Acute liver failure (ALF) is a life-threatening condition associated with significant morbidity and mortality. Raised intracranial tension because of cerebral edema (CE) and increased cerebral blood flow with subsequent herniation is a major cause of mortality in ALF (1,2). Both cytotoxic cerebral edema (CCE) and vasogenic cerebral edema (VCE) are seen in ALF, with the former being predominant (3). In addition, studies have shown a higher prevalence of the “systemic inflammatory response syndrome” (SIRS) in patients with ALF (4). The SIRS is a response to the presence of proinflammatory cytokines (PCs) such as interleukin (IL)-6, IL-1β, and tumor necrosis factor (TNF)-α, the levels of which are raised in patients with ALF (5).
Ammonia and inflammation are the 2 main pathogenetic mechanisms of encephalopathy and CE in ALF (6). A positive correlation between PCs and blood ammonia/brain glutamine has been shown in adults with ALF on magnetic resonance spectroscopy (1H-MRS) (7). Diffusion tensor imaging (DTI) has shown persistence of CE in follow-up at 3 weeks despite apparent clinical recovery. This suggests that recovery of CE lags behind the clinical recovery of encephalopathy in ALF (8).
In addition, significant loss of volume of mammillary bodies (MBs) in absence of change in other brain areas has been observed in adults with ALF and acute on chronic liver failure (9). A significant improvement in MBs volume was seen compared to baseline in patients with ALF who survived their illness (9). As symmetrical atrophy of both MBs is seen in Wernicke-Korsakoff syndrome because of thiamine deficiency (TD) (10) and liver is the main storage site for thiamine, the authors speculated that changes in MB volume in ALF may be because of acute TD. Owing to complexity, unavailability, and cost of tests for measurement of blood thiamine (11), most studies including the above have not estimated the thiamine levels.
The temporal pattern of these changes and relative contribution of each of these factors in pathogenesis of CE is, however, still unknown. Most studies have addressed these factors in isolation and at one time point. There is lack of information regarding the recovery pattern of ALF, that is, order and time frame of improvement of inflammation (cytokines), hyperammonemia, CE, liver functions, and clinical functioning (neuropsychological status).
There is scarcity of literature regarding the pathogenesis of CE and role of inflammation in children with ALF. The single pediatric study of ALF evaluated 5 cases and showed higher glutamine and lactate concentrations in brain white matter in ALF than in controls with in vivo MR spectroscopy (12). In children the brain continues to develop in terms of microstructure and function whereas it is almost static in adults and so changes in children may be different from that in adults.
This is the first prospective, sequential, pilot study done with the aim of understanding cerebral changes in ALF children at diagnosis by using advanced MR imaging methods (1H-MRS, DTI), estimation of blood PCs and thiamine levels, and evaluating the sequence and timing of recovery of brain edema, glutamine, PCs, liver functions, and neurocognitive functions in these children.
The present prospective study was done from August 2010 to May 2011 in the Departments of Pediatric Gastroenterology and Radiology, Sanjay Gandhi Postgraduate Institute, Lucknow, India. Ethical approval for the study was taken from the institutional ethics committee. Patients with acute viral hepatitis (AVH)–related ALF diagnosed as per the Pediatric Acute Liver Failure Study Group criteria were evaluated (13). The etiological diagnosis of AVH was based on positive viral serology for hepatitis A (HAV) (IgM anti-HAV), hepatitis E (HEV) (IgM anti-HEV), or hepatitis B (HBV) (HBsAg and IgM anti-HBc). Patients with the evidence of chronic liver disease on clinical, laboratory, or radiological evaluation, acute liver injury due to nonviral causes, that is, hepatotoxic drugs, and other comorbid systemic diseases, and those referred for liver transplantation, were excluded. The controls were taken from a study on screening of asymptomatic children for neurocysticercosis. Children who did not have neurocysticercosis were further evaluated by neuropsychological tests (NPTs), and blood sample was taken for cytokine and thiamine estimation. Controls were evaluated at a single time point only. Informed consent was taken from the parents of the participating children.
Standard medical management was given to all of the patients in intensive care unit. Details of demographic, clinical, and laboratory parameters including liver function tests (LFTs) were noted. Grade of hepatic encephalopathy (HE), jaundice to HE interval, complications such as clinical features of CE, bleeding from any site, sepsis, and renal failure were recorded.
