of this study was to assess the association between sleep regularity, that is, the day-to-day consistency in sleep-wake patterns and the risk of incident dementia and related brain MRI endophenotypes.

METHODS

We used Cox proportional hazard models to investigate the relationships between sleep regularity and incident dementia in 88,094 UK Biobank participants. The sleep regularity index (SRI) was calculated as the probability of being in the same state (asleep/awake) at any 2 time points 24 hours apart, averaged over 7 days of accelerometry.

RESULTS

The mean age of the sample was 62 years (SD = 8), 56% were women, and the median SRI was 60 (SD = 10). There were 480 cases of incident dementia over a median 7.2 years of follow-up. Following adjustments for demographic, clinical, and genetic confounders (APOE ε4), there was a nonlinear association between the SRI and dementia hazard (p [global test of spline term] < 0.001) with hazard ratios (HRs) following a U-shape pattern. HRs, relative to the median SRI, were 1.53 (95% CI 1.24-1.89) for participants with SRI at the 5th percentile (SRI = 41) and 1.16 (95% CI 0.89-1.50) for those with SRI at the 95th percentile (SRI = 71). In a subset with brain MRI (n = 15,263), gray matter and hippocampal volume tended to be lowest at the extremes of the SRI.

DISCUSSION

Sleep regularity displayed a U-shaped association with risk of incident dementia. Irregular sleep may represent a novel dementia risk factor.

Associations between the use of common nonsteroidal anti-inflammatory drugs, genetic susceptibility and dementia in participants with chronic pain: A prospective study based on 194,758 participants from the UK Biobank

OBJECTIVE

To investigate the potential relationship between common nonsteroidal anti-inflammatory drugs (NSAIDs), genetic susceptibility and all-cause dementia (ACD), Alzheimer's disease (AD), and vascular dementia (VD) among individuals experiencing chronic pain.

METHODS

This study was based on 194,758 chronic pain participants form UK biobank with a median follow-up of 13.7 years. Participants were categorized into different NSAIDs painkiller regimen groups: No NSAIDs group, Aspirin group, Ibuprofen group, Paracetamol group, and 2-3 NSAIDs group. Cox proportional risk models were used to examine the correlation between regular NSAIDs usage and the risk of ACD, AD, and VD. In addition, we further performed subgroup analyses and sensitivity analyses.

RESULTS

1) Compared to the No NSAIDs group, the aspirin group (HR = 1.12, 95% CI:1.01-1.24, P < 0.05), the paracetamol group (HR = 1.15, 95% CI:1.05-1.27, P < 0.01), and the 2-3 NSAIDs group (HR = 1.2, 95% CI:1.08-1.33, P < 0.05) showed a higher risk of ACD. Furthermore, the 2-3 NSAIDs group was also associated with a higher risk of VD (HR = 1.39, 95% CI: 1.08-1.33, P < 0.05). 2) At high dementia GRS participants with chronic pain, the paracetamol group (HR = 1.2, 95% CI: 1.03-1.43, P < 0.05) and the NSAIDs group (HR = 1.3, 95% CI: 1.07-1.59, P < 0.05) were associated with a higher risk of ACD compared to the no painkiller group. 3) There was no significant association between ibuprofen use and higher risk of dementia.

CONCLUSION

In individuals with chronic pain, the use of aspirin and paracetamol was associated with a higher risk of ACD, whereas the use of ibuprofen was not significantly associated with a higher risk of ACD.

Associations of screen-based sedentary activities with all cause dementia, Alzheimer's disease, vascular dementia: a longitudinal study based on 462,524 participants from the UK Biobank

BACKGROUND

Current drug treatments for dementia aren't effective. Studying gene-environment interactions can help develop personalized early intervention strategies for Alzheimer's disease (AD). However, no studies have examined the relationship between screen-based sedentary activities and genetic susceptibility to AD risk, and further understanding of the causal relationship is also crucial.

