Abstract
Background
To investigate the association between longitudinal sleep duration trajectories and the risk of cardio-renal-metabolic multimorbidity (CRMM) in a population of postmenopausal women.
Methods
A total of 2,859 postmenopausal women from the China Health and Retirement Longitudinal Study (CHARLS) were included, with 4 repeated measurements of night sleep duration collected from 2011 to 2018. Group-based trajectory modeling was adopted to identify distinct sleep duration trajectory patterns. Multivariable logistic regression was used to examine the associations between sleep trajectories and incident CRMM ascertained between 2018 and 2020. Subgroup analyses and sensitivity analyses were further conducted to validate the robustness of the findings.
Results
Three sleep duration trajectory patterns were identified: normal-stable (33.9% of the cohort), normal-fluctuating decline (52.0%), and short-rapid decline (14.1%). During a median follow-up of 2 years, the primary outcome (incident CRMM) occurred in 108 participants (3.8%). Compared with normal-stable group, those in normal-fluctuating decline and short-rapid decline groups had an increased risk of CRMM (ORs of 2.30 [95%CI, 1.35–3.91] and 3.75 [95%CI, 1.99–7.07], respectively) after adjustment for multiple covariates. Subgroup and sensitivity analyses confirmed the robustness of these associations, which were independent of key baseline characteristics and missing data.
Conclusion
In postmenopausal women, short and declining sleep duration trajectories are strong, independent risk factors for CRMM. Public health and clinical interventions should prioritize maintaining stable, adequate sleep to mitigate multimorbidity risk in this population.
Supplementary Information
The online version contains supplementary material available at https://doi.org/10.1186/s12889-026-28622-4.
Keywords: Cardio-renal-metabolic multimorbidity, Sleep duration trajectory, Postmenopausal women, CHARLS
Introduction
Over the past three decades, the global prevalence of cardiovascular diseases (CVD) has surged dramatically from 311 million in 1990 to 626 million in 2023 [1]. Concurrently, the number of individuals suffering from other chronic conditions such as diabetes and chronic kidney disease has also increased dramatically, reaching 526 million and 674 million respectively [2]. Moreover, with the intensification of global aging, the risk of progression from single cardio-renal-metabolic disease (CRMD) to cardio-renal-metabolic multimorbidity (CRMM, coexistence of ≧ 2 CRMDs) increases. Epidemiological data indicates that 25% of elderly patients are affected by CRMM [3]. Its pathophysiology involves a vicious cycle of insulin resistance, systemic inflammation, and endothelial dysfunction, leading to multi-organ injury [4]. The development of any component disease of CRMD triggers interactions between these conditions, accelerating the progression to CRMM, which results in additive organ damage and amplifies pathological mechanisms, resulting in a distinct clinical phenotype [5]. And compared with single CRMD, those with CRMM experience a higher rate of functional decline, more complex complications and an increased all cause mortality [6]. Consequently, the prevalence and disease burden of CRMM have surpassed the traditional scope of chronic diseases, emerging as an urgent public health issue that requires the establishment of a systematic prevention and control framework [7].
Sex-specific factors significantly shape the presentation and progression of individual CRMD, with disease mechanisms involving hormonal regulation and interconnected pathways including the renin-angiotensin system, oxidative stress, inflammatory responses, vascular pathology, and metabolic dysfunction [8]. It is widely acknowledged that postmenopausal women demonstrate markedly elevated cardiovascular risk profiles characterized by central adiposity, insulin resistance, lipid abnormalities, and impaired endothelial function [9]. Nevertheless, epidemiological and clinical characteristics of CRMM in this demographic remain inadequately investigated, highlighting an important knowledge gap that warrants further scientific exploration.
Adequate nighttime sleep is crucial for health [10], yet 40%–69% of postmenopausal women have sleep disturbances [11]. Aging alters sleep architecture—reducing slow-wave sleep, raising sympathetic activity, and weakening circadian rhythms [12]—and menopausal women also experience sharp drops in estrogen and progesterone, both sleep-regulating [11]. This endocrine shift may amplify age-related sleep risks, making sleep duration especially critical for this group. Insufficient nighttime sleep within this population has been significantly associated with an elevated risk of coronary heart disease, type 2 diabetes and kidney disease [13–15]. Nevertheless, the relationship between sleep duration and CRMM remains unexplored. The existing evidence predominantly relies on single-time sleep assessments, which fail to capture longitudinal sleep patterns and the cumulative impact of sustained sleep deficiency. Given that sleep represents a dynamic and modifiable behavioral factor, there is a clear need to investigate long-term trajectories of nighttime sleep and their association with CRMM, particularly among postmenopausal women.
To address this research gap, the present study utilized data from the China Health and Retirement Longitudinal Study (CHARLS) to identify distinct longitudinal trajectories of nighttime sleep duration among postmenopausal women and examine their association with CRMM risk. Based on repeated sleep duration measurements from 2011 to 2018, group-based trajectory modeling was applied to classify participants according to their long-term sleep patterns. The relationship between these trajectory groups and the incidence of CRMM was then evaluated using multivariable logistic regression, with sensitivity and subgroup analyses conducted to test the robustness of the findings.
Methods
Study design and population
The study used data from the CHARLS, a nationally representative longitudinal survey of Chinese adults aged 45 years and above. This comprehensive database provides detailed information on demographic characteristics, socioeconomic factors, lifestyle, health status and biomarker measurements. CHARLS was launched in 2011 with biennial follow-up assessments, resulting in five completed waves by 2020. To ensure a representative sample, participants were selected using a multi-stage, stratified probability sampling method proportional to population size, covering 150 counties or districts and 450 villages or urban communities across 28 provinces.
Based on the CHARLS data, our study initially identified 4,601 female participants with self-reported menopause in 2011. The following exclusion criteria were applied: (1) participants without complete five-wave follow-up records (n = 636); (2) those with missing nighttime sleep duration data across waves 1 to 4 (n = 414); (3) individuals diagnosed with CRMM at or before wave 4 (n = 440); (4) participants without information of CRMM during wave 1 to 5 (n = 252). Finally, a total of 2,859 participants were included for the analysis of the association between longitudinal nighttime sleep duration trajectories and incident CRMM (Fig. 1).
