Abstract
Background
Physical, psychological, and cognitive multimorbidity (PPC-MM), defined as the coexistence of conditions in at least two domains of physical conditions, depressive symptoms, and cognitive impairment, poses a major challenge to healthy aging. However, its potential risk factors remain unclear. As a key component of health, sleep habits are closely associated with various chronic conditions. This study aimed to investigate the longitudinal associations between sleep habits (night-time sleep duration, nap duration, and sleep quality) and the occurrence and progression of PPC-MM.
Methods
We enrolled 7586 participants aged ≥45 years from the China Health and Retirement Longitudinal Study (CHARLS). All participants had no PPC-MM at baseline and provided self-reported data on sleep habits. The relationships between sleep habits and PPC-MM were evaluated using Cox proportional hazards regression and multi-state models.
Results
Over a 7-year follow-up, 3239 participants (42.6%) developed PPC-MM. Short night-time sleep (<7 h/night) and poor sleep quality (>1 d/week) were related to a higher risk of incident PPC-MM, with the strongest associations observed for sleeping <5 h/night (HR = 1.52, 95% CI: 1.37–1.70) and restless sleep on 5−7 d/week (HR = 1.47, 95% CI: 1.32–1.62). PPC-MM risk showed a U-shaped relationship with night-time sleep duration. In addition, joint associations of sleep duration and sleep quality on PPC-MM were observed. Multi-state model revealed very short night-time sleep (<5 h/night) and poor sleep quality (>1 d/week) were related to transitions from baseline to dual-domain comorbidity, whereas frequent restless sleep (5−7 d/week: HR = 1.40, 95% CI: 1.06–1.86) were associated with direct transition to triple-domain multimorbidity.
Conclusion
Short night-time sleep duration and poor sleep quality were associated with an elevated risk of PPC-MM incidence and progression, with evidence of a joint adverse association. Such disturbances may serve as early indicators of PPC-MM risk. These findings suggest that sleep habits may serve as potential risk markers of multidomain health deterioration, thereby informing strategies to support healthy aging in middle-aged and older adults.
Keywords: Multimorbidities, Physical-psychological-cognitive, Sleep duration, Sleep quality, Healthy aging
1. Introduction
With the accelerating pace of global population aging, multimorbidity has become an increasingly prominent public health concern [1]. Multimorbidity affects an estimated 37.2% of adults globally and is especially common in those aged 60 years and above [2]. Multimorbidity is commonly defined as the coexistence of two or more long-term health conditions, encompassing chronic physical diseases as well as depressive symptoms and cognitive impairment [3]. However, the conceptual scope and operational definition of multimorbidity vary substantially across studies [4,5]. Most research has focused primarily on chronic physical conditions, leaving the cross-domain coexistence of physical diseases, depressive symptoms, and cognitive impairment, and its potential health consequences, relatively underexplored. To capture this multidimensional pattern, previous studies have used the framework of physical, psychological, and cognitive multimorbidity (PPC-MM), defined as the coexistence of conditions across at least two of these three domains [6,7]. Within this framework, the psychological domain has commonly been represented by depressive symptoms, given their relevance to late-life health and their use in previous PPC-MM studies. This domain-based definition was intended to capture the coexistence and progression of multidimensional health problems among middle-aged and older adults. Therefore, identifying potential risk factors associated with the development and progression of PPC-MM may provide evidence for health risk identification and integrated management in aging populations.
Sleep is an essential component of health, encompassing multiple dimensions, including sleep duration, sleep quality, and sleep continuity or efficiency [8]. In middle-aged and older adults, night-time sleep duration, sleep quality, and daytime napping represent distinct but complementary aspects of sleep behavior. Sleep duration reflects the quantitative aspect of sleep, whereas sleep quality captures individuals’ subjective perception of their sleep experience [9]. Night-time sleep is commonly assessed as the principal sleep episode, whereas napping reflects a distinct daytime sleep behavior that may be associated with underlying health status and provide complementary information on overall sleep habits [10,11]. Accumulating evidence suggests that adverse sleep habits are associated with multiple health outcomes in middle-aged and older adults. Short sleep duration and poor sleep quality have been linked to cardiometabolic risk factors such as obesity, hypertension, and type 2 diabetes [12], while evidence from a systematic review reported associations between long napping and several cardiovascular risk factors and mortality [13]. In addition, a 4-year longitudinal study in Korean adults suggested that poor sleep quality and short sleep duration were independently associated with an increased risk of depression [14]. A multicenter study from Nordic countries also revealed that disturbed sleep and abnormal sleep duration were associated with cognitive function [15]. These findings may be partly attributable to common pathophysiological pathways involved in sleep disturbances. Sleep disturbance has been associated with systemic low-grade inflammation and dysregulation of the hypothalamic-pituitary-adrenal (HPA) axis [16,17]. Inflammatory activation and stress-system dysregulation are not confined to a single disease process, but may contribute concurrently to the development and progression of chronic physical conditions, depressive symptoms, and cognitive decline [[18], [19], [20]]. Accordingly, sleep-related indicators may be linked to the co-occurrence of impairments across physical, depressive, and cognitive domains through these pathways. However, current evidence largely focuses on sleep behaviors in relation to single health outcomes, with limited attention to multidomain multimorbidity and its progression over time.
In our study, using data from the China Health and Retirement Longitudinal Study (CHARLS), we examined the relationships between sleep habits, including night-time sleep duration, nap duration, and sleep quality, with incident PPC-MM and its progression among middle-aged and older Chinese adults. We hypothesized that adverse sleep habits, particularly shorter night-time sleep duration and poorer sleep quality, would be associated with a higher risk of incident PPC-MM. We also examined whether these sleep habits were associated with PPC-MM progression.
