See “Expanded Neurologic Assessment in Pediatric Acute Liver Failure: An Important Initial Step” by Squires on page 394.
Neurologic complications of pediatric acute liver failure (ALF) are a major determinant of outcome (1). Pediatric ALF results in death or need for liver transplantation in nearly 50% of cases (2). Among children with ALF who develop hepatic encephalopathy (HE), fewer than 20% with HE survive without liver transplantation (3). Early detection of neurologic injury in children with ALF may enable interventions to prevent neurologic morbidity and improve outcomes, before neurologic injury becomes irreversible.
Pediatric ALF is defined as a hepatic-based coagulopathy that is not corrected by parenteral administration of vitamin K; lack of evidence of chronic liver disease; and prothrombin time (PT) ≥20 seconds; or international normalized ratio (INR) ≥2, based on the consensus of the pediatric (P)ALF Study Group (1,2,4,5). In the presence of HE, the thresholds for PT (15) and INR (1.5) may be reduced. More important, HE is not required for entry into the PALF study in patients with more severe coagulopathy because of consensus among investigators that HE is difficult to diagnose in children.
Although the presence and severity of HE affects outcome, and may support the diagnosis of ALF in children, the criteria for grading degrees of encephalopathy in children are challenging. HE in adults is graded on a continuum from no or minimal evidence of neurologic dysfunction to coma (6). Data are lacking on the use of other modalities, including EEG and neuroimaging, to detect or grade the severity of PALF (7,8). The incidence of nonconvulsive seizures (NCS) in children with HE is not known, although both pediatric and adult studies indicate a high frequency of NCS in critically ill patients (9–12). The mechanisms linking ALF and HE are not fully understood, and therapeutic targets have not been identified or validated (13).
We developed a clinical pathway for the neurologic evaluation of children with ALF upon admission to the pediatric intensive care unit (PICU), with the long-term goal of improving the early detection of neurologic compromise. This pathway included a combination of brain imaging, EEG, and serial neurologic examinations. We performed a retrospective analysis of the association between these measures, clinical signs of HE and survival after ALF. This is an initial step toward improving our ability to identify neurologic deterioration in children with ALF.
Patients and Clinical Data
In 2008, the neurocritical care service, with the Divisions of Hepatology, Gastroenterology, and Nutrition and Critical Care at Children's Memorial Hospital, established a set of clinical guidelines for the neurologic evaluation of children admitted to the PICU with ALF (Fig. 1). The definition of ALF used at our center is that used by the PALF study (2). These patients were evaluated by a neurologist at the time of ICU admission and thereafter twice per day. The HE score was discussed with the attending hepatologist and assigned together. EEG and brain imaging (magnetic resonance imaging [MRI] or computed tomography [CT]) were obtained at the time of admission or later, if the initial neurologic examination was normal. All of the patients were treated with L-carnitine (100 mg · kg−1 · day−1) (14–16). Normothermia was actively maintained to keep temperatures 36° to 37°C (17,18). The purpose of this approach was to improve earlier recognition of neurologic compromise and therefore to proceed to intracranial pressure (ICP)–therapy, if clinically indicated. No changes were made in the clinical management of ALF, and laboratory studies were obtained according to routine clinical practice.
The demographic information, EEG, neuroimaging, and laboratory results and clinical course of all of the children admitted to the Children's Memorial Hospital PICU who met criteria for the PALF study from January 2008 to August 2011 were collected. In the original design of this clinical pathway, neonates were not included and no data from the neonatal ICU were collected. This retrospective study was approved by the local institutional review board.
Medical Management of ALF
ALF was defined using the criteria established by the PALF investigators (2), as summarized above, consistent with the requirements for eligibility for the PALF study. The treatment followed standard critical care guidelines for the management of pediatric ALF (19). The ICP-directed therapy and placement of ICP monitors were initiated for patients with stage III or stage IV encephalopathy who were eligible for liver transplant (LT). The criteria for transplant included ongoing liver synthetic failure and the presence of at least stage II HE without expectation of improvement with available therapies.
