Serum ammonia is a strong prognostic factor for patients with acute-on-chronic liver failure

Ammonia is thought to be central to the pathogenesis of hepatic encephalopathy (HE), but its prognostic role in acute-on-chronic liver failure (ACLF) is still unknown. We aimed to determine the association between serum ammonia level and short-term prognosis in ACLF. Furthermore, we performed an in-depth evaluation of the independent effect of serum ammonia level on the short-term prognosis of hepatitis B virus (HBV) reactivation-induced ACLF patients. We identified 174 patients as part of prospective observational studies in patients with ACLF. Plasma ammonia levels were measured on admission, and several prognostic scores were used to determine the prognostic effect of ammonia. The 28-day patient survival was determined. Receiver operating characteristic analysis was used to identify the cut-off points for ammonia values, and multivariable analysis was performed using the Cox proportional hazard regression model. Plasma ammonia was significantly higher in nonsurvivors (83.53 ± 43.78 versus 67.13 ± 41.77 µmol/L, P = 0.013), and ACLF patients with hyperammonemia had significantly higher 28-day mortality than those without hyperammonemia. Ammonia was also closely related to ACLF grade (P < 0.001) and organ failure, including liver (P = 0.048), coagulation (P < 0.001) and brain (P < 0.001). HBV reactivation serves as the main precipitating factor in the ACLF population. Subgroup analysis showed that ammonia is also a strong prognostic factor in the HBV reactivation-induced ACLF population. Ammonia level is closely correlated with failure of other organs and is an independent risk factor for mortality in ACLF and the special population defined as HBV reactivation-related ACLF. Based on the results from our study, we measured serum ammonia in the population with ACLF, which strongly indicates their prognosis. It serves as an important biomarker and a therapeutic target.

Multiple studies have highlighted that hyperammonemia plays a critical role in the development of hepatic encephalopathy (HE) in patients with liver cirrhosis and other liver diseases. A large amount of serum ammonia escapes liver metabolism in acute liver failure (ALF) patients, and high ammonia concentrations are closely related to a high incidence of cerebral edema and herniation 1 . Hyperammonemia may aggravate liver injury by impairing liver cells, accelerating immune dysfunction, activating hepatic stellate cell proliferation and reducing liver recovery 2 . Although ammonia is not closely related to HE grade/coma score in patients with liver cirrhosis, those patients with Grade 4 HE have higher infection and systemic inflammation in vivo as they had a higher systemic inflammation score and SOFA score 3 . Ammonia is highly related to the risk and frequency of HE episodes, while glycerol phenylbutyrate can decrease the level of ammonia and risk of HE episodes regardless of the basal level of ammonia 4 . Regardless of ACLF severity, patients with HE had higher mortality. Moreover, they demonstrated that ammonia, abnormal cerebral oxygen consumption and systemic inflammation may play critical roles in HE episodes 5 . In a recent study, Shalimar et al. included 498 patients with liver cirrhosis and showed that the serum ammonia level is highly correlated with the severity of HE and the incidence of 28 www.nature.com/scientificreports/ infection; (5) acceptance of liver transplantation; and (5) no detection of serum ammonia. HBV reactivation was defined as a ≥ 2 log increase in the HBV DNA level from a previously stable baseline level or a level ≥ 100 IU/ mL in patients whose HBV DNA had been undetectable or ≥ 20,000 IU/mL in those negative for HBV DNA at baseline 11,12 . We used CLIF criteria for further classification 13 . Briefly, liver failure (TBil ≥ 12 mg/dL), brain failure (grade III-IV HE), kidney failure (serum creatinine ≥ 2 mg/dL), coagulation failure (INR ≥ 2.5 or platelet count ≤ 20 × 109/L), circulation failure (mean arterial pressure < 70 mmHg or treatment with a vasoactive agent), and lung failure (PaO2/FiO2 ≤ 200 or SpO2/FiO2 ≤ 214) were recorded. This study was approved by the Clinical Research Ethics Committees of the First Affiliated Hospital, Zhejiang University School of Medicine. After exclusion, a total of 174 patients were recruited, and their baseline venous ammonia was measured at admission. We started the follow-up of all included patients at the onset of hospital admission and obtained all prognostic information from medical records or telephone contact after they were discharged. The primary endpoint of the study was 28-day mortality, and we recorded their survival time up to 28 days.
