Lunch 3

Lunch 3 – Neonatal Liver Failure

Pre-Test:

1. What is the definition of neonatal liver failure?

2. What are the 5 categories of diseases that should systemically be ruled out?

3. What should be laboratory tests at presentation?

4. What is the empiric treatment to initiate after safety laboratories sent?

5. Who should make up the initial working up team? Lunch 3

Overview:

1. Neonatal liver failure is a rare but often fatal event.

2. Encephalopathy is very difficult to diagnose thus diagnosis depends on presence of coagulopathy unresponsive to vitamin K associated with other signs of liver dysfunction (jaundice, abnormal aminotransferases, GGT).

3. Usually neonatal liver failure is part of broader systemic disorder.

4. Goals of clinical management include:

a. Establish diagnosis

b. Provide supportive care

c. Decide if liver transplant is appropriate

5. Challenge for physician is to achieve these goals simultaneously in rapidly changing clinical environment.

6. Differential diagnosis for neonatal liver failure.

a. Infectious

i. Herpes simplex 1 and 2

ii. HSV 6

iii. Adenovirus

iv. Hepatitis B

v. Sepsis/urosepsis

vi. Coxsackie virus B/enterovisruses

b. Metabolic

i. Mitochondrial

ii. Beta oxidation defects

iii. Hereditary fructose intolerance

iv. Galactosemia

v. Organic acidemias

vi. Disorders of glycosylation not ALF

c. Toxic Lunch 3

i. Acetaminophen (paracetamol)

ii. others

d. Infiltrative

i. Leukemia

ii. Hemangiomatosis not ALF unless heart failure. Do give consumptive coagulopathy

iii. Severe myelodysplasia

e. Anatomic/perfusion

i. Fetal heart disease

ii. Placental fetal thrombotic vasculopathy

iii. Maternal shock

iv. Arrhythmia due to maternal lupus

f. Other

i. Hemophagocytic lymphohistiocytosis

ii. Neonatal iron storage

7. For many patients liver transplant, with a long term survival of over 60%, is the only therapeutic option, yet in some disorders liver transplant will not be therapeutic. Only 24% infants survived without orthotopic liver transplantation. Lunch 3

Case Scenario 1.

Day 1#

2-day-old male born in 34 week of gestation, presented with hepatosplenomegaly, abdominal distention.

PMHx: Oligohydramnios, thickened left ventricle, placental edema on fetal US at 32 weeks of pregnancy.

FHx: previous child born at 36 weeks of pregnancy, died from “sepsis” at 1 week of life.

PE: critically ill infant with a rectal temperature of 98.2°F, HR - 125 bpm, RR- 25 breaths per minute. BP- 87/43. Capillary refill 2s, Intubated on ventilator. No dysmorphic features. Abdomen distended liver palpable at 6 cm below rib edge, with prominent left lobe, spleen palpable 4 cm below rib edge.

AST-127, ALT 154, LDH 420, albumin 2.1, Bilirubin: total 12.2, conjugated 6.2, unconjugated 5.2 Creatinine 1.0, BUN 35, Glucose 45

PT 45.3, PTT-67.0, D Dimers – 6.5, fibrinogen 56 mg/dl WBC - 5.2 (34% lymphs, 62% segs), Hb-7.6, Platelets- 45

Lunch 3

Case Scenario 2.

Day 1#

1-month-old male presented with fever of 103.2 F. jaundice, hepatosplenomegaly, abdominal distention.

PMHx: normal vaginal delivery

FHx: unremarkable

PE: critically ill infant with a rectal temperature of 103.2°F, HR - 165 bpm, RR- 25 breaths per minute. BP-85/45. Capillary refill 3s, O2 saturation was 93%. No dysmorphic features. Abdomen distended liver palpable at 4 cm below rib edge, spleen palpable 2 cm below rib edge.

AST-3570, ALT 1534, LDH 12180, albumin 3.0, Bilirubin: total 7.2, conjugated 4.2, unconjugated 2.2

PT 24.3, PTT-75.0, D Dimers – 5.5, fibrinogen 50 mg/dl WBC - 4.2 (30% lymphs, 62% atypical lymphs, 5% segs), Hb-9.4, Platelets- 22

Lunch 3

Case Scenario 3.

Day 1#

10-day-old, male infant with a 1-day history of decreased activity, poor feeding and respiratory distress following a normal vaginal delivery.

PMHx: One week before delivery, mother developed fever and received antibiotics. No history of prolonged rupture of membranes or genital herpes.

FHx: unremarkable

PE: critically ill infant with a rectal temperature of 95.2°F, HR - 135 bpm, RR- 35 breaths per minute. BP-not recordable. Capillary refill 6s, O2 saturation was 65%. Central cyanosis, brief respiratory pauses and poor peripheral perfusion. Abdomen not distended liver palpable at 2 cm below rib edge, spleen not palpable.

WBC - 2.1 (2% segs, 9% bands, 89% lymphs), Hb-10.1, platelets-103 PT- 51.1, INR-11.7, PTT>120, fibrinogen 72 AST- 5985, ALT-1636, albumin-1.0 g/dl, Bilirubin, total 2.5, conjugated 1.0, unconjugated 1.2 Lactate-58.6 Blood and urine cultures were negative. A chest radiograph showed bilateral pulmonary infiltrates.

Lunch 3

Case Scenario 4.

Day 1#

3 month old child presenting for cholestasis noticed by the parents

Past history: full term child and normal pregnancy Left temporal stroke on DOL #1 attributed to omphalitis Achieving normal developmental milestones since Had transient FTT between 2-3 month of age, now resolved.

Family history negative – first child to a young couple

PE: Well nourished, active and interactive Mild icterus, no rashes Mild hepatomegaly at 2 cm below the RCM Non-focal neurol exam, but low neck tone for age

Labs: AST 250, ALT 200, C Bili 4, U Bili 1.5, Alc P 350, GGT 80, Alb 2.7, Normal kidney profile, Gluc 80, WBC 25000 (75% neutrophils, 23% Lymp), Plt 180, INR 4, PTT 37, NH3 55 Lunch 3

Summary

Familial hemophagocytic lymphohistiocytosis (HLH) is an autosomal recessive disorder of immune regulation characterized by fever, splenomegaly, cytopenia, hyper-triglyceridemia, hypo-fibrinogenemia,and hyper-ferritinemia, however not all patients may fulfill all the criteria. Additional clinical features, such as lymphadenopathy and central nervous system involvement, were found in about one third of cases, and skin rash only in a quarter. Typically disease occurs during the first 2 years of life, congenital presentation is rare. Clinicians face difficulty of confirming the diagnosis of HLH. However, establishing the diagnosis has important implications for genetic counseling and family planning. HLH should be considered in the setting of perinatal liver failure. Recently several genetic defects have been described in patients with familiar HLH: CD45, FHL2 (PRF1) and FHL3 (MUNC13-4). The effect of mutations on the protein structure correlates well with the severity of the clinical picture, especially with early-onset and defective NK cell function. HLH in neonatal period is fatal without bone marrow transplant. Medical treatment initiated early can stabilize the liver failure and helps to bridge to hematopoietic stem cell transplant the overall (survival is 58%). Liver transplantation alone is not curative and early follow up with a BMT remains a risky proposition.

Mitochondrial depletion syndrome (MDS) refers to a heterogeneous group of mitochondrial disorders characterized by a reduction of the mtDNA copy number in affected tissues. This disorder can present as a hepatocerebral form, a myopathic form, a benign later-onset form, or a cardiomyopathic form. The hepatocerebral form of MDS is usually characterized by onset <6 months of age, early progressive liver failure, muscle hypotonia, hyperreflexia, irritability, and death by 12 months of age from liver failure. Since the synthesis of mtDNA is not cell cycle regulated, a constant supply of deoxyribonucleotide triphosphate (dNTP) is needed for mitochondria to survive and replicate. DNA polymerase gamma (POLG), deoxyguanosine kinase (DGUOK), thymidine kinase 2 (TK2), and most recently MPV17, have emerged as being responsible for mtDNA depletion syndromes(low mtDNA copy number). Liver transplantation is generally contra-indicted.

Neonatal haemochromatosis (NH) is a rare syndrome in which congenital cirrhosis or fulminant hepatitis in early infancy is associated with marked iron deposition in the liver and extrahepatic tissues. NH frequently is preceded by oligohydramnios, placental edema, and intrauterine growth retardation. Stillbirth or premature birth is common. The cause of the condition is ill understood. Neonatal hemochromatosis is not associated with genetic defects of known genes associated with iron metabolism. There is equal sex incidence. There is strong indication that maternal antibodies participate in fetal liver injury. There is 80% incidence of NH in children of mothers with history of previous child with NH. The liver is generally shrunken and bile stained with extensive fibrosis and nodular regeneration; there is massive loss of hepatocytes but surviving cells show giant cell or pseudoglandular transformation with focal nodular regeneration and varying degrees of cholestasis. Little inflammation is usually present and most of the iron deposition is found in the hepatocytes. Extrahepatic iron is seen in the acinar cells of the pancreas and minor salivary glands as well as the proximal renal tubule, adrenal cortex, thyroid, and myocardium. Iron depositis typically spare the reticulo-endothelial system. Presence of extra-hepatic iron (usually small salivary gland) is diagnostic for NH. IVIG prevents mortality from NH in mothers with history of previous child with NH. Liver transplantation can be a life saving surgery in those untreated mothers/infants diads.

Acetaminophen acetylcystine protein adducts were present in 12.5% of samples from children with acute liver failure of indeterminate cause, and 9.6% of samples from children with acute Lunch 3 liver failure that was attributed to other causes. Clinicians should maintain high level of suspicion even when there is no clear history of acetaminophen ingestion.

Neonatal herpes is a rare disorder affecting newborn infants, but it is one of the most severe infections acquired during the perinatal period. Treatment is available therefore, neonatal HSV infection should be considered in the differential diagnosis of each severely ill neonate. Early empirical treatment with acyclovir is justified if there is clinical suspicion of herpes infection. Multiple cases are reported of surviving neonates as long as treatment is initiated early. Liver transplantation is contraindicated in the acute viremic phase. Journal of Pediatric Gastroenterology and Nutrition 38:542–544 © May 2004 Lippincott Williams & Wilkins, Philadelphia

Case Report Neonatal Ischemic Liver Failure: Potential Role of the Ductus Venosus

*Jean Bergounioux, †Ste´phanie Franchi-Abella, §Sophie Monneret, ‡Sandrine Essouri, and *Emmanuel Jacquemin

*Hepatology Unit, †Radiology Unit, ‡Intensive Care Unit, Department of Pediatrics, Assistance Publique–Hoˆpitaux de Paris, Bicêtre University Hospital, Le Kremlin Bicêtre, and §Pediatric Intensive Care Unit, Dijon University Hospital, Dijon, France

Transient acute liver dysfunction in the neonate sec- by both hepatomegaly and ascites. The spleen was not ondary to hemodynamic failure, or shock liver syndrome palpable. Chest radiography showed cardiomegaly and (SLS), is a well-recognized complication of reduced he- no parenchyma abnormality. Echocardiography showed patic blood flow in low cardiac output states (1,2). The normal heart structures but a dilated left ventricle with peculiar pattern of liver blood flow in neonates, charac- decreased shortening fraction (21%). There was no sig- terized by a large intrahepatic shunt through the ductus nificant abnormality of heart rhythm or repolarization on venosus and low arterial blood supply, produces an electrocardiogram. Blood hemoglobin was 10 g/dL, the asymmetric response of the right and left lobes to isch- platelet count was 60,000 /mm3, and blood glucose was emia (3). This phenomenon has been described in an 0.3 g/L. Results of serum liver tests are summarized in extreme form in deceased neonates and animal models in Table 1. Liver Doppler ultrasonography was asymmetric, which necrosis of the right half of the liver develops in showing hyperechogenicity of the right part of the liver response to shock (4–8). We report a neonate with SLS up to the median hepatic vein and clear evidence for secondary to acute cardiac failure. Imaging obtained at patency of the ductus venosus (Fig. 1). Magnetic reso- the acute phase of SLS showed, for the first time in vivo, nance imaging scan of the abdomen performed on day 1 signs compatible with right liver ischemic injury, sup- of admission to confirm ultrasonography findings (Fig. porting previous experimental data in lambs on liver vas- 2) showed an obvious asymmetry of the signal in the cularization during perinatal distress. right and left lobes. The right lobe signal was increased compared to the left. Infectious, metabolic, toxic, and CASE REPORT malignant causes of neonatal liver failure were excluded, as previously reported (9). A defect of mitochondrial This neonate was born after 39 weeks gestation. The fatty acid oxidation was excluded by measuring the ac- pregnancy was uneventful until a few days before deliv- ylcarnitine profile in plasma and organic acid profile in ery when the mother reported decreased fetal activity. urine. There was no evidence for traumatic liver injury. Emergency cesarean section was performed because fe- Dopamine, dobutamine, volume expansion, transfusion tal distress, with fetal tachycardia (greater than 200 bpm) of packed red blood cells, platelets and fresh frozen noted on ultrasonographic study. At birth, the neonate plasma, antibiotics, and intravenous 10% dextrose re- was apneic and hypotonic. APGAR score was 3 at 1 sulted in a rapid improvement of the hemodynamic con- minute and 6 at 5 minutes. Pulse oxymetry showed oxy- dition (10). Urine output was absent for the first 12 hours gen saturations below 80%. The neonate was immedi- and returned to normal with hemodynamic supportive ately intubated and ventilated under 25% FiO2. Heart therapy. Seizures of the lips and upper limbs with bilat- rate and pulse oxymetry gradually improved and blood eral spikes on electroencephalogram required phenytoin pressure was 58/37 mmHg. The abdomen was enlarged and phenobarbital for 10 days. The child was extubated on day 8, and inotropic support therapy was discontinued on day 10. Cholestatic jaundice was noted from day 8 as Address correspondence and reprint requests to Professeur liver function progressively returned to normal (Table 1). Emmanuel Jacquemin, De´partement de Pe´diatrie, He´patologie Pe´di- atrique, Centre Hospitalier Universitaire de Biceˆtre, 78, rue du Ge´ne´ral Subsequent Doppler ultrasonographies showed normal- Leclerc, 94275 Le Kremlin Biceˆtre Cedex, France (e-mail: emmanuel. ization of shortening fraction and progressive closure of [email protected]). the ductus venosus with disappearance of hyperecho-

542 NEONATAL ISCHEMIC LIVER FAILURE 543

TABLE 1. Evolution of serum liver tests in a neonate with ischemic liver failure

Day 1 Day 8 Day 12 Day 24 Month 2 Month 5 Month 12 PT (N > 70%) 19 43 72 100 100 100 89 Factor V (N > 80%) 21 75 100 100 ND ND ND Factor II (N > 70%) 9 39 42 72 ND ND ND Fibrinogen (N: 2–4 gr/l) 0.3 1.3 2.2 2.8 ND ND ND AST (N < 55 U/l) 1,035 32 55 97 50 40 39 ALT (N < 40 U/l) 461 47 24 70 54 24 23 GGTN<25U/l) 511 160 261 263 198 13 9 Total/conjugated bilirubin (N < 17/10 ␮mol/l) ND 132/61 131/92 111/80 11/5 4/0 5/0 Bile acids (N < 15 ␮mol/l) ND 133 ND ND 43 11 ND Ursodeoxycholic acid therapy Onset End

ס gamma-glutamyltranspeptidase; ND ס alanine aminotransferase; GGT ס aspartate aminotransferase; ALT ס prothrombin time; AST ס PT not determined. genicity of the right liver. The child was discharged from the hospital at age 24 days. The persistent cholestatic jaundice was treated with ursodeoxycholic acid from 12 days to 5 months of age (11). At age 12 months all clinical, biochemical, and liver ultrasound examinations were normal.

DISCUSSION

The case highlights for the first time in vivo that dur- ing the perinatal period the right half of the liver is sus- ceptible to ischemia during cardiocirculatory stress (3,4). This case also confirms the favorable natural history of SLS and shows that it may be complicated by transient neonatal cholestasis (11). This case presents the typical clinical phases previ- ously described as the natural history of SLS in children (1,2). Liver dysfunction associated with a sudden, tran- sient elevation in serum transaminase levels occurred early in the course of circulatory failure and improved

FIG. 2. (A) Liver magnetic resonance imaging coronal imaging (T1 weighted) performed on day 1 of life in a neonate with liver FIG. 1. Representative liver Doppler ultrasonography performed failure. It shows an asymmetric signal with an intense signal of during the first week of life in a neonate with liver failure. On a the right liver (star). (B) Liver magnetic resonance imaging frontal longitudinal scan through the liver at the level of umbilical vein, imaging (T2 weighted) performed on day 1 of life in a neonate there is evidence for patent ductus venosus (Doppler recording) with liver failure. It shows an asymmetric signal with an increased joining the left portal branch and the inferior vena cava. DV, duc- signal of the right liver (star). The left ventricle (arrow) appears tus venosus; IVC, inferior vena cava; LPB, left portal branch. distended with an enlarged atrium.

J Pediatr Gastroenterol Nutr, Vol. 38, No. 5, May 2004 544 BERGOUNIOUX ET AL. rapidly as the hemodynamic condition was corrected by ductus venosus secondary to cardiocirculatory failure supportive therapy (1,2,10). The acute phase was fol- (15). Although we have no direct proof for such a mecha- lowed by persistent cholestatic jaundice that lasted sev- nism, this hypothesis is reinforced by the parallel evolu- eral weeks (11–13). The origin of this patient’s cardiac tion between liver function test improvement and ductus failure is unclear. Congenital and metabolic cardiopa- venosus closure. In the clinical setting of neonatal liver thies and liver diseases were excluded. The favorable failure (9) and in the presence of perinatal distress, we outcome argues against a metabolic origin. Although believe that such a condition associated with liver imag- several postnatal recordings of cardiac rhythm showed ing showing asymmetric liver is strongly indicative of no abnormality, fetal tachycardia or transient and acute ischemic liver injury. rhythm abnormality appear the most likely origins of the Acknowledgments: The authors thank Dr. Dalila Habes initial cardiac dysfunction. (He´patologie Pe´diatrique, Hoˆpital de Bicêtre, Paris, France) for During fetal life, the liver receives blood flow from taking care of the patient and Professor Olivier Bernard (He´pa- different sources: the hepatic artery, the portal vein, and tologie Pe´diatrique, Hoˆpital de Bicêtre, Paris, France) for criti- the umbilical vein. Hepatic arterial blood flow has been cally reading the manuscript. shown to be quite low, representing only about 2% of total liver blood flow (3). The portal branch of the left REFERENCES hepatic lobe directly arises from the umbilical vein, then 1. Jacquemin E, Saliba E, Blond MH, et al. Liver dysfunction and the ductus venosus arises, and the umbilical vein arches acute cardiocirculatory failure in children. Eur J Pediatr 1992;151: 731–4. to the right lobe, where it is joined by the portal vein. 2. Garland JS, Werlin SL, Rice TB. Ischemic hepatitis in children: Using radioactive microspheres in fetal lambs, it has diagnosis and clinical course. Crit Care Med 1988;16:1209–12. been shown that about 50% of umbilical venous blood 3. Rudolph AM. Hepatic and ductus-venosus blood flows during fetal flow passes through the ductus venosus, bypassing the life. Hepatology 1983;3:254–8. hepatic parenchyma (7). This proportion increases to 4. Gruenwald P. Degenerative changes in the right half of the liver resulting from intrauterine hypoxia. Am J Clin Pathol 1949;19: 65% during fetal hypoxemia (7,8), with the ductus ve- 801–13. nosus blood flow preferentially distributed to the ascend- 5. Itskovitz J, Goetzman B, Rudolph AM. Effects of hemorrhage on ing aorta and contributing to higher oxygenation of heart umbilical venous return and oxygen delivery in fetal lambs. Am J and brain (14). After birth, the ductus venosus gradually Physiol 1982;242:H543–8. 6. Edelstone DI, Rudolph AM, Heymann MA. Effects of hypoxemia closes. In the event of perinatal hypoxemia and circula- and decreasing umbilical flow on liver and ductus venosus blood tory failure, the ductus venosus may remain patent, lead- flows in fetal lambs. Am J Physiol 1980;238:H656–63. ing to a significant reduction in blood flow to the right 7. Rudolph AM, Heymann MA. Circulatory changes during growth liver, compared with the left liver (3,5,6,8). Both liver in the fetal lamb. Circ Res 1970;26:289–99. lobes have similar oxygen consumption. Since the left 8. Reuss ML, Rudolph AM. Distribution and recirculation of umbili- cal and systemic blood flow in fetal lambs during hypoxia. J De- lobe is supplied mainly by umbilical venous blood with velop Physiol 1980;2:71–84. an oxygen saturation of 85%, and the right lobe mainly 9. Durand P, Debray D, Mandel R, et al. Acute liver failure in in- by portal venous blood with an oxygen saturation of fancy: a 14-year experience of a pediatric liver transplantation about 70%, the right lobe is already functioning under center. J Pediatr 2001;139:871–6. 10. Kram HB, Evans T, Bundage B, et al. Use of dobutamine for low oxygen supply conditions (3,8). These consider- treatment of shock liver syndrome. Crit Care Med 1988;16:644–5. ations have raised the possibility that the right lobe might 11. Jacquemin E, Lykavieris P, Chaoui N, et al. Transient neonatal be more susceptible to hypoxia and ischemia in the fetus cholestasis: origin and outcome. J Pediatr 1998;133:563–7. subjected to asphyxia or cardiovascular stress (3). This 12. Shneider B, Maller E, Van Marter L, et al. Cholestasis in infants was considered a possible explanation for the greater supported with extracorporeal membrane oxygenation. J Pediatr 1989;115:462–5. tendency for necrosis of the right liver lobe found at 13. Vajro P, Amelio A, Stagni A, et al. Cholestasis in newborn infants autopsy in neonates dying from perinatal asphyxia (4). In with perinatal asphyxia. Acta Paediatr 1997;86:895–8. the current report, liver imaging performed during the 14. Edelstone DI, Rudolph AM. Preferential streaming of ductus ve- acute phase of liver failure showed asymmetric signals nosus blood to the brain and heart in fetal lambs. Am J Physiol 1979;237:H724–9. from the right and left parts of the liver. This finding is 15. Araki T, Kamada M, Okamoto Y, et al. Coil embolization of a completely consistent with ischemia and necrosis of the patent ductus venosus in a 52-day-old girl with congenital heart right liver produced by intrahepatic shunting through the disease. Ann Thorac Surg 2003;75:273–5.

J Pediatr Gastroenterol Nutr, Vol. 38, No. 5, May 2004 Seminars in Neonatology (2003) 8, 393–401 Seminars in NEONATOLOGY

www.elsevierhealth.com/journals/siny Review article Neonatal liver failure

Patricia McClean*, Suzanne M. Davison

Children's Liver and GI Unit, St James's, University Hospital, Beckett Street, Leeds LS9 7TF, UK

KEYWORDS Summary Liver failure in the neonatal period is challenging to diagnose and manage, Neonatal liver failure; and still carries a high mortality. With ongoing developments in the field of metabolic Viral hepatitis; disorders and antiviral therapy, and the ability to offer liver transplantation to small Neonatal babies, an overall survival of 40% has been achieved. Early recognition of liver failure, haemochromatosis; good supportive care and prompt referral to a paediatric liver transplant centre are Mitochondrial essential elements in improving the outcome for these babies. Decisions about hepatopathy contra-indications to and timing of transplantation are complex as many of the disease processes are still evolving in the neonatal period, and extrahepatic disease, which cannot be corrected by a transplant, may appear later. © 2003 Elsevier Ltd. All rights reserved.

