Patients with minimal hepatic encephalopathy show impaired mismatch negativity correlating with reduced performance in attention tests

Authors


  • Potential conflict of interest: Nothing to report.

  • This work was supported by grants from Ministerio de Ciencia Innovacion (SAF2008-00062 and CSD2008-00005, to V.F.; and FIS 06/0065 and PS09/00806, to C.M.), from Consellería Educación (ACOMP-2009-025, ACOMP/2009/191, PROMETEO-2009-027, and ACOMP2010-220), AP-092/09, AP-043-10, and AP-028/10 from Conselleria Sanitat, Generalitat Valenciana, and from Fundacion Mutua Madrileña (to C.M.).

Abstract

Attention deficit is an early event in the cognitive impairment of patients with minimal hepatic encephalopathy (MHE). The underlying mechanisms remain unclear. Mismatch negativity (MMN) is an auditory event-related potential that reflects an attentional trigger. Patients with schizophrenia show impaired attention and cognitive function, which are reflected in altered MMN. We hypothesized that patients with MHE, similarly to those with schizophrenia, should show MMN alterations related with attention deficits. The aims of this work were to assess whether (1) MMN is altered in cirrhotic patients with MHE, compared to those without MHE, (2) MMN changes in parallel with performance in attention tests and/or MHE in a longitudinal study, and (3) MMN predicts performance in attention tests and/or in the Psychometric Hepatic Encephalopathy Score (PHES). We performed MMN analysis as well as attention and coordination tests in 34 control subjects and in 37 patients with liver cirrhosis without MHE and 23 with MHE. Patients with MHE show reduced performance in selective and sustained attention tests and in visuomotor and bimanual coordination tests. The MMN wave area was reduced in patients with MHE, but not in those without MHE. In the longitudinal study, MMN area improved in parallel with performance in attention tests and PHES in 4 patients and worsened in parallel in another 4. Logistic regression analyses showed that MMN area predicts performance in attention tests and in PHES, but not in other tests or critical flicker frequency. Receiver operating characteristic curve analyses showed that MMN area predicts attention deficits in the number connection tests A and B, Stroop tasks, and MHE, with sensitivities of 75%-90% and specificities of 76%-83%. Conclusion: MMN area is useful to diagnose attention deficits and MHE in patients with liver cirrhosis. (HEPATOLOGY 2012;)

Approximately 33%-50% of patients with liver cirrhosis without clinical symptoms of encephalopathy show minimal hepatic encephalopathy (MHE), which can be unveiled using psychometric tests or neurophysiological analysis.1-4 Patients with MHE show attention deficits and mild cognitive impairment. MHE reduces quality of life and is associated with increased risk of suffering with work, driving, and home accidents as well as clinical hepatic encephalopathy (HE) and reduced life span.5-10

Attention deficits are an early manifestation of MHE.11-16 Amodio et al.16 reported that MHE affects primarily selective attention control. Weissenborn et al.15 reported that patients with MHE show dysfunction in all attention subsystems. The brain areas involved in the attention system and the alterations in attention in MHE were previously summarized.15, 16 However, how MHE alters attention systems, which components are affected, and the underlying mechanisms remain unknown.

Measurement of event-related potentials from electroencephalography (EEG) recordings allows for quantifying the neuronal processes associated with sensorial and cognitive events. Mismatch negativity (MMN) is an auditory event-related potential elicited when a sequence of repetitive standard sounds is interrupted infrequently by deviant “oddball” stimuli. The MMN is a measure of cortical activity in response to the deviant sound and reflects an automatic, memory-based, comparison process.17-22 It can be rapidly assessed, elicited while the individuals are performing other tasks or sleeping, and reflects preattentive sensory memory and involuntary attention.17

The area under the MMN wave in frontal electrodes is reduced in patients with schizophrenia, compared to controls, and the area correlates with the degree of cognitive impairment.18 Baldeweg et al.18 suggested that altered MMN in schizophrenia reflects an impaired attentional trigger, which would be a consequence of deficits in N-methyl-D-aspartate (NMDA) receptor-dependent neural processes underlying it.

These and other studies19-22 support that, in schizophrenia, alterations in neurotransmission associated with NMDA receptors lead to impaired attention and cognitive function, which are reflected in altered MMN, and result in impairment in everyday functioning, including sustained attention impairment.

