Introduction

Abnormalities in cerebral blood flow (CBF) are implicated in the pathophysiology of increasingly large number of neurological and neurosurgical disorders ₁. The most obvious evidence of this awareness is the importance now given to maintaining an adequate cerebral perfusion pressure (CPP) in a wide variety of conditions. It is surprising that in spite of this knowledge, measurement of CBF has not become a routine tool in the management of patients with neurological illnesses. While CBF studies are increasingly the subject of research and publication, the vast majority of institutions caring for these patients do not have the equipment to perform this measurement, and only a small number of centers that have the capability use it in routine management. This is due to many factors including cost, ease of performance, concerns about safety and the current lack of ability to integrate the results into a standard patient care protocol.

The list of diseases in which CBF abnormalities are implicated is growing as our knowledge of the underlying processes increases. Ischemia, besides the being the obvious pathology in stroke, is now also known to have a significant effect on outcome in head injuries ₂, ₃, ₄, subarachnoid hemorrhage ₅, ₆ and in raised intracranial pressure (ICP) due to various causes. Hyperemia can have equally deleterious effects in head injuries ₇, and in patients who have had endovascular or surgical treatment for carotid artery disease or arteriovenous malformations ₈. Besides these primarily neurological diseases, abnormalities in CBF have been detected in psychiatric disorders ₉, ₁₀ and implicated in several systemic illnesses with neurological manifestations such as hepatic and renal failure ₁₁.

The awareness that CBF abnormalities are an important cause of pathophysiology has resulted in the need for an ability to measure the flow. Measurement will enable the physician to decide whether the cause of a deficit is abnormal flow, whether ischemic or hyperemic. It can also be used to assess the effectiveness of therapy aimed at modifying the abnormal state and therefore as a guide to further intervention.

Measurement of cerebral blood flow

Numerous techniques exist which are currently used clinically to estimate or measure CBF. These can be roughly classified into direct and indirect methods:

Direct

1) Kety and Schmidt performed the first measurements of CBF in humans using nitrous oxide and the Fick principle ₁₂. Since nitrous oxide is inert and diffuses rapidly into the brain, CBF could be calculated if its concentration could be simultaneously measured in the arterial blood entering the head and the venous blood leaving it, using a previously established blood-brain partition coefficient. This method is difficult to perform, as it requires multiple samples of arterial and internal jugular venous blood. It does not yield regional measurements but is presumed to be a measurement of global CBF; this also can be questioned because the variation in the dural sinuses and the torcula call into question the reliability of a sample taken from one jugular vein. A major advantage of this technique is that simultaneous measurement of blood gases and metabolites can be performed to assess cerebral metabolism.

2) 133Xe: the use of radioisotopes to measure cerebral blood flow was first reported by Ingvar et al ₁₃ in 1961, who used 85Kr to measure CBF in exposed brain. Later 133Xe was found to be more effective because it is a gamma emitter and can be measured extracranially. It is also safer because of a shorter half-life. The initial techniques were complicated involved injecting the isotope directly into the carotid artery, but the process was considerably simplified in 1967 by Obrist et al ₁₄, ₁₅ who developed a technique of inhaling the isotope as a gas, and even further simplified now by administering it as a solution in an intravenous injection. The procedure involves knowledge of the arterial concentration of the isotope (measured in end-tidal air) and the concentration of 133Xe in the brain, measured by multiple extracranial scintillation detectors. The procedure is easy to perform and can be carried out in the ICU with very little radiation hazard. It is however handicapped by a comparatively poor spatial resolution, especially for deeper brain structures (the "look through'' artifact ₁₆. This handicap has been overcome with the development of computing techniques and detectors that can provide three-dimensional measurements, but at the same time the process loses its main advantages of lower cost and mobility.

3) Positron emission tomography (PET): This technique yields the most information on the adequacy of cerebral circulation, as it can yield excellent three-dimensional images of not only the flow in any region of the brain but also the local metabolism to assess the appropriateness of this flow (quantitative and functional images)₁. It is however expensive for routine use and in addition requires the presence of an in-house cyclotron to produce the necessary positron emitting tracers.

