Associations of leisure-time physical activity with muscle strength and functional fitness domains among community-dwelling older adults: a cross-sectional study

Abstract

Background

Declines in muscle strength and functional fitness in older adults are associated with multiple adverse outcomes, yet the associations between leisure-time physical activity(LTPA) and multidimensional objective functional measures remain unclear.

Methods

The linked source sample comprised 466 community-dwelling older adults who completed both the questionnaire survey and physical fitness assessments and were successfully matched at the individual level. Analyses were conducted using analysis-specific complete-case samples, with 418 participants included in the primary grip strength model. The primary exposure was total weekly LTPA, and the secondary exposure was average daily sedentary time (ST). The primary outcome was grip strength, and the secondary outcomes included the 30-s chair stand test, eyes-closed balance, choice reaction time, and a composite functional fitness Z score. Ordinary least squares (OLS) linear regression with HC3 robust standard errors was used, with sequential adjustment for sociodemographic, lifestyle, and health-related covariates. LTPA tertile analyses and sensitivity analyses were also conducted.

Results

In the fully adjusted model, each additional 60 min/week of LTPA was associated with higher grip strength (β = 0.064 kg, 95% confidence interval (CI) 0.009 to 0.120). This association persisted after further adjustment for ST, and the direction of the association remained broadly consistent after excluding current smokers or participants with a greater chronic disease burden. In contrast, no clear independent associations were observed between LTPA and the 30-s chair stand test, eyes-closed balance, choice reaction time, or composite functional fitness Z score. Tertile analyses showed no consistent trends. ST was likewise not independently associated with grip strength or any functional fitness measure.

Conclusions

Among community-dwelling older adults, higher LTPA levels were associated with slightly greater grip strength, although the absolute effect size was small and its clinical significance remains uncertain. ST was not consistently associated with grip strength or functional fitness measures in this study. Across objective functional outcomes in older adults, the association of LTPA may be more pronounced for the muscle strength domain than for other functional fitness domains.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12877-026-07733-y.

Background

As global population ageing accelerates, gains in life expectancy have not been matched by comparable gains in healthy life expectancy, and declining skeletal muscle function has become a major underlying contributor to health deterioration in later life [1]. Reduced muscle strength and impaired functional capacity are closely associated with adverse outcomes, including falls and disability [2]. Accordingly, muscle strength and physical performance have become key components of geriatric assessment. The 2019 Asian Working Group for Sarcopenia (AWGS) consensus recognized low muscle strength and poor physical performance as key elements in the diagnostic pathway for sarcopenia, with handgrip strength and chair-stand performance as commonly used indicators [3]. The AWGS 2025 update further reframed sarcopenia within a broader muscle-health perspective, distinguishing muscle strength from physical performance outcomes [4]. This framework supports our separate evaluation of grip strength and functional fitness domains in older adults. Among the available strength measures, grip strength is widely regarded as a practical surrogate of overall muscle strength and functional reserve because it is simple, inexpensive, and reasonably reproducible. Previous studies have shown that lower grip strength is associated with higher mortality risk and adverse cognitive outcomes [5–7]. However, functional status in older adults cannot be adequately captured by grip strength alone. The 30-s chair stand test, balance, and reaction time reflect different dimensions of function, including lower-extremity performance, postural control, and cognitive-motor processing, respectively. It is therefore necessary to examine both muscle strength and functional fitness in parallel [8–10].

Physical activity is considered one of the most important and modifiable behavioral determinants of healthy ageing [11]. A systematic review and meta-analysis including 124 studies and 230,174 older adults found that total physical activity and moderate-to-vigorous physical activity were associated with a lower risk of sarcopenia, although the associations between different movement behaviors and individual sarcopenia components were not consistent [12]. Previous studies have often assessed physical activity according to total volume or intensity, whereas fewer studies have examined LTPA separately in relation to objective muscle-strength and functional-fitness outcomes. In older adults, particularly those no longer in paid work, LTPA may partly overlap with total physical activity; however, it remains a useful behavioral exposure because it reflects voluntary activity that can be promoted in community settings [13]. A study of community populations from six low- and middle-income countries showed that low LTPA was associated with sarcopenia, with an adjusted odds ratio of 1.85 (95% confidence interval (CI) 1.29–2.65) [14]. A cross-sectional analysis from PREDIMED-Plus likewise reported that moderate-to-vigorous LTPA was associated with a lower prevalence of sarcopenia and better lower-extremity strength, whereas no consistent association was observed for light-intensity activity [15]. At the same time, the magnitude of the association between physical activity and functional outcomes may vary across domains. Previous studies suggest that leisure activity is associated with grip strength, lower-extremity strength, and aerobic endurance, but not significantly associated with some other fitness dimensions [16]. The independent role of ST also remains uncertain. A systematic review and meta-analysis published in BMC Geriatrics found that sedentary behavior was independently associated with sarcopenia (pooled odds ratio 1.36, 95% CI 1.18–1.58), whereas another systematic review of community-dwelling older adults concluded that evidence on interrupting ST in relation to health outcomes remains limited and that the effects of sedentary behavior are not easily disentangled from those of physical activity [17].

