Clinical trial: oral zinc in hepatic encephalopathy

Authors


Dr Y. Takuma, Department of Gastroenterology, Kurashiki Central Hospital, 1-1-1 Miwa, Kurashiki, Okayama 710-8602, Japan.
E-mail: takuma@enjoy.ne.jp

Abstract

Aliment Pharmacol Ther 2010; 32: 1080–1090

Summary

Background  Hepatic encephalopathy has a negative effect on patient health-related quality of life (HRQOL). Zinc supplementation has been effective with regard to altered nitrogen metabolism.

Aim  To investigate the effectiveness of oral zinc supplementation on hepatic encephalopathy and HRQOL.

Methods  Seventy-nine cirrhotic patients with hepatic encephalopathy were randomized to receive 225 mg of polaprezinc in addition to standard therapies of a protein-restricted diet including branched-chain amino acid and lactulose, or to continue only standard therapies for 6 months. The change of HRQOL by Short Form-36, hepatic encephalopathy grade, laboratory parameters, and neuropsychological (NP) tests were compared at baseline and at 6 months. We also evaluated via multivariate analysis whether zinc supplementation and clinical variables correlated with the changes in physical component scale (PCS) and mental component scale (MCS) between the two visits.

Results  Zinc supplementation significantly improved the PCS (P = 0.04), but not the MCS (P = 0.95). Zinc supplementation significantly decreased hepatic encephalopathy grade and blood ammonia levels (P = 0.03 and P = 0.01), and improved Child-Pugh score and NP tests compared with standard therapy (P = 0.04 and P = 0.02). In multivariate analysis, zinc supplementation was significantly associated with improvement in PCS (P = 0.03), whereas it was not significantly associated with change in MCS (P = 0.98).

Conclusion  Zinc supplementation is effective in hepatic encephalopathy and consequently improves patients HRQOL.

Introduction

Hepatic encephalopathy (HE) is a common neuropsychiatric complication of liver disease affecting about 20–30% patients with cirrhosis.1 HE in patients with liver cirrhosis represents a significant reduction in health-related quality of life (HRQOL) and a reversible decline in cognitive function.2 The pathogenesis of HE is not completely understood, but ammonia is considered to play a key role.3 Most studies have demonstrated that reduction in blood ammonia level affected improvements in HE grade and neuropsychological tests. Improvement in cognitive functions may lead to improvement in HRQOL. HRQOL is a major outcome in the evaluation of treatments for HE because HE itself is a poor prognostic factor in cirrhotic patients. Whether treatment of HE by a reduction in ammonia level is associated with health benefits has not been well established.

Synthetic disaccharides such as lactulose and lactitol have been used as the main agents to treat HE in advanced cirrhosis. A recent randomized control trial suggested that lactulose improved HRQOL in minimal HE patients;4 however, some patients were known to be unresponsive to these agents, because it minimized ammonia absorption, but did not affect ammonia detoxification. Branched-chain amino acid (BCAA) may be another therapy for HE. Some studies5, 6 have shown that oral BCAA supplements reduced significantly the incidence of complications including HE and improved HRQOL in advanced cirrhotic patients.

In advanced cirrhosis, amino acid imbalance due to functional disorder of ammonia disposal via urea synthesis in the liver, and consumption of BCAA caused by ammonia disposal via glutamine synthesis in the skeletal muscle was observed.7 However, BCAA alone sometimes increased the blood ammonia levels in protein-intolerant patients.8

Zinc supplementation was considered to be another important therapeutic option of HE, because its supplementation significantly improved HE, which had been refractory to protein restriction, and lactulose.9–11 Liver cirrhotic patients with HE are associated with a high incidence of zinc deficiency, which contributes to nitrogen metabolism disorder.12 The causes of low serum zinc levels in advanced cirrhotic patients are thought to be poor dietary intake via protein-restricted diet, impaired intestinal absorption, and excessive urinary losses. Two major organs are involved in ammonia metabolism: the liver, in which ammonia is converted to urea via ornithine transcarbamylase (OTC), and the skeletal muscle, where ammonia is disposed with glutamic acid via glutamine synthetase.13 In animal models, zinc deficiency decreases the activity of OTC, and zinc supplementation produces a remarkable increase in hepatic OTC.14 Zinc deficiency was also reported to impair the activity of muscle glutamine synthetase,15 which leads to hyperammonaemia. Therefore, zinc supplementation in addition to standard therapies with lactulose and a protein-restricted diet including BCAA may enhance the hepatic conversion of amino acids into urea,10 decrease serum ammonia, and consequently improve HRQOL.

