Differentiation of minimal and grade 1 hepatic encephalopathy using neurophysiological, neuropsychological, magnetic resonance spectroscopic, and systemic inflammatory parameters

Abstract

BACKGROUND

Hepatic encephalopathy (HE) is traditionally classified as overt and minimal HE (MHE). The International Society for HE and Nitrogen Metabolism clubbed MHE and grade 1 HE as “covert” HE. However, these appear to be distinct entities.

AIM

To compare inflammatory markers, magnetic resonance spectroscopy (MRS), Bispectral index (BIS), spectral electroencephalography (EEG), and natural history of MHE and grade 1 HE.

METHODS

In this prospective study, 150 patients with cirrhosis were categorized into three groups: No HE (n = 50), MHE (n = 50), and grade 1 HE (n = 50). Patients were assessed using the West Haven criteria, the psychometric HE score, critical flicker frequency, MRS, inflammatory markers (cytokines and endotoxins), BIS, and spectral EEG. All patients were followed up for HE progression, complications, hospitalizations, and mortality over 12 months.

RESULTS

Patients with grade 1 HE showed significantly higher levels of ammonia (164.08 ± 30.12 vs 123.42 ± 19.97, P < 0.001), endotoxins (0.60 ± 0.05 vs 0.47 ± 0.04, P < 0.001), and cytokines than those with MHE. MRS in grade 1 HE revealed reduced myoinositol (-0.2 vs -0.11, P < 0.001) and choline levels (0.38 ± 0.33 vs 0.42 ± 0.18, P < 0.001) and elevated glutamine levels (1.01 ± 0.38 vs 0.73 ± 0.23, P < 0.001) compared to MHE. BIS values were lower in grade 1 HE compared to MHE (72.9 ± 5.49 vs 81.48 ± 3.9, P < 0.001). Spectral EEG metrics effectively distinguished MHE from grade 1 HE with high diagnostic accuracies. Model for end-stage liver disease, interleukin-12, BIS, and spectral EEG changes were independent predictors of survival.

CONCLUSION

MHE and grade 1 HE exhibit distinct inflammatory, neuropsychological, and neurophysiological profiles, as well as differences in their survival.

Key Words: Hepatic encephalopathy; Neurospectroscopy; Bispectral index; Systemic inflammation; Endotoxins; Critical flicker frequency; Psychometric hepatic encephalopathy score; Serum ammonia

Core Tip: Minimal hepatic encephalopathy (MHE) and grade 1 hepatic encephalopathy (HE) are often collectively referred to as covert HE. In this prospective study, we compared neuropsychological, neurophysiological, inflammatory, and magnetic resonance spectroscopic parameters in cirrhotic patients with no HE, MHE, and grade 1 HE. Grade 1 HE exhibited significantly higher inflammatory markers, greater cerebral metabolic changes, and more severe Bispectral index and electroencephalography abnormalities than MHE. Grade 1 HE was independently associated with an increased risk of progression to overt HE, hospitalization, and death. These findings support that MHE and grade 1 HE are distinct entities.

INTRODUCTION

Hepatic encephalopathy (HE) is a reversible neuropsychiatric syndrome caused by metabolic dysfunction in liver disease[1]. It causes significant hospitalization and mortality in patients with cirrhosis[2]. HE is graded according to the West Haven criteria into four stages based on severity. Minimal HE (MHE) is characterised by impaired neuropsychological and neurophysiological performance without overt signs. MHE affects quality of life and functioning and progresses to overt HE in approximately half of patients over 3 years, necessitating early detection[3,4]. The discrimination between MHE and grade 1 HE is unsatisfactory[5,6].

Given the challenges in diagnosing grade 1 HE, the International Society for HE and Nitrogen Metabolism classified HE into covert (MHE and grade 1 HE) and overt (grades 2-4) forms[2,7-9]. International Society for HE and Nitrogen Metabolism proposed “covert HE (CHE)” because grade 1 HE was difficult to diagnose owing to varying criteria and limited clinical applicability[2,7-9]. Evidence suggests that MHE and grade 1 HE may be distinct entities with different trajectories[10]. Grade 1 HE shows higher hospitalization and mortality rates than MHE, suggesting that these are separate entities according to the West Haven criteria[2,7-11].

CHE is a heterogeneous entity, including patients diagnosed with one or two abnormal neuropsychometric and/or neurophysiological tests, or including both patients with MHE and grade 1 HE[12-16]. Magnetic resonance spectroscopy (MRS) assesses cerebral metabolites in vivo and shows decreased choline and myoinositol levels and increased glutamine levels in HE[17-19]. These findings correlate with MHE and HE grades[17-19]. Inflammatory cytokines, such as tumor necrosis factor-alpha (TNF-α), modulate the neurotoxic effects of hyperammonemia[20]. TNF-α, interleukin-6 (IL-6), IL-18, and endotoxin values differ between MHE and grades 1-4 HE, correlating with HE severity and cognitive dysfunction[20-22]. Spectral electroencephalography (EEG) analysis, a computerized frequency distribution, aids in the diagnosis of MHE[23]. The mean dominant frequency (MDF), theta, and delta activity powers differ in MHE, grade 1 HE, and grade 2-4 HE[24]. The Bispectral index (BIS), used to monitor consciousness during surgery, helps grade HE and diagnose MHE[25,26]. This study aims to compare the clinical characteristics, inflammatory profiles, MRS, BIS, EEG parameters, and natural history and treatment outcomes of cirrhotics with MHE and grade 1 HE.

MATERIALS AND METHODS

Study population and data collection

This prospective study was conducted from November 2023 to April 2025 at the Department of Gastroenterology, Govind Ballabh Pant Institute of Postgraduate Medical Education and Research (New Delhi, India). Consecutive patients with liver cirrhosis aged from 18 years to 79 years, were screened for eligibility based on clinical features, liver function tests, ultrasonography, FibroScan, and histopathology (if available). After obtaining informed written consent, 150 patients were enrolled in this study. Patients were categorized into three groups: No HE (n = 50), MHE (n = 50), and grade 1 HE (n = 50), based on the psychometric HE score (PHES) and West Haven criteria (Figure 1). None of the patients had prospects for liver transplantation. The exclusion criteria were contraindications to MRS (e.g., magnetic resonance imaging-incompatible devices, claustrophobia), pregnancy, history of HE grade ≥ 2, recent (< 1 month) gastrointestinal bleeding, spontaneous bacterial peritonitis, hepatorenal syndrome, acute-on-chronic liver failure, recent (within 6 weeks) use of lactulose, rifaximin, L-ornithine-L-aspartate, probiotics, anti-inflammatory drugs (non-steroidal anti-inflammatory drugs, corticosteroids), or psychotropic medications, neurological or psychiatric illness, recent history of alcohol intake (in the last 3 months), poor vision, history of transjugular intrahepatic portosystemic shunt or shunt surgery, and ongoing infections or hepatocellular carcinoma.