Although previous studies suggest that physical activity is related to muscle health and functional status in older adults, relatively few studies in community settings have simultaneously used quantifiable LTPA as the exposure and examined grip strength separately from multidimensional objective functional fitness indicators. Existing evidence has largely been framed around sarcopenia as a global construct, total physical activity, or a single functional outcome. One study from six low- and middle-income settings focused primarily on the association between low LTPA and sarcopenia [13], a longitudinal study from Korea emphasized physical activity in relation to maintenance of grip strength [18], and a community-based study from Taiwan examined overall physical activity patterns in relation to physical function [19]. Even when different activity types have been considered, the strength of their associations with grip strength, lower-extremity strength, and aerobic endurance has not been consistent, and neuromotor indicators such as reaction time have rarely been included [15]. Research on sedentary behavior has likewise focused mainly on isotemporal substitution models or sarcopenia outcomes, and its independent associations with LTPA, muscle strength, and multidimensional functional fitness in community populations remain incompletely understood [20]. Even in more recent studies that have incorporated both physical activity and functional fitness measures, exposures have generally been based on total physical activity and outcomes have often relied on composite indicators, with less frequent parallel evaluation of grip strength, chair stand performance, balance, and reaction time [21]. Examining the associations of quantifiable LTPA with these objective functional indicators in community-dwelling older adults may therefore help identify more precisely the functional domains most closely related to LTPA.

Against this background, the present study examined the associations of LTPA with muscle strength and individual functional fitness domains in community-dwelling older adults. We focused primarily on the association between LTPA and grip strength and further evaluated its associations with the 30-s chair stand test, eyes-closed balance, choice reaction time, and a composite functional fitness score. We also treated ST as a secondary exposure to explore its independent associations with these outcomes. Based on previous evidence and the aims of the present study, we hypothesized that higher levels of LTPA would be associated with better muscle strength and functional fitness, whereas longer ST would be associated with poorer functional status. By examining the cross-sectional associations of LTPA and ST with muscle strength and multidimensional functional fitness in community-dwelling older adults, this study seeks to provide evidence to inform functional maintenance and health promotion in later life.

Methods

Study design and setting

This was a cross-sectional study [22] conducted among community-dwelling older adults in Shihezi, an inland city in northwestern China. The study data were derived from two databases established during the same survey cycle and linked at the individual level: an older adult questionnaire database and an older adult physical fitness testing database. Data were collected from June to November 2025. The questionnaire database provided information on demographic characteristics, lifestyle, health status, fall risk, and cognitive status, whereas the physical fitness testing database provided objective measures of muscle strength and functional fitness. All community-dwelling older adults who completed both the questionnaire survey and physical fitness testing during the study period were eligible for database linkage and constituted the linked source sample after individual-level data integration.

Participants and analytic sample

After individual-level linkage of the questionnaire database and the physical fitness testing database, 466 unique community-dwelling older adults were successfully matched and constituted the initial linked source sample. Eligibility for database linkage required meeting the study age criterion for older adults and completing both the questionnaire survey and the physical fitness testing during the study period. Duplicate records were checked before analysis, and no duplicate study IDs were identified.

Before analysis, values coded as missing or invalid were treated as missing according to the data-cleaning procedures used in this study. Analyses were then performed using complete cases for the variables required in each specific model, and no missing-data imputation was conducted. Accordingly, the final analytic sample size differed across analyses: 416 participants were included in the baseline comparisons by LTPA tertiles, 418 in the primary grip strength model, 417 in the models for the other functional fitness outcomes, and 431 in the secondary analyses of ST.

The linked source sample comprised community-dwelling older adults aged 60 years and above who were living independently in urban and rural communities in Shihezi during the study period. Participants were recruited through community health centers and residential committees. Eligibility criteria included: (1) age ≥ 60 years; (2) ability to complete the questionnaire survey independently or with assistance; (3) ability to participate in physical fitness testing without contraindications; and (4) provision of written informed consent. Exclusion criteria included: (1) severe cognitive impairment preventing informed consent or questionnaire completion; (2) acute illness or unstable chronic conditions contraindicating physical fitness testing; (3) severe mobility limitations preventing participation in fitness assessments; and (4) incomplete data linkage between the questionnaire and physical fitness databases.

The main item-level missing or invalid values in the linked source sample were observed for LTPA (n = 24), sleep duration (n = 25), ST (n = 14), grip strength (n = 3), the 30-s chair stand test (n = 3), eyes-closed balance (n = 1), choice reaction time (n = 2), and the composite functional fitness Z score (n = 4); these counts were not mutually exclusive.

Assessment of LTPA and ST

Leisure-time physical activity

LTPA was the primary exposure and was assessed using questionnaire items adapted from the International Physical Activity Questionnaire (IPAQ) [23, 24]. Participants reported the frequency (days per week) and duration (minutes per session) of light-, moderate-, and vigorous-intensity LTPA during a typical week in the past month. Light-intensity activities included slow walking and stretching; moderate-intensity activities included brisk walking, tai chi, and recreational cycling; vigorous-intensity activities included jogging, swimming, and ball sports. For each intensity level, weekly activity time was calculated as frequency × duration, then summed across intensity levels to derive total weekly LTPA time (min/week).

LTPA was treated as a continuous variable in the primary analysis, with each 60 min/week increment entered into regression models. In descriptive analyses, LTPA was further categorized into tertiles based on the distribution in the analytic sample for baseline comparisons and trend analyses.The continuous specification was prespecified as the primary analysis because LTPA was originally measured in minutes per week, and this approach preserves exposure information and avoids arbitrary cut-points.

