Trending autoregulatory indices during treatment for traumatic brain injury

Evaluation and Application of Ultra-Low-Resolution Pressure Reactivity Index in Moderate or Severe Traumatic Brain Injury

BACKGROUND

The pressure reactivity index (PRx) has emerged as a surrogate method for the continuous bedside estimation of cerebral autoregulation and a predictor of unfavorable outcome after traumatic brain injury (TBI). However, calculation of PRx require continuous high-resolution monitoring currently limited to specialized intensive care units. The aim of this study was to evaluate a new index, the ultra-low-frequency PRx (UL-PRx) sampled at ∼0.0033 Hz at ∼5 minutes periods, and to investigate its association with outcome.

METHODS

Demographic data, admission Glasgow coma scale, in-hospital mortality and Glasgow outcome scale extended at 12 months were extracted from electronic records. The filtering and preparation of time series of intracranial pressure (ICP), mean arterial pressure and cerebral perfusion pressure (CPP), and calculation of the indices (UL-PRx, Δ-optimal CPP), were performed in MATLAB using an in-house algorithm.

RESULTS

A total of 164 TBI patients were included in the study; in-hospital and 12-month mortality was 29.3% and 38.4%, respectively, and 64% of patients had poor neurological outcome at 12 months. On univariate analysis, ICP, CPP, UL-PRx, and ΔCPPopt were associated with 12-month mortality. After adjusting for age, Glasgow coma scale, ICP and CPP, mean UL-PRx and UL-PRx thresholds of 0 and +0.25 remained associated with 12-month mortality. Similar findings were obtained for in-hospital mortality. For mean UL-PRx, the area under the receiver operating characteristic curves for in-hospital and 12-month mortality were 0.78 (95% confidence interval [CI]: 0.69-0.87; P <0.001) and 0.70 (95% CI: 0.61-0.79; P <0.001), respectively, and 0.65 (95% CI: 0.57-0.74; P =0.001) for 12-month neurological outcome.

CONCLUSIONS

Our findings indicate that ultra-low-frequency sampling might provide sufficient resolution to derive information about the state of cerebrovascular autoregulation and prediction of 12-month outcome in TBI patients.

Emulating clinical pressure waveforms in cell culture using an Arduino-controlled millifluidic 3D-printed platform for 96-well plates

High blood pressure is the primary risk factor for heart disease, the leading cause of death globally. Despite this, current methods to replicate physiological pressures in vitro remain limited in sophistication and throughput. Single-chamber exposure systems allow for only one pressure condition to be studied at a time and the application of dynamic pressure waveforms is currently limited to simple sine, triangular, or square waves. Here, we introduce a high-throughput hydrostatic pressure exposure system for 96-well plates. The platform can deliver a fully-customizable pressure waveform to each column of the plate, for a total of 12 simultaneous conditions. Using clinical waveform data, we are able to replicate real patients' blood pressures as well as other medically-relevant pressures within the body and have assembled a small patient-derived waveform library of some key physiological locations. As a proof of concept, human umbilical vein endothelial cells (HUVECs) survived and proliferated for 3 days under a wide range of static and dynamic physiologic pressures ranging from 10 mm Hg to 400 mm Hg. Interestingly, pathologic and supraphysiologic pressure exposures did not inhibit cell proliferation. By integrating with, rather than replacing, ubiquitous lab cultureware it is our hope that this device will facilitate the incorporation of hydrostatic pressure into standard cell culture practice.

Dynamic detection and reversal of myocardial ischemia using an artificially intelligent bioelectronic medicine

Myocardial ischemia is spontaneous, frequently asymptomatic, and contributes to fatal cardiovascular consequences. Importantly, myocardial sensory networks cannot reliably detect and correct myocardial ischemia on their own. Here, we demonstrate an artificially intelligent and responsive bioelectronic medicine, where an artificial neural network (ANN) supplements myocardial sensory networks, enabling reliable detection and correction of myocardial ischemia. ANNs were first trained to decode spontaneous cardiovascular stress and myocardial ischemia with an overall accuracy of ~92%. ANN-controlled vagus nerve stimulation (VNS) significantly mitigated major physiological features of myocardial ischemia, including ST depression and arrhythmias. In contrast, open-loop VNS or ANN-controlled VNS following a caudal vagotomy essentially failed to reverse cardiovascular pathophysiology. Last, variants of ANNs were used to meet clinically relevant needs, including interpretable visualizations and unsupervised detection of emerging cardiovascular stress. Overall, these preclinical results suggest that ANNs can potentially supplement deficient myocardial sensory networks via an artificially intelligent bioelectronic medicine system.

