Successful Treatment of a South African Pediatric Case of Acute Liver Failure Caused by Perinatal Transmission of Hepatitis B

Abstract

We report on the successful treatment of a South African infant with hepatitis B virus (HBV)–induced acute liver failure using lamivudine with no evidence of clinical resistance. Perinatal HBV transmission occurred despite timely HBV vaccination at 6, 10 and 14 weeks, as per South African vaccination schedule, highlighting the need to introduce the birth-dose HBV vaccine in South Africa.

Since the introduction of universal hepatitis B vaccination in infancy, the prevalence of hepatitis B infection and incidence of hepatitis B virus (HBV)–induced acute liver failure (ALF) has declined.¹ However, in South Africa, where antenatal HBV screening is not practiced and the first dose of the HBV vaccine is given at 6 weeks of age, perinatal transmission of HBV still occurs.^2,3

Infants born to HBV-infected mothers who are negative for hepatitis B envelope antigen (HBeAg negative) and positive for the antibodies to the envelope antigen (anti-HBe positive) are typically at lower risk of perinatal infection.⁴ However, when perinatal transmission occurs, these infants are at higher risk of ALF,¹ especially if they are infected with mutant HBV strains with certain key mutations in the viral basal core promoter (BCP) or precore (PC) regions.⁵ Currently, other than liver transplantation, there are no guidelines for the treatment of HBV-induced ALF in infants.

We report on a case of ALF in a 3-month-old infant, caused by perinatal infection with a BCP/PC HBV mutant strain. This infant was successfully treated with lamivudine.

CLINICAL PRESENTATION

A 3-month-old male infant with recent-onset jaundice was referred to Tygerberg Hospital, a referral center in Cape Town, South Africa. He was irritable and deeply jaundiced. His liver and spleen were not palpable and there was no ascites. Tendon reflexes were increased. Further systemic examination was normal. His international normalized ratio (INR) on admission was 4.9, total and conjugated bilirubin were 322 µmol/L and 218 µmol/L, respectively, alanine aminotransferase 356 U/L, aspartate aminotransferase 219 U/L and albumin 25 g/L. HIV infection was excluded by laboratory testing.

Supportive treatment for ALF was commenced including N-acetyl-cysteine pending serum acetaminophen levels, nutritional support, prophylaxis for intestinal hemorrhage and neuroprotective measures. After receiving vitamin K, the patient’s INR decreased to 2.7; however, he remained irritable.

A search for the cause of liver failure excluded drug-induced liver injury and the most important inherited metabolic diseases.

Further testing for viral infections (Epstein-Barr virus, hepatitis viruses and herpes simplex viruses 1 and 2) was performed. He was positive for hepatitis B surface antigen (HBsAg), a marker of active HBV infection despite receiving vaccinations against HBV at 6 and 10 weeks of age as per the South African Expanded Programme on Immunization schedule. He was also positive for anti-HBe and negative for HBeAg. His viral load, as determined by real-time polymerase chain reaction using the automated COBAS AmpliPrep/COBAS TaqMan HBV Test, was 40,000 IU/mL.

The infant’s mother’s HBV status was unknown as routine HBV screening for pregnant women is not done in South Africa. She was therefore tested to identify the source of the infection and had a similar serologic profile as her child although her viral load was low (268 IU/mL). She was also negative for HIV-1 infection.

The infant was started on lamivudine 4 days after admission at a dose of 4 mg/kg/12 h (day 1). During the week after initiation of treatment, he gradually improved, and on day 7, his INR had decreased to 1.93. On day 10, his liver function improved and the dose of lamivudine was reduced to 4 mg/kg/d. On day 14, his INR had decreased to 1.6, albumin increased to 30 g/L, alanine aminotransferase decreased to 47 U/L and total bilirubin decreased to 119 µmol/L. He was subsequently discharged. Two months after treatment was started, the infant’s liver biochemistry and INR returned to normal and he was systemically well.

In the following months, he continued to thrive and within 6 months had cleared HBsAg and had undetectable levels of HBV DNA and protective levels of anti-HBs. Lamivudine treatment was stopped after 7 months. Sixteen months after initial presentation, there was no evidence of viral relapse and the anti-HBs titer was 41.3 IU/L. Liver biochemistry remained normal.

