Association between sedentary time and intrinsic capacity among community-dwelling older adults: evidence from a prospective BLINDSCE cohort

Abstract

While existing studies have linked declines in intrinsic capacity (IC) to adverse health outcomes, the role of potentially modifiable lifestyle factors in this pathway, especially sedentary behavior, remains critically underexplored. Using data from the Beijing Longitudinal Disability Survey in Community Elderly (BLINDSCE) cohort (2023–2024), this study investigated both cross-sectional and 1-year longitudinal associations between sedentary time and IC in community-dwelling older adults. Of the 2044 participants (≥ 65 years) enrolled at baseline, 1576 completed 1-year follow-up assessments through face-to-face interviews. Multivariable linear regression analyses revealed that each additional hour of daily sedentary time was associated with a 1.18-point lower baseline IC score (95% CI: −1.38, −0.98) and an accelerated 0.48-point greater IC decline over 1 year (95% CI: −0.72, −0.24). Exposure-response analyses showed a linear relationship between sedentary time and 1-year IC change. Significant interaction effects were observed between sedentary time and baseline IC level, moderate-to-vigorous physical activity (MVPA), and daily internet use. These findings provide empirical support for reducing sedentary behavior, ensuring adequate MVPA, promoting moderate digital engagement, and implementing IC-stratified interventions to promote healthy aging—thereby operationalizing the WHO Integrated Care for Older People (ICOPE) framework through actionable public health measures.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1007/s11357-025-01983-1.

Introduction

Given the global aging population, the World Health Organization (WHO) recommends clinical screening for adults aged 65 years and older who are at risk, aiming to preserve functional capacity and social well-being as part of healthy aging strategies [1]. To address this, the WHO proposed the function-oriented concept of intrinsic capacity (IC)—defined as the composite of an individual’s physical and mental capacities at a given time—as a key indicator of healthy aging [1, 2]. While most studies on IC have linked its decline to adverse outcomes such as disability and mortality [3], the modifiable lifestyle determinants driving this deterioration remain poorly understood [4].

Limiting sedentary behavior and increasing physical activity (PA) are recognized as pivotal modifiable lifestyle factors for maintaining motor function and muscle strength in older adults, thereby promoting health [5, 6]. The WHO 2020 guidelines explicitly recommend minimizing sedentary behavior and enhancing moderate-to-vigorous PA (MVPA) among older adults [7]. However, these recommendations primarily rely on evidence derived from studies involving younger adults [7, 8], despite documented age-related variations in exposure-response relationships [9]. More importantly, the exposure-response relationship between sedentary time and health outcomes in older adults remains unclear, thereby impeding the establishment of evidence-based thresholds for this population [7].

Existing evidence associates sedentary behavior with detrimental impacts on specific IC domains—including locomotion [6], cognition [10, 11], and vitality [6]—as well as age-related conditions such as diabetes [12], sarcopenia [13], disability, and mortality [14]. Nevertheless, the relationship between sedentary time and overall IC remains inadequately elucidated. This research gap holds significant clinical relevance, as IC provides a comprehensive framework for health assessment and management in older adults, facilitating a paradigm shift from disease-centric to function-oriented care approaches [1]. To our knowledge, a single study has investigated the association between sedentary behavior and composite IC score; however, its generalizability may be constrained by the geographically restricted sample (Spain) and the absence of exposure-response analysis [15]. Further research is warranted to clarify this relationship across diverse aging populations.

Using a community-based cohort of older adults, this study aimed to investigate both cross-sectional and 1-year longitudinal associations between sedentary time and IC, including exposure-response relationships. Additionally, we assessed whether these associations were moderated by baseline IC, MVPA, and internet use [15, 16]. We hypothesized that prolonged sedentary time would be associated with poorer IC cross-sectionally and greater IC decline over 1 year, and that these relationships would be modified by baseline IC, MVPA, and internet use.