Figure 1 Study flowchart. MRI: Magnetic resonance imaging; HE: Hepatic encephalopathy; GI: Gastrointestinal; SBP: Subacute bacterial peritonitis; HRS: Hepatorenal syndrome; HCC: Hepatocellular carcinoma; ACLF: Acute-on-chronic liver failure; LOLA: L-ornithine L-aspartate; NSAIDs: Non-steroidal anti-inflammatory drugs; TIPSS: Transjugular intrahepatic portosystemic shunt; PHES: Psychometric hepatic encephalopathy score; MHE: Minimal hepatic encephalopathy; CFF: Critical flicker frequency; MRS: Magnetic resonance spectroscopy; EEG: Electroencephalogram.

All patients underwent clinical assessment and laboratory investigations, including a hemogram, renal and liver function tests, serum electrolytes, international normalized ratio, random blood sugar, arterial ammonia levels, FibroScan, ultrasound, and contrast-enhanced computed tomography abdomen, if indicated. Disease severity was evaluated using the Child-Turcotte-Pugh (CTP) and model for end-stage liver disease (MELD) scores.

West Haven criteria was used to identify grade 1 HE, followed by PHES to identify MHE and no HE amongst those without grade 1 HE

Grade 1 HE was diagnosed clinically using the West Haven criteria[27]. MHE was diagnosed using the PHES, which includes the number/figure connection tests (A and B), the digit symbol test, the serial dotting test, and the line-tracing test. Patients who did not meet the grade 1 HE criteria underwent psychometric tests for PHES calculation. Those with PHES ≤ -5 were classified as having MHE, and those with PHES > -5 as having no MHE[28]. After diagnosis of grade 1 HE, MHE, and no HE, patients and investigators were blinded to the diagnosis for further investigations, treatment, and follow-up. PHES assessment was also performed in patients with grade 1 HE.

Critical flicker frequency measurement

The critical flicker frequency (CFF) was measured in all patients using a HEPAtonorm analyzer (R&R Medical Business Freiburg GmbH, Freiburg, Germany). The analyzer evoked an intrafoveal light stimulus with a defined pulse of light at a wavelength of 650 nm, a luminance of 270 cd/m², and a luminous intensity of 5.3 mcd. In this apparatus, the frequency of red light, initially generated as a high-frequency pulse (60 Hz), gives the patient the impression of a steady light, which then gradually reduces until the patient perceives a flicker. CFF was measured in a quiet, semi-darkened room after 5 minutes of adaptation. The patients were first instructed and trained on the procedure. The patient then placed the HEPAtonorm analyzer headset on their eyes. If the patient used eyeglasses, they could still wear them while using the headset. CFF threshold measurements were performed using intrafoveal stimulation with a luminous diode, decreasing the light frequency from 60 Hz. The CFF threshold was determined as the frequency at which the impression of the fused light changed to a flickering one. The frequency was recorded. The mean of the eight readings was considered the threshold value.

Inflammatory marker measurement

Blood samples were collected at baseline and stored at -80 °C after centrifugation. Cytokine (TNF-α, IL-6, IL-18, IL-10, IL-12, and IL-1β) and serum endotoxin levels were measured using enzyme-linked immunosorbent assay kits.

MR spectroscopy

Patients underwent brain MRS using a 3.0 Tesla PHILIPS INGENIA scanner. Single-voxel short echo time spectroscopy was performed in the medial occipital gray and parietal white matter regions. The quantified metabolites included myo-inositol, choline, creatinine, and glutamate.

BIS recording

BIS was recorded using a BIS monitor (Medtronic, 710 Medtronic Parkway, Minneapolis, MN 55432-5604, United States)[14,18]. The BIS QuatroTM sensor strip (Medtronic, 710 Medtronic Parkway, Minneapolis, MN 55432-5604, United States) was used for the study. Patients were instructed to remain calm, keep their eyes closed, avoid falling asleep, and try to keep their facial muscles completely relaxed. The BIS was recorded in a quiet environment for 10 minutes.

Spectral EEG recording and analysis

Spontaneous closed-eye EEG activity was recorded using a Nicolet V32 (Natus Medical Incorporated, 3150 Pleasant View Road, Middleton, WI 53562, United States) with Natus software (version 19). Attention was paid to avoid somnolence and artefacts. A standard 21-channel cap was used, and the electrodes were placed according to the International ‘10-20’ system. The ground was Fpz and the reference was Oz. After visual inspection to exclude artefacts, the EEG tracing was assessed by spectral analysis of 2-second epochs over a 60-second period. Spectral analysis was conducted on the derivation P3-P4 in the frequency range of 1-25.5 Hz. The spectral variables considered were as follows: (1) The MDF, which is the mean frequency weighted by the power of each frequency band; (2) The relative power in the beta (13.5-25.5 Hz), alpha (8.5-13 Hz), theta (4-8 Hz), and delta (1-3.5 Hz) bands; and (3) EEG Alterations: Grade one EEG alteration was defined as theta ≥ 35% with MDF > 7.3 Hz, grade two as MDF ≤ 7.3 Hz with delta < 44%, and grade three as MDF ≤ 7.3 Hz with delta ≥ 44%, as defined by Amodio et al[23].

All patients received standard care with lactulose therapy and underwent monthly evaluations for HE progression, gastrointestinal bleeding, infections, hospitalization, and mortality. Patients were reassessed at 12 months using the West Haven criteria, PHES, ammonia levels, CTP, and MELD scores.