Neonatal liver failure (NLF) is rare and may be Table 1 Differential diagnosis of a prolonged prothrombin difficult to recognize initially, as jaundice can be a time in the neonate late feature. Coagulopathy, unresponsive to intra- venous vitamin K, is always present, although this is Congenital bleeding disorder not uncommon in ill neonates (Table 1). Some Afibrinogenaemia conditions, such as neonatal haemochromatosis Dysfibrinogenaemia (NH), are due to chronic in-utero liver disease and Deficiencies of individual clotting factors present as decompensated cirrhosis with low Acquired bleeding disorder levels of albumin and normal transaminases, whilst Disseminated intravascular coagulation Vitamin K deficiency others have the typical features of acute liver fail- Antibodies/inhibitors affecting coagulation 1 ure of perinatal onset with high transaminases. NLF Hypoglycaemia, hyperammonaemia and encepha- Drug induced lopathy (often difficult to define in a neonate) are For normal laboratory ranges of clotting parameters in the common, although these are seen in many sick newborn see the work by Williams et al.53 infants. The diagnosis of liver failure must be con- sidered in any neonate with coagulopathy.2 If other features of liver dysfunction are absent or non- specific, measurement of the individual clotting factors will show low levels of II, V, VII, IX and X, Aetiology and normal or elevated levels of factor VIII in Table 2 lists most of the recognized causes of NLF. infants where the coagulopathy is due to liver Published series from paediatric liver centres will disease. Fibrinogen and factors XI and XII are not include cases that resolve spontaneously (e.g. frequently normal but may be decreased. hypoxic/ischaemic), respond quickly to treatment (e.g. bacterial infection), or die early.2,3 Infection * Corresponding author. Tel.: +44-113-2066689; fax: +44-113- 2088891 and metabolic disorders are the two main causes of E-mail address: [email protected] (P. NLF. Many of these diseases may also present as a McClean). less fulminant neonatal hepatitis syndrome (see

1084-2756/03/$ - see front matter © 2003 Elsevier Ltd. All rights reserved. doi:10.1016/S1084-2756(03)00095-2 394 P. McClean, S.M. Davison

Table 2 Aetiology of NLF characteristic pathological feature of infants with acute viral infections, toxic or ischaemic injury and Durand et al.2 Aw et al.3 some metabolic diseases. This results in initially (<1 year) (<4 weeks) high serum transaminases, but as the degree of n=80 n=33 necrosis progresses, the transaminase levels Infection fall, the liver shrinks in size and coagulopathy and Hepatitis B 6 0 hyperbilirubinaemia worsens. Recovery is heralded HSV 1 and 2 2 5 HHV6 4 0 by a sustained fall in prothrombin time in associ- Enterovirus 0 0 ation with falling transaminases and bilirubin Bacterial 0 1 levels. Metabolic In contrast, in some metabolic conditions, such Tyrosinaemia 1 12 1 as fatty acid oxidation defects and mitochondrial Mitochondrial 17 0 respiratory chain defects, there is little cell necro- Urea cycle 2 1 sis and the liver failure is at a subcellular level. Galactosaemia 2 3 Fatty acid oxidation 0 0 Histologically, there may be diffuse hepatic steato- HFI 1 0 sis and/or swelling of hepatocytes, and clinically IE bile salts 0 0 there is hepatomegaly with moderate elevation of CDGS 0 0 transaminases and minimal to moderate jaundice. NH 13 16 Often, with prompt and appropriate medical treat- Infiltrative/storage ment, the liver recovers completely, but if the HLH 3 4 liver disease is chronic and has progressed to Leukaemia 1 0 cirrhosis, as occurs in many mitochondrial Tumours 0 0 Niemann–Pick C 0 0 disorders, recovery will not occur. Infiltrative and storage disorders result in Other hepatosplenomegaly with raised transaminases and Drugs 1 1 AIH 3 0 bilirubin. Finally, NH is a typical example of decom- Hypocortisolism 0 1 pensated cirrhosis, presenting at birth with a small Hypoxic/ischaemic 0 0 liver, normal transaminases, mildly raised bilirubin, Unknown 13 0 markedly reduced serum albumin and, possibly, Most described causes of NLF are listed in the first column. features of portal hypertension. Columns 2 and 3 give the number of cases in the series by Durand et al., who reported all infants under 1 year old with liver failure and Aw et al. who documented neonatal Investigations cases only. HSV, herpes simplex virus; HHV6, human herpes virus 6; HFI, hereditary fructose intolerance; IE bile salts, The initial investigations necessary to establish inborn error of bile salt synthesis; CDGS, congenital the cause of NLF are shown in Table 3. It is import- disorder of glycosylation syndrome; HLH, haemophagocytic ant to recognize medically treatable causes early, lymphohistiocytosis; AIH, autoimmune hepatitis. such as galactosaemia or tyrosinaemia, before pro- ceeding to more invasive procedures. Specific in- vestigations are discussed under the relevant disease headings. Performing a liver biopsy is paper by Roberts). This paper will only describe hazardous due to coagulopathy. The potential diseases presenting primarily as NLF. benefits must be balanced against this risk. Some units advocate an open surgical biopsy in these Clinicopathological features circumstances following vigorous correction of coagulopathy. Presentation at birth implies an intra-uterine insult such as congenital infection, NH or mitochondrial Management disorders. A later presentation may be related to infection or a metabolic condition unveiled by There are no randomized trials of the following the introduction of feeding. A detailed obstetric interventions in neonates. Recommendations are history, including information on consanguinity, extrapolated from experience in paediatric prac- previous miscarriages and neonatal deaths, is tice. Intravenous dextrose to maintain normogly- important. caemia may require concentrations of 20–50% via There are various patterns of presentation which central venous access. Hyponatraemia usually re- are not mutually exclusive and may progress to flects hyperaldosteronism and/or fluid retention: fibrosis or cirrhosis. Hepatocyte necrosis is the fluids should be restricted and excessive sodium Neonatal liver failure 395

Table 3 General investigation of NLF 5 ml/kg of 20% human albumin solution), and di- uretic therapy [oral spironolactone or intravenous Blood potassium canrenoate (1 gϭ0.7 g spironolactone)]. Haematology If causing respiratory compromise, percutaneous Full blood count and film, Blood group and Coombs test drainage may be required. Ventilatory support Prothrombin time, partial thromboplastin time, fibrinogen should be considered early for neurological as well D dimers/fibrin degradation products as respiratory deterioration. Inotropic support Biochemistry may be required, and as peripheral vasodilation Urea and electrolytes, creatine kinase, amylase may accompany liver failure, vasoconstrictors such Bilirubin (unconjugated/conjugated), transaminases, GT, alkaline phosphatase, albumin as noradrenaline should be considered. Acidosis, Acid–base balance reflecting hepatic and/or renal dysfunction, sepsis Glucose, lactate, ammonia or a metabolic disorder may require bicarbonate Cholesterol, triglycerides, free fatty acids, hydroxy correction. butyrate Ferritin and transferrin saturation Hepatic encephalopathy manifests as irritabil- Plasma amino acids ity, poor sucking or excessive somnolence. Manage- Galactose-1-phosphate uridyl transferase ment includes restriction of protein intake to 2 g/ Carnitine/acyl carnitines kg/24 h, and enteral lactulose. Cerebral oedema Cortisol (9am) Serum bile salts may accompany encephalopathy; fluid overload Transferrin iso-electric focusing should be strictly avoided. Convulsions or deterio- Alpha fetoprotein rating conscious level should prompt imaging to Toxicology including paracetamol exclude intracerebral haemorrhage. Microbiology Feeds should be withheld until galactosaemia, Bacterial/viral culture and PCR detection tyrosinaemia and urea cycle defects have been Viral serology: mother and infant (IgM in infant may be excluded. Withdrawal of lactose or fructose from negative for several weeks) infants with galactosaemia or hereditary fructose Other intolerance, respectively, leads to a dramatic im- Storage for DNA provement in clinical symptoms. Enteral feeding Urine should be recommenced as soon as possible, al- Biochemistry though this may be compromised by fluid restric- Amino acids, organic acids including succinyl acetone, tion or poor tolerance, necessitating temporary orotic acid pH, ketones, reducing substances parenteral nutrition. Toxicology Thrombocytopenia, coagulopathy and dissemi- Urinary bile salts nated intravascular coagulation contribute to the Microbiology risk of bleeding. Ranitidine should be prescribed Bacterial/viral culture and PCR detection for gastric protection. Intravenous vitamin K (300 µg/kg/24 h) is used to ensure adequate sub- Other samples for viral culture/PCR viral detection Stool/rectal swab strate for coagulation factors. Active bleeding Guthrie card blood spots (HSV diagnosis)28,29 requires correction of thrombocytopenia and co- Nasopharyngeal secretions agulopathy. However, as the trend in the pro- Vesicle fluid thrombin time or international normalized ratio Cerebrospinal fluid (INR) provides the best indicator of hepatic func- Eye swab Ascitic fluid tion and recovery, coagulopathy is not routinely corrected, unless severe. Radiology Antibiotics effective against Gram-negative or- Chest X-ray ganisms, streptococci and listeria should be admin- Echocardiography Doppler ultrasound scan of abdomen istered even in the absence of overt sepsis, together with antifungal prophylaxis. Intravenous PCR, polymerase chain reaction. acyclovir (30 mg/kg/24 h) should be continued un- til herpes simplex virus (HSV) infection has been excluded. administration should be avoided. Fluid restriction Intravenous n-acetylcysteine may improve sur- is also indicated for deteriorating oliguric renal vival in liver failure by enhancing tissue oxygen- impairment, which may ultimately require dialysis ation.4 Although firm evidence is lacking, its use in support and for control of ascites. Ascites may supportive therapy is becoming widespread. In pre- respond to optimizing serum albumin (infusing term infants, the pharmacokinetics and excretion 396 P. McClean, S.M. Davison of intravenous n-acetylcysteine depend on weight distension due to hepatosplenomegaly and ascites. and gestational age.5 Convulsions may reflect meningoencephalitic The role of plasmapheresis has diminished as involvement (in one-third of cases). techniques in transplantation have advanced. In The outcome ranges from spontaneous recovery one study of 49 children, including neonates, the to a rapidly fatal fulminant course. Mortality was major benefit from plasmapheresis was improved estimated to be as high as 80% in 1986.11 The role of coagulation, with no effect on neurological compli- the oral antiviral agent pleconaril in neonatal infec- cations.6 There is increasing experience of extra- tion is being evaluated in a multicentre study.12 Of corporeal support systems such as the molecular three infants with life-threatening enteroviral adsorbent recirculation system (MARS)7 in children, hepatitis treated in an open study, two recovered but none in the neonatal period. completely.13

Herpes simplex virus infection Liver transplantation Herpes simplex virus (HSV) 1 and 2 are usually The role of orthotopic liver transplantation (OLT) in acquired due to exposure to infected maternal the management of acute liver failure in adults and genital secretions or lesions at delivery, although children is well established. However, there are intra-uterine and postnatal infection may occur. only a few small series describing the particular Risk of transmission is highest in seronegative medical and surgical challenges of OLT in neonates mothers with primary infection at delivery. Ma- or small infants.8,9 Infants with multi-organ failure, ternal viral shedding, however, is frequently uncontrolled sepsis, generalized mitochondrial dis- asymptomatic. Delivery by Caesarean section sig- orders or haemophagocytic lymphohistiocytosis nificantly reduces the risk of neonatal infection.14 should not be considered for liver transplantation Of 186 neonates with HSV infection, 60% pre- because they will succumb to their underlying dis- sented after day 5.15 HSV infection may involve the ease. Infants with NH or viral-induced liver failure skin, eyes, mucous membranes, brain, lung and are currently the most common groups undergoing liver. Liver failure may occur with disseminated OLT.8,9 In spite of the introduction of techniques disease or as the only manifestation; onset may which use part of larger livers (reduced, split or be sudden with lethargy, circulatory collapse and monosegment grafts), there are still difficulties mild jaundice. Absence of typical skin lesions is obtaining suitable donor organs, and the death rate common.15 on the waiting list remains high; 31% in one series.3 Standard treatment is intravenous acyclovir The mortality after transplant is 40–50% compared 30 mg/kg/24 h in three divided doses for 10–21 with 30–40% in older children or adults trans- days. However, there is evidence that 60 mg/kg/ planted for acute liver failure. Complications 24 h is associated with improved survival.16 Acyclo- include vascular thrombosis, sepsis and haemor- vir resistance occurs in 0.3% of immunocompetent rhage, but the incidence of acute rejection in this and 4–7% of immunocompromised patients,17 and age group is very low.8,9 may emerge during therapy.18,19 Cross-resistance occurs to penciclovir and famciclovir. Alternative antiviral agents include foscarnet and cidofovir, Infectious causes of neonatal liver but have increased toxicity. A novel group of anti- failure viral agents has been described that inhibit viral helicase–primase enzymes; potential advantages Enterovirus observed in animal models include superior efficacy and reduced resistance.20 Enteroviruses are small, single-stranded RNA vi- Despite acyclovir therapy, the clinical course is ruses, comprising polio-, Coxsackie A and B and often rapid deterioration, with death from multi- Echoviruses. Severe infection may occur in neo- organ failure within days. Of 59 infants with dis- nates with multi-organ involvement, including seminated disease treated with acyclovir between hepatic necrosis. Echovirus, particularly serotype 1981 and 1997, survival was 47%. OLT should be 11, is the most frequently identified virus. It considered. is postulated that Echovirus causes hepatic fail- ure by vascular rather than direct hepatocyte Hepatitis B virus infection damage.10 Symptoms typically develop between day 4 and Liver failure due to fulminant HBV infection is 7, with fever, lethargy, poor feeding and abdominal now rare due to screening and immunization Neonatal liver failure 397 policies.2,21 In the World Health Organization and salivary glands but spares the reticulo- European Region, 41 of 51 countries have imple- endothelial system. The aetiology of this condition mented universal immunization.22 In the UK, where is unclear, and it most likely represents a clinico- this has not been implemented, universal HBV pathological endpoint of different in-utero insults screening of all women during pregnancy since April to the fetal liver, including viral infection, 2000 has facilitated identification of at-risk infants. immunological mechanisms and an inherited pre- Fulminant HBV infection is particularly associ- disposition.30 ated with transmission from a ‘low risk’ surface- There is no evidence that these babies have the antigen-positive mother who is e-antigen negative HFE gene for hereditary haemochromatosis seen in and e-antibody positive. The diagnosis should be adults. Family studies have indicated both an auto- considered irrespective of immunization history. somal recessive and a maternal mode of inherit- Diagnosis is made by the detection of HBV DNA in ance. In some families, once a child is born with NH, peripheral blood. all subsequent children have been affected, and Spontaneous recovery may occur, but progres- this has occurred irrespective of having different sive deterioration may necessitate consideration fathers.30 Recurrent re-activation of maternal viral for OLT. Experience of antiviral strategies for infection or immunological factors have also been fulminant hepatitis B in infancy and childhood is postulated for maternal transmission. Early results limited. In adults, a potential role for lamivudine suggest that repeated immunoglobulin admin- in fulminant hepatitis B has been suggested.23 istration during subsequent pregnancies in these Pharmacokinetics and safety of lamivudine in mothers can significantly ameliorate the disease in 24 neonates has been established. the offspring.31 Infants with NH are often premature and/or Human herpes virus 6 small for dates. The obstetric history may reveal previous miscarriages or stillborn infants. Oligo- Human herpes virus 6 (HHV6) as a cause of neonatal hydramnios or polyhydramnios can complicate the liver failure (NLF) has been reported occasionally. pregnancy. The usual presentation is of acute de- Of two infants presenting at day 3 and day 5 of compensation of endstage liver disease in the first life,25 symptoms included fever, hypotonia, leth- 24 h of life with hypoglycaemia, coagulopathy, argy and shock. Both had raised hepatic trans- hypoalbuminaemia and ascites. Jaundice does de- aminases and thrombocytopenia. One recovered velop subsequently, but the serum transaminases spontaneously and the other died at day 15. Four are often within the normal range. Occasionally, infants with HHV6 in Durand et al's2 series under- infants present with a more protracted course and a went OLT. Ganciclovir is effective treatment,26 milder coagulopathy. Abnormalities of serum bile although resistance after prolonged exposure may acids suggestive of delta 4-3-oxosteroid 5-beta- occur.27 reductase deficiency have been associated with a particularly severe presentation of NH with a bad Other viruses prognosis. It has been suggested that this may reflect severe hepatocellular failure per se, and the Adenovirus and parvovirus have also been associ- infants do not respond to treatment with bile ated with NLF. It is likely that other viruses may salts.32 also be responsible. In one series, three of nine The diagnosis is reached by excluding other rec- infants had ‘nonA–nonB’ liver failure,8 and in an- ognized causes of NLF and demonstrating evidence other series, aetiology was undetermined in 13 of of iron overload. Much weight has been placed on 80 (16%) infants.2 finding high serum ferritin levels in this condition, but this is a frequent finding in neonatal liver disease of any aetiology.33 A95–100% saturation of transferrin is probably more specific for NH.34,35 Metabolic causes of NLF Extrahepatic iron deposition can be demonstrated in the salivary glands of the lip, although ensuring 36 Neonatal haemochromatosis an adequate sample is difficult. Decreased signal intensity on T2-weighted images on magnetic reso- NH or neonatal iron storage disease is a rare dis- nance imaging identifies iron in the liver and other order of abnormal iron storage presenting as liver tissues, and demonstrates a lack of siderosis in the failure within the first weeks of life. Iron accumu- spleen.37 Liver histology shows varying degrees of lates in the fetal liver, pancreas, heart, thyroid hepatocyte loss, stromal collapse and fibrosis with 398 P. McClean, S.M. Davison

Table 4 Anti-oxidant cocktail for NH symptoms are common with lethargy, hypotonia, poor feeding and vomiting. Some of these features N-acetylcysteine 140 mg/kg orally, then 70 mg/kg 4 hourly may be early signs of neurological involvement. for 19 doses More specific features of liver disease are conju- Selenium 2–3 µg/kg/day intravenously over 24 h Alpha tocopheryl polyethylene glycol succinate gated hyperbilirubinaemia, coagulopathy, ascites 20–30 IU/kg/24 h orally and moderately raised transaminases (between two Prostaglandin E1 0.4–0.6 µg/kg/h intravenously for 2–4 and 12 times normal). Hypoglycaemia is common weeks and may be due to secondary inhibition of oxidation Desferrioxamine 30 mg/kg/24 h intravenously over 8 h of fatty acids. Involvement of other organs result until ferritin <500 µg/l in neurological symptoms, myopathy, proximal renal tubular dysfunction, hypertrophic cardio- myopathy, and haematological and gastrointestinal disorders.42 Elevated blood lactate should raise suspicion of regenerative nodules. Grade 3–4 siderosis with a respiratory chain defect, particularly if it rises sparing of the Kupffer cells is a diagnostic further either postprandially or following an feature.31 intravenous glucose load. In the presence of Overall, the prognosis is poor with a mortality renal tubular dysfunction, plasma lactate may be rate of 75% in a recent review.30 General supportive lower, but urinary lactate is raised. Ketone bodies management is instigated as soon as the diagnosis are usually raised, with the plasma 3-OH- of NLF is recognized. There have been some reports butyrate:acetoacetate ratio often greater than 2. of spontaneous recovery. Three infants responded Intermediates of the tricarboxylic acid cycle plus to an anti-oxidant cocktail (Table 4) described in 3-methyl-glutaconic and 3-methylglutaric acid 42 1993.38 Success with this treatment may be more may be detected in urine. likely in a milder subgroup if commenced early.39 Evidence of involvement of other organs should Liver transplantation has been successful in NH but be sought by evaluating renal tubular function, with a high mortality on the waiting list (25–64%) echocardiography, and visual and auditory evoked and post-transplant (40–60%).30,35,39 responses. Neurological disease may be implied by a raised cerebrospinal fluid (CSF):plasma lactate ratio, an elevated CSF protein or abnormalities on Mitochondrial respiratory chain disorders magnetic resonance imaging. Liver histology almost invariably shows steatosis and fibrosis, which The mitochondrial respiratory chain consists of may have progressed to micronodular cirrhosis. five protein complexes, plus ubiquinone and cyto- Cholestasis, hepatocyte necrosis and increased iron chrome c, located on the inner mitochondrial staining may be present. Electron microscopy membrane. Other oxidative reactions within the reveals increased numbers of abnormal mitochon- mitochondria (the tricarboxylic acid cycle and fatty dria. Muscle may be histologically normal or show acid oxidation) generate reduced cofactors, such as steatosis. Ragged red fibres are rare in infancy, but if present, these are very suggestive of a flavin adenine dinucleotide (FADH2) and nicotina- mide adenine dinucleotide (NADH), which pass respiratory chain defect. electrons down the respiratory chain resulting in The definitive diagnosis is made by measuring the formation of ATP. This process is called oxidat- the enzymatic activity of respiratory chain com- ive phosphorylation. NLF has been recognized in plexes in affected tissues. Muscle is traditionally deficiencies of complex I, III and IV, multiple com- used as it is safer to obtain, and demonstration of plex deficiencies and in mitochondrial DNA extrahepatic disease in a patient with NLF is a (mtDNA) depletion syndrome. contra-indication to OLT.43 However, due to the Mitochondrial disorders of the electron transport variability of abnormal mtDNA within tissues proteins in the liver can present as NLF or as gradu- (heteroplasmy), and the wide range of normal ally progressive liver disease which can suddenly values, the results can be difficult to interpret.40 decompensate.40 Most infants have extrahepatic Liver biopsies are potentially hazardous, and ab- features, although a few patients with respiratory normalities may be artifactual due to the severity chain defects isolated to the liver have been de- of the liver damage. mt DNA depletion syndrome scribed.41 Typically, severe liver failure develops is diagnosed by demonstrating a low ratio of in the first few weeks of life. Some cases have mtDNA to nuclear DNA using Southern blotting evidence of prenatal liver disease. Non-specific techniques. Neonatal liver failure 399

Supportive treatment is usually the only thera- monitoring of plasma amino acids, urinary suc- peutic option. The use of ubiquinone, riboflavin cinylacetone, serum AFP, ophthalmological ex- and chloroacetate have not been shown to affect amination and hepatic imaging are important. prognosis in patients presenting with NLF. Liver Infants who do not respond to NTBC, or in whom transplantation has been successful in a few hepatocellular carcinoma is suspected, are con- patients with isolated liver disease and a less fulmi- sidered for OLT. This removes the risk of hepato- nant presentation41,44, but many patients succumb cellular carcinoma in the future, and the children to neurological disease after transplant. Therefore, can return to a normal diet, but the renal tubular Thompson et al.43 recommend that any evidence defect may not be corrected.49 of extrahepatic disease is a contra-indication to OLT. Genetic counselling is difficult as many of these Haemophagocytic lymphohistiocytosis defects are sporadic. However, mtDNA originates from the ovum and maternal inheritance has been Haemophagocytic lymphohistiocytosis (HLH) is a 45 reported. mtDNA depletion syndrome is probably rare disorder involving inappropriate activation of due to a defect of a nuclear gene that codes for macrophages. It is divided into a primary, familial replication of the mitochondrial genome. Consan- form and a secondary form usually triggered by guinity is common and the mode of inheritance infection in an immunocompromised host. The pri- 46 appears to be autosomal-recessive. mary form can present as NLF with hepatosplenom- egaly, markedly abnormal liver function tests and a Tyrosinaemia type 1 high serum ferritin, which may cause confusion with NH.50 Other diagnostic clues are fever, raised Hereditary tyrosinaemia type 1 (HT1), a recessive triglycerides, hypofibrinogenaemia and cytopenia. condition, is caused by a deficiency of fumaryl- The diagnosis is usually confirmed by evidence acetoacetate hydoxylase (FAH), resulting in of haemophagocytosis in a bone marrow aspirate. accumulation of toxic metabolites, fumaryl- Initial management includes chemotherapy, usually acetoacetate and maleylacetoacetate, and their dexamethasone and etoposide, but long-term 51 reduced derivatives succinylacetoacetate and survival requires a bone marrow transplant. 52 succinylacetone. These are thought to be respon- In one series, 5-year survival was 21%. OLT sible for liver and proximal renal tubular dys- is contra-indicated due to recurrence in the 50 function, and porphyria-like crises. The clinical graft. presentation is variable but, in infants less than 6 months old, HT1 causes acute liver failure. Coagu- lopathy is a dominant feature and has been de- Conclusions scribed in the absence of other signs of liver failure.47 The serum transaminases are only mildly NLF is an uncommon but challenging condition. It raised and some infants do not develop jaundice. must be considered early in the differential diag- Plasma tyrosine, phenylalanine and methionine nosis of coagulopathy in the newborn. Infection are raised, but these can be elevated in liver and metabolic disorders are the most common disease. The presence of succinyl acetone in the aetiologies. The overall mortality rate in the two urine is diagnostic. Very high levels of alpha feto- series described in this paper was 60%. Sepsis, protein (AFP) are typical. The diagnosis can be haemorrhage and multi-organ failure were the confirmed by measuring FAH activity in skin fibro- main causes of death. In each series, 24% of blasts or liver cells. patients survived with their native liver following 2-(2-Nitro-4-trifluoromethylbenzoyl)-1-3-cyclo- intensive supportive management and appropriate hexanedione (NTBC), a bleaching herbicide, has specific medical therapy.2,3 In one series, five of 16 been used to block the formation of toxic metabo- infants listed for a liver transplant died before an lites in patients with HT1 since 1992. Plasma tyro- organ became available,3 but in both series, the sine and phenylalanine remain raised unless the survival after transplant was 50%. Children with patient's dietary intake is restricted. In 80 mitochondrial respiratory chain defects may patients commenced on NTBC before 6 months of succumb to previously unrecognized neurological age, the response rate was 90%, but there are disease post-transplant. The recognition and man- still concerns that the future risk of developing agement of these infants require intensive hepatocellular carcinoma, a well-recognized co-operation between neonatal, hepatology and complication of HT1, is not removed.48 Ongoing transplant teams. 400 P. McClean, S.M. Davison

