Fulminant Acute Hepatitis in the Third Trimester: A Case Report

Fulminant Acute Hepatitis in the Third Trimester: A Case Report

 

Dr Rahul Nilkanth Patil 1, Dr Madhura Mandlik 1, Dr Kimaya A Mali 2, Dr Jyotsna Dwivedi 2

 

  1. Department of Obstetrics and Gynaecology, James Cook University Hospital, Middlesbrough, United Kingdom.
  2. Department of Obstetrics and Gynaecology, Seth GS Medical College and KEM Hospital, Mumbai, India.

 

*Correspondence to: Dr Rahul Nilkanth Patil, Department of Obstetrics and Gynaecology, James Cook University Hospital, Middlesbrough, United Kingdom.

Copyright

© 2026 Dr Rahul Nilkanth Patil. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Received: 03 November 2026

Published: 19 November 2026

       DOI: https://doi.org/10.5281/zenodo.24368346


Abstract

 

Background: Acute liver failure in the third trimester is usually attributed to acute fatty liver of pregnancy (AFLP) or HELLP syndrome, both requiring prompt delivery. Fulminant hepatitis B acquired during pregnancy is rarely described, and guidance on delivery with a uterine scar and severe coagulopathy is limited.

Case presentation: A 22-year-old woman (gravida 2, para 1, previous caesarean section) whose HBsAg test had been non-reactive on a rapid card test at 10 weeks presented at 35 weeks with fever, jaundice and grade II hepatic encephalopathy. AST was 2,117 U/L, ALT 1,606 U/L, bilirubin 9.4 mg/dL and INR 5.23 (MELD 33). HBsAg and IgM anti-HBc were reactive, HBeAg was non-reactive and HBV DNA was 19,700 IU/mL. Other viral and pregnancy-specific causes were excluded. She received tenofovir, N-acetylcysteine, supportive care and 16 units each of fresh frozen plasma and cryoprecipitate for overt disseminated intravascular coagulation. The fetus died in utero. Labour was induced with a single 0.5 mg dose of intracervical dinoprostone gel, and she delivered vaginally a fresh stillborn male baby of 1.905 kg 8 hours later. There was no uterine rupture or haemorrhage. She survived without liver transplantation and was discharged with improving liver function.

Conclusions: A negative booking HBsAg, particularly on a rapid card test, does not exclude hepatitis B later in pregnancy. IgM anti-HBc should be part of the first-line work-up of third-trimester liver failure. Early tenofovir, indication-driven correction of coagulopathy and a carefully supervised low-dose induction were associated with a good maternal outcome even with a scarred uterus.

Keywords: acute liver failure; hepatitis B virus; pregnancy; third trimester; stillbirth; tenofovir; previous caesarean section; induction of labour.

 

Fulminant Acute Hepatitis in the Third Trimester: A Case Report

Introduction

Acute liver failure (ALF) is rare in pregnancy, but when it occurs two lives are at stake and decisions have to be made quickly. The first question is always the cause. Some conditions, such as acute fatty liver of pregnancy (AFLP) and HELLP syndrome, are produced by the pregnancy itself and settle only after delivery. Others, such as viral hepatitis, simply happen to occur during pregnancy, and delivery does not cure them. In a large North American registry, only about half of the pregnant women with ALF had a pregnancy-specific disease, and outcomes for both mother and baby depended largely on the underlying cause [1]. Current European guidance therefore stresses rapid identification of the cause before delivery is planned [2].

In India the picture is different, because viral hepatitis is common. Hepatitis E and hepatitis B together cause around 60% of acute liver failure [3]. Hepatitis E is the virus most feared in pregnancy, accounting for the majority of fulminant cases in pregnant women [4,5]; a recent report from Delhi again showed how quickly it can lead to maternal liver failure and fetal death [6]. Hepatitis B is less frequent, but still causes about one in six cases of fulminant hepatic failure in pregnancy in neighbouring Pakistan [7]. Encouragingly, a 20-year Indian experience found that pregnant women with ALF do no worse than non-pregnant patients when managed in a specialist centre [8].