Primary outcomes were: Incidence of HE progression (grade ≥ 2), mortality, serum levels of cytokines and endotoxin, cerebral concentrations of myoinositol, choline, creatinine, and glutamate, BIS values and EEG parameters (MDF, relative powers of alpha, beta, delta and theta frequency bands, and grade of EEG alteration), in no HE, MHE and grade 1 HE. The secondary outcomes were the development of complications, such as gastrointestinal bleeding, infections, and hospitalizations.

Statistical analysis

Categorical variables are expressed as n (%), and continuous data as mean ± SD or median (interquartile range). Normality was assessed using the Shapiro-Wilk test. Comparisons were performed using analysis of variance with post-hoc Tukey’s test or Kruskal-Wallis test with Dunn’s test. Receiver operating characteristic (ROC) analysis was used to evaluate the diagnostic performance of the cytokines, MRS peaks, BIS, EEG parameters, CFF, and ammonia levels. Kaplan-Meier analysis was used to assess survival and complication-free outcomes, and the log-rank test was used to compare groups. Cox regression (univariate and multivariate) was used to identify predictors of complications and mortality, and Poisson regression was used to assess the significant risk factors affecting the number of hospitalizations. All variables showing statistical significance on univariable analysis (P < 0.05) were entered into the multivariable model. Statistical analyses were performed using SPSS v25.0 (IBM, Chicago, IL, United States). P < 0.05 was considered significant. This study was approved by the Institutional Ethics Committee and was conducted in accordance with the principles outlined in the Declaration of Helsinki.

RESULTS

Demographics

Of 457 consecutive cirrhotic patients, 150 meeting the inclusion criteria were enrolled and stratified into three groups: No HE (n = 50), MHE (n = 50), and grade 1 HE (n = 50) groups. The mean age was 45 ± 10 years, and 85% of the patients were male. The etiology was alcohol-related (40.7%), metabolic dysfunction-associated steatotic liver disease (16%), hepatitis B (11.3%), and others (Table 1). Using the West Haven criteria, of the 50 patients with grade 1 HE, symptoms included lack of awareness in 15 (30%), euphoria or anxiety in 18 (36%), shortened attention span in 24 (48%), impaired attention in 25 (50%), and multiple symptoms in 32 (64%).

Table 1 Baseline characteristics, n (%)/mean ± SD.

Characteristic	No HE (n = 50)	Minimal HE (n = 50)	Grade 1 HE (n = 50)	P value
Age, years	44.28 ± 9.15	46.7 ± 11.02	44.18 ± 11.29	0.401
Gender, male	41 (82)	42 (84)	45 (90)
Etiology
Alcohol	18 (36)	17 (34)	26 (52)	0.264
MASLD	9 (18)	10 (20)	5 (10)
HBV	3 (6)	10 (20)	4 (8)
Idiopathic	5 (10)	3 (6)	4 (8)
HCV	4 (8)	2 (4)	4 (8)
Others	11 (22)	8 (16)	7 (14)
Baseline investigations
Hemoglobin (g/dL)	9.32 ± 2.17	8.81 ± 1.62	8.91 ± 1.86	0.363
TLC (cells/μL)	5612.6 ± 2970	6606 ± 4144	7804 ± 6057	0.582
Platelet count (cells/μL)	97180 ± 67191	100240 ± 47215	100200 ± 53333	0.549
Blood urea (mg/dL)	31.58 ± 11.15	37.24 ± 12.95	38.26 ± 12.63	0.015
Serum creatinine (mg/dL)	0.89 ± 0.23	1.02 ± 0.25	1.03 ± 0.3	0.013
Total protein (g/dL)	7.4 ± 0.66	6.82 ± 0.67	6.65 ± 0.91	< 0.001
Serum albumin (g/dL)	3.3 ± 0.57	2.86 ± 0.49	2.79 ± 0.51	< 0.001
Total bilirubin (mg/dL)	2.14 ± 2.26	4.1 ± 3.74	4.42 ± 3.89	< 0.001
AST (U/L)	86.14 ± 82.93	89.64 ± 40.5	110.8 ± 76.02	0.005
ALT (U/L)	46.3 ± 54.15	43.54 ± 18.73	49.84 ± 31.1	0.188
ALP (U/L)	157.96 ± 56.93	161.82 ± 86.38	160.52 ± 62.11	0.943
Sodium (mEq/L)	134.66 ± 5.02	133.54 ± 5.09	131.9 ± 4.59	0.02
Potassium (mEq/L)	4.15 ± 0.42	4.16 ± 0.62	4.15 ± 0.62	0.997
INR	1.23 ± 0.21	1.48 ± 0.37	1.65 ± 0.55	< 0.001
Serum ammonia (μg/dL)	87.18 ± 19.12	123.42 ± 19.97	164.08 ± 30.12	< 0.001
Baseline prognostic scores, PHES and CFF
MELD	11.28 ± 3.17	15.8 ± 4.36	17.58 ± 5.8	0.001
CTP score	7.9 ± 1.09	10.3 ± 1.56	11.16 ± 1.5	0.001
CTP grade (A/B/C), n	1/44/5	0/12/38	0/7/43	0.001
PHES, median (interquartile range)	-1 (-2 to 1)	-5 (-6 to -5)	-8 (-9 to -7)	< 0.001
CFF (Hz)	42.82 ± 2.99	37.52 ± 2.07	34 ± 3.55	< 0.001
Inflammatory markers
TNF-α (pg/mL)	25.33 ± 10.81	47.08 ± 13.99	64.31 ± 28,39	< 0.001
IL-6 (pg/mL)	22.68 ± 16.79	69.92 ± 42.23	113.92 ± 46.13	< 0.001
IL-18 (pg/mL)	171.53 ± 61.58	348.01 ± 86.26	519.41 ± 178.71	< 0.001
Endotoxin (EU/mL)	0.37 ± 0.05	0.47 ± 0.04	0.6 ± 0.05	< 0.001
IL-10 (pg/mL)	8.19 ± 2.19	13.72 ±2	16.43 ± 1.86	< 0.001
IL-12 (pg/mL)	6.9 ± 2.23	12.97 ± 2.11	17.77 ± 2.74	< 0.001
IL-1β (pg/mL)	5.1 ± 15.6	10.62 ± 1.81	14.95 ± 2.27	< 0.001
MRS peaks
Myoinositol (mmol/L), median (interquartile range)	0.1 (0.02-0.2)	-0.11 (-0.2 to -0.08)	-0.2 (-0.3 to -0.16)	< 0.001
Choline (mmol/L)	0.71 ± 0.18	0.42 ± 0.18	0.38 ± 0.33	< 0.001
Glutamine (mmol/L)	0.41 ± 0.15	0.73 ± 0.23	1.01 ± 0.38	< 0.001
Bispectral index
Bispectral index value	91.18 ± 3.43	81.48 ± 3.9	72.9 ± 5.49	< 0.001
Spectral EEG analysis
MDF (Hz)	12.9 ± 4.66	7.89 ±1.19	5.13 ± 1.57	< 0.001
Relative power of frequency bands
Beta (%)	24.54 ± 8.73	12.76 ± 5.97	10.3 ± 5.91	< 0.001
Alpha (%)	48.5 ± 13.16	24 ± 10.6	16 ± 8.15	< 0.001
Theta (%)	11.6 ± 4.34	30.08 ± 12.98	24.7 ±12.9	< 0.001
Delta (%)	13.74 ± 6.73	31.42 ± 14.23	46.96 ± 17.56	< 0.001
Grade of EEG alteration
Grade 0	50 (100)	23 (46)	3 (6)	< 0.001
Grade 1	0 (0)	11 (22)	0 (0)
Grade 2	0 (0)	12 (24)	17 (34)
Grade 3	0 (0)	4 (8)	30 (60)

