Pediatric fulminant hepatic failure in endemic areas of hepatitis B infection: 15 years after universal hepatitis B vaccination



To investigate the role of hepatitis B virus (HBV) infection in pediatric fulminant hepatic failure (FHF) after the launch of universal HBV vaccination, the authors analyzed the data from patients with FHF collected from a nationwide collaborative study group. Children aged 1 month to 15 years who were diagnosed with FHF (62 males and 33 females) between 1985–1999 were included. HBV infection (hepatitis B surface antigen [HBsAg] and/or immunoglobulin M hepatitis B core antibody [IgM anti-HBc] seropositive) accounted for 46% (43 of 95 cases) of all the cases of FHF. The average annual incidence of FHF in the time period 1985–1999 was 0.053/100,000 in the group of patients ages 1–15 years and 1.29/100,000 in those patients age < 1 year. Approximately 61% (58 of 95 cases) of all FHF cases were infants. The percentage of HBV infection was found to be higher in infants (57%) compared with children ages 1–15 years (27%) (P = 0.004). The incidence rate ratio of those patients age < 1 year to those ages 1–15 years was 54.2 for HBV-positive FHF and 15.2 for HBV-negative FHF. Maternal HBsAg was found to be positive in 97% of the infants with HBV-positive FHF, and hepatitis B e antigen (HBeAg) was found to be negative in 84% of these infants. Approximately 74% of all HBV-positive FHF patients and 81% of the infantile HBV-positive patients had been vaccinated. In conclusion, within the first 15 years of universal vaccination, HBV was found to rarely cause FHF in children age > 1 year but remained a significant cause of FHF in infants. HBV-positive FHF was prone to develop in infants born to HBeAg-negative, HBsAg-carrier mothers; these infants had not received hepatitis B immunoglobulin according to the vaccination program in place. (HEPATOLOGY 2004;39:58–63.)

Fulminant hepatic failure (FHF) is a clinical disease involving acute and severe impairment of liver function, encephalopathy, and severe coagulopathy.1, 2 The cause of FHF has been reported to vary largely with location. The most frequent causes worldwide are reported to be viral hepatitis A, hepatitis B, and non-A non-B hepatitis, which are reported to account for 60%–70% of all cases.3–5 The prevalence of hepatitis B virus (HBV) infection in previous reports of FHF has reportedly ranged from 10%–65%.6–9 The mortality rate from FHF resulting from HBV has been reported to be as high as 61%–77%.10–12

Taiwan is known to be a highly endemic area of HBV infection; approximately 15%–20% of the population are reported to be chronic HBV carriers.13, 14 HBV infections are believed to cause 80% of the adult-onset and nearly 100% of the childhood cases of hepatocellular carcinoma (HCC).15–18 Before the implementation of hepatitis B vaccination, 65% of childhood fulminant hepatitis was reportedly caused by HBV infection. Of those affected children, patients age < 1 year were reported to account for 65% of the cases.10 Mother-to-infant transmission and blood transfusions were the major routes of infection. Most of the donors and mothers were hepatitis B surface antigen (HBsAg) positive and hepatitis B e antibody (Anti-HBe) positive.10, 19–21

A universal vaccination program was launched in Taiwan in 1984.22, 23 Before the vaccination program, the HBsAg positive rate in children age < 12 years was 9.8%. After the institution of the program, the HBsAg carrier rate in children decreased steadily: 4.8% at 5 years, 1.3% at 10 years, and 0.7% at 15 years.24–27 The annual rate of incidence of childhood HCC decreased from 0.52 to 0.13 per 100,000 children after the vaccination program.28 A trend demonstrating a decrease in the number of cases of acute hepatitis resulting from HBV also was noted.29 Moreover, the mortality rate from hepatitis in infants was reported to decline.30

In light of the high mortality rates found in children with HBV-associated diseases such as FHF and HCC, it is necessary to evaluate the progress of universal vaccination in the control of fulminant hepatitis. In this study, we recruited pediatric FHF patients in Taiwan during the first 15 years of the vaccination program to assess the current impact of HBV infection in these children.


HBV, hepatitis B virus; FHF, fulminant hepatic failure; HBsAg, hepatitis B surface antigen; anti-HBc, hepatitis B core antibody; HBeAg, hepatitis B e antigen; HCC, hepatocellular carcinoma; HBIG, hepatitis B immunoglobulin.

Subjects and Methods

Taiwan has a population of 22 million people, with approximately 330,000 births reported each year. During the period of the current study, there were a total of 9 tertiary referral centers in Taiwan, with a total bed count of 14,800. This study was conducted by the Childhood Fulminant Hepatic Failure Study Group, which was joined by the nine tertiary referral centers covering the northern, central, and southern parts of Taiwan. All hospitals included in the current study were qualified through the availability of the facilities for pediatric intensive care and staff members specializing in pediatric gastroenterology.