Methods

Study design and participants

The Beijing Longitudinal Disability Survey in Community Elderly (BLINDSCE) is an ongoing prospective cohort study designed to identify risk factors and develop predictive models for disability among community-dwelling adults aged 65 years and older from urban and rural areas of Beijing, China [17]. The baseline survey was conducted from August to December 2023, during which comprehensive face-to-face data collection was conducted, including assessments of sociodemographic characteristics, lifestyle factors, functional status, related outcomes, physical measurements, and blood draws. A 1-year follow-up assessment was conducted from July to December 2024, during which functional status and related outcomes were reassessed through face-to-face (77%) or telephone interviews (7%). This cohort study (NCT06863727) was approved by the Ethics Committee of Xuanwu Hospital, Capital Medical University (No. [2023]129), and written informed consent was obtained from all participants or their proxies.

The BLINDSCE cohort enrolled community-dwelling older adults (≥ 65 years) residing in Beijing. Exclusion criteria included severe mental disorders, serious medical conditions preventing study participation, and receipt of long-term professional treatment (e.g., hospitalization or rehabilitation) for physical and cognitive impairments. For the cross-sectional analysis between sedentary time and IC associations, we included 2044 participants with complete data on both sedentary time and IC, excluding those with extreme sedentary time values (> 15 h/d) or missing covariates. Among them, 1576 participants who completed 1-year face-to-face assessments were included for exploratory longitudinal analyses (Supplementary Fig. 1).

Assessment of sedentary time, PA, and internet use

Sedentary time was collected through the standardized questionnaire item “Approximately how long do you sit each day?” [10], with responses categorized into tertile-based groups. PA was evaluated using the short form of the International Physical Activity Questionnaire, which captures weekly time spent in vigorous-intensity activities, moderate-intensity activities, walking, and sitting. MVPA was calculated as the sum of weekly minutes spent in vigorous and moderate activities, categorized as ≥ 150 min/week (meeting WHO guidelines) or < 150 min/week [7]. Internet use frequency was self-reported using three categories: almost every day, almost every week, or not often. For this study, “internet use” specifically refers to daily engagement in digital activities including instant messaging, news browsing, video streaming, or online gaming—common digital behaviors among older adults [16].

Assessment of IC

Following WHO’s Integrated Care for Older People (ICOPE) framework, IC consists of five key domains (cognition, locomotion, psychology, vitality, and sensory) [18]. Each domain was measured using the validated instrument: (1) cognition—Montreal Cognitive Assessment Basic Version (MoCA-B; score range 0–30, higher is better) [19]; (2) locomotion—Short Physical Performance Battery (SPPB; range 0–12, higher is better) [20]; (3) psychology—5-item Geriatric Depression Scale (GDS-5; ranging 0–5, higher is worse) [21]; (4) vitality—grip strength measured using a Camry EH101 dynamometer (kg, maximum value from two measurements per hand) [20]; and (5) sensory—self-reported vision and hearing impairments (each impairment scored as 1, no impairment as 0; total range 0–2). Each domain score was rescaled from 0 (worst) to 100 (best) using the “percentage of maximum possible score” method, stratified by sex [22, 23]. To align the direction of effect, the original scores of psychological and sensory domains were weighted as −1 before rescaling. The composite IC score was calculated as the arithmetic mean of five domain scores, namely the IC mean score (hereafter termed the IC score). For the sensitivity analysis, an IC sum score was derived by summing dichotomized scores of five domains (range: 0 (worst) to 5 (best); see Supplementary Methods for details) [3].

Covariates

Potential confounders were identified and selected through a directed acyclic graph (DAG) approach, adhering to the “Evidence Synthesis for Constructing Directed Acyclic Graphs” guideline [24] and integrating findings from systematic reviews of sedentary and IC-related factors [25]. According to the DAG results (Supplementary Fig. 2), the following were identified as covariates for adjustment: demographic characteristics (age, sex, marital status, residency), socioeconomic factors (education, occupation, income level, social isolation), lifestyle behaviors (smoking, alcohol, sleep duration, internet use, MVPA), and health status (multimorbidity, polymedication, body mass index) (Supplementary Methods).

Statistical analysis

Descriptive statistics are presented as frequencies (percentages) for categorical variables and medians (interquartile ranges) for continuous variables. Multivariable linear regression models were used to examine both cross-sectional and longitudinal associations between sedentary time and IC. For exploratory longitudinal analysis, the outcome variable was defined as the 1-year change in IC score (follow-up score minus baseline score). The 1-year change in IC was regressed on sedentary time, adjusting for baseline IC scores and covariates identified using the DAG. This follow-up adjusted for baseline analysis helps mitigate reverse causation and provides an estimate approximating the direct causal effect [26] (Supplementary Methods). To assess multicollinearity, generalized variance inflation factors (GVIFs) were calculated for all variables in both baseline and longitudinal models. All GVIF values remained below 2.0 (Supplementary Table 1), suggesting negligible multicollinearity.

