Intracranial pressure (Monitoring and imaging) Part 2

Measurement techniques of intracranial pressure

In sedated patients with TBI, continuous ICP monitoring is recommended, and can only be achieved by direct invasive measurement. he indications for ICP monitoring are discussed below.

Different methods of monitoring ICP have been described. he gold standard for ICP monitoring is a catheter inserted into the lateral ventricle (usually via a small right frontal burr hole) and connected to an external pressure transducer. he reference point for the external pressure transducer is the foramen of Monro, which in practice is equated to the external auditory meatus. he advantages ofventricular catheters are the feasibility of repetitive calibration, withdrawal of CSF to treat elevated ICP and that it is a low-cost method. Disadvantages include difficulty with insertion in patients with brain swelling and small ventricles. Furthermore, the external pressure transducer needs to be moved to accommodate patient head movement so that an appropriate reference can be assured. Intraventricular catheters have infection rates between 2 and 27%, with significant attendant morbidity and mortality. Factors identified to increase the risk of external ventricular drain (EVD)-related infections are duration of catheterization, frequency of EVD manipulation (CSF sampling), insertion technique and intraventricular haemorrhage. In particular, after intraventricular haemorrhage, the catheters may become blocked .

The most common location for ICP monitoring nowadays is the brain parenchyma using intraparen-chymal probes. In these devices, a miniature strain gauge pressure sensor, a semiconductor strain gauge or fibre optic or pneumatic technology is used to transduce pressure. In the miniature strain gauge pressure sensor, ICP results in a change in resistance, and with the fibre optic, ICP results in a change in reflection of the light beam. he pneumatic system uses a catheter with an air pouch at the tip, and transmits the pressure over the catheter to the electronic hardware. he pneumatic system is able to calibrate itself repeatedly, but miniature pressure sensor and fibre-optic devices cannot be recalibrated in vivo, and a small drift of the zero reference may occur. However, this drift is generally considered negligible. he complication rate of intra-parenchymal probes is very low, with infection rate or the risk of major bleeding being below 2%.

As a substitute for the intraparenchymal probes, subarachnoid, subdural and epidural probes can be used, but the accuracy of these devices is lower. In an acute setting, estimation of ICP through a lumbar drain is generally not recommended, as the accuracy is limited and brain swelling or a space-occupying lesion can cause brain herniation.

The concept of cerebral perfusion pressure

Cerebral perfusion pressure is defined as the difference between cerebral arterial pressure and pressure in the cerebral venous bed just before the outlet to the major venous sagittal sinuses, i.e. in cortical or bridging veins. Cerebral perfusion pressure is the pressure driving blood to flow through the cerebrovascular bed. As the pressure in bridging veins is difficult to measure and can be approximated by ICP, the clinical definition of CPP is:

CPP = mean arterial pressure (MAP) – mean ICP

Too low a CPP causes ischaemia, while too high a CPP causes hyperaemia. he ability of the cerebrovascular bed to autoregulate CBF depends on CPP, and there are defined upper and lower CPP limits within which such autoregulation can operate. herefore, decreasing CPP is particularly dangerous after head injury: it both decreases the driving force for cerebral blood to flow and compromises autoregulation. Cerebral perfusion pressure-oriented therapy has been introduced to decrease the risk of ischaemia in post-injury care. he distribution of mean CPP for different outcome groups after head injury matches that of mean ICP. In patients who die, the mean CPP is significantly lower. Other outcome groups have similar CPP levels.

Intracranial pressure monitoring

The first continuous manometric monitoring of ICP was performed by Guillaume and Janny in 1951. Further studies, describing various waves of ICP, were conducted by Lundberg, who is regarded as the founder of contemporary continuous ICP monitoring.

Hie indications for ICP monitoring vary from centre to centre and include head injury, intracerebral haemorrhage, subarachnoid haemorrhage, hydrocephalus, ischaemic stroke, hypoxic brain injury with cerebral oedema, meningitis/encephalitis and hepatic encephal-opathy. For traumatic head injury, established guidelines exist (e.g. Brain Trauma Foundation Guidelines). here is also support for the use of ICP monitoring in selected patients with intracranial haemorrhage and subarach-noid haemorrhage. While routine ICP monitoring in all patients with hemispheric stroke has no impact on management or outcome, it may be useful in selected patients with severe intracranial hypertension and/or midline shift. Similarly, it may have a role in selected patients with severe meningoencephalitis or fulminant hepatic failure. In all instances, the benefits of ICP monitoring in the detection and management of intra-cranial hypertension must, in each individual patient, be weighed against the clinical and economic costs of such monitoring: in addition to the cost of the devices used, the technique requires special medical and nursing expertise for device insertion and maintenance, and has associated risks of haemorrhage or infection.