Inclusion and Exclusion Criteria for FHF Cases.

Between January 1985 to December 1999, all children ages birth to 15 years who were admitted for FHF were included, regardless of the outcome. This study was conducted retrospectively through review of hospital admission records. FHF was defined as the development of hepatic encephalopathy and coagulopathy within 8 weeks of disease onset in patients without known preexisting liver diseases. Children age < 1 month were not recruited to exclude cases caused by perinatal insults and inherited metabolic defects. Those with acute hepatic failure secondary to bacterial sepsis, drug intoxication, or total parenteral nutrition also were excluded. All patients included had hyperbilirubinemia with a direct bilirubin level > 2 mg/dL and coagulopathy with a prothrombin time that was above twice the normal value.

A thorough search for known hepatitis viruses A to C, cytomegalovirus, and the Epstein–Barr virus was conducted. HBV infection was defined by HBsAg and/or immunoglobulin M hepatitis B core antibody (IgM anti-HBc) seropositivity.

Nationwide Hepatitis B Vaccination Program.

In Taiwan, 16% of pregnant mothers reportedly were HBsAg carriers.23, 31 The Universal HBV Vaccination Program in Taiwan began on July 1, 1984. Hepatitis B immunogobulin (HBIG) was administered to newborns of HBeAg-positive, HBsAg-carrier mothers within 24 hours after birth. The details of the vaccination program were described previously.23, 28, 32 The vaccine coverage rate was as high as 95%.26

Vaccination History and Maternal HBV Status.

A vaccinated case was defined as the completion of 3 doses of the HBV vaccine in infants age > 6 months, or as having received appropriate doses by schedule in infants age < 6 months of age. Maternal HBV status was recorded as HBsAg and HBeAg/Anti-HBe positivity.

Incidence of FHF in Taiwanese Children.

The year-end population statistics for children in the current study were obtained from the annual reports regarding demographic statistics published by the Ministry of the Interior. The incidence rates of FHF recorded between 1985 and 1999 therefore were derived from the numbers of the above-mentioned registered cases and the age-specific population data obtained from year-end population statistics.

Data From the Nationwide Mortality Registry.

To compare nationwide pediatric hepatitis mortality rate before and after the vaccination program, we used the National Mortality Registry System to study the mortality associated with all causes of hepatitis in children age 1–15 years between 1980–1999. All death certificates were reviewed and coded by medical registrars according to the Manual of the International Statistical Classification of Diseases, Injuries, and Causes of Death in the central office of the National Death Certification System at the Department of Health.

Statistical Analysis.

Incidence rates and 95% confidence intervals were calculated using the person-year approach. Actual person-year calculations were derived by multiplying the total number of children alive by the number of years each individual had survived during the prespecified time span. Comparisons of the frequencies of incidence among the pre-specified time periods were made using the chi-square person-year approach.33 All comparisons were two sided, with P values < 0.05 considered to be statistically significant. The Stata package (Stata 7; Stata Corporation, College Station, TX) was used for all statistical analysis. The incidence rate ratio (IRR) was derived using Poisson regression to compare the differences between age cohorts.



Between 1985 and 1999, a total of 95 pediatric patients were diagnosed with FHF (male:female ratio of 62:33). HBV infection accounted for 45% of all cases (43 of 95 cases). Among the 52 HBV-negative cases, Wilson's disease was diagnosed in 6 cases, acute EBV infection was diagnosed in 4 cases, and infection-associated hemophagocytic syndrome was diagnosed in 2 cases. The remaining 40 cases were idiopathic. The case numbers and incidence of HBV-positive FHF and HBV-negative FHF are shown in Table 1. Overall, the average annual incidence of all FHF cases between 1985–1999 was 0.053 per 100,000 in the group of patients ages 1–15 years and 1.29 per 100,000 in the group of patients age < 1 year.

Table 1. Incidence Rates and Rate Ratios between Age Groups of HBV-Positive FHF and HBV-Negative FHF Within the 15 Years of the Universal Vaccination Program
 Year 1985–1999 Case no. (incidence per 100,000)IRR (<1 yr vs 1–15 yr) (95% CI)P Value
  1. NOTE. Total person-years: 0–1 yrs: 4,484,057; 1–15 yrs: 69,096,282.

  2. Abbreviations: HBV, hepatitis B virus; FHF, fulminant hepatic failure; IRR, incidence rate ratio; +, positive; −, negative.

