Partial CO2 rebreathing indirect Fick technique for non-invasive measurement of cardiac output.

OBJECTIVE: Evaluation in animals of a non-invasive and continuous cardiac output monitoring system based on partial carbon-dioxide (CO2) rebreathing indirect Fick technique. METHODS: We have developed a non-invasive cardiac output (NICO) monitoring system, based on the partial rebreathing method. The partial rebreathing technique employs a differential form of the Fick equation for calculating cardiac output (QT) using non-invasive measurements. Changes in CO2 elimination (deltaVCO2) and partial pressure of end-tidal CO2 (deltaPETCO2) in response to a brief period of partial rebreathing are used to measure pulmonary capillary blood flow (Q(PCBF)). A non-invasive estimate of anatomic and intrapulmonary shunt fraction (Q(S)/Q(T)), based on oxygen saturation from pulse oximetry (SpO2) and inspired oxygen concentration (FIO2), is added to compute total cardiac output [Q(T) = Q(PCBF)/(1 - Q(S)/Q(T))]. The performance of the NICO was compared with iced 5% dextrose bolus thermodilution cardiac output (TDco) measurements in 6 dogs. Cardiac output was varied using dobutamine, and halothane, and by clamping of the inferior vena cava. Two hundred and forty-six (n = 246) paired measurements of TDco and NICO over a range of cardiac outputs (TDco range = 0.60-8.87 l/min) were compared using Bland-Altman analysis and weighted correlation coefficient. RESULTS: The Bland-Altman technique yielded a NICO precision of +/- 0.70 l/min (13.8%) with a mean bias of -0.07 l/min (-1.4%) compared to TDco. The weighted correlation coefficient between TDco and NICO values was: r = 0.93 (n = 246). CONCLUSION: The partial CO2 rebreathing technique for measurement of cardiac output is non-invasive, automated, and based on the well accepted Fick principle. The limits of agreement between NICO and TDco is within the recommended value for NICO to be a clinically acceptable method for cardiac output measurement. The results of this canine study show that NICO performed as well, and in some cases better, than other currently available non-invasive cardiac output techniques over a wide range of cardiac outputs.

Evaluation of a thoracic bioimpedance cardiac output monitor during cardiac catheterization.

OBJECTIVE: To evaluate the accuracy and precision of an advanced thoracic bioimpedance cardiac output monitor by comparing it with conventional thermodilution. DESIGN: Prospective data collected from 47 patients undergoing routine cardiac catheterization. The new bioimpedance system differs from its predecessors in electrode system configuration, advanced signal processing, use of a modified Kubicek equation, and a reliable estimate of left ventricular ejection time from the time derivative bioimpedance signals. SETTING: A cardiac catheterization laboratory in a university affiliated teaching hospital. PATIENTS: A series of 47 relatively homogenous patients undergoing routine cardiac catheterization for suspected cardiac disease. MEASUREMENTS AND MAIN RESULTS: The data from the first 20 patients was used to determine optimal values for coefficients in the bioimpedance cardiac output equations. The coefficients found were used when the system was tested in the subsequent 27 patients. For the last 27 patients, a total of 80 simultaneous pairs of cardiac output measurements were made by conventional thermodilution and by thoracic bioimpedance. The mean difference between the two methods was -0.31 L/min and the standard deviation of the differences was (0.76 L/min). The correlation coefficient was r2 = .72 (p < .001). CONCLUSIONS: The correlation between conventional thermodilution and thoracic bioimpedance cardiac output estimates was good and the standard deviation of the differences was lower than that reported for commercially available devices. The system can be used in the cardiac catheterization lab for reliable and continuous noninvasive measurement of cardiac output.

Evaluation of a new advanced thoracic bioimpedance device for estimation of cardiac output.