Machine Learning-Based Continuous Intracranial Pressure Prediction for Traumatic Injury Patients

Structured Abstract—Objective: Abnormal elevation of intracranial pressure (ICP) can cause dangerous or even fatal outcomes. The early detection of high intracranial pressure events can be crucial in saving lives in an intensive care unit (ICU). Despite many applications of machine learning (ML) techniques related to clinical diagnosis, ML applications for continuous ICP detection or short-term predictions have been rarely reported. This study proposes an efficient method of applying an artificial recurrent neural network on the early prediction of ICP evaluation continuously for TBI patients. Methods: After ICP data preprocessing, the learning model is generated for thirteen patients to continuously predict the ICP signal occurrence and classify events for the upcoming 10 minutes by inputting the previous 20-minutes of the ICP signal. Results: As the overall model performance, the average accuracy is 94.62%, the average sensitivity is 74.91%, the average specificity is 94.83%, and the average root mean square error is approximately 2.18 mmHg. Conclusion: This research addresses a significant clinical problem with the management of traumatic brain injury patients. The machine learning model data enables early prediction of ICP continuously in a real-time fashion, which is crucial for appropriate clinical interventions. The results show that our machine learning-based model has high adaptive performance, accuracy, and efficiency.

Non-Invasive Hemodynamics Monitoring System Based on Electrocardiography via Deep Convolutional Autoencoder

This study evaluates cardiovascular and cerebral hemodynamics systems by only using non-invasive electrocardiography (ECG) signals. The Massachusetts General Hospital/Marquette Foundation (MGH/MF) and Cerebral Hemodynamic Autoregulatory Information System Database (CHARIS DB) from the PhysioNet database are used for cardiovascular and cerebral hemodynamics, respectively. For cardiovascular hemodynamics, the ECG is used for generating the arterial blood pressure (ABP), central venous pressure (CVP), and pulmonary arterial pressure (PAP). Meanwhile, for cerebral hemodynamics, the ECG is utilized for the intracranial pressure (ICP) generator. A deep convolutional autoencoder system is applied for this study. The cross-validation method with Pearson’s linear correlation (R), root mean squared error (RMSE), and mean absolute error (MAE) are measured for the evaluations. Initially, the ECG is used to generate the cardiovascular waveform. For the ABP system—the systolic blood pressure (SBP) and diastolic blood pressures (DBP)—the R evaluations are 0.894 ± 0.004 and 0.881 ± 0.005, respectively. The MAE evaluations for SBP and DBP are, respectively, 6.645 ± 0.353 mmHg and 3.210 ± 0.104 mmHg. Furthermore, for the PAP system—the systolic and diastolic pressures—the R evaluations are 0.864 ± 0.003 mmHg and 0.817 ± 0.006 mmHg, respectively. The MAE evaluations for systolic and diastolic pressures are, respectively, 3.847 ± 0.136 mmHg and 2.964 ± 0.181 mmHg. Meanwhile, the mean CVP evaluations are 0.916 ± 0.001, 2.220 ± 0.039 mmHg, and 1.329 ± 0.036 mmHg, respectively, for R, RMSE, and MAE. For the mean ICP evaluation in cerebral hemodynamics, the R and MAE evaluations are 0.914 ± 0.003 and 2.404 ± 0.043 mmHg, respectively. This study, as a proof of concept, concludes that the non-invasive cardiovascular and cerebral hemodynamics systems can be potentially investigated by only using the ECG signal.

