Biliary Atresia in Children: Update on disease mechanism, therapies, and patient outcomes

Introduction

Biliary atresia (BA) is a progressive, obstructive cholangiopathy with neonatal onset. While BA is a rare disease, it remains the leading indication for pediatric liver transplantation as there are no effective medical therapies to prevent or slow disease progression. The prevalence of BA varies by geographic region with an incidence of 1 in 6,000 to 8,000 live births in Taiwan and up to 1 in 19,000 live births in Canada with intermediate rates reported in other areas of the world.^1–3 A higher incidence of BA has been reported in patients with female versus male gender, premature infants, and in Asian and black versus white infants.^4–7 BA is most often an isolated defect but can be associated with other congenital abnormalities in up to 16% of cases of which more than half commonly have laterality defects (particularly splenic anomalies) and are referred to as syndromic BA.⁸ In the present review, we highlight disease biomarkers and therapies as they relate to our understanding of BA pathogenesis and outcomes.

Definition and diagnosis of biliary atresia

BA is the most common cause of obstructive cholestasis in the neonate in which infants develop scleral icterus, clay-colored acholic stools, and jaundice that persists beyond the first two weeks of life. Direct or conjugated bilirubin remains the primary screening laboratory test for BA, with elevated values shown to occur within the first 2 days of life.^9,10 Levels of gamma-glutamyl transferase (GGT) and matrix metalloproteinase-7 (MMP-7), a marker of biliary epithelial injury, also have an established role in the diagnosis of BA, and new candidate biomarkers continue to emerge (Box 1).^11–22 Imaging by abdominal ultrasound is considered one of the first-tier tests in a targeted investigation of BA to identify alternate anatomic etiologies of obstructive cholestasis and may also detect features supportive of BA (e.g. abnormal gallbladder, absent common bile duct, vascular anomalies such as preduodenal portal vein, polysplenia or asplenia, triangular cord sign).⁹ Other imaging modalities such as hepatobiliary scintigraphy, magnetic resonance cholangiopancreatography, endoscopic retrograde cholangiopancreatography, and percutaneous transhepatic cholangiogram may also be considered with varying levels of sensitivity and specificity.^9,23

Box 1: Serum biomarkers in BA.

Biomarker	Sensitivity (Sn) and Specificity (Sp)
Direct or conjugated Bilirubin screening algorithm10	Sn 100%, Sp 99.9%
GGT > 250 IU/L11	Sn 83%, Sp 71 %
GGT > 300 IU/L12	Sn 40%, Sp 98%
Age 61–90 days with GGT > 303 IU/L13	Sn 83%, Sp 82%
GGT/AST > 212	Sn 81%, Sp 72%
MMP-714,15	Sn 96%, Sp 91%
Candidate biomarkers:  Bile acid profiles18,19,22  MicroRNAs20,21  Cytokines: IL-33, IL-18, IL-815–17	NA

Liver biopsy (Figure 1) remains a central step in diagnosing BA although incorporation of new biomarkers or laparoscopic evaluation may reduce the need for liver biopsy before intraoperative cholangiogram (IOC).^9,15,23–26 Once a high index of suspicion is established, IOC is performed and is considered positive if a patent extrahepatic biliary tree is not visualized. Subsequently, Kasai Portoenterostomy (KPE) is performed as the primary treatment to promote biliary drainage. IOC paired with KPE also helps to define the specific anatomic variant of BA that may have prognostic implications.²⁷

Figure 1.

Liver biopsy from an infant with BA at diagnosis that underwent liver transplant before 2 years of age. Histologic findings include expansion of the portal tract with stromal edema, prominent fibrosis, bile duct proliferation (black arrows), and bile plugs (white arrows). White star delineates a branching artery.

Pathophysiology of Disease Onset and Progression

Our understanding of early disease pathogenesis suggests that susceptible infants with possible genetic predisposition experience an in-utero insult, leading to stimulation of innate and adaptive immunity, bile duct injury with impaired bile flow, accumulation of bile acids, and epithelial damage with alteration of intercellular junctions. While the exact triggering event remains unknown, evidence supports an in-utero onset of BA including identification of prenatal gallbladder anomalies, abnormally low GGT levels in amniotic fluid, and the presence of elevated direct/conjugated bilirubin levels within 24–48 hours of birth.^28,29 Ongoing research suggests that there is variable contribution among the dominant contributing factors of disease pathogenesis thereby leading to different BA phenotypes despite a common clinical presentation.

