BACKGROUND: In order to tailor therapy in heart failure, a solution might be to develop sensitive and reliable markers that can predict response in individual patients or monitor effectiveness of therapy. AIMS: To evaluate neurohumoral factors as markers for left-ventricular (LV) antiremodelling from metoprolol treatment in patients with chronic LV systolic heart failure. METHODS: Forty-one subjects randomised to placebo or metoprolol were studied with magnetic resonance imaging and blood samples to measure LV dimensions and ejection fraction, epinephrine, norepinephrine, plasma renin activity, aldosterone, atrial (ANP) and brain natriuretic peptides, arginine-vasopressin and endothelin-1 at baseline, 5 weeks and 6 months after randomisation. RESULTS: Baseline ANP was identified as sole independent marker for changes in LV end-diastolic (deltaLVEDVI: r=-0.70, P=0.002), and end-systolic (deltaLVESVI: r=-0.53, P=0.03) volumes during metoprolol treatment. Change in ANP during the study was an independent marker for deltaLVEDVI: r=0.66, P=0.004, and deltaLVESVI: r=0.69, P=0.002 in the entire metoprolol group, but at the individual patient level, results were less clear. CONCLUSION: The pre-treatment plasma level of ANP may be a predictor of LV antiremodelling from treatment with metoprolol in patients with chronic heart failure. However, the potential for individual neurohumoral monitoring of the effects on LV dimensions during beta-blockade appears limited.