A randomized control study was carried out to determine the influence of zinc supplementation on HRQOL and psychomotor performance of HE patients unresponsive to standard therapies.

Materials and methods

Design of the study

Ours is a prospective randomized study comparing the effect of zinc supplementation on HE and HRQOL in patients with decompensated liver cirrhosis treated with BCAA and synthetic disaccharides. The length of the study is 6 months.

Enrolment was from December 2005 to September 2007, and conducted at the National Hospital Organization Iwakuni Clinical Center. Written informed consent was obtained from all patients. This study was conducted according to the Declaration of Helsinki and approved by the investigation and ethics committee of the hospital.

Setting and participants

Inclusion criteria of this study were as follows: (i) patients with liver cirrhosis diagnosed histologically or clinically at least 3 months before enrolment, and under treatment with oral BCAA granules (Livact Granules; Ajinomoto Co., Inc., Tokyo, Japan; 12 g per day of BCAA; l-isoleucine 2856 mg, l-leucine 5712 mg, and l-valine 3432 mg); (ii) hyperammonaemia (fasting venous ammonia concentration at fasting ≧70 μg/dL); and (iii) patients with grade 1 or 2 recurrent episodic HE unresponsive to standard therapies with 30–60 mL per day of lactulose and a protein-restricted diet (1.0 g/kg/day of protein including BCAA preparation) for at least 4 weeks.

Exclusion criteria were as follows: (i) incurable hepatocellular carcinoma (HCC) (curable HCC conforming to the Milan criteria,16 and receiving treatments with radiofrequency ablation therapy, or surgical resection); (ii) alcoholic cirrhosis with inability to abstain from alcohol for at least 1 year before enrolment; (iii) severe complications of cirrhosis including serum albumin concentration <2.5 g/dL, serum total bilirubin concentration ≧3.0 mg/dL, or recurrent HE graded 3 or higher; (iv) use of antibiotics, drugs affecting psychometric performances (benzodiazepines, antiepileptics, psychotropic drugs, or narcotics); (v) history of gastrointestinal bleeding; (vi) history of shunt surgery, transjugular intrahepatic portosystemic shunt, or balloon-occluded retrograde transvenous obliteration; (vii) age >80 years; and (viii) organic neurological diseases such as subdural haematoma, Wernicke’s disease, encephalitis, other metabolic abnormalities, and drug intoxications.17

Male patients were considered to have alcohol-related cirrhosis if daily alcohol intake was more than 80 g; female patients, over 30 g for more than 5 years, if testing showed no viral, metabolic, or immunological cause. Viral hepatitis was diagnosed when HCV antibody or hepatitis B surface antigen was positive.

Randomization and interventions

Patients enrolled in this study were randomly assigned in a 1:1 ratio either to receive 225 mg (containing 51 mg of zinc and 174 mg of l-carnosine) per day of polaprezinc (Zeria Pharmaceutical Co., Ltd., Tokyo, Japan) in addition to ongoing standard therapies of a protein-restricted diet including BCAA and lactulose (zinc group), or to continue only standard therapies (control group) for 6 months. All other medications (diuretics, lactulose, and BCAA) and a protein-restricted diet were continued unchanged throughout.

The polaprezinc is a drug provided as granules, and it was really difficult to make the placebo with similar taste and dosage of the form.

For these reasons, the study was not blinded, and randomization was performed using computer-generated tables to assign patients to receive either polaprezinc or standard treatment. Minimization method to adjust the baseline imbalance between 2 groups in the composition of the HE grade (1 vs. 2) and that between inpatients and outpatients was used. Treatment allocation was concealed from the scorer (M.H.), but not from the principal investigator (Y.T.).

