GDF‐15 is a better complimentary marker for risk stratification of arrhythmic death in non‐ischaemic, dilated cardiomyopathy than soluble ST2

Abstract Growth differentiation factor (GDF)‐15 and soluble ST2 (sST2) are established prognostic markers in acute and chronic heart failure. Assessment of these biomarkers might improve arrhythmic risk stratification of patients with non‐ischaemic, dilated cardiomyopathy (DCM) based on left ventricular ejection fraction (LVEF). We studied the prognostic value of GDF‐15 and sST2 for prediction of arrhythmic death (AD) and all‐cause mortality in patients with DCM. We prospectively enrolled 52 patients with DCM and LVEF ≤ 50%. Primary end‐points were time to AD or resuscitated cardiac arrest (RCA), and secondary end‐point was all‐cause mortality. The median follow‐up time was 7 years. A cardiac death was observed in 20 patients, where 10 patients had an AD and 2 patients had a RCA. One patient died a non‐cardiac death. GDF‐15, but not sST2, was associated with increased risk of the AD/RCA with a hazard ratio (HR) of 2.1 (95% CI = 1.1‐4.3; P = .031). GDF‐15 remained an independent predictor of AD/RCA after adjustment for LVEF with adjusted HR of 2.2 (95% CI = 1.1‐4.5; P = .028). Both GDF‐15 and sST2 were independent predictors of all‐cause mortality (adjusted HR = 2.4; 95% CI = 1.4‐4.2; P = .003 vs HR = 1.6; 95% CI = 1.05‐2.7; P = .030). In a model including GDF‐15, sST2, LVEF and NYHA functional class, only GDF‐15 was significantly associated with the secondary end‐point (adjusted HR = 2.2; 95% CI = 1.05‐5.2; P = .038). GDF‐15 is superior to sST2 in prediction of fatal arrhythmic events and all‐cause mortality in DCM. Assessment of GDF‐15 could provide additional information on top of LVEF and help identifying patients at risk of arrhythmic death.

primary prevention with implantable cardioverter defibrillators (ICD) are based on the degree of LV ejection fraction (LVEF ≤ 30%-40%) reduction. 3,4 However, patients with ischaemic heart disease seem to have more benefit from prophylactic ICD implantation. [5][6][7] Furthermore, the recently published DANISH trial has demonstrated no mortality benefit from prophylactic ICD implantation in non-ischaemic DCM. 8 Although LVEF is currently the best marker for risk stratification, it lacks specificity and many DCM patients with ICDs never receive appropriate therapies. 9,10 It is well known that VT also occurs in patients with borderline or mildly reduced LVEF. 3 Noninvasive testing and LVEF could not reliably identify patients with DCM at risk of fatal VTs. 11 Therefore, the identification of additional markers for arrhythmia risk stratification of patients with DCM is essential.
Growth differentiation factor (GDF)-15 and soluble ST2 (sST2) are well-established prognostic markers for mortality in acute and chronic heart failure [12][13][14][15][16][17] as well as in acute myocardial infarction. 18 Inflammation and myocardial fibrosis, with subsequent ventricular remodelling and impairment of systolic function, are important pathophysiological mechanisms for VTs in patients with DCM. 19,20 Both GDF-15 and sST2 show strong correlations with myocardial stress and fibrosis, [21][22][23] and have been associated with sudden cardiac death in DCM. 24,25 The aim of the study was a head-to-head comparison of GDF-15 and sST2 for prediction of arrhythmic death (AD) and all-cause mortality in patients with non-ischaemic DCM.

| Study participants
This was a prospective, longitudinal cohort study. A total of 52 consecutive patients with non-ischaemic DCM were included in the study in 2002 and 2003 at the Medical University of Vienna, Austria. Patients were eligible to participate if they were aged ≥18 years, had a LVEF of ≤50%, had recently undergone coronary angiography with ventriculography as standard use of care independently from study participation, echocardiography, MRI at the physician's discretion, had no history of sustained ventricular arrhythmia or permanent atrial fibrillation and were not dependent on ventricular pacing. Ambulatory ECG, Holter recordings and exercise tests were performed at baseline in all patients. New York Heart Association (NYHA) classification was documented for each patient at the baseline according to ESC and AHA Heart Failure Guidelines: NYHA Class I: no limitation of physical activity; NYHA Class II: slight limitation of physical activity in which ordinary physical activity leads to fatigue, palpitation, dyspnoea, or anginal pain; the person is comfortable at rest; Class III: marked limitation of physical activity in which less-than-ordinary activity results in fatigue, palpitation, dyspnoea, or anginal pain; the person is comfortable at rest; Class IV: inability to carry on any physical activity without discomfort but also symptoms of heart failure or the anginal syndrome even at rest, with increased discomfort if any physical activity is undertaken. 4,26 Patients were followed up between 2003 and 2013. The study was approved by the local ethics committee, and all participants gave written informed consent.

