Changes in Sleep Patterns, Genetic Susceptibility, and Incident Cardiovascular Disease in China

This cohort study explores associations of genetic risk and 5-year changes in sleep patterns with incident cardiovascular disease among retired adults in China.


'Introduction
Cardiovascular disease (CVD) is a major cause of morbidity and mortality worldwide. 1In 2019, CVD accounted for approximately one-third of all deaths globally, with more than 40% of deaths in China. 2 As the CVD burden continues rising in almost all countries, identifying modifiable risk factors for CVD prevention is urgent.
Several studies 6,[8][9][10] have proposed the development of healthy sleep patterns, assessing sleep as a multidimensional construct, and have shown inverse associations of healthy sleep patterns with the risk of CVD.However, most of these studies have only used a single measurement, which might not adequately reflect the association of overall sleep with CVD because sleep habits may change over time.Only 1 study 10 of 9309 participants living in Europe showed that maintaining healthy sleep patterns over 2 to 5 years was associated with a lower risk of CVD and coronary heart disease (CHD), but not stroke.However, the study 10 was conducted among middle-aged people who typically adjusted their sleep patterns around work schedules.Evidence from retired, older people with natural sleep patterns is still lacking.
In addition to lifestyle factors, genetic factors are also associated with CVD. 11,12Several studies [13][14][15][16][17] have explored the joint association of genetic risk and lifestyle with the risk of incident CVD outcomes; these studies showed that individuals adhering to healthy lifestyles had a lower risk of CHD or stroke, even among those at high genetic risk.It is unknown whether maintaining favorable sleep patterns over time is associated with a lower risk of CHD or stroke among individuals with higher genetic susceptibility.
To fill the evidence gap, we collected sleep information at 2 time points approximately 5 years apart, and prospectively explored the long-term outcomes of changes in sleep patterns on the subsequent incidence of CVD outcomes among middle-aged and older Chinese retirees.We further investigated how the 5-year changes in sleep patterns interact and combine with CVD-related genetic variants for the risk of CVD outcomes.

Study Population
This cohort study was approved by the Ethics and Human Participants Committees of Tongji Medical College, Huazhong University of Science and Technology, and Dongfeng General Hospital.This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.The Dongfeng-Tongji cohort is an ongoing prospective cohort study in Shiyan, China. 18A total of 27 009 retired workers from the Dongfeng Motor Corporation (DMC) were recruited at the baseline survey between September 2008 and June 2010, and among them, 24 175 participated in the first follow-up survey in 2013.Participants completed standardized questionnaires and medical examinations in both surveys.We excluded participants with diagnosed CHD, stroke, cancer, or severely abnormal electrocardiogram findings before the first follow-up and those with missing sleep information at baseline or the first follow-up survey (eFigure 1 in Supplement 1).Among eligible participants, a subgroup of participants had available genetic data collected.Written informed consent was obtained from all participants.

JAMA Network Open | Cardiology
Sleep Patterns, Genetic Susceptibility, and Incident Cardiovascular Disease

Changes in Sleep Patterns
Sleep information was self-reported and collected through standardized questionnaires at baseline from 2008 to 2010 and at the first follow-up in 2013.Details of the assessment are provided in the eMethods in Supplement 1.To assess overall sleep, we created a sleep score based on 4 low-risk sleep factors, details of which were presented in our previous study. 60][21][22][23] Sleep factors were treated as dichotomous variables (low risk coded as 1 and high risk coded as 0), and sleep score was the sum of all the sleep factors ranging from 0 to 4.
To determine the changes in sleep patterns from baseline to the first follow-up, considering the sample size and associated statistical power, we defined sleep patterns as unfavorable (sleep score Յ2) and favorable (sleep score Ն3) and then divided participants into 4 groups: persistent unfavorable, favorable-unfavorable (ie, transitioning from favorable to unfavorable), unfavorablefavorable (ie, transitioning from unfavorable to favorable), and persistent favorable.