Patients with MHE also show impaired attention (including sustained attention) and cognitive function, which result in impairment in everyday functioning. Altered neurotransmission associated with NMDA receptors is a main contributor to cognitive impairment in animal models of HE.23-26 It is, therefore, likely that altered NMDA-receptor neurotransmission in the cortex could also contribute to attention deficits in MHE. This should be reflected in alterations in MMN.

We hypothesized that patients with MHE, similarly to those with schizophrenia, should show alterations in MMN, which would be related with attention deficits. The aim of this work was to assess whether (1) MMN is altered in cirrhotic patients with MHE, compared to those without MHE and to controls without liver disease, (2) MMN changes in parallel with performance in attention tests and/or with MHE in a longitudinal study; and (3) MMN predicts performance in attention tests and/or in the Psychometric Hepatic Encephalopathy Score (PHES).

We performed MMN analysis and attention tests in 34 controls without liver disease, 37 patients with liver cirrhosis without MHE, and 23 with MHE. We used the Stroop and Map search tests to assess selective attention and the Elevator Counting test to assess sustained attention as well as visuomotor and bimanual coordination tests. We analyzed, in the same patients, the critical flicker frequency, proposed as an alternative method to detect MHE.

Abbreviations

CFF, critical flicker frequency; DST, digit symbol test; EEG, electroencephalography; HE, hepatic encephalopathy; LTT, line-tracing test; MHE, minimal hepatic encephalopathy; MMN, mismatch negativity; NCT-A, number connection test A; NCT-B, number connection test B; NMDA, N-methyl-D-aspartate; SD, serial dotting test; PHES, the Psychometric Hepatic Encephalopathy Score; ROC, receiver operating characteristic.

Patients and Methods

Patients and Controls.

Sixty patients with liver disease and 34 controls were enrolled after written consent. Inclusion criteria comprised the following: Patients were recruited between March 2007 and April 2010 from the outpatient clinics at Hospital Clinico Universitario and Hospital Arnau de Vilanova, in Valencia, Spain, and were included if they had clinical, biochemical, and histological evidence of liver cirrhosis. For controls, liver disease was discarded by clinical, analytical, and serologic analysis. All subjects were volunteers. Patients were excluded if they had clinical evidence of overt HE, as measured by the West Haven criteria,27 decompensate diabetes, renal dysfunction, hyponatremia, neurological disease, severe cardiovascular disease, or antibiotic use. Patients had to be abstinent from alcohol for 6 months before the study. Patients were not on any specific therapy for HE.

The study protocol conformed to the ethical guidelines of the 1975 Declaration of Helsinki28 and was approved by the ethical committee of the hospital.

After performing the psychometric tests, patients were classified as without MHE or with MHE (see below). The study included, therefore, three groups: (1) control subjects; (2) patients without MHE; and (3) patients with MHE. Composition of the groups, age, and etiology of the disease are given in Tables 1A and 1B. Table 2 shows the analytical data.

Table 1A. Characteristics of Controls and Patients in the First Study
CharacteristicControlPatients Without MHEPatients With MHE
  1. Values are expressed as mean ± standard deviation.

  2. Abbreviations: MHE, minimal hepatic encephalopathy; M, male; F, female; HBV, hepatitis B virus; HCV, hepatitis C virus; MELD, Model for End Stage Liver Disease.

Total individuals (M/F)34 (23/11)37 (28/9)23 (14/9)
Age51 ± 1155 ± 864 ± 10
Alcohol 2514
HBV 10
HCV 78
HBV+alcohol 11
HCV+alcohol 30
Child Pugh A/B/C 30/7/014/8/1
MELD 10 ± 39 ± 3
Ascites 35

1B

Table 1B. Characteristics of Patients in Longitudinal Study
CharacteristicPatients Without MHEPatients With MHE
  1. Abbreviations: MHE, minimal hepatic encephalopathy; M, male; F, female; HBV, hepatitis B virus; HCV, hepatitis C virus; MELD, Model for End Stage Liver Disease.

Total individuals (M/F)31 (24/7)14 (7/7)
Exitus (M/F)0/14/1
Age57 ± 866 ± 7
Alcohol219
HBV10
HCV65
HBV+alcohol10
HCV+alcohol20
Child Pugh A/B/C24/7/09/5/0
MELD9 ± 311 ± 3
Ascites14
Table 2. Clinical Data of the Subjects in the First Study
 Normal RangeControlPatients Without MHEPatients With MHE
  • Values are expressed as mean ± standard deviation.

