Epidemiology of liver failure in Asia-Pacific region

1.1 Concept of liver failure

It is important to understand the concept of liver failure because it can manifest in a number of phenotypes. If the onset of liver dysfunction or jaundice is accompanied by encephalopathy within days or weeks, it is called acute liver failure (ALF).¹ If chronic liver disease is rapidly worsened (within 4 weeks) by an acute event such as a hepatitis B or autoimmune flare or excessive alcohol intake, this is called acute-on-chronic liver failure (ACLF).^2-4 Early recognition of ACLF before the onset of sepsis and extrahepatic insults (such as renal, circulatory and respiratory failure) is important to help prioritize organ-specific interventions.⁵ The syndrome of ACLF is potentially reversible² and with early attenuation of acute precipitating event, liver reserve improves, fibrosis regresses and portal pressure decreases. The clinical course and outcomes of ACLF patients are well differentiated from decompensated cirrhosis patients developing acute decompensation (AD) because of an acute event. It is extremely important to clearly distinguish all of these phenotypes as management approaches, prognostic patterns, time to onset of complications and mortality widely differ in each distinct phenotype.

In this review, we shall discuss in detail the current as well as changing epidemiological trends of common liver failure syndromes, ALF and ACLF in the Asia-Pacific region.

4.1 Hepatitis A

Less than 1% of hepatitis A (HAV) related patients develop ALF.²¹ However, data on the epidemiology of HAV-related ALF are limited. The Asian population is no longer uniform in terms of HAV exposure profile, and there is geographic heterogeneity in the aetiology of ALF by region.

In South Korea, rapidly changing socio-economic conditions are causing HAV outbreaks in adolescents and young adults who do not have antibodies to HAV (anti-HAV) due to prior childhood infection. Since 2000, hepatitis A has replaced hepatitis B as the leading cause of acute viral hepatitis in Korean adults.²² HAV accounts for up to 80% of all acute hepatitis infections in South Korea.²³ As a result, the number of HAV ALF cases requiring LT is increasing in South Korea as well. Also, HAV-related ALF in South Korea has been found to progress faster than ALF caused by other causes.²⁴

In China, 47 of the more than 300 000 infected patients died after a massive HAV outbreak in Shanghai. Immunization of the HAV vaccine with improved socio-economic indicators has reduced the incidence of HAV ALF.²⁵

In India, some reports show that adult HAV infections leading to ALF have gradually increased from 1.7% to much higher over the last 30 years.²⁶ HAV was the most common cause of paediatric acute liver failure (PALF) in a large study of paediatric ALF in India. HAV accounted for 45.9% of the total PALF of 285 people.²⁷ The coexistence of dengue and HAV ALF has also been reported from India and may present clinical scenarios that are difficult to manage due to the complex course of the disease.²⁸

Unlike many countries in the Asia Pacific region, HAV rarely causes acute hepatitis-related deaths in Japan. Hepatitis A accounted for 2.3% of all deaths from acute hepatitis in 2015.²⁹ Improved hygiene and living standards have changed the epidemiology of acute viral hepatitis in Japan. In a national study of ALF and late-onset liver failure (LOHF) in Japan, recent poor outcomes in HAV-ALF patients compared to previously reported patients were primarily an increase in age-related metabolic disease.³⁰

Lack of clean water, poor sanitation and poor access to medical facilities across the country, social and political turmoil, and numerous refugees are the main sources of sporadic cases of HAV and HEV in Pakistan. Despite the asymptomatic course of most children, HEV ALF was reported in 30 of 232 children with HAV assessed at Aga Khan University Hospital (Karachi), 1991–1998. The mortality rate was 36.7%.³¹

Acute HAV is rare in countries with very high HDI, but is primarily found in travellers to endemic areas, men who have sex with men and people who inject drugs. HAV is a common cause of death associated with acute hepatitis in Australia. Hepatitis A accounted for 25.9% of all Australian acute hepatitis deaths in 2015.²⁹ Infections are usually associated with high-risk groups such as nursery infants, injected drug users, men having sex with men and travellers to countries where infections are common.

A study by Fijiwara and colleagues from Japan suggested that low viral load and high substitution rates in 50-untranslated regions of the viral genome increase the likelihood of HAV ALF.³² A recent study in South Korea reported that the HAV genotype affects the severity of liver disease. Korean HAV subgenotype IIIA is also associated with ALF, in contrast to previous studies suggesting that all HAV-related ALF patients have subgenotype IA.³³ No clear results were obtained in studies of the differences between patients with a benign HAV course and those with a fulminant course.

The largest study of HAV-ALF patients from Japan, India and the United Kingdom, as well as primarily from South Korea, found that ALFA scores, including age, bilirubin, INR, ammonia, creatinine and haemoglobin levels, had predicted adverse outcomes.³⁴ In a Korean study, MELD and SIRS scores of >23.5 and >3 and above, respectively, were also associated with the development of ALF.²⁷ For paediatric HAV ALF, the PedsHAV model is a simple dynamic bedside prognostic model based on three objective parameters: (1) the presence of grade 3–4 HE, (2) an INR greater than 3.1 and (3) jaundice to HE interval >10 days a simple bedside dynamic HAV specificity based on jaundice.³⁵ In a large study by Jung et al. from South Korea, post-transplant outcomes for HAV ALF patients were worse than for HBV ALF patients. HAV recurrence after LT for HAV ALF is common and often fatal.³⁶ Acute pancreatitis and recurrence of HAV were significant risk factors for graft loss and mortality.

