Studies Made by Simulating Systole at Necropsy. VI. Estimation of Cardiac Stroke Volume From the Ballistocardiogram

Abstract

Ballistocardiograms, and estimations of central and peripheral blood pressure were secured while systole was simulated in cadavers, after first raising arterial pressure to normal or experimental levels by femoral perfusion. In 5 cadavers blood was used as perfusion fluid in 51 simulated systoles; in 14 cadavers water was used in 137 simulated systoles, conducted at both normal and abnormal levels of blood pressure. Stroke volumes estimated from ballistocardiograms were compared with known values measured in the injection syringe. The original formulae provide a reasonably good estimate of stroke volume. Tanner''s formula has a smaller standard deviation about the regression, but the absolute values estimated are too high. Improvement in estimations was sought by simple and multiple regression equations. The inclusion of a term for body surface area as a measure of size of the subject improves the estimate, the standard deviation in experiments in which blood was used falling to 8 cc. The addition of a term for age causes only slight further improvement. It is concluded that the ballistocardiogram provides a satisfactory but rather rough method of estimating stroke volume as long as the record is normal in form. It is at its best at detecting changes of stroke volume in individuals, and so should be useful in the testing of drugs and other therapeutic agents. When the ballistocardiogram is abnormal in form, estimates of stroke volume cannot be relied upon.