Severe hepatitis B in pregnancy appears in the literature mainly as reports of liver transplantation during pregnancy [9,10] and as flares of long-standing infection, which carry a high short-term mortality [10,11]. Fulminant hepatitis B newly acquired during the pregnancy, in a woman who had tested negative at booking, is seldom described. Such a patient raises two practical problems. Her illness can look very much like AFLP, which would prompt an emergency delivery. And if the baby dies, the obstetrician must decide how to deliver a woman whose blood does not clot and whose uterus carries a scar.

We report a young woman who developed fulminant acute hepatitis B at 35 weeks after a negative HBsAg test in early pregnancy. She lost her baby but survived without liver transplantation, after early tenofovir, careful correction of her coagulopathy and a vaginal delivery induced with a single low dose of prostaglandin. The report follows the CARE guideline [12].

Case Presentation

A 22-year-old woman, gravida 2, para 1, with one living child, was referred at 35 weeks’ gestation from a district hospital to our tertiary centre in February 2022 with deranged liver and renal function tests. Her only previous uterine surgery was a lower-segment transverse caesarean section for fetal distress in labour, 36 months before this conception, with no wound infection or uterine extension recorded. She had booked for antenatal care in early pregnancy; at 10 weeks, a rapid card test for HBsAg performed at a private facility was non-reactive, and the pregnancy had been uneventful until this illness. She had no known liver disease, previous jaundice, blood transfusion or alcohol use; she had never received immunosuppressive drugs, corticosteroids or chemotherapy; and she had not taken any prescription, over-the-counter or herbal medicines before the illness.

She had had fever with chills, repeated vomiting and yellow discolouration of the eyes and skin for 3 days. For 1 day she had been increasingly drowsy and confused. At the referring hospital she was found to be HBsAg-reactive, with a total bilirubin of 5.62 mg/dL and a platelet count of 120 ×10?/L, and she was transferred.

On arrival she was drowsy but arousable and disoriented, consistent with West Haven grade II hepatic encephalopathy. Plantar responses were flexor and there were no focal neurological signs. She was deeply icteric. Her blood pressure was 96/60 mmHg, pulse 84/min, oxygen saturation 99% and random blood glucose 89 mg/dL. She had no stigmata of chronic liver disease and no bleeding. The uterus corresponded to 34 weeks, with a healed previous caesarean scar that was not tender, and the fetal heart rate was normal. She was admitted to the intensive care unit and managed jointly by the obstetric, medical (gastroenterology), haematology and critical care teams.

 

Investigations

The investigations showed severe hepatocellular necrosis with coagulopathy, together with a serological profile of acute HBV infection. Days are counted from the day of admission to our centre (Day 1); values obtained at the referring hospital are labelled “Referral”. Aminotransferases peaked on Day 1 and by Day 5 had fallen by 93% (AST, from 2,117 to 152 U/L) and 72% (ALT, from 1,606 to 448 U/L). Bilirubin stayed between 7 and 10 mg/dL. Table 1 shows the trends.

 

Other results:

  • HBV markers: HBsAg reactive (ELISA); IgM anti-HBc reactive; total anti-HBc reactive; HBeAg non-reactive; anti-HBe reactive; HBV DNA 19,700 IU/mL. HBsAg had been non-reactive on a rapid card test at 10 weeks, indicating infection acquired during pregnancy (see Discussion).
  • Other causes excluded: IgM anti-HAV and IgM anti-HEV non-reactive; anti-HCV and HIV non-reactive; SARS-CoV-2 RT-PCR negative; blood culture sterile; malaria test negative.
  • Metabolic and urine: random blood glucose 73–89 mg/dL (no hypoglycaemia); serum uric acid 2.4 mg/dL; alkaline phosphatase 577 U/L (consistent with the third trimester); serum cholesterol 109 mg/dL and triglycerides 79 mg/dL. Serum ammonia was not measured. Urine dipstick showed no proteinuria.
  • Ultrasonography (Days 1–2): liver 11 cm, with no increased echogenicity (“bright liver”) and no ascites; no features of cirrhosis, splenomegaly or portal hypertension; both kidneys normal in size and echotexture. One live intrauterine fetus with biometry of 35 weeks 2 days.
  • Ultrasonography (Day 7): intrauterine fetal death with biometry of 35 weeks 1 day, and oligohydramnios (amniotic fluid index 4–5 cm).
  • Ultrasonography (after delivery): no retained products of conception.