Environmental factors

A viral trigger causing biliary obstruction in-utero has long been considered a possible etiological factor for the development of BA. The well-established murine model of BA relies on rhesus rotavirus inoculation within the first 24 hours of life.³⁰ Parallel studies in humans suggest some, but not all, cases of BA may be associated with a viral insult. Viruses implicated in BA include cytomegalovirus (CMV), reovirus, rotavirus, Epstein-Barr virus, and human papillomavirus.³¹ CMV has been the most frequently analyzed virus with positive rates of detection among BA patients varying by seasonality but overall ranging from 20–60%.³¹ A role for the gut microbiome in the pathogenesis of BA is also emerging although further studies are needed to define how precise microbial metabolites or microbial-driven immune modulation are implicated in BA and if probiotics may be beneficial.³²

After outbreaks of a disease similar to BA were reported in newborn lambs in Australia, a previously unknown Dysphania plant isoflavonoid, now named biliatresone, was hypothesized to be the suspected agent.³³ Murine extrahepatic bile ducts treated with biliatresone demonstrated disruption of cholangiocyte apical polarity, loss of monolayer integrity, increased permeability, and luminal obstruction mediated by decreased levels of glutathione and SOX17.³⁴ A more recent in vivo study of mice injected with biliatresone confirmed biliary obstruction with associated inflammation and fibrosis.³⁵ Additionally, these BA mice demonstrated evidence for oxidative stress including altered hepatic glutathione levels similar to sequencing analysis in human tissues implicating glutathione metabolism in BA outcomes.^35,36

Immune factors

Various immune mechanisms have been proposed including immune dysregulation, autoimmunity, and susceptibility of the neonate’s immature immune system. While the precise interplay between these processes has not been fully established, extensive research in the field has identified key components of the innate and adaptive immune responses in both the early and late stages of BA (Figure 2).^37,38 Innate immunity plays a key role in the early immune response to pathogens through pathogen recognition receptors (PRR), in particular toll-like receptors (TLRs) (Figure 2A). TLRs are upregulated in BA and recognize pathogen associated molecular patterns (PAMPs) or stimuli released from apoptotic/necrotic cells, i.e. damage associated molecular patterns (DAMPs).³⁹ Upon TLR activation type 1 interferons are released and initiate a complex cascade of immune signaling by tumor necrosis factor alpha (TNF-α), IL-1, IL-6, IL-8, and IL-15.^37,38 Cholangiocytes, macrophages, and dendritic cells play a primary role in this early immune response that results in neutrophil recruitment and activation of adaptive immunity.

Figure 2.

A. Early immune mechanism of BA is characterized by activation of innate immunity via PAMPs and DAMPs with subsequent stimulation of the adaptive immune response. B. Later stages of BA exhibit a Th2/Th17 phenotype with IL-33/IL-13 driven bile duct proliferation and fibrosis.

CCL2 – C-C motif chemokine ligand 2; DAMP - damage associated molecular pattern; DC – dendritic cell; ECM – extracellular matrix; IFN – interferon; IL – interleukin; NK – natural killer cell; PAMP - pathogen associated molecular pattern; TLR – toll-like receptor; TNF - tumor necrosis factor; Tregs - T regulatory cells

Within the adaptive immune arm, oligoclonal expansion of both T cells and B cells has been observed and various autoantibodies have been identified in infants with BA supporting a possible role for antigen-driven immune stimulation.^40–45 Most infants with BA exhibit a dominant Th1 immune response early in disease although Th17 T cell immunity has been demonstrated in various studies and Th2 responses may be involved in cystic BA.^46,47 Human and murine studies have also demonstrated a decrease in frequency and function of T regulatory cells (Tregs), further supporting the premise for immune dysregulation in the setting of immature neonatal immunity.^42,48 Despite evidence of oligoclonality and the presence of various autoantibodies, a critical antigen-independent role for B cells has also been shown.⁴⁹

The hepatic immune response at later stages of BA commonly exhibits a Th2 immune phenotype characterized by low-level inflammation and oxidative injury, bile duct proliferation, and progressive fibrosis (Figure 2B).³⁸ Important at this stage of disease is the hepatocyte-derived alarmin IL-33 that triggers IL-13 driven activation of hepatic stellate cells and progressive hepatic fibrosis.^38,50 Macrophages mediate both the pro-restorative and maladaptive responses in cholestatic liver disease, and distinct transcriptional subsets have been identified in pediatric cholestatic liver disease at the time of transplant.^38,51 Further delineating the role of profibrotic macrophages and the IL-33/IL-13 axis will help identify therapeutic targets and prevent progression to cirrhosis.

