The association between sarcopenia, sedentary behavior, and the motor cognitive risk syndrome: A Mendelian randomization study

Abstract

Recent observational studies have suggested a potential association between sarcopenia and motor cognitive risk syndrome (MCR), but the underlying mechanism of action has not been elucidated, and perhaps sedentary behavior may play a key role in the association. To further investigate this association, this study utilized publicly available genome-wide association study data. First, a bidirectional Mendelian randomization (MR) approach was employed to examine the relationship between sarcopenia-related phenotypes (including appendicular lean mass [ALM] [N = 450,243], handgrip strength [HGS] [N = 256,523], and walking pace [WP] [N = 459,915]) with MCR (N = 22,593). Subsequently, a two-step MR approach was employed to investigate whether sedentary behavior (nonwork computer use) (N = 422,218) mediates this causal association. Positive MR results showed that all relevant phenotypes of sarcopenia had a significant positive correlation with cognitive function, including ALM (odds ratios [OR] = 1.10, 95% confidence intervals [CIs]: 1.04–1.15; P < .01), left HGS (OR = 1.22, 95% CI: 1.02–1.47; P = .02), right HGS (OR = 1.22, 95% CI: 1.01–1.47; P = .03), and WP (OR = 3.06, 95% CI: 1.99–4.71; P < .01). The results of the reverse MR analysis indicate that cognitive function is positively associated only with WP (OR = 1.02, 95% CI: 1.00–1.04; P = .01), while there was no significant correlation with ALM, left, and right HGS. Mediation analysis further revealed that recreational computer use partially mediated the associations between ALM, HGS, and WP and cognitive function. The results of this study suggest that declines in ALM, HGS, and WP significantly increase the risk of developing MCR. Meanwhile, the onset of MCR may further compromise gait speed, potentially creating a vicious cycle. In addition, recreational computer use has an important mediating role in the association between sarcopenia and MCR. In patients with sarcopenia, excessive recreational computer use appears to accelerate cognitive impairment, potentially aggravating the risk of developing MCR. Future recommendations suggest that middle-aged and older adults, particularly those with sarcopenia, should appropriately reduce sedentary behavior. Such interventions can serve as early strategies to prevent cognitive decline, thereby enhancing overall quality of life.

1. Introduction

The global population is aging rapidly. Consequently, the public health burden of dementia is escalating quickly. By 2050, the number of confirmed cases worldwide is projected to exceed 152.8 million.^[1] Currently, there are no specific therapies for overt dementia. Therefore, identifying early warning signs and establishing effective intervention targets are central research priorities.^[2] In this context, motoric cognitive risk syndrome (MCR) has gained significant academic attention. It is considered a clinically valuable prodromal state of dementia. MCR refers to the coexistence of subjective cognitive complaints and bradykinesia in older adults.^[3] Evidence indicates that MCR serves as an independent risk factor for cognitive decline and the onset of dementia.^[4–6] Furthermore, MCR significantly increases the incidence of adverse outcomes. These outcomes include frailty, falls, disability, and mortality.^[7–9] The pathophysiology of MCR is complex. It involves multidimensional factors such as socioeconomic status,^[10] cardiovascular burden,^[11] metabolic abnormalities,^[12,13] and psychological stress.^[14] However, muscle health is emerging as a focal point in etiological research. Sarcopenia is particularly important within this network of associations. Sarcopenia is a degenerative disorder of aging. It is characterized by the loss of skeletal muscle mass, a decline in strength, and physical functional impairment.^[15] Its potential role in MCR development cannot be overlooked.^[16]

In recent years, researchers have established a significant association between sarcopenia and cognitive impairment. This link is evident in neurodegenerative diseases such as mild cognitive impairment (MCI) and Alzheimer disease.^[17–19] Sugimoto et al^[20] further revealed a marked increase in sarcopenia prevalence during the progression from amnestic MCI to Alzheimer disease. However, existing research remains predominantly focused on dementia and MCI stages. Evidence regarding the association between sarcopenia and MCR remains limited. Furthermore, the precise mechanisms underlying sarcopenia’s role in MCR pathogenesis remain poorly elucidated. Recent theoretical advances based on the “Muscle-Brain Axis” offer a novel perspective on this issue. Skeletal muscles secrete myokines, such as brain-derived neurotrophic factor (BDNF) and irisin.^[21] These myokines play crucial roles in neuroprotection and cognitive regulation. This suggests a direct link. Impaired skeletal muscle function may contribute to MCR onset by disrupting myokine-mediated signaling pathways.