4) Thermal diffusion flow probes: This is an invasive procedure that measures flow by estimating the temperature gradient between two plates on the surface of the brain ₁₇. It measures the blood flow only in the one area of cortex underlying it, and is difficult to extrapolate to a global picture. In addition it can yield false estimations if it is even slightly displaced. The technique does however yield continuous values, and has been used intraoperatively.

5) Stable xenon computed tomography: This technique will be discussed in detail later.

Indirect

1) SPECT: Single photon emission computed tomography, except when it is performed with 133Xe, yields images of the blood flow in various regions of the brain and can detect relative differences in flow, but it does not yield absolute quantitative data.

2) MRI: The utilization of perfusion and diffusion weighted imaging to assess adequacy of flow is spreading rapidly. It has been used to determine whether tissue in the region of a stroke is irreversibly damaged ₁₈, and to assess CBF in vasospasm following SAH ₁₉.

3) Transcranial Doppler (TCD): TCD measures the velocity and characteristics of flow in the major arteries of the brain. While this data is useful and can be interpreted in several ways in various vascular diseases, this technique also cannot measure the actual volume of blood flowing through a particular region of the brain.

4) Cerebral perfusion pressure (CPP): Simultaneous measurement of the mean arterial pressure (MAP) and the intracranial pressure (ICP) gives the pressure head perfusing the brain. Abnormalities in the CPP can give some information regarding the CBF but it is not regional or quantitative.

5) Jugular venous measurements: Estimation of the concentration of oxygen and various metabolites in the jugular vein such as lactate yields information on the adequacy of the cerebral circulation. An increased arteriovenous difference in oxygen concentration could be interpreted as increased oxygen extraction by the brain due to inadequate perfusion. This study is presumed to be a global estimate of the adequacy of CBF, but variations in dural sinus and torcular anatomy can yield unreliable estimates.

6) Regional oxygen saturation (rSO2): The oxygen saturation of blood in the underlying brain (rSO2) can be measured transcutaneously by near-infrared spectroscopy (NIRS). This technique can be used to judge the adequacy of blood flow in the frontal region bilaterally, but has yet to be proved reliable in routine clinical use. It is useful to monitor relative differences and to follow a trend.

7) Direct tissue measurement: Modern technology has developed fine probes that can be inserted into the brain to measure the concentration of oxygen, local temperature and other variables, which indirectly indicate the perfusion. Microdialysis catheters to measure ions, lactate and glucose perform a similar function ₂₀. The problem with this technique again is that the data obtained is only for that particular region of the brain and general extrapolation is difficult.

8) Clinical examination: Neurological deterioration generally occurs when the CBF has dropped to about half of normal, and becomes progressively worse with further decrease. This is a comparatively crude indicator and is often detected too late for successful intervention.

9) Electrophysiology: Neuronal activity is progressively decreased with reduction of blood flow. Electroencephalography (EEG) and evoked potentials (EP) have been used with varying success ₁ to assess the adequacy of blood flow.

History and principles of stable xenon CT CBF measurement

Drayer et al reported the first measurement of CBF using stable xenon in 1978 ₂₁, though the fact that in xenon is a radiodense substance was reported in 1966 ₂₂. The blood-brain partition coefficient for xenon was defined in 1978 ₂₃.

The improvements in CT scan technology in combination with the remarkable developments in the modern computer capabilities have made the process of acquiring CBF data at multiple levels much more user-friendly and economical both in terms of time and in cost of equipment.

The Fick principle of indicator dilution forms the basis for almost all techniques to measure the CBF. Simply put, it states that the concentration of a nonmetabolized, freely diffusible indicator absorbed in a tissue per unit time can be deduced from the difference of concentrations of the indicator in arterial and venous blood. This concentration per unit time in turn depends on the rate at which the indicator is delivered to the tissue (i.e. the blood flow) and the partition coefficient between the blood and the tissue. Thus, knowing the concentration of indicator in the tissue, the time for which the tissue has been exposed to the indicator and its partition coefficient, the blood flow for that particular tissue can be calculated at a given time.