Genetic susceptibility

While BA does not follow a mendelian pattern of inheritance and a single genetic variant has not been identified, various studies suggest a role for genetic susceptibility variants. Additionally, HLA associations are not present within BA, supporting alternate genetic influencers of immune function.⁵² Of interest in syndromic BA is identification of variants of PKD1L1, a gene associated with ciliary development and laterality determination.⁵³ Additional ciliary gene defects have been implicated in both syndromic and nonsyndromic BA.⁵⁴ A summary of genes that may be associated with BA are shown in Box 2.^55,56

Box 2: Genes implicated in BA.

Gene	Function
PKD1L1, CFC1, ZIC3, ZEB2, FOXA2, HNF1B, KIF3B, TTC17 53,54,55	Regulation of laterality, ciliogenesis, and development
ARF6, EFEMP155	Cytoskeleton and extracellular matrix modeling
ADD3 55	Cell contact and membrane structure
GPC1, JAG1 55,56	Hedgehog and Wnt signaling pathways
STIP1, REV1 55	Stress response and DNA repair

Much work has also focused on defining transcriptional phenotypes that relate to patient outcomes. Moyer et al characterized BA patients into inflammatory or fibrotic groups by molecular profiling and found the fibrotic gene signature to be associated with reduced transplant-free survival.⁵⁷ Similarly, Luo et al identified a 14-gene signature that predicted 2-year transplant-free survival, in which patients with poor outcome had enrichment for genes involved in fibrosis whereas those with good outcome expressed higher levels of genes involved in glutathione metabolism.³⁶ Overall, these studies suggest that while there is no established causal gene in BA, various susceptibility genes and specific gene pathways may be involved in pathogenesis and outcome. Larger human studies and parallel mechanistic studies in animals will help identify the role of specific genes.

Therapeutic options

Surgical intervention with KPE at the time of diagnosis remains the primary treatment strategy for BA. Due to our inability to prevent ongoing hepatic injury, BA remains the leading indication for pediatric liver transplantation. Isolated liver transplant without prior KPE can also achieve favorable outcomes and may be considered, particularly in BA patients with late diagnoses.⁵⁸ Medical management after KPE includes prevention and treatment of complications such as cholangitis and optimization of nutrition and fat-soluble vitamin supplementation.^59–64 Treatment of additional complications of portal hypertension and timing for liver transplantation parallel standard medical care as in other etiologies of end-stage liver disease and have been reviewed comprehensively by Sundaram et al.⁶⁵ Below we highlight previously studied and emerging medical therapies that target specific pathways of disease pathogenesis.

Investigational medical therapies

Generalized immune modulating therapies previously tested in BA have included steroids and intravenous immunoglobulin (IVIG). Theoretical benefits of using steroids include improving biliary inflammation to promote choleresis, however, their benefit in BA remains unclear. Davenport et al studied the effect of oral prednisolone after KPE and found that while there was improvement in jaundice in the early post-operative period, particularly for infants less than 70 days old, there was no significant reduction in the need for liver transplantation.^66,67 The utility of steroids was most extensively examined in the multicenter, double-blind Steroids in Biliary Atresia Randomized Trial (START), in which infants received a 13-week course of either high-dose steroids or placebo starting within 72 hours after KPE.⁶⁸ Results from START showed no difference in total bilirubin at 6 months post-KPE or 2-year survival with native liver in those that received steroids. Furthermore, the treatment group experienced impaired growth and a shorter time to first serious adverse event.^68,69 While a small clinical benefit could not be excluded by the authors, the results did not establish a role for routine use of high-dose steroids after KPE.

IVIG has been associated with clinical benefits in many inflammatory and autoimmune diseases and can interfere with phagocytosis by cells of the innate immune system, neutralize autoantibodies, and modulate the adaptive immune response.⁷⁰ In murine BA, administration of high-dose immunoglobulin resulted in decreased bilirubin, less bile duct inflammation and obstruction, and lower levels of cytokines associated with CD4+ Th1-mediated inflammation, although overall survival did not differ.⁷¹ However, a prospective multi-center open-label human trial (PRIME) found no improvement in bilirubin levels at 90 days post-KPE or 1-year survival with native liver in participants that received IVIG after KPE compared to a placebo-arm group.⁷²