Fig. 1.
Flow chart of patient selection
Assessment of nighttime sleep duration
Self-reported data on nightly sleep duration were obtained from the CHARLS follow-up questionnaire. Participants reported nightly sleep duration using the question: “In the past month, how many hours of sleep do you usually get per night?” This item is a core metric in sleep epidemiology research, and its reliability and validity have been well established in multiple large-scale cohort studies [16]. We identified trajectories of night-time sleep duration based on repeated measurements taken in 2011, 2013, 2015 and 2018 (wave 1 to 4). Additionally, to ensure analytical stability, only participants with at least four valid sleep measurements across these waves were included, to minimize the potential interference from short-term fluctuations on trajectory modeling.
Outcomes
The endpoint of the study was CRMM, defined as the presence of ≥ 2 CRMD, which included CVD, T2DM, and chronic kidney disease (CKD). The specific diagnostic criteria are as follows: CVD was assessed using a standardized questionnaire that asked participants whether a doctor had ever diagnosed them with heart disease (including angina, myocardial infarction, heart failure, and coronary heart disease). T2DM and CKD were defined based on self-reported physician diagnosis. These methods have been widely applied and validated in the CHARLS [17–18].
Covariates
The covariates in this study included demographic characteristics (age, location, marital status and education level), health behaviors (smoking status, drinking status) and health-related factors (chronic disease including baseline stroke, heart problems, dyslipidemia, diabetes and hypertension), which were collected via questionnaires at baseline. Physical examinations (body mass index: calculated as weight (Kg) divided by the square of height (m), waist circumference, blood pressure) and biochemical measurements (fasting blood glucose, glycated hemoglobin, total cholesterol, triglycerides, high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, serum creatinine, uric acid and C-reactive protein) were collected at baseline.
Statistical analysis
Conformity to normal distribution was evaluated for continuous variables using the Kolmogorov–Smirnov test. Continuous variables were described as mean ± standard deviation (SD) or median (interquartile range [IQR]), whereas categorical variables were expressed as the frequency and percentage. A one-way ANOVA test or a non-parametric Kruskal-Wallis test was used to make comparisons for continuous variables, while the χ2-test or Fisher’s exact test was applied for categorical variables. We addressed missing data using multiple imputation by chained equations. Five imputed datasets were generated, and Rubin’s rules were applied to combine the final analysis results.
Group-based trajectory modeling (GBTM) was performed to identify different sleep duration patterns (2011–2018). Models with 2 to 6 trajectory groups were estimated using linear, quadratic, and cubic functional forms. The optimal model was selected based on the lowest Bayesian Information Criterion (BIC) and Akaike Information Criterion (AIC) values, while ensuring each trajectory group comprised at least 5% of the study population. We determined the optimal number and shape of trajectories using the following criteria: the lowest Bayesian information criterion; an average posterior probability of assignment of at least 0.70 for each group; and the inclusion of at least 5% of participants within each group. Notably, although models with 4, 5 and 6 trajectories had lower AIC and BIC values reflecting better fitting effects, they were discarded due to excessively small subgroups below 5% proportion. Finally, the three-trajectory model meeting all above criteria was determined as the final solution. Detailed model indices are shown in Supplementary Tables 1–3.
Multivariable logistic regression analysis was employed to examine the association between different sleep duration trajectories and CRMM. The results were shown as odds ratio (ORs) and 95% confidence intervals (CIs). Subgroup analysis and interaction tests were performed to investigate whether the association between sleep duration trajectories and CRMM occurrence varied according to the state of the covariates (age, waist, smoking status, drinking status, BMI, DBP, HbA1c and dyslipidemia). Sensitivity analyses were performed in participants with menopausal age ≥ 40 years and in participants with complete-case to confirm the robustness of the results. All statistical analyses were conducted using R software (version 4.4.0), and two-sided P value < 0.05 was considered statistically significant.
Result
Baseline characteristics of enrolled population
In this study, a total of 2,859 participants were enrolled based on the inclusion and exclusion criteria (Fig. 1). The mean age of the study population was 59 years, and most of them were married and lived in rural areas. We identified three distinct trajectories of night sleep duration with quadratic order terms from 2,859 individuals in the 2011 to 2018 waves. The results of the model fitting were as follow: BIC = 46,146.73; AIC = 46,069.27; and entropy value = 0.617 (Table S2). The detailed information of sleep duration trajectories is shown in Fig. 2 and Table S3. A total of 969 (33.9%) participants were characterized as maintaining normal sleep duration throughout the follow-up (named normal-stable group; mean range, 8 h); 1,486 (52.0%) individuals started with normal sleep duration at baseline, but experienced a decrease during the follow-up period (named normal-decline group; mean sleep duration decrease from 6 to 5.5 h); and 404 (14.1%) individuals began with short sleep duration, but experienced a further decrease during follow-up period (named short-rapid decline group; mean sleep duration decrease from 4 to 3 h).
Fig. 2.
Trajectories of night sleep duration from 2011 to 2018
As shown in Table 1, participants in the short-rapid decline trajectory group were slightly older than those in the normal-stable and normal-fluctuating decline groups (P < 0.001). Residential location and educational levels also differed, with more rural residents and lower educational levels in the short-rapid decline group (P = 0.006). No significant differences were observed in marital status, smoking status, drinking status, or prevalence of diabetes and hypertension among groups (P > 0.05). Clinically, the short-rapid decline group had lower BMI and slightly smaller waist circumference, lower diastolic blood pressure, but higher HbA1c and HDL-C level compared with the other groups (P < 0.05). Overall, participants in the short-rapid decline group were slightly older and more rural with lower education but tended to have a healthier metabolic profile compared with the other groups.
Table 1.
