Abstract
Objective
To examine the association of changes in loneliness with subsequent cardiometabolic comorbidity (CMM) among middle-aged and older Chinese and South Korean adults.
Methods
We used the harmonized individual-level data from the China Health and Retirement Longitudinal Study (CHARLS, n = 9381) from China and the Korean Longitudinal Study of Aging (KLoSA, n = 5052) from South Korea. In both CHARLS and KLoSA, loneliness was measured using a single item from the 10-item Center for Epidemiological Studies Depression Scale (CESD-10) at baseline and in the second survey. CMM was defined as the presence of two or more cardiometabolic conditions, including diabetes, heart disease, and stroke, based on physician-diagnosed self-report. Within each cohort, we used the multivariable Cox proportional hazards models to estimate adjusted hazard ratios (aHRs) with 95% confidence intervals (CIs) of incident CMM according to changes in loneliness (never, initiated, relieved, and persistent) over 7 years (CHARLS) or 6 years (KLoSA) of follow-up.
Results
In CHARLS, initiated (aHR 1.42, 95%CI 1.14–1.78), relieved (aHR 1.40, 95%CI 1.16–1.70), and persistent (aHR 2.03, 95%CI 1.64–2.51) loneliness were associated with an increased likelihood of experiencing CMM. In KLoSA, both relieved (aHR 1.72, 95%CI 1.07–2.76) and persistent (aHR 1.86, 95%CI 1.21–2.88) loneliness were significantly associated with CMM, whereas the initiated loneliness showed no significant association (aHR 1.25, 95%CI 0.76–2.07).
Conclusions
Changes in loneliness were associated with an increased risk of subsequent CMM in both China and South Korea, with the strongest associations observed among individuals experiencing persistent loneliness. These findings indicate that loneliness is a dynamic and potentially modifiable risk factor for cardiometabolic multimorbidity across different sociocultural contexts. Early identification and targeted interventions addressing loneliness may contribute to the prevention of CMM among middle-aged and older adults.
Supplementary Information
The online version contains supplementary material available at 10.1007/s40520-026-03331-5.
Keywords: Loneliness, Dynamic change in loneliness, Cardiometabolic multimorbidity, Cross-nation study, East asia
Introduction
In the context of global aging, multimorbidity has emerged as a growing public health concern [1]. Cardiometabolic multimorbidity (CMM), a common and high disease burden comorbidity, is defined as the coexistence of two or more cardiometabolic conditions, most commonly diabetes, heart diseases, and stroke [2, 3]. There are approximately 10% of the global population who are affected by CMM [4, 5], which is associated with an increased risk of cognitive decline, dementia, and all-cause mortality, posing significant challenges to health care systems [2, 6, 7]. Therefore, exploring the modifiable factors of CMM is essential.
Loneliness is prevalent across all age groups, particularly among middle-aged and older adults [8, 9]. The WHO Commission on Social Connection has highlighted loneliness as a widespread issue with serious yet under-recognized impacts on health, well-being, and society [10]. While prior studies have examined associations between loneliness at a single time point and cardiometabolic conditions, such as diabetes [11] and cardiovascular disease (CVD) [12–14], little is known about how loneliness changes over time influences cardiometabolic conditions [15]. Considering the dynamic nature of loneliness, shaped by aging and life transitions, this perspective is crucial [8, 16]. As far as we know, only one study has investigated loneliness change in relation to cardiometabolic outcomes, focusing solely on CVD among Chinese adults aged 45 years and older [17]. However, the generalizability of these findings beyond China remains unclear. Moreover, CMM provides a more comprehensive representation of health burden in aging populations than single conditions alone [1]. Examining CMM better reflects the real-world clinical burden and public health impact of cardiometabolic diseases [3].
Therefore, to fill the above research gaps and provide more comprehensive evidence about loneliness and cardiometabolic comorbidity in Asia, we used the harmonised data from two Asian prospective cohorts and aimed to explore the association of change in loneliness with subsequent CMM among middle-aged and older adults in China and South Korea.
Methods
Study design and participants
Public datasets were obtained from harmonized data files in the Gateway to Global Aging Data [18]. For a cross-national comparison, we used data from similar time durations in waves 1–5 of CHARLS (2011, 2013, 2015, 2018, and 2020) and waves 4–8 of KLoSA (2012, 2014, 2016, 2018, and 2020). CHARLS 2011 and KLoSA 2012 were treated as baseline. The study designs of both cohorts have been described in detail elsewhere [19, 20]. Briefly, CHARLS and KLoSA are sister studies with comparable survey protocols, enrolling nationally representative, community-dwelling adults aged ≥ 45 years in China and South Korea, respectively, with follow-up surveys conducted every 2–3 years. CHARLS was approved by the Biomedical Ethics Review Committee of Peking University (IRB00001052-11015), and KLoSA by the National Statistical Office (Approval No. 33602). Written informed consent was obtained from all participants. The study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.
