Table of Contents:
sST2 and Galectin-3 have been shown to be pathogenetically involved in cardiac remodeling and several studies indicate a prognostic value of both proteins in heart failure (HF). Prolactin is also described as a predictor of HF, although the current data situation is very limited.
We investigated the role of the proteins in patients with dilated cardiomyopathy (DCM) as a non-ischemic cause of HF.
Therefore, we included 262 subjects with DCM enrolled in the sub-project 9a of the German Competence Network Heart Failure between December 2004 and December 2008.
Inclusion criteria were age between 18 and 70, left ventricular ejection fraction (LVEF) <45% and left ventricular end-diastolic diameter (LVEDD) ≥117% according to the formula of Henry. Coronary heart disease, valvular heart disease and arterial hypertension accompanied by end organ damage applied as criteria for exclusion.
The patients were classified into different etiologies. Viral cardiomyopathy was present if viral DNA or RNA was verified in endomyocardial biopsy. The term inflammatory cardiomyopathy was quantitative immunohistologically specified either as foci of lymphocytes and/or ≥ 14 lymphocytes and macrophages/mm2 or if the endomyocardial biopsy was positive for conventional histopathological criteria using the Dallas classification.
Subjects with viral evidence or inflammation or both we classified as the group inflammatory and/or viral DCM (DCMi⋎viral).
Patients with positive family history regarding dilated heart disease using the Mestroni criteria were categorized in the group familial DCM (fDCM).
Subjects without viral or inflammatory evidence and without familial background formed the group idiopathic DCM.
Clinical 1-year and 5-year follow-up tests were carried out. The primary endpoints were all-cause mortality (ACM), cardiac mortality (CM) and the combined endpoint cardiac failure (CF), here defined as CM, heart transplantation, reanimation, defibrillation or adequate ICD intervention.
sST2 and Galectin-3 values of the time of inclusion were measured for all 262 subjects, Prolactin concentrations for 166 subjects using quantikine ELISA kits of R&D Systems.
Statistical analyses were performed using the software R.
In order to estimate correlations between sST2, Galectin-3 and Prolactin and selected covariables two-sided univariate linear regression analyses were conducted and the correlation coefficient r and the corresponding p-value reported.
The association between the biomarkers and ACM, CM and CF was assessed with a univariate Cox-regression model and Hazard-Ratios (HR), 95% confidence intervals (CI) and p-values were given.
In the total group with all patients, analyses were performed using the biomarkers as continuous variables and a quartile model was created as well. In subgroups classified according to etiology and in a further analysis subdivided into sex-specific groups, the biomarkers were investigated as continuous variables.
Based on Cox-regression a multivariate model was created for the endpoints ACM and CF in the overall group to adjust for age, sex, body mass index (BMI), NYHA class, QRS duration, LVEF, LVEDD, etiology and diabetes mellitus.
Linear regression analysis showed a weak positive relationship between sST2 and Galectin-3 (r=0.143 p=0.021). Further correlations insisted between sST2 and white blood cell count (r=0.166 p=0.007), LVEF (r=-0.165 p=0.007) and B-type natriuretic peptide (BNP) (r=0.362 p<0.001). Between sST2 and the variables Prolactin, age, BMI, C-reactive protein (CRP), creatinine and LVEDD no significant correlation was noticed.
Besides the association with sST2, Galectin-3 correlated with the variables Prolactin (r=0.261 p<0.001), age (r=0.127 p=0.039), white blood cell count (r=0.155 p=0.012) and creatinine (r=0.198, p=0.001).
Except for the correlation with Galectin-3, no further association was observed between Prolactin and any other variable.
With regard to the survival analysis of the total group of patients, higher levels of sST2 were a predictor of ACM in the univariate quartile model and as a continuous variable sST2 was a predictor of the endpoints ACM (HR=1.05 CI=1.03-1.07 p<0.001), CM (HR=1.03 CI=1.00-1.06 p=0.04) and CF (HR=1.04 CI=1.02-1.07 p=0.001).
In the multivariate model, sST2 remained significant for the endpoint ACM (HR=1.04 CI=1.02-1.07 p=0.003).
Split into sex-specific subgroups, in the group of female patients higher sST2 values were associated with adverse outcomes for ACM (HR=1.06 CI=1.01-1.12 p=0.022), CM (HR=1.07 CI=1.00-1.13 p=0.042) and CF (HR=1.06 CI=1.02-1.11 p=0.009). In the group of male patients, a significant result was only achieved for the endpoint ACM (HR=1.05 CI=1.02-1.07 p<0.001).
Divided according to etiology, in the group of patients with idiopathic DCM, sST2 was associated with ACM (HR=1.04 CI=1.01-1.07 p=0.019). In the group of patients with DCMi⋎viral, the endpoints ACM (HR=1.10 CI=1.05-1.17 p<0.001), CM (HR=1.10 CI=1.02-1.18 p=0.013) and CF (HR=1.08 CI=1.01-1.14 p=0.021) were significant whereas in the group of patients with fDCM no significant associations between sST2 and any of the endpoints were observed.
Survival analysis of Galectin-3 as a continuous variable neither has shown significant results for the overall group nor for any of the subgroups. In the quartile model, Galectin-3 was significant for ACM and CM, whereby their third quartiles were associated with better outcomes.
Univariate, Prolactin was only significant in the subgroup of female patients for the endpoints ACM (HR=0.7 CI=0.52-0.95 p=0.023) and CM (HR=0.69 CI=0.49-0.99 p=0.044). Multivariate, higher Prolactin levels were significantly associated with a lower event rate of ACM (HR=0.90 CI=0.81-1.00 p=0.009).
In conclusion, the study revealed that sST2 is a convenient predictive biomarker in patients with non-ischemic HF. Further studies are needed to assess if sST2 improves discrimination and reclassification analysis in non-ischemic HF when attached to clinically established biomarkers.
The findings of the total group with all patients, that intermediate levels of Galectin-3 allow for better prognosis are in contrast with other studies and should be proved in further trials.
Data regarding Prolactin in connection with HF are very limited. Therefore, further research is necessary to clarify the prognostic role of Prolactin in patients with HF.