Use of Advanced Machine-Learning Techniques for Noninvasive Monitoring of Hemorrhage

Prediction of Occult Hemorrhage in the Lower Body Negative Pressure Model: Initial Validation of Machine Learning Approaches

INTRODUCTION

Detection of occult hemorrhage (OH) before progression to clinically apparent changes in vital signs remains an important clinical problem in managing trauma patients. The resource-intensiveness associated with continuous clinical patient monitoring and rescue from frank shock makes accurate early detection and prediction with noninvasive measurement technology a desirable innovation. Despite significant efforts directed toward the development of innovative noninvasive diagnostics, the implementation and performance of the newest bedside technologies remain inadequate. This poor performance may reflect the limitations of univariate systems based on one sensor in one anatomic location. It is possible that when signals are measured with multiple modalities in multiple locations, the resulting multivariate anatomic and temporal patterns of measured signals may provide additional discriminative power over single technology univariate measurements. We evaluated the potential superiority of multivariate methods over univariate methods. Additionally, we utilized machine learning-based models to compare the performance of noninvasive-only to noninvasive-plus-invasive measurements in predicting the onset of OH.

MATERIALS AND METHODS

We applied machine learning methods to preexisting datasets derived using the lower body negative pressure human model of simulated hemorrhage. Employing multivariate measured physiological signals, we investigated the extent to which machine learning methods can effectively predict the onset of OH. In particular, we applied 2 ensemble learning methods, namely, random forest and gradient boosting.

RESULTS

Analysis of precision, recall, and area under the receiver operating characteristic curve showed a superior performance of multivariate approach to that of the univariate ones. In addition, when using both invasive and noninvasive features, random forest classifier had a recall 95% confidence interval (CI) of 0.81 to 0.86 with a precision 95% CI of 0.65 to 0.72. Interestingly, when only noninvasive features were employed, the results worsened only slightly to a recall 95% CI of 0.80 to 0.85 and a precision 95% CI of 0.61 to 0.73.

CONCLUSIONS

Multivariate ensemble machine learning-based approaches for the prediction of hemodynamic instability appear to hold promise for the development of effective solutions. In the lower body negative pressure multivariate hemorrhage model, predictions based only on noninvasive measurements performed comparably to those using both invasive and noninvasive measurements.

Non-invasive biomarkers for detecting progression toward hypovolemic cardiovascular instability in a lower body negative pressure model

Occult hemorrhages after trauma can be present insidiously, and if not detected early enough can result in patient death. This study evaluated a hemorrhage model on 18 human subjects, comparing the performance of traditional vital signs to multiple off-the-shelf non-invasive biomarkers. A validated lower body negative pressure (LBNP) model was used to induce progression towards hypovolemic cardiovascular instability. Traditional vital signs included mean arterial pressure (MAP), electrocardiography (ECG), plethysmography (Pleth), and the test systems utilized electrical impedance via commercial electrical impedance tomography (EIT) and multifrequency electrical impedance spectroscopy (EIS) devices. Absolute and relative metrics were used to evaluate the performance in addition to machine learning-based modeling. Relative EIT-based metrics measured on the thorax outperformed vital sign metrics (MAP, ECG, and Pleth) achieving an area-under-the-curve (AUC) of 0.99 (CI 0.95-1.00, 100% sensitivity, 87.5% specificity) at the smallest LBNP change (0-15 mmHg). The best vital sign metric (MAP) at this LBNP change yielded an AUC of 0.6 (CI 0.38-0.79, 100% sensitivity, 25% specificity). Out-of-sample predictive performance from machine learning models were strong, especially when combining signals from multiple technologies simultaneously. EIT, alone or in machine learning-based combination, appears promising as a technology for early detection of progression toward hemodynamic instability.

Digitalization in orthopaedics: a narrative review

Advances in technology and digital tools like the Internet of Things (IoT), artificial intelligence (AI), and sensors are shaping the field of orthopaedic surgery on all levels, from patient care to research and facilitation of logistic processes. Especially the COVID-19 pandemic, with the associated contact restrictions was an accelerator for the development and introduction of telemedical applications and digital alternatives to classical in-person patient care. Digital applications already used in orthopaedic surgery include telemedical support, online video consultations, monitoring of patients using wearables, smart devices, surgical navigation, robotic-assisted surgery, and applications of artificial intelligence in forms of medical image processing, three-dimensional (3D)-modelling, and simulations. In addition to that immersive technologies like virtual, augmented, and mixed reality are increasingly used in training but also rehabilitative and surgical settings. Digital advances can therefore increase the accessibility, efficiency and capabilities of orthopaedic services and facilitate more data-driven, personalized patient care, strengthening the self-responsibility of patients and supporting interdisciplinary healthcare providers to offer for the optimal care for their patients.

Features derived from blood pressure and intracranial pressure predict elevated intracranial pressure events in critically ill children

Clinicians frequently observe hemodynamic changes preceding elevated intracranial pressure events. We employed a machine learning approach to identify novel and differentially expressed features associated with elevated intracranial pressure events in children with severe brain injuries. Statistical features from physiologic data streams were derived from non-overlapping 30-min analysis windows prior to 21 elevated intracranial pressure events; 200 records without elevated intracranial pressure events were used as controls. Ten Monte Carlo simulations with training/testing splits provided performance benchmarks for 4 machine learning approaches. XGBoost yielded the best performing predictive models. Shapley Additive Explanations analyses demonstrated that a majority of the top 20 contributing features consistently derived from blood pressure data streams up to 240 min prior to elevated intracranial events. The best performing prediction model was using the 30-60 min analysis window; for this model, the area under the receiver operating characteristic window using XGBoost was 0.82 (95% CI 0.81-0.83); the area under the precision-recall curve was 0.24 (95% CI 0.23-0.25), above the expected baseline of 0.1. We conclude that physiomarkers discernable by machine learning are concentrated within blood pressure and intracranial pressure data up to 4 h prior to elevated intracranial pressure events.

Development and Validation of a Dynamic Prediction Model for Massive Hemorrhage in Trauma

Objectives

Early warning prediction of massive hemorrhages can greatly reduce mortality in trauma patients. This study aimed to develop and validate dynamic prediction models for massive hemorrhage in trauma patients.

Methods

Based on vital signs (e.g., heart rate, respiratory rate, pulse pressure, and peripheral oxygen saturation) time-series data and the gated recurrent unit algorithm, we characterized a group of models to flexibly and dynamically predict the occurrence of massive hemorrhages in the subsequent T hours (where T = 1, 2, and 3). Models were evaluated in terms of accuracy, sensitivity, specificity, positive predictive value, negative predictive value, F1 score, and the area under the curve (AUC).

Results

Results show that of the 2205 trauma patients selected for model development, a total of 265 (12.02%) had a massive hemorrhage. The AUCs of the model in the 1-h-group, 2-h-group, and 3-h-group were 0.763 (95% CI: 0.708-0.820), 0.775 (95% CI: 0.728-0.823), and 0.756 (95% CI: 0.715-0.797), respectively. Finally, the models were used in a web calculator and information system for the hospital emergency department.

Conclusions

This study developed and validated a group of dynamic prediction models based on vital sign time-series data and a deep-learning algorithm to assist medical staff in the early diagnosis and dynamic prediction of a future massive hemorrhage in trauma.

