Decompensation occurring in a patient with chronic liver disease (CLD) may be due to either the progression of the underlying liver disease or superimposed acute insults such as acute viral hepatitis (AVH), and intake of hepatotoxic drugs or toxins. The superimposed acute insult causes exacerbation of liver injury and presents with a spectrum of manifestations ranging from deranged liver function tests to decompensation with liver failure and death (1). The decompensation due to the acute insult may be the first presentation of liver disease in patients with silent or unrecognized underlying CLD.
All of the cases of CLD with an acute insult are termed as acute on chronic liver disease (ACLD). Of these, subjects who develop liver failure that is directly attributable to the acute insult are termed as acute on chronic liver failure (ACLF). A consensus definition for ACLF has been given by the Asian Pacific Association for the Study of the Liver (APASL) (1) that defines patients in whom there is rapid decompensation of CLD because of an acute hepatic insult and presents with ascites or encephalopathy within 4 weeks of developing coagulopathy and elevated total bilirubin. This attributes the decompensation to the acute insult rather than to the slow decompensation occurring in the natural course of CLD.
In developing countries endemic for AVH (hepatitis A virus [HAV] and hepatitis E virus [HEV]), ACLD becomes a common diagnostic consideration whenever liver disease patients present with decompensation. The importance of a correct diagnosis in this setting lies in the following: the acute insult needs to be identified and treated promptly; the possibility of underlying CLD needs to be explored in a child presenting with jaundice and decompensation for the first time; and preventive measures such as vaccination may be undertaken in CLD cases, to prevent decompensation related to superadded AVH.
The studies in adults show that HEV is the most common viral insult in ACLD (2–7) and the Model for End-stage Liver Disease (MELD) score predicts the 3-month mortality (2). In patients admitted in the intensive care unit, Acute Physiology and Chronic Health Evaluation (APACHE) score and Sequential Organ Failure Assessment (SOFA) score have also been found to be useful in prognostication (8–10). The literature is sparse with regard to ACLD in children and restricted to only a few case reports (4). The etiology of CLD and the acute insult including infection with hepatotropic viruses in children is likely to be different from that in adults (7). The aim of the present study was to characterize the clinical presentation, etiology, outcome, and determinants of short-term mortality in children with ACLD: both ACLF and non-ACLF.
PATIENTS AND METHODS
Retrospective data were collected from the electronic health records of Sanjay Gandhi Postgraduate Institute of Medical Sciences, Lucknow, India. Children admitted in Pediatric Gastroenterology from January 2000 to January 2010 were evaluated. All of the patients presenting with an acute insult superimposed on a CLD were considered as ACLD cases. ACLF was defined as per the APASL consensus statement as the occurrence of an acute hepatitic insult in a previously diagnosed or undiagnosed CLD causing jaundice (total serum bilirubin >5 mg/dL) and coagulopathy (international normalized ratio [INR] >1.5) along with ascites and/or hepatic encephalopathy (HE) within 4 weeks (1). Children fulfilling the APASL definition were classified as the ACLF group. Children with an acute insult superimposed over a CLD, but not fulfilling the definition of ACLF in terms of the cutoff values for total bilirubin and INR or the presence of either ascites or encephalopathy and the time frame of their onset within 4 weeks, were classified as non-ACLF ACLD cases. Thus, for the study purpose ACLD was divided into ACLF and non-ACLF.
CLD was diagnosed by biopsy evidence of chronic liver disease (chronic hepatitis or cirrhosis) or ultrasonographic evidence of nodular liver with portosystemic collaterals along with esophageal varices (≥grade II) on upper gastrointestinal endoscopy (11). Etiology of underlying CLD was determined as per standard criteria (12–16). The patients with acute insult due to AVH who had anti-smooth muscle antibody positivity were retested after 6 months for persistence of this antibody. If anti-smooth muscle antibody was still positive and other features of autoimmune liver disease were present, the etiology of CLD was taken to be autoimmune liver disease. When no known etiology could be found, the CLD was labeled as cryptogenic. The diagnosis of acute insult was based on history of drug/herbal intake in case of hepatotoxic drugs/herbs or, in case of viral infections, on enzyme-linked immunosorbent assay positivity for HEV (IgM anti-HEV), HAV (IgM anti-HAV), and hepatitis B virus (IgM anti-HBc ± HBsAg), either alone or in combination.