2. Methods
2.1. Study population
Data were obtained from the CHARLS, a population-based cohort of Chinese adults. Participants were selected using a multistage stratified sampling design with probability proportional to size. Information on sociodemographic characteristics, health conditions, and lifestyle factors was obtained through standardized face-to-face interviews using structured questionnaires [21]. We analyzed data from four survey waves conducted in 2011, 2013, 2015, and 2018. The 2011 survey served as the baseline, and the 2013, 2015, and 2018 waves were treated as follow-up assessments, corresponding to approximately 2, 4, and 7 years after baseline. In total, 17,708 participants were recruited at baseline. Participants were excluded if they were younger than 45 years (n = 777), had PPC-MM at baseline (n = 7068), lacked information on sleep habits, covariates, or had abnormal BMI values (<10 or >60 kg/m²) (n = 1721), or lacked follow-up information on PPC-MM (n = 556). Finally, 7586 participants remained for the final analysis (Fig. S1a).
2.2. Sleep habits
Sleep habits were assessed at baseline in 2011, including night-time sleep duration, nap duration, and sleep quality. Night-time sleep duration was collected using the question, “During the past month, how many hours of actual sleep did you get at night?” and grouped into <5, 5−7, 7−9 (reference), and ≥9 h/night. Nap duration was self-reported based on the question, “During the past month, how long did you take a nap after lunch?” and classified as 0, 0−30 (reference), 30−60, 60−90, and >90 min/day. Sleep quality was classified as <1 (reference), 1−2, 3−4, and 5−7 d/week based on the weekly frequency of restless sleep.
2.3. Physical, psychological, and cognitive multimorbidity
PPC-MM was defined as the coexistence of conditions in at least two of three domains: physical conditions, depressive symptoms, and cognitive impairment. The operational definition of PPC-MM and the assessment of its component domains were based on previous studies [22,23], and detailed criteria are presented in Supplementary Table S1.
Physical conditions were defined as the presence of at least one self-reported physician-diagnosed chronic disease, including hypertension, diabetes, chronic lung disease, heart disease, stroke, cancer, or arthritis.
Depressive symptoms were assessed using the 10-item Center for Epidemiologic Studies Depression Scale (CESD-10) [24], which has been widely used and shown to have satisfactory reliability and validity among middle-aged and older Chinese adults [24]. Scores range from 0 to 30, and participants with scores ≥10 were classified as having depressive symptoms, which were used to operationalize the psychological domain of PPC-MM in the present study.
Cognitive function was assessed across three domains, including episodic memory (immediate and delayed recall), executive function (serial 7 s subtraction task), and temporal orientation. Participants were stratified by age (45–49, 50–54, 55–59, 60–64, 65–69, 70–74, 75−79, and ≥80 years). For each cognitive domain, the age-group-specific mean and standard deviation were calculated. Cognitive impairment was defined as a score more than 1.5 standard deviations below the corresponding age-group-specific mean in at least one of the three domains [25].
PPC-MM was further classified into four patterns: physical-psychological multimorbidity (P1P2), physical-cognitive multimorbidity (P1C), psychological-cognitive multimorbidity (P2C), and physical-psychological-cognitive multimorbidity (P1P2C).
2.4. Covariates assessment
Covariates comprised sociodemographic characteristics, lifestyle factors, and BMI. The sociodemographic variables covered age (45−54, 55−64, 65−74, and ≥75 years), sex (male/female), marital status (married/unmarried), education level (primary school or below/middle school/high school or above), and residence (urban/rural). Lifestyle-related variables included drinking status (yes/no) and smoking status (yes/no). BMI was derived as weight (kg) divided by height squared (m²) and classified into four groups based on Chinese standards: underweight (<18.5 kg/m²), normal weight (18.5−23.9 kg/m²), overweight (24.0–27.9 kg/m²), and obesity (≥28.0 kg/m²).
2.5. Statistical analysis
Baseline characteristics were described according to baseline sleep habits and the occurrence of PPC-MM during follow-up. Categorical variables were reported as frequencies and percentages, and group differences were evaluated using the chi-square test. Cox proportional hazards regression was employed to estimate the relationship between sleep habits and PPC-MM. Three models were fitted. Model 1 was unadjusted, Model 2 adjusted for age and sex, and Model 3 further adjusted for education level, smoking, drinking, marital status, residence, and BMI. Results are presented as hazard ratios (HRs) with 95% confidence intervals (CIs). All sleep habits and covariates were assessed at baseline. Event time was defined as the interval from the 2011 baseline to the follow-up wave at which PPC-MM was first observed, and participants without PPC-MM were censored at their last available follow-up wave.
The dose-response relationships between sleep duration and PPC-MM risk were investigated using restricted cubic spline (RCS) analysis. We evaluated the combined associations of sleep duration and quality with PPC-MM risk. These analyses were based on the Model 3.
To further investigate the relationships between sleep habits and transitions in PPC-MM status, we constructed a multi-state model (Fig. S1b). The multi-state model extends the competing risks framework and enables the evaluation of transitions between different disease states. In this analysis, death was not treated as a direct outcome, but rather as a competing event. For the multi-state analyses, PPC-MM status was updated at each available survey wave from 2011 to 2018, and transition time was defined by the wave in which a new PPC-MM state was first observed. Transitions from baseline were measured from 2011, whereas transitions from dual-domain to triple-domain PPC-MM were measured from the first wave in which dual-domain PPC-MM was observed. Baseline to dual-domain comorbidity, baseline to physical-psychological-cognitive multimorbidity, and dual-domain comorbidity to physical-psychological-cognitive multimorbidity were identified as the three primary transitions. We further identified seven transitions by classifying dual-domain multimorbidity into specific patterns. Schoenfeld residuals were used to assess the proportional hazards assumption, and no infraction was identified.
Several sensitivity analyses included: (1) missing data in sleep habits and covariates were addressed using multiple imputation with chained equations; (2) participants with physical, psychological, or cognitive disorders at baseline were excluded; (3) the CESD-10 score was recalculated after excluding the sleep-related item, and (4) subgroup analyses were conducted stratified by sex and age (<65 vs. ≥65 years).
The main analyses examined the associations of baseline sleep habits with incident PPC-MM. Multi-state transition analyses, RCS, and joint analyses were conducted to further characterize these associations and PPC-MM progression. To account for multiple testing, p-values from all Cox regression models were adjusted using the Benjamini-Hochberg false discovery rate (FDR) method, with statistical significance determined based on FDR p-values. R statistical software (version 4.5.0) was used for all analyses.
