The cerebral hemodynamic and metabolic changes observed in acute liver failure are divided into five different phases and are presented in Fig. 32.2 [20]. These phases are independent of the grades of encephalopathy. They are characterized as follows:

Phase 1 (Coupled): Demand (CMRO₂) and supply (CBF) are matched as demonstrated by normal AJDO₂ difference. In this phase the ICP is normal, response of CBF to arterial CO₂ changes (CO₂ reactivity) is intact. Cerebral vascular resistance is normal and there is no brain swelling.

Phase 2 (Uncoupled): Demand (CMRO₂) is low but the supply (CBF) in spite of being low, is more than the demand of the brain (relative hyperemia) as demonstrated by narrow AJDO₂. In this phase ICP is still normal and CO₂ reactivity is still intact. However, cerebral vascular resistance is decreased although there is still no brain swelling.

Phase 3 (Uncoupled): Demand (CMRO₂) is low but the supply (CBF) is high (absolute hyperemia) as demonstrated by narrower AJDO₂. CO₂ reactivity, although reduced, is still intact. However, ICP is high, cerebral vascular resistance is further decreased and brain swelling may be present.

Phase 4: Demand (CMRO₂) is low and the supply is low (CBF), as demonstrated by normal AJDO₂. The AJDO₂ is deceptively normal. Supply is low because of very high ICP and the blood flow to the brain is reduced because of extramural compression of the cerebral vessels by cerebral swelling. CO₂ reactivity is markedly reduced. Cerebral vascular resistance is normal and the brain is swollen. These observations confirm that ICP monitoring is extremely important in management of these patients. It is ICP which distinguishes coupled Phase 1 from the seemingly coupled Phase 4, which is close to herniation.

Phase 5 (Impending herniation or herniation): Shows minimal CBF with very high ICP.

It is uncommon to see all five phases in one patient, because (1) the natural disease progression is highly variable. The phases may change from one to the other within hours, or in days, depending on the etiology and other compounding factors like other organ system involvement; (2) impact and timing of liver transplantation; and (3) the phase at which the patient was referred, as the patient may have already passed through some of the phases prior to being transferred to the hospital. The first four phases of cerebral hemodynamic changes are reversible. Therefore, it is crucial to monitor the cerebral hemodynamics in order to institute proper therapy before irreversible changes occur.

CBF can be determined using the intravenous Xe-133 clearance technique [21] and/or stable xenon-CT scan method [22]. The intravenous Xe-133 clearance technique is the preferred method as (1) it can be determined at the bed side, and the risks associated with moving a hemodynamically unstable patient from intensive care unit setting to radiology suite can be eliminated; and (2) this technique can be performed more frequently. The limitation of this technique is that it solely measures CBF. Stable xenon-CT scan determinations of CBF also provide information regarding intracranial pathology which may help to determine the cause of the coma; i.e. cerebral swelling, space occupying lesion, and cerebral hemorrhage. However, to determine CBF by this technique the patient has to be transported to the CT scanner. It may be dangerous to transport a critically ill patient. Therefore, this technique should only be used in select circumstances.

Since acute liver failure is a metabolic disorder, the CBF changes are global. Interestingly, low CBF can even be seen in higher grades of coma (grade 3 and 4) [23–25]. Low CBF most often is not ischemia, as these patients have very low CMRO₂. Usually, low blood flow is coupled and is a good prognostic sign unless the patient is in phase 4 of cerebral hemodynamic and metabolic changes of the brain, where blood flow is low because of massive swelling and intracranial hypertension [20]. High cerebral blood flows, which are usually seen in grade 4 coma, are associated with cerebral swelling and intracranial hypertension and have a poor outcome [23, 26].

AJDO₂ is determined by analyzing the difference in oxygen content of systemic arterial blood and the jugular venous blood. Jugular venous samples can be obtained by inserting a catheter in the internal jugular vein [27, 28]. The position of the catheter tip must be confirmed by lateral roentgenogram of the head and neck. If the tip of the catheter is not at the level of jugular bulb, the blood samples can be contaminated by extra cranial blood and will give false results. Blood samples can also be contaminated if the blood samples are drawn rapidly.

AJDO₂ is a good bed side clinical tool to determine (1) changes in global cerebral blood flow and (2) the adequacy of CBF in relation to cerebral metabolic rate of oxygen consumption (CMRO₂). This is only applicable if the CMRO₂ remains unchanged. Since AJDO₂ = CMRO₂/CBF, monitoring and managing the AJDO₂ can eliminate repeated determination of CBF [29].