The clinical manifestations at admission and the presence of complications such as HE, spontaneous bacterial peritonitis (SBP), and renal failure were recorded. The West Haven system was used to grade HE as grades I to IV (17). Renal failure was diagnosed in the presence of decreased urine output (<1 mL · kg−1 · h−1) and elevation of serum creatinine above the age-specific normal range. SBP was diagnosed in the presence of a polymorphonuclear cell count of >250 cells/mm3 with or without positive bacterial culture of ascitic fluid. The liver function tests (total serum bilirubin, INR, alanine aminotransferase [ALT], aspartate aminotransferase [AST], and serum albumin) at admission were also noted. Esophageal and gastric varices observed by upper gastrointestinal endoscopy were graded according to the Conn grading system (11). Pediatric end-stage liver disease (PELD) score was calculated from the United Network for Organ Sharing (UNOS) Web site for children <12 years of age (18). Weight was considered unreliable due to presence of ascites. The Revised Wilson Predictive Index score was calculated in all of the patients with underlying Wilson disease (WD) (19).
All of the children of ACLD, whether ACLF or non-ACLF, were treated by a standard protocol as is used for patients with CLD in a sick state. Cefotaxime and cloxacillin were the first-line empirical antibiotics in patients with suspected sepsis that were modified later according to culture sensitivity reports. Piperacillin-tazobactam was used as the empirical antibiotic in patients with suspected sepsis who had prolonged hospital stay outside before admission to our center. In case of secondary deterioration with suspicion of nosocomial infections, repeat blood cultures for bacterial and fungal infections, urine cultures, chest x-rays, and cultures from central venous catheters were done. Until reports of these cultures were available, piperacillin-tazobactam or meropenem along with vancomycin and fluconazole was administered. Confirmed fungal sepsis with normal renal functions was treated with amphotericin B. Enteral feeds by Ryles tube was the preferred mode of nutrition. For HE, initially lactulose was used, and for last 6 months of the study period, rifaximin was used for gut sterilization
SPSS version 17.0 was used for statistical analysis (SPSS, Chicago, IL). Values are expressed as median with range. Mann-Whitney test was used to compare continuous variables. Chi square test and Fisher exact test were used for analyses of categorical variables. Survival rates were compared using Kaplan-Meier method with log rank (Mantel-Cox) comparison of cumulative survival. A P value of <0.05 was considered significant.
Thirty-six children had presented with an acute insult superimposed over CLD (ACLD; 20 boys; median age 9.5; range 3–15 years). Of these 36 cases, 17 children qualified for ACLF (8 boys; median age 10; range 3–15 years) and 19 children as non-ACLF (12 boys; median age 8; range 3–15 years). The etiology of underlying CLD and acute insult in all the cases of ACLD is shown in Table 1. Of the 36 children with ACLD, CLD had been previously diagnosed in 14% children (n-5) and all of the these children could be categorized as non-ACLF. The etiology of CLD in these 5 cases was autoimmune liver disease (AILD) in 2, and biliary atresia, primary sclerosing cholangitis (PSC), and chronic hepatitis B infection in 1 case each. Both children with AILD had cirrhosis on liver biopsy, of which 1 had presented with ascites and encephalopathy along with grade II esophageal varices and the other case had persistent jaundice, hepatosplenomegaly, and grade I esophageal varices. Both patients were receiving treatment for AILD. The child with chronic HBV infection was asymptomatic with histologic features of chronic hepatitis. The child with biliary atresia had a successful Kasai portoenterostomy performed at 75 days of age and, on follow-up, developed HEV infection at 5 years of age. The child with PSC had presented with pruritus and organomegaly. He had biopsy and magnetic resonance cholangiopancreatography evidence of PSC. There was no evidence of cirrhosis at the time of biopsy; however, he had grade II esophageal varices. In the other 86% (n = 31) CLD was diagnosed for the first time during their presentation with a superimposed acute insult (ACLF 17 and non-ACLF 14).
