Catalog of Regulatory Science Tools to Help Assess New Medical Devices
This regulatory science tool presents datasets for characterizing three dynamic attributes of pressure-based cardiac output monitoring systems that apply an algorithm to intra-radial arterial blood pressure waveforms for cardiac output and stroke volume variation measurements.
Technical Description
This RST is a database tool consisting of nine mock circulation loop (MCL)-generated datasets for characterizing three dynamic attributes of pressure-based cardiac output monitoring systems. These attributes include:
i) cardiac output response time,
ii) cardiac output resolution, and
iii) stroke volume variation resolution.
Each dataset of this RST includes non-clinically generated central flow (representative of aortic flow), central pressure (representative of aortic pressure) and peripheral pressure (representative of radial pressure) waveforms collected during conditions representative of three hemodynamic states (cardiogenic, normovolemic, and hyperdynamic shock). Each hemodynamic state was simulated with the MCL, based on the clinical data reported for the relevant population. Overall, nine datasets were generated (3 hemodynamic states each including 3 types of tests) that contain approximately 7.5 hours of synchronized peripheral pressure, central flow, and central pressure waveforms.
The whole database can be used either by applying it directly to an isolated algorithm (software approach) or the entire system (including sensors), and can be implemented by replicating the MCL-generated pressure waveforms using a pressure pulse generator connected to the system’s pressure transducer (Farahmand et al., 2023).
Intended Purpose
The purpose of this RST, a non-clinical open-access MCL-generated database tool, is to help characterize three dynamic attributes of hemodynamic monitoring devices that apply an algorithm to an intra-radial arterial pressure waveform to monitor cardiac output. The RST includes scenarios to characterize the following attributes: cardiac output response time, cardiac output resolution, and stroke volume variation resolution. This RST is intended to be used during device development of intra-radial arterial pressure-based cardiac output monitoring algorithms as a non-clinical method to characterize these attributes. The tool may also assist during device development of applications that may use intra-radial arterial pressure-derived cardiac output measurements from these types of devices.
Testing
The MCL-generated relationships between central pressure and peripheral pressure were verified against clinical data (Farahmand et al., 2023). Central flow waveforms were generated with a semi-triangular morphology and were linearly scaled for simulating these hemodynamic states. For each state, stepwise changes in the MCL flow were simulated and datasets were collected for characterizing pressure-based cardiac output systems.
The database was tested with successful execution to characterize the dynamic attributes of a benchtop pulse contour monitoring system using a physical pressure pulse generator.
Limitations
While this RST facilitates testing of certain characteristics that are challenging to address in clinical studies, such as the response to a small change in cardiac output, it is not a replacement for accuracy testing in clinically relevant patient populations with appropriate clinical reference methods.
The MCL from which the RST was generated is a bench-top model of cardiovascular components and therefore may not represent all physiological relationships. A known limitation of the current RST is that the MCL does not simulate the nonlinear relationship between arterial compliances and pressure, which can affect the testing of certain algorithms. The MCL also significantly attenuated spectral components above 4.5 Hz. For many applications the lost pressure contents in this spectral range may have negligible impact on the cardiac output measurement performance, but this limitation should be taken into consideration depending on the design of the algorithm the database is intended to be applied to and the potential effects of higher frequency components in the arterial pressure waveform. The database may not, for example, be representative of algorithms that use the systolic upslope as this parameter and related parameters may be affected by the loss of spectral pressure content above 4.5 Hz. Stroke volume variation was simulated with fixed simulated breathing rates that may not be representative of expected variability during spontaneous breathing.
Supporting Documentation
- Farahmand, M., Bodwell, E., D'Souza, G. A., Herbertson, L. H., & Scully, C. G. (2023). Mock circulatory loop generated database for dynamic characterization of pressure-based cardiac output monitoring systems. Computers in biology and medicine, 160, 106979. https://doi.org/10.1016/j.compbiomed.2023.106979
A detailed description of the database can be found on GitHub
The database package is also accessible on GitHub.
The following publications provide additional background on the approach to generate arterial pressure waveforms with a mock circulatory loop and application of arterial pressure waveforms to testing with a physical pressure pulse generator.
- Packy, A., D'Souza, G. A., Farahmand, M., Herbertson, L., & Scully, C. G. (2022). Simulating Radial Pressure Waveforms with a Mock Circulatory Flow Loop to Characterize Hemodynamic Monitoring Systems. Cardiovascular engineering and technology, 13(2), 279–290. https://doi.org/10.1007/s13239-021-00575-2
- Farahmand, M., Mirinejad, H., & Scully, C. G. (2021). Model-based approach to investigate equipment-induced error in pressure-waveform derived hemodynamic measurements. Physiological measurement, 42(11), 10.1088/1361-6579/ac38be. https://doi.org/10.1088/1361-6579/ac38be