Baseline characteristics of participants according to sleep duration trajectory groups
| Variables | Sleep duration trajectory (N = 2859) | |||
|---|---|---|---|---|
| Normal-stable (n = 969) |
Normal-fluctuating decline (n = 1486) | Short-rapid decline (n = 404) | P-value | |
| Age, years | 59.0 (54.0, 64.0) | 59.0 (55.0, 64.0) | 60.5 (56.0, 66.0) | < 0.001 |
| Location of Residential Address | 0.006 | |||
| Urban, n (%) | 314 (32.4) | 524 (35.3) | 109 (27.0) | |
| Rural, n (%) | 655 (67.6) | 962 (64.7) | 295 (73.0) | |
| Married, n (%) | 850 (87.7) | 1290 (86.8) | 340 (84.2) | 0.206 |
| Educational level | < 0.001 | |||
| No formal education, n (%) | 640 (66.1) | 932 (62.7) | 306 (75.7) | |
| Primary school, n (%) | 155 (16.0) | 258 (17.4) | 62 (15.3) | |
| Middle school, n (%) | 120 (12.4) | 202 (13.6) | 26 (6.5) | |
| High school or above, n (%) | 53 (5.5) | 94 (6.3) | 10 (2.5) | |
| Smokers, n (%) | 59 (6.1) | 90 (6.1) | 26 (6.4) | 0.960 |
| Drinkers, n (%) | 117 (12.1) | 179 (12.0) | 53 (13.1) | 0.833 |
| Chronic disease | ||||
| Diabetes, n (%) | 89 (9.2) | 126 (8.5) | 25 (6.2) | 0.187 |
| Hypertension, n (%) | 244 (25.2) | 365 (24.7) | 90 (22.4) | 0.534 |
| Heart disease, n (%) | 65 (6.7) | 120 (8.1) | 39 (9.7) | 0.159 |
| Dyslipidemia, n (%) | 71 (7.4) | 131 (8.9) | 34 (8.5) | 0.403 |
| Stroke, n (%) | 14 (1.4) | 29 (2.0) | 7 (1.7) | 0.644 |
| Chronic lung disease, n (%) | 50 (5.2) | 101 (6.8) | 33 (8.2) | 0.085 |
| Liver disease, n (%) | 21 (2.2) | 40 (2.7) | 17 (4.2) | 0.108 |
| Kidney disease, n (%) | 13 (1.3) | 42 (2.8) | 13 (3.2) | 0.030 |
| Clinical examination | ||||
| BMI, kg/m2 | 23.4 (21.2, 26.4) | 23.3 (21.1, 25.9) | 22.6 (20.2, 25.2) | < 0.001 |
| Waist, cm | 85.0 (78.2, 92.0) | 85.0 (77.8, 91.8) | 84.0 (77.1, 90.2) | 0.151 |
| SBP, mmHg | 129.0 (117.0, 145.0) | 129.0 (116.0, 146.0) | 127 (115.5, 147.0) | 0.995 |
| DBP, mmHg | 75.0 (67.0, 84.0) | 75.0 (68.0, 83.0) | 73.0 (66.0, 82.0) | 0.013 |
| Pulse pressure, mmHg | 71.0 (65.0, 78.0) | 72.0 (66.0, 78.0) | 72.0 (66.0, 79.8) | 0.076 |
| Laboratory parameters | ||||
| Total cholesterol, mg/dL | 196.8 (172.8, 221.9) | 195.8 (171.7, 221.8) | 202.4 (175.7, 227.0) | 0.062 |
| LDL-C, mg/dL | 119.1 (98.5, 141.9) | 119.8 (96.3, 143.0) | 120.8 (101.7, 148.3) | 0.219 |
| HDL-C, mg/dL | 49.5 (41.0, 59.9) | 51.8 (42.5, 61.5) | 53.4 (44.5, 62.6) | 0.001 |
| Triglycerides, mg/dL | 114.2 (80.5, 163.7) | 110.6 (80.5, 158.4) | 108.0 (78.8, 146.0) | 0.136 |
| Creatinine, mg/dL | 0.7 (0.6, 0.8) | 0.7 (0.6, 0.8) | 0.7 (0.6, 0.8) | 0.751 |
| Blood urea nitrogen, mg/dL | 14.8 (12.4, 17.6) | 14.7 (12.3, 17.5) | 15.1 (12.6, 18.2) | 0.402 |
| CRP level, mg/dL | 0.9 (0.5, 1.8) | 0.9 (0.5, 1.9) | 1.1 (0.6, 1.9) | 0.445 |
| FBG level, mg/dL | 101.9 (94.3, 110.6) | 101.2 (94.1, 110.5) | 101.3 (94.0, 110.1) | 0.634 |
| HbA1c, % | 5.1 (4.9, 5.4) | 5.1 (4.9, 5.4) | 5.2 (5.0, 5.5) | 0.044 |
| Serum uric acid, mg/dL | 3.9 (3.3, 4.6) | 3.9 (3.3, 4.5) | 3.8 (3.2, 4.4) | 0.226 |
| Sleep duration, h | ||||
| Year 2011 | 8.0 (7.0, 8.0) | 6.0 (5.0, 7.0) | 4.0 (3.0, 5.0) | < 0.001 |
| Year 2013 | 8.0 (7.0, 8.0) | 6.0 (5.0, 7.0) | 4.0 (3.0, 5.0) | < 0.001 |
| Year 2015 | 8.0 (7.0, 8.5) | 6.0 (5.0, 7.0) | 3.0 (2.4, 4.0) | < 0.001 |
| Year 2018 | 8.0 (7.0, 8.0) | 5.5 (4.5, 6.5) | 3.0 (2.4, 4.0) | < 0.001 |
BMI Body mass index, SBP Systolic blood pressure, DBP Diastolic blood pressure, LDL-C Low density lipoprotein-cholesterol, HDL-C High density lipoprotein-cholesterol, CRP C-reactive protein, FBG Fasting blood glucose, HbA1c Glycated hemoglobin
Longitudinal association between sleep duration trajectories and incident CRMM
During a median follow-up of 2 years, a total of 108 incident CRMM cases were identified, with an overall incidence rate of 3.8%. The incidence rates varied significantly across the sleep trajectory groups (Fig. 3). The normal-stable group had the lowest incidence (2.0%, 19/969), followed by the normal-fluctuating decline group (4.3%, 64/1486), while the short-rapid decline group exhibited the highest incidence (6.2%, 25/404).
