Performance of a machine-learning algorithm to predict hypotension in mechanically ventilated patients with COVID-19 admitted to the intensive care unit: a cohort study

The Hypotension Prediction Index (HPI) is a commercially available machine-learning algorithm that provides warnings for impending hypotension, based on real-time arterial waveform analysis. The HPI was developed with arterial waveform data of surgical and intensive care unit (ICU) patients, but has never been externally validated in the latter group. In this study, we evaluated diagnostic ability of the HPI with invasively collected arterial blood pressure data in 41 patients with COVID-19 admitted to the ICU for mechanical ventilation. Predictive ability was evaluated at HPI thresholds from 0 to 100, at incremental intervals of 5. After exceeding the studied threshold, the next 20 min were screened for positive (mean arterial pressure (MAP) < 65 mmHg for at least 1 min) or negative (absence of MAP < 65 mmHg for at least 1 min) events. Subsequently, sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and time to event were determined for every threshold. Almost all patients (93%) experienced at least one hypotensive event. Median number of events was 21 [7–54] and time spent in hypotension was 114 min [20–303]. The optimal threshold was 90, with a sensitivity of 0.91 (95% confidence interval 0.81–0.98), specificity of 0.87 (0.81–0.92), PPV of 0.69 (0.61–0.77), NPV of 0.99 (0.97–1.00), and median time to event of 3.93 min (3.72–4.15). Discrimination ability of the HPI was excellent, with an area under the curve of 0.95 (0.93–0.97). This validation study shows that the HPI correctly predicts hypotension in mechanically ventilated COVID-19 patients in the ICU, and provides a basis for future studies to assess whether hypotension can be reduced in ICU patients using this algorithm.

SEMI INVASIVE CARDIAC OUTPUT MONITORING IN A CARDIAC PATIENT UNDERGOING TOTAL THYROIDECTOMY USING SMARTPHONE APP CAPSTESIA.

Introduction: Advanced hemodynamic monitoring is need of today especially in patients with limited cardiac reserve. With the advent of smartphones & specially designed applications, hemodynamic monitoring becomes quite easy. Materials & Methods: Patient was pre – medicated with Inj. Fentanyl & inj. Glycopyrrolate, induced with Inj. Etomidate & Inj. Vecuronium and maintained with mixture ofIsourane, Nitrous Oxide & Oxygen. An arterial line was secured in Left Radial Artery. We used the CAPSTESIA app to take picture of the arterial waveform using a smartphone. Demographic data of the patient was fed in the app. App used it's pre- fed algorithm to give the real time Cardiac Output, Pulse Pressure variations, Cardiac Index based upon the arterial waveform. Results: Using the application we were able to monitor the cardiac output of the patient in real time using semi- invasive means. It enabled us to regulate the uid management of the patient and avoid any adverse cardiac events (hypotension). With Pulse Pressure variation also available in real time, we were able to restrict use of vasopressors since the Left Ventricle Ejection Fraction of the patient was 35 % on ECHO. Surgery was conducted without any untoward event. Patient was successfully extubated and sent to PACU. Conclusions:Advanced hemodynamic monitoring is time consuming using manual methods. We found the smartphone app CAPSTESIA pretty useful for semi-invasive hemodynamic monitoring of the Cardiac Output, Pulse Pressure variation, Cardiac Index,etc in real time.

Cardiac Output Evaluation on Septic Shock Patients: Comparison between Calibrated and Uncalibrated Devices during Vasopressor Therapy

