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 differencesspiro/control: −1.77 (95% confidence interval, CI −2.94 to −0.59) and −2.52 (95% CI −4.46 to −0.58) mL/m2; interaction pacross-tertiles = 0.005; interaction pthird 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 differencesspiro/control: −0.47 (95% CI −0.66 to −0.28) and −0.31 (95% CI −0.59 to −0.04) ng/L; interaction pacross-tertiles = 0.09; interaction pthird 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). Abstract Patients at risk of heart failure from the HOMAGE clinical trial were classified according to the baseline degree of myocardial collagen cross-linking, non-invasively assessed by the serum collagen type I C-terminal telopeptide (CITP) to matrix metalloproteinase-1 (MMP-1) ratio (CITP:MMP-1). As highly cross-linked collagen fibres are more resistant to degradation and CITP is a cross-linked peptide, for a given MMP-1 quantity less CITP will be released and, subsequently, serum CITP:MMP-1 will be lower. Whereas patients with low collagen cross-linking (high CITP:MMP-1) benefit from the cardioprotective effects of treatment with spironolactone on left atrial remodelling [i.e. a decrease in left atrial volume index (LAVI)] and on N-terminal pro-B-type natriuretic peptide (NT-proBNP) levels, these beneficial effects are not found in patients with higher collagen cross-linking (low CITP:MMP-1).