Association Between C-Reactive Protein-Triglyceride Glucose Index and All-Cause Mortality and Exacerbation in Adult Asthma: NHANES, 1999–2010 and 2015–2018

Abstract

Purpose

This study investigates the association between the C-reactive protein-triglyceride-glucose index (CTI), indicative of systemic inflammation and insulin resistance, and all-cause mortality as well as asthma exacerbations in adults with asthma.

Methods

Data from eight cycles of the US National Health and Nutrition Examination Survey (1999–2010 and 2015–2018) were analyzed. Adults aged ≥ 20 years were included, excluding those who were missing key variables. This resulted in 13,972 participants, of whom 1,832 had asthma. Participants were grouped into CTI tertiles. Participants self-reported exacerbations, and survival status was linked with the National Death Index. Complex survey weighting was applied to ensure national representativeness. Weighted multivariate logistic regression and Cox proportional hazards models assessed associations while adjusting for demographics, socioeconomic factors, body mass index, smoking, alcohol consumption, comorbidities, and laboratory indicators.

Results

The fully adjusted model revealed no significant association between CTI and asthma prevalence (odds ratio [OR], 1.07; 95% confidence interval [CI], 0.99–1.16; P = 0.078) but showed a strong association with all-cause mortality (hazard ratio [HR], 1.43; 95% CI, 1.12–1.81; P = 0.004). Patients in the highest CTI tertile exhibited a more than two-fold higher risk of mortality compared to those in the lowest tertile (HR, 2.22; 95% CI, 1.24–3.98; P = 0.007). Furthermore, higher CTI was associated with asthma exacerbation (OR, 1.23; 95% CI, 1.01–1.50; P = 0.039).

Conclusions

In this nationally representative sample, higher CTI was associated with higher all-cause mortality and higher odds of recent asthma exacerbations among adults with asthma, while no significant association with asthma prevalence was observed. CTI may help inform prognostic stratification in asthma and prospective studies are warranted.

INTRODUCTION

Asthma is a widespread chronic respiratory disease characterised by significant clinical heterogeneity. Its pathophysiological mechanisms involve complex interactions between systemic inflammation and metabolic abnormalities.¹,² According to the latest Global Burden of Disease statistics, there are now more than 350 million people worldwide living with asthma. This figure continues to rise, particularly in low- and middle-income countries.³ An acute asthma exacerbation can have a severe impact on patients’ quality of life. It also frequently results in visits to the emergency department and hospitalisations, placing a significant financial and resource burden on the healthcare system.³,⁴ The development of easily accessible, inexpensive, and predictive biomarkers is of great clinical and public health significance. The early identification of high-risk patients, the implementation of individualised treatment, and the prevention of acute episodes of disease are all possible with these biomarkers.

In recent years, growing evidence suggests that metabolic dysregulation and systemic inflammation jointly contribute to asthma pathophysiology, particularly in metabolic-related phenotypes such as obese asthma.⁵,⁶,⁷ Traditional biomarkers, including C-reactive protein (CRP) and triglyceride-glucose (TyG), have been used to reflect inflammatory and metabolic status, respectively; however, each has limitations when used individually.⁸ CRP may be influenced by acute inflammatory events,⁹ while TyG does not capture systemic immune activation.¹⁰,¹¹,¹²,¹³,¹⁴ To overcome these limitations, C-reactive protein-triglyceride-glucose composite index (CTI) has been proposed as an integrated biomarker capable of summarizing both metabolic and inflammatory burden.¹⁵,¹⁶,¹⁷,¹⁸,¹⁹,²⁰,²¹ Despite its theoretical advantages, the prognostic value of CTI in adults with asthma remains unclear. Therefore, we aimed to evaluate the association of CTI with asthma prevalence, exacerbations, and all-cause mortality in a nationally representative cohort.

Therefore, this research plan will utilise nationally representative data from the National Health and Nutrition Examination Survey (NHANES), combining multiple survey cycles from 1999–2010 and 2015–2018. The aim is to assess the association between CTI levels in adult asthma patients and the risk of all-cause mortality and exacerbation status. This will provide a basis for formulating individualised asthma management strategies in the future.

MATERIALS AND METHODS

Source of data

NHANES is a nationally representative research programme conducted to assess the health and nutritional status of adult and paediatric populations in the United States. This survey systematically collects data on demographics, socioeconomic characteristics, dietary habits and various health-related metrics, thereby providing substantial support for epidemiological studies and the formulation of public health policy. De-identified NHANES data are publicly accessible for use by researchers and the general public. Further information is available at www.cdc.gov/nchs/nhanes/. The NHANES study protocol was reviewed and approved by the Ethics Review Board of the National Center for Health Statistics (NCHS). Ethical approval for the data collection periods analyzed in this study corresponds to Protocol #98-12, Protocol #2005-06, Continuation of Protocol #2011-17, and Protocol #2018-01. Furthermore, informed consent was obtained from all participants prior to their inclusion in the survey during the data collection phases. This study was conducted in full compliance with the strict data usage and confidentiality guidelines established by the National Center for NCHS for the use of NHANES data. This study was conducted in full compliance with the strict data usage and confidentiality guidelines established by the NCHS for the use of NHANES data.