Measurement of parameters/data collection. We  ACLF treatments. The ACLF patients were managed according to established guidelines 8,17,18 . To eliminate or control precipitating factors/complications, patients who were HBV DNA positive were immediately given nucleoside analogs daily according to their previous usage of nucleoside analogs. Patients using hepatotoxic drugs or actively drinking alcohol were required to stop using or abstain from alcohol. Patients with bacterial infection were immediately treated with empirical antibiotic therapy, and adjustment of antibiotic therapy was based on bacterial culture and antibiotic sensitivity tests. In addition, weight-based intravenous albumin was used, especially in patients with spontaneous bacterial peritonitis (SBP); furthermore, all patients with acute variceal bleeding received IV somatostatin, proton pump inhibitors, and antibiotic prophylaxis. For those with uncontrolled hemorrhage resulting from pharmacological therapy, Sengstaken-Blackmore tube or urgent therapeutic endoscopy were performed. Diagnostic abdominocentesis was performed in patients with ascites to examine whether SBP was present. Those with moderate ascites were treated with restriction of sodium intake and/ or diuretics. Paracentesis combined with IV albumin was used in those with large or refractory ascites. Patients with renal failure were treated with IV albumin, vasoconstrictors and even renal replacement therapy. Patients with HE were given l-ornithine aspartate, lactulose, antibiotics, and discontinuation of potential precipitating events. Fluid replacement was performed in patients with mean arterial pressure (MAP) < 70 mmHg, and vasoactive agents were used when necessary. Oxygen therapy was performed in patients with decreased PaO 2 or SpO 2 . The choice of nasal catheter, mask, or venturi mask oxygen inhalation or mechanical ventilation was based on the severity of respiratory dysfunction. Nutritional support is also the basis of all patients.
Statistical analysis. Prognostic models used in predicting the 28-day mortality of ACLF patients included the Model for End-Stage Liver Disease (MELD); MELD sodium (MELD-Na); the integrated MELD (iMELD); CLIF Consortium Organ Failure score (CLIF-C OFs); and CLIF-Consortium-ACLF (CLIF-C-ACLF) score. The MELD score (range 6-40) was calculated as follows: 9 www.nature.com/scientificreports/ Continuous variables that were normally distributed were expressed as the mean ± standard deviation, and other variables were expressed as the median with interquartile ranges. Categorical data are presented as proportions. Comparison of demographics and clinical features was performed using the chi-square or Fisher's exact test for categorical variables. Continuous data between two groups were compared by Student's t test or Mann-Whitney's U test. Candidate variables (P < 0.10) after a bivariate analysis were entered into multivariable analysis by the Cox proportional hazard regression model 2 . Receiver operating characteristic curves were used to identify the cut-off points for baseline ammonia values. The Kaplan-Meier method was used to generate survival curves. The data were analyzed using SPSS statistics software (version 19.0; SPSS, Chicago, IL) and GraphPad software (version 5.0.1; MedCalc Software, Ostend, Belgium).

Results
Patients' baseline characteristics. Baseline demographic characteristics are shown in Table 1.
Serum ammonia levels at admission were further compared among subjects with various degrees of organ injury according to the CLIF-OF scoring system (Fig. 2). Serum ammonia levels were comparable in subjects with different levels of Cr (P = 0.766) (Fig. 2a), while they were higher in subjects with liver failure (P = 0.048) (Fig. 2b), more severe coagulation and brain injuries (P < 0.001) (Fig. 2c,d). Moreover, serum ammonia levels were higher in subjects with ACLF grade 2 and grade 3 (P < 0.001) (Fig. 2e).
In the group with ammonia ≥ 92.5 µmol/L, ferritin, aspartate aminotransferase (AST), WBC, INR, and PT levels were higher than those the group with ammonia ≤ 92.5 µmol/L (P < 0.05), while TSH, TG, and VLDL were  www.nature.com/scientificreports/ lower than those in the group with ammonia ≤ 92.5 µmol/L (P < 0.05). The prognostic scores, including MELD, iMELD, CLIF-C OF and CLIF-C ACLF, were higher in patients with ammonia ≥ 92.5 µmol/L than in those with ammonia ≤ 92.5 µmol/L (P < 0.01). Patients with higher levels of ammonia had increased mortality at 28 days compared to those with normal ammonia; the phenomenon can also be observed after classification of the population into HBeAg-positive subgroup and HBeAg-negative subgroup. The survival curves stratified by ammonia level are presented in Fig. 4.    (Table 6).