4.2 Hepatitis B

Hepatitis B is associated with significant liver-related morbidity and mortality in the Asia Pacific region. The main routes of hepatitis B transmission of HBV in most Asian countries include mother-to-child horizontal transmission, unsafe injections and blood product transfusions and intravenous drug use. While in Japan, sexual transmission contributes to most HBV infections.³⁷ Acute hepatitis B infection as the cause of ALF has been reported in a small number of patients from multiple centres. The actual incidence of HBV ALF is low (~1%).³⁸ In various series, hepatitis B infection as cause of ALF is based on positive HBsAg and IgM anti-Hbc as well as quantitative HBV DNA polymerase chain reaction assay. Severe reactivation of hepatitis B in patients with previously undiagnosed chronic hepatitis B may mimic severe acute hepatitis B and ALF. Therefore, most studies of HBV ALF also included patients with severe reactivation and ACLF. HBV-ALF generally has a higher core antibody titre for immunoglobulin hepatitis M and a lower viral load.³⁹ These two forms are separate entities, each with its own cause and consequences. HBV-ALF has high mortality and 50–60% of patients die without transplantation. Early transplantation improves short-term and long-term survival outcomes.

In India, HBV infection is a rare cause of ALF and its incidence is decreasing over time. In various series, Hepatitis B infection as a cause of ALF is highly variable ranging from 0.8% and 39%. In a series by Tandon et al. (1984), 33% of ALF were due to HBV infection.¹⁴ In another Indian series (2000), the prevalence of HBsAg was observed in only 10.5% of ALF cases.⁴⁰ In this study, upto 40% of patients with estimated non-AE ALF had detectable HBV DNA despite being anti-HBc negative and it was suspected that HBV core variants might have contributed in HBV ALF causation. In a study by Bhatia et al. (2007) involving 249 pregnant women presenting with ALF, only 2 (0.8%) had acute HBV infection in comparison to 6.1% in non-pregnant females.⁴¹ Similarly, Barkotoki et al. reported that acute HBV as ALF cause in 2 of 160 (1.25%) pregnant women.⁴² In a study by Das et al., 8 of 255 (3.13%) ALF cases were due to HBV infection, whereas 16.5% had combined HBV and HEV infection.⁴³ From various reports, HBV infection does not appear to be the usual cause of ALF cases in India. Acute hepatitis B accounted for 55% of all acute hepatitis deaths in India in 2015.²⁹ However, most of these include patients of hepatitis B reactivation presenting as acute liver failure.

Unlike United States and Europe, precore (G1896A) and core promoter mutations (A1762T / G1764A) are associated with HBV ALF patients in Asia.^{44, 45} A large cross-sectional study in Japan showed the association of ALF with mutations in the genotypes Bj/B1, A1762T / G1764A, G1896A, G1899A and A2339G.⁴⁶ There is a significant increase in viral replication by introducing either the G1896A or core A1762T / G1764A mutations into Bj clones in an in vitro transfection study. Recently, T1961V / C1962D was found to cause S21 substitution of core protein was found to be associated with ALF.⁴⁷ PreS2-deficient surface gene mutations are known to be associated with ALF. These mutants were characterized by misaggregation with viral retention and high replication capacity in hepatocytes.^{48, 49}

In a prospective study of 110 consecutive ALF adult patients from South Korea, 37% had HBV-related ALF. Adult emergency LDLT can be performed quickly and safely in patients with ALF significantly improving survival. The 1-year survival rate for adult LDLT patients was 85%. Of the 55 patients listed but not transplanted, 45 (82%) died within a median of 7 days (range 1–190 days).⁵⁰

Acute hepatitis B infection caused 76.7% of acute hepatitis-related deaths in China in 2015,²⁹ significantly reducing the incidence of HBV-ALF. Owing to universal HBV vaccination at birth in some countries and an increasing uptake of HBV vaccination in past two decades in others, some highly endemic areas in Asia have experienced a substantial decline in HBV ALF incidence. The low incidence even in highly endemic areas such as China and Taiwan suggests that the genetic traits of the contributing host are probably rare.

