Abstract
Background
Insulin resistance (IR) and arthritis are strongly associated, and the triglyceride–glucose (TyG) index combinations with obesity indicators [including TyG–BMI (glucose triglyceride–body mass index), TyG–WC (glucose triglyceride–waist circumference), and TyG–WHtR (glucose triglyceride–waist height ratio)] has recently been recognized as a more effective indicator for assessing IR. However, there is a lack of research on its association with arthritis, and it is also important to assess in different populations.
Methods
The analysis utilized data from the China Health and Retirement Longitudinal Study (CHARLS) and the National Health and Nutrition Examination Survey (NHANES). Arthritis diagnosis relied on self-reporting confirmed by physicians. The association of TyG–BMI, TyG–WC, and TyG–WHtR with arthritis was analyzed through weighted logistic regression models, and exploring nonlinear effects with restricted cubic spline (RCS) models. Secondary and sensitivity analyses included receiver operating characteristic curve (ROC) analysis, comparisons of z score-related odds ratios, subgroup analyses, and multiple imputation.
Results
The study involved 6141 CHARLS participants and 17,091 NHANES participants. Adjusting for confounding variables, TyG–BMI and TyG–WHtR demonstrate a positive correlation with arthritis prevalence in both CHARLS (TyG–BMI: OR = 1.02, 95% CI 1.00–1.04; TyG–WHtR: OR = 1.13, 95% CI 1.03–1.24) and NHANES (TyG–BMI: OR = 1.07, 95% CI 1.06–1.08; TyG–WHtR: OR = 1.50, 95% CI 1.40–1.60). RCS regression analysis demonstrated a significant nonlinear association. ROC analysis indicated that TyG–BMI and TyG–WHtR were superior to TyG for the diagnosis of arthritis in both CHARLS and NHANES.
Conclusions
TyG–BMI and TyG–WHtR demonstrate a positive correlation with arthritis prevalence in both Chinese and the U.S. populations, displaying superior diagnostic relevance compared to TyG.
Supplementary Information
The online version contains supplementary material available at 10.1186/s40001-024-01992-4.
Keywords: TyG, TyG–BMI, TyG–WC, TyG–WHtR, Arthritis, NHANES, CHARLS
Background
Arthritis is an inflammatory joint disease that presents as an acute or chronic conditions, characterized primarily by joint pain and stiffness. These symptoms typically worsen with age, accompanied by common manifestations such as joint stiffness, cartilage degradation, and pathological features like mononuclear cell infiltration, inflammation, synovial swelling, and mass formation [1, 2]. Osteoarthritis (OA) and rheumatoid arthritis (RA) stand out as the two most prevalent types of arthritis globally. In 2017, the age-standardized prevalence of RA showed a 7.4% global increase. Meanwhile, the prevalence of OA surpassed 520 million cases in 2019, emerging as one of the most prevalent disabling health conditions [3, 4]. However, effective medications to halt arthritis progression are still lacking, and the high prevalence of arthritis and its associated complications, such as joint injuries and disability, pose a growing burden on public health systems worldwide.
Common risk factors for arthritis include genetic predisposition, environment, family history, and behavioral factors such as smoking [5, 6]. The prevalent strategy in arthritis prevention involves mitigating exposure to these risk factors. Additionally, arthritis is associated with a high prevalence of insulin resistance (IR), a recognized risk factor for diabetes mellitus (DM) [7]. Moreover, arthritis can exacerbate the development of IR and DM [8]. Therefore, identifying indicators capable of assessing IR is crucial for the early detection and prognosis of arthritis.
The triglyceride–glucose (TyG) index is calculated by using the equation Ln [fasting triglyceride (mg/dL) × fasting glucose (mg/dL)/2], the utility of which as an index for the assessment of IR has gained widespread attention and recognition in recent years [9]. Compared with TyG, the assessment of IR is more accurate when TyG is combined with obesity indicators including TyG–BMI (glucose triglyceride–body mass index), TyG–WC (glucose triglyceride–waist circumference), and TyG–WHtR (glucose triglyceride–waist height ratio) [9, 10]. Although previous epidemiological investigations of the association between TyG and arthritis exist in the literature, and have mainly focused on the US population [11]. More importantly, there have been no studies examining the association between TyG index combinations with obesity indicators and arthritis. Considering the imperative for more dependable evidence concerning the association between IR and arthritis, it is essential to evaluate the association between TyG index combinations with obesity indicators and arthritis, preferably in representative and diverse populations to make the results more generalizable and representative.
Consequently, the purpose of this study was to investigate the association between TyG index combinations with obesity indicators and arthritis by using data from two national population-based studies: the China Health and Retirement Longitudinal Study (CHARLS) and the National Health and Nutrition Examination Survey (NHANES), with large representative samples in the China and the United States.
Methods
Study design and population
The CHARLS and NHANES are nationally representative prospective studies conducted in the China and United States, respectively. They employ complex, multistage probability sampling design in both data collection and research methodology to guarantee the national representativeness of the survey sample.
CHARLS, established in 2011, is a comprehensive biennial survey targeting Chinese residents aged 45 and older. It spans 450 villages and communities across 28 provinces in China, including autonomous regions and municipalities. CHARLS collects nationally representative data on various topics, including demographics, family, health status, cognition, healthcare, work, retirement, finances, and housing, and serves as a vital resource for understanding the socioeconomic landscape and health dynamics among older adults in China [12].