12. Sawyer MH. Enterovirus infections: diagnosis and Practice points treatment. Pediatr Infect Dis J 1999;18:1033–40. 13. Aradottir E, Alonso EM, Shulman ST. Severe neonatal entero- • NLF should be considered in any neonate viral hepatitis treated with pleconaril. Pediatr Infect Dis J 2001;20:457–9. with a coagulopathy which does not respond 14. Brown Z, Wald A, Morrow R et al. Effect of serologic status to intravenous vitamin K. and Caesarean delivery on transmission rates of herpes • Jaundice and raised transaminases are not simplex virus from mother to infant. JAMA 2003;289:203–9. always present in infants with in-utero onset 15. Kimberlin DW, Lin CY, Jacobs RF et al. Natural history of of cirrhosis. neonatal herpes simplex virus infections in the acyclovir • era. Pediatrics 2001;108:223–9. Infection and metabolic diseases are the 16. Kimberlin DW, Lin CY, Jacobs RF et al. Safety and efficacy of most frequent causes of NLF. high-dose intravenous acyclovir in the management of neo- • Feeds should be withheld until the results of natal herpes simplex virus infections. Pediatrics 2001; galactose-1-phosphate uridyl transferase and 108:230–8. urine organic and amino acids are available. 17. Bacon TH, Levin MJ, Leary JJ et al. Herpes simplex virus • resistance to acyclovir and penciclovir after two decades of Monitor and maintain normoglycaemia. antiviral therapy. Clin Microbiol Rev 2003;16:114–28. 18. Oram RJ, Marcellino D, Strauss D et al. Characterization of an acyclovir-resistant herpes simplex virus type 2 strain isolated from a premature neonate. J Infect Dis 2000; 181:1458–61. Research directions 19. Levin MJ, Weinberg A, Leary JJ et al. Development of • Genetics of mitochondrial cytopathies. Can acyclovir resistant herpes simplex virus early during the treatment of herpes neonatorum. Pediatr Infect Dis J 2001; the ratio of mutant to normal mtDNA be 20:1094–7. influenced? 20. Kleymann G, Fischer R, Betz UA et al. New helicase–primase • Pathogenesis of NH. inhibitors as drug candidates for the treatment of herpes simplex disease. Nat Med 2002;8:392–8. 21. Kao JH, Hsu HM, Shau WY et al. Universal hepatitis B vaccination and the decreased mortality from fulminant References hepatitis in infants in Taiwan. J Paediatr 2001;139:349–52. 22. Van Damme P, Vorsters A. Hepatitis B control in Europe by 1. Shneider BL. Neonatal liver failure. Curr Opin Pediatr 1996; universal vaccination programmes: the situation in 2001. J 8:495–501. Med Virol 2002;67:433–9. 2. Durand P, Debray D, Mandel R et al. Acute liver failure in 23. Santantonio T, Mazzola M, Pastore G. Lamivudine is safe and infancy: a 14-year experience of a pediatric liver transplan- effective in fulminant hepatitis B. J Hepatol 1999;30:551. tation center. J Pediatr 2001;130:871–6. 24. Moodley D, Pillay K, Naidoo K et al. Pharmacokinetics of 3. Aw MM, Rela M, Heaton ND et al. Neonatal liver failure—a zidovudine and lamivudine in neonates following coadmin- ten year experience [abst]. J Pediatr Gastroenterol Nutr istration of oral doses every 12 hours. J Clin Pharmacol 2001;32:381A. 2001;41:732–41. 4. Harrison PM, Wendon JA, Gimson AE et al. Improvement by 25. Mendel I, de Matteis M, Bertin C et al. Fulminant hepatitis in acetylcysteine of hemodynamics and oxygen transport in neonates with human herpesvirus 6 infection. Pediatr Infect fulminant hepatic failure. N Engl J Med 1991;324:1852–7. Dis J 1995;14:993–7. 5. Ahola T, Fellman V, Laaksonen R et al. Pharmacokinetics of 26. Tokimasa S, Hara J, Osugi Y et al. Ganciclovir is effective for intravenous N-acetylcysteine in pre-term newborn infants. prophylaxis and treatment of human herpesvirus-6 in allo- Eur J Clin Pharmacol 1999;55:645–50. geneic stem cell transplantation. Bone Marrow Transplant 6. Singer AL, Olthoff KM, Kim H et al. Role of plasmapheresis in 2002;29:595–8. the management of acute hepatic failure in children. Ann 27. Manichanh C, Olivier-Aubron C, Lagarde JP et al. Selection Surg 2001;234:418–24. of the same mutation in the U69 protein kinase gene of 7. Stange J, Mitzner S, Risler T et al. Molecular adsorbent human herpesvirus-6 after prolonged exposure to ganciclo- recycling system (MARS). Clinical results of a new membrane vir in vitro and in vivo. J Gen Virol 2001;82:2767–76. based blood purification system for bioartificial liver 28. Ilona LF, Pia O, Marianne F et al. Detection of herpes support. Artif Organs 1999;23:319–30. simplex virus DNA in dried blood spots making a retrospec- tive diagnosis possible. J Clin Virol 2003;26:39–48. 8. Bonatti H, Muiesan P, Connelly S et al. Hepatic transplan- 29. Barbi M, Binda S, Primache V et al. Use of Guthrie cards for tation in children under 3 months of age: a single centre's the early diagnosis of neonatal herpes simplex virus disease. experience. J Pediatr Surg 1997;32:486–8. Pediatr Infect Dis J 1998;17:251–2. 9. Noujaim HM, Mayer DA, Buckels JA et al. Techniques for and 30. Kelly AL, Lunt PW, Rodrigues F et al. Classification and outcome of liver transplantation in neonates and infants genetic features of neonatal haemochromatosis: a study of weighing up to 5 kilograms. J Pediatr Surg 2002;37:159–64. 27 affected pedigrees and molecular analysis of genes im- 10. Wang J, Atchison RW, Walpusk J et al. Echovirus hepatic plicated in iron metabolism. J Med Genet 2001;38:599–610. failure in infancy: report of four cases with speculation on 31. De Boissien D, Knisely AS. Neonatal hemochromatosis. Suchy the pathogenesis. Pediatr Dev Pathol 2001;4:454–60. FJ, Sokol RJ, Balistreri WF, editors. Liver disease in 11. Modlin JF. Perinatal echovirus infection: insight from a children. Philadelphia: Lippincott, Williams and Wilkins, literature review of 61 cases of serious infection and 16 2001;641–7. outbreaks in nurseries. Rev Infect Dis 1986;8:918–26. Neonatal liver failure 401

32. Siafakas CG, Jonas MM, Peraz-Atayde AR. Abnormal bile acid 43. Thompson M, McKiernan P, Buckels J et al. Generalised metabolism and neonatal hemochromatosis: a subset with mitochondrial cytopathy is an absolute contraindication to poor prognosis. J Pediatr Gastroenterol Nutr 1997; orthotopic liver transplant in childhood. J Pediatr Gastroen- 25:321–6. terol Nutr 1998;26:478–81. 33. Lee WS, McKiernan PJ, Kelly DA. Serum ferritin level in 44. Dubern B, Broue P, Dubuisson C et al. Orthotopic liver neonatal fulminant liver failure [letter]. Arch Dis Child Fetal transplantation for mitochondrial respiratory chain disor- Neonatal Ed 2001;85:F226. ders: a study of 5 children. Transplantation 2001;71:633–7. 34. Vohra P, Haller C, Emre S et al. Neonatal hemochromatosis: 45. Rotig A, Bessis J-L, Romero N et al. Maternally inherited the importance of early recognition of liver failure. J duplication of the mitochondrial genome in a syndrome of Pediatr 2000;136:537–41. proximal tubulopathy, diabetes mellitus and cerebellar 35. Sigurdsson L, Reyes J, Kocoshis SA et al. Neonatal hemo- ataxia. Am J Hum Genet 1992;50:364–70. chromatosis: outcomes of pharmacologic and surgical 46. Bakker HD, Scholte HR, Dingemans KP, et al. Depletion of therapies. J Pediatr Gastroenterol Nutr 1998;26:85–9. mitochondrial deoxyribonucleic acid in a family with fatal 36. Knisely AS, O'Shea PA, Stocks JF et al. Oropharngeal and neonatal liver disease. J Pediatr 1996;128:683–7. upper respiratory tract mucosal-gland siderosis in neonatal 47. Croffie JM, Gupta SK, Chung SKF et al. Tyrosinemia type 1 hemochromatosis: an approach to biopsy diagnosis. J should be suspected in infants with severe coagulopathy Pediatr 1988;113:871–4. even in the absence of other signs of liver failure. Pediatrics 37. Hayes AM, Jaramillo D, Levy HL et al. Neonatal hemochro- 1999;103:675–8. matosis: diagnosis with MR imaging. Am J Roentgenol 1992; 48. Holme E, Lindstedt S. Nontransplant treatment of 159:623–5. tyrosinemia. Clin Liver Dis 2000;4:805–14. 38. Shamieh I, Kibort PK, Suchy FJ et al. Antioxidant therapy for 49. Mohan N, McKiernan P, Preece MA et al. Indications and neonatal iron storage disease [abst]. Pediatr Res 1993; outcome of liver transplantation in tyrosinaemia type 1. Eur 33:109A. J Pediatr 1999;158(Suppl 2):S49–54. 39. Flynn DM, Mohan N, McKiernan P et al. Progress in treatment 50. Parizhskaya M, Reyes J, Jaffe R. Hemophagocytic syndrome and outcome for children with neonatal haemochromatosis. presenting as acute hepatic failure in two infants: clinical Arch Dis Child Fetal Neonatal Ed 2003;88:F124–7. overlap with neonatal hemochromatosis. Pediatr Develop 40. Cormier-Daire V, Chretien D, Rustin P et al. Neonatal and Pathol 1999;2:360–6. delayed-onset liver involvement in disorders of oxidative 51. Sung L, King SM, Carcao M et al. Adverse outcomes in phosphorylation. J Pediatr 1997;130:817–22. primary hemophagocytic lymphohistiocytosis. J Pediatr He- 41. Goncalves I, Hermans D, Chretien D et al. Mitochondrial matol Oncol 2002;24:550–4. respiratory chain defect: a new etiology for neonatal 52. Arico M, Janka G, Fischer A et al. Hemophagocytic lympho- cholestasis and early liver insufficiency. J Hepatol 1995; histiocytosis. Report of 122 children from the International 23:290–4. Registry. Leukaemia 1996;10:197–203. 42. Munnich A, Rotig A, Chretien D et al. Clinical presentations 53. Williams MD, Chalmers EA, Gibson BES. The investigation and laboratory investigations in respiratory chain and management of neonatal haemostasis and thrombosis. deficiency. Eur J Pediatr 1996;155:262–74. Br J Haematol 2002;119:295–309. ACUTE LIVER FAILURE IN CHILDREN: THE FIRST 348 PATIENTS IN THE PEDIATRIC ACUTE LIVER FAILURE STUDY GROUP

ROBERT H. SQUIRES,JR, MD, BENJAMIN L. SHNEIDER, MD, JOHN BUCUVALAS, MD, ESTELLA ALONSO, MD, RONALD J. SOKOL, MD, MICHAEL R. NARKEWICZ, MD, ANIL DHAWAN, MD, PHILIP ROSENTHAL, MD, NORBERTO RODRIGUEZ-BAEZ, MD, KAREN F. MURRAY, MD, SIMON HORSLEN, MD, MARTIN G. MARTIN, MD, PHD, M. JAMES LOPEZ, MD, PHD, HUMBERTO SORIANO, MD, BRENDAN M. MCGUIRE, MD, MAUREEN M. JONAS, MD, NADA YAZIGI, MD, ROSS W. SHEPHERD, MD, KATHLEEN SCHWARZ, MD, STEVEN LOBRITTO, MD, DANIEL W. THOMAS, MD, JOEL E. LAVINE, MD, PHD, SAUL KARPEN, MD, PHD, VICKY NG, MD, DEIRDRE KELLY, MD, NANCY SIMONDS, RN, AND LINDA S. HYNAN,PHD

Objectives To determine short-term outcome for children with acute liver failure (ALF) as it relates to cause, clinical status, and patient demographics and to determine prognostic factors. Study design A prospective, multicenter case study collecting demographic, clinical, laboratory, and short-term outcome data on children from birth to 18 years with ALF. Patients without encephalopathy were included if the prothrombin time and international normalized ratio remained >20 seconds and/or >2, respectively, despite vitamin K. Primary outcome measures 3 weeks after study entry were death, death after transplantation, alive with native liver, and alive with transplanted organ. Results The cause of ALF in 348 children included acute acetaminophen toxicity (14%), metabolic disease (10%), autoim- mune liver disease (6%), non-acetaminophen drug-related hepatotoxicity (5%), infections (6%), other diagnosed conditions (10%); 49% were indeterminate. Outcome varied between patient sub-groups; 20% with non-acetaminophen ALF died or underwent liver transplantation and never had clinical encephalopathy. Conclusions Causes of ALF in children differ from in adults. Clinical encephalopathy may not be present in children. The high percentage of indeterminate cases provides an opportunity for investigation. (J Pediatr 2006;148:652-8)

cute liver failure (ALF) is a dramatic clinical syndrome in which previously healthy children rapidly lose hepatic function and become critically ill within 1,2 3-6 From the University of Pittsburgh, Mt. Sinai days. Over the last 25 years, single center experiences or general re- Medical Center, Cincinnati Children’s Hos- A1,2,7 views of pediatric ALF from Europe and North America identified a variety of pital, Northwestern University, University of Colorado, Kings College-London (En- infectious, metabolic, cardiovascular, and drug-related causes, as well as indeterminate gland), University of California-San Fran- 4-6 cases. These pediatric studies used the adult definition of ALF, which requires the cisco, University of Texas Southwestern presence of hepatic encephalopathy (HE) within 8 weeks of the development of Medical Center, University of Washington, University of Nebraska, University of Cali- clinical jaundice. Unfortunately, HE is difficult to assess in many infants and children fornia at Los Angeles, University of Michi- and may not be essential to the diagnosis of ALF in children.3,8 gan, Drexel University, University of Ala- bama-Birmingham, Children’s Hospital of The outcome for children with ALF remains poor for infants under 1 year of Boston, Washington University, Johns Hop- age,3 patients with accidental acetaminophen overdose,9 and Wilson’s disease pre- kins University, Columbia University, Uni- 10 versity of Southern California, University of senting with hepatic HE. Spontaneous recovery (ie, survival without transplanta- California at San Diego, Baylor College of 5,6 tion) remains between 15% to 20% for those with severe HE. Given the ongoing Medicine, Hospital for Sick Children (Can- shortage of donor livers,11 development of a reliable prognostic score will be useful in ada), University of Birmingham (England). Supported by NIH grant RO1-DK58369- allocating organs to the most needy patients. 01; and by M01-RR00069, M01-RR00037, The Pediatric Acute Liver Failure (PALF) study group was formed in 1999 to and M01 RR08084 from the General Clin- develop a database that would facilitate an improved understanding of the pathogen- ical Research Center Program of the Na- tional Center for Research Resources of esis, treatment, and outcome of ALF in children. These data will also serve to identify the NIH. factors that will help to predict the likelihood of death or need for liver transplanta- Submitted for publication Jul 17, 2005; last tion. revision received Oct 3, 2005; accepted Dec 19, 2005. Reprint requests: Robert H. Squires, Jr, MD, University of Pittsburgh, Children’s Hospital of Pittsburgh, 3705 Fifth Ave, Pittsburgh, PA 15213. E-mail: [email protected]. 0022-3476/$ - see front matter AIH Autoimmune hepatitis INR International normalized ratio Copyright © 2006 Elsevier Inc. All rights ALF Acute liver failure PALF Pediatric acute liver failure reserved. HE Hepatic encephalopathy TTMC Tukey-type multiple comparison 10.1016/j.jpeds.2005.12.051

652 Table I. Assessment of encephalopathy for young children: birth to age 3 years Grade Clinical Asterixis/reflexes Neurologic signs Early (I and II) Inconsolable crying, sleep reversal, inattention to task Unreliable/normal or hyperreflexic Untestable Mid (III) Somnolence, stupor, combativeness Unreliable/hyperreflexic Most likely untestable Late (IV) Comatose, arouses with painful stimuli (IVa) or no Absent Decerebrate or decorticate response (IVb)

Table II. Patient demographics Acetaminophen Indeterminate All Others (%) (%) (%) ␹2 P value Total (N ϭ 348) 48 (14) 169 (49) 131 (38) Gender .0002 Female (n ϭ 181) 38 (79) 78 (46) 65 (50) Male (n ϭ 167) 10 (21) 91 (54) 66 (50) Age Ͻ 3.0 (n ϭ 127) 2 (4) 68 (40) 57 (44) Ͻ.0001 ୑ 3.0 (n ϭ 221) 46 (96) 101 (60) 74 (56) Race .0275* Black (n ϭ 54) 5 (10) 32 (19) 17 (13) Asian (n ϭ 30) 5 (10) 14 (8) 11 (8) Hispanic (n ϭ 53) 3 (6) 32 (19) 18 (14) Native American (n ϭ 4) 0 (0) 2 (1) 2 (2) Other (n ϭ 21) 3 (6) 10 (6) 8 (6) White (n ϭ 186) 32 (67) 79 (47) 75 (57)

*Test of white versus nonwhite.

METHODS Data Coordinating Center for review by the principal inves- tigator for the pediatric study. If discrepancies were identified, Organization the site was queried, and on resolution, data were then en- The PALF study group began as an adjunct to the tered into the PALF database. The National Institutes of National Institutes of Health–sponsored, adult-focused ALF Health provided a Certificate of Confidentiality to the study Study Group (William Lee, MD, Principal Investigator). The and IRB approval was secured at each site before patient PALF study group now consists of 24 active pediatric sites, 21 enrollment. within the United States, 1 in Canada, and 2 in the United Kingdom. Working groups of pediatric hepatologists estab- Subjects of the Study lished definitions for ALF and various diagnostic categories. Representatives from all participating centers approved final Patients from birth through 18 years of age were eligible recommendations from the working groups. Patient enroll- for enrollment if they met the following entry criteria for the ment began in December 1999. PALF study: (1) children with no known evidence of chronic liver disease, (2) biochemical evidence of acute liver injury, and (3) hepatic-based coagulopathy defined as a prothrombin Data Collection time (PT) Ն 15 seconds or international normalized ratio Following informed consent from a parent or legal (INR) Ն 1.5 not corrected by vitamin K in the presence of guardian, demographic, clinical and laboratory information clinical HE or a PT Ն20 seconds or INR Ն2.0 regardless of were recorded daily for 7 days. In most patients, an additional the presence or absence of clinical HE. A standard adult aliquot of serum or plasma was collected on each of the 7 Ϫ clinical coma grade scale was used for older children, and a study days, frozen at 70° C, and then shipped to the Data coma grade scale was adapted for infants and children Ͻ 4 Coordinating Center located at the University of Texas years old (Table I).12 Southwestern Medical Center in Dallas, Texas. Diagnostic evaluation and medical management were consistent with the standard of care at each site. As in the adult study, our Diagnostic Categories primary outcome measures determined at 3 weeks after entry Diagnostic criteria for acute acetaminophen toxicity into the study included death, death after transplantation, included a toxic serum acetaminophen level on the basis of the alive with native organ, and alive with transplanted organ. Rumack nomogram13 or a history of an acute ingestion of 100 Completed data forms were forwarded under code to the mg/kg within a 24-hour period. The diagnosis of autoim-

Acute Liver Failure In Children: The First 348 Patients In The Pediatric Acute Liver Failure Study Group 653 Figure 1. Age of the patient at entry into the PALF study. Figure 2. Calendar month when patients were entered into the PALF study. Patients were divided into three diagnostic cagtegories: acetaminophen toxicity, indeterminate, and all others with an established mune hepatitis (AIH) was established if a patient had 1 or diagnosis. more positive autoantibody test results (anti-nuclear antibody Ն 1:80, smooth muscle antibody Ն 1:20, liver-kidney micro- Ն somal antibody 1:20) and no evidence of serologically bilirubin, peak PT, maximum coma grade). These models 14 defined viral hepatitis. Non-acetaminophen drug-induced also included sex (male vs female) and age (Ͻ3vsՆ 3 years). hepatitis was diagnosed if a temporal relationship between SPSS V12.0 and SAS V9.1 were used in all analyses. The exposure to a suspected drug and the onset of ALF was assumptions for all statistical tests were checked for violations. established and other common causes were excluded. Hepa- Statistical significance was set at P Ͻ .05. titis A, B, or C infection was confirmed serologically or by polymerase chain reaction. Evidence of other viral infections RESULTS required a positive immunoglobulin M antibody, evidence of Between December 1999 and December 2004, 348 virus in liver tissue, or a positive polymerase chain reaction. children were enrolled; patient demographics are outlined in Metabolic diseases were diagnosed by laboratory tests (eg, Table II. The median dose of acetaminophen ingested was alpha-1-antitrypsin phenotype of ZZ), analysis of liver tissue 183 mg/kg (range 19.2 to 734.1). There was an association (eg, mitochondrial enzyme defect), or analysis of cultured skin between cause and sex (P ϭ .0002), with the percentage of fibroblasts (eg, fatty acid oxidation defect). If the site inves- females significantly higher in the acetaminophen group com- tigator suspected an infection or metabolic disease but lacked pared with both non-acetaminophen groups (TTMC P Ͻ supporting evidence or if a specific diagnosis could not be .05, 79% vs 46% or vs 60%, respectively). The association established, the final diagnosis was registered as indetermi- between the three etiologic categories and race (white vs nate. non-white) was found to be significant (P ϭ .0275) with more whites in the acetaminophen versus the indeterminate group Statistical Methods (TTMC P Ͻ .05, 67% vs 47%; respectively). Children under Diagnostic categories were defined as acetaminophen, 3 years of age accounted for 36.5% (127/348) of patients indeterminate and all others in whom a specific diagnosis was (Figure 1). ALF caused by acetaminophen and those of in- determined. Age was dichotomized into patients younger determinate cause appear to occur more commonly during the than 3 years of age and those 3 years of age and older. Race cooler months, but this did not reach statistical significance (P was dichotomized into white versus non-white. ϭ .2240) (Figure 2). All associations between pairs of dichotomous or di- HE at admission to the study and peak HE during the chotomized variables (2-way tables) were conducted with ␹2 subsequent 7 days stratified by age and diagnosis is captured analyses. For those ␹2 analyses found significant, post hoc in Table III. HE was present more frequently in the com- Tukey-type multiple comparison tests for proportions bined non-acetaminophen groups than the acetaminophen (TTMC) were performed.15 The 2 independent samples pro- group (56.6% [164/290] vs 39.6% [19/48]; P ϭ .0425). In portions test with correction was used to compare proportions addition, the number of patients with development of HE for two groups. Two different logistic regression models were during the 7-day study period was greater in those with used to predict death or transplantation in the 2 non-acet- non-acetaminophen ALF compared with the acetaminophen aminophen groups with data available at admission (coma group (68.4% [203/297] vs 39.6% [19/48]; P ϭ .0002). Pa- grade, PT, and total bilirubin) and peak measures within the tients in all diagnostic and age categories experienced wors- first 7 days of hospitalization or before transplantation (peak ening HE, although this occurred more commonly in older

654 Squires et al The Journal of Pediatrics • May 2006 Table III. Encephalopathy on admission and at peak during the 7-day study period stratified by age and diagnosis category Acetaminophen (%) Indeterminate (%) All Others (%) Admission coma grade Age Ͻ 3.0 Ն 3.0 yr Age Ͻ 3.0 Ն 3.0 yr Age Ͻ 3.0 Ն 3.0 yr 0(Nϭ 155) 0 (0) 29 (63) 27 (42) 38 (38) 30 (55) 31 (44) 1–2 (N ϭ 138) 0 (0) 12 (26) 34 (53) 45 (45) 20 (36) 27 (38) 3(Nϭ 31) 1 (50) 4 (9) 2 (3) 12 (12) 5 (9) 7 (10) 4(Nϭ 14) 1 (50) 1 (2) 1 (2) 5 (5) 0 (0) 6 (8) Peak coma grade Ͻ3.0 Ն 3.0 Ͻ 3.0 Ն 3.0 Ͻ 3.0 Ն 3.0 0(Nϭ 123) 0 (0) 29 (63) 21 (31) 25 (25) 26 (46) 22 (30) 1–2 (N ϭ 134) 0 (0) 12 (26) 36 (54) 40 (40) 21 (38) 25 (34) 3(Nϭ 39) 1 (50) 1 (2) 4 (6) 15 (15) 8 (14) 10 (14) 4(Nϭ 49) 1 (50) 4 (9) 6 (9) 21 (21) 1 (2) 16 (22)

Table IV. Clinical characteristics by diagnostic category and age category Diagnosis Age group Acetaminophen Indeterminate All others <3yr >3yr (%) (%) (%) ␹2 (%) (%) ␹2 P value 221 ؍ n 127 ؍ P value n 131 ؍ n 169 ؍ n 48 ؍ n Ascites (n ϭ 78) 2 (4) 35 (21) 41 (31) .0004 42 (33) 36 (16) .0003 Seizure (n ϭ 23) 0 (0) 13 (8) 10 (8) .1394 6 (5) 17 (8) .2833 Ventilation Support (n ϭ 145) 8 (17) 75 (44) 62 (47) .0007 62 (49) 83 (38) .0402 Pressor Support (n ϭ 82) 5 (10) 33 (20) 44 (34) .0012 33 (26) 49 (22) .4198 Hemofiltration (n ϭ 33) 3 (6) 11 (7) 19 (15) .0457 8 (6) 25 (11) .1244 Plasmapheresis (n ϭ 35) 3 (6) 21 (12) 11 (8) .3301 8 (6) 27 (12) .0772 Red cell transfusion (n ϭ 146) 7 (15) 81 (48) 58 (44) .0002 79 (62) 67 (30) Ͻ.0001 Fresh frozen plasma (n ϭ 221) 20 (42) 122 (72) 79 (60) .0003 90 (71) 131 (59) .0306 patients (24.0% [29/121] vs 34.6% [75/217]). Other clinical analysis. Survival and need for liver transplantation varied and management features stratified by diagnosis and age are depending on the diagnosis. Spontaneous recovery was great- listed in Table IV. Development of ascites, need for ventilator est in children with acetaminophen toxicity (45/48; 94%), and blood pressure support, and requirement for red blood worst for those with non-acetaminophen drug-induced liver cell and plasma infusions were more likely to develop in injury (7/17; 41%) and with indeterminate cause (73/169; patients within the 2 non-acetaminophen groups than in the 43%). Patient outcomes on the basis of admission and peak acetaminophen group. In comparisons by age category, those HE are outlined in Table VII. Patients who never had HE in the younger age group were more likely to have develop- were more likely to experience spontaneous recovery than ment of ascites, require ventilator support, and infusions of those who did (78.9% vs 40.1%; P Ͻ .0001). In contrast, red blood cells and fresh frozen plasma. patients with development of stage III or IV HE had a Overall, a specific cause of ALF was not identified in spontaneous recovery rate of only 33% and 22%, respectively. 49% of patients and 54% of children less than 3 years of age. Logistic regression analysis (Table VIII) to predict death or (Table V; available at www.jpeds.com) acetaminophen toxic- liver transplantation identified total bilirubin Ն 5 mg/dL, ity accounted for only 14% of all patients, with 96% of these INR Ն 2.55, and HE to be risk factors if present on admis- cases occurring in older patients. Specific viruses, drugs, tox- sion. The logistic regression model with peak values was ins, and metabolic disorders are also listed in Table V. Inter- similar to that for admission, with increasing predictive values estingly, only 3 patients with acute hepatitis A infection, 1 (odds ratios) for all variables. patient with hepatitis C, and no patients with hepatitis B were identified in this cohort. DISCUSSION The short-term outcome for each diagnostic category is This report of the first 348 children in the PALF data described in Table VI. When all cases are considered, the set highlights a number of important observations: (1) HE is association between outcome and sex was significant (P ϭ not an absolute requirement to establish the diagnosis of ALF .0428) with spontaneous recovery for females higher than for in children; (2) a specific diagnosis was not made in almost males (60% vs 46%, respectively), but this association disap- half of all infants and children; (3) the causes of ALF in pears when the acetaminophen group is removed from the children differ from those seen in adults,16 with children

