Arterial waveform morphomics during hemorrhagic shock

• Published in: European journal of trauma and emergency surgery (Munich : 2007) 2021-04, Vol.47 (2), p.325-332
• Main Authors: Wasicek, Philip J., Teeter, William A., Yang, Shiming, Hu, Peter, Gamble, William B., Galvagno, Samuel M., Hoehn, Melanie R., Scalea, Thomas M., Morrison, Jonathan J.
• Format: Article
• Language: English
• Subjects: Animals; Blood Pressure; Critical Care Medicine; Emergency medical care; Emergency Medicine; Hemodynamics; Hemorrhage; Hemorrhagic shock; Hypotension; Intensive; Medicine; Medicine & Public Health; Original Article; Shock, Hemorrhagic; Sports Medicine; Surgery; Surgical Orthopedics; Swine; Traumatic Surgery
• ISSN: 1863-9933; 1863-9941
• DOI: 10.1007/s00068-019-01140-2

Purpose The arterial pressure waveform is a composite of multiple interactions, and there may be more sensitive and specific features associated with hemorrhagic shock and intravascular volume depletion than systolic and/or diastolic blood pressure (BP) alone. The aim of this study was to characterize the arterial pressure waveform in differing grades of hemorrhage. Methods Ten anesthetized swine (70–90 kg) underwent a 40% controlled exponential hemorrhage. High-fidelity arterial waveform data were collected (500 Hz) and signal-processing techniques were used to extract key features. Regression modeling was used to assess the trend over time. Short-time Fourier transform (STFT) was utilized to assess waveform frequency and power spectrum density variance. Results All animals tolerated instrumentation and hemorrhage. The primary antegrade wave (P1) was relatively preserved while the renal (P2) and iliac (P3) reflection waves became noticeably attenuated during progressive hemorrhage. Several features mirrored changes in systolic and diastolic BP and plateaued at approximately 20% hemorrhage, and were best fit with non-linear sigmoidal regression modeling. The P1:P3 ratio continued to change during progressive hemorrhage ( R 2  = 0.51). Analysis of the first three harmonics during progressive hemorrhage via STFT demonstrated increasing variance with high coefficients of determination using linear regression in frequency ( R 2  = 0.70, 0.93, and 0.76, respectively) and power spectrum density ( R 2  = 0.90, 0.90, and 0.59, respectively). Conclusions In this swine model of volume-controlled hemorrhage, hypotension was a predominating early feature. While most waveform features mirrored those of BP, specific features such as the variance may be able to distinguish differing magnitudes of hemorrhage despite little change in conventional measures.