Evaluations of HE grade and HRQOL

During the selection period, the patients with an overt episode of HE were treated as in-patients or out-patients depending on the severity of HE episodes. In particular, out-patients were evaluated at every weekly interval. Recurrent episodic HE was defined as at least three episodes of overt HE in a 1-month period despite continuous treatment with lactulose and a protein-restricted diet. Diagnosis and grading of the overt HE were based on the clinical and neuropsychological abnormalities using the West Haven Criteria18 including state of consciousness, intellectual function, personality behaviour and neuromuscular abnormalities. Finally, diagnosis and grading of HE were decided by two senior hepatologists after consultation with an expert neurologist.

Number connection test (NCT)19 and digit symbol test (DST)20 were applied as neuropsychological (NP) tests for HE. For NCT-A, patients were asked to connect figures from 1 to 20 serially that were scattered on a page as quickly as possible (time limit: 120 s). For NCT-B, patients were asked to connect figures from 1 to 10 with ten Japanese characters scattered on a page as quickly as possible (time limit: 180 s). A low score indicated good performance. DST was included in the Wechsler adult intelligence scale-revised. The individuals were given a list of digits from 1 to 9 associated with symbols and asked to fill in the blanks with symbols that corresponded to each number. The test scores were the total correctly sequential matched symbols with those within a 60-s interval (maximum no. questions: 40). A high score indicated good performance.

Abnormal values of NCT-A, NCT-B, and DST were determined based on upper and lower 10th percentiles, which are regarded as outliers in healthy Japanese subjects obtained at 5-year intervals.21 NCT-A, NCT-B, and DST results were affected by age, but not by gender, facility, or education (most of the Japanese population receives compulsory education for at least 9 years).

The primary endpoints are 6-month changes of the physical component scale (PCS) in the SF-36 version 2 questionnaires;22, 23 these outcomes were chosen as a previous study24 indicated that PCS is more likely affected by HE than mental health domains. SF-36 was used to gather information on general quality of life on eight domains of health:25 (i) Physical functioning (PF), (ii) Role-physical (RP), (iii) Bodily pain (BP), (iv) General health (GH), (v) Vitality (VT), (vi) Social functioning (SF), (vii) Role-emotional (RE) and (viii) Mental health (MH). The score on each scale ranges from 0 to 100, with lower scores indicating poor health or greater disability. Two standardized summary scores, the PCS and the mental component scale (MCS), were calculated from SF-36 raw data.26 It has also proven useful for QOL evaluation in patients with advanced liver disease.24 The secondary endpoints are 6-month changes in HE grade, other clinical parameters, NP tests and any symptoms included in adverse events.

Follow-up

Clinical evaluation and laboratory values, adverse events and adherence to trial medication were assessed in both groups at 1-month intervals during the study period. Both the patients and their relatives were instructed on the importance of contacting medical staff immediately in the event of any alteration in the patient’s mental state between the scheduled visits. An independent scorer, unaware of assignment, assessed participants via a battery of NP tests, and the SF-36 questionnaire at baseline and at 6 months. All patients received dietary assessment and advice from a study dietician at baseline and 3 months.

Patients were regarded as responders if they had at least 1 degree of improvement in HE grade at the end of study period. In addition, patients were regarded as non-responders if they had aggravation or no change in HE grade at the end of study period.

Statistical analysis

For the randomized cohort, we determined that a sample size of 39 patients in each arm was needed to detect a 20% of PCS score increased in zinc treatment compared with standard treatment with a 5% (2-tailed) for α error and 80% for β error.27

Standard deviation (s.d. = 13.0) and 20% (5.9 of changes in PCS score) of expected effect size were derived from our preliminary study (unpublished data).

The Mann–Whitney U-test, chi-square test, and Fisher’s exact test were used to analyse the differences in baseline data between the two groups. Spearman’s rank correlation was used to evaluate association among the two summary scores of the SF-36 and continuous clinical parameters. anova was performed to detect association between the two summary scores of the SF-36 and clinical binary data.

In an attempt to explore independent predictors of the 6-month changes in two summary scores of the SF-36, all univariate variables were entered into stepwise backward multiple linear regression analyses, and variables with P values less than 0.1 were retained in the multivariate model. A comparison of serum laboratory values before and after treatment was studied using Student’s paired t-test, and on the rate of abnormal NP tests before and after treatment using McNemar’s test. Change in each clinical parameter between two groups was analysed using anova. Study outcomes were analysed using an intention-to-treat (ITT) population. A P value of less than 0.05 was considered statistically significant.

