Biliary atresia

INTRODUCTION — Biliary atresia (BA) is a progressive, idiopathic, fibro-obliterative disease of the extrahepatic biliary tree that presents with biliary obstruction exclusively in the neonatal period [1]. Although the overall incidence is low (approximately 1 in 10,000 to 20,000 live births [2-7]), BA is the most common cause of neonatal jaundice for which surgery is indicated and the most common indication for liver transplantation in children.

TYPES OF BILIARY ATRESIA — Infants with BA can be grouped into three categories:

●BA without any other anomalies or malformations – This pattern is sometimes referred to as perinatal BA and occurs in 70 to 85 percent of infants with BA [1,8,9]. Typically, these children are born without visible jaundice, but within the first two months of life, jaundice develops and stools become progressively acholic.

●BA associated with laterality malformations – This pattern is also known as biliary atresia splenic malformation (BASM) or "embryonal" biliary atresia and occurs in 10 to 15 percent of infants with BA [8-11]. The laterality malformations include situs inversus, asplenia or polysplenia, malrotation, interrupted inferior vena cava, and cardiac anomalies. Data suggest that children with BASM have poorer outcomes compared with those with perinatal BA, possibly due to the associated cardiac abnormalities [10-12]. (See "Heterotaxy (isomerism of the atrial appendages): Anatomy, clinical features, and diagnosis", section on 'Liver'.)

●BA associated with other congenital malformations – This occurs in the remaining 5 to 10 percent of BA cases; associated congenital malformations include intestinal atresia, imperforate anus, kidney anomalies, and/or cardiac malformations [9,13-15].

Regardless of the type of BA, in each case, the histology and cholangiogram are characteristic. The histology typically shows inflammation, portal tract fibrosis, cholestasis, and bile duct proliferation; the cholangiogram demonstrates loss of patency of the extrahepatic bile ducts.

Because of the cholangiographic findings, BA was previously termed "extrahepatic biliary atresia" to distinguish it from "intrahepatic biliary atresia," which was a histopathologic grouping of disorders in which the intrahepatic bile ducts are primarily affected. However, disorders primarily affecting the intrahepatic bile ducts are now categorized by the genetic, metabolic, or infectious cause rather than by the histologic finding. (See "Causes of cholestasis in neonates and young infants".)

PATHOGENESIS — The cause of BA is unknown, although several mechanisms have been implicated, as outlined below. In some patients, several of these mechanisms may contribute to the development of BA. Conversely, BA may be the common phenotype that can be caused by a variety of injuries to the biliary system occurring in the perinatal period.

Viral etiologies — Clustering of cases of BA in time and space suggest a possible viral etiology. As an example, an analysis of 249 cases of BA over a 16-year period in New York State demonstrated seasonal patterns of incidence that varied by region in the state; in New York City, the risk of BA was highest for infants born during the spring months, whereas outside of New York City, infants born in the autumn months had a higher risk [16]. To date, a specific virus has not been implicated; investigations have failed to identify associations with several specific viral infections including cytomegalovirus, reovirus, and group C rotavirus [17-20]. However, one study reported that infants positive for cytomegalovirus immunoglobulin M (IgM) have reduced clearance of jaundice following Kasai hepatoportoenterostomy (HPE), suggesting that such infants are a distinct prognostic subgroup [21].

Toxic etiologies — The clustering of cases of BA is also consistent with the possibility of a toxin-mediated inflammatory response. The strongest evidence for this hypothesis comes from three reported outbreaks of BA in lambs in Australia in 1964, 1988, and 2007 [22]. In each outbreak, in a time of drought, ewes that gave birth to affected lambs had grazed on lands that had previously been flooded. A significant number of offspring were thin, jaundiced, had acholic stools, and eventually died, and autopsy revealed a diagnosis of BA. The hypothesized mechanism is that the pregnant ewes ingested a toxin when grazing on lands previously submerged. A novel isoflavonoid toxin was isolated from the Dysphania plant, which was harvested in Australia at the site of a recent outbreak. This toxin caused severe damage to the extrahepatic biliary tree in a zebrafish model and also loss of cilia in neonatal mouse cholangiocytes. This evidence suggests that an environmental toxin may be implicated in some cases of BA [23].

