Change in ejection fraction and long-term mortality in adults referred for echocardiography

Aims
We investigated long-term mortality associated with changes in left ventricular ejection fraction (LVEF) in a large, real-world patient cohort.

Methods and results
A total of 117 275 adults (63 ± 16 years, 46% women) had LVEF quantified by the same method ≥6 months apart. This included 17 343 cases (66 ± 15 years, 48% women) being initially investigated for heart failure (HF). During 3.3 [interquartile range (IQR) 1.7–6.0] years from first to last echocardiogram, median change in LVEF was −1 (IQR −8 to +5) units from a baseline of 62% (IQR 54–69%). During subsequent 7.6 (IQR 4.3–10.1) years of follow-up, 11 397 (9.7%) and 34 101 (29.1%) cases died from cardiovascular disease and all causes, respectively. Actual 5-year, all-cause mortality increased from 12% to 29% among those with the smallest to the largest decrease in LVEF (from <5 units to >30 units); the adjusted risk of cardiovascular-related mortality increased two- to eightfold beyond a >10-unit decline in LVEF (vs. minimal change; P < 0.001 for all comparisons). Among those initially investigated for HF (32% with initial LVEF <50%), the adjusted hazard ratio for cardiovascular-related mortality ranged from 0.35 [95% confidence interval (CI) 0.28–0.49] to 4.21 (95% CI 3.30–5.22) for a >30-unit increase to >30-unit decline in LVEF (vs. minimal change; P < 0.001 for both comparisons). A distinctive, bi-directional plateau of improved vs. worsening mortality was evident around a final LVEF of 50% to 55%.

Conclusions
These data, derived from a large, heterogeneous cohort of adults being followed up with echocardiography, suggest that modest LVEF changes (particularly around an LVEF of 50–55%) may be of clinical significance.