HBV+ FHF43  
 <1 yr33 (0.74)54.2 (26.1–123.2)<.01
 1–15 yr10 (0.014)  
HBV− FHF52  
 <1 yr25 (0.56)15.2 (8.5–27.2)<.01
 1–15 yr27 (0.039)  

Age and Sex Differences.

We analyzed the data regarding patients with FHF using 2 age groups: those patients age < 1 year and those ages 1–15 years. Approximately 61% of the patients with FHF (58 of 95 patients) were age < 1 year, and 57% of these patients (33 of 58 patients) were HBV-positive. The percentage of HBV infections in those patients ages 1–15 years was 27% (10 of 37 patients), which was lower than that reported in those patients age < 1 year (P = 0.004). Of the total number of HBV-positive patients, 77% (33 of 43 patients) were infants. The IRR of those patients age < 1 year to those age 1–15 years was 54.2 (P < 0.001) in the HBV-positive FHF group and 15.2 (P < 0.001) in the HBV-negative FHF group (Table 1). The rates of HBV-positive FHF in the years 1985–1989, 1990–1994, and 1995–1999 were 42%, 70%, and 52%, respectively, in infants (P = 0.24) and 40%, 33%, and 13%, respectively, in children ages 1–15 years (P = 0.21).

We further compared the different incidences of age-stratified HBV-positive and HBV-negative FHF in different sexes (Table 2). The male-to-female incidence ranged from 1.17–2.88. Male preponderance was found to be statistically significant only in HBV-positive FHF cases occurring in patients age < 1 year.

Table 2. Gender Differences in the Incidence of HBV-positive and HBV-negative FHF in Different Age Groups
 Male Case No. (incidence per 100,000)Female Case No. (incidence per 100,000)IRR (male vs. female) (95% CI)P Value
  1. Abbreviations: HBV, hepatitis B virus; FHF, fulminant hepatic failure; IRR, incidence rate ratio; +, positive; −, negative.

HBV+ FHF    
 <1 yr25 (1.07)8 (0.37)2.88 (1.30–6.39)0.009
 1–15 yr6 (0.017)4 (0.012)1.40 (0.39–5.00)0.60
HBV− FHF    
 <1 yr16 (0.69)9 (0.42)1.64 (0.72–3.71)0.24
 1–15 yr15 (0.042)12 (0.036)1.17 (0.55–2.49)0.69

Maternal HBV Status.

In the 43 HBV-positive FHF patients, maternal HBsAg and HBeAg status was available for 35 cases and 26 cases, respectively. HBsAg was positive in 89% of cases (31 of 35 cases) and HBeAg was negative in 81% of cases (21 of 26 cases). Because the cases occurring in patients age < 1 year were most likely to be the result of infection through maternal-infant transmission (thereby making maternal status a relevant factor in this group), we also calculated the incidence rate of infants with mothers who tested positive for HBsAg and HBeAg. Maternal HBsAg was positive in 97% of cases (28 of 29 cases) and HBeAg was negative in 84% of cases (21 of 25 cases).

Vaccination History.

Of 43 HBV-positive FHF patients, 39 had available vaccination records, 29 of whom (74%) were vaccinated. Seven of the 10 unvaccinated children were born between 1985–1986, 1 child was born in 1988, and the remaining 2 children were born in 1990. After 1991, no child was unvaccinated. Of a total of 33 infants age < 1 year with HBV-positive FHF, 31 had a record of vaccination. Twenty-five infants (81%) received 1–3 doses of the HBV vaccine, depending on the age of onset; 3 patients received 1 dose, 20 patients received 2 doses, and 2 patients received 3 doses of vaccine. In this study, a total of 29 infants and children had acquired HBV-positive FHF despite compliance with the vaccination schedules.


The total mortality rate of the patients with FHF was 74%. The mortality rate was lower in the HBV-positive patients compared with the HBV-negative patients (65% vs. 83%; P = 0.05).

Nationwide Mortality Registry of Hepatitis.

The average annual hepatitis mortality rate in infants before and after the vaccination program decreased from 5.36 to 1.71 per 100,000 in the population.30 The average annual mortality rate of children ages 1–15 years between 1980–1984 and 1985–1999 were 0.27 and 0.10, respectively. The IRR from the time period 1985–1999 to 1980–1984 was 0.37 (P < 0.01).


The results of the current study revealed that after implementation of universal vaccination, HBV-related FHF had not been eradicated. Age was found to be a significant factor affecting the incidence of FHF. Although HBV-positive FHF was rare in children age ≥ 1 year, it still persisted as the major cause of FHF in infants age < 1 year. Approximately 57% of all infant cases of FHF could be attributed to HBV infection; this rate was lower than the rate reported before the hepatitis B vaccination program was implemented (82%).10

To our knowledge, the age distribution of pediatric FHF has not been well defined to date. The results of the current study show that the incidence of FHF was much higher in infants (1.29 per 100,000) compared with children ages 1–15 years (0.053 per 100,000). The definition of FHF in infants might be different from that used for the adult population because encephalopathy can be absent or difficult to diagnose in infants, and also because acute liver failure can cause death without overt symptoms of encephalopathy. Therefore, the actual number of infant cases could be higher than reported.