NHANES, conducted jointly by the Centers for Disease Control and Prevention (CDC) and the National Center for Health Statistics (NCHS), comprises a sequence of surveys aimed at evaluating the health and nutritional status of a representative subset of the noninstitutionalized population in the United States. These surveys collect comprehensive data on various health-related subjects, encompassing demographics, socioeconomic factors, dietary patterns, and health-related information, which are obtained through household interviews. Subsequently, blood samples are acquired at a mobile examination center (MEC).
As shown in Fig. 1, the study included 17,708 Chinese participants with blood samples collected from 2011–2012 (wave 1) and included 55,081 U.S. adults aged 20 years and older from 10 consecutive cycles of NHANES data covering the years 1999 to 2018. Exclusion criteria were applied, leading to the removal of individuals with missing data on arthritis, TyG index combinations with obesity indicators (including TyG–BMI, TyG–WC, and TyG–WHtR), demographics data [including age, sex, race, marital status, education level, household income or poverty income ratio (PIR), drinking, smoke, hypertension, total cholesterol, and high-density lipoprotein (HDL)]. Finally, our complete analysis encompassed 6141 CHARLS participants and 17,091 NHANES participants.
Fig. 1.
Flow diagram of the screening of the Chinese and the U.S. participants in CHARLS 2011–2012 and NHANES 1999–2018
Assessment of TyG, TyG–BMI, TyG–WC, and TyG–WHtR
The TyG index was formulated as ln [fasting triglyceride (mg/dL) × fasting glucose (mg/dL)/2]. BMI was computed by weight (kg)/height squared (m2), while WC was the waist circumference (cm), and WHtR was determined by waist circumference (cm)/height (cm). The outcomes were calculated by the following formula: TyG–WC = TyG × waist circumference; TyG–WHtR = TyG × WHtR; TyG–BMI = TyG × BMI [13].
Diagnosis of arthritis
Self-reported arthritis diagnosed by physicians is commonly utilized in epidemiological studies, with data regarding the outcome obtained through a questionnaire [14]. Participants who answered affirmatively to the question “Have you been diagnosed with arthritis or rheumatism by the doctor?” (CHARLS) or “Has a doctor or other health professional ever told you that you had arthritis?” (NHANES) were defined as having arthritis. Additionally, in NHANES, participants with arthritis were further queried: “Which type of arthritis was it?” those selecting “Rheumatoid arthritis” were diagnosed with RA, while those selecting “Osteoarthritis” (2005–2010) or “Osteoarthritis or degenerative arthritis” (2011–2018) were diagnosed with OA.
Covariates
Informed by published research and clinical judgment, several variables were identified as potential confounders, including age, sex, race/ethnicity, marital status, education level, household income/PIR, drinking status, smoke, hypertension, total cholesterol, and HDL [11, 13]. For consistency between the CHARLS and NHANES, age was analyzed as a continuous variable in logistic regression, while in subgroup analyses it was categorized as < 45, 45–64, and ≥ 65 years. Sex was categorized as female and male. Marital status was grouped as married or other. Education level was divided into three categories: less than high school, high school or equivalent, and above high school. Self-reported drinking and smoking status was categorized as never, former, and now. The diagnosis of hypertension occurred when the systolic reading ≥ 140 mmHg, or when the diastolic reading ≥ 90 mmHg.
Statistical analysis
Complex sampling design and weights were considered in our analyses. Participant characteristics were calculated based on the presence or absence of arthritis, with categorical variables was conveyed in terms of numerical counts and percentage frequencies (%), and continuous variables as means and standard error (SE). The Chi-squared test with Rao and Scott's second-order correction was utilized to analyze categorical data. Meanwhile, the Wilcoxon rank-sum test for complex survey samples was employed to assess divergent disparities in continuous variables, specifically for non-normal distributions, while the t-test was utilized for normally distributed data.
The multivariable weighted logistic regression models were employed to examine the association between TyG index combinations with obesity indicators and arthritis, and participants in NHANES were further explored the association between TyG index combinations with obesity indicators and arthritis types (RA, OA). Results were expressed as odds ratios (OR) and 95% confidence intervals (CI). Model 1 was the crude model, did not account for any covariates. Model 2 was adjusted for age, sex, race/ethnicity, marital status, education level, household income/PIR, drinking status, smoke, hypertension, total cholesterol, and HDL. The linear trends among TyG, TyG–BMI, TyG–WC, and TyG–WHtR quartiles were assessed by considering each quartile as a continuous variable. Additionally, possible nonlinear effects were modeled using restricted cubic spline (RCS) models with 3 knots at 10%, 50%, and 90%.
Secondary analyses involved conducting multivariable weighted logistic regression and RCS to explore the association of TyG and arthritis. The diagnostic efficacy of TyG, TyG–BMI, TyG–WC, and TyG–WHtR for arthritis was evaluated using receiver operating characteristic curve (ROC) analysis, with the area under the curve (AUC) calculated to ascertain sensitivity and specificity in predicting arthritis. Consistent with prior research, we conducted a comparison of z score-related odds ratios for arthritis between TyG–BMI, TyG–WC, and TyG–WHtR, as compared to TyG, utilizing a 2-independent-samples t-test based on bootstrapped estimates (n = 1000) [15]. In sensitivity analyses, subgroup analyses were conducted based on age, sex, race/ethnicity, marital status, education level, PIR/household income, drinking status, smoke, hypertension, and CRP (divided into < 8.2 mg/L and ≥ 8.2 mg/L groups). Furthermore, with the objective of reducing the impact of missing variables on the results, we explored missing values through multiple imputation with 5 replications and a chained equation approach.
