Insulin resistance and sarcopenia: a prognostic longitudinal link to stroke risk in middle-aged and elderly Chinese population

Abstract

Background

Stroke is the leading cause of death in middle-aged and elderly people in China. Insulin resistance (IR) and sarcopenia are both closely associated with metabolic diseases. However, the relationship between these two indicators and stroke has not been fully investigated. The aim of this study was to investigate the relationship between IR and sarcopenia and the risk of new-onset stroke.

Methods

Using longitudinal data from the China Health and Retirement Longitudinal Study (CHARLS) from 2011 to 2018, Cox proportional hazards models were used to determine the association between IR surrogate indicators and sarcopenia status with stroke incidence.

Results

In the present study, during a median 7 years of follow-up, we included 7009 middle-aged and elderly residents, of whom 515 presented with stroke incidence. After adjustment for potential confounders, both baseline IR surrogates and sarcopenia independently predicted stroke risk. In addition, co-morbidities had a higher risk of stroke than other groups. The positive association between TyG-WC and sarcopenia on stroke risk was particularly significant [HR (95% CI): 2.03 (1.52, 2.70)]. In subgroups of different ages and sexes, the combination of IR and sarcopenia is associated with the highest risk of stroke.

Conclusions

We found that IR and sarcopenia synergistically increase the incidence of stroke in older adults. This finding provides new perspectives for stroke detection and intervention and highlights the importance of early detection and management of IR and sarcopenia in older adults.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12889-024-20214-4.

Introduction

Stroke, caused by cerebral circulation disorders, is the fourth leading cause of death worldwide [1]. According to the Global Burden of Diseases Study 2019, the prevalence of stroke in China in 2019 increased sharply by 147.5% compared to 1990 [2], underscoring the urgent need for stroke prevention measures. Despite interventions targeting known risk factors for stroke, such as hypertension, diabetes, and smoking, stroke remains a leading cause of death and disability among the middle-aged and elderly in China. This suggests that other under-recognized risk factors may contribute to stroke risk.

Insulin resistance (IR), an essential pathological process in type 2 diabetes mellitus (T2DM), is clearly associated with hypertension and coronary atherosclerotic heart disease [3, 4]. Given the shared risk factors between stroke and cardiovascular disease, IR may also be a risk factor for stroke. One study has demonstrated that IR may increase stroke risk by impairing cerebrovascular function and accelerating atherosclerosis progression [5]. However, clinically diagnosing IR is challenging due to the complexity and time-consuming nature of the gold standard for estimating IR [6]. Consequently, novel indices for IR assessment based on biochemical indicators and anthropometric components, such as triglyceride-glucose (TyG) [7] and TyG-body mass index (TyG-BMI) [8], are receiving increasing attention. Several studies have demonstrated an association between TyG and stroke risk [9, 10], but further investigation is required to determine the relationship between other IR indices and stroke.

Sarcopenia, the age-related loss of muscle mass and strength, is prevalent among the elderly, posing a significant threat to their quality of life and increasing the risk of falls, disability, and even death [11]. Stroke patients are more susceptible to sarcopenia compared to those who have not had a stroke [12, 13], while this correlation seems to be more marked in the moderate-to-severe stroke population [14]. And it has also been shown that sarcopenia is an independent predictor of functional outcome after stroke [15]. Notably, sarcopenia exhibits a significant association with IR, and factors such as obesity, inflammatory agents, and hormonal shifts may explain their mutual relationship [16]. IR can potentially disrupt muscle protein homeostasis, leading to decreased muscle mass and function, while sarcopenia may exacerbate this imbalance by reducing the muscle’s ability to metabolize glucose. IR and sarcopenia promote physical deterioration through interactive pathways, then the co-occurrence of IR and sarcopenia may result in additive effects on the risk of stroke than either condition alone, particularly in middle-aged and older populations already at increased risk. However, the exact relationship between this combined effect and stroke risk is not fully understood. Hence, investigating the impact of IR and sarcopenia on stroke development in middle-aged and elderly individuals is crucial for enhancing stroke prevention and treatment strategies, potentially paving the way for a novel approach to health management in this demographic.

Methods

Study design and participants

This study is a longitudinal cohort study aimed at examining the synergistic impact of IR and sarcopenia on stroke incidence in a middle-aged population. Participants were recruited from the China Health and Retirement Longitudinal Study (CHARLS) baseline survey, which was conducted between 2011 and 2012 and covered 150 counties/districts, 450 villages/urban communities in 28 provinces across China. To date, CHARLS has completed five rounds of data collection, including the 2011 baseline survey and four subsequent annual follow-up surveys from 2011 to 2020.The study was reviewed and approved by the Biomedical Ethics Committee of Peking University (IRB 00001052–11015), and all participants signed an informed consent form. Detailed information on the CHARLS study design has been described in the literature [17].

