Biomarker sST2 in Adults with Transposition of the Great Arteries Palliated by Mustard Procedure: A Five-Year Follow-up

The Mustard procedure was an early cardiac surgery for transposition of the great arteries (TGA). Despite being successful, it has been associated with long-term arrhythmias and heart failure. A key factor complicating management in adults with congenital heart disease (CHD) is the deficiency of biomarkers predicting outcome. Soluble suppression of tumorogenicity-2 (sST2) is secreted by cardiomyocytes in response to mechanical strain and fibrosis. We hypothesized that adults with a Mustard procedure would have higher levels of sST2 than healthy individuals, and this would correlate with clinical outcome. We performed a single-center study in patients managed during childhood with a Mustard procedure versus age-matched controls. Clinical and demographic data were collected and biomarkers (sST2, cTnI, BNP, lipid panel, insulin, and glucose) were obtained. There were 18 patients (12 male) in the Mustard cohort and 18 patients (6 male) in the control group (22–49 years, mean of 35.8 vs. mean 32.6 years, respectively, p = ns). Nine Mustard subjects were NYHA class II, and 9 subjects were class III. The control group was asymptomatic. sST2 in the Mustard group was elevated in 56% vs. 17% in controls (p = 0.035). Of the Mustard subjects with elevated sST2, 60% had elevated cTnI and BNP, and 90% had low HDL. Over five years, the Mustard patients with elevated sST2 values had greater medication use, arrhythmias, hospitalizations, and ablation/pacer implantations than Mustard subjects with normal sST2. Mustard subjects with elevated sST2 had other biomarker abnormalities and clinically worse outcomes. Thus, sST2 may add a predictive value to cardiac-related morbidity and mortality.


Introduction
Transposition of the great arteries (D-TGA) is a cyanotic congenital heart disease in which the aorta and pulmonary artery are transposed. The Mustard procedure was one of the earliest cardiovascular surgeries allowing survival into adulthood [1,2] and was used widely until its replacement in the mid-1980s by the arterial switch operation [3]. The abandonment of the Mustard operation in favor of the arterial switch was due to its long-term adverse outcomes including arrhythmias, baffle leaks and obstruction, and right ventricular dysfunction leading to heart failure (HF) [4]. However, the number of adult Mustard patients continues to increase [5], and HF is of great concern in this population. One important problem in the Mustard population has been the difficulty in using echocardiography of the systemic right ventricle to follow function over time [6].
Biomarkers represent a valuable tool needed to allow earlier detection of HF and risk stratification. Soluble 1 3 suppression of tumorigenicity 2 (sST2) is a new biomarker of cardiac stress and HF severity [7][8][9][10][11]. It is predictive of outcome and mortality [6] in adults with chronic HF and signals the presence of cardiac fibrosis and adverse remodeling [4]. Soluble ST2 has strong prognostic value and aids in risk stratification to identify adult HF patients at risk of mortality or re-hospitalization [4]. To date, no studies have focused specifically on the value of sST2 in individuals with D-TGA who have undergone the Mustard procedure. In this study, we tested the hypothesis that adults with D-TGA palliated with a Mustard procedure would have higher levels of sST2 than healthy controls and that it would correlate with clinical outcomes.

Methods
This study was designed as a retrospective cohort investigation in which serum sST2 concentrations were measured in a cohort of subjects who had undergone a Mustard operation for D-TGA and an age-matched healthy control group. The Mustard cohort was followed for up to five years to assess their clinical outcome. The biomarker results were not known to the clinicians caring for these patients during the 5-year follow-up period. The Institutional Review Board at Indiana University Health approved the protocol and the subsequent procedures were followed in accordance with institutional guidelines.
A chart review identified 45 patients with D-TGA who had undergone palliation with a Mustard procedure. We excluded patients with any recent (past 3 months) surgical or interventional procedures. Of those remaining, eighteen subjects with a Mustard procedure provided written informed consent. All Mustard subjects were assessed by the NYHA functional classification, and all had recent examinations and EKG. Eighteen healthy control subjects were recruited from the cardiology clinic and community. Criteria for healthy individuals were defined as no known heart disease or recent (past 3 months) surgery, as well as no major disease process such as asthma, COPD, or chronic inflammation [12]. Because control subjects were recruited for blood draw only, no biometric data were obtained for them. A five-year follow-up of the Mustard cohort was conducted to determine current health status.
Bloods were drawn in the non-fasted state with EDTA anticoagulation. Soluble ST2 levels were obtained using Critical Diagnostics Presage ST2 Assay kits (Critical Diagnostics, San Diego, CA, USA). The assay kit measures sST2 by enzyme-linked immunosorbent assay (ELISA) in a 96-well microtiter plate format. The assay was calibrated with seven standards spanning the range of 3.1 to 200.0 ng/ ml. The sST2 cut point was set at 35 ng/ml for these analyses [13,14] above which mortality risk has been shown to be significantly higher in patients with HF [15]. The assay was performed at the Herman B. Wells Center for Pediatric Research, Indianapolis. For cardiac troponin I (cTnI), brain natriuretic peptide (BNP), lipid panel, and insulin, the clinical laboratories at Indiana University Health were utilized. The lipid panel included high density lipoprotein (HDL), low density lipoprotein (LDL), Triglycerides, and total cholesterol. Data were analyzed using standard statistical software (Excel and GraphPad Prism version 9.3). Continuous variable data are expressed as mean ± standard error of the mean (SEM) unless otherwise noted. Differences between unpaired groups were analyzed using the Mann-Whitney test unless otherwise noted, and correlations were evaluated using Spearman's regression coefficient. A two-tailed p value < 0.05 was considered statistically significant.