4.3 Hepatitis E

HEV is considered to be the most common cause of acute viral hepatitis in Asia. Seroprevalence of HEV varied widely across countries and populations. The highest recorded seroprevalence was 77.7% in the lowland community of Laos.⁵¹ On the other hand, the lowest rate among Malaysian urban blood donors was reported to be 2%.⁵² However, even in Singapore, a developed country with a high-income economy, seroprevalence (range 10.5%–22.3%) appeared to be relatively high.⁵³ HEV is recognized as an emerging infectious disease with increasing public health threats worldwide.⁵⁴

HEV genotypes 1 and 2 cause over 20 million infections each year in Asia and Africa resulting in approximately 3.4 million acute hepatitis, 70 000 ALF-related deaths and 3000 stillbirths.^{55, 56}

In India, hepatitis E is the most frequent aetiology cause of acute viral hepatitis (10–40% of acute hepatitis) and ALF (15–45% of ALF) except the northeastern India.^{57, 58} Almost 0.1–1% HEV-infected individuals develop ALF (all genotype 1) with a mortality rate of 30–70%.^{42, 59} Many large hepatitis E epidemics have been reported from India—Delhi (1955: 29300), Aurangabad (1961: 865), Siliguri (1966: 4287), Ahmedabad (1974: 2572), Kanpur (1990: 79091), Nellore (2008: 23915 case). The median age of onset of HEV ALF is lower than other causes of ALF. Interestingly, HEV ALF patients have better survival outcomes (55.1%) than ALF patients with other causes such as anti-tuberculosis treatment (30.0%), NonA-NonE virus (38.1%), HBV (35.9%). The most common cause of ALF during pregnancy in India is HEV (60%).⁶⁰

In China, HEV accounted for 15.4% of all acute hepatitis death.²⁹ Extensive epidemiological surveillance reveals that HEV genotype 4 is the predominant type.⁶¹ Acute HEV genotype 4 infections are a zoonotic disease that is transmitted by eating infected meat and handling pigs. China has made progress in preventing HEV infection through improved food and water hygiene and vaccination.⁶²

As per 2015 GHE dataset, HEV accounted for 29.7% of all deaths from acute hepatitis in Pakistan.²⁹ HEV ALF in pregnant women are more frequent (10–22%) than in men and non-pregnant women (1–2%) and has high maternal mortality rates ranging from 36% to 71%.²³

In Indonesia in 2015, 15.8% of all deaths from acute hepatitis were due to acute HEV infection.²⁹ There were no reported deaths from acute hepatitis E in 2015 GHE dataset. HEV rarely cause acute hepatitis-related mortality in Japan and South Korea.⁶³ Improvements in sanitation and living standards have contributed to a change in the epidemiology of acute viral hepatitis in South Korea.

Pregnant women are susceptible for severe HEV infection, in particular ALF (13 times more in pregnant women than in age-matched men and nonpregnant women).⁴⁰ HEV ALF in pregnancy is characterized by a short jaundice to encephalopathy interval, higher grades of encephalopathy, cerebral oedema and substantial foetal and perinatal mortality. Pregnant women are more likely to develop ALF caused by HEV.⁶⁴ Recent studies by Khuroo et al. have shown a mortality rate of pregnant women with HEV-related ALF upto 75%. However, once ALF develops, maternal outcomes are similar across all trimesters of pregnancy. Pregnant women with sarcopenia may be at higher risk of developing HEV ALF. Pregnancy-related changes in sex hormone levels, impaired immune system including reduced expression of toll-like receptor (TLR)3/TLR7/TLR9, Impaired monocyte macrophage function and high concentrations of cytokines might contribute to increased HEV predisposition of pregnant women.^65-67

4.4 Other viruses including dengue virus

In studies from India, almost 25% adults with ALF have undetermined cause.^{42, 68} In children, the aetiology remains unexplained in 10–20%. Many viruses including Dengue virus, HSV, EBV, VZV might rarely cause ALF but these are not regularly suspected or tested. Most patients with dengue fever have self-limiting mild hepatitis but in almost 5–10% of cases, it may lead to severe hepatitis, defined by an increase of more than 10-fold over transaminase.⁶⁹ Such severe dengue hepatitis is associated with high viral load. However, dengue-related ALF is rare. More recently, many cases of dengue-related ALF were observed in both children and adults during the Indian seasonal epidemic, including cases of mixed infections.⁷⁰ In an Indian study, dengue hepatitis progressed to ALF in 0.35%.⁷¹ The mean age of dengue-related ALF was 29 years. The median time from fever to the onset of ALF was 7.5 days and ALF was associated with high mortality (>50%).⁷² In a study from Malaysia by Tan et al., eight of 155 adult (5.2%) with dengue fever progressed to ALF.⁷³ In a Thai study, MELD score, baseline PH and lactate were independent predictors of ALF in patients with dengue fever.⁷⁴ Early detection of patients with dengue at risk of ALF is important to improve patient outcomes.