The patients were evaluated as per the protocol shown in Table 1. Hemodynamically unstable or mechanically ventilated children were not shifted for imaging studies. Blood sample for cytokines and thiamine was drawn on the day of neuroimaging. Patients were followed up after discharge and repeat evaluations were done on an outpatient basis as shown in Table 1. All of the patients were given a multivitamin preparation at discharge. The person conducting the neurocognitive test (NCT) was unaware of the findings of MR imaging.
Biochemical Tests and Viral Serology
Biochemical tests were performed using standard automated techniques. Commercially available enzyme-linked immunosorbent assay (ELISA) test kits were used for determining the viral etiology.
Serum Proinflammatory Cytokines
Serum TNF-α and IL-6 were quantified using commercially available ELISA kits (R&D Systems Inc, Minneapolis, MN).
Quantification of Blood Thiamine Level
Blood samples (2 mL) were collected in heparin tubes and stored at −80 °C. The details of the procedure are described elsewhere (14,15). Fluorescence was measured by multimode reader (BioTek Instrument Inc, Winooski, VT).
The Revised Amsterdamse Kinder Intelligentie test, a battery of 9 subsets of tests adapted for Indian children to detect abnormalities in neuropsychological functions was used (16). The procedure of performing these tests and functions assessed by them are described in detail elsewhere (17).
Magnetic Resonance Imaging
Patients with ALF underwent imaging after clinical stabilization. Patients were transported for imaging accompanied by staff trained in resuscitation and critical care with life-support facilities on standby. Brain MR imaging was performed on a 3.0-tesla MR scanner (HDXT, Signa, General Electric Technologies Systems, Milwaukee, WI) using an 8-channel head coil. Routine MR imaging, DTI, and 1H-MR spectroscopy studies were performed as described elsewhere (17,18).
DTI Data Analysis and Mammillary Bodies Volume Quantification
Regions of interest(s) (ROIs) were placed on posterior and anterior limbs of internal capsules (PLIC and ALIC), bilateral caudate nuclei (CN), putamen (P), thalami, globus pallidus (GP), cingulate gyrus (CG), frontal and occipital white matter (FWM, OWM), genu, and splenium (S) in patients and healthy controls to quantify fractional anisotropy (FA), mean diffusivity (MD), and spherical anisotropy (CS). The size of the ROIs varied from 2 × 2 to 6 × 6 pixels, with shape varying from elliptical to rectangular.
MBs volume was measured independently by 2 separate observers who were blinded to the clinical details. These computations and ROI analysis were done by in-house developed JAVA-based volume analysis tool (9). The total acquisition time for magnetic resonance imaging (MRI) examination was 16 minutes.
All demographic and laboratory parameters are expressed as mean ± SD. For comparison between groups, we used the chi-square test with Yates correction or the Fisher exact test for categorical variables and Student t test (paired and unpaired) for quantitative variables. SPSS version 17.0 (SPSS Inc, Chicago, IL) was used for statistical analysis and P value of <0.05 was taken as significant.
A total of 11 patients with ALF (5 boys, mean age 7.8 ± 2.7years) owing to AVH and 8 healthy controls (6 boys, mean age 8.8 ± 2.5years) were evaluated. Of the 11 patients with ALF, 1 died and 8 of 10 patients were reevaluated after discharge (first follow-up) and 6 of these 8 patients were further followed up (second follow-up) as shown in Table 1. All of the patients had HE at admission (HE grade I-4, II-1, III-4, IV-2), and the mean jaundice to HE interval was 10.6 ± 12.1 days. Two patients had spontaneous bacterial peritonitis; 1 patient each had mucosal bleed and renal failure. AVH was because of HAV in 7 (63.6%), HBV in 1(9.1%), and coinfection with HAV and HEV in 3 (27.3%) cases. The average time period between encephalopathy onset and discharge from the hospital was 15.8 ± 7.5 days. The first follow-up after discharge was at 43.5 ± 26.9 days and within 3 months after onset of encephalopathy, whereas the second follow-up was at 157.3 ± 52.3 days after discharge and at 4 to 7 months of onset of HE.