Acute Liver Failure In Children: The First 348 Patients In The Pediatric Acute Liver Failure Study Group 655 Table VI. Short-term (21 days) outcome of children with ALF

Not transplanted Transplanted ␹2 Alive Dead Alive Dead p-value Age .2141 Ͻ 3.0 (n ϭ 127) 67 (53) 24 (19) 33 (26) 3 (2) Ն 3.0 (n ϭ 221) 119 (54) 25 (11) 72 (33) 5 (2) Sex .0428 Female (n ϭ 181) 109 (60) 19 (10) 50 (28) 3 (2) Male (n ϭ 167) 77 (46) 30 (18) 55 (33) 5 (3) Diagnosis Ͻ.0001 Acetaminophen (n ϭ 48) 45 (94) 1 (2) 1 (2) 1 (2) Other Dx Categories (n ϭ 300) 141 (47) 48 (16) 104 (35) 7 (2) Dx details Acetaminophen (n ϭ 48) 45 (94) 1 (2) 1 (2) 1 (2) Indeterminate (n ϭ 169) 73 (43) 18 (11) 71 (42) 7 (4) Autoimmune (n ϭ 22) 12 (55) 3 (14) 7 (32) 0 Infectious (n ϭ 20) 10 (50) 5 (25) 5 (25) 0 Non-APAP drug induced liver 7 (41) 5 (29) 5 (29) 0 disease (n ϭ 17) Metabolic (n ϭ 36) 16 (44) 8 (22) 12 (33) 0 Other (n ϭ 20) 10 (50) 6 (30) 4 (20) 0 Shock (n ϭ 16) 13 (81) 3 (19) 0 0 For the Non-APAP patients Age .0884 Ͻ 3.0 (n ϭ 125) 66 (53) 23 (18) 33 (26) 3 (2) Ն 3.0 (n ϭ 175) 75 (43) 25 (14) 71 (41) 4 (2) Sex .2274 Female (n ϭ 143) 74 (52) 18 (13) 49 (34) 2 (1) Male (n ϭ 157) 67 (43) 30 (19) 55 (35) 5 (3)

Table VII. Patient outcome based on admission and peak encephalopathy

Not transplanted Transplanted ␹2 Alive (%) Dead (%) Alive (%) Dead (%) P value* Admission Coma Grade 0(nϭ 155) 102 (66) 14 (9) 35 (23) 4 (3) .0002 1–2 (n ϭ 138) 59 (43) 20 (14) 57 (41) 2 (1) 3(nϭ 31) 11 (35) 9 (29) 10 (32) 1 (3) 4(nϭ 14) 8 (57) 3 (21) 2 (14) 1 (7) Peak Coma Grade 0(nϭ 123) 97 (79) 9 (7) 16 (13) 1 (1) Ͻ.0001 1–2 (n ϭ 134) 65 (49) 12 (9) 55 (41) 2 (1) 3(nϭ 39) 13 (33) 8 (21) 18 (46) 0 (0) 4(nϭ 49) 11 (22) 18 (37) 15 (31) 5 (10)

*Test compares spontaneous recovery (alive–not transplanted) to the other 3 groups combined. having more indeterminate cases and fewer acetaminophen and, similar to other studies in the post-liver transplantation and viral-induced cases; and (4) short-term outcome varied era, only 25% of our children with a peak HE of grade 3-4 among diagnostic groups. had a spontaneous recovery. However, it is equally important HE is difficult to assess in children and, in fact, may to note that of 79 children with non-acetaminophen ALF never become clinically apparent in the setting of ALF.3 who never had clinically detectable HE, death (8/79) or liver However, coagulopathy is an independent risk factor for transplant (8/79) occurred in 20%. Our data support a defi- death or need for liver transplantation in ALF.17 Therefore nition of pediatric ALF that does not require HE. we chose to include children without HE in our study, but An indeterminate cause of ALF was assigned to 54% of only when a significant uncorrectable coagulopathy was children Ͻ 3 years of age and 49% overall. Factors that may present. HE remains an important predictor of outcome18 influence the intensity of the diagnostic evaluation in children

656 Squires et al The Journal of Pediatrics • May 2006 Table VIII. Logistic regression results predicting death or liver transplantation at 3 weeks Based upon Admission and Peak Measures

95.0% C.I. for OR Wald Odds Ratio Lower Upper p-value Admission values Overall Test - Coma .0012 Coma Grade 1–2 versus rest 2.83 1.55 5.16 .0007 Coma Grade 3–4 versus rest 2.96 1.22 7.16 .0160 INR Ն 2.55 2.04 1.15 3.62 .0150 Total Bilirubin Ն 5.0 mg/dL 10.69 5.49 20.85 Ͻ.0001 Model fit statistic: Hosmer-Lemeshow P ϭ .30 Peak values Overall Test for Max Coma Ͻ.0001 Max Coma 1–2 versus rest 3.60 1.95 6.66 Ͻ.0001 Max Coma 3–4 versus rest 6.92 3.40 14.07 Ͻ.0001 Max INR Ն 2.55 3.36 1.91 5.91 Ͻ.0001 Max Total Bilirubin Ն 5.0 mg/dL 8.62 4.28 17.34 Ͻ.0001 Model fit statistic: Hosmer-Lemeshow P ϭ .91 with ALF include prioritization of the etiologic possibilities, involving prolonged or inappropriate dosing, so-called “ther- blood volumes required for diagnostic studies, and the rapid apeutic misadventures,”23 are not easily captured by this study. evolution of disease to transplant or death. Thus all potential Acetaminophen-protein adducts are formed when the usual diagnostic studies were not performed on each patient. The mechanisms of acetaminophen metabolism and excretion are indeterminate group may include patients who were “under- exhausted, and the reactive acetaminophen metabolite binds evaluated” for known causes of ALF, as well as those with to important intracellular proteins, resulting in cell death. novel infectious, immune, autoimmune, metabolic, or genetic Detection of these adducts in serum may serve as a biomarker disorders. of acetaminophen toxicity.24 An infectious agent was identified in only 6% of pa- Non-acetaminophen drug-related ALF was recognized tients in this series. Herpes simplex virus and Epstein Barr only in the older age group in our series. Drug-related hep- virus were the most common identifiable infections in chil- atotoxicity is relatively common in children, particularly those dren Ͻ3 years and Ն3 years, respectively. Hepatitis A and B taking neuroleptic medications, yet ALF is rare.25,26 The are commonly associated with ALF in adults19; however, we mechanism of injury leading to ALF is believed to be an identified only 3 cases of hepatitis A, one case of hepatitis C, idiosyncratic reaction in most cases; however, children with and no cases of hepatitis B. Nevertheless, these infections are ALF related to valproic acid should be evaluated for an common causes of ALF in children living in endemic areas underlying mitochondrial disorder.27 In addition, polymor- where hepatitis A can represent up to 40% of ALF cases.20 phisms of genes associated with drug detoxification or cyto- Respiratory viruses, enterovirus, or perhaps medications used kine expressions may enhance a patient’s susceptibility to liver for symptomatic treatment of these conditions might be im- injury.28,29 plicated given the surge of cases in the winter months; how- Patient outcome was influenced by a number of factors ever, these viruses were rarely identified. including age, diagnosis, the degree of HE, and severity of the AIH presenting as ALF accounted for 6% of patients, coagulopathy. The risk of death or liver transplantation was occurred in all age groups and should therefore be considered highest among children Ͻ3 years of age. Although the num- early in the diagnostic evaluation to enable timely initiation of bers are relatively small, patients with grade IV HE at enroll- corticosteroid treatment.21 A metabolic cause for ALF was ment experienced a higher rate of spontaneous recovery than established in 18% of children Ͻ3 years of age. Unfortunately, those who progressed to grade IV during the course of the diagnostic criteria for several conditions are not well estab- study (50% vs 20%). At the same time, 20% of children who lished, and special attention to proper collection and transport never experienced clinical HE either died or received a liver of biological specimens to specialized research laboratories is transplant. Logistic regression analysis identified total biliru- needed. Wilson disease and defects or deficiencies in mito- bin Ն 5 mg/dL, INR Ն 2.55, and HE to be risk factors to chondrial function and metabolism (ie, mitochondrial hepa- predict death or liver transplantation. topathy) were the most common metabolic conditions iden- In summary, this multicenter, multinational database tified in our study. has confirmed that acetaminophen-induced ALF has an ex- Acute acetaminophen toxicity is the most common cellent outcome when HE is absent,30 demonstrated that the identifiable cause of ALF in children Ն3 years old (21%), but causes of ALF in children are age-dependent and differ from the frequency is even higher in adults (40%).22 Instances those in adults, and identified AIH as an important cause of

Acute Liver Failure In Children: The First 348 Patients In The Pediatric Acute Liver Failure Study Group 657 ALF in all ages of children. Unfortunately, most cases of ALF 14. Gregorio GV, Portmann B, Reid F, Donaldson PT, Doherty DG, in children are indeterminate. Therefore improvement in McCartney M, et al. Autoimmune hepatitis in childhood: a 20-year experi- ence. Hepatology 1997;25:541-7. diagnosis will require a focused search for treatable causes that 15. Zar JH. Biostatistical analysis. Upper Saddle River: Prentice Hall; prioritizes diagnostic conditions known to cause ALF. Newer 1999. techniques to identify children with an underlying metabolic 16. Ostapowicz G, Fontana RJ, Schiodt FV, Larson A, Davern TJ, Han disease, acetaminophen toxicity, and immune dysregulation SH, et al. Results of a prospective study of acute liver failure at 17 tertiary care will likely improve our ability to establish a diagnosis in these centers in the United States. Ann Intern Med 2002;137:947-54. seriously ill children. 17. Huo TI, Wu JC, Sheng WY, Chan CY, Hwang SJ, Chen TZ, et al. Prognostic factor analysis of fulminant and subfulminant hepatic failure in an area endemic for hepatitis B. J Gastroenterol Hepatol 1996;11:560-5. Acknowledgments available at www.jpeds.com. 18. Dhiman RK, Seth AK, Jain S, Chawla YK, Dilawari JB. Prognostic evaluation of early indicators in fulminant hepatic failure by multivariate REFERENCES analysis. Dig Dis Sci 1998;43:1311-6. 1. Riely CA. Acute hepatic failure in children. Yale J Biol Med 19. Schiodt FV, Davern TJ, Shakil AO, McGuire B, Samuel G, Lee WM. 1984;57:161-84. Viral hepatitis-related acute liver failure. Am J Gastroenterol 2003;98: 2. Russell GJ, Fitzgerald JG, Clark JH. Fulminant hepatic failure. J Pe- 448-53. diatr 1987;111:313-9. 20. Shah U, Habib Z, Kleinman RE. Liver failure attributable to hepatitis 3. Durand P, Debray D, Mandel R, Baujard C, Branchereau S, Gauthier A virus infection in a developing country. Pediatrics 2000;105:436-8. F, et al. Acute liver failure in infancy: a 14-year experience of a pediatric liver 21. Mieli-Vergani G, Vergani D. Autoimmune hepatitis in children. Clin transplantation center. J Pediatr 2001;139:871-6. Liver Dis 2002;6:335-46. 4. Psacharopoulos HT, Mowat AP, Davies M, Portmann B, Silk DB, 22. Lee WM. Acute liver failure in the United States. Semin Liver Dis Williams R. Fulminant hepatic failure in childhood: an analysis of 31 cases. 2003;23:217-26. Arch Dis Child 1980;55:252-8. 23. Heubi JE, Barbacci MB, Zimmerman HJ. Therapeutic misadventures 5. Devictor D, Desplanques L, Debray D, Ozier Y, Dubousset AM, with acetaminophen: hepatotoxicity after multiple doses in children. J Pediatr Valayer J, et al. Emergency liver transplantation for fulminant liver failure in 1998;132:22-7. infants and children. Hepatology 1992;16:1156-62. 24. Muldrew KL, James LP, Coop L, McCullough SS, Hendrickson HP, 6. Rivera-Penera T, Moreno J, Skaff C, McDiarmid S, Vargas J, Ament Hinson JA, et al. Determination of acetaminophen-protein adducts in mouse ME. Delayed encephalopathy in fulminant hepatic failure in the pediatric liver and serum and human serum after hepatotoxic doses of acetaminophen population and the role of liver transplantation. J Pediatr Gastroenterol Nutr using high-performance liquid chromatography with electrochemical detec- 1997;24:128-34. tion. Drug Metab Dispos 2002;30:446-51. 7. Whitington PF, Alonso EM. Fulminant hepatitis in children: evidence 25. Arnon R, DeVivo D, Defelice AR, Kazlow PG. Acute hepatic failure for an unidentified hepatitis virus. J Pediatr Gastroenterol Nutr in a child treated with lamotrigine. Pediatr Neurol 1998;18:251-2. 2001;33:529-36. 26. Chitturi S, George J. Hepatotoxicity of commonly used drugs: nonste- 8. Baker A, Alonso ME, Aw MM, Ciocca M, Porta G, Rosenthal P. roidal anti-inflammatory drugs, antihypertensives, antidiabetic agents, anti- Hepatic failure and liver transplant: Working Group report of the second convulsants, lipid-lowering agents, psychotropic drugs. Semin Liver Dis World Congress of Pediatric Gastroenterology, Hepatology, and Nutrition. 2002;22:169-83. J Pediatr Gastroenterol Nutr 2004;39(Suppl 2):S632-9. 27. Schwabe MJ, Dobyns WB, Burke B, Armstrong DL. Valproate-in- 9. Gyamlani GG, Parikh CR. Acetaminophen toxicity: suicidal vs. acci- duced liver failure in one of two siblings with Alpers disease. Pediatr Neurol dental. Crit Care 2002;6:155-9. 1997;16:337-43. 10. Loudianos G, Gitlin JD. Wilson’s disease. Semin Liver Dis 28. Aithal GP, Ramsay L, Daly AK, Sonchit N, Leathart JB, Alexander G, 2000;20:353-64. et al. Hepatic adducts, circulating antibodies, and cytokine polymorphisms in 11. Ojo AO, Heinrichs D, Emond JC, McGowan JJ, Guidinger MK, patients with diclofenac hepatotoxicity. Hepatology 2004;39:1430-40. Delmonico FL, et al. Organ donation and utilization in the USA. Am J 29. Watanabe I, Tomita A, Shimizu M, Sugawara M, Yasumo H, Koishi Transplant 2004;4(Suppl 9):27-37. R, et al. A study to survey susceptible genetic factors responsible for trogli- 12. Whitington PF, Alonso AE. Fulminant hepatitis and acute liver failure. tazone-associated hepatotoxicity in Japanese patients with type 2 diabetes In: DA D, editor. Paediatric liver disease. Oxford: Blackwell; 2003. p. mellitus. Clin Pharmacol Ther 2003;73:435-55. 107-26. 30. Bernal W, Wendon J, Rela M, Heaton N, Williams R. Use and 13. Rumack BH. Acetaminophen overdose in children and adolescents. outcome of liver transplantation in acetaminophen-induced acute liver fail- Pediatr Clin North Am 1986;33:691-701. ure. Hepatology 1998;27:1050-5.

658 Squires et al The Journal of Pediatrics • May 2006 Site co-investigators: Dominic Dell Olio (Birmingham, United Laura Krawczuk (Boston), Rosemary Nagy (St. Louis), Mary Kingdom), Emre Sukru (Mt. Sinai) Kay Alford (Baltimore), Stephanie Johnson (Los Angeles), Zana Parman (San Diego), Jaymee Scott (Houston). Study Coordinators: Beverly Bernard (Pittsburgh), Jeanna Zalsos DSMB Members: Michael W. Fried (Chair), John Barnard, (Mt. Sinai), Susan Krug (Cincinnati), Sue Kelly (Northwest- Katryn Furuya, Edmund A. Gehan, Caroline A. Riley, Lewis W. ern), Terri Fisher (Denver), Hazel Senz (Denver), Rachel Tay- Teperman. lor (Kings College), Lily Luu (San Francisco), Melissa Young (Seattle), Sharon Kochanowicz (Omaha), Nadia Tayeh (Ann National Institutes of Health-NIDDK: Patricia Robuck, Jay H. Arbor), Kurt Freer (Drexel), Pam Davis (Birminham, AL), Hoofnagle, Eduwad Doo, Rebecca Torrance.

Acute Liver Failure In Children: The First 348 Patients In The Pediatric Acute Liver Failure Study Group 658.e1 Table V. Final diagnosis in children with ALF Age category < 3 > 3 Diagnosis (%) (%) Total (%) Acetaminophen (n ϭ 48) 2 (2) 46 (21) 48 (14) Indeterminate (n ϭ 169) 68 (54) 101 (46) 169 (49) Autoimmune (n ϭ 22) 6 (5) 16 (7) 22 (6) Infectious (n ϭ 20) 9 (7) 11 (5) 20 (6) Adenovirus (n ϭ 2) 1 (1) 1 (0) 2 (1) Cytomegalovirus (n ϭ 1) 1 (1) 0 (0) 1 (0) Epstein-Barr virus (n ϭ 6) 1 (1) 5 (2) 6 (2) Enterovirus (n ϭ 1) 1 (1) 0 (0) 1 (0) Hepatitis A (n ϭ 3) 0 (0) 3 (1) 3 (1) Hepatitis C (n ϭ 1) 0 (0) 1 (0) 1 (0) Herpes simples virus (n ϭ 6) 5 (4) 1 (0) 6 (2) Non-APAP drug induced liver 1 (1) 16 (7) 17 (5) disease (n ϭ 17) Mushroom (n ϭ 2) 0 (0) 2 (1) 2 (1) Anesthetic (n ϭ 1) 0 (0) 1 (0) 1 (0) Bactrim (n ϭ 1) 0 (0) 1 (0) 1 (0) Cylert (n ϭ 1) 0 (0) 1 (0) 1 (0) Cytoxan/Dilantin (n ϭ 1) 0 (0) 1 (0) 1 (0) Dilantin (n ϭ 1) 0 (0) 1 (0) 1 (0) INH (n ϭ 2) 0 (0) 2 (1) 2 (1) Iron (n ϭ 1) 0 (0) 1 (0) 1 (0) Methotrexate (n ϭ 1) 0 (0) 1 (0) 1 (0) Minocycline (n ϭ 1) 0 (0) 1 (0) 1 (0) Pravastatin (n ϭ 1) 0 (0) 1 (0) 1 (0) Valproate (n ϭ 3) 1 (1) 2 (1) 3 (1) Metabolic (n ϭ 36) 23 (18) 13 (6) 36 (10) Alpha-1 antitrypsin (n ϭ 1) 1 (1) 0 (0) 1 (0) Fatty acid oxidation defect 4 (3) 0 (0) 4 (1) (n ϭ 4) Galactosemia (n ϭ 2) 2 (2) 0 (0) 2 (1) Fructose intolerance (n ϭ 1) 1 (1) 0 (0) 1 (0) Mitochondrial disorder (n ϭ 4) 2 (2) 2 (1) 4 (1) Niemann-Pick type C (n ϭ 1) 1 (1) 0 (0) 1 (0) Respiratory chain defect (n ϭ 7) 7 (6) 0 (0) 7 (2) Reyes syndrome (n ϭ 1) 0 (0) 1 (0) 1 (0) Tyrosinemia (n ϭ 4) 4 (3) 0 (0) 4 (1) Urea cycle defect (n ϭ 2) 1 (1) 1 (0) 2 (1) Wilson disease (n ϭ 9) 0 (0) 9 (4) 9 (3) Other (n ϭ 20) 11 (9) 9 (4) 20 (6) Budd-Chiari (n ϭ 2) 0 (0) 2 (1) 2 (1) Hemophagocytic syndrome 2 (2) 2 (1) 4 (1) (n ϭ 4) Leukemia (n ϭ 2) 1 (1) 1 (0) 2 (1) Neonatal iron storage disease 6 (5) 0 (0) 6 (2) (n ϭ 6) Veno-occlusive disease (n ϭ 6) 2 (2) 4 (2) 6 (2) Shock (n ϭ 16) 7 (6) 9 (4) 16 (5) Total 127 (36) 221 (64) 348 (100)

658.e2 Squires et al The Journal of Pediatrics • May 2006 Mutations in the MPV17 Gene are Responsible for Rapidly Progressive Liver Failure in Infancy

Lee-Jun C. Wong,1 Nicola Brunetti-Pierri,1 Qing Zhang,1 Nada Yazigi,2 Kevin E. Bove,2,3 Beverly B. Dahms,4 Michelle A. Puchowicz,5 Ignacio Gonzalez-Gomez,6,7 Eric S. Schmitt,1 Cavatina K. Truong,1 Charles L. Hoppel,5 Ping-Chieh Chou,1 Jing Wang,1 Erin E. Baldwin,8 Darius Adams,9 Nancy Leslie,10 Richard G. Boles,8,11 Douglas S. Kerr,5 and William J. Craigen1

MPV17 is a mitochondrial inner membrane protein of unknown function recently recog- nized as responsible for a mitochondrial DNA depletion syndrome. The aim of this study is to delineate the specific clinical, pathological, biochemical, and molecular features associ- ated with mitochondrial DNA depletion due to MPV17 gene mutations. We report 4 cases from 3 ethnically diverse families with MPV17 mutations. Importantly, 2 of these cases presented with isolated liver failure during infancy without notable neurologic dysfunction. Conclusion: We therefore propose that mutations in the MPV17 gene be considered in the course of evaluating the molecular etiology for isolated, rapidly progressive infantile hepatic failure. (HEPATOLOGY 2007;46:1218-1227.)

itochondrial disorders are established causes of basis of currently available information, most inherited liver failure in early childhood. However, the mitochondrial diseases in infants are due to mutations in Mreal incidence is likely to be underestimated nuclear genes encoding proteins with specific functions because of an underappreciation of this group of disor- targeted to the mitochondria rather than primary muta- ders, phenotypic heterogeneity, and difficulty in the in- tions in the mitochondrial genome (mitochondrial DNA vestigations required to achieve the diagnosis. On the [mtDNA]) itself. Mutations in nuclear-encoded genes involved in mtDNA replication and maintenance of mtDNA integrity, including DNA polymerase gamma Abbreviations: CS, citrate synthase; ⌬Ct, difference in the threshold cycle num- (POLG), deoxyguanosine kinase (DGUOK), thymidine bers of the nuclear gene and mtDNA; DGUOK, deoxyguanosine kinase; FeCN, ferricyanide; 1H-MRS, proton magnetic resonance spectroscopy; MRI, magnetic kinase 2 (TK2), and most recently MPV17, have emerged resonance imaging; mtDNA, mitochondrial DNA; NA, not available; NADH, as being responsible for mtDNA depletion syndromes nicotinamide adenine dinucleotide; NNH, Navajo neurohepatopathy; ORO, oil (low mtDNA copy number). MtDNA depletion typically red-O; PCR, polymerase chain reaction; POLG, DNA polymerase gamma; SD, standard deviation; SDH, succinate dehydrogenase; TK2, thymidine kinase 2; TM, causes respiratory chain dysfunction with prominent neu- transmembrane. rological, muscular, and hepatic involvement.1 Mutations From the 1Department of Molecular and Human Genetics, Baylor College of in the MPV17 gene that encodes a mitochondrial inner 2 Medicine, Houston, TX; Division of Gastroenterology, Hepatology, and Nutrition membrane protein were initially identified in 3 families and 3Division of Pathology, Cincinnati Children’s Hospital Medical Center, Cin- 2 cinnati, OH; 4Department of Pathology, and 5The Center for Inherited Disorders of with infantile hepatic mtDNA depletion. The pheno- Energy Metabolism, Rainbow Babies and Children’s Hospital, Case Western Re- type in these patients was characterized by severe liver serve University School of Medicine, Cleveland, OH; 6Department of Pathology and failure, hypoglycemia, growth retardation, neurological Laboratory Medicine, Childrens Hospital Los Angeles, Los Angeles, CA; 7University of Southern California, Keck School of Medicine, Los Angeles, CA; 8Division of symptoms, and multiple brain lesions during the first year Medical Genetics and the Saban Research Institute, Childrens Hospital Los Angeles, of life.2 Subsequently, a presumed founder MPV17 mu- Los Angeles, CA; 9Department of Pediatrics, Section of Genetics and Metabolism, tation was also identified in patients with Navajo neuro- Albany Medical Center, Albany, NY;10Division of Genetics and Metabolic Disor- ders, Cincinnati Children’s Hospital Medical Center, Cincinnati, OH; and 11De- hepatopathy (NNH; MIM 256810), an autosomal partment of Pediatrics, University of Southern California, Keck School of Medicine, recessive disorder prevalent in the Navajo population of Los Angeles, CA. the southwestern United States.3,4 This disorder exhibits Received March 19, 2007; accepted May 2, 2007. severe sensory and motor neuropathy, corneal anesthesia Address reprint requests to: Lee-Jun C. Wong, Ph.D., Department of Molecular and Human Genetics, Baylor College of Medicine, One Baylor Plaza, NAB 2015, and scarring, cerebral leukoencephalopathy, failure to Houston, TX 77030. E-mail: [email protected]; fax: 713-798-8937. thrive, recurrent metabolic acidosis with intercurrent ill- Copyright © 2007 by the American Association for the Study of Liver Diseases. ness, and liver disease.3,5 Decreased respiratory chain ac- Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/hep.21799 tivity and mtDNA depletion were demonstrated in liver Potential conflict of interest: Nothing to report. biopsy samples from 2 NNH patients.6 A homozygous

1218 HEPATOLOGY, Vol. 46, No. 4, 2007 WONG ET AL. 1219

Table 1. Summary of the Clinical and Molecular Findings in Patients with Known MPV17 Mutations

Mutations Age of Lactic Liver Failure Neurological Patient Sex Ethnicity Onset Outcome Acidosis Hypoglycemia Failure To Thrive Symptoms Allele 1 Allele 2

This study M Middle Eastern 2 months Death at 5 months ϩϩϩϩϩW69X W69X (case 1) This study M Middle Eastern Birth Liver transplant; ϩϪϩϩϪW69X W69X (case 2) death at 6 months This study F Hispanic 8 months Death at 19 ϩϪϩϩϩR50W R50W (case 3) months This study M Caucasian 1 month Death at 3 months ϩϩϩϩϪc.263_265del3 c.234_242del9 (case 4) Spinazzola M Caucasian NA Death before 12 NA NA ϩ NA NA R50Q R50Q et al.2 months Spinazzola M Caucasian NA Liver transplant; NA ϩϩϩϪR50Q R50Q et al.2 living at 4 years Spinazzola F Caucasian NA Living at 9 years NA ϩϪϩϩR50Q R50Q et al.2 Spinazzola M Moroccan NA Death in the first NA NA ϩ NA NA N166K N166K et al.2 months of life Spinazzola F Caucasian NA Death in the first NA NA ϩ NA NA c.116-141del25 R50W et al.2 months of life Karadimas M Navajo 1-12 months Death by cirrhosis ϩϩϩϩϩR50Q R50Q et al.4 at 16 years Karadimas F Navajo 6 months Death by cirrhosis ϩϩϩϩϩR50Q R50Q et al.4 at 20 years Karadimas M Navajo 2-3 years Death by cirrhosis ϩϪϩϩϩR50Q R50Q et al.4 at 15 years Karadimas F Navajo 6 months Liver transplant; ϩϪϩϩϩR50Q R50Q et al.4 living at 12 years Karadimas F Navajo 1 month Liver transplant; ϪϪϩϩϩR50Q R50Q et al.4 death at 2 years by sepsis Karadimas F Navajo 4 months Hepatocellular ϪϩϩϩϩR50Q R50Q et al.4 carcinoma; liver transplant; living at 21 years

F indicates female; M, male; and NA, not available.