The flowchart of study design and participant selection is shown in Fig. 1. In CHARLS, 17,106 participants aged ≥ 45 years were recruited at baseline (wave 1, 2011). Of these, 7725 were excluded due to missing information on loneliness at baseline and the second survey (n = 4784), participants with CMM during assessment of change in loneliness (n = 388), or missing CMM data during follow-up/loss to follow-up (n = 2553). In KLoSA, 7486 participants aged ≥ 45 years were enrolled at baseline (wave 4, 2012), of whom 2434 were excluded for missing loneliness data at baseline or the second survey (n = 647), prevalent CMM during loneliness assessment (n = 342), or missing CMM data during follow-up/loss to follow-up (n = 1445). Consequently, 14,433 participants were included in the final analysis (9381 from CHARLS and 5052 from KLoSA).
Fig. 1.
Study design and participants. A. Study design; B. Participant selection . Abbreviation: CMM = Cardiometabolic multimorbidity; CHARLS = China Health and Retirement Longitudinal Study; KLoSA: Korean Longitudinal Study of Aging
Assessment of loneliness
In both CHARLS and KLoSA, loneliness was assessed at baseline and in the second survey using a single item from the 10-item Center for Epidemiological Studies Depression Scale (CESD-10): “In the past week, how often did you feel lonely?“. Participants were classified as lonely if they reported feeling lonely occasionally (1–2 days), frequently (3–4 days), or most of the time (5–7 days). Those who reported feeling lonely infrequently or never (< 1 day) were classified as not lonely [21, 22].
Assessment of change in loneliness
In the current study, change in loneliness was summarized into the following four types based on loneliness status from baseline and the second survey:
(1) Never loneliness: no loneliness at baseline and in the second survey.
(2) Initiated loneliness: no loneliness at baseline and loneliness in the second survey.
(3) Relieved loneliness: loneliness at baseline and no loneliness in the second survey.
(4) Persistent loneliness: loneliness at baseline and in the second survey.
In all analyses, the never loneliness group was the reference.
Assessment of CMM
Consistent with previous research [2, 23], CMM was defined as the presence of two or more cardiometabolic conditions, including diabetes, heart disease, and stroke, based on physician-diagnosed self-report. Specifically, each cardiometabolic condition was operationalized using its corresponding survey item. Diabetes was defined based on the question, “Have you been diagnosed with diabetes or high blood sugar by a doctor?”. Heart disease was defined using the question, “Have you been diagnosed with heart attack, coronary heart disease, angina, congestive heart failure, or other heart problems by a doctor?” Stroke was defined based on the question, “Have you been diagnosed with stroke by a doctor?”.
Assessment of covariates
Baseline covariates included age (continuous), sex (male or female), residence (urban or rural), education (less than lower secondary [primary], upper secondary & vocational training [secondary], or tertiary), and marital status (married/cohabited vs. others, including separated, divorced, widowed, or never married). Smoking status was defined as non-current smokers or current smokers. Drinking status was also classified as no drinking frequently (less than monthly) or drinking frequently (monthly drinking or more). Social activity was classified as inactive (not monthly participation in any of the social activities) or active (monthly participation in at least one of the social activities). Body mass index (BMI) was calculated as weight in kilograms divided by height in meters squared (kg/m²) [24], and further classified as underweight (< 18.5), normal (18.5–22.9), overweight (23–24.9.9), and obese (≥ 25) based on the Asia-Pacific Obesity diagnosis criteria [25]. Self-reported physician diagnoses chronic disease, including hypertension (yes or no), cancer (yes or no), and chronic lung disease (yes or no).