Detection of a Stroke Volume Decrease by Machine-Learning Algorithms Based on Thoracic Bioimpedance in Experimental Hypovolaemia

Compensated shock and hypovolaemia are frequent conditions that remain clinically undetected and can quickly cause deterioration of perioperative and critically ill patients. Automated, accurate and non-invasive detection methods are needed to avoid such critical situations. In this experimental study, we aimed to create a prediction model for stroke volume index (SVI) decrease based on electrical cardiometry (EC) measurements. Transthoracic echo served as reference for SVI assessment (SVI-TTE). In 30 healthy male volunteers, central hypovolaemia was simulated using a lower body negative pressure (LBNP) chamber. A machine-learning algorithm based on variables of EC was designed. During LBNP, SVI-TTE declined consecutively, whereas the vital signs (arterial pressures and heart rate) remained within normal ranges. Compared to heart rate (AUC: 0.83 (95% CI: 0.73–0.87)) and systolic arterial pressure (AUC: 0.82 (95% CI: 0.74–0.85)), a model integrating EC variables (AUC: 0.91 (0.83–0.94)) showed a superior ability to predict a decrease in SVI-TTE ≥ 20% (p = 0.013 compared to heart rate, and p = 0.002 compared to systolic blood pressure). Simulated central hypovolaemia was related to a substantial decline in SVI-TTE but only minor changes in vital signs. A model of EC variables based on machine-learning algorithms showed high predictive power to detect a relevant decrease in SVI and may provide an automated, non-invasive method to indicate hypovolaemia and compensated shock.

Compensatory reserve and pulse character: Enhanced potential to predict urgency for transfusion and other life-saving interventions after traumatic injury

BACKGROUND

Field triage of trauma patients requires timely assessment of physiologic status to determine resuscitative needs. Vital signs and rudimentary assessments such as pulse character (PC) are used by first responders to guide decision making. The compensatory reserve measurement (CRM) has demonstrated utility as an easily interpretable method for assessing patient status. We hypothesized that the ability to identify injured patients requiring transfusion and other life-saving interventions (LSI) using a measurement of pulse character could be enhanced by the addition of the CRM.

METHODS

We performed a prospective observational study on 300 trauma patients admitted to a level I trauma center. CRM was recorded continuously after device placement on arrival. Patient demographics, field and trauma resuscitation unit vital signs, therapeutic interventions, and outcomes were collected. A field SBP <100 mmHg was utilized as a surrogate for abnormal PC as previously validated. A patient with a CRM threshold value of <60% was considered clinically compromised with a risk of onset of decompensated shock. Data were analyzed to assess the capacity of CRM and pulse character separately or in combination to predict LSI defined as need for transfusion, intubation, tube thoracostomy, or operative/ angiographic hemorrhage control.

RESULTS

An improvement in the predictive capability for LSI, transfusion, or a composite outcome was demonstrated by the combination of CRM and PC compared to either measure alone.

CONCLUSIONS

Combining PC assessment with CRM has the potential to enhance the recognition of injured patients requiring life-saving intervention thus improving sensitivity of decision support for prehospital providers.

AI-Enabled Advanced Development for Assessing Low Circulating Blood Volume for Emergency Medical Care: Comparison of Compensatory Reserve Machine-Learning Algorithms

The application of artificial intelligence (AI) has provided new capabilities to develop advanced medical monitoring sensors for detection of clinical conditions of low circulating blood volume such as hemorrhage. The purpose of this study was to compare for the first time the discriminative ability of two machine learning (ML) algorithms based on real-time feature analysis of arterial waveforms obtained from a non-invasive continuous blood pressure system (Finometer^®) signal to predict the onset of decompensated shock: the compensatory reserve index (CRI) and the compensatory reserve metric (CRM). One hundred ninety-one healthy volunteers underwent progressive simulated hemorrhage using lower body negative pressure (LBNP). The least squares means and standard deviations for each measure were assessed by LBNP level and stratified by tolerance status (high vs. low tolerance to central hypovolemia). Generalized Linear Mixed Models were used to perform repeated measures logistic regression analysis by regressing the onset of decompensated shock on CRI and CRM. Sensitivity and specificity were assessed by calculation of receiver-operating characteristic (ROC) area under the curve (AUC) for CRI and CRM. Values for CRI and CRM were not distinguishable across levels of LBNP independent of LBNP tolerance classification, with CRM ROC AUC (0.9268) being statistically similar (p = 0.134) to CRI ROC AUC (0.9164). Both CRI and CRM ML algorithms displayed discriminative ability to predict decompensated shock to include individual subjects with varying levels of tolerance to central hypovolemia. Arterial waveform feature analysis provides a highly sensitive and specific monitoring approach for the detection of ongoing hemorrhage, particularly for those patients at greatest risk for early onset of decompensated shock and requirement for implementation of life-saving interventions.

Central Hypovolemia Detection During Environmental Stress-A Role for Artificial Intelligence?

The first step to exercise is preceded by the required assumption of the upright body position, which itself involves physical activity. The gravitational displacement of blood from the chest to the lower parts of the body elicits a fall in central blood volume (CBV), which corresponds to the fraction of thoracic blood volume directly available to the left ventricle. The reduction in CBV and stroke volume (SV) in response to postural stress, post-exercise, or to blood loss results in reduced left ventricular filling, which may manifest as orthostatic intolerance. When termination of exercise removes the leg muscle pump function, CBV is no longer maintained. The resulting imbalance between a reduced cardiac output (CO) and a still enhanced peripheral vascular conductance may provoke post-exercise hypotension (PEH). Instruments that quantify CBV are not readily available and to express which magnitude of the CBV in a healthy subject should remains difficult. In the physiological laboratory, the CBV can be modified by making use of postural stressors, such as lower body "negative" or sub-atmospheric pressure (LBNP) or passive head-up tilt (HUT), while quantifying relevant biomedical parameters of blood flow and oxygenation. Several approaches, such as wearable sensors and advanced machine-learning techniques, have been followed in an attempt to improve methodologies for better prediction of outcomes and to guide treatment in civil patients and on the battlefield. In the recent decade, efforts have been made to develop algorithms and apply artificial intelligence (AI) in the field of hemodynamic monitoring. Advances in quantifying and monitoring CBV during environmental stress from exercise to hemorrhage and understanding the analogy between postural stress and central hypovolemia during anesthesia offer great relevance for healthy subjects and clinical populations.

Predictive and diagnosis models of stroke from hemodynamic signal monitoring

This work presents a novel and promising approach to the clinical management of acute stroke. Using machine learning techniques, our research has succeeded in developing accurate diagnosis and prediction real-time models from hemodynamic data. These models are able to diagnose stroke subtype with 30 min of monitoring, to predict the exitus during the first 3 h of monitoring, and to predict the stroke recurrence in just 15 min of monitoring. Patients with difficult access to a CT scan and all patients that arrive at the stroke unit of a specialized hospital will benefit from these positive results. The results obtained from the real-time developed models are the following: stroke diagnosis around 98% precision (97.8% sensitivity, 99.5% specificity), exitus prediction with 99.8% precision (99.8% Sens., 99.9% Spec.), and 98% precision predicting stroke recurrence (98% Sens., 99% Spec.). Graphical abstract depicting the complete process since a patient is monitored until the data collected is used to generate models.

Early Detection of Hypotension Using a Multivariate Machine Learning Approach

INTRODUCTION

The ability to accurately detect hypotension in trauma patients at the earliest possible time is important in improving trauma outcomes. The earlier an accurate detection can be made, the more time is available to take corrective action. Currently, there is limited research on combining multiple physiological signals for an early detection of hemorrhagic shock. We studied the viability of early detection of hypotension based on multiple physiologic signals and machine learning methods. We explored proof of concept with a small (5 minutes) prediction window for application of machine learning tools and multiple physiologic signals to detecting hypotension.

