Abstract
This study aimed to explore the relationship between C-reactive protein-triglyceride glucose index (CTI) levels and osteoarthritis (OA) using cross-sectional data from the National Health and Nutrition Examination Survey conducted between 2017 and 2020. The analysis included 3885 individuals from the National Health and Nutrition Examination Survey 2017 to 2020 cohort. Multivariate logistic regression models, along with smoothed curve fitting techniques, were employed to assess the association between CTI and OA. In addition, subgroup analyses were performed to further investigate the relationship between CTI levels and OA. Of the 3885 participants, 515 (14.36%) had a diagnosis of OA. After adjusting for potential confounders, a positive association between CTI levels and OA incidence was observed (odds ratio = 1.3417, 95% confidence interval: 1.1409–1.5777, P = .0004). The highest quartile of CTI showed a significant increase in OA risk (odds ratio = 1.5369, 95% confidence interval: 1.0605–2.2352, P = .0237), with the lowest quartile serving as the reference group. Smooth curve fitting analysis revealed a nonlinear positive correlation between CTI and OA prevalence, with a potentially key inflection point at a CTI value of 10.4266. Subgroup analyses indicated that the educational level was the only variable where a significant interaction between CTI and OA was found (P for interaction = .03). Our findings suggest a significant positive association between CTI levels and OA risk, with a critical threshold identified at 10.4266, which may hold clinical relevance for predicting OA risk. This study underscores the potential of CTI as an innovative, easily accessible inflammatory biomarker for assessing OA risk in adults. However, further prospective studies are required to confirm a causal relationship between CTI and OA.
Keywords: C-reactive protein-triglyceride glucose index, cross-sectional study, National Health and Nutrition Examination Survey, osteoarthritis
1. Introduction
A chronic degenerative condition affecting joints, osteoarthritis (OA) involves structural breakdown of cartilage, growth of bony projections, and remodeling of underlying bone tissue.[1] Globally, it now constitutes the second most frequent cause of persistent physical impairment, second only to coronary artery disease.[2] Recent epidemiological data estimate that over half a billion individuals experience OA-related disabilities, positioning it as the 15th major contributor to global disablement.[3] While causal pathways remain ambiguous, predisposing elements span biological variables (advancing age, hormonal profiles) and mechanical stressors (joint instability, trauma history).[4,5] Population-specific preventive frameworks emphasizing timely screening and personalized interventions are critical for mitigating disease progression.
The triglyceride glucose (TyG) index, a relatively recent measure for assessing insulin resistance (IR), has demonstrated strong correlations with various cardiovascular conditions, including hypertension, atherosclerosis, stroke, and other adverse cardiovascular events.[6,7] IR refers to the body’s diminished sensitivity to insulin, impairing its ability to facilitate glucose uptake from the bloodstream into peripheral tissues, which results in disrupted glucose and lipid metabolism.[8] In addition, emerging evidence suggests that inflammation may serve as an important risk factor for OA, further complicating its pathophysiology.[9]
While the TyG index has been explored in predicting the risk of OA, it often fails to account for the role of inflammation. Given the significant link between IR, inflammation, and OA progression, it is vital to consider a composite measure that incorporates both these factors. The “C-reactive protein-triglyceride glucose index” (CTI) is a novel metric that integrates inflammatory status with IR, providing a more comprehensive assessment of disease risk.[10] Recent studies have highlighted CTI’s utility in predicting survival outcomes for cancer patients in both Chinese and American cohorts, demonstrating its value in cancer mortality risk stratification.[10,11,12,13] Furthermore, CTI is a simple and cost-effective measure, requiring no additional tests beyond routine laboratory work, which reduces the financial burden on both patients and healthcare systems.[14]
The quantitative expression of acute-phase reactant C-reactive protein (CRP) reflects chronic inflammatory burdens within physiological systems, with heightened levels observed in degenerative joint disorders like OA.[15] Mechanical and molecular investigations reveal OA pathogenesis involves tripartite dynamics: biomechanical attrition (age-related degeneration), polymorphism-driven susceptibility (genetic load), adipose-mediated endocrine dysregulation (obesity effects), and cytokine microenvironments that trigger progressive articular matrix deterioration (primary pathogenic cascade).[16–18] The cellular environment in OA is highly complex, involving interactions between various cell types and the cytokines they produce. These cytokines establish complex inflammatory networks that are activated through endocrine, autocrine, and paracrine mechanisms. Recent research has uncovered a connection between IR and inflammatory pathways in OA, suggesting that both factors contribute to the disease’s progression in intricate ways.[19]
Existing research exhibits a conspicuous omission regarding CTI’s potential connection with OA progression. Focused on OA cohorts, the current inquiry pioneers the exploration of CTI-driven predictive modeling for OA susceptibility through advanced analysis of National Health and Nutrition Examination Survey (NHANES) 2017 to 2020 epidemiological records. By performing an innovative analytical framework, this work advances the formulation of optimized OA prevention protocols and personalized therapeutic interventions, underscoring the critical role of timely risk identification and mitigation in healthcare delivery.