Kelcz et al defined the blood-brain partition coefficient for xenon in 1978 ₂₃. Since xenon is radiodense the CT scan can directly measure its concentration in the brain and determination of the venous concentration is not required. The concentration of xenon in arterial blood can be determined from end-tidal xenon measured by a thermoconductivity analyzer. Therefore with the ability to measure the concentration of xenon in the blood and brain, the time for which xenon has been administered and knowing the blood-brain partition coefficient for xenon, the CBF can be calculated, using the Kety-Schmidt formula modified for xenon.

Technical considerations

The first CBF measurements performed with stable xenon and CT required the subjects to breathe a high concentration of the gas, which is a narcotic anesthetic at over 70%. The study itself was time consuming and the calculations for the CBF at a single level took several hours ₂₄. Developments in CT scanner and computer technology have contributed to make the process dramatically more user-friendly: now the patient has to breathe only 28% xenon, the study takes only a few minutes and the computer yields CBF values at multiple levels in the brain within minutes of the patient being removed from the scanner (Diversified Diagnostic Products, Houston).

The accuracy of the technique has been confirmed in experiments with baboons, where a high degree of correlation was obtained between xenon CT CBF measurement and radiolabeled microsphere technique over a large range of flow values ₂₅.

High concentrations of xenon were required in the early studies due to the fact that xenon is not a very good contrast agent and early scanners required a very high signal-to-noise ratio to reliably detect the contrast enhancement by the gas. This had several undesirable side effects, chief of which were the anesthetic effects of the gas and xenon induced increase in CBF. As the CT detector systems improved, the concentration of xenon required has dropped to the current 28%, which still yields an acceptable signal-to-noise ratio. A potential source of error when measuring CBF is the increased blood flow as a result of xenon inhalation, known as flow activation. In earlier studies with higher concentrations of xenon and longer duration of exposure to the gas this was as much as 30%, but current protocols have been demonstrated to have a much smaller effect ₂₄.

There has also been considerable development in protocols to perform CBF studies with stable xenon, which has paralleled the improvement in CT scan equipment and computers. Two differing types of protocols are currently in use, either only a wash-in study or a wash-in/wash-out study (wash-out studies alone are not satisfactory). With the advent of more sophisticated computers the merits and demerits of these protocols have been extensively studied with computer simulations ₂₆. The net result of all this development has however been beneficial with the concentration of xenon and the time for which it has to be inhaled both having decreased considerably.

Effects of xenon and safety considerations

1) Anesthetic: Xenon at sufficient concentration is a narcotic anesthetic. Cullen et al defined the minimum alveolar concentration of xenon to be 71%, at which half of all patients would be anesthetized. 50% xenon was also found to increase the pain threshold by 15% in volunteers ₂₇. Xenon has several properties of an ideal anesthetic ₂₈ including negligible cardiovascular and pulmonary effects, rapid induction and emergence, sufficient analgesic and hypnotic effect with 30% oxygen and absence of metabolism. The patient experiences a feeling of drunkenness, and the subjective experience of whether the patient found the feeling pleasant or otherwise depends on the individual makeup of the subject and the adequacy of patient preparation. A significant number of patients also feel a sensation similar to paresthesias in the limbs.

2) Cerebral vasodilatation: Xenon is known to cause cerebral vasodilatation, in common with other narcotic anesthetics such as halothane. The resulting ICP increase can be dangerous in patients with decreased intracranial compliance ₂₉. However, an analysis of blood pressure, CPP and ICP during actual CBF measurements showed an average increase in ICP of about 6mmHg and a decrease of CPP of about 9mmHg. This factor needs to be kept in mind when ordering xenon CT CBF studies. The other question is of the accuracy of the measurement in the presence of this xenon induced flow activation. While this was a definite factor in early measurements, it has been estimated that with current protocols the measured CBF is accurate and there is less than 5% effect due to flow activation. ₂₄.

