Introduction

Asthma, a common chronic non-communicable disease worldwide, is marked by airflow obstruction causing breathing difficulties and wheezing1. During the latter half of the twentieth century, the global prevalence of asthma increased rapidly2. Although the overall prevalence has decreased slightly from 3.9 to 3.6% since 1990, asthma remains the second most common chronic respiratory disease globally3. There is a pervasive lack of diagnosis and inadequate treatment, with both childhood and adult asthma imposing a significant public health burden4,5.

The C-reactive protein to albumin ratio (CAR) represents a novel biomarker that is gradually coming into the research spotlight6. The CAR integrates the prognostic values of C-reactive protein (CRP), an acute phase reactant that rises in the presence of inflammation7, and albumin, a negative acute phase reactant that declines in concentration in reaction to inflammation8. Studies have confirmed that elevated CAR is associated with several adverse outcomes, including cancers9,10,11, sepsis6,12,13, COVID-1914,15,16, and cardiovascular diseases17,18,19.

Inflammation is known to play a key role in the pathogenesis and progression of asthma20,21. To date, no studies have specifically examined CAR and asthma prevalence and mortality. Given the importance of identifying reliable biomarkers for asthma management, this study aims to analyze the effect of CAR on the prevalence of asthma and the mortality from all causes and respiratory diseases in patients with asthma.

Methods

Study population

The National Health and Nutrition Examination Survey (NHANES) is a study employing a complex sampling design to assess the health and nutritional status of the U.S. population. It collects participants’ basic information, questionnaire responses, physical examination results, and various laboratory test data. The survival status of participants in the NHANES as of December 31, 2019, a prerequisite for the conduct of the cohort study, can be obtained by the NHANES through interfacing with the National Center for Health Statistics (NCHS) death data.

Data from NHANES for the years 2001–2018 were utilized in this study. Participants who lacked complete CAR data (C-reactive protein, albumin), were less than 20 years of age, lacked pertinent asthma survey information, were pregnant, or lost to follow-up were excluded from this study (Fig. 1). As all study participants provided informed consent and the NCHS Institutional Review Board approved this study, no additional informed consent or ethical review was required. All methods were performed in accordance with the relevant guidelines and regulations.

Fig. 1
figure 1

Participant Selection Flowchart. NHANES National health and nutrition examination survey, CAR C-reactive protein to albumin ratio.

Full size image

C-reactive protein to albumin ratio

In this study, CAR was calculated as CRP (mg/L) divided by ALB (g/L). Both CRP and ALB were collected at the Mobile Examination Center and then sent to a laboratory for analysis. Before collection, blood samples were screened based on certain exclusion criteria. Furthermore, NHANES conducted quality control on the collection process and test results.

Asthma

With reference to existing research22,23, Individuals who responded “yes” to the question “Has a doctor or other health professional ever told you that you have asthma?” within the questionnaire were considered to have asthma.

Covariates

Several potential covariates (demographics, living status, disease status, biochemical indicators) were included in our study. Demographic information included sex, age, race, poverty level (non-poverty: poverty-income ratio > 1, poverty: poverty-income ratio < 1), marital status, and education. Lifestyle status included smoking, drinking alcohol, and body mass index. Disease status consisted of hypertension, coronary heart disease (CHD), diabetes mellitus, neoplasm, stroke, and rheumatoid arthritis. CHD, stroke, and rheumatoid arthritis disease status were provided by the questionnaire, while hypertension and diabetes mellitus status were determined by combining the questionnaire and actual measurements of blood pressure, fasting blood glucose, oral glucose tolerance test, and glycated hemoglobin. The biochemical indicators used were TC and AST.