2. Materials and methods
2.1. Study design and participants
The NHANES, conducted by the Centers for Disease Control and Prevention, is a biennial initiative designed to evaluate the health and dietary status of the US population. This extensive survey encompasses multiple components, such as face-to-face interviews, physical assessments, self-reported questionnaires, and laboratory tests. Data collection is based on a multistage probability sampling method, ensuring a representative sample of the population. Before participating in any interviews or examinations, individuals are required to provide informed consent. The procedures involved in NHANES are reviewed and approved by the National Center for Health Statistics Research Ethics Review Board, ensuring compliance with the US Department of Health and Human Services Policy for the Protection of Human Research Subjects. For further details on the methodology and data sources of the NHANES survey, the official website (http://www.cdc.gov/nchs/nhanes/index.htm) provides an in-depth overview.
In this study, we utilized data from the NHANES database, specifically from 4 consecutive survey cycles spanning from 2017 to 2020. The original cohort consisted of 15,560 participants. However, individuals lacking information on OA (n = 6354) or the CTI (n = 5321) were excluded from the analysis. Consequently, the final study population comprised 3885 participants, as shown in Figure 1.
Figure 1.
Flowchart of sample selection. CTI = C-reactive protein-triglyceride glucose index, NHANES = National Health and Nutrition Examination Survey.
2.2. Definition of OA
The NHANES database identifies OA status primarily through a self-reported questionnaire located in the “medical conditions” section. Initially, participants are asked, “Has a doctor or healthcare provider ever informed you that you have arthritis?” Individuals who answer “no” to this question are excluded from further analysis in the present study. Those who affirmatively respond are then queried, “What type of arthritis was diagnosed?” Participants who reported having “osteoarthritis” are classified as having OA, while those who specified other types of arthritis, such as “rheumatoid arthritis” or “psoriatic arthritis,” or failed to provide an answer, are categorized as non-OA individuals. The reliability of this self-reported method for identifying OA has been corroborated by previous research, which supports its accuracy in determining OA status.
2.3. Calculation of CTI
The CTI was computed using the following formula: CTI = 0.412 × ln (CRP) + TyG, where TyG is defined as ln ([fasting triglyceride (TG) in mg/dL] × [fasting plasma glucose in mg/dL]/2). Blood samples were obtained in the morning following an 8.5-hour fasting period and were analyzed in laboratories certified by the National Center for Health Statistics. Serum TG levels were quantified using Roche Diagnostics chemical analyzers, while fasting plasma glucose was measured through a Beckman-Coulter Diagnostics laboratory analyzer, utilizing the oxygen rate method. CRP concentrations were assessed using latex-enhanced nephelometry. For further information regarding the laboratory techniques, the NHANES official website provides comprehensive details. Generally, an elevated CTI is indicative of higher levels of inflammation and IR.
2.4. Covariables
For this research, the variables chosen were based on their known correlation with OA and their inclusion in previous similar investigations. These factors involve a broad range of individual characteristics, lifestyle habits, and health conditions. Specifically, we considered factors such as age, liver function markers (including aspartate aminotransferase and alanine aminotransferase [ALT]), blood lipid levels (high-density lipoprotein [HDL] and low-density lipoprotein [LDL]), as well as glycohemoglobin and general hemoglobin (HB) concentrations. In addition, missing data for covariates were handled as follows: covariates with missing data >20% were excluded from the analysis. For covariates with <20% missing data, continuous variables were imputed using mean imputation, and categorical variables were imputed using category probability imputation.
We categorized gender as either male or female, and classified ethnicity/race into several categories: Mexican American, non-Hispanic White, non-Hispanic Black, other Hispanic, and others. Educational level was divided into 3 distinct groups: below high school, high school diploma holders, and those with education beyond high school. Smoking habits were categorized into 2 groups: individuals who had smoked at least 100 cigarettes throughout their lives, compared with those who had not smoked. Drinking behavior was divided into 2 groups based on whether they consumed <12 alcoholic drinks per year or more than 12 drinks per year. Marital status was classified into 2 conditions: those living alone versus those with a partner. Diabetes was categorized as either present or absent.
Body mass index (BMI) was categorized into 3 groups: low to normal weight (<25), overweight (25–29.9), and obese (≥30). Family income-to-poverty ratio (PIR) was classified into 3 categories, namely, <1.3, 1.3 to 3.49, and ≥3.5, which is consistent with previous NHANES studies rather than using tertiles. Further details on the methods used to measure these variables can be found on the official Centers for Disease Control and Prevention website at www.cdc.gov/nchs/nhanes/.
2.5. Statistical analysis
In this study, continuous data are reported as means and standard deviations, while categorical data are expressed as percentages. For comparing continuous variables, the Kruskal–Wallis test is used, while categorical variables are analyzed using the chi-squared (χ2) test. To explore the connection between CTI and OA, we built 3 separate logistic regression models. The first model is unadjusted, while the second model accounts for age, gender, and ethnicity. The third model includes additional covariates, such as PIR, BMI, education level, marital status, smoking, diabetes, alcohol use, and several biochemical markers like ALT, aspartate aminotransferase, HB, glycohemoglobin, HDL, and LDL.