MATERIALS AND METHODS

Multivariate physiological signals from a preexisting dataset generated by an experimental hemorrhage model were employed. These experiments were conducted previously by another research group and the data made available publicly through a web portal. This dataset is among the few publicly available which incorporate measurement of multiple physiological signals from large animals during experimental hemorrhage. The data included two hemorrhage studies involving eight sheep. Supervised machine learning experiments were conducted in order to develop deep learning (viz., long short-term memory or LSTM), ensemble learning (viz., random forest), and classical learning (viz., support vector machine or SVM) models for the identification of physiological signals that can detect whether or not overall blood loss exceeds a predefined threshold 5 minutes ahead of time. To evaluate the performance of the machine learning technologies, 3-fold cross-validation was conducted and precision (also called positive predictive value) and recall (also called sensitivity) values were compared. As a first step in this development process, 5 minutes prediction windows were utilized.

RESULTS

The results showed that SVM and random forest outperform LSTM neural networks, likely because LSTM tends to overfit the data on small sized datasets. Random forest has the highest recall (84%) with 56% precision while SVM has 62% recall with 82% precision. Upon analyzing the feature importance, it was observed that electrocardiogram has the highest significance while arterial blood pressure has the least importance among all other signals.

CONCLUSION

In this research, we explored the viability of early detection of hypotension based on multiple signals in a preexisting animal hemorrhage dataset. The results show that a multivariate approach might be more effective than univariate approaches for this detection task.

Physiology of Human Hemorrhage and Compensation

Hemorrhage is a leading cause of death following traumatic injuries in the United States. Much of the previous work in assessing the physiology and pathophysiology underlying blood loss has focused on descriptive measures of hemodynamic responses such as blood pressure, cardiac output, stroke volume, heart rate, and vascular resistance as indicators of changes in organ perfusion. More recent work has shifted the focus toward understanding mechanisms of compensation for reduced systemic delivery and cellular utilization of oxygen as a more comprehensive approach to understanding the complex physiologic changes that occur following and during blood loss. In this article, we begin with applying dimensional analysis for comparison of animal models, and progress to descriptions of various physiological consequences of hemorrhage. We then introduce the complementary side of compensation by detailing the complexity and integration of various compensatory mechanisms that are activated from the initiation of hemorrhage and serve to maintain adequate vital organ perfusion and hemodynamic stability in the scenario of reduced systemic delivery of oxygen until the onset of hemodynamic decompensation. New data are introduced that challenge legacy concepts related to mechanisms that underlie baroreflex functions and provide novel insights into the measurement of the integrated response of compensation to central hypovolemia known as the compensatory reserve. The impact of demographic and environmental factors on tolerance to hemorrhage is also reviewed. Finally, we describe how understanding the physiology of compensation can be translated to applications for early assessment of the clinical status and accurate triage of hypovolemic and hypotensive patients. © 2021 American Physiological Society. Compr Physiol 11:1531-1574, 2021.

Multimodal tensor-based method for integrative and continuous patient monitoring during postoperative cardiac care

Patients recovering from cardiovascular surgeries may develop life-threatening complications such as hemodynamic decompensation, making the monitoring of patients for such complications an essential component of postoperative care. However, this need has given rise to an inexorable increase in the number and modalities of data points collected, making it challenging to effectively analyze in real time. While many algorithms exist to assist in monitoring these patients, they often lack accuracy and specificity, leading to alarm fatigue among healthcare practitioners. In this study we propose a multimodal approach that incorporates salient physiological signals and EHR data to predict the onset of hemodynamic decompensation. A retrospective dataset of patients recovering from cardiac surgery was created and used to train predictive models. Advanced signal processing techniques were employed to extract complex features from physiological waveforms, while a novel tensor-based dimensionality reduction method was used to reduce the size of the feature space. These methods were evaluated for predicting the onset of decompensation at varying time intervals, ranging from a half-hour to 12 h prior to a decompensation event. The best performing models achieved AUCs of 0.87 and 0.80 for the half-hour and 12-h intervals respectively. These analyses evince that a multimodal approach can be used to develop clinical decision support systems that predict adverse events several hours in advance.

Hospital Admission Location Prediction via Deep Interpretable Networks for the Year-Round Improvement of Emergency Patient Care

OBJECTIVE

This paper presents a deep learning method of predicting where in a hospital emergency patients will be admitted after being triaged in the Emergency Department (ED). Such a prediction will allow for the preparation of bed space in the hospital for timely care and admission of the patient as well as allocation of resource to the relevant departments, including during periods of increased demand arising from seasonal peaks in infections.

METHODS

The problem is posed as a multi-class classification into seven separate ward types. A novel deep learning training strategy was created that combines learning via curriculum and a multi-armed bandit to exploit this curriculum post-initial training.

RESULTS

We successfully predict the initial hospital admission location with area-under-receiver-operating-curve (AUROC) ranging between 0.60 to 0.78 for the individual wards and an overall maximum accuracy of 52% where chance corresponds to 14% for this seven-class setting. Our proposed network was able to interpret which features drove the predictions using a 'network saliency' term added to the network loss function.

CONCLUSION

We have proven that prediction of location of admission in hospital for emergency patients is possible using information from triage in ED. We have also shown that there are certain tell-tale tests which indicate what space of the hospital a patient will use.

SIGNIFICANCE

It is hoped that this predictor will be of value to healthcare institutions by allowing for the planning of resource and bed space ahead of the need for it. This in turn should speed up the provision of care for the patient and allow flow of patients out of the ED thereby improving patient flow and the quality of care for the remaining patients within the ED.

Combat medic testing of a novel monitoring capability for early detection of hemorrhage

BACKGROUND

Current out-of-hospital protocols to determine hemorrhagic shock in civilian trauma systems rely on standard vital signs with military guidelines relying on heart rate and strength of the radial pulse on palpation, all of which have proven to provide little forewarning for the need to implement early intervention prior to decompensation. We tested the hypothesis that addition of a real-time decision-assist machine-learning algorithm, the compensatory reserve measurement (CRM), used by combat medics could shorten the time required to identify the need for intervention in an unstable patient during a hemorrhage profile as compared with vital signs alone.

METHODS

We randomized combat medics from the Army Medical Department Center and School Health Readiness Center of Excellence into three groups: group 1 viewed a display of no simulated hemorrhage and unchanging vital signs as a control (n = 24), group 2 viewed a display of simulated hemorrhage and changing vital signs alone (hemorrhage; n = 31), and group 3 viewed a display of changing vital signs with the addition of the CRM (hemorrhage + CRM; n = 22). Participants were asked to push a computer key when they believed the patient was becoming unstable and needed medical intervention.

RESULTS

The average time of 11.0 minutes (95% confidence interval, 8.7-13.3 minutes) required by the hemorrhage + CRM group to identify an unstable patient (i.e., stop the video sequence) was less by more than 40% (p < 0.01) compared with 18.9 minutes (95% confidence interval, 17.2-20.5 minutes) in the hemorrhage group.

CONCLUSION

The use of a machine-learning monitoring technology designed to measure the capacity to compensate for central blood volume loss resulted in reduced time required by combat medics to identify impending hemodynamic instability.

LEVEL OF EVIDENCE

Diagnostic, level IV.

A Pilot Study using the Compensatory Reserve Index to evaluate individuals with Postural Orthostatic Tachycardia syndrome

PURPOSE

The diagnosis of Postural Orthostatic Tachycardia syndrome traditionally involves orthostatic vitals evaluation. The Compensatory Reserve Index is a non-invasive, FDA-cleared algorithm that analyses photoplethysmogram waveforms in real time to trend subtle waveform features associated with varying degrees of central volume loss, from normovolemia to decompensation. We hypothesised that patients who met physiologic criteria for Postural Orthostatic Tachycardia syndrome would have greater changes in Compensatory Reserve Index with orthostatic vitals.