 

 

Prognostic assessment

Using the admission bilirubin (9.4 mg/dL) and INR (5.23) on Day 1, and the first creatinine measured in our laboratory (0.6 mg/dL on Day 2, set to 1.0 as the formula requires), the MELD score was 33 [13]. Because MELD was derived in chronic liver disease, and creatinine is physiologically low in pregnancy, it was used only as an approximate guide. She did not meet the King’s College criteria for liver transplantation in non-paracetamol acute liver failure [14]. Her INR never exceeded 6.5, and she met only one of the five additional criteria (INR >3.5; prothrombin time 54 s). She was 22 years old, the cause was hepatitis B rather than an unfavourable aetiology, jaundice preceded encephalopathy by about 2 days, and peak bilirubin was

9.7 mg/dL (threshold 17.5 mg/dL). Liver transplantation was discussed with the hepatology team on Day 2; because she did not meet the criteria and her coagulation and aminotransferases were improving, she was not listed.

Differential diagnosis

The diagnosis was acute liver failure due to acute hepatitis B acquired during pregnancy. She had coagulopathy (INR >1.5) and encephalopathy within 8 weeks of the onset of jaundice, with no previous liver disease.

  • AFLP was the main alternative, because it requires prompt delivery. Table 3 lists the Swansea criteria [15]. She met five of the 12 criteria that could be assessed (vomiting, encephalopathy, raised bilirubin, raised aminotransferases and coagulopathy), below the threshold of six that suggests AFLP. Abdominal pain, polydipsia or polyuria, hypoglycaemia, raised uric acid, leucocytosis, renal impairment, and ascites or a bright liver on ultrasound were all absent. Serum ammonia was not measured and liver biopsy was not performed. Aminotransferases above 2,000 U/L are also unusual in AFLP.
  • HELLP syndrome and pre-eclampsia were excluded: she remained normotensive (96/60 mmHg on admission), urine dipstick showed no proteinuria, and her platelet count never fell below 100 ×10?/L (lowest 102 ×10?/L), so the Tennessee criteria for HELLP syndrome were not met [16].
  • Other viral hepatitis: hepatitis E, the commonest cause of ALF in pregnant women in India [4,5], and hepatitis A were excluded by negative IgM serology, and hepatitis C and HIV by negative serology. Herpes simplex virus PCR was not performed (see Limitations).
  • Drug- or herb-induced liver injury was excluded by a careful history: she had taken no medicines or herbal remedies.
  • Sepsis and malaria: blood cultures were sterile and the malaria test was negative.
  • Reactivation of chronic or occult hepatitis B was considered unlikely, because HBsAg had been non-reactive at 10 weeks, IgM anti-HBc was reactive, she had no exposure to immunosuppression, and ultrasonography showed no chronic liver disease (see Discussion).

 

Treatment

Treatment had three aims: suppress the virus, support the failing liver and brain, and correct the coagulopathy enough to allow a safe delivery. After consultation with the gastroenterology team, tenofovir disoproxil fumarate (TDF) 300 mg once daily was started. TDF was chosen for its potency, high barrier to resistance and established safety in pregnancy.