In head injury, guidelines have been established and periodically updated (http://www.braintrauma. org). Patients with traumatic head injury who present with a low Glasgow Coma Score (GCS) of <8 and CT abnormalities such as haematomas, contusions, swelling, herniation or compressed basal cisterns should receive ICP monitoring. Also patients with a GCS of <8 and normal CT scan may be considered for ICP monitoring if they are >40 years and uni- or bilateral motor posturing or systolic ABP under 90 mmHg occur. hese recommendations may be modified from centre to centre. Nevertheless, any diagnostic or therapeutic intervention has to be considered individually according to the patient’s age, personal medical history and assumed will. In monitored patients, thresholds for treatment are generally set at 20-25 mmHg.

Monitoring requirements and ICP treatment thresholds in other conditions with persistent high ICP, such as chronic hydrocephalus or idiopathic intracra-nial hypertension (also known as pseudotumour cere-bri), need to be decided in context. he rise in ICP in these conditions signifies a disturbance of CSF circulation due to an increase in resistance of CSF outflow or increased cerebral venous pressure. In both diseases, overnight ICP monitoring and assessment of the cere-brovascular volume compensatory reserve through CSF infusion may add some additional diagnostic information, especially in patients who suffer under persisting or recurring symptoms after shunt insertion.

Normal ICP varies with age and body position. In the upright position, ICP is negative with an approximate mean of -4 mmHg but not exceeding -10 mmHg. Furthermore, in the absence of disease, ICP may rise during coughing, sneezing or a Valsalva manoeuvre up to 50 mmHg without noticeable neurological impairment. herefore, it is the interaction of raised ICP with other intracranial pathology that leads to the pathological consequences, rather than elevated ICP per se.

There are a few distinctive patterns of ICP usually seen in recordings after TBI (Fig. 4.4). In head-injured patients, low and stable pressure is characterized by a low and stable mean ICP over time (around 15 mmHg) and a low pulse amplitude. his pattern can be observed during the first 6-8 h after head injury without or with minimal signs of brain swelling or a space-occupying lesion in cross-sectional imaging of the brain (Fig. 4.4a). When ICP is elevated over 20 mmHg, the pulse waveform may express vasogenic waves with a high amplitude due to reduced intracranial compliance. A further increase in ICP may lead to secondary cerebral insults that may cause cerebral oedema that again may increase ICP (Fig. 4.4b).

Various waves of ICP may be observed in clinical practice. Lundberg plateau waves (Fig. 4.4c) undoubtedly have a vasogenic origin, and the underlying model was elegantly described as a ‘vasodilatory cascade’ by Rosner. According to his theory, any vasodilatory stimulus produces an initially small increase in CBV. Under conditions of a weak pressure-volume compensatory reserve, this small change in CBV produces a substantial rise in ICP, leading to a decrease in CPP, further autoregulatory-induced rises in CBV, a further rise in ICP, and so on, until a point is reached where autoregulator vasodilation is maximal. At the plateau phase, both CPP and CBF are reduced (Fig. 4.4c). he plateau wave may be terminated spontaneously when a vasoconstrictory stimulus occurs and reverses the cascade to reduce ICP. Such spontaneous termination may be the consequence of a Cushing response, which elevates the MAP and reverses the vasodilation associated with autoregulation. Short-term hyperventilation or a bolus of hypertonic saline may be administered to terminate the plateau wave. Plateau waves are not associated with a worse outcome after head inj ury, unless they last excessively long – longer than 30 min. herefore, active termination of long plateau waves is extremely important in the neurocritical care of TBI patients.

Fig. 4.4. Examples of intracranial pressure (ICP) recordings after head trauma: (a) Low and stable ICP; (b) elevated and stable ICP; (c) plateau waves of ICP; (d) elevation of ICP due to hyperaemia. ABP, arterial blood pressure; CPP, cerebral perfusion pressure; FV, blood flow velocity in the middle cerebral artery; SjO2, jugular bulb saturation.

Another source of temporary ICP elevation may be hyperaemia. An increase in ICP is initiated as in a plateau wave by vasodilation (increase in CBV) and is maintained by a rapid increase in brain oedema. Vessels stay dilated as ICP increases and CPP decreases, but, contrary to plateau waves, CBF stays elevated (Fig. 4.4d). Finally, and most frequently, waves and irregular patterns seen in ICP monitoring are associated with transients in ABP.

