Plasma SMOC2 Predicts Prognosis in Patients with Heart Failure: A Prospective Cohort

Introduction

Heart failure (HF) is a complex clinical syndrome with high prevalence and mortality caused by physiological and neurohormonal changes caused by abnormal cardiac structure or function, which has seriously increased the global medical burden.¹ Wheezing, edema, arrhythmias, and cardiac enlargement are some of the clinical manifestations of HF.

Left ventricular ejection fraction (LVEF), ventricular structure, body mass index(BMI), and biomarkers such as brain natriuretic peptide, troponin, and C-reactive protein are commonly used clinically to assess heart disease.^2,3 There is still much to be discovered about the prognosis of HF, as it is an extremely complex pathophysiological process.

Secreted modular calcium-binding protein 2 (SMOC2) is a member of the cysteine-rich acidic secretory protein (SPARC) in the stromal cell protein family, which contains an extracellular calcium-binding (EC) domain and a typical domain with 10 cysteines.⁴ SMOC2 interacts with matrix proteins, cell surface receptors, cytokines, proteasomes, and other effector molecules, which can regulate the interaction of cells and cell matrix, as well as the activity of vascular growth factors.^5,6 The SMOC2 protein may serve as a marker for HF, and interventions targeting SMOC2 may be effective in treating heart disorders. For example, Wilk et al⁷ in 2007 studied mRNA expressions in left and right ventricular samples without HF and right heart failure samples from patients with pulmonary arterial hypertension (PAH) and after LVAD implantation and revealed SMOC2 to be an interesting target for HF. Moreover, we have previously shown that SMOC2 was significantly increased in the process of myocardial fibrosis in mice, which could be reversed by SMOC2 down-regulation.⁸

There has yet to be a clinical study evaluating the association between plasma SMOC2 and HF severity and prognosis. The study presented here was undertaken to address the relationship between plasma SMOC2 levels and the severity of HF, as well as its diagnostic value in the prognosis of patients with end-stage heart abnormalities.

Study Design

In this prospective cohort study, inpatients that were diagnosed with ischemic heart disease in the Central Hospital of Enshi Tujia and Miao Autonomous Prefecture were recruited from July 2020 to December 2021. Ethics approval was obtained from the Ethics Committee of the Central Hospital of Enshi Tujia and Miao Autonomous Prefecture.

Participants

The demographic and clinical data of the patients were taken from their original medical records. The inclusion criteria of the present study were: 1. Those who were diagnosed with ischemic cardiomyopathy (ICM) with LVEF≤50%. 2. Those who were aged 18 or older. 3. Those who were willing to cooperate with our follow-up objective. Criteria for diagnosis of ICM: clinical diagnosis of coronary heart disease or angiography of coronary stenosis greater than 75% with left ventricular dilatation and reduced systolic function(LVEF≤50%), excluding other causes of left ventricular dilatation and reduced systolic function.⁹ This study excluded patients with the following clinical conditions: acute myocardial infarction (MI), thyroid dysfunction, malignant tumors, inflammatory bowel disease, severe arrhythmias, etc. Clinical data were collected prospectively along with baseline characteristics. All patients were followed up.

Follow-Up and Study End Points

From discharge until the end of follow-up, patients were seen by outpatient clinical consultations or telephone interviews every 3 months. During follow-up, the primary endpoint event was all-cause re-hospitalization, and secondary endpoint events included all-cause death, new arrhythmias (including atrial fibrillation, ventricular premature, ventricular tachycardia, ventricular fibrillation), or heart rhythm abnormalities etc., and all these had been recorded by electrocardiogram or holter electrocardiogram. The telephone follow-up was conducted by a medical professional who had been trained for two weeks and had one year of experience. A second attempt was made to contact patients who were not reachable in the first round. Upon being unable to reach the patient’s family members for a third time, treatment was discontinued and the patient was recorded as missing. In addition, a clinical endpoint event committee was established in order to ensure that endpoint events were objective and traceable.