Genetic etiologies — Genetic factors may play a causative role in the small subgroup of patients with biliary atresia splenic malformations (BASM), as suggested by the following observations:

●Variants in the CFC1 gene, which is involved in determining laterality during fetal development, have been associated with BASM syndrome. A heterozygous mutation in CFC1 was found in 5 out of 10 infants with BASM [24]. The frequency of this mutation in infants with BASM is twice that found in a population of healthy individuals. Thus, CFC1 mutations may predispose to BASM but are not sufficient to cause the disease.

●Biallelic or heterozygous variants in the PKD1L1 gene (polycystin 1-like 1) were found in 8 of 67 individuals with BASM syndrome [25]. PKD1L1 is expressed in primary cilia and is involved in determination of laterality.

●A heterozygous deletion of the FOXA2 gene (forkhead box 2A) has been reported in a family with heterotaxy, panhypopituitarism, and BA [26].

Other than these subgroups of patients with BASM, genetic factors may not play a direct causative role in the development of most cases of BA; this is suggested by the observation that monozygotic twins usually have a discordant phenotype. Nonetheless, genetics may still play a role in disease susceptibility. Association studies have identified a few genomic loci with increased susceptibility to BA [27-29]:

●EFEMP1 was identified as a candidate susceptibility gene for BA in a genome-wide association study [30]. EFEMP1 encodes an extracellular matrix protein that is expressed in cholangiocytes and portal fibroblasts.

●A risk locus on chromosome 10q24.2 was identified in multiple genome-wide association studies in both Asian and White cohorts [27]. The ADD3 gene (adducin 3) is located in this region, and follow-up studies have demonstrated that variants altering ADD3 expression influence risk for BA [31]. In addition, loss of add3 in a zebrafish model leads to biliary developmental abnormalities [32].

With advances in next-generation sequencing technology, additional candidate genes may be identified in this patient population.

Epigenetic factors have also been postulated as important factors impacting biliary development and the pathogenesis of BA. Both microRNA and DNA methylation have been studied in animal models and humans [33,34]. (See "Causes of cholestasis in neonates and young infants", section on 'Alagille syndrome'.)

Immunologic etiologies — Immune dysregulation, either as a primary disorder or as the result of infectious or genetic triggers, has been implicated in various studies.

●A high concentration of maternal chimeric cells has been found in the portal and sinusoidal areas of patients with BA, suggesting that maternal lymphocytes cause bile duct injury through a graft-versus-host immune response [35].

●Coordinated activation of genes involved with lymphocyte differentiation, particularly those associated with T helper 1 immunity, has been identified in liver samples from infants with BA [36].

●Polymorphisms that enhance expression of the CD14 gene, which plays a role in the recognition of bacterial endotoxin, have been associated with BA and idiopathic neonatal cholestasis [37].

CLINICAL FEATURES — Most infants with BA are born at full term, have a normal birth weight, and initially thrive and seem healthy.

Prenatal findings — Most infants with BA have normal results on prenatal monitoring, including ultrasonography. Occasionally, prenatal ultrasound identifies an abnormality suggesting BA, including a cyst in the liver hilum or heterotaxy syndrome [38].

If these signs are present, the infant should be further evaluated by ultrasonography after birth and also have serial monitoring for clinical and laboratory signs of BA during the first few weeks and months of life, especially for a rise in conjugated bilirubin. (See 'Abdominal ultrasound' below and 'Laboratory studies' below.)

Signs and symptoms

●Jaundice is the first sign of BA. Initially, the jaundice may be seen only in the sclerae. The onset of jaundice occurs any time from birth up to eight weeks of age, and it is highly unlikely to appear later.

●Some infants have acholic stools. Acholic stools often go unrecognized because the stools are pale but not white and the stool color can vary on a daily basis. To help parents or caregivers distinguish between normal and acholic stools, printed "stool color cards," websites, and a free smartphone application (PoopMD for iPhone or Android) have been developed [39,40]. In a study from Japan that included more than 300,000 newborn infants, stool color cards completed by the parents had a sensitivity of 76.5 percent and specificity of 99.9 percent, respectively, for identifying infants with BA [41].

