Abstract
Background
CHG index, a novel metabolic biomarker calculated from total cholesterol, high-density lipoprotein cholesterol, and fasting blood glucose, is associated with type 2 diabetes and cardiovascular diseases. However, its association with the risk of new-onset sarcopenia in middle-aged and older adults remains unclear.
Methods
We included participants aged ≥50 years from the CHARLS 2011–2012, with follow-up in 2015−2016. Cumulative CHG was derived from measurements in 2011 and 2015. New-onset sarcopenia was diagnosed according to the 2025 AWGS consensus. Multivariable Cox regression and logistic regression were used to examine the associations of baseline and cumulative CHG with new-onset sarcopenia. ROC curves were employed to compare the predictive capability of CHG and TyG.
Results
A total of 5580 participants were included in the baseline analysis. During the 4-year follow-up, 590 (10.6%) participants developed sarcopenia. Compared with the lowest quartile (Q1) of baseline CHG, the highest quartile (Q4) was significantly associated with an increased risk of sarcopenia (HR = 1.287, 95% CI: 1.022−1.619) in the fully adjusted model. K-means clustering categorized 3287 participants into two distinct groups based on cumulative CHG trajectories. Compared with the low-stable group (Cluster 1, n = 1783), the high-declining group (Cluster 2, n = 1504) had a significantly higher risk of new-onset sarcopenia (OR = 1.25, 95% CI: 1.02–1.45). RCS analysis revealed a linear association between baseline CHG and sarcopenia risk. The predictive performance of the baseline and cumulative CHG index was modestly higher than that of the baseline and cumulative TyG index.
Conclusion
Elevated baseline and cumulative CHG indices are independent risk factors for new-onset sarcopenia in middle-aged and older Chinese adults.
Keywords: CHG, Sarcopenia, CHARLS
1. Introduction
Sarcopenia, a progressive and generalized skeletal muscle disorder characterized by the loss of muscle mass and muscle strength, with physical performance considered an outcome measure [1]. As a core determinant of physical disability, falls, frailty and poor quality of life in older adults, sarcopenia also confers an increased risk of major adverse cardiovascular events and all-cause mortality [[2], [3], [4]]. The global prevalence of sarcopenia in adults aged 60 years and above ranges from 10% to 27% [[5], [6], [7]]. Middle-aged and older adults represent the key population for sarcopenia prevention, as muscle mass begins to decline at a rate of 1%–2% per year after the age of 50 [8,9].
Sarcopenia is a multifactorial disorder, and metabolic dysfunction and insulin resistance (IR) are recognized as core pathogenic mechanisms [10,11]. Insulin plays a pivotal role in regulating skeletal muscle anabolism and catabolism by promoting muscle protein synthesis and inhibiting proteolysis; IR impairs this regulatory effect, leading to progressive muscle mass loss [12]. In addition, abnormal lipid metabolism can induce chronic low-grade inflammation, oxidative stress and mitochondrial dysfunction in skeletal muscle cells, further exacerbating muscle atrophy [13,14]. Over the past decade, numerous metabolic composite indices have been developed to evaluate IR and metabolic status, among which the triglyceride-glucose (TyG) index and its derived indicators are the most widely studied and validated [14,15]. These indices have been demonstrated to predict the onset and progression of sarcopenia [[16], [17], [18], [19]].
Recently, the Cholesterol, High density lipoprotein, and Glucose (CHG) index, a novel metabolic composite indicator, has been proposed for the diagnosis of T2DM [20]. Subsequent cohort studies have further confirmed that the CHG index is an independent predictor of cardiovascular diseases (CVD), as it integrates three key metabolic parameters (total cholesterol (TC), HDL-C and fasting blood glucose (FBG)) to more comprehensively reflect systemic glucose, lipid metabolism and IR status [[21], [22], [23]]. However, to date, no study has elucidated the association between the baseline and cumulative CHG index and the risk of new-onset sarcopenia in middle-aged and older adults.
In view of the above research gaps, we conducted a longitudinal cohort study using data from the China Health and Retirement Longitudinal Study (CHARLS). This study aims to evaluate the relationship between the baseline and cumulative CHG index and the risk of new-onset sarcopenia in middle-aged and older adults, and compare the predictive performance of the baseline and cumulative CHG index with that of the baseline and cumulative TyG index for new-onset sarcopenia.
2. Methods
2.1. Study design
The data for this study was derived from the CHARLS database (http://charls.pku.edu.cn/en), which is freely available to the public and focuses on middle-aged and older adults aged 45 years and above in China, providing reliable and comprehensive information on the population’s health status, economic and social conditions, and related influencing factors [24]. The CHARLS baseline survey was conducted in 2011 (Wave 1), recruiting 17,708 participants from 10,257 households across 150 counties/districts and 450 villages nationwide, with follow, up surveys conducted biennially in 2013, 2015, 2018 and 2020 respectively. The current study adopted the data from the first wave (2011–2012) as the baseline and the third wave (2015–2016) as the study endpoint, as these two waves included complete blood sample detections required for CHG index calculation and physical function assessments necessary for sarcopenia diagnosis, which are the core research variables of this study.
2.2. Study population
No formal sample size calculation was conducted for this study due to its nature as a secondary analysis of CHARLS cohort data, and the final sample size was determined after rigorous application of strict eligibility criteria to ensure the reliability and validity of subsequent statistical analyses. We initially included 17,708 participants from the CHARLS 2011 baseline survey, and participants who met the following conditions were excluded: (1) Participants aged less than 50 years at baseline; (2) Participants without complete blood biochemical data (TC, HDL-C, and FBG) required for CHG index calculation at baseline; (3) Participants lacking complete data related to sarcopenia diagnosis at baseline or follow-ups; (4) Participants with incomplete CHG index data at Wave 3; (5) Participants with a baseline diagnosis of sarcopenia. After applying the above exclusion criteria, a total of 5580 participants were included in the baseline CHG analysis of this study, and a total of 3287 participants were included in the cumulative CHG analysis of this study. The CHARLS cohort was approved by the Ethics Review Committee of Peking University (IRB0000105211015), and all participants provided written informed consent at the time of enrollment, with the detailed flowchart of the participant inclusion and exclusion process presented in Fig. 1 of this study.