Notably, sarcopenia likely impacts MCR through multiple biological pathways. Sedentary behavior serves as a pivotal “behavior-metabolism” hub. Unhealthy lifestyles act as a shared risk factor for both sarcopenia and cognitive impairment.^[22] They also serve as a pathological bridge connecting the 2. Sarcopenia causes a decline in skeletal muscle mass and dysfunction in the lower limbs. This physically restricts the mobility of older adults. Consequently, they are forced into prolonged sedentary states.^[23] This passive sedentary pattern is not merely a clinical consequence. It actively exacerbates MCR progression through specific biological circuits. On one hand, insufficient muscle contraction suppresses myokine release (e.g., irisin). This leads to the downregulation of BDNF and impairs hippocampal plasticity.^[24] On the other hand, prolonged sitting is associated with chronic low-grade inflammation (e.g., elevated interleukin-6 and C-reactive protein). It also causes vascular endothelial damage. These factors accelerate the accumulation of white matter hyperintensities. This disrupts the neural regulatory network governing gait. Simultaneously, it accelerates muscle breakdown.^[25] Therefore, we must deepen our understanding of the connections between sarcopenia, sedentary behavior, and MCR. This is crucial for elucidating pathophysiological mechanisms. It is also vital for developing early intervention strategies.

However, these mechanisms are currently inferred largely from observational studies. Such studies cannot fully eliminate confounding factors. They also struggle with reverse causality. Therefore, it is difficult to confirm that sedentary behavior truly mediates the pathway from sarcopenia to MCR. To address this limitation, this study employs Mendelian randomization (MR). MR uses genetic variants (single nucleotide polymorphism [SNP]) as instrumental variables (IVs). This design mimics a randomized controlled trial.^[26] This approach minimizes confounding bias and strengthens causal inference. Consequently, this study aims to systematically explore the causal relationship between sarcopenia and MCR using MR. We focus particularly on the mediating role of sedentary behavior.

2. Materials and methods

2.1. Study design

This study employs an MR design for causal inference based on pooled genome-wide association study (GWAS) data. By introducing SNPs as IVs, a causal chain is established that directly links exposure to the outcome without confounding, thereby replicating the strength of evidence from randomized controlled trials in observational data. To ensure robust inference, instrument selection strictly adheres to 3 fundamental principles: First, the association hypothesis stipulates that the IVs must exhibit a strong association with sarcopenia-related characteristics or cognitive function. Second, the independence hypothesis requires that the IVs be unrelated to potential confounders affecting the exposure–outcome relationship. Finally, the exclusivity hypothesis posits that the IVs can influence the outcome only through the aforementioned exposure factor, thereby eliminating interference from other pleiotropic pathways.

Sarcopenia-related phenotypic indicators are defined according to the European Working Group criteria, employing appendicular lean mass (ALM), handgrip strength (HGS), and walking speed (WP) to characterize muscle mass, strength, and physical function, respectively.^[27] Given that cognitive function is a core feature of MCR, it was employed as a surrogate measure for MCR.^[28] Additionally, leisure sedentary behavior (represented by common leisure-time computer use) was included as a potential mediator.

Using GWAS summary data, the study employed a bidirectional two-sample MR approach to systematically investigate the association between sarcopenia-related phenotypes and MCR. To further elucidate potential mediating effects, a two-step MR analysis was used. First, the causal effect of sarcopenia-related phenotypes on indicators of sedentary behavior was assessed. Next, the causal effect of sedentary behavior on MCR was analyzed. The mediation effect, representing the extent to which the relationship between sarcopenia and MCR is explained by sedentary behavior, was then quantified using the product method (Fig. 1).

Figure 1.

Flow chart of Mendelian randomization analysis. ALM = appendicular lean mass, HGS = handgrip strength, WP = walking pace.

2.2. Data sources

The data were derived from summary statistics of multiple large-scale GWAS. The ALM data were extracted from a GWAS study comprising 450,243 participants from the UK Biobank, in which measurements were obtained using the Tanita BC 418ma bioimpedance analyzer and validated by dual-energy X-ray absorptiometry.^[29] The HGS data originated from a GWAS meta-analysis comprising 256,523 European-descendant participants aged 60 years or older. Maximum grip strength was measured using a Jamar J00105 hydraulic hand dynamometer, with low grip strength defined as <30 kg for men and <20 kg for women.^[27] The WP data were obtained from the UK Biobank public database, which included 459,915 European-descendant participants. Walking pace was self-reported via questionnaire (slow: <3 mph; medium: 4 mph; fast: >4 mph).^[29] The MCR data were integrated from the latest GWAS findings. Given that cognitive function is a core assessment criterion for MCR, the study approximated the evaluation using the most representative GWAS dataset for cognitive function, which included 22,593 European-descendant participants.^[30] The sedentary behavior data were sourced from the UK Biobank, comprising 422,218 individuals of European ancestry (45.7% male), with a mean age of (57.4 ± 8.0) years at initial assessment and a mean daily leisure computer use duration of (1.0 ± 1.2) hours^[31] (Table 1).

Table 1.

Summary data from this study.