Participants

This study utilized a merged dataset derived from eight cycles of the NHANES, covering the periods 1999–2010 and 2015–2018. The survey cycles from 2011 to 2014 were excluded from this analysis because quantitative data for CRP, an essential component for calculating the CTI, were not available during this period. The initial dataset comprised 81,385 participants. As the study focused on adults, participants aged under 20 years (n = 37,633) were excluded from the initial analysis. To ensure the integrity of the primary exposure data and the validity of the complex survey weighting, listwise deletion (complete case analysis) was applied to handle missing data. More specifically, 25,209 participants were excluded due to missing values for key variables, including asthma status, fasting plasma glucose (FPG), fasting triglycerides (TGs), CRP, and survival covariates. Furthermore, 4,571 participants were excluded due to missing data on sample weights or other essential covariates. Consequently, the final cohort comprised 13,972 participants, consisting of 1,832 individuals with asthma and 12,140 without asthma. Among the asthma group, 517 participants were identified as having asthma exacerbations. All procedures for participant exclusion and cohort establishment were conducted in strict accordance with the pre-defined flowchart (Fig. 1), with the aim of enhancing the reproducibility of the study and minimising selection bias.

Fig. 1. Flowchart of participant selection and exclusion.

NHANES, National Health and Nutrition Examination Survey.

Definition of asthma

Asthma status was determined based on self-reported data from the NHANES Medical Conditions Questionnaire. Participants were identified as having asthma if they answered “Yes” to the question: “Has a doctor or other health professional ever told you that you have asthma?” (Question 1). Consequently, the reported prevalence represents the lifetime prevalence (‘Ever Asthma’). Among those with diagnosed asthma, disease stability was further stratified based on the occurrence of acute attacks in the previous 12 months, assessed by the question: “Have you had an asthma attack in the past year?” (Question 2). Participants who responded “No” to Question 2 were classified as having stable asthma, while those who responded “Yes” were classified as having unstable asthma (asthma exacerbation). This classification identifies individuals who experienced acute exacerbation episode in the preceding year. It should be acknowledged that this classification is derived from self-reported data and may therefore be subject to recall or reporting bias. However, this approach has been widely used in multiple NHANES studies and is substantiated by existing literature, supporting its suitability for identifying asthma stability and acute exacerbation within the scope of this investigation.²²,²³,²⁴

Definition of CTI

CTI was utilized as a composite metric for the evaluation of systemic inflammation and insulin resistance, and was calculated according to the following formula: CTI = 0.412 × ln(CRP) + ln([TG × FPG]/2). The abbreviation “ln” is shorthand for the natural logarithm. CRP concentrations are expressed in mg/L, while TG and FPG levels are reported in mg/dL.¹⁶,²⁵ The CTI values for all participants were stratified into tertiles (T1, T2, T3) based on the overall distribution. The first tertile (T1), comprising participants with the lowest CTI values, was designated as the reference group to assess the associations between CTI levels, study outcomes, and any potential dose-response relationship.

Follow-up and endpoints

Data pertaining to the vital status of participants, the date of their demise, and the underlying cause of death were obtained through a linkage process between the NHANES and the National Death Index (NDI). The follow-up period for each participant was defined as the time from the baseline examination until the date of death or the study cutoff date (31 December 2019), whichever occurred first. In accordance with the foregoing definition, the calculation of follow-up time was undertaken, with all-cause mortality constituting the primary study endpoint. All-cause mortality was defined as death from any cause, as identified by the corresponding code (010) in NDI.

Study covariates

The selection of covariates was informed by clinical relevance and previous literature, with the objective of minimising potential confounding. Baseline covariates were obtained from standardized questionnaires, physical examinations, and laboratory tests conducted within the NHANES protocol. Key demographic and socioeconomic covariates were incorporated as follows: age was treated as a continuous variable (years); sex and race/ethnicity were categorical, with the latter comprising Mexican American, Other Hispanic, non-Hispanic White, non-Hispanic Black, and Other categories. Socioeconomic status was assessed using the poverty income ratio (PIR). This continuous measure was categorized into three groups: low (PIR ≤ 1.0), middle (PIR 1.1–3.0), and high (PIR > 3.0). Educational attainment was categorized into a binary variable: “below university level” and “university level or above.” The body mass index (BMI) was calculated as the weight in kilograms divided by the height in kg/m². The categorisation of BMI was conducted in accordance with the standard clinical guidelines, which delineate the following categories: normal or underweight (< 25), overweight (25.0–30.0), and obese (≥ 30.0). Smoking status was operationalized based on lifetime smoking history and current use, defining three mutually exclusive categories: never smokers (individuals who had smoked ≤ 99 cigarettes in their lifetime); former smokers (those who had smoked ≥ 100 cigarettes but were not smoking at the time of assessment); and current smokers (those who had smoked ≥ 100 cigarettes and reported current smoking). The consumption of alcohol was divided into five distinct categories: never drinkers (< 12 lifetime drinks), former drinkers (≥ 12 lifetime drinks but none in the past year), and current drinkers—further classified as light (≤ 1 female/≤ 2 male daily drinks), moderate (1–3 female/2–4 male daily drinks), or heavy (≥ 4 female/≥ 5 male daily drinks) based on past-year intake. These categorizations allow for clear differentiation of exposure levels while accounting for sex-specific differences. The clinical comorbidities that were assessed—namely hypertension, diabetes mellitus (DM), cancer, cardiovascular disease (CVD), congestive heart failure (CHF), stroke, and emphysema-were based on self-reported physician diagnoses. This approach is consistent with previous NHANES investigations and facilitates direct comparability. The following laboratory parameters were quantified by the NHANES central laboratory: high-density lipoprotein (HDL), white blood cell count (WBC), hemoglobin (Hb), platelet count (PLT), alanine aminotransferase (ALT), total bilirubin (TBIL), uric acid (UA), creatinine (Cr), serum sodium (Na+), and serum potassium (K+). Laboratory parameters were primarily analyzed as continuous variables, incorporated into models either on their original scale or following appropriate transformation to meet statistical assumptions. Categorical variables were encoded according to their predefined groupings. To ensure the robustness of the statistical models, we assessed potential multicollinearity among these covariates using the Generalized Variance Inflation Factor (GVIF). As presented in Supplementary Tables S1, 2, S3, all GVIF values were less than 2, indicating that multicollinearity was negligible and did not compromise model stability.

