TRED-HF Recovery Relapse Prediction Model

TRED-HF Recovery / Relapse Prediction — explanation and clinical context The TRED-HF randomized pilot trial examined phased withdrawal of all heart failure pharmacotherapy in carefully selected patients with previously dilated cardiomyopathy who had asymptomatic recovery of left ventricular function (LVEF ≥50%), normalised indexed LV volumes, and low NT-proBNP at baseline; approximately 40–44% of participants relapsed within months after withdrawal, showing that recovery is often remission rather than permanent cure. Analyses derived from the TRED-HF dataset have repeatedly highlighted baseline NT-proBNP and rises in heart rate after withdrawal as important harbingers of relapse; higher baseline NT-proBNP and early increases in heart rate during medication reduction were associated with greater relapse risk. Other features suggested as informative include residual LV dilation (indexed LVEDV), and the extent/intensity of background disease-modifying therapy—likely reflecting underlying disease severity.

Because the original TRED-HF publication did not provide an externally validated point-by-point risk equation for bedside use, this calculator implements an explicit, transparent, approximate logistic estimator based on those reported predictors for educational demonstration only. Clinical interpretation summary: A higher estimated probability in this model reflects a higher likelihood that withdrawal of therapy would be followed by objective relapse (reduction in LVEF, increase in LV volume, rise in NT-proBNP or clinical HF). Given the limited sample size of TRED-HF, the heterogeneity of recovered DCM, and changes in contemporary therapy (ARNi, SGLT2i widely used after TRED-HF), any numeric estimate should be interpreted cautiously—this tool is not a substitute for individualized clinical judgement, imaging, and biomarkers in shared decision-making.