Statistical analysis

Considering the weighted analysis better reflects the true distribution of the overall population, the characteristics of all participants were evaluated according to the CAR quartiles. Continuous variables describing the total population characteristics were expressed as weighted means and standard deviations. In Categorical variables, the weighted percentages were employed. Comparisons between quartiles were conducted in t-tests or chi-square analyses. The characteristics of asthma patients were assessed based on survival status, with continuous variables expressed as means and standard deviations or medians and interquartile ranges and compared using t-tests. Categorical variables were described using frequencies and percentages, with differences assessed using chi-square tests. Logistic regression models, adjusted for multiple covariates, were employed to explore the associations between CAR and asthma prevalence. Three progressively adjusted models were used. Similarly, Cox proportional hazards models evaluated the relationship between CAR and mortality outcomes. Since the NHANES data is a nationally representative sample and not clustered by study centers or states, random effects were not included in our Cox models. Kaplan–Meier survival curves and log-rank tests were used to compare survival rates across CAR quartiles. Restricted cubic spline (RCS) analysis explored potentially non-linear relationships and identified threshold effects. Subgroup analyses assessed consistency across different groups. Model predictive accuracy over 3-, 5-, and 10-year periods was evaluated using time-dependent receiver operating characteristic (ROC) curves and the area under the curve (AUC). Random survival forest models were employed to demonstrate the importance of CAR. Sensitivity analyses were conducted using unweighted logistic regression and updated Cox models. The R software (version 4.3.3) was employed to facilitate all statistical analyses in this study.

Results

Baseline characteristics

The study included 33,565 participants, 4391 of whom had asthma. Mean age was 47.32 ± 16.87 years, with 48.9% male and 51.1% female. The prevalence of asthma was 13.7%. As shown in Table 1, participants were classified into four groups based on their weighted CAR quartiles: Q1: < 0.017; Q2: 0.017–0.043; Q3: 0.043–0.103; Q4: > 0.103. A comparison of the highest and lowest quartiles of the CAR group revealed that the latter had higher proportions of Mexican Americans, non-Hispanic whites, females, and smokers, and participants' mean age, BMI, and biochemical indicators showed significant trends with increasing CAR quartile. These subjects also had a significantly increased prevalence of hypertension, coronary heart disease, diabetes, cancer, stroke, and rheumatoid.

Table 1 Quartiles of C-reactive protein to albumin ratio, the participants' fundamental characteristics.
Full size table

As of December 31, 2019, 628 (14.3%) all-cause mortality events occurred in 4,391 asthma patients, with a mean follow-up period of 77.0 (34.8, 117) months. Table 2 shows the characteristics of asthma patients after grouping them according to their survival status up to the cut-off date. All-cause mortality was higher among asthmatics with higher mean age, male, Non-Hispanic Black, unmarried, and low education rates, and individuals with hypertension, coronary heart disease, diabetes, cancer, stroke, and rheumatoid continued to have higher all-cause mortality rates. Moreover, AST, CRP, albumin, and CAR were significantly elevated in asthmatics who died compared with survivors.

Table 2 According to the grouping based on survival status, the basic characteristics of asthma participants.
Full size table

Association between the CAR and asthma

This study has findings indicating a positive fluctuation between CAR levels and the prevalence of asthma (Table 3). Furthermore, a significant dose–response relationship was observed across all models. In the final model, Model 3, an increase of 0.1 in CAR was found to be associated with an increase in the odds of asthma by 5% (OR 1.03, 95% CI 1.02–1.05). This association became increasingly pronounced as the CAR quartiles were considered. Specifically, after adjusting for demographic factors in Model 2 and including additional variables such as poverty status and health conditions in Model 3, the association remained robust. participants in the top quartile of CAR (> 0.103) exhibited a 49% higher likelihood of asthma relative to the 1st quartile (Q4: OR 1.49, 95% CI 1.30–1.70). Moreover, a significant trend (P for trend < 0.001) of increasing asthma risk with rising CAR levels was evident in all three models.

Table 3 The association between the CAR and asthma.
Full size table

CAR and survival status of asthma patients

Kaplan–Meier survival curve analysis demonstrated a notable correlation between CAR levels and all-cause mortality among asthma patients (Fig. 2), as indicated by a Log-rank test result of P < 0.0001. Furthermore, asthmatics in the higher quartiles (Q3 and Q4) demonstrated substantially lower survival rates compared to the overall asthma patient population.

