Abstract
Background & Aims:
Patients with cirrhosis are growing older. The overlap between minimal hepatic encephalopathy (MHE) and pre-dementia mild cognitive impairment (MCI) could affect quality of life (QOL). We investigated the performance of elderly patients with cirrhosis on tests for MHE and MCI and their effects on QOL.
Methods:
We recruited outpatients with cirrhosis (n=109) and without cirrhosis (controls, n=100), 65 y or older, at 4 centers (derivation cohort). All study participants were assessed for psychometric hepatic encephalopathy score (PHES), EncephalApp score, and QOL. MCI was tested in patients with cirrhosis using the repeatable battery for assessment of neuropsychological status and assigned to the following groups: unimpaired, MCI-only, MHE-only, and MCI+MHE. We created adjusted norms to detect MHE using PHES and EncephalApp scores from the controls. Findings were validated using data from a separate cohort of 77 patients with cirrhosis (mean age, 69.49±4.36y; 72% men) at the same study sites.
Results:
Controls were older but were more educated, performed better cognitively, and had better QOL. Among patients with cirrhosis, age, education, model for end-stage liver disease score, EncephalApp score, and QOL were similar but PHES and repeatable battery for assessment of neuropsychological status differed among sites. In the derivation cohort, presence of MHE, with or without MCI, associated with poor QOL, which was lowest in the MCI+MHE group. When we adjusted for age, sex, and education, 49% of patients with cirrhosis had MHE based on the EncephalApp and 8% had MHE based on the PHES. A similar pattern (49% MHE based on EncephalApp and 6% MHE based on PHES) was found in a validation cohort.
Conclusions:
In a multi-center study of patients with cirrhosis (older than 65 y) and controls, the presence of MHE, regardless of MCI, associated with poor cognition and QOL. We created adjusted norms that defined the high sensitivity of EncephalApp for detection of MHE in older individuals and validated it in a separate cohort.
Keywords: dementia, aging, memory, daily function, transplant selection
Introduction
Patients with cirrhosis are growing older and increasing survival means that these patients are often diagnosed with cirrhosis at a later age1–4. This also implies a greater burden of co-morbid conditions that can negatively impact prognosis2. Cirrhosis and co-morbid conditions can negatively impact quality of life (QOL), which can also worsen with age5, 6. Cirrhosis-associated altered cognition and impaired QOL is usually related to hepatic encephalopathy (HE)7. The minimal HE (MHE) stage can only be diagnosed using specialized tests and does not affect memory7. However, with advancing age, it is important to analyze the potential overlap with age-related mild cognitive impairment (MCI), which involves memory impairment8. A single-center study showed significant difference in the gut-liver-brain axis between cirrhotic patients with/without MCI, but a multi-center approach is required9. Also, the performance of specific MHE-testing strategies requires investigation in individuals >65 years10.
Our aims were to (a) determine the performance of cirrhotic patients on tests focused on MHE and MCI compared to controls (b) determine the impact on QOL and (c) define norms for MHE tests in this population, in patients with cirrhosis and non-cirrhotic controls who were >65 years of age.
Methods:
We included outpatients with/without cirrhosis between 65–95 years from four centers [Virginia Commonwealth University (VCU), McGuire VA Medical Center (VAMC), University of Arkansas and Mayo Clinic, Phoenix] after IRB approval. Cirrhosis was diagnosed using liver biopsy, radiologic or elastographic evidence, or varices or portal. We excluded patients with prior/current overt HE, lactulose/rifaximin therapy, color-blindness, psycho-active drug use apart from >3 months opioids or anti-depressants11. Controls were recruited via word of mouth or advertisements, were between 65–95 years, without significant medical history (Supplementary data for details) and not on psychoactive drugs. After consent, mini-mental status exam (MMSE) was administered and only those with MMSE>25 continued12.
Cognitive testing:
MHE tests were Psychometric hepatic encephalopathy score (PHES) and EncephalApp Stroop10, 13. PHES has 5 tests (number connection A/B, digit symbol, serial dotting and line tracing). A lower composite score is considered impaired7. EncephalApp consists of an easier “Off State” where subjects have to identify the correct color of pound signs presented and the more difficult “On State” where there is an incongruity between the written word, and the color of the letters used to spell the stimulus. Outcomes were OffTime, OnTime, number of runs in both states to complete 5 correct runs and OffTime+OnTime.
In cirrhotic patients, we also administered Repeatable Battery for Assessment of Neuropsychological Status (RBANS) that assesses 5 cognitive domains (i.e., immediate memory, delayed memory, attention, visuo-spatial ability and language skill). Cortical domains are delayed memory and language while attention and visuo-spatial ability are considered subcortical. Higher RBANS scores indicate better cognitive performance14.
Evaluation of MCI vs MHE vs both:
A board-certified neuropsychologist (JBW) evaluated the RBANS of all cirrhotic patients and classified them according to prior publications into unimpaired, pattern consistent with MHE only (impaired subcortical [attention and visuo-spatial] and relatively normal cortical [delayed memory and language] scores), MCI only (poor performance on cortical but not subcortical), and a group classification (MCI+MHE) where subjects had difficulty on both cortical/subcortical measures of RBANS9, 15. We compared clinical variables and the effect on QOL based on this classification.