We also performed separate analyses based on gender, BMI, PIR, smoking, alcohol use, marital status, and education level to identify any potential effect modifiers. Odds ratios (ORs) with 95% confidence intervals (CIs) were computed to assess the prevalence of OA relative to CTI levels. To better model the relationship between CTI and OA occurrence, we applied a smooth curve fitting technique. All statistical analyses were carried out using DecisionLinnc1.0 software (Hangzhou Yuantong Information Technology Co., Ltd., Hangzhou, Zhejiang, China, http://www.statsape.com), with a threshold for statistical significance set at a P value <.05.
3. Results
3.1. Baseline characteristics of the participants
Table 1 presents a comprehensive summary of the baseline characteristics of participants, categorized by quartiles of the CTI. The analysis included a total of 3885 individuals, comprising 1876 males (49.41%) and 2009 females (50.86%), with a mean age of 50.91 ± 17.37 years. Among the participants, 515 were diagnosed with OA. The average CTI for the entire cohort was 8.97 ± 0.87.
Table 1.
The baseline characteristics of participants.
| Variable | Overall (N = 3885) | Quartiles of CTI | P value | |||
|---|---|---|---|---|---|---|
| Q1 (N = 971) | Q2 (N = 971) | Q3 (N = 971) | Q4 (N = 972) | |||
| Age, mean (SD) | 50.91 (17.37) | 44.89 (17.93) | 51.11 (17.59) | 53.57 (16.66) | 54.08 (15.67) | <.001 |
| PIR, mean (SD) | 2.63 (1.57) | 2.82 (1.61) | 2.70 (1.60) | 2.55 (1.53) | 2.46 (1.52) | .003 |
| BMI, mean (SD) | 30.05 (7.61) | 25.40 (5.11) | 29.11 (6.17) | 31.26 (7.29) | 34.41 (8.45) | <.001 |
| ALT, mean (SD) | 22.16 (19.43) | 18.75 (14.99) | 20.42 (12.75) | 23.49 (26.71) | 25.98 (19.49) | <.001 |
| AST, mean (SD) | 21.87 (15.20) | 21.52 (15.11) | 20.94 (9.73) | 22.47 (20.42) | 22.56 (13.53) | .245 |
| HB, mean (SD) | 14.02 (1.57) | 13.90 (1.54) | 14.05 (1.52) | 14.06 (1.63) | 14.06 (1.60) | <.001 |
| Glycohemoglobin, mean (SD) | 5.88 (1.15) | 5.43 (0.54) | 5.64 (0.68) | 5.81 (0.83) | 6.64 (1.75) | <.001 |
| HDL, mean (SD) | 53.77 (16.02) | 63.37 (16.54) | 56.00 (15.69) | 50.87 (13.60) | 44.85 (11.74) | <.001 |
| LDL, mean (SD) | 108.32 (35.67) | 98.01 (30.19) | 109.09 (35.04) | 114.51 (36.09) | 111.68 (38.68) | <.001 |
| Gender, n (p%) | ||||||
| Female | 2009.00 (51.71%) | 507.00 (52.21%) | 495.00 (50.98%) | 495.00 (50.98%) | 512.00 (52.67%) | .106 |
| Male | 1876.00 (48.29%) | 464.00 (47.79%) | 476.00 (49.02%) | 476.00 (49.02%) | 460.00 (47.33%) | |
| Race, n (p%) | ||||||
| Mexican American | 496.00 (12.77%) | 94.00 (9.68%) | 119.00 (12.26%) | 122.00 (12.56%) | 161.00 (16.56%) | .006 |
| Non-Hispanic Black | 1312.00 (33.77%) | 312.00 (32.13%) | 310.00 (31.93%) | 325.00 (33.47%) | 365.00 (37.55%) | |
| Non-Hispanic White | 404.00 (10.40%) | 60.00 (6.18%) | 101.00 (10.40%) | 118.00 (12.15%) | 125.00 (12.86%) | |
| Other Hispanic | 979.00 (25.20%) | 295.00 (30.38%) | 278.00 (28.63%) | 227.00 (23.38%) | 179.00 (18.42%) | |
| Other races | 694.00 (17.86%) | 210.00 (21.63%) | 163.00 (16.79%) | 179.00 (18.43%) | 142.00 (14.61%) | |
| Education level, n (p%) | ||||||
| Above high school | 2217.00 (57.07%) | 620.00 (63.85%) | 560.00 (57.67%) | 543.00 (55.92%) | 494.00 (50.82%) | <.001 |
| High school | 913.00 (23.50%) | 221.00 (22.76%) | 230.00 (23.69%) | 227.00 (23.38%) | 235.00 (24.18%) | |
| Under high school | 755.00 (19.43%) | 130.00 (13.39%) | 181.00 (18.64%) | 201.00 (20.70%) | 243.00 (25.00%) | |
| Marital, n (p%) | ||||||
| Living alone | 1609.00 (41.42%) | 432.00 (44.49%) | 388.00 (39.96%) | 384.00 (39.55%) | 405.00 (41.67%) | .049 |
| Living with a partner | 2276.00 (58.58%) | 539.00 (55.51%) | 583.00 (60.04%) | 587.00 (60.45%) | 567.00 (58.33%) | |
| PIR-group, n (p%) | ||||||
| <1.30 | 1014.00 (26.10%) | 227.00 (23.38%) | 251.00 (25.85%) | 261.00 (26.88%) | 275.00 (28.29%) | .002 |
| 1.30–3.49 | 1678.00 (43.19%) | 392.00 (40.37%) | 399.00 (41.09%) | 440.00 (45.31%) | 447.00 (45.99%) | |