Fig. 2
figure 2

Kaplan–Meier survival curve analysis of all-cause mortality based on CAR quartiles in patients with asthma. CAR in asthma patients Quartiles Q1: < 0.023; Q2: 0.023–0.059; Q3: 0.059–0.149; Q4: > 0.149. Abbreviations: CAR C-reactive protein to albumin ratio, KM Kaplan–Meier survival curve.

Full size image

Subsequently, three stepwise-adjusted Cox proportional hazards models were utilized to evaluate the correlation between CAR and mortality in asthma patients (Table 4). The findings demonstrated a significant association between CAR levels and all-cause mortality and mortality from respiratory diseases. Specifically, in Model 3 (detailed in Supplementary Tables S1 and S2), each 0.1 increase in CAR was associated with a 7% mortality risk increase (all-cause mortality: HR 1.07, 95% CI 1.05–1.09; respiratory mortality: HR 1.07, 95% CI 1.00–1.13). Furthermore, patients with the highest CAR quartile (> 0.149) exhibited a more than twofold elevated risk of all-cause mortality (HR 2.20, 95% CI 1.67–2.89) and respiratory disease mortality (HR 2.56, 95% CI 1.30–5.38) compared to asthmatics in the 1st quartile. All model trend checks were statistically significant. Figure 3 demonstrates the nonlinear relationship between CAR and the all-cause mortality risk in asthmatics (P for nonlinearity < 0.001). A threshold value of CAR at 0.1357 was identified as pivotal, with CAR's impact on mortality risk being more substantial below this point (Table 5).

Table 4 Association between CAR and all-cause mortality and respiratory disease mortality in asthma patients.
Full size table
Fig. 3
figure 3

Restricted cubic spline analysis of the association between CAR and the risk of all-cause mortality in asthma patients. Abbreviations: HR hazard rate, CI confidence interval, CAR C-reactive protein to albumin ratio.

Full size image
Table 5 Inflection point and segmentation effects of CAR.
Full size table

Subgroup analysis and model evaluation

As indicated in Table 6, subgroup analysis suggests a potential interaction effect between hypertension and the CAR and all-cause mortality among asthma patients (P for interaction = 0.011). Specifically, individuals without hypertension demonstrated a heightened risk of all-cause mortality. Each 0.1 increase in CAR is associated with a 15% increase in the risk of asthma mortality (HR 1.15, 95% CI 1.08–1.21). The correlation between CAR and all-cause mortality in asthmatics was relatively consistent across the other subgroups.

Table 6 Subgroup analysis by CAR on asthma mortality.
Full size table

ROC curves (Fig. 4) indicated that the fully adjusted model exhibited significantly greater predictive capacity than the unadjusted model. Furthermore, predictive accuracy was observed to improve with longer follow-up periods, and AUC values demonstrated an upward trend over time. Figure 5 illustrates the significance of CAR in forecasting all-cause mortality in asthma patients within the context of a randomized survival forest model.

Fig. 4
figure 4

Comparative ROC curves for all-cause and respiratory disease mortality over different periods. (A) 3-year ROC curve for all-cause mortality, (B) 5-year ROC curve for all-cause mortality, (C) 10-year ROC curve for all-cause mortality, (D) 3-year ROC curve for respiratory disease mortality, (E) 3-year ROC curve for respiratory disease mortality, (F) 3-year ROC curve for respiratory disease mortality. Abbreviations: ROC receiver operating characteristic, AUC Area under the curve.

Full size image
Fig. 5
figure 5

Variable importance in random survival forest models for all-cause and respiratory disease mortality. (A) Variable importance for all-cause mortality, (B) Variable importance for respiratory disease mortality. Incorporate the same variables as Model 3. Abbreviations: CAR C-reactive protein to albumin ratio, CHD coronary heart disease, BMI Body mass index, TC total cholesterol, AST aspartate aminotransferase.