Creating norms for PHES and EncephalApp:
Standard deviations according to healthy control performance in the four centers, adjusted for age, gender and years of education, were created, and a performance below -4, based on composite standard deviation (SD) scores, spread across all subtests as is customary for PHES, was considered to be MHE16. Cognitive performance on PHES and EncephalApp was also adjusted for alcohol etiology, diabetes and metabolic syndrome, none of which affected the performance. PHES performance was also further studied with -1 through -3 SD as well compared to controls. MHE by PHES was defined as an abnormal threshold based on these adjusted norms. For EncephalApp, OffTime+OnTime was used to define normal performance in controls, which was adjusted by age, educational attainment and gender10, 17, since etiology, diabetes and metabolic syndrome impact performance. Agreement between EncephalApp and MHE-PHES was assessed via concordance using straight agreement across diagonals. Discriminatory capacity was assessed using AUC with sensitivity and specificity. In addition, a logistic regression was performed to assess the predictive power of EncephalApp (without age/gender adjustment) compared to MHE-PHES.
All subjects were also administered the validated Sickness Impact Profile (SIP), for QOL18. It has a total score, psychosocial and physical indices and a high score indicates poorer QOL.
Validation of PHES and EncephalApp norms:
A separate validation cohort of patients with cirrhosis were recruited from the same sites, who were administered PHES and EncephalApp. Regression models generated using the controls were again used to determine concordance, sensitivity, specificity and prevalence of MHE using PHES and EncephalApp.
Results:
A total of 100 non-cirrhotic and 109 cirrhotic subjects were included in the derivation cohort. Controls were older but were less likely to be men and were better educated than patients with cirrhosis (Table 1). The mean MELD score of patients with cirrhosis was 10.3±3.5 and the major etiologies were NASH (n=45), alcohol (n=15), and hepatitis C (n=18), and followed by other (n=31) causes. The most common co-morbid condition was diabetes not on insulin or with end- organ damage, found in 29 (27%) of cirrhotic patients compared to 7 (7%, p<0.001) of controls. The only other major co-morbid condition was hypertension, the prevalence of which was statistically similar across groups (controls 29, 29%, vs 39, 36% p=0.29). As shown in table 1, patients with cirrhosis had worse QOL and cognitive function on PHES and EncephalApp compared to controls. This extended to all sub-tests of PHES apart from Number connection-A and all outcomes for EncephalApp except for On State runs. Most sub-tests of PHES, RBANS and EncephalApp were correlated with QOL (Supplementary table 1). None of the patients developed overt HE during the examination and study.
Table 1:
Comparison between controls and patients with cirrhosis
| mean±SD unless otherwise specified | Controls (n=100) | Cirrhosis (n=109) | P value |
|---|---|---|---|
| Age (years) | 74.14 ± 6.46 | 70.32 ± 4.31 | <0.0001 |
| Gender (M/F) | 42 / 58 | 56 / 53 | 0.03 |
| Education (years) | 15.54 ± 3.01 | 14.01 ± 3.04 | <0.001 |
| MMSE (of 30) | 28.72 ± 1.78 | 28.50 ± 1.63 | 0.41 |
| Psychometric Hepatic Encephalopathy Score | |||
| Number connection A | 42.7 ± 32.1 | 48.1 ± 26.5 | 0.19 |
| Number connection B | 104.6 ± 73.1 | 128.3 ± 86.1 | 0.05 |
| Digit Symbol | 56.6 ± 18.1 | 46.6 ± 16.0 | <0.0001 |
| Serial Dotting | 59.6 ± 23.9 | 90.0 ± 43.7 | <0.0001 |
| Line tracing time | 88.4 ± 49.1 | 115.4 ± 45.5 | <0.0001 |
| Line tracing errors | 58.1 ± 41.5 | 27.9 ± 28.1 | <0.0001 |
| Line tracing errors+time | 146.6 ± 48.2 | 143.3 ± 51.0 | 0.63 |
| EncephalApp | |||
| OffTime | 83.2 ± 23.6 | 93.9 ± 26.5 | 0.04 |
| OnTime | 106.9 ± 40.5 | 124.1 ± 54.3 | 0.01 |
| Median Runs in Off State | 5.0 (1.0) | 6.0 (2.0) | 0.03 |
| Median Runs in On State | 6.0 (2.0) | 6.0 (3.0) | 0.22 |
| EncephalApp Off+OnTime | 192.0 ± 62.9 | 218.1 ± 78.2 | 0.01 |
| Quality of life (High=poor) | |||
| SIP total score | 4.3 ± 7.1 | 9.8 ± 9.2 | <0.0001 |
| SIP Psychoscial domain | 3.7 ± 8.5 | 9.1 ± 11.8 | <0.0001 |
| SIP Physical domain | 3.4 ± 6.2 | 8.3 ± 10.3 | <0.0001 |
SIP: Sickness Impact Profile, MMSE: mini-mental status exam. A high score on Digit symbol indicates good performance while high score or time in the remaining cognitive tests indicates the opposite. Statistical comparison performed using t-tests or Chi-square tests as appropriate.
