Circulating Biomarkers and Cardiac Structure and Function in Rheumatoid Arthritis

Background: Rheumatoid arthritis (RA) increases the risk for abnormalities of the cardiac structure and function, which may lead to heart failure (HF). Studying the association between circulating biomarkers and echocardiographic parameters is important to screen patients with RA with a higher risk of cardiac dysfunction. Aim: To study the association between circulating biomarkers and echocardiographic parameters in patients with RA. Methods: Echocardiography was performed in 355 patients with RA from RA Porto cohort and the associations between echocardiographic characteristics and 94 circulating biomarkers were assessed. These associations were also assessed in the Metabolic Road to Diastolic Heart Failure (MEDIA-DHF) [392 patients with HF with preserved ejection fraction (HFpEF)] and the Suivi Temporaire Annuel Non-Invasif de la Santé des Lorrains Assurés Sociaux (STANISLAS) (1,672 healthy population) cohorts. Results: In the RA Porto cohort, mean age was 58 ± 13 years, 23% were males and mean RA duration was 12 ± 10 years. After adjustment and multiple testing correction, left ventricular mass index (LVMi), left atrial volume index (LAVi), and E/e′ were independently associated with biomarkers reflecting inflammation [i.e., bone morphogenetic protein 9 (BMP9), pentraxin-related protein 3 (PTX3), tumor necrosis factor receptor superfamily member 11a (TNFRSF11A)], extracellular matrix remodeling [i.e., placental growth factor (PGF)], congestion [i.e., N-terminal pro-brain natriuretic peptide (NT-proBNP), adrenomedullin (ADM)], and myocardial injury (e.g., troponin). Greater LVMi [hazard ratio (HR) (95% CI) per 1 g/m2 = 1.03 (1.02–1.04), p < 0.001], LAVi [HR (95% CI) per 1 ml/m2 = 1.03 (1.01–1.06), p < 0.001], and E/e′ [HR (95% CI) per 1 = 1.08 (1.04–1.13), p < 0.001] were associated with higher rates of cardiovascular events. These associations were externally replicated in patients with HFpEF and asymptomatic individuals. Conclusion: Circulating biomarkers reflecting inflammation, extracellular matrix remodeling, congestion, and myocardial injury were associated with underlying alterations of cardiac structure and function. Biomarkers might be used for the screening of cardiac alterations in patients with RA.


INTRODUCTION
Rheumatoid arthritis (RA) is a chronic condition characterized by systemic inflammation affecting nearly 1% of the population (1). As a consequence of the underlying proinflammatory and profibrotic state, patients with RA have a high risk of progressing toward cardiac structural and functional abnormalities [e.g., left ventricular hypertrophy (LVH), left atrial enlargement, and diastolic dysfunction] (2,3). These cardiac alterations may increase the risk of developing heart failure (HF) (4).
Circulating biomarkers may provide relevant information about the underlying pathophysiological processes associated with both the RA and cardiac alterations, even when asymptomatic (5,6). Determining the circulating biomarkers that correlate with the cardiac structure and function of patients with RA may help to better identify individuals at a risk of developing symptomatic HF and implement biomarker-based screening strategies. In this regard, consistent associations between inflammatory biomarkers and abnormal diastolic function were reported in HF with preserved ejection fraction (HFpEF) likely because HFpEF is a condition characterized by high expression of inflammatory markers (4,7).
The RA Porto cohort (ClinicalTrials.gov Identifier: NCT03960515) provides a good opportunity to explore the circulating biomarkers associated with echocardiographic parameters in RA and the prognostic value of echocardiographic parameters (8). Additionally, the associations between biomarkers and echocardiographic parameters were also assessed in a cohort of patients with HFpEF of the Metabolic Road to Diastolic Heart Failure (MEDIA-DHF) cohort study (ClinicalTrials.gov identifier: NCT02446327) (9) and a cohort of an initially healthy population in the Suivi Temporaire Annuel Non-Invasif de la Santé des Lorrains Assurés Sociaux (STANISLAS) (ClinicalTrials.gov identifier: NCTO1391442) (10).

