Associations of High-Sensitivity Cardiac Troponin T, D-Dimer, and N-Terminal Pro–B-Type Natriuretic Peptide with Subclinical Cardiovascular Dysfunction in the General Population: A Retrospective Cross-Sectional Study

Introduction

Cardiovascular diseases (CVDs) represent a major public health challenge globally. Projections indicate that coronary heart disease will continue to be the leading cause of mortality in the coming decades, while hypertension will remain the primary cardiovascular (CV) risk factor contributing to death rates.¹ Although substantial progress in medical science and clinical practice has led to more effective treatments, prevention remains essential. Preventive strategies emphasize lifestyle modification, optimal management of CV risk factors,² and the development of methods to identify patients exhibiting subclinical CV dysfunction.^3,4

Three parameters play a key role in the laboratory diagnosis of CVDs: serum cardiac troponin concentration as a marker of myocardial injury,⁵ serum D-dimer concentration as a marker of thrombosis,⁶ and serum B-natriuretic peptide concentration as a marker of myocardial overload.⁷ The specific guidelines for employing these biomarkers in diagnosing particular disease entities are extensively detailed in the literature.^8–10 Nevertheless, the utility of these markers, when measured for screening purposes outside a clinical context in which acute CVD is suspected, for evaluating CV dysfunction remains uncertain.

The role of biomarkers in screening for CV system health and CV risk has been analyzed in scientific research. For example, a large population-based study demonstrated that routinely measured cardiac troponin and serum B-natriuretic peptide concentrations are independently associated with all-cause and CV mortality in individuals with diabetes.¹¹ In our previous research, we analyzed the association between routinely measured cardiac troponin T levels and CV parameters in relation to hypertension diagnosis¹² and metabolic syndrome as defined by various diagnostic criteria.¹³ Currently, we want to investigate all three biomarkers concurrently to assess their relationship with parameters reflecting selected aspects of CV dysfunction (cardiac structure and function, subclinical atherosclerosis, arterial stiffness).

The purpose of this study was to analyze the relationship between resting high-sensitive serum cardiac troponin T concentration (hs-cTnT), D-dimer, and N-terminal pro–B-type natriuretic peptide (NT-proBNP), measured without clinical symptoms of acute illness, and selected parameters related to subclinical CV dysfunction in patients without features of acute disease.

Materials and Methods

Study Population

This study involved a retrospective analysis of medical records from patients hospitalized at the Department of Internal Medicine, Angiology and Physical Medicine, Medical University of Silesia in Katowice, Poland. The analysis included individuals hospitalized between January 2023 and May 2024 who underwent measurement of hs-cTnT, D-dimer, and NT pro-BNP during hospitalization, as well as CV system assessments. The inclusion and exclusion criteria are presented in the Table 1.

Table 1 Inclusion and Exclusion Criteria for the Study

Assessment of the Cardiovascular System

Transthoracic echocardiography (TTE) was conducted using a Vivid E9 ultrasound machine (GE HealthCare Ultrasound, Chicago, Illinois, USA) equipped with the M5S-D transducer. Standard imaging projections were acquired, and all measurements were performed in accordance with the current guidelines.¹⁴ The following parameters obtained by TTE were included in the final analysis: left ventricular ejection fraction (LV EF), left atrial volume index (LAVI), left ventricular mass index (LVMI), the ratio of the E velocity value in the mitral inflow profile to the E′ velocity value of the mitral annulus motion in tissue Doppler echocardiography (the average value of the ratio calculated for the lateral and septal regions) (E/E′), and tricuspid regurgitation maximal velocity (TRV_max).

Doppler ultrasound was conducted using a Samsung RS80 EVO device (Samsung Medison Co., Ltd., Republic of Korea). Carotid ultrasound utilized a linear probe LA4-18B, while ultrasonography of the lower extremity arteries employed a linear probe LA3-12A. All examinations and measurements adhered to established recommendations.^15–17 Intima-media thickness (IMT) was measured bilaterally in the common carotid artery (cIMT), common femoral artery (cfIMT), and superficial femoral artery (sfIMT). The mean value of each parameter from the right and left sides was calculated for the final analysis. All measurements were performed manually.