  • Abbreviations: MHE, minimal hepatic encephalopathy; GOT, glutamic-oxaloacetic transaminase; GPT, glutamic-pyruvic transaminase; GGT, gamma-glutamyl transferase.

  • *P ≤ 0.05;

  • P < 0.01;

  • P < 0.001.

  • §

    P < 0.05, values significantly different in patients with and without MHE.

Total individuals 343723
GOT (mU/mL)(1-37)20 ± 473 ± 5683 ± 58
GPT (mU/mL)(1-41)18 ± 677 ± 2490 ± 24
GGT (mU/mL)(10-49)27 ± 586 ± 60106 ± 64
Uric acid (mg/dL)(2.5-7.0)4.0 ± 1.06.2 ± 2.05.7 ± 2.3
Creatinine (mg/dL)(0.5-1.3)0.9 ± 0.11.1 ± 0.21.2 ± 0.2
Cholesterol (mg/dL)(140-200)172 ± 22175 ± 44167 ± 55
Triglycerides (mg/dL)(40-160)95 ± 32111 ± 64119 ± 64
Bilirubin (mg/dL)(0.1-1.0)0.6 ± 0.21.7 ± 0.72.3 ± 0.6
Albumin (g/dL)(3.5-5.0)4.4 ± 0.23.7 ± 0.62.9 ± 0.6§
Prothrombin time (seconds) 13 ± 1.324 ± 430 ± 4
Fibrinogen (g/L)(2-4)3.1 ± 1.03.3 ± 1.33.6 ± 1.2
Alkaline phosphatase (mU/mL)(50-250)147 ± 53216 ± 77314 ± 96§
Erythrocytes(4.2-6.1)4.6 ± 0.44.3 ± 0.73.4 ± 0.6
Leucocytes(4.8-10.8)6.5 ± 1.36 ± 2.65.5 ± 2.0
Neutrophils (%)(55-75)55 ± 754 ± 659 ± 9
Lymphocytes (%)(17-45)35 ± 629 ± 10*27 ± 9*
Monocytes (%)(2-8)6.0 ± 1.38.4 ± 3.010 ± 2.6
Eosinophils (%)(1-4)3.3 ± 2.02.4 ± 1.21.7 ± 1.0
Basophils (%)(0.05-0.5)0.5 ± 0.20.6 ± 0.30.6 ± 0.1

Follow-up Study and Outcomes.

To assess whether MMN changes in parallel with MHE and/or performance in attention tests, we performed a longitudinal follow-up study 10-18 months after the first study. In total, 31 of 37 patients without MHE were included in the follow-up. Two patients were not included because they underwent liver transplantation, 1 died, and 3 did not want to collaborate. A total of 14 of 23 patients with MHE were included in the follow-up. Four patients did not want to collaborate and another 5 died (only 1 by complication of liver cirrhosis). All patients in the longitudinal study were stable, with no clinical or therapeutic changes. Parameters remained stable during the follow-up time, with no incident derived from diuretics, digestive hemorrhage, or taking antibiotics.

Diagnosis of MHE.

MHE was diagnosed using the PHES, which is recommended as the “gold standard.”2 PHES comprises five psychometric tests: the digit symbol test (DST), number connection test A (NCT-A), number connection test B (NCT-B), the serial dotting test (SD), and the line-tracing test (LTT).29, 30 The score in each of the tests and the PHES were calculated by adjusting for age and education level by means of Spanish normality tables freely available since 2004 at http://www.redeh.org. Patients were classified as having MHE when the score was less than −4 points.29

Critical Flicker Frequency.

Critical flicker frequency (CFF) has been proposed as an alternative procedure for detection of MHE in cirrhotic patients.31, 32 CFF was measured as described previously.32

Stroop Test.

We used a color-word version of the Stroop task33 to assess selective attention. Each individual performed sequentially the congruent, neutral, and incongruent tasks, with 45 seconds per task. If a subject gives a wrong response, he or she must repeat the item. The number of items correctly named was quantified and adjusted by age, according to Spanish normality tables.34

Map Search.

The Map Search subtest version A of the everyday attention test was administered according to the manual to assess selective attention.35 The score is the number of items of 80 found in 2 minutes, and scaled-score equivalents of raw scores for four age bands are assigned to each subject.35

Elevator Counting: Sustained Attention.