METHODS

This study included 462,524 participants from the UK Biobank with a follow-up of 13.6 years. Participants' screen-based sedentary activities time was categorized into three groups based on recorded time: ≥ 4 h/day, 2-3 h/day, and ≤ 1 h/day. Cox proportional risk models were used to analyze the association between computer use/TV viewing groups and the risk of all-cause dementia, AD and vascular dementia (VD). Generalized linear model (GLM) were used to examine the relationship between screen-based sedentary activities and brain structure. Bidirectional Mendelian randomization (MR) was used to validate the causal relationship between TV viewing and AD.

RESULTS

Compared to TV viewing ≤ 1 h/day, 1)TV viewing 2-3 h/day was correlated with a higher risk of all-cause dementia (HR = 1.09, 95% CI:1.01-1.18, P < 0.05), and TV viewing ≥ 4 h/day was associated with a higher risk of all-cause dementia (HR = 1.29, 95% CI: 1.19-1.40, P < 0.001), AD (HR = 1.25, 95% CI:1.1-1.42, P < 0.001), and VD (HR = 1.24, 95% CI: 1.04-1.49, P < 0.05); 2) TV viewing ≥ 4 h/day was correlated with a higher AD risk at intermediate (HR = 1.34, 95% CI: 1.03-1.75, P < 0.001) and high AD genetic risk score (AD-GRS) (HR = 2.18, 95% CI: 1.65-2.87, P < 0.001);3) TV viewing 2-3 h/day [β = (-94.8), 95% CI: (-37.9) -(-151.7), P < 0.01] and TV viewing ≥ 4 h/day [β = (-92.94), 95% CI: (-17.42) -(-168.46), P < 0.05] were correlated with a less hippocampus volume. In addition, a causal effect of TV viewing times was observed on AD analyzed using MR Egger (OR = 5.618, 95%CI:1.502-21.013, P < 0.05).

CONCLUSIONS

There was a causal effect between TV viewing time and AD analyzed using bidirectional MR, and more TV viewing time exposure was correlated with a higher AD risk. Therefore, it is recommended that people with intermediate and high AD-GRS should control their TV viewing time to be less than 4 h/ day or even less than 1 h/day.

High-Quality Sleep Mitigates ABCA7-Related Generalization Deficits in Healthy Older African Americans

BACKGROUND

Both sleep deficiencies and Alzheimer's disease (AD) disproportionately affect older African Americans. Genetic susceptibility to AD further compounds risk for cognitive decline in this population. Aside from APOE ɛ4, ABCA7 rs115550680 is the strongest genetic locus associated with late-onset AD in African Americans. While sleep and ABCA7 rs115550680 independently influence late-life cognitive outcomes, we know too little about the interplay between these two factors on cognitive function.

OBJECTIVE

We investigated the interaction between sleep and ABCA7 rs115550680 on hippocampal-dependent cognitive function in older African Americans.

METHODS

One-hundred fourteen cognitively healthy older African Americans were genotyped for ABCA7 risk (n = 57 carriers of risk "G" allele; n = 57 non-carriers), responded to lifestyle questionnaires, and completed a cognitive battery. Sleep was assessed via a self-reported rating of sleep quality (poor, average, good). Covariates included age and years of education.

RESULTS

Using ANCOVA, we found that carriers of the risk genotype who reported poor or average sleep quality demonstrated significantly poorer generalization of prior learning-a cognitive marker of AD-compared to their non-risk counterparts. Conversely, there was no genotype-related difference in generalization performance in individuals who reported good sleep quality.

CONCLUSION

These results indicate that sleep quality may be neuroprotective against genetic risk for AD. Future studies employing more rigorous methodology should investigate the mechanistic role of sleep neurophysiology in the pathogenesis and progression of AD associated with ABCA7. There is also need for the continued development of non-invasive sleep interventions tailored to racial groups with specific AD genetic risk profiles.

