Biliary Atresia as a Disease Starting In Utero: Implications for Treatment, Diagnosis, and Pathogenesis

Abstract

Biliary atresia (BA) is the most common reason for pediatric liver transplant. BA’s varied presentation, natural history, and treatment with the KP have been well described; however, when BA starts relative to birth has not been clearly defined. In this review, we discuss laboratory, imaging, and clinical data which suggest that most if not all forms of BA may start before birth. This early onset has implications in terms of delivering treatments earlier and identifying possible factors underlying BA’s etiology.

Introduction

Biliary atresia (BA), a disease characterized by obstruction of the extrahepatic bile ducts, is generally thought to encompass a spectrum of diseases rather than a single disease entity. BA’s most common presentation was described as early as 1817 in the Western medical literature, in an infant with “jaundice…apprehend[ing] some incurable state of the biliary apparatus.”¹ Since then, many diverse BA presentations have been recognized, including with gall bladder abnormalities, bile duct cysts, cytomegalovirus (CMV) infection, prematurity, syndromic features, and/or laterality defects. In addition, to explain these variations, different etiologies such as genetic changes, infections, autoimmune dysregulation, or toxins have been proposed.

However, despite this diversity, all BA forms are united by key features. First, all BA cases have atretic extrahepatic bile ducts which obstruct bile flow out of the liver. Relieving this obstruction is the focus of the Kasai portoenterostomy (KP), a procedure developed in the 1950s which connects the liver and intestines directly in an attempt to establish bile flow.² Second, all untreated BA cases develop cirrhosis, progressing to end-stage liver disease and need for liver transplant within the first two years of life. Even with treatment, disease progression often continues and explains why BA is the most common indication for liver transplant in pediatrics despite having an incidence of only 1 in 5,000-20,000 (Table 1).³

Table 1.

Leading indications for pediatric transplantation in US, 2009-2018

Liver Transplant
Age ≤1 year
1. BA	59.0%
2. Hepatoblastoma	2.7%
3. Idiopathic Acute Liver Failure	2.3%
4. TPN-Associated Liver Disease	2.2%
5. Alpha-1-Antitrypsin Disease	1.2%
Age 0-17 years
1. BA	30.7%
2. Hepatoblastoma	6.4%
3. Idiopathic Acute Liver Failure	4.0%
4. TPN-Associated Liver Disease	3.1%
5. Alagille Syndrome	2.9%
All Solid Organ Transplant (Heart, Lung, Liver, Kidney, Small Bowel, Pancreas, Intestine)
Age ≤1 year
1. BA	31.0%
2. Congenital Heart Defect†	24.8%
3. Dilated Cardiomyopathy	13.7%
4. Hepatoblastoma	1.4%
5. Idiopathic Acute Liver Failure	1.2%
Age 0-17 years
1. Congenital Heart Defect†	9.7%
2. BA	9.2%
3. Dilated Cardiomyopathy	8.4%
4. Renal Dysgenesis/Agenesis	5.3%
5. Focal Glomerular Sclerosis	4.5%

^*Data collected from United Network for Organ Sharing online publicly-available database.³

^†Includes all congenital heart defects with or without prior surgery.

In this review, another possible unifying factor for BA is explored: time of onset relative to birth. We use laboratory, imaging, and clinical data to argue that most if not all cases may start before birth. We then discuss the implications of an in utero onset, in terms of delivering treatments earlier and identifying possible factors underlying BA’s etiology.

Timing BA’s Onset

“Congenital” BA “Congenital” forms of BA, which account for approximately 16% of cases, are thought to start before birth.^4,5 These BA forms are associated with major malformations (such as congenital heart disease, kidney disease, or genetic syndromes) with or without laterality defects (such as splenic abnormalities). One suggestion is that these BA forms begin in the first trimester, because bile duct development and many of BA’s laterality defects occur at this time. For example, bile ducts first develop as an outgrowth of the ventral foregut at gestational age (GA) 4 weeks, then remodel at the porta hepatis to form a passage to the liver at GA 9 weeks, and finally begin to support bile flow at GA 12-13 weeks.⁶ During the same period, laterality defects are thought to occur, such as development of left-right axis reversal (GA 2-3 weeks), polysplenia/asplenia (GA 3-6 weeks), pre-duodenal portal vein (GA 4-8 weeks), and interrupted vena cava (GA 6-8 weeks).