Comparison of patients between sites:
Table 2 demostrates that patients with cirrhosis from all sites had statistically similar age, MELD score, education, MMSE and etiologies of cirrhosis. As expected there were more men in the VAMC. Patients from VAMC and Arkansas had worse cognitive performance on most subparts of PHES and RBANS than the remaining two groups, but EncephalApp results were similar. Diabetes and hypertension prevalence were not significantly different between sites. When norms were applied, this translated into more MHE by PHES in the VAMC and Arkansas group compared to the rest, but no difference in MHE by EncephalApp. QOL was statistically similar across sites.
Table 2:
Comparison of Patients with Cirrhosis across Sites in the Derivation Cohort
| mean±SD unless otherwise specified | VCU (n=37) |
VA (n=34) |
Mayo (n=10) |
Arkansas (n=28) |
|---|---|---|---|---|
| Age (years) | 70.3±3.6 | 71.2±5.7 | 71.8±7.9 | 69.9±3.0 |
| MELD score | 10.8±4.1 | 10.3±4.0 | 10.0±3.2 | 9.5±3.0 |
| Etiology(Hepatitis C/Alcohol/NASH/Rest) | 7/3/15/12 | 6/9/13/6 | 2/1/5/2 | 3/2/12/11 |
| Gender (Men/Women)*** | 18/19 | 32/2 | 7/3 | 8/19 |
| Education (years) | 14.8±2.9 | 13.5±3.3 | 14.6±2.7 | 13.5±2.1 |
| MMSE (of 30) | 29.1 ±1.0 | 28.3±1.3 | 28.8±1.1 | 28.4±2.0 |
| Psychometric Hepatic Encephalopathy Score | ||||
| Number connection A*** | 36.3±14.5 | 56.6±33.4 | 36.6±7.8 | 41.9±14.4 |
| Number connection B** | 88.4±40.2 | 159.7±122.1 | 99.2±49.2 | 109.7±44.0 |
| Digit Symbol | 53.3±12.1 | 45.0±19.7 | 53.4±10.8 | 45.8±13.6 |
| Serial Dotting*** | 60.2±14.4 | 86.5±36.8 | 47.9±7.7 | 126.1 ±31.9 |
| Line tracing errors+time* | 137.0±29.0 | 166.4±87.2 | 126.7±31.3 | 123.6±25.6 |
| RBANS percentile (high=good) | ||||
| RBANS total*** | 49.4±26.2 | 27.0±22.8 | 51.8±28.4 | 25.7±19.7 |
| Visuo-spatial*** | 49.5±31.2 | 32.5±31.8 | 72.1 ±27.9 | 19.1 ±22.0 |
| Immediate memory | 46.8±24.8 | 31.2±28.8 | 44.5±34.6 | 32.1 ±25.0 |
| Delayed memory | 50.9±26.9 | 31.9±27.5 | 41.1 ±31.8 | 36.0±24.2 |
| Language | 45.8±21.2 | 33.4±16.0 | 48.4±19.6 | 43.5±17.0 |
| Attention** | 63.0±27.0 | 41.9±31.1 | 47.8±31.4 | 35.4±30.8 |
| EncephalApp (high=poor) | ||||
| OffTime | 89.7±22.8 | 96.1 ±36.0 | 81.0±14.7 | 87.5±11.3 |
| OnTime | 110.3±37.1 | 133.5±104.1 | 101.7±22.3 | 123.1±28.2 |
| Median Runs in Off State | 6.0 | 5.0 | 6.0 | 6.0 |
| Median Runs in On State* | 6.0 | 7.0 | 5.0 | 7.0 |
| EncephalApp Off+OnTime | 195.4±52.8 | 229.6±138.3 | 182.7±36.8 | 210.7±36.7 |
| MHE on EncephalApp | 20 (54%) | 20 (59%) | 8 (80%) | 11 (39%) |
| MHE on PHES* | 3 (8%) | 8 (23%) | 0 (0%) | 7 (25%) |
| Quality of life (high=poor) | ||||
| SIP total | 9.5±10.5 | 11.9±9.8 | 6.4±6.4 | 9.0±9.5 |
| SIP Psycho-social | 8.9±11.9 | 11.5±13.6 | 3.5±3.0 | 8.6±10.9 |
| SIP Physical | 8.1±10.7 | 9.9±10.3 | 5.8±8.8 | 7.6±10.4 |
| Neuropsych classification | ||||
| Unimpaired/MCI/MHE/Both | 15/7/10/5 | 7/6/8/13 | 0/5/4/1 | 8/7/7/6 |
SIP: Sickness Impact Profile, MMSE: mini-mental status exam, RBANS: repeatable battery for the assessment of neuro-psychological status, MCI: mild cognitive impairment, MHE: minimal hepatic encephalopathy, NASH: non-alcoholic steatohepatitis.
p<0.01,
p<0.001
ANOVA/Kruskal-Wallis between 4 groups. A high score on digit symbol indicates good performance while high score or time in the remaining tests indicates the opposite. Statistical comparison performed using t-tests or Chi-square tests as appropriate.