3. Results
3.1. Baseline characteristics of participants
Baseline characteristics stratified by incident PPC-MM status are summarized in Table 1. The mean (±SD) age of participants was 58.1 (±9.2) years, and 4069 participants (53.6%) were male. During follow-up from the 2011 baseline through the 2013, 2015, and 2018 survey waves, 3239 participants (42.6%) developed PPC-MM. Participants who developed PPC-MM were generally older, more frequently female, had lower educational attainment, and were more likely to live in rural areas than those who did not develop PPC-MM. They also differed significantly in marital status, smoking status, drinking status, and BMI (all p < 0.05). Baseline characteristics categorized by sleep duration and sleep quality are presented in Supplementary Tables S2–S3.
Table 1.
Baseline characteristics of participants according to the status of PPC-MM (n = 7586).
| Total (N = 7586) | No PPC-MM (N = 4347) | PPC-MM (N = 3239) | p-value | |
|---|---|---|---|---|
| Age groups (years) | <0.001 | |||
| 45−54 | 2988(39.4) | 1761(40.5) | 1227(37.9) | |
| 55−64 | 2806(37.0) | 1530(35.2) | 1276(39.4) | |
| 65−74 | 1356(17.9) | 776(17.9) | 580(17.9) | |
| ≥75 | 436(5.7) | 280(6.4) | 156(4.8) | |
| Sex | <0.001 | |||
| Male | 4069(53.6) | 2610(60.0) | 1459(45.0) | |
| Female | 3517(46.4) | 1737(40.0) | 1780(55.0) | |
| Marital status | 0.019 | |||
| Non-married | 1073(14.1) | 579(13.3) | 494(15.3) | |
| Married | 6513(85.9) | 3768(86.7) | 2745(84.7) | |
| Education level | <0.001 | |||
| Primary school or below | 4681(61.7) | 2333(53.7) | 2348(72.5) | |
| Middle school | 1834(24.2) | 1223(28.1) | 611(18.9) | |
| High school or above | 1071(14.1) | 791(18.2) | 280(8.6) | |
| Smoking status | <0.001 | |||
| No | 4334(57.1) | 2351(54.1) | 1983(61.2) | |
| Yes | 3252(42.9) | 1988(45.9) | 1254(38.8) | |
| Drinking status | <0.001 | |||
| No | 4429(58.4) | 2399(55.2) | 2030(62.7) | |
| Yes | 3157(41.6) | 1948(44.8) | 1209(37.3) | |
| Residence | <0.001 | |||
| Rural | 4603(60.7) | 2428(55.9) | 2175(67.2) | |
| Urban | 2983(39.3) | 1919(44.1) | 1064(32.8) | |
| BMI (kg/m2, Chinese standard) | 0.028 | |||
| <18.5 | 415(5.5) | 228(5.2) | 187(5.8) | |
| 18.5−23.9 | 4051(53.4) | 2381(54.8) | 1670(51.6) | |
| 24−27.9 | 2265(29.9) | 1274(29.3) | 991(30.6) | |
| ≥28 | 855(11.3) | 464(10.7) | 391(12.1) |
BMI: body mass index; PPC-MM: physical, psychological, and cognitive.
3.2. Association of sleep habits and PPC-MM
Longitudinal associations between baseline sleep habits and incident PPC-MM across the 2013, 2015, and 2018 follow-up waves are summarized in Table 2. In fully adjusted analyses, individuals who slept <5 h/night exhibited a higher risk of PPC-MM (HR = 1.52, 95% CI: 1.37–1.70), followed by those sleeping 5−7 h/night (HR = 1.12, 95% CI: 1.04–1.22). No significant association was found for ≥9 h/night.
Table 2.
The association between sleep habits and risk of PPC-MM during follow-up (n = 7586).
| Events/n | Model 1a | Model 2b | Model 3c | |
|---|---|---|---|---|
| HR (95% CI) | HR (95% CI) | HR (95% CI) | ||
| Night-time sleep duration(h/night) | ||||
| <5 | 457/815 | 1.69(1.52,1.88)*** | 1.63(1.46,1.81)*** | 1.52(1.37,1.70)*** |
| 5−7 | 1124/2637 | 1.11(1.03,1.20)* | 1.11(1.02,1.20)* | 1.12(1.04,1.22)** |
| 7−9 | 1385/3508 | Ref | Ref | Ref |
| ≥ 9 | 273/626 | 1.14(1.01,1.30)* | 1.13(0.99,1.29) | 1.06(0.93,1.20) |
| Nap duration(min/d) | ||||
| 0 | 1494/3377 | 1.03(0.94,1.14) | 1.01(0.92,1.11) | 0.94(0.86,1.04) |
| 0−30 | 563/1319 | Ref | Ref | Ref |
| 30−60 | 709/1762 | 0.93(0.83,1.04) | 0.96(0.86,1.08) | 0.92(0.83,1.04) |
| 60−90 | 107/271 | 0.89(0.73,1.10) | 0.94(0.77,1.16) | 0.90(0.74,1.11) |
| >90 | 366/857 | 0.98(0.86,1.12) | 1.03(0.90,1.17) | 0.97(0.85,1.11) |
| Sleep quality (restless days/week) | ||||
| <1 | 1730/4560 | Ref | Ref | Ref |
| 1−2 | 625/1293 | 1.39(1.27,1.52)*** | 1.33(1.22,1.46)*** | 1.35(1.23,1.48)*** |
| 3−4 | 408/815 | 1.46(1.32,1.63)*** | 1.37(1.23,1.53)*** | 1.37(1.23,1.53)*** |
| 5−7 | 476/918 | 1.53(1.39,1.70)*** | 1.47(1.33,1.63)*** | 1.47(1.32,1.62)*** |
HR, hazard ratio; CI, confidence interval.
Unadjusted model.
Adjusted for age, sex.
Adjusted for age, sex, marital status, educational level, residence, smoking status, drinking status, and BMI.