Clinical features at presentation in the ACLD cases (n = 36) were jaundice in 94.4% cases (median duration 60 days; range 3 days–4 years), edema in 69% (median duration 20 days; range 1–60 days), ascites in 69% (median duration 28; range 2–65 days), and HE 36% (median duration 4; range 1–27 days). Ten non-ACLF cases had ascites, but they were not included in ACLF group as the bilirubin in 1 case was 3.9 mg/dL and in the remaining 9 cases the jaundice to ascites interval was >4 weeks. Two of these 10 cases with ascites also had HE (grade I and grade II in 1 case each), which again occurred >4 weeks after onset of jaundice. In the 2 cases in which jaundice was not present, transaminases were elevated and HEV serology was positive. Chronic liver disease was confirmed by liver biopsy in 16 cases (1 postmortem, 1 transjugular) and in the other 20 cases, CLD was diagnosed by the presence of greater than and equal to grade II esophageal varices on endoscopy and ultrasonographic features.
WD and autoimmune liver disease were the most common underlying etiology of CLD in the entire group (Table 1). Comparing the ACLF and non-ACLF groups, there was no significant difference in the frequency of WD (7/17 cases vs 3/19 cases, P = 0.13) or autoimmune liver disease (3/17 cases vs 5/19 cases, P = 0.69). One case of glycogen-storage disease (GSD) type 3 had fibrosis on liver biopsy that has the potential of progressing to cirrhosis and thus was included as CLD (20).
The acute insult was due to hepatotropic viruses in all of the cases except in 1 case in which it was due to herbal drug toxicity. None of our other 35 patients had a history of hepatotoxic drug intake. Among all acute viral insults (n-35), HEV alone or in combination with another virus was the most common superinfection (27/35), constituting 81% (13/16) and 73.6% (14/19) cases in ACLF and non-ACLF groups, respectively. There was no difference in the frequency of HAV, HEV, or acute HBV between the ACLF and non-ACLF groups. None of the 36 children evaluated in the present study had received immunization against hepatitis A or B in the past.
Overall, evidence of bacterial sepsis was present in 25% (9/36, 7 ACLF and 2 non-ACLF) cases. Four cases had SBP alone, 2 had SBP with culture-proven sepsis, 2 had pneumonia, and 1 had only blood culture positivity for Acinetobacter baumanni. Of the 6 cases of SBP, 4 died of liver failure and other 2 survived. Pneumonia in 2 cases improved during hospital admission, but both patients ultimately died due to liver failure. Three cases had blood culture–proven sepsis (Escherichia coli, Pseudomonas aeruginosa, and A baumanni in 1 case each). Of these 3 patients, the first 2 died of liver failure with sepsis and the third improved but was lost to follow-up. Thus, 6 of 9 cases with bacterial sepsis died (5/7 ACLF and 1/ 2 of non-ACLF).
All deaths except 2 were entirely attributable to liver failure. In 2 cases sepsis due to P aeruginosa and E coli was also contributory. In the case with E coli infection, there was coagulopathy with variceal gastrointestinal bleed and hypovolemic shock before death.
A comparison of clinical and laboratory parameters of ACLF (n = 17) and non-ACLF groups (n = 19) is shown in Table 2. ACLF group had significantly higher INR and serum bilirubin, and increased frequency of ascites and HE as per definition criteria (Table 2). Among ACLF group ascites alone was present in 6 cases, HE alone in 2 cases, and both HE and ascites in 9 cases. Among non-ACLF cases, 8 cases had only ascites, whereas 2 cases had both ascites and encephalopathy. The outcome of all ACLD children is shown in Figure 1. Excluding 3 cases who were lost to follow-up after discharge from the hospital, the 3-month mortality of the ACLF group was significantly higher than that of the non-ACLF group (59% vs 11%, P = 0.001). A Kaplan-Meier survival analysis for 90 days, which included the 3 cases lost to follow-up, showed a mean survival in ACLF of 55.3 ± 7.7 (95% confidence interval [CI] 40.3–70.8) days, which was lower than the mean survival in non-ACLF group of 82 ± 5.3 days (95% CI 71.5–92.5 days, P = 0.003) (Fig. 2). Liver transplantation could not be carried out in any of these children.
The comparison between survivors and nonsurvivors among all ACLD children (n = 33) with 3-month follow-up is shown in Table 3. There was no significant difference between the groups with regard to age, sex, serum creatinine, ALT, AST, and albumin. Nonsurvivors had significantly higher INR and total serum bilirubin; increased frequency of ascites, HE, and gastrointestinal bleeding; and lower serum sodium level. There was no difference in the presence or grade of esophageal varices between survivors and nonsurvivors (no varices 6/21 vs 2/11, small varices 12/21 vs 7/11, large varices 3/21 vs 2/11, P = 0.8).
