Hepatic encephalopathy (HE) is a brain dysfunction

Transcription

1 HEPATOLOGY, VOL. 67, NO. 2, 2018 AMERICAN ASSOCIATION FOR THE STUDY OFLIVERD I S E ASES L-Ornithine L-Aspartate in Bouts of Overt Hepatic Encephalopathy Sandeep Singh Sidhu, 1 Barjesh Chander Sharma, 2 Omesh Goyal, 1 Harsh Kishore, 1 and Navpreet Kaur 3 High-quality data on the efficacy of L-ornithine L-aspartate (LOLA) in patients with cirrhosis and bouts of overt hepatic encephalopathy (OHE) are missing. We evaluated the efficacy of intravenous LOLA in the reversal of bouts of OHE in patients with cirrhosis. In this prospective, double-blind, randomized, placebo-controlled trial conducted at two tertiary care institutes in India, 370 patients with cirrhosis and bouts of OHE were screened. After exclusion, 193 (52.16%) patients were randomized to receive either intravenous infusions of LOLA (n 5 98), 30 g daily, or placebo (n 5 95) for 5 days. Standard of care treatment (including lactulose and ceftriaxone) was given in both groups. Randomization was done centrally ( com/). All study personnel were blinded to the treatment assignment. Fasting venous ammonia levels were estimated daily from 0 to 5 days. Serum tumor necrosis factor-alpha, interleukins, hemogram, and liver and renal function tests were performed at days 0 and 5. Primary outcome was mental state grade at day 5 of treatment. The grade of OHE was significantly lower in the LOLA group (compared to placebo) on days 1-4 but not on day 5. The mean time taken for recovery was lower in the LOLA group compared to the placebo group ( versus days, P ; 95% confidence interval to ). Venous ammonia at day 5 and length of hospital stay were significantly lower in the LOLA group. No significant difference in interleukins was seen between the groups. Conclusion: In patients with bouts of OHE, intravenous LOLA (as an addon therapy to lactulose and ceftriaxone) significantly improves the grade of OHE over days 1-4, but not on day 5, and decreases venous ammonia, time of recovery, and length of hospital stay. (HEPATOLOGY 2018;67: ) SEE EDITORIAL ON PAGE 476 Hepatic encephalopathy (HE) is a brain dysfunction caused by liver insufficiency and/or portosystemic shunting; it manifests as a wide spectrum of neurological or psychiatric abnormalities ranging from subclinical alterations to coma. (1) Bouts of overt HE (OHE) are estimated to occur in 30%-40% of patients with liver cirrhosis. (2) OHE is a clinical landmark in these patients. Severe OHE in patients with cirrhosis is associated with a mortality of >50% in the first year alone. (3,4) The chief neurotoxin implicated in the development of HE is ammonia. (5) Ammonia is released from kidney, muscle, and intestine. The colonic urease bacteria and small intestine brush border mucosal enzymes convert ingested proteins into ammonia, which enters the portal vein. In cirrhosis, ammonia is not detoxified in the liver and enters the systemic circulation directly or by bypassing the liver through portosystemic shunts. Ammonia crosses the blood brain barrier and enters the astrocytes. The ammonia induces almost instantaneous release of reactive oxygen species/reactive nitrogen species and glutamine accumulation. Finally, astrocyte swelling/hyperhydration ensues, and this, in turn, disturbs neuron transmission, leading to HE. (6) However, the correlation between arterial ammonia concentration and the clinical manifestation of HE is often poor, (5) and single time point ammonia measurements are frequently noncontributory to the diagnosis of HE. (7) Absence of hyperammonemia, however, has a negative predictive value because patients with low ammonia have good liver function, no portosystemic shunt, and no gut dysbiosis. Abbrevations: CI, confidence interval; ELISA, enzyme-linked immunosorbent assay; HE, hepatic encephalopathy; IL, interleukin; LOLA, L-ornithine L-aspartate; MELD, Model for End-Stage Liver Disease; OHE, overt hepatic encephalopathy; OR, odds ratio; SD, standard deviation; TNFa, tumor necrosis factor-alpha. Received April 24, 2017; accepted July 25, Supported by Dayanand Medical College and Hospital. Copyright VC 2017 by the American Association for the Study of Liver Diseases. View this article online at wileyonlinelibrary.com. DOI /hep Potential conflict of interest: Nothing to report. 700

2 HEPATOLOGY, Vol. 67, No. 2, 2018 SIDHU ET AL. Bacteria and their by-products, termed endotoxins/lipopolysaccharides, can activate Kupffer cells and hepatic macrophages through activation of toll-like receptors such as toll-like receptor 4, generating a proinflammatory response with the production of tumor necrosis factor-alpha (TNF-a) and interleukins. (8) An important aim of HE treatment is reduction of the ammonia in the body by decreasing ammonia production and increasing its detoxification using nonabsorbable disaccharides such as lactulose, antibiotics such as rifaximin, oral branched-chain amino acids, intravenous L-ornithine L-aspartate (LOLA), and probiotics. (9) Dopaminergic agonists and benzodiazepine receptor antagonists (flumazenil) have also been evaluated