Division of patients with cirrhosis based on neuropsychological assessment:
As shown in table 3, roughly a quarter of patients with cirrhosis belonged to each group. There was no difference in age, gender, MELD score or etiology of cirrhosis between the subgroups. When RBANS performance was used alone as the basis for group diagnostic classification, as expected, MHE only and MCI+MHE group subjects were impaired on PHES and EncephalApp, while MCI and both groups were impaired on RBANS compared to the unimpaired group. Patients with both MCI and MHE cognitive profiles demonstrated significantly worse QOL on both the SIP psychosocial and physical domains. This burden was noticed more in patients with presence of MHE rather than presence of MCI.
Table 3:
Comparison within patients with cirrhosis based on RBANS-informed neuropsychological divisions
| mean±SD unless otherwise specified | Unimpaired (n=30) |
Only MCI (n=25) |
OnlyMHE (n=29) |
Both (n=25) |
|---|---|---|---|---|
| Age (years) | 70.6±3.2 | 70.9±4.4 | 70.3±5.4 | 69.9±4.8 |
| MELD score | 10.6±4.2 | 10.1±3.4 | 9.9±3.3 | 10.6±3.0 |
| Etiology (Hepatitis C/Alcohol/NASH/Rest) | 2/3/16/9 | 3/2/11/9 | 4/4/12/9 | 9/6/6/4 |
| Gender (Men/Women) | 17/13 | 19/6 | 15/14 | 16/9 |
| Education (years) | 13.8±2.5 | 15.1±3.2 | 13.7±3.0 | 13.0±3.3 |
| MMSE (of 30) | 28.9±1.3 | 28.4±1.9 | 28.6±1.7 | 27.9±1.4 |
| Psychometric Hepatic Encephalopathy Score | ||||
| Number connection A | 37.9±11.1 | 41.6±17.6 | 43.7±20.4 | 72.1±37.9*** |
| Number connection Bt | 86.3±28.5 | 104.5±65.6 | 117.5±45.5 | 205.4±129.7*** |
| Digit Symbol | 50.2±13.4 | 50.2±17.4 | 49.5±16.1 | 36.0±13.4*** |
| Serial Dotting | 82.2±34.1 | 82.4±42.2 | 80.04±36.4 | 120.0±52.1*** |
| Line tracing errors+time | 133.6±27.4 | 134.3±35.7 | 135.7±38.4 | 173.3±81.9* |
| RBANS percentile (high=good) | ||||
| RBANS total | 50.1±23.3 | 33.8±24.6 | 36.9±26.3 | 8.0±6.6*** |
| Visuo-spatial† | 54.0±30.8 | 41.0±32.3 | 30.5±31.6 | 19.0±25.3*** |
| Immediate memory | 48.4±20.5 | 28.4±28.5 | 47.4±27.2 | 10.0±13.1*** |
| Delayed memory | 60.6±21.1 | 23.0±20.2 | 49.9±20.5 | 7.6±9.6*** |
| Language | 43.9±19.6 | 41.6±16.3 | 44.1±22.1 | 28.8±17.6* |
| Attention† | 60.9±22.1 | 60.8±27.4 | 35.5±31.8 | 12.3±15.8*** |
| EncephalApp (high=poor) | ||||
| OffTime† | 83.3±11.2 | 87.2±22.1 | 95.3±23.7 | 113.4±37.0*** |
| OnTime† | 10.5.1±27.2 | 110.5±36.2 | 127.3±39.3 | 159.8±87.5*** |
| sMedian Runs in Off State | 5.0 | 6.0 | 6.5 | 5.0* |
| Median Runs in On State | 6.0 | 6.0 | 6.0 | 6.0* |
| EncephalApp Off+OnTime† | 184.8±27.1 | 194.6±57.4 | 232.5±67.8 | 273.1±118.9*** |
| Quality of life(high=poor) | ||||
| SlPtotal† | 4.2±6.8 | 6.7±7.4 | 9.7±10.0 | 11.0±12.3* |
| SIP Psychosocial† | 4.2±8.7 | 5.9±7.4 | 9.8±13.4 | 9.3±13.4* |
| SIP Physical† | 3.5±6.4 | 5.4±7.5 | 8.1±9.6 | 9.9±12.9* |
SIP: Sickness Impact Profile, MMSE: mini-mental status exam, RBANS: repeatable battery for the assessment of neuro-psychological status, MCI: mild cognitive impairment, MHE: minimal hepatic encephalopathy, NASH: non-alcoholic steatohepatitis.
p<0.05,
p<0.001 ANOVA/Kruskal-Wallis between 4 groups,
p<0.05 MHE-contarining (MHE only or both) groups vs rest. A high score on digit symbol indicates good performance while high score or time in the remaining tests indicates the opposite. Statistical comparison performed using t-tests or Chi-square tests as appropriate.