The Elevator Counting subtest version A was administered according to the manual35 to evaluate sustained attention.

Visuomotor Coordination.

This test consists of a board with a series of perforations of identical size arranged in six rows and six columns, perfectly aligned, but differently orientated, and a series of metal pieces that fits perfectly into the holes.36 The subject has to place the pieces, one by one, by rows into the perforations, until filling all the board. The test is performed twice, and total time is recorded.

Bimanual Coordination.

The test consists of moving a series of metallic pegs, placed in one half of a pegboard, to the other half of the board, in order, by performing the movements symmetrically and simultaneously with both hands.37 The operation is repeated twice in each direction, and time is recorded as an index of bimanual coordination.

MMN.

The stimulation protocol and the MMN analysis were performed, as previously described,18 using a device for evoked potentials (NeuropackM1, 8-channels; Nihon-Kohden, Tokyo, Japan) and software for evoked potentials modified for MMN. Pure sinusoidal tones (80-dB SPL, 5-ms rise/fall) were delivered binaurally via insert earplugs. A stimulus train consisted of a sequence of standard tones of one frequency and duration (10 ms), which were followed by an intertrain interval of 300 ms. The first tone of the next train (differing in frequency) corresponded to the “deviant.” Twelve frequencies, ranging in 50-Hz steps from 750 to 1,250 Hz, were used. The number of tones in each stimulus train varied randomly and could be 2, 4, 8, 16, or 36. A total of 4,500 stimuli and 400 deviants were delivered. During the 45-minute EEG recording, subjects watched a silent self-selected video film. EEG was recorded continuously from electrodes Fz, F3, F4, Cz, and left and right mastoids placed according to the international 10-20 system. The vertical electrooculogram was recorded from electrodes placed above the right eye and the right outer canthus. System bandpass was 0-70 Hz, with a digital sampling rate of 500 Hz. The ground electrode was placed on the central forehead and reference on the bridge of the nose. Data were analyzed as previously described.18

Statistical Analysis.

Values are given as mean ± standard error of the mean. Results were analysed by one-way analysis of variance followed by post-hoc Newman-Keuls test, using GraphPad Prism software (version 4.0; GraphPad Software, Inc., Cary, NC). Variables that were not previously age adjusted (e.g., bimanual coordination and visuomotor coordination) were compared between groups using univariate analysis of covariance with age included as covariate, followed by post-hoc Bonferroni. The probability level accepted for significance was P < 0.05. Bivariate correlations among variables were evaluated using the Pearson correlation test. Partial correlation coefficients, controlled by age, were also calculated for variables not previously age adjusted. Binary logistic regression analyses were performed to assess whether MMN area predicts MHE, attention, or coordination deficits. The cutoffs (mean of controls ± 2 standard deviations) were 28 for Stroop Incongruent: 3.12 and 2.37 minutes for visuomotor and bimanual coordination tests, respectively, and 0 for NCT-A and NCT-B tests. Receiver operating characteristic (ROC) curves were then performed to determine sensitivity and specificity. Analyses were performed using SPSS software (version 17.0; SPSS, Inc., Chicago, IL), and two-sided P values <0.05 were considered significant.

Results

Latency and amplitude of MMN waves were similar in controls and patients with or without MHE (Fig. 1A,B). Latencies were 212 ± 5, 224 ± 8, and 213 ± 10 ms in controls, patients without MHE, and patients with MHE, respectively. Amplitudes were 5.4 ± 0.5, 5.1 ± 0.6, and 5.0 ± 0.8 μV in controls, patients without MHE, and patients with MHE, respectively.

Figure 1.

MMN area was reduced in patients with MHE. The figure represents the individual data for each patient or control. (A) Latency, (B) amplitude, and (C) area of the MMN wave. Values significantly different are indicated by asterisks. *P < 0.05; **P < 0.01.

In contrast, MMN area was reduced in patients with MHE, compared to controls (P < 0.01) and patients without MHE (P < 0.05). Areas were 167 ± 29, 120 ± 17, and 49 ± 4 μV/ms in controls, patients without MHE, and patients with MHE, respectively (Fig. 1C).

Performance in the Stroop test of selective attention was also assessed. In the congruent task (Fig. 2A), controls read 108 ± 3 words in 45 seconds. Patients without MHE read fewer words (94 ± 4; P < 0.05), and patients with MHE showed a strong reduction in number of words (77 ± 5), which was lower than for controls (P < 0.001) and patients without MHE (P < 0.05).