Fig. 3.
Comparison of CRMM endpoints among three groups
To evaluate the association between sleep duration trajectories and incident CRMM among postmenopausal women, three logistic regression models were developed (Table 2). According to the model 1 (the crude model), the risk of CRMM increased by 125% (OR = 2.25, 95%CI: 1.34–3.78, P = 0.002) and 230% (OR = 3.30, 95%CI: 1.79–6.06, P < 0.001) in the normal-fluctuating decline group and the short-rapid decline group, respectively, compared with the normal-stable group. After adjusting for sociodemographic factors, lifestyle behaviors and anthropometric measures (Model 2), the normal-fluctuating decline group had a 1.21 times higher odd (OR = 2.21, 95% CI: 1.31–3.74, P = 0.003) of developing CRMM than the normal-stable trajectory group, while the short-rapid decline group had a 2.60 times higher odd (OR = 3.60, 95% CI: 1.94–6.69, P < 0.001). After adjusting for all potential confounders in Model 3, the risk of CRMM remained significantly higher in both the normal-fluctuating decline and short-rapid decline groups compared with the normal-stable trajectory group (ORs of 2.30 [95%CI, 1.35–3.91] and 3.75 [95%CI, 1.99–7.07], respectively). These results indicated that more adverse sleep patterns were associated with progressively higher risks of CRMM in postmenopausal women.
Table 2.
Association between trajectories of sleep duration with the incidence of CRMM
| Model 1 (OR, 95%CI) | P | Model 2 (OR, 95%CI) | P | Model 3 (OR, 95%CI) | P | |
|---|---|---|---|---|---|---|
| Sleep duration trajectory | ||||||
| Normal-stable | Ref | Ref | Ref | |||
| Normal-fluctuating decline | 2.25 (1.34, 3.78) | 0.002 | 2.21 (1.31, 3.74) | 0.003 | 2.30 (1.35, 3.91) | 0.002 |
| Short-rapid decline | 3.30 (1.79, 6.06) | < 0.001 | 3.60 (1.94, 6.69) | < 0.001 | 3.75 (1.99, 7.07) | < 0.001 |
Model 1: unadjusted model
Model 2: adjusted for age, marital status, education level, location, smoking status, drinking status, waist, BMI, DBP, and Pulse pressure
Model 3: adjusted for the variables in Model 2, as well as for HbA1c, triglycerides, LDL-C, uric acid and dyslipidemia
OR Odds Ratio, 95% CI 95% Confidence Interval, BMI Body mass index, DBP Diastolic blood pressure, HbA1c Glycated hemoglobin, LDL-C Low density lipoprotein-cholesterol, CRMM Cardio-renal-metabolic multimorbidity
Subgroup analysis
Subgroup analyses were conducted to assess the consistency of the association between sleep duration trajectories and the risk of CRMM across populations with different baseline characteristics (Table 3). The association patterns observed in the primary analysis remained largely consistent in most subgroups, including those stratified by age, BMI, DBP, and metabolic health markers.
Table 3.
Multivariable-adjusted odds ratios for the association between sleep duration trajectory and CRMM by subgroups
| Case/total | Normal-stable | Normal-fluctuating decline | Short-rapid decline | P for interaction | |
|---|---|---|---|---|---|
| Age, years | 0.250 | ||||
| ≥ 60 | 1345/2859 | Ref | 2.39 (1.12, 5.12) | 1.53 (0.79, 2.98) | |
| <60 | 1514/2859 | Ref | 5.06 (1.81, 14.14) | 3.80 (1.58, 9,12) | |
| BMI, kg/m2 | 0.162 | ||||
| ≥24 | 1214/2859 | Ref | 3.71 (1.70, 8.08) | 5.49 (2.26, 13.37) | |
| <24 | 1645/2859 | Ref | 1.26 (0.57, 2.77) | 2.13 (0.88, 5.15) | |
| DBP, mmHg | 0.121 | ||||
| ≥80 | 1014/2859 | Ref | 2.86 (0.89, 9.17) | 3.59 (1.50, 8.60) | |
| <80 | 1845/2859 | Ref | 3.41 (1.64, 7.07) | 1.62 (0.83, 3.15) | |
| Waist, cm | 0.155 | ||||
| ≥85 | 1415/2859 | Ref | 3.26 (1.63, 6.51) | 5.10 (2.30, 11.28) | |
| <85 | 1444/2859 | Ref | 1.07 (0.44, 2.60) | 1.56 (0.50, 4.91) | |
| Smokers | 0.672 | ||||
| Yes | 215/2859 | Ref | 1.29 (0.23, 7.30) | 3.45 (0.54, 21.95) | |
| No | 2644/2859 | Ref | 2.36 (1.37, 4.08) | 3.27 (1.71, 6.23) | |
| Drinkers | 0.562 | ||||
| Yes | 445/2859 | Ref | 1.04 (0.24, 4.42) | 1.93 (0.38, 9.85) | |
| No | 2414/2859 | Ref | 2.50 (1.43, 4.37) | 3.63 (1.88, 7.01) | |
| Dyslipidemia | 0.478 | ||||
| Yes | 237/2859 | Ref | 1.38 (0,41, 4.59) | 3.54 (0.92, 13.59) | |
| No | 2622/2859 | Ref | 2.44 (1.37, 4.36) | 3.19 (1.60, 6.35) | |
| HbA1c, % | 0.473 | ||||
| ≥6.5 | 83/2859 | Ref | 7.98 (0.88, 72.41) | 9.72 (0.66, 142.65) | |
| <6.5 | 2776/2859 | Ref | 2.03 (1.18, 3.51) | 3.17 (1.69, 5.97) | |
The above models were adjusted for age, marital status, education level, location, smoking status, drinking status, waist, BMI, DBP, Pulse pressure, HbA1c, triglycerides, LDL-C, uric acid and dyslipidemia
The model was not adjusted for the stratification variable in each case
BMI Body mass index, DBP Diastolic blood pressure, HbA1c Glycated hemoglobin, LDL-C Low density lipoprotein-cholesterol, CRMM Cardio-renal-metabolic multimorbidity
To explore menopause-related vulnerability, we used 60 years as a cutoff to roughly distinguish a hormonally unstable phase from a stable phase among postmenopausal women. Although the interaction term did not reach statistical significance (P for interaction = 0.250), the results showed that sleep duration decline was more strongly associated with CRMM risk during the earlier postmenopausal period (OR = 5.06, 95% CI: 1.81–14.14; OR = 3.80, 95% CI: 1.58–9.12), when hormonal fluctuations are more pronounced.
