Shaun M. Davidson

Clinical Utilisation of Respiratory Elastance (CURE): Pilot Trials for the Optimisation of Mechanical Ventilation Settings for the Critically Ill

Abstract Current practice in determining Mechanical Ventilation (MV) settings is highly variable with little consensus, forcing clinicians to rely on general approaches and clinical intuition. The Clinical Utilisation of Respiratory Elastance (CURE) system was developed to aid clinical determination of important MV settings by providing real-time patient-specific lung condition information at the patient bedside. The pilot clinical trials to investigate the performance and efficacy of this system are currently being carried out in the Christchurch Hospital ICU, New Zealand. This paper presents the CURE clinical trial protocol and its initial findings from the two patients recruited to date. In particular, this paper focuses on CUREs ability to determine patient-specific responses in real time to PEEP changes and recruitment manoeuvres (RM). The results from this study demonstrate the potential for CURE Soft to improve the reliability and ease with which clinicians make decisions about MV settings in the ICU.

Real-Time Breath-to-Breath Asynchrony Event Detection using Time-Varying Respiratory Elastance Model

Abstract Asynchronous events (AE) occur during mechanical ventilation (MV) therapy when the patients breathing is not synchronised with the ventilator support. Frequent AE indicates sub-optimal ventilation therapy and may lead to further complications. Asynchrony Index (AI) gives the percentage of AEs as a percentage of total breaths, but is only assessed via manual scrutiny. Thus, there is a need to automate AE detection in real-time. A model-based approach using time-varying elastance to detect AEs is developed and retrospectively assessed in MV patients. Data from 14 mechanically ventilated respiratory failure patients, enrolled in an observational study in Christchurch Hospital, New Zealand were used to investigate the performance of the method. Patient data is sorted according to the ventilation mode used, and AI is calculated for each episode separately. The model-based approach accurately identifies AEs, and shown not to give false positive readings when compared to manual detection (gold standard). None of the ventilation modes give significantly different AI levels (P > 0.05). AI decreases when ventilation mode changes and increases overall time indicate worsen patient-ventilator interaction. The model-based method is able to successfully and accurately calculate AI. Real time use of this metric will enable patients with sub-optimal ventilator settings to be automatically identified for the first time and the settings adjusted as necessary, improving the efficacy of mechanical ventilation therapy, and providing a quantified metric to help guide MV care.

The dimensional reduction method for identification of parameters that trade-off due to similar model roles

Parameter identification is an important and widely used process across the field of biomedical engineering. However, it is susceptible to a number of potential difficulties, such as parameter trade-off, causing premature convergence at non-optimal parameter values. The proposed Dimensional Reduction Method (DRM) addresses this issue by iteratively reducing the dimension of hyperplanes where trade off occurs, and running subsequent identification processes within these hyperplanes. The DRM was validated using clinical data to optimize 4 parameters of the widely used Bergman Minimal Model of glucose and insulin kinetics, as well as in-silico data to optimize 5 parameters of the Pulmonary Recruitment (PR) Model. Results were compared with the popular Levenberg-Marquardt (LMQ) Algorithm using a Monte-Carlo methodology, with both methods afforded equivalent computational resources. The DRM converged to a lower or equal residual value in all tests run using the Bergman Minimal Model and actual patient data. For the PR model, the DRM attained significantly lower overall median parameter error values and lower residuals in the vast majority of tests. This shows the DRM has potential to provide better resolution of optimum parameter values for the variety of biomedical models in which significant levels of parameter trade-off occur.

Minimally invasive estimation of ventricular dead space volume through use of Frank-Starling curves