Near-infrared Spectroscopy-derived Cerebral Autoregulation Indices Independently Predict Clinical Outcome in Acutely Ill Comatose Patients

OBJECTIVE

Outcome prediction in comatose patients with acute brain injury remains challenging. Regional cerebral oxygenation (rSO2) derived from near-infrared spectroscopy (NIRS) is a surrogate for cerebral blood flow and can be used to calculate cerebral autoregulation (CA) continuously at the bedside from the derived cerebral oximetry index (COx). We hypothesized that COx derived thresholds for CA are associated with outcomes in patients with acute coma from neurological injury.

METHODS

A prospective cohort study was conducted in 88 acutely comatose adults with heterogenous brain injury diagnoses who were continuously monitored with COx for up to 3 consecutive days. Multivariable logistic regression was performed to investigate association between averaged COx and short (in-hospital and 3 mo) and long-term (6 mo) outcomes.

RESULTS

Six month mortality rate was 62%. Median COx in nonsurvivors at hospital discharge was 0.082 [interquartile range, IQR: 0.045 to 0.160] compared with 0.042 [IQR: -0.005 to 0.110] in survivors (P=0.012). At 6 months, median COx was 0.075 [IQR: 0.27 to 0.158] in nonsurvivors compared with 0.029 [IQR: -0.015 to 0.077] in survivors (P=0.02). In the multivariable logistic regression model adjusted for confounders, average COx ≥0.05 was associated with both in-hospital mortality (adjusted odds ratio [OR]=2.9, 95% confidence interval [CI]=1.15-7.33, P=0.02), mortality at 6 months (adjusted OR=4.4, 95% CI=1.41-13.7, P=0.01), and severe disability (modified Rankin Score ≥4) at 6 months (adjusted OR=4.4, 95% CI=1.07-17.8, P=0.04). Area under the receiver operating characteristic curve for predicting mortality and severe disability at 6 months were 0.783 and 0.825, respectively.

CONCLUSIONS

Averaged COx ≥0.05 is independently associated with short and long-term mortality and long-term severe disability in acutely comatose adults with neurological injury. We propose that COx ≥0.05 represents an accurate threshold to predict long-term functional outcome in acutely comatose adults.

Clinical Decision Support for Traumatic Brain Injury: Identifying a Framework for Practical Model-Based Intracranial Pressure Estimation at Multihour Timescales

BACKGROUND

The clinical mitigation of intracranial hypertension due to traumatic brain injury requires timely knowledge of intracranial pressure to avoid secondary injury or death. Noninvasive intracranial pressure (nICP) estimation that operates sufficiently fast at multihour timescales and requires only common patient measurements is a desirable tool for clinical decision support and improving traumatic brain injury patient outcomes. However, existing model-based nICP estimation methods may be too slow or require data that are not easily obtained.

OBJECTIVE

This work considers short- and real-time nICP estimation at multihour timescales based on arterial blood pressure (ABP) to better inform the ongoing development of practical models with commonly available data.

METHODS

We assess and analyze the effects of two distinct pathways of model development, either by increasing physiological integration using a simple pressure estimation model, or by increasing physiological fidelity using a more complex model. Comparison of the model approaches is performed using a set of quantitative model validation criteria over hour-scale times applied to model nICP estimates in relation to observed ICP.

RESULTS

The simple fully coupled estimation scheme based on windowed regression outperforms a more complex nICP model with prescribed intracranial inflow when pulsatile ABP inflow conditions are provided. We also show that the simple estimation data requirements can be reduced to 1-minute averaged ABP summary data under generic waveform representation.

CONCLUSIONS

Stronger performance of the simple bidirectional model indicates that feedback between the systemic vascular network and nICP estimation scheme is crucial for modeling over long intervals. However, simple model reduction to ABP-only dependence limits its utility in cases involving other brain injuries such as ischemic stroke and subarachnoid hemorrhage. Additional methodologies and considerations needed to overcome these limitations are illustrated and discussed.

Forecasting intracranial hypertension using multi-scale waveform metrics

OBJECTIVE

Acute intracranial hypertension is an important risk factor of secondary brain damage after traumatic brain injury. Hypertensive episodes are often diagnosed reactively, leading to late detection and lost time for intervention planning. A pro-active approach that predicts critical events several hours ahead of time could assist in directing attention to patients at risk.