Figure 2.

Performance in the Stroop and Map Search tests of patients with and without MHE. The figure represents the individual data for each patient or control. (A) Data for the Stroop congruent task, (B) neutral task, (C) incongruent task, and (D) Map Search test. Values significantly different are indicated by asterisks. *P < 0.05; **P < 0.01; ***P < 0.001.

In the neutral task (Fig. 2B), control subjects named 80 ± 3 colors. Patients without MHE named fewer colors (67 ± 3; P < 0.01) and patients with MHE named 53 ± 5, which was lower than for controls (P < 0.001) and patients without MHE (P < 0.05).

In the incongruent task (Fig. 2C), controls named 45 ± 2 colors. Patients without MHE named fewer colors (37 ± 2; P < 0.01) and patients with MHE named 30 ± 2, which was lower than for controls (P < 0.001) and patients without MHE (P < 0.05).

Visual selective attention was evaluated by performing the Map Search. In the 2-minute Map Search test (Fig. 2D), control subjects obtained a scaled score of 9.7 ± 0.8. The score was not affected in patients without MHE (7.9 ± 0.5). Patients with MHE showed a reduction in score (5.7 ± 0.8), which was lower than for controls (P < 0.01) and for patients without MHE (P < 0.05).

In the Elevator Counting test, all controls and patients without MHE got the maximal score of 7. Four of the eleven patients with MHE who performed the test obtained lower scores (4, 5, 6, and 6, respectively), indicating impaired sustained attention.

In the bimanual coordination test, control subjects completed the task in 1.7 ± 0.1 minutes. Patients without MHE needed 2.1 ± 0.1 minutes. Patients with MHE showed a reduction in bimanual coordination. They needed 2.4 ± 0.3 minutes, which was higher than for control subjects (P < 0.05, first study; P < 0.001, follow-up study) and for patients without MHE in follow-up study (P < 0.001)(Fig. 3A).

Figure 3.

Performance in the bimanual (A) and visuomotor (B) coordination tests of patients with and without MHE. The figure represents the individual data for each patient or control. Values significantly different are indicated by asterisks. *P < 0.05; ***P < 0.001.

In the visuomotor coordination test, controls completed the task in 2.2 ± 0.1 minutes. Score was not affected in patients without MHE, who needed 2.5 ± 0.1 minutes. Patients with MHE needed more time (3.4 ± 0.31 min; P < 0.05, first study; P < 0.001, follow-up study) (Fig. 3B).

Critical flicker frequency was not different in patients without MHE (41 ± 4 Hz; n = 36) than in controls (44 ± 4 Hz; n = 13). CFF was reduced (P < 0.001) in patients with MHE to 37 ± 4 Hz (n = 20).

Statistical correlations between the different parameters analyzed are shown in Table 3.

Table 3. Analysis of the Correlations Between the Different Parameters Analyzed in the First Study
Parameter MMN AreaPHES*CFFStroop Congruent*Stroop Neutral*Stroop Incongruent*Map Search*BCTVCTDST*NCT-A*NCT-B*SD*LTT*
  • P and r values of the correlation analysis are shown.

  • Abbreviations: MMN, mismatch negativity; CFF, critical flicker frequency; PHES, the Psychometric Hepatic Encephalopathy Score; DST, digit symbol test; NCT-A, number connection test A; NCT-B, number connection test B; SD, serial dotting test; LTT, line-tracing test; ns, not significant; BCT, bimanual coordination test; VCT, visuomotor coordination test.

  • *

    These variables were previously age adjusted when evaluated.

  • Partial correlations controlled by age are shown for these variables.

  • CFF bivariate correlations are shown because it did not correlate with age.