4.5 Drug-induced liver injury

In West and developed Asian countries, drugs induced liver injury (DILI) is gradually replacing viruses as predominant aetiology of ALF. Generally, DILI-related ALF is diagnosed after taking a thorough drug history including over-the-counter drugs in patients with negative serology for viral hepatitis A, B and E and autoimmune hepatitis markers. In Asia-Pacific regions, the true incidence of DILI-related ALF is not well described due to underreporting and under-recognition. Type of drugs leading to ALF varies geographically. Globally, Idiosyncratic DILI constitutes 5–13% of the ALF cases.^{17, 75}

Unlike the western countries, where paracetamol toxicity leads to ALF accounts for about 50% of all ALF, in Asia, it accounts for about 1% of all causes of drug-induced ALF.⁷⁶ Recently, there are multiple reports of DILI emphasized by new drugs such as Infliximab, immune checkpoint inhibitors, herbs and nutritional supplements.⁷⁷

In India, ~14% of all DILIs lead to ALF.⁷⁸ In a large study by Indian DILI Network (INDILI) combined antituberculosis drugs (ATDs) (46%), complementary and alternative medicines (14%), antiepileptics (first-generation drugs) (8%), antibiotics (6.5%), antiretroviral drugs (3.5%) and nonsteroidal anti-inflammatory drugs (NSAIDs) (2.6%) are the most common types of drugs that cause drug-induced ALF. In a large single-centre DILI study, nearly ~70% of all causes of DILI ALF were related to ATDs alone and ATDs are also the most common cause of liver transplant for DILI ALF. Potentially hepatotoxic first-line ATDs include isoniazid, rifampicin and pyrazinamide. ALF most often occurs within the first 2 months of initiation of ATDs and is particularly more common in women than men. Recent whole-genome studies have failed to confirm the importance of genome-wide and CYP2E1 polymorphisms as risk factors for ATDs ALF in the Indian population.⁷⁹ Some patients with ATDs ALF develop subacute hepatic failure that may resemble cirrhosis with clinical ascites and irregular liver surfaces and liver nodules on imaging. In one study, nearly 35% of patients hospitalized for ATDs DILI developed ALF resulting in high mortality rate of ~10%. ATDs ALF generally have worse outcomes than other causes of drug-related ALF.⁸⁰

In India, complementary and alternative medicines (CAMs) constitute the second leading cause of drug-induced ALF. These typically contain unlabelled, multi-ingredient mixtures, making identifying the ‘culprit’ substance in the preparation challenging. These are unregulated, under-monitored and accidently contaminated with volatile organic compounds and metals.⁸¹ Immunoallergic liver injury is caused by medications such sulphonamides, aromatic antiepileptic drugs and antiretroviral agents, which manifest with skin rashes, fever, lymphadenopathy, eosinophilia and other hypersensitivity features.⁸² These extrahepatic signs and symptoms can help with early recognition and drug cessation. Although controlled trials are lacking,⁸³ early use of steroids may be useful in patients with hypersensitive syndrome. Rodenticide (yellow phosphorus) induced intrinsic hepatotoxicity is a far more serious problem in South India than paracetamol.⁸⁴

In China, the majority of data on the occurrence of DILI comes from tertiary institutes.⁸⁵ Acute DILI is responsible for ~20% of ALF.⁸⁶ Furthermore, DILI subtypes and their relative incidence may differ across China due to differences in medication practices, variable adverse event reporting system and distinct ethnic groups and populations' specific genetic polymorphism of drug-metabolizing enzymes.^{87, 88} Traditional Chinese medicine (TCM) is growing more popular, and the risks associated with it are frequently overlooked or underestimated. Due to the large and expanding diversity of TCM, easy access to these drugs without a prescription and a lack of knowledge of drug safety issues, the incidence of TCM-related DILI ALF is increasing each year in China. Specific TCM implicated in liver injury include: Ba Jiao Lian (Dysosma pleianthum), Bol Gol Zhee (Fructus psoraleae), Chi R Yun (Breynia officinalis), Jin Bu Huan (Lycopodium serratum), Ma Huang (Ephedra) and Shou Wu Pian (Polygonum multiflorum). TCM (23%), antibiotics (17.6%), anti-cancer drugs (15%), hormonal drugs (14%), cardiovascular drugs (10%), NSAIDs (8.7%), immunosuppressive agents (4.7%) and neuropsychiatric drugs (2.6%) were the major causes of DILI in China, according to a recent report.⁸⁶

In Thailand, acetaminophen (35%) and ATDs (34.7%) were the most prevalent causes of DILI in a large population-based investigation. DILI was responsible for around 1300 admissions every year.⁸⁹ The death rates in the hospital and after 90 days were 3.4% and 17.2% respectively. Non-acetaminophen DILI, cirrhosis, the elderly and concurrent illnesses all increased the chance of mortality.

In South Korea, herbals and complementary and alternative medicine (CAMs) are the major causes of DILI (42%), with an estimated incidence of 12 per 100 000.⁹⁰ In Singapore, CAM was responsible for 81% of DILI.⁹¹ In Japan, herbs were the second most prevalent cause of DILI (17%).⁹²

5.1 Wilson disease (WD)

WD is an uncommon disease with a wide range of clinical manifestations. In 10–20% of individuals, WD manifests as ALF primarily in children and young adults.^{103, 104} Most individuals presenting as ALF-WD have cirrhosis on explant biopsy.^{105, 106} ALF with Coomb's negative haemolytic anaemia can help to diagnose WD. In ALF-WD, an alkaline phosphatase-to-total bilirubin ratio of 4 and an AST: ALT ratio of >2.2 have very high sensitivity and specificity.¹⁰⁷

In Japan, WD presenting as late-onset ALF that can be effectively treated with plasmapheresis, haemodiafiltration with zinc and chelation as bridge to liver transplant. In a study on acute severe WD conducted in India, 68 of 272 individuals with WD had liver failure. The average age was 14.4 years, and 73% of patients had unfavourable outcomes, especially those who presented with ascites.¹⁰⁸

The New Wilson's Index (NWI) has been used to forecast the need for long-term care. Although they need to be validated. Rising bilirubin, hepatic encephalopathy and acute haemolysis have been indicated as stronger predictors of LT than NWI, MELD and PELD. The Molecular Adsorbent Recirculating System (MARS) and total plasma exchange (TPE) are effective copper chelating therapies that can be used as a bridge to LT.

