Neurologic Monitoring

Robert C. Tasker

Mateo Aboy

Alan Graham

Brahm Goldstein

KEY POINTS

Increasing evidence suggests that neuromonitoring with various modalities can improve diagnosis, treatment, and outcome in critically ill and injured children.

To understand the current and future applications of neuromonitoring in the pediatric intensive care unit, it is necessary to understand the basic engineering and physiologic principles that underlie different monitoring technologies.

Because digital signal processing algorithms generally use a moving window of the physiologic signal to generate estimates, the clinical parameters displayed on the bedside monitor represent an average over past values of the signal and cannot respond instantaneously to alarm conditions. Thus, in a deteriorating patient, monitors typically generate alarms after the alarm condition has persisted for several seconds and following resuscitation typically lag a few seconds after the patient’s condition improves.

Our management of three disease states necessitates advanced neuromonitoring. These include continuous electroencephalogram coupled with digital video monitoring for status epilepticus in diagnosing nonconvulsive seizures, intracranial pressure monitoring in patients with severe traumatic brain injury, and intraoperative and postoperative neurophysiologic monitoring of children who undergo repair of congenital cardiac disease with nearinfrared spectroscopy technology.

Multimodality neuromonitoring may have the potential to improve overall patient outcome, but definitive pediatric studies are required.

Increasing evidence demonstrates tangible benefits from neurologic monitoring in a variety of PICU patients. Continuous

electroencephalogram (EEG) coupled with digital video monitoring for status epilepticus (SE) is routinely applied with good results in diagnosing and managing nonconvulsive seizures. Intracranial pressure (ICP) monitoring is the mainstay of our management in patients with severe traumatic brain injury (TBI). Near-infrared spectroscopy (NIRS) technology is used increasingly in the neonatal, cardiac surgical, and extracorporeal membrane oxygenation populations. Last, the practice of transcranial Doppler (TCD) insonation of intracranial vessels with determination of cerebral blood flow (CBF) velocity is a

practice being translated from neurocritical care in adults to PICU practice. Thus, it is important for PICU personnel to understand the underlying engineering and physiologic principles of such neuromonitoring.

SCIENTIFIC FOUNDATIONS

General Engineering Aspects of PICU Monitoring

Medical instrumentation systems are often composed of sensors, signal conditioning hardware and software, output displays, and auxiliary signals. Sensors are used to convert a physical measurement (i.e., a quantity, property, or condition of interest) into an electrical signal output. The sensors used for medical instrumentation purposes are designed to be minimally invasive and to respond to the source of energy present in the measure, while excluding all other sources as much as possible. Generally, the electrical signal produced by these sensors cannot be connected directly to the output display device. Signal conditioning and processing, such as amplification and analog filtering, are typically required (1). Additionally, the sensor outputs are analog signals and must be converted to digital form before they can be processed using more advanced digital signal processing techniques. Analog-to-digital (A/D) conversion involves signal conditioning, antialiasing filtering, uniform sampling, and quantization. Table 58.1 provides definitions for the engineering terms used in this chapter. For an in-depth discussion of medical instrumentation, such as sensors, biopotential electrodes and amplifiers, blood pressure, flow, and volume measurement equipment, chemical biosensors, and imaging systems, the reader is referred to a general textbook on biomedical engineering (1).

Most signals are filtered with analog integrated circuits before A/D conversion. These frequency-selective filters are usually linear bandpass or highpass filters used to remove drift, prevent aliasing, and reduce noise; they can distort the waveform morphology due to nonlinear phase in the passband or removal of signal frequencies.

During A/D conversion, analog signals are sampled at a rate determined by the manufacturer (i.e., the sampling rate or sampling frequency). To accurately represent the signal on the patient’s monitor display, the sampling rate must be high enough so that a linear interpolation between the sample points results in a visually smooth and representative signal. To achieve this, the sampling rate should be at least 10 times higher than the highest-frequency component of the prefiltered signal. For most physiologic signals that vary with the cardiac and respiratory cycles, a sampling rate of 100 Hz is sufficient. Due to the impulsive nature of the QRS complex, electrocardiogram (ECG) signals have a higher bandwidth than most other physiologic signals encountered in the PICU and require a higher sampling rate of at least 250 Hz to accurately represent different segments of the ECG waveform. EEGs also require a higher sampling rate, although this signal is infrequently displayed directly on PICU patient monitors.

