Sarcopenia as a causal mediator in aging, obesity and central obesity related outcomes: A comprehensive analysis of NHANES, CHARLS and ELSA.

BACKGROUNDDepressive symptoms differ from clinical depression. However, the relationship between depressive symptom trajectories and stroke risk across diverse geographic regions remains unclear.AIMTo address the gap in the existing understanding of the relationship between depressive symptom trajectories and stroke risk, the current study utilized three representative cohorts.METHODSIn this study, we used three representative cohorts from Asia, Europe, and the Americas: China Health and Retirement Longitudinal Study (CHARLS), English Longitudinal Study of Ageing (ELSA), and Health and Retirement Study (HRS). Depressive symptoms were assessed using the 8-item Center for Epidemiological Studies Depression scale and categorized into somatic and cognitive-affective subtypes. The trajectories of depressive symptoms were monitored over four surveys starting from baseline and classified into five distinct states: persistently low, decreasing, fluctuating, increasing, and consistently high. Self-reported physician diagnoses were used to evaluate the subsequent stroke events. Hazard ratios (HRs) and 95% confidence intervals (95%CIs) were computed using Cox proportional-risk models adjusted for potential confounding factors.RESULTSA total of 7990 participants from CHARLS (females: 52.3%, mean age: 63.4 years), 5642 participants from ELSA (females: 56.2%, mean age: 63.7 years), and 12260 participants from HRS (females: 61.4%, mean age: 64.7 years) participated in this study. The median follow-up periods were 5 years for CHARLS, 8 years for ELSA, and 10 years for HRS. In comparison with the persistently low trajectory, consistently high and fluctuating trajectories of total depressive symptoms increased the risk of stroke in all three cohorts (CHARLS: HR = 1.80, 95%CI: 1.36-2.38; ELSA: HR = 1.50, 95%CI: 1.02-2.21; HRS: HR = 1.45, 95%CI: 1.29-1.62 for consistently high; CHARLS: HR = 1.47, 95%CI: 1.14-1.90; ELSA: HR = 1.44, 95%CI: 1.17-1.77; HRS: HR = 1.26, 95%CI: 1.13-1.41 for fluctuating). Increasing trajectories enhanced the risk in the European cohort (ELSA: HR = 1.71, 95%CI: 1.06-2.74), while decreasing trajectories did not increase stroke risk in any cohort. For somatic depressive symptoms, consistently high and fluctuating trajectories increased the risk of stroke across all cohorts (CHARLS: HR = 2.16, 95%CI: 1.67-2.79; ELSA: HR = 1.94, 95%CI: 1.34-2.81; HRS: HR = 1.79, 95%CI: 1.49-2.15 for consistently high; CHARLS: HR = 1.35, 95%CI: 1.20-1.62; ELSA: HR = 1.56, 95%CI: 1.27-1.92; HRS: HR = 1.33, 95%CI: 1.20-1.46 for fluctuating). Increasing trajectories only increased the risk in the European cohort (ELSA: HR = 1.95, 95%CI: 1.11-3.43), while decreasing trajectories did not increase stroke risk in the European and American cohorts. For cognitive-affective depressive symptoms, consistently high and fluctuating trajectories increased the risk in the Asian and European cohorts (CHARLS: HR = 2.06, 95%CI: 1.52-2.81; ELSA: HR = 1.25, 95%CI: 1.02-1.54 for consistently high; CHARLS: HR = 1.63, 95%CI: 1.23-2.16; ELSA: HR = 1.58, 95%CI: 1.11-2.24 for fluctuating). Increasing trajectories increased the risk only in the American cohort (HRS: HR = 14.67, 95%CI: 1.87-114.91).CONCLUSIONConsistently high and fluctuating trajectories of total and somatic depressive symptoms were associated with an increased risk for stroke across all populations. Consistently high, fluctuating, and increasing trajectories of cognitive-affective symptoms pose a risk for certain populations. These findings highlight the importance of targeted interventions for managing depressive symptoms as potential strategies for stroke prevention, particularly in regions where specific symptom trajectories are prevalent.