Variables

The following information was collected and recorded: age, sex, height, weight, history of diabetes, history of arrhythmia, history of smoking, history of drinking wine, history of PCI heart failure duration, BMI, NYHA function class, systolic blood pressure (SBP), diastolic blood pressure (DBP), heart rate (HR), medications(including anti-platelet therapy, Statins, β-blocker, ACEI/ARNI/ARB), and length of hospital stay. The blood was collected to detect the clinical biochemical parameters, consisting of high-density lipoprotein (HDL), low density lipoprotein (LDL), total cholesterol (TC), triglycerides (TG), creatinine (Cr), hemoglobin(Hb) and hemoglobin (Hb). Left atrial diameter (LA), left ventricle end-diastolic diameter (LVEDd), right ventricular end-diastolic diameter (RVEDd), right atrial diameter (RA), fractional shortening obtained from short axis images (FS), and Left ventricular ejection fraction (LVEF) were measured using echocardiography. LVEF was measured using Modified Simpson’s algorithm. All laboratory data was collected by the same laboratory to the same standard. Roche Elecsys ^®NT-proBNP (Roche Diagnostics GmbH) was used for NT-proBNP determination. The SMOC2 was detected using a double antibody sandwich assay (Shanghai Enzyme-linked Biotechnology Co., Ltd., China).

Statistics Analysis

Participants were divided into three groups for further analysis according to the tertiles of SMOC2 concentration at admission. A continuous variable is presented as mean + standard deviation (SD) (Gaussian distribution) or median (range) (skewed distribution), and a categorical variable as Numbers and percentages. A one-way ANOVA, Chi-square test, and Kruskal–Wallis H-test were used to detect differences among three SMOC2 groups, where appropriate. The univariate and multivariate Cox proportional hazard models were employed to determine the correlation between the SMOC2 level and prognosis of the individuals included in the study. In model I, no variables were adjusted. In model II, age and sex were adjusted. All covariate adjustments were taken into account in model III. HR ratio and 95% confidence interval (CI) were calculated. The robustness of our results was verified by conducting a sensitivity analysis. We convert the SMOC2 level into a categorical variable and calculate the p for trend to verify the results obtained using the SMOC2 level as a continuous variable and check the possibility of nonlinearity. Since Cox proportional hazards regression model-based models are often associated with incapacity to handle nonlinear models, the nonlinearity between SMOC2 and the prognosis of HF was performed using a Cox proportional hazard regression model with cubic spline functions and smooth curve fitting (penalized spline method). Furthermore, a Pearson correlation analysis was performed on the relationship between SMOC2 and NT - proBNP, EF, LVEDd and length of hospital stay.

R statistical programming language (version 3.4.3), and EmpowerStats were applied to perform statistical analysis. A bilateral p < 0.05 was considered to be statistically significant.

Results

Patients Characteristics

During the study period of January to December 2021, 252 continuous participants with HF were enrolled, with 74 participants being excluded because of exclusion criteria. As a result, 188 participants were included in the study. Within an average follow-up time of 489.5 ± 88.3 days, 53 patients (28.2%) were all-cause re-hospitalization, 14(7.4%) patients died, and 37(19.7%) subjects developed malignant arrhythmias. Patients were categorized based on their SMOC2 tertiles into three groups. Table 1 presents their demographic and clinical characteristics.

Table 1 The Demographic and Clinical Characteristics of All Participants

Correlation Analysis

A Pearson correlation analysis was performed to determine the correlation between SMOC2 and NT-proBNP, LVEDd, EF, length of hospital stay, etc., in HF. The results showed that the level of SMOC2 was positively correlated with NT-proBNP(R=0.680, p<0.001), LVEDd(R=0.361, p<0.001), and length of hospital stay(R=0.415, p<0.001) and was negatively correlated with EF(R=−0.479, p<0.001) at baseline. Figure 1 depicts the correlation analysis results.

