ICP reflects the global pressure in the intracranial vault. Measurement of ICP remains the most commonly performed type of monitoring in patients with acute brain injury and is routinely used to guide medical care. Elevated intracranial pressure can compromise cerebral blood flow and lead to brain herniation. The most common devices for ICP measurement are intraventricular catheters and fiberoptic intraparenchymal monitors. Intraventricular catheters are surgically placed into the frontal horn of the lateral ventricle and connected to an external pressure transducer. They are highly accurate and allow for therapeutic drainage of CSF. The primary disadvantage is a high rate of infection (5–10%) and difficult placement in case of small or compressed ventricles. Intraparenchymal monitors consist of a small catheter with a fiberoptic microtransducer at the tip. They have a lower complication rate and are not at risk for catheter occlusion or leakage. Their main drawbacks include the inability to drain CSF for diagnostic or therapeutic purposes and the potential to lose accuracy over several days, as catheters cannot be recalibrated after initial placement [1]. The decision to place an ICP monitor is generally made when a patient is suspected to have elevated ICP (imaging and exam findings) and coma (GCS of 8 or less). Normal values range between 7 and 15 mmHg. The threshold that defines intracranial hypertension is uncertain but generally values above 20–25 mmHg are considered pathological [2]. It is important to consider that when little or no CSF volume is left due to brain swelling, compartmentalized intracranial hypertension may exist. Uniformly distributed ICP requires that CSF can circulate freely between all its natural pools, thus equilibrating pressure everywhere. In circumstances where CSF becomes trapped in isolated areas, ventricular catheters may not reflect ICP in all intracranial compartments.

Continuous measurement of ICP can be use to calculate cerebral perfusion pressure (CPP). Under normal circumstances, the cerebral vasculature has the intrinsic ability to maintain a stable blood flow despite changes in cerebral perfusion pressure, a mechanism known as cerebral autoregulation [3, 4]. This mechanism ensures that the cerebral blood flow matches the brain’s metabolic demands and protects it from hypo- or hyperperfusion. After acute brain injury, this precise control of cerebral blood flow is frequently impaired, and as a result, acute changes in systemic pressure are passively transmitted to the cerebral circulation. This may lead to insufficiently low cerebral blood flow causing ischemia, or conversely, too high flow causing intracranial hypertension and cerebral edema. Although the optimal CPP for a given patient may vary, it is generally kept between 60 and 110 mmHg.

In addition to absolute pressure measurements, analysis of the ICP waveform can provide clues about reduced brain compliance (Δ volume/Δ pressure), which often precedes frank ICP elevations [5]. As compliance falls, the second peak of the ICP waveform (P2 or tidal wave) becomes elevated relative to the first peak (P1 or percussion wave) giving the wave a more rounded appearance (Fig. 20.2).

Elevated ICP is consistently associated with poor outcome and increased mortality [6]. Refractory ICP is associated with a drastic increase in risk of death [7]. However, a recent randomized controlled trial failed to demonstrate a benefit from ICP monitoring after traumatic brain injury [8]. This trial compared two management strategies, in which treatment was triggered by either ICP monitoring or by a combination of physical exam findings and neuroimaging. It must be emphasized that evaluation and treatment of elevated ICP was fundamental to both groups in this trial, and although there was a trend towards lower mortality and more efficient care, ICP monitoring alone may not be enough to enhance TBI outcome. Care of severe TBI is complex with several mechanisms contributing to secondary brain injury. Consequently, it may require additional monitors (i.e. brain multimodality monitoring) to gain better insight into patient specific pathophysiology and provide more targeted care in order to improve outcome.

