Association between the CRP-triglyceride-glucose index and chronic obstructive pulmonary disease: A cross-sectional study based on NHANES 2015 to 2018

Chronic obstructive pulmonary disease (COPD), characterized by persistent airflow limitation, ranks as the third leading global cause of mortality. [ 1 , 2 ] With the rising global prevalence of smoking, both the incidence and mortality of COPD exhibit an escalating trend. Projections indicate that annual COPD-attributable deaths will exceed 4.5 million by 2030, [ 3 , 4 ] while the Global Burden of Disease Study forecasts a surge in global COPD cases to approximately 592 million by 2050, posing a substantial public health challenge. [ 5 ] Consequently, identifying effective early-warning biomarkers is crucial for implementing clinical interventions to improve prognosis. Inflammation, nutritional imbalance, and immune dysregulation play pivotal roles in COPD pathogenesis. [ 6 , 7 ] Existing studies have validated associations between COPD and various individual blood biomarkers. For instance, exacerbated insulin resistance (IR) correlates significantly with a reduced percentage of predicted forced expiratory volume in 1 second and a percentage of predicted forced vital capacity in COPD patients. [ 8 ] Currently, the triglyceride–glucose index (TyG) serves as a widely adopted surrogate for IR assessment. [ 9 ] Elevated C-reactive protein (CRP) levels are observed in COPD patients, with high baseline CRP independently predicting increased long-term mortality. [ 10 , 11 ] Nevertheless, these individual biomarkers demonstrate limited explanatory power for the multifactorial pathophysiology of COPD. Given the substantial disease burden and mechanistic complexity, developing novel composite biomarkers for early risk identification and stratification represents an urgent clinical need. The CRP–triglyceride–glucose index (CTI), integrating TyG and high-sensitivity CRP (hs-CRP), emerges as a novel indicator concurrently reflecting IR and systemic inflammation. Initially proposed by Ruan et al [ 12 ] in 2022, CTI has been validated for predicting survival in cancer populations. As a laboratory-based indicator, CTI offers computational simplicity without imposing additional burdens on patients. Recent studies establish associations between CTI and conditions including coronary heart disease, [ 13 ] hepatic fibrosis, [ 14 ] and endometriosis. [ 15 ] However, the correlation between CTI and COPD remains largely unexplored. Despite growing evidence linking IR and systemic inflammation to COPD, most prior studies have focused on single biomarkers, such as TyG or CRP, in isolation. However, COPD is increasingly recognized as a systemic disease characterized by complex metabolic–inflammatory interactions. Composite biomarkers that simultaneously capture these interrelated pathways may better reflect the underlying pathophysiological mechanisms and improve risk stratification. The CTI, which integrates hs-CRP and TyG, has emerged as a novel indicator of combined inflammatory and metabolic dysregulation. Although CTI has been associated with cardiovascular diseases, liver fibrosis, and cancer prognosis, its relationship with COPD in the general population has not been investigated. Therefore, the present study aimed to evaluate the association between CTI and COPD prevalence using nationally representative data from National Health and Nutrition Examination Survey (NHANES) 2015 to 2018. We specifically sought to examine whether higher CTI levels are independently associated with COPD, assess potential dose–response relationships between CTI and COPD using restricted cubic splines (RCS), explore whether this association varies across demographic and socioeconomic subgroups, and compare the discriminatory performance of CTI with that of individual biomarkers. Based on existing evidence regarding the interplay between systemic inflammation, IR, and pulmonary dysfunction, we hypothesized a priori that elevated CTI levels would be positively associated with COPD prevalence in the adult population.

Conceptualization: Peng Sun, Ziqi Sang, Dongkai Zhao. Data curation: Peng Sun, Chunlei Ma, Zheng Liang. Formal analysis: Peng Sun, Peng Dai. Funding acquisition: Junnan Liu, Dongkai Zhao. Investigation: Peng Sun, Lingyu Xu. Methodology: Peng Sun, Chunlei Ma. Validation: Peng Sun, Ziqi Sang, Xupeng Huang. Writing – original draft: Peng Sun, Peng Dai, Lingyu Xu, Zheng Liang. Writing – review & editing: Peng Sun, Chunlei Ma, Junnan Liu, Dongkai Zhao.

