A103-07 Early Study on the Spectrum of Fluid Responsiveness Using Physiologic Waveforms

Abstract Rationale Intravenous fluid bolus therapy is ubiquitous in inpatient clinical settings for the management of various physiological states. Fluid responsiveness has been commonly studied as a binary event using clinical metrics such as cardiac output, stroke volume variation, and pulse pressure variation. Thresholds have been developed to categorize patients as responders vs non-responders. However, there are few studies investigating the continuous nature of fluid responsiveness and its dynamic physiologic development. We evaluated the hemodynamic effects of intravenous fluid bolus therapy utilizing continuous waveform data from ECG leads and photoplethysmography. We analyzed hemodynamic changes before, during, and after fluid bolus administration. Methods We utilized a large database from University of Pittsburgh containing clinical and waveform data from over 22,000 hospitalized patients from 2018-2023. Using a random subset of approximately 500 patients, we isolated isotonic crystalloid fluid bolus therapy (1000 mL) while avoiding concurrent administration of pharmacologic therapies that may impact hemodynamics, such as vasopressors. We then extracted and analyzed waveform (256 Hz) and numeric(1 Hz) physiologic data surrounding bolus times. We performed simple feature extraction to obtain hemodynamic parameters such as heart rate, R-R interval, blood pressure. We visualized the change in these parameters and tested the statistical significance of the differences the non-parametric Wilcoxon signed-rank test. Results The change in hemodynamic parameters were calculated before, during, and after bolus administration. The median change in heart rate before, during, and after bolus administration were 0.15, -2.22, and -2.72 bpm, respectively. The p-values for the Wilcoxon signed-rank test for before vs during, during vs after, and before vs after were 0.001, 0.95, and 0.001, respectively. The median change in mean non-invasive blood pressure during and after bolus administration compared to before (i.e. baseline) were 3.6 and 2.4 mmHg, respectively. The p-values for the Wilcoxon signed-rank test for before vs during, during vs after, and before vs after were 0.02, 0.49, and 0.20, respectively. Conclusions In patient’s receiving 1000 mL of isotonic crystalloid, there was a statistically significant change in hemodynamic parameters such as heart rate and blood pressure during bolus administration, providing evidence that fluid responsiveness is a dynamic process that can be actively measured during fluid resuscitation. Future efforts will seek to develop an integratedmultidimensional index of fluid responsiveness. This abstract is funded by: National Institutes of Health