Participants were selected for the study based on CHARLS data from 2011 to 2018. Inclusion criteria included individuals aged > 45 years who were free of stroke at baseline. Exclusion criteria were individuals with incomplete information on stroke during follow-up, missing baseline data on sarcopenia status, missing blood-based bioassays, missing information on age and sex, and other incomplete demographic information. Finally, we included 7009 participants who did not report a history of stroke in 2011 and who also completed follow-up in 2013, 2015, and 2018 in this study. The detailed inclusion and exclusion process are shown in Fig. 1.

Fig. 1.

Flow diagram for participants included in the study. China Health and Retirement Longitudinal Study, CHARLS

Assessment of sarcopenia status

In this study, we employed the 2019 Asian Working Group for Sarcopenia (AWGS) criteria to define sarcopenia [18]. Participants were evaluated for hand strength using the Yuejian WL-1000 dynamometer, with measurements taken twice for the dominant and nondominant hands, and the maximum value recorded. In accordance with the AWGS 2019 criteria, diagnostic thresholds for sarcopenia were established at < 28 kg for men and < 18 kg for women. Muscle mass was then estimated by employing the Appendicular Skeletal Muscle Mass (ASM) formula, which considers factors such as weight, height, sex, and age. A previous study have demonstrated substantial concordance between the ASM formula and double X-ray absorptiometry (DXA) [19]. In accordance with previous studies [20–22], individuals with height-adjusted muscle mass (ASM/Ht²) below 20% of the population were classified as having low ASM. The values of < 5.31 kg/m² for women and < 7.05 kg/m² for men were considered indicative of low muscle mass. Low physical performance was defined as walking speed < 1.0 m/s or 5-time chair stand test ≥ 12 s. Sarcopenia was diagnosed when there was low muscle mass plus low physical performance or low muscle strength. Individuals exhibiting solely low muscle strength or physical performance are classified as having possible sarcopenia; whereas a decline in all three factors indicates severe sarcopenia. Accordingly, the study population was divided into three groups: non-sarcopenia (n = 4399), possible sarcopenia (n = 1902), and sarcopenia (n = 708), with individuals with severe sarcopenia included in the sarcopenia group.

Assessment of stroke events

The primary outcome was the incidence of stroke. The time to event was defined as the duration from the baseline assessment to the first occurrence of stroke, censoring, or the end of the study period, whichever came first. Person-time at risk was calculated for each participant from the date of baseline assessment until the data of stroke occurrence, death, loss to follow-up, or the end of the study period. The outcome of this study was dependent on participants’ self-reported physician diagnosis. Patients with stroke were identified by asking “Have you been diagnosed with stroke by a doctor?” in the “Health Status and Functioning” section of the CHARLS questionnaire. If participants answered yes, they were recorded as stroke cases. Furthermore, “When was stroke first diagnosed or known by yourself?” was used to determine the exact time of stroke onset. This methodology was similar to that employed in previous studies [23, 24], thereby ensuring consistency and accuracy in the diagnosis of stroke.

Calculation of IR surrogate indicators

In addition, the present study employed IR alternative indicators. Fasting plasma glucose (FPG), triglyceride (TG), and high-density lipoprotein cholesterol (HDL-C) were quantified by drawing venous blood from the subjects after an overnight fast. The subjects’ height, weight, and waist circumference (WC) were measured using the SecaTM 213 height meter, OmronTM HN-286 scale, and a soft ruler, respectively. Six surrogate indicators for consideration as IR were estimated by validated equation [8, 25–27].

$$\mathit{TyG}=\mathit{Ln}\left[\left(\mathit{TG},\times ,,\mathit{FPG}\right),/,2\right]$$ 1

$$TyG-BMI=TyG\times BMI\left(B,M,I,,=,,w,e,i,g,h,t,,\left(k,g\right),/,h,e,i,g,h,t^2,,\left(m^2\right)\right)$$ 2

$$TyG-WC=TyG\times WC$$ 3

$$TyG-WHtR=TyG\times WHtR(WHtR=WC\left(c,m\right)/Height\left(c,m\right)$$ 4

$$TG/HDL-C=TG/HDL-C$$ 5

$$METS-IR=Ln\left[\left(2,\times ,F,P,G\right),,+,,T,G\right]\times BMI/Ln\left(H,D,L,-,C\right)$$ 6

Given the absence of uniform criteria for categorizing these IR surrogate indicators, we adopted the median value within our study population as a cutoff to distinguish between high and low levels.