The LFT values of the patients with ALF at diagnosis and follow-up are shown in Table 2. There was significant improvement in all parameters of liver functions at first follow-up in comparison to the values at diagnosis. At the second follow-up, there was complete normalization of the liver functions.
Both TNF-α and IL-6 were significantly higher in patients with ALF at diagnosis and on first and second follow-up as compared to controls (Fig. 1). There was significant reduction in TNF-α (40.1 ± 8.9 vs 29.3 ± 8.8 pg/mL; P = 0.01) and IL-6 values (29.2 ± 14.4 vs 17.1 ± 5.3 pg/mL; P = 0.04) at first follow-up in comparison to that at diagnosis. The TNF-α and IL-6 values were insignificantly lower at second follow-up in comparison to that at first follow-up.
Patients with ALF had lower thiamine levels at diagnosis in comparison to controls (55.2 ± 6.7 vs 81.8 ± 10.2 nmol/L; P = 0.01). It showed significant improvement at first and second follow-up with values not significantly lower than controls (Fig. 1).
1H-Magnetic Resonance Spectroscopy
Patients with ALF had higher brain Glx (23.2 ± 3.4 vs 15.3 ± 2.7 [ppm]; P = 0.01) and lower choline (1.9 ± 0.36 vs 2.6 ± 0.6 [ppm]; P = 0.01) than controls at diagnosis. N-Acetyl aspartate and myoinositol values were similar in controls and patients with ALF (at diagnosis and follow-up, data not given). At first follow-up, Glx (16.5 ± 3.7 [ppm]; P = 0.5] was similar but the choline (2.0 ± 0.4 [ppm]; P = 0.04] was still significantly lower than controls.
Diffusion Tensor Imaging
ALF at Diagnosis Versus Control
The DTI metrics (MD and CS) in patients with ALF (diagnosis and first follow-up) and controls are shown in Fig. 2 A and B.
Significantly decreased MD value was observed in 6 regions, that is, PLIC, P, S, thalamus, OWM, and spectroscopic voxel (SV), in patients with ALF at diagnosis. Significantly increased CS value was seen in FWM, GP, S, ALIC, PLIC, CN, CG, and SV. Significantly decreased FA value was present in FWM, GP, S, genu, and SV.
ALF at First Follow-up Versus Control
There was significant improvement in the DTI metrics at follow-up, with all brain areas showing MD and FA values similar to those of controls. The CS values had reduced and were similar to those of controls in 7 of 8 involved areas at diagnosis. Significantly higher CS value than control at first follow-up was present only in CN.
Mammillary Body Volume
There was a significant reduction in the right (0.26 ± 0.06 mm3; P = 0.001) and left (0.22 ± 0.05 mm3; P = 0.001) MBs volume in patients with ALF at diagnosis as compared to controls (right 0.42 ± 0.07; left 0.43 ± 0.03 mm3). It showed significant increase at first follow-up (left = 0.35 ± 0.03; right = 0.37 ± 0.02 mm3) but still was significantly lower than controls.
The neuropsychological assessment was done in patients with ALF at both the follow-up visits. Neuropsychological assessment was not done in ALF children at diagnosis as all of them were in overt hepatic encephalopathy of various grades (I-IV). At first follow-up, that is, after discharge and no apparent HE on clinical examination, the patients with ALF performed poorly in 8 of 9 administered tests (Fig. 3). The performance was at par with controls only for the verbal meaning test. At second follow-up, the scores showed improvement in all the administered tests as compared to first follow-up. In 7 tests, the performance scores were similar to those of controls.
We have shown presence of CE with raised brain glutamine and blood PCs (TNF-α and IL-6) at diagnosis in children with ALF. The MBs were significantly smaller without structural changes in other brain areas along with reduced thiamine levels. At follow-up, at 6 weeks, despite absence of overt HE, there was impairment of cognitive functions with near-normalization of CE. The MBs showed significant increase in volume with improvement in thiamine levels. There was significant improvement in LFT with reduction in the PCs; however, these continued to be abnormal with low brain choline levels. At second follow-up, at nearly 6 months, the children showed normalization of liver functions with largely normal NCT. The cytokines had decreased significantly but still not normalized and thiamine levels were nearly normal. This suggests that recovery in ALF is prolonged, with various parameters normalizing in different time frames.