All analyses were conducted using R (version 4.2.3) with the survey package (version 4.2-1), and Free Software Foundation statistics software (version 1.8). Statistical significance was determined based on 2-sided P < 0.05.
Results
Characteristics of the participants
Table 1 presents the demographic characteristics of a sample comprising 247.38 million individuals in China, with a weighted mean age of 60.91 years (SE, 0.31), and 74.95 million individuals in the United States, with a weighted mean age of 46.10 years (SE, 0.25). Within this sample, 99.63 million individuals in China and 14.61 million individuals in the U.S. were identified as having arthritis. Notably, individuals with arthritis in both the China and the U.S. exhibited elevated levels of total cholesterol, HDL, TyG, TyG BMI, TyG–WC, and TyG–WHtR.
Table 1.
Characteristics of the Chinese and the U.S. participants in CHARLS 2011–2012 and NHANES 1999–2018
| Characteristics | CHARLS | NHANES | ||||||
|---|---|---|---|---|---|---|---|---|
| Total | Without arthritis | With arthritis | P value | Total | Without arthritis | With arthritis | P value | |
| Weighted population, n [in millions] | 247.38 | 158.75 | 88.63 | 74.95 | 60.34 | 14.61 | ||
| Age, mean (SE), years | 60.91 (0.31) | 60.70 (0.38) | 61.28 (0.37) | 0.21 | 46.10 (0.25) | 43.02 (0.24) | 58.83 (0.28) | < 0.001 |
| < 45 | / | / | / | 36.47 (48.66) | 34.25 (56.76) | 2.23 (15.23) | ||
| 45–64 | 155.41 (64.5) | 100.61 (65.33) | 54.79 (63.02) | 26.59 (35.48) | 19.52 (32.35) | 7.07 (48.39) | ||
| ≥ 65 | 85.54 (35.50) | 53.39 (34.67) | 32.15 (36.98) | 11.89 (15.86) | 6.57 (10.89) | 5.31 (36.38) | ||
| Sex, n [in millions] (%) | < 0.001 | < 0.001 | ||||||
| Female | 130.93 (52.93) | 79.00 (49.76) | 51.93 (58.59) | 37.65 (50.23) | 28.91 (47.92) | 8.73 (59.79) | ||
| Male | 116.45 (47.07) | 79.76 (50.24) | 36.70 (41.41) | 37.30 (49.77) | 31.43 (52.08) | 5.87 (40.21) | ||
| Race/ethnicity, n [in millions] (%) | / | < 0.001 | ||||||
| Mexican American | / | / | / | 6.06 (8.09) | 5.55 (9.20) | 0.51 (3.50) | ||
| Non-Hispanic Black | / | / | / | 7.41 (9.88) | 6.22 (10.32) | 1.18 (8.10) | ||
| Non-Hispanic White | / | / | / | 52.86 (70.52) | 41.13 (68.17) | 11.72 (80.23) | ||
| Other Hispanic | / | / | / | 3.88 (5.17) | 3.45 (5.71) | 0.43 (2.94) | ||
| Other | / | / | / | 4.75 (6.34) | 3.99 (6.6) | 0.76 (5.23) | ||
| Marital status, n [in millions] (%) | 0.88 | < 0.001 | ||||||
| Married | 200.64 (81.11) | 128.89 (81.19) | 71.75 (80.96) | 43.00 (57.37) | 33.81 (56.03) | 9.19 (62.92) | ||
| Other | 46.74 (18.89) | 29.87 (18.81) | 16.88 (19.04) | 31.95 (42.63) | 26.53 (43.97) | 5.42 (37.08) | ||
| Education level, n [in millions] (%) | 0.079 | < 0.001 | ||||||
| Less than high school | 226.44 (91.53) | 143.52 (90.41) | 82.91 (93.55) | 11.67 (15.57) | 9.06 (15.01) | 2.61 (17.89) | ||
| High school or equivalent | 18.38 (7.43) | 13.38 (8.43) | 4.99 (5.63) | 17.74 (23.67) | 14.16 (23.46) | 3.58 (24.53) | ||
| Above high school | 2.57 (1.04) | 1.85 (1.16) | 0.72 (0.81) | 45.54 (60.76) | 37.13 (61.53) | 8.41 (57.58) | ||
| Drinking status, n [in millions] (%) | 0.02 | < 0.001 | ||||||
| Never | 152.41 (61.61) | 96.31 (60.66) | 56.10 (63.30) | 7.93 (10.58) | 6.29 (10.43) | 1.64 (11.22) | ||
| Former | 21.42 (8.66) | 12.99 (8.18) | 8.43 (9.51) | 10.33 (13.78) | 7.45 (12.34) | 2.88 (19.71) | ||
| Now | 73.55 (29.73) | 49.46 (31.15) | 24.07 (27.19) | 56.69 (75.64) | 46.60 (77.23) | 10.09 (69.07) | ||
| Household income/PIR, mean (SE) | 25,942.25 (1,287.69) | 27,850.59 (1,535.37) | 22,523.98 (1,241.55) | 0.002 | 3.06 (0.03) | 3.07 (0.03) | 3.02 (0.05) | 0.37 |
| Smoke, n [in millions] (%) | < 0.001 | < 0.001 | ||||||
| Never | 148.27 (59.94) | 91.34 (57.53) | 56.94 (64.24) | 40.07 (53.46) | 33.74 (55.91) | 6.34 (43.36) | ||
| Former | 23.36 (9.44) | 16.07 (10.12) | 7.29 (8.23) | 19.00 (25.35) | 13.94 (23.09) | 5.06 (34.66) | ||