Results
Of the eighteen subjects with Mustard procedure in the analysis, twelve were male and six were female with an age range of 30-46 years. In the Mustard group, nine (50%) subjects were assigned to NYHA class II, and nine (47%) subjects to class III. The control group was asymptomatic and was matched for age to the Mustard group with a range of 22-49 years and included six males and 12 females.
The average age of the Mustard subjects was 34.9 years (± 1.1 years) and was not different from the control cohort (32.6 ± 1.7 years). All subjects in the Mustard cohort were Caucasian versus the control cohort with 78% Caucasian, 17% Hispanic, and 5% Asian. The Mustard cohort was 67% male compared to 33% male in the control cohort. A male predominance is known for D-TGA [16], and no attempt was made to control for gender or race in this study.
Biomarker averages for Mustard subjects compared to controls are shown in Table 1. Soluble ST2 levels were significantly elevated in the Mustard group when compared to control groups with an average of 42.93 ± 6.90 versus 29.71 ± 1.24, respectively (p = 0.0071). Of the Mustard patients, 56% (n = 10) had elevated levels while 17% (n = 3) had elevated levels in the control group (p = 0.035). Levels of HDL, total cholesterol, and cholesterol/HDL ratio in the Mustard subjects compared to the control group were also significantly different (Table 1). Both HDL and total cholesterol were low in the Mustard group, whereas the total cholesterol to HDL ratio was elevated in Mustard subjects. Triglycerides and insulin levels were not different between the groups.
Cardiac troponin I (cTnI) was significantly elevated in the Mustard cohort when compared to controls (Fig. 1A), with the average Mustard cTnI at 0.21 ± 0.14 ng/mL and the average control cTnI at < 0.03 ng/mL (p = 0.008). None of the control subjects had an elevated cTnI level (defined as ≥ 0.03 ng/mL), whereas 39% (n = 7) of the Mustard cohort had an elevated cTnI level (p = 0.008). BNP was also significantly elevated in the Mustard group (Fig. 1B) with an average of 145.90 ± 55.07 pg/mL, while the control subjects were at 30.28 ± 5.01 pg/mL (p < 0.0001). None of the control subjects had an elevated BNP level, whereas 44% (n = 7) of the Mustard cohort had a BNP level > 100 pg/mL (p < 0.001). Of the seven (out of 16 samples) Mustard subjects with elevated BNP levels, 5 (71%) also had elevated sST2 levels. The BNP assay for two Mustard subjects was not available.
A significant correlation between the elevation of sST2 within the Mustard cohort and markers of cardiac injury or strain was noted ( Table 2). Of the Mustard subjects with elevated sST2 (n = 10), 60% had elevated cardiac troponin I (cTnI) compared with only 12% of those without elevation of sST2 (p = 0.042). Sixty percent of the Mustard subjects with elevated sST2 also had a significant elevation of BNP (> 100 pg/mL) compared with 12% of Mustard subjects without elevated sST2 (p = 0.037). Additionally, 90% of the patients with elevated sST2 levels had low HDL, whereas only 50% of the Mustards without sST2 elevation had depressed HDL levels (< 40 mg/dL) but this did not reach significance (p = 0.06).
We also examined the correlation between clinical parameters and elevated sST2 levels. There were no differences between Mustard patients with elevated sST2 (n = 10) and those without elevation of sST2 (n = 8) for use of cardiac medications, arrhythmias, or T-wave abnormalities such as  inversion or flattened T-waves in in the left chest leads on EKG. Of the 18 Mustard subjects, 9 were in class II of the NYHA scale, and 9 were classified as III. There was no significant relationship of NYHA classification and sST2 levels. Six of the NYHA class II subjects and 4 of the class III subjects had elevated sST2 levels (33 and 22% respectively).
In-depth correlations between biochemical measures in the Mustard population and the control subjects were explored. Table 3 shows a significant positive relationship between sST2 and cTnI with a correlation by Pearson's r of p < 0.0001 (R 2 = 0.839). Other significant correlations were seen between LDL and triglycerides, LDL, and total cholesterol (p = 0.0001, R 2 = 0.822), and the total cholesterol/HDL ratio. Interestingly, significant correlation was observed between glucose levels, triglycerides, and the cholesterol/ HDL ratio. No significant correlations were found for insulin levels.
Clinical outcomes of the Mustard subjects were determined at the most recent 5 years of follow-up (Table 4). Of those with elevated sST2, there was an increase in cardiac medication (n = 1) and arrhythmias (n = 2) although these did not reach significance. Of the Mustard patients without elevated sST2, there was no increase in cardiac medication usage and a small increase in arrhythmias (n = 1). Of the subjects with elevated sST2 levels, 70% had at least one hospitalization across 5 years and 100% (n = 10) had either an ablation, pacemaker implantation, or both. Of those without elevated levels, 50% had at least one hospitalization over 5 years and 75% (n = 6) had an ablation, pacemaker, or both. One subject died but did not have elevated sST2 levels, and one subject with elevated sST2 was transplanted.

