Beat to beat and observer variation of the electrocardiogram

• Published in: The American heart journal Vol. 75; no. 4; pp. 465 - 473
• Main Authors: Fischmann, Eugene, Cosma, John, Pipberger, Hubert V.
• Format: Journal Article
• Language: English
• Published: United States Mosby, Inc 01.04.1968
• Subjects: Computers; Electrocardiography; Heart Diseases - diagnosis; Humans; Methods
• ISSN: 0002-8703; 1097-6744
• DOI: 10.1016/0002-8703(68)90005-7

The BBV of selected ECG measurements was studied in an all male series consisting of 29 subjects with and 29 without heart disease. Nine consecutive beats of each of the 58 3-lead orthogonal records, selected to show minimal base-line disturbance, were converted by computer into enlarged digital plots. A single reader manually marked P and QRS onsets and P, QRS, and T endings in the plots, and the timing of these 5 points, punched on cards, was fed to a digital computer to obtain the measurements required for the BBV comparison. Two beats in each record, independently marked on two consecutive occasions by two physician and two technician readers, were similarly processed to supply information concerning beat-to-beat and repeat observer variation. BBV of instantaneous vectors increased with the spatial velocity of QRS and was therefore greatest in the QRS midportion. It varied also with the dependence of a measurement on QRS end point recognition, so that spatial maximum vectors showed the least, instantaneous vectors the greatest, and scalar lead amplitudes intermediate variation. Using the peak of R X for time reference instead of the less well defined QRS end points reduced BBV. The four observer, beat-to-beat variation approximates the 9 beat single observer BBV. Inconsistent wave recognition is an important cause of BBV. Other factors causing BBV are respiration, aberrant conduction, other so far unelucidated physiologic factors, base-line shifts, and rapid transients within ECG complexes. Occasionally, in loops with, two or more widely divergent near-maximal vectors, extreme spurious BBV appears when the maximum keeps changing from one of these vectors to another. In the presence of spatial vectors, which are near perpendicular to one of the three planes, small BBV of the spatial vector may result in extreme BBV of its planar projection. Spatial vectors are, however, not affected by this type of error. Two technicians performed wave end point timing more consistently than two physician readers.