Frailty is associated with the progression of prediabetes to diabetes and elevated risks of cardiovascular disease and all-cause mortality in individuals with prediabetes and diabetes: Evidence from two prospective cohorts

Cardiometabolic multimorbidity (CMM) is increasingly common among middle-aged and older adults, but there remains a lack of simple risk indicators that can simultaneously reflect inflammatory and metabolic burden. The C-reactive protein-triglyceride-glucose index (CTI) has been proposed as a novel biomarker that integrates insulin resistance and inflammatory status. This study aimed to assess the association of CTI and its cumulative exposure with CMM. This study conducted the primary analyses using the 2011 baseline cross-sectional data and the 2011-2020 prospective follow-up data from the China Health and Retirement Longitudinal Study (CHARLS), and additionally performed longitudinal supplementary analyses by calculating cumulative CTI over 2011-2015 within CHARLS. An external longitudinal validation was conducted using the 2002-2012 cohort from the UK English Longitudinal Study of Ageing (ELSA), and cross-population cross-sectional replication was further performed using the 2001-2010 data from the US National Health and Nutrition Examination Survey (NHANES). Multivariable logistic regression, Cox proportional hazards models, Kaplan-Meier curves, and restricted cubic spline analyses were used to examine the associations of CTI with the prevalence and incident risk of CMM, and cumulative CTI from 2011 to 2015 was calculated in CHARLS for supplementary analyses; meanwhile, receiver operating characteristic (ROC) curves were used to evaluate the predictive performance of CTI for 4-year CMM occurrence, and prespecified subgroup analyses were conducted to test the robustness of the findings. In the CHARLS prospective cohort, among 10,863 participants free of CMM at baseline, 1698 incident CMM cases (15.6%) occurred during approximately 9years of follow-up. In the primary model, each 1-unit increase in baseline CTI was associated with a 71% higher risk of CMM (HR = 1.71, 95% CI 1.39-2.11), and the hazard ratio comparing the highest with the lowest quartile of CTI was 1.33 (95% CI 0.94-1.88); in the 2015 baseline subcohort, each 1-unit increase in cumulative CTI was associated with an HR of 1.02 (95% CI 1.00-1.05). Cross-sectional analyses in CHARLS and NHANES both showed a significant positive association between CTI and CMM prevalence, with a dose-response relationship. After adding CTI to a model containing traditional risk factors, the area under the receiver operating characteristic curve (AUC) for 4-year CMM prediction increased from 0.753 to 0.778. In the external longitudinal validation, 3129 participants from the ELSA cohort were included, among whom 406 incident CMM cases occurred during follow-up (13.0%). In the primary model, each 1-unit increase in CTI was associated with an HR of 1.59 (95% CI 0.99-2.55), and participants in the highest quartile had a higher risk than those in the lowest quartile (HR = 2.22, 95% CI 1.06-4.62), with a significant trend test (P for trend = 0.027), indicating that the direction of the association was consistent with that observed in CHARLS. CTI and its cumulative exposure level are closely associated with both the prevalence and incidence risk of CMM, and these findings were further supported by validation in the independent longitudinal ELSA cohort. As a simple composite index derived from routinely measured indicators, CTI may help identify individuals at high risk of CMM and provide a reference for early risk stratification and intervention in middle-aged and older populations.

Arthritis is a common and debilitating condition, particularly in older adults. Cardiorespiratory fitness (CRF) is a known indicator of cardiovascular health and has been associated with various chronic conditions. This study explores the association between estimated cardiorespiratory fitness (eCRF) and arthritis in older adults. Data were obtained from the English Longitudinal Study of Ageing (ELSA), Health and Retirement Study (HRS), National Health and Nutrition Examination Survey (NHANES), and China Health and Retirement Longitudinal Study (CHARLS). Logistic regression was used to examine the relationship between eCRF levels and arthritis across different cohorts, adjusting for various confounders. Sensitivity analysis was performed to assess the robustness of our results. Higher eCRF was consistently associated with lower odds of arthritis in Western populations. In Eastern populations, although a negative correlation was observed, the significance was relatively weaker. Each 1-SD increase in eCRF was associated with 40% lower odds of arthritis in ELSA (OR 0.60, 0.54-0.67), 31% lower odds in HRS (OR 0.69, 0.63-0.75), 32% lower odds in NHANES (OR 0.68, 0.63-0.73), and 11% lower odds in CHARLS (OR 0.89, 0.77-1.03). In the Western cohorts, compared to the low eCRF group, the moderate eCRF and high eCRF groups had significantly lower odds of arthritis (all p < 0.05). However, in the Eastern cohorts, only the moderate eCRF group showed a statistically significant association with lower odds of arthritis (OR 0.79, 0.65-0.97, p = 0.025). In these cross-sectional analyses, eCRF was associated with the presence of self-reported, physician-diagnosed arthritis among older adults, with the association being weaker in the Eastern cohort than in the Western cohorts. Future prospective research should investigate the complex and potentially different mechanisms linking eCRF with arthritis in Eastern and Western populations.

