Biomarkers to Predict the Response to Cardiac Resynchronization Therapy

Ward Heggermont; Angelo Auricchio; Marc Vanderheyden


Europace. 2019;21(11):1609-1620. 

In This Article

Cardiac Biomarkers in Resynchronization Therapy

Brain-derived natriuretic peptide (BNP) and its amino-terminal linked counterpart (NT-pro-BNP) are widely accepted diagnostic and prognostic markers in HF patients.[1] Therefore, their potential to predict a response to resynchronization therapy has been investigated. In 2013, Brenyo et al. investigated the predictive value of BNP in a subgroup of the landmark Multicenter Automatic Defibrillator Implantation with Cardiac Resynchronization Therapy (MADIT-CRT) trial study, with mildly symptomatic HF.[3,4] Both baseline and 1 year of follow-up BNP levels were assessed in 1197 patients. As expected, elevated baseline BNP was associated with a 68% increased risk of HF or death (P = 0.007) in patients allocated to CRT-D as well as in those receiving implantable cardioverter-defibrillator (ICD)—in the ICD-only group, the increased risk of HF or death was 58% (P = 0.02). Importantly however, at 1 year of follow-up, patients who received CRT showed significantly greater reductions in BNP levels (26%) compared to ICD-only patients (8% increase, P = 0.005 for the difference). Moreover, in the CRT-D group, lower 1-year BNP levels were associated with a significantly lower risk of HF or death, compared to the subgroup where BNP levels remained high. Furthermore, the echocardiographic response to CRT-D was the highest in patients maintaining or attaining low BNP levels at 1 year of follow-up.[3] In the same patient group, elevated baseline and follow-up BNP levels were independent predictors of increased risk for ventricular tachyarrhythmias [ventricular tachycardia/ventricular fibrillation (VT/VF)], whereas reduced BNP levels following CRT-D implantation coincided with a lower incidence of VT/VF during follow-up.[5] Further evidence for the usefulness of BNP as a predictor of the CRT response was obtained in a prospective follow-up study of 267 HF patients with an average LVEF of 25% ± 8% (mean + SD) undergoing CRT-D implantation.[6] Both individually and in combination, baseline high-sensitivity troponin T (hsTnT) and BNP values—in a multivariate Cox regression model including age, NYHA class, LVEF, and QRS duration—were independent predictors of outcome, defined as death or HF hospitalization. A risk category based on the elevation of two, only one, or none of the investigated biomarkers, proved a significant predictor of outcome, with respective hazard ratios (HRs) of 7.34 [95% confidence interval (CI) 2.48–21.69] and 2.50 (95% CI 1.04–6.04) for high- and intermediate risk groups. Also, in a smaller study in which 105 HF patients (68% men, aged 65.4 ± 10.1 years) were followed for BNP levels and inflammatory markers, lower BNP levels were observed in both the objective responders to CRT [defined as a reduction of ≥15% in left ventricular end-systolic volume (LVESV)] and in the subjective response to CRT (defined as an improvement of ≥10 points on the patient-reported Kansas City Cardiomyopathy Questionnaire).[7] Altogether, these studies provide convincing evidence for the use of BNP as a predictor of the response to CRT.

Soluble suppressor of tumorigenicity 2 (sST2) is a protein that is encoded by the IL1R1 gene (interleukin 1 receptor 1) and it basically reflects adverse cardiac remodelling and fibrosis. Its interesting role as a biomarker in HF has been described elsewhere,[8] but its involvement in the CRT response prediction was also assessed in a subpopulation of the MADIT-CRT trial (n = 410, NYHA Class I/II). In multivariate-adjusted models, elevated baseline sST2 was associated with an increased risk of death, death or HF, and death or ventricular arrhythmia, even when adjusting for baseline BNP levels.[9] Furthermore, lower baseline sST2 levels coincided with a greater risk reduction with CRT-D (P = 0.006). This was confirmed upon measuring sST2 serially over time. Unfortunately, no data are available about serial measurements of sST2 after CRT and how baseline sST2 levels predict the response to CRT.

Galectin-3, also a marker of fibrosis, is a soluble beta-galactoside-binding lectin that has a regulatory function in fibrosis, tissue repair, and inflammation.[10] Again, this biomarker was tested in a subgroup of the MADIT-CRT population (n = 654, NYHA I/II)[11] with non-fatal HF events or death as study endpoints. Patients having a baseline galectin-3 level in the upper quartile of the distribution, also had a 65% reduction in the occurrence of the primary endpoint if given CRT-D (HR 0.35, 95% CI 0.19–0.67). In contrast, patients having lower galectin-3 levels at baseline only had a 25% non-significant risk reduction (HR 0.75, 95% CI 0.51–1.11), meaning that the patients with the highest risk for HF had the greatest benefit of CRT implantation. Baseline galectin-3 level was an independent predictor of outcome (HR 1.55, 95% CI 1.01–2.38; P = 0.043).[11] Similar results were obtained in a subpopulation of the CARE-HF trial,[12] showing galectin-3 as an independent predictor of worse overall cardiovascular outcome, however not predicting the response to CRT as a separate outcome parameter.[13] Although caution needs to be taken when interpreting all these results together, it seems that sST2 and galectin-3, markers of fibrosis and to some extent ventricular remodelling, better reflect the ongoing processes during successful resynchronization therapy, compared to NT-pro-BNP, BNP, or troponin. However, a head-to-head comparison in a prospective follow-up study would be helpful to determine the most specific cardiac biomarker related to the CRT response.

The abovementioned biomarkers (NT-pro-BNP, hsTnT, galectin-3, and sST-2) were also tested for their predictive value in estimating the improvement of mitral regurgitation (MR) after implanting a CRT device in 132 patients. From the BIOCRT study,[14] it appeared that higher galectin-3 levels at the time of CRT implantation conveyed a MR non-improvement (status quo or worsening) after 6 months. Although these patients also had higher hsTnT at baseline, after multivariate analysis only galectin-3 prevailed as a statistically significant predictor of MR evolution. Conversely, in the patients who had an improvement in MR, absolute levels of NT-pro-BNP and sST2 were lower at follow-up however without reaching statistical significance.[14]

Figure 1.

Central figure illustrating the presumed pathophysiologic pathways involved in left ventricular reverse remodelling, and hence the response to cardiac resynchronization therapy. BNP, brain natriuretic peptide; CRT, cardiac resynchronization therapy; CT, carboxyterminal; hs, high-sensitive; LV, left ventricle; miR, microRNA; N/L, neutrophil/lymphocyte; NT, amino-terminal; RV, right ventricle; ST, soluble factor of tumorigenicity; TIMP, tissue inhibitor of metalloproteinase.