Sedentary time

ST was the secondary exposure and was assessed using a questionnaire item asking participants to estimate the total time spent sitting or reclining during waking hours on a typical day, including time spent watching television, reading, using a computer, and sitting during transportation [24, 25]. Responses were recorded in hours and minutes and converted to average daily sedentary time (min/day). ST was modeled as a continuous variable, with each 60 min/day increment entered into regression models. ST was included only in supplementary analyses and was not treated as a co-primary exposure alongside LTPA.

Assessment of muscle strength and functional fitness dimensions

Primary outcome: muscle strength

The primary outcome was muscle strength, assessed objectively using handgrip strength. Grip strength testing was performed using a calibrated hand dynamometer following standardized procedures [26, 27]. Participants stood upright with their feet naturally apart and their arms hanging naturally at the side of the body. The handle spacing of the dynamometer was adjusted to a comfortable position before testing. Participants held the dynamometer with their stronger hand and squeezed it with maximal effort. The test was performed twice, and the maximum value was recorded in kilograms (kg).

Secondary outcomes: functional fitness dimensions

The secondary outcomes were the 30-s chair stand test, eyes-closed balance, choice reaction time, and a composite functional fitness Z score, all obtained from standardized physical fitness testing. The 30-s chair stand test was performed following the protocol described by Jones et al. [27], which has demonstrated good reliability in older adults. Participants sat on a standard chair (seat height 43 cm) with arms folded across the chest and were instructed to stand up and sit down as many times as possible within 30 s. The number of completed repetitions was recorded. Eyes-closed balance was assessed by asking participants to stand on one leg with eyes closed for as long as possible, with duration recorded in seconds (maximum 60 s). Choice reaction time was measured using a standardized reaction-time testing device [28, 29]. During the test, participants placed the middle finger on the start key and waited for the signal. When a signal key was activated with auditory and visual cues, participants pressed the corresponding key as quickly as possible. The test included five signal responses and was performed twice, with the best result recorded in seconds. Lower values indicated faster reaction time and better cognitive-motor response performance [30, 31].

Composite functional fitness Z score

The composite functional fitness Z score was calculated using grip strength, the 30-s chair stand test, eyes-closed balance, and choice reaction time. For each component, sex-specific Z scores were calculated using the mean and sample standard deviation among participants with non-missing data for that component in the linked source sample: Because lower values of choice reaction time indicate better performance, the standardized reaction time score was reverse-coded before inclusion in the composite score. The composite score was then calculated as the mean of the four standardized component scores:

Participants with missing values in any of the four component tests were not assigned a composite functional fitness Z score. Higher composite scores indicated better overall functional fitness.

Covariates

Covariates were selected a priori based on previous literature and data availability. Sociodemographic covariates included age, sex, educational level (primary or below, middle/high school, college or above), marital status (married/cohabiting vs. other), living arrangement (living alone vs. living with others), and urban-rural residence. Lifestyle covariates included sleep duration, current smoking, and current alcohol consumption. Health-related covariates included the number of self-reported physician-diagnosed chronic diseases, fall risk score, and cognitive screening score.

The fall risk score and cognitive screening score were derived from questionnaire-based items by summing affirmative responses; higher scores indicated greater fall-related risk and more cognitive-related problems, respectively. The cognitive screening score was used as a screening indicator rather than as a clinical diagnosis of cognitive impairment. Both scores were included in the regression models as continuous covariates. Sparse categorical levels were combined when necessary to ensure stable model estimation.

Statistical analysis

Continuous variables are presented as mean ± standard deviation; for variables with skewed distributions, the median (P25, P75) is also reported. Categorical variables are presented as number (%). Baseline characteristics were summarized according to LTPA tertiles. For between-group comparisons, one-way analysis of variance was used for approximately normally distributed continuous variables, the Kruskal-Wallis test for skewed variables, and the chi-square test or Fisher’s exact test for categorical variables. All analyses were based on complete-case analysis, with no imputation of missing values. Because the extent of missingness differed slightly across outcome variables, the final sample size varied modestly across models.

The primary analysis used ordinary least squares (OLS) linear regression to examine the association between LTPA and grip strength, with grip strength as the primary outcome and LTPA as the primary exposure, modeled per 60 min/week increment. To improve the robustness of the estimates, heteroskedasticity-consistent HC3 robust standard errors were used [32]. Covariates were introduced sequentially in prespecified models: the crude model was unadjusted; Model 1 adjusted for age and sex; Model 2 further adjusted for educational level, marital status, living arrangement, and urban-rural residence; and Model 3 further adjusted for sleep duration, current smoking, current alcohol consumption, number of chronic diseases, fall risk score, and cognitive screening score. LTPA tertiles were also used for baseline comparisons and supplementary trend tests, but the main conclusions were based on the continuous exposure models.

Secondary analyses used the same multivariable modeling strategy to evaluate the associations of LTPA with the 30-s chair stand test, eyes-closed balance, choice reaction time, and the composite functional fitness Z score. ST was included as a secondary exposure, modeled per 60 min/day increment, for supplementary analyses of the same outcomes. To assess the robustness of the primary findings, ST was additionally included in the main grip strength model for mutual adjustment, and the primary analysis was repeated after excluding current smokers and after excluding participants with three or more chronic diseases. Supplementary models and trend tests based on LTPA tertiles were also conducted. All tests were two-sided, and statistical significance was defined as P < 0.05. Diagnostic checks were conducted for the fully adjusted OLS models. Linearity and model specification were assessed by adding a quadratic term for the exposure variable to each fully adjusted model. Residual normality was evaluated using the Jarque-Bera test, and heteroskedasticity was assessed using the Breusch-Pagan test. Multicollinearity was assessed using the variance inflation factor. Influential observations were examined using Cook’s distance, including the maximum Cook’s distance and the number of observations with Cook’s distance greater than 4/n. Because departures from homoscedasticity and residual normality may occur in observational data, all main OLS models used HC3 robust standard errors, and the estimates were interpreted with appropriate caution. All statistical analyses were performed in Python, primarily using the pandas and statsmodels packages.

