Individualized estimation of the central aortic blood pressure waveform: a comparative study.

This paper presents a comparative study on the relative performance of alternative strategies for estimating an individualized central aortic blood pressure (BP) waveform. Based on the transmission line representation of the central aortic-radial arterial line, a fully individualized (ITF 1), two partially individualized (ITF 2 and ITF 3 ), and a fully nonindividualized (NITF) transfer functions (i.e., frequency-dependent central aortic-radial BP relationships) were constructed using experimental data collected from nine swine subjects. The central aortic BP waveforms estimated by these transfer functions were compared against their measured gold standards, with the root-mean-squared waveform error and the absolute errors associated with systolic and pulse pressures as performance measures. Overall, the advantage of the individualized over the nonindividualized approach was modest but significant. The superiority of the individualized approach to its nonindividualized counterpart was increasingly pronounced under nonnominal or extreme physiologic conditions, as the subject's pulse transit time deviated from the averaged nominal value. The results suggest that the use of a fully individualized transfer function (ITF 1) is strongly recommended for nonnominal physiologic conditions, whereas a partially (ITF 2 and ITF 3) or even fully nonindividualized transfer function (NITF) may also yield acceptable performance under nominal physiologic conditions.