Research traits	GWAS id	Population source	Sample size	SNP count
Exposure
ALM	ebi-a-GCST90000025	European	450,243	18,071,518
Left HGS	ukb-b-7478	European	461,026	9,851,867
Right HGS	ukb-b-10215	European	461,089	9,851,867
WP	ukb-b-4711	European	459,915	9,851,867
Mediating
Leisure computer use	ukb-b-4522	European	360,895	9,851,867
Outcome
Cognitive Function	ieu-b-4838	European	22,593	6,719,661

ALM = appendicular lean mass, GWAS = genome-wide association study, HGS = handgrip strength, SNP = single nucleotide polymorphism, WP = walking pace.

Although the data on sarcopenia, MCR, and sedentary behavior were derived from individuals of European ancestry, differences in sample size, study design, and data-collection protocols may have introduced population heterogeneity. To minimize the impact of population stratification, only GWAS summary statistics from European-ancestry samples were included during variable selection.

2.3. Selection of IVs

In the screening of IVs, a genome-wide significance threshold (P < 5 × 10^‐8) was applied to select highly correlated SNPs. Due to the limited number of SNPs associated with cognitive function (only 3), the screening criteria were appropriately relaxed to P < 1 × 10^<−6 to ensure statistical power. Although relaxing the threshold may increase the risk of type I errors, subsequent linkage disequilibrium analysis and F-statistic validation (F > 10) ensured the strength of the IVs.^[32] To eliminate the influence of linkage disequilibrium, SNPs that did not meet the criteria (r² = 0.001; distance > 10,000 kb) were excluded using R software. This process effectively reduced bias due to covariance and addressed nonindependence among IVs, thereby significantly reducing the likelihood of false-positive results.^[33] Additionally, the F-statistic was calculated for each SNP, and weak IVs (F < 10) were excluded.^[34] Finally, allele harmonization was performed to ensure consistency in the effect alleles, avoiding strand mismatches and improving the accuracy of causal inference.^[35]

To control for potential pleiotropy, sensitivity analyses, such as MR-Egger and MR-PRESSO, were employed to detect horizontal pleiotropy and to validate the robustness of the findings.

2.4. MR analysis

To systematically elucidate the potential association between sarcopenia-related phenotypes and MCR, this study employed a two-way two-sample MR analysis. In terms of statistical strategy, we primarily used the random-effects inverse-variance weighting (IVW) method for causal estimation. Given the potential heterogeneity among SNPs, we prioritized the random-effects model over the fixed-effects model because it allows greater variability in effect sizes. This model provides more conservative confidence intervals by adjusting standard errors, effectively reducing the risk of type I errors while ensuring robust results.^[36] To verify the robustness of the results, supplementary analyses were performed using the weighted median method and MR-Egger regression. The weighted median method remains robust even when up to 50% of the IVs are invalid, while MR-Egger regression can effectively identify and adjust for potential directional pleiotropic bias.^[37] Furthermore, to address potential weak-instrument bias, the MR-RAPS method was applied to correct parameters and improve estimation precision.^[38]

To ensure the reliability of the findings, systematic sensitivity analyses were conducted: the intercept term of MR-Egger regression was tested to assess pleiotropic effects (P < .05 was considered indicative of significant pleiotropy); Cochran Q test was used to evaluate heterogeneity; and leave-one-out analysis and scatter plots were employed to verify result stability. In the two-step MR analysis, a rigorous mediation framework was established. The total effect (β0) of sarcopenia-related traits on cognitive function was estimated using the IVW method. The effect of sedentary behavior on cognitive function (β2) and the effect of sarcopenia-related phenotypes on sedentary behavior (β1) were separately assessed. The mediation effect was calculated as β1×β2, the direct effect as β0′ = β0 ‐ β1×β2, and the mediation proportion as β1×β2/β0. A significant mediation effect was determined if the following criteria were met: the IVW analysis P-values for the exposure-outcome, mediator-outcome, and exposure-mediator relationships were all <.05; and the mediation proportion exceeded 5%.

Additionally, the MR Steiger method was used for directional testing, excluding SNPs suggesting reverse causality before reanalysis. Associations were quantified using odds ratios (OR) and 95% confidence intervals (95% CIs), with significance set at P < .05. An OR > 1 implies a positive association (increased likelihood), whereas an OR < 1 implies a negative association (decreased likelihood); the 95% CI provides a measure of the estimate’s precision and reliability. All statistical analyses were performed in R software (version 4.4.2; The R Foundation for Statistical Computing, Vienna, Austria), primarily using the TwoSampleMR (version 0.6.8; MRC Integrative Epidemiology Unit, University of Bristol, Bristol, United Kingdom), and MR-PRESSO (version 1.0; The Charles Bronfman Institute for Personalized Medicine, Icahn School of Medicine at Mount Sinai, New York) packages for data processing.