Assessment of covariates

At the study baseline, professionals collected comprehensive data on participants, including demographic information, health behaviors, chronic disease history, drug utilization, and blood test results. Demographic information encompassed age, sex, area of residence (rural or urban), and education level (elementary school and below, secondary school, and college and above). Health-related factors included height, weight, BMI, WC, WHtR, and smoking and drinking habits. The dataset regarding chronic illness incidence and drug utilization encompasses diagnostic and therapeutic insights for hypertension, diabetes, dyslipidemia, and kidney disease, along with a detailed account of medication usage for each of these conditions. The chronic disease such as diabetes and dyslipidemia were identified based on participant’s self-reports of a physician diagnosis, following the same approach used for the evaluation of stroke events in the CHARLS database. We did not utilize biochemical indicators or diagnostic tests for the definition of these conditions. All data were collected according to strict scientific and ethical standards, ensuring the accuracy of the study and the privacy of the participants.

Statistical analysis

In this study, we initially assessed the distributional characteristics of continuous variables. For continuous variables with skewed distributions, we expressed as median and interquartile range (IQR). For continuous variables with normal distributions, we used mean and standard deviation (SD) to represent the data. Categorical variables were presented as numbers and percentages. Subjects were stratified into four groups based on sarcopenia status and IR surrogate indexes levels at baseline: (1) non-sarcopenia and low IR levels; (2) high IR levels only; (3) possible sarcopenia and sarcopenia only; (4) coexistence of sarcopenia and high IR levels. To compare baseline characteristics between the different groups, we applied the χ2 test and the Mann-Whitney U test. Subsequently, we employed the Cox proportional hazards model to assess the relationship between baseline sarcopenic status, IR levels, and the incidence of stroke, calculating hazard ratios (HRs) and their 95% confidence intervals (CIs).

For the longitudinal analyses, participant with missing follow-up information on stroke(n = 1491) were excluded from the analysis. This decision was based on several factors: the potential of noise and bias if such data were included as right-censored, the non-random nature of the missing data, and the need to maintain the integrity and reliability of the Cox proportional hazards model assumptions. This approach aligned with established practices in similar epidemiological studies [28, 29] to ensure the robustness and validity of our findings. we constructed three stepwise-adjusted Cox regression models to control for potential confounders. Model 1 was unadjusted; Model 2 was adjusted for age and sex; while Model 3 was further adjusted for systolic, diastolic, smoking, drinking, hypertension, diabetes, heart disease, dyslipidemia, hypertension medications, diabetes medications and dyslipidemia medications, based on model 2. For multivariate analysis, variables that were significant in the univariate analysis (P < 0.05), were clinically relevant and exhibited no multicollinearity(Variance Inflation Factors, VIF<10) were selected. All statistical analyses were conducted using SPSS version 27.0 software (SPSS, Inc., Chicago, IL, USA) and R Statistical Software (Version 4.2.2). In all cases, a p-value of less than 0.05 was considered statistically significant.

Results

Baseline characteristics of the participants

Seven thousand and nine participants free of stroke at baseline were included in the final analysis. The median age of the participants was 57 years, and 3,169 (45.21%) were men. During a median 7 years of follow-up,515 new cases of stroke were recorded, representing 7.35% of the total. Six alternative IR indicators, including TyG, TyG-BMI, TyG-WC, TyG-WHtR, TG/HDL-C, and METS-IR, exhibited significantly elevated levels in patients with new-onset stroke compared to the control group without stroke (P < 0.001). This suggests that these proxies are associated with an increased risk of stroke. The prevalence rates are 10.10% for sarcopenia, 27.14% for possible sarcopenia and 62.76% for non-sarcopenia. In comparison to patients with no history of stroke, those with a diagnosis of stroke were observed to be older, to have higher values of BMI and WC, and to have a greater prevalence of hypertension, diabetes, and dyslipidemia, as well as a higher frequency of medication use for related disorders (P < 0.001). A detailed account of these demographic and clinical characteristics is presented in Table 1. Based on the results of the normality test, all continuous variables were non-normally distributed, so we have expressed these variables as median (IQR) in Table 1. This approach is more appropriate for skewed data because it more accurately reflects central tendency and variability and is less affected by extreme values.

Table 1.

