Comparison and co-relation of invasive and noninvasive methods of ejection fraction measurement

J Natl Med Assoc. 2007 Nov;99(11):1227-8, 1231-4.


Background: Accurate estimation of left ventricular ejection fraction (LVEF) has assumed great significance in the era of automatic implantable cardioverter defibrillators (AICDs), and a low EF may be one of the sole deciding factor in determining AICD implantation in certain patient populations.

Aim: There are various methods, invasive and noninvasive, which can help calculate EF. We sought to conduct a retrospective study comparing EF estimation by invasive (angiography) and noninvasive methods [MUGA (multiple-gated acquisition), echocardiography (echo), single-photon emission computed tomography (SPECT)] in 5,558 patients in our hospital from 1995-2004.

Methods and results: EF was estimated by > or = 1 method (angiography, MUGA, echo, SPECT) within a one-month period. Values for the four tests in 5,558 patients were as follows: angiography mean 46.2, range 20-75, standard deviation (SD) 13.1; MUGA mean 45.7, range 20-70, SD 11.6; echo mean 45.7, range 22-70, SD 11.2; and SPECT mean 54.4, range 30-75, SD 11.9. Excellent positive correlations were found among all four tests as follows: angiography and MUGA, correlation coefficient (r) = 0.97, angiography and echo r = 0.96, angiography and SPECT r = 0.94, MUGA and echo r = 0.97, MUGA and SPECT r = 0.94, and echo and SPECT r = 0.94. Values for SPECT were significantly higher than for angiography, echo and MUGA (p < 0.001). The arithmetic difference between angiography and MUGA (mean 0.50, range -5.0-5.0) and the arithmetic difference between angiography and echo (mean 0.52, range -5.0-15.0) were similar (p = 0.59). The arithmetic difference between SPECT and angiography (mean 8.2, range -15.0-20.0) was significantly larger than the arithmetic difference between angiography and echo (p < 0.001).

Conclusions: All the four methods used to estimate EF corelate well with each other. However, values estimated during stress testing by SPECT overestimate EF and are significantly higher as compared to MUGA, echo and angiography. Estimation of EF by MUGA, echo or angiography should be preferred over SPECT, especially when that patient warrants intervention. We conclude that the overestimation of EF by SPECT may deprive some deserving patients of the survival benefit afforded by ICD.

Publication types

  • Comparative Study

MeSH terms

  • Coronary Angiography
  • Databases, Factual
  • Defibrillators, Implantable*
  • Gated Blood-Pool Imaging*
  • Heart Ventricles / diagnostic imaging
  • Humans
  • Retrospective Studies
  • Stroke Volume*
  • Tomography, Emission-Computed, Single-Photon*
  • Ultrasonography
  • Ventricular Function*