More recent studies include use of immune cell subset specific therapies in BA. A limited study examining the effect of B-cell depleting agents in BA showed that one dose of rituximab was safe and well-tolerated, however, long term clinical outcomes were not reported.⁴² A possible beneficial role for hematopoietic stem cell recruitment via granulocyte-colony stimulating factor (GCSF) in liver disease has been supported by various studies in adults. Based on this experience, a clinical trial using GCSF in patients with BA is ongoing with initial phase 1 data demonstrating safety and perhaps some improvement in early biliary drainage and frequency of cholangitis.⁷³

Additional therapeutic targets in BA include interrupting ongoing hepatic injury induced through oxidative injury and bile-acid toxicity. N-acetylcysteine (NAC) is an antioxidant that has been shown to improve hepatic injury and fibrosis, reduce biliary obstruction, and increase survival in murine BA.³⁶ A single center, open label, phase 2 trial is currently investigating whether NAC given after KPE may improve bile flow in humans.⁷⁴ Ursodiol has various proposed mechanisms to reduce cholestatic liver injury and is standard supportive medical therapy after KPE. Newer agents that inhibit the ileal apical sodium-dependent bile acid transporter to interrupt enterohepatic bile acid recirculation include maralixibat and odevixibat, both with ongoing clinical trials in BA. ⁷⁵ Lastly, farnesoid x receptor agonists control metabolic homeostasis and inhibit bile acid synthesis and may be considered as future therapeutic options for pediatric cholestatic liver disease.⁷⁵

Predicting Prognosis in Biliary Atresia

Prognosis in BA remains difficult to predict due to disease heterogeneity and contribution of multiple clinical variables. Ongoing work to strengthen screening practices and develop targeted therapies for modifiable clinical factors is needed to further improve patient outcomes.

Patient variables

A defining clinical variable associated with duration of transplant-free survival is the age at which KPE is performed. KPE performed before 30–60 days of life improves outcomes, however, late KPE can still achieve biliary drainage.^2,76 Center expertise, patient anatomy and BA subtype can also impact patient outcome.⁷⁷ BA characterized by atresia of the hepatic duct and porta hepatis (Ohi type II and III) and syndromic BA have been associated with lower rates of transplant free survival.²⁷ Growth failure after KPE is another well-established risk factor for death or transplant and highlights the critical importance of accurate anthropometric measurements to institute early nutritional support.⁶¹ More recently, cardiomyopathy has been identified as a prevalent comorbidity in infants with BA and is associated with higher risk for serious adverse event and peri-transplant death.⁷⁸ In a study population of children with cirrhosis in BA of which 71% of children were less than 1 year of age, presence of ascites, a lower serum sodium, higher bilirubin, and increased Pediatric End-stage Liver Disease score were associated with higher mortality.⁷⁹ A summary of key clinical variables and their impact on prognosis is illustrated in Figure 3.

Figure 3.

Cumulative data on the overall prognosis of BA demonstrates about 50% of patients require liver transplantation by 2 years of age.

CMV – cytomegalovirus; KPE - Kasai portoenterostomy

Biochemical markers

One of the most widely studied factors in the prognosis of BA is the bilirubin level after KPE.⁸⁰ Total bilirubin < 2 mg/dL by 3 months post-KPE is associated with greater transplant-free survival at 2 years of age (86%) compared to those with bilirubin ≥ 2 mg/dL (20%). In addition, infants with a total bilirubin ≥ 2 mg/dL have a higher likelihood of developing complications of liver disease including ascites, hypoalbuminemia, or coagulopathy. Elevated GGT levels and the aspartate aminotransferase-to-platelet ratio index (APRI) after KPE have been associated with a lower rate of survival with native liver.^15,81 Other studied but not yet widely established markers for poor prognosis after KPE in BA include positive CMV serology, high IL-8, reduced number of T regulatory cells, low IL12B levels, increased serum hyaluronic acid, and elevated MMP-7 levels.^81–84

Hepatic phenotype

The extent of hepatic fibrosis at diagnosis of BA has also been associated with worse prognosis. Russo et al identified that a higher stage of fibrosis as well as the findings of ductal plate configuration, and moderate to severe bile duct injury on diagnostic needle liver biopsies correlated with a higher risk of transplantation.²⁵ Non-invasive modalities to measure fibrosis such as APRI or liver stiffness by transient elastography have also demonstrated prognostic utility.⁸¹ Furthermore, transcriptional profiling of the liver in BA demonstrated that a fibrotic gene signature is associated with poor outcome whereas a more inflammatory gene signature and increased gene expression for pathways of glutathione metabolism have been associated with higher rates of transplant-free survival.^36,57 Different hepatic immune phenotypes have also been associated with patient outcomes. For example, high amounts of macrophages and IL-17+ immune cells in the portal tracts of infants with BA have been associated with worse outcome, demonstrating the clinical relevance of different immune phenotypes.^46,85 While specific transcriptional or immune signatures have not been established in clinical use, further research may better define their prognostic role in BA.