Baseline patient characteristics grouped by outcomes

Variables	All (n = 7009)	Non-stroke (n = 6494)	Stroke (n = 515)	P
Age, years, median(IQR)	57(51,64)	57(51,64)	61(55,67)	<0.001
Men, n(%)	3169(45.21)	2920(44.96)	249(48.35)	0.137
Rural, n(%)	4671(66.64)	4338(66.80)	333(64.66)	0.321
Educational level, n (%)				0.071
Elementary school or bellow	4869(69.47)	4492(69.17)	377(73.20)
Secondary school	2065(29.46)	1929(29.70)	136(26.41)
College and above	75(1.07)	73(1.12)	2(0.39)
BMI, Kg/m2, median(IQR)	23.20(20.93,25.80)	23.13(20.90,25.73)	24.01(21.69,26.66)	<0.001
WC, cm, median(IQR)	85.00(78.00,92.00)	84.50(78.00,91.80)	88.10(80.60,95.00)	<0.001
WHtR, median(IQR)	0.54(0.49,0.58)	0.53(0.49,0.58)	0.56(0.51,0.60)	<0.001
Blood pressure, mmHg
Systolic, median(IQR)	125.50(113.00,140.00)	125(113.00,139.50)	135(120.00,152.00)	<0.001
Diastolic, median(IQR)	74.00(66.50,82.50)	74.00(66.50,82.00)	78.50(71.00,87.50)	<0.001
Pulse, median(IQR)	71.50(65.00,78.00)	71.00(65.00,78.00)	72.00(66.00,79.00)	0.012
Smoking, n (%)	2666(38.04)	2452(37.76)	214(41.55)	0.088
Drinking, n (%)	2695(38.45)	2478(38.16)	217(42.14)	0.074
Comorbidities, n (%)
Hypertension	1712(24.43)	1463(22.53)	249(48.35)	<0.001
Diabetes	384(5.48)	335(5.16)	49(9.51)	<0.001
Heart disease	778(11.10)	677(10.43)	101(19.61)	<0.001
Dyslipidemia	629(8.97)	534(8.22)	95(18.45)	<0.001
Kidney disease	382(5.45)	352(5.42)	30(5.83)	0.697
History of medication use, n (%)
Hypertension medications	1256(17.92)	1066(16.42)	190(36.89)	<0.001
Dyslipidemia medications	337(4.80)	285(4.39)	52(10.10)	<0.001
Diabetes medications	236(3.37)	199(3.06)	37(7.18)	<0.001
FBG, mg/dL, median(IQR)	102.24(94.32,112.68)	102.06(94.14,112.50)	104.22(95.94,115.92)	<0.001
HbA1c, median(IQR)	5.10(4.90,5.40)	5.10(4.90,5.40)	5.20(4.90,5.50)	0.008
TC, mg/dL, median(IQR)	190.59(167.78,214.95)	190.59(167.40,214.95)	193.30(168.17,218.43)	0.143
TG, mg/dL, median(IQR)	105.32(75.23,153.11)	104.43(74.34,151.34)	116.82(84.96,163.73)	<0.001
HDL-C, mg/dL, median(IQR)	49.48(40.59,59.92)	49.87(40.98,60.31)	46.78(39.05,56.06)	<0.001
LDL-C, mg/dL, median(IQR)	114.43(93.94,137.24)	114.43(93.94,136.86)	117.52(93.94,140.34)	0.101
TyG, median(IQR)	8.59(8.22,9.03)	8.58(8.21,9.03)	8.74(8.36,9.11)	<0.001
TyG-BMI, median(IQR)	199.73(176.06,229.13)	198.93(175.49,228.23)	211.94(185.29,239.30)	<0.001
TyG-WC, median(IQR)	730.08(655.24,817.60)	726.41(653.06,814.84)	775.41(688.43,849.98)	<0.001
TyG-WHtR, median(IQR)	4.64(4.14,5.19)	4.62(4.13,5.17)	4.88(4.39,5.43)	<0.001
TG/HDL-C, median(IQR)	2.10(1.32,3.56)	2.07(1.30,3.51)	2.42(1.61,4.06)	<0.001
METS-IR, median(IQR)	34.37(29.78,40.28)	34.16(29.64,39.99)	36.44(31.61,42.48)	<0.001
Sarcopenia status				<0.001
Non-sarcopenia	4399(62.76)	4145(63.83)	254(49.32)
Possible sarcopenia	1902(27.14)	1699(26.16)	203(39.42)
Sarcopenia	708(10.10)	650(10.01)	58(11.26)

IQR Interquartile range, BMI Body mass index, WC Waist circumference, WHtR The ratio of waistline to height, FPG Fasting plasma glucose, HbA1c Glycosylated hemoglobin, TC Total cholesterol, TG Triglyceride, HDL-C High-density lipoprotein cholesterol, LDL-C Low-density lipoprotein cholesterol, TyG Triglyceride glucose, METS-IR The metabolic score for insulin resistance