“Acquired” BA It is tempting to think the remaining 84% of BA cases—the “acquired” forms—start at a time after birth. First, infants with “acquired” BA usually appear healthy as newborns. Only after the first weeks or months of life do they develop concerning symptoms such as persistent icterus and/or pale stools. In line with a postnatal acquisition, twins may also appear healthy at birth. However, later, one twin (but not the other) develop signs of disease.⁷

Second, the most accepted animal model of BA is caused by a postnatal insult. This model involves infecting newborn BALB/c mouse pups with Rhesus rotavirus (RRV) Type A via intraperitoneal injection.⁸ Within days, the virus localizes to bile duct cells and promotes immune-mediated bile duct injury. Pups quickly develop jaundice and pale stools, and their liver histology shows changes which mimic some of the changes seen in livers from humans with BA. Hence, researchers have speculated that because a postnatal insult causes BA-like disease in mice, a postnatal insult could also cause BA in humans.

Third, signs of liver disease only occur after the first weeks of life. For example, common liver injury makers such as aspartate aminotransferase (AST) and alanine aminotransferase (ALT) are normal or near-normal initially.⁹ AST and ALT subsequently rise to exceed reference intervals by four weeks of life, signifying the eventual development of liver injury. Similarly, liver fibrosis indicators such as liver stiffness (as measured by transient elastography) are normal in the first weeks of life.¹⁰ Only at later time points does liver stiffness exceed reference intervals, suggesting that the progressive liver fibrosis characteristic of BA begins in the postnatal period.

Biliary Changes as a Marker of Disease Onset

Timing BA’s onset with elevated AST, ALT, or liver stiffness assumes that liver injury is the primary insult. However, in BA, the primary insult is obstruction of the extrahepatic bile ducts. Obstruction leads to bile retention in the liver, which subsequently leads to hepatocellular injury. Bile-induced liver injury may be limited initially, because of lower bile production or in utero placental clearance of bile acids by the mother. Bile-induced liver injury would increase after birth, when bile acid production increases in response to feeding. With this reasoning, onset of extrahepatic bile duct obstruction—not onset of liver injury—may be a better indicator of when BA starts.

Multiple lines of evidence suggest that extrahepatic bile duct obstruction, and hence “acquired” BA, starts before birth:

1. Infants with BA have elevated direct or conjugated bilirubin (DB or Bc) levels at birth.

Because DB or Bc rises with biliary obstruction, elevated DB/Bc levels can be used to time onset of BA. In an initial study, DB/Bc levels were elevated in 34 newborns who were later diagnosed with “acquired” BA (earliest measurement was at 1 hour of life).¹¹ The Bc levels in these infants continued to rise in the first few days after birth. Subsequent studies confirmed that newborn DB/Bc levels were high in diverse forms of BA, including those occurring in twin pregnancies, premature infants, and infants with syndromic/laterality defects (CMV infection was not determined in these studies).^12,13 Based on these birth DB/Bc observations, we conclude that obstruction is present in utero in all or almost all infants later diagnosed with BA. We have yet to encounter an exception, suggesting that any infants with normal newborn DB/Bc levels later diagnosed with BA would comprise an uncommon subgroup.

2. Infants with BA have biliary abnormalities on fetal ultrasound.

Abdominal ultrasound (AUS) can also be used to time onset of biliary abnormalities in BA. Numerous studies have documented biliary abnormalities in the fetus, supporting the notion that BA starts before birth. For example, in one case report, a routine fetus AUS at GA 21 found an abnormal biliary cyst.¹⁴ After birth, the infant was found to have obstruction of extrahepatic bile ducts and diagnosed with cystic BA. Other infants later diagnosed with cystic BA had biliary cysts found by fetal AUS at GA 19-32 weeks.¹⁵