Study Population
This study is reported as per the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guideline (Supplementary Table 1 (8). Demographic parameters, medical history, physical examination, laboratory findings and treatments, and RA-related information (i.e., RA duration period) were collected. Patient median follow-up time was 1,459 days (4.0 years). Cause of death and hospitalization were independently adjudicated. For the current analysis, we included patients with available data on echocardiogram and biomarkers. This study was conducted in accordance with the Declaration of Helsinki and approved by the Institutional Ethics Committees and all the participants signed a written informed consent prior to entry into this study.
The MEDIA-DHF cohort study included 626 patients with HFpEF diagnosis according to the 2007 European Society of Cardiology (ESC) consensus recommendations between 2012 and 2014 as previously published (9). A total of 392 patients with HFpEF with available data on echocardiographic parameters and biomarkers were included.
The STANISLAS cohort study is a single-center familial longitudinal population-based cohort from the Nancy region in France as previously reported (10). Among 1,705 participants undergoing the 4th cohort visit, 1,679 participants had available both the echocardiographic and biomarker data.
The data used for this report can be assessed by other authors upon reasonable request to the corresponding author.

Echocardiography
Echocardiographic parameters were acquired by one cardiologist that was blinded to clinical data and circulating biomarker values of patient by using the Philips R iE33 Ultrasound Machine, Philips, Bothell, WA. Cardiac chamber and systolic function [i.e., left ventricular ejection fraction (LVEF), left ventricular mass index (LVMi), left atrial volume index (LAVi), and inferior vena cava (IVC) diameter] were assessed according to the international recommendations (12,13). Pulmonary arterial systolic pressure (PASP) was estimated by using peak tricuspid regurgitation (TR) velocity and IVC compliance and diameter (13). Diastolic function was assessed from the mitral inflow pattern by pulsedwave Doppler. Mitral annular early diastolic velocity (e ′ ) was assessed at the septal and lateral sites of the mitral annulus by using tissue Doppler imaging. E/A ratio, e ′ mean, and E/e ′ mean ratio were calculated (14). LVH was defined by a LVMi > 115 g/m² in males or > 95 g/m² in females (12). The measurement reproducibility is shown in Supplementary Table 2.

Circulating Proteomic Biomarkers
A large biomarker panel with 92 biomarkers from a wide range of pathophysiological domains was measured (Olink R CVDII panel). An overview of biomarkers, their full names,   (15). In addition to the Olink R CVDII panel, in the RA Porto cohort, N-terminal pro-brain natriuretic peptide (NT-proBNP) and high-sensitivity troponin T (hsTnT) were measured by the Elecsys (Roche Diagnostics R , GmbH, Penzberg, Germany) NT-proBNP and troponin T (Gen 5 STAT test). In the MEDIA-DHF and the STANISLAS cohorts, NT-proBNP and high-sensitivity troponin I (hsTnI) were, respectively, measured by the Olink R CVDIII and organ damage panel.

Statistical Analysis
Categorical variables are described as frequencies (percentages), while continuous variables are described as mean ± SD or median (25th and 75th percentiles) depending on the variable distributions. The multivariable linear regression analysis was performed to test the associations between the circulating biomarkers and echocardiographic parameters. Models included relevant confounders as previously shown (3,16): age, sex, body mass index (BMI), diabetes, systolic blood pressure, estimated glomerular filtration rate (eGFR) [the Chronic Kidney Disease Epidemiology Collaboration formula (17)], and RA duration. In the MEDIA-DHF and the STANISLAS cohorts, the above confounders excluding RA duration were adjusted to assess the associations between biomarkers and echocardiographic parameters. Multiple testing correction for false discoveries was set at 5% (false discover rate (FDR) < 0.05), as described by Benjamini and Hochberg (18).
In the RA Porto cohort, the primary outcome was the composite of cardiovascular mortality or hospitalization for cardiovascular reasons. Hospitalization for cardiovascular

Association of Circulating Biomarkers With Echocardiographic Parameters
The circulating biomarkers associated with echocardiographic parameters are shown in the Supplementary Tables 4-6.
Similar associations were found in population of patients with HFpEF and asymptomatic individuals (Supplementary Table 8).

Echocardiographic Parameters, Circulating Biomarkers, and Clinical Outcomes
In patients with RA during a median follow-up of 48 months, the primary outcome (n = 37, 10.2%) occurred more frequently in patients with abnormalities of cardiac structure and function (Figure 1) p < 0.001] were associated with higher rates of cardiovascular events. After adjustment for potential confounders, the poor prognosis associated with greater LVMi and higher E/e ′ persisted (p < 0.05; Table 3).
Furthermore, higher levels of TNF-related apoptosisinducing ligand receptor 2 (TRAILR2) and CTSL1 were significantly associated with higher risk of the outcome after adjustment for echocardiographic parameters plus clinical confounders ( Table 4).