Polygenic Risk Score
The genotyping arrays used in this study were Illumina Infinium OmniZhongHua-8 chips.Detailed information on the genotyping process was described elsewhere. 24The reference panel in genotype imputation was 3931 samples of individuals with East Asian heritage from the 1000 Genomes Project phase III and the SG10K Project.The single-nucleotide variations (SNVs) selected for calculating the polygenic risk scores (PRS) for CHD and stroke were from the recently published and validated CHD PRS 25 (comprising 540 SNVs) and stroke PRS 26 (comprising 534 SNVs); of these SNVs, 533 SNVs for CHD (eTable 1 in Supplement 1) and 527 SNVs for stroke (eTable 2 in Supplement 1) reached the criteria of a call rate greater than 95%, minor allele frequency greater than 0.001, P value greater than 1 × 10 −6 in Hardy-Weinberg equilibrium tests, and a mean R 2 imputation quality greater than 0.3 in the Dongfeng-Tongji cohort data.Each SNV was coded as 0, 1, or 2, according to the number of risk alleles it carried.The weighted PRS for CHD and stroke was calculated separately using the following equation: where SNV was coded as 0, 1, or 2 according to the number of risk alleles it carried and the β-coefficient for each SNV was obtained from the previously published CHD and stroke PRS. 25,26The CHD PRS and stroke PRS followed normal distribution (eFigure 2 in Supplement 1).The PRS were then classified into low (quintile 1), intermediate (quintile 2-4), or high (quintile 5) genetic risk groups and showed good stratification capability for CHD and stroke (eTable 3 in Supplement 1).

Outcomes
All participants could be tracked for morbidity and mortality through the DMC health care system and death certificates.The primary outcome in this study was incident CVD, assessed until December 31,

Covariates
At baseline and the first follow-up survey, demographic characteristics (age, sex, and education), lifestyle (drinking status, smoking status, and physical activity), and medical history were collected by trained interviewers using standardized questionnaires.Physical examination, including weight, standing height, blood pressure, fasting glucose, and blood lipid levels, were measured by trained

JAMA Network Open | Cardiology
Sleep Patterns, Genetic Susceptibility, and Incident Cardiovascular Disease physicians (see the eMethods in Supplement 1 for details).The covariates used for adjustment in this study were those collected at the first follow-up survey in 2013.

Statistical Analysis
The characteristics of participants at follow-up in 2013 were described as mean (SD) for continuous variables and percentages for categorical variables.We calculated the follow-up time for each participant from the date of recruitment at the first follow-up survey to the date of diagnosis of CVD, death, or the censoring date (December 31, 2018), whichever came first.We applied Cox proportional hazard regression models to calculate the hazard ratios (HRs) and 95% CIs.Potential covariates adjusted in the models were age; sex; education level; regular exercise; drinking status; smoking status; body mass index (BMI); hypertension; hyperlipidemia; diabetes; and family history of CVD, CHD, or stroke (in the corresponding analysis).We assessed the assumption of proportional hazards with a test based on Schoenfeld residuals 29 ; the nonsignificant results suggested that the models met the assumption.We evaluated HRs and 95% CIs for CVD, CHD, and stroke according to changes in sleep patterns, with a persistent unfavorable sleep pattern being the reference group.We also examined the associations of the changes in sleep patterns with the risk of incident CVD, CHD, and stroke stratified by age (<65 and Ն65 years) and sex (male or female).
In the subgroup of participants with genetic data, we assessed the association of changes in sleep patterns with incident CHD and stroke stratified by each genetic risk group of CHD and stroke, with the potential effect modification tested by including a multiplicative interaction term between the changes in sleep patterns and the genetic risk group.When stratified by PRS, the number of cases among the groups of favorable-unfavorable and unfavorable-favorable sleep patterns was relatively small.Given that both groups represent changed sleep patterns and their associations with CVD risk were similar, we combined them into 1 group in the subsequent analyses to ensure more stable estimates.We further explored the joint associations of changes in sleep patterns and genetic risk of CHD and stroke with incident CHD and stroke, using high genetic risk combined with persistent unfavorable sleep patterns as the reference group.We also performed the linear trend test by treating the joint category as a continuous variable.
We conducted several sensitivity analyses to test the robustness of the results.To reduce the possibility of reverse causation, we repeated the primary analyses after excluding participants with the event of interest observed within the first year of follow-up.To examine whether the results were sensitive to the use of hypnotics, we also reran the primary analyses after excluding participants who reported very poor sleep quality with frequent use of hypnotics.All statistical analyses were performed using R software version 4.2.2 (R Project for Statistical Computing) in November 2023.
Statistical significance was defined as a 2-sided P value < .05.