To further characterize the potential exposure-response relationships between sedentary time and IC scores, we used restricted cubic splines (RCS) with 3 knots, adjusting for all predefined covariates. Additionally, an RCS interaction model incorporating the baseline IC × sedentary time term was constructed to evaluate effect modification by the initial IC score. The model was implemented using an R package (“interactionRCS”), and the estimated β was derived from the corresponding linear regression model containing this interaction term. The associations between sedentary time and IC were also calculated, stratified by baseline IC level (< median vs. ≥ median) [27].

Subgroup analyses were conducted to examine potential effect heterogeneity across age, sex, residency, education, and multimorbidity—factors previously identified as associated with IC in older Chinese populations [16, 28]. Sensitivity analyses included the following: (1) reanalysis using an alternative IC operationalization—IC sum scores (aggregation of domain sum scores) [3]; (2) restriction to high-capacity individuals by removing individuals with severe to complete disability (WHO Disability Assessment Schedule 2.0 score > 12) and serious dementia (MoCA-B score < 10) [28–30], to minimize reverse causality; and (3) multiple imputation using R package “mice” to address attrition during follow-up and missing data on baseline IC and covariates. All models for subgroup analyses and sensitivity analyses were adjusted for all covariates. All statistical analyses were performed using R (version 4.3.3), with a two-sided significance level of P < 0.05.

Results

Characteristics of the study population

Among the 2044 participants at baseline, the median age was 72.0 (9.0) years, with 60.1% being female, 67.3% residing in urban communities, and 74.7% having multimorbidity. The median sedentary time was 3.0 (3.0) hours per day (h/d). At baseline, the median IC mean score was 68.3 (21.1) out of 100, and the median IC sum score was 3.0 (2.0) out of 5 (Table 1), suggesting a relatively high level of IC in the study population (Supplementary Fig. 3). Of these participants, 1576 (77.1%) completed the 1-year follow-up face-to-face assessment, and 1350 (66.1%) provided complete data across all five IC domains (Supplementary Fig. 1). While most demographic and socioeconomic characteristics were comparable between the cross-sectional and longitudinal samples, retained participants in the longitudinal analysis demonstrated healthier lifestyle patterns and higher functional capacity than those lost to follow-up (Supplementary Table 2).

Table 1.

Baseline characteristics of study participants

Characteristic	Sample (n = 2044)
Age, median (IQR), y	72.0 (9.0)
Female, no. (%)	1229 (60.1)
Married, no. (%)	1563 (76.5)
Urban residence, no. (%)	1376 (67.3)
Education, no. (%)
Below high school	633 (31.0)
High school or above	1411 (69.0)
Occupation, no. (%)
Employed	1411 (69.0)
Peasant	517 (25.3)
Unemployed	116 (5.7)
Income level, no. (%)
< 2000￥/mo	683 (33.4)
2000–5000￥/mo	553 (27.1)
5000–10000￥/mo	706 (34.5)
> 10,000￥/mo	102 (5.0)
Social isolation, no. (%)	1681 (82.2)
Current smoking, no. (%)	222 (10.9)
Current drinking, no. (%)	393 (19.2)
Sleep duration, median (IQR), h/d	7.0 (2.2)
Daily internet use, no. (%)	1050 (51.4)
MVPA, median (IQR), min/week	210.0 (840.0)
Multimorbidity, no. (%)	1527 (74.7)
Polymedication, no. (%)	632 (30.9)
BMI, median (IQR), kg/m2	25.1 (4.8)
Measurement for IC domaina
Cognition: MoCA-B, 0–30, median (IQR)	23.0 (9.0)
Locomotion: SPPB, 0–12, median (IQR)	11.0 (4.0)
Psychology: GDS-5, 0–5, median (IQR)	1.0 (2.0)
Vitality: GS, median (IQR), kg	21.1 (11.0)
Sensory, no. (%)
Vision deficit, no. (%)	1039 (50.8)
Hearing deficit, no. (%)	547 (26.8)
IC sum score, 0–5b, median (IQR)	3.0 (2.0)
IC mean score, 0–100b, median (IQR)	68.3 (21.1)
Sedentary time, median (IQR), h/d	3.0 (3.0)