MMN arear = 0.389ns0.314ns0.4680.364ns−0.475ns0.3520.422nsns
P = 0.006 0.029 0.0020.034 0.016 0.0150.003  
PHES*r =  0.4000.4290.3900.4090.412−0.632−0.6460.4420.6890.7660.5260.699
P =  0.0010.0050.0110.0070.010<0.0001<0.0001<0.0001<0.0001<0.0001<0.0001<0.0001
CFFr =   nsnsnsnsnsns0.296ns0.409ns0.386
P =         0.017 0.001 0.002
Stroop congruent*r =    0.6730.5790.526−0.558−0.538nsns0.362nsns
P =    <0.0001<0.0001<0.00010.0020.003  0.020  
Stroop neutral*r =     0.7520.416−0.636−0.618ns0.466nsnsns
 P =     <0.00010.008<0.0001<0.0001 0.003  
Stroop incongruent*r =      0.353−0.384−0.413ns0.3090.321nsns
P =      0.0380.0440.029 0.0470.04  
Map Search*r =       −0.474−0.397ns0.4260.360nsns
P =       0.0110.036 0.0120.031  
BCTr =        0.843−0.604−0.543−0.513ns−0.514
P =        <0.00010.0010.0030.005 0.005
VCTr =         −0.650−0.501−0.544ns−0.450
P =         <0.00010.0070.003 0.016

To assess whether MMN changes in parallel with MHE and/or performance in attention tests, we performed a longitudinal follow-up study. The effects of MHE on MMN latency, amplitude, and area and on performance on the Stroop, Map Search, and bimanual and visuomotor coordination tests were the same as in the first study (Figs 1-4).

Figure 4.

Follow-up of changes in different parameters in patients who improved (A-C) or worsened (D-F) in the second study.

In the follow-up study, 5 patients with MHE remained in MHE, 5 died, and 4 improved. Three of these patients (PR51, A41, and A28) improved the PHES because of improved performance in attention tests and also showed increased MMN area (Fig. 4; Table 4). In 1 patient (PR27) who improved PHES because of better motor coordination without changes in attention tests, MMN area was not significantly altered (Table 4; Fig. 4).

Table 4. Changes in Different Parameters in Patients Who Improved (PR51, PR27, A41, and A28) or Worsened (A40, PR41, A49, and A23) in the Second Study
Patient No.Dates of First and Second StudyMMN AreaPHESCFFStroop CongruentStroop NeutralStroop IncongruentMap SearchBCTVCTDSTNCT-ANCT-BSDLTT
  1. MMN, mismatch negativity; CFF, critical flicker frequency; PHES, the Psychometric Hepatic Encephalopathy Score; DST, digit symbol test; NCT-A, number connection test A; NCT-B, number connection test B; SD, serial dotting test; LTT, line-tracing test; ns, not significant; BCT, bimanual coordination test; VCT, visuomotor coordination test.

PR51First study (April 2010)53.8−43786813161.932.30−20−200
Second study (March 2011)146037.888793172.022.4800000
PR27First study (Jan. 2010)108−441.2122845581.821.88000−2−2
Second study (March 2011)126−141.51119257121.782.08000−10
A41First study (March 2010)49−440.584724082.022.850−1−2−10
Second study (April 2011)80−239.589693582.002.930−1−100
A28First study (Nov. 2009)43.3−535.593673181.982.37−1−1−300
Second study (April 2011)162037.991623191.902.3500000
A40First study (April 2010)1840431138445   00000
Second study (June 2011)42−239867339   −10−100
PR41First study (Jan 2010)119039&00000
Second study (April 2011)45.8−2411046131101.82.40−1−100
A49First study (March 2010)88.314145392.452.9000100
Second study (June 2011)59−239.4443792.332.880−1−100
A23First study (Sep. 2009)59−34600−1−1−1
Second study (June 2011)25−8418661343.755.48−1−1−1−2−3

Four patients who did not show MHE in the first study (A40, PR41, A49, and A23) showed worse performance in attention tests in the second study, with reduced PHES that reached −8 (MHE) in 1 of them (A23). MMN area was reduced in these patients in parallel with deterioration of attention (Table 4; Fig. 4).

These data show that MMN area changed (i.e., increases or decreases), from the first to the second study, in parallel with changes (i.e., improvement or worsening) in performance in attention tests in the same patients.

Logistic regression analyses show that MMN area predicts performance in attention tests NCT-A (P = 0.002; 95% CI = 1.015-1.071), NCT-B (P < 0.0001; 95% CI = 1.010-1.035), and Stroop incongruent (P = 0.023; 95% CI = 1.003-1.030) and in the PHES (P < 0.001; 95% CI = 1.017-1.062). MMN area does not predict performance in visuomotor or bimanual coordination, in the Map Search, or in CFF.

ROC curves identified a cutoff of MMN area of 67 μV/ms, which predicts the presence of attention deficits in the NCT-A, NCT-B, and Stroop task, and MHE (PHES) with sensitivities of 90%, 75%, 80%, and 83%, respectively, and specificities of 76%, 83%, 76%, and 77%, respectively.