Ethical considerations

The study protocol was reviewed and approved by the Ethics Committee of Shihezi People’s Hospital (approval No. KYLL No. (2015) 11). All participants provided written informed consent before the questionnaire survey and physical fitness testing. The study was conducted in accordance with the Declaration of Helsinki and relevant ethical standards.

Results

Derivation of the analytic sample

A total of 466 unique participants were successfully linked between the questionnaire and physical fitness testing databases and formed the initial linked source sample. After applying analysis-specific complete-case criteria, 416 participants were available for baseline comparisons by LTPA tertiles, 418 for the primary analysis of LTPA and grip strength, 417 for the analyses of LTPA and the other functional fitness outcomes, and 431 for the secondary analyses of ST.

Participant characteristics

For the baseline comparisons by LTPA tertiles, 416 older adults were included and categorized into three groups: 208 in T1, 83 in T2, and 125 in T3. As shown in Table 1, demographic characteristics, lifestyle factors, and most health-related indicators were broadly comparable across the LTPA groups. No statistically significant between-group differences were observed for age, sex, education, marital status, living arrangement, residence, sleep duration, smoking status, alcohol consumption, number of chronic diseases, cognitive screening score, or ST. Fall risk score differed across the three groups (P = 0.029). The mean age of the sample was 66.1 years, and 59.9% were female. The majority of participants resided in rural areas (93.3%).

Table 1.

Selected baseline characteristics of participants according to tertiles of LTPA

Characteristic	Overall (n = 416)		T1 (≤ 0 min/wk) (n = 208)	T2 (0–420 min/wk) (n = 83)	T3 (> 420 min/wk) (n = 125)	P value
Sociodemographic characteristics
Age, years		66.05 ± 4.80	66.12 ± 4.94	66.64 ± 4.54	65.54 ± 4.71	0.262
Sex, n (%)
Male		167 (40.1)	79 (38.0)	34 (41.0)	54 (43.2)	0.633
Female		249 (59.9)	129 (62.0)	49 (59.0)	71 (56.8)
Education level, n (%)
Primary or below		95 (22.8)	48 (23.1)	24 (28.9)	23 (18.4)	0.109
Middle/High school		234 (56.2)	125 (60.1)	40 (48.2)	69 (55.2)
College or above		87 (20.9)	35 (16.8)	19 (22.9)	33 (26.4)
Marital status, n (%)
Married/cohabiting		360 (86.5)	181 (87.0)	66 (79.5)	113 (90.4)	0.076
Other		56 (13.5)	27 (13.0)	17 (20.5)	12 (9.6)
Living arrangement, n (%)
Living alone		48 (11.5)	21 (10.1)	14 (16.9)	13 (10.4)	0.236
Living with others		368 (88.5)	187 (89.9)	69 (83.1)	112 (89.6)
Residence, n (%)
Urban		28 (6.7)	20 (9.6)	3 (3.6)	5 (4.0)	0.063
Rural		388 (93.3)	188 (90.4)	80 (96.4)	120 (96.0)
Lifestyle and health-related characteristics
Sleep duration, h/day		7.59 ± 1.14	7.59 ± 1.17	7.71 ± 1.21	7.51 ± 1.04	0.472
Current smoker, n (%)
No		355 (85.3)	175 (84.1)	68 (81.9)	112 (89.6)	0.243
Yes		61 (14.7)	33 (15.9)	15 (18.1)	13 (10.4)
Current drinker, n (%)
No		318 (76.4)	158 (76.0)	65 (78.3)	95 (76.0)	0.904
Yes		98 (23.6)	50 (24.0)	18 (21.7)	30 (24.0)
No. of chronic diseases		1.06 ± 1.23	1.05 ± 1.18	1.29 ± 1.52	0.91 ± 1.08	0.095
Fall risk score		0.75 ± 1.38	0.59 ± 1.26	1.05 ± 1.61	0.83 ± 1.38	0.029
Cognitive screening score		0.42 ± 0.92	0.39 ± 0.83	0.42 ± 0.91	0.47 ± 1.05	0.73
ST, min/day		160.12 ± 101.19	165.99 ± 106.97	152.89 ± 90.22	155.16 ± 98.31	0.492

Notes: Values are presented as mean ± standard deviation or n (%)

Fall risk score and cognitive screening score were questionnaire-derived summary scores and were analyzed as continuous variables; higher scores indicate higher fall risk and a greater number of cognitive-related problems, respectively

The cognitive screening score was used as a screening indicator rather than as a clinical diagnosis of cognitive impairment

P values were derived from one-way ANOVA or χ² test, as appropriate

LTPA and muscle strength

Among participants with available grip strength data, LTPA modeled as a continuous variable was positively associated with grip strength, and the direction of this association remained consistent after sequential adjustment for covariates (Table 2). In the crude model, each additional 60 min/week of LTPA was associated with a mean increase of 0.155 kg in grip strength. Although the effect estimate was attenuated after adjustment for age and sex, the association remained statistically significant in the fully adjusted model (β = 0.064, 95% (CI 0.009–0.120, P = 0.023). In the tertile analysis, the estimated grip strength values in both T2 and T3 were higher than those in T1, but the differences did not reach statistical significance in any adjusted model. After full adjustment, the regression coefficients for T2 and T3 relative to T1 were 0.679 kg and 0.684 kg, respectively, and the linear trend test was not significant (P for trend = 0.350). Overall, the association between LTPA and grip strength was evident primarily in the continuous exposure models, whereas the grouped analysis did not indicate a clear dose-response gradient (Fig. 1).