HE: Hepatic encephalopathy; MASLD: Metabolic dysfunction-associated steatotic liver disease; HBV: Hepatitis B virus; HCV: Hepatitis C virus; TLC: Total leukocyte count; ALP: Alkaline phosphatase; ALT: Alanine aminotransferase; AST: Aspartate aminotransferase; INR: International normalized ratio; MELD: Model for end-stage liver disease; CTP: Child Turcotte Pugh; PHES: Psychometric hepatic encephalopathy score; CFF: Critical flicker frequency; TNF: Tumor necrosis factor; IL: Interleukin; MRS: Magnetic resonance spectroscopy; EEG: Electroencephalogram; MDF: Mean dominant frequency.

Baseline laboratory investigations

Serum bilirubin, creatinine and international normalized ratio were higher in grade 1 HE, while serum sodium levels were lower in grade 1 HE compared to MHE and no HE. The MELD and CTP scores progressively increased from no HE to MHE to grade 1 HE (Table 1).

Neuropsychological and neurophysiological evaluation

As expected, the median PHES scores progressively worsened as the HE grade increased, from -1 in the no-HE group to -5 in the MHE group and -8 in the grade 1 HE group (P < 0.001). Serum ammonia levels progressively increased from no HE (87.18 ± 19.12 μg/dL) to MHE (123.42 ± 19.97 μg/dL) to grade 1 HE (164.08 ± 30.12 μg/dL), while the CFF values progressively declined from no HE (42.82 ± 2.99 Hz) to MHE (37.52 ± 2.07 Hz) to grade 1 HE (34.00 ± 3.55 Hz) (Table 1).

Inflammatory marker profile

TNF-α, IL-6, IL-18, IL-10, IL-12, IL-1β, and endotoxin levels were higher in grade 1 HE than in MHE and no HE (P < 0.001) (Table 1). According to ROC analysis, IL-1β (area under the ROC: 0.993) best predicted MHE, and endotoxin (area under the ROC: 0.989) predicted grade 1 HE (Figure 2).

Figure 2 Receiver operating characteristic curves of inflammatory markers for predicting minimal hepatic encephalopathy. A: Tumor necrosis factor-alpha [area under the receiver operating characteristic (AUROC): 0.934], interleukin-6 (IL-6) (AUROC: 0.864), IL-18 (AUROC: 0.962), endotoxin (AUROC: 0.976), IL-10 (AUROC: 0.976), IL-12 (AUROC: 0.984), IL-1β (AUROC: 0.993); B: Grade 1 hepatic encephalopathy [tumor necrosis factor-alpha (AUROC: 0.729)], IL-6 (AUROC: 0.751), IL-18 (AUROC: 0.812), endotoxin (AUROC: 0.989), IL-10 (AUROC: 0.833), IL-12 (AUROC: 0.921), IL-1β (AUROC: 0.938). TNF-α: Tumor necrosis factor-alpha; IL: Interleukin.

MRS

MRS showed lower myoinositol and choline peaks and higher glutamine peaks in patients with grade 1 HE than in those with MHE or without HE (P < 0.001). The myoinositol peak best predicted MHE, whereas the glutamine peak best predicted grade 1 HE (Table 1, Supplementary Figure 1).

BIS

BIS values were significantly lower in grade 1 HE than in MHE and no HE (Table 1).

Spectral EEG analysis

On spectral EEG analysis, the MDF was highest in no HE (12.9 ± 4.66 Hz), followed by MHE (7.89 ± 1.19 Hz), and lowest in grade 1 HE (5.13 ± 1.57 Hz) (Table 1). MDF predicted MHE and grade 1 HE at cut-off points of ≤ 8.9 and ≤ 6.6 Hz, respectively (Figure 3). Significant differences in β, α, θ, and δ relative powers were observed across the no HE, MHE, and grade 1 HE groups (Table 1). θ best predicted MHE at a cutoff of > 18, and δ predicted grade 1 HE at a cutoff of > 40 (Figure 3). All patients without HE had grade 0 EEG alteration, whereas patients with grade 1 HE had grade 2 or 3 EEG alterations. Patients with MHE had any grade of EEG alteration (Supplementary Figure 2).

Figure 3 Receiver operating characteristic curves of Bispectral index, mean dominant frequency, relative powers of electroencephalogram frequencies, and magnetic resonance spectroscopy peak values in predicting minimal hepatic encephalopathy. A: Bispectral index [area under the receiver operating characteristic (AUROC: 0.911)], mean dominant frequency (AUROC: 0.876), beta (AUROC: 0.876), alpha (AUROC: 0.876), delta (AUROC: 0.876), theta (AUROC: 0.886), myoinositol (AUROC: 0.911), choline (AUROC: 0.876), glutamine (AUROC: 0.886); B: Grade 1 hepatic encephalopathy [Bispectral index (AUROC: 0.911)], mean dominant frequency (AUROC: 0.876), beta (AUROC: 0.876), alpha (AUROC: 0.876), delta (AUROC: 0.876), theta (AUROC: 0.886), myoinositol (AUROC: 0.739), choline (AUROC: 0.681), glutamine (AUROC: 0.741). BIS: Bispectral index; MDF: Mean dominant frequency.