Cox regression analysis of baseline IR and sarcopenia on follow-up stroke

Table 2 and Supplementary Table 1 provide a detailed illustration of the association between the six IR surrogate indicators (TyG, TyG-BMI, TyG-WC, TyG-WHtR, TG/HDL-C, and METS-IR) and the risk of stroke, as well as the status of sarcopenia. In a multivariate analysis, a significant association was observed of TyG (per SD increment) with the risk of incident stroke[HR (95% CI): 1.27 (1.05,1.52)]. Similarly, other IR surrogates demonstrated a statistically significant correlation with increased stroke risk, including TyG-BMI [HR (95% CI): 1.10 (1.01,1.20)], TyG-WC [HR (95% CI): 1.12 (1.02,1.22)], TyG-WHtR [HR (95% CI): 1.12 (1.02,1.23)], and METS-IR [HR (95% CI):1.27 (1.18,1.37)]. However, the TG/HDL-C result was attenuated, and no significant relationship was found [HR (95% CI): 1.06 (0.99,1.14)]. In the mutual adjustment models, TyG, TyG-BMI, TyG-WC, TyG-WHtR and METS-IR remained statistically significant. When sarcopenia status was categorized into three types, it was found that possible sarcopenia was associated with an increased risk of stroke compared with the non-sarcopenia population, after adjusting for other confounders [HR (95% CI): 1.49 (1.23,1.81)]. Nevertheless, the risk of stroke in the sarcopenia population did not reach statistical significance [HR (95% CI): 1.30 (0.95,1.77)]. Furthermore, when sarcopenia was simplified into two categories, namely non-sarcopenia/possible sarcopenia and sarcopenia combined groups, a notable increase in stroke risk was observed among the latter [HR (95% CI): 1.45 (1.20,1.75)]. In addition, a significant association was observed of possible sarcopenia and sarcopenia with the risk of incident stroke in the multivariate analyses even though adjustments were made for IR surrogate indicators. These results indicate that IR and sarcopenia status are important predictors of stroke hazard.

Table 2.

Relationship of IR surrogate indicators or sarcopenia status with stroke risk

Characteristics	Crude model		Adjusted model		P for interactionc
	HR(95%CI)	P	HR(95%CI)	P
IR surrogate indicators
TyG (per SD)	1.26 (1.16,1.36)	< 0.001	1.13 (1.03,1.23)	0.006	0.221
TyG-BMI (per SD)	1.27 (1.18,1.37)	< 0.001	1.10 (1.01,1.20)	0.035	0.384
TyG-WC (per SD)	1.35 (1.25,1.47)	< 0.001	1.12 (1.02,1.22)	0.013	0.222
TyG-WHtR (per SD)	1.33 (1.23,1.45)	< 0.001	1.12 (1.02,1.23)	0.015	0.240
TG/HDL-C (per SD)	1.12 (1.06,1.19)	< 0.001	1.06 (0.99,1.14)	0.087	0.164
METS-IR (per SD)	1.27 (1.18,1.37)	< 0.001	1.11 (1.02,1.21)	0.017	0.596
Sarcopenia statusa
Non-sarcopenia	1.00(Reference)		1.00(Reference)
Possible sarcopenia	1.90 (1.58,2.28)	< 0.001	1.49(1.23,1.81)	< 0.001
Sarcopenia	1.44 (1.08,1.91)	0.013	1.30 (0.95,1.77)	0.101
Sarcopenia statusb
Non-sarcopenia	1.00(Reference)		1.00(Reference)
Possible sarcopenia and sarcopenia	1.77 (1.49,2.11)	< 0.001	1.45 (1.20,1.75)	< 0.001

Adjusted model: adjusted for Age, Sex, Systolic, Diastolic, Smoking, Drinking, Hypertension, Diabetes, Heart disease, Dyslipidemia, Hypertension medications, Diabetes medications and Dyslipidemia medications. IR surrogate indicators is analyzed as a continuous variable in the table