for the treatment of HE. (10,11) The effectiveness of nonabsorbable disaccharides in the management of HE is well established. (12,13) Lactulose is metabolized by bacteria in the colon to acetic and lactic acids, which reduce colonic ph, decrease survival of urease-producing bacteria in the gut, and facilitate conversion of NH 3 to NH 1 4. The cathartic effect of lactulose also increases fecal nitrogen waste. LOLA, the salt of the natural amino acids ornithine and aspartate, acts through the mechanism of substrate activation to detoxify ammonia. In clinical trials (5,7) intravenous infusions of LOLA and oral LOLA showed a statistically significant effect with respect to reduction in HE grade, reduction of blood ammonia, and positive effects on psychomotor function in patients with cirrhosis and minimal or chronic grade 1 HE versus placebo. However, there is lack of highquality data on the efficacy of LOLA in patients with bouts of OHE, which is one of the major causes of hospital admissions and resource use in patients with decompensated cirrhosis. The aim of this prospective, randomized, placebocontrolled trial was to evaluate the efficacy of intravenous LOLA in the reversal of OHE in patients with cirrhosis. Patients and Methods This prospective, double-blind, randomized, controlled trial was conducted at the Gastroenterology Department in two tertiary care centers: Dayanand Medical College and Hospital (Ludhiana, Punjab, India) and Govind Ballabh Pant Institute of Postgraduate Medical Education and Research (New Delhi, India). Patients were enrolled from December 2013 to November Inclusion criteria were patients with cirrhosis aged years and having bouts of OHE grade 2-4, according to the West Haven criteria, with or without precipitating factors. Diagnosis of cirrhosis was based on clinical, biochemical, radiological, and/or histological criteria. Exclusion criteria were terminally ill patients, advanced cardiac or pulmonary disease, presence of underlying chronic renal failure (serum creatinine > 3 mg/dl), neurodegenerative disease (including head injury and drug intoxication) or major psychiatric illness, use of sedatives or antidepressants, pregnancy or breast-feeding, hepatocellular carcinoma, and acuteon-chronic liver failure. LABORATORY ASSESSMENT Fasting venous ammonia levels were estimated using a blood ammonia meter (PocketChem BA) at baseline and then daily for the next 5 days. The PocketChem BA ammonia meter is an apparatus for quick bedside blood ammonia estimation. Serum concentrations of interleukin-1b (IL-1b), IL-6, IL-10, and TNF-a were measured at days 0 and 5 using a commercially available enzyme linked immunosorbent assay (ELISA) system (Diaclone SAS, Besancon, France). The venous blood of patients was withdrawn, and serum was stored immediately at 70 8C until ELISA was done. The results of ELISA were expressed in picograms per ARTICLE INFORMATION: From the 1 Department of Gastroenterology, Dayanand Medical College and Hospital, Ludhiana, Punjab, India; 2 Department of Hepatology, Institute of Liver & Biliary Sciences, New Delhi, India; 3 Department of Biochemistry, Dayanand Medical College and Hospital, Ludhiana, Punjab, India. ADDRESS CORRESPONDENCE AND REPRINT REQUESTS TO: Sandeep Singh Sidhu, M.B.B.S., M.D., D.M. Department of Gastroenterology, Dayanand Medical College and Hospital Ludhiana, Punjab, India Tel: sandeepsidhu1963@gmail.com 701

3 SIDHU ET AL. HEPATOLOGY, February 2018 milliliter. The lower limit of detection was 1 pg/ml. Hemogram, liver and renal function tests, prothrombin time index, and fasting blood sugar were performed at days 0 and 5. Child-Turcotte-Pugh score and Model for End-Stage Liver Disease (MELD) score for stage of cirrhosis were calculated at days 0 and 5. Etiology of cirrhosis was considered to be alcohol if there was a history of significant alcohol intake (20-30 g/day for women and g/day for men for 10 years). Each patient was tested for hepatitis B surface antigen and anti hepatitis C antibody using a thirdgeneration commercial ELISA. Wherever indicated, autoimmune hepatitis was diagnosed using antinuclear antibody, anti smooth muscle antibody, and antibody to liver/kidney microsome; and hemochromatosis was diagnosed using serum iron, total iron binding capacity, ferritin, and transferrin saturation. Each patient had an abdominal ultrasonography done. Upper gastrointestinal endoscopy was done, if indicated. Written informed consent was obtained from blood relatives or others holding the power of attorney because the patients were encephalopathic at enrollment. The study protocol was approved by the institutional ethics committee, and it conformed to the ethical guidelines of the 1975 Declaration of Helsinki. The trial was registered with ClinicalTrials.gov (NCT ). ASSIGNMENT AND MASKING Consecutive patients diagnosed as having bouts of OHE were randomized into two groups (LOLA or placebo). Each investigator was provided with the code of the group assignment and randomization schedule for each patient; in the event of an emergency, the code was to be used to