BackgroundThe remnant cholesterol inflammatory index (RCII) is a novel metric that combines remnant cholesterol and high-sensitivity C-reactive protein, reflecting the metabolic and inflammatory risk. This study investigates the association between RCII and long-term risks of all-cause and cause-specific mortality in middle-aged and elderly populations in the US and China.MethodWe analyzed data from the National Health and Nutrition Examination Survey (NHANES) and the China Health and Retirement Longitudinal Study (CHARLS), including 7,565 and 12,932 participants aged 45 years and older, respectively. The participants were categorized into quartiles based on natural log-transformed RCII (lnRCII) values. Kaplan–Meier survival analysis, Cox proportional hazards models, restricted cubic splines (RCS) and mediation analysis were used to examine the relationship between lnRCII and mortality outcomes, adjusting for potential covariates.ResultThe mean age of the participants was 59.90 ± 10.44 years (NHANES) and 58.64 ± 9.78 years (CHARLS), with 53.28% and 52.50% female, respectively. Kaplan–Meier survival analysis showed that higher lnRCII quartiles (≥ 0.79 in NHANES, ≥ -0.13 in CHARLS) were significantly associated with increased all-cause mortality risk (p < 0.001). Each standard deviation (SD) increase in lnRCII corresponded to a higher risk of all-cause mortality, and the hazard ratios (HRs) and 95% confidence interval (CI) were 1.29 (95% CI: 1.21–1.36) in NHANES and 1.26 (95% CI: 1.15–1.38) in CHARLS. In NHANES, lnRCII was also associated with elevated risks of cardiovascular mortality (HR = 1.21, 95% CI: 1.08–1.35) and cancer mortality (HR = 1.30, 95% CI: 1.09–1.55). RCS analysis indicated a J-shaped relationship between lnRCII and both all-cause and cardiovascular mortality, and a linear association with cancer mortality. Mediation analysis showed that systolic blood pressure and fasting plasma glucose partially mediated these associations. Subgroup analyses suggested a stronger association between lnRCII and all-cause mortality in middle-aged US participants (p for interaction = 0.010).ConclusionsElevated RCII levels are significantly associated with increased all-cause mortality risk middle-aged and elderly populations in both the US and China. In the US population, RCII is also associated with increased risks of cardiovascular and cancer mortality. By integrating metabolic and inflammatory risk factors, RCII may serve as a valuable tool for mortality risk stratification and clinical decision-making.

Depression is highly prevalent among older adults and has consistently been identified as an independent risk factor for incident stroke. Most previous cohort studies have relied on a single baseline measure of depressive symptoms. However, accumulating evidence suggests that the dynamic progression of depression-its persistence, remission, or emergence over time-may differentially affect cerebrovascular risk. Furthermore, existing research has predominantly been limited to individual national cohorts, thereby restricting the generalizability of findings across diverse sociocultural and healthcare contexts. To address these limitations, we analyzed depressive symptom trajectories over multiple waves and their association with subsequent stroke onset in three large, prospective aging cohorts: the Health and Retirement Study (HRS), the English Longitudinal Study of Ageing (ELSA), and the China Health and Retirement Longitudinal Study (CHARLS). In this study, Depressive symptoms were evaluated biennially using the Center for Epidemiologic Studies Depression Scale (CES-D) for participants in HRS and ELSA and CHARLS. Through rule-based trajectory modeling, four distinct patterns of symptom burden trajectory were identified: No, Decreasing, Increasing, and Consistently high. Stroke events were ascertained via self-reported physician diagnosis in HRS and ELSA, and through medical record linkage in CHARLS. For each cohort, Logistic regression models were applied to estimate Odds Ratios (ORs) and 95% Confidence Intervals (CIs). for stroke risk across the trajectory groups, using the persistent low group as the reference. The models were adjusted for potential confounders, including demographic variables, health behaviors, and history of chronic disorders. Furthermore, stratified analyses were performed across subgroups to explore potential heterogeneity in the associations between depressive symptom trajectories and stroke risk. The study included 4587 participants from HRS, 4879 from ELSA, and 7792 from CHARLS. During the follow-up period, there were 376, 137, and 133 incident strokes reported in the HRS, ELSA, and CHARLS cohorts, respectively. The "Consistently high" depressive-symptom trajectory was associated with a significantly increased risk of stroke across all three cohorts (CHARLS: Odds Ratio [OR] 2.56, 95% Confidence Interval [CI] 1.62-4.04; ELSA: OR 2.96, 95% CI 1.28-6.84; HRS: OR 1.56, 95% CI 1.03-2.36) compared to the "No" group. The "Decreasing" trajectory did not show a significant association with stroke, whereas the "Increasing" trajectory was linked to a moderate elevation in stroke risk in the ELSA and HRS cohorts. Older adults exhibiting persistently high or increasing depressive symptoms are at a significantly elevated risk of stroke. Continuous monitoring and early intervention targeting these high-risk depression trajectories may offer a novel strategy for primary stroke prevention.