Statistical analysis

All statistical analyses were conducted accounting for the complex multistage probability sampling design of NHANES. To ensure national representativeness and correct for non-response bias in the fasting subsample (required for CTI calculation), we utilized the Fasting Subsample Weights.²⁶ Data were combined from eight survey cycles: 1999–2000, 2001–2002, 2003–2004, 2005–2006, 2007–2008, 2009–2010, 2015–2016, and 2017–2018. New sample weights were constructed for the combined dataset in accordance with NCHS analytical guidelines. Specifically, for participants in the 1999–2002 period (which used 4-year weights), the weights were rescaled by multiplying by 2/8; for participants in the remaining six cycles (which used 2-year weights), the weights were divided by 8. The complex survey design was specified using the recalculated weights, strata (SDMVSTRA), and primary sampling units (SDMVPSU). Continuous variables were summarized as weighted means with standard errors for normally distributed variables, or as weighted medians with interquartile ranges for variables with skewed distributions. Categorical variables were presented as weighted counts and weighted percentages. Between-group comparisons were performed using weighted linear regression for continuous variables and the Rao-Scott χ² test for categorical variables. CTI was analyzed both as a continuous variable and tertiles (T1–T3), with T1 serving as the reference. Asthma prevalence was estimated as the weighted proportion of participants reporting physician-diagnosed asthma. All prevalence estimates incorporated the NHANES examination sample weights, strata, and primary sampling units to account for the complex survey design and to produce nationally representative estimates of the U.S. adult population. To assess the relationships of CTI with asthma prevalence, acute asthma exacerbation, and all-cause mortality, we conducted stratified regression analyses using the survey package in R (R Foundation for Statistical Computing, Vienna, Austria). Weighted multivariable logistic regression models (svyglm) were applied for prevalence and exacerbation outcomes, while weighted Cox proportional hazards models (svycoxph) were used for all-cause mortality. For each outcome, three progressively adjusted models were constructed: an unadjusted model (Model 1); a model adjusted for age and sex (Model 2); and a fully adjusted model (Model 3). Model 3 included adjustments for race/ethnicity, educational attainment, PIR, BMI, smoking status, alcohol consumption, comorbidities (hypertension, DM, cancer, CVD, CHF, stroke, and emphysema), and laboratory biomarkers (HDL, WBC, Hb, PLT, ALT, TBIL, UA, Cr, Na, and K). Potential multicollinearity among covariates was assessed using the GVIF, with values < 2 considered acceptable. To evaluate potential nonlinear dose-response relationships, Restricted Cubic Splines (RCSs) with three knots (at the 10th, 50th, and 90th percentiles) were employed. For survival analysis, unadjusted survival probabilities across CTI tertiles were visualized using Kaplan-Meier curves, and differences were assessed using the log-rank test. Subgroup analyses were performed to evaluate the consistency of associations across key demographic and clinical strata (age, sex, smoking status, BMI, education, PIR, and major comorbidities). Analysis was performed using R 4.2.2 (http://www.Rproject.org) and the Free Statistics software (version 2.3; Beijing FreeClinical Medical Technology Co., Ltd, Beijing, China). A two-tailed P value of less than 0.05 was considered statistically significant.

RESULTS

Participant characteristics

The final analytical cohort included 13,972 participants, representing an estimated weighted population of approximately 177.6 million U.S. adults after applying NHANES sampling weights. Within this weighted population, the prevalence of physician-diagnosed asthma was 14.1%, corresponding to approximately 24.9 million individuals nationwide. As shown in Table 1, significant differences in baseline characteristics were observed between participants with and without asthma. Compared to participants without asthma, those with asthma had a higher proportion of females (58.3% vs. 50.0%, P < 0.001), a lower mean age (weighted mean age 44.4 vs. 46.6 years, P < 0.001), a higher mean BMI (29.9 vs. 28.5 kg/m², P < 0.001), and a different ethnic distribution, characterized by a lower proportion of non-Hispanic Whites (4.3% vs. 8.3%). Patients with asthma had a significantly higher burden of comorbidities, including hypertension (27.9% vs. 23.6%), diabetes (9.2% vs. 7.7%), CVD (6.1% vs. 4.7%), CHF (4.1% vs. 1.6%), stroke (4.0% vs. 2.2%), cancer (11.2% vs. 8.3%), and emphysema (5.7% vs. 1.1%) (all P < 0.05). This was characterized by elevated white blood cell (7.09 vs. 6.74 ×10⁹/L) and PLTs (261.6 vs. 254.8 ×10⁹/L), lower Hb levels (14.32 vs. 14.49 g/dL, P < 0.001), and a higher median CRP level (2.40 vs. 1.80 mg/L, P < 0.001). Although the TyG index did not differ between the groups (P = 0.094), the CTI was higher in patients with asthma (9.01 vs. 8.86, P < 0.001). Although the absolute difference in mean CTI between groups was small, it reached statistical significance in this large sample.