Statistical analyses
Continuous variables with normal distributions were presented as means with standard deviations (SD), while categorical variables were described as counts and percentages. Group comparisons were conducted using χ² tests for categorical variables, one-way ANOVA for normally distributed continuous variables, and the Kruskal-Wallis H test for variables not meeting normality assumptions. The association between changes in loneliness and incident CMM was examined using multivariable Cox proportional hazards models to estimate adjusted hazard ratios (aHRs) with 95% confidence intervals (CIs). The proportional hazards assumption was evaluated with Schoenfeld residuals and found to be satisfied (P > 0.05). Kaplan-Meier curves with log-rank tests were applied to compare CMM-free survival across loneliness-change groups in CHARLS (2013–2020, 7 years) and KLoSA (2014–2020, 6 years). To clarify the impact of loneliness change, three models with increasing levels of adjustment were fitted: Model 1 was unadjusted; Model 2 adjusted for age and sex; and Model 3 additionally adjusted for residence, education, marital status, smoking, drinking, social activity, BMI, hypertension, cancer, and chronic lung disease. Meanwhile, these associations were further examined through stratified analyses across all covariates, and likelihood ratio tests were performed by comparing models with and without interaction terms to assess their statistical significance. Moreover, three sensitivity analyses were considered: (1) To address potential bias from missing data, multiple imputation was applied. (2) We altered the measurement timing of loneliness by redefining changes in loneliness using alternative survey waves. Specifically, changes in loneliness were measured between Waves 2 and 3 in CHARLS and between Waves 5 and 6 in KLoSA. Incident CMM was then ascertained during subsequent follow-up periods, namely between Waves 4 and 5 in CHARLS and between Waves 7 and 8 in KLoSA. The same exclusion criteria and statistical models as in the main analyses were applied. (3) Depressive symptoms were further adjusted for in Model 3. Depressive symptoms were assessed using the CESD-10 scale, excluding the loneliness item, and were included as a continuous variable with total scores ranging from 0 to 27. All analyses were performed in Stata MP version 18, and two-sided P < 0.05 was considered statistically significant.
Results
In CHARLS, the never, initiated, relieved, and persistent loneliness groups had 5690 (65.6%), 1050 (11.2%), 1627 (17.3%), and 1024 (10.9%) participants, respectively. In KLoSA, the never, initiated, relieved, and persistent loneliness groups had 2646 (52.4%), 804 (15.9%), 705 (14.0%), and 897 (17.8%) participants, respectively. In CHARLS, compared with never loneliness, participants with persistent loneliness were more likely to be older, female, rural residents, with primary education, not married/cohabitated, non-current smokers, not drinking frequently, inactive social activity, underweight and normal weight, and had hypertension and chronic disease (all P < 0.05). In KLoSA, although similar findings were observed, some statistical differences were still found (Table 1).
Table 1.
Participants’ baseline characteristics by the change in loneliness
| Characterisitics | CHARLS (n = 9381) | KLoSA (n = 5052) | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Never loneliness | Initiated loneliness | Relieved loneliness | Persistent loneliness | P-value a | Never loneliness | Initiated loneliness | Relieved loneliness | Persistent loneliness | P-value a | |
| Number (%) | 5680 (65.6) | 1050 (11.2) | 1627 (17.3) | 1024 (10.9) | / | 2646 (52.4) | 804 (15.9) | 705 (14.0) | 897 (17.8) | / |
| Age, mean (SD), years | 57.5 (8.4) | 58.2 (9.0) | 58.7 (8.9) | 59.2 (8.5) | < 0.001 | 63.0 (8.4) | 65.2 (8.7) | 65.8 (9.3) | 66.9 (9.3) | < 0.001 |
| Female, n (%) | 2788 (49.1) | 605 (57.6) | 971 (59.7) | 654 (63.9) | < 0.001 | 1414 (53.4) | 505 (62.8) | 442 (62.7) | 589 (65.7) | < 0.001 |
| Urban residence, n (%) | 2180 (38.4) | 301 (28.7) | 447 (27.5) | 286 (27.9) | < 0.001 | 1997 (75.5) | 576 (71.6) | 506 (71.8) | 658 (73.4) | 0.063 |
| Education, n (%) | < 0.001 | < 0.001 | ||||||||