Table 2.

Associations of LTPA with grip strength: continuous and tertile-based analyses

Panel A. LTPA as a continuous variable (per 60 min/week increase; n = 418)
Model	β (kg)	95% CI	P value
Crude	0.155	0.056 to 0.253	0.002
Model 1	0.067	0.008 to 0.125	0.025
Model 2	0.059	0.002 to 0.116	0.044
Model 3	0.064	0.009 to 0.120	0.023

Panel B. LTPA tertiles (reference: T1 ≤ 0 min/wk; n = 418)
Model	T2 vs. T1, β (95% CI)	P value	T3 vs. T1, β (95% CI)	P value	P for trend
Crude	0.843 (-1.453 to 3.139)	0.471	1.652 (-0.466 to 3.770)	0.126	0.13
Model 1	0.650 (-0.999 to 2.300)	0.439	0.639 (-0.743 to 2.020)	0.364	0.385
Model 2	0.485 (-1.151 to 2.121)	0.56	0.473 (-0.910 to 1.856)	0.502	0.522
Model 3	0.679 (-1.010 to 2.368)	0.43	0.684 (-0.692 to 2.060)	0.329	0.35

Notes: Values are regression coefficients (β) and 95% confidence intervals from ordinary least squares regression with HC3 robust standard errors

Model 1 was adjusted for age and sex

Model 2 was additionally adjusted for education, marital status, living arrangement, and residence

Model 3 was further adjusted for sleep duration, current smoking, current drinking, number of chronic diseases, fall risk score, and cognitive screening score

Fig. 1.

Adjusted muscle strength and functional fitness across LTPA tertiles. Legend: Adjusted mean values for grip strength (A) and functional fitness composite Z-score (B) across tertiles of LTPA. T1, T2, and T3 were defined as ≤0, 0–420, and >420 min/week, respectively. Error bars indicate 95% confidence intervals. P for trend was derived from tertile-based models

LTPA and other functional fitness dimensions

In the fully adjusted models, no clear independent associations were observed between LTPA and the other functional fitness dimensions apart from grip strength (Table 3). For each additional 60 min/week of LTPA, the regression coefficients for the 30-s chair stand test, eyes-closed balance time, and the composite functional fitness Z score were all positive, but the 95% CIs all crossed zero: β = 0.0203 (95% CI -0.0116 to 0.0522, P = 0.211), β = 0.0634 (95% CI -0.0307 to 0.1575, P = 0.186), and β = 0.0036 (95% CI -0.0034 to 0.0106, P = 0.318), respectively. The association with choice reaction time was likewise not statistically significant (β = 0.0024, 95% CI -0.0045 to 0.0092, P = 0.499). Overall, there was no consistent statistical evidence of an association between LTPA and these functional fitness outcomes (Fig. 2).

Table 3.

Associations of LTPA with other functional fitness dimensions (fully adjusted models; per 60 min/week increase; n = 417)

Outcome	β	95% CI	P value
30-s chair stand, repetitions	0.0203	-0.0116 to 0.0522	0.211
Eyes-closed balance, s	0.0634	-0.0307 to 0.1575	0.186
Choice reaction time, s†	0.0024	-0.0045 to 0.0092	0.499
Functional fitness composite Z-score	0.0036	-0.0034 to 0.0106	0.318

Notes: Values are regression coefficients (β) and 95% confidence intervals from fully adjusted ordinary least squares regression models with HC3 heteroscedasticity-consistent robust standard errors

Models were adjusted for age, sex, education, marital status, living arrangement, residence, sleep duration, current smoking, current drinking, number of chronic diseases, fall risk score, and cognitive screening score

† Higher values indicate longer reaction time

Fig. 2.

LTPA and muscle strength / functional fitness dimensions. Legend: Fully adjusted regression coefficients (β) and 95% confidence intervals for the associations of LTPA with grip strength and functional fitness outcomes, expressed per 60 min/week increase in LTPA. The dashed vertical line indicates the null value. Models were adjusted for demographic, lifestyle, and health-related covariates

ST and muscle strength / functional fitness dimensions

In the secondary analyses, the fully adjusted models showed no independent association between ST and grip strength or any functional fitness indicator. For each additional 60 min/day of ST, the regression coefficient was − 0.1093 kg for grip strength (95% CI -0.4424 to 0.2239, P = 0.519) and 0.0067 for the composite functional fitness Z score (95% CI -0.0257 to 0.0390, P = 0.686). The directions of association were also inconsistent across the individual functional indicators, and none reached statistical significance, including the 30-s chair stand test (β=-0.1214, 95% CI -0.3008 to 0.0581, P = 0.184), eyes-closed balance time (β = 0.7486, 95% CI -0.3189 to 1.8160, P = 0.169), and choice reaction time (β=-0.0048, 95% CI -0.0176 to 0.0079, P = 0.455). Overall, there was no consistent statistical evidence of an association between ST and muscle strength or any functional fitness dimension (Additional file 1: Table S1).