METHODS

Orthostatic vitals and Compensatory Reserve Index values were assessed in individuals previously diagnosed with Postural Orthostatic Tachycardia syndrome and healthy controls aged 12-21 years. Adolescents were grouped for comparison based on whether they met heart rate criteria for Postural Orthostatic Tachycardia syndrome (physiologic Postural Orthostatic Tachycardia syndrome).

RESULTS

Sixty-one patients were included. Eighteen percent of patients with an existing Postural Orthostatic Tachycardia syndrome diagnosis met heart rate criteria, and these patients had significantly greater supine to standing change in Compensatory Reserve Index (0.67 vs. 0.51; p<0.001). The optimal change in Compensatory Reserve Index for physiologic Postural Orthostatic Tachycardia syndrome was 0.60. Patients with physiologic Postural Orthostatic Tachycardia syndrome were more likely to report previous diagnoses of anxiety or depression (p = 0.054, 0.042).

CONCLUSION

An accurate diagnosis of Postural Orthostatic Tachycardia syndrome may be confounded by related comorbidities. Only 18% (8/44) of previously diagnosed Postural Orthostatic Tachycardia syndrome patients met heart rate criteria. Findings support the utility of objective physiologic measures, such as the Compensatory Reserve Index, to more accurately identify patients with true autonomic dysfunction.

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.

Resuscitation After Hemorrhagic Shock in the Microcirculation: Targeting Optimal Oxygen Delivery in the Design of Artificial Blood Substitutes

Microcirculatory preservation is essential for patient recovery from hemorrhagic shock. In hemorrhagic shock, microcirculatory flow and pressure are greatly reduced, creating an oxygen debt that may eventually become irreversible. During shock, tissues become hypoxic, cellular respiration turns to anaerobic metabolism, and the microcirculation rapidly begins to fail. This condition requires immediate fluid resuscitation to promote tissue reperfusion. The choice of fluid for resuscitation is whole blood; however, this may not be readily available and, on a larger scale, may be globally insufficient. Thus, extensive research on viable alternatives to blood has been undertaken in an effort to develop a clinically deployable blood substitute. This has not, as of yet, achieved fruition, in part due to an incomplete understanding of the complexities of the function of blood in the microcirculation. Hemodynamic resuscitation is acknowledged to be contingent on a number of factors other than volume expansion. The circulation of whole blood is carefully regulated to optimize oxygen delivery to the tissues via shear stress modulation through blood viscosity, inherent oxygen-carrying capacity, cell-free layer variation, and myogenic response, among other variables. Although plasma expanders can address a number of these issues, hemoglobin-based oxygen carriers (HBOCs) introduce a method of replenishing the intrinsic oxygen-carrying capacity of blood. There continue to be a number of issues related to HBOCs, but recent advances in the next-generation HBOCs show promise in the preservation of microcirculatory function and limiting toxicities. The development of HBOCs is now focused on viscosity and the degree of microvascular shear stress achieved in order to optimize vasoactive and oxygen delivery responses by leveraging the restoration and maintenance of physiological responses to blood flow in the microcirculation. Blood substitutes with higher viscous properties tend to improve oxygen delivery compared to those with lower viscosities. This review details current concepts in blood substitutes, particularly as they relate to trauma/hemorrhagic shock, with a specific focus on their complex interactions in the microcirculation.

Supervised Machine Learning Applied to Automate Flash and Prolonged Capillary Refill Detection by Pulse Oximetry

Objective

Develop an automated approach to detect flash (<1.0 s) or prolonged (>2.0 s) capillary refill time (CRT) that correlates with clinician judgment by applying several supervised machine learning (ML) techniques to pulse oximeter plethysmography data.

Materials and Methods

Data was collected in the Pediatric Intensive Care Unit (ICU), Cardiac ICU, Progressive Care Unit, and Operating Suites in a large academic children’s hospital. Ninety-nine children and 30 adults were enrolled in testing and validation cohorts, respectively. Patients had 5 paired CRT measurements by a modified pulse oximeter device and a clinician, generating 485 waveform pairs for model training. Supervised ML models using gradient boosting (XGBoost), logistic regression (LR), and support vector machines (SVMs) were developed to detect flash (<1 s) or prolonged CRT (≥2 s) using clinician CRT assessment as the reference standard. Models were compared using Area Under the Receiver Operating Curve (AUC) and precision-recall curve (positive predictive value vs. sensitivity) analysis. The best performing model was externally validated with 90 measurement pairs from adult patients. Feature importance analysis was performed to identify key waveform characteristics.

Results

For flash CRT, XGBoost had a greater mean AUC (0.79, 95% CI 0.75–0.83) than logistic regression (0.77, 0.71–0.82) and SVM (0.72, 0.67–0.76) models. For prolonged CRT, XGBoost had a greater mean AUC (0.77, 0.72–0.82) than logistic regression (0.73, 0.68–0.78) and SVM (0.75, 0.70–0.79) models. Pairwise testing showed statistically significant improved performance comparing XGBoost and SVM; all other pairwise model comparisons did not reach statistical significance. XGBoost showed good external validation with AUC of 0.88. Feature importance analysis of XGBoost identified distinct key waveform characteristics for flash and prolonged CRT, respectively.

Conclusion

Novel application of supervised ML to pulse oximeter waveforms yielded multiple effective models to identify flash and prolonged CRT, using clinician judgment as the reference standard.

Tweet

Supervised machine learning applied to pulse oximeter waveform features predicts flash or prolonged capillary refill.

Comparative Analysis on Machine Learning and Deep Learning to Predict Post-Induction Hypotension

Hypotensive events in the initial stage of anesthesia can cause serious complications in the patients after surgery, which could be fatal. In this study, we intended to predict hypotension after tracheal intubation using machine learning and deep learning techniques after intubation one minute in advance. Meta learning models, such as random forest, extreme gradient boosting (Xgboost), and deep learning models, especially the convolutional neural network (CNN) model and the deep neural network (DNN), were trained to predict hypotension occurring between tracheal intubation and incision, using data from four minutes to one minute before tracheal intubation. Vital records and electronic health records (EHR) for 282 of 319 patients who underwent laparoscopic cholecystectomy from October 2018 to July 2019 were collected. Among the 282 patients, 151 developed post-induction hypotension. Our experiments had two scenarios: using raw vital records and feature engineering on vital records. The experiments on raw data showed that CNN had the best accuracy of 72.63%, followed by random forest (70.32%) and Xgboost (64.6%). The experiments on feature engineering showed that random forest combined with feature selection had the best accuracy of 74.89%, while CNN had a lower accuracy of 68.95% than that of the experiment on raw data. Our study is an extension of previous studies to detect hypotension before intubation with a one-minute advance. To improve accuracy, we built a model using state-of-art algorithms. We found that CNN had a good performance, but that random forest had a better performance when combined with feature selection. In addition, we found that the examination period (data period) is also important.

Evidence for misleading decision support in characterizing differences in tolerance to reduced central blood volume using measurements of tissue oxygenation

BACKGROUND

The physiological response to hemorrhage includes vasoconstriction in an effort to shunt blood to the heart and brain. Hemorrhaging patients can be classified as "good" compensators who demonstrate high tolerance (HT) or "poor" compensators who manifest low tolerance (LT) to central hypovolemia. Compensatory vasoconstriction is manifested by lower tissue oxygen saturation (StO₂ ), which has propelled this measure as a possible early marker of shock. The compensatory reserve measurement (CRM) has also shown promise as an early indicator of decompensation.