FDR p < 0.05.
FDR p < 0.01.
FDR p < 0.001.
Participants with poor sleep quality were shown to have an increased risk of PPC-MM. The corresponding HRs (95% CIs) for 1−2, 3−4, and 5−7 days/week were 1.35 (1.23–1.48), 1.37 (1.23–1.53), and 1.47 (1.32–1.62) in comparison to those who reported restless sleep on <1 day/week. However, nap duration was not significantly associated with PPC-MM risk (p > 0.05).
A U-shaped association was found between night-time sleep duration and PPC-MM risk based on RCS (p for overall < 0.001; p for nonlinear < 0.001). The lowest risk was observed at approximately 7−8 h/night, with risk increasing at both shorter and longer sleep durations. In contrast, nap duration showed no significant overall or nonlinear association with PPC-MM risk (Fig. 1).
Fig. 1.
Dose-response associations of night-time sleep duration (A) and nap duration (B) with PPC-MM risk (n = 7586).
3.3. Joint association of sleep habits with PPC-MM
The joint associations between sleep habits and PPC-MM risk are shown in Fig. 2. Relative to the reference group (7−9 h/night and <1 day/week of restless sleep), most other groups were associated with an increased PPC-MM risk. In the fully adjusted model, among participants sleeping <5 h/night, the HRs (95% CIs) were 2.00 (1.57–2.55), 1.85 (1.49–2.29), and 1.59 (1.37–1.85) for those reporting restless sleep on 1−2, 3−4, and 5−7 d/week, respectively. A similar pattern was observed for participants sleeping 5−7 h/night. Among participants sleeping ≥9 h/night, only those reporting restless sleep on 1−2 d/week showed an increased risk of PPC-MM (HR = 1.63, 95% CI: 1.21−2.20).
Fig. 2.
Joint associations of night-time sleep duration and sleep quality (A) and nap duration and sleep quality (B) with PPC-MM risk (n = 7586).
In contrast, the joint associations of nap duration and sleep quality with PPC-MM risk were less consistent. Among individuals with poor sleep quality (5−7 d/week of restless sleep), several nap duration categories were associated with increased PPC-MM risk. The highest risk was observed in participants reporting restless sleep on 5−7 d/week and napping for 0−30 min/day (HR = 1.61, 95% CI: 1.26–2.07). Participants with good sleep quality (<1 d/week of restless sleep) showed no significant associations between nap duration and PPC-MM risk. These results imply that good sleep quality may attenuate the association between unfavorable sleep duration and PPC-MM risk.
3.4. Association between sleep habits and transitions in PPC-MM status
From the 2011 baseline through the 2013, 2015, and 2018 follow-up waves, 1562 (20.6%) developed from baseline to physical-psychological multimorbidity (P1P2), 852 (11.2%) to physical-cognitive multimorbidity (P1C), 426 (5.6%) to psychological-cognitive multimorbidity (P2C), and 399 (5.3%) directly to physical-psychological-cognitive multimorbidity (P1P2C). In addition, 113 (7.2%), 77 (9.0%), and 36 (8.5%) participants further progressed from P1P2, P1C, and P2C to P1P2C, respectively. Baseline characteristics according to PPC-MM patterns are presented in Supplementary Table S4. To further illustrate the occurrence and progression of PPC-MM over time, the cumulative distribution of PPC-MM status is shown in Supplementary Fig. S2.
As shown in Table 3, sleep habits were associated with transitions in PPC-MM status. Compared with 7−9 h/night, very short night-time sleep duration was linked to an elevated risk of transition from baseline to dual-domain multimorbidity (<5 h/night: HR = 1.51, 95% CI = 1.35–1.70). Poor sleep quality was similarly linked to transition from baseline to dual-domain comorbidity with HRs (95% CIs) of 1.36 (1.24–1.50), 1.41 (1.26–1.58), and 1.45 (1.30–1.62) for restless sleep on 1−2, 3−4, and 5−7 d/week, respectively. Restless sleep 5−7 d/week was also associated with a direct transition from baseline to P1P2C (HR = 1.40, 95% CI: 1.06–1.86).
Table 3.
The association between sleep habits and the transitions of PPC-MM during follow-up (n = 7586).
| HR (95% CI) for the transitions of PPC-MM |
|||
|---|---|---|---|
| Baseline to dual comorbidity (P1P2 or P1C or P2C) (N = 2840) | Baseline to P1P2C (N = 399) | Dual comorbidity to P1P2C (N = 226) | |
| Night-time sleep duration(h/night) | |||
| <5 | 1.51 (1.35, 1.70) *** | 1.43 (1.06, 1.93) | 1.22 (0.83, 1.80) |
| 5−7 | 1.11 (1.02, 1.21) | 1.15 (0.92, 1.44) | 1.32 (0.98, 1.78) |
| 7−9 | Ref | Ref | Ref |
| ≥ 9 | 1.09 (0.95, 1.25) | 0.85 (0.58, 1.27) | 0.87 (0.51, 1.48) |
| Nap duration(min/d) | |||
| 0 | 0.94 (0.84, 1.04) | 1.02 (0.77, 1.36) | 0.89 (0.62, 1.28) |
| 0−30 | Ref | Ref | Ref |
| 30−60 | 0.92 (0.82, 1.03) | 1.06 (0.77, 1.45) | 0.86 (0.57, 1.31) |
| 60−90 | 0.91 (0.73, 1.14) | 0.85 (0.45, 1.61) | 1.01 (0.45, 2.27) |
| >90 | 0.99 (0.86, 1.13) | 0.91 (0.61, 1.36) | 1.04 (0.64, 1.69) |
| Sleep quality (restless days/week) | |||
| <1 | Ref | Ref | Ref |
| 1−2 | 1.36 (1.24, 1.50) *** | 1.14 (0.87, 1.49) | 0.84 (0.57, 1.23) |
| 3−4 | 1.41 (1.26, 1.58) *** | 1.06 (0.77, 1.47) | 1.24 (0.86, 1.78) |
| 5−7 | 1.45 (1.30, 1.62)*** | 1.40 (1.06, 1.86)* | 1.23 (0.86, 1.77) |
HR, hazard ratio; CI, confidence interval; P1P2: Physical-psychological multimorbidity; P1C: Physical-cognitive multimorbidity; P2C: Psychological-cognitive multimorbidity; P1P2C: Physical-psychological-cognitive multimorbidity.