There was evidence of both CCE (reduced MD) and VCE (increased CS) in ALF children. The FA was reduced in few brain areas and we feel that it is due to the dilution effect caused by increased extracellular component of CE rather than due to microstructural changes. Both CCE and VCE have been reported in ALF (19) and the cytotoxic component is believed to be the predominant one (4,20). In our study, 6 brain areas showed CCE whereas 8 areas showed presence of VCE. This may be because we did not evaluate children with higher grades of HE and waited for them to clinically stabilize before performing imaging. In an animal study of ALF, Cauli et al (21) have shown that changes in brain are not uniform, more brain areas are involved with increasing grades of HE, and CCE predominates in grade III-IV HE. The relative contribution of cytotoxic and vasogenic component may also depend on the rapidity of the insult, that is, etiology and type (hyperacute/acute/subacute) of liver failure. Chavarria et al (22) studied an animal model of ALF with rapid development of coma and found only CCE whereas Cauli et al (21) found both CCE and VCE.
At follow-up after a mean duration of 45 days, complete reversal of CCE (normal MD) and near-normalization of VCE (increased CS in only 1 brain area) were seen. Rai et al (8) had shown that there is incomplete reversal of both CCE and VCE in adults with ALF at 3 weeks. Our data suggest that it may take >6 weeks for CE to normalize in ALF.
The brain glutamine was significantly raised at diagnosis in ALF and showed normalization at first follow-up. This supports the hypothesis that the increased ammonia in ALF is detoxified in the brain by the amidation of glutamate to glutamine, which accumulates in the astrocytes and causes astrocytic swelling and CE (23,24). Although we have not measured the blood ammonia levels, the raised brain glutamine appears to be an indirect evidence of increased cerebral ammonia.
The PCs (TNF and IL-6) were markedly increased in patients with ALF at diagnosis, which is consistent with earlier observations (5,24). Raised levels of PC have been shown to increase the cerebral blood flow and correlate with intracranial pressure in ALF (24). They may also directly affect astrocyte function (25) and upregulate inducible nitric oxide syntheses and evoke oxidative stress in these cells (26). Animal studies have shown that there is increased production of IL-6/TNF-α/IL-1β in the brain in ALF (27). Cerebral production of cytokines has also been shown in patients with ALF with raised intracranial tension (5). Release from the necrotic liver and increased production secondary to stimulation of the immune system by infection are the other postulated mechanisms for raised PCs in ALF. Significant reduction of CE with reduced cytokines and unchanged ammonia after temporizing hepatectomy in patients with ALF (28) and reduced CE in anhepatic than hepatic ALF animals (29) provides support to the fact that the necrotic liver is the source of these cytokines. The much higher cytokines at diagnosis and persistent increase even after normalization of CE support the hypothesis that the liver recovering from the injury is probably the source of increased cytokines in these patients.
There was significant reduction in the MBs volume in patients with ALF at diagnosis in comparison to controls that showed incomplete recovery at first follow-up. Concomitant with the MBs changes, blood thiamine levels showed a significant reduction at diagnosis and thereafter improvement at first follow-up. As thiamine is stored primarily in the liver (30), we believe that the rapid loss of liver cell mass in ALF results in TD and MBs changes. Biochemical evidence of TD in ALF has been observed in the past (31). The classical disorder caused by TD is Wernicke encephalopathy, and MBs atrophy is a consistent feature (32). Rapid shrinkage of MBs within 2 weeks in acute WE has been shown by sequential MR studies (33). This is in accordance with our observation as the duration of illness (jaundice) in patients before imaging was 24.0 ± 17.2 days. Our finding of loss of MBs volume is consistent with earlier observations in adults with ALF (9); however, this study confirms the TD along with reduction of MBs volume, which was not done in the earlier study (9). The improvement in thiamine levels seen at first follow-up could be attributed to the thiamine supplemented in the multivitamin, improved appetite and dietary intake, and recovery of liver functions; however, the multivitamin supplementation contained only 2 mg/day of thiamine, which is much less than the therapeutic doses recommended in cases with established TD.
