Increased Plasma Concentrations of Soluble ST2 Independently Predict Mortality but not Cardiovascular Events in Stable Coronary Heart Disease Patients: 13-Year Follow-up of the KAROLA Study

Purpose: sST2 (soluble suppression of tumorigenicity 2), a member of the interleukin-1 family, has been suggested to play a role in cardiac remodeling and inflammatory signaling. We assessed the association between sST2 in patients with stable coronary heart disease (CHD) with multiple cardiovascular outcomes and total mortality, simultaneously controlling for a large number of potential confounders.

Methods: Plasma concentrations of sST2 (ELISA, Critical Diagnostics) were measured at baseline in a cohort of 1081 patients. The Cox-proportional hazards model was used to determine the prognostic value of sST2 on a combined cardiovascular disease (CVD) endpoint, on cardiovascular death, and on total mortality after adjustment for covariates.

Results: The median sST2 level was 28.9 ng/mL (IQR 23.8, 35.1) (mean age at baseline 58.9 years, 84.6% male). sST2 concentration was positively correlated with inflammatory markers and emerging risk factors, e.g., cystatin C, N-terminal probrainnatriuretic peptide (NT-proBNP), high-sensitivity (hs)-Troponin T and I, mid-regional pro-atrial natriuretic peptide (MR-proANP), and growth differentiation factor 15 (GDF-15). Results after short- and long-term (4.5 and 12.3 years, respectively) follow-up (FU) displayed no statistically significant association with the combined endpoint of non-fatal and fatal CVD events when the top quartile (Q4) of sST2 concentration was compared to the bottom quartile (Q1). A relationship during long-term FU was seen with CVD mortality even after multivariable adjustments including clinical risk variables (HR 1.65; 95% CI 1.02-2.86), but not in a fully adjusted model whereas, in contrast, it was still highly significant after short-term FU (HR (5.97 (95%CI 1.32-27.06)). In addition, the sST2 concentration was still strongly associated with total mortality in the fully adjusted model including clinical variables and cystatin C based estimated glomerular filtration rate, NT-proBNP, hsCRP and hs-TnI comparing Q4 vs Q1 during long-term FU (HR of 1.48 (95% CI 1.03-2.13)) and short-term FU (HR 3.06 (95% CI 1.29-7.24)).

Conclusions: Elevated levels of sST2 concentration in stable CHD patients may independently predict short- and long-term risk for fatal CVD events and total mortality but not non-fatal CVD events.