OBJECTIVE: This study is an evaluation of a new thoracic bioimpedance cardiac output monitoring system which incorporates a modified form of the Kubicek equation and a method of estimating the left ventricular ejection time from the time derivative bioimpedance signals. METHODS: The performance of the new system was compared with conventional thermodilution in a porcine model. One hundred and ninety nine (n = 199) paired measurements of thermodilution cardiac output (TDCO) (range 1.20-18.00 L/min) and thoracic bioimpedance cardiac output (BICO) were collected in 7 pigs. The bioimpedance measurements were adjusted for the animal's weight and chest circumference, thus compensating for the differences in the anatomy of pigs when compared to humans. Data were compared using weighted correlation coefficient and Bland-Altman analysis. RESULTS: The weighted correlation coefficient between TDCO and BICO values was 0.87 (n = 199). The Bland-Altman technique yielded a precision of the device of +/-1.69 L/min with a bias of 0.11 L/min. CONCLUSION: The results from the porcine study show that the new system performed well over a wide range of cardiac outputs, comparing favorably with data from other new bioimpedance cardiac output devices currently available in the market.

Evaluation in volunteers of the VIA V-ABG automated bedside blood gas, chemistry, and hematocrit monitor.

OBJECTIVE: To evaluate the VIA V-ABG (VIA Medical Corp.) point-of-care blood gas and chemistry monitor in healthy human volunteers, with particular emphasis on the measurement of blood gases. METHODS: Experimental conditions were varied by intermittently subjecting volunteers to either isocapnic hypercapnia (end-tidal (ET), PETCO2 = 50+/-2 mmHg, ETPO2 = 130+/-5 mmHg) or isocapnic hypoxia (PETCO2 = 42+/-2, PETO2 + 45+/-2 mmHg) in addition to room air breathing. Measurements by the VIA V-ABG device were compared with paired samples and measurements performed by two ABL Radiometers (505 and 500). Analysis of results includes bias and precision plots and comparison of results with minimal performance criteria as established by CLIA. RESULTS: Nineteen volunteers yielded 222 matched samples. The range of values were 7.32-7.61 for pH, 20.9-51.6 mmHg for PCO2, 27.9-184.5 mmHg for PO2, 134-141 mmol/l for Na, 3.1-4.1 mmol/l for K, and 30.0-50.4% for hematocrit. Bias and precision (+/-2 sd) for pH was 0.01 and 0.04, for PCO2 was 0.4 and 4.8, for PO2 was 1.0 and 17.0, for Na was -0.3 and 5.2, for K was 0.1 and 0.2, and for Hct was 2.0 and 5.4. CONCLUSIONS: Over the range of blood gas values assessed, blood gas measurements by the VIA V-ABG device were clinically acceptable and met minimal performance criteria utilizing current Medicare CLIA proficiency standards. Performance criteria were also met by the VIA V-ABG device for Na, K, and Hct measurements but the range of values was too narrow to allow characterization of clinical acceptability. The VIA V-ABG device appears to perform well compared with the results which have been published for other point-of-care devices. Comparison between different studies investigating point-of-care devices is difficult due to several factors (range of values measured, comparison device, population studied, etc.). Some of these instruments, including the VIA V-ABG device, may serve quite well as point-of-care devices to perform certain tests at the bedside. Whether or not any of these devices can substitute for traditional laboratory blood gas and chemistry measurements remains an issue that is not adequately studied.

Fundamentals of feedback control: PID, fuzzy logic, and neural networks.

Feedback controllers have been shown to bring blood pressure, muscle relaxation, inhalation drug concentration, and ventilation to the target level and keep it at the target as quickly and as accurately as can a well-trained clinician. Feedback control is an effective and convenient clinical tool for optimizing the day-to-day delivery of anesthetics, reducing induction time, delivering a minimum amount of drug, and avoiding costly delays from failing to keep the patient in a desired state.

Fuzzy logic for model adaptation of a pharmacokinetic-based closed loop delivery system for pancuronium.