3) Respiration: Respiratory depression with xenon inhalation was reported anecdotally early in the use of xenon as a contrast agent. Xenon has been shown to have some effect on respiration, thought to due to direct action on the medulla. In a large number of xenon CT CBF measurements analyzed by Latchaw et al ₃₀ the incidence of short periods of apnea (greater than 10 seconds) was 3.6%. However most of these were less than15 seconds, and in no case was it prolonged enough to require ventilatory assistance. Even in patients with apnea, a verbal command to breathe was generally obeyed and very few studies were terminated due to apnea.

4) Nausea and vomiting: This was first reported as early as 1955 during xenon narcosis, and the incidence reported by Latchaw et al ₃₀ is 0.4%. This can be dangerous in a patient with an altered sensorium, an unguarded airway and a tight fitting mask, and needs to be kept in mind while establishing study protocols.

5) Heart rate and blood pressure: Xenon has very little effect on the heart rate, myocardial contractility and blood pressure, and is therefore considered a useful anesthetic in situations where cardiovascular stability is needed ₂₈. Since the concentrations used for CBF studies are less than half of the anesthetic dose, no cardiovascular effects should occur.

6) Neurological sequelae: Headache and change in sensorium has been reported in less than 1% of all patients studied ₃₀. The incidence of seizures reported during studies is about 0.2%, and these were mostly in patients who had a prior history of seizures. There are also reports of occasional changes in the neurological status of the patient.

Study protocol

Centers measuring CBF with the stable xenon CT must establish definite protocols for the procedure, both for ventilated and nonventilated patients. Personnel must be familiar with all the equipment and procedures necessary to perform the study successfully and to interpret the resulting data.

Indication

Any patient in whom neurological symptoms or signs are possibly due to abnormal cerebral blood flow (hyperemia or ischemia), follow up of these patients and to monitor the effects of corrective therapy.

Absolute contraindications:

1) Patient cannot tolerate FiO2 of < 60%, which is the highest value that can be administered by the Enhancer (Diversified Diagnostic Products, Houston), for at least 30 minutes. Even if some compromise in ventilation is accepted, the CBF values will not reflect the actual condition of the patient in the ICU, and the chances of deterioration are high.

2) Patient with unguarded airway, depressed mental status and full stomach.

3) Ventilated patients with a peak inspiratory pressure more than 50 cmH2O.

4) Ventilated patients with a tidal volume less than 250 ml cannot currently undergo the study due to design limitations of the Enhancer.

Relative contraindications (increased risk and / or unreliable study)

1) Unstable blood pressure with the resulting risks of moving the patient; the xenon itself has no reported effect on the blood pressure.

2) Uncontrolled ICP or borderline CPP: the effects of xenon at the concentrations used are minimal, but caution is advisable.

3) Restless patients who cannot be adequately sedated: movement causes severe artifacts and may render the study unusable.

4) Patients with external fixation devices or other artifact causing objects within or around the skull.

5) Patients known to have had seizures previously have a slightly increased risk (0.3%) of seizures during the procedure, though this usually occurs in patients with inadequate seizure control.

6) Ventilated patients with a PEEP of more than 10 cmH2O may not tolerate the procedure well.

Management of nonventilated patients

1) An absolute prerequisite of the procedure is that the patient lies still while repeated scans are taken at the chosen levels. Inhalation of xenon can produce sensations of drunkenness or of losing control, as well as sensations similar to paresthesias in the limbs, and the patient is much more likely to lie still if adequately prepared. Allowing the patient to get used to the tightly fitting face mask before entering the CT gantry is recommended. Verbal reassurance during the procedure is also useful.

2) The patient must have been starving for at least 4 hours before the procedure, to minimize the risk of vomiting and aspiration.

3) Sedation may be required if the patient is incapable of lying still for the duration of the study. Sedated patients should be monitored closely due to the increased risk of respiratory depression and aspiration.