Fig. 1.
Flowchart of participants selection. Participants with sarcopenia diagnosed at Wave 2 were excluded from the cumulative CHG analysis due to the absence of Wave 3 biochemical data required for cumulative CHG calculation; however, they were retained as incident cases in the baseline Cox regression analysis.
2.3. Assessment of CHG index
All blood testing in the CHARLS cohort was conducted at the Youanmen Center for Clinical Laboratory at Capital Medical University, with strict quality control measures implemented throughout the testing process, and notably, the coefficients of variation for TC, HDL-C and FBG were all less than 2%, indicating a high level of reliability and consistency in the blood biochemical measurements. The calculation formula for the CHG index used in this study was defined as Ln [TC (mg/dL) × FBG (mg/dL)/(2 × HDL-C (mg/dL))], and the cumulative CHG index, which reflects the long-term CHG level changes of participants during the follow-up period, was calculated as (CHG2011 + CHG2015)/2 × follow-up time (2015−2011).
2.4. Assessment of endpoint events
The primary outcome of this study was the occurrence of new-onset sarcopenia among participants during the 4-year follow-up period from 2011–2012 to 2015–2016, with strict data quality control implemented for all endpoint assessment indicators to ensure the accuracy of the study results. Sarcopenia was diagnosed based on the 2025 Asian Working Group for Sarcopenia (AWGS) Consensus criteria, which requires the coexistence of low skeletal muscle mass plus low muscle strength, with low skeletal muscle mass defined by height-adjusted appendicular skeletal muscle mass (ASM) values and low muscle strength assessed by hand grip strength measurements. Participants with no diagnosis of sarcopenia at Wave 1 but meeting the above sarcopenia diagnostic criteria at Wave 2 or Wave 3 were classified as having new-onset sarcopenia, and all participants included in the cumulative CHG analysis completed the full 4-year follow-up to ensure data continuity and the reliability of the observed endpoint outcomes.
Low muscle mass was defined by height-adjusted criteria: for middle-aged adults (50–64 years), height-adjusted ASM (ASM/height2, kg/m2) <7.6 kg/m2 in men and <5.7 kg/m2 in women; for older adults (≥65 years), height-adjusted ASM < 7.0 kg/m2 in men and <5.7 kg/m2 in women. Low muscle strength thresholds were defined as <34.0 kg in middle-aged men, <20.0 kg in middle-aged women, <28.0 kg in older men, and <18.0 kg in older women.
2.5. Covariates
Based on previous relevant studies and clinical influencing factors, potential confounders were collected at baseline through standardized questionnaires, physical examinations, and laboratory tests, and were included in the subsequent adjustment models. Demographic characteristic confounders included sex, age, marital status, and educational level; marital status was dichotomized into married/cohabitating and other marital statuses, while educational level was categorized into three grades: below high school, high school, and college or above. Health-related confounders encompassed anthropometric indicators and lifestyle factors, as well as chronic comorbidities and relevant medication use: anthropometric indicators included body mass index (BMI, calculated as weight divided by height squared, kg/m2) and waist circumference (cm); lifestyle factors included sleep duration, smoking status, drinking status and physical activity, with smoking and drinking status each categorized into two groups (non-current smoking/drinking and current smoking/drinking), and physical activity categorized into never, mild, moderate, and vigorous; chronic comorbidities included hypertension, diabetes, and cancer; lipid, lowering medication use, antidiabetic drugs, and antihypertensive drugs were additionally included as a confounder, categorized as current use or non-current use. Physical activity was assessed via the CHARLS baseline questionnaire, which queries the frequency and duration of walking and moderate‑ and vigorous‑intensity activities. Based on the IPAQ scoring protocol, participants were categorized into never, mild, moderate, and vigorous levels. Hypertension was defined as systolic blood pressure/diastolic blood pressure ≥ 140 mmHg/90 mmHg, self-reported history of hypertension, or current use of antihypertensive medications. Diabetes was defined as fasting blood glucose ≥ 126 mg/dL, self-reported history of diabetes, or current treatment with insulin and/or oral hypoglycemic agents. Laboratory examination confounders included glycated hemoglobin (HbA1c, %), which was measured from fasting venous blood samples collected at baseline, with detection accuracy guaranteed by strict internal quality control and external proficiency testing. All chronic comorbidity diagnoses and medication use were confirmed by combining self-reported medical history and medication use to improve the accuracy of the data.