Table 1. Baseline characteristics of adults in National Health and Nutrition Examination Survey 1999–2010 and 2015–2018.

Characteristics		Total (n = 177624699.0)	Asthma		P value
			No (n = 152,657,856.0)	Yes (n = 24,966,843.0)
Sex					< 0.001
	Male	86,757,536.6 (48.8)	76,341,454.8 (50.0)	10,416,081.8 (41.7)
	Female	90,867,162.4 (51.2)	76,316,401.3 (50.0)	14,550,761.2 (58.3)
Age (yr)		46.31 ± 16.53	46.63 ± 16.56	44.37 ± 16.20	< 0.001
Race/ethnicity					< 0.001
	Non-Hispanic White	13,779,951.2 (7.8)	12,708,891.9 (8.3)	1,071,059.3 (4.3)
	Non-Hispanic Black	8,632,826.8 (4.9)	7,199,045.2 (4.7)	1,433,781.6 (5.7)
	Mexican American	126,244,362.1 (71.1)	108,453,193.9 (71.0)	17,791,168.2 (71.3)
	Other race—including multi-racial	18,361,898.4 (10.3)	15,358,424.3 (10.1)	3,003,474.1 (12.0)
	Other Hispanic	10,605,660.4 (6.0)	8,938,300.7 (5.9)	1,667,359.7 (6.7)
Education level					0.458
	Below college	73,140,730.9 (41.2)	63,138,309.6 (41.4)	10,002,421.3 (40.1)
	Above college	104,483,968.1 (58.8)	89,519,546.5 (58.6)	14,964,421.6 (59.9)
BMI (kg/m2)		28.68 ± 6.53	28.49 ± 6.36	29.86 ± 7.39	< 0.001
Alcohol consumption					0.002
	Never	19,157,454.5 (10.8)	17,268,154.0 (11.3)	1,889,300.5 (7.6)
	Former	25,726,210.4 (14.5)	21,703,351.3 (14.2)	4,022,859.1 (16.1)
	Mild	65,499,546.1 (36.9)	56,490,202.0 (37.0)	9,009,344.1 (36.1)
	Moderate	29,992,841.0 (16.9)	25,473,061.2 (16.7)	4,519,779.8 (18.1)
	Heavy	37,248,647.0 (21.0)	31,723,087.5 (20.8)	5,525,559.4 (22.1)
Smoking status					< 0.001
	Never smoker	92,961,648.3 (52.3)	81,414,368.6 (53.3)	11,547,279.7 (46.3)
	Former smoker	46,210,459.9 (26.0)	38,973,125.3 (25.5)	7,237,334.6 (29.0)
	Current smoker	38,452,590.7 (21.6)	32,270,362.1 (21.1)	6,182,228.7 (24.8)
Family PIR		3.07 ± 1.61	3.10 ± 1.60	2.90 ± 1.65	0.001
Hypertension					0.001
	No	134,694,787.0 (75.8)	116,684,295.7 (76.4)	18,010,491.2 (72.1)
	Yes	42,929,912.0 (24.2)	35,973,560.3 (23.6)	6,956,351.7 (27.9)
DM					0.090
	No	163,620,418.6 (92.1)	140,939,674.5 (92.3)	22,680,744.1 (90.8)
	Yes	14,004,280.4 (7.9)	11,718,181.5 (7.7)	2,286,098.9 (9.2)
Cancer					0.003
	No	162,232,614.2 (91.3)	140,054,101.6 (91.7)	22,178,512.7 (88.8)
	Yes	15,392,084.7 (8.7)	12,603,754.5 (8.3)	2,788,330.3 (11.2)
CVD					< 0.024
	No	168,908,337.8 (95.1)	145,474,423.9 (95.3)	23,433,913.9 (93.9)
	Yes	8,716,361.2 (4.9)	7,183,432.2 (4.7)	1,532,929.1 (6.1)
CHF					< 0.001
	No	174,113,558.3 (98.0)	150,158,777.8 (98.4)	23,954,780.5 (95.9)
	Yes	3,511,140.7 (2.0)	2,499,078.3 (1.6)	1,012,062.5 (4.1)
Stroke					< 0.001
	No	173,203,434.2 (97.5)	149,245,024.0 (97.8)	23,958,410.2 (96.0)
	Yes	4,421,264.8 (2.5)	3,412,832.0 (2.2)	1,008,432.7 (4.0)
Emphysema					< 0.001
	No	174,584,589.1 (98.3)	151,041,256.5 (98.9)	23,543,332.5 (94.3)
	Yes	3,040,109.9 (1.7)	1,616,599.5 (1.1)	1,423,510.4 (5.7)
HDL		53.73 ± 15.74	53.78 ± 15.77	53.38 ± 15.52	0.428
WBC		6.79 ± 1.92	6.74 ± 1.89	7.09 ± 2.08	< 0.001
Hb		14.47 ± 1.46	14.49 ± 1.46	14.32 ± 1.43	< 0.001
PLT		255.78 ± 65.15	254.83 ± 64.88	261.59 ± 66.52	0.001
ALT		21.00 (16.00–29.00)	21.00 (16.00–29.00)	20.00 (16.00–28.00)	0.099
TBIL		12.04 ± 5.13	12.17 ± 5.16	11.25 ± 4.86	< 0.001
UA		324.46 ± 82.27	324.75 ± 82.28	322.63 ± 82.21	0.427
Cr		75.97 ± 19.77	76.23 ± 19.85	74.38 ± 19.24	0.001
Na		139.20 ± 2.24	139.24 ± 2.23	138.97 ± 2.29	< 0.001
K		4.06 ± 0.33	4.07 ± 0.33	4.04 ± 0.33	0.031
Glucose		104.19 ± 28.84	104.08 ± 28.50	104.84 ± 30.88	0.389
Triglyceride		133.37 ± 116.19	132.63 ± 114.96	137.89 ± 123.39	0.179
CRP		1.90 (0.80–4.40)	1.80 (0.71–4.20)	2.40 (0.98–5.80)	< 0.001
TyG		8.63 ± 0.66	8.63 ± 0.66	8.66 ± 0.67	0.094
CTI		8.88 ± 0.94	8.86 ± 0.94	9.01 ± 0.98	< 0.001
CTI					0.001
	T1	66,603,405.7 (37.5)	58,148,633.0 (38.1)	8,454,772.7 (33.9)
	T2	57,969,059.1 (32.6)	49,927,968.8 (32.7)	8,041,090.3 (32.2)
	T3	53,052,234.2 (29.9)	44,581,254.2 (29.2)	8,470,980.0 (33.9)
Status					0.985
	Alive	156,766,785.1 (88.3)	134,735,926.4 (88.3)	22,030,858.7 (88.2)
	Dead	20,857,913.9 (11.7)	17,921,929.7 (11.7)	2,935,984.2 (11.8)
OS		132.54 ± 71.34	133.29 ± 71.40	127.93 ± 70.81	0.037