Previous studies found that frailty was an important risk factor for cardiovascular disease (CVD). However, previous studies only focused on baseline frailty status, not taking into consideration the changes in frailty status during follow-up. The aim of this study was to investigate the associations of changes in frailty status with incident CVD. This study used data of three prospective cohorts: China Health and Retirement Longitudinal Study (CHARLS), English Longitudinal Study of Ageing (ELSA), and Health and Retirement Study (HRS). Frailty status was evaluated by the Rockwood frailty index and classified as robust, pre-frail, or frail. Changes in frailty status were assessed by frailty status at baseline and the second survey which was two years after the baseline. Cardiovascular disease was ascertained by self-reported physician-diagnosed heart disease (including angina, heart attack, congestive heart failure, and other heart problems) or stroke. Cox proportional hazard models were used to calculate the hazard ratio (HR) and 95% confidence interval (95% CI) after adjusting for potential confounders. A total of 7116 participants from CHARLS (female: 48.6%, mean age: 57.4 years), 5303 from ELSA (female: 57.7%, mean age: 63.7 years), and 7266 from HRS (female: 64.9%, mean age: 65.1 years) were included according to inclusion and exclusion criteria. The median follow-up periods were 5.0 years in the CHARLS, 10.7 years in the ELSA, and 9.5 years in the HRS. Compared with stable robust participants, robust participants who progressed to pre-frail or frail status had increased risks of incident CVD (CHARLS, HR = 1.84, 95% CI: 1.54-2.21; ELSA, HR = 1.53, 95% CI: 1.25-1.86; HRS, HR = 1.59, 95% CI: 1.31-1.92). In contrast, frail participants who recovered to robust or pre-frail status presented decreased risks of incident CVD (CHARLS, HR = 0.62, 95% CI: 0.47-0.81; ELSA, HR = 0.49, 95% CI: 0.34-0.69; HRS, HR = 0.70, 95% CI: 0.55-0.89) when compared with stable frail participants. These decreased risks of incident CVD were also observed in pre-frail participants who recovered to robust status (CHARLS, HR = 0.66, 95% CI: 0.52-0.83; ELSA, HR = 0.65, 95% CI: 0.49-0.85; HRS, HR = 0.71, 95% CI: 0.56-0.91) when compared with stable pre-frail participants. Different changes in frailty status are associated with different risks of incident CVD. Progression of frailty status increases incident CVD risks, while recovery of frailty status decreases incident CVD risks.