5.2 Vascular causes

Budd-Chiari syndrome (BCS) is a very rare cause of ALF accounting for 0.9–1.5% of all ALFs.^{109, 110} Complete thrombosis of all three major hepatic veins at the same time resulting in broad and heterogeneous intrahepatic ischaemia and severe necrosis, in the absence of sufficient draining collaterals leads to BCS ALF. It is associated with a short clinical course and high mortality, needing prompt diagnosis and treatment. In a major population-based study of BCS epidemiology in South Korea, the prevalence was shown to be greater than in Western studies.¹¹¹ However, there is a scarcity of BCS ALF data from Asia. The biggest case series came from the US Acute Liver Failure (USALF) study group, which included 19 instances with BCS presenting as ALF.

Doppler US reveals inferior vena cava compression, no flow and non-visualization of HVs and lack of collaterals. A ratio of aspartate transaminase (AST) to alanine transaminase (ALT) of >1 indicates ALF caused by BCS. Peak ALT and high peak creatinine level are also linked to bad outcomes and suggest that there is more ischaemia.¹¹² In BCS ALF, MELD and conventional King's College Hospital (KCH) criteria liver failure are not accurate to predict prognosis. Anticoagulation should be started as soon as the diagnosis is established, along with a thorough evaluation for malignancy or thrombophilia. When compared to previous published data, mortality in BCS ALF has improved in recent years with the use of TIPS and/or LT.^{10, 113} Despite LT, survival rates are dismal (~30–40%).

5.3 ALF during pregnancy

In Asia, the actual incidence of ALF in pregnancy is unclear. A large number of studies from India have found that ALF is more frequent among pregnant women than in the general population.^{114, 115} This is most likely attributable to a greater rate of HEV infection during pregnancy, as well as certain pregnancy-related disorders including acute fatty liver of pregnancy (AFLP) and haemolysis, elevated liver enzymes, low platelets (HELLP) syndrome, which causes ALF in the third trimester of pregnancy. Furthermore, in low-income nations, pregnant women's susceptibility to viral infections was heightened by physiological stress combined with inadequate nutrition. In India and Pakistan, viral hepatitis is the most prevalent cause of ALF in pregnancy; HEV infection accounts for more than half of all instances of icteric hepatitis in the pregnant population, with a frequency of 60–90% in pregnant women with ALF¹¹⁶ (see above). Conditions that are particular to pregnancy, such as AFLP and HELLP, should be checked out.

Ascites and presence of hypertension provide clue for pregnancy-related liver disorders and can help distinguish it from viral hepatitis.^{117, 118} Previous pregnancy losses, pregnancy in the third trimester, numerous gestations, a male foetus and primigravid status may all be signs of AFLP. Itching, pedal oedema and hypertension are all signs that might indicate HELLP syndrome. The prognosis of ALF in pregnancy is identical to that of the general population, and it is unaffected by the cause of ALF or the pregnant trimester.¹¹⁸

6.1 Definition

There is no unified definition of ACLF. The most appropriate diagnostic criteria for ACLF are still a debate. The Asian Pacific Association for the Study of the Liver (APASL) criteria and the European Association for the Study of the Liver (EASL) criteria are two seminal, consensus-based and generally recognized definitions. APASL defines ACLF as ‘Acute hepatic insult manifesting as jaundice (serum bilirubin >5 mg/dl) and coagulopathy (international normalized ratio [INR] >1.5) complicated within 4 weeks by clinical ascites and/or encephalopathy in a patient with previously diagnosed or undiagnosed chronic liver disease/cirrhosis, that is associated with a high 28-day mortality’.² To give severity ratings to ACLF, the EASL definition needs an initial decompensation (hepatic encephalopathy, gastrointestinal haemorrhage, ascites or bacterial infection) followed by one or more organ failures (OFs).³

Even in Asia, country-specific ACLF definitions are often utilized. ACLF is defined according to the Chinese Society of Hepatology as ‘an acute decompensation in liver function in patients with previously diagnosed or undiagnosed chronic liver disease, manifesting within 4 weeks as severe jaundice (serum total bilirubin >10 mg/ml) and coagulopathy (prothrombin activity >40%), complicated by ascites and/or hepatic encephalopathy’.¹¹⁹ Within the disease entity, Japanese individuals also have liver failure caused by gastrointestinal haemorrhage.¹²⁰