2.6. Statistical analysis
All statistical analyses were performed using R software (version 4.5.1), with a two- sided P < 0.05 considered statistically significant. Participants were divided into four quartile subgroups (Q1: the lowest, Q4: the highest) by baseline CHG index, and participants were categorized into two distinct groups based on CHG index trajectories using the K–means clustering method ('factoextra' package); the elbow method was used to confirm the optimal binary clustering result. Cox proportional hazards regression models were constructed to analyze the association of continuous and quartile-stratified baseline CHG index with new-onset sarcopenia risk, with the Schoenfeld residual test verifying the proportional hazards assumption. Three models were established: Model 1 (unadjusted), Model 2 (adjusted for sex, age, marital status, educational level, smoking status, drinking status, sleep duration, physical activity), and Model 3 (fully adjusted by adding BMI, waist circumference, hypertension, diabetes, cancer, lipid-lowering, antidiabetic, antihypertensive medication use, and HbA1c on the basis of Model 2). Results were presented as hazard ratios (HR) and 95% confidence intervals (CI), with trend P values calculated using the median of each baseline CHG index quartile. Restricted cubic spline (RCS) regression models were used to explore the potential nonlinear association between continuous baseline CHG index and new-onset sarcopenia risk. The optimal number of knots was determined by the smallest Akaike Information Criterion (AIC): AIC values for 3, 4, and 5 knots were 4212.7, 4215.3, and 4318.6, respectively, thus 3 knots were used. Nonlinearity was assessed by corresponding P values. For the RCS analysis of cumulative CHG index, the AIC values for 3, 4, and 5 knots were 4100.5, 4203.2, and 4405.9, respectively, and 3 knots were also adopted as the optimal number. Multivariable binary logistic regression analysis was applied to evaluate the association of K–means binary-clustered and quartile-stratified cumulative CHG index with new-onset sarcopenia risk, using the same three hierarchical adjustment models as the Cox regression analysis, with results presented as odds ratios (OR) and 95% CI and trend P values calculated by the median of each cumulative CHG index quartile. RCS regression models were also used to explore the potential nonlinear association between continuous cumulative CHG index and new-onset sarcopenia risk, with the optimal knot number and nonlinearity determined by the same method as above. Receiver operating characteristic (ROC) curves were plotted to quantify and compare the predictive performance of baseline/cumulative CHG index and baseline/cumulative TyG index for new-onset sarcopenia, with the area under the ROC curve (AUC) as the core indicator. Subgroup and interaction analyses were conducted across age groups (50–64 years, ≥65 years), sex, diabetes status, hypertension status and BMI levels (<24 kg/m2, 24–28 kg/m2, ≥28 kg/m2); interaction terms of CHG index (baseline/cumulative) and subgroup variables were added to the fully adjusted Cox and logistic models, with interaction P values calculated to identify effect modification. Sensitivity analyses were performed to verify result robustness: all regression and predictive analyses were repeated using the non-imputed original dataset and additionally using the 2019 AWGS sarcopenia diagnostic criteria.
3. Results
3.1. Participants characteristics at baseline
A total of 5580 participants were included in the final analysis. The baseline characteristics of the study population stratified by quartiles of the CHG index are presented in Table 1. The median age of participants was 60.0 (IQR: 55.0–66.0) years, and 2513 (45.0%) were male. Compared to participants in the lowest CHG quartile (Q1), those in the highest quartile (Q4) were older and had higher levels of FBG, CRP, and TyG (all P < 0.001). Participants in the higher CHG quartiles were more likely to be female and non-current smokers, and have a higher prevalence of diabetes (all P < 0.001).
Table 1.
Participants characteristics at baseline.
| Variables | CHG_quartile (n = 5580) |
P value | |||
|---|---|---|---|---|---|
| Q1 (n = 1395) | Q2 (n = 1395) | Q3 (n = 1395) | Q4 (n = 1395) | ||
| Age, years | 60.00 (55.00, 66.00) | 60.00 (56.00, 66.00) | 59.00 (55.00, 65.00) | 60.00 (56.00, 66.00) | 0.001 |
| BMI, kg/m2 | 21.82 (20.13, 23.87) | 22.14 (20.13, 24.43) | 22.41 (20.48, 24.65) | 22.35 (20.45, 24.48) | <0.001 |
| Waist Circumference, cm | 80.00 (75.20, 85.70) | 81.00 (76.00, 87.00) | 81.60 (76.40, 88.20) | 81.85 (76.43, 88.00) | <0.001 |
| Sleep duration, h | 6.00 (5.00, 8.00) | 6.00 (5.00, 8.00) | 6.00 (5.00, 8.00) | 6.00 (5.00, 8.00) | 0.037 |
| FBG, mg/dL | 92.16 (85.68, 98.64) | 99.27 (94.50, 104.94) | 103.95 (98.28, 110.83) | 117.90 (107.10, 136.80) | <0.001 |
| HbA1c, % | 5.40 (5.20, 5.60) | 5.20 (5.00, 5.40) | 5.00 (4.80, 5.20) | 4.90 (4.60, 5.20) | <0.001 |
| CRP, mg/dL | 0.90 (0.50, 1.89) | 0.86 (0.50, 1.73) | 0.96 (0.52, 2.07) | 0.96 (0.54, 1.91) | <0.001 |
| TyG | 8.36 (8.05, 8.70) | 8.48 (8.16, 8.83) | 8.61 (8.29, 8.99) | 8.82 (8.42, 9.33) | <0.001 |
| Sex, n (%) | <0.001 | ||||
| Male | 748 (53.6%) | 664 (47.6%) | 571 (40.9%) | 530 (38.0%) | |
| Female | 647 (46.4%) | 731 (52.4%) | 824 (59.1%) | 865 (62.0%) | |
| Marital Status, n (%) | 0.001 | ||||
| Married | 1254 (89.9%) | 1256 (90.0%) | 1236 (88.6%) | 1196 (85.7%) | |
| Others | 141 (10.1%) | 139 (10.0%) | 159 (11.4%) | 199 (14.3%) | |
| Education level, n (%) | <0.001 | ||||
| Below High School | 890 (63.8%) | 913 (65.4%) | 926 (66.4%) | 938 (67.2%) | |
| High School | 325 (23.3%) | 312 (22.4%) | 309 (22.2%) | 307 (22.0%) | |
| College or Above | 180 (12.9%) | 170 (12.2%) | 160 (11.5%) | 150 (10.8%) | |
| Smoking Status, n (%) | <0.001 | ||||
| Current Smoking | 490 (35.1%) | 431 (30.9%) | 406 (29.1%) | 385 (27.6%) | |
| Non-current Smoking | 905 (64.9%) | 964 (69.1%) | 989 (70.9%) | 1010 (72.4%) | |
| Drinking Status, n (%) | 0.017 | ||||
| Current Drinking | 423 (30.3%) | 453 (32.5%) | 466 (33.4%) | 501 (35.9%) | |
| Non-current Drinking | 972 (69.7%) | 942 (67.5%) | 929 (66.6%) | 894 (64.1%) | |
| Physical activity, n (%) | 0.012 | ||||
| Never | 291 (20.9%) | 278 (19.9%) | 265 (19.0%) | 252 (18.1%) | |
| Mild | 522 (37.4%) | 530 (38.0%) | 539 (38.6%) | 547 (39.2%) | |
| Moderate | 401 (28.7%) | 407 (29.2%) | 412 (29.5%) | 418 (30.0%) | |
| Vigorous | 181 (13.0%) | 180 (12.9%) | 179 (12.8%) | 178 (12.7%) | |
| Hypertension, n (%) | 0.119 | ||||
| Yes | 677 (48.5%) | 698 (50.0%) | 724 (51.9%) | 734 (52.6%) | |
| No | 718 (51.5%) | 697 (50.0%) | 671 (48.1%) | 661 (47.4%) | |
| Diabetes, n (%) | <0.001 | ||||
| Yes | 119 (8.5%) | 126 (9.0%) | 195 (14.0%) | 413 (29.6%) | |
| No | 1276 (91.5%) | 1269 (91.0%) | 1200 (86.0%) | 982 (70.4%) | |
| Cancer, n (%) | 0.743 | ||||
| Yes | 10 (0.7%) | 11 (0.8%) | 13 (0.9%) | 8 (0.6%) | |
| No | 1385 (99.3%) | 1384 (99.2%) | 1382 (99.1%) | 1387 (99.4%) | |
FBG, fasting blood glucose; CHG, cholesterol-HDL-glucose index; BMI, body mass index; HbA1c, glycated hemoglobin; CRP, C-reactive protein; TyG, triglyceride-glucose index.