DISCUSSION
This study shows that cardiac structure (i.e., LVMi and LAVi) and diastolic function (i.e., E/e ′ ) of patients with RA were associated with circulating biomarkers reflecting inflammation (i.e., BMP9, PTX3, and TNFRSF11A), adverse extracellular matrix remodeling (i.e., PGF), congestion (i.e., NT-proBNP and ADM), and myocardial injury (i.e., hsTnT). Alterations of these echocardiographic parameters were associated with a worse prognosis. In addition, higher levels of inflammatory biomarkers (i.e., TRAILR2 and CTSL1) were associated with poorer prognosis in RA, independent of these cardiac alterations. These associations were also found in a population of patients with HFpEF and asymptomatic individuals. Together, these findings suggest that biomarker-guided screening may be relevant to detect cardiac structure and functional alterations in patients with RA, who are at a high risk of cardiovascular events.
The inflammatory biomarker associations described here make biologic and pathophysiological sense. We observed that a greater LVMi was associated with lower BMP9 levels and higher NT-proBNP and hsTnT levels. BMP9, a member of transforming growth factor-β1, as a potent anti-inflammatory cytokine, is expressed in the blood and heart tissues in HF. This protein inhibits cardiac fibrosis and reduces LV mass in an experimental model of HF (19). In addition, NT-proBNP and hsTnT are strongly associated with abnormalities of left ventricle and with cardiovascular morbidity and/or mortality even in pre-clinical settings (20,21).
E/e ′ reflects left ventricular filling pressures and carries relevant prognostic value (22,23). Several biomarkers reflecting systemic inflammation (e.g., TNFRSF11A, PTX3, CTSL1, and CD4) were increased in patients with RA, which may reflect a link between RA-associated inflammation and diastolic dysfunction (7). Other biomarkers associated with higher E/e ′ included BNP/NT-proBNP, hsTnT, ADM, PGF, SPON-2, and THBS2. As E/e ′ reflects left ventricular filling pressures and diastolic stiffness, it is not surprising its association with these biomarkers. For example, ADM is a vasoactive peptide synthesized by endothelial and vascular smooth muscle cells, which is increased by endothelial dysfunction and volume overload (24). PGF, a member of the vascular endothelial growth factor family, is expressed in myocytes and is increased by pressure and volume overload (25). SPON-2 expression has been associated with inflammatory processes and cardiac fibrosis (26). THBS2 has been associated with the severity of congestion in HFpEF (27). Concordantly, LAVi, reflecting left ventricular filling pressure, was also associated with many inflammatory biomarkers (e.g., CD4 and IL4RA). Interestingly, TRAILR2 and CTSL1 expressed inflammation and apoptosis and were related to inflammatory status in RA (28,29). Therefore, our observations showing higher levels of these biomarkers were associated with a higher incident outcome that supported an adverse prognosis of enhanced levels of inflammatory biomarkers in patients with RA.
Several algorithms combining LVMi, LAVi, and E/e ′ have been proposed for diagnosing or grading diastolic function (14,30). Our findings suggest that inflammatory biomarkers and widely available biomarkers as BNP/NT-proBNP and hsTnT should be used for the screening of cardiac alterations in patients with RA and asymptomatic individuals. These biomarkers may be used for an early detection of asymptomatic cardiac dysfunction and prompt an adequate control of cardiovascular and RA-related risk factors (e.g., hypertension, diabetes, and RA activity).

LIMITATIONS
This study had several limitations. First, all the three cohorts are observational cohort studies; thus, causality cannot be inferred. Second, some patients in the RA Porto cohort did not undergo echocardiogram and were excluded from the final analysis. Third, the RA Porto and the STANISLAS cohorts came from a single center and some of the findings may be influenced by local echocardiographers or practice patterns. Fourth, abnormalities of cardiac structure and function were reported to be independent of traditional cardiac risk factors and coronary artery disease in patients with RA (31, 32); however, cardiovascular comorbidities (i.e., coronary artery disease and HF) may influence the associations between echocardiographic parameters and biomarkers. We observed consistent associations between biomarkers and echocardiographic parameters across the different populations (i.e., patients with RA, those with HFpEF, and asymptomatic individuals) and biomarkers associated with study outcomes. However, external validation in other cohorts of patients with RA is needed.

CONCLUSION
Circulating biomarkers reflecting inflammation, extracellular matrix inflammation, remodeling, congestion, and myocardial injury were associated with underlying alterations of cardiac structure and function. Biomarkers might be used for the screening of cardiac alterations in patients with RA.