All results were adjusted for age, sex, marital status, educational level, residence, smoking status, drinking status, and BMI.
FDR p < 0.05.
FDR p < 0.001.
When dual-domain multimorbidity was further classified into specific patterns (Supplementary Table S5), sleeping <5 h/night was significantly associated with transitions from baseline to P1P2 (HR = 1.52, 95% CI = 1.29–1.78), and P2C (HR = 2.52, 95% CI = 1.96–3.24). Sleeping 5−7 h/night was associated only with transition to P1P2 (HR = 1.26, 95% CI = 1.12–1.40). After FDR correction, no statistically significant associations were observed between nap duration and specific transition patterns. Poor sleep quality remained primarily associated with transitions from baseline to P1P2 and P2C, with the strongest association observed for the transition to P2C among those reporting restless sleep on 5−7 d/week (HR = 2.24, 95% CI = 1.74–2.89).
3.5. Subgroup analysis and sensitivity analysis
Following multiple imputation of missing data in sleep habits and covariates, the sample size increased to 8984. The associations were in line with the main analysis (Supplementary Tables S6−7, Supplementary Figs. S3−5). Sensitivity analyses excluding participants with baseline physical conditions, depressive symptoms, or cognitive impairment, as well as analyses excluding the sleep-related item from the CESD-10, further supported the association between very short night-time sleep duration (<5 h/night) and increased PPC-MM risk, although the association for 5−7 h/night was not retained in these analyses (Supplementary Tables S8−11, Supplementary Figs. S3−5). In the analysis restricted to participants without baseline physical, depressive, or cognitive conditions, nap duration of 30−60 min/day was associated with lower PPC-MM risk (HR = 0.71, 95% CI: 0.56−0.90). Subgroup analyses demonstrated that the associations between sleep habits and PPC-MM were not materially modified by age and sex, and no significant interactions were detected (Supplementary Table S12).
4. Discussion
In this nationally representative cohort study, we found that short night-time sleep duration and poor sleep quality at baseline were associated with an increased risk of PPC-MM during follow-up. Moreover, a U-shaped pattern was found between night-time sleep duration and PPC-MM risk, whereas nap duration showed no significant association in the primary analyses. We also observed joint associations of sleep duration and sleep quality with PPC-MM risk. In the multi-state model analyses, very short night-time sleep duration (<5 h/night) and poor sleep quality (>1 d/week) were associated with the transition from no multimorbidity to dual-domain multimorbidity. In addition, frequent restless sleep (5−7 d/week) was associated with direct transition from no multimorbidity to triple-domain multimorbidity. Notably, unlike existing studies that predominantly focus on single health outcomes, this study adopts a multidimensional perspective that integrates physical conditions, depressive symptoms, and cognitive impairment, providing evidence of the associations between sleep habits and multimorbidity.
Our findings indicate that a higher risk of PPC-MM was associated with short night-time sleep duration and poor sleep quality. These associations may reflect multiple biological pathways. First, chronic sleep disorders and short sleep duration have been associated with systemic inflammation, as reflected by elevated levels of interleukin-6 and C-reactive protein [26,27]. Such inflammation contributes to the pathogenesis of chronic physical conditions and has also been implicated in depressive symptoms and cognitive impairment [18,28,29]. Second, sleep disturbances may disrupt the hypothalamic-pituitary-adrenal (HPA) axis and alter cortisol cycles, thereby impairing stress and emotion regulation [30]. In addition, insufficient sleep may reduce the clearance of neurotoxic metabolites, particularly β-amyloid, thereby potentially accelerating neurodegenerative processes [31,32]. However, the biological pathways linking sleep habits to PPC-MM remain unclear, and further studies are needed to elucidate the relative contributions and interactions of underlying mechanisms.
Night-time sleep duration showed a U-shaped association with PPC-MM risk, in line with previous findings. This is broadly consistent with previous evidence demonstrating similar non-linear patterns between sleep duration and various health outcomes, such as cardiovascular disease [33], type 2 diabetes [34], depression [35], and cognitive impairment [36]. Building on prior findings based on single health outcomes, our study further demonstrates that this U-shaped association is present in PPC-MM, suggesting that sleep duration may serve as a marker of broader multidomain health risk.
In the primary analyses, nap duration showed no significant association with PPC-MM risk. Evidence on the health effects of napping remains inconsistent. A meta-analysis of 44 cohort studies reported that habitual napping was related to a lower risk of cognitive impairment but an increased risk of cardiovascular disease, all-cause mortality, and several other adverse health outcomes [37]. Other studies have reported both increased risks of cognitive decline [38] and depression [39], whereas a longitudinal study in Chinese adults revealed that a lower risk of depression with longer nap duration [40]. These inconsistencies may be partly explained by differences in nap categorization, cultural context, and baseline health status across populations. After excluding participants with baseline conditions, napping for 30−60 min was associated with a decreased PPC-MM risk, suggesting a potential protective effect among relatively healthy individuals.
Sleep duration and sleep quality are two core dimensions of sleep behavior [8]. We observed joint associations of these dimensions with PPC-MM risk. Specifically, short night-time sleep (<7 h/night) was associated with a higher risk of PPC-MM among participants with poor sleep quality, whereas these associations were less evident among those with better sleep quality (<1 day/week of restless sleep). A similar but less consistent pattern was observed for nap duration. These findings suggest that the effect of sleep habits may be influenced by the combined profile of multiple sleep characteristics, with sleep quality potentially playing an important role. Previous studies have reported similar patterns. For instance, a cohort study based on the CHARLS indicated that sleep duration and sleep quality jointly influenced multimorbidity progression [41]. Similarly, a prospective study from the Netherlands found that the combination of short sleep duration and poor sleep quality was related to increased cardiovascular risk [42]. Further longitudinal studies are necessary to clarify the joint effects of multidimensional sleep characteristics on multidomain health outcomes.