In this paper, we investigate the ability of fuzzy to adapt the parameters of a pharmacokinetic and pharmacodynamic model-based controller for the delivery of the muscle relaxant pancuronium. The system uses the model to control the rate of drug delivery and uses feedback from a sensor which measures muscle relaxation level to adapt the model using fuzzy logic. The control strategy administers mini-bolus doses of pancuronium and modulates the magnitude and time interval between the bolus doses to maintain a patient's muscle relaxation within an allowable range specified by the user. Before each new dose is given, the fuzzy logic adaptation scheme uses the error between the predicted patient response and the measured response to adapt the model. The system was tested using computer simulation by varying the parameters of the model by 50% from their nominal values. It was also evaluated in a clinical trial of five patients undergoing surgical procedures lasting 5 h or longer.

Clinical analysis of the flexor hallucis brevis as an alternative site for monitoring neuromuscular block from mivacurium.

STUDY OBJECTIVES: To compare the flexor hallucis brevis, which is responsible for flexion of the great toe, to the adductor pollicis as a site for monitoring the onset and recovery from neuromuscular block after an intubating dose of mivacurium chloride. DESIGN: Prospective patient-controlled study. SETTING: University teaching hospital. PATIENTS: 10 ASA physical status I and II adults (age 18 to 55 years, 6 women, 4 men) scheduled for elective procedures requiring muscle relaxation for tracheal intubation. MEASUREMENTS AND MAIN RESULTS: Patients were monitored at the adductor pollicis and the flexor hallucis brevis during the onset and recovery of neuromuscular block, which was administered to facilitate tracheal intubation. All subjects were given mivacurium 0.2 mg/kg over 30 seconds. Their train-of-four (TOF) response was continually monitored at both sites until the patient recovered from the intubating dose to a TOF ratio of 0.75. The time to onset of neuromuscular block, recovery of the first TOF response, and recovery to a TOF ratio of 0.75 were compared between the two monitoring sites using the Wilcoxon signed rank test. Following administration of the intubating dose of mivacurium, the loss of all twitch response occurred 1.2 minutes sooner at the adductor pollicis than at the flexor hallucis brevis (p < 0.02). Reappearance of the first twitch occurred 0.49 minutes slower at the adductor pollicis, although this difference was not statistically significant. The time to recovery to a TOF ratio of 0.75 at the adductor pollicis was slower by 2.83 minutes (p = 0.046). CONCLUSIONS: Due to its lag behind the adductor pollicis, the flexor hallucis brevis is not a good indicator of when to intubate the trachea during the onset of neuromuscular block; however, its faster recovery may make it useful for monitoring deep neuromuscular block intraoperatively or during recovery when the adductor pollicis TOF response still shows complete blockade.

An integrated graphic data display improves detection and identification of critical events during anesthesia.

OBJECTIVE: To show that an integrated graphic data display can shorten the time taken to detect and correctly identify critical events during anesthesia. METHODS: We developed a graphic display which presents 30 anesthesia-related physiologic variables as shapes and colors, rather than traditional digits and waveforms. To evaluate the new display, we produced four critical events on a computer-based anesthesia simulator and asked two groups of five anesthesiologists to identify the events as quickly as possible. One group observed the new display while the other group viewed a traditional cardiovascular monitor with digital and waveform displays. RESULTS: The group which observed the integrated graphic display saw changes caused by inadequate paralysis 2.4 min sooner, and changes caused by a cuff leak 3.1 min sooner than those observing the traditional display. The integrated display group correctly identified the reason for the change 2.8 min sooner for inadequate paralysis, 3.1 min sooner for cuff leak and 3.1 min sooner for bleeding. These differences were all statistically significant. CONCLUSIONS: The results show that some simulated critical events are detected and correctly identified sooner, when an anesthesiologist views an integrated graphic display, rather than a traditional digital/waveform monitor.

Theoretical analysis of non-invasive oscillometric maximum amplitude algorithm for estimating mean blood pressure.