Fetal AUS also has identified gall bladder abnormalities in infants later diagnosed with BA. These studies are performed routinely in some countries, though identification of the gall bladder is not part of the fetal ultrasound algorithm in the United States. In one Israeli study, a review of second trimester AUS images from 25 patients with BA identified gall bladder abnormalities in six cases.¹⁶ In another French study, AUS at GA 22 weeks of a twin pregnancy detected gall bladder abnormalities in one fetus but not the other.¹⁷ The fetus with gall bladder abnormalities was later found to have BA, whereas the other twin had an uneventful infancy. A third AUS study reported gall bladder abnormalities as early as GA 15 weeks in a fetus who was thought to have BA based on autopsy of the terminated pregnancy.¹⁸

3. Infants with BA have abnormal gamma-glutamyl transferase (GGT) levels in their amniotic fluid.

GGT levels in the amniotic fluid can help identify early biliary obstruction. With normal biliary development, GGT levels are elevated at GA 18-19 weeks. At this early stage before the anal sphincter is fully developed, the gut is in continuity with the amniotic sac. GGT, originating in the liver and passing through the biliary system into the intestines, can flow freely to the amniotic fluid. With obstruction, on the other hand, GGT cannot pass out of the biliary system. Hence, low GGT amniotic fluid levels at GA 18-19 suggest very early biliary obstruction.

Studies have reported low amniotic fluid GGT levels in infants later diagnosed with various conditions affecting the biliary tract, including cystic fibrosis, Trisomy 21, and BA.¹⁹ For example, one report identified three fetuses with low amniotic fluid GGT levels at GA 18-19.²⁰ After birth, two of the cases were confirmed to have BA, whereas the third case had severe common bile duct stricture. Similarly, amniotic fluid GGT levels were sampled at GA 18-19 weeks from the twin pregnancy with gall bladder abnormalities mentioned earlier.¹⁷ Levels were appropriately elevated in the unaffected twin. In contrast, levels were inappropriately low in the affected twin with BA, suggesting biliary obstruction was present at or before GA 18-19 weeks.

4. BA is more common in premature infants.

Multiple studies have reported a higher incidence of prematurity in infants with BA.^21,22 While the mechanism is unknown, the findings support the notion that biliary obstruction develops at some time during fetal development. One explanation is that biliary obstruction occurs very early, leading to fetal distress and a higher likelihood of preterm birth. An alternative explanation is that infants are born prematurely and then “acquire” BA in a critical window when the biliary system is especially vulnerable. During this time window, most infants would still be developing in utero.

Implications of an in Utero Onset

Treatment

The most clinically-relevant implication of an in utero onset is that treating all infants with BA shortly after birth is possible and attainable. Studies from around the world correlate KPs performed at ≤30 days of life (DoL) with the best outcomes (Figure 1). In a Canadian study (n=312), KPs performed at ≤30 DoL (n=21) required the fewest liver transplants over a 10 year follow-up period.²³ Similarly, in a French study (n=695), KPs performed at ≤30 DoL (n=59) were associated with lowest need for transplant over a 15-year period.²⁴ The one exception is from a United States site (n=82), where KPs performed ≤30 DoL (n=9) had poorer outcomes.²⁵ Whether these nine infants had more severe disease, and hence were identified earlier, was not clear.

Figure 1.

Earlier KPs correlate with higher transplant-free survival. Age cut-offs among studies varied, but all showed better outcomes when the KP was performed earlier. Studies included were those with ≥100 total participants and ≥10 participants with KP ≤30 days. Percentages based on data from Karrer FM et al. (1990) Journal of Pediatric Surgery 25:1076-80, Schreiber RA et al. (2007) Journal of Pediatrics 151:659-65, Serinet M-O et al. (2009) Pediatrics 123:1280-86, and Shneider BL et al. (2006) Journal of Pediatrics 148:467-74.