Discussion

The main contribution of this work, as discussed below, is the identification of a new tool, the determination of MMN area, that is useful to diagnose and follow the course of attention deficits and MHE in patients with liver cirrhosis.

The data reported also show that patients who do not show MHE, as detected using the PHES, already have some psychomotor slowing, as reflected by the reduced number of words and colors in the congruent and neutral tasks of the Stroop and increased time in the bimanual coordination test. This indicates that there are some mild neurological alterations not detected with the PHES and are detected by other procedures. This agrees with a report38 showing that ataxia, tremor, and slowing of finger movements are early markers for cerebral dysfunction in cirrhotic patients, even before alterations in performance in the PHES become detectable. This suggests that the PHES battery detects some “subtypes of MHE,” but not others.

Patients with MHE show much stronger alterations in the Stroop tasks and in bimanual coordination than patients without MHE. Moreover, they show other alterations not present in patients without MHE, including reduced area in the MMN wave, reduced performance in Map Search and elevator tests, indicating impairment of selective and sustained attention, respectively, and reduced performance in the visuomotor coordination test. This supports that patients with MHE have remarkable attention deficits.

Reduction of MMN area in patients with MHE is specifically associated with reduced performance in attention tests, but not with other alterations, such as motor coordination. This is supported by the results of patients who improved or worsened in the follow-up study. Patients PR51, A41, and A28 had MHE, mainly the result of impairment of attention (mainly NCT-B; Table 4). In the follow-up, they improved in attention tests, resulting in resolution of MHE and normalization of MMN area, which increased from 49 ± 3 to 130 ± 25. In contrast, patient PR27 did not show impairment in attention tests or in the MMN area (108.5) in the first study, and MHE was caused by impaired motor coordination, of which improvement led to resolution of MHE in the second study without changes in MMN area. This supports that reduction of MMN area in patients with MHE is associated with reduced performance in attention tests, but not with other alterations, such as motor coordination. Moreover, in the second study, MMN area was reduced in those patients (A40, PR41, A49, and A23) showing worsened performance in attention tests (Table 4; Fig. 4).

MMN area selectively predicts performance in attention tests and MHE, as shown by logistic regression analyses. ROC curves identified a cutoff of MMN area of 67 μV/ms, which predicts attention deficits and MHE with good sensitivities and specificities. Therefore, MMN is a good procedure, using routine assessment in neurophysiological settings, to diagnose attention deficits and MHE and follow their course in patients with liver cirrhosis.

Patients with MHE show a wide array of neurologic-neuropsychiatric alterations, including reduced attention, psychomotor slowing, reduced motor coordination, and so on. Each alteration is the result of impairment of different neuronal circuits and processes, of which modulation involves different brain areas, neurotransmitter systems, and mechanisms. Also, different pathogenic mechanisms could be involved in the different neurological alterations in the same patient. This is nicely illustrated by a recent report39 showing that in patients with MHE, alterations in the PHES performance strongly correlated with elevated inflammatory markers, but not with increased ammonia. However, EEG abnormalities correlated with high ammonia levels, but not with inflammation. This shows that different cerebral and neurological alterations are the result of different mechanisms. This has been demonstrated in more detail in animal models of MHE. Hypokinesia is caused by increased extracellular glutamate in substantia nigra,40 whereas impairment of learning a Y maze task is the result of reduced function of the glutamate/nitric oxide/cGMP pathway in the cerebellum.41

The MMN wave is generated by multiple neuronal elements. Latency depends on the neurons with faster response. Amplitude represents the maximum response of the sum of all neurons responding at the same time point. The area represents the accumulated response of all neurons from the beginning of the wave until its return to basal levels. In patients with MHE, latency and amplitude are not altered, but the area is reduced, indicating that the neurons respond in a similar way to control subjects, but a lower number of neurons are activated and during shorter periods.

Impairment of MMN in patients with MHE could be caused by similar mechanisms as in patients with schizophrenia. Understanding the mechanisms leading to attention deficits in MHE may help to design treatments to eliminate these deficits.

In summary, the data reported show that MMN is a good procedure, using routine neurophysiological techniques, to diagnose attention deficits and MHE with good sensitivity and specificity and follow their course in patients with liver cirrhosis.

Ancillary