Statistical analyses were performed using spss 16.0J for Windows (SPSS Inc. and Microsoft Corp., Chicago, IL, USA).

Results

Trial enrolment

During the study period, 156 cirrhosis patients with overt HE were screened; 79 (51%) meeting eligibility criteria were enrolled (Figure 1). Thirty-nine patients with HE were assigned to zinc supplementation in addition to standard therapy, and 40 patients continued only standard therapy for 6 months. Reasons for exclusion (77 patients: 49%) included refusal to participate (6 patients), advanced HCC (21 patients), complicated chronic recurrent HE grade 3 or higher (10 patients), recent gastrointestinal bleeding (10 patients), serum albumin concentration <2.5 g/dL (8 patients), serum total bilirubin concentration ≧3.0 mg/dL (6 patients), history of recent alcohol intake (5 patients), recent antibiotic use (5 patients), recent use of drugs affecting psychomotor performance (3 patients), unfit to perform NP tests due to severe involuntary movements such as tremor (2 patients) and history of balloon-occluded retrograde transvenous obliteration (1 patient).

Figure 1.

 Flowchart demonstrating participants throughout the study procedures.

Baseline characteristics

Clinical and demographic characteristics are shown in Table 1. Causes of cirrhosis were chronic viral hepatitis (58 patients), alcoholic liver injury (13 patients), primary biliary cirrhosis (3 patients), autoimmune hepatitis (2 patients) and cryptogenic cirrhosis (3 patients). There were no statistically significant differences between the zinc and control groups with respect to baseline characteristics. Zinc deficiency (plasma zinc concentration <70 μg/dL) was observed in 38 patients (97%) in the zinc group, and 38 patients (95%) in the control group.

Table 1.   Characteristics of the zinc group and control group at baseline
CharacteristicTotal (n = 79)Control group (n = 40)Zinc group (n = 39)P value
  1. MELD, model for end-stage liver disease; HE, hepatic encephalopathy; HCC, hepatocellular carcinoma.

  2. Values expressed as mean ± s.d. The Mann–Whitney U-test, Chi-square test and Fisher’s exact test were used to analyse the differences in background and biochemical data between the two groups. Modified Child-Pugh score: qualitative HE assessment was excluded from the Child-Pugh score (a range of 4–12).

Gender (male/female)40/3923/1717/220.26
Age (years)66.5 ± 6.666.5 ± 7.466.5 ± 5.70.49
Hospital admission (in-patients/out-patients)18/619/319/300.95
Cause of cirrhosis (viral/alcoholic/others)58/13/831/4/527/9/30.25
Child-Pugh score8.0 ± 1.57.9 ± 1.48.1 ± 1.60.81
Child-Pugh classification (A/B/C)15/49/157/26/78/23/80.85
Modified Child-Pugh score6.0 ± 1.55.9 ± 1.46.1 ± 1.60.81
MELD score11.8 ± 3.211.4 ± 3.012.3 ± 3.30.24
Ascites (present/absent)26/5314/2612/270.69
HE Grade (1/2)49/3025/1524/150.93
No. of HE episodes per patient3.7 ± 1.23.7 ± 1.23.8 ± 1.20.84
Oesophageal varices (present/absent)52/2723/1729/100.11
History of HCC (present/absent)34/4520/2014/250.26
Total bilirubin (mg/dL)1.5 ± 0.71.4 ± 0.71.6 ± 0.70.19
Albumin (g/dL)3.3 ± 0.53.3 ± 0.53.3 ± 0.50.95
Prothrombin time (%)70.4 ± 12.470.4 ± 12.170.4 ± 12.90.80
Ammonia (μg/dL)133.4 ± 52.8129.1 ± 54.9137.8 ± 50.80.25
Platelets (×104/μL)10.4 ± 4.510.8 ± 4.910.0 ± 4.10.58
Zinc (μg/dL)50.3 ± 11.551.6 ± 13.348.9 ± 9.30.29
Fischer ratio1.7 ± 0.81.8 ± 1.01.7 ± 0.60.97
Use of diuretics (yes/no)50/2928/1222/170.25
Use of beta-blockers (yes/no)9/703/376/330.31

Baseline values of individual domains and summary scores of SF-36 did not differ significantly between groups (Table 2).