EASL and APASL criteria show considerable discrepancy in ACLF enrolments and events. The majority of individuals with EASL-alone ACLF had normal liver function, indicating that new liver-specific ACLF criteria are needed.¹²¹ APASL ACLF is simple to use and closely follows ALF criteria. It also identifies individuals with significant short-term mortality (>50% at 90 days) and events that do not appear to be related to acute decompensated cirrhosis, as laboratory criteria must be satisfied before an ascites or HE diagnosis can be made.² The APASL definition uses parameters based on INR and total bilirubin to identify liver-specific damage, and there are prognostic gradations (AARC APASL score) that may help patients choose liver-directed therapy.⁵ The AARC criteria apply to ACLF in its early stages, with a golden window for reversing disease development.¹²² The EASL ACLF definition is more difficult to duplicate and necessitates a specific set-up. Because the criteria need triggering factors anterior to OFs, such as hepatic encephalopathy, SBP and ascites, EASL ACLF episodes are not clearly distinguishable from acute decompensated cirrhosis. The face and construct validity of the current EASL ACLF definition are problematic because a considerable number of individuals with EASL ACLF do not have evident liver damage (bilirubin level 5 mg/dl). Indeed, renal failure was the most often recognized OF across all EASL grades, whereas liver failure was one of the least prevalent. This suggests that non-hepatotropic insults could lead to high short-term mortality seen in many EASL patients. More patients with significant short-term mortality who have liver-specific impairment need to be identified, which will improve the potential utility of allocating liver-directed interventions or recognizing medical futility.

6.2 ACLF Epidemiology and recent trends

Epidemiological data on ACLF, particularly in Asia are lacking. Depending on the parameters utilized, the incidence and mortality of ACLF vary substantially. The existing statistics do not account for a wide range of cirrhosis aetiologies and are mostly based on individuals treated in tertiary care centres. It is likely that with longer periods of cirrhosis, there's a time-dependent increase in ACLF incidence, which varies depending on the aetiology of liver disease and treatment.¹²³ In the landmark EASL study, 30.9% of cirrhotic patients admitted to the hospital had ACLF at the time of admission or developed it within 28 days.¹²⁴ Within 28 days, 10.8% of individuals who did not have ACLF at the time of inpatient admission developed it. Piano et al. observed a 14% chance of ACLF in outpatients with cirrhosis within 12 months.¹²⁵

Although the aetiology of ACLF has seen paradigm shift in recent years, with more alcohol insults and fewer chronic HBV infections, the latter remains the predominant cause in many Asian nations.¹²⁶ ACLF is most common in patients with alcohol-related liver damage. Many individuals fulfil the criteria for ACLF solely because they have alcoholic hepatitis, which has a lower short-term mortality rate in the literature than ACLF in general.¹²⁷ ACLF is more common in patients with HCV or NAFLD, but they also had the highest short-term mortality rates.¹²¹ Increased BMI has recently been identified as a risk factor for ACLF,¹²⁸ and it is probable that more obese patients have higher ACLF mortality. Further research is needed to determine the influence of treating hepatitis C, losing weight and quitting drinking on the incidence and mortality of ACLF in the above-mentioned cohorts. Recently Japanese group reported that among the 501 patients enrolled into a nationwide survey,¹²⁹ 183 patients (37%) were diagnosed as having ACLF. The aetiologies of the cirrhosis and acute insults were alcohol intake/abuse in 114 (62%) and 75 (41%) patients respectively.

According to Mezzano et al, the global prevalence of ACLF was estimated to be 35% in a meta-analysis of 30 selected cohort studies, with the lowest frequency in East Asia (15%).¹²³ A recent prospective research in China found that 29.3% of patients had ACLF (as defined by the EASL-CLIF criteria), with transplant-free survival rates of 58.4% and 32.2% at 28 and 90 days respectively.¹³⁰ Overall, ACLF data from East Asia are sparse and usually focuses on a specific aetiology, such as alcohol-related ACLF in Korea and HBV-related ACLF in China.

ACLF is becoming less common in Eastern China, with a better prognosis. In a recent population-based cross-sectional study encompassing 1934 ACLF patients from East China over a 10-year period, ACLF incidence rate has decreased from 3.35 to 2.06 per 100 000 population per year in 2014. Moreover, during the study period, the 28-day mortality rate for ACLF patients fell from 50.39% in 2005 to 35.44% in 2014.¹³¹

The data from 1470 hospitalized patients with chronic liver disease and acute deterioration from 21 university hospitals in South Korea were evaluated, and the prevalence of ACLF according to the AARC criteria was 9.5%. Patients with ACLF had greater 28-day and 90-day death rates than those without ACLF. The total mortality rate for the APASL ACLF group was 40.5% at 28 days and 49.2% at 90 days.¹³²

Pre-ACLF was recently coined as a term to identify individuals who developed ACLF within 28–90 days after being admitted to the hospital for treatment of an acute decompensation episode. In a Chinese study, 120 pre-ACLF patients developed ACLF during their initial hospitalization, and 98 developed ACLF between the first discharge and 3-month follow-up period.¹³³ The main studies on aetiology of ACLF in Asian-Pacific region are shown in Table 2.