●Most infants have dark urine because of bilirubin excretion into the urine. Parents or caregivers may not recognize that dark yellow urine or a yellow-stained diaper is abnormal in an infant.

●If the jaundice has gone unnoticed and the child's disease has progressed, there may be a firm, enlarged liver and splenomegaly.

Laboratory studies

●Symptomatic infants – When laboratory studies are performed in infants after they come to attention because of one of the above symptoms, they reveal elevations in direct and/or conjugated bilirubin and mild or moderate elevations in serum aminotransferases, with a disproportionately increased gamma-glutamyl transpeptidase (GGTP). If coagulopathy is present at diagnosis, it is most likely due to vitamin K deficiency.

●Presymptomatic infants – If laboratory studies are performed shortly after birth (before the infant develops symptoms due to BA), mild elevations of conjugated bilirubin are seen. As an example, in a study of 346 infants who had direct bilirubin measured within the first three days of life, direct bilirubin >0.45 mg/dL was a good predictor of later diagnosis of BA (sensitivity 100 percent, specificity 15.4 percent) [42]. Among more than 4000 infants who had direct bilirubin measured between 3 and 60 days of life, direct bilirubin >1.0 mg/dL was a good predictor of BA (sensitivity 100 percent, specificity 77 percent). Similar findings were noted in other cohorts of neonates [43,44]. These observations suggest that infants with mildly elevated conjugated or direct bilirubin levels in the perinatal period should be followed closely and evaluated for the possibility of BA.

These findings also raise the possibility that routine measurements of conjugated or direct bilirubin during the first few days of life could be used as a screening tool for BA [45,46]. In a study of such an approach, more than 120,000 newborn infants were screened for persistent elevations in conjugated or direct bilirubin [47]. All infants were screened once during the birth hospitalization (within the first 60 hours of life), and those infants with a positive initial screen were tested again around two weeks of age. This screening procedure accurately identified seven infants with BA (sensitivity 100 percent, 95% CI 56-100 percent). It also identified 112 infants who did not have BA, many of whom were ultimately diagnosed with other conditions that explained the abnormal result. This study suggests that conjugated or direct bilirubin is a promising approach to newborn screening for BA, but further studies are needed to fully assess the diagnostic yield and cost-effectiveness of this measure and its potential effect on patient outcome.

EVALUATION — Infants with suspected BA should be evaluated as rapidly as possible because the success of the surgical intervention (hepatoportoenterostomy [HPE], the Kasai procedure) diminishes progressively with older age at surgery [48]. (See 'Predictors of the need for transplantation' below.)

The evaluation process involves a series of serologic, laboratory, urine, and imaging tests. The order of diagnostic tests is prioritized based on testing for treatable diseases first, such as biliary obstruction, infections, and some metabolic diseases. Because the timing of surgery is crucial for infants with BA, the evaluation should be completed as quickly as possible, and it is sometimes appropriate to proceed with surgical exploration, even if all of the test results have not returned. In our practice, we typically proceed with each of the steps below if the infant is less than six weeks of age. For those patients six weeks and older, we still try to complete a full evaluation as rapidly as possible (eg, three to four days). Our rationale is that some diseases, such as Alagille syndrome or alpha-1-antitrypsin deficiency, can mimic many of the findings of BA.

Laboratory testing — Evaluation of a young infant with cholestasis includes blood tests to assess hepatic function and rule out metabolic disease and other etiologies (table 1). Further detail about the laboratory evaluation of an infant with cholestasis is given in a separate topic review. (See "Approach to evaluation of cholestasis in neonates and young infants", section on 'Laboratory studies'.)

Research is in progress to identify biomarkers to help distinguish BA from other causes of neonatal cholestasis. In particular, preclinical studies suggest that elevated serum levels of interleukin-33 [49] or matrix metalloproteinase-7 (MMP-7) [50-54] may be useful for this purpose, with positive and negative predictive values above 90 percent. MMP-7 values may vary with age, comorbidities such as congenital heart disease, and other factors. Proposed thresholds vary among these studies, and further validation is needed to determine how MMP-7 levels can be efficiently used in the diagnostic algorithm for neonatal cholestasis at different ages.