Although adverse sleep trajectories showed stronger associations with CRMM in individuals with higher BMI (≥ 24 kg/m²; Short-rapid decline OR: 5.49 vs. 2.13) or larger waist circumference (≥ 85 cm; Low-rapid decline OR: 5.10 vs. 1.56), none of the subgroups reached statistical significance (all P for interaction > 0.05). This suggests that the increased CRMM risk associated with non-stable sleep trajectories is robust and independent of these baseline metabolic and lifestyle factors.
Sensitivity Analyses
Sensitivity analyses were conducted to assess the robustness of the primary findings, demonstrating consistent results under different analytical conditions (Table 4). The analysis which excluded participants with premature menopause (menopausal age < 40 years, Model I), showed significantly increased risks for both the normal-fluctuating decline (OR = 2.22, 95% CI: 1.30–3.80, P = 0.004) and short-rapid decline groups (OR = 3.59, 95% CI: 1.85–6.97, P < 0.001). Similarly, a complete-case analysis including only participants with complete data for all covariates (Model II), yielded consistent results with ORs of 2.64 (95% CI: 1.37–5.09, P = 0.004) and 3.48 (95% CI: 1.57–7.71, P = 0.002) for the respective groups. These findings confirm that the observed associations are robust and not substantially influenced by menopausal age or missing data.
Table 4.
Relationship between sleep duration trajectory and CRMM in different sensitivity analyses
| Exposure | Model I (OR, 95%CI) | P | Model II (OR, 95%CI) | P |
|---|---|---|---|---|
| Sleep duration trajectory | ||||
| Normal-stable | Ref | Ref | ||
| Normal-fluctuating decline | 2.22 (1.30, 3.80) | 0.004 | 2.64 (1.37, 5.09) | 0.004 |
| Short-rapid decline | 3.59 (1.85, 6.97) | < 0.001 | 3.48 (1.57, 7.71) | 0.002 |
Model I was sensitivity analysis in participants without menopausal age < 40 years. We adjusted age, marital status, education level, location, smoking status, drinking status, waist, BMI, DBP, Pulse pressure, HbA1c, triglycerides, LDL-C, uric acid and dyslipidemia
Model II was sensitivity analysis in participants without missing data on any of the covariates included in the fully adjusted model (Model 3). We adjusted age, marital status, education level, location, smoking status, drinking status, waist, BMI, DBP, Pulse pressure, HbA1c, triglycerides, LDL-C, uric acid and dyslipidemia
OR Odds Ratio, 95% CI 95% Confidence Interval, BMI Body mass index, DBP Diastolic blood pressure, HbA1c Glycated hemoglobin, LDL-C Low density lipoprotein-cholesterol, CRMM Cardio-renal-metabolic multimorbidity
Discussion
The association between longitudinal sleep duration trajectories and incident CRMM among postmenopausal women remains poorly understood. In this longitudinal study, we used the group-based trajectory model and identified three distinct nighttime sleep trajectories-normal-stable (33.9%), normal-fluctuating decline (52.0%), and short-rapid decline (14.1%)-among 2,859 postmenopausal women from the CHARLS study. Notably, the normal-stable group maintained an average sleep duration of about 8 h over the 2011–2018 period, which is consistent with the commonly recommended optimal sleep range [10, 19]. Furthermore, only a small proportion of participants in this cohort reported consistently long sleep duration (sleep duration > 8 h), precluding the identification of a stable long‑sleep trajectory. These findings are consistent with the distinctive sleep characteristics of postmenopausal women [11].
Night sleep duration is recognized as important factors influencing the development and progression of CRMM. A large cohort study by Wang et al. confirmed that insufficient sleep duration increased the risk of CRMM in Chinese and European populations [20]. Similarly, Sabia et al. also showed an association between shorter sleep duration and an increased risk of multimorbidity in Whitehall II cohort study [21]. Most previous studies have focused on the association between sleep duration and individual diseases among postmenopausal women [13–15], whereas its relationship with multimorbidity remains poorly understood. Consistent with evidence emphasizing age- and population-specific interpretations of sleep [22], we identified three trajectories in CHARLS postmenopausal women: a normal-stable group (mean ≈ 8 h), a normal-fluctuating decline group (decreasing from ~ 6 h to < 6 h), and a short-rapid decline group (starting < 5 h and dropping to 3–4 h). Notably, these short and declining patterns were strong, independent predictors of CRMM, underscoring that dynamic sleep changes—beyond static duration—are critical for early risk stratification and intervention in this population.
To our knowledge, this study fills a critical gapby providing the first evidence on the association between longitudinal sleep duration patterns and CRMM within the specific context of postmenopausal women. Most previous studies on the relationship between sleep duration and health outcomes among postmenopausal women have been based on a single measurement [14–15, 23], paying little attention to how sleep duration changes over time, which failed to adequately capture the chronicity of exposure to poor sleep. However, we all know that sleep is not a static trait, but rather a lifestyle behavior that can potentially be modified. The decline of ovarian function in postmenopausal women leads to decreased levels of sex hormones and increased levels of gonadotropins, which could result in the occurrence of sleep disorders [24]. Consequently, the sleep patterns of postmenopausal women evolve over time along with many other physiological changes. A key strength of this study is its ability to characterize trajectories of nightly sleep duration in postmenopausal women based on up to four assessments conducted over an eight-year period, and it showed that different trajectories predicted the associated differential risks of CRMM. This stratification captured dynamic changes of sleep duration, which would enable more accurate discrimination of at-risk individuals than a single observation.