Baseline and changes in cardiometabolic index and incident cardiovascular disease in two prospective cohorts

Previous studies have shown that estimated glucose disposal rate (eGDR) is an important risk factor for heart disease. However, previous studies did not consider changes in eGDR during follow-up. This study aimed to investigate associations of changes in eGDR with incident heart disease. This study used data from 2 prospective cohorts: the China Health and Retirement Longitudinal Study (CHARLS) and the English Longitudinal Study of Ageing (ELSA). Changes in eGDR were assessed by eGDR at baseline and at the second survey using K-means clustering analysis, and the total eGDR was also calculated. Heart disease was assessed by self-reported physician-diagnosed heart disease (including angina, heart attack, congestive heart failure, and other heart problems). Cox proportional hazards models were used to calculate the hazard ratio (HR) and 95% confidence interval (95% CI) after adjustment for potential confounders. A total of 4929 participants from CHARLS (mean age: 58.6 years, male: 47.1%) and 2404 from ELSA (mean age: 63.3 years, male: 42.7%) were included for the main analyses. The median follow-up was 5.0 years in the CHARLS and 9.9 years in the ELSA. In the CHARLS, the HRs and 95% CIs for incident heart disease were 1.65 (1.29–2.11) for class 2 (high decreasing class), 1.39 (1.09–1.77) for class 3 (consistently moderate class), and 1.86 (1.49–2.31) for class 4 (consistently low class) compared with class 1 (consistently high class). In the ELSA, the HRs and 95% CIs for incident heart disease were 1.46 (1.12–1.91) for class 2, 1.45 (1.15–1.82) for class 3, and 1.68 (1.32–2.15) for class 4, compared to class 1. For total eGDR, lower levels of total eGDR were associated with an increased risk of incident heart disease in CHARLS and ELSA. A negative linear relationship between total eGDR and the risk of incident heart disease was observed using the multivariable adjusted restricted cubic spine models. Different changes in eGDR are associated with different risks of incident heart disease. Future research is needed to develop precise prevention strategies to maintain high levels of eGDR as well as tailored interventions to improve eGDR levels in cardiovascular practice.

Internet use and cognitive trajectories in middle-aged and older people: A longitudinal mixed effects analysis of two cohorts.

Bidirectional transitions of frailty states in middle-aged and older adults: Findings from two longitudinal cohorts studies.

Independent and combined effects of depressive symptoms and cardiometabolic risk factors on dementia incidence: a cross-country comparison in England, the United States and China.

ObjectiveOlder adults are more vulnerable to digital exclusion, which has been associated with psychological distress. This study investigated the relationship between digital exclusion and loneliness among older adults across three countries using three longitudinal surveys.Design and measurementsDigital exclusion was defined as self-reported non-use of the internet. Loneliness was assessed using the Three-Item Loneliness Scale (T-ILS). We employed Generalized Estimating Equations (GEE) with binary logistic regression and propensity score matching (PSM) to examine the association between digital exclusion and loneliness, adjusting for covariates including Age; Gender; Education; Marital status; Employment status; Cohabitation with children; Self-rated health; and Income.Setting and participantsNationally representative samples of older adults were obtained from three longitudinal studies: the China Health and Retirement Longitudinal Study (CHARLS), the Health and Retirement Study (HRS), and the English Longitudinal Study of Ageing (ELSA). The analysis included 39,190 participants (87,256 observations) across the three studies.ResultsSubstantial cross-national disparities in digital exclusion rates were observed: CHARLS (96.20%), HRS (52.13%), and ELSA (33.54%). In the fully adjusted model (Model 3), digital exclusion was significantly associated with loneliness in all three studies (CHARLS: OR = 1.22; HRS: OR = 1.16; ELSA: OR = 1.30). These associations remained statistically significant after propensity score matching (CHARLS: OR = 1.33; HRS: OR = 1.23; ELSA: OR = 1.23).ConclusionOur findings indicate that a substantial proportion of older adults experience digital exclusion, particularly in China. Digital exclusion demonstrates a positive association with loneliness, suggesting that enhancing digital inclusion may serve as a critical strategy for alleviating loneliness and mitigating psychological distress in ageing populations.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12877-025-06337-2.