Notably, even before these studies, Morio Kasai had argued for performing the procedure earlier. Kasai reasoned that intervening early would capitalize on the microscopic ducts at the liver hilum in BA, through which the KP depends on for bile flow.² Waiting longer, in contrast, risked destroying these ducts and losing any conduit for bile flow out of the liver. United States surgeons were initially doubtful that microscopic ducts could support bile flow, and generally opted to delay intervention with the hopes that the ducts would increase in size. Given this skepticism, United States surgeons did not fully adopt the KP until nearly 20 years after its development, a conversion the future Surgeon General colorfully recounted in a 1975 editorial.²⁶

Similar to surgical treatment, medical treatment may also be more successful if given earlier. Most post-KP medical treatments focus on limiting or reversing liver injury; hence, it follows that they may be more efficacious when there is less initial liver injury. Table 2 lists current medical therapies that have been or are being investigated in registered clinical trials. Steroids had no benefit in a multi-center, double-blind, placebo-controlled randomized trial, though small observational studies argue DB/Bc clearance may be improved in a subset of patients.²⁷ Similarly, IVIg after KP did not improve DB/Bc clearance or short-term transplant-free survival.²⁸ Other therapies aim to reduce bile-induced liver damage, by decreasing bile acid production, decreasing the bile acid pool, or increasing bile flow. One potential confounder is that these studies often include infants up to 90 DoL, which may mask potential benefits in younger infants with less liver injury at the outset.

Table 2.

Current Investigative Therapies for BA

Therapy (Trial ID*)	Class	Rationale
Lactobacillus casei rhamnosus ()	Probiotic	Colonizes intestine post-KP → prevents growth of harmful bacteria → prevents occurrence of ascending cholangitis
Steroids (, )	Immunosuppressant	Antagonizes immune system → reduces immune-mediated bile duct injury → intact bile ducts allow for more bile flow
Desflurane, Sevoflurane ()	Anesthetic	Compares anesthetics which can induce liver damage → less liver → better post-operative course
Pentoxifylline ()	Methylxanthine	Non-selectively antagonizes phosphodiesterase/adenosine signaling → modulates immune system
Intravenous Immunoglobulin ()	Immunomodulant	Modulates immune system → reduces immune-mediated bile duct injury → intact bile ducts allow for more bile flow
Vancomycin ()	Antibiotic	Affects the microbiome → microbiome changes lead to beneficial effects in the liver → less liver injury
Meloxicam ()	Anti-inflammatory	Antagonizes COX-2 → reduces inflammatory response after KP → prevents liver fibrosis
GCSF ()	Immunomodulant	Stimulates hematopoietic stem cell (HSC) production → HSCs engraft in the liver → HSCs reduce fibrosis and cirrhosis
Bone Marrow Mononuclear Stem Cells ()	Stem cells	Generate liver stem cells → stem cells populate injured liver → hepatocytes and bile duct cells made
N-Acetylcysteine ()	Choleretic, Anti-oxidant	Stimulates glutathione production in hepatocytes → glutathione secreted into bile → bile flow stimulated via osmosis
Obetocholic Acid (EudraCT 2014-004693-42)	FXR agonist/bile acid analogue	Stimulates FXR signaling → multiple responses in the liver, including repression of bile acid synthesis and liver fibrosis

^*NCT trials registered at clinicaltrials.gov; EudraCT trial registered at EU Clinical Trials Register

Diagnosis

While early KPs correlate with better outcomes, KPs are performed on average after 60 DoL in the United States.²⁹ The problem is not with surgeon reluctance as it may have been in the 1950s and 1960s; rather, the problem is with primary care physicians (PCPs) and specialists identifying infants in a timely manner. As mentioned earlier, PCPs face the challenge of quickly referring infants who may appear normal at birth. Specialists must evaluate younger infants who may have normal markers of liver injury, such as AST, ALT, and liver stiffness. In both cases, focusing on biliary changes which are already present at birth can help accelerate the BA diagnosis (Figure 2).

Figure 2.

Approach to diagnose BA. PCPs use clinical signs and/or screening to identify and refer infants. Specialists then perform a series of tests to exclude BA, grouped as commonly ordered tests (a normal result does not necessarily exclude BA), exclusionary tests (a normal result excludes BA), and tests for other diseases. If BA cannot be excluded, invasive gold-standard testing is performed. *tests which detect biliary abnormalities, which can help in early diagnoses; ^†screening performed by parents as well; ^‡ tests which detect liver injury; ^§must be performed expeditiously.