Creation of norms:
We created the regression formulae adjusted to age, gender and education for healthy controls for each PHES subtest and for EncephalApp OffTime+OnTime. The regression formulae are:
Digit symbol=92.14189–0.86879*age+2.10844*education-9.51050*male, Numberconnection-A=54.37625+0.45152*age-3.27626*education+12.19085*male, Number connectio-B=129.19630+1.66005*age-10.25998*education+27.28603*male, Serial
Dotting test=124.45505–0.28774*age-3.01174*education+7.96783*male, Line tracing errors+time=130.59226+0.83839*age-3.36657*education+14.81727*male and EncephalApp Stroop=82.13076+2.95483*age-7.78722*education+31.67838*male.
In all cases, older age, being male, and a lower education was associated with poor cognitive performance. Neither alcohol etiology, diabetes or hypertension were associated with cognition. Of the 100 controls, 46 were from VCU, 7 VAMC, 23 from Mayo and 25 from Arkansas. Their ages and educational level and MMSE values were similar apart from gender (Supplementary table 3). After adjusting for age, gender, and education the sites did not differ on DST (p = 0.13), NCTB (p = 0.21), SDT (p = 0.09), LTTotal (p = 0.43), NCTA (p=0.06) or EncephalApp OffTime+OnTime (p=0.07). Power analysis for controls was performed (Supplementary table 4). Based on norms defined among controls adjusted for age, gender and education and using a -4 cut-off, we found that only 6% of the patients with cirrhosis were positive for MHE on PHES. On the other hand, age-, gender- and education-adjusted performance on EncephalApp found that 53 (49%) patients with cirrhosis were impaired. The concordance between the two tests was 53%. However, if we consider the PHES to be the gold standard, when comparing EncephalApp the sensitivity of the Stroop test was 1.00 [95% CI (0.93, 1.00)] and the specificity of the Stroop was 0.50 [95% CI (0.40, 0.60)]. When we fit a logistic regression model predicting MHE-PHES by Stroop, the AUC for the model was 0.97 (95% CI [0.89, 0.99]). When evaluating this model at the Youden Index, which was OffTime+OnTime>269 seconds, the sensitivity was 1.00 (95% CI [0.61, 1.00]) and the specificity was 0.89 (95% CI [0.81, 0.94]).
Due to this low prevalence, we also determined cut-offs at other levels (supplementary table 2). As expected, lowering the SD deviations required would increase the prevalence of MHE on PHES from 6% at below -4 to 29% at below -1 along with increase in concordance with EncephalApp from 53% to 65% at -1 levels. The AUC of EncephalApp predicting PHES-related MHE was lowest at below -1 and highest at below -4 levels.
Validation cohort:
We included 77 cirrhotic patients (30 VCU, 22 VA, 20 Mayo, 5 AR), with a mean age of 69.49±4.36 years, 72% men and median education of 14 years (IQR 10–21). The MELD score was 10.0±4.0, and most had NASH (n=28), followed by alcohol (n=21), hepatitis C (n=18) and others (n=10), cirrhosis etiologies. Apart from more men in this cohort, remaining demographics were similar to the derivation cohort. Compared to the controls, the validation patients were younger, less educated and had more men, which was similar to the comparison between the derivation cohort and controls. Cognitive performance on individual PHES tests were: digit symbol (43.51±13.17 raw score), number connection A (53.79±27.35 seconds), number connection B (130.36±70.84 seconds), serial dotting (80.11±36.53 seconds) and line tracing time+errors (138.75±40.63). EncephalApp OffTime+OnTime was 215.19±80.08 seconds. All test results were statistically similar to the derivation cohorts and were significantly more impaired compared to controls.
MHE diagnosis on PHES and EncephalApp in validation cohort:
Based on cut-offs generated in by healthy controls, on PHES, only 8% in the validation cohort (n=6) were positive for MHE. This can be contrasted to 49% (n=38) who had EncephalApp MHE with a concordance of 58% (n=45). However, if we consider the PHES to be the true diagnosis and EncephalApp to be a test, the sensitivity of the EncephalApp is 1.00(95% CI [0.61,1.00]) and the specificity is 0.55(95% CI [0.43,0.66]). On the logistic regression model predicting MHE-PHES by EncephalApp, the AUC is 0.94 (0.86, 0.98). At the Youden Index (OffTime+OnTime>244 seconds), the sensitivity is 1.00(0.61, 1.00) with specificity is 0.89(0.79, 0.94). Proportion of patients with PHES-MHE(8% vs 6%, p=0.71) and EncephalApp (49% vs 49%, p=0.91) were similar to the derivation cohort.
Discussion:
With the increasing age of patients with cirrhosis, it is often difficult to interpret the cognitive impairment that is epidemic in this population6, 16. This can result in poor QOL, burden of patients and caregivers and difficulty in differentiating between age-related compared to HE-related alterations in cognitive performance6. Our multi-center experience demonstrates that older cirrhotic patients have a worse QOL and cognitive dysfunction on multiple domains compared to non-cirrhotic subjects. Moreover, elderly cirrhotics with both MCI and MHE demonstrate worse QOL compared to MCI alone. Finally, we determined norms for PHES and EncephalApp, which were then applied in a separate cirrhosis cohort.
