| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
LETTER TO THE EDITOR
First, our long time interval analysis technique for monitoring average CO over 
6-min (or shorter) intervals should be adequate to resolve many important circulatory pathologies (sepsis, moderate hemorrhage, cardiac tamponade, catecholamine tachyphalaxis, etc.). For catastrophic circulatory events that develop within a few minutes (arrhythmias, massive pulmonary embolism, etc.), standard monitoring modalities such as cardiac rhythm, heart rate, and blood pressures will reveal the event, and it is speculative that beat-to-beat CO monitoring would be of additional value.
Second, the results of our technique in animals, healthy subjects, and critically ill patients indicate that its underlying assumption of a constant arterial compliance (within a subject in the acute setting) does not preclude reliable measurement of relative changes in CO (1, 2). Nonetheless, future work is justified in implementing a nonlinear arterial compliance relationship such as that used in the study of Wesseling et al. (5) to investigate whether it further improves the accuracy of our technique.
Third, a concern about intrabeat analysis techniques for monitoring CO is that they must directly confront extremely complex wave reflections, which dominate over the short time intervals of analysis. Lumped parameter models, in particular, completely ignore the significant wave phenomena that occur within a beat (e.g., retrograde components) and therefore run the risk of introducing serious CO estimation errors. Nevertheless, it may be possible that a given intrabeat analysis such as the Modelflow method (5) may ultimately prove to be sufficiently accurate for clinical use. However, a long time interval analysis should be, at least in principle, more accurate in terms of estimating changes in average CO, because it 1) extracts all possible CO information from the ABP waveform and 2) circumvents the reflected waves by analyzing interbeat ABP variations in which the confounding effects of wave reflections cease to be a major factor (3).
Ultimately, we are in agreement with Drs. van Lieshout and Jansen (4): an analysis of various pulse contour analysis techniques is speculative until direct comparisons are performed. It is preferable to have ideal testing data with gold standard measurements and a panoply of patient characteristics, physiological states, and arterial measurement locations. However, we perceive value in direct comparison using other clinical data provided that 1) less restrictive testing ultimately leads to a substantially larger volume of comparison data and 2) authors and readers bear in mind the limitations of any given testing data set. Pulse contour analysis techniques will of course appear less accurate when there is error in the reference standard, but their comparative accuracies should be less affected by noise in the reference method. Overall, it may be more informative, and certainly more practical, to compare pulse contour analysis techniques through several different datasets and hundreds of subjects rather than idealized testing in a smaller number of subjects.
FOOTNOTES
Address for reprint requests and other correspondence: R. Mukkamala, Dept. of Electrical and Computer Engineering, Michigan State Univ., 2120 Engineering Bldg. East Lansing, MI 48824 (e-mail: rama{at}egr.msu.edu)
REFERENCES
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