MOLECULAR PRESENTATION

The surface/polymerase, BCP/PC and core regions of the viral genome were sequenced. The mother–child pair was infected with HBV genotype D, the second most common genotype observed in South Africa.⁶ Both were infected with HBV strains with a single nucleotide mutation in the PC region, from G to A at nucleotide 1896 (G1896A), producing a stop codon (W28*), which would have effectively stopped HBeAg production.⁷ Three other well-characterized mutations in the BCP region (T1753C/A1762T/G1764A) were also found (Fig. 1).^8–10

FIGURE 1.: Organization of the HBV genome (genotype D) and relative positions and significance of the mutations observed in the patient.^{8 , 9} The nt numbering of the genome is based on the unique EcoRI restriction enzyme site shown. The different overlapping open reading frames encoded by the genome, designated as S, core, polymerase and X, are indicated by the arrows. nt numbers designate the boundaries of each ORF with position 1 mapped at the EcoRI site. nt indicates nucleotide. Edited with permission from Hunt et al.¹⁰

DISCUSSION

Origin of ALF

Our patient was infected with HBV genotype D, with several well-characterized mutations in the BCP/PC region that have previously been reported in infants with ALF.^5,11

In the present case, there are 2 potential explanations for the fulminant outcome observed. First, the HBV variants isolated from the patient possessed the T1753C/A1762T/G1764A BCP triple mutation that had previously been described in vitro to increase viral replication.⁸ This could have caused a rapid rise in the viral load despite the presence of anti-HBe.

Second, the HBV variants also had a G1896A mutation in the PC region that would have stopped HBeAg production through the insertion of a stop codon. The absence of the tolerogen HBeAg would have caused the infant’s immune system to recognize the virus and mount a vigorous immune response to clear it.¹²

Hence, the high viral replication observed, combined with the absence of HBeAg, may have contributed to an exaggerated immune response in our patient and his subsequent ALF.

Treatment of ALF Due to HBV

Infants with ALF caused by HBV infection typically are well at birth and only present between 3 and 4 months of age,¹¹ delaying the diagnosis of HBV infection, particularly where there is a record of complete HBV vaccination (as was the case for our patient).

Two previous reports have documented recovery from ALF caused by HBV infection in young infants after treatment with lamivudine at initial doses of 8 and 3 mg/kg.^13,14 We chose to initially treat our patient using the same dose administered to HIV-infected infants (4 mg/kg/12 h) because of the severity of the infection. This dose is well tolerated in HIV-infected infants and we did not observe any evidence of bone marrow suppression (white cell count: 8 × 10⁹/L and hemoglobin: 10 mmol/L). After liver function improved, the dose was reduced to the dose recommended for children with chronic HBV infection (4 mg/kg/d). Both the previously reported infants and our patient made an uneventful recovery after initiation of lamivudine therapy^13,14 but because of the small number of patients reported, it is not possible to confidently assert that the recovery was because of the antiviral treatment. However, lamivudine therapy did not lead to failure of clearance and therefore chronic hepatitis B, which has been previously reported and was therefore a concern clinically.

Although viral resistance is an important limitation of therapy with lamivudine, in the case of our patient, lamivudine resistance was not anticipated because of his low baseline viral load and was felt to be less important than reducing the impact of HBV replication in the short term. Our patient remains well with no clinical evidence of active HBV infection; he remains HBsAg negative 3  years after presentation. In addition, neither of the 2 previous reports of infants administered lamivudine for the management of acute HBV developed clinical resistance to lamivudine and both infants cleared the infection before 9 months.^13,14

Diamanti et al¹⁵ treated a 4-month-old female Italian infant with HBV-induced ALF with tenofovir prepared from crushed tablets for 92 days until viral clearance was documented. The authors argue that tenofovir may be the better choice because of the low risk of viral resistance and superior antiviral activity. However, given the ease of administration, the low risk of short-term side effects and the good response to treatment, we believe lamivudine may be an acceptable therapy, particularly in developing countries where access to close monitoring of renal function, which is required when tenofovir is used, may be more difficult.

Failure of Present South African HBV Vaccination Policy

In South Africa, the first dose of HBV vaccine is administered at 6 weeks with 2 subsequent doses at 10 and 14 weeks. Although this has reduced the prevalence of HBV infection in children,¹⁶ there are growing concerns that the present vaccine schedule is not sufficient to prevent perinatal transmission of HBV.^2,3,17

Furthermore, as pregnant women are not screened for HBV, no immunoprophylaxis (either hepatitis B immune globin or HBV vaccine) is administered to HBV-exposed infants at birth; mothers with high viral loads and therefore at high risk of transmitting HBV to their infants perinatally remain untreated, despite the availability of potent antiviral therapy such as tenofovir, thereby maintaining cycles of transmission and chronic infection.

Despite the lower transmission risk for infants born to anti-HBe positive/HBeAg negative mothers,⁴ for those born to mothers infected with BCP/PC HBV mutants, perinatal transmission can lead to ALF,¹ chronic HBV infection and early-onset hepatocellular carcinoma.¹⁸ The current study emphasizes that there is a need to change the current vaccination policy in South Africa and introduce HBV birth-dose vaccine.

CONCLUSIONS

To our knowledge, this is the first reported case of perinatal transmission of BCP/PC HBV mutant strains leading to ALF in Southern Africa. The patient was treated with lamivudine without progression to chronic hepatitis B. Virus clearance was documented up to 3 years after the acute infection.

Routine HBV vaccination failed to protect this infant from perinatal HBV infection supporting the need to implement an HBV screening program for pregnant women, treat those at high risk of perinatal transmission and to provide early immunoprophylaxis, especially HBV birth-dose vaccine, to HBV-exposed infants.