PCPs can screen infants for signs of Bc retention, secondary to bile duct obstruction that starts developing in utero. In Japan, Taiwan, and Canada, PCPs along with parents screen for pale stools with the stool color card and/or a mobile device application.^30-32 Pale stools occur when gut bacteria have no Bc to metabolize, and are present in 77%, 83%, and 97% of infants with BA by 30, 45, and 60 DoL, respectively.^30,31 Another approach is to screen newborns with DB/Bc measurements, which are elevated at birth in BA.^12,13 Larger studies are needed to assess newborn DB/Bc screening’s efficacy, as well as further explore DB/Bc screening’s cost-effectiveness in light of initial modeling of high costs compared to potential benefits.³³ Newborn DB/Bc testing would be an addition to current recommendations in the United States, which call for DB/Bc testing only in infants appearing jaundiced at the 2-3 week well-child visit.³⁴ These recommendations are limited because: (1) jaundice is not always obvious at 2-3 weeks in BA, and (2) jaundice is also present in approximately 15% of healthy newborns.

By focusing on signs of biliary injury which are already present at birth, specialists can perform tests which can exclude BA in the first weeks of life. For example, DB/Bc levels are always elevated with biliary obstruction, so a normal DB/Bc in the newborn period makes BA very unlikely. Elevated DB/Bc levels have two caveats in their interpretation. First, DB/Bc levels can be slightly elevated initially, rise, plateau, and then slightly decrease in BA.³⁵ Hence, elevated DB/Bc levels refer to values above an upper limit of normal (ULN) derived with standard methods from a healthy population, rather than absolute ULNs as suggested in some algorithms.³⁶ Second, while a normal DB/Bc at any point may exclude BA, consistently elevated DB/Bc values occur in BA as well as other diseases. Slightly elevated DB levels can also occur in infants without liver disease, because DB assays are less precise than Bc assays and will also detect delta and some unconjugated bilirubin.³⁷

Imaging can also help identify infants who may have BA, by detecting defects in the biliary system that developed before birth. For example, AUS identifies gall bladder abnormalities and/ or biliary cysts. Importantly, while gall bladder abnormalities are common, BA may present with a normal gall bladder in a minority of cases.³⁸ In some cases, the gall bladder is intact or only slightly affected, and the biliary obstruction is located proximally at the hepatic duct. Magnetic resonance cholangiopancreatography (MRCP) is another noninvasive modality that may provide better images of the biliary tree but requires anesthesia to perform.³⁹

Functional tests which assess bile duct patency can exclude BA in younger infants who have bile duct obstruction but few other symptoms. For example, BA’s gold-standard test—the intraoperative cholangiogram (IOC)—directly outlines the biliary tract by surgically isolating the gall bladder and injecting it with dye. An abnormal IOC diagnoses BA, but the test is highly invasive and reserved for infants with a high suspicion of having BA. Less invasive tests assessing patency which can be performed before IOC include the hepatobiliary iminodiacetic acid (HIDA) scan (follows a tracer from liver to intestine), endoscopic retrograde cholangiopancreatography (ERCP) (injects dye into the ampulla of Vater via endoscopy), and the percutaneous cholangiogram (PTC) (injects dye into the gall bladder via a percutaneous approach). Patency can also be assessed indirectly by measuring bilirubin or bile acids in the intestine (duodenal aspirates), with presence of bilirubin or bile acids inconsistent with BA.⁴⁰ A normal HIDA scan, ERCP, PTC, or duodenal aspirate can exclude BA and make an IOC unnecessary. However, the tests have important limitations. HIDA scans have a high false positive rate and consume time if pretreatment with agents that stimulate bile flow are required. ERCP, PTC, and duodenal aspirates are challenging to perform, and ERCP and PTC require anesthesia.