Data are presented as mean ± standard error, median (interquartile range) or number (%).

P values were calculated using analysis of variance for normally distributed continuous variables, Kruskal-Wallis test for non-normally distributed continuous variables, and chi-square or Fisher’s exact test for categorical variables, as appropriate. P values indicate differences across all people with or without asthma.

BMI, body mass index; PIR, poverty income ratio; DM, diabetes mellitus; CVD, cardiovascular diseases; CHF, congestive heart failure; HDL, high-density lipoprotein; WBC, white blood cell count; Hb, hemoglobin; PLT, platelet count; ALT, alanine aminotransferase; TBIL, total bilirubin; UA, uric acid; Cr, creatinine; Na+, serum sodium; K+, serum potassium; CRP, C-reactive protein; TyG, triglyceride-glucose; CTI, C-reactive protein-triglyceride-glucose composite index; OS overall survival.

The association between CTI and asthma

Clinical characteristics by CTI tertiles

To further characterize the clinical phenotype associated with elevated CTI, we stratified the asthma population by CTI tertiles (Supplementary Table S4). A distinct metabolic-inflammatory profile was observed as CTI levels increased. Patients in the highest tertile (T3) were significantly older (mean age 48.8 vs. 40.1 years) and had a substantially higher BMI (33.8 vs. 25.8 kg/m²) compared to those in the lowest tertile (T1). The burden of comorbidities accumulated progressively with CTI levels; specifically, the prevalence of hypertension and DM increased from 16.5% and 2.7% in T1 to 39.8% and 19.4% in T3, respectively (both P < 0.001). Moreover, markers of systemic inflammation showed significant upward trends across tertiles. Crucially, crude survival outcomes differed markedly: the unadjusted mortality rate rose sharply from 5.1% in the lowest tertile to 20.1% in the highest tertile (P < 0.001).

CTI and the prevalence of asthma

The association between CTI and the prevalence of asthma was evaluated using weighted multivariate logistic regression (Table 2). A significant positive correlation was observed between CTI and asthma risk in both unadjusted and partially adjusted models. Specifically, in Model 1 and Model 2, odds ratios (Ors) per unit increase in CTI were 1.17 (95% confidence interval [CI], 1.10–1.24; P < 0.001) and 1.23 (95% CI, 1.16–1.31, P < 0.001), respectively. Furthermore, when the tertiles were compared, a significantly higher risk was observed in the T3 group than in the T1 group in both Model 1 and Model 2. The corresponding OR in Model 1 was 1.31 (95% CI, 1.15–1.48; P = 0.001), while in Model 2, it was 1.43 (95% CI, 1.26–1.63; P < 0.001). However, in Model 3, for which full adjustment was performed for potential confounders including ethnicity, PIR, educational attainment, BMI, smoking status, alcohol consumption, major comorbidities and laboratory indicators, the association was significantly reduced and was no longer statistically significant. Specifically, the OR per unit increase in CTI was 1.07 (95% CI, 0.99–1.16; P = 0.078), whereas the OR for the T3 group compared to the T1 group was 1 (95% CI, 0.83–1.20; P = 0.993).

Table 2. Weighted odds ratios (95% confidence intervals) for the prevalence of asthma according to the tripartite of CTI among adults in National Health and Nutrition Examination Survey 1999–2000 and 2015–2018.

CTI		Model 1	P value	Model 2	P value	Model 3	P value
Continuous		1.17 (1.10–1.24)	< 0.001	1.23 (1.16–1.31)	< 0.001	1.07 (0.99–1.16)	0.078
Tertiles
	T1	1 (Ref)		1 (Ref)		1 (Ref)
	T2	1.11 (0.97–1.27)	0.134	1.21 (1.05–1.38)	0.008	1.02 (0.86–1.2)	0.852
	T3	1.31 (1.15–1.48)	< 0.001	1.43 (1.26–1.63)	< 0.001	1 (0.83–1.2)	0.993
P trend			< 0.001		< 0.001		0.992

Model 1: unadjusted analysis.