Q1: <5.1; Q2: 5.1–5.7; Q3: 5.7–6.3; Q4: ≥6.3.
3.2. Association between baseline CHG and new-onset sarcopenia
During the 4-year follow-up period, 590 (10.6%) participants developed new-onset sarcopenia. The cumulative incidence of sarcopenia was significantly higher in participants with elevated baseline CHG levels (log-rank P = 0.039; Fig.2).
Fig. 2.
K-M plot of cumulative incidence of new-onset sarcopenia based on CHG levels. CHG, cholesterol-HDL-glucose index.
3.3. Association between baseline CHG and new-onset sarcopenia
Table 2 presents the results of the multivariable Cox regression analysis. In the fully adjusted model (Model 3), each one-unit increase in the continuous CHG index was associated with a 35% increased risk of new-onset sarcopenia (HR = 1.350, 95% CI: 1.122−1.627). When baseline CHG was assessed as quartiles, the risk increased progressively. Compared with the lowest quartile (Q1), participants in the highest quartile (Q4) had a significantly higher risk of sarcopenia (HR = 1.287, 95% CI: 1.022−1.619), with a significant trend across quartiles (P for trend = 0.021).
Table 2.
Cox analysis of association between baseline CHG and new-onset sarcopenia.
| Variable | Model 1 |
Model 2 |
Model 3 |
|||
|---|---|---|---|---|---|---|
| HR (95%CI) | P value | HR (95%CI) | P value | HR (95%CI) | P value | |
| CHG_continuous | 1.400 (1.164−1.683) | <0.001 | 1.353 (1.125−1.628) | <0.001 | 1.350 (1.122−1.627) | 0.002 |
| CHG_quartile | ||||||
| Q1 | Ref | Ref | Ref | |||
| Q2 | 1.046 (0.822−1.331) | 0.717 | 1.025 (0.805−1.304) | 0.833 | 1.016 (0.798−1.293) | 0.886 |
| Q3 | 1.131 (0.891−1.435) | 0.311 | 1.101 (0.867−1.397) | 0.423 | 1.110 (0.873−1.410) | 0.386 |
| Q4 | 1.355 (1.08−1.701) | 0.009 | 1.304 (1.039−1.637) | 0.022 | 1.287 (1.022−1.619) | 0.032 |
| P for trend | 0.006 | 0.015 | 0.021 | |||
CHG, cholesterol-HDL-glucose index; HR, hazard ratio; CI, confidence interval.
Q1: <5.1; Q2: 5.1–5.7; Q3: 5.7–6.3; Q4: ≥6.3.
Model 1 was unadjusted.
Model 2 was adjusted for sex, age, marital status, educational level, smoking status, drinking status, sleep duration,physical activity.
Model 3 was adjusted for sex, age, marital status, educational level, smoking status, drinking status, sleep duration, physical activity, BMI, waist circumference, hypertension, diabetes, cancer, lipid-lowering, antidiabetic, antihypertensive medication use, and HbA1c.
3.4. RCS analysis of association of baseline CHG with new-onset sarcopenia risk
The RCS regression demonstrated a linear dose-response relationship between baseline CHG levels and the risk of new-onset sarcopenia (P for nonlinearity > 0.05; P for overall < 0.001, Fig.3).
Fig. 3.
RCS analysis for CHG and new-onset sarcopenia risk. Adjusted for sex, age, marital status, educational level, smoking status, drinking status, sleep duration, physical activity, BMI, waist circumference, hypertension, diabetes, cancer, lipid-lowering, antidiabetic, antihypertensive medication use, and HbA1c.
3.5. Predictive capability of baseline CHG and TyG for predicting new-onset sarcopenia
The time-dependent ROC analysis showed that the baseline CHG index had a moderate predictive ability for new-onset sarcopenia, with an area under the curve (AUC) of 0.646 (95% CI: 0.560−0.696). This predictive performance was higher than that of the baseline TyG index (AUC = 0.599; Fig.4), with P of DeLong Test of 0.031.
Fig. 4.
Time-dependent ROC Curve for AUC of CHG and TyG. Adjusted for sex, age, marital status, educational level, smoking status, drinking status, sleep duration, physical activity, BMI, waist circumference, hypertension, diabetes, cancer, lipid-lowering, antidiabetic, antihypertensive medication use, and HbA1c.