Several pathophysiological mechanisms may be involved in the relationship between nightly sleep duration trajectories and CRMM in this population. Firstly, previous studies have revealed that insomnia with objective short sleep duration is associated with a higher risk of hypertension [25]. The underlying mechanism of this association may involve physiological hyperarousal [26], which can activate the hypothalamic-pituitary-adrenal (HPA) axis and consequently attenuate the nocturnal blood pressure (BP) dipping [27–28]. Recent studies have demonstrated that elevated nocturnal BP is associated with asymptomatic target organ damage, as well as an increased risk of cardiovascular and renal events [29–30]. Multiple mechanisms appear to be involved in the adverse metabolic outcomes of sleep disorders, including impaired cellular insulin sensitivity, a shift in skeletal muscle energy metabolism towards non-glucose oxidation, increased sympathetic nervous system activity and altered gut microbiota composition [31–33]. Notably, even maintaining a healthy diet could not reduce the risk of T2DM among those with habitual short sleep duration [34]. Therefore, it suggests that the sleep duration is a significant factor influencing CRMM.
Several other aspects of our research require attention. Recently, Liao et al. founded that the association between sleep duration and health outcomes varied across different age groups [10]. To explore whether this age-dependent pattern extends to postmenopausal women, we conducted subgroup analyses in the present study. Consistent with their findings, we found that shortened sleep duration was significantly associated with an elevated risk of CRMM across both age subgroups. However, due to the sleep characteristics of postmenopausal women, the long sleep trajectory group was absent in this cohort. Moreover, although the interaction term did not reach statistical significance, sleep duration decline was more strongly associated with CRMM risk during the earlier postmenopausal period. These findings suggest that menopause-related hormonal disruption may amplify the adverse impact of sleep deterioration, particularly in the hormonally unstable phase after menopause, highlighting the need for closer monitoring and early intervention in this vulnerable subgroup. This association merits further investigation in future prospective studies.
Although subgroup analyses revealed numerically stronger associations between adverse sleep trajectories and CRMM among participants with higher BMI (≥ 24 kg/m²) or abdominal obesity, the interaction tests were not statistically significant. Therefore, these findings should be interpreted as exploratory. While insufficient sleep and obesity are known to reciprocally exacerbate each other via pathways involving appetite dysregulation, sleep apnea, and visceral adipose-driven inflammation [35–37], our results suggest that the impact of sleep duration on cardiorenal-metabolic health is robust and largely independent of baseline metabolic status. Future studies with larger sample sizes are needed to confirm whether a synergistic effect exists. Nevertheless, optimizing sleep duration remains a critical, independent target for multimorbidity prevention in postmenopausal women.
Some limitations of our study need to be acknowledged. First, as this study only included Chinese participants, the findings may not be generalizable to other ethnicities. Further studies are necessary to validate these findings in other ethnicities. Second, the effect of sleep quality on CRMM in postmenopausal women was not evaluated, which warrants further investigation. Third, sleep duration was assessed by self-report rather than by objective methods such as actigraphy, accelerometry, or polysomnography, which may introduce recall bias and affect trajectory classification. Future studies using objective sleep measurements may help validate the present findings, as recent trajectory-based sleep research has shown the value of objectively measured sleep-related parameters in longitudinal risk assessment [38]. Finally, the relatively short follow-up period in this study limits the assessment of long-term associations between sleep trajectories and CRMM progression.
Conclusion
This study provides new evidence of the relationship between night sleep duration trajectories and incident CRMM among postmenopausal women. This prospective cohort study demonstrates that specific longitudinal sleep duration trajectories—particularly short and decline patterns—serve as strong, independent predictors for the development of CRMM among postmenopausal women. Public health and clinical interventions should prioritize maintaining stable, adequate sleep to mitigate multimorbidity risk in this population.
Supplementary Information
Acknowledgements
We extend our appreciation to all participants and investigators involved in the CHARLS study.
Abbreviations
- CRMM
Cardio-renal-metabolic multimorbidity
- CHARLS
The China Health and Retirement Longitudinal Study
- CVD
Cardiovascular diseases
- CRMD
Cardio-renal-metabolic disease
- T2DM
Type 2 diabetes mellitus
- HbA1c
Glycated hemoglobin
- CKD
Chronic kidney disease
- SD
Standard deviation
- OR
Odds ratio
- CI
Confidence intervals
- GBTM
Group-based trajectory modeling
- BIC
Bayesian Information Criterion
- AIC
Akaike Information Criterion
- BMI
Body mass index
- DBP
Diastolic blood pressure
- BP
Blood pressure
- HDL-C
High-density lipoprotein cholesterol
Authors’ contributions
Junwei Xu: Conceptualization, Formal analysis, Writing – original draft; Minchun Zhang: Data curation, Formal analysis, Visualization, Writing – original draft; Rongzhi Liu: Methodology, Data curation, Writing – original draft; Piaorong Zeng: Data curation, Formal analysis, Writing – review; Shuting Liao: Visualization, Software, Writing – review; Zixiang Lin: Methodology, Software, Writing – review; Bihan Huang: Data curation, Writing – review; Yixin Yao: Visualization, Writing – review; Kuihua Xie: Project administration, Writing – review; Shaoyuan Chen: Supervision, Writing – review & editing; Ling Long: Funding acquisition, Project administration, Writing – review & editing; Yuxi Huang: Supervision, Project administration, Writing – review & editing.
Funding
The present study received support from the Major Science and Technology Project of the Nanshan District Health System in Shenzhen City (NSZD2025009). National Natural Science Foundation of China (82501034), the China Postdoctoral Science Foundation (2025M772393), and the Natural Science Foundation of Shenzhen (JCYJ20250604175753070).
Data availability
More information regarding obtaining data for research use can be found at the CHARLS database (https://charls.pku.edu.cn).
Declarations
Ethics approval and consent to participate
Not applicable. The CHARLS study was conducted in accordance with the Declaration of Helsinki and approved by the Institutional Review Board of Peking University (Approval No. IRB00001052-11015). All research procedures were conducted in accordance with the ethical guidelines of the Declaration of Helsinki. Written informed consent was obtained from all participants.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Junwei Xu, Rongzhi Liu and Minchun Zhang contributed equally to this work.