Previous reports had observed an infant preponderance (36%–65% of all pediatric cases) of FHF caused by HBV.10, 34 It is interesting to note that in the current study, the infant preponderance was more exaggerated in HBV-positive cases compared with HBV-negative cases. The IRR of infants age < 1 year to children age 1–15 years was 54.2 in the HBV-positive group compared with 15.2 in the HBV-negative group. It is well recognized that fulminant hepatitis B primarily develops in two groups of patients with different sources of infections. One such group is infants, with the infants acquiring the infection from their mothers, and has a peak age of onset of 3–6 months; the other group are patients of an older age, in which infection occurs via a horizontal route (i.e., blood transfusion).10, 34 Cases caused by horizontal infection also could be prevented by the screening of blood products.

Limitations of the current study included incomplete vaccination records and lack of information concerning maternal HBsAg/HBeAg status. Despite the fact that 4 of the 43 HBV-positive patients did not have an HBV vaccination record, the majority of children acquiring HBV-positive FHF in the postvaccination era had received vaccinations appropriate for their ages. However, many infants (23 of 25 cases) developed FHF before the age of 6 months (i.e., before completion of a full course of vaccination). The development of HBV-positive FHF earlier than the age at which full protection from vaccination can be achieved is possible. The question of whether infants born to HBsAg-positive/HBeAg-negative mothers were at a higher risk of developing FHF deserves further investigation. Before the universal vaccination program, we had observed that FHF caused by HBV occurred particularly in infants born to HBeAg-negative, HBsAg-carrier mothers.10, 19, 20 The present study demonstrated that this trend continued after implementation of the universal vaccination program. This result is biased because the vaccination program only administered HBIG to infants born of HBeAg-positive mothers, and only 25 of 33 HBV-positive infants had maternal HBeAg/anti-HBe data available. It could be possible that vaccination alone was not effective in preventing HBV infection in these children who did not receive HBIG. Conversely, we and other investigators have observed fulminant hepatitis B infection in infants despite appropriate prophylaxis of HBIG and HBV vaccines.35 Previous studies have shown that under the current vaccination program, 1%–3% of the infants born to HBeAg-negative, HBsAg-carrier mothers and 3.3%–7.4% of infants born to HBeAg-positive, HBsAg-carrier mothers became HBsAg carriers.23, 36, 37 The risk of developing FHF in these infants was unknown.

The results of the current study did not provide direct evidence that vaccination has reduced the incidence of HBV-induced FHF in children. Nevertheless, indirect evidence was provided by the Nationwide Mortality Registry Database showing that in infants age < 1 year, the yearly mortality rates from hepatitis had decreased from 5.36 to 1.71 per 100,000 in the years 1975–1984 and 1985–1998.30 We also showed in the current study that in children ages 1–15 years, the annual mortality rate had decreased from 0.27 to 0.10 per 100,000. The decrease in cases also was supported by the percentage of HBV-related cases in infants and children (57% and 27%, respectively, in the current study, compared with 82% and 67%, respectively, in the previous study, before the vaccination program10). Although the mortality database recruited patients with hepatitis resulting from all causes, the source of the reduction was most likely the decrease in the number of HBV-positive FHF cases.

Male predominance has been well documented in patients with HCC.38 It also has been demonstrated that the decrement of childhood HCC after universal vaccination was more significant in males than in females.39 In the current study, male preponderance was found to be significant only in infants with HBV-positive FHF, and not in HBV-negative FHF patients or in the group of patients ages ≥ 1 year. The reason remains to be clarified.

Within the first 15 years of universal vaccination, HBV was rarely was found to be the cause of FHF in children age ≥ 1 year, whereas it remained a major cause of FHF in infants. These infants most likely were infected perinatally from their HBeAg-negative, HBsAg-carrier mothers. Under the vaccination program in place during the duration of the current study, only HBV vaccines, and not HBIG, were given to those infants. The administration of HBIG to all infants born of HBsAg-positive mothers might be considered, and its effect on the prevention of infant FHF deserves further investigation.


The authors thank the statistics office and Shu-Fang Chen, Center for Disease Control, Department of Health, Taiwan, for their kind assistance; Mr. Kevin Kao for editing the article; Dr. Wen-Yi Shau and Dr. Ping-Ing Lee for important advice; and Dr. Mao-Meng Tiao for data collection.