R50Q mutation in the MPV17 gene was identified in all The clinical findings of 4 cases caused by MPV17 mu- patients with NNH.4 tations are summarized in Table 1 and are compared with We describe the clinical, biochemical, pathological and the findings from patients with documented MPV17 mu- molecular features of 4 additional patients from 3 ethni- tations reported in the literature.2,4 cally distinct families with MPV17 mutations. All these Case 1. The patient was the first child of healthy con- cases presented with rapidly progressive liver failure dur- sanguineous parents of Middle Eastern descent. The fam- ing infancy. Importantly, in 2 patients, neurological dys- ily history was significant for liver disease in 2 members of function was not noted. the family and for premature deaths in infancy of 2 other individuals (Fig. 1). He presented on day 2 of life with Patients and Methods focal seizures, cyanosis, and apnea, which was presumed Patients to be secondary to a documented hypoglycemic episode. Samples from the 4 patients included in this study were A head computed tomography at that time revealed an referred to the Mitochondrial Diagnostics Laboratory at the ischemic area in the distribution of the left middle cere- Baylor College of Medicine for biochemical and molecular bral artery. At the age of 2 months, the patient was noted studies. The initial workup included biochemical assays of to have poor weight gain, irritability, and jaundice. On respiratory chain function, determination of the mtDNA physical examination, he had hepatosplenomegaly. The copy number, and sequence analysis of the DGUOK and neurological examination at that time revealed age-appro- POLG1 genes. Because mutations were not detected in ei- priate muscle tone. Laboratory findings included low al- ther of these 2 genes despite abnormal electron transport bumin, high conjugated bilirubin, mild increases in chain activities and reduced mtDNA content, we searched aminotransferases, elevated blood lactate of 6.4 mmol/L for mutations in the MPV17 gene, which was recently re- (normal: 0.7-2.1), increased plasma methionine and ty- ported to be responsible for a previously unrecognized he- rosine (15-fold and 10-fold above the upper limit of the patic mtDNA depletion syndrome.2 normal range, respectively), and a 2-3–fold increase of 1220 WONG ET AL. HEPATOLOGY, October 2007

Fig. 1. Pedigree and MPV17 sequence analysis of the presented cases. Squares represent males, circles represent females, and diamonds indicate that sex information is not available. A closed symbol indicates that the individual was affected with the disease, and strike-through indicates that the individual expired. Arrows indicate the patients studied. Double horizontal lines indicate consanguineous mating. glutamine and alanine. There was evidence of a coagu- tor development and neurological examination were nor- lopathy with reduced antithrombin III, protein C, and mal, and there was no evidence of other organ system protein S. The coagulopathy and cholestasis progressed in involvement. Brain MRI, 1H-MRS, an electroencephalo- parallel throughout the course of the disease, with inter- gram, retinal fundoscopy, and an echocardiogram were all mittent spontaneous improvements lasting days to weeks, normal. The patient received a liver allograft at the age of until the final irreversible decline in liver function and 5 months but died 1 month later following multiorgan death at the age of 5 months. With the exception of the failure associated with sepsis. residual changes related to the previously noted infarct, Case 3. This patient was the daughter of an indige- the brain magnetic resonance imaging (MRI) was consid- nous language–speaking Hispanic couple from an iso- ered normal, and proton magnetic resonance spectros- lated region in Mexico. She presented at the age of 8 copy (1H-MRS) did not show an elevated lactate peak. months with failure to thrive and at 14 months with hep- The brain autopsy showed a single vessel thrombosis without evidence of demyelination. atomegaly, moderate elevations in aminotransferases Case 2. This infant was the younger brother of case 1. (aspartate aminotransferase, 352 U/L, and alanine amino- He was found by newborn screening to have increased transferase, 155 U/L), and hyperbilirubinemia (total 9.1 plasma tyrosine, and there was evidence of liver disease mg/dL, conjugated 5.6 mg/dL). At 18 months, she devel- from the newborn period. By 10 weeks of age, he exhib- oped rapidly progressive liver insufficiency (prothrombin ited failure to thrive, jaundice, elevated aminotrans- time of 36.2 seconds, with albumin level of 3.0 g/dL). ferases, and a coagulopathy refractory to treatment. Liver Hypotonia and loss of previously attained skills were function rapidly deteriorated, with progressive hepato- noted, as the previously developmentally normal patient megaly and cholestasis. He had lactic acidosis that peaked became increasingly encephalopathic and minimally in- at 9 mmol/L. Notably, up to 5 months of age, psychomo- teractive. She had lactic acidosis, with blood lactic acid at HEPATOLOGY, Vol. 46, No. 4, 2007 WONG ET AL. 1221

9 mmol/L. Five days prior to her death from liver failure Respiratory Chain Enzyme Analysis at the age of 19 months, brain MRI revealed cytotoxic Spectrophotometric analysis of the respiratory chain edema involving the deep and subcortical white matter. complexes was performed on liver samples from patients 2 The 1H-MRS revealed glutamine and lactate at 3 and 10 (explanted liver), 3, and 4 and on muscle (vastus lateralis times the upper limit of normal. of quadriceps femoris) from patients 1 and 4. The liver Case 4. The patient was the child of healthy, unre- and muscle samples were immediately frozen in liquid lated Caucasian parents. He developed hypoglycemia and nitrogen after collection, stored at Ϫ80°C, and shipped lactic acidosis on day 1 of life. At 4 weeks of age, he had on dry ice. For patient 4, mitochondria were isolated from persistent lactic acidemia (4.9 mmol/L) and hypoalbu- a fresh muscle sample for polarographic assays of oxidative minemia. At 6 weeks of age, he developed recurrent phosphorylation. Respiratory chain complexes were also episodes of hypoglycemia requiring large doses of in- assayed from the fresh homogenate.7 Additional muscle travenous glucose. An echocardiogram and chest radio- and liver specimens were obtained from patient 4 post graph were reported as normal. Plasma tyrosine was mortem within 2 hours of death. increased. Urinary p-hydroxyphenylacetic, p-hydroxy- The electron transport chain enzymes were assayed at phenyllactic, lactic, and short-chain dicarboxylic acids 30°C with a temperature-controlled spectrophotometer were elevated without detectable succinylacetone. (Ultraspec 6300 Pro, Biochrom, Ltd., Cambridge, En- Liver failure developed and rapidly progressed, with gland). Each assay was performed in duplicate. The activ- elevated aminotransferases, extremely high ␣-fetopro- ities of nicotinamide adenine dinucleotide (NADH): tein (157,000 ␮g/L), a coagulopathy and hematemesis, ferricyanide (FeCN) reductase, succinate dehydrogenase (SDH), rotenone sensitive complex IϩIII (NADH:cyto- jaundice, hypoalbuminemia, ascites, edema, decreased ϩ urine output, and severe lactic acidosis, with blood chrome c reductase), complex II III (succinate:cyto- lactate up to 28 mmol/L. The child died at 3 months of chrome c reductase), and complex IV (cytochrome c age from progressive liver failure. At autopsy, the most oxidase) were measured with appropriate electron accep- tors and donors.8,9 The increase or decrease in the absor- striking finding was a very enlarged (285 g versus ex- bance of cytochrome c at 550 nm was measured for pected 136 g), yellow, fatty liver. complex IϩIII, complex IIϩIII, or complex IV. The ac- tivity of NADH:FeCN reductase was measured by the Tissue Histological and Electron Microscopy Studies oxidation of NADH at 340 nm. For SDH, the reduction All samples were taken after parental informed con- of 2,6-dichloroindophenol at 600 nm was measured. Ci- sent. Liver specimens were obtained through open biopsy trate synthase (CS) was used as a marker for the mito- in cases 1, 3, and 4 and through explantation in case 2. chondrial content. Enzyme activities are expressed with Patients 3 and 4 died shortly after the biopsy was per- respect to both the total protein and CS activity. formed. The liver specimens were fixed in 10% buffered formaldehyde, paraffin-embedded, and stained with he- Molecular Studies matoxylin and eosin, trichrome, reticulin, and periodic Total nuclear DNA and mtDNA were extracted from acid–Schiff diastase and for iron. A fresh frozen sample peripheral blood leukocytes and skin fibroblast cultures was stained with oil red-O (ORO) for lipids in cases 2 and with commercially available DNA isolation kits (Gentra 4. Case 4 had a muscle biopsy that was subjected to a Systems Inc., Minneapolis, MN). The mtDNA copy histochemical study using the following methods: nico- number in various tissues was measured by the real-time tinamide adenine dinucleotide–tetrazolium reductase, quantitative polymerase chain reaction (PCR) method ATPase at both pH 9.4 and pH 4.3, cytochrome c oxi- with specific primers for the mitochondrial transfer dase, and ORO. For electron microscopy, biopsy tissues RNALeu(uur) gene (mtF3212-3231 and mtR3319-3300; were fixed in 2% or 3% buffered glutaraldehyde, post- see http://www.mitomap.org).10 As a control, a nuclear fixed in 1% osmium tetroxide, dehydrated in graded con- single-copy gene, ␤-2-microglobulin, was amplified with centrations of ethyl alcohol followed by propylene oxide, primers ntF 589-613 and ntR 674-653 in the 3ЈUTR and embedded in Epon 1-␮ semithin sections, which (GI:37704390).10 The real-time quantitative PCR was were stained with methylene blue and azure II and coun- performed twice for each reaction. The 10-␮L PCR con- terstained with sodium borate and basic fuchsine for light tained 5X iTaqSYBR Green Supermix with 5-carboxy-X- microscopy orientation. Ultrathin sections were stained rhodamine (Bio-Rad, Hercules, CA), 300 nmol/L of each with uranyl acetate and Sato’s lead citrate, mounted on primer, and 0.1-1.0 ng of the total genomic DNA extract. copper grids, and examined with a Phillips CM-12 or The real-time PCR conditions were 2 minutes at 50°C Zeiss 912 transmission electron microscope. and 10 minutes at 95°C, followed by 45 cycles of 15 1222 WONG ET AL. HEPATOLOGY, October 2007 seconds of denaturation at 95°C and 30 seconds of an- the appearance in paraffin sections. Canalicular and cyto- nealing/extension at 63°C. The fluorescent signal inten- plasmic bile stasis was common. The prevalence of hepa- sity of the PCR products was recorded and analyzed on a tocyte nuclear immunoreactivity for MIB-1 (Ki67 7900HT fast real-time PCR system (Applied Biosystems, monoclonal antibody) supported the evidence of regener- Foster City, CA) with SDS version 2.2.2 software. The ative activity, such as scattered oval-shaped regenerative mtDNA content, called the relative mtDNA copy num- hepatocytes (Fig. 2K). Isolated hepatocyte necrosis was a ber, was calculated from the difference in the threshold minor feature in all but case 3. Degenerating and necrotic cycle numbers of the nuclear gene and mtDNA (⌬Ct). hepatocytes lacking obvious lipid vacuoles exhibited un- The amount of mtDNA per cell equals 2(2Ϫ⌬Ct), account- usual coarse eosinophilic cytoplasmic granularity. In gen- ing for the 2 copies of the ␤-2-microglobulin gene in each eral, liver inflammation was minimal to absent, and the cell nucleus. Through a comparison of the mtDNA con- liver architecture was characterized by delicate periportal tent of the patient to that of the control, mtDNA deple- and pericellular fibrosis coexisting with areas of intralob- tion could be determined. For muscle, the control value ular collapse (Fig. 2A,D,F,G). was derived from a pool of 12 samples from age-matched Ultrastructural studies revealed nonspecific changes individuals, and for liver, the control value was from a related to cholestasis. In most hepatocytes, increased pool of 3 age-matched liver specimens. residual bodies, decreased glycogen, increased ran- Sequence-specific oligonucleotide primers linked to domly scattered rough endoplasmic reticulum (often M13 universal primers were designed to amplify all cod- dissociated from mitochondria), and abundant non– ing exons of the MPV17 gene and at least 50 nucleotides membrane-bound minute lipid droplets were ob- of the flanking introns. PCR products generated with Fast served. Mitochondrial abnormalities were easily Start DNA polymerase (Roche, Indianapolis, IN) were detected in almost all hepatocytes and included in- purified on ExcelaPure 96-well ultrafiltration PCR puri- creased numbers per cell, abnormal variations in the fication plates (Edge BioSystems, Gaithersburg, MD). Se- size and shape, distorted tubular dilatation of the cris- quencing reactions were performed with the BigDye tae, and absence of dense matrix granules (Fig. Terminator cycle sequencing kit (version 3.1), purified 2B,C,E,H). The accumulation of an abnormally dense with Performa DTR 96-well V3 short plates (Edge Bio- matrix with the displacement of the cristae was a prom- Systems), and analyzed on an ABI3730XL automated inent focal feature (Fig. 2C,H). Megamitochondria DNA sequencer with Sequencing Analysis Software ver- due to a massive accumulation of the matrix were sion 5.1 (Applied Biosystems). DNA sequences were an- found only in case 3 (Fig. 2H). There were no paracrys- alyzed with Mutation Surveyor version 2.6.1, and the talline inclusions in the matrix or in the intracristal MPV17 Genebank sequence (NT_022184 gene ID space. Interestingly, the mitochondria in other cell 4358) was used as the reference. types within the liver and in the muscle (cases 1 and 4) had a normal structure (data not shown). The brain Results was examined at autopsy in 2 patients (cases 1 and 4). Histopathological Studies Both had Alzheimer type II astrocytosis, which is con- The liver samples for histology (4 cases) and electron sistent with hepatic encephalopathy (data not shown). microscopy (3 cases) were reviewed. Light microscopic Patient 1 also had an old ischemic infarct in the left features at the time of tissue collection were similar in all temporal lobe. The size, character, and advanced age of samples and were typical for liver failure due to a mito- this lesion were consistent with a major arterial occlu- chondriopathy (Fig. 2). In all cases, hepatocyte cytoplasm sion during the perinatal period. There were no fea- was severely expanded and contained coarse, evenly dis- tures of mitochondrial encephalopathy. persed eosinophilic granules, numerous small lipid vacu- oles, or both (Fig. 2A,F,I,J). The largest granules Biochemical Analysis measured up to 4-5 ␮m in diameter and most likely rep- The activities of mitochondrial enzymes, including re- resented abnormal mitochondria (Fig. 2A,K). Distinctive spiratory chain complexes, are summarized in Table 2. In bright eosinophilic granules not associated with lipid liver, the CS, NADH:FeCN reductase, and SDH activi- vacuoles were seen in a minority of the hepatocytes in 3 ties, which are not encoded by mtDNA, were found to be cases and were very prevalent in case 3 (Fig. 2A,J). The increased in all 3 cases studied, in agreement with mito- ORO staining showed most hepatocytes contained lipid chondrial proliferation, whereas complexes I/III and IV, droplets of various sizes, from extremely fine to medium- which include subunits encoded by mtDNA, were found diameter vacuoles (data not shown). The number of hepa- to be deficient in all 3 patients after correction for in- tocytes containing lipid exceeded expectations based on creased CS activity (Table 2). From the muscle biopsy of Fig. 2. Hepatic histology and ultrastructure in 4 cases of MPV17 disease. Case 1. (A) Hepatocytes are swollen, with coarsely granular cytoplasm in which vacuoles are focally visible. Several hepatocytes contain bright eosinophilic granules (arrows) and no visible lipid vacuoles. (B) Hepatocyte cytoplasm contains mildly pleomorphic mitochondria with dilated cristae amid numerous minute lipid droplets. A higher magnification shows a detail of the mitochondrial cristal dilatation (inset). (C) Striking disparities in the matrix density in adjacent mitochondria and a lack of dense matrix granules accompany abnormal profiles of cristae. Case 2. (D) Swollen hepatocytes with coarse cytoplasmic granules are more common than hepatocytes with visible vacuoles. P indicates portal. (E) Mitochondrial abnormalities in 2 adjacent hepatocytes are subtle, including mild pleomorphism with cytoplasmic invaginations (inset), dilated cristae, and absent matrix granules. Case 3. (F) Swollen granular hepatocytes contain few visible lipid vacuoles and mold neighboring cells. Note the scattered foci of interlobular collapse (arrow). C indicates central vein. (G) Periportal, pericentral, and pericellular fibrosis. (H) Gigantic megamitochondria (*) have markedly expanded the matrix, which displaces dilated cristae to the periphery. A detail of mildly enlarged pleomorphic mitochondria with dilated cristae is shown (inset). Case 4. (I) Side by side are hepatocytes with prominent small lipid droplets and others with nonvacuolated, swollen coarsely granular cytoplasm. (J) A hepatocyte with bright coarse eosinophilic granules (arrow) is located to the left of a canalicular bile plug. (K) Amid coarsely granular and very finely vacuolated hepatocytes, there are 2 small cells with oval nuclei (arrow), which are interpreted as regenerating hepatocytes. (A,D,F,I,J,K) Hematoxylin and eosin stain. (G) Trichrome stain. (B,C,E,H) Electron micrographs. 1224 WONG ET AL. HEPATOLOGY, October 2007

Table 2. Mitochondrial Enzyme Activities in Patients with MPV17 Mutations

Activity of the Enzymatic Complex*

Tissue and NADH:FeCN Subject CS Activity Reductase I؉III SDH II؉III IV

Liver Case 2 155 381 (154; 63) 19.2 (19; 8) 31.5 (179; 73) 11 (93; 38) 7.7 (29; 12) Case 3 366 765 (310; 53) 48.3 (47; 8) 87.6 (497; 85) NA 18.8 (71; 12) Case 4 193 760 (308; 100) 61.8 (60; 19) 35.6 (202; 66) 17.6 (149; 48) 25.5 (97; 31) Controls (n ϭ 10) Mean Ϯ SD 62.7 Ϯ 25.8 247 Ϯ 127 188 Ϯ 58.7 17.6 Ϯ 5.3 11.8 Ϯ 3.2 26.4 Ϯ 10.8 Range 31.6–117 58.5–438 84.5–281 11.1–26.5 4.8–15.7 7.5–42.8 Muscle Case 1 3.49 3.39 (16; 45) 0.27 (11; 29) 0.09 (7; 20) 0.51 (43; 118) 0.31 (9; 25) Case 4 2.57 0.75 (4; 15) 0.54 (21; 78) 0.27 (22; 81) 0.21 (18; 67) 0.16 (5; 19) Controls (n ϭ 10) Mean Ϯ SD 9.6 Ϯ 4.8 20.7 Ϯ 8.7 2.5 Ϯ 1.3 1.2 Ϯ 0.3 1.2 Ϯ 0.5 3.4 Ϯ 1.4 Range 2.9–17 5.8–35.5 0.6–4.5 0.9–1.5 0.3–1.9 1–5

NA indicates not available, and SD, standard deviation. *The activities are given as micromoles per minute per gram of wet weight for muscle and as nanomoles per minute per milligram of protein for liver. The first number in the parentheses is the percentage of the control mean, and the second number in the parentheses is the percentage of the control mean corrected to CS activity. Case 4 muscle was postmortem tissue. The other tissues were all biopsies. case 1 and the postmortem muscle sample of case 4, all 3). Although only measured in 1 patient (case 2), the enzyme activities, including CS, were reduced, indicating reduction of the mtDNA content in blood was found to defects in mitochondrial biogenesis or a generalized tissue be mild (50% of the age-matched mean). deterioration, although after correction to reduced CS, Deleterious mutations in the MPV17 gene were de- complex IV remained deficient (Table 2). However, in tected in all 4 cases (Fig. 1 and Table 1). Three novel fresh muscle obtained from case 4 a few hours before mutations were identified in 3 patients (cases 1, 2, and death, the activities of respiratory chain complexes I (ro- 4). Case 3 was found to harbor a homozygous R50W tenone sensitive), II, III, and IV and CS, NADH:FeCN mutation that has previously been reported.2 The reductase, SDH, carnitine palmitoyl transferases I and II, R50W mutation changes the basic amino acid arginine and enzymes of fatty acid oxidation were all within the to the hydrophobic aromatic amino acid tryptophan. normal control range. Mitochondria isolated from the This mutation is predicted to be more drastic than the same specimen showed reduced oxidation of tetramethyl- arginine-to-glutamine (R50Q) change observed in enephenyldiamine plus ascorbate (52% of the control NNH,4 as supported by the observation that in a yeast mean), which suggested reduced function of at least com- model devoid of the MPV17 ortholog, partial pheno- plex IV in the intact muscle mitochondria. typic correction was noted upon transfection with a con- Molecular Analysis struct containing R50Q, but not with R50W.2 The The mtDNA content was measured in all available previously unreported W69X mutation was observed in pa- tissues (Table 3). The liver mtDNA content in patients 2, tients 1 and 2 and is expected to be severe because the trun- 3, and 4 was severely reduced to 3.9%, 4.5%, and 3% of cated mutant protein, if stably synthesized, is missing two- the mean of the age and tissue-matched controls, respec- thirds of the C-terminal portion (Fig. 3). Patient 4 was found tively. The mtDNA content in muscle was also reduced, to be a compound heterozygote for 2 in-frame dele- but not as severely as in liver: 12.6% and 8.4% of the tions (c.263_265del3 and c.234_242del9) of 3 and 9 matched controls for patients 1 and 4, respectively (Table nucleotides.

Table 3. Quantification of mtDNA in Patients with MPV17 Mutations

Patient 1 2 3 4 Controls

Liver mtDNA NA 148 (3.9%) 171 (4.5%) 113 (3%) 3787 Muscle mtDNA 430 (12.6%) NA NA 267 (8.4%) 3197 Blood mtDNA NA 157 (50%) NA NA 312

The results are expressed as the copy number per haploid nuclear gene. The percentage of the control is shown in parentheses. The control value of the liver mtDNA copy number was obtained through the pooling of 3 age-matched liver specimens. The control value of the muscle mtDNA copy number was obtained through the pooling of 12 age-matched muscle samples. NA indicates not available. HEPATOLOGY, Vol. 46, No. 4, 2007 WONG ET AL. 1225

Fig. 3. Schematic representation of the MPV17 protein based on bioinformatic pre- diction models (http://www.pir.uniprot.org/ cgi-bin/upEntry?idϭQ53SY2) and relative localization of all the mutations so far re- ported. The novel mutations identified in this study are shown in red. The inner mi- tochondrial membrane (green) lies between the mitochondrial matrix and the inter- mitochondrial membrane.