Regression diagnostics

Diagnostic checks for the fully adjusted OLS models are presented in Additional file 1: Table S3. In the primary model examining LTPA and grip strength, the quadratic term for LTPA was not statistically significant (P = 1.000), suggesting no clear evidence of non-linearity in the primary association. Across the other fully adjusted models, quadratic terms were generally not statistically significant, except for the LTPA model for eyes-closed balance (P = 0.035), and this finding was interpreted cautiously given the absence of a statistically significant main association for this outcome.

Jarque-Bera tests indicated departures from residual normality in all fully adjusted models (all P < 0.001). The Breusch-Pagan test suggested heteroskedasticity in the primary LTPA–grip strength model (P = 0.013), whereas most other models did not show statistically significant evidence of heteroskedasticity. Maximum variance inflation factor values were high across the fully adjusted models, suggesting potential multicollinearity in the full covariate design matrix. Cook’s distance diagnostics also indicated influential observations in several models where the statistic was available. Therefore, the OLS regression results were interpreted cautiously, and statistical inference was based on HC3 robust standard errors.

Sensitivity analyses

The sensitivity analyses broadly supported the main findings (Additional file 1: Table S2). In the fully adjusted model, each additional 60 min/week of LTPA was associated with a mean increase of 0.063 kg in grip strength (95% CI 0.008–0.119, P = 0.026). After further inclusion of ST in the same model, the association remained essentially unchanged (β = 0.061, 95% CI 0.006–0.117, P = 0.031), whereas ST remained not independently associated with grip strength (P = 0.566). The restricted analyses showed the same direction of association. After excluding current smokers, the association between LTPA and grip strength remained positive, although the statistical evidence was attenuated to borderline significance (β = 0.062, 95% CI -0.000 to 0.123, P = 0.050). After excluding participants with three or more chronic diseases, the association remained statistically significant (β = 0.064, 95% CI 0.005–0.123, P = 0.033). By contrast, in the tertile-based analyses, neither T2 nor T3 differed significantly from T1, and the trend test was not statistically significant (P for trend = 0.395).

Discussion

Principal findings

Previous observational studies have suggested that higher levels of physical activity are associated with better muscle strength and physical performance in older adults or adults entering older age [33, 34]. However, much of the available evidence has focused on customary or total physical activity, accelerometer-derived moderate-to-vigorous physical activity, or global physical performance measures, with relatively limited attention to the independent associations of leisure-time physical activity (LTPA) with specific objective functional fitness domains [35, 36]. For sedentary behavior, associations with physical capability and functional performance have also been inconsistent, and some observed associations were attenuated after accounting for moderate-to-vigorous physical activity. Against this background, the present study showed that LTPA was independently and positively associated with muscle strength in older adults, as indicated by grip strength. This association persisted after sequential adjustment for demographic characteristics, lifestyle factors, and health status, and remained broadly consistent across multiple robustness analyses. By contrast, no clear independent associations were observed between LTPA and the 30-s chair stand test, eyes-closed balance, choice reaction time, or the composite functional fitness score, suggesting that the main signal in this study was concentrated in muscle strength rather than across all functional domains. We did not assume that self-reported LTPA was inherently more accurate than self-reported ST; rather, both exposures were subject to measurement error and should be interpreted cautiously. In addition, ST was not independently associated with grip strength or other functional fitness indicators in this study.

Interpretation of the associations of LTPA with muscle strength and functional fitness dimensions

Most previous studies have focused on total physical activity or composite functional outcomes, with relatively limited discussion of the differential associations between LTPA and specific objective fitness domains, and longitudinal evidence based on repeated measurements also remains scarce [6, 11]. Grip strength is different in this regard. It is one of the most commonly used and reproducible objective indicators of muscle strength in older populations, and it is also a core measure in the Asian AWGS diagnostic pathway for defining low muscle strength and possible sarcopenia. It is therefore methodologically and clinically plausible that the primary signal of LTPA in this study emerged first in grip strength rather than across all functional tests [3, 5]. From a behavioral perspective, LTPA may better capture sustained, self-initiated activity performed in daily life. Such exposure may not be sufficient to induce concurrent changes in balance control, reaction speed, or multidimensional functional performance, but may be reflected earlier in the maintenance of muscle strength. This interpretation is broadly consistent with recent literature. In a network meta-analysis of older adults with sarcopenia, exercise interventions showed a relatively stable benefit for handgrip strength, while longitudinal cohorts from Korea and community-based cohorts from Japan likewise suggested moderate continuous associations between physical activity and grip strength, with lower-intensity activity appearing to relate first to grip strength, whereas broader fitness benefits were more often observed at higher activity intensities [37, 38]. Accordingly, this pattern supports a modest continuous association rather than a clear tertile-based threshold effect, as categorizing skewed LTPA data may reduce statistical power and increase within-group heterogeneity.