Our results demonstrate that there are impairments in cognitive function on PHES and EncephalApp between subjects with cirrhosis compared to controls. All subtests of PHES except number connection A were impaired in cirrhotic subjects, EncephalApp OffTime and OnTime were both impaired in this population. These patterns reflect what is seen in younger patients with cirrhosis and demonstrate the generalizability of this cognitive pattern of performance, and the tools used to assess it, in an elderly population19. These findings were maintained despite the patients being younger than control subjects, and less educated.
A major reason for assessing and understanding altered cognition in patients with cirrhosis is the negative impact cognitive impairment can have on QOL20. Indeed, we found that patients with cirrhosis had worse QOL than controls on almost every aspect of the SIP. Therefore, despite differences in gender, education and age between the groups, patient-reported outcomes were more keenly felt in cirrhotic patients.
Cognitive domains impaired in MHE are related to visuo-spatial function, cognitive flexibility and psychomotor speed, while memory changes beyond working memory are usually not encountered21, 22. On the other hand, MCI unrelated to MHE is typically associated with impairment in learning and delayed memory13, which has often been diagnosed using RBANS. However, the interaction of these conditions is important to study since pre-dementia MCI is unlikely to reverse with HE-related therapies, or with liver transplant23, 24. The division of cirrhotic subjects into one of four categories (i.e., unimpaired, MHE, MCI, or both MCI+MHE) was performed solely based on RBANS performance15, which has a relatively long track record of use. The fact that this classification method successfully distinguished PHES and EncephallApp performance, speaks to their concurrent validity. MHE patients had worse EncephalApp performance compard to MCI patients but there was no significant difference in EncephalApp between MCI and unimpaired patients. Despite a statistically similar demographic and cirrhosis profile, it is striking that those with the combination of both amnestic and non-amnestic cognitive impairment (the MCI+MHE group) had worse QOL than the remaining groups.
Neuropsychological performance is a measure of brain function. Patients with the combination of MCI and MHE, likely manifest a greater acculmulation of neuropathology. The lower perceived QOL in these subjects likely reflects an individual’s awareness regarding their associated neurobehavioral decline. Of note, patients with MHE were more likely to be impaired on QOL compared to those with MCI alone and unimpaired patients. Cortical tone and arousal, which are the foundation skills for all cognitive ability, is dependent on the integrity of subcortical structures25. For example, if an individual has an inability to attend to a task due to decreased cortical tone or arousal, they will perform poorly on any cognitive measure, regardless of whether the task emphasizes concentration, learning, memory, executive ability or motor skills26. For this reason, it is not surprising that a cognitive profile characterized by subcortical cognitive impairment (i.e., MHE only) would be associated with signficant decline in QOL.
A prior study examining the cognitive profiles of patients with cirrhosis demonstrated a similar pattern of impairment on these disease-specific neuropsychological measures9. The current study extends these findings to an older age group, provides concurrent validity for these assessment tools with this specific medical population, and may help with generalization since this used a multi-center US-based population with validation cohorts. When individual centers were compared, as expected the VAMC cirrhotic patients were more likely to be men. Patients from Arkansas and the VAMC tended to perform worse on cognitive tests despite a similar age, education and MELD score. While rates of MHE on PHES varied, rates on EncephalApp remained stable between sites. This is important for further generalization and could be due to the relatively higher age and gender dependence of the PHES compared to EncephalApp. This, however, was accompanied by statistically similar QOL, which trended to also be worse in the VAMC patients. It highlights the important relationship between QOL and cognition20.
Most published studies have focused on relatively younger patients with cirrhosis, with a cut-off of 65 years or younger. Therefore norms based on older patients for PHES and EncephalApp, especially in the US, are lacking. Using prior studies and adjusting for variables that affect results such as age, gender and education, we found that a minority of elderly patients with cirrhosis had MHE on PHES10, 27. On the other hand, almost half of the cirrhotic patients exhibited poor performance on EncephalApp, which is in line with prior studies of younger patients10. The rate improved if we reduced the threshold for considering PHES abnormal with respect to controls. Moreover, these were further validated with similar lower prevalence of MHE on PHES and higher on EncephalApp in a separate validation cohort. Regardless of the cut-off, EncephalApp showed excellent AUC when PHES was used as the gold standard. These differences in the relative prevalence of MHE by EncephalApp vs PHES is intriguing. PHES, which involves multiple paper-pencil tests and could reflect baseline poor performance of these tests with age, where the relative difference between subjects with and without cirrhosis would diminish due to other co-morbid conditions and has multiple tests to take into account27. However, alcohol, diabetes and hypertension, were not associated with these results in our cohorts28. On the other hand, EncephalApp involves electronic presentations of multiple stimuli that depend on the psychomotor speed and cognitive flexibility and is only one test10, 17. Since greater differences were seen in the OnTime compared to the OffTime condition, the results suggest that impairment in the frontal lobe inhibitory circuits may drive these differences. Of note, number connection-B, which assesses cognitive flexibility, was also impaired in patients with cirrhosis. Both of these cognitive measures assess the integrity of the prefrontal cortex. Prior studies, have previously demonstrated that cirrhosis selectively impacts prefrontal integrity24, 29, 30.