Discussion and have been previously observed in other forms of mtDNA depletion due to molecular defects in other nu- Mitochondrial respiratory chain disorders are often clear genes such as DGUOK13 and POLG (Alpers syn- recognized on the basis of multisystemic involvement that drome).14 The ultrastructure of MPV17-deficient typically includes neurological findings. Liver complica- mitochondria is characterized by dilated and distorted tions are typically considered a late feature of a multisys- cristae and by central accumulation of the matrix, which tem mitochondrial disorder in which neuromuscular displaces the cristae toward the periphery of the organelle disease is prominent. In contrast, the clinical presentation and causes enlargement in some instances to an extreme of our 4 cases of mitochondrial disease due to MPV17 degree. There is also diminished prominence of matrix mutations was dominated by rapidly progressive liver fail- granules along with a disturbance of the normally inti- ure in infancy, and neuromuscular disease was not appre- mate relationship of a rough endoplasmic reticulum with ciated (cases 2 and 4), was attributed to a neonatal stroke mitochondria. In aggregate, these changes also are not (case 1), or was of lesser severity than the hepatic disease (case 3). unique for MPV17-deficient hepatocytes but constitute a In the past few years, there has been a substantial in- pattern that is highly suggestive of a respiratory chain 13,15 crease in the understanding of the molecular basis of disorder. Thus, our detailed pathological examina- mtDNA depletion syndromes. There are at least 3 nuclear tion does not support the existence of a specific patholog- genes currently known to be responsible for liver mtDNA ical hallmark for MPV17 hepatopathy. depletion: POLG,11 DGUOK,12 and the recently discov- As shown in cases 1 and 4, histochemical, ultrastruc- ered MPV17.2 Given the paucity of cases, there are still tural, and respiratory chain studies of skeletal muscle can unanswered questions regarding MPV17 mitochondrial be uninformative, whereas similar assays on liver samples hepatopathy, including specific clinical, pathological, and appear more helpful (Table 2). The observation that biochemical features and appropriate testing strategies. mtDNA is depleted in both muscle and liver tissue is From a clinical perspective, similarly to the cases re- diagnostically useful and fairly specific, yet why both tis- ported by Spinazzola et al.,2 the onset of liver disease in sues exhibit mtDNA depletion but not similar enzymo- our patients occurred early, primarily in the first year of logic deficits remains a puzzle. life. The presence of lactic acidosis and hypoglycemia are To date, only 4 mutations in the MPV17 gene have useful diagnostic features, as is the detection of elevated been reported (Table 1 and Fig. 3). The R50Q mutation tyrosine by expanded newborn screening. However, these has been observed in NNH4 and in 1 Caucasian family.2 findings are very nonspecific. A previously reported mutation at the same amino acid With respect to the pathologic features of MPV17- position, the R50W mutation,2 was seen also in 1 of our associated liver disease, mixed macrovesicular and mi- cases, and this suggests that this codon, which includes a crovesicular steatosis and cholestasis have been reported CpG dinucleotide, may be a hot spot for mutations. The in the livers of patients with NNH and liver failure prior R50Q mutation appears to be associated with longer sur- to the discovery of the causative role of the MPV17 gene.6 vival, and the onset of the liver failure may be preceded by At the histological level, all 4 of our cases exhibited similar neurological involvement (Table 1). In contrast, all the features, including swollen granular hepatocytes, mi- other MPV17 mutations are associated with a more severe crovesicular steatosis, and focal pericellular and periportal and rapidly progressive disease. We have also detected 3 fibrosis. Taken together, these findings are not specific previously unreported mutations: W69X and 2 deletions 1226 WONG ET AL. HEPATOLOGY, October 2007

(c.263_265del3 and c.234_242del9; Fig. 3). R50 and stroke that, on the basis of autopsy findings, was incon- W69 are 2 of the 6 invariant amino acids in MPV17 that sistent with mitochondrial dysfunction. It is plausible that are conserved in yeast, fish, flies, frogs, mice, and humans. the stroke may have been the result of an altered antico- Molecular modeling of the MPV17 protein has revealed agulation cascade due to liver disease because, similar to that this 176 amino acid protein is predicted to contain 4 that of his sibling, his newborn screen was clearly abnor- transmembrane (TM) spans: TM1 from amino acid 18- mal and suggestive of a perinatal onset of liver dysfunc- 38, TM2 from 53-73, TM3 from 94-114, and TM4 from tion. Although neurological deterioration would likely 131-151, with short flanking hydrophilic intermembrane have occurred in our patients had they survived through and matrix regions (Fig. 3). The R50 residue is located infancy, our experience highlights the importance of sus- within the intermitochondrial membrane space. The pecting an mtDNA depletion syndrome even in the ab- W69 residue is located within the second TM span (Fig. sence of involvement of other systems. 3). The W69X mutation is expected to be severe because In an era when liver transplantation is available for liver the truncated mutant protein is missing two-thirds of the failure, mitochondrial cytopathies represent a special C-terminal portion of the MPV17 protein, and this allele challenge. With scarce organ availability, poor outcomes may well undergo nonsense-mediated decay of the mes- in those patients undergoing transplantation, and a lack senger RNA. The in-frame deletion c.234_242del9 of effective medical treatment for the underlying mito- (p.79_81del3aa) in patient 4 occurred in the region be- chondrial disorder, transplantation is usually contraindi- tween TM2 and TM3 and led to the deletion of a putative cated in the presence of multiorgan involvement from a protein kinase C phosphorylation site (Fig. 3). mitochondrial disorder. Although there has been no sys- From the small number of patients described so far, it tematic evaluation of liver transplantation in such pa- is clear that MPV17 mutations lead to mtDNA depletion. tients, it appears that transplantation does not improve However, the specific function of the MPV17 protein survival and may in fact precipitate multiorgan deteriora- remains unknown. The mouse model of MPV17 defi- tion because of the added stress associated with the pro- ciency generated more than 10 years ago has not been cedure. However, this may not be true in the case of instructive in elucidating the role of this gene in human mitochondrial disorders with predominant liver involve- disease.16 In contrast to humans, the absence of Mpv17 in ment. There are reports of at least 4 cases of MPV17 the mouse is compatible with survival to adulthood and, hepatopathy who received liver transplantation with rel- with the exception of an age-dependent hearing loss, is atively acceptable outcomes.2,4 The inability to predict without significant clinical problems, at least in a con- the timing and the severity of extrahepatic organ involve- trolled laboratory environment.17 ment remain major issues in the management of these The biosynthetic and detoxifying properties of the liver diseases. are highly dependent on adenosine triphosphate, and it is The differing time of onset and severity of neurological possible that liver failure may be triggered by environ- deterioration in reported cases of MPV17 mitochondrial mental agents that MPV17-deficient hepatocytes are not cytopathy merit frank discussion in the course of evaluat- able to detoxify. Thus, Mpv17Ϫ/Ϫ mice may not exhibit ing a child for transplantation candidacy. As we experi- liver damage because they are not exposed to these envi- enced in a subset of our patients, neurologic involvement ronmental factors in a laboratory setting. can clearly be absent at the initial presentation of liver Given its localization within the mitochondrial inner failure. When one is faced with a mitochondrial disorder, membrane and the similarity with the hepatic phenotype clinical deterioration due to the stress of transplantation, caused by DGUOK and TK2 mutations, it is conceivable infections, and bleeding should also be emphasized and that MPV17 participates in maintaining the deoxyribo- anticipated if transplantation is considered. Unfortu- nucleotide pool necessary for mtDNA synthesis. In fact, it nately, therapeutic support of mitochondrial functions in has been shown that mtDNA integrity is severely affected those instances is still not available. A better understand- in patients with a deficiency of succinyl-CoA synthetase, ing of the role of MPV17 in mitochondrial functions which is tightly associated in a complex with nucleoside could shed light on the pathogenesis of the disorder and diphosphate kinase, an enzyme that is crucial in maintain- potentially inform new treatments. ing the homeostasis of ribonucleotides and deoxyribonu- Finally, our first 2 patients showed evidence of liver cleotides.18,19 disease in the expanded newborn screen. As we develop Our cases illustrate that liver failure may be the only more knowledge about this disorder, early detection and presenting feature at the time of disease onset, as reflected possible early treatment might delay or prevent entirely by the observation that 2 of our 4 cases initially exhibited the development of severe liver disease and improve out- normal neurological examinations. Case 1 had a neonatal comes. With the increasing availability of appropriate HEPATOLOGY, Vol. 46, No. 4, 2007 WONG ET AL. 1227

DNA diagnostics, prenatal testing for the causative gene 7. Purchowicz MA, Varnes ME, Cohen BH, Friedman NR, Kerr DS, Hoppel can be offered to affected families. CL. Oxidative phosphorylation analysis: assessing the integrated func- tional activity of human skeletal muscle mitochondria—case studies. Mi- In summary, we report the occurrence of novel tochondrion 2004;4:377-385. MPV17 mutations in 3 families from different ethnic 8. Vu TH, Sciacco M, Tanji K, Nichter C, Bonilla E, Chatkupt S, et al. groups. Liver failure occurred in 2 of the 4 patients in the Clinical manifestations of mitochondrial DNA depletion. Neurology 1998;50:1783-1790. absence of neurological findings, and therefore we recom- 9. Enns GM, Hoppel CL, DeArmond SJ, Schelley S, Bass N, Weisiger K, et mend that testing for mtDNA depletion be considered in al. Relationship of primary mitochondrial respiratory chain dysfunction to infants and children with isolated liver failure. Quantita- fiber type abnormalities in skeletal muscle. Clin Genet 2005;68:337-348. tive testing of the mtDNA content provides an attractive 10. Bai RK, Wong LJ. Simultaneous detection and quantification of mito- chondrial DNA deletion(s), depletion, and over-replication in patients approach to establishing a diagnosis, but its sensitivity and with mitochondrial disease. J Mol Diagn 2005;7:613-622. specificity remain unknown. The availability of DGUOK, 11. Naviaux RK, Nguyen KV. POLG mutations associated with Alpers’ syn- POLG, and MPV17 molecular analyses can help establish drome and mitochondrial DNA depletion. Ann Neurol 2004;55:706-712. 12. Mandel H, Szargel R, Labay V, Elpeleg O, Saada A, Shalata A, et al. The the etiology of liver disease in a timely fashion in acutely ill deoxyguanosine kinase gene is mutated in individuals with depleted hepa- patients. tocerebral mitochondrial DNA. Nat Genet 2001;29:337-341. 13. Mandel H, Hartman C, Berkowitz D, Elpeleg ON, Manov I, Iancu TC. Acknowledgment: We thank Dr. Mark Cohen and The hepatic mitochondrial DNA depletion syndrome: ultrastructural Dr. Gretta Jacobs, Case Western Reserve University, for changes in liver biopsies. HEPATOLOGY 2001;34:776-784. their assistance with the pathological descriptions. We 14. Simonati A, Filosto M, Savio C, Tomelleri G, Tonin P, Dalla Bernardina appreciate the cooperation of the families of these patients B, et al. Features of cell death in brain and liver, the target tissues of in supporting these investigations. progressive neuronal degeneration of childhood with liver disease (Alpers- Huttenlocher disease). Acta Neuropathol (Berl) 2003;106:57-65. 15. Bioulac-Sage P, Parrot-Roulaud F, Mazat JP, Lamireau T, Coquet M, References Sandler B, et al. Fatal neonatal liver failure and mitochondrial cytopathy 1. Dimauro S. Mitochondrial medicine. Biochim Biophys Acta 2004;1659: (oxidative phosphorylation deficiency): a light and electron microscopic 107-114. study of the liver. HEPATOLOGY 1993;18:839-846. 2. Spinazzola A, Viscomi C, Fernandez-Vizarra E, Carrara F, D’Adamo P, 16. Zwacka RM, Reuter A, Pfaff E, Moll J, Gorgas K, Karasawa M, et al. The Calvo S, et al. MPV17 encodes an inner mitochondrial membrane protein glomerulosclerosis gene Mpv17 encodes a peroxisomal protein producing and is mutated in infantile hepatic mitochondrial DNA depletion. Nat reactive oxygen species. EMBO J 1994;13:5129-5134. Genet 2006;38:570-575. 17. Meyer zum Gottesberge AM, Reuter A, Weiher H. Inner ear defect similar 3. Holve S, Hu D, Shub M, Tyson RW, Sokol RJ. Liver disease in Navajo to Alport’s syndrome in the glomerulosclerosis mouse model Mpv17. Eur neuropathy. J Pediatr 1999;135:482-493. Arch Otorhinolaryngol 1996;253:470-474. 4. Karadimas CL, Vu TH, Holve SA, Chronopoulou P, Quinzii C, Johnsen 18. Kowluru A, Tannous M, Chen HQ. Localization and characterization SD, et al. Navajo neurohepatopathy is caused by a mutation in the MPV17 of the mitochondrial isoform of the nucleoside diphosphate kinase in gene. Am J Hum Genet 2006;79:544-548. the pancreatic beta cell: evidence for its complexation with mitochon- 5. Singleton R, Helgerson SD, Snyder RD, O’Conner PJ, Nelson S, Johnsen drial succinyl-CoA synthetase. Arch Biochem Biophys 2002;398:160- SD, et al. Neuropathy in Navajo children: clinical and epidemiologic fea- 169. tures. Neurology 1990;40:363-367. 19. Elpeleg O, Miller C, Hershkovitz E, Bitner-Glindzicz M, Bondi-Rubin- 6. Vu TH, Tanji K, Holve SA, Bonilla E, Sokol RJ, Snyder RD, et al. Navajo stein G, Rahman S, et al. Deficiency of the ADP-forming succinyl-CoA neurohepatopathy: a mitochondrial DNA depletion syndrome? HEPATOL- synthase activity is associated with encephalomyopathy and mitochondrial OGY 2001;34:116-120. DNA depletion. Am J Hum Genet 2005;76:1081-1086. Neonatal Hemochromatosis: It’s OK To Say “NO” to Antioxidant-Chelator Therapy Mike A. Leonis and William F. Balistreri

“All who drink of this remedy recover in a short over, isolated hepatic siderosis is not specific for NH time but rather is observed in the normal term infant liver Except those whom it does not help, who all die. and more prominently in a multitude of liver diseases.4 It is obvious, therefore, that it fails only in incurable In addition to the standard supportive care given to cases” — Galen (circa 100 A.D.) any patient with acute liver failure, contemporary med- ical management of patients with NH typically has See Article on Page 1417 included the use of an antioxidant-chelator cocktail. eonatal hemochromatosis (NH) is a rare medical This strategy followed a preliminary report, in 1993, N condition presenting as severe liver disease in the which suggested that the use of this cocktail within the newborn period. The natural history of patients with first 14 days of life spared three newborns with NH 5 NH is grim; when not stillborn, most infants present in from otherwise certain demise. In theory, the antioxi- ␣ the first days or weeks of life with clinical features of dants used (D- -tocopheryl polyethylene glycol succi- liver failure. Within the first few weeks of life, most nate, N-acetylcysteine, prostaglandin-E1 and selenium) succumb to the complications of end-stage liver disease provide protection to body tissues from the oxidant if not successfully rescued by liver transplantation. burden of the iron overload state, while the iron chela- Given its rarity and severity, NH has been difficult to tor desferrioxamine provides an efficient means of elim- study and frustrating to treat. Rarely, spontaneous inating toxic circulating iron via the kidneys.1,2,6 Given recovery without disease-specific medical intervention these theoretical considerations and the dire outcomes or liver transplantation is observed in patients with NH. of patients with NH, it is understandable that this cock- In this issue of Liver Transplantation, Rodrigues and tail has been almost universally administered immedi- co-workers describe the presentation, management and ately after an infant is diagnosed to have NH. Anecdotal outcomes of the largest series of patients with NH experience, including our own, is that some patients reported to date: 19 infants evaluated over a 12-year that have received this cocktail have recovered shortly time span at the King’s College Hospital, London. Very after its initiation. Whether there is any real clinical few centers in the world have reported on more than a benefit derived from the use of this cocktail or whether few patients with NH.1-3 theses anecdotal associations are true, true and unre- NH is a syndrome of likely diverse etiology that lated, has remained a controversial and open question.2 leads to accumulation of iron in the liver and extrahe- To this point, perhaps the greatest impact of the patic sites throughout the body. The diagnosis requires current report will be the observation that there was no a constellation of clinical features in addition to those observed benefit to the use of the antioxidant-chelator typically found in neonatal liver failure. In the current cocktail in a large series of patients with NH. The analysis, the authors reasonably require at least two outcome of the 10 patients who received the antioxi- additional clinical features to arrive at the diagnosis of dant-chelator cocktail was virtually identical to that of NH: a positive family history, high serum ferritin levels, the nine patients who did not receive the cocktail prior histological confirmation of hepatic siderosis sparing the reticuloendothelial system and extrahepatic sidero- sis, and/or magnetic resonance imaging confirming Abbreviation: NH, neonatal hemochromatosis. extrahepatic involvement. The authors remind us, how- From the Cincinnati Children’s Hospital Medical Center, Cincin- nati, OH. ever, that a markedly elevated serum ferritin level, Received June 20, 2005; accepted June 21, 2005. although suggestive and consistent with the diagnosis of Address reprint requests to Mike A. Leonis, Division of Gastroenter- NH, is not a specific or sensitive marker to use in ology, Hepatology & Nutrition, Cincinnati Children’s Hospital Medical differentiating NH from other forms of neonatal acute Center, Cincinnati, OH 45229. Telephone: 513-736-3151; FAX: 513- liver failure.1,4 Additional markers of iron metabolism, 636-7805; E-mail: [email protected] Copyright © 2005 by the American Association for the Study of such as elevated transferrin iron saturation, may be Liver Diseases more helpful in the diagnosis, although this has not Published online in Wiley InterScience (www.interscience.wiley.com). been validated in a large number of patients.2,4 More- DOI 10.1002/lt.20541

Liver Transplantation, Vol 11, No 11 (November), 2005: pp 1323-1325 1323 1324 Leonis and Balistreri to 1993. Five patients from both groups proceeded to of the results reported by Rodrigues and colleagues, liver transplantation, and of the remaining non-trans- when considering whether to use some or any of the planted patients, only one patient from each group components of the antioxidant-chelator cocktail in survived with medical support only. In a previously patients with NH. reported smaller series of patients with NH, Flynn and Two broad hypotheses have been entertained colleagues suggested that early use of the cocktail regarding the etiology of NH.6 The first hypothesis is (within the first several days of life) in mild cases of NH that NH represents a clinical phenotype that is the end (the subset of patients with only a moderate elevation of result of a wide range of unique primary hepatic insults serum ferritin and less prolonged prothrombin time) that secondarily leads to abnormal fetal handling of may be of benefit.3 However, only two of five patients placentally derived iron. This hypothesis is based in part in the series of Flynn et al. treated with the antioxidant on an association of NH with certain genetic syn- cocktail survived, and the age at initiation of treatment dromes, metabolic liver diseases, or infectious hepatiti- (on the second day of life) and degree of biochemical des.4 abnormalities of these survivors were similar to three of The second hypothesis is that abnormal maternal- eight similarly treated patients reported by Sigurdsson fetal transfer, fetal handling, or fetal storage of iron and colleagues, all of whom died without liver trans- leads to an iron overload state that secondarily causes plantation.2,3 Moreover, in the current report, the sub- severe fetal liver injury; this is manifest as fetal loss or set of patients with the mildest degree of liver failure, in liver failure in the live born. Accumulating evidence whom the cocktail was started within the first 1-2 days suggests that this theory may be applicable in a large of life, all required transplantation or died before that percentage of patient sibships with NH and that NH could be achieved. Thus, unfortunately, there are no likely represents a gestational alloimmune disease pro- convincing data that show a beneficial effect of using cess. In this scenario, a mother having given birth to a the antioxidant-chelator cocktail in newborns with healthy child generates an antibody directed against a NH. fetal antigen to which she was immune-naive and which Ideally, to best address the clinical merits of the is involved in a critical part of fetal iron homeostasis. antioxidant-chelator cocktail prospectively in patients During subsequent pregnancies, fetuses are exposed to with NH, a randomized, placebo-controlled study that maternal IgG directed against this target fetal antigen, involves multiple pediatric liver referral centers would and this reaction subsequently leads to disturbed fetal need to be performed. Given the rarity of NH, perform- iron homeostasis.7 Supporting evidence for this ing an appropriately powered study would be a daunt- hypothesis includes (1) a high recurrence rate of up to ing task. Therefore, it is reasonable that the authors 80% in families following the birth of an affected child used historical controls in their evaluation, despite with healthy siblings being born prior to the proband potential limitations of such a research design. Of note, (similar to Rh incompatibility, for example); (2) a strik- the use of historical controls in this report should have ing pattern of maternal inheritance; and 3) the presence biased the results in favor of a beneficial clinical effect of in the sera of affected, but not unaffected, women of the cocktail after 1993, as earlier diagnosis and recog- immunoreactivity to a 32-kDa protein found in fetal nition of patients with NH and better contemporary liver, but not child or adult liver.7,8 Whitington and pre-liver transplantation management of patients with colleagues showed a dramatic reduction in the severity acute liver failure, would have favored survival in the and recurrence of NH born to mothers whose most treated group compared to the untreated group. recent gestation was affected with proven NH, when Importantly, several of the pharmacological agents they were treated in the latter half of their pregnancy in the cocktail pose potential dangers to the infant with with weekly IVIG.9 NH. For example, prostaglandin-E1 can lead to persis- There has been general success with liver transplan- tent patency of the ductus arteriosus and subsequent tation in patients with NH. Rates of successful trans- pulmonary fluid overload, a problem especially con- plantation and long-term graft function in liver trans- cerning for the youngest of newborns.2 This concern plant recipients appears to be good, especially given the was noted by the authors and led to modification of the extremely young age of the patient population and the cocktail components in their youngest patients wherein severity of the patient’s illness at the time of transplan- the ductus was noted to be patent. Desferrioxamine is a tation.1,2 Although reaccumulation of iron in the donor siderophore that may predispose infants to infection. liver has been reported in a patient transplanted for These potentially serious risks have to be weighed NH,10 in most cases this appears not to have a long- against the lack of substantiated clinical benefit, in light term negative impact on liver graft function. Editorial 1325

Obviously, much needs to be learned about the Knisely AS. Neonatal hemochromatosis: outcomes of pharma- pathophysiologic disturbances leading to the NH phe- cologic and surgical therapies. J Pediatr Gastroenterol Nutr 1998;26:85-89. notype and the appropriateness of medical manage- 3. Flynn DM, Mohan N, McKiernan P, Beath S, Buckels J, Mayer ment specifically tailored to patients with NH. The D, et al. Progress in treatment and outcome for children with current report reinforces the notion that the use of the neonatal haemochromatosis. Arch Dis Child Fetal Neonatal Ed antioxidant-chelator cocktail in this patient population 2003;88:F124-F127. may not be of benefit. At present, liver transplantation 4. Murray KF, Kowdley KV. Neonatal hemochromatosis. Pediat- rics 2001;108:960-964. in carefully selected patients (i.e., those without serious 5. Shamieh I, Kibort PK, Suchy FJ, Freese DK. Antioxidant ther- neurological sequelae) is the only reasonable approach apy for neonatal iron storage disease (NISD) [Abstract]. Pediatr that will impact the outcome for these newborns. The Res 1993;33:109A. encouraging studies of Whitington and colleagues,9 if 6. Knisely AS, Mieli-Vergani G, Whitington PF. Neonatal hemo- validated by others, would provide a life-saving medical chromatosis. Gastroenterol Clin North Am 2003;32:877-889, vi-vii. treatment strategy for already identified “carrier” moth- 7. Whitington PF, Malladi P. Neonatal hemochromatosis: is it an ers, as well as a useful paradigm to direct future research alloimmune disease? J Pediatr Gastroenterol Nutr 2005;40:544- investigations to identify at-risk mothers in an attempt 549. to reduce the number of spontaneous NH. 8. Verloes A, Temple IK, Hubert AF, Hope P, Gould S, Debauche C, et al. Recurrence of neonatal haemochromatosis in half sibs born of unaffected mothers. J Med Genet 1996;33:444-449. References 9. Whitington PF, Hibbard JU. High-dose immunoglobulin dur- ing pregnancy for recurrent neonatal haemochromatosis. Lancet 1. Rodrigues F, Kallas M, Nash R, Cheeseman P, D’Antiga L, Rela 2004;364:1690-1698. M, et al. Neonatal hemochromatosis-medical treatment vs. 10. Egawa H, Berquist W, Garcia-Kennedy R, Cox K, Knisely AS, transplantation: the Kings experience. Liver Transpl 2005;11: Esquivel CO. Rapid development of hepatocellular siderosis 1417-1424. after liver transplantation for neonatal hemochromatosis. Trans- 2. Sigurdsson L, Reyes J, Kocoshis SA, Hansen TW, Rosh J, plantation 1996;62;1511-1513. Pediatr Transplantation 2005: 9: 640–645 Copyright Ó 2005 Blackwell Munksgaard Pediatric Transplantation DOI: 10.1111/j.1399-3046.2005.00357.x Neonatal hemochromatosis: Fetal liver disease leading to liver failure in the fetus and newborn

Whitington PF, Kelly S, Ekong UD. Neonatal hemochromatosis: Fetal Peter F. Whitington, Susan Kelly and liver disease leading to liver failure in the fetus and newborn. Udeme D. Ekong Pediatr Transplantation 2005: 9: 640–645. Ó 2005 Blackwell Munksgaard Department of Pediatrics, Northwestern University Feinberg School of Medicine, Children's Memorial Abstract: Acute liver failure in the newborn is relatively rare but often Hospital, The Siragusa Transplantation Center, fatal. The broadest definition of acute liver failure is failure of the vital Chicago, IL, USA functions of the liver occurring within weeks or a few months of the onset of clinical liver disease. Therefore, by definition, any liver failure Key words: neonatal hemochromatosis – acute liver in the newborn can be construed to be acute liver failure. A second failure – fetus – newborn – alloimmune component of the general definition of acute liver failure is the lack of known preexisting liver disease. In the case of neonatal acute liver Peter F. Whitington, Children's Memorial Hospital Box 57, 2300 Children's Plaza, Chicago, IL 60614, USA failure, preexisting disease would by definition be liver disease that Tel.: +773 880 4643 affects the fetus. Almost nothing is known about fetal onset liver Fax: +773 975 8671 failure, and there is no literature addressing the subject. This review E-mail: [email protected] will address fetal liver disease that leads to liver failure in the fetus or newborn. Accepted for publication 22 February 2005

Acute liver failure in the newborn is relatively fetal liver disease that leads to liver failure in the rare but often fatal. In the broadest definition of fetus or newborn. acute liver failure, ÔacuteÕ means occurring within At what point in time relative to term that liver 8 wks of the onset of clinical liver disease. dysfunction or failure is typically detected may be Therefore, by definition, any liver failure in the an indicator of when the onset of liver injury newborn can be construed to be ÔacuteÕ. ÔFailureÕ occurred. Fig. 1 demonstrates conceptually that means loss of function to the degree that life is various liver diseases affecting the newborn will imperiled. In most definitions, failure is deter- produce liver failure at different gestational ages. mined by the onset of hepatic encephalopathy. Whereas, metabolic liver diseases tend to produce However, encephalopathy may be difficult to liver failure weeks to months after birth and detect in the newborn, so liver failure is more infectious diseases produce liver failure typically often defined in terms of synthetic failure (i.e. weeks after birth, neonatal hemochromatosis is a coagulopathy) and metabolic failure (i.e. hypo- disease that clearly affects the fetus. Several lines glycemia). Another assumption of the general of evidence date the onset of severe liver injury in definition of acute liver failure is the lack of neonatal hemochromatosis to about 24 wk. Intra- known preexisting liver disease. In the case of uterine growth restriction, oligohydramnios, and neonatal acute liver failure, preexisting disease fetal distress are recognized in the vast majority of would by definition be liver disease that affects proven cases of neonatal hemochromatosis (1). the fetus. Almost nothing is known about fetal Thus, neonatal hemochromatosis may be a model onset liver failure, and there is no literature for understanding the consequences and compli- addressing the subject. This review will address cations of severe fetal liver disease.

