Central Venous Pressure Waveform Analysis Predicts Perfusion in a Porcine Model of Endotoxemic Hypotension and Resuscitation

Abstract

Introduction: Cardiac output (CO), a proxy for perfusion, is a critical measure during sepsis and resuscitation. The understudied venous system has nuanced physiology that confers valuable information. In this investigation, deconvolution of the central venous pressure (CVP) waveform by the fast Fourier transformation (FFT) algorithm is performed to derive a CO surrogate in a “sepsis-like” porcine model of endotoxemic hypotension and resuscitation. Methods: Nine pigs were anesthetized, catheterized and intubated. A lipopolysaccharides infusion protocol was employed to precipitate low systemic vascular resistance (SVR) hypotension. Four subsequent crystalloid boluses (10 cc/kg) were given, after which thermodilution-CO, CVP and the CVP waveform were collected, the latter undergoing FFT analysis. The amplitude of the fundamental frequency of CVP waveform’s cardiac wave (“f0-CVP”) was obtained. Mean CVP, f0-CVP and CO were plotted over the course of the boluses to determine whether f0-CVP tracked with CO better than CVP itself. Results: Distributive hypotension to a 25% mean arterial pressure decrement was achieved, with decreased SVR (918±227 [standard deviation] dynes/seconds/cm-5 vs 685±180 dynes/seconds/cm-5; P = 0.038). Over four boluses, slopes of linear regression lines of CVP (r2=0.27; P < 0.001), f0-CVP (r2=0.15; P = 0.01) and CO (r2=0.08; P = 0.06) were 0.447, 0.148 and 0.178, respectively, demonstrating that f0-CVP constitutes a more accurate CO surrogate than CVP. Conclusion: FFT analysis of CVP waveforms offers a novel paradigm to interpret readily obtainable hemodynamic information. It offers potential for real-time utility in assessing volume responsiveness in sepsis and resuscitation, while further offering potential translation to the less-invasive peripheral venous pressure waveform.