3.6. Logistic regression analysis on the relationship between cumulative CHG and new-onset sarcopenia
We further explored the effect of long-term CHG burden on sarcopenia risk among 3287 participants with complete follow-up data, during which 209 (6.4%) developed new-onset sarcopenia. Using K-means clustering based on cumulative CHG changes (Fig. 5, Fig. S1, Table S2), participants were categorized into two distinct groups. Cluster 1 (n = 1783) was characterized by consistently low and stable CHG levels (from 4.88 ± 0.25–4.90 ± 0.26), while Cluster 2 (n = 1504) exhibited higher baseline CHG levels that declined over time (from 6.37 ± 0.27–5.74 ± 0.26). The baseline characteristics of these two clusters are detailed in Table S3. Baseline characteristics of participants included in and excluded from the cumulative CHG analysis were compared (Table S4). No significant between-group differences were observed in all major demographic, anthropometric, metabolic, and lifestyle variables (all P > 0.05), indicating that the included and excluded participants had well-balanced baseline profiles, and no obvious healthy survivor bias or informative dropout was present in the cumulative analysis.
Fig. 5.
The K-means clustering results. For the first cluster (n = 1783), CHG levels ranged from (4.88 ± 0.25) in 2011 to (4.90 ± 0.26); for the second cluster (n = 1504), its CHG range from (6.37 ± 0.27) in 2011 to (5.74 ± 0.26) in 2015.
Logistic regression analysis revealed that, compared with the stable low-level group (Cluster 1), the high-declining group (Cluster 2) had a significantly higher risk of new-onset sarcopenia (OR = 1.25, 95% CI: 1.02–1.45) after full adjustment (Table 3). When analyzed by cumulative CHG quartiles, participants in the highest quartile (Q4, ≥18.6) had a 1.52-fold increased risk of sarcopenia compared to those in the lowest quartile (Q1, <16.2) (95% CI: 1.14–2.06; P for trend = 0.004).
Table 3.
Logistic regression analyses of the association between cumulative CHG and new-onset sarcopenia.
| Characteristic | Model 1 |
Model 2 |
Model 3 |
|||
|---|---|---|---|---|---|---|
| OR (95%CI) | P value | OR (95%CI) | P value | OR (95%CI) | P value | |
| Cluster1 (n = 1783) | Ref | Ref | Ref | |||
| Cluster2 (n = 1504) | 1.27 (1.04, 1.45) | <0.001 | 1.26 (1.03, 1.44) | <0.001 | 1.25 (1.02, 1.45) | <0.001 |
| CumCHG_quartile | ||||||
| Q1 | Ref | Ref | Ref | |||
| Q2 | 1.28 (0.95, 1.73) | 0.110 | 1.31 (0.96, 1.77) | 0.076 | 1.31 (0.96, 1.78) | 0.077 |
| Q3 | 1.51 (1.13, 2.03) | 0.006 | 1.51 (1.13, 2.04) | 0.007 | 1.51 (1.13, 2.04) | 0.008 |
| Q4 | 1.53 (1.14, 2.05) | 0.005 | 1.52 (1.14, 2.05) | 0.004 | 1.52 (1.14, 2.06) | 0.005 |
| P for trend | 0.003 | 0.003 | 0.004 | |||
| Per SD increase | 1.20 (1.09, 1.33) | <0.001 | 1.19 (1.09, 1.32) | <0.001 | 1.19 (1.09, 1.32) | <0.001 |
CHG, cholesterol-HDL-glucose index; OR, odds ratio; CI, confidence interval; SD, standard deviation.
Q1: <16.2; Q2: 16.2–17.4; Q3: 17.4–18.6; Q4: ≥18.6.
Model 1 was unadjusted.
Model 2 was adjusted for sex, age, marital status, educational level, smoking status, drinking status, sleep duration, physical activity.
Model 3 was adjusted for sex, age, marital status, educational level, smoking status, drinking status, sleep duration, physical activity, BMI, waist circumference, hypertension, diabetes, cancer, lipid-lowering, antidiabetic, antihypertensive medication use, and HbA1c.
3.7. RCS analysis of association of cumulative CHG with new-onset sarcopenia risk
RCS analysis confirmed a linear association between cumulative CHG and sarcopenia risk (Fig. S2).
3.8. Predictive capability of cumulative CHG and TyG for predicting new-onset sarcopenia
The predictive performance of the cumulative CHG index (AUC = 0.680, 95% CI: 0.596−0.753) was higher than that of the cumulative TyG index (Fig. S3), with P of DeLong Test of 0.026.
3.9. Subgroup and sensitivity analyses
Subgroup analyses showed that the positive associations of both baseline and cumulative CHG with new-onset sarcopenia were generally consistent across strata of age, sex, hypertension, and BMI (all P for interaction > 0.05; Fig.6). Association between baseline CHG and sarcopenia risk was statistically significant among participants younger than 65 years (HR = 1.182, 95%CI: 1.066–1.311), those with BMI < 24 kg/m2 (HR = 1.093, 95%CI: 1.006–1.187), individuals with and without diabetes (HR = 1.189 and 1.409, respectively), those with and without hypertension (HR = 1.121 and 1.214, respectively), and males (HR = 1.219, 95%CI: 1.083–1.372); however, the association did not reach statistical significance in participants aged 65 years or above, those with BMI 24–28 kg/m2 or ≥28 kg/m2, or females, with all interaction P values exceeding 0.05, indicating no significant effect modification. For cumulative CHG, the association with sarcopenia risk was statistically significant in participants aged < 65 years (OR = 1.341, 95%CI: 1.129–1.579), those with BMI < 24 kg/m2 (OR = 1.170, 95%CI: 1.025–1.336), individuals with or without diabetes, those with hypertension (OR = 1.342, 95%CI: 1.121–1.605), and males (OR = 1.245, 95%CI: 1.027–1.517). The association was not significant in participants aged ≥65 years, those with BMI 24–28 or ≥28 kg/m2, those without hypertension, or females. All interaction P values were >0.05, indicating no significant effect modification across subgroups. The robustness of our findings was confirmed in sensitivity analyses. The results remained consistent after excluding participants with missing covariate data and when sarcopenia was redefined using the 2019 AWGS diagnostic criteria (Tables S5, S6).