| Primary | 4980 (87.7) | 977 (93.1) | 1509 (92.8) | 955 (93.3) | 1363 (51.5) | 528 (65.7) | 457 (64.8) | 597 (66.6) | ||
| Secondary | 628 (11.1) | 66 (6.3) | 103 (6.3) | 62 (6.1) | 968 (36.6) | 219 (27.2) | 185 (26.2) | 241 (26.9) | ||
| Teriary | 72 (1.3) | 7 (0.7) | 15 (0.9) | 7 (0.7) | 315 (11.9) | 57 (7.1) | 63 (8.9) | 59 (6.6) | ||
| Married/Cohabitated, n (%) | 5374 (94.6) | 927 (88.3) | 1375 (84.5) | 744 (72.7) | < 0.001 | 2345 (88.6) | 626 (77.9) | 546 (77.5) | 610 (68.0) | < 0.001 |
| Current smokers, n (%) | 1824 (32.1) | 305 (29.1) | 461 (28.3) | 287 (28.0) | 0.002 | 442 (16.7) | 115 (14.3) | 108 (15.3) | 123 (13.7) | 0.114 |
| Drinking frequently, n (%) | 1047 (19.5) | 157 (15.6) | 226 (14.6) | 113 (11.6) | < 0.001 | 556 (21.0) | 136 (16.9) | 132 (18.7) | 142 (15.8) | 0.002 |
| Active social activity, n (%) | 2730 (48.1) | 450 (42.9) | 728 (44.8) | 444 (43.4) | 0.001 | 1923 (73.7) | 519 (65.9) | 484 (69.0) | 509 (57.6) | < 0.001 |
| BMI, n (%) | < 0.001 | 0.001 | ||||||||
| Underweight | 262 (5.2) | 66 (7.0) | 107 (7.6) | 83 (9.6) | 53 (2.0) | 26 (3.3) | 28 (4.1) | 38 (4.5) | ||
| Normal | 2068 (41.3) | 414 (43.8) | 611 (43.4) | 411 (47.5) | 1098 (42.1) | 327 (41.9) | 282 (40.9) | 395 (46.5) | ||
| Overweight | 1025 (20.5) | 192 (20.3) | 279 (19.8) | 153 (17.7) | 820 (31.5) | 230 (29.5) | 210 (30.5) | 223 (26.3) | ||
| Obese | 1657 (33.1) | 273 (28.9) | 412 (29.2) | 218 (25.2) | 636 (24.4) | 198 (25.4) | 169 (24.5) | 193 (22.7) | ||
| Hypertension, n (%) | 1251 (22.1) | 276 (26.4) | 393 (24.2) | 270 (26.5) | 0.001 | 862 (32.6) | 318 (39.6) | 269 (38.2) | 386 (43.0) | < 0.001 |
| Cancer, n (%) | 41 (0.7) | 8 (0.8) | 12 (0.7) | 6 (0.6) | 0.963 | 84 (3.2) | 47 (5.9) | 43 (6.1) | 51 (5.7) | < 0.001 |
| Chronic lung disease, n (%) | 409 (7.2) | 104 (9.9) | 165 (10.2) | 128 (12.5) | < 0.001 | 56 (2.1) | 19 (2.4) | 18 (2.6) | 27 (3.0) | 0.490 |
CHARLS = China Health and Retirement Longitudinal Study; KLoSA: Korean Longitudinal Study of Aging; SD = Standard deviation; BMI = Body mass index
a The groups’ differences among characteristics were performed using ANOVA for continuous variables, and the chi-square test for categorical variables
Missing information: 10 for age, 1 for smoking, 470 for drinking, 5 for social activity, 1150 for BMI, 30 for hypertension, 14 for cancer, 11 for chronic lung disease (CHARLS), and 71 for social, 126 for BMI, 2 for cancer (KLoSA)
In the current study, a total of 896 (9.55%) and 149 (2.95%) participants developed CMM in the CHARLS and KLoSA studies during the follow-up period. Figure 2 shows Kaplan-Meier survival curves by change in loneliness group, indicating that the groups significantly differed in incident CMM (log-rank tests: P < 0.001). For the incidence rate per 1000 person-years, participants with never loneliness, initiated loneliness, relieved loneliness, and persistent loneliness were 11.55, 16.61, 16.20, and 21.79 (CHARLS); 3.49, 4.82, 6.95, and 7.94 (KLoSA).
Fig. 2.
Kaplan-Meier plots of CMM-free survival by change in loneliness with the log-rank test. A. CHARLS; B. KLoSA. Abbreviation: CMM = Cardiometabolic multimorbidity; CHARLS = China Health and Retirement Longitudinal Study; KLoSA: Korean Longitudinal Study of Aging
Table 2 summarizes the results of multivariable adjusted Cox proportional hazards models to explore the association of change in loneliness and CMM among the CHARLS and KLoSA studies. After adjusting covariates (Model 3), we found that participants who experienced initiated (aHR 1.42, 95%CI 1.14–1.78), relieved (aHR 1.40, 95%CI 1.16–1.70), and persistent (aHR 2.03, 95%CI 1.64–2.51) loneliness had higher risks of experiencing CMM than participants with never loneliness in CHARLS. Differently, in the KLoSA study, the evident associations were found in two groups of change in loneliness: relieved (aHR 1.72, 95%CI 1.07–2.76) and persistent (aHR 1.86, 95%CI 1.21–2.88) loneliness, whereas the initiated loneliness showed no significant association (aHR 1.25, 95%CI 0.76–2.07). The association between changes in loneliness and CMM did not differ across any covariates (all P for interaction > 0.05) (Tables S1 and S2). Moreover, we repeated the main analyses in all sensitivity analyses and found similar findings (Table S3-S5)
Table 2.