She received supportive care in the form of:

  • oxygen by non-rebreather mask, stepped down to room air as she improved
  • intravenous N-acetylcysteine for 5 days
  • lactulose 30 mL, titrated to two or three soft stools a day, and rifaximin 550 mg twice daily
  • 20% human albumin 100 mL daily
  • intravenous pantoprazole 40 mg daily
  • empirical piperacillin-tazobactam 4.5 g and metronidazole 500 mg every 8 hours, for fever with chills
  • ursodeoxycholic acid 300 mg twice daily and B-complex vitamins
  • close glucose monitoring, avoiding sedatives.

 

 

Coagulopathy

She had overt disseminated intravascular coagulation (DIC) without clinical bleeding. On Day 2, the prothrombin time was 43.5 s, APTT 50.1 s, fibrinogen 138 mg/dL and D-dimer 14.3 µg/mL, with raised fibrin degradation products; fibrinogen fell to 82 mg/dL on Day 5. The ISTH overt-DIC score was 5 on Day 2 and 6 on Day 5 [17], although this score is not validated in pregnancy or liver failure. Guidance on acute liver failure advises against giving plasma solely to correct the INR in patients who are not bleeding [18]. Blood products were given here because overt DIC with a falling fibrinogen was present in a viable pregnancy at 35 weeks, when spontaneous labour, placental abruption or an emergency caesarean section could occur at any time.

Under haematology guidance, she received four sets of 4 units of fresh frozen plasma and 4 units of cryoprecipitate (16 units of each in total; Table 1).  

The sets on Days 1, 2 and 5 were each triggered by fibrinogen below 150 mg/dL or a rising INR; the fourth set was given on Day 7, before induction of labour, after the INR had risen again to 2.68 on Day 6. The targets were fibrinogen of at least 150 mg/dL, INR of about 2.0 before induction and platelets above 50

×10?/L. PT/INR, APTT, fibrinogen, D-dimer and platelets were measured daily, and fluid balance and neurological status were charted to detect volume overload. Viscoelastic testing was not available. There was no bleeding, transfusion reaction or pulmonary oedema.

 

Obstetric management

The pregnancy was initially continued, because delivery does not improve liver failure of viral origin and the fetus was alive. Fetal well-being was monitored by daily intermittent auscultation and daily cardiotocography. On Day 7 she reported reduced fetal movements, and ultrasonography the same day confirmed intrauterine fetal death with oligohydramnios (amniotic fluid index 4–5 cm).

The team chose induction of labour rather than repeat caesarean section, because her coagulopathy and encephalopathy made a laparotomy hazardous. The cervix was unfavourable (Bishop score 3). A transcervical Foley balloon was offered but declined by the patient, and misoprostol was avoided because of the uterine scar. After counselling and written consent, which included the risk of uterine rupture and the possible need for laparotomy or hysterectomy, a further four units each of fresh frozen plasma and cryoprecipitate were given (16 units of each in total). Before induction, the INR was 2.03 and fibrinogen 164 mg/dL.

Labour was induced with a single dose of intracervical dinoprostone gel 0.5 mg, the lowest available dose; no further doses and no oxytocin were needed. A senior obstetrician supervised throughout, four units of blood were cross-matched, the anaesthetist was informed and an operating theatre was kept ready. Maternal pulse and blood pressure were recorded every 30 minutes, together with contraction frequency, scar tenderness, vaginal bleeding and haematuria, on a partograph. She delivered vaginally a fresh stillborn male baby weighing 1.905 kg, 8 hours after induction. The liquor was meconium-stained and the placenta appeared normal on gross examination. Active management of the third stage was used and blood loss was 400 mL; there were no clinical signs of scar dehiscence, and no uterine exploration, hysterectomy or red-cell transfusion was needed. The parents declined a fetal post-mortem and placental histology. Lactation was suppressed by mechanical methods (a firm breast binder and avoidance of expression), without drugs.