Results
The final cohort included 15   favorable sleep patterns had a 16% lower risk of CHD (HR, 0.84; 95% CI, 0.76-0.92)and 34% lower risk of stroke (HR, 0.66; 95% CI, 0.54-0.82).The associations of changes in sleep patterns with CVD, CHD, and stroke were consistent in subgroup analyses stratified by age and sex (eTable 4 in Supplement 1).

Table 2 presents the PRS-stratified analyses of the association of 5-year changes in sleep
patterns with incident CHD and stroke.Sleep patterns were not associated with risk of incident CHD among those with low or high genetic risk (P for interaction = .36).Persistent favorable sleep patterns were associated with a 15% lower risk of CHD in the intermediate genetic risk subgroup (aHR, 0.85; 95% CI, 0.74-0.98).For stroke, there was also no evidence of stroke PRS being a modifier (P for interaction = .23).Among individuals with persistent favorable sleep patterns, those at low genetic risk had a lower but not statistically significant risk of incident stroke (aHR, 0.77; 95% CI,

Discussion
In this prospective cohort study, individuals with persistent favorable sleep patterns over 5 years had the lowest risk of incident CVD, CHD, and stroke during the subsequent 5 years.There was no evidence of effect modification by genetic susceptibility to CHD or stroke.Joint subgroups of changes in sleep patterns and genetic risk were inversely associated with the risk of CHD and stroke, with the lowest risk found in individuals at low genetic risk who maintained favorable sleep patterns over 5 years.
Little is known about the associations of longitudinal changes in sleep patterns with subsequent risk of CVD outcomes.We found that retired Chinese adults with persistent favorable sleep patterns

Figure 1
Figure 1 illustrates the associations of the 5-year changes in sleep patterns with the subsequent risk of incident CVD, CHD, and stroke.Compared with persistent unfavorable sleep patterns, the risk of CVD was lower for those with favorable-unfavorable sleep patterns (adjusted HR [aHR], 0.85; 95% CI, 0.78-0.93),those with unfavorable-favorable sleep patterns (aHR, 0.84; 95% CI, 0.77-0.93),and those with persistent favorable sleep patterns (aHR, 0.80; 95% CI, 0.73-0.87).Changes in sleep patterns were also associated with lower risk of CHD and stroke; participants with persistent

Table 1 .
Characteristics of the Study Participants at the First Follow-Up Survey in 2013 According to Changes in Sleep PatternsFigure 2 shows the joint associations of 5-year changes in sleep patterns and genetic risk of CHD and stroke with the subsequent risk of incident CHD and stroke.We observed a generally monotonic association of decreasing PRS and changes in sleep patterns (from persistent unfavorable to a Body mass index was calculated as weight in kilograms divided by height in meters squared.Figure 1. Associations of Changes in Sleep Patterns With Risk of Incident Cardiovascular Disease (CVD)

Table 2 .
Associations of Changes in Sleep Patterns With Coronary Heart Disease and Stroke, Stratified by the Genetic Risk Group Genetic Variants Included in the CHD PRS eTable 2. Genetic Variants Included in the Stroke PRS eFigure 2. Distributions of CHD PRS and Stroke PRS eTable 3. Associations of CHD PRS and Stroke PRS With Risk of Incident CHD and Stroke eTable 4. Associations of Changes in Sleep Patterns With Risk of Incident CVD According to Age and Sex eTable 5. Associations of Changes in Sleep Patterns With Risk of Incident CVD After Excluding Events Occurred Within the First Year of Follow-Up eTable 6. Associations of Changes in Sleep Patterns With Risk of Incident CVD After Excluding Participants Reporting Very Poor Sleep Quality With Frequent Use of Hypnotics JAMA Network Open | Cardiology Sleep Patterns, Genetic Susceptibility, and Incident Cardiovascular Disease a The Cox regression models were adjusted for age, sex, education level, smoking status, drinking status, regular exercise, body mass index, hypertension, diabetes, hyperlipidemia, and family history of coronary heart disease or stroke (in the corresponding analysis).JAMA NetworkOpen.2024;7(4):e247974.doi:10.1001/jamanetworkopen.2024.7974(Reprinted) April 23, 2024 10/10 Downloaded from jamanetwork.comby guest on 04/25/2024