Incidence of CMM according to the change in loneliness
| Case, Number | Incidence Rate, per 1000 Person-Years | HR (95% CI) | |||
|---|---|---|---|---|---|
| Model 1 | Model 2 | Model 3 | |||
| CHARLS | |||||
| Change in loneliness | |||||
| Never loneliness | 449 | 11.55 | 1.00 [Ref.] | 1.00 [Ref.] | 1.00 [Ref.] |
| Initiated loneliness | 118 | 16.61 | 1.44 (1.18–1.76.18.76) *** | 1.38 (1.12–1.69.12.69) ** | 1.42 (1.14–1.78.14.78) ** |
| Relieved loneliness | 179 | 16.20 | 1.40 (1.18–1.67.18.67) *** | 1.34 (1.12–1.59.12.59) *** | 1.40 (1.16–1.70.16.70) *** |
| Persistent loneliness | 150 | 21.79 | 1.89 (1.57–2.27.57.27) *** | 1.77 (1.47–2.14.47.14) *** | 2.03 (1.64–2.51.64.51) *** |
| KLoSA | |||||
| Change in loneliness | |||||
| Never loneliness | 55 | 3.49 | 1.00 [Ref.] | 1.00 [Ref.] | 1.00 [Ref.] |
| Initiated loneliness | 23 | 4.82 | 1.38 (0.85–2.25.85.25) | 1.29 (0.79–2.10.79.10) | 1.25 (0.76–2.07.76.07) |
| Relieved loneliness | 29 | 6.95 | 1.99 (1.27–3.12.27.12) ** | 1.78 (1.13–2.81.13.81) * | 1.72 (1.07–2.76.07.76) * |
| Persistent loneliness | 42 | 7.94 | 2.27 (1.52–3.40.52.40) *** | 1.96 (1.30–2.96.30.96) *** | 1.86 (1.21–2.88.21.88) ** |
CMM=Cardiometabolic multimorbidity; CHARLS=China Health and Retirement Longitudinal Study; KLoSA: Korean Longitudinal Study of Aging; HR=Hazard ratio; CI=Confidence interval.Mode 1 is a crude model. Model 2 adjusted for age and sex. Model 3 adjusted age, sex, residence, education, marital status, smoking, drinking, social activity, BMI, hypertension, cancer, and chronic lung disease.
Discussion
This study investigated the association between changes in loneliness and the risk of incident CMM in two large Asian cohorts. We found that both relieved and persistent loneliness were significantly associated with a higher risk of subsequent CMM, whereas the effect of initiated loneliness was evident only in the CHARLS cohort. These findings underscore the importance of considering the dynamic nature of loneliness in understanding its long-term cardiometabolic implications.
To our knowledge, no prior research has prospectively examined the relationship between changes in loneliness and CMM. Previous studies have largely focused on baseline loneliness and single cardiometabolic conditions. For example, analyses from the English Longitudinal Study of Ageing demonstrated that loneliness predicted type 2 diabetes, while studies in the USA, South Korea, and the UK consistently reported baseline loneliness to be associated with increased risk of CVD (adjusted HRs ranging from 1.05 to 1.16) [13, 14]. The only study addressing loneliness transitions examined 8463 Chinese adults and showed that increased loneliness was associated with higher risks of overall CVD and heart disease (HRs of 2.44 and 2.34, respectively) [17]. Our results extend this evidence in two key ways: first, by broadening the outcome from single conditions to CMM, which better reflects the real clinical burden of aging populations; and second, by using harmonized data from both China and Korea, thereby enhancing the generalizability of findings across different Asian settings.
Although family structures in China and South Korea are quite similar, with people tending to be more interdependent at the family level [26, 27], in the current study, initiation of loneliness was linked to CMM onset in CHARLS but not in KLoSA. It should be noted, however, that the number of CMM cases in the initiated loneliness group in KLoSA was small, which may have limited the statistical power to detect a significant association. Despite this limitation, we also consider cultural and social structural factors that may influence the impact of newly initiated loneliness. A sudden change in subjective social health may have impacts that depend on the social structure. In more collectivist countries, social bonds and family roles come with stronger expectations, so a sudden loss of these ties can be more stressful, and without other support, may further lead to worse health outcomes [28]. Differences in social structural resources may also play a role. Compared to CHARLS, the KLoSA sample had higher education levels (Table 1) and access to a well-developed universal healthcare system, which can help individuals manage health risks and buffer the effects of sudden social frailty, such as the initiation of loneliness. Previous evidence supports this interpretation, and it suggests that socioeconomic inequalities were associated with a higher risk of experiencing psychological and cognitive multimorbidities [29]. Overall, while aging itself brings biological risks, the effect of newly initiated loneliness on CMM seems shaped by both cultural norms and social resources.
