Abstract
Introduction
Cardiac output (CO), the volume of blood ejected by the heart per minute, is of critical importance in delivering oxygen and nutrients to vital organs [1]. A fall in CO is one of the main reasons of death in intensive care units and for complications during/after major surgery. For this reason, it is routinely monitored to guide management, but accurate measurements of CO require specialist equipment and invasive access. Non-invasive alternatives that could transform critical care outside of the intensive care unit, if they were sufficiently accurate, perform poorly when tracking within-patient changes in CO that are used to inform patient management. We present an algorithm based on physical principles that show great promise in overcoming these limitations.
Methods
The algorithm first extracts the key pressures of a measure pressure waveform. A corresponding blood flow waveform is generated through an iterative algorithm to minimise errors between the measured pressure features and those generated from a validated simplified model of the arterial tree linking pressure to flow [3]. We tested the algorithm with invasive data obtained in 29 patients (19 men, age 45±15 years old, BP 117±18/63±12 mmHg, mean±SD) during general anaesthesia. We also compared the performance of the algorithm against 11 other algorithms previously
published – some used by existing commercial devices - to 1) estimate single CO measurements (n=106) and 2) track within-subject variations in CO (n=29).
Results & Discussion
The estimates yielded by the algorithm were highly correlated with reference measures (Pearson coefficient, r=0.93) with a small deviation after calibration (SD=0.42 L/min). Comparatively, the second-best algorithm had a Pearson coefficient r=0.77 and a standard deviation twice as high (SD=0.83 L/min). Importantly, our algorithm was the only one to accurately track within-subject variations in CO (Pearson coefficient, r=0.85), despite variations over a relatively small range. By contrast, many existing methods showed no or little correlation with the “gold-standard” measure which is in accordance with the findings of other studies [3].
Conclusion
Our algorithm provided the most accurate estimates and trends of all the algorithms tested. We now need to test it prospectively on wider ranges of variation.
References
1. Vlachopoulos C, O’Rourke M, Nichols WW. McDonald’s Blood Flow in Arteries:
Theoretical, Experimental and Clinical Principles. CRC Press; 2011.
2. Vennin et al. Identifying Hemodynamic Determinants of Pulse Pressure. Hypertension, 70(6), 1176-1182; 2017
3. Caillard et al. Comparison of cardiac output measured by oesophageal Doppler
ultrasonography or pulse pressure contour wave analysis. British Journal of Anaesthesia, 114(6), 893-900; 2015.
Cardiac output (CO), the volume of blood ejected by the heart per minute, is of critical importance in delivering oxygen and nutrients to vital organs [1]. A fall in CO is one of the main reasons of death in intensive care units and for complications during/after major surgery. For this reason, it is routinely monitored to guide management, but accurate measurements of CO require specialist equipment and invasive access. Non-invasive alternatives that could transform critical care outside of the intensive care unit, if they were sufficiently accurate, perform poorly when tracking within-patient changes in CO that are used to inform patient management. We present an algorithm based on physical principles that show great promise in overcoming these limitations.
Methods
The algorithm first extracts the key pressures of a measure pressure waveform. A corresponding blood flow waveform is generated through an iterative algorithm to minimise errors between the measured pressure features and those generated from a validated simplified model of the arterial tree linking pressure to flow [3]. We tested the algorithm with invasive data obtained in 29 patients (19 men, age 45±15 years old, BP 117±18/63±12 mmHg, mean±SD) during general anaesthesia. We also compared the performance of the algorithm against 11 other algorithms previously
published – some used by existing commercial devices - to 1) estimate single CO measurements (n=106) and 2) track within-subject variations in CO (n=29).
Results & Discussion
The estimates yielded by the algorithm were highly correlated with reference measures (Pearson coefficient, r=0.93) with a small deviation after calibration (SD=0.42 L/min). Comparatively, the second-best algorithm had a Pearson coefficient r=0.77 and a standard deviation twice as high (SD=0.83 L/min). Importantly, our algorithm was the only one to accurately track within-subject variations in CO (Pearson coefficient, r=0.85), despite variations over a relatively small range. By contrast, many existing methods showed no or little correlation with the “gold-standard” measure which is in accordance with the findings of other studies [3].
Conclusion
Our algorithm provided the most accurate estimates and trends of all the algorithms tested. We now need to test it prospectively on wider ranges of variation.
References
1. Vlachopoulos C, O’Rourke M, Nichols WW. McDonald’s Blood Flow in Arteries:
Theoretical, Experimental and Clinical Principles. CRC Press; 2011.
2. Vennin et al. Identifying Hemodynamic Determinants of Pulse Pressure. Hypertension, 70(6), 1176-1182; 2017
3. Caillard et al. Comparison of cardiac output measured by oesophageal Doppler
ultrasonography or pulse pressure contour wave analysis. British Journal of Anaesthesia, 114(6), 893-900; 2015.
| Original language | English |
|---|---|
| Title of host publication | BioMedEng18 Conference Proceedings |
| Publisher | Imperial College London (University of London) |
| Pages | 328 |
| Number of pages | 1 |
| Publication status | Published - 2018 |
| Event | BioMedEng18 - Imperial College London, London, United Kingdom Duration: 6 Sept 2018 → 7 Sept 2018 |
Conference
| Conference | BioMedEng18 |
|---|---|
| Country/Territory | United Kingdom |
| City | London |
| Period | 6/09/2018 → 7/09/2018 |