Figure 1 The correlation analysis of plasma SMOC2 between NT-proBNP, LVEDd, EF and length of hospital stay.

Univariate Analysis of the Relationship Between Variables and Rehospitalization Rate

Rehospitalization rates and potential confounders were identified using a univariate analysis. According to the criterion of p < 0.1, the following variables were considered in multivariate Cox regression: age, sex, history of smoking, history of AF, history of diabetes mellitus, medications, Cr, Hb, NT-proBNP, EF, DBP, HR, heart failure duration, length of hospital stay, NYHA function class, SMOC2, LA, LVEDd, RA, and FS. The results of the Univariate analysis are shown in Table 2.

Table 2 The Results of the Univariate Analysis

Multivariate COX Regression Analysis of SMOC2 and Readmission Risk of HF Patients

To examine the association between SMOC2 and the rate of readmission in HF patients, we used three models in multivariate COX regression analysis. A stable linear regression was observed when the SMOC2 level was a continuous variable. The fully adjusted model showed readmission risk increased by 2% (HR = 1.020, 95% CI:1.012–1.655, p=0.017). When SMOC2 was used as a categorical variable, we found a statistical difference between the Middle and High groups compared to the Low group in all three models (p < 0.05). In addition, the p for trend also showed a linear correlation across the three models (p < 0.05). The multivariate COX regression analysis is shown in Table 3. The Kaplan-Meier survival curve is shown in Figure 2.

Table 3 The Results of Multivariate COX Regression Analysis of SMOC2 and Readmission Risk of HF Patients

Figure 2 Kaplan–Meier curves demonstrating re-hospitalization stratified by tertiles of SMOC2.

Smooth Curve Fitting and Threshold or Saturation Effect Analysis

Furthermore, we performed a smooth curve fit of a weighted generalized addition model to detect the nonlinear association between SMOC2 and readmission risk during follow-up and further confirmed the results. In patients with HF, plasma SMOC2 was not associated with readmission risk in a nonlinear manner. In addition, we compared two fitted models using threshold or saturation effect analysis to explain the association between SMOC2 and rehospitalization. Ultimately, we found no threshold or saturation effect (p = 0.082). The smooth curve fitting and threshold or saturation effect analysis were shown in Figure 3 and Table 4.

Table 4 The Threshold or Saturation Effect Analysis

Figure 3 The smooth curve fitting analysis of plasma SMOC2 and re-hospitalization. Notes: Relationship between plasma SMOC2 and re-hospitalization. The red line represents the smooth curve fit between variables. In comparison, blue bands represent the 95% CI. Age, sex, history of smoking, history of AF, NT-proBNP, EF, DBP, HR, heart failure duration, length of hospital stay, NYHA function class, SMOC2, LA, LVEDd, and RA were adjusted.

Multivariate COX Regression Analysis of SMOC2 and All-Cause Mortality/ New-Onset Arrhythmias in HF Patients

In order to further clarify the relationship between plasma SMOC2 and the prognosis of patients with HF, we analyzed the relationship between SMOC2 and all-cause mortality and new-onset arrhythmias in this cohort study. Cox regression analysis and Kaplan-Meier survival curve analysis showed that SMOC2 was not associated with the risk of mortality (p=0.42) or new arrhythmias (p=0.69) in patients with HF. The Kaplan-Meier survival curves are shown in Figures 4 and 5.

Figure 4 The smooth curve fitting analysis of plasma SMOC2 and all-cause mortality. Notes: Age, sex, history of smoking, history of AF, NT-proBNP, EF, DBP, HR, heart failure duration, length of hospital stay, NYHA function class, SMOC2, LA, LVEDd, and RA were adjusted.

Figure 5 The smooth curve fitting analysis of plasma SMOC2 and new-onset arrhythmias. Notes: Age, sex, history of smoking, history of AF, NT-proBNP, EF, DBP, HR, heart failure duration, length of hospital stay, NYHA function class, SMOC2, LA, LVEDd, and RA were adjusted.