| Now | 75.75 (30.62) | 51.35 (32.35) | 24.40 (27.53) | 15.88 (21.19) | 12.67 (21.00) | 3.21 (21.98) | ||
| Hypertension, n [in millions] (%) | 0.011 | < 0.001 | ||||||
| No | 133.48 (53.96) | 88.13 (55.51) | 45.34 (51.16) | 48.40 (64.58) | 42.31 (70.12) | 6.09 (41.68) | ||
| Yes | 113.91 (46.04) | 70.62 (44.49) | 43.28 (48.84) | 26.55 (35.42) | 18.03 (29.88) | 8.52 (58.32) | ||
| Total cholesterol, mean (SE), mg/dL | 192.56 (0.90) | 191.42 (1.03) | 194.59 (1.23) | 0.033 | 195.68 (0.50) | 194.71 (0.52) | 199.72 (1.02) | < 0.001 |
| HDL, mean (SE), mg/dL | 50.57 (0.45) | 50.24 (0.49) | 51.15 (0.61) | 0.16 | 53.81 (0.21) | 53.49 (0.23) | 55.12 (0.42) | 0.002 |
| TyG, mean (SE) | 8.71 (0.02) | 8.70 (0.02) | 8.74 (0.02) | 0.11 | 8.62 (0.01) | 8.58 (0.01) | 8.78 (0.02) | < 0.001 |
| TyG–BMI, mean (SE) | 204.27 (1.33) | 203.20 (1.35) | 206.19 (1.74) | 0.022 | 247.89 (0.83) | 242.86 (0.87) | 268.68 (1.88) | < 0.001 |
| TyG–WC, mean (SE) | 739.14 (4.59) | 736.21 (5.03) | 744.40 (5.31) | 0.14 | 850.28 (2.36) | 834.87 (2.44) | 913.90 (4.87) | < 0.001 |
| TyG–WHtR, mean (SE) | 4.70 (0.03) | 4.66 (0.03) | 4.77 (0.03) | < 0.001 | 5.03 (0.01) | 4.92 (0.01) | 5.47 (0.03) | < 0.001 |
All means and SEs for continuous variables and numbers and percentages for categorical variables were weighted
CHARLS: China Health and Retirement Longitudinal Study; HDL: high-density lipoprotein; NHANES: National Health and Nutrition Examination Survey; PIR: poverty income ratio; SE: standard error; TyG:, triglyceride–glucose; TyG–BMI: glucose triglyceride–body mass index; TyG–WC: Glucose triglyceride–waist circumference; TyG–WHtR: glucose triglyceride–waist height ratio
a“Household income/PIR” refers to “household income” for CHARLS participants, and refers to “PIR” for NHANES participants
Association of TyG–BMI, TyG–WC, and TyG–WHtR with arthritis
As shown in Fig. 2, following adjustment for confounding variables, there was a significant association between an increase prevalence of arthritis and both TyG–BMI (per 10-point increase) and TyG–WHtR in both CHARLS (TyG–BMI: OR = 1.02, 95% CI 1.00–1.04; TyG–WHtR: OR = 1.13, 95% CI 1.03–1.24) and NHANES (TyG–BMI: OR = 1.07, 95% CI 1.06–1.08; TyG–WHtR: OR = 1.50, 95% CI 1.40–1.60). Moreover, individuals in the fourth quartile of both TyG–BMI (CHARLS: OR = 1.33, 95% CI 1.03–1.71; NHANES: OR = 2.96, 95% CI 2.45–3.57) and TyG–WHtR (CHARLS: OR = 1.38, 95% CI 1.11–1.71; NHANES: OR = 2.68, 95% CI 2.17–3.31) exhibited a higher prevalence of arthritis compared to those in the first quartile, with statistically significant trend test.
Fig. 2.
Association of TyG–BMI, TyG–WC, and TyG–WHtR with arthritis of the Chinese and the U.S. participants in CHARLS 2011–2012 and NHANES 1999–2018. CHARLS: China Health and Retirement Longitudinal Study; CI: confidence interval; HDL: high-density lipoprotein; NHANES: National Health and Nutrition Examination Survey; OR: odds ratios; PIR: poverty income ratio; Q2: second quartile; Q3: third quartile; TyG: triglyceride–glucose; TyG–BMI: glucose triglyceride–body mass index; TyG–WC: glucose triglyceride–waist circumference; TyG–WHtR: glucose triglyceride–waist height ratio. aQuartile 1 is the reference category. bThe crude model did not adjust for covariates, while the adjusted model adjusted for age, sex, race/ethnicity, marital status, education level, household income/PIR, drinking status, smoke, hypertension, total cholesterol, HDL. cTyG–BMI was treated as a continuous variable with per 10-unit increase. dTyG–WC was treated as a continuous variable with per 100-unit increase
When TyG–WC was treated as a continuous variable, a significant positive association between TyG–WC and the prevalence of arthritis was observed in NHANES (OR = 1.27, 95% CI 1.22–1.32), while the association was not significant in CHARLS (OR = 1.05, 95% CI 1.00–1.12). Results were similar after quartile classification of TyG–WC, and the trend test was not significant in CHARLS.