Of the 10 cases of WD, 5 died, 3 survived, and 2 were lost to follow-up. The median score of the Revised Wilson Predictive Index of all patients at diagnosis was 13.5 (6–18). In the 5 patients who died within 3 months, the scores were 13, 13, 14, 16, and 18. In the 3 children who were alive, the scores were 6, 13, and 13. Two patients, both with a score of 16, were lost to follow-up.
PELD score was calculated in 21 of 26 children younger than 12 years, of which 3-month follow-up was available in 19 of 21 cases. In 5 cases, PELD could not be computed due to nonavailability of height. PELD score in the 7 nonsurvivors was higher than that in the 12 survivors (30 vs 11.5, P = 0.003). A PELD score >25.5 could predict death in ACLD with a sensitivity of 100% and specificity of 83.3%. The area under the receiver operating curve was 0.91. All of the patients with a PELD score >33 died (n = 3).
Among the 17 cases of ACLF, 10 children (58.8%) died before 3 months with a median time of death of 28.5 (10–80) days, and the other 5 were alive after a median duration of follow-up of 767 (100–872) days. There were no differences in the age or sex distribution of the survivors and nonsurvivors. None of the individual clinical or laboratory parameters could discriminate the 3-month survivors from the nonsurvivors. The PELD score among ACLF group was calculated in 9 patients (survivors: 4, nonsurvivors: 5) with nonsurvivors having a higher PELD score ([34 [26–52]) compared with the survivors (21 [8–25], P = 0.01).
We have shown that superinfection with hepatotropic viruses A, B, and E occurs in children with chronic liver disease, which is termed as ACLD. A majority (86%) of ACLD manifest clinically with liver dysfunction for the first time. Despite ongoing CLD it appears that these cases manifest only when there is an acute insult with a hepatotropic virus. Two subsets of patients with ACLD were identified: ACLF (47%) and non-ACLF (53%). We did not find any differences in etiology of either superimposed AVH or underlying CLD among the 2 groups.
In our series WD and AILD were the most common causes of CLD, and we had only 1 case of biliary atresia (non-ACLF). The reason for the small number of biliary atresia cases is that the follow-up of our biliary atresia cases is not long enough to assess exposure of these cases to acute insults because of infection with hepatotropic viruses. There were no differences between etiology of acute insult and outcome between the different groups of underlying CLD. Seven of 10 cases of WD and 6 of 8 cases of AILD had HEV superinfection alone and 5 of 10 cases of WD died in comparison to 3 of 8 cases of AILD.
Among hepatotropic viruses, HEV was the most common cause of acute insult in 64% cases followed by HAV in 17%, HAV plus HEV in 11%, and acute HBV in 6%. Overall HEV infection was the cause in 75% of our children that is similar to what is seen in adults with ACLD (67%). HAV is the most common etiology of AVH in children in developing countries (21,22). In our recently published study on AVH we found HAV in 57%, HAV plus HEV combined in 31%, and HEV and acute HBV in 6% each among children younger than 15 years (23). These data bring out a clear predominance of HEV superinfection in ACLD even in the pediatric age group in contrast to the higher frequency of HAV in AVH. Ramachandran et al (4) have shown in adult CLD a higher seroconversion to HEV (56%) compared to age-matched healthy controls and thereby have proposed a higher predilection to HEV infection. It has been shown that the chances of being HEV seropositive (IgG anti-HEV) is significantly greater (34.8%) in patients with CLD as compared with controls (10.5%) (24).
The reasons for the increased frequency of HEV over HAV in ACLD cases are still unclear. The above 2 studies have proposed a predisposition to HEV infection in subjects with CLD (4,24). In addition, the HEV predominance in our study may be caused by a higher likelihood of decompensation with HEV superinfection in CLD than that with HAV, thus manifesting as ACLD.