Background and objectivesAs the global population ages, frailty and pain have become two significant health issues that impact the quality of life in older adults. Previous studies have not thoroughly explored the relationship between them. This study aims to investigate the longitudinal association between frailty and pain using data from prospective cohorts in China (CHARLS), the United Kingdom (ELSA), and the United States (HRS).MethodsThis study utilized data from three prospective cohort studies: the China Health and Retirement Longitudinal Study (CHARLS), the English Longitudinal Study of Ageing (ELSA), and the Health and Retirement Study (HRS). Frailty status was assessed using the Rockwood frailty index and categorized into robust, pre-frail, and frail. Pain was evaluated by self-reports. Pain degrees were categorized into mild, moderate and severe. Pain areas were grouped into four main areas: head and neck, trunk, limbs, oral. Generalized linear mixed-effects models were employed to analyze the longitudinal relationship between frailty and pain while adjusting for covariates, including gender, age, marital status, education level, sleep quality, smoking, drinking, hypertension, and diabetes.ResultsAccording to the inclusion and exclusion criteria, 10,624 participants from CHARLS (47% female, mean age: 60.76 years), 4945 participants from ELSA (52.2% female, mean age: 70.05 years), and 11,439 participants from HRS (55.8% female, mean age: 69.28 years) were included in the subsequent analysis. Compared to robust individuals, those in pre-frail and frail states showed a significantly increased risk of experiencing pain. In all three cohorts, pre-frail individuals had a 3.82-fold increased likelihood of pain compared to robust individuals (OR = 3.82, 95%CI = 3.51–4.15, p-value < 0.001, CHARLS), 4.29-fold (OR = 4.29, 95%CI = 3.74–4.93, p-value < 0.001, ELSA), and 4.17-fold (OR = 4.17, 95%CI = 3.81–4.57 p-value < 0.001, HRS). Frail individuals had a 10.44-fold increased likelihood of pain (OR = 10.44, 95%CI = 9.05–12.04, p-value < 0.001, CHARLS), 10.14-fold (OR = 10.14, 95%CI = 8.05–12.76, p-value < 0.001, ELSA), and 13.27-fold (OR = 13.27, 95%CI = 11.71–15.03, p-value < 0.001, HRS).ConclusionThis study demonstrates that frailty significantly impacts the risk of pain, the degree of pain, and the areas of pain. And this association is consistently observed across older populations in different countries. Future pain management strategies should incorporate frailty assessments to mitigate the adverse effects of pain on the health of older adults.

Background: Elevated triglyceride-glucose (TyG) related indices, such as TyG-body mass index, TyG-waist-to-height ratio, TyG-waist circumference was associated with higher risk of cardiovascular disease (CVD). Body roundness index (BRI) is superior to traditional anthropometric indices in predicting metabolic syndrome. However, the association between the TyG-BRI, as a new metabolic indicator, and CVD incidence and whether its predicting effect of CVD incidence is better than other TyG related indices remains unknown. Method: The datasets analyzed in our study were derived from two nationally representative prospective cohort studies: English Longitudinal Study of Ageing (ELSA) and China Health and Retirement Longitudinal Study (CHARLS). TyG was calculated as ln [TG (mg/dL) × FBG (mg/dL)/2]; The TyG-BRI index is determined by TyG index * BRI index. The participants were classified into four groups (Q1, Q2, Q3, and Q4) by the quartiles of TyG-BRI index. We performed Cox proportional hazards models after adjusting for potential confounders to analyze the association between the TyG-BRI index and CVD incidence. Restricted cubic spline models (RCS) were used to explore the non-linear relationship between TyG-BRI and CVD incidence. Receiver operating characteristic (ROC) curve analysis to evaluate and compare the predictive performance of TyG-BRI and TyG-related indices for CVD assessment Results: A total of 3,256 participants from ELSA (female: 54.0%, age more than 60: 63.7%) and 8,323 participants from CHARLS (female: 53.5%, age more than 60: 41.6%) were included in the analysis. The median follow-up periods were 12 years in the ELSA and 7 years in the CHARLS. After adjusting for potential confounding factors, the highest TyG-BRI group (Q4) had an increased risk of CVD compared with the Q1 group. (ELSA, HR 1.81, 95% CI 1.09–2.98; CHARLS, HR 1.57, 95% CI 1.32–1.87). An inverted U-shaped association was identified between TyG-BRI and CVD during the examination of nonlinear relationships (both P &lt;0.05). TyG - BRI has a higher AUC of 0.557 (95% CI: 0.531 - 0.583) in ELSA and AUC of 0.581 (95% CI: 0.565 - 0.596) in CHARLS than other TyG-related indices in predicting CVD incidence. Conclusions: Elevated TyG-BRI levels was associated with higher risk of incident CVD. TyG-BRI offers a new tool for early risk identification, and TyG-BRI had a better predictive ability than other TyG- related indices in predicting CVD incidence.