Table 2.   SF-36 and NP tests of the zinc group and control group at baseline
SF-36Total (n = 79)Control group (n = 40)Zinc group (n = 39)P value
  1. PF, physical functioning; RP, role-physical; BP, bodily pain; GH, general health; VT, vitality; SF, social functioning; RE, role-emotional; MH, mental health; PCS, physical component scale; MCS, mental component scale; NCT, number connection test; DST, digit symbol test; NP test ,neuropsychological test.

  2. Values expressed as mean ± s.d. The Mann–Whitney U-test was used to analyse the differences between the two groups.

PF65.1 ± 26.365.4 ± 28.164.9 ± 24.60.67
RP53.9 ± 32.650.5 ± 33.957.4 ± 31.30.38
BP61.1 ± 24.065.4 ± 23.256.7 ± 24.40.08
GH41.6 ± 19.241.9 ± 18.541.4 ± 20.10.86
VT49.8 ± 22.350.9 ± 24.548.7 ± 20.00.67
SF69.1 ± 25.966.9 ± 28.271.5 ± 23.30.55
RE58.5 ± 32.454.4 ± 33.362.8 ± 31.20.24
MH65.1 ± 22.263.3 ± 24.267.1 ± 20.20.56
PCS35.2 ± 13.235.9 ± 13.234.5 ± 13.30.64
MCS44.9 ± 11.643.3 ± 12.746.5 ± 10.40.18
NCT-A (s)75.6 ± 28.872.6 ± 30.578.8 ± 27.00.32
NCT-A (abnormal/normal)53/2623/1730/90.07
NCT-B (s)143.6 ± 31.2141.6 ± 31.3145.8 ± 30.40.64
NCT-B (abnormal/normal)63/1631/932/70.61
DST (points)10.4 ± 3.210.9 ± 3.89.8 ± 2.40.38
DST (abnormal/normal)62/1731/931/80.83
Mean numbers of abnormal NP tests2.3 ± 1.02.1 ± 1.02.4 ± 1.00.22

Clinical course

Thirty-nine zinc-assigned patients (zinc group) completed the study; one patient discontinued due to an adverse event (nausea and vomiting), and the rest received the full amount of supplement prescribed during the study period. Forty non-zinc treatment patients (control group) were also enrolled. One developed liver failure caused by bleeding oesophageal varices, and died at 3 months, and one dropped out at 4 months (Figure 1). Ultimately, 38 control group patients received the full amount of supplement prescribed during the study period. Noncompliance with assignment was not observed in either group, and none was excluded from statistical analysis. During the study period, all other medications (diuretics, lactulose, and BCAA) were continued unchanged, but two patients in the zinc and five in the control group developed severe HE (grade 3 or 4), and these patients were treated with kanamycin in addition to these medications until remission state.

Impact of zinc supplementation on HRQOL

Figure 2 shows the impact of treatment with zinc supplementation on SF-36. RP and PCS in the zinc group significantly improved after 6 months (P = 0.04 and P = 0.02 respectively); however, there was no significant change in the control group. Significant improvements in PF, RP and PCS in the zinc group were found compared to the control group (P = 0.04, P < 0.01, and P = 0.04 respectively, anova); however, there was no significant improvement in other variables compared with the control group (P = 0.95, anova) (Figure 2).

Figure 2.

 The 6-month changes from baseline (mean ± S.E.M.) in individual domains and summary scores of health-related quality of life (SF-36) in patients with cirrhosis randomized to control (open bars) and zinc (solid bars) groups. *Statistically significant (< 0.05) compared with each value prior to treatment, as determined via Student’s paired t-test. Statistically significant (< 0.05), compared between zinc and control groups via anova. Abbreviations: PF, physical functioning; RP, role-physical; BP, bodily pain; GH, general health; VT, vitality; SF, social functioning; RE, role-emotional; MH, mental health; PCS, physical component scale; MCS, mental component scale.