Pulse wave velocity (PWV) and central blood pressure parameters, including central systolic blood pressure (cSBP), central diastolic blood pressure (cDBP), and central pulse pressure (cPP), as well as analogously peripheral blood pressure parameters (SBP, DBP, PP), were assessed using the Sphygmocor XCEL^® (AtCorMedical, Sydney, Australia).

Ankle-brachial index (ABI) was measured bilaterally using the Dopplex^® Ability Automatic ABI System (Huntleigh, Healthcare Ltd., Cardiff, UK). Toe-brachial index (TBI) was measured bilaterally with the Dopplex DMX Digital Doppler (Huntleigh, Healthcare Ltd., Cardiff, UK). The mean values of ABI and mean TBI were included in the final analysis.

Anthropometric Measurements

Body mass, height, waist circumference, and hip circumference were measured. These parameters were then used to calculate the body mass index (BMI) and waist–hip ratio (WHR).

Laboratory Tests

Laboratory analyses were conducted at the Laboratory of Specialist Hospital No. 2 in Bytom, which serves as the headquarters of the Department of Internal Medicine, Angiology and Physical Medicine. Blood samples were collected in the morning, at least 12 to 14 hours after the last meal. hs-cTnT was measured using electrochemiluminescence with the Elecsys Troponin T hs STAT kit (Roche Diagnostics GmbH, Mannheim, Germany; detection limit is 3 pg/mL). D-dimer concentrations were assessed by latex particle-enhanced immunoturbidimetry using the D-DI2, Tina-quant D-Dimer Gen.2 assay (Roche Diagnostics GmbH, Mannheim, Germany; detection limit is 0.15 μg FEU/mL). NT-proBNP was measured using electrochemiluminescence with the Elecsys proBNP II kit (Roche Diagnostics GmbH, Mannheim, Germany; detection limit is 10 pg/mL).

Statistical Analysis

Qualitative variables were reported as counts and percentages. The Shapiro–Wilk test and visual inspection of histogram were used to assess whether quantitative variables followed a normal distribution. For variables without significant deviation from normality, the mean and standard deviation (SD) were reported. For variables with distributions that did not follow normality, the median, first quartile (Q1), and third quartile (Q3) were reported. Correlations were examined using the Spearman rank correlation test. Statistical significance was defined as a p-value less than 0.05.

Next, based on the correlation analysis, a generalized linear model was constructed to better understand the identified relationships. First, three separate models were constructed, adjusted for sex, age, BMI, hypertension, diabetes, smoking, and one laboratory marker (hs-cTnT, D-dimer, or NT-proBNP as appropriate). Secondly, a model was built that analyzed all markers simultaneously (adjusted for sex, age, BMI, hypertension, diabetes, smoking, hs-cTnT, D-dimer, and NT-proBNP). Non-normal distributed variables were log-transformed.

Statistical analyses were performed using TIBCO Software Inc., Palo Alto, CA, USA (2017), Statistica (data analysis software system), version 13.

Results

Basic Characteristics of the Study Population

The final analysis included 94 individuals, the majority of whom were women (58.51%). Obesity was present in 30.85% of patients, while 39.36% were classified as overweight. Most participants (78.72%) were diagnosed with atherosclerosis, defined as the presence of atherosclerotic plaque in any vascular bed as confirmed by imaging tests. Asymptomatic atherosclerosis was observed in the vast majority of cases. Only a small number of individuals had a diagnosis of atherosclerotic CVD and have undergone invasive revascularization or experienced a CV event. Slightly more than half of the patients had carbohydrate metabolism disorders, with diabetes diagnosed in 28.72% and prediabetes in 22.34% of participants. The characteristics of the study population are summarized in Table 2.

Table 2 Basic Characteristics of the Study Population

Parameters for Assessment the Cardiovascular System

In the study population, 79.79% of individuals exhibited hs-cTnT levels within the normal range. A reduced LV EF (below 50%) was observed in 6.38% of participants, with no cases below 40%. Left ventricular myocardial hypertrophy (LVMI > 95 g/m² in women; LVMI > 115 g/m² in men) was identified in 30.9% of women and 5.1% of men. LAVI remained within the normal range (not exceeding 34 mL/m²) in 80.85% of cases. An elevated E/E′ value of 14.0 or higher was observed in only one participant. TRV_max exceeded 2.8 m/s in two individuals. Increased arterial stiffness, defined as PWV greater than 10.0 m/s, was detected in 39.7% of the study population. Peripheral arterial disease, indicated by ABI below 0.9, was found in 10.6% of participants. Table 3 provides a detailed summary of the CV system parameters assessed in the study population.