Classification of ALF and Collection of Laboratory Data
The clinical data were abstracted from the medical record. Causes of ALF were classified as indeterminate, acetaminophen toxicity, sepsis/ischemia-reperfusion injury, metabolic disease, hematology-oncology related, or other (2,4). The abstracted laboratory data included serum aspartate aminotransferase, alanine transaminase, PT, partial thromboplastin time, INR, ammonia, lactate, total bilirubin, creatinine, pH, and hemoglobin level, and was recorded until ALF resolved, LT, or death.
Neurologic Assessment and Monitoring
The clinical neurologic status and the staging of HE of each patient were assessed at least daily by a member of the neurocritical care service. The assessment and grading of HE was classified according to published criteria for children (2–4,6). Continuous (c)EEG or routine (r)EEG monitoring was obtained for all of the patients during the hospital stay. The selection of cEEG was made to increase the likelihood of detection of NCS, which may require up to 24 hours of EEG monitoring (10).
The digital EEG recordings were performed using the international 10–20 system of electrode placement according to the American Clinical Neurophysiology Society guidelines with electrocardiogram monitoring (XLTEK, Neuroworks System, Oakville, Ontario, Canada). Recordings were performed for a minimum of 4 hours. The length of study was based on ongoing EEG results and clinical status of the patient.
In accordance with published criteria, status epilepticus was defined as a seizure with or without clinical accompaniment lasting 10 minutes or longer, or recurrent seizures without interval return to baseline lasting 10 minutes or longer (20). Electrographic seizures were diagnosed when paroxysmal EEG patterns with a discrete onset, offset, and evolution were present. Periodic lateralized epileptiform discharges (PLEDS) were not considered an ictal pattern. NCS were defined as electrographic seizures without clear clinical correlate in accordance with established criteria (21). The nonconvulsive status epilepticus was defined as a prolonged NCS lasting longer than 15 minutes of frequent NCS during 1 hour without return to baseline.
All EEG tracings were reviewed in real time at a minimum of 4-hour intervals, by a pediatric epileptologist with American Board of Psychiatry and Neurology certification in neurophysiology, and were coded by standard criteria for ICU monitoring proposed by the American Clinical Neurophysiology Society (22). EEG abnormalities were classified as background abnormalities (focal and generalized slowing, attenuation, reduced organization, or complexity), interictal epileptiform abnormalities (focal spikes and/or sharp waves, generalized discharges), specific patterns (periodic lateralized epileptiform discharges, suppression bursts, triphasic waves, frontal intermittent rhythmic delta activity, and electrocerebral silence), and ictal abnormalities (electroclinical seizures, electrographic only seizures).
Neuroimaging, using head CT or MRI, was obtained within the first 48 hours of admission in patients in whom neurologic compromise was suspected. The use of hyperosmolar therapies (mannitol or hypertonic saline), plasmapheresis, and placement of ICP monitoring device was recorded. Outcomes were defined at the time of death or hospital discharge, as survival without LT, survival with LT, and death without LT.
For descriptive purposes, continuous variables are reported as mean ± standard error of the mean and discrete variables as n (%). The HE scores are reported medians with interquartile range (IQR). Differences among groups were tested by 1-way analysis of variance for continuous variables or by using the Fisher exact test or the Mantel-Haenszel χ2 test for categorical variables. Nonparametric data were analyzed by analysis of variance with Dunn correction. The differences were considered statistically significant if P < 0.05. The data were analyzed with SPSS (version 11.0.1, SPSS Inc, Chicago, IL).