Impact of zinc supplementation on clinical and laboratory parameters

In the zinc group, the Child-Pugh and modified Child-Pugh score, mean HE grade and mean HE episodes improved (P < 0.001, P = 0.02, P < 0.01, and P < 0.0001 respectively), serum albumin level and serum zinc level increased (P = 0.02, and P < 0.0001 respectively), and blood ammonia level decreased (P < 0.0001) (Table 3). Moreover, the rate of presence of ascites showed a tendency to decrease in the zinc group (P = 0.08, McNemar’s test). No significant changes in the serum laboratory values were observed in the control group. Significant difference in the Child-Pugh score, mean HE grade, mean HE episodes, blood ammonia level and serum zinc level was observed between the two groups at 6 months after therapy (P = 0.04, P = 0.03, P = 0.02, P = 0.01, and P < 0.001 respectively, anova). The rate of zinc deficiency decreased significantly after 6 months of zinc administration (P < 0.001, McNemar’s test). However, no significant change was observed in the control group (P = 0.10, McNemar’s test). HE improved in 21 patients (responders) in the zinc group, and 10 patients in the control group at the end of study period (54% vs. 26%, P = 0.03, Chi-square test).

Table 3.   Changes in clinical and laboratory parameters at 6-month follow-up
VariableControl groupZinc group
Baseline6 monthsBaseline6 months
  1. MELD, model for end-stage liver disease; HE, hepatic encephalopathy; NCT, number connection test; DST, digit symbol test. Modified Child-Pugh score, qualitative HE assessment was excluded from Child-Pugh score (a range of 4–12).

  2. Values expressed as mean ± s.d.

  3. * Statistically significant, compared with each value prior to treatment, as determined via Student’s paired t-test and McNemar’s test.

  4. † Statistically significant (< 0.05), compared between zinc and control groups via anova.

Child-Pugh score7.9 ± 1.47.8 ± 1.68.1 ± 1.67.2 ± 1.4*†
Modified Child-Pugh score5.9 ± 1.45.9 ± 1.46.1 ± 1.65.6 ± 1.4*
MELD score11.4 ± 3.011.6 ± 3.112.3 ± 3.311.8 ± 3.0
Ascites (present/absent)14/2616/2212/276/33
Average of the HE grade1.4 ± 0.51.3 ± 0.91.4 ± 0.50.9 ± 0.9*†
No. of HE episodes per patient3.7 ± 1.23.6 ± 2.33.8 ± 1.21.8 ± 1.8*†
Total bilirubin (mg/dL)1.4 ± 0.71.4 ± 0.81.6 ± 0.71.5 ± 0.6
Albumin (g/dL)3.3 ± 0.53.4 ± 0.53.3 ± 0.53.5 ± 0.6*
Prothrombin time (%)70.4 ± 12.171.3 ± 12.670.4 ± 12.970.6 ± 13.2
Ammonia (μg/dL)129.1 ± 54.9112.0 ± 56.3137.8 ± 50.890.4 ± 33.4*†
Platelets (×104/μL)10.8 ± 4.911.5 ± 6.110.0 ± 4.110.6 ± 6.7
Zinc (μg/dL)51.6 ± 13.356.5 ± 12.448.9 ± 9.372.3 ± 17.2*†
Zinc deficiency (present/absent)38/231/738/115/24*
Fischer ratio1.8 ± 1.01.9 ± 0.91.7 ± 0.61.8 ± 0.6
NCT-A (s)72.6 ± 30.570.6 ± 29.778.8 ± 27.063.6 ± 22.4*†
NCT-A (abnormal/normal)23/1722/1630/916/23*
NCT-B (s)141.6 ± 31.3139.0 ± 34.9145.8 ± 30.4127.1 ± 30.5*†
NCT-B (abnormal/normal)31/924/1432/719/20*
DST (points)10.9 ± 3.810.8 ± 3.910.3 ± 3.211.4 ± 2.8*†
DST (abnormal/normal)31/926/1231/821/18*
Mean numbers of abnormal NP tests2.1 ± 1.01.9 ± 1.22.4 ± 1.01.4 ± 1.3*†

Furthermore, 16 patients in the zinc group improved to grade 0 HE, and six patients in the control group improved to grade 0 HE at the end of study period (41% vs. 16%, P = 0.02, Chi-square test).