Outcome and follow-up

She survived without liver transplantation. Her sensorium improved steadily and oxygen was weaned off. After delivery the INR fell from 2.03 to 1.17 within 5 days without further blood products, and fibrinogen rose to 203 mg/dL. Ultrasonography showed no retained products. At discharge on Day 15 she was conscious and oriented; AST and ALT had returned to normal (43/29 U/L), INR was 1.14, albumin was 3.3 g/dL and creatinine 1.0 mg/dL. Bilirubin remained raised at 9.0 mg/dL, consistent with slowly resolving cholestasis after severe acute hepatitis.

She was discharged on tenofovir 300 mg once daily, rifaximin 550 mg twice daily, ursodeoxycholic acid 300 mg twice daily, lactulose, pantoprazole and multivitamins, with follow-up planned in the gastroenterology and obstetric clinics. She did not return for follow-up, and no post-discharge liver function tests or HBV serology are available.

 

Discussion

This case carries four messages. Hepatitis B can be acquired during pregnancy, so a negative booking test does not exclude it later. In the third trimester it can closely imitate AFLP. A modest viral load in an HBeAg-negative patient does not mean mild disease. And when the cause is viral, a scarred uterus and severe coagulopathy need not force a caesarean section.

 

Hepatitis B as a cause of liver failure in late pregnancy

In South Asian series, hepatitis E causes most liver failure in pregnancy, while hepatitis B accounts for roughly 7–17% [4,7]. Pregnancy itself does not seem to worsen survival in specialist centres [8], but pregnant women with hepatitis B-related liver failure still have a high short-term mortality; in one series of 26 such women, 30.8% died within 28 days [11]. Several reported survivors of fulminant hepatitis B in pregnancy needed a liver transplant [9,10]. Our patient had a MELD score of 33 and grade II encephalopathy, yet survived with medical treatment alone. As described above, she did not meet the King’s College criteria for transplantation at any point [14]. Her rapidly falling aminotransferases were an early sign of recovery.

Primary infection or reactivation?

The non-reactive HBsAg at 10 weeks, followed by reactive HBsAg and IgM anti-HBc at 35 weeks, strongly suggests primary infection acquired during pregnancy. Universal antenatal screening is designed to identify chronic carriers so that their babies can be protected. It is a single test, however, and cannot detect infection acquired later in pregnancy. The source of her infection could not be established, and her partner and family were not screened.

We considered whether this could instead be reactivation of unrecognised chronic or occult infection. Five findings favoured primary infection. First, HBsAg changed from non-reactive to reactive within the same pregnancy. Second, IgM anti-HBc was reactive. Third, she had no previous liver disease and had never received immunosuppressive drugs, corticosteroids or chemotherapy, the usual triggers of reactivation. Fourth, ultrasonography showed no features of chronic liver disease, splenomegaly or portal hypertension. Fifth, HBV DNA was modest (19,700 IU/mL) despite aminotransferases above 2,000 U/L, a pattern consistent with immune-mediated clearance in fulminant acute infection, whereas reactivation is usually preceded by a marked rise in HBV DNA. However, the booking test was a rapid card test, which is less sensitive than laboratory immunoassays at low HBsAg concentrations; low-titre IgM anti-HBc can occur in severe flares of chronic infection; and anti-HBc was not tested at booking. Reactivation of occult infection therefore cannot be excluded with complete certainty.

 

The diagnostic trap: the Swansea criteria are not specific

AFLP requires urgent delivery, so mislabelling viral liver failure as AFLP can lead to an emergency caesarean in a woman whose blood does not clot. In the North American registry, the Swansea criteria were met by all women with pregnancy-specific liver disease and by almost all women with other causes of liver failure [1]. Biopsy-based data from India have also questioned how well the criteria predict the microvesicular fat of AFLP [19], and clinical and laboratory features alone could not separate hepatitis E from AFLP in pregnant women [20]. In our patient, who met five of 12 assessable criteria (Table 3), aminotransferases above 2,000 U/L, normal glucose and a low uric acid were the bedside clues, and IgM anti-HBc settled the diagnosis. We suggest that every pregnant woman with liver failure has a full viral panel at presentation (hepatitis A, B with IgM anti-HBc, C and E, and herpes simplex), even when a booking test was negative.