This study was designed as a nationally representative cross-sectional analysis based on data from the NHANES 2015 to 2018 cycles. NHANES is an ongoing, stratified, multistage probability survey conducted by the National Center for Health Statistics (NCHS) to assess the health and nutritional status of the noninstitutionalized civilian population of the United States. The survey combines in-home interviews, standardized physical examinations, and laboratory assessments conducted in mobile examination centers (MECs). Given the complex sampling design of NHANES, all analyses incorporated examination sampling weights, clustering, and stratification variables in accordance with the analytic guidelines provided by the Centers for Disease Control and Prevention (CDC). The present study evaluated the cross-sectional association between the CTI and self-reported COPD. All participants provided written informed consent, and the NHANES protocol was approved by the Research Ethics Review Board of the CDC. NHANES is a cross-sectional study targeting the noninstitutionalized civilian population in the United States. Trained interviewers and medical examiners collected self-reported data via computer-assisted personal interviewing systems. All participants provided written informed consent, and the study protocol was approved by the Research Ethics Review Board of the CDC. From the 2015 to 2016 and 2017 to 2018 NHANES cycles, data from 19,225 participants were initially extracted. After sequentially excluding individuals aged <20 years (N = 7937), pregnant women (N = 125), those with missing CTI calculation data (N = 6477), participants lacking COPD diagnostic records (N = 6), and those with incomplete covariate information (N = 1304), the final analytical cohort comprised 3442 eligible participants (Fig. 1 ). Flowchart of study participants. COPD = chronic obstructive pulmonary disease, CTI = C-reactive protein–triglyceride–glucose index, NHANES = National Health and Nutrition Examination Survey. NHANES 2015 to 2018 includes 2 consecutive survey cycles (2015–2016 and 2017–2018). Data collection was conducted continuously during this period across multiple locations in the United States using a MEC system. Participants were recruited through a stratified, multistage probability sampling framework to ensure national representativeness. Exposure (CTI components including fasting plasma glucose [FPG], triglycerides [TGs], and hs-CRP), covariates, and outcome (self-reported COPD diagnosis) were collected during the same survey cycle, reflecting a cross-sectional design without longitudinal follow-up. Laboratory measurements were obtained during standardized morning examination sessions following fasting protocols, and questionnaire data were collected through structured in-home interviews. As this was a secondary analysis of publicly available de-identified data, no additional recruitment or follow-up was performed for the present study. Participants were drawn from the NHANES 2015 to 2018 cycles, which employed a stratified, multistage probability sampling design to obtain a nationally representative sample of the noninstitutionalized civilian population of the United States. Eligibility criteria for the present analysis were as follows: age ≥ 20 years; availability of complete laboratory measurements required for CTI calculation (FPG, TGs, and hs-CRP); availability of data on COPD status; and complete information on prespecified covariates. From an initial 19,225 participants in the 2015 to 2018 cycles, we excluded individuals younger than 20 years (n = 7937), pregnant women (n = 125), participants with missing CTI-related laboratory data (n = 6477), individuals lacking COPD diagnostic information (n = 6), and those with incomplete covariate data (n = 1304). After applying these eligibility criteria, a total of 3442 participants were included in the final analytic sample (Fig. 1 ). COPD status was determined based on self-reported physician diagnosis obtained from the NHANES Medical Conditions Questionnaire. Participants who answered “yes” to the question “Has a doctor or other health professional ever told you that you have COPD?” were classified as having COPD. As this was a cross-sectional analysis of NHANES data, no matching procedures were applied, and no longitudinal follow-up was performed. The primary outcome was the presence of COPD. COPD status was determined based on self-reported physician diagnosis obtained from the NHANES Medical Conditions Questionnaire. Participants who answered affirmatively to the question “Has a doctor or other health professional ever told you that you have COPD?” were classified as having COPD. As spirometry data were not used for case definition in the present study, COPD was defined based on self-reported diagnosis. The primary exposure variable was the CTI, a composite biomarker reflecting IR and systemic inflammation. CTI was calculated using the following formula: where FPG and TGs were measured in milligrams per deciliter (mg/dL), and hs-CRP was measured in milligrams per liter (mg/L). Laboratory measurements were conducted under