^aSarcopenia status is classified into three categories

^bSarcopenia status is classified into two categories

^cInteraction between IR surrogate indicators and sarcopenia status^b

The combination of IR and sarcopenia increased the risk of stroke

Supplementary Figs. 1–6 showed the cumulative incidence of stroke based on sarcopenia status and IR surrogate indexes levels at baseline (Log rank test P < 0.001). Table 3 demonstrated the crude and multivariable-adjusted associations between the primary exposure and stroke risk. To illustrate, consider the combination of TyG and sarcopenia status. Incident stroke rates of the 4 groups based on the combination of IR and sarcopenia were 6.89,10.84,9.77 and 18.13 per 1000 person-years in non-sarcopenia/low TyG, possible sarcopenia and sarcopenia/low TyG, non-sarcopenia/high TyG, and possible sarcopenia and sarcopenia/high TyG groups, respectively. In the unadjusted model, the other three groups had significant relationship to incident stroke relative to non-sarcopenia/low TyG. These differences remained statistically significant in Model 2, which was adjusted for age and sex. Further, after adjusting for potential confounders, including blood pressure, smoking and drinking habits, history of chronic disease, and medication use in Model 3, we found that possible sarcopenia and sarcopenia/high TyG was associated with a 1.82-fold higher hazard of experiencing stroke [HR (95% CI): 1.82 (1.40, 2.37)]. Furthermore, the combination of other IR surrogates with sarcopenia demonstrated a similar pattern. After adjusting for all covariates, participants with possible sarcopenia and sarcopenia accompanied by high TyG-BMI, high TyG-WC, high TyG-WHtR, high TG/HDL-C, and high METS-IR were more likely to increase the risk of stroke. The respective HRs (95% CI) were 1.86 (1.41, 2.47), 2.03 (1.52, 2.70), 1.98 (1.51, 2.62), 1.90 (1.47, 2.47), and 1.91 (1.44, 2.53). These findings underscore the significance of IR and sarcopenia in the assessment of stroke risk.

Table 3.

Association of Sarcopenia status and IR surrogate indicators with new onset of stroke

Group	Case	Incidence rate (per1000 person-year)	Model 1 HR (95%CI)	Model 2 HR (95%CI)	Model 3 HR (95%CI)
Non-sarcopenia & Low TyG	104	6.89	1.00(Reference)
Possible sarcopenia and sarcopenia & Low TyG	95	10.84	1.59(1.20,2.10)**	1.40 (1.05,1.86)*	1.30 (0.97,1.73)
Non-sarcopenia & High TyG	150	9.77	1.43(1.11,1.83)**	1.46 (1.13,1.87)**	1.16 (0.90,1.50)
Possible sarcopenia and sarcopenia & High TyG	166	18.13	2.70(2.11,3.45)***	2.44(1.89,3.16)***	1.82(1.40,2.37)***
Non-sarcopenia & Low TyG-BMI	90	6.02	1.00(Reference)
Possible sarcopenia and sarcopenia & Low TyG-BMI	110	11.84	2.00(1.51,2.64)***	1.70(1.27,2.26)***	1.63 (1.22,2.17)**
Non-sarcopenia & High TyG-BMI	164	10.58	1.78(1.37,2.30)***	1.92(1.48,2.49)***	1.37 (1.05,1.79)**
Possible sarcopenia and sarcopenia & High TyG-BMI	151	17.49	2.98(2.30,3.87)***	2.86(2.19,3.75)***	1.86(1.41,2.47)***
Non-sarcopenia & Low TyG-WC	81	5.26	1.00(Reference)
Possible sarcopenia and sarcopenia & Low TyG-WC	106	11.93	2.31(1.73,3.08)***	2.00(1.49,2.70)***	1.90(1.41,2.55)***
Non-sarcopenia & High TyG-WC	173	11.49	2.22(1.70,2.88)***	2.24(1.72,2.92)***	1.62 (1.23,2.12)**
Possible sarcopenia and sarcopenia & High TyG-WC	155	17.12	3.34(2.55,4.37)***	3.03(2.29,4.00)***	2.03(1.52,2.70)***
Non-sarcopenia & Low TyG-WHtR	95	6.02	1.00(Reference)
Possible sarcopenia and sarcopenia & Low TyG-WHtR	91	10.86	1.83(1.37,2.44)***	1.63 (1.21,2.18)**	1.56 (1.16,2.10)**
Non-sarcopenia & High TyG-WHtR	159	10.83	1.82(1.41,2.35)***	1.94(1.50,2.51)***	1.42 (1.09,1.85)**
Possible sarcopenia and sarcopenia & High TyG-WHtR	170	17.77	3.03(2.36,3.90)***	2.92(2.24,3.82)***	1.98(1.51,2.62)***
Non-sarcopenia & Low TG/HDL-C	103	6.81	1.00(Reference)
Possible sarcopenia and sarcopenia & Low TG/HDL-C	97	10.77	1.60 (1.21,2.11)**	1.38 (1.04,1.84)*	1.28 (0.96,1.71)
Non-sarcopenia & High TG/HDL-C	151	9.84	1.46 (1.13,1.87)**	1.48 (1.16,1.91)**	1.20 (0.93,1.55)
Possible sarcopenia and sarcopenia & High TG/HDL-C	164	18.40	2.77(2.17,3.55)***	2.51(1.94,3.23)***	1.90(1.47,2.47)***
Non-sarcopenia & Low METS-IR	87	5.83	1.00(Reference)
Possible sarcopenia and sarcopenia & Low METS-IR	113	12.11	2.11(1.59,2.79)***	1.79(1.34,2.39)***	1.70(1.27,2.27)***
Non-sarcopenia & High METS-IR	167	10.75	1.87(1.44,2.42)***	1.98(1.52,2.56)***	1.44 (1.10,1.88)**
Possible sarcopenia and sarcopenia & High METS-IR	148	17.23	3.03(2.33,3.95)***	2.85(2.18,3.74)***	1.91(1.44,2.53)***

