Association of outcome with left ventricular volumes and ejection fraction measured with two- and three-dimensional echocardiography in patients referred for routine, clinically indicated studies

Objectives: We sought to analyze if left ventricular (LV) volumes and ejection fraction (EF) measured by three-dimensional echocardiography (3DE) have incremental prognostic value over measurements obtained from two-dimensional echocardiography (2DE) in patients referred to a high-volume echocardiography laboratory for routine, clinically-indicated studies. Methods: We measured LV volumes and EF using both 2DE and 3DE in 725 consecutive patients (67% men; 59 ± 18 years) with various clinical indications referred for a routine clinical study. Results: LV volumes were significantly larger, and EF was lower when measured by 3DE than 2DE. During follow-up (3.6 ± 1.2 years), 111 (15.3%) all-cause deaths and 248 (34.2%) cardiac hospitalizations occurred. Larger LV volumes and lower EF were associated with worse outcome independent of age, creatinine, hemoglobin, atrial fibrillation, and ischemic heart diseases). In stepwise Cox regression analyses, the associations of both death and cardiac hospitalization with clinical data (CD: age, creatinine, hemoglobin, atrial fibrillation, and ischemic heart disease) whose Harrel's C-index (HC) was 0.775, were augmented more by the LV volumes and EF obtained by 3DE than by 2DE parameters. The association of CD with death was not affected by LV end-diastolic volume (EDV) either measured by 2DE or 3DE. Conversely, it was incremented by 3DE LVEF (HC = 0.84, p < 0.001) more than 2DE LVEF (HC = 0.814, p < 0.001). The association of CD with the composite endpoint (HC = 0.64, p = 0.002) was augmented more by 3DE LV EDV (HC = 0.786, p < 0.001), end-systolic volume (HC = 0.801, p < 0.001), and EF (HC = 0.84, p < 0.001) than by the correspondent 2DE parameters (HC = 0.786, HC = 0.796, and 0.84, all p < 0.001) In addition, partition values for mild, moderate and severe reduction of the LVEF measured by 3DE showed a higher discriminative power than those measured by 2DE for cardiac death (Log-Rank: χ2 = 98.3 vs. χ2 = 77.1; p < 0.001). Finally, LV dilation defined according to the 3DE threshold values showed higher discriminatory power and prognostic value for death than when using 2DE reference values (3DE LVEDV: χ2 = 15.9, p < 0.001 vs. χ2 = 10.8, p = 0.001; 3DE LVESV: χ2 = 24.4, p < 0.001 vs. χ2 = 17.4, p = 0.001). Conclusion: In patients who underwent routine, clinically-indicated echocardiography, 3DE LVEF and ESV showed stronger association with outcome than the corresponding 2DE parameters.