standardized protocols at NCHS-certified laboratories. CTI was analyzed both as a continuous variable and as a categorical variable divided into quartiles (Q1–Q4). Based on prior literature and biological plausibility, the following variables were considered potential confounders: age (continuous and categorical), sex (male and female), race/ethnicity (Mexican American, Other Hispanic, non-Hispanic White, non-Hispanic Black, and Other), educational level, marital status, poverty-to-income ratio (PIR), smoking status (never, former, and current), alcohol consumption status (never, former, and current), and physical activity level. These covariates were selected a priori based on established associations with COPD and metabolic–inflammatory status. We prespecified sex, age group, race/ethnicity, educational level, marital status, and PIR as potential effect modifiers. Interaction terms were tested in subgroup analyses to assess heterogeneity of associations across these groups. The CTI was calculated based on postprandial blood glucose (PBG), TGs, and CRP using the following formula: ln(FPG [mg/dL] × TG [mg/dL])/2 + 0.412 × ln(CRP [mg/L]). [ 12 ] All blood specimens were collected under standardized protocols requiring ≥8.5 hours of fasting prior to morning venipuncture, with samples transported to the NCHS-certified laboratories for analysis. Serum TG levels were quantified via enzymatic colorimetric assays using Roche Diagnostics modular biochemistry systems (Modular P and Cobas 6000), while FPG concentrations were determined by oxygen rate method on Beckman Coulter DxC800 automated analyzers. hs-CRP was measured through a latex-enhanced immunoturbidimetric methodology (Siemens Healthcare; historically CSL Behring). COPD diagnosis was ascertained based on self-reported physician confirmation from the standardized NHANES Medical Conditions Questionnaire, wherein participants responding affirmatively to the question “Has a doctor or other health professional ever told you that you have COPD?” were classified as COPD cases, and all others were classified as non-COPD controls. Consistent with prior studies, [ 16 – 18 ] we included clinically relevant covariates comprising demographic and lifestyle factors collected via self-administered questionnaires. This encompassed age, sex, race/ethnicity, educational attainment, marital status, family PIR (categorized as 3.50), smoking status (classified as current, former, or never smoker), alcohol consumption status (categorized as current, former, or never drinker), and physical activity levels. Further details regarding these covariates are accessible at https://www.cdc.gov/nchs/nhanes . All statistical analyses were performed using R version 4.4.3 (The R Foundation for Statistical Computing). Given the complex, multistage probability sampling design of NHANES, all primary analyses incorporated examination sampling weights, primary sampling units, and stratification variables in accordance with CDC analytic guidelines to obtain nationally representative estimates. [ 14 ] Baseline characteristics were summarized as weighted means (standard errors) for continuous variables and unweighted frequencies (weighted percentages) for categorical variables. Differences between COPD and non-COPD groups were assessed using weighted t tests for continuous variables and weighted chi-square tests for categorical variables. Weighted multivariable logistic regression models were constructed to estimate odds ratios (ORs) and 95% confidence intervals (CIs) for the association between CTI and COPD prevalence. CTI was analyzed both as a continuous variable and as quartiles. Three models were specified: Model 1: unadjusted; Model 2: adjusted for age, sex, and race/ethnicity; and Model 3: fully adjusted for age, sex, race/ethnicity, education level, marital status, PIR, smoking status, alcohol consumption, and physical activity. Confounding was addressed by including prespecified covariates selected a priori based on clinical relevance and prior literature. Trend tests across CTI quartiles were performed by modeling the median value of each quartile as a continuous variable. RCS regression with 3 knots placed at weighted percentiles of CTI was used to evaluate potential nonlinear associations between CTI and COPD prevalence. Nonlinearity was assessed using Wald tests. Prespecified subgroup analyses were conducted stratified by sex, age group, race/ethnicity, education level, marital status, and PIR. Interaction terms between CTI (continuous) and subgroup variables were included in fully adjusted models to formally test effect modification. Participants with missing exposure or outcome data were excluded from the primary analysis. To evaluate potential bias due to missing covariate data, multiple imputation by chained equations was performed, generating 5 imputed datasets. Regression analyses were repeated in the imputed datasets and pooled using Rubin rules. Several sensitivity analyses were conducted, including unweighted logistic regression models to assess the influence of sampling weights and analyses based on multiply imputed