Fig. 6.
Forest plot of subgroup analyses of association of baseline and cumulative CHG with new-onset sarcopenia risk. A: subgroup analysis of association of baseline CHG with new-onset sarcopenia risk; B: subgroup analysis of association of cumulative CHG with new-onset sarcopenia risk.
4. Discussion
In this study, we examined the relationship between the CHG index and new-onset sarcopenia in middle-aged and older Chinese adults. The key findings are threefold. First, higher baseline CHG levels were significantly associated with increased risk of developing sarcopenia over 4 years of follow-up, independent of established risk factors. Second, participants with elevated but declining CHG trajectories still faced greater sarcopenia risk compared to those with consistently low levels. Third, the CHG index demonstrated modestly better predictive performance than the traditional TyG index for identifying individuals at risk of sarcopenia.
Our results align with current understanding of sarcopenia as a metabolic disorder. IR is widely recognized as a core driver of muscle loss. When insulin signaling is impaired, skeletal muscle cells fail to take up glucose efficiently, and the balance between protein synthesis and degradation shifts toward net loss [25,26]. The CHG index captures this metabolic disturbance by integrating three clinically accessible parameters: TC, HDL-C, and FBG. Beyond insulin resistance, lipid abnormalities may directly contribute to muscle deterioration. Elevated TC and reduced HDL-C, both reflected in the CHG calculation, are associated with chronic low-grade inflammation and oxidative stress [27,28]. Pro-inflammatory cytokines such as TNF-α and IL-6 activate ubiquitin-proteasome pathways in muscle tissue, accelerating protein breakdown [29,30]. The CHG index may therefore serve as a composite marker that captures multiple interconnected pathways—glucose dysregulation, lipid abnormalities, and their downstream inflammatory consequences—that collectively promote sarcopenia development.
We observed that CHG outperformed TyG in predicting sarcopenia. This difference likely stems from the additional information CHG provides about cholesterol metabolism. While TyG reflects triglyceride and glucose homeostasis, CHG incorporates TC and HDL-C, both of which are relevant to muscle cell membrane integrity and mitochondrial function [21,31]. Given that mitochondrial dysfunction is increasingly recognized as a key feature of age-related muscle decline, CHG may more comprehensively capture the metabolic health of muscle tissue.
A particularly noteworthy finding was the elevated risk in Cluster 2—participants whose CHG levels started high but declined over time. We believe this seemingly counterintuitive observation warrants careful interpretation. Several explanations may account for this pattern. First, the concept of metabolic memory likely applies here. Prolonged exposure to adverse metabolic conditions during the period when CHG was elevated may have caused lasting damage to muscle tissue, such as mitochondrial DNA mutations, epigenetic modifications favoring inflammation, or depletion of satellite cell pools [[32], [33], [34]]. These changes may not be fully reversible even if metabolic parameters subsequently improve. Second, despite the declining trend, the cumulative metabolic burden over the entire follow-up period remained substantially higher in Cluster 2 than in Cluster 1. The persistently elevated risk may reflect this greater total exposure rather than the trajectory direction per se. Third, we cannot exclude the possibility of survivorship bias or informative censoring. Some high-risk individuals with extreme metabolic disturbances may have been lost to follow-up due to death or serious illness, leaving behind a healthier subset whose CHG levels appeared to improve but whose underlying susceptibility to sarcopenia remained elevated.
To our knowledge, no published studies have directly examined CHG in relation to sarcopenia risk. However, our findings are consistent with research using related metabolic indices. Studies from the United States and Korea have reported positive associations between TyG index and low muscle mass [35,36]. Our study extends this line of evidence by introducing a potentially more comprehensive indicator and by examining cumulative effects, which better reflect long-term metabolic burden than single-time-point measurements.
Subgroup analyses showed consistent associations across different age groups, sex, hypertension status, and BMI categories, suggesting that CHG’s utility as a risk marker is broadly applicable in middle-aged and older populations. The consistent lack of significant interactions across most subgroups suggests that CHG is a robust risk marker applicable to diverse populations. For baseline CHG, the significant associations observed in younger participants and those with lower BMI may reflect earlier metabolic disturbances that precede overt clinical conditions. For cumulative CHG, the stronger associations in older participants and those with hypertension likely reflect the cumulative damage of long-term metabolic stress on muscle health. The borderline interaction for age in cumulative CHG analysis (P = 0.064) suggests that the effect of long-term CHG exposure may be more pronounced in older adults, possibly due to longer exposure time and age-related vulnerability to metabolic stress. Sensitivity analyses using non-imputed data and alternative sarcopenia definitions confirmed the robustness of our primary results.
Several limitations should be acknowledged when interpreting our results. First, the relatively short follow-up period of 4 years may not fully capture the slow, progressive nature of sarcopenia development. Second, sarcopenia diagnosis in CHARLS relies on BIA rather than DXA, which is considered the gold standard, potentially introducing measurement variability. Third, although we adjusted for physical activity level, the categorization was based on self-report rather than objective accelerometry, which may not fully capture activity intensity or individual variability. Furthermore, residual confounding from unmeasured factors such as detailed dietary intake cannot be entirely ruled out. Fourth, the modest AUC values indicate that CHG alone has limited utility as a standalone screening tool; its value likely lies in combination with other clinical risk factors for risk stratification. Fifth, our sample is restricted to Chinese adults aged 50 and above, and extrapolation to other populations requires caution. Sixth, attrition analysis revealed that participants excluded from the mean CHG analysis had marginally higher baseline CHG levels, suggesting a potential for healthy survivor bias that may have attenuated the observed associations. Additionally, we were unable to adjust for certain chronic conditions—such as chronic obstructive pulmonary disease, chronic kidney disease, and liver disease—because their self-reported physician-diagnosed prevalence in the CHARLS baseline was low (approximately 3–4%) and subject to misclassification. Nevertheless, the omission of these comorbidities may represent a source of residual confounding.