This paper develops a means of more easily and less invasively estimating ventricular dead space volume (Vd), an important, but difficult to measure physiological parameter. Vd represents a subject and condition dependent portion of measured ventricular volume that is not actively participating in ventricular function. It is employed in models based on the time varying elastance concept, which see widespread use in haemodynamic studies, and may have direct diagnostic use. The proposed method involves linear extrapolation of a Frank-Starling curve (stroke volume vs end-diastolic volume) and its end-systolic equivalent (stroke volume vs end-systolic volume), developed across normal clinical procedures such as recruitment manoeuvres, to their point of intersection with the y-axis (where stroke volume is 0) to determine Vd. To demonstrate the broad applicability of the method, it was validated across a cohort of six sedated and anaesthetised male Pietrain pigs, encompassing a variety of cardiac states from healthy baseline behaviour to circulatory failure due to septic shock induced by endotoxin infusion. Linear extrapolation of the curves was supported by strong linear correlation coefficients of R = 0.78 and R = 0.80 average for pre- and post- endotoxin infusion respectively, as well as good agreement between the two linearly extrapolated y-intercepts (Vd) for each subject (no more than 7.8% variation). Method validity was further supported by the physiologically reasonable Vd values produced, equivalent to 44.3–53.1% and 49.3–82.6% of baseline end-systolic volume before and after endotoxin infusion respectively. This method has the potential to allow Vd to be estimated without a particularly demanding, specialised protocol in an experimental environment. Further, due to the common use of both mechanical ventilation and recruitment manoeuvres in intensive care, this method, subject to the availability of multi-beat echocardiography, has the potential to allow for estimation of Vd in a clinical environment.

Modelling of the nonlinear end-systolic pressure-volume relation and volume-at-zero-pressure in porcine experiments

The End-Systolic Pressure-Volume Relation (ESPVR) is generally modelled as a linear relationship between P and V as cardiac reflexes, such as the baroreflex, are typically suppressed in experiments. However, ESPVR has been observed to behave in a curvilinear fashion when cardiac reflexes are not suppressed, suggesting the curvilinear function may be more clinically appropriate. Data was gathered from 41 vena cava occlusion manoeuvres performed experimentally at a variety of PEEPs across 6 porcine specimens, and ESPVR determined for each pig. An exponential model of ESPVR was found to provide a higher correlation coefficient than a linear model in 6 out of 7 cases, and a lower Akaike Information Criterion (AIC) value in all cases. Further, the exponential ESPVR provided positive V0 values in a physiological range in 6 out of 7 cases analysed, while the linear ESPVR produced positive V0 values in only 3 out of 7 cases, suggesting linear extrapolation of ESPVR to determine V0 may be flawed.

Pre-ejection period, the reason why the electrocardiogram Q-wave is an unreliable indicator of pulse wave initialization

OBJECTIVE Pulse wave velocity measurements are an indicator of arterial stiffness and possible cardiovascular dysfunction. It is usually calculated by measuring the pulse transit time (PTT) over a known distance through the arteries. In animal studies, reliable PTT measures can be obtained using two pressure catheters. However, such direct, invasive methods are undesirable in clinical settings. A less invasive alternative measure of PTT is pulse arrival time (PAT), the time between the Q-wave of an electrocardiogram (ECG) and the arrival of the foot of the beats pressure waveform at one pressure catheter. Since the Q-wave signifies the start of ventricular contraction, PAT includes the pre-ejection period (PEP), a time where no blood is ejected. Thus, inter- or intra- subject variation in PEP could result in poor correlation between pulse arrival time (PAT) and the desired pulse transit time (PTT). APPROACH This study looks at the relationship between PAT and PTT, over a range of common critical care therapies and determines the effect of PEP on PAT as a possible surrogate of PTT in a critical care environment. The analysis uses data from five porcine experiments, where ECG, aortic arch and abdominal aortic pressure were measured simultaneously, over a range of induced hemodynamic conditions. RESULTS The resulting correlations of PAT verse PTT varied within pigs and across interventions (r 2  =  0.32-0.69), and across pigs (r 2  =  0.05-0.60). Variability was due to three main causes. First, the interventions themselves effect PEP and PTT differently, second, pig specific response to the interventions, and third, inter- and intra- pig variability in PEP, independent of PTT. SIGNIFICANCE The overall analysis shows PAT is an unreliable measure of PTT and a poor surrogate under clinical interventions common in a critical care setting, due to intra- and inter- subject variability in PEP.