Patient Outcomes

Overall, about half of patients with BA require a liver transplant by 2 years of age, and most are transplanted by early adulthood (Figure 3).^86,87 Rates of graft and patient survival in BA after liver transplantation are excellent with a recent report from Society of Pediatric Liver Transplantation (SPLIT) of graft and patient survival at 90% and 97% respectively.⁸⁸ At 10 years post-transplant in the SPLIT registry, more than 80% of patients had normal liver and kidney function, 89% had height z-scores greater than the 3^rd percentile, and 94% had weight z-scores for age greater than the 3^rd percentile.⁸⁹ In addition, there is growing recognition of the need to improve outcomes in patients that survive with their native liver beyond the age of 2 years. Patients surviving without transplant continue to experience complications of chronic liver disease including clinical manifestations of cirrhosis, portal hypertension, cholangitis, and pruritus. Prognostic models have recently been developed to risk stratify poor outcomes in patients with their native liver after the age of 2 years.⁹⁰ Cumulative incidence of liver transplant or death was 23.7% and predicted by increased total bilirubin, low albumin, decreased platelet count, and a history of either cholangitis or ascites. ⁹⁰

Children with BA are also at risk for impairments in neurodevelopmental outcomes and decreased quality of life. Multiple studies have shown variable levels of impairment in motor and/or language skills in infants with BA at diagnosis up to the time of transplant as well as school-age children with or without transplant.^91–93 Importantly, a 2013 cross-sectional study from the Childhood Liver Disease Research Network found that BA patients had poorer health-related quality of life compared to healthy children across all domains.⁹⁴ Similar findings have been validated by other groups and highlight the need for ongoing multidisciplinary support for this vulnerable patient population.

Summary

BA remains the leading cause of neonatal obstructive jaundice and is the most common indication for pediatric liver transplantation. While outcomes after liver transplantation for BA are excellent, there remains an unmet need to develop medical therapies to prolong transplant-free survival and reduce the incidence of adverse events in patients surviving with their native liver. Significant research is needed to establish new therapies for BA that account for the multifactorial nature of disease progression and heterogeneity of immune phenotypes.

Synopsis:

Biliary atresia is a rare disease but remains the most common indication for pediatric liver transplantation as there are no effective medical therapies to slow progression after diagnosis. Variable contribution of genetic, immune, and environmental factors contributes to disease heterogeneity among patients with biliary atresia. Developing a deeper understanding of the disease mechanism will help develop targeted medical therapies and improve patient outcomes.

Clinics care points:

• New diagnostic biomarkers may facilitate earlier diagnosis to improve outcomes after KPE.

• Variable contribution of immune, genetic, and environmental factors accounts for disease heterogeneity in BA and may influence response to medical therapies.

• Risk stratification of outcomes for BA patients before and after transplant is critical to reduce occurrence of serious adverse events, prolong survival with native liver, and improve neurodevelopmental outcomes.

Key points:

• Biliary atresia (BA) remains the leading indication for pediatric liver transplantation as there are no established medical therapies to delay progression of liver disease after Kasai Portoenterostomy.

• Despite a common clinical presentation of disease, the multifactorial nature of disease pathogenesis contributes to phenotypic heterogeneity and variable rate of disease progression.

• There remains an unmet need to develop BA-specific medical therapies and improve outcomes for patients with and without survival with their native liver.

Abbreviations:

APRI

aspartate aminotransferase-to-platelet ratio index

BA

biliary atresia

CCL2

C-C motif chemokine ligand 2

CMV

cytomegalovirus

DAMP

damage associated molecular pattern

DC

dendritic cell

ECM

extracellular matrix

GGT

gamma-glutamyl transferase

GCSF

Granulocyte-colony stimulating factor

IFN

interferon

IL

interleukin

IOC

intraoperative cholangiogram

IVIG

intravenous immunoglobulin

KPE

Kasai portoenterostomy

MMP-7

matrix metalloproteinase-7

NAC - N

acetylcysteine

NK

natural killer cell

PAMP

pathogen associated molecular pattern

PRR

pathogen recognition receptor

TLR

toll-like receptor

TNF

tumor necrosis factor

Tregs

T regulatory cells