| ≥3.50 | 1193.00 (30.71%) | 352.00 (36.25%) | 321.00 (33.06%) | 270.00 (27.81%) | 250.00 (25.72%) | |
| BMI group, n (p%) | ||||||
| Normal | 982.00 (25.28%) | 510.00 (52.52%) | 234.00 (24.10%) | 167.00 (17.20%) | 71.00 (7.30%) | <.001 |
| Obese | 1654.00 (42.57%) | 154.00 (15.86%) | 367.00 (37.80%) | 474.00 (48.82%) | 659.00 (67.80%) | |
| Overweight | 1249.00 (32.15%) | 307.00 (31.62%) | 370.00 (38.11%) | 330.00 (33.99%) | 242.00 (24.90%) | |
| Drink, n (p%) | ||||||
| No | 319.00 (8.21%) | 87.00 (8.96%) | 66.00 (6.80%) | 89.00 (9.17%) | 77.00 (7.92%) | .529 |
| Yes | 3566.00 (91.79%) | 884.00 (91.04%) | 905.00 (93.20%) | 882.00 (90.83%) | 895.00 (92.08%) | |
| Diabetes, n (p%) | ||||||
| No | 3256.00 (83.81%) | 917.00 (94.44%) | 878.00 (90.42%) | 828.00 (85.27%) | 633.00 (65.12%) | <.001 |
| Yes | 629.00 (16.19%) | 54.00 (5.56%) | 93.00 (9.58%) | 143.00 (14.73%) | 339.00 (34.88%) | |
| Smoke, n (p%) | ||||||
| No | 2217.00 (57.07%) | 604.00 (62.20%) | 597.00 (61.48%) | 533.00 (54.89%) | 483.00 (49.69%) | .002 |
| Yes | 1668.00 (42.93%) | 367.00 (37.80%) | 374.00 (38.52%) | 438.00 (45.11%) | 489.00 (50.31%) | |
| OA, n (p%) | ||||||
| No | 3370.00 (86.74%) | 889.00 (91.56%) | 853.00 (87.85%) | 831.00 (85.58%) | 797.00 (82.00%) | .011 |
| Yes | 515.00 (13.26%) | 82.00 (8.44%) | 118.00 (12.15%) | 140.00 (14.42%) | 175.00 (18.00%) | |
Data are summarized as mean ± SD for continuous variables or as proportions for categorical variables.
Bold values indicate statistical significance with P < .05.
ALT = alanine aminotransferase, AST = aspartate aminotransferase, BMI = body mass index, CTI = C-reactive protein-triglyceride glucose index, HB = hemoglobin, HDL = high-density lipoprotein, LDL = low-density lipoprotein, OA = osteoarthritis, PIR = income-to-poverty ratio, SD = standard deviation.
Individuals in the higher CTI quartiles were generally older, exhibited higher levels of educational attainment, and had a greater prevalence of diabetes. In addition, non-Hispanic Black individuals were more likely to fall into the higher CTI groups. The data also revealed that participants with elevated CTI levels were more likely to engage in smoking, live with a partner, and have higher rates of obesity. Moreover, those with higher CTI values demonstrated elevated levels of key biomarkers, including ALT, HB, glycohemoglobin, HDL, and LDL. These findings suggest a significant association between higher CTI levels and a range of demographic, behavioral, and metabolic factors.
3.2. Associations between the CTI and the risk of OA
Table 2 outlines the association between CTI and OA, analyzed across 3 distinct models. In all models, an increase in CTI was consistently linked with an elevated risk of OA. In the unadjusted model (model 1), a continuous rise in CTI showed an OR of 1.4575 (95% CI: 1.3122–1.6195; P < .0001). In model 2, after adjusting for age, gender, and ethnicity, the OR decreased slightly to 1.3169 (95% CI: 1.1705–1.4816; P < .0001). In model 3, further adjustments for factors like BMI, PIR, educational background, laboratory results, and lifestyle behaviors showed that the OR remained statistically significant at 1.3417 (95% CI: 1.1409–1.5777; P = .0004), reinforcing the persistent relationship between increased CTI and higher OA risk. When CTI was classified into quartiles, the Q4 displayed consistently higher OA risk across all models. In model 1, the OR for Q4 was 2.3805 (95% CI: 1.8067–3.1598; P < .0001). In model 2, this OR decreased to 1.7189 (95% CI: 1.2793–2.3237; P = .0004), and in model 3, the OR further reduced to 1.5369 (95% CI: 1.0605–2.2352; P = .0237). These results reinforce the strong association between higher CTI levels and an increased risk of OA, even after controlling for demographic, clinical, and lifestyle factors. The trend analysis confirmed that the positive relationship between CTI and OA was statistically significant across all models, with all P values for trend being <.05.