Beat-by-Beat Estimation of the Left Ventricular Pressure–Volume Loop Under Clinical Conditions

This paper develops a method for the minimally invasive, beat-by-beat estimation of the left ventricular pressure–volume loop. This method estimates the left ventricular pressure and volume waveforms that make up the pressure–volume loop using clinically available inputs supported by a short, baseline echocardiography reading. Validation was performed across 142,169 heartbeats of data from 11 Piétrain pigs subject to two distinct protocols encompassing sepsis, dobutamine administration and clinical interventions. The method effectively located pressure–volume loops, with low overall median errors in end-diastolic volume of 8.6%, end-systolic volume of 17.3%, systolic pressure of 19.4% and diastolic pressure of 6.5%. The method further demonstrated a low overall mean error of 23.2% predicting resulting stroke work, and high correlation coefficients along with a high percentage of trend compass ‘in band’ performance tracking changes in stroke work as patient condition varied. This set of results forms a body of evidence for the potential clinical utility of the method. While further validation in humans is required, the method has the potential to aid in clinical decision making across a range of clinical interventions and disease state disturbances by providing real-time, beat-to-beat, patient specific information at the intensive care unit bedside without requiring additional invasive instrumentation.

A Minimally Invasive Method for Beat-by-Beat Estimation of Cardiac Pressure-Volume Loops.

This paper develops a minimally invasive means of estimating a patient-specific cardiac pressure-volume loop beat-to-beat. This method involves estimating the left ventricular pressure and volume waveforms using clinically available information including heart rate and aortic pressure, supported by a baseline echocardiography reading. Validation of the method was performed across an experimental data set spanning 5 Piétrain pigs, 46,318 heartbeats and a diverse clinical protocol. The method was able to accurately locate a pressure-volume loop, identifying the end-diastolic volume, end-systolic volume, mean-diastolic pressure and mean-systolic pressure of the ventricle with reasonable accuracy. While there were larger percentage errors associated with stroke work derived from the estimated pressure-volume loops, there was a strong correlation (average R value of 0.83) between the estimated and measured stroke work values. These results provide support for the potential of the method to track patient condition, in real-time, in a clinical environment. This method has the potential to yield additional information from readily available waveforms to aid in clinical

A comparison between four techniques to measure cardiac output

Cardiac output is an important variable when monitoring hemodynamic status. In particular, changes in cardiac output represent the goal of several circulatory management therapies. Unfortunately, cardiac output is very difficult to estimate, either in experimental or clinical settings. The goal of this work is to compare four techniques to measure cardiac output: pressure-volume catheter, aortic flow probe, thermodilution, and the PiCCO monitor. These four techniques were simultaneously used during experiments of fluid and endotoxin administration on 7 pigs. Findings show that, first, each individual technique is precise, with a relative coefficient of repeatability lower than 7 %. Second, 1 cardiac output estimate provided by any technique relates poorly to the estimates from the other 3, even if there is only small bias between the techniques. Third, changes in cardiac output detected by one technique are only detected by the others in 62 to 100 % of cases. This study confirms the difficulty of obtaining a reliable clinical cardiac output measurement. Therefore, several measurements using different techniques should be performed, if possible, and all such should be treated with caution.

Stroke Volume Estimation using Aortic Pressure Measurements and Aortic Cross Sectional Area: Proof of Concept

Accurate Stroke Volume (SV) monitoring is essential for patient with cardiovascular dysfunction patients. However, direct SV measurements are not clinically feasible due to the highly invasive nature of measurement devices. Current devices for indirect monitoring of SV are shown to be inaccurate during sudden hemodynamic changes. This paper presents a novel SV estimation using readily available aortic pressure measurements and aortic cross sectional area, using data from a porcine experiment where medical interventions such as fluid replacement, dobutamine infusions, and recruitment maneuvers induced SV changes in a pig with circulatory shock. Measurement of left ventricular volume, proximal aortic pressure, and descending aortic pressure waveforms were made simultaneously during the experiment. From measured data, proximal aortic pressure was separated into reservoir and excess pressures. Beat-to-beat aortic characteristic impedance values were calculated using both aortic pressure measurements and an estimate of the aortic cross sectional area. SV was estimated using the calculated aortic characteristic impedance and excess component of the proximal aorta. The median difference between directly measured SV and estimated SV was -1.4ml with 95% limit of agreement +/- 6.6ml. This method demonstrates that SV can be accurately captured beat-to-beat during sudden changes in hemodynamic state. This novel SV estimation could enable improved cardiac and circulatory treatment in the critical care environment by titrating treatment to the effect on SV.