The mean value of each NP test was improved by zinc administration (NCT-A, P < 0.0001; NCT-B, P < 0.001; DST, P < 0.01, anova) and differences between the two groups at 6 months after treatment were also statistically significant (NCT-A, P < 0.01; NCT-B, P < 0.01; DST, P = 0.02, anova) (Table 3).

Moreover, the frequency of an abnormal result of each NP test decreased significantly after 6 months of zinc administration (NCT-A, P < 0.001; NCT-B, P < 0.01; DST, P = 0.02, McNemar’s test). However, no significant change was observed in the control group (NCT-A, P = 0.71; NCT-B, P = 0.06; DST, P = 0.44, McNemar’s test). Significant change in the mean numbers of abnormal NP tests was observed between the zinc and control groups (P = 0.04, anova).

Associations of HRQOL and clinical variables

Univariate associations between clinical variables and each of the 6-month changes in the two summary scores of SF-36 (ΔPCS and ΔMCS) are analysed.

In univariate analysis, zinc supplementation and blood ammonia level were significantly associated with ΔPCS (P = 0.02, and P = 0.04 respectively), whereas no significant variable was associated with ΔMCS.

In multivariate analysis, zinc supplementation was the only factor that affected ΔPCS (P = 0.03), whereas no significant variable affected ΔMCS (Table 4).

Table 4.   Predictors of change in PCS and MCS (n = 77)
Dependent variableIndependent variableNonstandardized beta coefficient (95% CI)P value
  1. CI, confidence intervals; HE, hepatic encephalopathy; PCS, physical component scale; MCS, mental component scale; ΔPCS, change in PCS between two visits; ΔMCS, change in MCS between two visits.

  2. The clinical variables were considered either continuous or binary variables.

  3. Multivariate analysis; P values of stepwise backward multiple linear regression analysis are reported.

  4. * Significant (< 0.05).

ΔPCSZinc supplementation*5.36 (0.47, 10.25)0.03
 Ammonia (μg/dL)5.78 (−0.80, 12.36)0.08
ΔMCSAge (years)0.26 (−0.044, 0.57)0.09
 Cause of cirrhosis (viral)−0.73 (−0.34, 9.05)0.07

Discussion

In this study, zinc supplementation in addition to standard treatment clearly demonstrated improved liver function, HE, NP tests, and HRQOL especially PCS, in patients with decompensated liver cirrhosis.

Several reports describe zinc supplementation improving psychometric performance with a reduction in blood ammonia level in HE patients.10, 11 In addition, combinations of zinc and conventional therapies such as a protein-restricted diet including BCAA preparation or lactitol have been reported as effective therapies for HE.

Hayashi et al.28 reported that combination treatment with BCAA and zinc supplements decreased blood ammonia level more than BCAA treatment alone in cirrhotic patients during the study period. They inferred from this result that zinc administration increased ability to metabolize ammonia in the liver as compared with the nitrogen load by BCAA supplementation.

Katayama29 reported that while either lactitol or zinc alone reduced ammonia levels to about 70% of pre-treatment concentrations, combination treatment reduced them to about 50%. Synthetic disaccharides are effective in reducing blood ammonia by mainly inhibiting absorption of ammonia from the intestine, and are known to improve NP tests.30, 31

This synergism of two agents for reducing ammonia caused by different mechanisms seems to be effective in patients unresponsive to standard therapies only.

Zinc supplementation in addition to rifaximin, a minimally absorbed antibiotic, may be more effective in refractory HE, because a recent study32 showed the superiority of rifaximin therapy over treatment with lactulose alone in severe recurrent HE patients.

Our report is the first study confirming that zinc therapy is an effective treatment for HE patients who were unresponsive to standard therapies in terms of improving patient HRQOL.

In this study, zinc supplementation improved not only NP test results, but also the HRQOL in patients with HE. These results indicated that improvement in HRQOL was linked to improvement in cognitive functions. Moreover, in multivariate analysis, zinc supplementation was significantly associated with improvement in PCS, whereas zinc supplementation was not significantly associated with change in MCS.

There are two possible explanations for these results. The first is improvement of physical manifestations in HE patients via zinc supplementation. Physical manifestations in HE patients presumably were tremor, ataxia, asterixis and ascites; consequently, improvements in these manifestations via zinc supplementation might correlate with change in PCS.