It should be stressed that ICP is more than a number; it has its mean value and dynamic, variable components. Diagnostically relevant information is included in both factors. A instantaneous ICP value of 8 mmHg does not preclude the possibility that the ICP 2 min later may be 80 mmHg. Only inspection of properly monitored time trends, ideally supported by computer analysis, warning about poor compensatory reserve, autoregulation failure, etc., may help in the interpretation of ICP with confidence.

Non-invasive intracranial pressure estimation

Hie middle cerebral artery is a large vessel with elastic walls. It can be considered a membrane transducer able to detect changes in transmural pressure and, with knowledge of the MAP, allow computation of ICP. Unfortunately the elastic properties of the membrane are unknown and may be variable in time. herefore, the value of the calibration coefficient and its linearity and stability in time are unknown. Nevertheless, it is known that CPP affects the shape of the blood flow velocity waveform, which can be non-invasively visualized using TCD. However, the arterial pulse waveform, heart rate, tension of arterial carbon dioxide, distal vascular resistance and age can all affect the blood flow velocity waveform and confound the estimation of ICP.

Some simple formulas have been proposed in the past to assess non-invasive CPP (nCPP) from ABP and blood flow velocity (FV) waveforms. Out of these, one particular method has reached possibly satisfactory accuracy (error <10 mmHg in >80% of measurements):

where FVd is diastolic blood flow velocity and FVm is mean flow velocity.

Hie availability of non-invasive estimates of CPP are useful both to estimate absolute CPP and to monitor changes in CPP with time. he 95% confidence limit for estimation of CPP is 12 mmHg. Although this seems to be satisfactory for CPP, such a precision would be not good enough to estimate ICP.

Non-invasive measurement of ICP can be carried out using various methods: tympanic membrane displacement, time-of-flight ultrasound through the skull, change in skull diameter, change in blood FV in the straight sinus or analysis of the pulse waveform of TCD. he pulsatility index (PI) increases with rising ICP. Prediction of absolute ICP using PI is not accurate enough as many other factors may influence PI (arterial pulse, heart rate, arterial carbon dioxide tension (PaCO2), vascular tone, proximal stenosis, spasm, etc.). Estimation of ICP can be carried out using the formula proposed by Aaslid:

where F1 and A1 are first harmonic components of fl ow velocity and arterial pressure pulse waveforms, respectively, which gives a 95% confidence limit of around ±25 mmHg.

Hie moving-average model of transmission between ABP and ICP, modified by the relationship between ABP and FV, gives a mean absolute error of around 6 mmHg (see Fig. 4.5). he method is based on analysis of a large data base of patients with homogeneous pathology undergoing full ICP, ABP and FV direct monitoring and is most probably pathology-dependent. Changes in PaCO2 , spasm and proximal stenosis are confounding factors.

Fig. 4.5. Example of dynamic noninvasive monitoring of intracranial pressure (ICP), using arterial blood pressure (ABP) and blood flow velocity (FV) waveform. The pale grey tracing shows the non-invasive estimate, and the grey trace shows simultaneously measured ICP during transient intracranial hypertension.

Non-invasive ICP techniques allow a wide insight into intrahemispherical pressure gradients. As long as the CSF communicates freely between different fluid cavities within the brain, there should not be any substantial differences in regionally measured ICP. his fact has been employed to use ultrasonographic or MRI measurement of the optic nerve sheath diameter to estimate ICP. Measured 3 mm behind the globe, an optic nerve sheath diameter >6 mmHg makes intra-cranial hypertension highly likely, while a diameter of <5 mm makes it very unlikely.

Direct measurements of pressure in two CSF compartments are rarely performed. However, following head injury, intrahemispheric pressure gradients have been reported. here are intrahemispheric pressure gradients in non-invasive CPP associated with mid-line shift, side of contusion (assessed using CT) or side of craniectomy. Surprisingly, non-invasive estimates of nCPP and nICP indicate that CPP is greater on the side of the contusion or expanding brain in the case of midline shift or on the side of craniectomy. his may support the hypothesis that the interhemispheric differences in ICP are the consequence not of brain tissue volume expansion but of a vascular expansion. his hypothesis may be further supported by the fact that cerebral autoregulation is worse on the side of contusion or brain expansion. Asymmetry in nCPP and nICP correlates with worse outcome following head injury.