Abdominal ultrasound — Evaluation of biliary anatomy begins with an ultrasound. The main utility of the ultrasound is to exclude other anatomic causes of cholestasis (ie, choledochal cyst, intrahepatic bile duct dilatation, and gallstones) (see "Causes of cholestasis in neonates and young infants", section on 'Biliary cysts'). In addition, several findings (if present) support the diagnosis of BA, including absence of the gallbladder, irregularly shaped gallbladder, or polysplenia [38]. When a detailed ultrasonographic protocol is used, additional features can be identified to support the diagnosis of BA, including the "triangular cord" sign, gallbladder contractility, absence of the common bile duct, and an enlarged hepatic hilar lymph node [38,55-60]. The triangular cord sign is a triangular echogenic density seen just above the porta hepatis on ultrasound scan. Its presence is highly suggestive of BA [61]; in a meta-analysis, the specificity of this sign was 0.95 (95% CI 0.92-0.97) and sensitivity was 0.68 (95% CI 0.57-0.78) [38].

Hepatobiliary scintigraphy — Patency of the extrahepatic biliary tree can be further assessed by hepatobiliary scintigraphy, but results are not definitive. If there is strong suspicion for BA, such as in a patient with acholic stools, excretion of tracer is unlikely to occur and it is preferred to proceed directly to liver biopsy without performing hepatobiliary scintigraphy. Although some centers may use phenobarbital to enhance radioisotope excretion, we do not use phenobarbital, because it will delay diagnosis and does not obviate the need for liver biopsy.

Results are interpreted as follows:

●Failure of tracer excretion suggests BA but does not exclude other diseases.

●Conversely, if scintigraphy demonstrates definite excretion of the tracer from the liver to the small bowel, patency is established and BA is very unlikely. In this case, an intraoperative cholangiogram is not necessary. However, a scan with delayed or attenuated excretion does not definitively exclude BA and further evaluation is warranted [62]. A publication reports six cases of infants who had biliary patency on hepatobiliary scintigraphy but were ultimately diagnosed with BA [62]. Similarly, if excretion is noted on a scan that is performed when the infant is very young (ie, less than six weeks old) and cholestasis persists, the scan should be repeated one to two weeks later because the disease process may progress during the neonatal period.

Liver biopsy — We perform a liver biopsy in the majority of infants with suspected BA for two reasons. One purpose is to identify histologic changes consistent with obstruction that warrant surgical exploration. The other is to differentiate BA from other causes of intrahepatic cholestasis, which would not need surgical exploration.

Biopsy findings that indicate another etiology include bile duct paucity (Alagille syndrome), periodic acid-Schiff (PAS)-positive diastase resistant granules (consistent with alpha-1 antitrypsin deficiency), loss of MDR3 staining (suggestive of PFIC3), or giant cell hepatitis without proliferation of ducts. Characteristic histologic features of BA include expanded portal tracts with bile duct proliferation, portal tract edema, fibrosis and inflammation, and canalicular and bile duct bile plugs.

The earliest histologic changes associated with BA may be relatively nonspecific, and biopsies done too early may result in a false negative [63]. At times, it is necessary to repeat a liver biopsy at an older age (eg, two to three weeks later). The sensitivity of liver biopsy in the diagnosis of BA is reported to be approximately 88 percent [64]. Therefore, infants without classic findings on liver biopsy but with a typical presentation for BA and no apparent alternative diagnosis may warrant an operative cholangiogram.

Histologic findings alone cannot help to distinguish BA from other causes of obstruction, such as choledochal cyst or external compression. Therefore, any evidence of obstruction mandates a definitive cholangiogram, as well as imaging studies (ie, ultrasonography, which is typically performed with the liver biopsy).

Cholangiogram — If the above steps in the evaluation support the diagnosis of BA, the infant should be taken to the operating room. The first step is an intraoperative cholangiogram, which is the gold standard in the diagnosis of BA. It is essential that patency be investigated both proximally into the liver and distally into the bowel to determine whether BA is present. If the intraoperative cholangiogram demonstrates biliary obstruction (ie, if the contrast does not fill the biliary tree or reach the intestine), the surgeon should perform a Kasai HPE at that time. In some cases, the cholangiogram cannot be performed, because the gallbladder and biliary tree are atretic; in this case, the surgeon makes the diagnostic decision based on visual inspection of the biliary tree.