We have attempted to maintain homogeneity in our study population by restricting it to only acute hepatic insults as described by APASL definition that does not include sepsis as a cause for decompensation. There is evidence to suggest that nonhepatic insults such as sepsis contribute to mortality in ACLF, thereby proposing its inclusion as an acute insult in ACLF definition (25,26). Bacterial infection was present in 25% (9/36, 7 ACLF and 2 non-ACLF) of our cases, of which 6 patients died (5/7 ACLF and 1/2 of non-ACLF). Among patients with bacterial infection (n = 9), there was a predominance of ACLF group in terms of frequency (7/9) and mortality (5/6). Analysis of acute insult contributing to ACLF (n = 17) suggests that 9 were due to viruses alone, 7 due to viruses together with bacterial infection, and 1 due to drugs. Our data suggest that sepsis plays a major role in ACLF and perhaps should be included in the list of acute insults.
We evaluated the Revised Wilson Predictive Index in our WD cases. All nonsurvivors had a score of greater than and equal to 11 but 2 of 3 survivors also had a score of greater than and equal to 11. However, these scores may not correctly reflect prognosis as parameters used for calculation of WD score are also affected by the superimposed AVH over underlying WD present in our ACLD cases versus only WD.
ACLD should be suspected in any child who presents with jaundice, edema, ascites, and/or HE. Alternatively, ACLD presents with raised transaminases of recent onset with previously known CLD. We showed that the presence of ascites, HE, gastrointestinal bleeding, lower serum sodium, higher INR, and serum bilirubin would determine higher chances of mortality (Table 3). We have also objectively shown that a PELD score >25.5 (Fig. 3) in such children could predict death with a sensitivity of 100% and specificity of 83.3%.
Among all cases of ACLD, the PELD scoring system reliably differentiates the children with higher 3-month mortality. We suggest using the PELD value of >25.5 as the cutoff to detect cases that are less likely to respond to medical therapy alone. Even in the ACLF group, which by itself is a high-mortality group, the PELD scores were significantly higher in the nonsurvivors than in the survivors (Fig. 3), thereby offering a tool to identify patients who would require early liver transplantation.
The patients in the present study are drawn from a population with high endemicity of both hepatitis A and E. Therefore, our results cannot be generalized to other populations except other resource-constrained countries with similar endemicity. Even though this pattern and magnitude of viral infection may not be seen in developed countries where hepatitis E infection is presumed to be rare, our study is relevant for developed countries for 2 reasons. First, immigration and increase in global travel may bring patients with ACLF due to HEV to these countries (27). Second, HEV infection is not unknown in developed countries. It is felt that HEV infection being largely subclinical is underdiagnosed in developed nations with much of the seropositivity being wrongly ascribed to lack of specificity of the serological assay (28–30). There is evidence to suggest that HEV-induced hepatitis may also be wrongly labeled as drug-induced liver injury (31).
Our study has the limitation of having a small sample size and representing cases with only “hepatic” acute insult. This restricts the applicability of the results of our study to the selected population. Large, multicentric studies are required for both validation of our results and developing a predictive model to compute chances of survival for individual PELD scores.
We substantiate in our study that ACLF in children has a higher mortality of 59% compared to 11% in non-ACLF group. This reconfirms that the APASL definition of ACLF is useful for identifying high-risk cases among all pediatric cases of ACLD.
The patients who survived the 3-month period following presentation with ACLF went on to have a good long-term survival (n = 5, 767 days), thereby indicating that they have retained good liver function after recovery from the acute insult. This highlights the importance of prevention of superinfection with hepatotropic viruses among patients with CLD and also the need of intensive management during the acute phase in these children.
In conclusion, we have distinctly defined superinfection with hepatotropic viruses on chronic liver disease in children manifesting as ACLD: ACLF and non-ACLF. Hepatitis E virus is the most common superinfection even in the pediatric age group in our population. The APASL definition of ACLF identifies children with poorer outcome among the ACLD group. The PELD score is useful in differentiating survivors and nonsurvivors.