A observation of our study is that there was significant impairment on NCT in children with ALF at first follow-up despite absence of overt HE and this resolved at the second follow-up. This may be explained by 2 factors: elevated PCs and loss of volume of MBs. Cytokines can stimulate vagal afferents that affect brainstem, limbic, and hypothalamic structures to produce “illness behavior” characterized by depressed mood and cognitive impairment (34). MBs are known to play an important role in memory and cognitive functions (35) as these are the major relay nuclei with limbic and extralimbic connections. As the MBs were still showing loss of volume as compared to controls at first follow-up, we believe these may be an important contributor to the poor performance on NCT in these children at first follow-up. This also suggests that thiamine supplementation in therapeutic doses may lead to faster recovery of the subtle neurocognitive abnormality in patients with ALF at follow-up and needs further evaluation.
The mildly elevated TNF-α and IL-6 at the second follow-up despite normalization of liver functions are difficult to explain and need confirmation in future studies with large sample size. It may be attributed to the process of liver regeneration as cytokines, especially IL-6, are important for replication of differentiated hepatocytes and hematopoietic cells in the liver (36). Also studying the serial cytokine profile at diagnosis and follow-up in ALF children who have undergone liver transplantation may be useful to resolve this issue in future.
Small sample size, noninclusion of severe grades of (III-IV) HE, lack of MR imaging at the second follow-up, and not estimating the blood ammonia levels at the different time points in the ALF cases are some of the limitations of this study; however, this study has the advantage of being the first sequential study in children with ALF owing to single etiology (AVH), which has evaluated brain changes with both 1H-MRS and DTI, PCs and NPT at different time points of diagnosis and recovery. We have also shown loss of volume of MBs in association with low thiamine level and its recovery at follow-up for the first time. There is a need to study larger number of ALF patients with different grades of HE at multiple time points of acute illness until complete neurocognitive recovery to understand the exact pathogenetic process in ALF.
In conclusion, this is the first longitudinal study in children with ALF that shows both CCE and VCE. The MBs are small due to TD. The recovery in ALF is prolonged and goes well beyond the apparent clinical recovery of HE. CE and brain glutamine recover first, followed by normalization of NPT and liver functions. Raised cytokines and loss of MBs volume probably contribute to the neurocognitive abnormality seen at first follow-up. The TD in these children gradually recovers at serial follow-up, suggesting that thiamine supplementation may of value during the acute illness. Persistence of raised cytokines up to 6 months after insult suggests possible contribution from liver regeneration.
1. Acharya SK, Dasarathy S, Kumar TL, et al. Fulminant hepatitis in a tropical population: clinical course, cause, and early predictors of outcome. Hepatology
2. Ware AJ, D’Agostino AN, Combes B. Cerebral edema: a major complication of massive hepatic necrosis. Gastroenterology
3. Blei AT. The pathophysiology of brain edema in acute liver failure. Neurochem Int
4. Rolando N, Wade J, Davalos M, et al. The systemic inflammatory response syndrome in acute liver failure. Hepatology
5. Jalan R, Olde Damink SW, Deutz NE, et al. Moderate hypothermia prevents cerebral hyperemia and increase in intracranial pressure in patients undergoing liver transplantation for acute liver failure. Transplantation
6. O’Beirne JP, Chouhan M, Hughes RD. The role of infection and inflammation in the pathogenesis of hepatic encephalopathy and cerebral edema in acute liver failure. Nat Clin Pract Gastroenterol Hepatol
7. Gupta RK, Yadav SK, Rangan M, et al. Serum proinflammatory cytokines correlate with diffusion tensor imaging derived metrics and 1H-MR spectroscopy in patients with acute liver failure. Metab Brain Dis
8. Rai V, Nath K, Saraswat VA, et al. Measurement of cytotoxic and interstitial components of cerebral edema in acute hepatic failure by diffusion tensor imaging. J Magn Reson Imaging
9. Mandal PK, Yadav SK, Saraswat VA, et al. Mammillary body atrophy in acute liver failure and acute-on-chronic liver failure of nonalcoholic etiology. Metab Brain Dis
10. Charness ME, DeLaPaz RL. Mammillary bodies atrophy in Wernicke's encephalopathy antemortem identification using magnetic resonance imaging. Ann Neurol
11. Fei GQ, Zhong C, Jin L, et al. Clinical characteristics and MR imaging features of nonalcoholic Wernicke encephalopathy. Am J Neuroradiol
12. Sijens PE, Alkefaji H, Lunsing RJ, et al. Quantitative multivoxel 1H MR spectroscopy of the brain in children with acute liver failure. Eur Radiol
13. Bucuvalas J, Yazigi N, Squires RH Jr. Acute liver failure in children. Clin Liver Dis
14. Tabrizi AB. A cloud point extraction-spectroflurimetric method for determination of thiamine in urine. Bull Korean Chen Soc
15. Wyatt DT, Nelson D, Hillman RE. Age-dependent changes in thiamine concentrations in whole blood and cerebrospinal fluid in infants and children. Am J Clin Nutr
16. Khire U, Bleichrodt N, Hoksbergen Rene AC, et al. Indian Child Intelligence Test (ICIT), Technical Manual, Adaptation of Revised Amsterdamse Kinder Intelligentie Test.