ROC curves

ROC curve analysis identified distinct cut-off values for predicting MHE and grade 1 HE. Compared with MHE, grade 1 HE required higher thresholds for ammonia, inflammatory cytokines (TNF-α, IL-6, IL-18, IL-10, IL-12, IL-1β), endotoxin, and glutamine, reflecting greater systemic inflammation and metabolic derangement. Conversely, lower cut-offs for myoinositol, choline, BIS, MDF, β and α power, and more pronounced increases in δ and θ activity were observed in grade 1 HE. Overall, progression from MHE to grade 1 HE was characterized by worsening biochemical, inflammatory, metabolic, and neurophysiological abnormalities (Figures 2 and 3, Supplementary Figure 3).

Survival analysis

Gastrointestinal bleeding, infections, hospitalization, and HE progression (grade ≥ 2) were significantly higher in grade 1 HE than in MHE and no HE after 12-month follow-up (Table 2). Kaplan-Meier analyses showed reduced overall survival, progression-free survival, complication-free survival, and hospitalization-free survival in patients with grade 1 HE compared to those with MHE and no HE. Over 12 months of follow-up, overall survival (90% in no HE, 72% in MHE, and 58% in grade 1 HE groups), progression-free survival (80% in no HE, 52% in MHE, and 38% in grade 1 HE groups), complication-free survival (66% in no HE, 46% in MHE, and 26% in grade 1 HE groups), and hospitalization-free survival (72% in no HE, 56% in MHE, and 32% in grade 1 HE groups) were lower in grade 1 HE (Figure 4).

Figure 4 Kaplan-Meier curves comparing no hepatic encephalopathy, minimal hepatic encephalopathy, and grade 1 hepatic encephalopathy for overall survival, progression of hepatic encephalopathy (≥ 2) free survival, complications-free survival, and hospitalization-free survival. HE: Hepatic encephalopathy.

Table 2 Complications and mortality over 1 year, n (%).

	No HE (n = 50)	Minimal HE (n = 50)	Grade 1 HE (n = 50)	P value
GI bleed	10 (20)	11 (22)	15 (30)	< 0.001
Infections	4 (8)	7 (14)	13 (26)	< 0.001
Progression of HE (to grade ≥ 2)	10 (20)	24 (48)	31 (62)	< 0.001
More than one complication	8 (16)	17 (34)	30 (60)	< 0.001
Hospitalisation	14 (28)	22 (44)	34 (68)	< 0.001
Death	5 (10)	14 (28)	21 (42)	< 0.001

GI: Gastrointestinal; HE: Hepatic encephalopathy.

Regression analysis

After multivariate analysis, the predictors of complications were the MELD score, grade 1 EEG alteration, and grade 1 HE (Table 3). The predictors of HE progression were grade 1 HE, MELD score, IL-12 level, BIS, and delta relative power (Table 4), and grade 1 HE, MELD score, CFF, and IL-12 level were predictors of mortality (Table 5). IL-12 level and grade 1 HE were predictors of hospitalization (Table 6 and Figure 5).

Figure 5 Independent risk factors for complications, hepatic encephalopathy progression (≥ grade 2), hospitalization, and mortality on multivariate analysis. HE: Hepatic encephalopathy; MELD: Model for end-stage liver disease; EEG: Electroencephalogram; IL: Interleukin; BIS: Bispectral index; CFF: Critical flicker frequency.

Table 3 Univariate and multivariate regression analysis to identify significant risk factors of complications.

Parameters	Univariate hazard ratio by the Cox regression model		Multivariate hazard ratio
	HR (95%CI)	P value	HR (95%CI)	P value
Age	0.989 (0.968-1.011)	0.328
Hemoglobin	0.966 (0.865-1.080)	0.545
TLC	1.000 (1.000-1.000)	0.026	1.000 (1.000-1.000)	0.597
Platelets	1.000 (1.000-1.000)	0.223
Blood urea	1.017 (1.000-1.035)	0.051
Creatinine	4.429 (2.159-9.086)	< 0.0001	1.434 (0.546-3.765)	0.464
Total protein	0.631 (0.486-0.818)	0.001	0.920 (0.660-1.283)	0.623
Serum albumin	0.489 (0.327-0.732)	0.001	0.950 (0.539-1.676)	0.860
Total bilirubin	1.085 (1.041-1.131)	0.0001	0.954 (0.866-1.052)	0.345
AST	1.006 (1.003-1.008)	< 0.0001	1.004 (0.998-1.010)	0.166
ALT	1.008 (1.002-1.014)	0.007	0.998 (0.986-1.010)	0.998
ALP	1.001 (0.999-1.004)	0.279
Na	0.949 (0.908-0.991)	0.018	0.971 (0.920-1.025)	0.283
K	1.014 (0.678-1.515)	0.947
INR	1.716 (1.253-2.350)	0.001	0.349 (0.159-0.766)	0.009
Serum ammonia	1.008 (1.003-1.014)	0.002	0.993 (0.983-1.004)	0.233
CFF	0.911 (0.868-0.975)	0.0002	0.933 (0.855-1.019)	0.122
MELD	1.110 (1.076-1.145)	< 0.0001	1.209 (1.104-1.322)	< 0.0001
Myoinositol peak	0.154 (0.045-0.525)	0.003	2.330 (0.212-25.657)	0.490
Choline peak	0.334 (0.126-0.887)	0.028	0.994 (0.291-3.396)	0.992
Glutamine peak	2.004 (1.170-3.433)	0.011	0.949 (0.392-2.300)	0.908
TNF-α	1.013 (1.006-1.021)	0.0003	0.995 (0.981-1.010)	0.517
IL-6	1.005 (1.001-1.09)	0.012	0.998 (0.992-1.004)	0.513
IL-18	1.001 (1.000-1.002)	0.008	1.000 (0.997-1.003)	0.939
Endotoxin	43.315 (6.576-285.318)	0.0001	3.181 (0.002-4085.938)	0.751
IL-10	1.136 (1.071-1.204)	< 0.0001	1.145 (0.950-1.381)	0.156
IL-12	1.083 (1.038-1.130)	0.0002	0.823 (0.676-1.002)	0.053
IL-1β	1.101 (1.049-1.156)	0.0001	1.171 (0.967-1.418)	0.107
Gender
Female	1
Male	1.598 (0.799-3.196)	0.185
CTP grade
A	1
B	1654.812 (0.000-4.83E+60)	0.913
C	5289.119 (0.000-1.54E+61)	0.899
Grade of HE
No HE	1
Minimal HE	1.756 (0.957-3.223)	0.069
Grade 1 HE	3.018 (1.696-5.373)	0.0002	3 (1.25-7.20)	0.014