Contributor Information
Shaoyuan Chen, Email: moter2002@qq.com.
Ling Long, Email: longl3@mail.sysu.edu.cn.
Yuxi Huang, Email: huangyx298@alumni.sysu.edu.cn.
References
- 1.Global, Regional, and National Burden of Cardiovascular Diseases and Risk Factors in 204 Countries and Territories, 1990–2023. J Am Coll Cardiol. 2025. 10.1016/j.jacc.2025.08.015. [DOI] [PubMed]
- 2.Global incidence,prevalence, years lived with disability (YLDs), disability-adjusted life-years (DALYs), and healthy life expectancy (HALE) for 371 diseases and injuries in 204 countries and territories and 811 subnational locations, 1990–2021: a systematic analysis for the Global Burden of Disease Study 2021. Lancet. 2024;403(10440):2133–61. 10.1016/s0140-6736(24)00757-8. [DOI] [PMC free article] [PubMed]
- 3.Hua R, Yang A, Gao L, Chow E, Ji W, Cheung YT. Social determinants of health and progression to cardio-renal-metabolic multimorbidity and mortality in people with prediabetes: A prospective cohort study. Diabetes Obes Metab. 2025. 10.1111/dom.70067. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Ndumele CE, Rangaswami J, Chow SL, Neeland IJ, Tuttle KR, Khan SS, et al. Cardiovascular-Kidney-Metabolic Health: A Presidential Advisory From the American Heart Association. Circulation. 2023;148(20):1606–35. 10.1161/cir.0000000000001184. [DOI] [PubMed] [Google Scholar]
- 5.Ostrominski JW, Arnold SV, Butler J, Fonarow GC, Hirsch JS, Palli SR, et al. Prevalence and Overlap of Cardiac, Renal, and Metabolic Conditions in US Adults, 1999–2020. JAMA Cardiol. 2023;8(11):1050–60. 10.1001/jamacardio.2023.3241. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Svensson T, Saito E, Svensson AK, Melander O, Orho-Melander M, Mimura M, et al. Association of Sleep Duration With All- and Major-Cause Mortality Among Adults in Japan, China, Singapore, and Korea. JAMA Netw Open. 2021;4(9):e2122837. 10.1001/jamanetworkopen.2021.22837. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Marassi M, Fadini GP. The cardio-renal-metabolic connection: a review of the evidence. Cardiovasc Diabetol. 2023;22(1):195. 10.1186/s12933-023-01937-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Guldan M, Unlu S, Abdel-Rahman SM, Ozbek L, Gaipov A, Covic A, et al. Understanding the Role of Sex Hormones in Cardiovascular Kidney Metabolic Syndrome: Toward Personalized Therapeutic Approaches. J Clin Med. 2024;13(15). 10.3390/jcm13154354. [DOI] [PMC free article] [PubMed]
- 9.Nappi RE, Chedraui P, Lambrinoudaki I, Simoncini T. Menopause: a cardiometabolic transition. Lancet Diabetes Endocrinol. 2022;10(6):442–56. 10.1016/s2213-8587(22)00076-6. [DOI] [PubMed] [Google Scholar]
- 10.Liao J, Shi Y, Li Y, Han D. Impact of age on sleep duration and health outcomes: Evidence from four large cohort studies. Sleep Med. 2025;129:140–7. 10.1016/j.sleep.2025.02.024. [DOI] [PubMed] [Google Scholar]
- 11.Maki PM, Panay N, Simon JA. Sleep disturbance associated with the menopause. Menopause. 2024;31(8):724–33. 10.1097/gme.0000000000002386. [DOI] [PubMed] [Google Scholar]
- 12.Scullin MK, Bliwise DL. Sleep, cognition, and normal aging: integrating a half century of multidisciplinary research. Perspect Psychol Sci. 2015;10(1):97–137. 10.1177/1745691614556680. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Sands-Lincoln M, Loucks EB, Lu B, Carskadon MA, Sharkey K, Stefanick ML, et al. Sleep duration, insomnia, and coronary heart disease among postmenopausal women in the Women’s Health Initiative. J Womens Health (Larchmt). 2013;22(6):477–86. 10.1089/jwh.2012.3918. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Shadyab AH, Kritz-Silverstein D, Laughlin GA, Wooten WJ, Barrett-Connor E, Araneta MR. Ethnic-specific associations of sleep duration and daytime napping with prevalent type 2 diabetes in postmenopausal women. Sleep Med. 2015;16(2):243–9. 10.1016/j.sleep.2014.11.010. [DOI] [PubMed] [Google Scholar]
- 15.Tong J, Li C, Hu J, Teng Y, Zhou Y, Tao M. Association of sleep characteristics with renal function in menopausal women without recognized chronic kidney disease. Front Psychiatry. 2022;13:1024245. 10.3389/fpsyt.2022.1024245. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Li M, Wang N, Dupre ME. Association between the self-reported duration and quality of sleep and cognitive function among middle-aged and older adults in China. J Affect Disord. 2022;304:20–7. 10.1016/j.jad.2022.02.039. [DOI] [PubMed] [Google Scholar]
- 17.Liu J, Liu Z, Zhou Y, Liu Y, Song W, Wang N, et al. The joint effect of triglyceride-glucose related indices and depression on cardio-renal-metabolic multimorbidity among middle-aged and older Chinese adults. J Affect Disord. 2025;382549–57. 10.1016/j.jad.2025.04.072. [DOI] [PubMed]
- 18.Zhang N, Liu X, Wang L, Zhang Y, Xiang Y, Cai J, et al. Lifestyle factors and their relative contributions to longitudinal progression of cardio-renal-metabolic multimorbidity: a prospective cohort study. Cardiovasc Diabetol. 2024;23(1):265. 10.1186/s12933-024-02347-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.O’toole CK, Song Z, Anagnostakis F, Yang Z, Tian YE, Duggan MR, et al. Sleep chart of biological ageing clocks in middle and late life. Nature. 2026. 10.1038/s41586-026-10524-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Wang H, Tang S. Exploring the relationship between sleep patterns and cardio-renal-metabolic multimorbidity in Chinese and European populations. J Health Popul Nutr. 2026;45(1). 10.1186/s41043-026-01313-x. [DOI] [PMC free article] [PubMed]