5. Conclusion
This study demonstrates that elevated baseline and cumulative CHG indices are independently associated with increased risk of new-onset sarcopenia in middle-aged and older Chinese adults. The CHG index shows modestly better predictive performance than the traditional TyG index. Notably, individuals with high baseline CHG remain at elevated risk even if their levels subsequently decline, underscoring the importance of early and sustained metabolic management.
CRediT authorship contribution statement
Yan Chen: collecting information, writing – original draft, writing – review and editing; Lingli Gao: visualization; Jing Gao and Xiaodong Feng: supervision, proof–red the manuscript.
Ethics declaration
Not applicable.
Clinical trial number
Not applicable.
Declaration of Generative AI and AI-assisted technologies in the writing process
During the preparation of this work, the authors did not use any generative AI or AI-assisted technologies in the writing process or for the creation of figures, images, or artwork.
Funding
This work was supported by the National Key Research and Development Program of China [grant number 2023YFC3503705].
Data availability
The original contributions presented in the study are included in the article. Further inquiries can be directed to the corresponding author.
Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgments
Thanks to the editors and reviewers for their hard work and important comments.
Footnotes
Supplementary material related to this article can be found, in the online version, at doi:https://doi.org/10.1016/j.jnha.2026.100889.
Appendix A. Supplementary data
The following is Supplementary data to this article:
References
- 1.Chen L.-K., Hsiao F.-Y., Akishita M., Assantachai P., Lee W.-J., Lim W.S., et al. A focus shift from sarcopenia to muscle health in the Asian Working Group for Sarcopenia 2025 Consensus Update. Nat Aging. 2025;5:2164–2175. doi: 10.1038/s43587-025-01004-y. [DOI] [PubMed] [Google Scholar]
- 2.Boonpor J., Pell J.P., Ho F.K., Celis-Morales C., Gray S.R. In people with type 2 diabetes, sarcopenia is associated with the incidence of cardiovascular disease: a prospective cohort study from the UK Biobank. Diabetes Obes Metab. 2024;26:524–531. doi: 10.1111/dom.15338. [DOI] [PubMed] [Google Scholar]
- 3.Wathanavasin W., Banjongjit A., Avihingsanon Y., Praditpornsilpa K., Tungsanga K., Eiam-Ong S., et al. Prevalence of sarcopenia and its impact on cardiovascular events and mortality among dialysis patients: a systematic review and meta-analysis. Nutrients. 2022;14:4077. doi: 10.3390/nu14194077. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Purnamasari D., Tetrasiwi E.N., Kartiko G.J., Astrella C., Husam K., Laksmi P.W. Sarcopenia and chronic complications of type 2 diabetes mellitus. Rev Diabet Stud. 2022;18:157–165. doi: 10.1900/RDS.2022.18.157. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Ibrahim K., Mullee M.A., Cox N., Russell C., Baxter M., Tilley S., et al. The feasibility and acceptability of assessing and managing sarcopenia and frailty among older people with upper limb fracture. Age Ageing. 2021;51 doi: 10.1093/ageing/afab252. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.von Haehling S., Morley J.E., Anker S.D. An overview of sarcopenia: facts and numbers on prevalence and clinical impact. J Cachexia Sarcopenia Muscle. 2010;1:129–133. doi: 10.1007/s13539-010-0014-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Petermann-Rocha F., Balntzi V., Gray S.R., Lara J., Ho F.K., Pell J.P., et al. Global prevalence of sarcopenia and severe sarcopenia: a systematic review and meta-analysis. J Cachexia Sarcopenia Muscle. 2022;13:86–99. doi: 10.1002/jcsm.12783. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Liu S., Zhang L., Li S. Advances in nutritional supplementation for sarcopenia management. Front Nutr. 2023;10 doi: 10.3389/fnut.2023.1189522. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Alghannam A.F., Alharbi D.S., Al-Hazzaa H.M. Sarcopenia of ageing: does a healthier lifestyle matter in reversing the trajectory? A brief narrative review and a call for action in Saudi Arabia. Saudi J Med Med Sci. 2024;12:10–16. doi: 10.4103/sjmms.sjmms_54_23. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Bilski J., Pierzchalski P., Szczepanik M., Bonior J., Zoladz J.A. Multifactorial mechanism of sarcopenia and sarcopenic obesity. Role of physical exercise, Microbiota and myokines. Cells. 2022;11 doi: 10.3390/cells11010160. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Liu Z.J., Zhu C.F. Causal relationship between insulin resistance and sarcopenia. Diabetol Metab Syndr. 2023;15:46. doi: 10.1186/s13098-023-01022-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Tessari P. Stepwise discovery of insulin effects on amino acid and protein metabolism. Nutrients. 2024;16:119. doi: 10.3390/nu16010119. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Huh J.Y., Cho J., Bak H.R. Association between high-density lipoprotein cholesterol levels and sarcopenia in elderly koreans. J Clin Med. 2025:15. doi: 10.3390/jcm15010183. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Yao J., Xu J., Sun Y., Ji Y., Lou H., Wu X., et al. Association of triglyceride-glucose index and atherogenic index of plasma with sarcopenia across different glucose metabolism states: the mediating role of oxidative stress and chronic inflammation. Eur J Med Res. 2025;31:27. doi: 10.1186/s40001-025-03550-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Mustakin Triglyceride - glucose (TyG) index as a screening tool in community settings for early detection of type 2 diabetes risk: a systematic review. Ejifcc. 2026;37:149–158. [PMC free article] [PubMed] [Google Scholar]
- 16.Chen Y., Liu C., Hu M. Association between triglyceride-glucose index and sarcopenia in China: a nationally representative cohort study. Exp Gerontol. 2024;190 doi: 10.1016/j.exger.2024.112419. [DOI] [PubMed] [Google Scholar]