A theoretical analysis is performed to evaluate the effect of arterial mechanical and blood pressure pulse properties on the accuracy of non-invasive oscillometric maximum amplitude algorithm (MAA) estimates of the mean blood pressure obtained using air-filled occlusive cuffs. Invasively recorded blood pressure pulses, selected for their varied shapes, are scaled to simulate a wide range of blood pulse pressures (diastolic blood pressure minus systolic blood pressure). Each scaled blood pressure pulse is transformed through an exponential model of an artery to create a series of blood volume pulses from which a simulated oscillometric waveform is created and the corresponding MAA estimate of the mean blood pressure and error (mean blood pressure minus MAA estimate) are determined. The MAA estimates are found to depend on the arterial blood pressure. The errors are found to depend on the arterial mechanical properties, blood pressure pulse shape and blood pulse pressure. These results suggest that there is no direct relationship between the mean blood pressure and MAA estimate, and that multiple variables may affect the accuracy of MAA estimates of the mean blood pressure obtained using air-filled occlusive cuffs.

Pharmacokinetic-based minibolus delivery as an alternative to continuous infusion for drugs that exhibit a biophase lag.

The presence of a biophase compartment in a pharmacokinetic model indicates that the response to an administered dose of drug is damped such that the time to peak effect occurs after the peak concentration in the bloodstream. This phenomenon, which is common to most intravenous anesthetic agents, can be exploited by a drug delivery method that administers minibolus doses of drug rather than a continuous infusion. Through analysis of the frequency response behavior of the biophase compartment, a bolus magnitude and dose frequency or interval (1/frequency) can be chosen such that the oscillation in drug effect is minimized even though the plasma concentration may be changing significantly with each supplemental dose. A pharmacokinetic and pharmacodynamic based method for calculating the bolus dose size and dosing interval is presented. The trade-off between dose interval and change in drug effect is exemplified through computer simulation of this strategy applied to delivery of the neuromuscular blocking agent pancuronium. The method provides a repetitive perturbation to the pharmacokinetic and pharmacodynamic system that can aid in model parameter identification during closed loop applications.

A preliminary laboratory investigation of air embolus detection and grading using an artificial neural network.

SUMMARY STATEMENT: Processed digitized Doppler signals abstracted from recordings during continuous air infusion in dogs were used to train a neural network to estimate air embolism infusion rates. BACKGROUND: Precordial Doppler is a sensitive technique for detecting venous air embolism during anesthesia, but it requires constant attentive listening. Since neural networks are particularly well suited to the task of pattern recognition, we sought to investigate this technology for detection and grading of air embolism. METHODS: Air was infused into peripheral veins of four anesthetized dogs at rates of 0.025, 0.05, 0.10, 0.25, 0.50 and 1.0 ml-1.kg-1.min-1 while digital recordings of the precordial Doppler ultrasound signal were collected. The frequency content of the recordings was determined by Fourier analysis. The output of the Fourier transform was the input to a neural network. The network was then trained to estimate the air infusion rate. RESULTS: The correlation coefficient between the size of the air embolism and the air infusion rate was greater than r2 = 0.93 for each of the four animals in the study when the network was trained using the data for all four dogs. When the data from a dog was withheld from the training set and used only for testing the correlation coefficients ranged from r2 = 0.75 to r2 = 0.27. For frequencies below 250 Hz, the acoustic energy tended to fall as the air infusion rate increased. The opposite occurred at frequencies above 325 Hz. CONCLUSIONS: Neural network processing of the precordial Doppler signal provides a quantitative estimate of the size of an air embolism.

Supraorbital artery as an alternative site for oscillometric blood pressure measurement.

OBJECTIVE: Noninvasive blood pressure measured from the superficial temporal artery has been shown to correlate well with pressure in the brachial artery. The supraorbital artery may be an even better site for monitoring blood pressure on the forehead because it originates from the internal carotid artery, and it is easier to locate anatomically. This study compares mean pressure measured oscillometrically over the supraorbital artery and at the upper arm. METHODS: Oscillometric signals from the supraorbital artery were recorded in 20 surgical patients under general anesthesia using a 2.5- x 1-cm bladder attached to the forehead with a self-adhesive pad. Blood pressure was measured simultaneously from the arm using a Dinamap 1846 blood pressure monitor, and the resulting data compared with the supraorbital artery measurements. RESULTS: The mean difference between 219 pairs of blood pressure measurements, from the forehead and the arm, was 3.8 mm Hg. The standard deviation of the differences was 7.4 mm Hg. The linear regression equation for the data was y = 0.98x + 3.25, with a standard error of estimate of 7.31 mm Hg. The correlation coefficient between the two measurements was 0.82. CONCLUSIONS: The results show that mean blood pressures measured oscillometrically from the supraorbital and brachial arteries agree and correlate well with each other. The supraorbital artery should be a good alternative site for blood pressure measurement.