Additionally, as depicted in Fig. 3, the results remained consistent when arthritis was stratified into RA and OA in NHANES. RCS regression analysis demonstrated a significant nonlinear association between TyG–BMI and arthritis, as well as between TyG–WHtR and arthritis, in both CHARLS and NHANES. When exploring the linear association between TyG–WC and arthritis, the linear relationship between CHARLS and NHANES was inconsistent (Fig. 4). Furthermore, the aforementioned findings remained robust even after employing multiple imputation (Supplementary Table 1, 2).
Fig. 3.
Association of TyG–BMI, TyG–WC, and TyG–WHtR with rheumatoid arthritis and osteoarthritis of the U.S. participants in NHANES 1999–2018. CI: confidence interval; HDL: high-density lipoprotein; NHANES: National Health and Nutrition Examination Survey; OR: odds ratios; PIR: poverty income ratio; Q2: second quartile; Q3: third quartile; TyG: triglyceride-glucose; TyG–BMI: glucose triglyceride–body mass index; TyG–WC: glucose triglyceride–waist circumference; TyG–WHtR, glucose triglyceride–waist height ratio. aQuartile 1 is the reference category. bThe crude model did not adjusted for covariates, while the adjusted model adjusted for age, sex, race/ethnicity, marital status, education level, PIR, drinking status, smoke, hypertension, total cholesterol, HDL. cTyG–BMI was treated as a continuous variable with per 10-unit increase. dTyG–WC was treated as a continuous variable with per 100-unit increase
Fig. 4.
Association of TyG–BMI, TyG–WC, and TyG–WHtR with arthritis of the Chinese and the U.S. participants in CHARLS 2011–2012 and NHANES 1999–2018 by RCS. a The model adjusted for age, sex, race/ethnicity, marital status, education level, household/PIR income, drinking status, smoke, hypertension, total cholesterol, HDL
Secondary and sensitivity analyses
This study further investigates the relationship between TyG and arthritis in CHARLS and NHANES. As illustrated in Supplementary Figs. 1 and 2, while a significant association was found between TyG and arthritis in NHANES (OR = 1.30, 95% CI 1.17–1.45), with consistent results observed after stratifying arthritis into OA and RA, the association was not significant in CHARLS (OR = 1.11, 95% CI 0.95–1.31). Moreover, RCS regression analysis also indicated an inconsistent linear relationship between CHARLS and NHANES (Supplementary Fig. 3). The diagnostic efficacy of TyG–BMI and TyG–WHtR for arthritis was higher than that of TyG in both CHARLS (AUC = 0.5239 for TyG–BMI; AUC = 0.5375 for TyG–WHtR; AUC = 0.5170 for TyG) and NHANES (AUC = 0.6088 for TyG–BMI; AUC = 0.6505 for TyG–WHtR; AUC = 0.5934 for TyG), whereas the diagnostic efficacy of TyG–WC (AUC = 0.5158) in CHARLS was comparable to that of TyG (Supplementary Fig. 4). Furthermore, TyG index combinations with obesity indicator were associated with a higher prevalence of arthritis compared to TyG, except for TyG–WC in CHARLS (Supplementary Table 3).
Additionally, in the subgroup analyses, no significant interactions were found within any subgroup in either the CHARLS or NHANES (Supplementary Tables 4, 5).
Discussion
In this study, TyG–BMI and TyG–WHtR demonstrate a positive correlation with arthritis prevalence in both CHARLS and NHANES, those in the fourth quartile exhibited higher arthritis prevalence compared to those in the first quartile, with a significant trend test. Consistent results were observed when stratifying arthritis into RA and OA in NHANES. Multiple interpolation proves the robust of the results. RCS regression analysis demonstrated a significant nonlinear association. However, while NHANES showed a significant association between TyG and arthritis, CHARLS did not, and RCS regression indicated an inconsistent linear relationship between the two datasets. Besides, the predictive power of TyG–BMI and TyG–WHtR for arthritis was higher than that of TyG in both CHARLS and NHANES, whereas the predictive power of TyG–WC in CHARLS was comparable to that of TyG.
Although research on the TyG index combinations with obesity indicators and arthritis is scarce, TyG has been extensively studied for its associations with arthritis and IR. Studies have established TyG as a reliable surrogate biomarker for IR [10, 16, 17]. In previous research involving arthritis patients, TyG has been recognized as a useful screening tool for IR in RA patients. Moreover, IR estimated by TyG has been associated with disease activity in RA patients [18, 19]. Yan et al. found a positive correlation between TyG and arthritis in NHANES, consistent with our findings [11]. However, our study did not identify an association between TyG and arthritis in CHARLS. Together with our findings, this suggests that TyG may not be as well correlated with arthritis diagnosis as TyG index combinations with obesity indicators (especially TyG–BMI and TyG–WHtR) in both the Chinese and US populations. Additionally, the association of TyG–WC with arthritis diagnosis was inconsistent in both populations, with its relevance for arthritis diagnosis being more pronounced in the US population.
Additionally, studies have indicated that the correlation of TyG index combinations with obesity indicators with IR surpasses that of TyG for assessing IR [13, 20, 21]. Given that arthritis and IR are closely related, our results may in line with these findings. While previous research has suggested that TyG index combinations with obesity indicators are strongly linked to non-alcoholic fatty liver disease, stroke, hyperuricemia, and psoriasis, studies on arthritis are lacking [22–25]. Based on CHARLS data, Li et al. concluded that TyG–BMI and TyG–WHtR serve as effective markers for predicting DM in the Chinese population, a condition often associated with insulin resistance [26]. Similarly, a Chinese community-based study found that TyG–BMI, TyG–WC, TyG–WHR, and TyG–WHtR were all independently associated with hypertension, although our study did not find an interaction for hypertension [27].