6.3 Defining the acute insult

ACLF can result from hepatotropic or non-hepatotropic acute insults. The severity and types of acute insults varies by geographic region and population studied. The following are the major acute insults that cause ACLF to develop:

6.4 Alcohol

Alcohol use has been identified as the most common acute hepatic insult (50%) causative for ACLF in Asia.² Over time, the proportion of people with severe alcoholic hepatitis has increased. According to a recent population-based study conducted in South Korea, alcohol use and high-risk drinking rates had remained high and the prevalence of alcohol-related liver illness has doubled in the last two decades.^{23, 134}

Obesity, diabetes and dyslipidaemia are prevalent metabolic co-morbidities that have a negative influence on the severity and prognosis of alcohol-related ACLF.¹³⁵ We do not know whether certain alcohol-induced ACLF is a distinct kind of ALD or simply a clinical development of severe alcoholic hepatitis. A study of hospitalized patients with alcohol-related cirrhosis with liver biopsy suggested an association between specific histological features (ductular bilirubinostasis and cholangiolitis) and ACLF.¹³⁶

Intense systemic inflammation, immunological dysfunction and a significant risk of infection define the pathophysiology and clinical coarse of these individuals.¹³⁷ Multiple organ failures are a common complication of AH and they are a primary cause of mortality. Compared to other causes of ACLF, patients with active excessive alcohol intake are younger and have more significant biological abnormalities (increased total bilirubin, leucocyte count and international normalized ratio). Alcohol ACLF exhibited a more severe phenotypic presentation, more incidence of organ failures and higher mortality than hepatitis B or hepatitis E-associated ACLF in a large Indian study by Shalimar et al.¹³⁸

6.5 Hepatitis B

Corticosteroids, cancer chemotherapy especially uses rituximab (anti-CD20)-based chemotherapy, and immunotherapy can lead to hepatitis B reactivation.^139-142 Both HBeAg-positive and -negative individuals might experience HBV reactivation. In a study based on severe hepatitis B reactivation, Jindal et al from India found that 70 of 151 patients with HBV reactivation (46.3%) presented with ACLF.¹⁴³ The prognosis is exceedingly bad if the illness has progressed to the point of ACLF, with 3-month death rates of roughly 50% without liver transplantation.¹⁴⁴

During biochemical exacerbation of chronic HBV infection, 23–38% of patients in Far Eastern areas develop jaundice and hepatic decompensation.¹⁴⁵ Patients with pre-existing cirrhosis and more significant hepatic dysfunction (prolonged prothrombin time, raised serum bilirubin and high Child-Pugh's score) had a greater risk of death, according to several other studies from Hong Kong, Taiwan and Japan.^146-148 In a study of HBeAg-positive noncirrhotic patients with acute exacerbation in Taiwan, 5.1% of exacerbation episodes resulted in ACLF, with blood HBV DNA level being the sole significant risk factor. With a sensitivity of 85.7% and specificity of 85.5%, a serum HBV DNA cut-off value of 1.55 × 10 (9) copies/ml indicated ACLF.¹⁴⁹

In chronic hepatitis B, many prediction models have been used to predict ACLF. Patients with cirrhotic HBV–ACLF had substantially greater 90-day transplant-free mortality than those without cirrhosis. Cirrhotic HBV–ACLF patients showed a considerably greater risk of 28-day death (1.02-to 1.94-fold) than non-cirrhotic patients, according to the Chinese Group for the Study of Severe Hepatitis B (COSSH) criteria, confirming the crucial role of cirrhosis in HBV–ACLF.¹⁵⁰

In numerous studies, the MELD score >30 predicted survival in CHB patients with ACLF.¹⁵¹ More recently, a novel predictive score for short-term mortality created by the COSSH group (0.741INR + 0.523HBV-SOFA+0.026age+0.003 TB) was proven to be superior to five previous scores based on both discovery and external validation studies.¹⁵⁰

6.6 Hepatitis E

The first report of HEV superinfection in chronic liver disease leading to severe hepatic decompensation was published in Pakistan in year 2002.¹⁵² Many studies have now recognized HEV as one of the primary cause of cirrhosis decompensation in Asia, where the virus is widespread.¹⁵³ The majority of these studies followed the APASL definition of ACLF, and HEV was the precipitating reason in 21% of cases (range 4–72%) of ACLF. Zhang et al from China reported the largest study demonstrating HEV as the acute insult of ACLF in 2010.¹⁵⁴ In this Chinese study comparing the HEV and HAV superinfections in chronic hepatitis B, liver failure (39.7 vs. 11.5%, p = .002) and death (33.8 vs. 1.9%, p < .001) were more common in the HEV group. HEV was the most prevalent acute insult resulting to ACLF in 61% of the study group, in a study by Dhiman et al from Lucknow, India.¹⁵³ In a recent Indian study, Shalimar et al found that 45 of 368 ACLF patients (12.2%) had HEV superinfection, which was different from alcohol-related ACLF.¹⁵⁵ HEV ACLF patients had the greatest survival rate. In locations where HEV is widespread, patients with chronic liver illness should take necessary measures against superinfection with HAV/HEV, such as drinking heated water and eating well-cooked food, because there is no vaccination against HEV in Asia except in China.