The study is limited by the relatively low number of patients per center, which was necessitated by the exclusion criteria. Moreover, healthy controls across sites were relatively similar in demographics and cognitive function. However, we then included a validation cohort also from the same centers to find similar demographic and cognitive patterns. We excluded patients with prior OHE and other severe co-morbid conditions in order to reduce confounding. It is likely that the number of patients abnormal on the PHES would increase if a greater proportion of OHE subjects were included in the study sample31. As expected in a real-world setting there were differences in demographics between centers, which were adjusted in statistical analyses. The study group division of MCI, MHE and Both categories were based on a single neuropsychologist assessment. However this classification demonstrated biological plausibility as it correctly distinguished PHES and EncephalApp performance. In addition, as expected the both (MCI+MHE) classification, in comparison with either the MCI or MHE groups, by virtue of greatest cognitive compromise, had the lowest QOL. The current results define the first step in the delineation of this complex situation where age-related and HE-related changes coexist. Further studies need to determine impact of treatments for MHE and study of other co-morbid conditions.
We conclude that in a multi-center of study of individuals over 65 years of age, there is worsening health-related quality of life in patients with cirrhosis, which is associated with poor cognition. While they co-exist in almost a quarter of patients with cirrhosis tested, minimal HE had a greater impact than pre-dementia mild cognitive impairment on health-related quality of life. EncephalApp Stroop was able to detect cognitive impairment adjusted for age, gender and educational attainment in this population using derivation and validation cohorts, which was relatively specific for MHE rather than MCI diagnosis. The use of this tool can potentially facilitate selection of higher risk elderly population with cirrhosis and facilitate the allocation of needed resources.
Supplementary Material
What You Need to Know.
Background:
The overlap between minimal hepatic encephalopathy (MHE) and pre-dementia mild cognitive impairment (MCI) could affect quality of life (QOL).
Findings:
In a multi-center study of patients with cirrhosis (older than 65 y) and controls, the presence of MHE, regardless of MCI, associated with poor cognition and QOL.
Implications for patient care:
We adjusted norms, the EncephalApp detects MHE in older individuals with a high level of sensitivity. These patients should be treated to improve QOL.
Grant Support:
partly supported by VA Merit Review 2I0CX001076, RO1HS025412 and R21TR002024 to JSB
List of Abbreviations:
- MHE
minimal hepatic encephalopathy
- MCI
mild cognitive impairment
- QOL
health-related quality of life
- PHES
psychometric hepatic encephalopathy score
- SIP
Sickness Impact Profile
- RBANS
Repeatable Battery for the Assessment of Neuropsychologcial Status
- MMSE
mini-mental status exam
- SD
standard deviation
- VCU
Virginia Commonwealth University
- VAMC
McGuire VA Medical Center
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Conflict of Interest: None for any author
References:
- 1.Asrani SK, Devarbhavi H, Eaton J, et al. Burden of liver diseases in the world. J Hepatol 2019;70:151–171. [DOI] [PubMed] [Google Scholar]
- 2.Kim IH, Kisseleva T, Brenner DA. Aging and liver disease. Curr Opin Gastroenterol 2015;31:184–91. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Koehler EM, Schouten JN, Hansen BE, et al. Prevalence and risk factors of non-alcoholic fatty liver disease in the elderly: results from the Rotterdam study. J Hepatol 2012;57:1305–11. [DOI] [PubMed] [Google Scholar]
- 4.Blazer DG, Wu LT. The epidemiology of at-risk and binge drinking among middle-aged and elderly community adults: National Survey on Drug Use and Health. Am J Psychiatry 2009;166:1162–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Kanwal F, Gralnek IM, Hays RD, et al. Health-related quality of life predicts mortality in patients with advanced chronic liver disease. Clin Gastroenterol Hepatol 2009;7:793–9. [DOI] [PubMed] [Google Scholar]
- 6.Rakoski MO, McCammon RJ, Piette JD, et al. Burden of cirrhosis on older Americans and their families: analysis of the health and retirement study. Hepatology 2012;55:184–91. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Vilstrup H, Amodio P, Bajaj J, et al. Hepatic encephalopathy in chronic liver disease: 2014 Practice Guideline by the American Association for the Study of Liver Diseases and the European Association for the Study of the Liver. Hepatology 2014;60:715–35. [DOI] [PubMed] [Google Scholar]
- 8.Sanford AM. Mild Cognitive Impairment. Clin Geriatr Med 2017;33:325–337. [DOI] [PubMed] [Google Scholar]
