Characterisation of circulating biomarkers before and after cardiac resynchronisation therapy and their role in predicting CRT response: the COVERT-HF study

Abstract

Aims: Cardiac resynchronisation therapy (CRT) is effective treatment for selected patients with heart failure (HF) but has ~30% non-response rate. We evaluated whether specific biomarkers can predict outcome.

Methods: A prospective single-centre pilot study of consecutive unselected patients undergoing CRT for HF between November 2013 and December 2015 evaluating cardiac extracellular matrix biomarkers and micro-ribonucleic acid (miRNA) expression before and after CRT assessing ability to predict functional response and survival. Each underwent three assessments (pre-implant, 6 weeks and 6 months postimplant) including: New York Heart Association (NYHA) class, echocardiography, electrocardiography, 6 min walk test (6MWT), Minnesota Living with Heart Failure Questionnaire (MLHFQ) and N-terminal pro-brain natriuretic peptide (NT-pro-BNP). Plasma markers of cardiac fibrosis assessed were: N-terminal pro-peptides of collagen I and III, collagen I C-terminal telopeptides (CTx) and matrix metalloproteinases (MMP-2 and MMP-9) as well as a panel of miRNAs (miRNA-21, miRNA-30d, miRNA-122, miRNA-133a, miRNA-210 and miRNA-486).

Results: A total of 52 patients were recruited; mean age (±SD) was 72.4±9.4 years; male=43 (82.7%), ischaemic aetiology=30 (57.7%), mean QRS duration=166.4±23.5 ms, left bundle branch block (LBBB) morphology = 39 (75.0%), mean NYHA=2.7±0.6, 6MWT=238.8±130.6 m, MLHFQ=46.4±21.3 and left ventricular ejection fraction (LVEF)=24.3%±8.0%. Mean follow-up=1.7±0.3 and 5.8±0.7 months. There were 27 (55.1%) functional responders (3 no definable 6-month response; 2 missed assessments and 1 long-term lead displacement). No marker predicted response, however, CTx and LBBB trended most towards predicting functional response.

Conclusion: No specific biomarkers reached significance for predicting functional response to CRT. CTx showed a trend towards predicting response and warrants further study.

Trial registration number: NCT02541773.

Keywords: cardiac resynchronization therapy; heart failure; micro-RNAs; non-response; vascular biomarkers.

Conflict of interest statement

Competing interests: None declared.

Figures

Figure 1

Patient recruitment, flow and outcomes. CRT, cardiac resynchronisation therapy; HF, heart failure.

Figure 2

Trends in functional variables, LV geometry and biomarker expression following CRT implantation in responders and non-responders. Trends represent the mean value of responders or non-responders. Differences over time and between response status were tested. 6MWT and LVEF are presented as mean (95% CI). PINP (ug/L) and MMP-2 (ug/L) are presented as median (CI 95%). CRT, cardiac resynchronisation therapy; LV, left ventricle; MMP-2, matrixmetalloproteinase-2; 6MWT, 6 min walk test; PINP, N-terminal pro-peptides of collagen I.

Figure 3

Bivariate correlation analysis of short-term and long-term changes following CRT between biomarkers versus functional and echocardiographic variables. Relative change applied to short-term (6 weeks) and long-term (6 months) reviews compared with the baseline assessments. Relative change was calculated by follow-up-Baseline/Baseline. Parametric or non-parametric bivariate correlation analysis performed dependent of continuous data distribution. All prespecified biomarkers compared with were 6MWT, QoL score, NT-pro-BNP, LVESV and LVEF. CRT, cardiac resynchronisation therapy; LV, left ventricle; MMP-2, matrixmetalloproteinase-2; 6MWT, 6 min walk test; NT-pro-BNP, N-terminal pro-brain natriuretic peptide; PINP, N-terminal pro-peptides of collagen I; PIIINP, N-terminal pro-peptides of collagen III.

Figure 4

Univariate and multivariate regression model of pre-CRT implant variables for prediction of functional response at 6 months. Forrest plot demonstrated the OR and 95% CI for parameters in univariate analysis. The table demonstrated the final step in the multivariate analysis. ECM, GDF-15 and NT-pro-BNP were logarithmically transformed for the prediction model. ECM, extracellular matrix; GDF-15, Growth Differentiation Factor-15; NT-pro-BNP, N-terminal pro-brain natriuretic peptide.

Figure 5

Variation between biomarker expression in peripheral and coronary sinus blood. PIIINP is expressed as mean±SD and underwent parametric comparison. PINP, MMP-2, hs-TnT, miR-30d, miR-133a and miR-486 were reported as median (range) and underwent non-parametric comparison. Comparisons were undertaken on the number of paired datasets available (n given for each comparison). hs-TnT, high-sensitivity Troponin-T; MMP-2, matrixmetalloproteinase-2; PINP, N-terminal pro-peptides of collagen I; PIIINP, N-terminal pro-peptides of collagen III.