Table 2.
Association between CTI and OA.
| Variable | OR (95% CI), P value | ||
|---|---|---|---|
| Model 1 | Model 2 | Model 3 | |
| CTI | 1.4575 (1.3122–1.6195), <.0001 | 1.3169 (1.1705–1.4816), <.0001 | 1.3417 (1.1409–1.5777), .0004 |
| Stratified by CTI quartiles | |||
| Q1 | Reference | Reference | Reference |
| Q2 | 1.4998 (1.1163–2.0236), .0075 | 1.1710 (0.8556–1.6080), .3262 | 1.0800 (0.7762–1.5073), .6491 |
| Q3 | 1.8265 (1.3722–2.4453), <.0001 | 1.3278 (0.9787–1.8099), .0703 | 1.2084 (0.8597–1.7050), .2782 |
| Q4 | 2.3805 (1.8067–3.1598), <.0001 | 1.7189 (1.2793–2.3237), .0004 | 1.5369 (1.0605–2.2352), .0237 |
| P for trend | 1.3149 (1.2081–1.4324), <.001 | 1.1965 (1.0911–1.3129), .0001 | 1.1578 (1.0293–1.3029), .0148 |
Model 1: adjusted for none.
Model 2: adjusted for race, age combined with race.
Model 3: adjusted for model 2 combined with education level, marital status, smoking behavior, alcohol consumption, diabetes, AST, ALT, HB, BMI, PIR, glycohemoglobin, HDL, and LDL.
Bold values indicate statistical significance with P < .05.
ALT = alanine aminotransferase, AST = aspartate aminotransferase, BMI = body mass index, CI = confidence interval, CTI = C-reactive protein-triglyceride glucose index, HB = hemoglobin, HDL = high-density lipoprotein, LDL = low-density lipoprotein, OA = osteoarthritis, OR = odds ratio, PIR = income-to-poverty ratio.
In order to investigate the potential for a nonlinear association between CTI and OA, methods such as generalized additive models and smooth curve fitting were applied (Fig. 2). The adjusted model revealed a significant potential inflection point at a CTI value of 10.4266 (refer to Table 3). For CTI values below this point, each increase of 1 unit in CTI resulted in a 1.232-fold rise in the probability of developing OA (OR = 1.232, 95% CI: 1.0323–1.4704, P = .0207). However, based on this analysis, once the CTI surpassed the 10.4266 threshold, the association became notably stronger, with a single unit increase in CTI leading to a 2.589-fold increase in the likelihood of OA (OR = 2.589, 95% CI: 1.4548–4.6074, P = .0012). These findings support a nonlinear and positive relationship between CTI and OA, indicating that OA risk increases at a much sharper rate as CTI exceeds this threshold.
Figure 2.
The relationship between the CTI and OA is demonstrated through smooth curve fitting methods utilizing a generalized additive model. Adjusted for age, sex, race, education level, marital status, smoking behavior, alcohol consumption, diabetes, AST, ALT, HB, BMI, PIR, glycohemoglobin, HDL, and LDL. The solid blue line illustrates the smooth curve fit between the variables, while the dashed gray line denotes the 95% CIs obtained from the fit. ALT = alanine aminotransferase, AST = aspartate aminotransferase, BMI = body mass index, CI = confidence interval, CTI = C-reactive protein-triglyceride glucose index, HB = hemoglobin, HDL = high-density lipoprotein, LDL = low-density lipoprotein, OA = osteoarthritis, OR = odds ratio, PIR = income-to-poverty ratio.
Table 3.
Threshold effect analysis.
| Variable | OR | 95% CI | P value | |
|---|---|---|---|---|
| CTI | Model 1 Line effect | 1.3380 | 1.1372–1.5741 | .0004 |
| Model 2 threshold | 10.4266 | – | – | |
| Model 2 < W effect | 1.2320 | 1.0323–1.4704 | .0207 | |
| Model 2 > W effect | 2.5890 | 1.4548–4.6074 | .0012 | |
| Log-likelihood ratio test | – | – | .0241 | |
Adjusted for age, gender, race, education level, marital status, smoking behavior, alcohol consumption, diabetes, AST, ALT, HB, BMI, PIR, glycohemoglobin, HDL, and LDL.
Bold values indicate statistical significance with P < .05.
ALT = alanine aminotransferase, AST = aspartate aminotransferase, BMI = body mass index, CI = confidence interval, CTI = C-reactive protein-triglyceride glucose index, HB = hemoglobin, HDL = high-density lipoprotein, LDL = low-density lipoprotein, OA = osteoarthritis, OR = odds ratio, PIR = income-to-poverty ratio.
3.3. Subgroup analysis
To assess the stability of the nonlinear positive association between CTI and OA across different demographic groups (Fig. 3), a subgroup analysis was carried out. The analysis revealed that the strength of this relationship differed notably depending on education level (P for interaction = .03). Specifically, those with higher education, high school graduates, and individuals with less than high school education had ORs (with 95% CI) of 1.44 (1.20–1.75), 1.92 (1.41–2.60), and 2.19 (1.48–3.24), respectively. However, no significant variation was observed when the data were stratified by factors such as sex, race, marital status, PIR, BMI, smoking, alcohol consumption, or diabetes status (all P > .05). These findings suggest that the nonlinear positive relationship between CTI and OA holds true across most subgroups, with the exception of differences in educational level. And given the exploratory nature of our subgroup analysis, these results should be interpreted with caution, and further validation through larger prospective studies is warranted.