identify the assigned group and randomization schedule. However, during the course of the study, no investigators were unblinded to treatment assignment. All study personnel were blinded to the treatment assignment for the duration of the study. Placebo and LOLA were similar in appearance and mode of administration. The study drugs LOLA and placebo were provided by Win-Medicare Pvt. Ltd. (India). Randomization was done centrally ( Patients were randomized to LOLA or placebo at a ratio of 1:1 using random permuted blocks of size 2 or 4, stratified by site. The code was revealed only after the recruitment, data collection, and analysis of results had been carried out. Enrollment of patients, assessing eligibility, and obtaining informed consent were carried out by one of the investigators (S.S.S., O.G., or B.C.S.). STUDY INTERVENTION Patients were randomized to receive either LOLA or placebo. LOLA was supplied as ampoules (each containing 5 g LOLA in 10 ml clear solution). Six ampoules of LOLA (30 g of drug) diluted in 440 ml of dextrose 5% were given as an intravenous infusion at a rate of 21 ml/hour (1.25 g/hour) over 24 hours for 5 days. Placebo vials (six vials containing 10 ml of sterile water) diluted in 440 ml of dextrose 5% were given as an intravenous infusion at a rate of 21 ml/hour over 24 hours for 5 days. Hyponatremic (serum Na 1 < 120 meq/l) patients were given LOLA in 0.9% saline. The duration of therapy of 5 days was chosen because other similar studies had given intravenous LOLA between 3 and 7 days. CLINICAL MONITORING All OHE patients defined as grades 2-4 per the modified West Haven criteria were admitted to the intensive care unit and shifted to ward once they reverted back to covert HE, defined as minimal HE and grade 1 HE. Continuous monitoring of vital signs was done. Neurologic monitoring was done twice daily, and grading of mental state was done per the modified West Haven criteria. Clinical responses were assessed as follows (14) : 1. Resolution: disappearance of the OHE clinical syndrome during the study 2. Improvement: a decrease in HE by one grade but not reaching covert HE 3. No Improvement: no improvement in HE 4. Failure: shift to a higher grade of HE DIET AND CONCURRENT THERAPY Patients were allowed to take a kcal/kg body weight diet: 50%-60% carbohydrate, 20%-30% vegetable protein (1-1.5 g/kg body weight), 10%-20% fats, branched chain amino acid enriched supplements, vitamins (A, D, E, K) and minerals (zinc, calcium, and magnesium) if found deficient, and <2 g salt/day in patients with ascites orally or through a Ryles tube in patients with higher grades of OHE/coma. Patients with severe hyponatremia (serum Na 1 < 120 meq/l) were treated with fluid (1-1.5 L) and salt restriction. Energy and nitrogen requirements in patients with HE are unlikely to differ substantially from those 702

4 HEPATOLOGY, Vol. 67, No. 2, 2018 SIDHU ET AL. recommended in patients with cirrhosis, viz kcal/g and g/kg protein daily. (15) STANDARD OF CARE TREATMENT Standard of care treatment for patients with cirrhosis and bouts of OHE was given to all patients in both groups. 1. Other potential causes for altered mental state, e.g., head injury or drug intoxication, were identified. 2. Precipitating causes of encephalopathy, such as sepsis, gastrointestinal bleeding, hypokalemia (serum potassium < 3.5 meq/l), alkalosis, azotemia, dehydration, diuretics, constipation, large protein intake, and psychoactive drugs, were identified and treated. 3. Lactulose syrup, ml in three divided doses, was given through a nasogastric tube or orally to produce two or three semiformed stools and/or lactulose retention enemas (300 ml lactulose ml water) were given twice daily. In patients with constipation, higher doses of oral lactulose (120 ml three times daily) along with lactulose enemas were given. 4. No other ammonia-lowering agent, such as rifaximin or probiotics, were given to these patients. 5. Prophylactic parenteral antibiotics were started at presentation in all patients intravenous cefotaxime 2 g twice daily (after sending blood, urine, and ascitic fluid cultures). Antibiotics were modified per the infection and based on the culture sensitivity, if clinically indicated. Antibiotics were continued until the patient completed neurologic recovery. OUTCOME MEASURES The primary outcome was mental state grade after 5 days of treatment. Secondary outcomes were change in blood ammonia levels at day 5 compared to day 0, change in serum cytokines levels (IL-1b, IL-6, IL-10, TNF-a [time frame 5 days]), rate of recovery from HE, length of hospital stay (time frame 4 weeks), and mortality (time frame 4 weeks). SAMPLE SIZE The sample size for this study was estimated using the results of a previous study, (16) which reported definite recovery of encephalopathy in 79% of patients in the LOLA group and 55.6% of patients in the placebo group. At 80% power and a 5% significance level (twosided), 71 patients in each group would be required to detect a 22% difference in proportions. Assuming a dropout rate of 15% approximately, we needed to enroll 