A total of 19 patients with ALF were admitted to the PICU between 2008 and 2011 and were managed following this approach. The demographic data are shown in Table 1. The mean age was 6.8 ± 1.5 years, with 13 girls (68%) and 6 boys (32%). In the majority of cases (74%) the etiology was indeterminate. The remaining cases included 2 (11%) acetaminophen toxicity, 1 (5%) hemophagocytic lymphohistiocytosis (HLH), 1 (5%) ischemic, and 1 (5%) autoimmune. A group of 10 patients (53%) survived to discharge without liver transplant (LT), 5 (26%) survived to discharge after LT, and 4 (21%) died. None of the patients who received an LT died.
The average hospital length of stay in the patients who survived without LT was 12.0 ± 1.0 days. In the group who survived with LT, length of stay was 20.0 ± 5.0 and in the patients who expired, 14.0 ± 10.0. None of the patients who survived without LT required mechanical ventilation (Table 2). Four patients were enrolled into a clinical trial of n-acetylcysteine and were treated with n-acetylcysteine or placebo as part of that trial. Of the patients who died, 50% required continuous venovenous hemodiafiltration, and all were treated with pressors.
Neuroimaging (CT or MRI) was performed on 15 patients (Table 3), on average 1 ± 1 day following admission to the PICU (Table 2), and was abnormal in only 2 (13%) cases. Among the 10 survivors without LT, only 1 of 7 MRIs (14%) was abnormal, showing restricted diffusion (Fig. 2A, B). In this case (case 1, Table 2), the patient had electrographic and clinical seizures. Among the 5 patients who underwent LT, neuroimaging (CT, MRI, or both) was obtained on admission to the ICU in 4 and was normal in every case. In the 4 patients who did not survive, acute neuroimaging on admission to the PICU was normal in 3 cases (75%). In the fourth case (HLH with respiratory failure), both CT and MRI showed ex vacuo ventricular dilatation without evidence of increased intracranial pressure.
EEGs (1 routine, 17 continuous) were obtained in 18 patients, on average within 1 ± 1 (median ± IQR) day of PICU admission (Table 3, additional detail can be found in the online-only supplemental table, http://links.lww.com/MPG/A294). The most common EEG recording abnormalities were in the background rhythms, which showed slowing or epileptiform discharges in 10 studies (59%) (Fig. 2C, D). In contrast, clinical seizures were uncommon and occurred in only 2 of the 19 cases (11%). One patient had electrographic seizures alone (Fig. 2E).
Among the 10 patients who survived without LT, only 1 patient (10%) had clinical seizures. Five (50%) had normal EEGs. In 3 (30%) the EEG showed mild slowing and in 2 (20%) epileptiform discharges were present. EEGs were obtained in all 5 patients who underwent LT. Three of studies were abnormal, with moderate and severe slowing and attenuation. These studies were obtained on the day of ICU admission. EEGs were obtained in 3 of the 4 patients who died without LT, within 2 ± 1 days of ICU admission. In all 3 cases (100%), the EEGs were profoundly abnormal, including electrographic seizures, triphasic waves, and an isoelectric background. The fourth patient died before an EEG could be completed.
There was no association between abnormal EEG recordings and neuroimaging findings in any of the 3 outcome groups (P > 0.05, Fisher exact test). Among survivors, both EEGs and imaging were obtained in 7 cases, all within a maximum of 72-hour admission to the PICU. In 4 of 7 cases (57%) the EEG was abnormal; however, only 1 of these 7 cases (14%) had any abnormal imaging findings (Fig. 2A, B). There was also no association between abnormal EEGs and neuroimaging in the LT group and those patients who died. Head CT or MRIs were obtained in 80% of the patients who underwent LT and all were normal. In contrast, the majority of EEGs (80%) in these patients were abnormal and were obtained within the same time frame (within 24 hours of ICU admission). Although all of the EEGs obtained in the patients who died were abnormal, only 1 head imaging study showed any abnormality (ventricular enlargement), and this was a patient with HLH, multiple organ failure, and cardiac arrests.