1. Sarin SK, Kumar A, Almeida AJ, et al. Acute-on-chronic liver failure: consensus recommendations of the Asian Pacific Association for the study of the liver (APASL). Hepatol Int
2. Radha Krishna Y, Saraswat VA, Das K, et al. Clinical features and predictors of outcome in acute hepatitis A and hepatitis E virus hepatitis on cirrhosis. Liver Int
3. Hamid SS, Atiq M, Shehzad F, et al. Hepatitis E virus superinfection in patients with chronic liver disease. Hepatology
4. Ramachandran J, Eapen CE, Kang G, et al. Hepatitis E superinfection produces severe decompensation in patients with chronic liver disease. J Gastroenterol Hepatol
5. Kumar M, Sharma BC, Sarin SK. Hepatitis E virus as an etiology of acute exacerbation of previously unrecognized asymptomatic patients with hepatitis B virus-related chronic liver disease. J Gastroenterol Hepatol
6. Acharya SK, Sharma PK, Singh R, et al. Hepatitis E virus (HEV) infection in patients with cirrhosis is associated with rapid decompensation and death. J Hepatol
7. Acharya SK, Batra Y, Bhatkal B, et al. Seroepidemiology of hepatitis A virus infection among school children in Delhi and north Indian patients with chronic liver disease: implications for HAV vaccination. J Gastroenterol Hepatol
8. Wehler M, Kokoska J, Reulbach U, et al. Short term prognosis in critically ill patients with cirrhosis assessed by prognostic scoring systems. Hepatology
9. Zimmerman JE, Wagner DP, Seneff MG, et al. Intensive care unit admissions with cirrhosis: risk-stratifying patient groups and predicting individual survival. Hepatology
10. Cholongitas E, Senzolo M, Patch D, et al. Risk factors, sequential organ failure assessment and model for end-stage liver disease scores for predicting short term mortality in cirrhotic patients admitted to intensive care unit. Aliment Pharmacol Ther
11. Conn HO. Ammonia tolerance in the diagnosis of esophageal varices. A comparison of endoscopic radiologic and biochemical techniques. J Lab Clin Med
12. Lok AS, McMahon BJ. Chronic hepatitis B. Hepatology
13. Ghany MG, Strader DB, Thomas DL, et al. Diagnosis, management and treatment of hepatitis C: an update. Hepatology
14. Vergani GM, Heller S, Jara P, et al. Autoimmune hepatitis. J Pediatr Gastroenterol Nutr
15. Roberts EA, Schilsky ML. Diagnosis and treatment of Wilson disease: an update. Hepatology
16. Valla DC. The diagnosis and management of the Budd-Chiari syndrome: consensus and controversies. Hepatology
17. Ong JP, Aggarwal A, Krieger D, et al. Correlation between ammonia levels and the severity of hepatic encephalopathy. Am J Med
19. Dhawan A, Taylor RM, Cheeseman P, et al. Wilson's disease in children: 37-year experience and revised King's score for liver transplantation. Liver Transpl
20. Demo E, Frush D, Gottfried M, et al. Glycogen storage disease type III-hepatocellular carcinoma a long-term complication? J Hepatol
21. Poddar U, Thapa BR, Prasad A, et al. Changing spectrum of sporadic acute viral hepatitis in Indian children. J Trop Pediatr
22. Yachha SK, Goel A, Khanna V, et al. Ascitic form of sporadic acute viral hepatitis in children: a distinct entity for recognition. J Pediatr Gastroenterol Nutr
23. Kumar A, Yachha SK, Poddar U, et al. Does co-infection with multiple viruses adversely influence the course and outcome of sporadic acute viral hepatitis in children? J Gastroenterol Hepatol
24. Atiq M, Shire NJ, Barrett A, et al. Hepatitis E virus antibodies in patients with chronic liver disease. Emerg Infect Dis
25. Duseja A, Chawla YK, Dhiman RK, et al. Non-hepatic insults are common acute precipitants in patients with acute on chronic liver failure (ACLF). Dig Dis Sci
26. Aithal GP. Defining ‘acute on chronic liver failure’: an identity crisis!. Indian J Gastroenterol
27. Sallie R, Chiyende J, Tan KC. Fulminant hepatic failure resulting from coexistent Wilson's disease and hepatitis E. Gut
28. Dalton HR, Bendall R, Ijaz S, et al. Hepatitis E: an emerging infection in developed countries. Lancet Infect Dis
29. Emerson SU, Purcell RH. Running like water—the omnipresence of hepatitis E. N Engl J Med
30. Purcell RH, Emerson SU. Hidden danger: the raw facts about hepatitis E virus. J Infect Dis
31. Dalton HR, Fellows HJ, Stableforth W, et al. The role of hepatitis E virus testing in drug-induced liver injury. Aliment Pharmacol Ther