3. Results

3.1. Association between sarcopenia-related phenotypes and MCR

In the MR analysis of sarcopenia-related phenotypes and MCR, a total of 431 SNPs associated with ALM and cognitive function were included based on the screening criteria, with F-statistics ranging from 29.87 to 779.56. For left-hand HGS and cognitive function, 126 SNPs were included (F-statistics: 29.77–191.54), while 136 SNPs were included for right-hand HGS and cognitive function (F-statistics: 29.89–231.73). For WP and cognitive function, 46 SNPs were included (F-statistics: 29.79–101.56). The F-statistics for all IVs exceed 10, indicating strong associations between the selected genetic variants and the exposure factor. This fully satisfies the association assumption of MR. This outcome effectively mitigates the risk of weak IV bias, ensuring the robustness and reliability of causal inference results.

The IVW analysis revealed significant positive correlation between the following traits and cognitive function: ALM (OR = 1.10, 95% CI: 1.04–1.15; P < .01), left HGS (OR = 1.22, 95% CI: 1.02–1.47; P = .02), right HGS (OR = 1.22, 95% CI: 1.01–1.47; P = .03), and WP (OR = 3.06, 95% CI: 1.99–4.71; P < .01). MR-Egger regression showed no significant association with ALM (OR = 1.00, 95% CI: 0.89–1.13; P = .93), left HGS (OR = 0.90, 95% CI: 0.44–1.87; P = .79), right HGS (OR = 0.99, 95% CI: 0.49–1.98; P = .98), and WP (OR = 1.83, 95% CI: 0.27–12.34; P = .53) showed no significant causal association with cognitive function. Results from the weighted median method indicated no significant association between ALM and bilateral HGS with cognitive function (both P > .05), but WP demonstrated a significant causal effect (OR = 2.37, 95% CI: 1.41–3.98; P < .01). Although MR-Egger regression and partial weighted median method results did not reach statistical significance, their effect directions were highly aligned with those of the IVW analysis, further validating the robustness of the findings (Table 2).

Table 2.

Results of the analysis of the causal association between the characteristics associated with sarcopenia and MCR.

Exposure	Outcome	SNP count	MR analysis results
			Method	OR (95% CI)	P
ALM	Cognitive function	431	IVW	1.10 (1.04–1.15)	<.01
			Weighted median	1.05 (0.97–1.14)	.16
			MR-Egger	1.00 (0.89–1.13)	.93
			RAPS	1.09 (1.03–1.15)	<.01
			MR-PRESSO	1.09 (1.04–1.15)	<.01
Left HGS	Cognitive function	126	IVW	1.22 (1.02–1.47)	.02
			Weighted median	1.21 (0.95–1.55)	.11
			MR-Egger	0.90 (0.44–1.87)	.79
			RAPS	1.25 (1.03–1.52)	.02
			MR-PRESSO	NA
Right HGS	Cognitive Function	136	IVW	1.22 (1.01–1.47)	.03
			Weighted Median	1.19 (0.94–1.50)	.13
			MR-Egger	0.99 (0.49–1.98)	.98
			RAPS	1.23 (1.01–1.51)	.03
			MR-PRESSO	NA
WP	Cognitive function	46	IVW	3.06 (1.99–4.71)	<.01
			Weighted Median	2.37 (1.41–3.98)	<.01
			MR-Egger	1.83 (0.27–12.34)	.53
			RAPS	2.92 (1.86–4.61)	<.01
			MR-PRESSO	NA

ALM = appendicular lean mass, HGS = handgrip strength, IVW = inverse-variance weighting, MCR = motoric cognitive risk syndrome, MR = Mendelian randomization, RAPS = robust adjusted profile score, SNP = single nucleotide polymorphism, WP = walking pace.

Sensitivity analysis detected no significant horizontal pleiotropy, suggesting that the IVs primarily influenced the outcome through the exposure factors. Additionally, the robustness of the results was further confirmed by the MR-RAPS and MR-PRESSO methods (Table 3). Cochran Q test indicated heterogeneity, which was effectively addressed by using a random-effects IVW model (Table 3). Furthermore, leave-one-out analysis showed that removing any single SNP did not substantially alter the overall effect estimates. Scatter plots visually presented the analytical results, further validating the robustness of the conclusions (Fig. 2).

Table 3.

Sensitivity analysis results.

Exposure	Outcome	SNP count	Cochran Q test		MR-Egger intercept test
			Q	P	Intercept	P
ALM	Cognitive function	431	540.197	.001	0.002	.098
Left HGS	Cognitive function	126	171.607	.003	0.003	.397
Right HGS	Cognitive function	136	205.331	<.001	0.002	.533
WP	Cognitive function	46	79.253	.001	0.004	.586
Cognitive function	ALM	10	30.555	<.001	0.005	.412
Cognitive function	Left HGS	10	21.154	.011	‐0.006	.112
Cognitive function	Right HGS	10	12.976	.163	‐0.006	.333
Cognitive function	WP	10	12.119	.206	‐0.004	.087
ALM	Casual computer use	502	576.784	<.001	<0.001	.170
Left HGS	Casual computer use	137	409.653	<.001	<‐0.001	.973
Right HGS	Casual computer use	126	494.877	<.001	<0.001	.474
WP	Casual computer use	50	282.244	<.001	<-0.001	.834
Casual computer use	Cognitive function	72	93.170	.040	‐0.001	.796

ALM = appendicular lean mass, HGS = handgrip strength, MR = Mendelian randomization, SNP = single nucleotide polymorphism, WP = walking pace.