To further characterize the associations between sleep habits and PPC-MM progression, we used multi-state models to examine transitions across different multimorbidity states. The transition from baseline to dual-domain multimorbidity was associated with very short night-time sleep duration and poor sleep quality. Additionally, a direct transition from baseline to triple-domain multimorbidity was related to frequent restless sleep. When dual-domain multimorbidity was further classified into specific patterns, short night-time sleep duration and poor sleep quality appeared to be more strongly related to transitions involving psychological components, particularly physical-psychological and psychological-cognitive multimorbidity. This preferential vulnerability may be explained by the fact that the brain’s emotion regulation networks, including the prefrontal cortex and amygdala, are particularly sensitive to sleep deprivation, making depressive symptoms an early and prominent manifestation of sleep-related health deterioration [43,44]. In contrast to these early-stage dynamics, most associations for transitions from dual-domain to triple-domain multimorbidity were not statistically significant. These estimates should be interpreted cautiously because the number of participants in several transition states was relatively small, and later transitions may be increasingly influenced by existing disease burden and other time-varying factors. Taken together, these findings suggest that sleep disturbances may serve as early risk markers or partly reflect subclinical stages of PPC-MM. However, the observational design of this study precludes causal inference. Further longitudinal studies with repeated sleep assessments and longer follow-up are needed to clarify the temporal relationship between sleep habits and PPC-MM progression.
The advantage of our study lies in its use of nationally representative CHARLS data and long-term follow-up to systematically examine the longitudinal associations between sleep habits and PPC-MM. By integrating physical, psychological, and cognitive domains, our findings extend previous evidence that largely focused on isolated health outcomes. In addition, our analyses revealed joint associations between sleep duration and sleep quality in relation to PPC-MM risk, highlighting the importance of considering multidimensional sleep characteristics. Multi-state models offered deeper insight into the dynamic transitions of PPC-MM patterns. Nevertheless, several limitations should be considered. First, given the observational design, causal relationships cannot be inferred. Residual confounding, including unmeasured use of sleep medications or anxiolytics, may remain despite adjustment for multiple covariates. Second, self-reported measures rather than objective assessments were used to evaluate sleep habits, which may introduce recall or misclassification bias, particularly among older adults. Future studies incorporating other objective sleep measures would help validate these findings. Third, our analyses relied on baseline sleep habits and did not account for time-varying trajectories during follow-up, which may have limited the evaluation of long-term sleep patterns and their associations with PPC-MM. Fourth, the physical domain was defined using seven common chronic conditions, which may not adequately represent the complexity of the somatic burden in later life. Fifth, the psychological domain was operationalized using depressive symptoms assessed by the CESD-10. Although this approach is consistent with previous PPC-MM studies, it does not capture broader psychological conditions, such as anxiety, stress, or other psychiatric symptoms. Sixth, cognitive impairment was defined using the cognitive measures available in CHARLS, which may not fully capture all cognitive domains and do not appear to have established normative data. Finally, as the study population consisted of middle-aged and older Chinese adults, the findings may not be directly generalizable to other populations.
5. Conclusion
In conclusion, this longitudinal cohort study demonstrates that short night-time sleep duration and poor sleep quality are prospectively related to increased PPC-MM risk, with evidence of a joint association. Notably, these sleep disturbances may serve as early indicators of PPC-MM risk and partly reflect its subclinical stages. Our findings suggest that assessment of sleep habits may support early PPC-MM risk identification and promote healthy aging in middle-aged and older adults.
Author contribution
Junyi Liu: Writing—original draft, Writing—review, and editing, Methodology, Formal analysis, Conceptualization. Xianhao Huang: Writing—original draft, Writing—review, and editing, Methodology, Formal analysis, Conceptualization. Yu Hu: Writing—review, and editing, Methodology, Validation. Cui Wang: Writing—review, and editing, Methodology, Validation. Xinru Hu: Writing—review, and editing, Methodology, Validation. Xijie Yu: Writing—review, and editing, Project administration, Methodology, Validation, Funding acquisition, Conceptualization.
Ethics approval and consent to participate
The study was approved by the Biomedical Ethics Review Board of Peking University (IRB00001052-11015), and written informed consent was obtained from all participants.
Declaration of Generative AI and AI-assisted technologies in the writing process
During the preparation of this work the authors used OpenAI ChatGPT in order to assist with language editing, phrasing, and clarity. After using this tool, the authors reviewed and edited the content as needed and take full responsibility for the content of the published article.
Funding information
This work was supported by grants from the National Natural Science Foundation of China (No. 82273294) and the 1.3.5 project for discipline of excellence, West China Hospital, Sichuan University (No. 2020HXFH008, No. ZYJC18003).
Data availability
The data that support the findings of this study are available from the CHARLS database (http://charls.pku.edu.cn/). Access to the data requires registration and approval.
Declaration of competing interest
All listed authors declare that they have no conflicts of interest.
Acknowledgements
The authors thank the CHALRS databases for providing open-access data, as well as all participants and research teams for their contributions.
Footnotes
Supplementary material related to this article can be found, in the online version, at doi:https://doi.org/10.1016/j.jnha.2026.100957.