Mechanisms of fetal liver disease Abbreviations: AFP, a-fetoprotein; ALT, alanine amino- The potential etiologies of fetal liver disease are transferase; AST, aspartate aminotransferase; CMV, cyto- megalovirus; HSV, herpes simplex virus; IVIG, intravenous shown in Table 1. Although there is very little immunoglobulin; MRI, magnetic resonance imaging; evidence to support the idea, infectious disease is TPGS, tocopherol polyethylene glycol succinate. generally considered to be the most important 640 Fetal and neonatal liver failure

neonatal liver failure (5–7). Furthermore, they Neonatal haemochromatosis rarely produce isolated liver disease. Perhaps the most compelling evidence that viral infection does not cause severe fetal liver injury comes Infectious hepatitis from experimental infection in farm animals. Fetal loss due to infection is economically important in horticulture, and therefore has been Incidence carefully studied. Various viruses that cause fetal Metabolic disease loss in farm animals have been shown to relat- ively spare the liver. They often cause severe placentitis and sometimes necroinflammatory disease of brain or lung (8–14). From all of this –4 –2 Term 246 evidence, one can conclude that infection is at Age at recognition (wk) most a rare cause for fetal liver failure. Fig. 1. The age at onset of liver failure relative to term and Ischemia/abnormal perfusion has been demons- the relative incidence of the major causes of neonatal liver trated as a cause for acute liver failure in other failure are demonstrated. Neonatal hemochromatosis often age groups, and also seems to be a rare cause for results in fetal loss from liver failure and the severe liver fetal liver failure (15). Cases of severe maternal disease in affected live-born babies can often be dated to early in the 3rd trimester of gestation. In contrast, infectious shock have been shown to cause fetal loss with hepatitis rarely causes liver failure in the fetus or newborn, severe hepatic necrosis (14). Furthermore, severe and metabolic liver diseases typically present with liver placental disease such as severe placentitis can failure weeks after birth (and essentially never in utero). cause hepatic necrosis, but it is usually in association with other organ injury such as global brain necrosis (16, 17). Rare cases of Table 1. Etiologies of fetal liver disease severe fetal heart disease, for example severe Infection (rare) arrhythmia due to maternal lupus (18), have been Herpes viruses shown to rarely affect the liver with ischemic Enteroviruses (perhaps) necrosis. One can conclude from these rare case Not hepatitis viruses reports that ischemia/abnormal perfusion is a Ischemia/abnormal perfusion (rare) rare cause of fetal liver failure. Maternal shock Placentitis Replacement with other tissue or cells (hepato- Fetal heart disease phthisis) is a rare cause of fetal liver failure as Hepatophthisis (rare) well. Several examples of this phenomenon exist. Myelodysplasia Severe myelodysplasia in Down syndrome is Congenital leukemia associated with a liver that is essentially over- Hemangiomatosis Hemophagocytic syndrome (perhaps) whelmed with hematopoietic elements, leaving Genetic and metabolic liver disease (very rare) no room for functioning hepatic parenchyma Neonatal Hemochromatosis (most common) (19). The same has been described in cases of congenital leukemia, where blasts replace the liver and cases of diffuse hemangiomatosis, where vascular tissue replaces hepatic paren- etiology, closely followed by genetic and meta- chyma (20, 21). Finally, among the known causes bolic diseases. However, when evidence is con- for neonatal liver failure, hemophagocytic syn- sidered, all of the etiologies listed in Table 1 are drome is prominent in some series (22). However, extremely rare causes of fetal liver injury, other there is little evidence that hemophagocytic than perhaps neonatal hemochromatosis. syndrome affects the fetus or causes fetal liver Evidence does not support the supposition that failure (23). In conclusion, hepatophthisis is a transplacental infection is a cause of fetal liver rare documented cause of fetal liver failure. failure. Most known viral infections that can Evidence that genetic and metabolic diseases affect the fetus, such as HSV infection, cause cause fetal liver failure is sparse although they are necrosis and inflammation of many tissues most common causes of liver failure in neonates and notably the placenta, central nervous system and infants (24). In general, the fetus seems to be skin (2–4). Careful search of the literature fails to spared the consequences of most metabolic identify cases of HSV infection or other infec- diseases presumably because of placental support tious agents causing severe isolated hepatic removing toxic metabolites and supplying needed necrosis in the fetus. Viruses such as CMV have substrates for metabolism. One cannot find not been identified as a cause of fetal or early examples of tyrosinemia, bile salt synthetic 641 Whitington et al. defects, mitochondrial defects, lysosomal dis- At present it is not clear whether neonatal eases, or other prominent metabolic diseases hemochromatosis represents a single disease entity causing fetal liver failure. Some rare genetic or a phenotypic expression of more than one syndromes mimic neonatal hemochromatosis in disease process (33). It is a syndrome with char- some ways in the fetus and newborn, but fetal acteristic biochemical and anatomic findings (1). liver failure has not been described (25). One can The typical biochemical findings in neonatal conclude from the absence of data that genetic hemochromatosis include an extremely high and metabolic diseases are at most a rare cause of serum ferritin level, usually exceeding 800 ng/ fetal liver failure. mL, and extremely high levels of AFP, usually exceeding 200 ng/mL. Affected babies almost Neonatal hemochromatosis always have hypoglycemia, extreme cholestasis, and liver synthetic failure, with factors V and VII Neonatal hemochromatosis is a condition affect- usually less than 10% of normal and albumin ing the fetus and newborn in which iron is usually less than 2 gm/dL. An unusual charac- abnormally distributed in the liver and extra teristic given the degree of liver injury is low or hepatic tissues in a distribution similar to that absent serum AST and ALT, the cause for which is seen in hereditary HFE-associated hemochroma- currently unknown. None of these findings are tosis (1). Liver damage is extensive and liver specific for or diagnostic of neonatal hemochro- failure is the dominant clinical feature. Neonatal matosis. hemochromatosis is the most frequently recog- The diagnosis of neonatal hemochromatosis nized cause of liver failure in neonates. Consid- rests on demonstrating iron deposition (siderosis) erable evidence exists to suggest that it is a in extra hepatic tissues, namely epithelial tissues gestational disease in which fetal liver injury is a and heart (1). MRI is an excellent approach to constant feature. demonstrating abnormal iron in the pancreas, It has been well established that the liver adrenals, heart, and other tissues (34, 35). The histology of affected newborns is not character- power of this test to diagnose or exclude neonatal ized by acute necrosis. Rather, there is marked hemochromatosis is currently unknown, results structural damage with intense fibrosis and having been reported in only a few cases. Fig. 2 cirrhosis in most cases. Indeed, hepatocellular demonstrates the typical findings in neonatal carcinoma has been found in some affected hemochromatosis. Demonstrating iron depos- newborns (26–29). Since most affected babies ition in extra hepatic tissues by histology is exhibit liver failure within the first day of life and challenging, but is the gold standard for diagno- die within the first few days, it seems unlikely if sis. Autopsy specimens demonstrate the various not impossible that cirrhosis and its complica- epithelial tissues in which iron is abnormally tions could develop unless the process predated deposited, but most of these tissues are relatively birth. Moreover, affected babies show evidence inaccessible. It has been shown that biopsy of the of fetal insult. Essentially all affected live born buccal mucosa that includes sub mucosal glands babies are small for gestational age. It seems, therefore, that some process has caused intra- uterine fetal growth restriction. Several live born babies have been demonstrated to have renal dysgenesis (30–32). Correlation with the process of normal renal development dates this arrest of renal development to about 24 wk gestation. It is believed that this final stage of renal development is dependant upon liver function, and therefore the dysplasia dates liver dysfunction to the late second and early third trimester. Finally, even though only a few cases of renal dysgenesis have been reported, essentially all babies with NH have oligohdramnios, which indicates there has Fig. 2. The MRI (T2-weighted) and the iron stained section been fetal oliguria. This is perhaps the intra- of buccal mucosa from a newborn with neonatal hemo- uterine equivalent of hepatorenal syndrome seen chromatosis are demonstrated. The MRI (left panel) dem- onstrates attenuated signal (dark) from the liver and in post-natal acute liver failure, but one can only pancreas relative to the spleen, indicating increased iron speculate in that regard given that the appropri- content. The buccal biopsy shows iron deposition in epi- ate tests of renal function cannot be performed in thelium of sub-mucosal glands. Both findings are diagnostic the fetus. of neonatal hemochromatosis in a baby with evidence of liver disease or liver failure.

642 Fetal and neonatal liver failure

Table 2. The ÔcocktailÕ for treating neonatal hemochromatosis The evidence that neonatal hemochromatosis is an alloimmune gestational disease is mainly N-Acetycysteine – 200 mg/kg/day, divided three times daily, by mouth for 17–21 doses empiric. It seems to fit the recurrence mode of a-TPGS – 25 IU/kg/d divided twice daily by mouth for 6 wk gestational alloimmune diseases. It has been Deferoxamine – 30 mg/kg/d i.v. infused over 8 h until the serum ferritin demonstrated that once a woman has an affected <500 lg/L baby, there is an approximate 80% likelihood Selenium – 3 lg/kg/d i.v. by continuous infusion for length of hospitalization that the next pregnancy will be affected with Prostaglandin E1 – 0.4 lg/kg/h i.v. increasing to 0.6 over 3–4 h. The infusion is maintained for 10 days severe neonatal hemochromatosis. It has also been shown in several cases that mothers can have affected babies with different fathers, but never has it been described that fathers can have can provide a diagnosis in a relatively non- affected half-siblings with different mothers. All invasive way (36). However, the sensitivity of the of these are consistent with a maternal disease approach is unknown. While its positive predic- affecting the fetus rather than a primary fetal tive value seems great, its negative predictive disease. That affected babies can fully recover value is probably poor. Fig. 2 also shows the with medical therapy is more evidence that this is diagnostic deposition of iron in sub-mucosal not a primary fetal liver disease. glands in a buccal biopsy from an affected There are several examples of similar recurrent neonate. disease affecting the fetus. In alloimmunity, brief Managing a neonate with liver failure from access of the fetal target antigen to the maternal any cause is challenging. With regard to neonatal circulation is required, as well as lack of the hemochromatosis, the first advance in treatment antigen in the maternal repertoire of Ôself-anti- came with the introduction of a cocktail of gensÕ. In the most commonly occurring alloim- chelation and antioxidants designed to remove mune diseases, the fetus expresses a dominant excess iron and provide protection against iron paternal allele for blood group antigens induced oxidant injury (37). The components of expressed on erythrocytes (hydrops fetalis) or this cocktail are given in Table 2. The use of this platelets (alloimmune thrombocytopenia) that cocktail has never been tested in clinical trial. Its many normal mothers lack. A rare cause is for use is based on a reasonable hypothesis that iron the mother to have a homozygous genetic defi- induced oxidant injury causes or contributes to ciency that results in her not expressing a the liver injury in neonatal hemochromatosis. common species-related protein and therefore The reported data show an approximate 20% developing an immune response to a fetus survival with medical therapy (38–40). It is not expressing that protein from the normal paternal clear whether this is greater than the survival that allele. This mechanism was demonstrated in a could be achieved with intensive care support case of antenatal membranous glomerulo- alone, but major centers continue to use the nephritis in which the mother was genetically cocktail as no other therapy has proven greater deficient in neutral endopeptidase, a protein benefit. It is reasonable to institute therapy as normally expressed in the kidney and present in early as the diagnosis of neonatal hemochroma- the circulation (48). A third mechanism of tosis is considered, and there is some evidence alloimmune disease involves sensitization to a that instituting therapy early may improve out- fetal antigen that significantly differs from the come (41), especially those presenting with milder isoform expressed in mature individuals. This disease (38). It is of interest that babies who do mechanism has been described as a cause of recover with medical therapy generally recover recurrent arthrogyposis multiplex congenita (49). completely and do not have chronic liver disease. Mothers of affected babies were shown to have Indeed, we have demonstrated histological recov- antibodies against fetal acetylcholine receptor, ery from extensive cirrhosis to normal histology but not against the adult form, and did not have in several patients. myasthenia gravis. Passage of antibodies to the Several possible pathogenic mechanisms have fetus resulted in fetal myasthenia, muscle weak- been considered for neonatal hemochromatosis. ness, reduced fetal movement, and contractures. Perhaps the most widely believed is that neonatal This represents a lapse of memory by the immune hemochromatosis represents a consequence of or system of what was ÔselfÕ during intrauterine life. common phenotype for any severe fetal liver Maternal autoimmunity has recently been impli- injury (42–44). Some groups are investigating it cated as a cause of fetal open neural tube defects as a probable genetic disease since it recurs in (50). In this circumstance, mothers of affected siblings (45, 46). We have theorized that this is an children were shown to have circulating anti- immune mediated gestational disease (47). bodies against a folate receptor. These antibodies 643 Whitington et al. when transported to the fetus interfered with the have been affected with neonatal hemochromato- action of folate in neural tube development. The sis we will be able to demonstrate a specific mothers themselves had no clinical illness, sug- mechanism for disease, and prove that neonatal gesting the function of the targeted receptor is hemochromatosis is a single disorder. It may also most important in fetal development. There are offer a test with greater diagnostic power than no data from which to single out one of these those currently available. Moreover, it may permit mechanisms as the cause of recurrent NH. Based a treatment trial based on antibody titer with the on several observations we favor the possibility focus of improving gestational therapy and that alloimmunity against a common fetal anti- making it more palatable. Finally, it may be gen is the cause. possible to recapitulate the disease in an animal Based on the hypothesis that neonatal hemo- model that would permit laboratory study to chromatosis is a gestational alloimmune disease, improve post-natal therapy for affected babies. we have undertaken a trial of treating women during pregnancy to prevent its recurrence (47). Conclusion In this study, women whose most recent preg- Fetal liver failure is unusual, although it inci- nancy ended in documented neonatal hemochro- dence and prevalence as a cause of fetal loss are matosis were treated with IVIG, 1 g/kg body unknown. Neonatal hemochromatosis is unique weight weekly from the 18th week until the end among the diseases known to cause liver failure of gestation. The outcomes of treated pregnan- in the newborn to actually cause severe fetal liver cies were compared to the outcomes of randomly injury. While the etiology of neonatal hemochro- selected previous affected pregnancies for each matosis remains unknown, there is growing woman, which were used as historical controls. evidence that many cases are the result of Fifteen women were treated through 16 preg- maternal alloimmunity directed at fetal liver. nancies, all of which progressed uneventfully and Ongoing investigation in this area may lead to resulted in live babies that were healthy appear- understanding of the pathogenesis of neonatal ing and had appropriate for gestational age birth hemochromatosis, which should improve diag- weights. None had oligohydramnios. Twelve nosis and management. It may also improve babies had evidence of liver involvement with understanding of fetal liver injury in general. neonatal hemochromatosis: 11 had elevated serum AFP and ferritin levels, and four babies References had clinical liver disease with coagulopathy. All babies survived with medical or no treatment and 1. Knisely AS, Mieli-Vergani G, Whitington PF. Neonatal are currently healthy. Gestational IVIG therapy hemochromatosis. Gastroenterol Clin North Am 2003: 32: was associated with improved infant survival 877–889, vi–vii. 2. Robertson NJ, McKeever PA. Fetal and placental pathology (p ¼ 0.0009). We concluded from this experience in two cases of maternal varicella infection. Pediatr Pathol that treatment with high-dose IVIG during 1992: 12: 545–550. gestation appears to have modified recurrent 3. Robb JA, Benirschke K, Mannino F, Voland J. Intrauterine neonatal hemochromatosis so that it was not latent herpes simplex virus infection: II. Latent neonatal lethal to the fetus or newborn; results, which infection. Hum Pathol 1986: 17: 1210–1217. 4. Sauerbrei A. Varicella-zoster virus infections in pregnancy. further suggest an alloimmune mechanism. Intervirology 1998: 41: 191–6. Our ongoing laboratory investigation is 5. Benirschke K, Mendoza GR, Bazeley PL. Placental and fetal focused at determining the alloimmune mecha- manifestations of cytomegalovirus infection. Virchows Arch B nism, including the target of alloimmunity. We Cell Pathol 1974: 16: 121–139. are using affected mothersÕ sera to identify 6. Becroft DM. Prenatal cytomegalovirus infection: epidemiol- potential fetal target antigens. We preliminarily ogy, pathology and pathogenesis. Perspect Pediatr Pathol 1981: 6: 203–241. demonstrated that affected mothersÕ sera at high 7. Griffiths PD. Cytomegalovirus and the liver. Semin Liver Dis titer recognizes a protein expressed in fetal liver, 1984: 4: 307–313. both human and mouse, but not in human child 8. Joo HS, Dee SA, Molitor TW, Thacker BJ. In utero infec- or adult mouse liver. Using the same experimen- tion of swine fetuses with infectious bovine rhinotracheitis virus tal conditions, sera from spouses of affected (bovine herpesvirus-1). Am J Vet Res 1984: 45: 1924–1927. Nielsen Ronsholt Sorensen mothers, nulliparous women, and unaffected 9. J, L, KJ. Experimental in utero infection of pig foetuses with porcine parvovirus (PPV). Vet multiparous women do not recognize this pro- Microbiol 1991: 28: 1–11. tein. Our direction is to purify, identify, and 10. Richardson C, Taylor WP, Terlecki S, Gibbs EP. Obser- clone the fetal protein with the focus on devel- vations on transplacental infection with bluetongue virus in oping a serologic test for detecting alloantibody sheep. Am J Vet Res 1985: 46: 1912–1922. in women. If such alloantibody can be determined 11. Shibata I, Hatano Y, Nishimura M, Suzuki G, Inaba Y. to be both sensitive and specific for women who Isolation of Getah virus from dead fetuses extracted from a 644 Fetal and neonatal liver failure

naturally infected sow in Japan. Vet Microbiol 1991: 27: 385– 31. Johal JS, Thorp JW, Oyer CE. Neonatal hemochromatosis, 91. renal tubular dysgenesis, and hypocalvaria in a neonate. 12. Waldvogel AS, Anderson GA, Phillips DL, Osburn BI. Pediatr Dev Pathol 1998: 1: 433–437. Infection of bovine fetuses at 120 daysÕ gestation with virulent 32. Morris S, Akima S, Dahlstrom JE, Ellwood D, Kent A, and avirulent strains of bluetongue virus serotype 11. Comp Falk MC. Renal tubular dysgenesis and neonatal hemochro- Immunol Microbiol Infect Dis 1992: 15: 53–63. matosis without pulmonary hypoplasia. Pediatr Nephrol 2004: 13. Munoz DP, Lager IA, Mersich S, et al. Foetal infections with 19: 341–344. bovine viral diarrhoea virus in Argentina. Br Vet J 1996: 152: 33. Collins J, Goldfischer S. Perinatal hemochromatosis: one 175–82. disease, several diseases or a spectrum? Hepatology 1990: 12: 14. Rogers FB, Rozycki GS, Osler TM, et al. A multi- 176–177. institutional study of factors associated with fetal death in in- 34. Marti-Bonmati L, Baamonde A, Poyatos CR, Monteagudo jured pregnant patients. Arch Surg 1999: 134: 1274–1277. E. Prenatal diagnosis of idiopathic neonatal hemochromatosis 15. Dahms BB, Boyd T, Redline RW. Severe perinatal liver dis- with MRI. Abdom Imaging 1994: 19: 55–56. ease associated with fet al thrombotic vasculopathy. Pediatr 35. Hayes AM, Jaramillo D, Levy HL, Knisely AS. Neonatal Dev Pathol 2002: 5: 80–85. hemochromatosis: diagnosis with MR imaging. Am J Roent- 16. Okumura M, Aparecida dos Santos V, Camargo ME, genol 1992: 159: 623–625. Schultz R, Zugaib M. Prenatal diagnosis of congenital Cha- 36. Knisely AS, O’Shea PA, Stocks JF, Dimmick JE. Oropha- gasÕ disease (American trypanosomiasis). Prenat Diagn 2004: ryngeal and upper respiratory tract mucosal-gland siderosis in 24: 179–181. neonatal hemochromatosis: an approach to biopsy diagnosis. 17. Chatterjee A, Chartrand SA, Harrison CJ, Felty- J Pediatr 1988: 113: 871–874. Duckworth A, Bewtra C. Severe intrauterine herpes simplex 37. Shamieh I, Kibort PK, Suchy FJ, Freese DK. Antioxidant disease with placentitis in a newborn of a mother with recurrent therapy for neonatal iron storage disease (NISD). Abstract. genital infection at delivery. J Perinatol 2001: 21: 559–564. Pediatr Res 1993: 33: 109A. 18. Brucato A, Cimaz R, Stramba-Badiale M. Neonatal lupus. 38. Flynn DM, Mohan N, McKiernan P, et al. Progress in Clin Rev Allergy Immunol 2002: 23: 279–299. treatment and outcome for children with neonatal haemo- 19. Ruchelli ED, Uri A, Dimmick JE, et al. Severe perinatal liver chromatosis. Arch Dis Child Fet al Neonatal Ed 2003: 88: disease and Down syndrome: an apparent relationship. Hum F124–F127. Pathol 1991: 22: 1274–1280. 39. Shneider BL. Neonatal liver failure. Curr Opin Pediatr 1996: 20. Enjolras O, Riche MC, Merland JJ, Escande JP. Manage- 8: 495–501. ment of alarming hemangiomas in infancy: a review of 25 cases. 40. Sigurdsson L, Reyes J, Kocoshis SA, Hansen TW, Rosh J, Pediatrics 1990: 85: 491–498. Knisely AS. Neonatal hemochromatosis: outcomes of phar- 21. Tsukagoshi H, Iwasaki Y, Toyoda M, et al. An autopsy case macologic and surgical therapies. J Pediatr Gastroenterol Nutr of systemic hemangiomatosis with honeycomb-like liver and 1998: 26: 85–89. focal splenic sarcomatoid changes. Intern Med 1998: 37: 847– 41. Vohra P, Haller C, Emre S, et al. Neonatal hemochroma- 852. tosis: the importance of early recognition of liver failure. 22. Rowbotham D, Wendon J, Williams R. Acute liver failure J Pediatr 2000: 136: 537–541. secondary to hepatic infiltration: a single centre experience of 42. Hoogstraten J, de Sa DJ, Knisely AS. Fetal liver disease 18 cases. Gut 1998: 42: 576–580. may precede extrahepatic siderosis in neonatal hemochroma- 23. Parizhskaya M, Reyes J, Jaffe R. Hemophagocytic syn- tosis. Gastroenterology 1990: 98: 1699–1701. drome presenting as acute hepatic failure in two infants: clinical 43. Kershisnik MM, Knisely AS, Sun CC, Andrews JM, overlap with neonatal hemochromatosis. Pediatr Dev Pathol Wittwer CT. Cytomegalovirus infection, fetal liver disease, 1999: 2: 360–366. and neonatal hemochromatosis. Hum Pathol 1992: 23: 1075– 24. Durand P, Debray D, Mandel R, et al. Acute liver failure in 1080. infancy: a 14-year experience of a pediatric liver transplanta- 44. Witzleben CL, Uri A. Perinatal hemochromatosis: entity or tion center. J Pediatr 2001: 139: 871–876. end result? Hum Pathol 1989: 20: 335–340. 25. Rapola J, Heikkila P, Fellman V. Pathology of lethal fetal 45. Kelly AL, Lunt PW, Rodrigues F, et al. Classification and growth retardation syndrome with aminoaciduria, iron over- genetic features of neonatal haemochromatosis: a study of load, and lactic acidosis (GRACILE). Pediatr Pathol Mol Med 27 affected pedigrees and molecular analysis of genes implica- 2002: 21: 183–193. ted in iron metabolism. J Med Genet 2001: 38: 599–610. 26. Jonas MM, Kaweblum YA, Fojaco R. Neonatal hemochro- 46. Cox TM, Halsall DJ. Hemochromatosis-neonatal and young matosis: failure of deferoxamine therapy. J Pediatr Gastroen- subjects. Blood Cells Mol Dis 2002: 29: 411–417. terol Nutr 1987: 6: 984–988. 47. Whitington PF, Hibbard JU. High-dose immunoglobulin 27. Oliveira MG, Fernandes A, Silva AC, Moreira R, during pregnancy for recurrent neonatal haemochromatosis. Azevedo A, Da Silva LJ. A case of neonatal haemochroma- Lancet 2004: 364: 1690–1698. tosis. Acta Paediatr 1998: 87: 102–104. 48. Debiec H, Guigonis V, Mougenot B, et al. Antenatal mem- 28. Egawa H, Berquist W, Garcia-Kennedy R, Cox K, Knisely branous glomerulonephritis due to anti-neutral endopeptidase AS, Esquivel CO. Rapid development of hepatocellular side- antibodies. N Engl J Med 2002: 346: 2053–2060. rosis after liver transplantation for neonatal hemochromatosis. 49. Riemersma S, Vincent A, Beeson D, et al. Association of Transplantation 1996: 62: 1511–1513. arthrogryposis multiplex congenita with maternal antibodies 29. Silver MM, Beverley DW, Valberg LS, Cutz E, Phillips inhibiting fet al acetylcholine receptor function. J Clin Invest MJ, Shaheed WA. Perinatal hemochromatosis. Clinical, 1996: 98: 2358–2363. morphologic, and quantitative iron studies. Am J Pathol 1987: 50. Rothenberg SP, da Costa MP, Sequeira JM, et al. Auto- 128: 538–554. antibodies against folate receptors in women with a pregnancy 30. Bale PM, Kan AE, Dorney SF. Renal proximal tubular complicated by a neural-tube defect. N Engl J Med 2004: 350: dysgenesis associated with severe neonatal hemosiderotic liver 134–142. disease. Pediatr Pathol 1994: 14: 479–489.