Figure 2.

Scatterplot of the analyzed results of the causal association of sarcopenia-related characteristics with MCR. MCR = motoric cognitive risk syndrome.

3.2. Association between MCR and sarcopenia-related phenotypes

A reverse MR analysis was employed to investigate the causal relationship between cognitive function and sarcopenia-related phenotypes. Based on the screening criteria, 4 SNP sets were included in the analysis. For cognitive function and ALM, 10 SNPs were included, with F-statistics ranging from 24.63 to 29.91; for cognitive function and left HGS, 10 SNPs were included, with F-statistics ranging from 24.63 to 29.91; for cognitive function and right HGS, 10 SNPs were included, with F-statistics ranging from 24.63 to 31.14; and for cognitive function and WP, 10 SNPs were included, with F-statistics ranging from 24.63 to 31.14. All IVs had F-values >10, indicating robust results.

IVW analysis demonstrated a significant causal association between cognitive function and WP (OR = 1.02, 95% CI: 1.00–1.04; P = .01). MR-Egger regression analysis further confirmed this finding (OR = 1.10, 95% CI: 1.02–1.18; P = .03). Although the weighted median method did not reach statistical significance (OR = 1.01, 95% CI: 0.99–1.04; P = .09), its effect direction was consistent with the IVW analysis. In contrast, no causal effects were observed for cognitive function on ALM (OR = 1.04, 95% CI: 0.99–1.09; P = .05), left HGS (OR = 1.00, 95% CI: 0.98–1.03; P = .53), and right HGS (OR = 1.00, 95% CI: 0.98–1.03; P = .41). Sensitivity analyses using MR-Egger and weighted median methods confirmed these findings (Table 4).

Table 4.

Results of the analysis of causal associations between MCR and characteristics related to sarcopenia.

Exposure	Outcome	SNP Count	MR analysis results
			Method	OR (95% CI)	P
Cognitive function	ALM	10	IVW	1.04 (0.99–1.09)	.05
			Weighted median	1.02 (0.98–1.05)	.17
			MR-Egger	0.95 (0.77–1.76)	.67
			RAPS	1.02 (0.99–1.06)	.09
			MR-PRESSO	1.02 (0.99–1.05)	.12
Cognitive function	Left HGS	10	IVW	1.00 (0.98–1.03)	.53
			Weighted median	0.99 (0.97–1.02)	.95
			MR-Egger	1.12 (0.44–1.87)	.79
			RAPS	1.25 (0.99–1.26)	.09
			MR-PRESSO	0.99 (0.97–1.02)	.94
Cognitive function	Right HGS	10	IVW	1.00 (0.98–1.03)	.41
			Weighted median	1.00 (0.98–1.03)	.56
			MR-Egger	1.12 (1.03–1.23)	.02
			RAPS	1.01 (0.98–1.04)	.33
			MR-PRESSO	NA
Cognitive function	WP	10	IVW	1.02 (1.00–1.04)	.01
			Weighted median	1.01 (0.99–1.04)	.09
			MR-Egger	1.10 (1.02–1.18)	.03
			RAPS	1.02 (1.00–1.04)	.01
			MR-PRESSO	NA

ALM = appendicular lean mass, CIs = confidence intervals, HGS = handgrip strength, IVW = inverse-variance weighting, MCR = motoric cognitive risk syndrome, MR = Mendelian randomization, RAPS = robust adjusted profile score, SNP = single nucleotide polymorphism, WP = walking pace.

Sensitivity analysis indicated no significant horizontal pleiotropy. To further validate the results, the MR-RAPS and MR-PRESSO methods were applied, and the findings remained consistent. Cochran Q test detected heterogeneity in the analyses of cognitive function with ALM and left HGS; therefore, a random-effects IVW model was used for adjustment (Table 3). Additionally, leave-one-out sensitivity analysis and scatter plots visually supported the conclusions, further enhancing the reliability of the findings (Fig. 3).

Figure 3.

Scatterplot of the results of the analysis of the causal association between MCR and the characteristics associated with sarcopenia. MCR = motoric cognitive risk syndrome.

3.3. Associations between sarcopenia, sedentary behavior, and MCR

This study employed a two-step MR analysis to systematically investigate the associations among sarcopenia-related phenotypes, leisure computer use, and cognitive function. For IV selection, strict screening criteria were applied, ultimately identifying the following: 502 SNPs for ALM and leisure computer use analysis (F-statistic range: 29.63–566.91), 137 SNPs for left HGS (F-statistic range: 9.88–324.45), 126 SNPs for right HGS (F-statistic range: 29.34–423.78), and 50 SNPs for WP (F-statistic range: 24.16–34.23). For the analysis of leisure computer use and cognitive function, 72 SNPs were included (F-statistic range: 26.45–131.36). All IVs had F-statistics >10, indicating sufficient statistical strength for the selected SNPs.