Recent advancements in technology allow direct measurement of regional blood flow via thermal diffusion (TD-rCBF). The probe consists of small catheter that contains two metal plates at its distal tip (about 5 mm apart). The distal plate is minimally heated thereby generating a constant spherical temperature field. Temperature variations at the proximal sensor are a measure of the tissue’s ability to transport heat and correlate with cerebral blood flow. The probe is usually inserted into the brain parenchyma through a multi-lumen bolt next to ICP and PbtO2 probes. It provides a continuous and quantitative measure of blood flow in a small volume surrounding the catheter tip. Regional CBF values have shown good agreement with xenon enhanced CT [9]. The technique has been used in combination with PbtO2 to optimize CPP after TBI and guide blood pressure management in SAH patients with vasospasm [10, 11]. While the technology is promising, the available data is limited without clearly defined ischemia thresholds. Although the risks of probe placement are small (1–2% risk of bleeding, infection) and comparable to the complication rate of other intraparenchymal catheters, the utility of rCBF monitoring is limited due to its invasive nature as well as its small sample volume with uncertainty about where to place the probes. Furthermore, the accuracy of TD-rCBF probes may be influenced by elevations in temperature.

Venous blood from the brain drains via the cerebral sinuses and jugular veins to the right atrium. Measurement of the oxygen saturation in the draining blood provides information about the balance between oxygen delivery and the cerebral metabolic demand. Simplistically, when metabolic demand exceeds supply the brain extracts more oxygen resulting in a decreased SjvO2.

For the measurement of SjvO2, a catheter is placed retrograde via the internal jugular vein into the jugular bulb (dilated portion of the jugular vein just below the base of the skull). It is important that the catheter tip is positioned beyond the inlet of the facial vein and inferior petrosal sinus to avoid contamination with oxygen-rich, extracerebral blood. Placement of the catheter tip in the jugular bulb should be confirmed with a lateral skull radiograph. The tip of the catheter should be at the level of the mastoid air cells [12].

O2 saturation can be measured continuously using a fiberoptic catheter or intermittently by drawing and analyzing a blood sample.

Under physiologic conditions, cerebral blood flow matches metabolic rate of oxygen (CMRO2) and the difference in oxygen content between arterial and jugular venous blood (AVDO2) remains constant. If arterial oxyhemoglobin saturation and hemoglobin concentration remain stable, the SjvO2 is a good approximation of the AVDO2. A low SjvO2 (i.e. jugular desaturation) indicates increased oxygen extraction from the blood, suggesting inadequate cerebral blood flow in relation to CMRO2. A high SjVO2 indicates that oxygen supply is greater than demand (Table 20.1).

Monitoring of jugular venous oximetry can be considered in comatose patients (GCS ≤8) at risk for cerebral ischemia such as severe traumatic brain injury or high-grade subarachnoid hemorrhage [2]. In these patients monitoring of SjvO2 allows early detection of ischemia and can be used to guide hyperventilation therapy [13–16]. Robertson et al. reported that jugular venous desaturation (SjvO2 <50%) were a frequent occurrence in patients with severe traumatic brain injury and in the majority of cases could be attributed to elevated intracranial pressure, hypocarbia and hypotension. Furthermore, the number of desaturations was associated with an increased risk of death and poor neurologic outcome [14, 16]. Hyperventilation therapy is frequently used in the management of intracranial pressure. Reductions in PaCO₂ lead to cerebral vasoconstriction resulting decreased intracranial volume and pressure. However, the reduction in blood flow can also cause cerebral hypoperfusion and thus precipitate or worsen ischemia. SvjO2 has been shown to correlate well with brain tissue oxygen (PbtO2) and SvjO2 monitoring can be useful to optimize hyperventilation therapy [15]. However, small regions of critically hypoperfused brain have been demonstrated even while SvjO2 remained above 50% [17].