Figure 3.
The results of subgroup analysis were adjusted for all covariates except effect modifier. BMI = body mass index, CI = confidence interval, OR = odds ratio, PIR = family income–poverty ratio.
4. Discussion
This cohort analysis identifies the CTI as a novel composite marker that reflects both inflammation and IR, serving as a potential marker of OA risk. Drawing from the NHANES database, this study included 11,381 participants between 2017 and 2020, comprising 2009 females (50.86%) and 1876 males (49.41%). Among the participants, 515 (14.36%) were diagnosed with OA. Our findings revealed a significant positive correlation between CTI levels and OA, suggesting that higher CTI values are associated with an increased risk of developing the condition. Notably, from the NHANES evidence, while the risk of OA increased gradually when CTI levels were below 10.4266, the risk escalated more sharply when CTI values exceeded this threshold.
OA is a prevalent disorder affecting millions globally, particularly impacting joints such as the knees, hips, hands, and spine.[20] As the most common form of arthritis, OA continues to be a subject of intense research, especially with the advancements in medical technology that have enhanced the understanding of its etiology and treatment options. The global burden of OA is increasing annually, with rising incidence rates and broader health implications.[21] Existing research highlights the role of inflammation in linking IR to OA outcomes. Inflammation is critical in the progression of OA, with multiple factors influencing the relationship between IR and the disease, including systemic inflammation, metabolic dysfunction, lipid abnormalities, and obesity.[22] These factors collectively contribute to cartilage degradation, synovial inflammation, and joint damage, thereby linking metabolic disorders to the onset and progression of OA.[23] Veronese et al further elucidated how hyperglycemia and IR, particularly in type 2 diabetes, reduce insulin receptor sensitivity, impairing chondrocyte function and accelerating OA progression.[24] In addition, Wang et al demonstrated that the metabolic disruptions in OA are closely tied to IR, with energy stagnation caused by the downregulation of AMP-activated protein kinase potentially offering a novel therapeutic target for OA.[25]
Furthermore, immunoinflammatory pathways play a crucial role in both the initiation and progression of OA, underscoring the importance of regulating inflammatory responses for effective disease management.[26] Inflammation within joint tissues drives cartilage degradation and synovitis through the activation of cartilage-degrading enzymes, apoptosis, and matrix protein alterations.[27,28] Clinical markers such as CRP and erythrocyte sedimentation rate are frequently used to assess systemic inflammation and are typically elevated in inflammatory joint conditions, including rheumatoid arthritis.[29] While OA was traditionally considered a noninflammatory disease, more recent studies have revealed that high-sensitivity CRP and erythrocyte sedimentation rate levels are significantly elevated in OA patients and are closely linked to disease progression and prognosis.[30,31]
Given that both IR and inflammation have been established as independent risk factors for OA, we hypothesize that CTI may serve as an effective predictive tool for identifying individuals at heightened risk for OA development. Our results align with previous studies, which suggest that elevated markers of inflammation, including CRP and the IR index, are strongly associated with OA. For example, Tang et al demonstrated that increased CTI levels correlate with a higher stroke risk in hypertensive individuals,[32] while Mei et al found a positive relationship between higher CTI levels and an increased risk of erectile dysfunction in American men.[33] Moreover, Yan et al also highlighted a robust positive correlation between CTI and erectile dysfunction, although this relationship remains underexplored.[34]
This investigation boasts several strengths. To begin with, the NHANES database ensures a robust and representative sample, enhancing the external validity and applicability of the findings. Furthermore, the analysis incorporated thorough adjustments for various covariates, boosting the credibility of the results. This research marks the first extensive study that uses the NHANES database to explore how CTI correlates with OA. However, the study is not without limitations. Its cross-sectional nature prevents it from establishing causal links between CTI and OA, needing prospective studies to confirm our findings. In addition, OA diagnoses were based on participant self-reports through surveys, while key confounders such as prior joint injury, occupational load, physical activity, and medication use were not available in the dataset and could not be included in our analysis. These factors may introduce recall biases. Meanwhile, it remains unclear whether CTI provides additional discriminatory value beyond well-established inflammatory or metabolic markers such as CRP, fasting glucose, and TGs. Future studies that directly compare the discriminatory power of CTI with these markers are needed to determine its added value. Moreover, the limitations of the NHANES database meant that not all potential variables affecting OA and inflammatory markers could be included, possibly affecting the completeness and size of the data set.
In conclusion, our study presents CTI as a promising marker for predicting the risk of OA, highlighting its potential utility in clinical practice for early detection and intervention. Future longitudinal studies are needed to further investigate the causality and clinical applicability of CTI in OA management.