METHODS

Fifty-one healthy volunteers (37% LT) were subjected to progressive lower body negative pressure (LBNP) as a model of controlled hemorrhage designed to induce an onset of decompensation. During LBNP, CRM was determined by arterial waveform feature analysis. StO₂ , muscle pH, and muscle H⁺ concentration were calculated from spectrum using near-infrared spectroscopy (NIRS) on the forearm.

RESULTS

These values were statistically indistinguishable between HT and LT participants at baseline (p ≥ 0.25). HT participants exhibited lower (p = 0.01) StO₂ at decompensation compared to LT participants.

CONCLUSIONS

Lower StO₂ measured in patients during low flow states associated with significant hemorrhage does not necessarily translate to a more compromised physiological state, but may reflect a greater resistance to the onset of shock. Only the CRM was able to distinguish between HT and LT participants early in the course of hemorrhage, supported by a significantly greater ROC AUC (0.90) compared with STO₂ (0.68). These results support the notion that measures of StO₂ could be misleading for triage and resuscitation decision support.

[Computer-assisted decision-making for trauma patients]

BACKGROUND

In the management of trauma patients in the resuscitation room many time-pressured and critical decisions must continuously be made in complex situations. Even experienced teams frequently make errors in this context. Computer-assisted decision-making systems can predict critical situations based on patient data continuously acquired online. Based on the calculated predictions these systems can suggest the next steps in managing the patient. This review summarizes the current literature on computer-assisted decision-making in the management of trauma patients.

OBJECTIVE

A literature review summarizing existing concepts and applications of computer-assisted decision-making support for the management of trauma patients.

METHODS

Narrative review article based on an analysis of literature in the German and English languages from the last 10 years.

RESULTS

There exist numerous computer-assisted decision-making systems in the field of trauma care. Several studies could show that computer-assisted decision-making can improve the outcome in the preclinical setting, the resuscitation room and in the intensive care unit. For further validation and implementation of these systems, information technological barriers have to be overcome, existing systems need to be adapted to current data protection regulations and large multicenter studies are necessary.

CONCLUSION

Computer-assisted decision-making can help to improve the management of trauma patients; however, before a ubiquitous implementation a number of technological and legislative barriers have to be overcome.

Bridging the gap between military prolonged field care monitoring and exploration spaceflight: the compensatory reserve

The concept of prolonged field care (PFC), or medical care applied beyond doctrinal planning timelines, is the top priority capability gap across the US Army. PFC is the idea that combat medics must be prepared to provide medical care to serious casualties in the field without the support of robust medical infrastructure or resources in the event of delayed medical evacuation. With limited resources, significant distances to travel before definitive care, and an inability to evacuate in a timely fashion, medical care during exploration spaceflight constitutes the ultimate example PFC. One of the main capability gaps for PFC in both military and spaceflight settings is the need for technologies for individualized monitoring of a patient's physiological status. A monitoring capability known as the compensatory reserve measurement (CRM) meets such a requirement. CRM is a small, portable, wearable technology that uses a machine learning and feature extraction-based algorithm to assess real-time changes in hundreds of specific features of arterial waveforms. Future development and advancement of CRM still faces engineering challenges to develop ruggedized wearable sensors that can measure waveforms for determining CRM from multiple sites on the body and account for less than optimal conditions (sweat, water, dirt, blood, movement, etc.). We show here the utility of a military wearable technology, CRM, which can be translated to space exploration.

Detecting central hypovolemia in simulated hypovolemic shock by automated feature extraction with principal component analysis

Assessment of the volume status by blood pressure (BP) monitoring is difficult, since baroreflex control of BP makes it insensitive to blood loss up to about one liter. We hypothesized that a machine learning model recognizes the progression of central hypovolemia toward presyncope by extracting information of the noninvasive blood pressure waveform parametrized through principal component analysis. This was tested in healthy volunteers exposed to simulated hemorrhage by lower body negative pressure (LBNP).

Fifty‐six healthy volunteers were subjected to progressive central hypovolemia. A support vector machine was trained on the blood pressure waveform. Three classes of progressive stages of hypovolemia were defined. The model was optimized for the number of principal components and regularization parameter for penalizing misclassification (cost): C. Model performance was expressed as accuracy, mean squared error (MSE), and kappa statistic (inter‐rater agreement).

Forty‐six subjects developed presyncope of which 41 showed an increase in model classification severity from baseline to presyncope. In five of the remaining nine subjects (1 was excluded) it stagnated. Classification of samples during baseline and end‐stage LBNP had the highest accuracy (95% and 50%, respectively). Baseline and first stage of LBNP demonstrated the lowest MSE (0.01 respectively 0.32). Model MSE and accuracy did not improve for C values exceeding 0.01. Adding more than five principal components did not further improve accuracy or MSE. Increment in kappa halted after 10 principal components had been added. Automated feature extraction of the blood pressure waveform allows modeling of progressive hypovolemia with a support vector machine. The model distinguishes classes between baseline and presyncope.

Measurement of compensatory reserve predicts racial differences in tolerance to simulated hemorrhage in women

BACKGROUND

The compensatory reserve measurement (CRM) has been established to accurately measure the body's total integrated capacity to compensate for physiologic states of reduced central blood volume and predict hemodynamic decompensation associated with inadequate tissue oxygenation. We previously demonstrated that African American (AA) women have a higher tolerance to reductions in central blood volume. Therefore, we tested the hypothesis that the CRM would identify racial differences during simulated hemorrhage, before the onset of traditional signs/symptoms.

METHODS

We performed a retrospective analysis during simulated hemorrhage using lower-body negative pressure (LBNP) in 23 AA (22 ± 1 years; 24 ± 1 kg/m) and 31 white women (WW) (20 ± 1 years; 23 ± 1 kg/m). Beat-by-beat blood pressure (BP) and heart rate (HR) were recorded during progressive lower body negative pressure to presyncope. The BP waveforms were analyzed using a machine-learning algorithm to derive the CRM at each lower body negative pressure stage.

RESULTS

Resting mean arterial BP (AA, 78 ± 3 mm Hg vs. WW, 74 ± 2 mm Hg) and HR (AA, 68 ± 2 bpm vs. WW, 65 ± 2 bpm) were similar between groups. The CRM progressively decreased during LBNP in both groups; however, the rate of decline in the CRM was less (p < 0.05) in AA. The CRM was 4% higher in AA at -15 mm Hg LBNP and progressively increased to 21% higher at -50 mm Hg LBNP (p < 0.05). However, changes in BP and HR were not different between groups.

CONCLUSION

These data support the notion that the greater tolerance to simulated hemorrhage induced by LBNP in AA women can be explained by their greater capacity to protect the reserve to compensate for progressive central hypovolemia compared with WW, independent of standard vital signs.

LEVEL OF EVIDENCE

Diagnostic test, level II.