Appendix A. Supplementary data
The following is Supplementary data to this article:
References
- 1.Wang Z., Zhu D., Zhang H., Wang L., He H., Li Z., et al. Recommendations and quality of multimorbidity guidelines: a systematic review. Ageing Res Rev. 2024;102 doi: 10.1016/j.arr.2024.102559. [DOI] [PubMed] [Google Scholar]
- 2.Chowdhury S.R., Chandra Das D., Sunna T.C., Beyene J., Hossain A. Global and regional prevalence of multimorbidity in the adult population in community settings: a systematic review and meta-analysis. EClinicalMedicine. 2023;57 doi: 10.1016/j.eclinm.2023.101860. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Xu X., Mishra G.D., Jones M. Evidence on multimorbidity from definition to intervention: an overview of systematic reviews. Ageing Res Rev. 2017;37:53–68. doi: 10.1016/j.arr.2017.05.003. [DOI] [PubMed] [Google Scholar]
- 4.Ho I.S., Azcoaga-Lorenzo A., Akbari A., Black C., Davies J., Hodgins P., et al. Examining variation in the measurement of multimorbidity in research: a systematic review of 566 studies. Lancet Public Health. 2021;6(8):e587–e597. doi: 10.1016/s2468-2667(21)00107-9. [DOI] [PubMed] [Google Scholar]
- 5.Johnston M.C., Crilly M., Black C., Prescott G.J., Mercer S.W. Defining and measuring multimorbidity: a systematic review of systematic reviews. Eur J Public Health. 2019;29(1):182–189. doi: 10.1093/eurpub/cky098. [DOI] [PubMed] [Google Scholar]
- 6.Ni Y., Zhou Y., Kivimäki M., Cai Y., Carrillo-Larco R.M., Xu X., et al. Socioeconomic inequalities in physical, psychological, and cognitive multimorbidity in middle-aged and older adults in 33 countries: a cross-sectional study. Lancet Healthy Longev. 2023;4(11):e618–e628. doi: 10.1016/s2666-7568(23)00195-2. [DOI] [PubMed] [Google Scholar]
- 7.Liu K., Luo H., E B, Kuang H., Zhang C., Guo X. Adverse childhood and adult experiences, intrinsic capacity decline, and subsequent physical-psychological-cognitive multimorbidity: a prospective cohort study from China. J Nutr Health Aging. 2026;30(5) doi: 10.1016/j.jnha.2026.100832. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Buysse D.J. Sleep health: can we define it? Does it matter? Sleep. 2014;37(1):9–17. doi: 10.5665/sleep.3298. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Nelson K.L., Davis J.E., Corbett C.F. Sleep quality: an evolutionary concept analysis. Nurs Forum. 2022;57(1):144–151. doi: 10.1111/nuf.12659. [DOI] [PubMed] [Google Scholar]
- 10.Zhang Z., Xiao X., Ma W., Li J. Napping in older adults: a review of current literature. Curr Sleep Med Rep. 2020;6(3):129–135. doi: 10.1007/s40675-020-00183-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Picarsic J.L., Glynn N.W., Taylor C.A., Katula J.A., Goldman S.E., Studenski S.A., et al. Self-reported napping and duration and quality of sleep in the lifestyle interventions and independence for elders pilot study. J Am Geriatr Soc. 2008;56(9):1674–1680. doi: 10.1111/j.1532-5415.2008.01838.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.St-Onge M.P., Grandner M.A., Brown D., Conroy M.B., Jean-Louis G., Coons M., et al. Sleep duration and quality: impact on lifestyle behaviors and cardiometabolic health: a scientific statement from the American Heart Association. Circulation. 2016;134(18):e367–e386. doi: 10.1161/cir.0000000000000444. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Sun J., Ma C., Zhao M., Magnussen C.G., Xi B. Daytime napping and cardiovascular risk factors, cardiovascular disease, and mortality: a systematic review. Sleep Med Rev. 2022;65 doi: 10.1016/j.smrv.2022.101682. [DOI] [PubMed] [Google Scholar]
- 14.Um Y.J., Kim Y., Chang Y., Jung H.S., Cho I.Y., Jeon S.W., et al. Association of changes in sleep duration and quality with incidence of depression: a cohort study. J Affect Disord. 2023;328:64–71. doi: 10.1016/j.jad.2023.02.031. [DOI] [PubMed] [Google Scholar]
- 15.Overton M., Skoog J., Laukka E.J., Bodin T.H., Mattsson A.D., Sjöberg L., et al. Sleep disturbances and change in multiple cognitive domains among older adults: a multicenter study of five Nordic cohorts. Sleep. 2024;47(3) doi: 10.1093/sleep/zsad244. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Irwin M.R. Sleep and inflammation: partners in sickness and in health. Nat Rev Immunol. 2019;19(11):702–715. doi: 10.1038/s41577-019-0190-z. [DOI] [PubMed] [Google Scholar]
- 17.Asarnow L.D. Depression and sleep: what has the treatment research revealed and could the HPA axis be a potential mechanism? Curr Opin Psychol. 2020;34:112–116. doi: 10.1016/j.copsyc.2019.12.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Furman D., Campisi J., Verdin E., Carrera-Bastos P., Targ S., Franceschi C., et al. Chronic inflammation in the etiology of disease across the life span. Nat Med. 2019;25(12):1822–1832. doi: 10.1038/s41591-019-0675-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Miller A.H., Raison C.L. The role of inflammation in depression: from evolutionary imperative to modern treatment target. Nat Rev Immunol. 2016;16(1):22–34. doi: 10.1038/nri.2015.5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Irwin M.R., Vitiello M.V. Implications of sleep disturbance and inflammation for Alzheimer’s disease dementia. Lancet Neurol. 2019;18(3):296–306. doi: 10.1016/s1474-4422(18)30450-2. [DOI] [PubMed] [Google Scholar]
- 21.Zhao Y., Hu Y., Smith J.P., Strauss J., Yang G. Cohort profile: the China Health and Retirement Longitudinal Study (CHARLS) Int J Epidemiol. 2014;43(1):61–68. doi: 10.1093/ije/dys203. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Zhou Y., Htun H.L., Kivimäki M., Liao Z., Xu X., Guo Y. Social frailty and the risk of progression from health to physical, psychological, and cognitive multimorbidity: a prospective multi-cohort study. BMC Med. 2026 doi: 10.1186/s12916-026-04761-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Zhou Y., Kivimäki M., Yan L.L., Carrillo-Larco R.M., Zhang Y., Cheng Y., et al. Associations between socioeconomic inequalities and progression to psychological and cognitive multimorbidities after onset of a physical condition: a multicohort study. EClinicalMedicine. 2024;74 doi: 10.1016/j.eclinm.2024.102739. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Chen H., Mui A.C. Factorial validity of the Center for Epidemiologic Studies Depression Scale short form in older population in China. Int Psychogeriatr. 2014;26(1):49–57. doi: 10.1017/s1041610213001701. [DOI] [PubMed] [Google Scholar]