The IVW analysis revealed significant positive associations between ALM (OR = 1.06, 95% CI: 1.04–1.07; P < .01), left HGS (OR = 1.05, 95% CI: 1.00–1.11; P = .029), right HGS (OR = 1.04, 95% CI: 1.02–1.07; P = .028), and WP (OR = 1.21, 95% CI: 1.04–1.41; P < .01) with leisure computer use. Additionally, leisure computer use (OR = 1.88, 95% CI: 1.49–2.30; P < .01) was positively associated with cognitive function. Although MR-Egger regression did not reach statistical significance for left and right HGS, the effect direction aligned with the IVW results, supporting the reliability of the findings (Table 5).

Table 5.

Results of the analysis of causal associations between sarcopenia-related characteristics, sedentary behavior and MCR.

Exposure	Outcome	SNP count	MR analysis results
			Method	OR (95% CI)	P
ALM	Casual computer use	502	IVW	1.06 (1.04–1.07)	<.01
			Weighted median	1.04 (1.02–1.05)	<.01
			MR-Egger	1.04 (1.00–1.07)	.01
			RAPS	1.05 (1.04–1.07)	<.01
			MR-PRESSO	1.05 (1.04–1.06)	<.01
Left HGS	Casual computer use	137	IVW	1.05 (1.00–1.11)	.02
			Weighted median	1.05 (1.00–1.11)	.04
			MR-Egger	1.06 (0.87–1.28)	.53
			RAPS	1.06 (1.00–1.12)	.03
			MR-PRESSO	1.06 (1.01–1.11)	.01
Right HGS	Casual computer use	126	IVW	1.04 (1.00–1.10)	.03
			Weighted median	1.03 (1.01–1.09)	.04
			MR-Egger	0.98 (0.81–1.17)	.84
			RAPS	1.04 (1.01–1.10)	.01
			MR-PRESSO	1.04 (0.99–1.08)	.05
WP	Casual computer use	50	IVW	1.21 (1.04–1.41)	<.01
			Weighted median	1.30 (1.15–1.47)	<.01
			MR-Egger	1.29 (0.69–2.41)	.41
			RAPS	1.19 (1.01–1.39)	.02
			MR-PRESSO	1.23 (1.11–1.37)	<.01
Casual computer use	Cognitive function	72	IVW	1.85 (1.49–2.30)	<.01
			Weighted median	1.62 (1.22–2.15)	<.01
			MR-Egger	2.14 (0.69–6.65)	.18
			RAPS	1.88 (1.49–2.36)	<.01
			MR-PRESSO	NA

ALM = appendicular lean mass, CIs = confidence intervals, HGS = handgrip strength, IVW = inverse-variance weighting, MCR = motoric cognitive risk syndrome, MR = Mendelian randomization, RAPS = robust adjusted profile score, SNP = single nucleotide polymorphism, WP = walking pace.

Sensitivity analyses indicated no significant horizontal pleiotropy detected by MR-Egger regression. Further validation using MR-RAPS and MR-PRESSO methods supported these conclusions. Cochran Q test suggested heterogeneity, but the results remained robust after correction using a random-effects model (Table 3). Leave-one-out analysis confirmed that the findings were not driven by any single SNP, reinforcing the reliability of the conclusions. Scatter plots visually presented the primary results, further enhancing the study’s credibility (Fig. 4).

Figure 4.

Scatterplot of the results of the analysis of the causal association between sarcopenia-related characteristics, sedentary behavior, and MCR. MCR = motoric cognitive risk syndrome.

3.4. Mediating role of sedentary behavior in the association between sarcopenia and MCR

Using a two-step MR approach, the study demonstrated that leisure computer use significantly mediated the relationship between sarcopenia and MCR. Mediation analysis revealed the following: in the causal pathway between ALM and cognitive function, leisure computer use accounted for a mediation effect of 0.038, explaining 40.76% of the total effect; for left HGS and cognitive function, the mediation effect was 0.035 (16.99%); for right HGS and cognitive function, it was 0.028 (13.93%); and for WP and cognitive function, it was 0.121 (10.86%) (Table 6). These results indicate that leisure computer use plays a substantial mediating role in the causal relationships between sarcopenia-related traits and cognitive function, with the highest explanatory power observed for ALM and cognitive function.

Table 6.

Intermediation effects and their share.

Exposure	Mediating	Outcome	β0	β1	β2	Direct effect	Mediating effect	Mediation ratio
ALM	Casual computer use	Cognitive function	0.094	0.062	0.618	0.056	0.038	40.76%
Left HGS	Casual computer use	Cognitive function	0.206	0.057	0.618	0.171	0.035	16.99%
Right HGS	Casual computer use	Cognitive function	0.204	0.046	0.618	0.176	0.028	13.93%
WP	Casual computer use	Cognitive function	1.121	0.197	0.618	1.000	0.121	10.86%

ALM = appendicular lean mass.