Prediction of Recovery From Severe Hemorrhagic Shock Using Logistic Regression

This paper implements logistic regression models (LRMs) and feature selection for creating a predictive model for recovery form hemorrhagic shock (HS) with resuscitation using blood in the multiple experimental rat animal protocols. A total of 61 animals were studied across multiple HS experiments, which encompassed two different HS protocols and two resuscitation protocols using blood stored for short periods using five different techniques. Twenty-seven different systemic hemodynamics, cardiac function, and blood gas parameters were measured in each experiment, of which feature selection deemed only 25% of the them as relevant. The reduced feature set was used to train a final logistic regression model. A final test set accuracy is 84% compared to 74% for a baseline classifier using only MAP and HR measurements. Receiver operating characteristics (ROC) curve analysis and Cohens kappa statistics were also used as measures of performance, with the final reduced model outperforming the model, including all parameters. Our results suggest that LRMs trained with a combination of systemic hemodynamics, cardiac function, and blood gas parameters measured at multiple timepoints during HS can successfully classify HS recovery groups. Our results show the predictive ability of traditional and novel hemodynamic and cardiac function features and their combinations, many of which had not previously been taken into consideration, for monitoring HS. Furthermore, we have devised an effective methodology for feature selection and shown ways in which the performance of such predictive models should be assessed in future studies.

Differentiating compensatory mechanisms associated with low tolerance to central hypovolemia in women

Women generally display lower tolerance to acute central hypovolemia than men. The measurement of compensatory reserve (CRM) is a novel metric that provides information about the sum total of all mechanisms that together work to compensate for the relative blood volume deficit. Hemodynamic decompensation occurs with depletion of the CRM (i.e., 0% CRM). In the present study, we hypothesized that the lower tolerance to progressive central hypovolemia reported in women can be explained by a faster reduction rate in CRM compared with men rather than sex differences in absolute integrated compensatory responses. Continuous, noninvasive measures of CRM were collected from 208 healthy volunteers (107 men and 85 women) who underwent progressive stepwise central hypovolemia induced by lower body negative pressure to the point of presyncope. Comparisons revealed shorter ( P < 0.01) times in female participants compared with male participants to reach 30% and 0% CRM. Similarly, the lower body negative pressure level, represented by the cumulative stress index, was less at 30% and 0% CRM in women compared with men ( P < 0.01). Changes in hemodynamic responses and frequency-domain data (oscillations in cerebral blood flow velocity and mean arterial blood pressure) were similar between men and women at 0% CRM ( P > 0.05). We conclude that compensatory responses to central hypovolemia in women were similar to men but were depleted at a faster rate compared with men. The earlier depletion of the compensatory reserve in women appears to be influenced by failure to maintain adequate cerebral oxygen delivery.

NEW & NOTEWORTHY We compared hemodynamic and metabolic responses in men and women to experimentally controlled reductions in central blood volume at physiologically equivalent levels of compensatory reserve. We corroborated previous findings that females have lower tolerance to central hypovolemia than males but demonstrated for the first time that compensatory responses are similar. Our findings suggest lower tolerance to central hypovolemia in women results from reaching critical cerebral delivery of oxygen faster than men.

Unmasking the Hypovolemic Shock Continuum: The Compensatory Reserve

Hypovolemic shock exists as a spectrum, with its early stages characterized by subtle pathophysiologic tissue insults and its late stages defined by multi-system organ dysfunction. The importance of timely detection of shock is well known, as early interventions improve mortality, while delays render these same interventions ineffective. However, detection is limited by the monitors, parameters, and vital signs that are traditionally used in the intensive care unit (ICU). Many parameters change minimally during the early stages, and when they finally become abnormal, hypovolemic shock has already occurred. The compensatory reserve (CR) is a parameter that represents a new paradigm for assessing physiologic status, as it comprises the sum total of compensatory mechanisms that maintain adequate perfusion to vital organs during hypovolemia. When these mechanisms are overwhelmed, hemodynamic instability and circulatory collapse will follow. Previous studies involving CR measurements demonstrated their utility in detecting central blood volume loss before hemodynamic parameters and vital signs changed. Measurements of the CR have also been used in clinical studies involving patients with traumatic injuries or bleeding, and the results from these studies have been promising. Moreover, these measurements can be made at the bedside, and they provide a real-time assessment of hemodynamic stability. Given the need for rapid diagnostics when treating critically ill patients, CR measurements would complement parameters that are currently being used. Consequently, the purpose of this article is to introduce a conceptual framework where the CR represents a new approach to monitoring critically ill patients. Within this framework, we present evidence to support the notion that the use of the CR could potentially improve the outcomes of ICU patients by alerting intensivists to impending hypovolemic shock before its onset.

Robust monitoring of hypovolemia in intensive care patients using photoplethysmogram signals

The paper presents a fingertip photoplethysmography based technique to assess patient fluid status that is robust to waveform artifacts and health variability in the underlying patient population. The technique is intended for use in intensive care units, where patients are at risk for hypovolemia, and signal artifacts and inter-patient variations in health are common. Input signals are preprocessed to remove artifact, then a parameter-invariant statistic is calculated to remove effects of patient-specific physiology. Patient data from the Physionet MIMICII database was used to evaluate the performance of this technique. The proposed method was able to detect hypovolemia within 24 hours of onset in all hypovolemic patients tested, while producing minimal false alarms over non-hypovolemic patients.

The Effect of Passive Heat Stress and Exercise-Induced Dehydration on the Compensatory Reserve During Simulated Hemorrhage

Compensatory reserve represents the proportion of physiological responses engaged to compensate for reductions in central blood volume before the onset of decompensation. We hypothesized that compensatory reserve would be reduced by hyperthermia and exercise-induced dehydration, conditions often encountered on the battlefield. Twenty healthy males volunteered for two separate protocols during which they underwent lower-body negative pressure (LBNP) to hemodynamic decompensation (systolic blood pressure <80 mm Hg). During protocol #1, LBNP was performed following a passive increase in core temperature of ∼1.2°C (HT) or a normothermic time-control period (NT). During protocol #2, LBNP was performed following exercise during which: fluid losses were replaced (hydrated), fluid intake was restricted and exercise ended at the same increase in core temperature as hydrated (isothermic dehydrated), or fluid intake was restricted and exercise duration was the same as hydrated (time-match dehydrated). Compensatory reserve was estimated with the compensatory reserve index (CRI), a machine-learning algorithm that extracts features from continuous photoplethysmograph signals. Prior to LBNP, CRI was reduced by passive heating [NT: 0.87 (SD 0.09) vs. HT: 0.42 (SD 0.19) units, P <0.01] and exercise-induced dehydration [hydrated: 0.67 (SD 0.19) vs. isothermic dehydrated: 0.52 (SD 0.21) vs. time-match dehydrated: 0.47 (SD 0.25) units; P <0.01 vs. hydrated]. During subsequent LBNP, CRI decreased further and its rate of change was similar between conditions. CRI values at decompensation did not differ between conditions. These results suggest that passive heating and exercise-induced dehydration limit the body's physiological reserve to compensate for further reductions in central blood volume.

Comparison of compensatory reserve and arterial lactate as markers of shock and resuscitation

BACKGROUND

During traumatic hemorrhage, the ability to identify shock and intervene before decompensation is paramount to survival. Lactate is extremely sensitive to shock, and its clearance has been demonstrated a useful gauge of shock and resuscitation status. Though lactate can be measured in the field, logistical constraints render it impractical in certain environments. The compensatory reserve represents a new clinical measurement reflecting the remaining capacity to compensate for hypoperfusion. We hypothesized the compensatory reserve index (CRI) would be an effective surrogate marker of shock and resuscitation compared to lactate.

METHODS

The CRI device was placed on consecutive patients meeting trauma center activation criteria and remained on the patient until discharge, admission, or transport to operating suite. All subjects had a lactate level measured as part of their routine admission metabolic analysis. Time-corresponding CRI and lactate values were matched in regards to initial and subsequent lactate levels. Mean time from lactate sample collection to data availability in the electronic medical record was calculated. Predictive capacity of CRI and lactate in predicting hemorrhage was determined by receiver-operator characteristic curve analysis. Correlation analysis was performed to determine if any association existed between changing CRI and lactate values.