datasets. Consistency of effect estimates across sensitivity analyses was evaluated to assess robustness. All statistical tests were 2-sided, and a P value <.05 was considered statistically significant. Continuous variables, including CTI and age, were initially analyzed as continuous measures in regression models to preserve statistical information. CTI was further categorized into quartiles (Q1–Q4) based on its weighted distribution in the study population. Quartile categorization was applied to facilitate interpretation of effect estimates across increasing exposure levels and to assess potential nonlinear associations. Tests for trend across quartiles were conducted by assigning the median value of each quartile as a continuous variable in the regression models. In addition, RCS regression with 3 knots was used to explore potential dose–response relationships and assess nonlinearity between CTI and COPD prevalence. No additional transformation was applied to CTI in regression analyses, as it was already constructed using natural logarithmic transformations of its components. Other covariates were analyzed according to their original measurement scales or predefined categorical groupings based on clinical relevance and NHANES documentation. Several potential sources of bias were considered and addressed in the present study. First, to minimize selection bias and enhance generalizability, we used NHANES sampling weights and incorporated stratification and clustering variables according to CDC analytic guidelines. This approach accounts for the complex multistage sampling design and reduces the risk of biased population estimates. Second, exposure variables (FPG, TG, and hs-CRP) were measured using standardized laboratory protocols in NCHS-certified facilities, thereby reducing measurement bias. However, COPD status was based on self-reported physician diagnosis, which may introduce misclassification bias. To mitigate this limitation, we used a standardized questionnaire administered uniformly to all participants. Third, to reduce confounding bias, we adjusted for a broad range of demographic, socioeconomic, and lifestyle variables selected a priori based on clinical relevance and prior literature. Subgroup and interaction analyses were conducted to assess potential effect modification. Fourth, to address potential bias due to missing data, we performed sensitivity analyses using multiple imputation with chained equations and compared results with complete-case analyses. Finally, given the cross-sectional design, reverse causation cannot be excluded. Therefore, findings were interpreted cautiously, and causal inference was avoided. The study sample size was determined by the number of eligible participants available in the NHANES 2015 to 2018 cycles. As this was a secondary cross-sectional analysis of a nationally representative survey, no a priori sample size calculation was performed. Instead, all participants meeting the predefined eligibility criteria were included to maximize statistical power and representativeness. After applying exclusion criteria, a total of 3442 participants were included in the primary analysis. Given the large sample size and the weighted design, the study was considered adequately powered to detect modest associations between CTI and COPD prevalence. All data used in this study were obtained from publicly available NHANES 2015 to 2018 datasets. NHANES employs standardized data collection procedures, including structured interviews, physical examinations, and laboratory assessments conducted in MECs. FPG, TGs, and hs-CRP were measured using standardized laboratory protocols. Blood samples were collected during morning MEC sessions after a minimum fasting period of 8.5 hours. All specimens were processed at NCHS-certified laboratories following strict quality control procedures. Serum TG levels were quantified using enzymatic colorimetric methods on Roche modular biochemistry analyzers. FPG concentrations were measured using the oxygen rate method on a Beckman Coulter DxC800 automated analyzer. hs-CRP was determined using a latex-enhanced immunoturbidimetric assay. The CTI was calculated using these laboratory values as previously described. All laboratory assessments were conducted uniformly across participants, ensuring comparability of exposure measurements between COPD and non-COPD groups. COPD status was determined based on self-reported physician diagnosis collected through the NHANES Medical Conditions Questionnaire. The same standardized question was administered to all participants during in-home interviews, ensuring consistency of outcome ascertainment across groups. Demographic and socioeconomic variables (age, sex, race/ethnicity, education, marital status, and PIR) were obtained through structured household interviews. Smoking status, alcohol consumption, and physical activity were also self-reported using standardized NHANES questionnaires. Because all exposure, outcome, and covariate data were collected using uniform protocols within the same survey cycles, measurement methods were comparable across participant subgroups.