- 9.Bajaj JS, Ahluwalia V, Steinberg JL, et al. Elderly patients have an altered gut-brain axis regardless of the presence of cirrhosis. Sci Rep 2016;6:38481. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Allampati S, Duarte-Rojo A, Thacker LR, et al. Diagnosis of Minimal Hepatic Encephalopathy Using Stroop EncephalApp: A Multicenter US-Based, Norm-Based Study. Am J Gastroenterol 2016;111:78–86. [DOI] [PubMed] [Google Scholar]
- 11.Bajaj JS, Thacker LR, Heuman DM, et al. Cognitive performance as a predictor of hepatic encephalopathy in pretransplant patients with cirrhosis receiving psychoactive medications: a prospective study. Liver Transpl 2012;18:1179–87. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Corrias M, Turco M, Rui MD, et al. Covert hepatic encephalopathy: does the mini-mental state examination help? J Clin Exp Hepatol 2014;4:89–93. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Weissenborn K, Ennen JC, Schomerus H, et al. Neuropsychological characterization of hepatic encephalopathy. J Hepatol 2001;34:768–73. [DOI] [PubMed] [Google Scholar]
- 14.Duff K, Hobson VL, Beglinger LJ, et al. Diagnostic accuracy of the RBANS in mild cognitive impairment: limitations on assessing milder impairments. Arch Clin Neuropsychol 2010;25:429–41. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Mooney S, Hasssanein TI, Hilsabeck RC, et al. Utility of the Repeatable Battery for the Assessment of Neuropsychological Status (RBANS) in patients with end-stage liver disease awaiting liver transplant. Arch Clin Neuropsychol 2007;22:175–86. [DOI] [PubMed] [Google Scholar]
- 16.Amodio P, Wenin H, Del Piccolo F, et al. Variability of trail making test, symbol digit test and line trait test in normal people. A normative study taking into account age-dependent decline and sociobiological variables. Aging Clin Exp Res 2002;14:117–31. [DOI] [PubMed] [Google Scholar]
- 17.Bajaj JS, Thacker LR, Heuman DM, et al. The Stroop smartphone application is a short and valid method to screen for minimal hepatic encephalopathy. Hepatology 2013;58:1122–32. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Bergner M, Bobbitt RA, Carter WB, et al. The Sickness Impact Profile: development and final revision of a health status measure. Med Care 1981;19:787–805. [DOI] [PubMed] [Google Scholar]
- 19.Bajaj JS, Heuman DM, Sterling RK, et al. Validation of EncephalApp, Smartphone-Based Stroop Test, for the Diagnosis of Covert Hepatic Encephalopathy. Clin Gastroenterol Hepatol 2014. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Montagnese S, Bajaj JS. Impact of Hepatic Encephalopathy in Cirrhosis on Quality-of-Life Issues. Drugs 2019;79:11–16. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Ortiz M, Jacas C, Cordoba J. Minimal hepatic encephalopathy: diagnosis, clinical significance and recommendations. J Hepatol 2005;42 Suppl:S45–53. [DOI] [PubMed] [Google Scholar]
- 22.Morgan MY, Amodio P, Cook NA, et al. Qualifying and quantifying minimal hepatic encephalopathy. Metab Brain Dis 2015. [DOI] [PubMed] [Google Scholar]
- 23.Bajaj JS, Kamath PS. The brain gets its say: Hepatic encephalopathy and its evolving role in transplant priority. Liver Transpl 2016;22:1319–20. [DOI] [PubMed] [Google Scholar]
- 24.Ahluwalia V, Wade J, Gilles H, et al. The Functional Basis for Cognitive Improvement after Liver Transplant: A Prospective Brain MRI Study, In HEPATOLOGY, WILEY-BLACKWELL 111 RIVER ST, HOBOKEN 07030-5774, NJ USA, 2015. [Google Scholar]
- 25.Montagnese S, Turco M, Amodio P. Hepatic encephalopathy and sleepiness: an interesting connection? J Clin Exp Hepatol 2015;5:S49–53. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Butterworth RF. Pathogenesis of hepatic encephalopathy: new insights from neuroimaging and molecular studies. J Hepatol 2003;39:278–85. [DOI] [PubMed] [Google Scholar]
- 27.Goldbecker A, Weissenborn K, Hamidi Shahrezaei G, et al. Comparison of the most favoured methods for the diagnosis of hepatic encephalopathy in liver transplantation candidates. Gut 2013;62:1497–504. [DOI] [PubMed] [Google Scholar]
- 28.Perez-Matos MC, Jiang ZG, Tapper EB. Factors That Affect Results of Psychometric Tests to Identify Patients With Minimal Hepatic Encephalopathy. Clin Gastroenterol Hepatol 2018;16:1836–1838. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Ahluwalia V, Wade JB, White MB, et al. Rifaximin Improves Brain Edema and Working Memory in Minimal Hepatic Encephalopathy: A Prospective fMRI Study. Hepatology 2012;56:162A. [Google Scholar]
- 30.Zhang XD, Zhang LJ, Wu SY, et al. Multimodality magnetic resonance imaging in hepatic encephalopathy: an update. World J Gastroenterol 2014;20:11262–72. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Arguedas MR, DeLawrence TG, McGuire BM. Influence of hepatic encephalopathy on health-related quality of life in patients with cirrhosis. Dig Dis Sci 2003;48:1622–6. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.