RESULTS

Receiver-operator characteristic (ROC) curves were generated and area under the curve was 0.8052 and 0.8246 for CRI and lactate, respectively. There was no significant difference in each parameter's ability to predict hemorrhage (p = 0.8015). The mean duration from lactate sample collection to clinical availability was 44 minutes whereas CRI values were available immediately. Analysis of the concomitant change in serial CRI and lactate levels revealed a Spearman's correlation coefficient of -0.73 (p < 0.01).

CONCLUSION

CRI performed with equivalent predictive capacity to lactate with respect to identifying initial perfusion status associated with hemorrhage and subsequent resuscitation.

LEVEL OF EVIDENCE

Diagnostic, Level II.

The effect of blood transfusion on compensatory reserve: A prospective clinical trial

BACKGROUND

Bleeding activates the body's compensatory mechanisms, causing changes in vital signs to appear late in the course of progressive blood loss. These vital signs are maintained even when up to 30% to 40% of blood volume is lost. Laboratory tests such as hemoglobin, hematocrit, lactate, and base deficit levels do not change during acute phase of bleeding. The compensatory reserve measurement (CRM) represents a new paradigm that measures the total of all physiological compensatory mechanisms, using noninvasive photoplethysmography to read changes in arterial waveforms. This study compared CRM to traditional vital signs and laboratory tests in actively bleeding patients.

METHODS

Study patients had gastrointestinal bleeding and required red blood cell (RBC) transfusion (n = 31). Control group patients had similar demographic and medical backgrounds. They were undergoing minor surgical procedures and not expected to receive RBC transfusion. Vital signs, mean arterial pressure, pulse pressure, hemoglobin and hematocrit levels, and CRM were recorded before and after RBC transfusion or the appropriate time interval for the control group. Receiver operator characteristic curves were plotted and areas under the curves (AUCs) were compared.

RESULTS

CRM increased 10.5% after RBC transfusion, from 0.77 to 0.85 (p < 0.005). Hemoglobin level increased 22.4% after RBC transfusion from 7.3 to 8.7 (p < 0.005). Systolic and diastolic blood pressure, mean arterial pressure, pulse pressure, and heart rate did change significantly. The AUC for CRM as a single measurement for predicting hemorrhage at admission was 0.79, systolic blood pressure was 0.62, for heart rate was 0.60, and pulse pressure was 0.36.

CONCLUSIONS

This study demonstrated that CRM is more sensitive to changes in blood volume than traditional vital signs are and could be used to monitor and assess resuscitation of actively bleeding patients.

LEVEL OF EVIDENCE

Care management, level II.

A machine-learning based analysis for the recognition of progressive central hypovolemia

OBJECTIVE

Traditional patient monitoring during surgery includes heart rate (HR), blood pressure (BP) and peripheral oxygen saturation. However, their use as predictors for central hypovolemia is limited, which may lead to cerebral hypoperfusion. The aim of this study was to develop a monitoring model that can indicate a decrease in central blood volume (CBV) at an early stage.

APPROACH

Twenty-eight healthy subjects (aged 18-50 years) were included. Lower body negative pressure (-50 mmHg) was applied to induce central hypovolemia until the onset of pre-syncope. Ten beat-to-beat and four discrete parameters were measured, normalized, and filtered with a 30 s moving window. Time to pre-syncope was scaled from 100%-0%. A total of 100 neural networks with 5, 10, 15, 20, or 25 neurons in their respective hidden layer were trained by 10, 20, 40, 80, 160, or 320 iterations to predict time to pre-syncope for each subject. The network with the lowest average slope of a fitted line over all subjects was chosen as optimal.

MAIN RESULTS

The optimal generalized model consisted of 10 hidden neurons, trained using 80 iterations. The slope of the fitted line on the average prediction was  -0.64 (SD 0.35). The model recognizes in 75% of the subjects the need for intervention at  >200 s before pre-syncope.

SIGNIFICANCE

We developed a neural network based on a set of physiological variables, which indicates a decrease in CBV even in the absence of HR and BP changes. This should allow timely intervention and prevent the development of symptomatic cerebral hypoperfusion.

Wearable technology for compensatory reserve to sense hypovolemia

Traditional monitoring technologies fail to provide accurate or early indications of hypovolemia-mediated extremis because physiological systems (as measured by vital signs) effectively compensate until circulatory failure occurs. Hypovolemia is the most life-threatening physiological condition associated with circulatory shock in hemorrhage or sepsis, and it impairs one's ability to sustain physical exertion during heat stress. This review focuses on the physiology underlying the development of a novel noninvasive wearable technology that allows for real-time evaluation of the cardiovascular system's ability to compensate to hypovolemia, or its compensatory reserve, which provides an individualized estimate of impending circulatory collapse. Compensatory reserve is assessed by real-time changes (sampled millions of times per second) in specific features (hundreds of features) of arterial waveform analog signals that can be obtained from photoplethysmography using machine learning and feature extraction techniques. Extensive experimental evidence employing acute reductions in central blood volume (using lower-body negative pressure, blood withdrawal, heat stress, dehydration) demonstrate that compensatory reserve provides the best indicator for early and accurate assessment for compromises in blood pressure, tissue perfusion, and oxygenation in resting human subjects. Engineering challenges exist for the development of a ruggedized wearable system that can measure signals from multiple sites, improve signal-to-noise ratios, be customized for use in austere conditions (e.g., battlefield, patient transport), and be worn during strenuous physical activity.

Accuracy of Blood Loss Measurement during Cesarean Delivery

Objective This study aims to compare the accuracy of visual, quantitative gravimetric, and colorimetric methods used to determine blood loss during cesarean delivery procedures employing a hemoglobin extraction assay as the reference standard. Study Design In 50 patients having cesarean deliveries blood loss determined by assays of hemoglobin content on surgical sponges and in suction canisters was compared with obstetricians' visual estimates, a quantitative gravimetric method, and the blood loss determined by a novel colorimetric system. Agreement between the reference assay and other measures was evaluated by the Bland-Altman method. Results Compared with the blood loss measured by the reference assay (470 ± 296 mL), the colorimetric system (572 ± 334 mL) was more accurate than either visual estimation (928 ± 261 mL) or gravimetric measurement (822 ± 489 mL). The correlation between the assay method and the colorimetric system was more predictive (standardized coefficient = 0.951, adjusted R2 = 0.902) than either visual estimation (standardized coefficient = 0.700, adjusted R2 = 00.479) or the gravimetric determination (standardized coefficient = 0.564, adjusted R2 = 0.304). Conclusion During cesarean delivery, measuring blood loss using colorimetric image analysis is superior to visual estimation and a gravimetric method. Implementation of colorimetric analysis may enhance the ability of management protocols to improve clinical outcomes.