HR: Hazard ratio; CI: Confidence interval; TLC: Total leukocyte count; ALP: Alkaline phosphatase; ALT: Alanine aminotransferase; AST: Aspartate aminotransferase; INR: International normalized ratio; CFF: Critical flicker frequency; MELD: Model for end-stage liver disease; TNF: Tumor necrosis factor; IL: Interleukin; CTP: Child Turcotte Pugh; HE: Hepatic encephalopathy.

Table 4 Univariate and multivariate regression analysis to identify significant risk factors of progression of hepatic encephalopathy (hepatic encephalopathy ≥ grade 2).

Parameters	Univariate hazard ratio by the Cox regression model		Multivariate hazard ratio
	HR (95%CI)	P value	HR (95%CI)	P value
Age	1.009 (0.986-1.032)	0.441
Hemoglobin	1.005 (0.888-1.138)	0.937
TLC	1.000 (1.000-1.000)	0.055
Platelets	1.000 (1.000-1.000)	0.163
Blood urea	1.019 (1.000-1.038)	0.056
Creatinine	3.800 (1.760-8.208)	0.0007	1.293 (0.431-3.880)	0.646
Total protein	0.634 (0.475-0.845)	0.002	0.951 (0.647-1.399)	0.800
Serum albumin	0.497 (0.315-0.784)	0.003	1.123 (0.590-2.139)	0.724
Total bilirubin	1.092 (1.043-1.144)	0.0002	0.924 (0.825-1.034)	0.169
AST	1.006 (1.003-1.008)	0.0002	1.004 (0.999-1.008)	0.523
ALT	1.003 (0.998-1.009)	0.221
ALP	1.002 (0.99-1.005)	0.279
Na	0.945 (0.901-0.992)	0.022	0.969 (0.914-1.028)	0.294
K	0.981 (0.625-1.540)	0.934
INR	1.925 (1.390-2.664)	0.0001	0.443 (0.202-0.969)	0.041
Serum ammonia	1.010 (1.004-1.017)	0.001	0.996 (0.985-1.008)	0.523
CFF	0.909 (0.862-0.959)	0.0004	0.942 (0.855-1.039)	0.232
MELD	1.122 (1.083-1.161)	< 0.0001	1.211	< 0.0001
Myoinositol peak	0.059 (0.015-0.234)	0.0001	0.284 (0.023-3.549)	0.329
Choline peak	0.211 (0.069-0.645)	0.006	0.687 (0.165-2.858)	0.606
Glutamine peak	2.227 (1.244-3.986)	0.007	0.740 (0.287-1.904)	0.532
TNF-α	1.012 (1.005-1.020)	0.002	0.997 (0.981-1.012)	0.668
IL-6	1.005 (1.001-1.010)	0.012	0.998 (0.991-1.004)	0.521
IL-18	1.001 (1.000-1.003)	0.014	1.000 (0.997-1.003)	0.972
Endotoxin	58.094 (6.844-493.151)	0.0002	0.731 (0.000-1717.320)	0.937
IL-10	1.149 (1.075-1.229)	< 0.0001	1.122 (0.906-1.388)	0.291
IL-12	1.089 (1.038-1.142)	0.0004	0.786 (0.623-0.990)	0.041
IL-1β	1.116 (1.057-1.179)	0.0001	1.142 (0.926-1.410)	0.215
Gender
Female	1
Male	1.357 (0.647-2.844)	00.419
CTP grade
A	1
B	1.46E+03 (0.000-1.20E+71)	0.927
C	5.57E+03 (0.000-4.57E+71)	0.914
Grade of HE
No HE	1
Minimal HE	2.783 (1.330-5.822)	0.007	1.499 (0.327-6.878)	0.603
Grade 1 HE	4.200 (2.056-8.580)	0.0001	2.23	00.02

HR: Hazard ratio; CI: Confidence interval; TLC: Total leukocyte count; ALP: Alkaline phosphatase; ALT: Alanine aminotransferase; AST: Aspartate aminotransferase; INR: International normalized ratio; CFF: Critical flicker frequency; MELD: Model for end-stage liver disease; TNF: Tumor necrosis factor; IL: Interleukin; CTP: Child Turcotte Pugh; HE: Hepatic encephalopathy.

Table 5 Univariate and multivariate regression analysis to identify significant risk factors of mortality.