Full size image

Sensitive analysis

Sensitivity analyses were conducted using unweighted logistic regression (Supplementary Table S3) and updated adjusted models (Supplementary Table S4) to examine the association between the CAR and the incidence and mortality of asthma. The updated adjusted models incorporated additional adjustments for aspartate aminotransferase, chronic bronchitis, and liver disease. The results consistently demonstrated that elevated CAR was positively correlated with an increased risk of asthma incidence, all-cause mortality, and respiratory disease mortality. These associations remained significant across different models and adjustments, thereby reinforcing the robustness of our findings, and highlighting CAR as a reliable predictor of mortality outcomes.

Discussion

The results of our study identified a significant association between the CAR and asthma prevalence, as well as all-cause and respiratory disease mortality in patients with asthma. Higher CAR levels were associated with an increased likelihood of asthma. Asthma patients with elevated CAR showed significantly higher both overall mortality and mortality due to respiratory disease. The mortality rate in the highest CAR quartile was more than twice that of the lowest quartile, underscoring CAR's value as a prognostic biomarker for asthma outcomes.

While there are no existing studies specifically examining the association between CAR and asthma, as well as the prognosis of asthmatics, this can be explained through the lens of systemic inflammation and nutritional status. CRP is a molecule synthesized primarily by hepatocytes that increases in concentration during inflammatory responses7. It is commonly used as a static measurement indicator and is correlated with the severity and prognosis of diseases24,25. Some studies have shown that CRP levels are significantly elevated in patients not receiving steroid treatment26 and in non-allergic asthma patients27, and CRP is negatively correlated with pulmonary function parameters such as FEV1/FVC ratio28,29. Conversely, albumin is a protein that is synthesized by the liver and is decreased during inflammation and malnutrition8. Factors such as malnutrition, fluid overload, inflammation, and liver dysfunction are all associated with hypoalbuminemia and adverse outcomes in various diseases30. Studies have also indicated that serum albumin levels are lower in asthma patients compared to the normal population31, and this difference is also present in asthmatic children32. These studies indirectly support our results, suggesting that CAR, as a combined marker, may provide comprehensive prognostic information. Additionally, the impact of CRP on albumin levels is relatively small, with a significant negative correlation only observed at extremely high CRP levels (500 mg/L)33. This implies that CAR can provide more reliable assessments of inflammation and nutritional status, as it is less affected by extreme inflammation.

We also revealed a potential nonlinear relationship between CAR and asthma outcomes. This nonlinear trend may be due to varying degrees of inflammation and nutritional deficiencies among patients with different CAR levels. When analyzed by subgroup, the effect of CAR on all-cause mortality was particularly significant among hypertensive patients (P for interaction = 0.011). The synergistic impact of hypertension and elevated CAR levels on asthma outcomes could be attributed to the cumulative effects of inflammation. Numerous studies have shown that hypertension increases oxidative stress and inflammation34,35. This finding highlights the necessity of considering hypertension when assessing the prognostic significance of CAR in patients with asthma.

CAR is likely associated with various diseases, as our analysis of non-asthmatic patients also found significant associations between CAR and mortality. However, the impact of CAR is particularly pronounced in asthmatic patients, where the highest CAR quartile was associated with a markedly greater risk of both all-cause and respiratory disease mortality compared to non-asthmatic patients (Supplementary Table S5). This underscores the research potential of CAR across different conditions, while also highlighting its unique prognostic value in the context of asthma.

The study is notable for its reliance on a large, nationally representative sample. However, there are also some limitations. The study population was limited to adults, and the reliance on self-reported asthma diagnosis may introduce classification bias. Additionally, CAR levels may be affected by various factors, including recent acute infections and nutritional status, which were not fully controlled for in this study. Our analysis did not account for the random effects of the NHANES dataset structure. However, given the nationally representative nature of NHANES, the likelihood of regional variability is minimized. To establish causality and explore the potential mechanisms of the association between CAR and asthma outcomes, future longitudinal studies are necessary.

Conclusion

The results of our study indicated a significant correlation between CAR and asthma prevalence, as well as all-cause and respiratory disease mortality in asthma patients. Higher CAR levels were correlated with an increased likelihood of asthma prevalence and significantly higher all-cause and respiratory disease mortality. The findings indicate that CAR is a valuable biomarker for predicting asthma outcomes.