4) Ambulatory patients should have a cardiac monitor and intermittent noninvasive blood pressure measurement. Respiration should be monitored with SpO2, and the computer screen provides real time end tidal CO2 (ETCO2) values and respiratory tracing in all patients. More intensive monitoring may be necessary for some patients depending on the clinical condition.

5) Patients with a history of seizures should be on adequate medication.

Management of ventilated patients

Figure 1

Figure 1

A respiratory therapist familiar with the xenon administration equipment (Fig 1) is necessary to safely perform the study. The equipment has a reservoir and carbon dioxide absorption apparatus to permit rebreathing and conserve xenon. This reservoir increases the compliance of the system markedly, requiring a significant increase in tidal volume and other ventilatory parameters. The compensation for this increased compliance can be determined in different ways, and the adequacy of the changes made should be monitored clinically, with the measurements on the ventilator and on a patient monitor. The increased compliance also makes it extremely difficult for a patient taking any spontaneous breaths to trigger the ventilator, and it is recommended that sufficient sedation or paralysis be administered to suppress any patient initiated breathing. The respiratory parameters being monitored to ensure the adequacy of ventilation (such as SpO2 and ETCO2) should closely approximate the status of the patient prior to being transferred for the study. It is also important to remember to reset the ventilator to the original parameters at the end of the study before reconnecting it directly to the patient.

Figure 2

Figure 2

1) The technician has to confirm the FiO2 of the patient in the ICU and set this value on the Enhancer (Fig. 2). The inspiratory concentration of xenon should be set at 28%. The Enhancer has to be in the ready mode before the patient is transferred from the ICU.

2) Before the patient is disconnected from the ventilator in the ICU the following parameters have to be recorded: tidal volume, respiratory rate, peak inspiratory pressure, plateau pressure, PEEP, FiO2 and ETCO2. These values should be obtained with the patient sufficiently sedated to make no respiratory efforts, and if sedation is insufficient the patient should be paralyzed before these values are recorded. The study should be performed with the ventilator in A/C mode and the patient not taking any spontaneous breaths in addition to the set ventilator rate.

Figure 3

Figure 3

Figure 4

Figure 4

3) The minimum monitoring required during this study would be EKG, noninvasive blood pressure and pulse oximetry. ETCO2 and a respiratory tracing are available from the computer (Fig. 3 and 4). In addition, invasive blood pressure and other cardiac function monitoring are often required.

4) On arrival at the CT room the patient should be connected directly to the ventilator there with the same parameters as in the ICU. Once hemodynamic and ventilatory parameters are satisfactory the scout CT is performed.

Figure 5

Figure 5

5) Without moving the patient at all, the ventilator is connected to the Enhancer (Fig. 5), and the Enhancer to the patient. The tidal volume is increased in increments of 100 ml initially and by 50 ml subsequently until the peak inspiratory pressure and the plateau pressure approach the original baseline values. The additional volume required is usually in the order of 200 to 350 ml, depending on the original tidal volume and pressures. Once the SpO2 and ETCO2 values are close to the baseline values, the study may proceed.

6) After the study is completed the Enhancer can be disconnected from the patient. Do not connect the ventilator directly to the patient at this point until the tidal volume and respiratory rate have been reset as otherwise barotrauma may result. The patient may now be transferred back to the ICU.

Xenon is washed out of the body through the lungs at a very rapid rate, so the recovery of the patient from the procedure is almost immediate. This also makes it possible to repeat a study after 20 minutes, if the effect of some intervention needs to be assessed.

Applications of stable xenon CT CBF measurement

CBF measurement with stable xenon CT has been described for numerous neurological and non-neurological diseases. These include head injuries, strokes, subarachnoid hemorrhage, psychiatry and behavioral disorders, migraine and research into the effects of drugs and techniques in anesthesia.