Discussion

Various outcomes result from HF, including heart failure recurrence, deterioration, arrhythmia, embolisms, and death. Numerous basic studies have identified SMOC2 as a member of SPARC that plays different roles in different systems. Studies on SMOC2, especially clinical studies, are scare, as are studies relating SMOC2 to heart diseases. SMOC2 is currently associated with lung diseases, rheumatoid arthritis, renal cell carcinomas, thyroid tumors, and asthma,^10–13 and can be used as a predictive biomarker for a variety of diseases. A previous study had compared the gene expression profiles of patients with HF (n=177) and non-HF (n=136) through bioinformatics analysis and identified 38 HF characteristic genes, and their results suggested that SMOC2 plays a significant role in HF.¹⁴ Additionally, SMOC2 is up-regulated in ISO-induced HF models, suggesting that SMOC2 may regulate myocardial remodeling.⁸

Main Findings

This prospective cohort study examined the association of plasma SMOC2 levels with severity and prognosis in 188 patients with ICM. During the 540-day follow-up, 28.2% of patients were re-hospitalized due to cardiac deterioration, 7.4% died, and 19.7% developed new-onset arrhythmias. There were three important findings from this study: Elevated plasma SMOC2 levels increased the risk of readmission in patients with HF (HR = 1.02, 95% CI:1.012–1.655). SMOC2 levels positively correlated with the severity of HF. SMOC2 was not associated with all-cause death or malignant arrhythmias during follow-up.

Firstly, to explore the utility of SMOC2 in evaluating and diagnosing the severity of HF, we used Pearson correlation analysis in order to examine the correlation between SMOC2 and indicators that represent the disease severity such as NT-proBNP, LVEDd, LA, EF, FS, and length of hospital stay. SMOC2 level was positively correlated with NT-proBNP, LVEDd, and length of hospital stay, and negatively correlated with EF and FS. This study showed that the higher the SMOC2 level, the more severe the decline of cardiac function, the more significant the enlargement of the left ventricle, and the longer the length of hospital stay in patients with heart disorders. SMOC2 is a marker of myocardial fibrosis,⁸ and it was confirmed for the first time in this study that SMOC2 can be used to evaluate the severity of HF, which will provide an important reference for molecular markers of HF in practice.

Furthermore, three COX regression models were used to examine the relation between plasma SMOC2 and prognosis in patients with HF. The results showed that the risk of readmission increased for increase in SMOC2 in the fully adjusted model. In order to detect the nonlinear relationship between SMOC2 and readmission rate, we used trend test, weighted generalized addition model for smooth curve fitting, threshold saturation effect. The results showed that there was no nonlinear association between SMOC2 and the risk of rehospitalization. Our study is the first to suggest that plasma SMOC2 is associated with prognoses in patients with cardiac dysfunction. The potential role of SMOC2 in heart conditions has been suggested by previous studies. Zhou et al¹⁵ used weighted gene co-expression network analysis (WGCNA) to identify new biomarkers and biological pathways and found that SMOC2 genes were highly correlated with HF caused by dilated cardiomyopathy. Previous reports also found that the high expression of SMOC2 was related to cardiac fibrosis in chronic Chagas disease cardiomyopathy.¹⁶ In this study, we found that SMOC2 may be related to the prognosis of patients with ICM. It is suggested that in addition to detecting NT-proBNP and cardiac ultrasound, SMOC2 may also be of great significance in clinical practice.

Thirdly, we attempted to elucidate the association between SMOC2 and all-cause mortality and new-onset arrhythmia in ICM patients by using Cox regression analysis and Kaplan-Meier survival curve analysis, and found that SMOC2 was not associated with the risk of mortality (p=0.42) or new-onset arrhythmia (p=0.69) in patients with heart failure. A larger prospective study may be required to further confirm this finding due to the limited participants included in this study and the short period of follow-up.

SMOC2 has been used as a new biomarker to predict the severity and outcome of different diseases. Studies show that SMOC2 is independent of prognosis in colorectal cancer patients.¹⁷ Schmidt et al¹⁸ found that SMOC2 is a promising non-invasive biomarker for renal fibrosis. At the same time, many studies have confirmed that SMOC2 can promote tissue fibrosis by regulating the transformation of fibroblasts into myofibroblasts, which can lead to fibrosis in lungs, kidneys, and other organs.^19,20 We also clearly elaborated that downregulation of SMOC2 can alleviate myocardial fibrosis by inhibiting the ILK/p38 signaling pathway through animal experiments, indicating that high expression of SMOC2 can lead to myocardial fibrosis.⁸ In the pathophysiology and progression of HF, myocardial fibrosis is an important mechanism for the occurrence and development of the cardiac remodeling. Fibrosis of the myocardium can lead to abnormal cardiac structure and left ventricular dysfunction.²¹ It has also been demonstrated in animal studies that inhibiting SMOC2 expression inhibits cardiomyocyte apoptosis, suggesting that high levels of SMOC2 can result in apoptosis in cardiomyocytes indirectly.²² Similarly, cardiomyocyte apoptosis is also involved in the pathophysiological process of heart failure. The loss of cardiomyocytes caused by non-regenerative cardiomyocytes can result in severe heart failure if non-regenerative cardiomyocytes are lost through apoptosis.²³ At present, some relevant studies have found that SMOC2 has a certain role in inflammation-mediated reactions. According to the latest studies, SMOC2 has a significant anti-inflammatory effect on rheumatoid arthritis and nerve transmission.^11,24 In a study on SMOC2 and the prevention of renal function loss in a mouse model of chronic kidney disease,²⁵ it was found that SMOC2 could play a central role in fibrogenesis and inflammation by activating a variety of signaling pathways and autophagy in vivo and in vitro, and therapeutic down-regulation of SMOC2 could play an anti-fibrosis and anti-inflammation role. The role of inflammation in cardiac remodeling has been studied for decades, and it has been confirmed that inflammation is involved in the occurrence and development of HF.²⁶

We can therefore reasonably assume that SMOC2 is highly correlated with HF progression and deterioration. A high level of SMOC2 expression can cause inflammation, myocardial fibrosis, and apoptosis, ultimately leading to ventricular dysfunction, heart enlargement, and adverse cardiac remodeling. Clinical manifestations include worsening symptoms of heart failure and deterioration of the condition, complications, and increased hospitalization rates.

Clinical Implications

This study assessed the relationship between SMOC2 and HF by analyzing the expression of plasma SMOC2 in patients with ICM. As a result of these findings, we will be able to build on previous pre-clinical studies and advance our understanding of the relationship between SMOC2 and HF and other cardiac conditions in the future. In addition to its complex etiology, HF is characterized by variable conditions, a long course of disease with many adverse outcomes, and a very poor prognosis. There is still a need to explore more biomarkers to predict and evaluate the prognosis of patients with heart failure and guide treatment. As a secreted protein, SMOC2 is widely distributed in many organs and tissues, and its plasma concentration is easy to obtain in clinic. As a result of this study, we found that SMOC2 levels were positively correlated with rehospitalizations in HF sufferings during follow-up, making it a good predictor in clinical practice. In addition to providing new treatment options, these results lay the groundwork for future study of deeper pathogenesis of HF.

Limitations

In this study, our prospective cohort study included a small sample size of patients with ICM and a relatively short follow-up time, resulting in few endpoints such as death and malignant arrhythmias, and failed to find the association between SMOC2 and mortality and malignant arrhythmias. More large-scale data and prospective studies are needed to confirm this, as residual confounding could affect regression analysis.

Conclusion

Increased plasma SMOC2 levels increase the risk of readmission in HF. Plasma SMOC2 is associated with the severity and deterioration of HF and is a good predictor of the risk of readmission in patients.