164 patients, i.e., 82 in each treatment arm. STATISTICAL ANALYSIS All analyses were based on the intention-to-treat population unless otherwise stated. Patients were analyzed according to the trial arm that they were randomized to, regardless of the treatment they actually received, study eligibility, or compliance postrandomization. All missing values were assigned the value of a high mental state grade 4 and as an average grade 2 HE, and calculations were performed. A sensitivity analysis was done by performing a complete cases analysis on the intention-to-treat population. Patients with missing data were excluded, invoking a missing-at-random assumption. A secondary analysis was performed on the primary outcome based on the per protocol population. The per protocol population included all randomized patients who remained eligible for the study and were compliant to the allocated treatment for at least 5 days postrandomization. Any patients who withdrew from the trial or treatment or who did not receive the allocated treatment were excluded. There was no interim analysis of the trial data, and the final analysis was performed after all follow-up data had been collected. The 95% confidence intervals (CIs) are presented with the significance of P values assessed based on a 5% significance level. All statistical tests and CIs were two-sided. No adjustment for multiplicity of outcomes was made. All statistical analyses were adjusted for center where possible. Outliers were identified by viewing boxplots and/or histograms of the outcome variables of interest and were queried at the data checking stage if an error was suspected. All analyses included outliers as standard. Distributional assumptions underlying the regression analyses were assessed by visual inspection of residual plots. Normality was examined by normal probability plots. If the distributional assumptions for the parametric approach were not satisfied, further data transformation (for achieving normality) or other suitable methods were considered. This was documented in the statistical report together with the reasoning supporting the action taken, if applicable. 703

5 SIDHU ET AL. HEPATOLOGY, February 2018 FIG. 1. Patient flowchart. Abbreviation: ITT, intention to treat. An ordinal regression model was fitted to the mental state grade variable, with trial arm as an explanatory factor (with placebo as the reference category), adjusting for baseline mental state grade and site. If, however, > 90% of outcome data belonged to only two response categories, then a logistic regression model was fitted instead, with trial arm, baseline mental state grade, and site as explanatory variables. A cross-tabulation of mental state at baseline and at 1-5 days against trial arm was also reported. A secondary analysis was performed using the same method as described above but based on the per protocol population, that is, only including those patients who were compliant to the medication they were randomized to, during the first 5 days of treatment. The secondary outcome measures of blood ammonia levels at 5 days and serum cytokine levels (IL-1b, IL- 6, IL-10, and TNF-a) at 5 days each were analyzed using multiple linear regression adjusting for baseline, trial arm, and center. These analyses were conducted as complete case analyses, and no imputation of missing values was performed. Rate of recovery from HE was assessed according to the time taken (in days) to recover from grade 2 or above mental state grade to grade 1 or below. A Cox regression analysis was used to compare time until recovery between trial arms, adjusting for center. Similarly, length of hospital stay until hospital discharge was compared between trial arms using linear regression analysis, adjusting for center. Any patients 704

6 HEPATOLOGY, Vol. 67, No. 2, 2018 SIDHU ET AL. TABLE 1. Baseline Characteristics of Patients Group Clinical Indices LOLA (n 5 98) Placebo (n 5 95) P Age Male 89 (90.8) 84 (88.4) 0.17 Past history Cirrhosis* 54 (55.1) 53 (55.8) 0.92 Bleeding 20 (20.4) 13(13.7) 0.06 HE 15 (15.3) 14 (14.7) 0.84 Diabetes mellitus 17 (17.4) 15 (15.7) 0.63 Hypertension 8 (8.1) 11 (11.5) 0.41 Etiology HCV 19 (19.4) 17 (17.8) 0.67 HBV 13 (13.3) 18 (18.9) 0.23 Alcohol 16 (16.3) 20 (21.1) 0.35 Alcohol 1 HCV 33 (33.7) 20 (21.1) Alcohol 1 HBV 6 (6.1) 14 (14.7) 0.01 NASH 5 (5.1) 5 (5.4) 0.98 Other 6 (6.1) 4 (4.4) 0.42 Precipitating factors Diuretics 18 (18.4) 21 (22.8) 0.59 Constipation 44 (44.9) 50 (52.6) 0.39 Gastrointestinal bleed 6 (6.1) 6 (6.3) 1.00 Hyponatremia 10 (10.2) 10 (10.5) 1.00 Hypokalemia 7 (7.1) 12 (12.6) 0.27 Sepsis 33 (33.6) 24 (25.2) 0.16 Alkalosis 3 (3.0) 1 (1.0) 0.31 Hematological and biochemical parameters Hemoglobin (g/dl) TLC (10 3 /ml) Neutrophils (%) PTI-INR Total bilirubin (mg/dl) Albumin Proinflammatory cytokines IL-1b (pg/ml) IL-6 (pg/ml) IL-10 (pg/ml) TNF (pg/ml) CTP classification A 13 (13.3) 14 (14.7) 0.71 B 36 (36.7) 41 (43.2) C 49 (50) 43 (43.3) MELD score Data are expressed as number (percentage) or mean 6 SD. *In the LOLA arm 45% and in the placebo arm 44% of patients were