Table 3 Descriptive Statistics of Parameters Relating to the Assessment of the Cardiovascular System Status in the Entire Study Population

Correlation Analysis

Correlation analysis demonstrated that hs-cTnT is significantly positively associated with LAVI, LVMI, E/E′, cIMT, cfIMT, sfIMT, SBP, cSBP, PP, cPP, and PWV. Significant negative correlations were observed between hs-cTnT and LV EF, DBP, cDBP, ABI, and TBI. The strongest correlations for hs-cTnT (│R│ > 0.5) were with PWV, PP, cPP, cIMT, cfIMT, and sfIMT, while the weakest correlations (│R│ < 0.3) were with LVMI, DBP, cDBP, and TBI. No significant correlation was identified between hs-cTnT and TRV_max.

NT-proBNP demonstrated significant positive correlations with LAVI, LVMI, E/E′, TRV_max, cIMT, cfIMT, sfIMT, PP, cPP, and PWV. Significant negative correlations were observed with DBP, cDBP, and ABI. For most parameters, the strength of correlation with NT-proBNP was weaker or comparable to that observed for hs-cTnT.

D-dimer exhibited significant correlations only with cIMT, cfIMT, sfIMT, PP, and cPP. In all cases, these correlations were weaker than those observed for hs-cTnT.

The complete results of the univariate correlation analysis between biochemical markers (hs-cTnT, NT-proBNP, D-dimer) and CV dysfunction parameters are provided in Table 4.

Table 4 Comparison of Correlations Between the Serum Concentration of High-Sensitivity Cardiac Troponin T, N-Terminal Pro–B-Type Natriuretic Peptide, and D-Dimer, and Parameters Related to Subclinical CV Dysfunction. Correlations Were Assessed Using the Spearman Rank Correlation Coefficient (R)

Multivariate Analysis

High-Sensitivity Cardiac Troponin T

Multivariate analysis indicated that the constructed model was significant for the following parameters: LAVI, LVMI, log(SBP), cDBP, log(cPP), PWV, TBI, cIMT, log(sfIMT), log(cfIMT), and log(E/E′). Complete results are provided in Table 5.

Table 5 Results of the Multivariate Analysis Model Significance Test for Explaining Individual Parameters. The Model Was Adjusted for Age, Sex, Body Mass Index (BMI), Hypertension, Diabetes, Smoking, and Troponin

Troponin contributed significantly to the explanation of cDBP and log(sfIMT), independently of confounding variables. For log(cPP), PWV, log(cfIMT), and log(E/E′), the effect of troponin on the model approached statistical significance (p < 0.1). Detailed results are shown in Table 6.

Table 6 Results of the Multivariate Analysis Model Test for Explaining Individual Variables (Troponin)

D-Dimer

Multivariate analysis indicated that the developed model was significant for the following parameters: LV EF, LAVI, LVMI, log(SBP), log(cPP), PWV, TBI, cIMT, log(sfIMT), log(cfIMT), and log(E/E′). Complete results are provided in Table 7.

Table 7 Results of the Multivariate Analysis Model Significance Test for Explaining Individual Parameters. The Model Was Adjusted for Age, Sex, Body Mass Index (BMI), Hypertension, Diabetes, Smoking, and D-Dimer

D-dimer contributed significantly to the explanation of log(cfIMT) and log(E/E′), independently of confounding variables. Complete results are provided in Table 8.

Table 8 Results of the Multivariate Analysis Model Test for Explaining Individual Variables (D-Dimer)

N-Terminal Pro–B-Type Natriuretic Peptide

Multivariate analysis indicated that the developed model was significant for the following parameters: LV EF, LAVI, LVMI, log(SBP), log(cPP), PWV, TBI, cIMT, log(sfIMT), log(cfIMT), and log(E/E′). Comprehensive results are provided in Table 9.

Table 9 Results of the Multivariate Analysis Model Significance Test for Explaining Individual Parameters. The Model Was Adjusted for Age, Sex, Body Mass Index (BMI), Hypertension, Diabetes, Smoking, and NT-proBNP

NT-proBNP contributed significantly to the explanation of LAVI, independently of confounding variables. For LV EF, the effect of NT-proBNP on model explanation approached statistical significance (p = 0.059). Detailed results are presented in Table 10.