4. Discussion

This study employed bidirectional two-sample MR analysis and two-step MR analysis to systematically investigate the bidirectional association between sarcopenia-related phenotypes and cognitive function, along with its potential mediating mechanisms. Forward MR analysis revealed significant positive correlations between muscle mass, grip strength, and WP with cognitive function, indicating that declines in muscle mass, muscle strength, and WP may be key factors contributing to cognitive impairment and increased MCR risk. Reverse MR analysis revealed a significant positive association between cognitive function and WP, suggesting that cognitive decline may impair WP. Mediator analysis further confirmed that sedentary behavior partially mediated the relationship between sarcopenia and cognitive function risk, providing new insights into the pathophysiological mechanisms underlying the association between sarcopenia and MCR.

Using MR analysis, this study establishes a causal link between sarcopenia-related traits and cognitive function. This finding validates previous cross-sectional research while offering new insights into the underlying pathological mechanisms.^[16] Muscle mass loss underlies sarcopenia and is known to accelerate executive function decline^[39]; indeed, it has shown greater predictive value than bone density in older women.^[40] However, both our analysis and prospective evidence indicate that impaired muscle function (strength and speed) is a more sensitive predictor of cognitive impairment.^[41] This indicates that anatomical changes (muscle mass) and functional performance (muscle function) may influence cognition through distinct pathways, collectively raising the risk of MCR. Notably, the significant association between HGS and cognitive decline strongly supports the “force-brain coupling” hypothesis. HGS is not merely a physical measure of limb strength but also a surrogate biomarker for central nervous system integrity. Historical longitudinal data show that both absolute declines in HGS and increased bilateral asymmetry significantly precede the onset of MCR.^[42] Quantitative analyses further indicate a 3% increase in MCR risk for every 1-kilogram decrease in HGS.^[16] Thus, muscle weakness serves as an early warning signal for neurodegenerative disease, often appearing before obvious cognitive symptoms and reflecting an early decline in the nervous system’s ability to recruit motor units.^[43]

However, among functional indicators, WP demonstrated unique significance in our study. Reverse MR analysis revealed a complex association: cognitive decline was associated with reduced WP but did not exert a direct causal effect on muscle mass. This “functional-structural dissociation” does not negate the connection between the 2, but rather underscores the cascading nature of neurodegenerative disease. As demonstrated by Tian et al’s multicohort meta-analysis, concurrent declines in gait speed and memory represent a fundamental early marker of neurodegeneration, rather than merely a musculoskeletal issue.^[44] Cognitive decline first disrupts motor planning and executive function, prioritizing impairment in WP (a functional indicator). Effects on muscle mass (a structural indicator) often lag, primarily mediated by cognitive impairment-induced emotional apathy and reduced physical activity.^[45] Furthermore, mediation analyses indicate that slowed WP is a key pathway through which sarcopenia contributes to cognitive impairment. This supports the “bidirectional vicious cycle” hypothesis: slowed WP increases cognitive load, while impaired cognition weakens motor control systems, exacerbating gait abnormalities.^[46] Evidence from the Korean Elderly Cohort Study corroborates this, demonstrating that WP plays a central mediating role between sarcopenia and cognitive impairment.^[47] Thus, slowed WP should not be viewed merely as a marker of aging but as a critical bridge connecting muscle pathology to cognitive decline.

While aging, metabolic disorders, and physical inactivity are often viewed as common background factors for sarcopenia and MCR.^[48–50] our study points to a specific biological pathway connecting them: the “muscle-brain axis.” As the body’s largest endocrine organ, skeletal muscle maintains secretory homeostasis that is crucial for the central nervous system.^[51] In sarcopenia, loss of muscle mass directly reduces the secretion of exercise-induced myokines, particularly irisin.^[51] Choi et al confirmed that irisin can cross the blood–brain barrier and upregulate BDNF in the hippocampus, promoting synaptic plasticity and neurogenesis.^[52] This mechanism explains the link between BDNF deficiency and cognitive decline.^[53,54] indicating that muscle status directly regulates the brain’s neurotrophic environment. Beyond neurotrophic support, chronic low-grade inflammation is another key driver. While previous literature has noted the role of inflammation.^[55] recent immunological evidence from the MCR cohort study by Merchant et al offers more precision: MCR development is accompanied by elevated pro-inflammatory factors (e.g., tumor necrosis factor-alpha) and a significant reduction in anti-inflammatory interleukin-10 levels.^[56] This immune imbalance may induce chronic microglial activation, causing neurotoxic damage in cortical regions responsible for motor control and executive function.^[56] edition analysis further establishes sedentary behavior as a pivotal mechanistic hub. Prolonged inactivity directly elevates the risk of cardiovascular and neurodegenerative diseases by impairing cerebral hemodynamics and accelerating neurotoxic β-amyloid deposition.^[57] On the metabolic front, it induces insulin resistance and disrupts the regulation of adipokines such as leptin^[58] and adiponectin,^[59] thereby forming a biological bridge linking sarcopenia to cognitive decline. Crucially, physical inactivity triggers “anabolic resistance,” blunting muscle responsiveness to protein synthesis signals.^[60] This ultimately forms a self-reinforcing vicious feedback loop: functional limitations caused by sarcopenia force patients into a sedentary lifestyle, which in turn exacerbates muscle loss and cerebrovascular burden, accelerating pathological cognitive decline.