Among the 10 patients who survived without LT, the HE score (median ± IQR) on admission was 1 ± 1 (Table 3). This was significantly different (P < 0.05) from those who died (3 ± 1), but not from those who survived with LT (2 ± 1). The peak HE score among survivors (1 ± 1) was significantly different (P < 0.01) from those who died (4 ± 0), but not from those who survived with LT (2 ± 1). Notably, of the 8 patients with an admission HE score of ≤1, the EEG was normal or showed mild slowing in 6 cases (75%).
There were significant differences in early (obtained within the first 48 hours of admission to the PICU) laboratory measures of liver injury between the patients who died and those who survived with or without LT (Fig. 3). The data are shown for individual days in the figure, but for these comparisons laboratory values for the first 48 hours after admission to the PICU were combined. Lactate (mean ± standard error of the mean, mEq/L) in the patients who died (9.9 ± 2.9) was significantly increased (P < 0.01) compared with patients who survived without LT (2.4 ± 0.5) (Fig. 3A). There was no significant difference between patients who survived without LT and those who underwent LT (2.3 ± 0.6). The aspartate aminotransferase levels (Fig. 3B) in patients who died (1019 ± 303 IU/L) were significantly lower compared with both patients who survived without LT (4417 ± 1605, P < 0.05), and those who underwent LT (5367 ± 1305, P < 0.01). A similar pattern was present for alanine transaminase values (Fig. 3C), which were significantly lower in the patients who died compared with each group of survivors. Values for INR (Fig. 3D) were significantly increased in patients who died (4.5 ± 0.8), compared with patients who underwent LT (2.0 ± 0.2, P < 0.05), but not with those who survived without LT (2.4 ± 0.2). There were no significant intergroup differences for ammonia or total bilirubin levels in the first 48 hours of ICU admission (Fig. 3E, F). Some studies (4,5) have derived a liver injury unit (LIU) score to predict outcome in children with ALF. We used the peak values for INR, ammonia, and total bilirubin to calculate the LIU score on our patients as described (4,5). The LIU score in patients who survived without LT (173 ± 20) was significantly lower (P < 0.05) compared with those who died (387 ± 62), but not with those who underwent LT (290 ± 47).
We examined the relation between abnormalities of EEG recording and outcome. For this analysis we combined the patients with normal EEGs with those with mild slowing, a total of 11 studies. We grouped together patients with EEGs showing moderate or severe slowing, epileptiform discharges, suppressed background, or electrographic seizures (n = 7). These patients were significantly more likely to die or require LT (P < 0.05, the Fisher exact test) (Table 4).
We next examined the combination of the HE score on admission with the results of the first EEG with respect to outcome (Table 4). A total of 16 patients were admitted with an HE score of ≤2. Of these, 11 patients had an EEG which was normal or showed mild slowing. Nine of these patients (82%) survived without LT. Five patients with admission HE scores ≤2 had abnormal EEGs on admission and 4 (80%) of these patients died or required LT (P < 0.05, Fisher exact test). There was no effect of etiology of ALF on this association. When nonindeterminate etiologies were combined into 1 group, analysis of the impact of etiology on the association of EEG and HE score with outcome showed no significance by the Mantel-Haenszel χ2 test. Similarly patients with either an admission HE score ≤2 or LIU score < 222, combined with a normal or mildly abnormal EEG, were more likely to survive without LT.
Neurologic morbidity is a major determinant of outcome following PALF, but early recognition of signs of neurologic compromise may be difficult in young children, including infants <1 year old who comprise 23% of cases of pediatric ALF (2). This study adds 2 key findings. First, children with a moderate abnormality of EEG background waveforms on admission to the PICU were significantly more likely to require LT or to die. Conversely, children with an HE score ≤2 on admission to the PICU and a normal or only mildly abnormal EEG were significantly more likely to survive without needing LT. These findings are an initial step toward distinguishing patients with ALF who may recover spontaneously from those who will require LT.