diagnosed with cirrhosis on admission with overt HE. Abbreviations: CTP, Child-Turcotte-Pugh; HBV, hepatitis B virus; HCV, hepatitis C virus; NASH, nonalcoholic steatohepatitis; PTI-INR, prothrombin time index-international normalized ratio; TLC, total leukocyte count. dying during hospital stay were excluded, to avoid biasing the treatment effect estimate. Adverse events were compared between trial arms using Fisher s exact test. All secondary outcomes were tested for superiority and used two-sided 95% CIs. Statistical analysis was mainly conducted using SAS software, version 9.3 (SAS Institute, Cary NC), except that R software was used to perform Fisher s exact tests. Results were checked using SPSS (Statistics for Windows, version 21.0; IBM Corp., Armonk, NY) and R software. All authors had access to the study data and reviewed and approved the final manuscript. Results During the study period, 370 patients with cirrhosis and OHE were screened. Out of these, 177 (47.82%) 705

7 SIDHU ET AL. HEPATOLOGY, February 2018 were excluded for various reasons: 83 (46.89%) were terminally ill, 16 (9.1%) had acute-on-chronic liver failure, 42 (23.7%) had hepatocellular carcinoma, 5 (2.9%) had Wilson s disease as the etiological factor of liver disease, 3 (1.7%) had advanced cardiac or pulmonary disease, 56 (31.6%) had presence of underlying chronic renal failure (serum creatinine > 3 mg/dl), 9 (5.08%) had neurodegenerative disease (including head injury and drug intoxication) or major psychiatric illness, 6 (3.5%) were on sedatives and antidepressants, and 1 (0.5%) had pregnancy. Out of these, 193 (52.16%) were randomized to receive either placebo or LOLA. The flow of patients in the study is shown in Fig. 1. BASELINE CHARACTERISTICS AND FLOW OF PATIENTS The baseline characteristics of the patients are shown in Table 1. Baseline characteristics were similar across treatment arms. At the end of day 5 of treatment, in the LOLA arm, 10 patients died or were discharged against medical advice compared to 11 patients in the placebo arm. PRIMARY ANALYSIS ON THE PRIMARY OUTCOME Intention-to-Treat Primary Analysis (Based on Total N 5 193, Including n 5 98 in LOLA Group and N 5 95 in Placebo Group) An ordinal proportional odds regression model was fitted to grade of HE at 5 days, adjusting for site and baseline grade of HE. Patients with missing values were assigned a high mental state grade of 4. The odds ratio (OR) for group 2 relative to group 1 was 1.44 (95% CI, ), with P There was no significant difference between group 2 and group 1 at 5- day follow-up. After replacing missing values with an average mental state grade of 2, our conclusions were unchanged (OR, 1.39; 95% CI, ; P ). Per Protocol Primary Analysis (Based on Total N 5 162, Including n 5 83 in Group 1 and n 5 79 in Group 2) The ordinal proportional odds regression model was fitted to grade of HE at 5 days, adjusting for site and baseline grade of HE. Patients with missing outcome values were excluded from the analysis. The OR for group 2 relative to group 1 was 1.73 (95% CI, ), with P There was no significant difference between group 2 and group 1 at 5-day follow-up. (Logistic regression analysis results on binary outcome HE grade >0 gave an OR of 1.96 [95% CI, ], with P ) A cross-tabulation of mental state at baseline and at 5 days against the trial arm has also been reported (Table 2). Analysis was performed based on the per protocol population according to the grade of HE on days 1-5. The ordinal proportional odds regression model was fitted to grade of HE at each time point, adjusting for site and grade of HE at baseline. The grade of HE was significantly lower over days 1-4 in the LOLA group. However, on day 5 of therapy, there were no significant differences between the two groups (Table 3). SECONDARY OUTCOME ANALYSES In sensitivity analysis, we removed outliers more than 4 standard deviations (SDs) above the overall mean (if any were present). 1. Venous Ammonia: In the LOLA group, the mean 6 SD venous ammonia level was mg/dl at baseline and mg/dl at day 5. In the placebo group the mean 6 SD venous ammonia level was mg/dl at baseline and mg/dl at day 5. Results from a linear regression analysis showed significant reduction in ammonia levels in the LOLA group compared to the placebo group (P < 0.001) (Table 4). 2. Length of Hospital Stay: The mean 6 SD length of hospital stay was days in the LOLA group and days in the placebo group (OR, 3.08; 95% CI ; P < ). 3. Rate of Recovery: In the LOLA group the rate of recovery (mean 6 SD) was days, while in the placebo group it was days, with P (95% CI to 0.202). The percentage of patients with resolution of OHE from day 1 to day 5 is shown in Fig Change in Interleukin Levels: There was insufficient evidence of a difference in IL-6 between groups. The results for the other secondary outcomes (IL-1b, IL-10, and TNF-a) were found to 706

