Objectives: Arterial wave reflection has traditionally been quantified from pressure and flow measurements using wave separation and wave intensity (WI) analysis. In the recently proposed reservoir-wave paradigm, these analyses are performed after dividing pressure into 'reservoir' and 'excess' components, yielding a modified wave intensity (WI(RW)). This new approach has led to controversial conclusions about the nature and significance of arterial wave reflection. Our aim was to assess whether WI or WI(RW) more accurately represent wave phenomena.
Methods: We studied two computer models (a simple network and a full model of the systemic arterial tree) in which all systolic forward waves and reflection properties were known a priori. Results of these models were compared with haemodynamic measurements in the ascending aorta of five adult sheep at baseline and after incremental arterial constriction.
Results: The key findings of model studies were that the reservoir-wave approach markedly underestimated or eliminated reflected compression waves, overestimated or artefactually introduced forward and backward expansion waves, and displayed nonphysical interactions between distal reflection sites and early systolic waves. These errors arose because, contrary to a key assumption of the reservoir-wave approach, reservoir pressure was not spatially uniform during systole. In-vivo results were qualitatively similar to model results, with baseline WI and WI(RW) suggesting that the arterial network was dominated by positive and negative wave reflection, respectively, while under all conditions, reflected WI(RW) compression waves were substantially smaller than corresponding WI waves.
Conclusion: We conclude that the reservoir-wave paradigm introduces error into arterial wave analyses.