Abbreviations: IQR, interquartile range; y, year; ￥/mo, Chinese yuan/month; h/d, hours/day; MVPA, moderate-to-vigorous physical activity; min/week, minutes/week; BMI, body mass index; IC, intrinsic capacity; MoCA-B, Montreal Cognitive Assessment Basic version; SPPB, Short Physical Performance Battery; GDS-5, 5-item Geriatric Depression Scale; GS, grip strength; kg, kilogram

^aHigher value indicates better function except for GDS-5 and Sensory

^bHigher value indicates better IC

Association between sedentary time and IC

Table 2 presents the cross-sectional and 1-year longitudinal associations between sedentary time and IC. In the fully adjusted model, each 1 h/d increase in sedentary time was associated with a 1.18 points (95% CI: −1.38, −0.98) reduction in IC scores at baseline and a 0.48 points (95% CI: −0.72, −0.24) acceleration in IC decline at the 1-year follow-up. Additionally, using ≤ 2 h/d as the reference group, the β values for participants with sedentary time between 2–4 h/d and > 4 h/d were −0.23 (−1.51, 1.05) and −4.79 (−6.14, −3.43), respectively, after adjustment for all covariates at the cross-sectional level. Longitudinally, both the 2–4 h/d and > 4 h/d groups exhibited significantly faster IC decline at the 1-year follow-up compared to the reference group (β = −1.63, −2.41, respectively, both P < 0.05).

Table 2.

Associations between sedentary time and IC mean scores

Sedentary time	Model 1a		Model 2b		Model 3c
	β (95% CI)	P value	β (95% CI)	P value	β (95% CI)	P value
Cross-sectional (n = 2044)
Per 1 h/d increase	−1.40 (−1.62, −1.18)	< 0.001	−1.28 (−1.49, −1.08)	< 0.001	−1.18 (−1.38, −0.98)	< 0.001
Tertile 1 (≤ 2 h/d)	Reference		Reference		Reference
Tertile 2 (2–4 h/d)	−0.44 (−1.92, 1.04)	0.562	0.47 (−0.85, 1.78)	0.487	−0.23 (−1.51, 1.05)	0.723
Tertile 3 (> 4 h/d)	−5.79 (−7.33, −4.24)	< 0.001	−5.01 (−6.40, −3.62)	< 0.001	−4.79 (−6.14, −3.43)	< 0.001
P for trend		< 0.001		< 0.001		< 0.001
Longitudinal (n = 1350)
Per 1 h/d increase	−0.45 (−0.69, −0.21)	< 0.001	−0.48 (−0.72, −0.24)	< 0.001	−0.48 (−0.72, −0.24)	< 0.001
Tertile 1 (≤ 2 h/d)	Reference		Reference		Reference
Tertile 2 (2–4 h/d)	−1.78 (−3.02, −0.54)	0.005	−1.51 (−2.75, −0.27)	0.017	−1.63 (−2.87, −0.38)	0.010
Tertile 3 (> 4 h/d)	−2.30 (−3.67, −0.92)	0.001	−2.36 (−3.74, −0.98)	< 0.001	−2.41 (−3.79, −1.03)	< 0.001
P for trend		< 0.001		< 0.001		< 0.001

Abbreviations: IC, intrinsic capacity; CI, confidence interval; h/d, hours/day

^aModel 1 adjusted for age and sex

^bModel 2 adjusted for factors in model 1 plus marriage, residence, education, occupation, income level, social isolation, smoking, drinking, internet use, sleep duration, multimorbidity, polymedication, and body mass index (BMI)

^cModel 3 adjusted for factors in model 2 plus moderate-to-vigorous physical activity (MVPA)

Association between sedentary time and the five IC domains

After adjusting for covariates, increased sedentary time was cross-sectionally associated with poorer function in all IC domains except sensory capacity, where a nonsignificant protective trend was observed. At 1-year follow-up, significant associations were observed between sedentary time and changes in three capacity domains: cognition, locomotion, and psychology. Although the longitudinal associations of sedentary time with vitality and sensory were not significant, the consistent negative estimates suggest a potential for adverse effects (Fig. 1 and Supplementary Table 3).