TyG index combinations with obesity indicators are frequently implicated in the onset or exacerbation of arthritis by promoting IR. Metabolic changes associated with inflammation, including alterations in glucose tolerance, are believed to play a significant role in this process. Cytokines such as TNF-α, IL-1β, and IL-6, which are commonly associated with increased systemic inflammation, are considered key factors. Among these, TNF-α plays a particularly significant role in the pathogenesis of arthritis. As part of the inflammatory cascade, TNF-α reduces tyrosine phosphorylation of insulin receptor and insulin receptor substrate 1 (IRS-1) kinases, while promoting serine phosphorylation of IRS-1. These actions of TNF-α contribute to the dysfunction of adipocytes and skeletal muscle cells, ultimately resulting in reduced insulin sensitivity and the onset of glucose metabolism disorders [28, 29].
More importantly, persistent inflammation, a hallmark characteristic of obesity, contributes to altered adipokine secretion and exacerbates IR in RA patients with longer disease duration [30, 31]. Obesity triggers adipose tissue to produce increased levels of adipokines, including leptin, which instigates metabolic alterations and exacerbates autoimmunity, thereby predisposing RA patients to metabolic comorbidities. Dysfunction in adipose tissue also results in heightened levels of cytokines such as IL-1β, IL-6, and TNF-α, which further propagate systemic inflammation [32].
To our knowledge, this study is the first to investigate the association between TyG index combinations with obesity indicators and arthritis within two large and representative samples. The rigorous quality control procedures implemented by CHARLS and NHANES in data collection, coupled with their sophisticated sampling designs, enabled us to evaluate this association in a sizable and diverse sample of adults in the China and United States. Furthermore, subgroup analyses and multiple imputation were employed to enhance the robustness and reliability of our findings.
The present study has several limitations that warrant consideration. Firstly, its cross-sectional design prohibits the establishment of causality between TyG index combinations with obesity indicators and arthritis, highlighting the need for further prospective studies and intervention trials. Secondly, the potential for confounding effects arising from measurement errors or residuals from unmeasured variables or unknown confounders cannot be entirely ruled out. Thirdly, the reliance on self-reported diagnosis of arthritis and certain covariates introduces the possibility of measurement errors and inaccuracies in data collection. Employing scientifically robust tools or modalities could enhance the objectivity of the data and their outcomes. Fourth, due to the lack of detailed arthritis typing (RA, OA) in the CHARLS database, we were unable to assess the association of TyG-related index with different types of arthritis in the Chinese population. Lastly, the inconsistency observed in the results of TyG–WC between the US and Chinese populations suggests that the applicability of TyG index combinations with obesity indicators may vary across regions, emphasizing the necessity for further investigations in other geographic areas.
Conclusions
TyG–BMI and TyG–WHtR demonstrate a positive correlation with arthritis prevalence in both Chinese and the U.S. populations, displaying superior diagnostic relevance compared to TyG. Future use for diverse populations has the potential to make them effective markers for early identification of arthritis risk and improved prognosis.
Supplementary Information
Acknowledgements
Not applicable.
Abbreviations
- AUC
Area under the curve
- CDC
Centers for Disease Control and Prevention
- CHARLS
China Health and Retirement Longitudinal Study
- CI
Confidence intervals
- DM
Diabetes mellitus
- HDL
High-density lipoprotein
- IR
Insulin resistance
- IRS-1
Insulin receptor substrate 1
- MEC
Mobile examination center
- NCHS
National Center for Health Statistics
- NHANES
National Health and Nutrition Examination Survey
- OA
Osteoarthritis
- OR
Odds ratios
- PIR
Poverty income ratio
- RA
Rheumatoid arthritis
- RCS
Restricted cubic spline
- ROC
Receiver operating characteristic curve
- SE
Standard error
- TyG
Triglyceride–glucose
- TyG–BMI
Glucose triglyceride–body mass index
- TyG–WC
Glucose triglyceride–waist circumference
- TyG–WHtR
Glucose triglyceride–waist height ratio
Author contributions
XZ and HXT had full access to all the data in the study and took responsibility for the integrity of the data and the accuracy of the data analysis. Conceptualization: XZ, HXT, HZ, FY. Formal analysis: XZ, HXT. Methodology: XZ, HXT. Project administration: HZ, FY. Supervision: HZ, FY. Visualization: XZ, HXT, JJC. Writing-original draft: All authors. Writing—review and editing: JJC, JYC, HFZ, TTQ.
Funding
This research was supported by grants from National Natural Science Foundation of China (No. 82102568; No. 82172432), Shenzhen Key Medical Discipline Construction Fund (No. SZXK023), Shenzhen “San-Ming” Project of Medicine (No. SZSM202211038), Shenzhen Science and Technology Program (No. ZDSYS20220606100602005; No. JCYJ20220818102815033; No. KCXFZ20201221173411031; No. JCYJ20210324110214040), Guangdong Basic and Applied Basic Research Foundation (No. 2022A1515220111; No. 2022B1515120046) and The Scientific Research Foundation of PEKING UNIVERSITY SHENZHEN HOSPITAL (No. KYQD2021099).
