Biomarker‐based assessment of collagen cross‐linking identifies patients at risk of heart failure more likely to benefit from spironolactone effects on left atrial remodelling. Insights from the <scp>HOMAGE</scp> clinical trial
Susana Ravassa, Begoña López, João Pedro Ferreira, Nicolas Girerd, Erwan Bozec, Pierpaolo Pellicori, Beatrice Mariottoni, Franco Cosmi, Mark R. Hazebroek, Job A.J. Verdonschot, Joe Cuthbert, Johannes Petutschnigg, María U. Moreno, Stéphane Heymans, Jan A. Staessen, Burkert Pieske, Frank Edelmann, Andrew L. Clark, John G.F. Cleland, Faı̈ez Zannad, Javier Dı́ez, Arantxa González
Abstract
Aims The HOMAGE randomized trial found that spironolactone reduced left atrial volume index (LAVI), E:A ratio, and a marker of collagen type I synthesis (procollagen type I C‐terminal propeptide) in patients at risk of heart failure (HF). Previous trials showed that patients with HF, preserved ejection fraction and low serum collagen type I C‐terminal telopeptide to matrix metalloproteinase‐1 ratio (CITP:MMP‐1), associated with high collagen cross‐linking, had less improvement in diastolic function with spironolactone. We evaluated the interaction between serum CITP:MMP‐1 and spironolactone on cardiac function in the HOMAGE trial. Methods and results Patients at risk of HF were randomized to spironolactone ( n = 260) or not ( n = 255). Blood sampling and echocardiography were done at baseline, one and nine months. CITP:MMP‐1 was used as an indirect measure of collagen cross‐linking. Higher baseline CITP:MMP‐1 (i.e. lower collagen cross‐linking) was associated with greater reductions in LAVI with spironolactone at both one ( p = 0.003) and nine ( p = 0.01) months, but no interaction was observed for E:A ratio. Spironolactone reduced LAVI after one and nine months only for those patients in the third tertile of CITP:MMP‐1 (estimated lowest collagen cross‐linking) [mean differences spiro/control : −1.77 (95% confidence interval, CI −2.94 to −0.59) and −2.52 (95% CI −4.46 to −0.58) mL/m 2 ; interaction p across‐tertiles = 0.005; interaction p third tertile = 0.008] with a similar trend for N‐terminal pro‐B‐type natriuretic peptide which was consistently reduced by spironolactone only in the lowest collagen cross‐linking tertile [mean differences spiro/control : −0.47 (95% CI −0.66 to −0.28) and −0.31 (95% CI −0.59 to −0.04) ng/L; interaction p across‐tertiles = 0.09; interaction p third tertile < 0.001]. Conclusions These findings suggest that, for patients at risk of HF, the effects of spironolactone on left atrial remodelling may be more prominent in patients with less collagen cross‐linking (indirectly assessed by serum CITP:MMP‐1).