functions, thereby defining subclinical hepatic encephalopa-Hepatic encephalopathy is a common problem in cirthy. 1 Neuropsychiatric testing is biased by education, intellirhosis. The pathogenesis of this complication of adgence, compliance, age, and learning effects. 2 Because early vanced liver disease still remains unclear. Magnetic resdiagnosis and treatment of subclinical hepatic encephalopaonance spectroscopy was used to assess prospectively thy have been suggested to be critical for the preservation of cerebral metabolism in 51 patients with histologically cerebral function, 3 efforts have been made to apply other, proven cirrhosis (Child-Pugh classes A, B, and C, 18, 18, more objective and reliable methods in the diagnosis of heand 15, respectively) and 36 healthy volunteers. Acpatic encephalopathy. cording to the results of psychometric tests, overt he-Using proton magnetic resonance spectroscopy (MRS), typpatic encephalopathy, subclinical encephalopathy, and ical cerebral metabolite patterns have been found in clinical no encephalopathy were found in 14, 21, and 16 patients, or subclinical hepatic encephalopathy. 2,4 The predominant respectively. Myoinositol/creatine ratios in gray (.36 { findings are a decreased myoinositol/creatine (Myo/Cr) ratio, 4 .17) and white (.35 { .22) matter voxel were reduced sigdecreased of choline/creatine (Cho/Cr) ratio, 4 and increased nificantly (P õ .0001) in cirrhotic patients compared glutamine/glutamate/creatine (Glx/Cr) ratio 4,5 in the brain. with healthy volunteers (gray matter, .51 { .11; white
Alterations of cerebral metabolism measured by MRS rematter, .64 { .16). In addition, patients showed a signifiturned to normal in most patients after liver transplantacant reduction (P Å .024) in white matter choline/cretion, [6][7][8] demonstrating that they are related to liver disease. atine ratio (.77 { .27) compared with controls (.92 { .25),
In magnetic resonance imaging (MRI), the presence of and glutamine/glutamate level was elevated in cirrhotic signal hyperintense basal ganglia in T1-weighted images appatients compared with controls (gray matter, P õ .0001; pears closely related to severe liver disease with both portalwhite matter, P Å .036). Changes in cerebral myoinositol systemic shunting 9 and hepatic encephalopathy. [10][11][12][13][14] Howand glutamine/glutamate levels correlated significantly ever, the latter finding has been questioned recently. 15 with the severity of hepatic encephalopathy (P õ .0001).
Pomier Layrargues et al. 16 and Krieger et al. 17 have demon-However, these metabolic alterations were also detected strated recently that the concentration of manganese, a parain patients without hepatic encephalopathy (normal magnetic substance, is elevated in bright basal ganglia.
psychometric test results). N-acetyl aspartate/creatine
The aim of the present study was to evaluate prospectively ratios did not differ between patients and controls. Magcerebral metabolism and morphological abnormalities in 51 netic resonance imaging detected bright basal ganglia in clinically well-defined patients with histologically proven cir-37 patients, which correlated significantly with portalrhosis with or without signs of hepatic encephalopathy using systemic shunting and elevation of glutamine/gluta-MRS and MRI. mate, but not with the degree of hepatic encephalopathy. In conclusion, magnetic resonance imaging and PATIENTS AND METHODS spectroscopy showed that alterations of cerebral metabolism are common in patients with cirrhosis, even with-
Patients and Volunteers.
Fifty-one patients with proven cirrhoout evidence of clinical or subclinical hepatic encephasis, 35 male and 16 female, aged 51 { 11 years (mean { SD) and 53 lopathy. (HEPATOLOGY 1997;25:48-54.)
{ 12 years, respectively, were prospectively included in the study. Patients with neurological disorders, diabetes mellitus, or renal failure were not enrolled in this study. Thirty-five patients had alcoholic cirrhosis, 8 patients had posthepatitic cirrhosis, and 3 had hemochro-In patients with cirrhosis, hepatic encephalopathy is a commatosis; the other 5 had cirrhosis caused by Wilson's disease (n Å mon complication. Although the clinical diagnosis of severe 2), primary biliary cirrhosis (n Å 2), and portal malformation (n Å portal-systemic encephalopathy is easily made, minor neuro-1). Severity of liver disease was staged according to the Child-Pugh psychiatric alterations are difficult to detect. Psychometric classification: 18 patients were classified as having Child A disease, tests have been used to detect subtle neuropsychiatric dys-18 as Child B, and 15 as Child C. In each patient a thorough clinical examination, psychometric testing, MRI, and MRS were performed within 24 hours. Gastroscopy was performed in all but two patients to assess the presence of esophageal varices. All patients were exam-Abbreviations: MRS, magnetic resonance spectroscopy; Myo/Cr, myoinositol/creatine; ined by abdominal ultrasound.
Cho/Cr, choline/creatine; Glx/Cr, glutamine/glutamate/creatine; MRI, magnetic resonance Thirty-six healthy volunteers (21 male, 15 female; aged 37 { 16 imaging; TR, repetition time; TE, echo time; NAA/Cr, N-acetyl aspartate/creatine. years and 39 { 17 years, respectively) served as controls.