5. Conclusion
This study reveals a notable positive connection between CTI and OA, underlining the importance of early intervention for individuals with elevated CTI to mitigate the risk of OA onset. Although a possible association between CTI and OA is suggested, the cross-sectional approach of the current study limits the ability to establish a clear cause-and-effect relationship. To strengthen these conclusions, further large-scale prospective studies are essential to validate the directionality and robustness of the association between CTI and OA.
Acknowledgments
We would like to thank NHANES for the database.
Author contributions
Conceptualization: Feng Zhou, Mingyang Huang, Shihai Wang.
Data curation: Feng Zhou, Mingyang Huang, Shihai Wang.
Formal analysis: Feng Zhou, Shihai Wang.
Investigation: Feng Zhou, Mingyang Huang, Shihai Wang.
Methodology: Feng Zhou, Mingyang Huang, Shihai Wang.
Resources: Feng Zhou, Mingyang Huang, Shihai Wang.
Software: Feng Zhou, Mingyang Huang, Shihai Wang.
Validation: Feng Zhou, Shihai Wang.
Visualization: Feng Zhou, Shihai Wang.
Project administration: Shihai Wang.
Supervision: Shihai Wang.
Writing – review & editing: Shihai Wang.
Writing – original draft: Feng Zhou, Mingyang Huang.
Abbreviations:
- ALT
- alanine aminotransferase
- BMI
- body mass index
- CI
- confidence interval
- CRP
- C-reactive protein
- CTI
- C-reactive protein-triglyceride glucose index
- HB
- hemoglobin
- HDL
- high-density lipoprotein
- IR
- insulin resistance
- LDL
- low-density lipoprotein
- NHANES
- National Health and Nutrition Examination Survey
- OA
- osteoarthritis
- OR
- odds ratio
- PIR
- income-to-poverty ratio
- TG
- triglyceride
- TyG
- triglyceride glucose
The NCHS Ethics Review Board examined and approved this study. In order to take part in this study, the patients/participants gave their written informed permission. This study uses anonymized, publicly available data obtained from the NHANES (www.cdc.gov/nchs/nhanes/).
The authors have no funding and conflicts of interest to disclose.
The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.
How to cite this article: Zhou F, Huang M, Wang S. Association between C-reactive protein-triglyceride glucose index and osteoarthritis in US adults: Evidence from NHANES (2017–2020). Medicine 2026;105:17(e48520).
Contributor Information
Feng Zhou, Email: 407212@csu.edu.cn.
Mingyang Huang, Email: 2451416251@qq.com.
References
- [1].Englund M. Osteoarthritis, part of life or a curable disease? A bird’s-eye view. J Intern Med. 2023;293:681–93. [DOI] [PubMed] [Google Scholar]
- [2].Speerin R, Slater H, Li L, et al. Moving from evidence to practice: models of care for the prevention and management of musculoskeletal conditions. Best Pract Res Clin Rheumatol. 2014;28:479–515. [DOI] [PubMed] [Google Scholar]
- [3].GBD 2021 Osteoarthritis Collaborators. Global, regional, and national burden of osteoarthritis, 1990-2020 and projections to 2050: a systematic analysis for the Global Burden of Disease Study 2021. Lancet Rheumatol. 2023;5:e508–22. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [4].Hartley A, Gregson CL, Paternoster L, Tobias JH. Osteoarthritis: insights offered by the study of bone mass genetics. Curr Osteoporos Rep. 2021;19:115–22. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [5].Wang T, He C. Pro-inflammatory cytokines: the link between obesity and osteoarthritis. Cytokine Growth Factor Rev. 2018;44:38–50. [DOI] [PubMed] [Google Scholar]
- [6].Hong S, Han K, Park CY. The triglyceride glucose index is a simple and low-cost marker associated with atherosclerotic cardiovascular disease: a population-based study. BMC Med. 2020;18:361. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [7].Ding X, Wang X, Wu J, Zhang M, Cui M. Triglyceride-glucose index and the incidence of atherosclerotic cardiovascular diseases: a meta-analysis of cohort studies. Cardiovasc Diabetol. 2021;20:76. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [8].Lee S-H, Park S-Y, Choi C. Insulin resistance: from mechanisms to therapeutic strategies. Diabet Metab J. 2021;46:15–37. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [9].Terkawi MA, Ebata T, Yokota S, et al. Low-grade inflammation in the pathogenesis of osteoarthritis: cellular and molecular mechanisms and strategies for future therapeutic intervention. Biomedicines. 2022;10:1109. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [10].Ruan G-T, Xie H-L, Zhang H-Y, et al. A novel inflammation and insulin resistance related indicator to predict the survival of patients with cancer. Front Endocrinol. 2022;13:905266. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [11].Zhao D-F. Value of C-reactive protein-triglyceride glucose index in predicting cancer mortality in the general population: results from National Health and Nutrition Examination Survey. Nutr Cancer. 2023;75:1934–44. [DOI] [PubMed] [Google Scholar]