| Biomarker measurements
Venous blood samples were obtained from each patient at study admission. EDTA blood samples were immediately centrifuged at 1500 g for 15 minutes, plasma was aliquoted and stored at À80°C for further use. The samples did not undergo any freeze-thaw cycles before the performance of the assays. Plasma concentration of GDF- sST2 was quantified using human ST2/IL-1 R4 DuoSet â ELISA Kit (R&D Systems). 28,29 ST2/IL-33R ELISA Kit has an intra-assay variability of 4.4%-5.6% and an inter-assay variability of 5.4%-7.1%. This ELISA has a sensitivity of 13.5 pg/mL, and it is specific for natural and recombinant human ST2 as well as free ST2 and IL-33 complexed ST2.

| Echocardiography
Echocardiographic examinations following a standard protocol 30 were conducted using a Vivid ultrasound machine (GE Healthcare, Milwaukee, WI, USA) or an ACUSON Sequia (Acuson, Mountain View, CA, USA). Following the practice at our institution, cardioversion was attempted in all patients. This time window, when the patient was in sinus rhythm, was used for echocardiographic examinations. Interventricular septal thickness was obtained at end-diastole from two-dimensional directed M-mode in the parasternal long axis, according to American Society of Echocardiography guidelines. 31,32 Two-dimensional apical 4-and 2-chamber views were used to calculate LVEF using the biplane Simpson's method. An LVEF <55% was considered abnormal. The severity of mitral regurgitation was evaluated semiquantitatively from the area of the regurgitant jet by colour Doppler.
LA anteroposterior diameter in parasternal long-axis view and LA major axis in apical 4-chamber view were used to calculate LA diameters.

| End-points
The primary end-point was time to AD or resuscitated cardiac arrest (RCA). All-cause mortality was the secondary end-point. Deaths were STOJKOVIC ET AL.
| 2423 categorized using an adapted form of the Hinkle classification. 33 Appropriate ICD therapy without VT acceleration that failed to save the patient's life at the time of arrhythmias was classified as AD. 34 An RCA was ventricular fibrillation or VT > 240 beats per minute leading to syncope before ICD therapy, and multiple slower VT episodes (electrical storm) leading to syncope and ICD discharge without ICD therapy-related acceleration. All other ICD therapies because of VT < 240 beats per minute were not taken as surrogate for AD. Study end-point data were collected through pre-planned ambulatory visits. In case of non-appearance study, end-point data were collected through treating physicians or patients 0 relatives, and in case of death through abduction, which is mandatory in Austria.   No patients were lost during follow-up. Cardiac death was observed in 20 patients, and 1 patient died a non-cardiac death. Of 20 cardiac deaths, 10 patients died an arrhythmic death and 10 patients died due to pump failure. Nine patients were implanted with a prophylactic ICD, and 10 patients were implanted with a CRT at study entry.

| Statistical analysis
The ICD therapy was unable to stop an electrical storm in 4 patients, and those patients died an AD. Reasons for not implanting an ICD during later follow-up in a patient with LVEF ≤ 30% were patients' refusal or the treatment policy of the attending physician. In addition, 2 ICD patients experienced an RCA. Overall, our primary endpoint was observed in 12 patients (10 AD and 2 RCA). Half of the patients with AD/RCA had a LVEF > 30%, and those patients were equally distributed in subgroups LVEF 31%-40% and 41%-50% (

| DISCUSSION
We compared the prognostic value of GDF-15 and sST2 for prediction of fatal ventricular arrhythmias and all-cause mortality in patients with non-ischaemic DCM. In contrast to sST2, GDF-15 was  Prophylactic ICD implantation in patients with non-ischaemic DCM did not improve long-term survival in a recently published DANISH trial. 8 However, almost twice as much patients died a sudden cardiac death in a control group as compared to the ICD group. 8 These recent findings highlight the importance to define an improved risk stratification strategy for prophylactic ICD implantation in patients with non-ischaemic DCM.
To best of our knowledge, this is the first study with a head-to-   37 Thus, the prognostic value of sST2 might be its ability to predict heart failure, rather than fatal VTs in patients with non-ischaemic DCM. 46,48 In agreement with previously published data, sST2 independently predicted all-cause mortality in the present study. Similarly, GDF-15 levels were also independently associated with all-cause mortality, as was shown previously. 14,17 Uric acid was previously shown to be a powerful prognostic marker in patients with chronic heart failure, and uric acid levels were associated with GDF-15. 14,49 In the present study, both GDF-15 and sST2 were superior to uric acid for prediction of all-cause mortality. In contrast to sST2, GDF-15 predicted allcause mortality independently of NT-proBNP. These results are in agreement with previous studies. 14,21 Finally, in a model including sST2 and GDF-15, only latter remained significantly associated with all-cause mortality in non-ischaemic DCM patients. These results suggest that in non-ischaemic heart failure, GDF-15 might be superior to sST2 in predicting all-cause mortality.