TABLE 58.1 ENGINEERING TERMS AND DEFINITIONS

Aliasing—The apparent conversion of high-frequency signals to low-frequency signals due to an insufficient sample rate

Analog-to-digital (A/D) conversion—The electrical conversion of an analog signal (often a voltage) to a digital representation that enables manipulation and processing by computers

Bandpass filter—A filter that eliminates low- and high-frequency components of a signal, but retains an intermediate range

Bandwidth—The range of frequencies spanned by a signal. When applied to bandpass filters, it describes the range of frequencies that are allowed to pass through the filter

Capacitance—A measure of the ability of a circuit element to store electrical charge. Elements with a large capacitance dampen or resist rapid fluctuations in voltage

Demodulation—The process of extracting an information-bearing signal from another signal; analogous to extracting a file of interest from a compressed or encrypted file

Hertz (Hz)—A measure of frequency; equivalent to cycles per second (cps)

High-frequency noise—Many types of artifact in physiologic signals contain significant power at high frequencies. This noise is often emitted by medical equipment near the patient

Highpass filter—A filter that eliminates low-frequency components of a signal but retains high-frequency components

Inductance—A measure of the ability of a circuit element to store energy in a magnetic field. Elements with a large inductance dampen or resist rapid fluctuations in current

Linear interpolation—The process of estimating a value of a signal or function between two intermediate values using a line between the two points

Low-frequency noise—Some types of artifact in physiologic signals contain significant power at low frequencies. This noise is often caused by patient movement

Lowpass filter—A filter that eliminates high-frequency components of a signal but retains low-frequency components

Modulation—The process of embedding an information-bearing signal in another signal; analogous to creating an encrypted or compressed file that contains a file of interest

Moving window—A technique for estimating an average quality of a signal continuously by averaging over period of the preceding 3-5 s. For example, the systolic blood pressure is usually calculated by averaging the systolic peaks over a moving window of 3-5 s

Nonlinear—Any system or device the behavior of which is governed by a set of nonlinear equations

Signal power—The power contained in a signal is generally defined as the square of the signal; often averaged over a moving window to create a smoothly varying continuous estimate

In addition to the sampling rate requirements, quantization requirements must be met to avoid quantization error. Each sample of the digital signal is represented by B bits and can only take one of 2^B levels. The quantization error is the error that results from using the quantized signal rather than the true signal amplitude.

Once the physiologic signals have been converted to digital form, more advanced digital signal processing algorithms are used to process these physiologic digital signals and extract clinically significant parameters. For instance, heart rate is estimated from the ECG signal using automatic QRS detection algorithms, and diastolic and systolic pressures are obtained from pressure signals. It is important for

the clinician to understand that digital signal processing algorithms generally use a moving window of the physiologic signal to generate estimates. These moving-window segments (signal frames) range from 3 to 10 seconds in duration. Consequently, the clinical parameters obtained represent an average over past values of the signal and monitors do not respond instantaneously to alarm conditions. Thus, patient monitors typically generate alarms after the alarm condition has persisted for several seconds. The obverse is also true. For example, after a successful resuscitation from cardiopulmonary arrest, the arterial oxygen saturation value will typically lag a few seconds after the patient’s cyanosis has resolved.

Common Clinical Questions

Specific problems in assessing the neurological state of PICU patients arise when sedative and muscle relaxant medications have been used as part of supportive care. Some of the typical questions that clinicians have about patients at risk of brain injury are:

Is the unresponsive and immobile patient awake or aware of pain from procedures?
Are repetitive movements purposeful responses, reflexes, or seizures?
Is this patient at risk of ongoing subclinical or nonconvulsive seizures?
Is the patient at risk of cerebral edema and intracranial hypertension?
Is the patient at risk of cerebral ischemia?
Is the brain functionally intact or irreversibly damaged?

These questions fall into two categories, diagnostic versus monitoring requirements. Both of these requirements are essential and usually complimentary. Diagnostic assessment focuses on documentation of a certain event, such as nonconvulsive seizures using the EEG or middle cerebral artery vasospasm using TCD. Theoretically, monitoring acquires continuous information that allows accurate detection of intracranial changes that will lead to more precise treatment. In this regard, monitoring should be undertaken if the following three conditions are satisfied:

Significant changes in the signal provide ongoing diagnostic information or occur at the stage of reversible injury.
Criteria are well established for identifying significant changes in the signal.
Identification of diagnostic information or new signal from monitoring results in a change of practice. In regard to information from monitoring, there is some intervention that bedside caregivers can correct so as to minimize injury.