It is important to emphasize that SjvO2 is a global measure of cerebral oxygenation and not very sensitive for detecting regional ischemia. For jugular venous desaturations <50% to occur, at least 13% (170 ml) of the brain has to become ischemic [18]. Further limitations include considerable variability in saturations measured from both sides of the brain [19]. The sensitivity to detect jugular venous desaturation can be increased by cannulating the side of the predominant lesion or for diffuse injury the side with the larger jugular foramen on CT imaging [20]. Alternatively, the dominant internal jugular vein can be determined by unilateral compression of the vessel and selection of the side that produces the greater rise in ICP as it likely represents the side of predominant venous drainage [21]. Contamination with extracranial blood can occur if the catheter is placed too proximally or in cases of intermittent sampling if blood is aspirated rapidly (>2 ml/min). Inaccuracies may also occur if catheter tip is thrombosed or impacted against the vessel wall. While the rate of infection and complications related to catheter insertion are rare, the incidence of subclinical thrombosis is reported in up to 40% of cases [22].

Continuous electroencephalopathy monitoring (cEEG) in the neurocritical intensive care unit is primarily used for the detection of non-convulsive seizures (NCSz) or status epilepticus (NCSE) in patients with unexplained changes in level of consciousness. EEG is also helpful for the characterization of sudden spells such as tremors, twitching, posturing, eye deviation or agitation. Other indications in include detection of cerebral ischemia, monitoring the level of sedation and titration of medication during anaesthetic coma as well as prognostication.

Seizures are seen in 10–30% of patients with acute brain injury [23, 24]. While the majority of them are not exhibiting motor features, subclinical seizures have been associated with increased intracranial pressure and disturbed brain metabolism possibly leading to secondary brain injury [25, 26]. Furthermore, patients with seizures have increased mortality and among patients with SAH seizure burden has been associated with poor functional and cognitive [27, 28].

There is limited data to support the use of cEEG compared to spot EEGs (approximately 30 min); however, spot EEG will not detect nonconvulsive seizures in about half of those having seizures compared to longer monitoring [23].

Another important application of EEG in neurocritical care is the detection of brain ischemia. Brain function is represented on EEG as an oscillating wave of various frequencies. EEG activity is conventionally divided into the following frequencies (number of waveforms per second):

Most of the waveforms are generated by pyramidal neurons within the cortex. These cells are very sensitive to hypoxia and ischemia and EEG abnormalities can be seen within seconds to minutes. With decreasing cerebral blood flow, the EEG progresses through predictable changes: (1) loss of faster beta frequencies, (2) slowing of background to theta and later delta range, and (3) background attenuation and finally suppression of all frequencies [29]. Therefore, EEG allows providers the opportunity to detect ischemia and to alter treatment before hypoperfusion leads to cell death and permanent injury. However, some of the early changes can be subtle and may not be appreciated on raw EEG. Mathematical processing of the raw data, also known as quantitative EEG (qEEG) analysis, can help to better visualize these changes. Using frequency analysis, the original EEG can be quantified in terms of frequency, amplitude and rhythmicity, which allows the calculation of numerical values or percentages. Furthermore, this digital form can compress long periods of EEG data into readable graphs allowing for prolonged monitoring of patients and aid detection of ischemia. The ideal qEEG measure to identify ischemia is still being debated and may depend of the clinical situation. A variety of parameters to measure slowing or attenuation have been suggested; they commonly include some ratio between fast and slow frequencies or a measure of the relative power (i.e. amount) of a specific frequency within the entire EEG spectrum.

Ischemia monitoring can be particularly useful in comatose or sedated patients when the clinical exam is limited. A reduction in alpha/delta ratio (8–13 Hz/1–4 Hz) and a decrease in the variability in relative alpha frequency have been shown to correlate with angiographic vasospasm or delayed cerebral ischemia (DCI) in patients with poor-grade SAH [30, 31]. For detection of focal ischemia, the relative delta percentage appears to provide the most robust correlation with CBF [32]. Clinically, qEEG has been correlated with stroke severity as measured by the National Institutes of Health Stroke Scale, infarct volume on MRI, and functional outcome [33–36]. Furthermore, qEEG can be used to assess treatment response, and given the correlation to cerebral perfusion pressure may be helpful to guide blood pressure management in individual patients after large-vessel acute ischemic stroke [37].

