Aims: The prognostic values of left ventricular ejection fraction (LVEF) during heart failure (HF) with acute decompensation or after optimal treatment have not been extensively studied. We hypothesized that posttreatment LVEF has superior predictive value for long-term prognosis than LVEF at admission does.
Methods and results: In Protocol 1, 428 acute decompensated HF (ADHF) patients with LVEF ≤35% in a tertiary medical center were enrolled and followed for a mean period of 34.7 ± 10.8 months. The primary and secondary end points were all-cause mortality and HF readmission, respectively. In total, 86 deaths and 240 HF readmissions were recorded. The predictive values of baseline LVEF at admission and LVEF 6 months posttreatment were analyzed and compared. The posttreatment LVEFs were predictive for future events (P = 0.01 for all-cause mortality, P < 0.001 for HF readmission), but the baseline LVEFs were not. In Protocol 2, the outcomes of patients with improved LVEF (change of LVEF: ≥+10%), unchanged LVEF (change of LVEF: -10% to +10%), and reduced LVEF (change of LVEF: ≤-10%) were analyzed and compared. Improved LVEF occurred in 171 patients and was associated with a superior long-term prognosis among all groups (P = 0.02 for all-cause mortality, P < 0.001 for HF readmission). In Protocol 3, independent predictors of improved LVEF were analyzed, and baseline LV end-diastolic dimension (LVEDD) was identified as a powerful predictor in ADHF patients (P < 0.001).
Conclusions: In patients with ADHF, posttreatment LVEF but not baseline LVEF had prognostic power. Improved LVEF was associated with superior long-term prognosis, and baseline LVEDD identified patients who were more likely to have improved LVEF. Therefore, baseline LVEF should not be considered a relevant prognosis factor in clinical practice for patients with ADHF.