Clinically Important Physiologic Systems and Signals

A limited number of physiologic systems may be monitored within the central nervous system (CNS). Table 58.2 provides a list of commonly used monitoring modalities, according to frequency of usage in the PICU.

The EEG has been widely available for decades to record and quantitatively measure brain electrical activity. Technology for monitoring ICP has also been widely available for decades, and now is accomplished with more accurate fiber-optic technology. Additional techniques involve the surrogates of CBF and include ultrasound and TCD, brain tissue oxygenation using NIRS, and jugular venous oxyhemoglobin saturation (SjvO₂). Additionally, monitoring of cellular and extracellular processes includes local brain tissue oxygen tension (PbtO₂) and extracellular fluid concentrations of glutamate, glucose, lactate, and pyruvate using microdialysis. Finally, the term multimodality neuromonitoring refers to the simultaneous use of different combinations of these methods to provide a more complete physiologic picture of CNS activity from cellular, to tissue, to organ level.

TABLE 58.2 MONITORING SYSTEMS

▪ SYSTEM	▪ TYPE	▪ FEATURES	▪ MEASURE	▪ APPLICATIONS	▪ USAGE
Clinical	Coma scale scores	Noninvasive Regional or global Intermittent	Change in clinical parameters	Very sensitive in awake, non-intubated patients	Common in all cases
ICP	Invasive pressure transducer	Invasive Regional Continuous	Change in pressure with known norms	Guides treatment in intracranial hypertension algorithm	Common in severe TBI
EEG	Brain activity (focal and global)	Noninvasive Focal or global Intermittent or continuous	Seizure detection, depth of anesthesia, and severity of injury	Guides treatment and prognostic	Common in seizure cases and encephalopathy
NIRS	Light absorption in near-infrared range	Noninvasive Regional Continuous	Estimate of frontal region cerebral venous oxygen saturation	Identification of unappreciated cerebral hypoxia; desaturation associated with outcome	Less common with most data in cardiac cases
TCD	Ultrasound of cerebral arteries	Noninvasive Regional Intermittent or continuous	Assessment of CBF-velocity; assessment of vasospasm	Noninvasive assessment of CBF-velocity, ICP and autoregulation	Rarely used in pediatrics since lack of norms
PbtO₂	Intraparenchymal electrode measurement	Invasive Focal Continuous	Focal assessment of oxygen tension	Level III recommendation in adult severe TBI, with certain thresholds	Rarely used in pediatrics
SjvO₂	Assessment of cerebral venous oxygen saturation	Invasive Global Intermittent or continuous	Global assessment of cerebral oxygen extraction	Level II recommendation in adult severe TBI, with certain thresholds	Rarely used in pediatrics
Pupillometry	Pupil assessment with standardized light source	Noninvasive Regional Intermittent	Quantified pupillary diameter and reactivity	Detects pupillary constriction even in small pupils	Rarely used in pediatrics
Cerebral microdialysis	Measurement of cerebral analytes	Invasive Focal Intermittent	Focal assessment of bioenergetics and tissue glutamate	Some recommendation for use in adults with severe TBI or poor-grade SAH	Rarely used in pediatrics
CBF, cerebral blood flow; EEG, electroencephalography; ICP, intracranial pressure; PbtO₂, partial pressure of oxygen in brain tissue; NIRS, near-infrared spectroscopy; SAH, subarachnoid hemorrhage; SjvO₂, oxygen-hemoglobin saturation in jugular vein (bulb); TBI, traumatic brain injury; TCD, transcranial Doppler.

APPLICATIONS IN THE PICU

The most commonly used forms of monitoring in the PICU are EEG and ICP. Of the other modalities described in Table 58.2,

NIRS and TCD have ongoing clinical evaluations in the pediatric critical care literature. The focus of this chapter will be on these four forms of monitoring.

Diagnostic EEG and Evoked Potentials

The EEG monitors the electrical activity of the brain observed via scalp electrical potentials. The sources of this electrical activity are the neurons located predominantly in the outermost layers of the cerebral cortex. Thus, the EEG follows the spatial
topography of the cortex. The information provided can sometimes be diagnostic of certain forms of encephalopathy (e.g., hepatic) or indicative of the severity of encephalopathy (2,3,4,5).

Vision, hearing, sensory, and motor functions may all be discretely monitored to assess peripheral and CNS

Only gold members can continue reading. Log In or Register to continue