- 25.Zhou Y., Kivimäki M., Holt-Lunstad J., Yan L.L., Zhang Y., Wang H., et al. Stressful life events in childhood and adulthood and risk of physical, psychological and cognitive multimorbidities: a multicohort study. EClinicalMedicine. 2025;83 doi: 10.1016/j.eclinm.2025.103225. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Irwin M.R., Wang M., Ribeiro D., Cho H.J., Olmstead R., Breen E.C., et al. Sleep loss activates cellular inflammatory signaling. Biol Psychiatry. 2008;64(6):538–540. doi: 10.1016/j.biopsych.2008.05.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Irwin M.R., Olmstead R., Carroll J.E. Sleep disturbance, sleep duration, and inflammation: a systematic review and meta-analysis of cohort studies and experimental sleep deprivation. Biol Psychiatry. 2016;80(1):40–52. doi: 10.1016/j.biopsych.2015.05.014. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Lin T., Liu G.A., Perez E., Rainer R.D., Febo M., Cruz-Almeida Y., et al. Systemic inflammation mediates age-related cognitive deficits. Front Aging Neurosci. 2018;10:236. doi: 10.3389/fnagi.2018.00236. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Berk M., Williams L.J., Jacka F.N., O’Neil A., Pasco J.A., Moylan S., et al. So depression is an inflammatory disease, but where does the inflammation come from? BMC Med. 2013;11:200. doi: 10.1186/1741-7015-11-200. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Balbo M., Leproult R., Van Cauter E. Impact of sleep and its disturbances on hypothalamo-pituitary-adrenal axis activity. Int J Endocrinol. 2010;2010 doi: 10.1155/2010/759234. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Shokri-Kojori E., Wang G.J., Wiers C.E., Demiral S.B., Guo M., Kim S.W., et al. β-Amyloid accumulation in the human brain after one night of sleep deprivation. Proc Natl Acad Sci U S A. 2018;115(17):4483–4488. doi: 10.1073/pnas.1721694115. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Owen J.E., Veasey S.C. Impact of sleep disturbances on neurodegeneration: Insight from studies in animal models. Neurobiol Dis. 2020;139 doi: 10.1016/j.nbd.2020.104820. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Yin J., Jin X., Shan Z., Li S., Huang H., Li P., et al. Relationship of sleep duration with all-cause mortality and cardiovascular events: a systematic review and dose-response meta-analysis of prospective cohort studies. J Am Heart Assoc. 2017;6(9) doi: 10.1161/jaha.117.005947. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Shan Z., Ma H., Xie M., Yan P., Guo Y., Bao W., et al. Sleep duration and risk of type 2 diabetes: a meta-analysis of prospective studies. Diabetes Care. 2015;38(3):529–537. doi: 10.2337/dc14-2073. [DOI] [PubMed] [Google Scholar]
- 35.Li X.L., Wei J., Zhang X., Meng Z., Zhu W. Relationship between night-sleep duration and risk for depression among middle-aged and older people: a dose-response meta-analysis. Front Physiol. 2023;14 doi: 10.3389/fphys.2023.1085091. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Ma Y., Liang L., Zheng F., Shi L., Zhong B., Xie W. Association between sleep duration and cognitive decline. JAMA Netw Open. 2020;3(9) doi: 10.1001/jamanetworkopen.2020.13573. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Yang Y.B., Zheng Y.B., Sun J., Yang L.L., Li J., Gong Y.M., et al. To nap or not? Evidence from a meta-analysis of cohort studies of habitual daytime napping and health outcomes. Sleep Med Rev. 2024;78 doi: 10.1016/j.smrv.2024.101989. [DOI] [PubMed] [Google Scholar]
- 38.Leng Y., Redline S., Stone K.L., Ancoli-Israel S., Yaffe K. Objective napping, cognitive decline, and risk of cognitive impairment in older men. Alzheimers Dement. 2019;15(8):1039–1047. doi: 10.1016/j.jalz.2019.04.009. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Li L., Zhang Q., Zhu L., Zeng G., Huang H., Zhuge J., et al. Daytime naps and depression risk: a meta-analysis of observational studies. Front Psychol. 2022;13 doi: 10.3389/fpsyg.2022.1051128. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Xie B., Wang J., Li X., Zhang J., Chen M. Association between daytime napping duration and depression in middle-aged and elderly Chinese: evidence from the China Health and Retirement Longitudinal Study (CHARLS): a cross-sectional study in China. Medicine (Baltimore) 2020;99(43) doi: 10.1097/md.0000000000022686. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Zhou Y., Ni Y., Jones M., Dai X., Lim C.C.W., Zhu A., et al. Sleep behaviors and progression of multimorbidity in middle-aged and older adults: a prospective cohort study from China. J Gerontol A Biol Sci Med Sci. 2023;78(10):1871–1880. doi: 10.1093/gerona/glad087. [DOI] [PubMed] [Google Scholar]
- 42.Hoevenaar-Blom M.P., Spijkerman A.M., Kromhout D., van den Berg J.F., Verschuren W.M. Sleep duration and sleep quality in relation to 12-year cardiovascular disease incidence: the MORGEN study. Sleep. 2011;34(11):1487–1492. doi: 10.5665/sleep.1382. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Yoo S.S., Gujar N., Hu P., Jolesz F.A., Walker M.P. The human emotional brain without sleep–a prefrontal amygdala disconnect. Curr Biol. 2007;17(20):R877–8. doi: 10.1016/j.cub.2007.08.007. [DOI] [PubMed] [Google Scholar]
- 44.Goldstein A.N., Walker M.P. The role of sleep in emotional brain function. Annu Rev Clin Psychol. 2014;10:679–708. doi: 10.1146/annurev-clinpsy-032813-153716. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The data that support the findings of this study are available from the CHARLS database (http://charls.pku.edu.cn/). Access to the data requires registration and approval.