Figure 1 presents the participant selection process. A total of 19,225 individuals participated in the NHANES 2015 to 2018 cycles. Of these, 7937 individuals aged <20 years were excluded. We further excluded 125 pregnant women, 6477 participants with missing laboratory data required for CTI calculation (FPG, TGs, or hs-CRP), 6 individuals without COPD diagnostic information, and 1304 participants with incomplete covariate data. After applying these exclusion criteria, 3442 participants were included in the final analytic sample and were incorporated into all primary analyses. As this was a cross-sectional study without longitudinal follow-up, no additional losses to follow-up occurred after inclusion in the analytic sample. Table 1 presents the baseline characteristics of the study population. A total of 3442 participants were enrolled, with a mean age of 49.91 years. The gender distribution was balanced, comprising 1710 males and 1732 females. Compared to non-COPD participants, individuals with COPD exhibited the following features: older age, a higher proportion of females, increased representation of non-Hispanic Black ethnicity, greater prevalence of widowed/divorced/separated marital status, lower educational attainment, reduced household income levels, and higher rates of current smoking and drinking (all P < .05). Additionally, COPD patients demonstrated significantly elevated CTI levels and a higher proportion of participants in the highest CTI quartile (both P < .05). Baseline characteristics of the study population are presented in Table 1 . The final analytic sample included 3442 participants (weighted mean age: 49.91 years; 50.79% female). Compared with participants without COPD, those with COPD were older, more likely to be female, and had higher proportions of current smokers and lower socioeconomic status indicators (all P  < .05). CTI levels were significantly higher among participants with COPD (9.29 vs 8.95; P < .001). Regarding missing data, exposure variables (FPG, TG, and hs-CRP) were complete for all included participants by design, as individuals with missing CTI components were excluded during sample selection. Missing data for covariates were addressed through complete-case analysis in the primary model, and the number of participants excluded due to incomplete covariate data was 1304 (Fig. 1 ). A multiple imputation sensitivity analysis was additionally performed to assess the impact of missing covariate data. As this was a cross-sectional study without longitudinal follow-up, no follow-up time was applicable. Among the 3442 participants included in the final analytic sample, 134 individuals were classified as having COPD, corresponding to a weighted prevalence of 3.31%. The distribution of COPD cases across CTI quartiles is presented in Table 1 . The proportion of COPD increased across increasing CTI quartiles, with 13.55% of COPD cases in Q1, 13.20% in Q2, 31.25% in Q3, and 41.55% in Q4 ( P for trend = .010). These findings indicate a higher frequency of COPD among participants with elevated CTI levels. Descriptive baseline characteristics of participants based on COPD status. Values are presented as weighted means (SE) and unweighted frequencies (weighted percentages, %). COPD = chronic obstructive pulmonary disease, CTI = C-reactive protein–triglyceride–glucose index, PIR = poverty-to-income ratio, Q = quartile. Table 2 presents the association between CTI and COPD under different adjustment models. In the unadjusted model (Model 1), higher CTI levels were significantly associated with increased odds of COPD (continuous CTI: OR = 1.02, 95% CI: 1.01–1.02). After adjusting for age, sex, and race/ethnicity (Model 2), the association remained statistically significant. In the fully adjusted model (Model 3), which additionally controlled for education level, marital status, PIR, smoking status, alcohol consumption, and physical activity, CTI remained positively associated with COPD (continuous CTI: OR = 1.01, 95% CI: 1.00–1.01). When CTI was categorized into quartiles based on weighted percentiles, participants in the highest quartile (Q4) had higher odds of COPD compared with those in the lowest quartile (Q1) in the fully adjusted model (OR = 1.04, 95% CI: 1.02–1.07; P for trend = .019). Given the cross-sectional design and relatively low prevalence of COPD (3.31%), the observed OR of 1.04 corresponds to a modest absolute difference in COPD prevalence across CTI categories. Covariates included in the fully adjusted model were selected a priori based on established