The effect of ventricular pacing on measurements of left ventricular function: a comparison between echocardiographic methods and with radionuclide ventriculography.

AIMS Different methods exist for measuring left ventricular function echocardiographically; each may be error prone due to the abnormal pattern of ventricular activation during pacing. METHODS AND RESULTS Echocardiography was undertaken on 307 patients with permanent pacemakers; a subset of 57 underwent radionuclide ventriculography. Intrinsic and paced beats were analysed for left ventricular function by: Simpson's bi-plane, Teicholz M-mode, wall-motion scoring and 'eyeball' assessment. Agreement between techniques and with radionuclide ventriculography were compared according to intrinsic or paced beats. Echocardiographic measures of ejection fraction give mean values 5% higher than radionuclide ventriculography (Simpson's 30+/-9%, vs. Teicholz 30+/-13% vs. radionuclide ventriculography 25+/-9%, p=0.03). Agreement between Simpson's, Teicholz and radionuclide ventriculography by Bland-Altman analysis showed poor agreement (Simpson's vs. Teicholz range (4xSD)=57%, Simpson's vs. radionuclide ventriculography=36%, Teicholz vs. radionuclide ventriculography=46%, p=0.02), the level of agreement deteriorates with ventricular pacing (Simpson's vs. Teicholz range=61%, Simpson's vs. radionuclide ventriculography=34%, Teicholz vs. radionuclide ventriculography=47%, p=0.02). The correlation between wall motion analysis and radionuclide ventriculography is moderately poor (all subjects r=0.58, ventricular pacing r=0.52, not pacing r=0.66). CONCLUSION Echocardiography and radionuclide ventriculography are the only non-invasive techniques to assess left ventricular function in the paced population. Results are poorly interchangeable and the accuracy of any comparison dependent on the underlying rhythm.