Data availability
The data derived from the National Health and Nutrition Examination Survey can be publicly accessed at https://wwwn.cde.gov/nchs/nhanes. And the data derived from the China Health and Retirement Longitudinal Study can be publicly accessed at https://charls.pku.edu.cn/. The data underpinning this article will provide access upon a request to the corresponding author.
Declarations
Ethics approval and consent to participate
The study protocol for NHANES underwent review and approval by the NCHS Research Ethics Review Committee, and all participants provided written informed consent. And all participants of CHARLS in the survey provided informed consent, and the study protocol received approval from the Ethical Review Committee of Peking University (IRB00001052-11,015).
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Xuan Zhang and Haoxian Tang have contributed to the work equally and should be regarded as co-first authors.
Contributor Information
Hui Zeng, Email: zenghui_36@163.com.
Fei Yu, Email: yufei89@pku.edu.cn.
References
- 1.Senthelal S, Li J, Ardeshirzadeh S, Thomas MA. Arthritis. Treasure Island: StatPearls Publishing; 2024. [PubMed] [Google Scholar]
- 2.Tang C-H. Research of pathogenesis and novel therapeutics in arthritis. Int J Mol Sci. 2019;20:1646. 10.3390/ijms20071646 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Long H, Liu Q, Yin H, Wang K, Diao N, Zhang Y, et al. Prevalence trends of site-specific osteoarthritis from 1990 to 2019: findings from the global burden of disease study 2019. Arthr Rheumatol. 2022;74:1172–83. 10.1002/art.42089 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Finckh A, Gilbert B, Hodkinson B, Bae S-C, Thomas R, Deane KD, et al. Global epidemiology of rheumatoid arthritis. Nat Rev Rheumatol. 2022;18:591–602. [DOI] [PubMed] [Google Scholar]
- 5.Petrovská N, Prajzlerová K, Vencovský J, Šenolt L, Filková M. The pre-clinical phase of rheumatoid arthritis: from risk factors to prevention of arthritis. Autoimmun Rev. 2021;20:102797. 10.1016/j.autrev.2021.102797 [DOI] [PubMed] [Google Scholar]
- 6.Deane KD, Demoruelle MK, Kelmenson LB, Kuhn KA, Norris JM, Holers VM. Genetic and environmental risk factors for rheumatoid arthritis. Best Pract Res Clin Rheumatol. 2017;31:3–18. 10.1016/j.berh.2017.08.003 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Chung CP, Oeser A, Solus JF, Avalos I, Gebretsadik T, Shintani A, et al. Prevalence of the metabolic syndrome is increased in rheumatoid arthritis and is associated with coronary atherosclerosis. Atherosclerosis. 2008;196:756–63. 10.1016/j.atherosclerosis.2007.01.004 [DOI] [PubMed] [Google Scholar]
- 8.Li J, Chen Y, Liu Q, Tian Z, Zhang Y. Mechanistic and therapeutic links between rheumatoid arthritis and diabetes mellitus. Clin Exp Med. 2023;23:287–99. 10.1007/s10238-022-00816-1 [DOI] [PubMed] [Google Scholar]
- 9.Lee J, Kim B, Kim W, Ahn C, Choi HY, Kim JG, et al. Lipid indices as simple and clinically useful surrogate markers for insulin resistance in the U.S. population. Sci Rep. 2021;11:2366. 10.1038/s41598-021-82053-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Er L-K, Wu S, Chou H-H, Hsu L-A, Teng M-S, Sun Y-C, et al. Triglyceride glucose-body mass index is a simple and clinically useful surrogate marker for insulin resistance in nondiabetic individuals. PLoS ONE. 2016;11: e0149731. 10.1371/journal.pone.0149731 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Yan Y, Zhou L, La R, Jiang M, Jiang D, Huang L, et al. The association between triglyceride glucose index and arthritis: a population-based study. Lipids Health Dis. 2023;22:132. 10.1186/s12944-023-01899-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Zhao Y, Hu Y, Smith JP, Strauss J, Yang G. Cohort profile: the China Health and Retirement Longitudinal Study (CHARLS). Int J Epidemiol. 2014;43:61–8. 10.1093/ije/dys203 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Dang K, Wang X, Hu J, Zhang Y, Cheng L, Qi X, et al. The association between triglyceride-glucose index and its combination with obesity indicators and cardiovascular disease: NHANES 2003–2018. Cardiovasc Diabetol. 2024;23:8. 10.1186/s12933-023-02115-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.March LM, Schwarz JM, Carfrae BH, Bagge E. Clinical validation of self-reported osteoarthritis. Osteoarthr Cartil. 1998;6:87–93. 10.1053/joca.1997.0098 [DOI] [PubMed] [Google Scholar]
- 15.Wu F, Juonala M, Jacobs DR, Daniels SR, Kähönen M, Woo JG, et al. Childhood non-HDL cholesterol and LDL cholesterol and adult atherosclerotic cardiovascular events. Circulation. 2024;149:217–26. 10.1161/CIRCULATIONAHA.123.064296 [DOI] [PubMed] [Google Scholar]
- 16.Son D-H, Lee HS, Lee Y-J, Lee J-H, Han J-H. Comparison of triglyceride-glucose index and HOMA-IR for predicting prevalence and incidence of metabolic syndrome. Nutr Metab Cardiovasc Dis. 2022;32:596–604. 10.1016/j.numecd.2021.11.017 [DOI] [PubMed] [Google Scholar]