Fig. 1.

Associations between sedentary time and IC domains. A Cross-sectional associations and B longitudinal associations. Models were adjusted for all covariates. *P < 0.05. Abbreviations: IC, intrinsic capacity; CI, confidence interval

Exposure-response relationship between sedentary time and IC

RCS analyses delineated the exposure-response relationship between sedentary time and IC (Fig. 2A, B). After full adjustment, the cross-sectional association exhibited a nonlinear, inverted J-shaped curve (P for nonlinearity < 0.001), with IC scores plateauing below 3 h/d of sedentary time and declining sharply beyond this threshold. Conversely, the 1-year longitudinal association was linear and exposure-dependent (P for nonlinearity = 0.332). Furthermore, the association between sedentary time and IC change was significantly modified by baseline IC scores (P for interaction = 0.006), as evidenced in Fig. 2C. Specifically, the adverse effect of sedentary time was pronounced in participants with low baseline IC (< 68.3 scores) but not significant in those with high baseline IC (≥ 68.3 scores) (P for interaction < 0.05) (Supplementary Table 4).

Fig. 2.

Exposure-response associations between sedentary time and IC mean scores. A Cross-sectional association, B longitudinal association, and C interaction of baseline IC on the longitudinal association. Models were fit using cubic restricted splines and adjusted for all covariates, with three knots at the 5th, 50th, and 95th percentiles. Abbreviations: IC, intrinsic capacity; CI, confidence interval; h/d, hours/day

Interaction of internet use and MVPA with sedentary time

There was an interaction between internet use and sedentary time in the fully adjusted model (P-interaction < 0.001 for cross-sectional analysis; P-interaction = 0.060 for 1-year longitudinal analysis, in Supplementary Table 5). As shown in Fig. 3, the adverse association between sedentary time and IC was more pronounced in non-daily internet users compared to daily users in both cross-sectional and 1-year longitudinal analyses. A similar moderating effect was found for MVPA (P-interaction < 0.001 for cross-sectional analysis; P-interaction = 0.078 for 1-year longitudinal analysis, in Supplementary Table 5). The association between sedentary time and IC scores varied according to MVPA status both cross-sectionally and at 1-year follow-up (Fig. 3). Sedentary time exhibited an adverse association with IC among participants not meeting MVPA recommendations (< 150 min/week), whereas no significant association was observed in those meeting the recommended MVPA level (≥ 150 min/week).

Fig. 3.

Associations between sedentary time and IC mean scores stratified by different statuses of internet use and MVPA. A Internet use-stratified analysis and B MVPA-stratified analysis. Model adjusted for all covariates. *P < 0.05. Abbreviations: IC, intrinsic capacity; CI, confidence interval; MVPA, moderate-to-vigorous physical activity; min/week, minutes/week

Subgroup analyses and sensitivity analyses

Subgroup analyses showed that the associations between sedentary time and IC were generally consistent across stratifications by sex, education level, and multimorbidity status (Supplementary Fig. 4). Notably, stratified analyses revealed that older individuals (aged ≥ 75 years) and urban residents experienced greater harmful effects of sedentary time on IC compared to younger and rural individuals both cross-sectionally and longitudinally (all P for interaction < 0.05 for age and urban-rural status). Additionally, the primary findings demonstrated robustness under various sensitivity analyses, including the use of an alternative IC sum score, restriction to high-functioning participants, and multiple imputation for missing data (Supplementary Tables 6–10; Supplementary Figs. 5–7).

Discussion

This study demonstrated significant adverse associations between prolonged sedentary time and IC among community-dwelling older adults at both cross-sectional and 1-year longitudinal levels. Exposure-response analyses revealed that sedentary time exhibited a dose-response relationship with 1-year IC decline, indicating progressive IC deterioration with increasing sedentary duration. The detrimental association of sedentary time was attenuated among participants with high baseline IC, those meeting MVPA recommendations, or those reporting daily internet use. Taken together, these findings underscore the critical importance of sedentary behavior reduction, adequate MVPA maintenance, moderate digital engagement, and IC-stratified preventive measures for healthy aging.