Table 10 Results of the Multivariate Analysis Model Test for Explaining Individual Variables (NT-proBNP)

Troponin, D-Dimer, and NT-proBNP Analyzed Together

The model, adjusted for sex, age, BMI, hypertension, diabetes, smoking, hs-cTnT, D-dimer, and NT-proBNP, demonstrated statistical significance for the following variables: LV EF, LAVI, LVMI, PWV, TBI, cIMT, log(E/E′), log(sfIMT), log(cfIMT), log(SBP), and log(cPP). The association with cDBP approached significance (p = 0.093). NT-proBNP accounted for variability in LAVI independently of confounding factors (β = 0.257; 95% CI: 0.052–0.462; p = 0.015). hs-cTnT explained the variability in cDBP [β = –0.3; 95% CI: (–0.582)–(–0.017); p = 0.038] and PWV (β = 0.239; 95% CI: 0.016–0.463; p = 0.036), independently of confounders. D-dimer accounted for the variability in log(E/E′) (β = 0.244; 95% CI: 0.039–0.45; p = 0.02) and log(cfIMT) (β = 0.169; 95% CI: 0.0001–0.339; p = 0.05), independently of confounding factors.

Discussion

The presented study indicates that, even among individuals in relatively good health and predominantly without overt CVD, randomly measured biomarkers such as hs-cTnT, D-dimer, and NT-proBNP demonstrate significant correlations with parameters associated with subclinical CV dysfunction, including heart structure and function, subclinical atherosclerosis, and arterial stiffness. The importance of biomarkers determined without acute illness has already been discussed in the literature. The primary innovative aspect of the present study is the simultaneous analysis of the relationships among hs-cTnT, D-dimer, and NT-proBNP, in addition to parameters associated with multiple aspects of CV dysfunction.

The present results indicate that the strongest correlations occurred between hs-cTnT and IMT values across multiple vascular beds. A Norwegian research group demonstrated a high predictive value of serum troponin concentration for assessing subclinical carotid atherosclerosis, although their study focused on troponin I.¹⁸ Alanis et al proposed that vascular aging may be a mechanism underlying troponin release in adults without clinically overt CVD.¹⁹

Faintuch et al reported an association between D-dimer levels, subclinical atherosclerosis, and arterial stiffness in morbidly obese patients. Their study identified a significant correlation between D-dimer and IMT (r = 0.524, p = 0.005), but no correlation with PWV.²⁰ These findings align with the present results, which demonstrate a significant correlation between serum D-dimer levels and IMT across all vascular beds examined. In multivariate analysis, this association remained significant for log-transformed cfIMT. Although the current study population did not include morbidly obese individuals, the proportion of overweight and obese participants was substantial. In contrast, Azuma et al found no significant relationship between D-dimer concentration and IMT in Japanese men aged 40–49 years. However, a borderline significant association was observed in white men, who exhibited a higher mean body mass index (BMI) (27.8 ± 4.2 kg/m² vs 24.1 ± 3.1 kg/m²; p < 0.01).²¹ Similarly, in a cohort of 100 patients with atherosclerosis and a mean age of 72.1 years, no significant correlation was observed between D-dimer concentration and IMT. Nevertheless, individuals with more severe atherosclerosis, as assessed by IMT and cardio-ankle vascular index (CAVI) (CAVI > 8.0 and IMT > 1.1 mm), exhibited significantly higher D-dimer concentrations compared to those with less severe disease (CAVI < 8.0 and IMT < 1.1 mm) (0.48 μg/mL vs. 0.32 μg/mL; p < 0.01).²² Furthermore, Zhou et al demonstrated that D-dimer exerts a pro-atherosclerotic effect on macrophage function, promoting the inflammatory response during macrophage-derived foam cell formation.²³

Another significant finding of this study is the positive association between D-dimer concentration and log-transformed E/E′, an established indicator of diastolic dysfunction. Catena et al similarly demonstrated that a prothrombotic state in patients with hypertension correlates with early left-ventricular diastolic impairment. In contrast, the present study included both hypertensive and normotensive individuals.²⁴