- [12].Ruan GT, Shi J-Y, Xie H-L, et al. Prognostic importance of an indicator related to systemic inflammation and insulin resistance in patients with gastrointestinal cancer: a prospective study. Front Oncol. 2024;14:1394892. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [13].Ruan GT, Deng L, Xie H-L, et al. Systemic inflammation and insulin resistance-related indicator predicts poor outcome in patients with cancer cachexia. Cancer Metabo. 2024;12:3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [14].Xu M, Zhang L, Xu D, Shi W, Zhang W. Usefulness of C-reactive protein-triglyceride glucose index in detecting prevalent coronary heart disease: findings from the National Health and Nutrition Examination Survey 1999-2018. Front Cardiovasc Med. 2024;11:1485538. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [15].Alexander LC Jr, McHorse G, Huebner JL, et al. A matrix metalloproteinase-generated neoepitope of CRP can identify knee and multi-joint inflammation in osteoarthritis. Arthritis Res Ther. 2021;23:226. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [16].Chen L, Zheng JJY, Li G, et al. Pathogenesis and clinical management of obesity-related knee osteoarthritis: impact of mechanical loading. J Orthop Transl. 2020;24:66–75. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [17].Kiełbowski K, Herian M, Bakinowska E, Banach B, Sroczyński T, Pawlik A. The role of genetics and epigenetic regulation in the pathogenesis of osteoarthritis. Int J Mol Sci. 2023;24:11655. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [18].Mocanu V, Timofte DV, Zară-Dănceanu C-M, Labusca L. Obesity, metabolic syndrome, and osteoarthritis require integrative understanding and management. Biomedicines. 2024;12:1262. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [19].Bradley D. The intriguing intersection of type 2 diabetes, obesity-related insulin resistance, and osteoarthritis. J Clin Endocrinol Metab. 2021;106:e2370–2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [20].Jordan JM. Osteoarthritis: new perspectives. Curr Opin Rheumatol. 2010;22:510–1. [DOI] [PubMed] [Google Scholar]
- [21].Weng Q, Chen Q, Jiang T, et al. Global burden of early-onset osteoarthritis, 1990–2019: results from the Global Burden of Disease Study 2019. Ann Rheum Dis. 2024;83:915–25. [DOI] [PubMed] [Google Scholar]
- [22].Nedunchezhiyan U, Varughese I, Sun AR, Wu X, Crawford R, Prasadam I. Obesity, inflammation, and immune system in osteoarthritis. Front Immunol. 2022;13:907750. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [23].Kulkarni P, Martson A, Vidya R, Chitnavis S, Harsulkar A. Pathophysiological landscape of osteoarthritis. Adv Clin Chem. 2020;100:37–90. [DOI] [PubMed] [Google Scholar]
- [24].Veronese N, Cooper C, Reginster JY, et al. Type 2 diabetes mellitus and osteoarthritis. Semin Arthritis Rheum. 2019;49:9–19. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [25].Wang J, Li J, Song D, et al. AMPK: implications in osteoarthritis and therapeutic targets. Am J Transl Res. 2020;12:7670–81. [PMC free article] [PubMed] [Google Scholar]
- [26].Herrero-Beaumont G, Pérez-Baos S, Sánchez-Pernaute O, Roman-Blas JA, Lamuedra A, Largo R. Targeting chronic innate inflammatory pathways, the main road to prevention of osteoarthritis progression. Biochem Pharmacol. 2019;165:24–32. [DOI] [PubMed] [Google Scholar]
- [27].Griffin T, Scanzello C. Innate inflammation and synovial macrophages in osteoarthritis pathophysiology. Clin Exp Rheumatol. 2019;37(Suppl 120):57–63. [PMC free article] [PubMed] [Google Scholar]
- [28].Li Z, Huang Z, Zhang H, et al. P2X7 receptor induces pyroptotic inflammation and cartilage degradation in osteoarthritis via NF-κB/NLRP3 crosstalk. Oxid Med Cell Longevity. 2021;2021:8868361. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [29].Kim MJ, Lee EB, Song YW, Park JK. Profile of common inflammatory markers in treatment-naïve patients with systemic rheumatic diseases. Clin Rheumatol. 2020;39:2899–906. [DOI] [PubMed] [Google Scholar]
- [30].Kondo F, Takegami Y, Ishizuka S, Hasegawa Y, Imagama S. The association of the progression of knee osteoarthritis with high-sensitivity CRP in community-dwelling people – the Yakumo study. Clin Rheumatol. 2021;40:2643–9. [DOI] [PubMed] [Google Scholar]
- [31].Zaki NA, Saleem TH, Said E, Hassan MH. Evidence of contributory role of inflammation in patients with knee osteoarthritis. SVU-Int J Med Sci. 2020;3:1–6. [Google Scholar]
- [32].Tang S, Wang H, Li K, et al. C-reactive protein-triglyceride glucose index predicts stroke incidence in a hypertensive population: a national cohort study. Diabetol Metab Syndrome. 2024;16:277. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [33].Mei Y, Li Y, Zhang B, et al. Association between the C-reactive protein-triglyceride glucose index and erectile dysfunction in US males: results from NHANES 2001-2004. Int J Impot Res. 2025;37:612–22. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [34].Yan Y, Zhou L, La R, et al. Is erectile dysfunction associated with osteoarthritis and rheumatoid arthritis? Insights from a population-based study. Sexual Med. 2024;12:qfae028. [DOI] [PMC free article] [PubMed] [Google Scholar]