Optimal sites for forehead oscillometric blood pressure monitoring.

OBJECTIVE: Blood pressure is usually measured noninvasively with a cuff on the arm of the leg. Circumstances exist, however, when an alternative site for blood pressure measurement is desirable. This study is designed to identify a location on the forehead where blood pressure can be reliably measured noninvasively. METHODS: We mapped the superficial temporal artery and/or the supraorbital artery in 65 volunteers and found a rectangular area where an adhesive pressure pad could be placed over each artery. Oscillometric signals were recorded from four different locations over the forehead in 19 of the 65 volunteers to compare the amplitude of the signal and mean blood pressure between locations. RESULTS: The course of the supraorbital artery is quite consistent. It passed through a 2.5- x 1-cm rectangular area on the forehead in all volunteers in which it was mapped. The medial border of the rectangle is 0.5-cm medial and 1-cm above the medial corner of the left eyebrow. The course of the superficial temporal artery differed remarkably from person to person. We could not find an area of reasonable size to cover the artery in all wounds. Mean blood pressures were the same in all forehead locations. The signal was the weakest on the center of the forehead and strongest directly over the superficial temporal artery. CONCLUSIONS: Our results show that the supraorbital artery, an end-artery of the internal carotid artery, which emerges through the supraorbital foramen and cross the forehead near the center, is the preferred site to monitor blood pressure noninvasively on the forehead with an adhesive pressure bladder.

Computers in critical care.

This article reviews the current state-of-the-art and future applications of computers in critical care, with particular attention to ventilator and drug-delivery applications. Automated charting, alerts and alarms, and tools for decision support (such as expert systems and closed-loop control) are discussed also.

A bolus plus continuous infusion protocol for controlling neuromuscular blockade during anesthesia.

Neuromuscular blockade is controlled during anesthesia by administering either bolus doses or a continuous infusion of a blocking agent. To test whether a constant infusion technique requires less attention and provides better control we used a computer to simulate neuromuscular blockade. Using the model we maintained 95% blockade with mivacurium, atracurium, and vecuronium. It required 1.2 changes per hour to maintain the blockade by continuous infusion; an average of 4.5 bolus per hour were required to maintain blockade by the bolus technique. When the bolus and continuous infusion techniques were combined, only 0.16 changes per hour were required. Atracurium was then given to ten patients during anesthesia, following the bolus plus continuous infusion protocol. After a bolus was given to obtain 100% twitch depression, for tracheal intubation, neuromuscular function was assessed by train-of-four stimulation of the ulnar or facial nerves by observing the resultant muscle movement. When the first twitch of the train-of-four returned, relaxation was maintained by continuous infusion. A bolus was given and the drug infusion rate was changed whenever the level of relaxation changed from the desired one twitch of the train-of-four. The infusion rate was adjusted only 1.12 +/- 0.79 times per hour. The desired level of muscle relaxation was easily controlled using the bolus plus continuous infusion protocol. The infusion scheme might be implemented in future drug infusion pumps.

A comparison of dynamic and isometric force sensors for train-of-four measurement using submaximal stimulation current.