Discussion
The primary finding of this study is that sST2 is significantly elevated in patients with D-TGA corrected by the Mustard procedure when compared with healthy individuals. In the Mustard cohort, 56% had an elevation of sST2 compared with only 17% of the control subjects. This is one of the few studies which has translated the value of sST2 in individuals with CHD, specifically D-TGA palliated with the Mustard procedure. Prior studies have also shown that in patients with complex CHD, sST2 is found to have significant association between adverse cardiovascular events [12], and these levels were elevated when compared to healthy individuals [17]. Control subjects, although defined as not having any cardiovascular disease but which also had elevated sST2 levels, could likely be explained by other factors, e.g., sST2 levels can also be elevated in cases of endurance exercise [18], obesity [19], previous surgeries, asthma, or other underlying illness. Other authors have recently noted the value and difficulty of identifying novel biomarkers for risk stratification with prognostic value in adults with a systemic right ventricle [20].
Of greater importance, we found that in the Mustard patients with elevated sST2, there were correlations with other biomarker abnormalities that are associated with clinical HF signs. Correlations between elevated sST2 and cTnI, as well as BNP, were seen. Cardiac troponin I elevation or leak is associated with ongoing myocardial damage [21]. Similar correlations were not found in the control subjects indicating the abnormal levels in the Mustard group reflect ongoing ventricular strain and/or injury.
When comparing averages of other biomarkers between Mustard and control groups, cholesterol levels were significantly different. Of the Mustard subjects with elevated sST2, 90% had abnormally low HDL levels and total cholesterol. The Mustard procedure has been known to cause liver impairment due to elevated systemic venous pressure from the atrial baffle [22] although depressed HDL levels have not been specifically noted before. Additionally, a 5-year follow-up demonstrated a trend in those Mustard subjects with elevated sST2 towards greater hospitalization, worsening arrhythmias, greater use of interventional studies to control arrhythmias, and greater medication usage, although none of these reached statistical significance due to the small cohort size. However, larger follow-up studies may include these outcomes in their analysis.

Conclusion
In this small cohort, patients with the Mustard operation had, on average, higher sST2 levels than healthy subjects. Importantly, in patients with elevated levels, there was also a significant association of sST2 with biomarker abnormalities. Limitations of this study include small cohort size, and lack of direct assessment for other conditions such as asthma, autoimmune diseases, or viral infections that may cause variation in the level of sST2. Additionally, we could not match the control group with the Mustard cohort for sex, which may introduce significant bias in the lipid and sST2 results [23]. sST2 is known to have a higher range in men (8.6-49.3 ng/mL) than women (7.2-33.5 ng/mL), and thus, diagnostic prediction would likely improve with a cohort matched for sex [24,25].
In the 5 years of follow-up, patients with elevated sST2 levels were associated with worsening clinical HF signs such as increased rates of hospitalization, complex arrhythmia management, and medication usage. These outcomes did not reach significance in this small study. Thus, sST2 may add predictive value to cardiac-related morbidity and mortality in patients with D-TGA palliated with Mustard procedure but larger studies are needed. Multivariate analysis in a larger cohort may allow better prediction of outcomes for these patients and improved clinical management.
Author Contributions HNF, HRM, and RMP wrote the manuscript. HNF and HRM contributed equally to this work. PMP performed the sST2 assays and reviewed the manuscript. HRM and EAS recruited subjects and performed chart review, and EAS reviewed the manuscript. RAH reviewed and edited the manuscript. RMP supervised the project. All authors have reviewed this manuscript.
Funding This project was funded in part with support from the Short-Term Training Program in Biomedical Sciences Grant, T35 HL 110854, from the National Institutes of Health (to HNF), 1P01HL134599 from NHLBI (to RMP), and the Indiana Clinical and Translational Sciences Institute, 5UL1TR002529 from NCATS. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.