Despite the many advantages of cEEG, it does have its drawbacks. The monitoring systems, technician training, and the costs of disposable equipment can be expensive. Frequent review of many hours of data is necessary and can be time consuming. Quantitative EEG has the promise to reduce some of this time; however, having clinicians specially trained to interpret qEEG results is essential.

Near-infrared spectroscopy (NIRS) is a non-invasive technique to measure the oxygen saturation of hemoglobin in the brain tissue. Delivered via optodes placed on the skin, near-infrared light has the ability to penetrate through scalp, bone and brain tissue up to a depth of about 3 cm. NIRS operates on the principle that the majority of near-infrared light is absorbed intracranially by oxygenated hemoglobin (HbO2) and de-oxygenated hemoglobin (Hb). Assuming constant scattering and applying knowledge of the absorption spectra for HbO2 and Hb, light attenuation can be converted into concentrations of HbO2 and Hb using the modified Lambert–Beer Law [38]. Their sum provides the total hemoglobin concentration (HbT) and the ratio HbO2/HbT equals the oxygen saturation of hemoglobin in brain tissue (StO2 or ScO2). In the setting of stable arterial oxygenation and cerebral metabolic rate, StO2 has been used as a surrogate for cerebral blood flow [39].

NIRS has many clinical applications. For example, NIRS can be used to monitor for changes in cerebral blood flow and oxygenation without the need for invasive probes and has been used to provide information regarding cerebral autoregulation after TBI, stroke and subarachnoid hemorrhage [40–42].

Despite the increased utilization of NIRS, there are some important limitations. First and foremost, there is concern for extracerebral contamination of the NIRS signal. In order to measure brain tissue oxygen saturation photons must travel through scalp, skull and dura, which contain various concentrations of blood and other tissue-derived chromophores (light absorbing molecules), which potentially confound the signal derived from the cerebral cortex [43]. In addition, NIRS light attenuation is not just the result of absorbtion by target chromophores (Hg and HgO2), but also light scattering. Bone, hair, subgaleal collections, subdural hematomas, and differences in areas of subarachnoid spaces can result in in nonlinear relationships between absorption and attenuation changes [44].

Cerebral microdialysis is used to measure brain tissue chemistry thus providing important information about brain metabolism and more specifically the adequacy of energy supply and cellular function. A thin catheter is placed through a cranial bolt into an area of interest in the brain parenchyma. The catheter is lined with a semi-permeable dialysis membrane and constantly perfused at a very low rate with an isotonic solution (e.g. artificial cerebrospinal fluid). Molecules below a certain size (usually 20 kDa) diffuse from the extracellular space through the membrane into the perfusion fluid, which is collected at regular intervals (e.g. every 60 min) and analyzed at the bedside using the manufacturer’s equipment [45]. The analysis usually includes concentrations of glucose, pyruvate, lactate, glutamate, and glycerol. Many other metabolites as well as exogenous substances such as administered drugs can be studied, however, their clinical utility remains to be determined.

Glucose, pyruvate, and lactate provide information about the available fuel source of the brain as well as the brain’s ability to go through aerobic metabolism. When energy is needed, glucose undergoes a series of enzymatic conversions known as glycolysis. Under aerobic conditions, the resulting pyruvate enters the citric acid cycle and gets metabolized to ATP. During hypoxia and ischemia, the end product of pyruvate is lactate resulting in an increased lactate to pyruvate ratio (LPR). Ischemia also results in the release of glutamate, a marker of metabolic distress. High glycerol levels originating from glycero–phospholipid containing cell membranes indicate cellular breakdown. Normal values for the metabolites have been established to help guide clinicians in interpreting this data (Table 20.2), however, variations over time and changes in response to therapeutic interventions may be more useful [46]. Furthermore, analyzing trends for multiple microdialysis markers rather than looking at individual metabolites may allow for a more meaningful interpretation. For example, a typical pattern of cerebral ischemia includes a marked decrease in brain glucose, elevated lactate, increase in LPR and lactate to glucose ratio (LGR) and a moderate decrease in pyruvate [45].