At a PaCO₂ of 40 mmHg, the normal AJDO₂ range is 5.1–8.3 vol.%. The normal AJDO₂ range for other levels of PaCO₂ can be calculated by increasing the AJDO₂ range by 3 % for a decrease of each mm Hg in PaCO₂ and vice-versa. A normal AJDO₂ indicates that the supply of CBF is closely coupled with the cerebral metabolic demand of the oxygen. AJDO₂ below the normal range demonstrates uncoupling, showing cerebral hyperperfusion relative to CMRO₂ (cerebral luxury perfusion), whereas values above the normal range, which also indicate uncoupling, are consistent with cerebral hypoperfusion relative to CMRO₂.

In acute liver failure, very few patients have normal AJDO₂ (coupling). A majority of patients have low AJDO₂ (uncoupling) which is most likely caused by depressed cerebral metabolism. Moreover, if the AJDO₂ and CBF both are very low, this may indicate minimal extraction of oxygen and irreversible brain injury as seen in phase 5 [20, 29].

CMRO₂ is calculated as product of CBF × AJDO₂/100. In acute liver failure CMRO₂ is depressed even in early phases of the disease. It is usually less than 50 % of normal [23, 24]. In some instances it can even be as low as 25 % of normal, and still the patients may recover without apparent neurological deficit. Low CMRO₂ may very well be an indication of the depressed active and basal metabolism of the brain. Unlike in head injury [29], CMRO₂ is not a predictor of outcome in acute liver failure [23, 24].

CO₂ Reactivity is the response of CBF to changes in arterial CO₂ tension. The normal response is 3 % CBF change per mm Hg change in PaCO₂. It is a relatively noninvasive technique. The response of CBF changes to CO₂ alterations provides a tool to predict the efficacy of hypo or hyperventilation from both the therapeutic and prognostic points of view [20]. In determining CO₂ reactivity certain precautions must be observed: (1) Systemic mean arterial pressure must be maintained (2) An observable change in CBF requires a minimum of 5 mmHg alteration in CO₂ (3) Prior to determining CO₂ reactivity, a baseline CBF needs to be assessed in order to ensure that a patient with low CBF is not hyperventilated and that a patient with high CBF is not hypoventilated.

CO₂ reactivity is well preserved in early phases (phase 1, 2) for hyper and hypoventilation [17, 20, 30]. However, in phase 3 and 4, although the vascular response to changes in PaCO₂ is preserved for hyperventilation, it is reduced for hypoventilation. This change in response suggests that as patients move towards higher phases, the vessels become more dilated and lose their vasomotor tone for further dilatation. It should be noted that hyperventilation can still be an effective therapy for reducing CBF in late phases as shown in Fig. 32.3.

CVR is calculated as (MAP − ICP)/CBF (normal = 3.6 mmHg/ml/100 g) CVR is normal in phase 1. When the cerebral vessels dilate the vascular resistance reduces as in phase 2. In phase 3, cerebral vascular resistance reduces further, however, increase in cerebral blood volume causes a rise in ICP [19]. In phase 4, because of cerebral swelling, CVR is increased to a point that the CBF is reduced [20].

ICP is one of the most important parameters to be monitored in acute liver failure. This is because 40 % of mortalities in this population are from brain herniation [31]. An ICP monitoring device should be inserted as soon as patient is intubated. It is challenging to insert an ICP monitoring device because of the coagulopathy associated with liver failure.

Prior to inserting an ICP monitor (1) computed tomography (CT scan) of the head should be performed in order to rule out any adverse intracranial pathology; (2) coagulation status must be optimized after evaluation by either following the conventional coagulation parameters (prothrombin time, partial thromboplastin time, platelet count, and fibrinogen degradation products) and/or thromboelastography (reaction time >6 min, alpha angle >50°, maximum amplitude >50 mm, and whole blood clot lysis time >300 min) [32, 33]; and (3) the airway must be secured.

There are three different sites for inserting intracranial pressure monitoring devices: epidural, subdural, and parenchymal. Out of these three sites, epidural placement is the most popular because of lowest incidence of hemorrhage (<4 %), and infection (<1 %) [34, 35]. Parenchymal placement has the highest incidence of both hemorrhage (13 %) and infection (4 %). While, the drawback of an epidural monitoring device is its unreliability in obtaining absolute ICP values, it does measure changes in ICP consistently.

The etiology of intracranial hypertension in ALF is unknown. However, it appears that toxins lead to cerebral vasodilation, which increases cerebral blood volume, followed by an increase in intracranial pressure, and increased capillary leak (vasogenic cerebral edema). Cerebral edema further compounds the intracranial pressure. Initially, the ICP increase precedes the development of cerebral swelling. Later, the increase in ICP is compounded by cerebral swelling. An increase in ICP is seen when the CBF is high and cerebral vascular resistance is low indicating increase in cerebral blood volume [36]. The precipitating factors for cerebral vasodilation and increase in cerebral blood volume are: (1) anemia and episodes of hypotension [37]; (2) hypoxia caused by pulmonary congestion; and (3) lactic acidosis [24].