Our findings indicate that a 1-h/d increase in sedentary time was associated with a 1-year IC decline of 0.48 scores. Although no established threshold exists for defining clinically meaningful IC deterioration, previous studies have shown that each 1-score decline in IC is associated with a 7–10% higher risk of disability [30] and a 5% higher mortality risk [31, 32] in this population. Based on these estimates, we speculate that each additional hour of daily sedentary time may be associated with an approximately 3.5% increase in the risk of adverse health outcomes later in life. Domain-specific analyses revealed consistent adverse associations between sedentary time and four core IC domains—locomotion, cognition, psychology, and vitality—both cross-sectionally and longitudinally, in line with existing evidence linking sedentary behavior to declines in physical performance and cognitive function [6, 10, 11]. However, no clear association was observed with sensory capacity, implying that it may represent a more distal component of IC compared to the other four domains [23]. Further research is needed to elucidate the hierarchy of IC domains.

This study is the first to explore the exposure-response relationship between sedentary time and IC. Existing research on the associations between sedentarism and aging-related markers has yielded inconsistent findings, with some studies reporting linear associations [10, 14, 33] and others reporting nonlinear associations [11, 12, 34]. Similarly, our study yielded divergent results: a nonlinear association cross-sectionally and a linear negative association longitudinally. This discrepancy may stem from differences across study designs. The cross-sectional findings may be influenced by reverse causality and between-individual differences. For instance, high-functioning individuals may naturally maintain moderate PA, leading to better IC scores at intermediate sedentary levels. In contrast, the longitudinal analysis, which examined the relationship between baseline sedentary time and subsequent IC change, controlled for time-invariant individual factors and therefore provided a more causal effect of sedentary behavior. This relationship was linear, suggesting that any increase in sedentary time is associated with 1-year IC decline, with no safe threshold, consistent with findings on the relationship of objectively measured sedentary time with all-cause mortality and chronic kidney diseases [14, 33]. In comparison, other studies have proposed specific thresholds, such as 3–4 h/d for mild cognitive impairment (MCI) [10], diabetes [12], and all-cause death [14] in older adults, and 6–10 h/day for cardiovascular disease, mortality [34], or dementia [11] in middle-aged or female populations. These inconsistencies may arise from heterogeneity in study populations, assessment of sedentary time (self-reported or objectively measured), statistical methodology (adjusted for PA or not), and outcome definitions. To date, there is no consensus on the existence or value of a sedentary behavior threshold, highlighting a critical gap for future research.

Another finding of our study was the modifying role of MVPA and internet use on the relationship between sedentary time and IC, aligning with previous reports [15, 35]. Our finding—meeting the WHO-recommended MVPA threshold (≥ 150 min/week) attenuated the harmful effect of sedentary behavior—reinforces and extends the evidence of the WHO 2020 guidelines on MVPA and sedentary time to a sample of community-dwelling older adults with a high prevalence of multimorbidity [7]. Furthermore, we provide novel evidence that daily internet use—such as online learning, instant messaging, and e-shopping—can mitigate the adverse association between prolonged sedentary time and IC. This finding aligns with evidence on the beneficial role of internet use for health markers [16, 36], thereby supporting digital engagement as a practical strategy to preserve IC, especially among older individuals with limited mobility or access to PA [36]. Notably, the observed associations between lifestyle behaviors and IC remained statistically significant after adjusting for social isolation, a prevalent and important factor influencing lifestyle behavior and health [37, 38]. This suggests that these associations are partly independent of social isolation, with potential mechanisms including physiological benefits of regular exercise, cognitive stimulation from online activities, mental health promotion, and prevention of chronic conditions [39–41].

Collectively, we recommend sedentary time reduction, sufficient MVPA, and routine daily internet use as cost-effective and accessible interventions to alleviate IC decline. Notably, the detrimental effect of sedentary behavior was more pronounced in older adults with low baseline IC, identifying this group as a high-risk population meriting prioritized interventions. These findings align with and expand upon the WHO ICOPE framework by underscoring the importance of early IC assessment for risk stratification and the necessity for personalized care pathways [2, 18]. Consequently, healthcare providers must integrate IC monitoring with tailored guidance on physical activity and digital engagement to promote healthy aging in community settings.