Significant positive correlations were observed between serum NT-proBNP levels and parameters associated with left ventricular myocardial mass (LVMI) and diastolic function (LAVI, TRV_max, and log-transformed E/E′). In multivariate analysis, only the association between NT-proBNP and LAVI remained significant. These findings are partially consistent with those reported by Yamazaki et al, who identified positive relationships between serum NT-proBNP concentration and LAVI (r = 0.353, p < 0.0001), LVMI (r = 0.336, p = 0.0002), and E/E′ (r = 0.412, p < 0.0001) in patients with end-stage renal disease undergoing dialysis.²⁵

An unexpected result from our study is the absence of a significant correlation between serum NT-proBNP levels and LV EF. This outcome may be attributed to the limited variation in LV EF values within our study population, as most participants exhibited values within the normal range. Gao et al reported similar findings in patients with unstable angina and preserved left ventricular systolic function; specifically, no significant correlation was observed between NT-proBNP and LV EF in individuals without diabetes (r = −0.109, p = 0.567), whereas a significant correlation was present in those with diabetes.²⁶ In contrast, among patients with diverse left ventricular systolic function and dyspnea, a clear relationship between NT-proBNP and LV EF has been demonstrated.²⁷

The serum NT-proBNP concentration showed a significant correlation with PWV in univariate analysis; however, this association was no longer significant after adjustment for confounding factors. Researchers from Denmark reported similar findings when analyzing data from 1872 apparently healthy participants in the MONItoring of trends and determinants in CARdiovascular disease (MONICA) study.²⁸

Limitations of the Study

The presented study has certain limitations. The study was cross-sectional, which allows for the identification of certain relationships but does not permit conclusions about causality. Due to the study’s retrospective design, data were missing in some cases. The study was conducted on a relatively small sample; therefore, it was not possible to perform additional subgroup analyses that accounted for coexisting CV risk factors across different constellations. It would be worthwhile to repeat similar studies in the future with a larger group. Moreover, a limitation of the study is its single-center nature. Another limitation is the lack of use of modern echocardiographic techniques, such as global longitudinal strain (GLS) measurement. It would be particularly valuable because the vast majority of study participants had normal LV EF, and GLS would allow for more precise diagnosis of subclinical left ventricular systolic dysfunction. Moreover, when interpreting the results of the study, it is worth noting that the study population was relatively homogeneous with respect to CV status, as the CV status assessment results fell within the normal range for most participants. On the one hand, it demonstrates that even among individuals with a relatively good CV condition, there are correlations between the studied markers and features of subclinical CV dysfunction. On the other hand, it narrows the possibilities for reaching a conclusion.

Conclusion

In the general population without signs of acute illness, screening markers, including hs-cTnT, NT-proBNP, and D-dimers, are associated with specific aspects of subclinical CV system dysfunction. However, the nature and strength of these associations vary considerably among the individual markers.

Univariate analysis demonstrates that hs-cTnT correlates with peripheral and central blood pressure (SBP, DBP, cSBP, cDBP), left ventricular systolic function (LV EF), left ventricular mass, and selected features of diastolic dysfunction (LVMI, LAVI, E/E′, but not TRV_max). Additionally, hs-cTnT is associated with features of subclinical atherosclerosis (cIMT, cfIMT, sfIMT, ABI, TBI) and arterial stiffness (PWV, PP, cPP). In multivariate analysis, significant relationships persist between hs-cTnT and both cDBP and sfIMT, while associations with PWV, cPP, cfIMT, and E/E′ approach borderline significance.

In univariate analysis, NT-proBNP correlates with peripheral and central blood pressure (DBP, cDBP, but not SBP and cSBP), left ventricular mass, and selected features of diastolic dysfunction (LVMI, LAVI, TRV_max, E/E′), features of subclinical atherosclerosis (cIMT, cfIMT, sfIMT, ABI, but not TBI), and arterial stiffness (PWV, PP, cPP). In multivariate analysis, the relationship between NT-proBNP and LAVI remains significant, while the relationship between NT-proBNP and LV EF is of borderline significance.

Univariate analysis shows that D-dimer serum levels correlate with selected features of subclinical atherosclerosis (cIMT, cfIMT, sfIMT) and arterial stiffness (PP, cPP, but not PWV). In multivariate analysis, significant associations persist between D-dimer serum levels and both cfIMT and E/E′.

Further research, particularly prospective multi-center observational studies with larger patient cohorts, is required to elucidate the observed relationships, their determinants, and their diagnostic and prognostic significance.