What the serology tells us

She was HBeAg-negative and anti-HBe-positive, with HBV DNA of only 19,700 IU/mL. In a Japanese study of 301 patients with acute hepatitis B, fulminant disease was independently linked to HBeAg negativity, bilirubin of 10 mg/dL or more and the precore G1896A mutation, and precore and core-promoter mutations were far more common in fulminant than in self-limited cases [21]. Fulminant acute hepatitis B is thought to result from a vigorous immune attack on infected liver cells, so by presentation the viral load is often already falling. A low viral load in this setting is therefore not reassuring. We could not sequence the virus, and early HBeAg seroconversion during a severe primary infection is an alternative explanation.

 

Antiviral therapy

No randomised trial has studied fulminant acute hepatitis B in pregnancy. Liver societies recommend a nucleos(t)ide analogue for severe acute hepatitis B or liver failure. A review of severe and fulminant acute hepatitis B reported survival of about 80% with early antiviral treatment and judged placebo-controlled trials unethical [22]. In hepatitis B-related acute-on-chronic liver failure, a 2023 meta-analysis found similar survival at 4, 12 and 48 weeks with tenofovir and entecavir [23], and a network meta-analysis supports nucleos(t)ide analogues in hepatitis B-related liver failure generally [24]. Tenofovir is the preferred agent in pregnancy, where it is widely used to prevent mother-to-child transmission [25], and European guidance supports its use [2]. We cannot prove that tenofovir changed her course, but it was started early, before any delivery decision.

 

The stillbirth

Fetal loss is frequent in fulminant hepatitis during pregnancy; fetal mortality was 50% in one series [20]. In our patient the cause of death could not be established, because the parents declined post-mortem examination and placental histology. The fetus was alive with normal biometry on admission and was monitored daily by auscultation and cardiotocography. It was stillborn without maceration, with meconium-stained liquor and a placenta that was normal on gross examination, and oligohydramnios was present when the death was diagnosed. Maternal fever, consumptive coagulopathy and the systemic effects of liver failure have been proposed as contributors to fetal loss in this setting, but we cannot attribute this death to any one of them.

In hindsight, one could argue for delivering a viable 35-week fetus earlier. Against this, delivery does not improve viral liver failure [2], and a caesarean section at an INR above 3.5 and fibrinogen below 150 mg/dL in a woman with encephalopathy carried a substantial risk to her life. We report this dilemma openly because others will face it.

 

Delivery with a uterine scar and coagulopathy

Once the fetus had died, the mother’s safety was the only priority. The RCOG recommends mifepristone with a prostaglandin for induction after late intrauterine death, but notes that there is not enough evidence for a specific misoprostol regimen beyond 28 weeks in women with a previous caesarean [26]. Prostaglandins increase the risk of scar rupture: PGE2 gel was associated with a uterine rupture rate of 2.9% in women with a previous caesarean, against 0.7% in spontaneous labour and 0.8% with a transcervical Foley catheter [27]. Among women with antepartum fetal death and a previous caesarean, rupture occurred in 2.4% overall and 3.4% of those induced [28]. Our patient had a single previous lower-segment transverse caesarean with an interval of 36 months and no complications, and she declined a Foley catheter. We weighed a 2–3% risk of rupture, which could be managed in a prepared theatre, against the near-certain bleeding of a laparotomy with uncorrected coagulopathy. After targeted replacement of fibrinogen and clotting factors, a single 0.5 mg dose of intracervical dinoprostone gel achieved vaginal delivery within 8 hours, without oxytocin, haemorrhage or scar complications. Keeping prostaglandin exposure to one low dose probably helped. Mifepristone priming is a reasonable addition whe