There are no reliable, non-invasive methods to accurately measure cardiac output (CO) in septic patients. MostCare (Vytech Health™, Vygon, Padova, Italy), is a beat-to-beat, self calibrated method for CO measurement based on continuous analysis of reflected arterial pressure waveforms. We enrolled 40 patients that were suffering from septic shock and requiring norepinephrine infusion to target blood pressure in order to to evaluate the level of agreement between a calibrated transpulmonary thermodilution device (PiCCO System, Pulsion Medical Systems, Feldkirchen, Germany) and the MostCare system in detecting and tracking changes in CO measurements related to norepinephrine reduction in septic shock patients,. PiCCO was connected to a 5 Fr femoral artery catheter and to a central venous catheter. System calibration was performed with 15 mL of cold saline injection over about 3 s. The MostCare device was connected to the artery catheter to analyze the arterial waveform. Before reducing norepinephrine infusion, the PiCCO system was calibrated, the MostCare waveform was optimized, and the values of the complete hemodynamic profile were recorded (T1). Norepinephrine infusion was then reduced by 0.03 mcg/Kg/min. After 30 min, a new calibration of PiCCO system and a new record on both monitors were performed (T2). Static measurements agreements were assessed using the Bland-Altman test, while trending ability was investigated using polar plot analysis. If volume expansion occurred, then related data were separately analyzed. At T1 mean the CO was 5.38 (SD 0.60) L/min, the mean difference was 0.176 L/min, the limits of agreement (LoA) was +1.39 and −1.04 L/min, and the percentage error (PE) was 22.6%; at T2 the mean CO was 5.44 (SD 0.73) L/min, the mean difference was 0.053 L/min, the LoA was +1.51 and −1.40, and the PE was 27%. After considering the volume expansion between T1 and T2, the mean CO at T1 was 5.39 L/min (SD 0.47), the LoA was +1.09 and −0.78 L/min, and the percentage error (PE) was 17%; at T2 the mean CO was 5.35 L/min (SD 0.81), the LoA was +1.73 and −1.52 L/min, and the PE was 30%. The polar plot diagram seems to confirm the trending ability of MostCare system versus the reference method. In septic patients, when the arterial waveform is accurate, MostCare and PiCCO transpulmonary thermodilution exhibit good agreement even after the reduction of norepinephrine and changes in vascular tone or volume expansion. MostCare could be a rapid to set, reliable, and useful tool to monitor hemodynamic variations in septic patients in emergency contexts where thermodilution methods or other advanced systems are not easily available.

Abstract 16334: An Artificial Intelligence Derived Method for Instantaneous Detection of Elevated Left Ventricular End Diastolic Pressure

Introduction: Instantaneous, non-invasive detection of an elevated left ventricular end-diastolic pressure (LVEDP) offers a significant benefit in diagnosis and treatment of heart failure. We recently proposed a systems approach, called cardiac triangle mapping (CTM), that uses intrinsic frequencies (IFs) of the arterial waveform and pre-ejection period (PEP) to map the global ventricular function (Pahlevan et al. Fluids 4.1 (2019): 16). Here, we tested the hypothesis that an elevated LVEDP can be detected using ECG and arterial pressure waveform by applying an artificial neural network (ANN) combined with CTM approach. Methods: This study included 46 patients (12 females, age 39-90 (66.4±9.9), BMI 20.2-36.8 (27.6±4.1)) who were scheduled for a clinical left heart catheterization or coronary angiogram at the Keck Medical Center of USC. Exclusion criteria were valvular heart disease, atrial fibrillation, or left bundle branch block. Invasive LVEDP and aortic pressure waveforms were measured using a 3F Millar transducer tipped catheter with simultaneous 3 channel ECG. The IFs were computed from pressure waveforms. PEPs were calculated as the time difference between the beginning of QRS and the uprising of the pressure waveform. A 3-layer network consisted of 6 input, 6 hidden and one output nodes was developed. LVEDP=18 mmHg was used as the cut-off for a binary outcome. Data from 34 patients were used to design the ANN (27 for training, 7 for validation). The model was tested on 12 additional patients. Results: Our results showed a specificity of 87% and a sensitivity of 96% in detecting an elevated LVEDP (Fig.1). Conclusions: Here, we demonstrated the proof-of-concept that an AI model based on reduced-order parameters (extracted from arterial waveform and ECG) can instantaneously detect an elevated LVEDP. Although our hemodynamic measurements were done invasively, all variables that are required for this AI-LVEDP calculation can be collected noninvasively.

Wearable Sensors Incorporating Compensatory Reserve Measurement for Advancing Physiological Monitoring in Critically Injured Trauma Patients

Vital signs historically served as the primary method to triage patients and resources for trauma and emergency care, but have failed to provide clinically-meaningful predictive information about patient clinical status. In this review, a framework is presented that focuses on potential wearable sensor technologies that can harness necessary electronic physiological signal integration with a current state-of-the-art predictive machine-learning algorithm that provides early clinical assessment of hypovolemia status to impact patient outcome. The ability to study the physiology of hemorrhage using a human model of progressive central hypovolemia led to the development of a novel machine-learning algorithm known as the compensatory reserve measurement (CRM). Greater sensitivity, specificity, and diagnostic accuracy to detect hemorrhage and onset of decompensated shock has been demonstrated by the CRM when compared to all standard vital signs and hemodynamic variables. The development of CRM revealed that continuous measurements of changes in arterial waveform features represented the most integrated signal of physiological compensation for conditions of reduced systemic oxygen delivery. In this review, detailed analysis of sensor technologies that include photoplethysmography, tonometry, ultrasound-based blood pressure, and cardiogenic vibration are identified as potential candidates for harnessing arterial waveform analog features required for real-time calculation of CRM. The integration of wearable sensors with the CRM algorithm provides a potentially powerful medical monitoring advancement to save civilian and military lives in emergency medical settings.