Initiation of loneliness was associated with CMM onset only in CHARLS, while both relieved and persistent loneliness were associated with a higher CMM risk in both CHARLS and KLoSA. For relieved loneliness, several explanations may account for the persistently elevated risk. First, although feelings of loneliness were reduced at the subsequent assessment, the duration of recovery may have been too short to reverse the cumulative psychological and physiological damage induced by earlier loneliness. Second, individuals classified as having relieved loneliness may still experience residual or subthreshold loneliness that is not fully captured by a single-item measure, leading to continued exposure to loneliness-related stress. Third, earlier experience of loneliness may already have lasting effects on psychological and physiological systems. Chronic stress during these periods can dysregulate the hypothalamic-pituitary-adrenal (HPA) axis, leading to elevated cortisol levels and sustained inflammatory activity [30]. It can also disturb autonomic nervous system function, affecting heart rate variability and vascular regulation, while contributing to metabolic changes such as insulin resistance and altered lipid profiles [31, 32]. In addition, loneliness can influence health behaviors, including reduced physical activity, poorer sleep quality, and less optimal dietary choices, which may persist even after loneliness is reportedly alleviated, thereby further compounding cardiometabolic risk [31, 33, 34]. Taken together, these pathways suggest that the subjective experience of loneliness, whether persistently ongoing or previously experienced, can have enduring effects on multiple biological and behavioral systems, providing plausible mechanisms linking loneliness to CMM across populations.
Although loneliness and depressive symptoms are closely related, they represent distinct psychosocial constructs. In the present study, the associations between changes in loneliness and CMM remained largely unchanged after additional adjustment for depressive symptoms, suggesting that loneliness may exert an independent effect on the development of cardiometabolic multimorbidity beyond depression. This finding is consistent with previous evidence indicating that loneliness may influence cardiometabolic health through behavioral, neuroendocrine, and inflammatory pathways that are not fully explained by depressive symptoms [13, 14].
Strengths and limitations
This study has several strengths. First, including two large, well-characterized cohorts from China and South Korea enhances the generalizability of the findings. Second, examining changes in loneliness over time provides novel insights into the dynamic nature of loneliness and its impact on CMM. Several limitations should also be noted. Loneliness was measured by self-report, which may be subject to recall bias or misclassification, and changes in loneliness were captured using only two surveys; more frequent assessments could provide a more precise picture of these changes. Similarly, CMM was based on physician diagnoses reported by participants, as in many previous studies [35, 36], which could introduce some degree of bias, although prior evidence indicates that about 77.5% of self-reported cardiometabolic conditions are confirmed in medical records [37]. The observational design of this study also prevents causal inference. To reduce the risk of reverse causality, we excluded cases of CMM reported at baseline and in the second survey, strengthening the temporal sequence. The number of CMM cases in the initiated loneliness group in KLoSA was relatively small, which may have limited statistical power to detect significant associations. Therefore, the non-significant finding in this subgroup should be interpreted cautiously. Loneliness was measured using a single item from the CESD-10 rather than a multi-item scale, such as the UCLA Loneliness Scale. This single-item measure may be less precise and could result in some misclassification of participants’ loneliness status. This limitation was due to the constraints of the harmonized datasets used in this study. Finally, despite adjustment for numerous confounders, residual confounding cannot be ruled out. Future research should aim to address these limitations.
Conclusions
In summary, this study highlights the importance of loneliness as a risk factor for CMM, with both relieved and persistent loneliness contributing to increased risk. While the impact of initiated loneliness appeared less consistent, historically or longer-term experienced loneliness showed consistent associations with CMM across China and South Korea. These results emphasize the value of monitoring loneliness over time in middle-aged and older adults and taking country-specific contexts into account to better understand its impact on cardiometabolic health. Clinically, loneliness should be recognized as a potentially modifiable risk factor and addressed proactively to help reduce the risk of CMM in this population.
Supplementary Information
Below is the link to the electronic supplementary material.
Acknowledgements
Thanks to the CHARLS and KLoSA teams, as well as all participants.
Author contributions
MW: Drafting of the manuscript; Acquisition, analysis, or interpretation of data; Critical revision of the manuscript for important intellectual content; and Statistical analysis. HW: Concept and design; Drafting of the manuscript; Acquisition, analysis, or interpretation of data; Critical revision of the manuscript for important intellectual content; Administrative, technical, or material support; and Supervision.
Funding
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Data availability
The data supporting this study’s findings are available from the CHARLS website: https://charls.charlsdata.com/pages/data/111/en.html and the KLoSA website: http://survey.keis.or.kr/eng/klosa/klosa01.jsp.
Declarations
Competing interests
The authors declare no competing interests.
Ethical approval
CHARLS was approved by the Biomedical Ethics Review Committee of Peking University (IRB00001052-11015), and KLoSA by the National Statistical Office (Approval No. 33602).
Human and animal rights
All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.
Informed consent
All participating studies were approved by Institutional Review Boards and the respondents provided written informed consent. Additional ethical approval was not required for this analysis of the anonymised data.