The use of HE scoring in children is adapted from that used for adults (23). For children and neonates in particular, distinguishing normal neurologic function and stages I or II of HE may not be reliable because the assignment of the early stages of HE depends on assessment of attention, memory, and affect (6). Accordingly, there is a need for improvement in neurologic assessment in infant and children with ALF based on a more sensitive clinical examination, neurophysiologic monitoring, or serum biomarkers. Our data suggest that the EEG is a promising tool to support the clinical examination as an early measure of declining neurologic function. We found that the EEG was a sensitive measure of neurologic dysfunction in these patients. Of the 16 patients with an HE score on admission of ≤2, only 5 (31%) were abnormal. In adults with hepatic failure, conventional EEGs have been shown to demonstrate latent HE in clinically normal patients (24). The most consistent finding in adults with HE is slowing of the mean dominant frequency (25). We found that EEG abnormalities, most frequently generalized slowing or epileptiform discharges, were a common finding among patients with both early and late stages of HE. In the PALF study (2) among the cases of nonacetaminophen-related ALF, 79 did not have clinically detected HE and 16 (20%) of these cases died or required LT. Our data suggest that the EEG may be a helpful adjunctive tool in assessing the presence and degree of HE in such children for whom early neurologic compromise may be difficult to detect. Notably, we performed EEGs within 24 hours of admission to the PICU so these results reflect data available early in the ICU course of ALF. In our clinical practice, we use the EEG background and changes in the background to assess risk for poor outcome and to assess changes in neurologic function, which may require escalation of neurologic care including more aggressive temperature control and early use of ICP-directed therapy.
Neither clinical nor electrographic seizures were common in our series, occurring in only 16% of cases. In adults, seizures have been reported in up to one-third of patients with ALF (26). The frequency of nonconvulsive (electrographic) seizures in the pediatric ALF population has not been studied in detail; however, convulsive or NCS are common complications of critical illness, occurring in up to 50% of adults undergoing continuous EEG (cEEG) monitoring (10,27), often with nonconvulsive status epilepticus (21). The pediatric data indicate that among children with structural brain injuries, prior in-hospital convulsive seizures or interictal EEG abnormalities, electrographic seizures occur in approximately one-third (28). Similarly, in children with acute encephalopathy, electrographic seizures occurred in 46% (12), for which the main risk factor was young age. Our finding of a low frequency of seizures despite the use of cEEG monitoring was unexpected.
Neuroimaging was performed in the majority of patients in this study, including patients with progressive decline in neurologic function, seizures, and abnormal EEG background. The lack of pathologic findings in these studies was striking. Studies in adults (29) have suggested that the ALF is associated with MRI evidence of cytotoxic edema based on the reduction in apparent diffusion coefficient. The reduction in the ratio of choline to creatine has been proposed as a marker for increased risk for mortality in ALF (8). Abnormal signal on fluid-attenuated inversion recovery and diffusion-weighted sequences, mostly in the thalami and deep brain structures were reported to be associated with plasma ammonia level (and indirectly with clinical outcome) in patients with HE following ALF (30). The majority of patients in these series are adults, and there are limited pediatric data.
In our study only 1 patient who received ICP monitoring had evidence of intracranial hypertension with ICPs persistently in the mid-20s and had no imaging evidence of cerebral edema. This discrepancy is similar to those of other studies in adults (31) with HE, in which conventional MRI techniques have not consistently shown T2-weighted signal intensity abnormalities indicating the presence of cerebral edema (32). Alternate magnetic resonance techniques, such as magnetization transfer imaging, functional MRI, magnetic resonance spectroscopy, diffusion tensor imaging, and diffusion-weighted imaging have also been applied to evaluate brain water content and cerebral metabolic changes in adults with hepatic failure, and may be more sensitive to these changes than conventional sequences (32–34).