1st ed. Mumbai: Jnana Prabodhini's Samshodhan Sanstha; 1992.
17. Yadav SK, Srivastava A, Srivastava A, et al. Encephalopathy assessment in children with extra-hepatic portal vein obstruction with MR, psychometry and critical flicker frequency. J Hepatol
18. Saksena S, Rai V, Saraswat VA, et al. Cerebral diffusion tensor imaging and in vivo proton magnetic resonance spectroscopy in patients with fulminant hepatic failure. J Gastroenterol Hepatol
19. Detry O, Roover AD, Honore P, et al. Brain edema and intracranial hypertension in fulminant hepatic failure: pathophysiology and management. World J Gastroenterol
20. 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
21. Cauli O, Lopez-Larrubia P, Rodrigo R, et al. Brain region selective mechanisms contribute to the progression of cerebral alterations in acute liver failure in rats. Gastroenterology
22. Chavarria L, Oria M, Romero-Gimenez J, et al. Diffusion tensor supports the cytotoxic origin of brain edema in a rat model of acute liver failure. Gastroenterology
23. Martinez-Hernandez A, Bell KP, Norenberg MD. Glutamine synthetase: glial localization in brain. Science
24. Jalan R, Olde Damink SW, Hayes PC, et al. Pathogenesis of intracranial hypertension in acute liver failure: inflammation, ammonia and cerebral blood flow. J Hepatol
25. Blei AT. Infection, inflammation and hepatic encephalopathy, synergism redefined. J Hepatol
26. Zhao ML, Liu JS, He D, et al. Inducible nitric oxide synthase expression is selectively induced in astrocytes isolated from the adult human brain. Brain Res
27. Jiang W, Desjardins P, Butterworth RE. Direct evidence for central proinflammatory mechanisms in rats with experimental acute liver failure: protective effect of hypothermia. J Cereb Blood Flow Metab
28. Jalan R, Pollok A, Shah SH, et al. Liver derived pro-inflammatory cytokines may be important in producing intracranial hypertension in acute liver failure. J Hepatol
29. Olafsson S, Gottstein J, Blei AT. Brain edema and intracranial hypertension in rats after total hepatectomy. Gastroenterology
30. Frank O, Luisada-Opper A, Sorrell MF, et al. Vitamin deficits in severe alcohol fatty liver of man calculated from multiple reference points. Exp Mol Pathol
31. Labadarios D, Rossouw JE, McConnell JB, et al. Thiamine deficiency in fulminant hepatic failure and effects of supplementation. Int J Vitam Nutr Res
32. Sheedy D, Lara A, Garrick T, et al. Size of mammillary bodies in health and disease: useful measurement in neuroradiological diagnosis of Wernicke's encephalopathy. Alcohol Clin Exp Res
33. Dap rile P, Gentile MA, Carmella A. Enhancement MR in the acute phase of WE. Am J Neuroradiol
34. Dantzer R, Konsman JP, Bluthe RM, et al. Neural and humoral pathways of communication from the immune system to the brain: parallel or convergent. Auton Neurosci
35. Vann SD, Aggleton JP. The mammillary bodies: two memory systems in one? Nat Rev Neurosci
36. Galun E, Axelrod JH. The role of cytokines in liver failure and regeneration: potential new molecular therapies. Biochim Biophys Acta