APPROACH

We developed a prediction framework that forecasts onsets of acute intracranial hypertension in the next 8 h. It jointly uses cerebral auto-regulation indices, spectral energies and morphological pulse metrics to describe the neurological state of the patient. One-minute base windows were compressed by computing signal metrics, and then stored in a multi-scale history, from which physiological features were derived.

MAIN RESULTS

Our model predicted events up to 8 h in advance with an alarm recall rate of 90% at a precision of 30% in the MIMIC-III waveform database, improving upon two baselines from the literature. We found that features derived from high-frequency waveforms substantially improved the prediction performance over simple statistical summaries of low-frequency time series, and each of the three feature classes contributed to the performance gain. The inclusion of long-term history up to 8 h was especially important.

SIGNIFICANCE

Our results highlight the importance of information contained in high-frequency waveforms in the neurological intensive care unit. They could motivate future studies on pre-hypertensive patterns and the design of new alarm algorithms for critical events in the injured brain.

Discrepancy Between Internal and External Intracranial Pressure Transducers: Quantification of an Old Source of Error in EVDs?

BACKGROUND

Intracranial pressure monitoring remains the foundation for prevention of secondary injury after traumatic brain injury and is most commonly performed using an external ventricular drain or intraparenchymal pressure monitor. The Integra Flex ventricular catheter combines an external ventricular catheter with a pressure transducer embedded in the tip of the catheter to allow continuous pressure readings while simultaneously draining cerebrospinal fluid. Discrepancies between measurements from the continuously reported internal pressure transducer and intermittently assessed and externally transduced ventricular drain prompted an analysis and characterization of pressures transduced from the same ventricular source.

METHODS

More than 500 hours of high-resolution (125 Hz) continuous recordings were manually reviewed to identify 73 hours of simultaneous measurements (clamped external ventricular drain) from internal and external transducers in patients with traumatic brain injury.

RESULTS

A significant positive bias was found in pressure readings obtained from external relative to internal measurements. The 2 methods of measurement generally correlated poorly with each other and variably. Although proportional bias was found with Bland-Altman analysis, coherence revealed rare shifts in the external transducer as a major source of discrepancy. Infrequent changes in the 0-level of the external transducer were found to be the primary source of discrepancy. Relative to the observed differences, no significant trend was observed over time between the 2 modalities.

CONCLUSIONS

This study suggests that the internal pressure transducer may be a more reliable estimate of intracranial pressure relative to bedside external transducers due to the inherent behavioral requirement of leveling.

Non-Invasive Blood Pressure Estimation from ECG Using Machine Learning Techniques

BACKGROUND

Blood pressure (BP) measurements have been used widely in clinical and private environments. Recently, the use of ECG monitors has proliferated; however, they are not enabled with BP estimation. We have developed a method for BP estimation using only electrocardiogram (ECG) signals.

METHODS

Raw ECG data are filtered and segmented, and, following this, a complexity analysis is performed for feature extraction. Then, a machine-learning method is applied, combining a stacking-based classification module and a regression module for building systolic BP (SBP), diastolic BP (DBP), and mean arterial pressure (MAP) predictive models. In addition, the method allows a probability distribution-based calibration to adapt the models to a particular user.

RESULTS

Using ECG recordings from 51 different subjects, 3129 30-s ECG segments are constructed, and seven features are extracted. Using a train-validation-test evaluation, the method achieves a mean absolute error (MAE) of 8.64 mmHg for SBP, 18.20 mmHg for DBP, and 13.52 mmHg for the MAP prediction. When models are calibrated, the MAE decreases to 7.72 mmHg for SBP, 9.45 mmHg for DBP and 8.13 mmHg for MAP.

CONCLUSION

The experimental results indicate that, when a probability distribution-based calibration is used, the proposed method can achieve results close to those of a certified medical device for BP estimation.