Previous studies have used a combination of laboratory values to predict outcome in ALF (5). Certain laboratory markers have been described to be predictive of death or need for LT in patients with PALF, including total bilirubin, PT/INR, and ammonia (4,5). In adults, persistently elevated arterial blood lactate levels despite adequate fluid resuscitation have been shown to be indicators of poor prognosis (19,35). We also found significant differences in LIU score, early lactate, and transaminases among the 3 groups. Whether these laboratory values provide additional discriminative value to HE score and EEG findings will require a larger, prospective study. The reduction in transaminases in the patients who died was unexpected and may reflect greater progression of liver injury at the time of transfer to our hospital.
The distribution of etiologies of ALF in this series is largely comparable with the distribution in the series of 348 cases from the PALF group (2) and suggests these results are applicable to other centers. In our series, 74% cases were indeterminate and 16% died, compared with 49% and 9%, respectively, in the PALF study group; however, only 26% of cases in our series were ≤3 years, compared with 36% in the PALF group study. As has been noted in previous PALF studies (4), the validity of admission measures of neurologic function in relation to severity of ALF is confounded by variability in the timing of presentation or referral of these patients. It is unlikely that any single factor or combination of factors will serve as reliable predictive tools for the multiple causes of ALF in children (6,36).
Neurologic morbidity is a major determinant of outcomes in pediatric ALF. Early identification of declining neurologic function would allow for earlier therapeutic interventions that may help minimize mortality. Larger, multicenter studies will be needed in order to determine biochemical, neurophysiologic, imaging, or clinical signs of neurologic injury, which identify pediatric patients with ALF who are at risk for significant neurologic injury and death. In this study, we found moderate or severe background signal abnormalities present in the EEG recording on admission to the PICU aid in identifying children at risk for poor outcome or requiring LT. These findings are an initial step toward identifying patients with ALF who may recover spontaneously from those who will require LT.
1. Ciocca M, Ramonet M, Cuarterolo M, et al. Prognostic factors in paediatric acute liver failure
. Arch Dis Child
2. Squires RH Jr, Shneider BL, Bucuvalas J, et al. Acute liver failure
in children: the first 348 patients in the pediatric acute liver failure
study group. J Pediatr
3. Rivera-Penera T, Moreno J, Skaff C, et al. Delayed encephalopathy
in fulminant hepatic failure in the pediatric population and the role of liver transplantation. J Pediatr Gastroenterol Nutr
4. Lu BR, Gralla J, Liu E, et al. Evaluation of a scoring system for assessing prognosis in pediatric acute liver failure
. Clin Gastroenterol Hepatol
5. Liu E, MacKenzie T, Dobyns EL, et al. Characterization of acute liver failure
and development of a continuous risk of death staging system in children. J Hepatol
6. Squires RH Jr. Acute liver failure
in children. Semin Liver Dis
7. Amodio P, Gatta A. Neurophysiological investigation of hepatic encephalopathy
. Metabolic Brain Dis
8. Saxena N, Bhatia M, Joshi YK, et al. Electrophysiological and neuropsychological tests for the diagnosis of subclinical hepatic encephalopathy
and prediction of overt encephalopathy
9. Jette N, Claassen J, Emerson RG, et al. Frequency and predictors of nonconvulsive seizures during continuous electroencephalographic monitoring in critically ill children. Arch Neurol
10. Claassen J, Mayer SA, Kowalski RG, et al. Detection of electrographic seizures with continuous EEG monitoring in critically ill patients. Neurology