In addition to focusing on biliary abnormalities, two additional approaches are used to diagnose BA. First, tests are performed to identify other causes of cholestasis, and if performed quickly, can help to efficiently exclude BA. For example, alpha-1-antitrypsin disease can be diagnosed by protease inhibitor typing, Alagille Syndrome by butterfly vertebrate on chest radiography and genetic testing, choledochal cysts by AUS, and diseases such as cystic fibrosis and galactosemia by state newborn screens. Looking for alternative diseases assumes infants will not have BA and changes found in other conditions. The assumption may not be entirely correct, however, as patients with BA have also been found to have gene mutations associated with other conditions such as Alagille Syndrome.⁴¹

Second, signs of hepatocellular changes are often used to evaluate infants for BA, though, as discussed earlier, may not be as apparent in younger infants. For example, AST, ALT, and liver stiffness are normal in the first weeks of life.^9,10 Serum matrix metalloproteinase 7 (MMP7) is another liver marker, which has excellent sensitivity and specificity for BA as demonstrated in a series of recent papers.^42-44 Most of these studies focused on older infants, and how MMP7 functions in infants younger than 30 DoL may be the subject of future investigations. The triangular cord sign is an additional liver marker suggestive of BA. The triangular cord is a >4 mm echogenic focus adjacent to the portal vein detectable by AUS. Recent studies suggest that the triangular cord may thicken over time in BA as more injury accrues, and as a result may be less commonly found in younger infants.⁴⁵

Liver biopsy is the most direct way to detect hepatocellular changes that occur secondary to bile duct obstruction. Liver biopsy in BA shows bile duct proliferation, bile duct plugging, portal fibrosis, and/or portal edema. Unlike other markers of liver injury, liver biopsy changes do not appear to be affected by age, though some have cautioned that changes may be less apparent in very young infants^34,46,47 The main limitation of liver biopsies is that they depend on a pathologist’s interpretation, which studies have shown can be quite variable.⁴⁸ As a result, some pathologists may be more stringent, and interpret a biopsy as consistent with extrahepatic obstruction only if many features are present. This raises the possibility of “false negative” reads, and cautions against using interpretations of liver injury on biopsy (at any age) as a foolproof way to exclude BA.

Etiology

An in utero onset also has implications for understanding the etiology of BA. To date, theories that have been proposed include genetic changes, infection, immune dysregulation, and/or toxins, as described in many excellent reviews on the subject.^49-51 BA’s etiological factor(s) must explain a disease that does not follow familial inheritance patterns and that is discordant in twins (including identical twins). In addition, if BA starts before birth at the level of the extrahepatic bile ducts, any etiological factor(s) must also target the biliary system in utero (Figure 3).

Figure 3.

Etiological Considerations for BA. By the second trimester, BA has already started as evidenced by gall bladder abnormalities and low amniotic fluid GGT levels. A BA pathogenic factor affects one but not both fetuses in twin pregnancies (affected fetus with atretic GB and low GGT in amniotic fluid). Proposed pathogenic factors include genetic changes, infection, immune dysregulation, and/or toxins.

Genetic changes could cause defects in bile duct development during development. For example, PKD1L1 variants which are predicted to affect protein function were recently identified in syndromic forms of BA. PKD1L1 controls cilia development, consistent with the hypothesis that cilia defects impair both bile flow and left-right axis determination.⁵² Similarly, aneuploidy of chromosome 22 (Cat’s eye syndrome) is associated with BA, further pointing to a genetic etiology.⁵³ Other clues for a genetic etiology include BA’s 3:2 female:male occurrence, and a higher incidence of polymorphisms in Adducin 3.^54,55 A genetic etiology, however, does not fully explain BA discordance in identical twins, unless mechanisms such as somatic mutations are also involved.