Another possible explanation is difficulties in measurement of the mental component of mild HE patients via SF-36. Impairment of concentration such as shortened attention span or arithmetic disturbance is a main symptom in Grade 1 HE. Although improvement in concentration may improve the mental aspect, the SF-36 mental component appears to focus on energy rather than concentration, and this may explain the reason that the mental scale does not change.

Hepatic encephalopathy, ascites, and Child-Pugh score have been known as variables associated with HRQOL in patients with decompensated liver cirrhosis.24, 33, 34 In particular, HE and Child-Pugh score negatively correlated with physical health more than mental health in the SF-36.24, 34–36 The results of zinc supplementation in this study were explanatory to the results of these reports.

In this study, serum albumin level significantly increased, and the rate of ascites’ presence showed a tendency to decrease in the zinc group. BCAA recovers the impaired turnover kinetics of albumin in cirrhotic patients;37 particularly, leucine is known to activate in vitro albumin synthesis through the mammalian target rapamycin (m-TOR).38 Reduction in BCAA in advanced liver cirrhosis was explained by enhanced consumption of BCAA for ammonia detoxification in skeletal muscle and for energy generation. Zinc supplementation enhanced ammonia detoxification in the liver, and consequently led to the alleviation of ammonia disposal in skeletal muscle. For this reason, zinc supplementation might result in a decrease of muscular BCAA consumption and consequently, the administered BCAA might be used for albumin synthesis, and lead to an increase in serum albumin level and a decrease in ascites. In this study, the Fischer ratio did not increase in the zinc group in spite of putative decrease in BCAA consumption, indicating the possibility of utilization of BCAA to synthesize albumin.

The Child-Pugh score, the modified Child-Pugh score and mean HE grade improved, and serum albumin level increased by zinc supplementation. However, no significant change was observed in MELD score. Complications and laboratory parameters such as HE and ascites, and serum albumin are included in the Child-Pugh score, while HE and ascites are not part of the MELD score. For this reason, the Child-Pugh score correlates better with HRQOL than the MELD score.35 It is necessary to understand the impact of liver disease on HRQOL because of the prolonged wait for transplantation. This result indicates that careful attention is required for patients awaiting transplant with advanced cirrhosis, because the authors found a poor correlation between HRQOL and MELD scores.

There are several reports describing that plasma zinc concentrations of cirrhotic patients return to normal after treatment with 200–600 mg per day of zinc sulphate or 600 mg per day of zinc acetate.10, 11, 28 In spite of the small zinc dose (51 mg per day), the polaprezinc supplementation increased serum zinc level significantly in this study. Polaprezinc, a synthesized agent N-(3-aminopropionyl)-l-histidinate zinc, is a chelate compound consisting of zinc and l-carnosine, used clinically as an anti-peptic-ulcer drug.39l-carnosine enhances a strong zinc absorption from the intestine in animal models,40 and 150 mg of polaprezinc (containing 34 mg of zinc per day) sustained human serum zinc levels higher than those of the 300 mg of zinc sulphate.41 Limitations of this study are that it was conducted in a short period and non-blinded fashion; furthermore, it was assessed via nondisease-specific instruments. Although treatment bias is unavoidable, it may be minimal because this study was objectively assessed by the same scorer unaware of the assignment, via a battery of validated NP tests and relevant instrument of the SF-36 questionnaire. In spite of a generic instrument of the SF-36, effectiveness in HRQOL after zinc therapy was confirmed. A disease-specific instrument such as the Chronic Liver Disease Questionnaire (CLDQ)42 may be more reliable to measure the HRQOL in cirrhotic patients; however, most of the data using CLDQ have been derived from non-Japanese patients; as such, little data for Japanese patients are available. A generic SF-36 is most widely used and valid in Japanese populations. Accordingly, SF-36 was adopted in this study.

Several studies43, 44 concluded that zinc supplementation failed to improve HE.

The different outcome between those studies and the present one may be due to the difference in zinc supplementation period or background of participants.

In conclusion, zinc supplementation may be an effective treatment for HE in terms of improving patient HRQOL. Prospective double-blind studies with large samples are necessary.

Acknowledgements

Declaration of personal and funding interests: None.

Ancillary