11. Saengpattrachai M, Sharma R, Hunjan A, et al. Nonconvulsive seizures in the pediatric intensive care unit
: etiology, EEG, and brain imaging findings. Epilepsia
12. Abend NS, Gutierrez-Colina AM, Topjian AA, et al. Nonconvulsive seizures are common in critically ill children. Neurology
13. Blei AT. Brain edema in acute liver failure
: can it be prevented? Can it be treated? J Hepatol
14. Wainwright MS, Kohli R, Whitington PF, et al. Carnitine treatment inhibits increases in cerebral carnitine esters and glutamate detected by mass spectrometry after hypoxia-ischemia in newborn rats. Stroke
15. Rebouche CJ. Kinetics, pharmacokinetics, and regulation of L-carnitine and acetyl-L-carnitine metabolism. Ann NY Acad Sci
16. Binienda ZK. Neuroprotective effects of L-carnitine in induced mitochondrial dysfunction. Ann NY Acad Sci
17. Diringer MN, Reaven NL, Funk SE, et al. Elevated body temperature independently contributes to increased length of stay in neurologic intensive care unit
patients. Crit Care Med
18. Badjatia N. Hyperthermia and fever control in brain injury. Crit Care Med
19. Bernal W. Lactate is important in determining prognosis in acute liver failure
. J Hepatol
20. Treiman DM. Electroclinical features of status epilepticus. J Clin Neurophysiol
21. Young GB, Jordan KG, Doig GS. An assessment of nonconvulsive seizures in the intensive care unit
using continuous EEG monitoring: an investigation of variables associated with mortality. Neurology
22. Hirsch LJ, Brenner RP, Drislane FW, et al. The ACNS subcommittee on research terminology for continuous EEG monitoring: proposed standardized terminology for rhythmic and periodic EEG patterns encountered in critically ill patients. J Clin Neurophysiol
23. Polson J, Lee WM. AAASLD position paper: the management of acute liver failure
24. Van der Rijt CC, Schalm SW, De Groot GH, et al. Objective measurement of hepatic encephalopathy
by means of automated EEG analysis. Electroencephalogr Clin Neurophysiol
25. Bajaj JS. Current and future diagnosis of hepatic encephalopathy
. Metab Brain Dis
26. Ellis AJ, Wendon JA, Williams R. Subclinical seizure activity and prophylactic phenytoin infusion in acute liver failure
: a controlled clinical trial. Hepatology
27. Vespa PM, Nenov V, Nuwer MR. Continuous EEG monitoring in the intensive care unit
: early findings and clinical efficacy. J Clin Neurophysiol
28. McCoy B, Sharma R, Ochi A, et al. Predictors of nonconvulsive seizures among critically ill children. Epilepsia
29. Ranjan P, Mishra AM, Kale R, et al. Cytotoxic edema is responsible for raised intracranial pressure in fulminant hepatic failure: in vivo demonstration using diffusion-weighted MRI in human subjects. Metab Brain Dis
30. McKinney AM, Lohman BD, Sarikaya B, et al. Acute hepatic encephalopathy
: diffusion-weighted and fluid-attenuated inversion recovery findings, and correlation with plasma ammonia level and clinical outcome. Am J Neuroradiol
31. D’Agostino D, Diaz S, Sanchez MC, et al. Management and prognosis of acute liver failure
in children. Curr Gastroenterol Rep
32. Rovira A, Alonso J, Cordoba J. MR imaging findings in hepatic encephalopathy
. Am J Neuroradiol
33. McPhail MJ, Patel NR, Taylor-Robinson SD. Brain imaging and hepatic encephalopathy
. Clin Liver Dis
34. Zhang LJ, Zhong J, Lu GM. Multimodality MR imaging findings of low-grade brain edema in hepatic encephalopathy
. Am J Neuroradiol
35. Macquillan GC, Seyam MS, Nightingale P, et al. Blood lactate but not serum phosphate levels can predict patient outcome in fulminant hepatic failure. Liver Transpl
36. Dhawan A. Etiology and prognosis of acute liver failure
in children. Liver Transpl
Keywords:© 2014 by European Society for Pediatric Gastroenterology, Hepatology, and Nutrition and North American Society for Pediatric Gastroenterology,
acute liver failure; electroencephalogram; encephalopathy; intensive care unit