associations with both COPD and metabolic–inflammatory status in prior literature. Association between CTI and COPD, weighted. Model 1: no covariates were adjusted. Model 2: age, gender, and ethnicity were adjusted. Model 3: all covariates were adjusted. CI = confidence interval, COPD = chronic obstructive pulmonary disease, CTI = C-reactive protein–triglyceride–glucose index, OR = odds ratio, Q = quartile. The dose–response relationship between the CTI and COPD prevalence was analyzed using RCS regression (Fig. 2 ). After adjusting for all covariates, a significant linear association was observed between CTI levels and COPD risk (nonlinear P = .319). Restricted cubic spline regression analysis. Models were adjusted for age, gender, ethnicity, education level, marital status, poverty-to-income ratio, smoking status, and alcohol status. CI = confidence interval, CTI = C-reactive protein–triglyceride–glucose index, OR = odds ratio. Subgroup and interaction analyses were performed to assess potential effect modification of the association between CTI and COPD across predefined demographic and socioeconomic factors, including age group, sex, race/ethnicity, marital status, educational level, and PIR (Fig. 3 ). In fully adjusted models, a statistically significant interaction was observed between CTI (continuous) and sex ( P for interaction = .016). Stratified analyses indicated that the positive association between CTI and COPD was stronger among males compared with females. No statistically significant interactions were identified for age group, race/ethnicity, marital status, educational level, or PIR (all P for interaction > .05). Subgroup stratified forest plot and interaction P values for the association between continuous CTI and COPD across demographic and socioeconomic characteristics. COPD = chronic obstructive pulmonary disease, CTI = C-reactive protein–triglyceride–glucose index. Note: Age group, gender, ethnicity, education level, marital status, poverty-to-income ratio, smoking status, and alcohol status were all adjusted, except for the variable itself. The discriminatory performance of the CTI for COPD was evaluated using receiver operating characteristic (ROC) curve analysis in comparison with single blood biomarkers, including the TyG, PBG, TG, and CRP (Fig. 4 ). Results demonstrated that CTI yielded the highest area under the ROC curve (AUC) value (AUC = 0.630); nevertheless, the overall predictive performance remained suboptimal. ROC curve analysis of the discriminatory performance of CTI, TyG, CRP, PBG, and TG for COPD. AUC = area under the ROC curve, COPD = chronic obstructive pulmonary disease, CRP = C-reactive protein, CTI = C-reactive protein–triglyceride–glucose index, PBG = postprandial blood glucose, ROC = receiver operating characteristic, TG = triglyceride, TyG = triglyceride–glucose index. To evaluate the robustness of the primary findings, 2 sensitivity analyses were conducted (Tables 3 and 4 ). First, unweighted logistic regression models were applied to assess the influence of sampling weights. The association between CTI and COPD remained directionally consistent with the weighted analyses. In the fully adjusted unweighted model, participants in the highest CTI quartile had higher odds of COPD compared with those in the lowest quartile ( P for trend = .008). Second, multiple imputation by chained equations was performed to address potential bias due to missing covariate data. In the fully adjusted model based on the imputed dataset (n = 4746), the association between CTI and COPD remained statistically significant, with participants in the highest quartile showing modestly higher odds of COPD compared with those in the lowest quartile (OR = 1.02, 95% CI: 1.01–1.04). Overall, the magnitude and direction of the associations were generally consistent across sensitivity analyses. Association between CTI and COPD, unweighted. Model 1: no covariates were adjusted. Model 2: age, gender, and ethnicity were adjusted. Model 3: all covariates were adjusted. CI = confidence interval, COPD = chronic obstructive pulmonary disease, CTI = C-reactive protein–triglyceride–glucose index, OR = odds ratio, Q = quartile. Association between CTI and COPD based on the imputed data set (N = 4746). Model 1: no covariates were adjusted. Model 2: age, gender, and ethnicity were adjusted. Model 3: all covariates were adjusted. CI = confidence interval, COPD = chronic obstructive pulmonary disease, CTI = C-reactive protein–triglyceride–glucose index, OR = odds ratio, Q = quartile.