An alternative yet valid approach used at some centers is to perform a percutaneous gallbladder cholangiogram or an endoscopic retrograde cholangiopancreatography (ERCP) [65-68]. These procedures are less invasive than intraoperative cholangiogram, but performance of these procedures in infants requires special expertise and equipment. Moreover, if BA is confirmed, the infant will still need to undergo an operation for treatment. The percutaneous gallbladder cholangiogram can only be performed in an infant with an identified gallbladder; the procedure is performed by an interventional radiologist. Alternatively, an ERCP can be used to demonstrate biliary patency in an infant. (See "Endoscopic retrograde cholangiopancreatography (ERCP) for biliary disease in children", section on 'Neonatal cholestasis'.)

DIAGNOSIS — The possibility of BA is suggested by the clinical presentation of neonatal conjugated hyperbilirubinemia and/or acholic stools. The suspicion of BA is strengthened by results of a variety of tests (typically ultrasound, hepatobiliary scan, and liver biopsy). The definitive diagnosis is made by intraoperative cholangiogram.

Differential diagnosis — In term infants, BA and neonatal hepatitis account for a majority of cases of neonatal cholestasis (table 2). A variety of genetic disorders account for most of the remainder. In premature infants, cholestasis more often results from total parenteral nutrition or sepsis. The range of causes and an approach to distinguishing among them are discussed separately. (See "Causes of cholestasis in neonates and young infants" and "Approach to evaluation of cholestasis in neonates and young infants".)

KASAI PROCEDURE — If BA is confirmed by cholangiogram, a Kasai procedure (hepatoportoenterostomy [HPE]) should be performed promptly. This operation is undertaken in the attempt to restore bile flow from the liver to the proximal small bowel (figure 1) [69]. For this procedure, a Roux-en-Y loop of bowel is created by the surgeon and directly anastomosed to the hilum of the liver, following excision of the biliary remnant and portal fibrous plate.

If the HPE is successful, the remaining small patent bile ducts will drain into the Roux limb and jaundice will start to resolve in the weeks following surgery. Even if bile flow is established and cholestasis improves, many patients will have slowly progressive liver disease despite undergoing the Kasai HPE procedure and the majority of patients with BA will ultimately require liver transplantation. At least 50 percent of patients who undergo HPE will require liver transplantation by two years of age as a result of primary failure of the HPE and/or growth failure. The patient's age at the time of HPE can predict native liver survival at later time points. As an example, among those patients who undergo HPE ≤30 days of life, the chance of native liver survival at four years of age is nearly 50 percent, whereas among those who underwent HPE between 31 and 90 days of life, the chance of native liver survival at four years of age is 36 percent [70]. (See 'Prognosis' below.)

If the HPE is unsuccessful, bile drainage is not achieved and the child remains jaundiced. If there is persistent jaundice or elevated serum bilirubin three months after the procedure, the patient should be referred for liver transplant evaluation. (See 'Liver transplantation' below and 'Predictors of the need for transplantation' below.)

Revision of a nonfunctioning HPE generally is not recommended. This is because a revisional procedure is unlikely to be effective if the original HPE did not achieve bile drainage and because the revisional procedure is likely to cause adhesions that increase the technical difficulty of a subsequent transplant procedure [71]. However, in patients in whom the initial HPE was successful, revisional HPE may be appropriate if the patient abruptly develops jaundice, or experiences recurrent episodes of cholangitis but has no other evidence of chronic liver disease. In one report of 24 patients who underwent revisional HPE for these indications, 75 percent achieved bile drainage and 46 percent survived with their native liver (mean follow-up 92 months) [72].