645 Journal of Pediatric Gastroenterology and Nutrition 42:282Y286 Ó March 2006 Lippincott Williams & Wilkins, Philadelphia

Neonatal Herpes Simplex Virus Infection Presenting as Acute Liver Failure: Prevalent Role of Herpes Simplex Virus Type I

*Anita Verma, MD, †Anil Dhawan, MD, *Mark Zuckerman, MRCPath, †Nedim Hadzic, FRCPCH, †Alastair J. Baker, FRCPCH, and †Giorgina Mieli-Vergani, MD

*Department of Medical Microbiology and Virology, Health Protection Agency, London, King’s College Hospital, London, UK; and †Paediatric Liver Centre, King’s College London School of Medicine at King’s College Hospital, London, UK

ABSTRACT sented with nonspecific symptoms of poor feeding and Background: Acute liver failure (ALF) in neonates is rare but lethargy within 2 weeks from birth. Seven of the 11 patients carries a high mortality without liver transplantation. Herpes had HSV-1 infection, 4 HSV-2. Only 2 patients who received simplex virus (HSV) is one of the microbes that more com- early treatment with intravenous acyclovir survived. monly causes ALF and is potentially treatable; hence, early Conclusions: HSV-related ALF in the neonatal period carries diagnosis and treatment are important to avoid progression to high morbidity and mortality and needs a high index of liver failure. suspicion so that life-saving treatment can be started promptly. Patients and Results: We have analysed retrospectively the Both HSV-1 and HSV-2 can cause severe neonatal infection. It case notes of 11 patients with HSV-induced ALF. A history of is important to recognise HSV infection in women of child- possible herpes infection was elicited in 5 parents, but HSV bearing age and their sexual partners. JPGN 42:282Y286, 2006. had not been suspected clinically. All patients were asympto- Key Words: neonatal herpes simplex infectionVacute liver matic when discharged from postnatal units and were pre- failureVacyclovir. Ó 2006 Lippincott Williams & Wilkins

INTRODUCTION proven HSV infection were reviewed. The diagnosis of ALF was based on a prolonged international normalised prothrom- Neonatal acute liver failure (ALF) is a rare condition bin ratio (INR 9 2) associated with elevated liver enzymes and that carries high mortality without liver transplantation. bilirubin. All testable causes of ALF in this age group were The most common causes of ALF in this age group are excluded by appropriate investigations. The diagnosis of HSV neonatal iron storage disease, inborn errors of metabo- infection was based on virus isolation, antigen detection by immunofluorescence and immunostaining or HSV DNA lism, hemophagocytic lymphocytosis and infection (1). detection by real-time polymerase chain reaction (PCR) in Among the infectious causes, herpes simplex virus swabs from vesicular lesions, eye, throat, stool, whole blood (HSV) is the most common (2). Infection with HSV is and postmortem tissue. treatable, hence early diagnosis and treatment are important to avoid progression to liver failure (3). The infection, however, is often suspected at a late stage RESULTS when treatment is no longer effective. In this retro- spective review of HSV-induced ALF in neonates, we The median maternal age at delivery was 25 years have attempted to identify anamnestic, clinical and (range 17Y42 years) and the median gestation age was 40 laboratory features able to increase the level of suspi- weeks (range 33Y42 weeks). Two mothers had a history cion for HSV infection. We also highlight the impor- of miscarriage. Antenatal fetal scan was normal in all. tance of recognising HSV infection in women of Seven neonates were delivered vaginally, whereas emer- childbearing age and their sexual partners. gency lower segment cesarean section (LSCS) was performed in 4 for fetal distress. The median (range) PATIENTS AND METHODS duration of rupture of membrane to delivery was 8 hours (2Y48 hours). In 3 patients, membranes were ruptured for The case records of 11 neonates referred to our specialist more than 23 hours before delivery. A history of possible liver service between 1995 and 2004 with ALF secondary to herpes infection was present in 5 parents (3 mothers and 3 partners), but no investigations to establish the etiology Received May 6, 2005; accepted December 20, 2005. had been carried out. Two mothers had a history of Address correspondence and reprint requests to Dr Anita Verma, Department of Medical Microbiology, HPA, London, King’s College genital lesions, treated as vaginal candidiasis in one, Hospital, Denmark Hill, London SE5 9RS, UK. (e-mail: anita.verma@ genital warts in the other and 1 mother had a history of kingsch.nhs.uk). recurrent cold sores. Of the 3 partners, 2 had recurrent

282

Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. NEONATAL HERPES SIMPLEX VIRUS INFECTION PRESENTING AS ACUTE LIVER FAILURE 283 cold sores and 1 had vesicular lesions on his hand 3 days Investigations and Management at before the birth of the baby. King’s College Hospital

Management of Patient at Their Local Hospital The clinical characteristics and laboratory investiga- tions of the patients are summarised in Tables 1Y3. The The clinical characteristics and investigations of the median age at referral to our unit was 12 days (range patients are summarised in Tables 1Y3. All patients 8Y24 days). A median of 6 days (range 2Y20) delay had normal birth weight (median 3.2 kg, range 2.3Y4.5 occurred between onset of symptoms and admission to kg), and Apgar scores at delivery were asymptomatic King’s College Hospital, where all patients were when discharged from the maternity unit. All pre- managed in the pediatric intensive care unit and sented with nonspecific symptoms of poor feeding and required assisted ventilation. Six also required renal lethargy at a median age of 5 days (range 2Y12 days). support by continuous veno-venous hemofiltration. Six patients had fever, 5 were jaundiced and 3 had a Antibiotics were changed to cefuroxime and amoxicillin vesicular rash. Two patients developed generalised and fluconazole was added in all according to the seizures. All were suspected to have bacterial sepsis ALF protocol. All were given intravenous acyclovir and started on broad-spectrum antibiotics (benzylpe- 20Y60 mg/kg/d, the lower dose being used in the presence nicillin, gentamicin and amoxicillin). Swabs were sent of renal impairment. Five patients showed a mild degree from the rash and eye discharge in 1 patient and from of hemophagocytosis on bone marrow examination. skin vesicular lesions in another to investigate possi- All patients had very high INR and LFTs (Table 2). ble HSV infection. Lumbar puncture was performed in Four patients had vesicular skin lesions, from which only 2 patients to rule out meningitis, CSF PCR being HSV-I was isolated in 2 patients and HSV-2 in 1 later reported positive for HSV-1 in one and negative patient. Nine patients underwent a lumbar puncture, in the other. All patients had abnormal liver function whereas in 2 patients it was not done because of severe tests (LFTs) (Table 2). Only one 8-day-old baby, who coagulopathy. CSF PCR was positive for HSV DNA had vesicular lesions, was started on intravenous in 5 patients, 4 having HSV-1 and 1 having HSV-2 acyclovir 1 day after admission and other 5 patients (Table 3). On abdominal ultrasound scan, 7 patients 1 day before transfer to our hospital. All patients had had ascites and 6 patients had hepatomegaly. Overall, severe coagulopathy and high transaminases and were 7 patients were infected by HSV-1 and 4 patients transferred to our specialist liver centre for manage- were infected by HSV-2 (Table 3). None was found ment of ALF. to have underlying metabolic disease, other infections or

TABLE 1. Clinical characteristics of patients

Local hospital King’s College Hospital Age at which Patient Age Age IV acyclovir Outcome No. (days) Clinical features (days) *Clinical features was started (days) (age in days) 1 8 Lethargy, fever, choking 14 ALF, RF, RDS, hepatomegaly 14 Died at day 18 episode and hepatomegaly and ascites 2 5 Drowsy, fever and hypotonia 11 ALF and uncontrolled bradycardia 10 Died at day 11 3 6 Poor feeding, lethargy, fever and 11 ALF, RF, generalised seizures 10 Died at day 12 generalised seizure and ascites 4 5 Lethargy and jaundice 8 ALF, GI bleed, RDS, hepatomegaly 9 Died at day 9 and ascites 5 12 Fever, vomiting, irritability and 14 ALF, RDS and unconscious 13 Died at day 15 poor feeding 6 3 Fever, jaundice and vesicle 9 ALF, drowsy and ascites 8 Died at day 14 on hand 7 2 Hypotonia, rash, lethargy and 9 ALF, RF, RDS, hepatomegaly 9 Died at day 11 poor feeding and ascites 8 3 Lethargy, jaundice and 10 ALF, RF and hepatomegaly 10 Died at day 11 poor feeding 9 5 Vesicular rash, poor feeding 16 ALF, RF, generalised seizures, 15 Died at day 17 and hypotonia hepatomegaly and ascites 10 4 Fever, jaundice and hepatomegaly 24 ALF and hepatomegaly 12 Well after LT 11 8 Lethargy, poor feeding 11 ALF, RF, RDS and ascites 9 Alive and well and jaundice

Pt = patient; ALF = acute liver failure; RF = renal failure; RDS = respiratory distress syndrome; IV = intravenous; GI = gastrointestinal; LT = liver transplantation. *All patients had diagnosis of ALF based on prolonged INR, elevated serum transaminases and bilirubin. All had varying degree of renal dysfunction.

J Pediatr Gastroenterol Nutr, Vol. 42, No. 3, March 2006

Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. 284 VERMA ET AL.

TABLE 2. Laboratory investigations on admission: LFTs and RFTs

Test (normal values) Local hospital, median (range) Referral hospital, median (range) INR (G1.2) 6.5 (1.6Y910) 10 (3.9Y915) GGT (17Y131 IU/L) 89 (38Y283) 103 (63Y284) ALP (48Y406 IU/L) 219 (56Y2100) 305 (94Y2300) ALT (6Y50 IU/L) 9100 (248Y13860) 9705 (908Y13806) Bilirubin (G20 Kmol/L) 56 (5Y359) 78 (24Y430) Creatinine (14Y48 Kmol/L) 64 (21Y168) 122 (60Y232) Urea (0.7Y5.7 mmol/L) 4.1 (2.1Y7.8) 5.4 (3.9Y10)

LFTs = liver function tests; RFTs = renal function test; INR = international normalised prothrombin ratio; GGT = gamma-glutamyl transpeptidase; ALP = alkaline phosphatase; ALT = aspartate aminotransferases. congenital abnormalities. Of the 11 patients, 9 died weeks of antiviral treatment.Aliverbiopsyperformed between 1 and 5 days after admission to our unit of at the age of 9 weeks showed postnecrotic fibrosis, multiorgan failure. Postmortem examination, carried out patchy coagulative necrosis of the parenchyma and in 8 patients, confirmed the diagnosis in all. profound hepatocellular and canalicular cholestasis. Only 2 patients survived. One underwent liver The patient is well, with normal growth but still had transplantationattheageof78days.Hehadbeen mildly elevated gamma-glutamyl transpeptidase treated with acyclovir for 2 weeks at the age of 12 114 IU/L (nv for age 1Y55 IU/L) 13 months after dis- days, 8 days after admission into his local hospital. charge. Longer follow-up is needed to determine the At the age of 30 days, HSV PCR on whole blood was possible consequences of the severe neonatal liver damage. positive whereas bacterial and viral cultures were negative. He was restarted on 60 mg/kg/d of acyclovir, DISCUSSION but because of worsening liver function, coagulopathy, persistent positivity of whole blood HSV DNA PCR In this retrospective review of a relatively large num- despite 4 weeks of treatment, underwent liver trans- ber of patients, the poor prognosis of HSV-related ALF in plantation. HSV DNA was confirmed by immunostain- the neonatal period is emphasised. A high index of sus- ing in the explanted liver (3) and became negative in picion for HSV infection among the medical profession is the whole blood 7 days after transplant on acyclovir needed to institute promptly life-saving treatment. treatment, which was continued for further 3 weeks. The mode of presentation of HSV infection in the The patient is well with normal LFTs 3 years after neonatal period is variable. The most common is that transplantation. The other patient was given intra- of a sepsis like picture at 5Y7 days of age, due to venous acyclovir at day 9 of life, 1 day after the onset disseminated disease, in which, however, skin lesions of symptoms and admission in his local hospital. His can be absent in up to 30% of cases, making diagnosis whole blood PCR became negative for HSV after 8 difficult (4). Other modes of presentation are isolated

TABLE 3. Laboratory investigations for the diagnosis of HSV

CSF analysis (WBC count/gram HSV culture positive on Pt No. Samples analysed for HSV infection stain/HSV PCR postmortem tissue HSV type 1 Eye swab positive for HSV-1 ++Pus cells/NOS/HSV-1 Liver, AF, lung, PF HSV-1 2 CSF PCR positive for HSV-1 2 WBC/NOS/HSV-1 Liver, AF, lung, PF HSV-1 3 ND No cells/NOS/Negative Liver, lung, PF HSV-1 4 Eye and TS negative for HSV 2 WBC/NOS/Negative Liver, AF, lung, PF HSV-2 5 CSF PCR positive for HSV-1 2 WBC/NOS/HSV-1 Liver HSV-1 6 Swab from hand vesicle positive for HSV-1 2 WBC/NOS/HSV-1 Liver, AF HSV-1 7 Swab from rash negative for HSV 2 WBC/NOS/Negative Liver HSV-2 8* Whole blood PCR positive for HSV-1 ND Liver HSV-1 9 Swab from rash and whole blood PCR ND ND HSV-1 positive for HSV-1 10 Whole blood and liver biopsy PCR No cells/NOS/Negative Explanted liver† HSV-2 were positive for HSV-2 11 Whole blood, AF and CSF PCR were No cells/NOS/HSV-2 NA HSV-2 positive for HSV-2. TS and scalp lesions culture positive for HSV-2

WBC = white blood cell; PCR = polymerase chain reaction; CSF = cerebrospinal fluid; HSV-1 = herpes simplex virus type 1; HSV-2 = herpes simplex virus type 2; NOS = no organisms seen; ND = not done; TS = throat swab; AF = ascitic fluid; PF = pleural fluid; NA = not applicable. *Only 1 patient had blood culture positive for Staphylococcus aureus and in remaining patients bacterial cultures were negative. †Immunostaining and PCR positive in explanted liver.

J Pediatr Gastroenterol Nutr, Vol. 42, No. 3, March 2006

Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. NEONATAL HERPES SIMPLEX VIRUS INFECTION PRESENTING AS ACUTE LIVER FAILURE 285 skin, eye and/or mouth (SEM) lesions, or encephalitis over, genital HSV infection in women may remain with or without SEM lesions (4). CNS involvement is undiagnosed, as symptoms of vaginal irritation may be present in about one third of the infected neonates (5). attributed to other causes such as candida. The diagnosis Hepatitis is a recognised feature of disseminated HSV may be missed unless visible vesicular lesions are pre- infection, although progression to ALF is rare (6,7). All sent and an appropriate sample is collected for virus 11 patients in this series presented nonspecifically with isolation or HSV DNA detection. Diagnosis of HSV poor feeding, lethargy and fever, and these symptoms infection in the neonate is based on cultures of fluid being widely recognised to be prodromes of bacterial from skin lesions, if present, or on swabs in viral trans- sepsis, they were treated promptly with broad-spectrum port medium from the eye, oropharynx, urine, stool or antibiotics in their local hospitals. The most striking rectum and blood (12). CSF, if coagulopathy permits, feature in all patients was the severity of liver disease should be sent for HSV DNA detection. The detection with high transaminases and coagulopathy (Table 2). It of HSV DNA in the blood may be helpful (3) but its is surprising that the infants’ symptoms did not elicit the significance needs to be clarified (20,21). suspicion of possible HSV infection despite a positive The role of antenatal screening is controversial. A HSV history in 5 parents, including one with genital surveillance of neonatal HSV infection undertaken by the lesions, a high rate of prolonged rupture of membranes British Paediatric Surveillance Unit between 1986 and and the presence of skin lesions in 3 babies. In only 1 1991 reported an estimated prevalence of 1.65 in 100,000 patient acyclovir was started a day after admission in live births. HSV-1 and HSV-2 were represented in equal the local hospital because of the presence of vesicles on proportions. Antenatal screening was felt to be unjusti- his scalp. These findings suggest a lack of awareness of fied, as the prevalence figure was considered too low. the risk factors, severity and modes of presentation of However, although not a standard practice, it has been neonatal HSV infection among health care workers. reported that prophylaxis with oral acyclovir late in In contrast to previous studies reporting that dissemi- pregnancy in mothers with HSV infection reduces viral nated infection is more commonly associated with HSV- shedding and frequency of recurrence, with consequent 2 than HSV-1 (70% versus 30%) (4), HSV-1 was the probable reduction of HSV transmission to babies during most frequent cause of infection in our series. Although labor (22,23). If these data were confirmed, antenatal most HSV-1 infections are oropharyngeal, this virus has screening would have important implications in the been reported to cause genital herpes in 20Y40% of prevention of infection. It is well recognised that iden- infected patients (8,9), whereas HSV-2 disease is almost tifying pregnant women or their partners with genital always genital (10). Only about 20% of people who HSV is important to counsel them about HSV trans- have been infected with HSV-2 has been reported to mission and the potentially disastrous consequences of have genital herpes disease (10). It is possible that neonatal infection. Our data suggest that appropriate sexual practices, including oral sex with an HSV-1- questioning and counseling should be offered also to positive partner, may result in a primary genital infec- parents with other sites of herpes infection. tion in seronegative women, with a consequent risk of Early initiation of antiviral therapy in the infected child intrapartum transmission to the infant (11). Neonatal improves the prognosis. Since acyclovir has become HSV infection can be acquired in utero, perinatally or available, and with early recognition of SEM lesions, postnatally, the perinatal modality being the most the rate of disseminated neonatal infection has decreased common (985%) (12). Most neonatal HSV-1 infections from 75% to 25% (12). The mortality rate in infants with are associated with maternal infection late in pregnancy encephalitis has reduced to 4%, although up to 69% of (16), although they may also be acquired postnatally survivors are left with neurologic sequelae (12). In a through breast-feeding from a nipple lesion, or from im- nonrandomised study, high dose intravenous acyclovir (60 mediate family members or hospital staff with sympto- mg/kg/d) for 21 days has been reported to reduce the matic or asymptomatic orolabial infections (14Y16). mortality of neonates with HSV infection (25). High dose HSV-2 vertical transmission occurs most frequently acyclovir for 14 days has been recommended by the when primary infection is acquired late in pregnancy American Academy of Pediatrics Committee on Infec- (17,18), whereas recurrent HSV-2 disease has a lower tious Diseases to treat neonatal SEM HSV disease (26). In incidence of transmission to the newborn. The retro- agreement with the above observations, the only 2 spective nature of this study does not allow us to neonates who survived in the present series had received establish the exact mode of transmission, but it is acyclovir early in the course of the disease. One of these conceivable that most infants were infected perinatally patients, who, despite early treatment, has required a liver or postnatally by their parents. transplant for persistent liver failure, confirms that HSV The diagnosis of neonatal HSV can be difficult. Up to infection is not a contraindication to liver transplantation, 60Y80% of women who deliver an HSV-infected infant if the damage is mainly hepatic and severe neurologic have no evidence of genital HSV infection at the time of damage has not occurred (3,24). For all other patients, by delivery, nor a past history of genital herpes or a sexual the time acyclovir treatment was instituted, liver and partner with a history of genital HSV (5,12,19). More- multiorgan failures were too advanced to be reversible.

J Pediatr Gastroenterol Nutr, Vol. 42, No. 3, March 2006

Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. 286 VERMA ET AL.

On the basis of this and previous reports (6,7), HSV 10. Corey L. The current trend in genital herpes: progress in pre- Y must be considered in the differential diagnosis of vention. Sex Transm Dis 1994;21:S38 44. 11. Brown ZA, Vontver LA, Benedetti J, et al. Effects on infants of a acutely ill newborns. Acyclovir treatment should be first episode of genital herpes during pregnancy. N Engl J Med started empirically in neonates with suspected sepsis, 1987;317:1246Y51. particularly in the presence of abnormal liver function, 12. Kimberlin DW. Neonatal herpes simplex infection. Clin Microbiol with the same solicitude with which antibiotics are pre- Rev 2004;17(1):1Y13. 13. Baldwin S, Whitley RJ. Intrauterine herpes simplex virus scribed, while the results of investigations are awaited. infection. Teratology 1989;39:1Y10. Because acyclovir can occasionally cause neutropenia 14. Nahmias A, Dowdle WR, Josey WE, et al. Newborn infec- and renal impairment in the neonate, close monitoring tions with herpes virus homonis types I and II. J Pediatr 1996; for these complications is advisable. 75:1194. In conclusion, both HSV-1 and HSV-2 cause neonatal 15. Yeager AS, Ashley RL, Corey L. Transmission of herpes simplex virus from father to neonate. J Pediatr 1983;103: infection. A high index of suspicion is required for early 905Y7. diagnosis in septic neonates to start promptly appro- 16. Francis DP, Hermann KL, MacMahon JR, et al. Nosocomial and priate antiviral therapy. In addition, it is important to maternally acquired herpesvirus hominis infections. Am J Dis recognise genital HSV infection in women of child- Child 1975;129:889. 17. Brown ZA. Genital herpes complicating pregnancy. Dermatol bearing age and their sexual partners. Clin 1998;16:805Y10. 18. Kesson AM. Management of neonatal herpes simplex virus in- fection. Paediatr Drugs 2001;3:81Y90. REFERENCES 19. Yeager AS, Arvin AM. Reasons for absence of a history of recurrent genital infections in mothers of neonates infected with 1. Bhaduri BR, Mieli-Vergani G. Fulminant hepatic failure. Paedi- herpes simplex virus. Pediatrics 1984;73:188Y93. atric aspects. Semin Liver Dis 1996;16(4):349Y55. 20. Diamond C, Mohan K, Hobson A, et al. Viremia in neonatal 2. Marion AW, Dhawan A. Acute liver failure. Indian J Pediatr herpes simplex virus infections. Pediatr Infect Dis J 1999;18: 2002;69(1):87Y91. 487Y9. 3. Twagira M, Hadzic N, Smith M, et al. Disseminated neonatal 21. Kimura H, Futamura M, Kito H, et al. Detection of viral DNA in herpes simplex virus (HSV) type 2 infection diagnosed by HSV neonatal herpes simplex virus infections: frequent and prolonged DNA detection in blood and successfully managed by liver presence in serum and cerebrospinal fluid. J Infect Dis 1991; transplantation. Eur J Pediatr 2004;163:166Y9. 164:289Y93. 4. Whitley RJ, Arvin AM. Herpes simplex virus infections. In: 22. Scott LL, Hollier LM, McIntire D, et al. Acyclovir suppression to Remington JS, Klein JO, eds. Infectious Diseases of the Fetus and prevent clinical recurrence at delivery after first episode of genital Newborn Infant, 4th ed. Philadelphia: Saunders, 1995:354Y76. herpes in pregnancy: an open label trial. Infect Dis Obstet Gynecol 5. Whitley RJ, Corey L, Arvin A, et al. Changing presentation of 2001;9:60Y75. herpes simplex virus infection in neonates. J Infect Dis 1988; 23. Braig S, Luton D, Sibony O, et al. Acyclovir prophylaxis in late 158:109Y16. pregnancy prevents recurrent genital herpes and viral shedding. 6. Benador N, Mannhardt W, Shcranz D, et al. Three cases of Eur J Obstet Gynecol Reprod Biol 2001;96:55Y8. neonatal herpes simplex virus infection presenting as fulminant 24. Lee WS, Kelly DA, Tanner MS, et al. Neonatal liver trans- hepatitis. Eur J Pediatr 1990;149:555Y9. plantation for fulminant hepatitis caused by herpes simplex virus 7. Greenes DS, Rowitch D, Thorne GM, et al. Neonatal herpes type 2. J Pediatr Gastroenterol Nutr 2002;35:220Y3. simplex virus infection presenting as fulminant liver failure. 25. Kimberlin DW, Lin CY, Jacobs RF, et al. Safety and efficacy Pediatr Infect Dis J 1995;4(3):242Y3. of high dose intravenous acyclovir in the management of neo- 8. Lafferty WE, Downey L, Celum C, et al. Herpes simplex virus natal herpes simplex virus infections. Pediatrics 2001;108: type 1 as a cause of genital herpes: impact on surveillance and 230Y8. prevention. J Infect Dis 2000;181:1454Y7. 26. American Academy of Pediatrics, Committee on Infectious Dis- 9. Ross JD, Smith IW, Elton RA. The epidemiology of herpes eases. Herpes simplex. In: Pickering LK, ed. 2000 Red Book, simplex types 1 and 2 infection of genital tract in Edinburgh 25th ed. Elk Grove Village, IL: American Academy of Pediatrics, 1978Y1991. Genitourin Med 1993;69:381Y3. 2000:309Y18.

J Pediatr Gastroenterol Nutr, Vol. 42, No. 3, March 2006

Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.