- 21.Sabia S, Dugravot A, Léger D, Ben Hassen C, Kivimaki M, Singh-Manoux A. Association of sleep duration at age 50, 60, and 70 years with risk of multimorbidity in the UK: 25-year follow-up of the Whitehall II cohort study. PLoS Med. 2022;19(10):e1004109. 10.1371/journal.pmed.1004109. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Liao J, Shi Y, Li Y, Han D. Cognitive-friendly sleep patterns in older Adults:A cross-regional HCAP study. Sleep Med. 2025;134(106684). 10.1016/j.sleep.2025.106684. [DOI] [PubMed]
- 23.Rissling MB, Gray KE, Ulmer CS, Martin JL, Zaslavsky O, Gray SL, et al. Sleep Disturbance, Diabetes, and Cardiovascular Disease in Postmenopausal Veteran Women. Gerontologist. 2016;56(Suppl 1):S54–66. 10.1093/geront/gnv668. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Baker FC, De Zambotti M, Colrain IM, Bei B. Sleep problems during the menopausal transition: prevalence, impact, and management challenges. Nat Sci Sleep. 2018;10:73–95. 10.2147/nss.S125807. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Lunyera J, Park YM, Ward JB, Gaston SA, Bhavsar NA, Muntner P, et al. A prospective study of multiple sleep dimensions and hypertension risk among white, black and Hispanic/Latina women: findings from the Sister Study. J Hypertens. 2021;39(11):2210–9. 10.1097/hjh.0000000000002929. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Li Y, Vgontzas AN, Fernandez-Mendoza J, Bixler EO, Sun Y, Zhou J, et al. Insomnia with physiological hyperarousal is associated with hypertension. Hypertension. 2015;65(3):644–50. 10.1161/hypertensionaha.114.04604. [DOI] [PubMed] [Google Scholar]
- 27.Mohammadi H, Rezaei M, Amiri SM, Rahimi Z, Mansouri K, Khazaie H. Sleep Architecture and Hypothalamic-Pituitary-Adrenal Activity in Paradoxical and Psychophysiological Insomnia. Basic Clin Neurosci. 2018;9(6):397–407. 10.32598/bcn.9.6.397. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Xu Y, Barnes VA, Harris RA, Altvater M, Williams C, Norland K, et al. Sleep Variability, Sleep Irregularity, and Nighttime Blood Pressure Dipping. Hypertension. 2023;80(12):2621–6. 10.1161/hypertensionaha.123.21497. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Kario K, Hoshide S, Mizuno H, Kabutoya T, Nishizawa M, Yoshida T, et al. Nighttime Blood Pressure Phenotype and Cardiovascular Prognosis: Practitioner-Based Nationwide JAMP Study. Circulation. 2020;142(19):1810–20. 10.1161/circulationaha.120.049730. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Park CH, Jhee JH, Chun KH, Seo J, Lee CJ, Park SH, et al. Nocturnal systolic blood pressure dipping and progression of chronic kidney disease. Hypertens Res. 2024;47(1):215–24. 10.1038/s41440-023-01368-x. [DOI] [PubMed] [Google Scholar]
- 31.Arnoriaga-Rodríguez M, Leal Y, Mayneris-Perxachs J, Pérez-Brocal V, Moya A, Ricart W, et al. Gut Microbiota Composition and Functionality Are Associated With REM Sleep Duration and Continuous Glucose Levels. J Clin Endocrinol Metab. 2023;108(11):2931–9. 10.1210/clinem/dgad258. [DOI] [PubMed] [Google Scholar]
- 32.Cedernaes J, Schönke M, Westholm JO, Mi J, Chibalin A, Voisin S, et al. Acute sleep loss results in tissue-specific alterations in genome-wide DNA methylation state and metabolic fuel utilization in humans. Sci Adv. 2018;4(8):eaar8590. 10.1126/sciadv.aar8590. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Reutrakul S, Van Cauter E. Sleep influences on obesity, insulin resistance, and risk of type 2 diabetes. Metabolism. 2018;84:56–66. 10.1016/j.metabol.2018.02.010. [DOI] [PubMed] [Google Scholar]
- 34.Nôga DA, Meth E, Pacheco AP, Tan X, Cedernaes J, Van Egmond LT, et al. Habitual Short Sleep Duration, Diet, and Development of Type 2 Diabetes in Adults. JAMA Netw Open. 2024;7(3):e241147. 10.1001/jamanetworkopen.2024.1147. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Kline CE, Chasens ER, Bizhanova Z, Sereika SM, Buysse DJ, Imes CC, et al. The association between sleep health and weight change during a 12-month behavioral weight loss intervention. Int J Obes (Lond). 2021;45(3):639–49. 10.1038/s41366-020-00728-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Naufel MF, Frange C, Andersen ML, Girão M, Tufik S, Beraldi Ribeiro E, et al. Association between obesity and sleep disorders in postmenopausal women. Menopause. 2018;25(2):139–44. 10.1097/gme.0000000000000962. [DOI] [PubMed] [Google Scholar]
- 37.Tasali E, Wroblewski K, Kahn E, Kilkus J, Schoeller DA. Effect of Sleep Extension on Objectively Assessed Energy Intake Among Adults With Overweight in Real-life Settings: A Randomized Clinical Trial. JAMA Intern Med. 2022;182(4):365–74. 10.1001/jamainternmed.2021.8098. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Liao J, Shi Y, Zhang B, Lv M, Li Y, Han D. Blood Pressure Trajectories and Cardiovascular Risk in Obstructive Sleep Apnea: A Dual-Cohort Analysis. J Sleep Res. 2026;e70366. 10.1111/jsr.70366. [DOI] [PubMed]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
More information regarding obtaining data for research use can be found at the CHARLS database (https://charls.pku.edu.cn).