Of note, no consensus exists regarding the operationalization of IC. Studies vary widely in their selection of domains, choice of indicators, assignment of domain weights, and methods for constructing a composite IC score. In the present study, we adopted the domains recommended in the WHO ICOPE guidelines and assessed each using validated instruments. However, limitations exist, such as reliance on self-reported measures for the sensory domain. Regarding IC scoring, while some studies derive factor scores by assigning weights to domains based on sample characteristics [42], others—including this study—calculate IC as an average or a sum of all domains using approaches such as percentage of maximum possible score or categorical variables by clinical cutoffs and summing [15, 23]. Although this unweighted method does not account for varying domain contributions, it offers a transparent, reproducible, and sample-independent means of generating a standardized IC score [22]. Nevertheless, potential within-domain clustering or interactive effects should not be overlooked [23]. Further research is needed to elucidate the complex interrelationships among IC domains.

Several limitations should be acknowledged. First, the generalizability of our findings may be limited by the specific demographics of our study cohort, which consisted exclusively of Chinese individuals, predominantly from urban areas, and had a high prevalence of multimorbidity. Second, the study population exhibited relatively high baseline IC levels—consistent with other community-based cohorts [23]. This may introduce selection bias due to the exclusion of participants with incomplete data or loss to follow-up, as those with lower IC are less likely to complete all assessments. Third, sedentary time and PA were assessed via self-report questionnaires, which are subject to recall bias and may underestimate true sedentary exposure [43]. Any resulting misclassification would probably attenuate the observed associations and, more importantly, obscure the true exposure-response relationship by potentially flattening the dose-response curve and blurring any sharp risk threshold [43, 44]. Fourth, the analysis did not account for domain-specific sedentary behaviors, types of internet use, or their longitudinal changes. Future research should investigate how different patterns and contexts of sedentary behavior influence IC trajectories and distinguish the potentially distinct roles of active (e.g., learning, communication) versus passive (e.g., video watching) internet engagement. Fifth, although we adjusted for MVPA and sleep duration, the interrelated feature of movement behaviors warrants further investigation using compositional data analysis to explore how time reallocation across the entire day influences IC. Additionally, while we adjusted for key confounders and conducted sensitivity analyses, residual confounding or reverse causality cannot be fully ruled out. Finally, the 1-year follow-up period is relatively short, but the ongoing longitudinal follow-up with methodological enhancements will address this limitation.

Conclusions

In conclusion, longer sedentary time was associated with poorer IC both cross-sectionally and over a 1-year follow-up in community-dwelling older adults. Exposure-response analyses indicated a linear relationship between sedentary time and 1-year IC change. These associations were moderated by baseline IC level, MVPA, and daily internet use. The findings underscore the need to prevent IC deterioration through optimized lifestyle strategies and IC-stratified interventions, aligning with and expanding the WHO ICOPE framework. However, generalizing these results to other populations requires caution due to the study’s focus on Chinese, predominantly urban participants with high multimorbidity rates.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contribution

Concept and design: CSQ, GYS. Acquisition, analysis, or interpretation of data: all authors. Drafting of the manuscript: CSQ. Critical review of the manuscript for important intellectual content: CSQ, GYS, LSY, HCB. Statistical analysis: CSQ, LSY, HCB. Obtained funding: GYS, YXL. Administrative, technical, or material support: GYS, YXL. Supervision: GYS. All authors read and approved the final manuscript.

Funding

This research was funded by Capital’s Funds for Health Improvement and Research [No: 2024-2G-20112] and Project for Innovation and Development of Beijing Municipal Geriatric Medical Research Center [No: 11000025T000003320658]. The funder had no role in the study design, data collection, analysis, interpretation of data, or manuscript writing.

Data availability

The data that support the findings of this study are not publicly available due to ethical restrictions but are available from the corresponding authors on reasonable request.

Declarations

Ethics approval

The study was approved by the Ethics Committee of Xuanwu Hospital, Capital Medical University (reference number: [2023]129). All participants (or their proxies) provided written informed consent. This trial was registered at ClinicalTrials.gov (identifier: NCT06863727).

Competing interests

The authors declare no competing interests.