Model 1: unadjusted; Model 2: adjusted for age, sex; Model 3: adjusted as model 2 with further adjustment for systolic, diastolic, smoking, drinking, hypertension, diabetes, heart disease, dyslipidemia, hypertension medications, diabetes medications, dyslipidemia medications

*P < 0.05, **P < 0.01, *** P < 0.001; IR surrogate category: divide into high value and low value groups based on the median

Subgroup analysis

Figure 2 further refines the analysis, demonstrating the significant impact of higher levels of the IR surrogate on stroke risk when combined with possible sarcopenia and sarcopenia status in different age and sex subgroups. The results of the analysis showed that individuals with higher levels of IR exhibited a higher risk of stroke across all age groups and genders. This risk correlation was particularly significant in the elderly population, which may be related to the increased prevalence of sarcopenia with age. Furthermore, sex-stratified analyses demonstrated that the effects of IR and sarcopenia on stroke risk were prevalent in both men and women, indicating the prevalence and importance of this risk factor.

Fig. 2.

Subgroup analysis of IR surrogate indicators and sarcopenia status with stroke risk. Hazard Ratio (HR) were adjusted for age, sex, systolic, diastolic, smoking, drinking, hypertension, diabetes, heart disease, dyslipidemia, hypertension medications, diabetes medications and dyslipidemia medications. Body mass index, BMI. Waist circumference, WC. The ratio of waistline to height, WHtR. Triglyceride glucose, TyG. The metabolic score for insulin resistance, METS-IR

Discussion

In this longitudinal study, we found a significant positive association between IR and sarcopenia on the risk of stroke in the middle-aged and elderly population. In particular, the combination of high IR and sarcopenia synergistically increased the potential stroke incidence in the target adults. Notably, those with possible sarcopenia and sarcopenia/high TyG-WC showed the strongest association with stroke compared with those in the non-sarcopenia/low TyG-WC group. To our knowledge, this is the first large national cohort study to clarify the combined effects of IR and sarcopenia on stroke risk.

The results of this study reinforce those of previous studies which have demonstrated a significant positive correlation between the TyG index and the risk of stroke [30–32]. Individuals with higher TyG levels have a correspondingly higher incidence of stroke, both in the general population and in patients with hypertension or diabetes [33–35]. Du et al. [36] indicated that the risk of ischemic stroke increased by 72% within the range of high TyG-BMI, based on two cross-sectional surveys in northeastern China. Another longitudinal study, also on 4583 subjects in CHARLS, showed that substantial changes in the TyG-BMI were independently associated with stroke risk in middle-aged and older adults [37]. Furthermore, Huang et al. [38] found that the risk of future stroke was 96% higher in the high TG and high WC groups than in the low TG and low WC groups. In our study, we also observed a positive correlation between TyG-WHtR and stroke events, which provides further evidence for the relationship between IR indicators and the risk of stroke. Regarding TG/HDL-C, a previous study have demonstrated its role as an independent risk factor for cardiovascular events [39]. However, the relationship between TG/HDL-C and the outcome of stroke was not consistent across all studies. The results by Tang et al. demonstrated that the TG/HDL-C level was linked to the risk of stroke after controlling for covariates (HR: 1.65, 95% CI: 1.28–2.13) [40]. It is noteworthy that in a prospective cohort study in Japan involving 12 districts [41], it was found that elevated levels of TG/HDL-C did not contribute to the prediction of the risk of stroke in the overall study population. Our results supported this latter idea. These results suggest a nonlinear relationship and a threshold effect between TG/HDL-C and stroke. Consequently, it is imperative to ascertain the optimal distribution of TG/HDL-C levels for the early intervention of patients with a potential for cryptogenic stroke to enhance their quality of life. METS-IR, a novel index for screening insulin sensitivity, has been validated in a large-scale cohort to be as diabetes predictor [30] as well as has been strongly associated with a variety of stroke risk factors [42–45]. Our results demonstrated that elevated METS-IR may signal a high risk of suffering stroke in the future, after adjustment for confounding factors such as smoking history and alcohol consumption.