OBJECTIVE: We studied the accuracy and repeatability of train-of-four (TOF) ratio measurements made from a dynamic piezoelectric sensor that records movement of the thumb in response to ulnar nerve stimulation compared with an isometric mechanomyogram that measures force of contraction of the adductor pollicis. METHODS: The study involved 10 patients whose level of neuromuscular block was held constant with an intravenous (IV) infusion of vecuronium bromide (0.4 to 1.0 micrograms/kg/min) (Organon, West Orange, NJ). The sensors were attached to opposite arms of each patient and simultaneous measurements of TOF ratio were taken at stimulation current levels of 50, 30, and 20 mA. RESULTS: In comparison to the TOF ratio measured at the maximal stimulation current (50 mA), the TOF ratio from the piezo sensor showed a bias and standard deviation of -0.13 +/- 0.24 when the stimulation current was reduced to 30 mA. At 20 mA, the bias and standard deviation was -0.24 +/- 0.28. The TOF ratio from the mechanomyogram showed a bias and standard deviation of 0.01 +/- 0.07 at 30 mA and 0.0 +/- 0.20 at 20 mA when compared with measurements made when the stimulation current was 50 mA. CONCLUSIONS: Both sensors showed diminished repeatability in TOF measurement with decreasing stimulation current. The data indicate that neither sensor is reliable for general monitoring of neuromuscular block at submaximal current levels. However, the individual patient results showed that some patients could be monitored accurately with both sensors, even at the lowest stimulation current levels.

Intelligent monitor for an anesthesia breathing circuit.

A competent breathing circuit is mandatory to the safe and effective delivery of oxygen and anesthetic gases to the patient. Studies have shown that failures in the circuit are the most likely causes of anesthetic mishaps. Unfortunately, the complexity of the system renders traditional monitoring methods ineffective. We have developed a hierarchical artificial neural network monitor that is capable of examining ventilator signals. It was trained to identify 23 faults in the breathing circuit during ventilator controlled breathing and 21 faults during spontaneous breathing. The networks correctly identified a fault condition in 92% and 83% of cases for ventilator and spontaneous data, respectively. The correct fault type was found in 76% and 68% of cases for ventilator and spontaneous data, respectively. Results show that the network met our criteria for a holistic, specific, and vigilant monitoring system.

An effectiveness study of a new piezoelectric sensor for train-of-four measurement.

We have developed an easy-to-use, noninvasive piezoelectric sensor for quantitative monitoring of neuromuscular block. In a clinical evaluation with 23 patients, the piezo sensor was objectively compared to a mechanomyogram (MMG) for its ability to measure train-of-four (TOF) ratio from the adductor pollicis. After administration of succinylcholine (120-200 mg intravenously [i.v.]) to facilitate intubation, neuromuscular block was maintained with vecuronium by either boluses (1-2 mg i.v.) or an infusion (0.4-1.0 micrograms.kg-1.min-1 i.v.). Paired measurements were made of the TOF ratio from both sensors over a complete range of block levels (8%-100%). The difference in the TOF ratio measurement between the sensors showed a bias of 0.018. The SD of the difference between the sensors was +/- 0.129. The limits of agreement, which define the range in which 95% of the differences between the sensor measurements lie, were from -0.24 to 0.275. The sensitivity of the piezo sensor for detecting recovery based on a TOF ratio greater than 0.70 was shown to be 0.74 with specificity of 0.91. Under the conditions tested, the piezo sensor was not as accurate as the MMG. However, it was able to predict recovery of neuromuscular block with better accuracy than shown previously by manual evaluation of the TOF ratio, making it a reasonable, convenient alternative for quantitative monitoring of recovery from neuromuscular block.

A breathing circuit alarm system based on neural networks.

OBJECTIVE: The objectives of our study were (1) to implement intelligent respiratory alarms with a neural network; and (2) to increase alarm specificity and decrease false-alarm rates compared with current alarms. METHODS: We trained a neural network to recognize 13 faults in an anesthesia breathing circuit. The system extracted 30 breath-to-breath features from the airway CO2, flow, and pressure signals. We created training data for the network by introducing 13 faults repeatedly in 5 dogs (616 total faults). We used the data to train the neural network using the backward error propagation algorithm. RESULTS: In animals, the trained network reported the alarms correctly for 95.0% of the faults when tested during controlled ventilation, and for 86.9% of the faults during spontaneous breathing. When tested in the operating room, the system found and correctly reported 54 of 57 faults that occurred during 43.6 hr of use. The alarm system produced a total of 74 false alarms during 43.6 hr of monitoring. CONCLUSION: Neural networks may be useful in creating intelligent anesthesia alarm systems.