8 HEPATOLOGY, Vol. 67, No. 2, 2018 SIDHU ET AL. TABLE 2. Cross-Tabulation Between Mental State at Day 5 and Trial Arm, Stratified by Baseline Mental State Grade Grade of HE at Day Total Baseline grade 2 HE LOLA Placebo Baseline grade 3 HE LOLA Placebo Baseline grade 4 HE LOLA Placebo Baseline all patients LOLA Placebo be highly sensitive to outliers and so are difficult to interpret. Patients with only partial follow-up data were censored in the analysis. Patients who died postrandomization were not considered as missing data in case this would introduce bias in the analysis but were instead considered the same as patients who did not recover within 5 days (i.e., censored at 5 days with no recovery having been observed). The negative outcome of mortality or failure to recover within 5 days was analyzed in a logistic regression model adjusting for center. The adjusted OR for group 2 versus group 1 was 1.21 (95% CI, ; P , N 5 181, n 5 91 in group 1 and n 5 90 in group 2). Therefore, there was insufficient evidence of a difference between groups. Eighteen out of 90 (20%) patients in group 2 died or failed to recover in 5 days compared to 16 of 91 patients (18%) in group 1. Six patients died in group 2 compared to only 1 in group 1. The two-sided Fisher s exact P value was nonsignificant for the difference in mortality between groups (P ). Numbers with mortality were too small to do multiple logistic regression analysis. POST HOC EXPLORATORY SUBGROUP ANALYSIS The analysis was performed on the per protocol population, according to MELD (<15 versus TABLE 3. Analysis Performed Based on the Per Protocol Population According to the Grade of HE on Days 1-5 OR for Group 2 Relative to Outcome Variable Group 1 95% CI P Grade of HE at day Grade of HE at day Grade of HE at day Grade of HE at day Grade of HE at day versus 20) and baseline HE grade subgroups. The ordinal proportional odds regression model was fitted to grade of HE at 5 days, adjusting for site subgroup; analysis according to the baseline grade of HE revealed a significant difference for grade 2 HE only (Table 5). FOLLOW-UP Patients who completed 5 days of LOLA/placebo treatment were followed up for 1 month. In the LOLA group, out of 83 patients, 16 (19.27%) expired (9 due to hematemesis, 6 due to sepsis, 1 due to myocardial infarction), 3 (3.61%) developed OHE, and 10 (8.3%) were lost to follow-up. In the placebo group, out of 79 patients, 15 (18.98%) expired (8 due to hematemesis, 4 due to sepsis, and 2 from unknown causes), 3 (3.79%) developed OHE, and 15 (18.98%) were lost to follow-up. ADVERSE EVENTS Various adverse events, such as nausea, vomiting, diarrhea, abdominal pain, fever, spontaneous bacterial peritonitis, insomnia, and headache, were compared between the LOLA and placebo arms. However there was no significant difference of adverse events between the patients in the LOLA and placebo arms (Table 6). FIG. 2. Percentage of patients with resolution of HE from day 1 to day

9 SIDHU ET AL. HEPATOLOGY, February 2018 TABLE 4. Results From a Linear Regression Analysis on Venous Ammonia and Interleukins at Day 5 Adjusted for Baseline and Center Lola (Mean 6 SD) Placebo (Mean 6 SD) Adjusted Difference (95% CI) P Ammonia (mg/dl) (LOLA 5 80, Placebo 5 78) ( ) < IL-1b (pg/ml) (LOLA 5 85, Placebo 5 75) ( 0.4 to 2.3) 0.15 IL-6 (pg/ml) (LOLA 5 83, Placebo 5 76) ( 8.7 to 15.4) 0.58 IL-10 (pg/ml) (LOLA 5 84, Placebo 5 76) ( 0.3 to 12.7) 0.06 TNF-a (pg/ml) (LOLA 5 83, Placebo 5 75) ( 8.95 to 6.17) Discussion The current study is a large prospective, doubleblind, placebo-controlled trial that was conducted at two major tertiary care centers in India. In this study 193 patients were randomized to receive LOLA or placebo. The standard medical treatment given to both groups included lactulose and prophylactic ceftriaxone. The grade of HE was significantly lower over days 1-4 in the LOLA group. However, on day 5 of therapy, there was no significant difference between the two groups. This was possibly due to the effects of LOLA peaking by the fourth day of therapy. Patients receiving intravenous infusion of LOLA had a significantly shorter recovery time and a significantly shorter length of hospital stay. The hospital stay was, however, longer than the time taken to recover from OHE because there were other complications of cirrhosis such as gastrointestinal bleeding and septicemia, which also required treatment. The levels of ammonia at day 5 were also significantly lower in the LOLA group. The side effects profile in the LOLA and placebo groups were similar. The mortality rate at 1-month follow-up was 19.3% in the LOLA group and 18.3% in the control group. The MELD score was 21 in both groups. The very poor liver function correlated with similar grade of HE in both patient groups at the end of 5 days of therapy and the equivalent high mortality at 1 month. TABLE 5. Post Hoc Exploratory Subgroup Analysis on the Per Protocol Population, According to MELD and Baseline HE Grade Subgroups Subgroup OR for Group 2 Relative to Group 1 95% CI P HE grade HE grade HE grade 3 or 4* MELD < MELD MELD > *Sample size was insufficient to consider the HE grade 4 subgroup on its own. Subgroup analysis according to the baseline grade of HE revealed a significant difference for grade 2 HE