The study revealed that, alongside IR, sarcopenia was an equally significant predictor of stroke risk. The results demonstrated that individuals with possible sarcopenia had a higher risk of stroke compared to the non-sarcopenia population. Gao et al. noted that sarcopenia is associated with an increased risk of cardiovascular disease (CVD) in middle-aged and elderly populations [22]. Secondly, a large cohort study by Chai et al. in Taiwan [46] also confirmed that sarcopenia was associated with a higher prevalence of stroke in patients with T2DM. However, in our study, when sarcopenia status was presented in three categories, after adjusting for covariates, we found that there was no statistically significant difference between sarcopenia and risk of stroke, which may be due to the limitation of the number of sarcopenic populations; meanwhile, the population of sarcopenia in our study was more predominantly female, whereas a cross-sectional cohort study in Korea [47] did not find a significant correlation between sarcopenia and the incidence of stroke in the female population. Thus, more prospective multicenter studies are need for further validation. In light of these findings, it is of primary and vital importance to implement targeted interventions for patients diagnosed with sarcopenia, especially for possible sarcopenia. These include physical therapy to increase muscle strength and function, and medication to improve muscle health, which is expected to reduce the incidence of stroke.

The hypothesis that IR and sarcopenia can synergistically increase the risk of incident stroke was proposed, and the present study demonstrated that people with high IR levels and suffering from possible sarcopenia and sarcopenia had a 1.82- to 2.03-fold higher risk of stroke when compared with people with non-sarcopenia and low IR levels. Furthermore, the reliability of this finding was verified in subgroup analyses of age (45–60 years old or ≥ 60 years old) and gender. These findings further confirm the role of IR and sarcopenia in stroke risk assessment and suggest that those at high risk should be targeted for early diagnosis and intervention, when developing stroke prevention strategies.

The precise operational mechanism of the synergistic effect remains unknown. However, several previous studies have yielded some enlightening observations. Skeletal muscle is the primary site of insulin-mediated glucose uptake in the postprandial state [48], and IR will result in abnormal muscle glucose metabolism, which in turn will lead to sarcopenia. With age, there is a fat redistribution in the body, with skeletal muscle producing most of fat infiltration. It has been demonstrated that loss of muscle mass is associated with intermuscular fat deposition in older adults [49]. Simultaneously, the accumulation of lipids in myocytes results in the activation of proinflammatory factors, inhibiting the insulin signaling pathway subsequently [50, 51]. The mutual enhancement of these factors ultimately contributes to the occurrence and development of stroke via metabolic disturbances, inflammatory processes, oxidative stress and vascular impairments [50, 52–54].

However, this study is subject to certain limitations. First, as a retrospective study, stroke diagnosis relied on subject recall, which may have introduced bias. Second, muscle mass, as one of the features for the diagnosis of sarcopenia, was not assessed using the criteria recommended by the 2019 AWGS, but was estimated using ASM formula, even though this formula has been widely used. Third, we were unable to monitor changes in IR levels and sarcopenia status over time, which limited our understanding of their dynamic role in stroke development. Future studies should consider incorporating more potential confounders and further exploring the long-term effects of changes in IR and sarcopenia status on stroke risk. The risk of stroke was found to be 1.82- to 2.03-fold higher in individuals with sarcopenia and high IR levels compared to those with non-sarcopenia and low IR levels. This finding was further validated through subgroup analyses stratified by age and gender.

Conclusions

In conclusion, this study demonstrates that IR and sarcopenia act in a synergistic manner to elevate the risk of stroke within the study group. The findings underscore the necessity for early identification and prompt intervention for individuals with sarcopenia and possible sarcopenia, particularly those with elevated IR levels. This may have significant implications for reducing the incidence of stroke in the future, and promote healthier aging among middle-aged and older individuals. Nevertheless, further prospective studies are still required to substantiate these findings and to elucidate the specific mechanisms underlying the interaction between IR and sarcopenia.

Authors’ contributions

Canhui Guo carried out the data collection, completed the statistical analysis and drafted the manuscript.Ling He were responsible for the study design, provided statistical support.Ling He and Yansong Tu revised of the manuscript. Chunyan Xu,Caifeng Liao and Hurong Lai assisted in data collection and statistical analysis. Chuyang Lin reviewed the analysis process and checked the results. Huaijun Tu conceived and supervised the project and contributed to manuscript preparation. All authors approved the final manuscript.

Funding

Not applicable.

Availability of data and materials

The data were collected from the 2011 and 2018 waves of the CHARLS survey, publicly available on the CHARLS website (http://charls.pku.edu.cn/).

Declarations

Ethics approval and consent to participate

The study was reviewed and approved by the Biomedical Ethics Committee of Peking University (IRB 00001052–11015), and all participants signed an informed consent form.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.