Genetic associations between sleep traits and cognitive ageing outcomes in the Hispanic Community Health Study/Study of Latinos

BACKGROUND

Sleep phenotypes have been reported to be associated with cognitive ageing outcomes. However, there is limited research using genetic variants as proxies for sleep traits to study their associations. We estimated associations between Polygenic Risk Scores (PRSs) for sleep duration, insomnia, daytime sleepiness, and obstructive sleep apnoea (OSA) and measures of cogntive ageing in Hispanic/Latino adults.

METHODS

We used summary statistics from published genome-wide association studies to construct PRSs representing the genetic basis of each sleep trait, then we studied the association of the PRSs of the sleep phenotypes with cognitive outcomes in the Hispanic Community Healthy Study/Study of Latinos. The primary model adjusted for age, sex, study centre, and measures of genetic ancestry. Associations are highlighted if their p-value <0.05.

FINDINGS

Higher PRS for insomnia was associated with lower global cognitive function and higher risk of mild cognitive impairment (MCI) (OR = 1.20, 95% CI [1.06, 1.36]). Higher PRS for daytime sleepiness was also associated with increased MCI risk (OR = 1.14, 95% CI [1.02, 1.28]). Sleep duration PRS was associated with reduced MCI risk among short and normal sleepers, while among long sleepers it was associated with reduced global cognitive function and with increased MCI risk (OR = 1.40, 95% CI [1.10, 1.78]). Furthermore, adjustment of analyses for the measured sleep phenotypes and APOE-ε4 allele had minor effects on the PRS associations with the cognitive outcomes.

INTERPRETATION

Genetic measures underlying insomnia, daytime sleepiness, and sleep duration are associated with MCI risk. Genetic and self-reported sleep duration interact in their effect on MCI.

FUNDING

Described in Acknowledgments.

Machine learning for the prediction of cognitive impairment in older adults

Objective

The purpose of this study was to develop and validate a predictive model of cognitive impairment in older adults based on a novel machine learning (ML) algorithm.

Methods

The complete data of 2,226 participants aged 60-80 years were extracted from the 2011-2014 National Health and Nutrition Examination Survey database. Cognitive abilities were assessed using a composite cognitive functioning score (Z-score) calculated using a correlation test among the Consortium to Establish a Registry for Alzheimer's Disease Word Learning and Delayed Recall tests, Animal Fluency Test, and the Digit Symbol Substitution Test. Thirteen demographic characteristics and risk factors associated with cognitive impairment were considered: age, sex, race, body mass index (BMI), drink, smoke, direct HDL-cholesterol level, stroke history, dietary inflammatory index (DII), glycated hemoglobin (HbA1c), Patient Health Questionnaire-9 (PHQ-9) score, sleep duration, and albumin level. Feature selection is performed using the Boruta algorithm. Model building is performed using ten-fold cross-validation, machine learning (ML) algorithms such as generalized linear model (GLM), random forest (RF), support vector machine (SVM), artificial neural network (ANN), and stochastic gradient boosting (SGB). The performance of these models was evaluated in terms of discriminatory power and clinical application.

Results

The study ultimately included 2,226 older adults for analysis, of whom 384 (17.25%) had cognitive impairment. After random assignment, 1,559 and 667 older adults were included in the training and test sets, respectively. A total of 10 variables such as age, race, BMI, direct HDL-cholesterol level, stroke history, DII, HbA1c, PHQ-9 score, sleep duration, and albumin level were selected to construct the model. GLM, RF, SVM, ANN, and SGB were established to obtain the area under the working characteristic curve of the test set subjects 0.779, 0.754, 0.726, 0.776, and 0.754. Among all models, the GLM model had the best predictive performance in terms of discriminatory power and clinical application.

Conclusions

ML models can be a reliable tool to predict the occurrence of cognitive impairment in older adults. This study used machine learning methods to develop and validate a well performing risk prediction model for the development of cognitive impairment in the elderly.