only. Subgroup analysis per the baseline MELD scores did not show any difference between the two groups. Besides hyperammonemia, inflammation is postulated to be the main driver of OHE. To determine the effect of inflammation: IL-1b, IL-6, IL-10, and TNF-a were analyzed. There was no significant difference in TNF-a and interleukins after completion of treatment between the treatment groups. However, other authors have shown increased levels of serum TNF-a, IL-1b, and IL-6 and their arteriovenous differences consistent with a brain cytokine efflux in patients with a bout of HE. This supports the presence of a central nervous system inflammatory component (neuroinflammation) in HE. (17) In patients with cirrhosis, reduced hepatocyte mass results in reduced urea and glutamine synthesis, which reduces the ammonia detoxification capacity. The effect of 30 g of amino acids in LOLA, which are absent in placebo, on urea synthesis and ammonia levels requires comment. Previously, protein restriction was considered a mainstay of treatment in liver disease with HE. Current research shows that there is no proven association between protein (nitrogen) intake and HE and that patients with protein restrictions often present with worse HE and other outcomes. (15) Disaccharides such as lactulose and rifaximin are effective in the treatment of bouts of OHE. (12,13) At the same time, several randomized controlled trials TABLE 6. Adverse Events Event LOLA (n 5 98) Placebo (n 5 95) Any event 53 (54.1%) 50 (52.6%) Nausea 10 (10.2%) 7 (7.3%) Flatulence 13 (13.3%) 11 (11.6%) Vomiting 1 (1.0%) 1 (1.0%) Diarrhea 2 (2.0%) 5 (52%) Abdominal pain 2 (2.0%) 5 (5.2%) Spontaneous bacterial peritonitis 0 (0.0%) 1 (1.0%) Body pain 38 (38.8%) 41 (43.2%) Insomnia 40 (40.8%) 45 (47.5%) Dizziness 4 (4.1%) 4 (4.2%) Fever 13 (13.3%) 2 (2.0%) 708

10 HEPATOLOGY, Vol. 67, No. 2, 2018 SIDHU ET AL. have compared the efficacy of LOLA on HE with placebo, (16,18-21) no-intervention control, or other treatments (lactulose and probiotic). (22,23) Most of these studies suggest that LOLA is more effective than placebo/no-intervention control in the improvement of HE. (7,16,18,21) There are heterogeneities among the quality and conclusions of these trials. The first study on LOLA was done by Kircheis et al. (18) on patients with cirrhosis, hyperammonemia (>50 lg/dl), and chronic persistent grade 1-2 HE, without precipitating factors. The patients on LOLA (20 g/day for 7 days) had a significantly greater decline in fasting venous ammonia levels along with a significant improvement in the grade of OHE compared to placebo. In a study by Ahmad et al., (16) 80 patients with cirrhosis, hyperammonemia, and OHE were enrolled. There was a significant improvement in the LOLA (20 g/day for 5 days) group with regard to improvement in mental state grade of HE and venous postprandial ammonia levels. Abid et al. (21) conducted a randomized, doubleblind, placebo-controlled, prospective study on 120 patients with OHE using intravenous infusions of 20 g LOLA for 3 days. They concluded that adjuvant therapy with LOLA is safe and effective in improving HE, especially with higher grades. Moreover, LOLA administration was associated with shorter duration of hospital stay. Three meta-analyses which evaluated the effectiveness of LOLA on HE have been published. The first meta-analysis, by Jiang et al., (24) including 212 patients from three randomized controlled trials, concluded that LOLA benefited patients with OHE (grade 1 or 2), whereas the data did not support the use of LOLA for patients with subclinical HE. Another meta-analysis, by Bai et al., (25) included eight randomized controlled trials with 646 patients and compared LOLA with placebo/no-intervention control. LOLA was significantly more effective in the improvement of HE in the whole group. A recent meta-analysis by Zhu et al. (26) concluded that LOLA treatment shows a trend in superiority for clinical efficacy among standard interventions for OHE. In a very recent review, the efficacy and safety of oral and parenteral LOLA in patients with cirrhosis and HE was emphasized. LOLA improved performance in psychometric tests as well as mental state gradation. (27) The limitation of the present study is that the results of subgroup analysis based on baseline MELD scores and grade of HE have to be interpreted cautiously as the type 1 error rate of observing a false-negative result is increased due to multiple testing. These results need to be verified in a future study with baseline stratification into different MELD groups and baseline grade of HE with a much larger number of HE patients, to give more power to the subgroup analysis. We conclude that intravenous infusion of LOLA (as an add-on therapy to lactulose and ceftriaxone) decreases the fasting venous ammonia level and improves the grade of OHE over days 1-4 but not at day 5. It hastens the recovery from OHE and shortens the length of hospital stay compared to placebo. Hence, we recommend that LOLA infusion be given for 4 days. LOLA is well tolerated. Acknowledgment: We thank Richard Parker (Department of Statistics, The University of Edinburgh, Scotland), who provided statistical support and advice to the study. REFERENCES 1) Vilstrup H, Amodio P, Bajaj JS, Cordoba J, Ferenci P, Mullen KD, et al. Hepatic encephalopathy in chronic liver disease. HEPATOLOGY 2014;60: ) Fleming KM, Aithal GP, Solaymani-Dodaran M, Card TR, West J. Incidence and prevalence of cirrhosis in the United Kingdom, : a general population-based study. J Hepatol 2008;49: ) Garcia-Martinez R, Simon-Talero M, Juan C. Prognostic assessment in patients with hepatic encephalopathy. Dis Markers 2011;31: ) Wong RJ, Gish RG, Ahmed A. Hepatic encephalopathy is associated with significantly increased mortality among patients awaiting liver transplantation. Liver Transpl 2014;20: ) Butterworth RF, Giguere JF, Michaud J, Lavoie J, Layrargues GP. Ammonia: key factor in the pathogenesis of hepatic encephalopathy. Neurochem Pathol 1987;6: ) Haussinger D, Schliess F. Pathogenetic mechanisms of hepatic encephalopathy. Gut 2008;57: ) Stauch S, Kircheis G, Adler G, Beckh K, Ditschuneit H, Gortelmeyer R, et al. Oral L-ornithine-L-aspartate therapy of chronic hepatic encephalopathy: results of a placebo-controlled double-blind study. J Hepatol 1998;28: ) Cirera I, Bauer TM, Navasa M, Vila J, Grande L, Taura P, et al. Bacterial translocation of enteric organisms in patients with cirrhosis. J Hepatol 2001;34: ) McGee RG, Bakens A, Wiley K, Riordan SM, Webster AC. Probiotics for patients with hepatic encephalopathy. Cochrane Database Syst Rev 2011:CD008716; doi: / CD pub2. 10) Als-Nielsen B, Gluud LL, Gluud C. Dopaminergic agonists for hepatic encephalopathy. Cochrane Database Syst Rev 2004: CD003047; DOI: / CD pub2. 11) Als-Nielsen B, Gluud LL, Gluud C. Benzodiazepine receptor antagonists for hepatic encephalopathy. Cochrane Database Syst Rev 2004:CD002798; DOI: / CD pub2. 12) Als-Nielsen B, Gluud LL, Gluud C. Nonabsorbable disaccharides for hepatic encephalopathy. Cochrane Database Syst Rev 2004:CD003044; doi: / CD pub4. 709

11 SIDHU ET AL. HEPATOLOGY, February ) Gluud LL, Vilstrup H, Morgan MY. Non-absorbable disaccharides versus placebo/no intervention and lactulose versus lactitol for the prevention and treatment of hepatic encephalopathy in people with cirrhosis. Cochrane Database Syst Rev 2016: CD003044; doi: / CD pub4. 14) Romero-Gomez M, Ampuero J. Deciphering the spectrum of low-grade hepatic encephalopathy in clinical practice. 2014;146: ) Amodio P, Bemeur C, Butterworth R, Cordoba J, Kato A, Montagnese S, et al. The nutritional management of hepatic encephalopathy in patients with cirrhosis: International Society for Hepatic Encephalopathy and Nitrogen Metabolism Consensus. HEPATOLOGY 2013;58: ) Ahmad I, Khan AA, Alam A, Dilshad A, Butt AK, Shafqat F, et al. L-Ornithine-L-aspartate infusion efficacy in hepatic encephalopathy. J Coll Physicians Surg Pak 2008;18: ) Jayakumar AR, Rama Rao KV, Norenberg MD. Neuro inflammation in hepatic encephalopathy: mechanistic aspects. J Clin Exp Hepatol 2015;5(Suppl. 1):S21-S28. 18) Kircheis G, Nilius R, Held C, Berndt H, Buchner M, Gortelmeyer R, et al. Therapeutic efficacy of L-ornithine-Laspartate infusions in patients with cirrhosis and hepatic encephalopathy: results of a placebo-controlled, double-blind study. HEP- ATOLOGY 1997;25: ) Schmid M, Peck-Radosavljevic M, Konig F, Mittermaier C, Gangl A, Ferenci P. A double-blind, randomized, placebocontrolled trial of intravenous L-ornithine-L-aspartate on postural control in patients with cirrhosis. Liver Int 2010;30: ) Sharma P, Sharma BC. Management of overt hepatic encephalopathy. J Clin Exp Hepatol 2015;5(Suppl. 1):S82-S87. 21) Abid S, Jafri W, Mumtaz K, Islam M, Abbas Z, Shah HA, et al. Efficacy of L-ornithine-L-aspartate as an adjuvant therapy in cirrhotic patients with hepatic encephalopathy. J Coll Physicians Surg Pak 2011;21: ) Mittal VV, Sharma BC, Sharma P, Sarin SK. A randomized controlled trial comparing lactulose, probiotics, and L-ornithine L-aspartate in treatment of minimal hepatic encephalopathy. Eur J Gastroenterol Hepatol 2011;23: ) Poo JL, Gongora J, Sanchez-Avila F, Castillo SK, Ramos GG, Zertuche MF, et al. Efficacy of oral L-ornithine-L-aspartate in cirrhotic patients with hyperammonemic hepatic encephalopathy. Results of a randomized, lactulose-controlled study. Ann Hepatol 2006;5: ) Jiang Q, Jiang XH, Zheng MH, Chen YP. L-Ornithine-Laspartate in the management of hepatic encephalopathy: a metaanalysis. J Gastroenterol Hepatol 2009;24: ) Bai ML, Yang Z, Qi X, Fan D, Han G. L-Ornithine-L-aspartate for hepatic encephalopathy in patients with cirrhosis: a metaanalysis of randomized controlled trials. J Gastroenterol Hepatol 2013;28: ) Zhu GQ, Shi KQ, Huang S, Wang LR, Lin YQ, Huang GQ, et al. Systematic review with network meta-analysis: the comparative effectiveness and safety of interventions in patients with overt hepatic encephalopathy. Aliment Pharmacol Ther 2015;41: ) Kircheis G. Current state of knowledge of hepatic encephalopathy (part V): clinical efficacy of L-ornithine-L-aspartate in the management of HE. Metab Brain Dis 2016;31: