Abstract
BACKGROUND: This study aimed to investigate the relationship between baseline soluble suppression of tumorigenesis-2 (sST2) concentration and the outcomes of heart failure (HF), atrial fibrillation (AF) or death in patients with coronary heart disease (CHD) with or without renal insufficiency (RI). METHODS: Between March 2011 and December 2015, 3454 patients with CHD from the Chinese PLA General Hospital were enrolled in this cohort study. The patients were followed up until October 2021. AF, HF, and death events were recorded. Associations between baseline sST2 concentrations and clinical outcomes were assessed using Kaplan-Meier (K-M) curves, and Cox regression and generalised additive models. Subgroup analysis were carried out between RI and non-RI groups. RESULTS: Among the patients with CHD (61.5 ± 11.8 years; 78.6 % men), 415 (12.02 %) had RI. During a median follow-up of 8.37 years, HF and AF were reported in 216 (6.25 %) and 174 (5.04 %) patients, respectively, and 297 (8.60 %) died. The K-M curves indicated that patients in the higher quartiles of sST2 concentrations were correlated with a poor survival rate of HF, AF, or death (all Ps < 0.001). Generalised additive model (GAM) demonstrated a nonlinear positive association between sST2 concentration and the risk of HF, AF, and death in CHD patients. The cut-off value of sST2 for predicting HF, AF and death were 32.1, 25.4 and 28.6 ng/mL, respectively. CHD patients with sST2 higher than the cut-off value had higher risks of HF (HR: 3.02, 95%CI: 2.24-4.05), AF (HR: 2.86; 95%CI: 2.10-3.90), and death (HR:2.11, 95%CI: 1.67-2.67). Furthermore, in patients with RI (12.02 %, n = 415), the prognostic value of sST2 over the cut-off value for HF and death remained unchanged (HR: 3.21 and 2.35; P < 0.05). In patients with CHD with or without RI, sST2 improved the area under the curve (AUC) of traditional risk models for predicting clinical endpoint events. CONCLUSIONS: The biomarker sST2 may be useful for predicting HF, AF, and death in patients with CHD. The predicted value was not affected by renal function.