In this nationally representative cross-sectional study of U.S. adults, we evaluated the association between the CTI and COPD prevalence. Consistent with our prespecified objectives, we observed that higher CTI levels were independently associated with greater odds of COPD after adjustment for demographic, socioeconomic, and lifestyle factors. A linear dose–response relationship was identified, and a statistically significant interaction by sex suggested a relatively stronger association among males. In addition, CTI demonstrated slightly higher discriminatory ability compared with individual biomarkers, although overall predictive performance remained limited. Notably, subgroup analyses indicated that the association was more pronounced in males. Furthermore, ROC curve analysis confirmed CTI’s superior discriminatory power for COPD prediction over individual biomarkers, including the TyG, TG, PBG, and CRP, validating its unique value as a composite indicator for COPD risk assessment. Although several recent studies have reported separate associations of the TyG (reflecting IR) and CRP with COPD, this study is the first to demonstrate the superior predictive value of their synergistic biomarker, the CTI. [ 19 – 22 ] This finding extends the theoretical framework of CTI proposed by Ruan et al, [ 12 ] which was initially applied to predict cancer prognosis, successfully establishing its utility in the respiratory disease domain. Our results robustly support the synergistic interplay between inflammatory responses and IR in COPD pathogenesis, revealing multifaceted pathophysiological mechanisms underlying COPD development. Specifically, IR driven by dysregulated glucose and lipid metabolism activates oxidative stress pathways, leading to mitochondrial dysfunction and protease/antiprotease imbalance, thereby directly damaging airway endothelial cells. [ 23 , 24 ] Concurrently, chronic inflammation exacerbates insulin signaling dysfunction through cytokine-mediated cascades (e.g., TNF-α and IL-6), forming a self-perpetuating “metabolism-inflammation” vicious cycle. [ 25 , 26 ] This observation aligns with prior evidence: IR status significantly correlates with accelerated lung function decline in COPD patients, while elevated baseline CRP independently predicts COPD mortality. [ 8 , 27 , 28 ] As a dual-pathway integrator, CTI provides critical molecular epidemiological evidence elucidating the systemic metabolic–inflammatory mechanisms in COPD. This study observed a significantly stronger association between CTI and COPD in males. This sex-specific disparity may be primarily driven by graded differences in smoking behaviors and occupational exposures. Previous studies consistently demonstrate that male participants exhibit substantially higher rates of current smoking, and male smokers consume more cigarettes per day than females, resulting in significantly greater cumulative tobacco exposure. [ 29 , 30 ] Pathophysiologically, tobacco smoke toxins amplify systemic inflammation by activating pathways in alveolar macrophages. [ 31 ] Concurrently, nicotine-induced IR appears more pronounced in heavy smokers. [ 32 , 33 ] Furthermore, males have greater occupational representation in jobs with elevated exposure risks to airborne particulates or chemical vapors. [ 34 , 35 ] These synergistic environmental exposures likely potentiate the deleterious effects of CTI-quantified metabolic–inflammatory dysregulation on pulmonary function in males, thereby manifesting as an intensified CTI–COPD relationship. [ 36 ] Overall, our findings suggest that higher CTI levels are associated with a higher prevalence of COPD in U.S. adults, which is consistent with the study objective of evaluating a composite biomarker reflecting both metabolic dysregulation and systemic inflammation. This observation aligns with prior epidemiologic evidence linking IR-related indices (e.g., glucose-lipid–based markers) and inflammatory biomarkers (e.g., CRP) to impaired lung function and COPD-related outcomes. CTI integrates these 2 pathophysiological domains and may therefore capture a broader systemic phenotype relevant to COPD. However, the magnitude of association observed in this study was modest, and the