Parameters	Univariate hazard ratio by the Cox regression model		Multivariate hazard ratio
	HR (95%CI)	P value	HR (95%CI)	P value
Age	1.025 (0.997-1.054)	0.081
Hemoglobin	0.917 (0.782-1.076)	0.289
TLC	1.000 (1.000-1.000)	0.473
Platelets	1.000 (1.000-1.000)	0.158
Blood urea	1.049 (1.025-1.074)	0.0001	1.008 (0.963-1.056)	0.726
Creatinine	7.596 (3.170-18.205)	< 0.0001	0.842 (0.133-5.346)	0.856
Total Protein	0.367 (0.251-0.537)	< 0.0001	0.495 (0.279-0.881)	0.017
Serum albumin	0.359 (0.197-0.655)	0.001	1.243 (0.502-3.082)	0.638
Total bilirubin	1.073 (1.008-1.143)	0.028	0.865 (0.725-1.031)	0.106
AST	1.003 (1.000-1.007)	0.039	1.001 (0.995-1.07)	0.740
ALT	1.002 (0.996-1.009)	0.517
ALP	0.998 (0.993-1.03)	0.544
Na	0.906 (0.852-0.963)	0.002	0.920 (0.852-0.994)	0.035
K	1.678 (0.994-2.834)	0.053
INR	1.948 (1.281-2.963)	0.002	0.338 (0.121-0.942)	0.038
Serum ammonia	1.012 (1.005-1.020)	0.001	0.994 (0.977-1.010)	0.449
CFF	0.895 (0.836-0.959)	0.002	0.895 (0.773-1.038)	0.142
MELD	1.132 (1.084-1.183)	< 0.0001	1.263	0.000
Myoinositol peak	0.080 (0.014-1.183)	0.005	0.759 (0.015-39.025)	0.891
Choline peak	0.245 (0.059-1.017)	0.053
Glutamine peak	2.521 (1.201-5.295)	0.015	0.692 (0.180-2.662)	0.593
TNF-α	1.015 (1.005-1.024)	0.003	1.003 (0.984-1.022)	0.777
IL-6	1.006 (1.000-1.011)	0.036	0.999 (0.990-1.007)	0.776
IL-18	1.001 (1.00-1.003)	0.137
Endotoxin	123.151 (7.777-1.95E+03)	0.0006	9445.964 (0.220-3.97E+08)	0.093
IL-10	1.149 (1.053-1.255)	0.002	1.022 (0.757-1.381)	0.886
IL-12	1.085 (1.021-1.154)	0.009	0.713 (0.532-0.956)	0.024
IL-1β	1.115 (1.040-1.195)	0.002	1.206 (0.901-1.614)	0.208
Gender
Female	1
Male	1.044 (0.438-2.486)	0.923
CTP grade
A	1
B	1180.167 (0.000-6.66E+95)	0.948
C	5819.475 (3.28E+96)	0.937
Grade of HE
No HE	1
Minimal HE	3.069 (1.105-8.524)	0.031	1.333 (0.165-10.745)	0.787
Grade 1 HE	5.155 (1.942-13.683)	0.001	2.34	0.008

HR: Hazard ratio; CI: Confidence interval; TLC: Total leukocyte count; ALP: Alkaline phosphatase; ALT: Alanine aminotransferase; AST: Aspartate aminotransferase; INR: International normalized ratio; CFF: Critical flicker frequency; MELD: Model for end-stage liver disease; TNF: Tumor necrosis factor; IL: Interleukin; CTP: Child Turcotte Pugh; HE: Hepatic encephalopathy.

Table 6 Univariate and multivariate regression analysis to identify significant risk factors of number of hospitalisations.

Parameters	Univariate incidence rate ratio by the Poisson regression model		Multivariate incidence rate ratio
	Incidence rate ratio (95%CI)	P value	Incidence rate ratio (95%CI)	P value
Age	0.994 (0.973-1.015)	0.574
Hemoglobin	0.949 (0.846-1.063)	0.365
TLC	1.000 (1.000-1.000)	0.115
Platelet count	1.000 (1.000-1.000)	0.323
Blood urea	1.021 (1.004-1.039)	0.017	1.003 (0.979-1.028)	0.817
Creatinine	3.012 (1.502-6.039)	0.002	1.267 (0.433-3.709)	0.666
Total Protein	0.650 (0.501-0.843)	0.001	0.934 (0.667-1.309)	0.692
Serum albumin	0.548 (0.364-0.826)	0.004	0.825 (0.487-1.395)	0.472
Total bilirubin	1.056 (1.008-1.107)	0.023	0.989 (0.892-1.097)	0.836
AST	1.002 (1.000-1.005)	0.097
ALT	1.002 (0.997-1.007)	0.356
ALP	1.001 (0.998-1.004)	0.418
Na	0.961 (0.922-1.001)	0.056
K	0.923 (0.622-1.368)	0.689
INR	1.573 (1.094-2.262)	0.014
Serum ammonia	1.008 (1.003-1.013)	0.003	0.999 (0.990-1.009)	0.917
CFF	0.919 (0.876-0.965)	0.001	0.848 (0.875-1.026)	0.186
MELD	1.069 (1.035-1.104)	0.0001	1.068 (0.975-1.171)	0.157
Myoinositol peak	0.206 (0.059-0.721)	0.013	4.348 (0.411-46.003)	0.222
Choline peak	0.513 (0.206-1.280)	0.152
Glutamine peak	2.053 (1.190-3.542)	0.010	1.081 (0.496-2.357)	0.844
TNF-α	1.011 (1.005-1.017)	0.0004	1.005 (0.995-1.014)	0.345
IL-6	1.006 (1.002-1.009)	0.003	1.002 (0.997-1.07)	0.378
IL-18	1.001 (1.000-1.002)	0.007	1.000 (0.997-1.002)	0.817
Endotoxin	31.751 (4.610-218.657)	0.0004	0.356 (0.001-245.301)	0.757
IL-10	1.123 (1.056-1.194)	0.0002	1.035 (0.872-1.228)	0.698
IL-12	1.085 (1.038-1.134)	0.0003	0.713 (0.754-1.072)	0.024
IL-1β	1.103 (1.050-1.159)	0.0001	1.170 (0.971-1.411)	0.100
Gender
Female
Male	0.981 (0.514-1.873)	0.954
CTP grade
A
B	1.72E+08 (0.000-7.12E+88)	0.841
C	3.44E+08 (0.000-1.71E+101)	0.857
Grade of HE
No HE
Minimal HE	1.056 (0.467-2.386)	0.896
Grade 1 HE	2.45 (1.22-4.90)	0.012	2.1 (1.24-3.56)	0.006

HR: Hazard ratio; CI: Confidence interval; TLC: Total leukocyte count; ALP: Alkaline phosphatase; ALT: Alanine aminotransferase; AST: Aspartate aminotransferase; INR: International normalized ratio; CFF: Critical flicker frequency; MELD: Model for end-stage liver disease; TNF: Tumor necrosis factor; IL: Interleukin; CTP: Child Turcotte Pugh; HE: Hepatic encephalopathy.

DISCUSSION

Patients with MHE and grade 1 HE show distinct clinical, neurophysiological, and neuropsychiatric features, complications, and mortality rates. This study highlights the prognostic relevance of MHE and grade 1 HE and the role of inflammatory markers, MRS, BIS, and spectral EEG in diagnosis and risk stratification. MHE and grade 1 HE are distinct entities that cannot be combined with CHE. There is no consensus on the testing strategies for CHE[9,29,30]. Flud and Duarte-Rojo[12] found confusion due to a lack of a formal CHE definition, uncertainty about diagnostic tools, and disagreement over whether one or two tests were needed for diagnosis. Studies have shown variations in the diagnostic criteria used for CHE[13,25-32].