Model 2: adjusted for age, sex.

Model 3: adjusted for sex, age, race/ethnicity, educational attainment, poverty income ratio, body mass index, smoking status, alcohol consumption, hypertension, diabetes mellitus, cancer, cardiovascular diseases, congestive heart failure, stroke, and emphysema, high-density lipoprotein, white blood cell count, hemoglobin, platelet count, alanine aminotransferase, total bilirubin, uric acid, creatinine, serum sodium, and serum potassium.

CTI, C-reactive protein-triglyceride-glucose composite index.

The association between CTI and all-cause mortality of asthma

We utilized weighted multivariate Cox proportional hazards models to assess the prognostic value of CTI for all-cause mortality among adults with asthma (Table 3).

Table 3. Weighted multiple Cox regression analysis of CTI in patients with asthma.

CTI		Model 1	P value	Model 2	P value	Model 3	P value
Continuous		1.67 (1.44–1.92)	< 0.001	1.45 (1.23–1.71)	< 0.001	1.43 (1.12–1.81)	0.004
Tertiles
	T1	1 (Ref)		1 (Ref)		1 (Ref)
	T2	2.15 (1.38–3.37)	< 0.001	1.44 (0.94–2.22)	0.095	1.28 (0.78–1.67)	0.331
	T3	3.96 (2.57–6.1)	< 0.001	2.25 (1.51–3.36)	< 0.001	2.22 (1.24–3.98)	0.007
P trend			< 0.001		< 0.001		0.003

Model 1: unadjusted analysis.

Model 2: adjusted for age, sex.

Model 3: adjusted for sex, age, race/ethnicity, educational attainment, poverty income ratio, body mass index, smoking status, alcohol consumption, hypertension, diabetes mellitus, cancer, cardiovascular diseases, congestive heart failure, stroke, and emphysema, high-density lipoprotein, white blood cell count, hemoglobin, platelet count, alanine aminotransferase, total bilirubin, uric acid, creatinine, serum sodium, and serum potassium.

CTI, C-reactive protein-triglyceride-glucose composite index.

The results showed that higher CTI levels were significantly linked to an increased risk of death, and this association remained consistent when adjusting for various factors: in the unadjusted model (Model 1), the hazard ratio (HR) for a one-unit increase in CTI was 1.67 (95% CI, 1.44–1.92; P < 0.001); after adjusting for age and sex (Model 2), the HR was 1.45 (95% CI, 1.23–1.71; P < 0.001); and in the fully adjusted model (Model 3), the HR was 1.43 (95% CI, 1.12–1.81; P = 0.004). Following full adjustment (Model 3), the highest tertile (T3) exhibited a significantly elevated risk of mortality compared with the lowest tertile (T1) (HR, 2.22; 95% CI, 1.24–3.98; P = 0.007), whereas the middle tertile (T2) demonstrated no statistically significant difference. Furthermore, the trend test based on tertiles revealed a significant dose-response relationship (Model 3 P for trend = 0.003).

Using RCSs based on the fully adjusted model, we characterized the dose-response association between continuous CTI and all-cause mortality (Fig. 2A-C). The analysis revealed no evidence of a non-linear relationship (P for non-linearity > 0.05). Consequently, the curve exhibited a linear trajectory, indicating that the risk of all-cause mortality increases monotonically with rising CTI levels (P for overall < 0.05). This finding reinforces the results obtained from the weighted Cox proportional hazards models (svycoxph).

Fig. 2. Dose-response association between CTI and all-cause mortality and asthma exacerbation risk in adults with asthma using restricted cubic spline regression. Analyses were performed using three sequential models: Model 1, unadjusted; Model 2, adjusted for age and sex; Model 3, additionally adjusted for race/ethnicity, educational attainment, poverty income ratio, body mass index, smoking status, alcohol consumption, hypertension, diabetes mellitus, cancer, cardiovascular diseases, congestive heart failure, stroke, and emphysema, high-density lipoprotein, white blood cell count, hemoglobin, platelet count, alanine aminotransferase, total bilirubin, uric acid, creatinine, serum sodium, and serum potassium. (A-C) Association of CTI with all-cause mortality in Model 1 (A), Model 2 (B), and Model 3 (C). (D-F) Association of CTI with asthma exacerbation risk in Model 1 (D), Model 2 (E), and Model 3 (F).

OR, odds ratio; CI, confidence interval; CTI, C-reactive protein-triglyceride-glucose composite index; HR, hazard ratio.

Fig. 3 presents the Kaplan-Meier survival curves stratified by CTI tertiles. The log-rank test indicated a statistically significant difference in survival probabilities among the groups (P < 0.001), with the highest CTI tertile showing markedly poorer unadjusted survival outcomes over the follow-up period.

Fig. 3. Kaplan-Meier survival analysis of the association between CTI levels and all-cause mortality in adults with asthma. Kaplan-Meier curves illustrate the cumulative survival probability stratified by CTI tertiles. The log-rank test demonstrated a significant difference in mortality risk across CTI groups (P < 0.05), with higher CTI levels associated with reduced survival.

CTI, C-reactive protein-triglyceride-glucose composite index.