A narrative review of the clinical application of pressure reactiviy indices in the neurocritical care unit

Pressure reactivity indices are used in clinical research as a surrogate marker of the ability of the cerebrovasculature to maintain cerebral autoregulation. The use of pressure reactivity indices in patients with neurological injury represents a potential to move away from population-based physiological targets used in guidelines to individualized physiological targets. The aim of this review is to describe the underlying principles and development of pressure reactivity indices, alongside a critique of how they have been used in clinical research, including their limitations. The primary source literature was identified from a database search of PUBMed and OVID online using the search terms "pressure reactivity index" and "pressure reactivity indices". The evidence base regarding pressure reactivity indices currently remains Level III. Pressure reactivity indices rely on the correlation (-1 to +1) between the arterial blood pressure and intracranial pressure, with negative values indicating intact cerebral autoregulation and positive values indicating dysfunctional cerebral autoregulation. Meaningful data is taken from summary measures and trends. The traumatic brain injury population feature most prominently in the literature. There is limited description of the potential confounding factors that may affect pressure reactivity indices, including physiological parameters and therapeutic interventions. Plotting a pressure reactivity index against a cerebral perfusion pressure can indicate an optimal cerebral perfusion pressure to individualise patient care. There is potential to over interpret optimal cerebral perfusion pressure targets when the values of pressure reactivity indices are close to zero. There is an association between pressure reactivity indices and neurological outcomes, however the use of pressure reactivity indices as a prognostication tool is to be challenged. Average values of cerebral perfusion pressure that are not close to averaged values of optimal cerebral perfusion pressure are also associated with poor outcome. Further research is required to ascertain whether targeting an optimal cerebral perfusion pressure may alter outcome.

Assessing Cerebral Hemodynamic Stability After Brain Injury

OBJECTIVE

Following brain injury, unstable cerebral hemodynamics can be characterized by abnormal rises in intracranial pressure (ICP). This behavior has been quantified by the RAP index: the correlation (R) between ICP pulse amplitude (A) and mean (P). While RAP could be a valuable indicator of autoregulatory processes, its prognostic ability is not well established and its validity has been questioned due to potential errors in measurement. Here, we test (1) whether RAP is a consistent measure of intracranial hemodynamics and (2) whether RAP has prognostic value in predicting hemodynamic instability following brain injury.

MATERIALS AND METHODS

RAP was tested in seven brain injured patients treated in a surgical intensive care unit. A sample of ICP data was randomly chosen and segmented into 1 hour periods. Hours were then categorized as either stable, which contained no sharp rises in ICP, or unstable, which contained ≥1 sharp rise-where a sharp rise is defined as ICP exceeding a mean slope of 0.15 mmHg/s. Equal numbers of stable and unstable segments were then selected for each patient. RAP was calculated as the Pearson's correlation coefficient between ICP pulse amplitude (AMP) and mean (mICP), determined in 6 second windows, according to established methods.

RESULTS

Results showed that (1) average AMP and ICP levels were similar between stable and unstable periods and (2) unstable periods were identified by RAP values exceeding 0.6 with an average positive predictive value of 74%.

CONCLUSIONS

We conclude that RAP can provide a valid measure of ICP dynamics, is not affected by sensor drift, and can better distinguish periods of instability than ICP or AMP alone.

Journal of clinical monitoring and computing 2016 end of year summary: monitoring cerebral oxygenation and autoregulation

In the perioperative and critical care setting, monitoring of cerebral oxygenation (ScO₂) and cerebral autoregulation enjoy increasing popularity in recent years, particularly in patients undergoing cardiac surgery. Monitoring ScO₂ is based on near infrared spectroscopy, and attempts to early detect cerebral hypoperfusion and thereby prevent cerebral dysfunction and postoperative neurologic complications. Autoregulation of cerebral blood flow provides a steady flow of blood towards the brain despite variations in mean arterial blood pressure (MAP) and cerebral perfusion pressure, and is effective in a MAP range between approximately 50–150 mmHg. This range of intact autoregulation may, however, vary considerably between individuals, and shifts to higher thresholds have been observed in elderly and hypertensive patients. As a consequence, intraoperative hypotension will be poorly tolerated, and might cause ischemic events and postoperative neurological complications. This article summarizes research investigating technologies for the assessment of ScO₂ and cerebral autoregulation published in the Journal of Clinical Monitoring and Computing in 2016.