In previous studies, under CHE grading, both patients with grade 1 HE, as defined by the West Haven criteria, and those with MHE, as defined by the PHES and/or CFF, were included[14-16]. Some studies employed two tests (neuropsychological and neurophysiological) or one test (neuropsychological or neurophysiological) to diagnose CHE, as previously used for diagnosing MHE[13,18,31-33]. Montagnese et al[9] reported that grade 1 HE, diagnosed using the West Haven criteria, is associated with HE episodes on follow-up, higher ammonia levels, alterations in psychometric and neurophysiological indices, EEG/PHES abnormalities, higher MELD scores, and worse prognosis, which corresponds with the results of our study.

Liver dysfunction (MELD and CTP scores) worsened with increasing HE severity. These scores are associated with progression to overt HE, complications, and mortality, thereby confirming their role in predicting HE risk[4]. The PHES, which is traditionally used to detect MHE, is inversely correlated with HE severity. Our study suggests that the PHES can aid in identifying grade 1 HE when interpreted at lower cut-offs, although further validation is required. Serum ammonia levels were significantly correlated with HE severity. ROC analysis confirmed the test's diagnostic utility, with high predictive accuracy for MHE and grade 1 HE. These findings align with the pathophysiology of HE, where ammonia-induced astrocyte swelling and cerebral edema impair neurotransmission[34]. However, due to overlap across clinical stages, ammonia should be used in conjunction with neurophysiological tools and biomarkers[35].

CFF testing, a simple, non-invasive tool reflecting cortical processing, showed outstanding performance in diagnosing minimal and grade 1 HE. With specific thresholds, CFF effectively stratified patients and provided high predictive value. Its functional assessment complements the pathophysiological role of ammonia, making CFF a clinically relevant diagnostic tool.

Neuroinflammation is a central pathogenic mechanism of HE. Cytokine and endotoxin levels increased with HE severity. IL-1β and endotoxins best discriminated between MHE and grade 1 HE. These results align with studies implicating cytokine cascades in HE and highlight the role of gut-derived endotoxemia in neuropsychiatric dysfunction[20,21].

In a previous study, the MRS values of myoinositol and choline were lower, and glutamine was higher in the MHE vs no HE group (P < 0.001)[19]. In a meta-analysis of 31 studies conducted by Zeng et al[18] the parietal region was the most reliable region for differentiating between patients with and without HE. The glutamine/glutamate ratio correlated with HE grades in all brain regions. Oeltzschner et al[36] reported mean visual Gsx/Cr peaks of 0.260 (0.071) and 0.235 (0.055) for MHE and grade 1 HE, respectively. The mean sensorimotor Gsx/Cr peaks were 0.2480 (0.052) for MHE and 0.2242 (0.045) for grade 1 HE. In the present study, decreased myoinositol and choline (osmotic and membrane markers) and increased glutamine (an ammonia detoxification marker) levels were observed in patients with varying HE severity. The myoinositol peak best predicted MHE, whereas the glutamine peak best predicted grade 1 HE. These findings reflect astrocytic swelling and osmoregulatory failure, which are key features of HE neuropathology[17,19,37]. The BIS has shown a good correlation with progressive grades of HE in previous studies. Different cut-offs have been obtained in different studies for each grade. Only one previous study had also included MHE patients[26]. This study showed that the BIS score correlated with progressive HE grades, with different cut-off values for MHE and grade 1 HE, as seen in our study.

Spectral EEG parameters exhibited excellent diagnostic performance for MHE and grade 1 HE. EEG grading correlated with HE grade. Therefore, EEG is a valuable tool for distinguishing between MHE and grade 1 HE. A previous study found that patients with MHE had lower MDF, higher theta power, and lower alpha power than those without MHE[38]. In the current study, the relative powers of beta and alpha were highest in the no HE group, followed by the MHE and grade 1 HE groups. Theta relative power was higher in MHE, and delta relative power was higher in grade 1 HE, as previously shown. Inflammatory markers, MR spectroscopy, BIS and spectral EEG can thus aid in the recognition of MHE and grade 1 HE, in addition to the already existing wide range of tests available[39].

Follow-up revealed differences between MHE and grade 1 HE. Patients with grade 1 HE showed higher rates of progression to overt HE, complications (infections and gastrointestinal bleeding), hospitalization, and mortality than those with minimal and no HE. These findings are consistent with a multicentre study from Europe and United States, by Gairing et al[40], which showed that MHE is associated with higher risk of progression to overt HE and poorer liver transplant free survival. Multivariate Cox regression analysis identified that grade 1 HE and MELD score were independent predictors of disease progression, complications, and mortality. Elevated IL-12 levels were paradoxically associated with a reduced risk of progression, potentially indicating a regulatory immune response. These findings suggest that HE has a multifactorial pathogenesis, and indicate that targeting inflammation and cerebral metabolism could improve outcomes[41]. This study has a few limitations: It is conducted at a tertiary care centre, and patients with early cirrhosis (CTP grade A) were underrepresented, likely due to referral bias. Some of the study parameters, such as PHES, require a basic education level, which was not assessed. However, the patients were informed in detail about each test, and the tests were performed only after proper understanding. This study represents the first comprehensive comparison of the natural history, survival rates, inflammatory markers, CFF, MRS characteristics, BIS, and EEG parameters in cirrhotic patients with no HE, MHE, or grade 1 HE. This highlights the differences between MHE and grade 1 HE, suggesting that MHE and grade 1 HE cannot be classified as CHE.

CONCLUSION

This study demonstrates that minimal and grade 1 HE are distinct clinical entities. Inflammatory markers, such as IL-1β and endotoxins, MRS metabolites, including myoinositol and glutamine, the BIS, and spectral EEG variables, provide objective diagnostic and prognostic differences between MHE and grade 1 HE. Patients with grade 1 HE showed a significantly higher risk of progression, complications, hospitalization, and mortality than those with MHE. Incorporating CFF, MRS, inflammatory profiling, BIS, and spectral EEG into clinical practice could enhance the differentiation between MHE and grade 1 HE, leading to targeted interventions and improved patient outcomes in patients with cirrhosis.