The physiology of blood loss and shock: New insights from a human laboratory model of hemorrhage

The ability to quickly diagnose hemorrhagic shock is critical for favorable patient outcomes. Therefore, it is important to understand the time course and involvement of the various physiological mechanisms that are active during volume loss and that have the ability to stave off hemodynamic collapse. This review provides new insights about the physiology that underlies blood loss and shock in humans through the development of a simulated model of hemorrhage using lower body negative pressure. In this review, we present controlled experimental results through utilization of the lower body negative pressure human hemorrhage model that provide novel insights on the integration of physiological mechanisms critical to the compensation for volume loss. We provide data obtained from more than 250 human experiments to classify human subjects into two distinct groups: those who have a high tolerance and can compensate well for reduced central blood volume (e.g. hemorrhage) and those with low tolerance with poor capacity to compensate.We include the conceptual introduction of arterial pressure and cerebral blood flow oscillations, reflex-mediated autonomic and neuroendocrine responses, and respiration that function to protect adequate tissue oxygenation through adjustments in cardiac output and peripheral vascular resistance. Finally, unique time course data are presented that describe mechanistic events associated with the rapid onset of hemodynamic failure (i.e. decompensatory shock). Impact Statement Hemorrhage is the leading cause of death in both civilian and military trauma. The work submitted in this review is important because it advances the understanding of mechanisms that contribute to the total integrated physiological compensations for inadequate tissue oxygenation (i.e. shock) that arise from hemorrhage. Unlike an animal model, we introduce the utilization of lower body negative pressure as a noninvasive model that allows for the study of progressive reductions in central blood volume similar to those reported during actual hemorrhage in conscious humans to the onset of hemodynamic decompensation (i.e. early phase of decompensatory shock), and is repeatable in the same subject. Understanding the fundamental underlying physiology of human hemorrhage helps to test paradigms of critical care medicine, and identify and develop novel clinical practices and technologies for advanced diagnostics and therapeutics in patients with life-threatening blood loss.

Integrated Compensatory Responses in a Human Model of Hemorrhage

Hemorrhage is the leading cause of trauma-related deaths, partly because early diagnosis of the severity of blood loss is difficult. Assessment of hemorrhaging patients is difficult because current clinical tools provide measures of vital signs that remain stable during the early stages of bleeding due to compensatory mechanisms. Consequently, there is a need to understand and measure the total integration of mechanisms that compensate for reduced circulating blood volume and how they change during ongoing progressive hemorrhage. The body's reserve to compensate for reduced circulating blood volume is called the 'compensatory reserve'. The compensatory reserve can be accurately evaluated with real-time measurements of changes in the features of the arterial waveform measured with the use of a high-powered computer. Lower Body Negative Pressure (LBNP) has been shown to simulate many of the physiological responses in humans associated with hemorrhage, and is used to study the compensatory response to hemorrhage. The purpose of this study is to demonstrate how compensatory reserve is assessed during progressive reductions in central blood volume with LBNP as a simulation of hemorrhage.

State-of-the-art monitoring in treatment of dengue shock syndrome: a case series

BACKGROUND

Early recognition and treatment of circulatory volume loss is essential in the clinical management of dengue viral infection. We hypothesized that a novel computational algorithm, originally developed for noninvasive monitoring of blood loss in combat casualties, could: (1) indicate the central volume status of children with dengue during the early stages of "shock"; and (2) track fluid resuscitation status.

METHODS

Continuous noninvasive photoplethysmographic waveforms were collected over a 5-month period from three children of Thai ethnicity with clinical suspicion of dengue. Waveform data were processed by the algorithm to calculate each child's Compensatory Reserve Index, where 1 represents supine normovolemia and 0 represents the circulatory volume at which hemodynamic decompensation occurs. Values between 1 and 0 indicate the proportion of reserve remaining before hemodynamic decompensation.

RESULTS

This case report describes a 7-year-old Thai boy, another 7-year-old Thai boy, and a 9-year-old Thai boy who exhibited signs and symptoms of dengue shock syndrome; all the children had secondary dengue virus infections, documented by serology and reverse transcriptase polymerase chain reaction. The three boys experienced substantial plasma leakage demonstrated by pleural effusion index >25, ascites, and >20 % hemoconcentration. They received fluid administered intravenously; one received a blood transfusion. All three boys showed a significantly low initial Compensatory Reserve Index (≥0.20), indicating a clinical diagnosis of "near shock". Following 5 days with fluid resuscitation treatment, their Compensatory Reserve Index increased towards "normovolemia" (that is, Compensatory Reserve Index >0.75).

CONCLUSIONS

The results from these cases demonstrate a new variation in the diagnostic capability to manage patients with dengue shock syndrome. The findings shed new light on a method that can avoid possible adverse effects of shock by noninvasive measurement of a patient's compensatory reserve rather than standard vital signs or invasive diagnostic methods.

Comparison of compensatory reserve during lower-body negative pressure and hemorrhage in nonhuman primates

Compensatory reserve was measured in baboons (n = 13) during hemorrhage (Hem) and lower-body negative pressure (LBNP) using a machine-learning algorithm developed to estimate compensatory reserve by detecting reductions in central blood volume during LBNP. The algorithm calculates compensatory reserve index (CRI) from normovolemia (CRI = 1) to cardiovascular decompensation (CRI = 0). The hypothesis was that Hem and LBNP will elicit similar CRI values and that CRI would have higher specificity than stroke volume (SV) in predicting decompensation. Blood was removed in four steps: 6.25%, 12.5%, 18.75%, and 25% of total blood volume. Four weeks after Hem, the same animals were subjected to four levels of LBNP that was matched on the basis of their central venous pressure. Data (mean ± 95% confidence interval) indicate that CRI decreased (P < 0.001) from baseline during Hem (0.69 ± 0.10, 0.57 ± 0.09, 0.36 ± 0.10, 0.16 ± 0.08, and 0.08 ± 0.03) and LBNP (0.76 ± 0.05, 0.66 ± 0.08, 0.36 ± 0.13, 0.23 ± 0.11, and 0.14 ± 0.09). CRI was not different between Hem and LBNP (P = 0.20). Linear regression analysis between Hem CRI and LBNP CRI revealed a slope of 1.03 and a correlation coefficient of 0.96. CRI exhibited greater specificity than SV in both Hem (92.3 vs. 82.1) and LBNP (94.8 vs. 83.1) and greater ROC AUC in Hem (0.94 vs. 0.84) and LBNP (0.94 vs. 0.92). These data support the hypothesis that Hem and LBNP elicited the same CRI response, suggesting that measurement of compensatory reserve is superior to SV as a predictor of cardiovascular decompensation.

A Signal Processing Approach for Detection of Hemodynamic Instability before Decompensation

Advanced hemodynamic monitoring is a critical component of treatment in clinical situations where aggressive yet guided hemodynamic interventions are required in order to stabilize the patient and optimize outcomes. While there are many tools at a physician’s disposal to monitor patients in a hospital setting, the reality is that none of these tools allow hi-fidelity assessment or continuous monitoring towards early detection of hemodynamic instability. We present an advanced automated analytical system which would act as a continuous monitoring and early warning mechanism that can indicate pending decompensation before traditional metrics can identify any clinical abnormality. This system computes novel features or bio-markers from both heart rate variability (HRV) as well as the morphology of the electrocardiogram (ECG). To compare their effectiveness, these features are compared with the standard HRV based bio-markers which are commonly used for hemodynamic assessment. This study utilized a unique database containing ECG waveforms from healthy volunteer subjects who underwent simulated hypovolemia under controlled experimental settings. A support vector machine was utilized to develop a model which predicts the stability or instability of the subjects. Results showed that the proposed novel set of features outperforms the traditional HRV features in predicting hemodynamic instability.

A disposable, flexible skin patch for clinical optical perfusion monitoring at multiple depths

Stable, relative localization of source and detection fibers is necessary for clinical implementation of quantitative optical perfusion monitoring methods such as diffuse correlation spectroscopy (DCS) and diffuse reflectance spectroscopy (DRS). A flexible and compact device design is presented as a platform for simultaneous monitoring of perfusion at a range of depths, enabled by precise location of optical fibers in a robust and secure adhesive patch. We will discuss preliminary data collected on human subjects in a lower body negative pressure model for hypovolemic shock. These data indicate that this method facilitates simple and stable simultaneous monitoring of perfusion at multiple depths and within multiple physiological compartments.