Subgroup analyses were conducted to evaluate the robustness of the association between the CTI index and mortality risk across varying populations (Fig. 4). We stratified participants by key demographic and clinical characteristics, including sex, age (< 60 vs. ≥ 60 years), BMI (< 25 vs. 25–30 vs. ≥ 30 kg/m²), education level (below university vs. university level or above), smoking status (never smokers vs. former smokers vs. current smokers), family PIR (< 1.1 vs. 1.1–3.0 vs. ≥ 3.0), and comorbidities (hypertension, DM, cancer, CVD, CHF, emphysema, and stroke). Interaction tests demonstrated no statistically significant modification of the effect by any of these stratifications (all P for interaction > 0.05). Interaction tests showed no statistically significant effect modification across these stratifications (all P for interaction > 0.05).

Fig. 4. Subgroup analysis of C-reactive protein-triglyceride-glucose composite index and all-cause mortality association in adults with asthma. Forest plot displays HRs across patient subgroups using fully adjusted Model 3, with comprehensive adjustment for sex, age, race/ethnicity, educational attainment, PIR, BMI, smoking status, alcohol consumption, hypertension, DM, cancer, CVD, CHF, stroke, and emphysema, high-density lipoprotein, white blood cell count, hemoglobin, platelet count, alanine aminotransferase, total bilirubin, uric acid, creatinine, serum sodium, and serum potassium.

HR, hazard ratio; CI, confidence interval; BMI, body mass index; PIR, poverty income ratio; DM, diabetes mellitus; CVD, cardiovascular diseases; CHF, congestive heart failure.

The association between CTI and asthma exacerbation

The relationship between CTI levels and the risk of asthma exacerbation (defined as individuals experiencing ≥ 1 acute attack in the past year) was evaluated using weighted logistic regression (Table 4). In the fully adjusted model (Model 3), elevated CTI levels were significantly associated with an increased likelihood of exacerbation. Specifically, a one-unit increase in CTI was associated with a 23% higher risk of reporting an asthma attack (OR, 1.23; 95% CI, 1.01–1.50, P = 0.039). As shown in Fig. 2D-F, RCS analysis demonstrated a linear dose–response relationship between CTI and the risk of exacerbation (P for overall < 0.05; P for non-linearity > 0.05).

Table 4. Weighted multifactorial logistic regression analysis of CTI in patients with stable asthma or asthma exacerbation.

CTI	OR (95%CI)	P value
Model 1	1.25 (1.09–1.43)	0.002
Model 2	1.29 (1.12–1.48)	< 0.001
Model 3	1.23 (1.01–1.5)	0.039

Model 1: unadjusted analysis.

Model 2: adjusted for age, sex.

Model 3: adjusted for sex, age, race/ethnicity, educational attainment, poverty income ratio, body mass index, smoking status, alcohol consumption, hypertension, diabetes mellitus, cancer, cardiovascular diseases, congestive heart failure, stroke, and emphysema, high-density lipoprotein, white blood cell count, hemoglobin, platelet count, alanine aminotransferase, total bilirubin, uric acid, creatinine, serum sodium, and serum potassium.

CTI, C-reactive protein-triglyceride-glucose composite index; OR, odds ratio; CI, confidence interval.

DISCUSSION

This study used the large-scale, nationally representative NHANES database to systematically examine the associations between CTI (a biomarker integrating systemic inflammation and metabolic parameters) and all-cause mortality and acute exacerbation risk in adult asthmatic patients for the first time. After adjusting for multiple potential confounders, no significant association was found between CTI levels and asthma prevalence. However, elevated CTI levels were found to be significantly associated with an increased risk of all-cause mortality and acute exacerbation in patients with asthma. Further validation of these findings was provided by Kaplan-Meier survival analysis with the log-rank test, which showed that individuals with higher CTI levels had significantly poorer survival outcomes. Although the absolute differences in CTI levels between groups were relatively small, these differences reached statistical significance in this large population, which may reflect a shift toward a higher inflammatory–metabolic burden among individuals with asthma. These results suggest that CTI has potential as a prognostic biomarker in clinical practice, assisting in the guidance of treatment decisions and the monitoring of disease progression.

In recent years, increasing attention has been given to the roles of inflammation and metabolism in asthma pathogenesis. Asthma is increasingly recognized as a complex “immuno-metabolic” syndrome where inflammation and metabolic dysregulation form a bidirectional, self-perpetuating loop. This inflammation involves multiple cell types, such as eosinophils, mast cells and Th2 lymphocytes, as well as their secreted cytokines, including interleukin (IL)-4, IL-5 and IL-13. These mediators facilitate the persistence of airway inflammation and critically influence the development of airway smooth muscle hyperresponsiveness and structural remodelling.⁵,⁶,²⁶ Meanwhile, there is accumulating evidence to suggest that systemic metabolic dysregulation plays a significant role in the pathophysiology of asthma. This occurs primarily through mechanisms such as altered immune cell activity and increased secretion of inflammatory mediators. Insulin resistance has been linked to heightened airway hyperresponsiveness, whereas disturbances in lipid metabolism may exacerbate disease severity through the modulation of critical inflammatory signaling pathways.⁷ While markers such as CRP and TyG index have been independently validated as prognostic indicators, utilizing them in isolation offers a fragmented physiological view.²⁷,²⁸,²⁹ A solitary marker may fail to capture the synergistic toxicity of concurrent metabolic and inflammatory insults, potentially underestimating risk in patients where these pathways operate simultaneously but subclinically.³⁰ Therefore, interpreting inflammatory and metabolic markers in an integrated manner is essential to capture the cumulative biological burden, thereby minimizing assessment bias and facilitating more precise risk stratification.

