ICM-2006-00940-Revised1-ESM 1 1
Contribution of arterial stiffness and stroke volume to peripheral pulse pressure in ICU patients: an arterial tonometry study.
Bouchra Lamia, MD(1), Jean-Louis Teboul, MD(1), Xavier Monnet, MD(1),
David Osman, MD(1), Julien Maizel, MD(1) , Christian Richard, MD(1), Denis Chemla, MD(2)
Electronic Supplementary Material (ESM) 1: The augmentation index.
From the foot of the systolic pressure wave to the peak systolic pressure wave, two components can be distinguished: the forward pressure wave and the reflected pressure wave. An inflection point allows separating these two components and the time to the peak/shoulder of the first pressure wave component (T1) during systole was automatically identified. The pressure at the inflection point (P1) indicates the beginning upstroke of the reflected wave. The augmentation (Ax) was defined as the difference between the systolic pressure and P1. The Ax was thus the amplitude of the reflected pressure wave. The augmentation index (AI) was automatically calculated as Ax / aortic pulse pressure. In 62 patients (93%), the systolic inflection point was clearly defined, thus allowing AI calculation. In the remaining five subjects, the inflection point could not be discerned.
Electronic Supplementary Material (ESM) 2: Total arterial compliance and stiffness
Total arterial compliance was estimated using the area method previously described by Liu et al. (9):
total arterial compliance = stroke volume index / [K (Pes – DAP)]
where Pes is end-systolic aortic pressure, DAP is diastolic aortic pressure and K is an area index obtained as follows: K = (As + Ad)/Ad
where As is the systolic area under the aortic pressure curve and Ad is the diastolic area under the aortic pressure curve (9) (Figure 2). Total arterial stiffness was calculated as 1/compliance.
The area method is a simple and accurate method that has gained wide clinical acceptance to estimate total arterial compliance in the time domain (2, 9, 10). It is admitted that the windkessel model essentially applies during the diastolic period whatever the model chosen for the overall systemic circulation i.e., windkessel, distributed linear or non-linear models (3, 9). Because the windkessel model has physiologically interpretable parameters and a limited number of unknown parameters, and fits the data exactly at low frequencies, it has been suggested that the estimation of total arterial compliance C can play a pivotal role in solving some aspects of hemodynamic problems. In its simplest two-element form (RC), the windkessel model allows a reasonably accurate prediction of aortic pulse pressure in experimental settings where only mean arterial pressure and aortic flow are given. The limitations of the model have been discussed, including the over-neglected influence of zero-flow pressure on compliance estimates (2, 9, 10).
Although compliance is widely distributed throughout the vascular tree, it is admitted that the elastic properties of the proximal aorta account for a substantial portion of compliance (3, 11). The compliance of the proximal aorta and large arteries mainly depends on the relative contribution of elastin and collagen. Wall stress is normally supported by the compliant elastin fibers within the media in the proximal aorta. At high pressure, wall stress is supported by the much stiffer collagen fibers within the aortic media. Thus, the pressure-volume relationship is expected to be non-linear: first, at low pressure, small pulse pressure corresponds to high volume increases (elastin recruitment); and second, at higher pressure, high pulse pressure correspond to small volume increases (collagen recruitment) (3, 11). However, linear pressure-volume relations are easier to use and rely on assumptions that appear reasonable when linear and non linear estimates of compliance are compared at pressures near mean aortic pressure (3, 9). Within this constraint, comparisons among patients may be performed to provide a first approximation of compliance (9).