Measurement of CBF proved that hyperventilation in acute injury caused ischemia ₃₁, but it can also be used to demonstrate hyperemia and indicate which patients can safely be ventilated to a lower pCO2 to control raised ICP. Muizelaar et al ₃₂ have demonstrated a technique by which regional cerebral blood volume can be measured using stable xenon CT CBF measurement as part of the process. CBF measurement has also been used to assess autoregulation following trauma, predict outcome ₃₃ and demonstrate ischemia due to post-traumatic vasospasm ₃₄.

There have been multiple uses for xenon CT demonstrated in stroke. Hyperacute CT scans in stroke may appear normal, but xenon CT demonstrated the abnormality in 100% of hemispheric stroke ₃₅, and can show which areas have suffered irreversible damage. This can in turn guide the use of thrombolytic therapy. The CBF study can be continued from the admission CT, saving time and improving early diagnosis. These studies have also been used to predict the risk of herniation and need for surgery ₃₆. Patients with TIAs or otherwise at risk for stroke can have their vascular reserve measured by xenon CT used with acetazolamide to cause vasodilatation (24). This has also been reported in moyamoya disease ³⁷.

Stable xenon CT has is primarily used in subarachnoid hemorrhage for the management of vasospasm ³⁸. It can be used for detection of vasospasm-induced ischemia, either primarily, or triggered by transcranial Doppler or neurological status. Once ischemia is detected, CBF measurement can be used to assess the effectiveness of therapy and guide the intensity.

Choice of measurement technique

All measurement techniques have their advantages and disadvantages as described briefly earlier. Therefore choosing a technique to assess CBF depends on numerous clinical, technical and economic factors.

Direct quantitative techniques do not provide knowledge of the functional state of the brain at the time of measurement. Since in most situations CBF is coupled to metabolism, the abnormality of flow detected in a particular case might be appropriate at that time. For example, if CBF is measured while the patient is in barbiturate coma or hypothermic, the low flow values obtained are appropriate and do not need any intervention, though they would be interpreted as severe ischemia without knowledge of the hypometabolic state.
The indirect techniques that use the measurement of cerebral metabolism to draw conclusions about the CBF suffer from other drawbacks. Jugular venous monitoring for oxygen consumption or other metabolites does not yield regional data, and several authors have questioned whether it can be said to yield truly global information. Direct tissue measurement of oxygen and various metabolites at one point in the brain is difficult to extrapolate globally. Cerebral oximetry can also be used currently mainly to follow trends in individual patients.

Transcranial Doppler measures only the flow velocity, but not the rate (ml per minute) or tissue perfusion (ml per 100 gms per minute), and therefore at best can be used to follow a trend in an individual patient and trigger further investigation or therapy. Neurological examination and electrophysiological testing are also only comparatively crude indicators of adequacy of CBF.

PET is the only test currently available that simultaneously yields quantitative regional blood flow data and the corresponding regional metabolism. The necessity for a local cyclotron to provide the positron emitting isotopes (which have a very short half-life) and other factors contribute to making the cost of the examination too high for routine use. This test also requires transport of the patient.

Stable xenon CT measurement of CBF yields the most accurate regional quantitative data. The measurement requires the patient to be transported to the CT scan, and this is often difficult or not possible in a critically ill patient. The other major drawback as mentioned earlier, is that with the phenomenon of flow-metabolism coupling, the value of low readings obtained in isolation is doubtful, especially if these low values are global in a comatose patient. The CBF may be appropriate for the metabolic activity at that time and attempts to raise the flow will be against autoregulation and may actually cause harm to the patient. Focal abnormalities are of much more value in diagnosis and planning of treatment. Therefore the xenon CT has to be interpreted taking into account the clinical and metabolic status of the patient. The addition of one of the metabolic indicators of CBF adequacy such as jugular venous oxygen saturation is useful, especially in comatose patients.

Thus the choice of technique is dependent on several factors including simplicity of performance, accuracy and clinical reliability, cost, and ease of repetition. There is no ideal method available, and a combination of techniques is probably the best option at the present time. Regardless of the method or combination of methods utilized, protocols must be developed and standardized to ensure the logical integration of this data into patient care.