The degree to which myocardial infarct site and size determine the electrocardiographic axis. Analysis of correlative...

The spatial electrical QRS axis was determined for 428 twelve-lead electrocardiograms from patients subsequently shown by postmortem dissection to have ventricular myocardial fibrosis or necrosis. Four 16-segment ventricular models of the heart were used to "predict" the spatial electrical QRS axis from known ventricular mass and deficit. The raw model I in ideal anatomic position and vectors perpendicularly outward for each muscle segment showed a mean correlation value of -0.494; model II was rotated in the chest to produce best fit with a correlation of 0.638, but the anatomic orientation was not reasonable; model III maintained original position and orientation but the vectors were scaled (correlation 0.780); and model IV, with nonperpendicular vectors, yielded a mean correlation of 0.793. The exceptions to good predictability formed a distinct subset largely composed of electrocardiograms with some form of variant intraventricular conduction (slight QRS widening, fascicular block patterns and "indeterminate frontal-plane axis").