4.1. Clinical implications and future outlook

This study shows the link between sarcopenia and MCR through MR analysis. It provides a basis for stopping cognitive decline by treating sarcopenia early. Clinically, finding sarcopenia early is a key step to prevent brain diseases. Doctors should not only focus on fall risks. They should also monitor walking speed and check for memory complaints. These steps help create a better way to find early cognitive loss.

Sedentary behavior is a key factor between sarcopenia and MCR. This finding suggests we should change how we give exercise advice. Many patients with sarcopenia feel tired easily. They often find heavy exercise difficult. Therefore, “reducing sitting time” is a more practical goal for them. Simple changes, such as standing or walking for 2 minutes every half hour, can help. These activities improve blood flow to the brain and lower inflammation.^[61] Medical advice should focus on these small, daily habits. This approach makes it easier for patients to follow the plan. It also helps break the bad cycle caused by sitting too much.

Public health plans should manage physical activity throughout the whole day. Community centers should screen for both muscle loss and memory issues. They can use grip strength and walking speed to find high-risk people. National health rules should also focus more on sitting time. Reducing sedentary time is just as important as increasing exercise.

Future plans should combine strength training, less sitting, and brain exercises. New trials should test if this combined way can reverse MCR. We also need to study how different exercises affect the brain. For example, we should look at how resistance and aerobic exercise work differently. Finding specific muscle factors will help us create better, personal rehab plans in the future.

5. Strengths and limitations

This study has several strengths: it is the first to reveal the bidirectional association between sarcopenia and MCR from a genetic perspective and to confirm the mediating role of sedentary behavior. These findings provide important clinical implications, suggesting that interventions targeting sarcopenia and sedentary behavior may help reverse MCR in its early stages and delay disease progression. Methodologically, the study employed multiple MR techniques and sensitivity analyses, thereby significantly reducing the influence of confounding factors and reverse causality and enhancing the reliability of the results. However, this study also has limitations. First, selecting only cognitive function as the key measure of MCR may introduce measurement bias. Future studies should incorporate more multidimensional assessment indicators. Second, there was some heterogeneity in the MR analysis, but the use of a random-effects IVW model provided more conservative estimates, and sensitivity analyses confirmed the stability of the results. Additionally, for certain outcome measures (e.g., cognitive function), the number of SNPs meeting genome-wide significance was insufficient, so the threshold was relaxed to <5 × 10^‐6. While this increases the number of IVs, it may also introduce potential confounders. Finally, the data were primarily sourced from the MRC-IEU and UK Biobank, which have inherent limitations, including measurement errors and sample selection bias. Future studies using multicenter data for validation would help further elucidate the mechanisms linking sarcopenia and MCR and strengthen the scientific validity of the MR findings.

6. Conclusion

In conclusion, this study establishes a robust causal framework linking sarcopenia to MCR, identifying genetically determined low muscle mass, reduced handgrip strength, and slower WP as significant risk factors. Notably, the revelation that sedentary behavior mediates this pathological pathway highlights a modifiable target for disrupting the trajectory from physical frailty to cognitive impairment. For the aging population, these findings underscore that preserving physical function is intrinsic to maintaining cognitive health. Consequently, there is a critical need to shift towards integrated management approaches that address both sarcopenia and cognitive decline simultaneously. Public health initiatives and clinical strategies should prioritize holistic interventions (specifically targeting reductions in sedentary behavior and improvements in muscle function) to effectively mitigate the dual burden of physical and cognitive dysfunction in older adults.

Acknowledgments

We would like to extend our sincere gratitude to all the Genome-Wide Association Studies (GWAS) consortia that have made the pooled data publicly accessible. Their commitment to open science was instrumental in advancing our research. We are also deeply appreciative of the numerous investigators and participants who contributed their time, expertise, and resources to these studies.

Author contributions

Data curation: Hao Peng.

Formal analysis: Yanping Song.

Funding acquisition: Qigang Chen.

Investigation: Na Yao.

Methodology: Zhijuan He.

Project administration: Hongbo Chen, Li Huang.

Resources: Yang Jiang.

Software: Pengcheng Li.

Supervision: Zhen Shen.

Writing – original draft: Xingxiao Yin.

Writing – review & editing: Xingxiao Yin.