LTPA was modeled primarily as a continuous variable to preserve statistical power and estimate dose-response associations without arbitrary cut-points. Tertile analysis was included as a supplementary descriptive approach. Grip strength was designated as the primary outcome because it is widely used, reproducible, and a core component of sarcopenia diagnostic criteria. Other functional fitness measures were secondary outcomes to evaluate whether LTPA associations extended beyond muscle strength. HC3 robust standard errors were used because heteroskedasticity was anticipated in observational data. Complete-case analysis was used rather than multiple imputation because missing data were modest and appeared unsystematic, and the primary aim was to estimate associations in the observed sample.

The absence of clear independent associations between LTPA and the other functional fitness dimensions does not necessarily indicate that they are entirely unrelated; rather, it more likely reflects differences in the physiological basis of these outcomes and in the training stimuli required to influence them [9, 39, 40]. The 30-s chair stand test reflects lower-extremity functional strength, endurance, and movement coordination; balance is influenced jointly by vestibular, visual, and proprioceptive input as well as central integration; and reaction time additionally involves attention, executive function, and processing speed. General leisure activity may therefore not exert effects on these more complex functional domains that are comparable in magnitude to those observed for grip strength [9]. Intervention evidence also supports this domain-specific interpretation. International consensus emphasizes that functional improvement in older adults often requires multicomponent training that includes resistance, balance, and other exercise elements [39, 41]. At the same time, functional training may improve balance and chair stand performance without necessarily altering grip strength, whereas cognitive-motor dual-task training may be more likely to yield benefits in balance and cognitive domains [39, 40]. For the composite functional fitness score, it is also important to consider that a composite indicator may dilute signals present in individual components. A prospective study using the World Health Organization (WHO) intrinsic capacity composite outcome found that moderate-to-vigorous physical activity (MVPA) and sedentary behavior were associated with long-term change; however, that study used accelerometer-based exposure assessment and incorporated multiple domains, including vitality, cognition, psychology, sensory function, and locomotion, into the composite score. This differs from the present study, which used self-reported LTPA and objective functional testing to construct its composite indicator, and some inconsistency in findings is therefore to be expected [42].

Although the association between LTPA and grip strength was statistically significant, the absolute effect size was small. In the fully adjusted model, each additional 60 min/week of LTPA was associated with only a 0.064 kg higher grip strength. There is currently no universally accepted minimal clinically important difference for grip strength in community-dwelling older adults in the context of cross-sectional physical activity research [43, 44]. Therefore, this finding should be interpreted as a modest population-level association rather than evidence of a clinically meaningful individual-level improvement in grip strength.

Interpretation of the null findings for ST

In this study, no independent associations were observed between ST and grip strength or any functional fitness dimension. One direct explanation for the null findings for ST lies in exposure measurement. Both LTPA and ST were based on self-reported questionnaire data and were therefore susceptible to recall bias and systematic misclassification. Self-reported physical activity may be overestimated because participants may have difficulty recalling activity duration accurately or may report socially desirable activity levels. Conversely, self-reported ST may be underestimated, particularly because sedentary behavior is often accumulated in fragmented bouts throughout the day and may be less salient to recall. This issue is especially relevant for the interpretation of the ST findings. If ST was systematically underestimated or imprecisely reported, the resulting measurement error may have attenuated the associations toward the null. Therefore, the absence of statistically significant associations between ST and grip strength or functional fitness in this study should not be interpreted as definitive evidence that sedentary behavior has no functional relevance in older adults [25, 45, 46]. Consistent with the direction of the present findings, the Lifestyle Interventions and Independence for Elders (LIFE) study likewise did not observe a clear association between objectively measured or self-reported ST and grip strength among community-dwelling older adults at increased risk of activity limitations [47]. More broadly, the effects of sedentary behavior in older adults are increasingly interpreted within a time-substitution framework rather than as a single exposure simply opposed to physical activity. A substitution-model study published in BMC Geriatrics suggested that replacing part of ST with light-intensity physical activity or MVPA may improve grip strength, gait speed, or chair stand performance, whereas a systematic review of breaking up ST concluded that the relevant evidence in community-dwelling older adults remains limited and that the effects of sedentary behavior are not easily disentangled from those of physical activity [17, 20]. Taken together, the present findings are more supportive of a role for active leisure behavior in maintaining muscle strength in older age than of the conclusion that total ST independently determines muscle strength or functional fitness. This interpretation also contrasts with a recent longitudinal study based on accelerometer assessment, which found that higher ST was associated with declines in intrinsic capacity, suggesting that differences in study design and exposure assessment may be an important source of inconsistency across studies [42].

Clinical and public health implications

From a clinical assessment perspective, older adults with lower levels of LTPA and poorer grip strength may represent a subgroup with relatively limited functional reserve and may warrant priority attention. In community screening or comprehensive geriatric assessment, grip strength is simple to administer and straightforward to interpret, and may be used alongside information on LTPA to help identify early functional vulnerability. From a health promotion perspective, LTPA is a relatively modifiable behavioral factor that can be more readily integrated into daily life. Although the cross-sectional design precludes causal inference, the findings nonetheless suggest that supporting older adults in maintaining or increasing sustained participation in LTPA may be a practical approach to preserving muscle strength. At the same time, LTPA was not equally associated with all functional domains, which suggests that interventions for healthy ageing should not treat functional fitness as a single homogeneous construct, but should instead distinguish more carefully among muscle strength, lower-extremity function, balance, and reaction capacity. Although the sample was drawn from a single city, the value of this study extends beyond providing a local description. More importantly, it offers a relatively specific signal: among community-dwelling older adults, the associations of LTPA appear to be concentrated in muscle strength rather than extending uniformly across all functional fitness domains. This finding may therefore still inform research in other older populations and the design of community-based interventions.