- 17.Khalaji A, Behnoush AH, Khanmohammadi S, Ghanbari Mardasi K, Sharifkashani S, Sahebkar A, et al. Triglyceride-glucose index and heart failure: a systematic review and meta-analysis. Cardiovasc Diabetol. 2023;22:244. 10.1186/s12933-023-01973-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Contreras-Haro B, Hernandez-Gonzalez SO, Gonzalez-Lopez L, Espinel-Bermudez MC, Garcia-Benavides L, Perez-Guerrero E, et al. Fasting triglycerides and glucose index: a useful screening test for assessing insulin resistance in patients diagnosed with rheumatoid arthritis and systemic lupus erythematosus. Diabetol Metab Syndr. 2019;11:95. 10.1186/s13098-019-0495-x [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Elazab SA, Elsayed WE, Alrahim NM, Elsaid MA, Akab SM, Mohammed Enayet AAE, et al. Relationship between triglyceride-glucose index and disease activity and subclinical atherosclerosis in rheumatoid arthritis. Curr Rheumatol Rev. 2023. 10.2174/0115733971259984230922054439. 10.2174/0115733971259984230922054439 [DOI] [PubMed] [Google Scholar]
- 20.Yan S, Wang D, Jia Y. Comparison of insulin resistance-associated parameters in US adults: a cross-sectional study. Hormones. 2023;22:331–41. 10.1007/s42000-023-00448-4 [DOI] [PubMed] [Google Scholar]
- 21.Han Y, Zhou Z, Zhang Y, Zhao G, Xu B. The association of surrogates of insulin resistance with hyperuricemia among middle-aged and older individuals: a population-based nationwide cohort study. Nutrients. 2023;15:3139. 10.3390/nu15143139 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Sheng G, Lu S, Xie Q, Peng N, Kuang M, Zou Y. The usefulness of obesity and lipid-related indices to predict the presence of Non-alcoholic fatty liver disease. Lipids Health Dis. 2021;20:134. 10.1186/s12944-021-01561-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Zhao Y, Zhang J, Chen C, Qin P, Zhang M, Shi X, et al. Comparison of six surrogate insulin resistance indexes for predicting the risk of incident stroke: The Rural Chinese Cohort Study. Diabetes Metab Res Rev. 2022;38: e3567. 10.1002/dmrr.3567 [DOI] [PubMed] [Google Scholar]
- 24.Wang H, Zhang J, Pu Y, Qin S, Liu H, Tian Y, et al. Comparison of different insulin resistance surrogates to predict hyperuricemia among U.S. non-diabetic adults. Front Endocrinol. 2022;13:1028167. 10.3389/fendo.2022.1028167 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Huang D, Ma R, Zhong X, Jiang Y, Lu J, Li Y, et al. Positive association between different triglyceride glucose index-related indicators and psoriasis: evidence from NHANES. Front Immunol. 2023;14:1325557. 10.3389/fimmu.2023.1325557 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Li X, Sun M, Yang Y, Yao N, Yan S, Wang L, et al. Predictive effect of triglyceride glucose-related parameters, obesity indices, and lipid ratios for diabetes in a Chinese population: a prospective cohort study. Front Endocrinol. 2022;13:862919. 10.3389/fendo.2022.862919 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Miao H, Zhou Z, Yang S, Zhang Y. The association of triglyceride-glucose index and related parameters with hypertension and cardiovascular risk: a cross-sectional study. Hypertens Res. 2023. 10.1038/s41440-023-01502-9. 10.1038/s41440-023-01502-9 [DOI] [PubMed] [Google Scholar]
- 28.Popa C, Netea MG, van Riel PLCM, van der Meer JWM, Stalenhoef AFH. The role of TNF-alpha in chronic inflammatory conditions, intermediary metabolism, and cardiovascular risk. J Lipid Res. 2007;48:751–62. 10.1194/jlr.R600021-JLR200 [DOI] [PubMed] [Google Scholar]
- 29.Nicolau J, Lequerré T, Bacquet H, Vittecoq O. Rheumatoid arthritis, insulin resistance, and diabetes. Joint Bone Spine. 2017;84:411–6. 10.1016/j.jbspin.2016.09.001 [DOI] [PubMed] [Google Scholar]
- 30.Cox AJ, West NP, Cripps AW. Obesity, inflammation, and the gut microbiota. Lancet Diabetes Endocrinol. 2015;3:207–15. 10.1016/S2213-8587(14)70134-2 [DOI] [PubMed] [Google Scholar]
- 31.Guin A, Sinhamahapatra P, Misra S, Choudhury Mazumder SR, Chatterjee S, Ghosh A. Incidence and effect of insulin resistance on progression of atherosclerosis in rheumatoid arthritis patients of long disease duration. Biomed J. 2019;42:394–402. 10.1016/j.bj.2019.01.007 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Van Raemdonck K, Umar S, Szekanecz Z, Zomorrodi RK, Shahrara S. Impact of obesity on autoimmune arthritis and its cardiovascular complications. Autoimmun Rev. 2018;17:821–35. 10.1016/j.autrev.2018.02.007 [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 data derived from the National Health and Nutrition Examination Survey can be publicly accessed at https://wwwn.cde.gov/nchs/nhanes. And the data derived from the China Health and Retirement Longitudinal Study can be publicly accessed at https://charls.pku.edu.cn/. The data underpinning this article will provide access upon a request to the corresponding author.