The vast majority of individuals with BA will eventually require liver transplantation. Nonetheless, the Kasai HPE obviates the need for liver transplantation in a substantial minority and significantly delays the liver transplantation for many others. If bile drainage is achieved, it is likely that transplantation will not be needed for years or decades. Patients who undergo HPE have survival rates with native liver of 35 to 50 percent at four years of age; this compares favorably with historical series from before the Kasai procedure was introduced in 1968, which reported a 10 percent survival rate at age three years of age [73,74]. (See 'Liver transplantation' below.)

Preemptive transplant (ie, transplant without prior HPE) is generally avoided because of the advantages of transplanting older, larger patients and because of the potential for improved transplantation therapies in the future. At one time, preemptive transplantation was used to avoid the difficulties of performing surgery on a patient with a prior HPE. With modern transplantation techniques, there is no surgical advantage to preemptive transplantation. Therefore, HPE is almost always performed first. In an infant who presents very late (>120 days) and with evidence of more advanced liver disease, such as ascites, preemptive liver transplant may be appropriate given the futility of HPE, although these decisions should be made on a case-by-case basis.

POSTOPERATIVE MANAGEMENT — Medical care following Kasai hepatoportoenterostomy (HPE) consists of the following interventions, as detailed below [75]:

●Choleretics

●Nutritional supplementation

●Fat-soluble vitamin supplementation

●Prevention of cholangitis

●Management of portal hypertension and its sequelae

Clinical evidence does not support routine use of glucocorticoids in the treatment of BA. (See 'Glucocorticoids (unproven benefit)' below.)

Choleretics — We suggest treating patients with ursodeoxycholic acid (UDCA) after Kasai HPE. This is standard practice in BA, although its clinical utility has not been definitively established. UDCA is a hydrophilic bile acid; when given by mouth, it shifts the balance of bile acids towards hydrophilic forms. This is thought to stabilize membranes and then reduce generation of free radicals, thus protecting mitochondria from damage.

The recommended dose of UDCA in BA ranges from 15 to 30 mg/kg/day and should not exceed 30 mg/kg/day. To avoid potential toxicity, UDCA therapy should be discontinued if the total bilirubin level rises above 15 mg/dL (257 micromol/L).

The most convincing human data regarding benefits of UDCA come from the group of patients with primary biliary cholangitis. In several large, randomized, double-blind, placebo-controlled trials in patients with primary biliary cholangitis, UDCA decreased plasma levels of aminotransferases, improved liver histology and quality of life, and decreased risk of death and need for liver transplantation (see "Overview of the management of primary biliary cholangitis"). However, in patients with primary sclerosing cholangitis, a randomized trial suggested negative effects of long-term, high-dose UDCA therapy [76].

In BA, observational studies suggest a number of possible benefits of UDCA treatment, ranging from enhanced weight gain to reduced episodes of cholangitis and improved bile flow, but definitive evidence from randomized trials is lacking [77,78].

Glucocorticoids (unproven benefit) — Clinical evidence does not support routine use of glucocorticoids in the treatment of BA [79,80]. This was shown in a randomized, placebo-controlled trial of glucocorticoid treatment in 140 infants with BA [81]. The glucocorticoids were given for 13 weeks (intravenous methylprednisolone 4 mg/kg/day for two weeks, followed by oral prednisolone 2 mg/kg/day for two weeks, then tapering), and outcomes were measured at 6 and 24 months post-HPE. No statistically significant benefit in bile drainage at six months post-HPE was observed in the infants treated with glucocorticoids as compared with the placebo group. In addition, no statistically significant improvement in survival with native liver at two years of age was observed in the treatment group. Moreover, the infants treated with glucocorticoids had significantly earlier onset of serious adverse events as compared with those given placebo. After HPE, the infants that had been treated with glucocorticoids had impaired growth parameters (length, weight, and head circumference) for at least six months after HPE [82]. Despite this evidence to the contrary, we are aware of centers in the United States, Europe, and Asia where steroids are still used, either routinely or in selected cases.

Nutrition — Nutritional problems in BA are common and difficult to overcome. Poor nutrition is a significant clinical problem and is one of the most common indications for liver transplantation. Nutrition needs for children with chronic cholestasis are summarized in the joint North American Society For Pediatric Gastroenterology, Hepatology and Nutrition and European Society for Paediatric Gastroenterology Hepatology and Nutrition position paper [83].