Strengths and limitations of the study

This study has several strengths. It used objective measures, including handgrip strength, the 30-s chair stand test, eyes-closed balance, and choice reaction time, to assess muscle strength and functional fitness in older adults. LTPA was modeled primarily as a continuous variable to preserve the information in weekly activity time, while tertile-based analyses were used as supplementary descriptive analyses to examine whether a categorical gradient or threshold pattern was present. The absence of a significant tertile trend therefore tempered the interpretation of the continuous finding and suggested that the association should be viewed as modest and continuous rather than as evidence of a clear threshold effect. The models adjusted systematically for potential confounders, including demographic characteristics, lifestyle factors, chronic disease burden, fall risk, and cognitive status. Sensitivity analyses were broadly consistent with the primary analyses, supporting the robustness of the main findings.

At the same time, the cross-sectional design does not permit causal inference, and reverse causation cannot be excluded. Both LTPA and ST were self-reported and may therefore be subject to recall bias and exposure misclassification. Self-reported LTPA may have been overestimated, whereas self-reported ST may have been underestimated; the latter is particularly important because measurement error in ST may have attenuated its associations with grip strength and functional fitness toward the null.

Regarding the choice of total LTPA duration rather than intensity-specific analyses, several considerations apply: (1) Although the questionnaire collected frequency and duration for light-, moderate-, and vigorous-intensity LTPA, intensity-specific analyses were not used as primary analyses because vigorous-intensity activity was sparse and the LTPA distribution was highly zero-inflated. The primary analysis therefore focused on total weekly LTPA, while future studies with larger samples could more robustly examine intensity-specific associations. (2) in this population of older adults, most LTPA was reported as light or moderate intensity, with few participants engaging in regular vigorous activity, limiting the statistical power for intensity-stratified analyses. (3) the primary research question concerned the overall dose-response relationship between leisure-time activity and functional outcomes, rather than differential effects by intensity. Future studies with larger samples and more detailed activity records could explore intensity-specific associations, particularly given evidence from other cohorts that moderate-intensity activity may have stronger associations with functional outcomes than light-intensity activity in older adults.The absence of a stable trend in the tertile analyses suggests that the findings are better interpreted in terms of continuous exposure. Because the sample was drawn from communities in a single city, caution is warranted when generalizing the findings to other older populations, and residual confounding cannot be completely ruled out.

Employment or retirement status was not collected in this study. Given that the study population comprised community-dwelling older adults, most of whom were likely retired or not in paid employment, LTPA may approximate total physical activity more closely in this group than in working-age populations.

The study did not collect information on the specific types, contexts, or social settings of LTPA, and occupational and transportation-related physical activity were not assessed. The fall risk and cognitive screening scores were derived from brief screening tools rather than comprehensive clinical assessments, and nutritional status, medication use, and inflammatory biomarkers were not measured. Future longitudinal and intervention studies are needed to clarify the temporal sequence and potential causal relationship between LTPA and muscle strength, and to determine whether different types, frequencies, and intensities of LTPA correspond to different patterns of functional benefit. Where feasible, future studies should incorporate objective activity-monitoring tools such as accelerometers to reduce self-report error and to further examine whether more refined and domain-specific associations exist between LTPA and muscle strength and other functional dimensions.

Conclusions

Among community-dwelling older adults, higher levels of LTPA were independently associated with slightly higher grip strength, although the absolute effect size was small and its clinical significance remains uncertain. This association persisted after sequential adjustment for covariates and remained broadly consistent in sensitivity analyses. By contrast, no clear independent associations were observed between LTPA and the 30-s chair stand test, eyes-closed balance, choice reaction time, or the composite functional fitness score, and no consistent associations were found between self-reported ST and muscle strength or any functional fitness dimension, although these null findings should be interpreted cautiously given the potential underestimation and measurement error of self-reported ST. These findings suggest that the associations of LTPA with functional status in older adults may be more concentrated in the muscle strength domain than across all functional fitness indicators. Combining information on LTPA with grip strength measurement may help support the early identification of reduced functional reserve in community-dwelling older adults, although the temporal sequence and modifiability of this association still require confirmation in longitudinal and intervention studies.

Abbreviations

ADL

Activities of daily living

AWGS

Asian Working Group for Sarcopenia

CI

Confidence interval

HC3

Heteroskedasticity-consistent HC3 robust standard errors

LTPA

LTPA

OLS

Ordinary least squares

ST

ST

Authors’ contributions

SJ drafted the manuscript. BG and JL was responsible for data collection and data management. TH contributed to data collection and data management. XF contributed to data entry and data management. GQ supervised the study, critically revised the manuscript, and confirmed the final version. All authors read and approved the final manuscript.

Funding

The authors received no specific funding for this work.

Data availability

The datasets generated or analysed during the current study are not publicly available because they contain information that could compromise participant privacy, but de-identified data are available from the corresponding author on reasonable request, subject to ethical approval and applicable data protection requirements.

Declarations

Ethics approval and consent to participate

The study protocol was approved by the Ethics Committee of Shihezi People’s Hospital (approval No. KYLL No. (2015) 11). Written informed consent was obtained from all participants before the questionnaire survey and physical fitness assessment.

Consent for publication

Not applicable. No identifying images or other personal or clinical details of participants are included in this manuscript.

Competing interests

The authors declare no competing interests.