overall discriminatory ability of CTI was limited. Given the cross-sectional nature of NHANES, the results should be interpreted as an association rather than evidence of causality, and reverse causation remains possible because COPD itself may contribute to chronic inflammation and metabolic disturbances. In addition, COPD was defined using self-reported physician diagnosis, which may introduce misclassification and could attenuate or distort effect estimates. We also acknowledge the multiplicity of analyses (e.g., quartile-based models, spline analyses, subgroup/interaction testing, and sensitivity analyses). Although results were generally directionally consistent, the sex interaction and subgroup patterns should be considered exploratory and require confirmation in independent cohorts. Taken together, these findings support a cautious interpretation that CTI may reflect a metabolic–inflammatory profile associated with COPD prevalence, but its clinical utility as a standalone screening tool is likely limited and warrants further validation in prospective studies with spirometry-defined COPD. Several limitations of this study should be acknowledged. First, due to the cross-sectional design, temporal relationships between CTI and COPD cannot be established. Reverse causation cannot be excluded, as systemic inflammation and metabolic alterations may occur secondary to COPD. This limitation may bias the observed association in either direction and precludes causal inference. Second, COPD status was defined based on self-reported physician diagnosis rather than spirometry-confirmed airflow limitation. Although this approach is commonly used in population-based studies, misclassification is possible. If nondifferential, such misclassification would likely attenuate the true association toward the null; however, differential misclassification cannot be completely excluded. Third, residual confounding may persist despite adjustment for multiple demographic, socioeconomic, and lifestyle variables. Factors such as environmental exposure, occupational hazards, or detailed smoking intensity were not fully captured, which could partially influence the observed effect estimates. Given the modest magnitude of association (OR ≈ 1.04), even small unmeasured confounding could materially impact the results. Fourth, missing covariate data resulted in the exclusion of a subset of participants in the primary analysis. Although multiple imputation analyses yielded similar results, selection bias due to incomplete data cannot be entirely ruled out. Finally, the discriminatory performance of CTI was limited (AUC = 0.630), suggesting that although statistically significant, its clinical utility as a standalone screening tool may be modest. This study was based on nationally representative NHANES data, incorporating sampling weights to reflect the noninstitutionalized civilian adult population of the United States. Therefore, the findings may be generalizable to the U.S. adult population aged ≥20 years. However, caution is warranted when extrapolating these results to other populations. Differences in genetic background, environmental exposures, healthcare access, smoking patterns, and metabolic risk profiles across countries may influence the relationship between CTI and COPD. In addition, institutionalized individuals and patients with severe comorbidities were not included in NHANES, which may limit applicability to hospitalized or high-risk clinical populations. Future prospective studies in diverse ethnic and geographic settings are needed to validate the external applicability of these findings.

This study demonstrates that CTI, a novel composite biomarker reflecting both IR and inflammatory status, is significantly and positively associated with the risk of COPD. Notably, this association is more pronounced in the male population. Furthermore, CTI exhibits superior predictive performance for COPD compared to traditional single biomarkers, offering an economical and accessible tool for early identification and risk stratification of high-risk individuals.

We sincerely thank Professor D.Z. for his guidance in this study. We also extend our gratitude to all participants for their contributions.