- 17.Zeng B., Xu Z., Xu Q., Lin C. Joint association of triglyceride-glucose index and body roundness index for sarcopenia risk: a national cohort study. Clin Nutr ESPEN. 2025;72 doi: 10.1016/j.clnesp.2025.11.161. [DOI] [PubMed] [Google Scholar]
- 18.Zhang Z., Chen X., Jiang N. The triglyceride glucose related index is an indicator of sarcopenia. Sci Rep. 2024;14 doi: 10.1038/s41598-024-75873-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Chen Y., Huang T., Wang J., Guo C., Gao J., Feng X. Association between C-reactive protein-triglyceride glucose index and risk of sarcopenia in older adults: findings from CHARLS and NHANES. Exp Gerontol. 2026;213 doi: 10.1016/j.exger.2025.113003. [DOI] [PubMed] [Google Scholar]
- 20.Li T., Li Y., Xiang N., Hu Y., Wang F., Meng C., et al. Cholesterol-HDL-glucose index and type 2 diabetes risk in middle-aged and older adults: evidence from a nationwide Chinese cohort. Medicine (Baltimore) 2025;104 doi: 10.1097/MD.0000000000044830. [DOI] [PubMed] [Google Scholar]
- 21.Mo D., Zhang P., Zhang M., Dai H., Guan J. Cholesterol, high-density lipoprotein, and glucose index versus triglyceride-glucose index in predicting cardiovascular disease risk: a cohort study. Cardiovasc Diabetol. 2025;24:116. doi: 10.1186/s12933-025-02675-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Zhang Y., Zhu Y., He Y., Yan Y., Zhao Q., Du Y. Association of cumulative cholesterol-HDL-glucose index with blood pressure changes and risk of new-onset hypertension in middle-aged and older adults: a cohort study. Cardiovasc Diabetol. 2026;25:49. doi: 10.1186/s12933-026-03081-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Guo Z., Xiong Z., Zhang W., Chen G., Xie M., Zhou Z., et al. Association between cholesterol, high-density lipoprotein, and glucose index and risks of cardiovascular and all-cause mortality in patients with calcific aortic valve stenosis: the ARISTOTLE cohort study. Cardiovasc Diabetol. 2025;24:393. doi: 10.1186/s12933-025-02906-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Cui C., Zhou B., Gao L., Wang D., Rowe A.C., Nascimento B.S., et al. Physical health and cognitive ability factors in predicting retirement adjustment based on machine learning approach: results from the China health and retirement longitudinal study. Front Psychol. 2025;16 doi: 10.3389/fpsyg.2025.1601723. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Whytock K.L., Goodpaster B.H. Unraveling skeletal muscle insulin resistance: molecular mechanisms and the restorative role of exercise. Circ Res. 2025;137:184–204. doi: 10.1161/CIRCRESAHA.125.325532. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Fujita S., Rasmussen B.B., Cadenas J.G., Grady J.J., Volpi E. Effect of insulin on human skeletal muscle protein synthesis is modulated by insulin-induced changes in muscle blood flow and amino acid availability. Am J Physiol Endocrinol Metab. 2006;291:E745–754. doi: 10.1152/ajpendo.00271.2005. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Ouyang W., Chiang H.-H., Hsu W.-L., Tsai M.-H., Huang C.-Y., Remaley A.T., et al. Dysfunctional high-density lipoprotein: an updated review. Front Cardiovasc Med. 2025;12 doi: 10.3389/fcvm.2025.1713387. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Vekic J., Stromsnes K., Mazzalai S., Zeljkovic A., Rizzo M., Gambini J. Oxidative stress, atherogenic dyslipidemia, and cardiovascular risk. Biomedicines. 2023;11:2897. doi: 10.3390/biomedicines11112897. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Huang Y., Wang C., Cui H., Sun G., Qi X., Yao X. Mitochondrial dysfunction in age-related sarcopenia: mechanistic insights, diagnostic advances, and therapeutic prospects. Front Cell Dev Biol. 2025;13 doi: 10.3389/fcell.2025.1590524. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Ma K., Huang F., Qiao R., Miao L. Pathogenesis of sarcopenia in chronic obstructive pulmonary disease. Front Physiol. 2022;13 doi: 10.3389/fphys.2022.850964. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Zhou H., Shi Y., Zhou X. Comparison of nine insulin resistance surrogates for predicting cardiovascular disease: a cohort study. Diabetol Metab Syndr. 2025;17:365. doi: 10.1186/s13098-025-01933-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Lehti M., Donelan E., Abplanalp W., Al-Massadi O., Habegger K.M., Weber J., et al. High-density lipoprotein maintains skeletal muscle function by modulating cellular respiration in mice. Circulation. 2013;128:2364–2371. doi: 10.1161/CIRCULATIONAHA.113.001551. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Wood N., Straw S., Scalabrin M., Roberts L.D., Witte K.K., Bowen T.S. Skeletal muscle atrophy in heart failure with diabetes: from molecular mechanisms to clinical evidence. ESC Heart Fail. 2021;8:3–15. doi: 10.1002/ehf2.13121. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Sfyri P., Matsakas A. Crossroads between peripheral atherosclerosis, western-type diet and skeletal muscle pathophysiology: emphasis on apolipoprotein E deficiency and peripheral arterial disease. J Biomed Sci. 2017;24:42. doi: 10.1186/s12929-017-0346-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Wei X., Liu D. Association of triglyceride-glucose index with sarcopenia: NHANES 2011–2014. Front Endocrinol (Lausanne) 2024;15 doi: 10.3389/fendo.2024.1452664. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Ahn S.H., Lee J.H., Lee J.W. Inverse association between triglyceride glucose index and muscle mass in Korean adults: 2008–2011 KNHANES. Lipids Health Dis. 2020;19:243. doi: 10.1186/s12944-020-01414-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The original contributions presented in the study are included in the article. Further inquiries can be directed to the corresponding author.