Infections could also cause bile duct obstruction in utero. One candidate virus is CMV, based on patients with BA that also have CMV nucleic acid in liver tissue, CMV-responsive T cells in blood, and/or anti-CMV immunoglobulins in serum.^56,57 CMV infections causing BA would differ from typical congenital CMV infections. First, they would spare the fetus of the intrauterine growth restriction, neurodevelopmental abnormalities, and hearing loss characteristic of congenital CMV infections. Second, in twin pregnancies, they would only affect one fetus whereas congenital CMV infections can affect both fetuses. Importantly, in utero CMV infections causing BA may respond to postnatal anti-viral therapy, as suggested in a recent study describing experiences at a large United Kingdom center.⁵⁸

Immune dysregulation in the form of autoimmunity seems less likely to initiate a focused attack of the biliary system in utero. This is because the fetal immune system generally exists to develop tolerance to maternal antigens rather than mount responses against antigens. Only in rare and severe diseases, such as in the Immune Dysregulation, Polyendocrinopathy, Enteropathy X-linked (IPEX) Syndrome, has fetal autoimmunity been described.⁵⁹ However, autoimmunity may drive disease progression postnatally, after the initial bile duct obstruction has occurred. Consistent with this, BA serum and liver studies show signs of autoimmunity, including decreased T regulatory lymphocytes and high expression of Th1 and Th2 cytokines.⁶⁰

In contrast, immune dysregulation in the form of alloimmunity could cause bile duct obstruction before birth. In this model, maternal antibodies cross the placenta and cause direct destruction of the fetal extrahepatic bile duct. Alloimmunity is supported by the finding of significantly more maternal T cells in the portal areas of BA liver biopsies versus biopsies from other pediatric diseases.⁶¹ However, one problem with the alloimmunity hypothesis is that BA occurs independent of an infant’s birth order. Other alloimmunity diseases, on the other hand, spare the first pregnancy (while the mother develops alloimmunity antibodies) and affects the remaining pregnancies.

Finally, toxins are an etiological factor that could explain bile duct damage in utero. A toxin—biliatresone—is responsible for causing BA-like disease in Australian sheep.⁶² Pregnant sheep feed on the biliatresone-containing Dysphania weed during drought seasons, and give birth to offspring with bile duct damage reminiscent of BA. Subsequent studies have shown that biliatresone damages bile duct cells in mice and zebrafish, via mechanisms antagonized by glutathione.^63,64 One unresolved issue with a maternally-ingested toxin such as biliatresone is that it would affect both fetuses in twin pregnancies, unlike BA which is discordant in twins.

Conclusion

Though BA is a diverse disease, BA has a few important unifying characteristics. One such characteristic is that most if not all forms of BA may start before birth. An in utero onset argues that treatment before 30 DoL is possible, which in turn should help delay or even prevent need for liver transplant. An in utero onset also provides guidance on how to diagnose BA earlier, by focusing on signs of bile duct obstruction rather than secondary signs of liver injury. Finally, an in utero onset adds to our understanding of the elusive factor(s) that cause bile duct injury before birth and trigger what ultimately in the postnatal period is identified as BA.

What is Known:

• Many forms of biliary atresia (BA) are thought to start after birth.

• Signs of liver injury, which are used to diagnose BA, only become apparent in the first weeks to months of life.

• As result, treatment with the Kasai portoenterostomy (KP) is after 60 days of life (DoL), on average.

What is new:

• Most if not all forms of BA may start before birth, based on laboratory, imaging, and clinical data.

• Given an in utero time of onset, earlier treatment with the KP, i.e., before 30-45 DoL, is possible and attainable.

• Signs of biliary obstruction, rather than liver injury, can be used to identify BA earlier.

• Etiological factor(s) for BA should account for potential biliary damage in utero.

Abbreviations:

BA

Biliary atresia

CMV

cytomegalovirus

KP

Kasai portoenterostomy

GA

Gestational age

RRV

Rhesus rotavirus

AST

Aspartate aminotransferase

ALT

Alanine aminotransferase

DB

Direct bilirubin

Bc

Conjugated bilirubin

AUS

Abdominal ultrasound

GGT

Gamma-glutamyl transferase

DoL

Days of life

PCPs

Primary care providers

ULN

Upper limit of normal

MRCP

Magnetic resonance cholangiopancreatography

IOC

Intraoperative cholangiogram

HIDA

hepatobiliary iminodiacetic acid

ERCP

Endoscopic retrograde cholangiopancreatography

PTC

Percutaneous transhepatic cholangiogram

MMP7

Matrix metalloproteinase 7