The accuracy of inferring left ventricular volume from dimension depends on the frequency of information needed to answer a given question.

Left ventricular volume is an important variable in cardiovascular physiology. Because volume cannot be transduced directly, by necessity people measure one or more cardiac dimensions, then implicitly or explicitly estimate volume. Despite its common use, this practice has never been adequately justified. Previous validations of the use of left ventricular dimensions to infer volume have ignored the complexity of the relation between dimensions and volume, particularly over a broad range of hemodynamic states, and have not considered the frequency content of the volume signal needed to answer a given physiological question. We show that, strictly speaking, the assumptions underlying this practice are false. The resulting errors are of little practical consequence under many circumstances, whereas they can be quite important under other circumstances. We consider the frequency content of the volume signal to organize our examples and findings. In this general framework, we find that, for most assessments of volume change from dimension change at or below a frequency of about the heart rate (e.g., beat-to-beat end-diastolic volume), there is no limitation, provided all relevant dimensions are measured. However, for information at frequencies above the heart rate (i.e., rapid changes in volume or dimension), it is probably not possible to draw accurate conclusions about left ventricular volume from dimension data.