6.7 Drugs and herbs

There is a scarcity of data on DILI in cirrhosis that leads to ACLF. In Asia, drugs are responsible for 1.8–5.7% of ACLF precipitating events. The culprit drug ranged from herbal or traditional remedies in the Chinese cohort to anti-tuberculosis drugs in the Indian cohort. Drugs were identified as a cause of ACLF in 10.5% of the 3132 ACLF patients in the AARC database.¹⁵⁶ Complementary and alternative treatments (71.7%), followed by combination ATDs (27.3%) were the most common insults. In a recent study, drug-induced ACLF patients had higher MELD score (30.2) and a higher 90-day mortality (46.5%) than nondrug-induced ACLF patients (38.8%) (p < .05). At the time of presentation, arterial lactate and total bilirubin were significant predictors of death.¹⁵⁶

6.8 Autoimmune hepatitis flare

In the Asia Pacific region, autoimmune hepatitis flare leading to ACLF is considered less frequent and therefore it is often neglected and the treatment delayed. Many people who were diagnosed with AIH ALF, in fact, have ACLF with the presence of cirrhosis on biopsy. The incidence of AIH flare as a cause of ACLF was found in 2.8% of the ACLF patient based on the AARC database.¹⁵⁷ AIH flare was seen in 2.5% of ACLF patients in an investigation by Gupta et al from India.¹⁵⁸ In roughly 17% of patients, DILI and acute viral hepatitis were observed to trigger an AIH flare. This necessitates a lower diagnostic threshold for transjugular liver biopsy. This is especially relevant because early steroid use or liver transplantation (MELD >27, hepatic encephalopathy grade III/IV) might shorten ICU stay and improve clinical outcomes. AIH-ACLF in AARC ACLF database from Asia suggests that lymphoid aggregates and perivenulitis are less common in AIH-ACLF.¹⁵⁷ However, extensive liver fibrosis, marked ductular reaction and extensive parenchymal collapse, as well as conventional autoimmune characteristics, are notable observations.

6.9 Bacterial infections

As per APASL ACLF consensus, bacterial infection is unlikely to be the principal cause for precipitating liver failure and hence does not constitute a definite hepatotropic acute insult. At present, we have inadequate data to suggest that infection per se results in jaundice and liver failure. Bacterial infection is more likely to precipitate extrahepatic organ failure(s).

6.10 COVID-19 infection and ACLF

Patients with cirrhosis have high rates of hepatic decompensation, ACLF and death from respiratory failure following severe acute respiratory syndrome coronavirus 2 (SARS-CoV2) infection. There is a stepwise increase in morbidity and mortality with increasing severity of cirrhosis as measured by Child–Pugh class.¹⁷⁶ The possible pathogenic mechanisms linking cirrhosis with severe COVID-19 lung disease include increased systemic inflammation, cirrhosis-associated immune dysfunction, coagulopathy and intestinal dysbiosis.¹⁷⁷ In a large multicentric COVID-19 Liver Injury Spectrum Study (APCOLIS study) from Asia, SARS-Cov-2 infection was found to cause significant liver injury in CLD patients, decompensating one-fifth of cirrhosis and worsening the clinical status of the already decompensated.¹⁷⁸ The CLD patients with diabetes and obesity are more vulnerable. Even in CLD patients not infected with SARS-CoV2, increased alcohol consumption, unhealthy eating habits and interruptions to hepatology services during the pandemic might have lead to an upward trend in liver disease incidence and severity.

6.11 Future perspectives and public health concerns

Efforts to eliminate viral hepatitis as a public health threat, together with the rapid increase in per-capita alcohol consumption in countries and the epidemic of obesity, are expected to change the spectrum of ALF and ACLF in the Asia-Pacific region in the near future. Preventive strategies for viral hepatitis-related liver disease include increasing access to clean drinking water and sanitation. All countries have implemented HBV vaccination programmes for neonates, although birth dose coverage is extremely suboptimal in some. Availability of screening tests for blood and tissue, donor recall policies and harm reduction strategies are in their initial stages in most countries. Many governments have put HBV and HCV drugs on their essential medicines lists and the availability of generic versions of these drugs has reduced costs. The increasing burden of alcohol-related liver diseases can be contained through government policies to limit consumption and promote less harmful patterns of alcohol use, which are in place in some countries but need to be enforced more strictly. Steps are needed to control obesity and MAFLD, including policies to promote healthy lifestyles and regulate the food industry. Inadequate infrastructure and insufficient health-care personnel trained in diagnosing and managing liver failure are issues that also need to be addressed in the Asia-Pacific region. There must be a renewed focus on prevention, early detection, timely referral and research into the best means to introduce and improve health interventions to reduce the burden of ALF and ACLF in the Asia-Pacific region. Lastly, we need to learn lessons from the current SARS-COV2 pandemic where interruptions to hepatology services during the pandemic led to an upward trend in incidence and severity of liver failure.