It is interesting that, in acute liver failure, papilledema has not been reported to be observed in conjunction with intracranial hypertension. In acute liver failure patients, high ICP’s (>40 mm H₂O) are tolerated much better in comparison to patients with head injury [38]. Once intracranial hypertension sets in, it is difficult to control unless and until the diseased liver is replaced by a normal functioning liver. With every new episode of intracranial hypertension, it becomes increasingly difficult to bring the ICP to its baseline value and the subsequent fluctuations are more pronounced. Unequal and dilated pupils are seen with extremely high ICP’s, but, fortunately, these can be reversed by aggressive medical management.

It is very difficult to transport these patients to a computed tomography (CT) scanner, as they may be (1) hemodynamically unstable, (2) on assisted ventilation, and (3) require multiple vasopressor infusions. However, it is essential to obtain a CT scan of the head in order to: (1) to establish baseline condition of the brain on admission to intensive care unit; (2) prior to insertion of ICP monitor; and (3) in any acute change in neurological status, particularly a sharp rise in ICP.

It is interesting that cerebral swelling appears mostly in phase 3 and 4 and is associated with very high blood volume [20, 23, 24, 37, 39, 40]. Presence of cerebral swelling in conjunction with high CBF indicates a poor prognosis [41]. Likewise loss of the gray–white matter interface signifies a poor prognosis. During these phases, if cardiovascular status becomes very unstable, liver transplantation may become contra-indicated.

CBFV can be measured by using transcranial Doppler ultrasonography (TCD) . TCD consists of a pulsed Doppler instrument operating at a low frequency (2 MHz), coupled with a computer that performs Fourier transformation of the complex waveform data for real-time spectral display. This technique is noninvasive, can be used at the bedside to determine CBFV in all the major intracranial vessels [42, 43].

TCD variables (Fig. 32.4) include systolic cerebral blood flow velocity (SFV), diastolic cerebral blood flow velocity (DFV), mean cerebral blood flow velocity (MFV), or [(SFV + DFV)/2]), and pulsatility index (PI), or [(SFV − DFV)/MFV]). Every variable of the waveform is important, however, diastolic cerebral blood flow velocity determines the period during which the brain receives its blood flow. Pulsatility index is believed to reflect cerebrovascular resistance and thus shows exponential correlation with ICP. In TCD monitoring, the observed trends over time are more revealing than any single set of measurements [44].

TCD can be used to estimate: (1) cerebral blood flow, (2) ICP, (3) CO₂ reactivity, and (4) effect of various therapeutic modalities on CBF, ICP, and CO₂ reactivity [45, 46]. Figure 32.2, shows sequential changes in TCD patterns of a patient with acute hepatic failure [47]. Phase 1: Decreased systolic cerebral blood flow velocity (normal = 61 cm/s), normal pulsatility index (normal = 0.90 ± 0.24). Phase 2: Normal systolic cerebral blood flow velocity with low pulsatility index. Phase 3: Increased systolic cerebral blood flow velocity with low pulsatility index. Phase 4: Decreased systolic cerebral blood flow velocity, with high pulsatility index. Phase 5: Negative diastolic flow or retrograde CBF (brain death). Figure 32.3 shows the effects of hyperventilation on TCD patterns of a patient with acute liver failure. It shows that as the patient is hyperventilated and the PaCO₂ is decreased from 44 to 27 mmHg the mean cerebral blood flow velocity reduces from 145 to 98 cm/s and at the same time the CBF determined by CT-xenon method also shows reduction from 80 to 60 ml/100 g. This indicates that; (1) the CO₂ reactivity demonstrated by TCD compares very well with the CT-xenon CBF determination method, (2) CO₂ reactivity is preserved even at very high CBF, though 50 % of normal (CO₂ reactivity 1.4 %) [48–50] demonstrating that hyperventilation is an effective therapeutic modality in reducing CBF. The diminished CO₂ reactivity is secondary to the vessels developing vasoparalysis. Figure 32.5 shows the effects of barbiturates on TCD patterns of a patient with acute liver failure. In a patient with high intracranial pressure (ICP = 34 mmHg, PI = 1.92) when barbiturates are administered, the ICP (28 mmHg) and PI (1.27) are reduced, provided the mean arterial pressure (MAP) is maintained at normal level (>80 mmHg). If the MAP reduces as an effect of barbiturates then ICP may further increase and cerebral perfusion may reduce to dangerously low levels. MAP should be maintained by administering vasopressors, alone or in conjunction with increased fluid volume