Discussion
Early in 2003, Nicolao et al. stated that there is only a rough and inconstant correlation between hyperammonemia and the clinical severity of HE 23 . EASL/AASLD clinical practice guidelines 17 state that high blood-ammonia levels do not add any diagnostic staging or prognostic value in HE patients with chronic liver disease. They also state that HE cannot be diagnosed according to serum ammonia level since the ammonia level is in the normal range in some patients with clinical HE and neurological abnormalities. Recently, another study debated this opinion and stated that serum ammonia is related to the clinical grade of HE and serves as a prognostic factor in patients with liver cirrhosis 2 . The high mortality of ACLF patients with hyperammonemia is potentially related to the toxicity of ammonia in the circulatory system, which extends beyond the brain 24 , including a role in the pathophysiology of portal hypertension 25 . It is worth noting that the usefulness of measuring serum ammonia in routine practice remains controversial, although multiple studies highlighted that ammonia is a prognostic factor in patients with liver cirrhosis. Based on the results from our study, we measured serum ammonia in the population with ACLF, which strongly indicates their prognosis. Hyperammonemia occurs in patients with liver cirrhosis due to reduced activity of urea cycle enzymes in the liver or portosystemic shunting 26 . Ravi et al. demonstrated that a higher level of ammonia at admission served as an important indicator of in-hospital survival in patients with alcoholic hepatitis 27 . A certain proportion of cirrhotic patients develop HE, 43% of HE patients die within 1 year, and the short-term mortality rate is extremely high in patients with more advanced grades of HE 13 . A quality meta-analysis from the Cochrane Database System Review revealed that the prevention and treatment of HE with l-ornithine l-aspartate significantly reduced the mortality of patients with cirrhosis with acute decompensation when compared to patients treated with placebo, and this treatment decreased the serum level of ammonia to improve the survival rates of these patients 28 . In addition, several studies found that the incidence of HE is closely related to the mortality of patients with chronic liver diseases 29,30 . However, they speculated that a significant correlation exists between increasing ammonia levels and the development of HE or the severity of HE. Two recent studies highlighted that ammonia levels on admission are important predictive factors of in-hospital mortality in decompensated cirrhosis 6,27 .
Currently, no study has directly analyzed the prognostic effect of ammonia in populations with APASL ACLF. Our study adds to these previous studies through inclusion of a defined ACLF population and further analysis www.nature.com/scientificreports/ of HBV reactivation-related ACLF, allowing for analysis of the relationship between ammonia and various prognostic factors in the population. We proved that hyperammonemia is a strong indicator in the prognosis of patients with ACLF, which indicated that lowering ammonia in these patients may prolong the survival time of ACLF regardless of the incidence of HE. Although the data were obtained retrospectively from the studies, all patients were prospectively recruited. In the total ACLF population, an ammonia level of ≥ 89 µmol/L is closely correlated with liver, coagulation, and brain failure, although our data did not find a relationship between higher ammonia and kidney, circulation or respiration failure. Although there is no difference in HE incidence between ACLF with ammonia level of ≥ 89 µmol/L and those without ammonia level of ≥ 89 µmol/L, we found that ammonia level is positively related to HE grade according to CLIF-OF scores. This is in accordance with the most recent study by Shalimar et al. 2 . HBV infection is still an important public health problem since 3/4 of people with HBV infection and chronic carriers with positive HBV surface antigens (HBsAg) are from China. On the other hand, nearly 50,000 people die because of HBV infection every year in China 31 . This reminds us that HBV reactivation is a life-threatening case in the current status. After we analyzed the data from our hospital, we found that most patients underwent progression into ACLF when they withdrew or accepted irregular usage of nucleoside analogs. In addition, HBV reactivation commonly occurs in other patients undergoing impairment of their antiviral immunity, such as chemotherapy, immunosuppressive treatment or biological therapy 32 . Without considering clinical signs of HE or other parameters in ACLF patients, a high level of ammonia indicates a significantly higher risk of death at 28 days. We found that ammonia is still a strong prognostic factor in HBV reactivation-induced ACLF and is closely related to more prognostic scores according to our study. The Kaplan-Meier survival curves further demonstrated that ammonia was still a strong prognostic factor after division of the population into subgroups. In vivo and in vitro studies showed that ornithine phenylacetate significantly decreased the cell death rate of hepatocytes and alleviated the progression of fibrosis after downregulation of serum ammonia levels 33 . Although the pathophysiology of HE is multifactorial and undetermined according to current studies, hyperammonemia, inflammation and genetic factors are three main pathways in the pathogenetic process. In bile duct-ligated www.nature.com/scientificreports/ rats, hyperammonemia-induced brain edema and brain swelling can be reduced by a reduction in ammonia 34 .
In another model of liver cirrhosis, downregulation of ammonia levels protected the brain from a subsequent challenge with lipopolysaccharide 35 . This indicates that ammonia is a potential goal of treatment to prolong the survival time of ACLF patients. However, there are still limitations in this study, as we appealed to expand the study population in future study. The cut-off of the actual ammonia level should be further determined according to additional larger studies around the world. In an Indian study 36 , a cut-off value of 124 µmol/L upon initial evaluation predicted death in 76% of cases. In 2007, Bernal et al. showed that an arterial ammonia level above 100 µmol/L is the cut-off that sensitively and specifically predicts the occurrence of severe HE in 70% of cases 37 . In 2019, another study demonstrated that an arterial ammonia level above 79.5 µmol/L is the cut-off that sensitively and specifically predicts the outcome of patients with liver cirrhosis 2 .
To this end, this study of ACLF patients adds significantly to the evidence that ammonia levels correlate not only with the failure of other organs but also with 28-day mortality. A reduction in ammonia levels may serve as a potential therapeutic target in patients with ACLF. www.nature.com/scientificreports/ Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creat iveco mmons .org/licen ses/by/4.0/.