Footnotes
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Barnett K, Mercer SW, Norbury M, Watt G, Wyke S, Guthrie B (2012) Epidemiology of multimorbidity and implications for health care, research, and medical education: a cross-sectional study. Lancet 380(9836):37–43 [DOI] [PubMed] [Google Scholar]
- 2.Zhou Y, Kivimaki M, Lim CCW, Carrillo-Larco RM, Qi S, Wu X, Xu X (2024) Bidirectional associations between cardiometabolic multimorbidity and depression and mediation of lifestyles a multicohort study. Jacc-Asia 4(9):657–671 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Busija L, Lim K, Szoeke C, Sanders KM, McCabe MP (2019) Do replicable profiles of multimorbidity exist? Systematic review and synthesis. Eur J Epidemiol 34(11):1025–1053 [DOI] [PubMed] [Google Scholar]
- 4.Cheng X, Ma T, Ouyang F, Zhang G, Bai Y (2022) Trends in the prevalence of cardiometabolic multimorbidity in the United States, 1999–2018. Int J Environ Res Public Health 19(8):4726. 10.3390/ijerph19084726 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Ling J, Koye D, Buizen L, Khunti K, Montvida O, Paul SK (2021) Temporal trends in co-morbidities and cardiometabolic risk factors at the time of type 2 diabetes diagnosis in the UK. Diabetes Obes Metab 23(5):1150–1161 [DOI] [PubMed] [Google Scholar]
- 6.Lyall DM, Celis-Morales CA, Anderson J, Gill JMR, Mackay DF, McIntosh AM, Smith DJ, Deary IJ, Sattar N, Pell JP (2017) Associations between single andmultiple cardiometabolic diseases and cognitive abilities in 474 129 UK biobank participants. Eur Heart J 38(8):577–583 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Fishbook BN, Brinton CD, Siever J, Klassen TD, Sakakibara BM (2022) Cardiometabolic multimorbidity and activity limitation: a cross-sectional study of adults using the Canadian longitudinal study on aging data. Fam Pract 39(3):455–463 [DOI] [PubMed] [Google Scholar]
- 8.Li R, Li RY, Xie JC, Chen JX, Liu S, Pan A, Liu G (2023) Associations of socioeconomic status and healthy lifestyle with incident early-onset and late-onset dementia: a prospective cohort study. Lancet Healthy Longev 4(12):693–702 [DOI] [PubMed] [Google Scholar]
- 9.Gerst-Emerson K, Jayawardhana J (2015) Loneliness as a public health issue: the impact of loneliness on health care utilization among older adults. Am J Public Health 105(5):1013–1019 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.From loneliness to social connection: charting a path to healthier societies [https://www.who.int/publications/i/item/978240112360
- 11.Hackett RA, Hudson JL, Chilcot J (2020) Loneliness and type 2 diabetes incidence: findings from the English longitudinal study of ageing. Diabetologia 63(11):2329–2338 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Valtorta NK, Kanaan M, Gilbody S, Hanratty B (2018) Loneliness, social isolation and risk of cardiovascular disease in the English longitudinal study of ageing. Eur J Prev Cardiol 25(13):1387–1396 [DOI] [PubMed] [Google Scholar]
- 13.Lee HH, Chen RJ, Okuzono SS, Kubzansky LD (2025) Loneliness and cardiovascular disease incidence: two cohorts of older adults in the USA and South Korea. Int J Epidemiol 54(3):dyaf050. 10.1093/ije/dyaf050 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Bu FF, Zaninotto P, Fancourt D (2020) Longitudinal associations between loneliness, social isolation and cardiovascular events. Heart 106(18):1394– [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Valtorta NK, Kanaan M, Gilbody S, Ronzi S, Hanratty B (2016) Loneliness and social isolation as risk factors for coronary heart disease and stroke: systematic review and meta-analysis of longitudinal observational studies. Heart 102(13):1009–1016 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.von Soest T, Luhmann M, Hansen T, Gerstorf D (2020) Development of loneliness in midlife and old age: its nature and correlates. J Personal Soc Psychol 118(2):388–406 [DOI] [PubMed] [Google Scholar]
- 17.Lu Y, Wang R, Norman J, Yu P (2024) Loneliness status transitions and risk of cardiovascular disease among middle-aged and older adults. Nutr Metab Cardiovasc Dis 34(3):718–725 [DOI] [PubMed] [Google Scholar]
- 18.Juster FT, Suzman R (1995) An overview of the health and retirement study. J Hum Resour 30:S7–S56
- 19.Zhao Y, Hu Y, Smith JP, Strauss J, Yang G (2014) Cohort profile: the China health and retirement longitudinal study (CHARLS). Int J Epidemiol 43(1):61–68 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Shin C (2019) Korean Longitudinal Study of Ageing. In: Gu D, Dupre ME (eds) Encyclopedia of Gerontology and Population Aging, edn. Springer International Publishing, Cham, pp 1–4
- 21.Victor CR, Bowling A (2012) A longitudinal analysis of loneliness among older people in Great Britain. J Psychol 146(3):313–331 [DOI] [PubMed] [Google Scholar]
- 22.Wang Y, Liu M, Yang F, Chen H, Wang Y, Liu J (2024) The associations of socioeconomic status, social activities, and loneliness with depressive symptoms in adults aged 50 years and older across 24 countries: findings from five prospective cohort studies. Lancet Healthy Longev 5(9):100618. 10.1016/j.lanhl.2024.07.001 [DOI] [PubMed] [Google Scholar]
- 23.Collaboration TERF (2015) Association of cardiometabolic multimorbidity with mortality. JAMA 314(1):52–60 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Hu H, Wang J, Han X, Li Y, Wang F, Yuan J, Miao X, Yang H, He M (2018) BMI, waist circumference and all-cause mortality in a middle-aged and elderly Chinese population. J Nutr Health Aging 22(8):975–981 [DOI] [PMC free article] [PubMed]
- 25.Organization WH (2000) The Asia-Pacific perspective: redefining obesity and its treatment. https://www.who.int/data/gho/data/themes/topics/topic-details/GHO/body-mass-index
- 26.Nisbett RE, Masuda T (2003) Culture and point of view. Proc Natl Acad Sci USA 100(19):11163–11170 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Ishii K, Miyamoto Y, Rule NO, Toriyama R (2014) Physical objects as vehicles of cultural transmission: maintaining harmony and uniqueness through colored geometric patterns. Pers Soc Psychol Bull 40(2):175–188 [DOI] [PubMed] [Google Scholar]
- 28.Beller J, Wagner A (2020) Loneliness and health: the moderating effect of cross-cultural individualism/collectivism. J Aging Health 32(10):1516–1527 [DOI] [PubMed] [Google Scholar]
- 29.Zhou YG, Kivimäki M, Yan LL, Carrillo-Larco RM, Zhang Y, Cheng YY, Wang H, Zhou MG, Xu XL (2024) Associations between socioeconomic inequalities and progression to psychological and cognitive multimorbidities after onset of a physical condition: a multicohort study. Eclin Med 74:102739. 10.1016/j.eclinm.2024.102739 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Zhu X, Tang HD, Dong WY, Kang F, Liu A, Mao Y, Xie W, Zhang XL, Cao P, Zhou WJ et al (2021) Distinct thalamocortical circuits underlie allodynia induced by tissue injury and by depression-like States. Nat Neurosci 24(4):542–553 [DOI] [PubMed] [Google Scholar]
- 31.Cacioppo JT, Hawkley LC, Crawford LE, Ernst JM, Burleson MH, Kowalewski RB, Malarkey WB, Van Cauter E, Berntson GG (2002) Loneliness and health: potential mechanisms. Psychosom Med 64(3):407–417 [DOI] [PubMed] [Google Scholar]
- 32.Hawkley LC, Cacioppo JT (2003) Loneliness and pathways to disease. Brain Behav Immun 17:S98–S105 [DOI] [PubMed] [Google Scholar]
- 33.Christiansen J, Lund R, Qualter P, Andersen CM, Pedersen SS, Lasgaard M (2021) Loneliness, social isolation, and chronic disease outcomes. Ann Behav Med 55(3):203–215 [DOI] [PubMed] [Google Scholar]
- 34.Perez LG, Siconolfi D, Troxel WM, Tucker JS, Seelam R, Rodriguez A, Shih RA (2022) D’Amico EJ: loneliness and multiple health domains: associations among emerging adults. J Behav Med 45(2):260–271 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Bai AY, Chen QS, Geldsetzer P, Gray M, Xie ZL, Zhang DQ, Baernighausen T, Hu YX, Jiang Y, Chen SM (2024) Functional dependency and cardiometabolic multimorbidity in older people: pooled analysis of individual-level data from 20 countries. Age Ageing 53(12):afae269. 10.1093/ageing/afae269 [DOI] [PubMed] [Google Scholar]
- 36.Han YT, Hu YZ, Yu CQ, Guo Y, Pei P, Yang L, Chen YP, Du HD, Sun DJY, Pang YJ et al (2021) Lifestyle, cardiometabolic disease, and Multimorbidity in a prospective Chinese study. Eur Heart J 42(34):3374– [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Xie WX, Zheng FF, Yan L, Zhong BL (2019) Cognitive decline before and after incident coronary events. J Am Coll Cardiol 73(24):3041–3050 [DOI] [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 supporting this study’s findings are available from the CHARLS website: https://charls.charlsdata.com/pages/data/111/en.html and the KLoSA website: http://survey.keis.or.kr/eng/klosa/klosa01.jsp.