Comparison of techniques for the measurement of left ventricular function following cardiac transplantation

Background: Assessment of graft function after cardiac transplantation is essential for patient management and clinical research. Previous studies have found that the left ventricular (LV) ejection fraction (EF) by echocardiography (echo), radionuclide ventriculography (RNV), and cardiovascular magnetic resonance (CMR) is discrepant in patients with heart failure. Method: Twelve patients underwent LV EF assessment by echo, angiography (angio), RNV and CMR one year following heart transplantation. The scans were analylzed independently in blinded fashion. Results: The mean EF was 63 +/- 6% by RNV 66 +/- 6% by CMR, 70 +/- 12% by angio, and 74 +/- 4% by echo. Significant differences were found between CMR and echo (p < 0.001), RNV and echo (p < 0.001), and RNV and angio (p < 0.05). The correlation between the techniques was poor (r = 0.3-0.6), and the scatter plots also suggested considerable variations between techniques. This was confirmed by the wide Bland-Altman limits of agreement (ranging from 22 to 45%). These were particularly wide for comparisons with angiography (43-45%). Conclusion: The EF measurement by echo, angio, RNV, and CMR are not interchangeable in patients following heart transplantation. The CMR and RNV provided the best agreement in EF and appear preferable for research studies. Echocardiography systematically overestimated LV EF and showed poor agreement with other techniques. Angiography overestimated LV function, and its routine use did not add to information gained from noninvasive studies.