CTI combines CRP and the TyG index, capturing the co-occurrence of systemic inflammation and metabolic dysregulation in asthma. The observed associations between higher CTI and adverse outcomes may reflect a shared metabo-inflammatory burden rather than the effect of either pathway alone. Several biological mechanisms may plausibly explain this relationship.³¹,³²,³³,³⁴,³⁵,³⁶,³⁷ First, metabolic abnormalities and chronic inflammation may increase oxidative stress, which can impair airway epithelial integrity and amplify inflammatory signaling. Second, oxidative stress has been linked to reduced responsiveness to glucocorticoids, potentially contributing to poor asthma control and a higher risk of exacerbations. Third, metabolic dysfunction may promote endothelial dysfunction and systemic vascular inflammation, which could contribute to increased cardiopulmonary risk. These pathways remain speculative and were not directly tested in the present study, but they provide biologically plausible explanations for the observed associations.

Our study observed a significant positive association between CTI and all-cause mortality. In the fully adjusted model, a one-unit increase in CTI was associated with a 43% increased risk of death. More strikingly, patients in the highest tertile (T3) exhibited a 2.22-fold higher mortality risk compared to those in the lowest tertile (T1), with a significant dose-response trend observed (P for trend = 0.003). This suggests that systemic inflammation and insulin resistance may be critical factors in asthma-related adverse outcomes. Consistent with the spline analysis, a linear dose–response relationship was observed between CTI and exacerbation risk, indicating a consistent association across the exposure range. To our knowledge, this study provides the first large-scale epidemiological evidence of an independent positive association between the composite biomarker of TyG (CTI) and the risk of acute asthma exacerbation. In the fully adjusted model, a one-unit increase in CTI was associated with a 23% higher risk of acute exacerbation, and a significant dose-response relationship was evident. While the magnitude of this risk is more modest compared to mortality, it remains statistically significant and clinically relevant.

In subgroup analyses, the association between CTI and all-cause mortality appeared more evident in certain strata, including females, individuals with overweight, and those with hypertension. However, no statistically significant interaction was detected across the examined subgroups (all P for interaction > 0.05), indicating that these stratified differences should be interpreted with caution.³⁸,³⁹,⁴⁰,⁴¹ Therefore, the observed variations across subgroups may reflect differences in sample size or underlying risk profiles rather than true effect modification. These findings should be considered exploratory and hypothesis-generating, and further studies are needed to clarify whether CTI has differential prognostic value across specific patient populations.

Overall, the present study demonstrates that higher CTI is significantly associated with an increased risk of both all-cause mortality and recent asthma exacerbations among adults with asthma. These findings highlight the potential clinical and prognostic relevance of CTI as an integrated marker reflecting the combined burden of metabolic dysregulation and systemic inflammation. Although specific asthma phenotypes were not examined in this study, CTI may provide insight into the heterogeneity of asthma, particularly in relation to immunometabolic pathways.³²,³³,³⁶ Future research should further explore the role of CTI in different asthma endotypes and evaluate whether modifying metabolic and inflammatory pathways may improve clinical outcomes.

This study has several limitations that should be acknowledged. First, although adjustments were made for numerous known confounding factors, including age, sex, and smoking status, the influence of unmeasured or uncontrollable variables—such as environmental exposures and adherence to pharmacological treatments—cannot be fully eliminated. Second, the diagnosis of asthma and identification of acute exacerbation were primarily based on self-reported data, an approach that may introduce misclassification or information bias. It should be noted, however, that the questionnaire employed in NHANES has been extensively validated in previous studies. Third, given that asthma-specific mortality is an extremely rare event within population-based datasets, the limited number of cases precluded a robust statistical evaluation of asthma-related death. Consequently, our mortality analysis was restricted to all-cause mortality. Fourth, due to a substantial amount of missing data regarding asthma exacerbations, the effective sample size was insufficient to support a robust tertile analysis. Consequently, stratification into tertiles for this outcome yielded statistically unstable results and was therefore excluded. Fifth, due to the retrospective nature of this study, the establishment of a definitive causal relationship between CTI and the outcomes of interest is limited. Finally, although the NHANES sample is widely regarded as nationally representative, it consists largely of individuals from the United States. Therefore, caution should be exercised when generalizing these findings to other ethnic populations or different healthcare systems.

In this nationally representative NHANES sample, higher CTI was associated with increased risks of all-cause mortality and recent asthma exacerbations among adults with asthma, while no significant association with asthma prevalence was observed. These findings suggest that CTI may serve as a prognostic marker reflecting the combined burden of metabolic dysregulation and systemic inflammation. Further prospective studies are needed to confirm its prognostic value and elucidate underlying mechanisms.

SUPPLEMENTARY MATERIALS

Supplementary Table S1

Collinearity assessment of covariates included in the multivariable logistic regression models analyzing the association between CTI and asthma prevalence (National Health and Nutrition Examination Survey 1999–2010 and 2015–2018)

Supplementary Table S2

Collinearity assessment of covariates included in the multivariate Cox proportional hazards regression models for the association between CTI and asthma

Supplementary Table S3

Collinearity assessment of covariates in the multivariable logistic regression analysis comparing patients with stable asthma and asthma exacerbation

Supplementary Table S4

Clinical characteristics and outcomes of patients with asthma in National Health and Nutrition Examination Survey 1999–2010 and 2015–2018