Caloric needs — Several factors contribute to malnutrition in patients with BA, including anorexia, malabsorption due to insufficient intraluminal bile salts, and chronic liver inflammation [84]. Because of malabsorption and metabolic alterations, the total caloric needs in infants with BA are approximately 130 to 150 percent of the recommended energy intake for healthy infants and children (see "Poor weight gain in children younger than two years in resource-abundant settings: Management", section on 'Estimation of energy requirements') [83]. To compensate for losses and catabolism, the expected protein needs are 3 to 4 g/kg/day in infants and 2 to 3 g/kg/day in children [85-87].

In order to meet these nutritional requirements, several strategies are utilized. These strategies are similar to those used for infants with other causes of growth failure, except that they should be implemented proactively because of the high rates of growth failure in patients with BA. (See "Poor weight gain in children younger than two years in resource-abundant settings: Management", section on 'Estimation of energy requirements'.)

●For infants, formulas are concentrated or expressed breast milk is fortified to provide additional energy. After Kasai HPE, the feed is typically designed to provide 24 kcal per ounce. If growth is inadequate, the feed may be increased to 27 kcal per ounce; additional energy content can be added in solid foods when the infant is old enough.

●High-energy supplements, such as glucose polymer powder (which provides 8 cal/teaspoon) or medium-chain triglyceride (MCT) oil (which provides 7.7 cal/mL), are used to fortify formula or solid foods [79,86]. MCT oil is useful because it is calorically rich, and it is readily absorbed by patients with cholestasis because it does not require micellar solubilization. To avoid essential fatty acid deficiency, long-chain triglycerides (LCT) should also be included, providing at least 3 percent of total calories [83]. A reasonable target is to provide 30 to 70 percent of fat calories as MCT. Formula with high MCT content is recommended in infants with chronic cholestasis. Supplemental MCT oil in the formula or foods can also be used.

●Despite these measures, many infants and children with BA require supplemental feeding by nasogastric tube because they are unable to take enough energy by mouth to meet their increased nutritional needs. Gastrostomy tubes are not recommended, because many patients develop portal hypertension, leading to gastric varices and the propensity to develop varices around the gastrostomy tube site. Candidates for nasogastric feeds are identified by poor weight gain and/or poor linear growth. Proactive management is recommended because the malnutrition may worsen the overall prognosis, with or without liver transplantation [79]. (See "Overview of enteral nutrition in infants and children".)

●Ultimately, some infants will have persistent malnutrition despite efforts to optimize enteral calories. Parenteral nutrition can be considered in these situations; however, the risks of serious bacterial infections need to be considered. Parenteral nutrition is usually limited to children who are already hospitalized with other complications of chronic liver disease and are awaiting liver transplant.

Fat-soluble vitamin supplements — All jaundiced infants and children with BA should be given supplements of fat-soluble vitamins (table 3). Once jaundice resolves and vitamins are replete, children can be transitioned to standard multivitamins. Nevertheless, routine monitoring of vitamin levels should continue. Vitamin levels should be monitored frequently (ie, several times in the first year), starting at the first month after HPE, in order to adjust supplements appropriately for deficiencies or toxicities. The frequency of laboratory monitoring depends in part on persistence or resolution of cholestasis.

Deficiencies of fat-soluble vitamins are common in patients with BA. In a series of 29 patients with BA, serologic deficiencies of vitamins A and E and radiographic evidence of vitamin D deficiency were reported despite establishment of bile flow by HPE [88]. Vitamin deficiencies occur despite recommended supplementation and are particularly common among patients with residual cholestasis after Kasai HPE (defined in this study as serum total bilirubin ≥2 mg/dL [34 micromol/L]) [89]. In one study, 81 percent of infants were found to be vitamin D deficient before HPE and vitamin D deficiency persisted post-HPE despite aggressive supplementation [90].

Infants with BA and prolonged jaundice may be deficient in vitamin K. Patients should receive supplementation with oral vitamin K and should be monitored for coagulopathy. Some infants may require parenteral vitamin K supplementation due to poor absorption of oral medications in the setting of severe cholestasis.

Complications — Children with successful bile drainage following HPE must be followed closely for complications, particularly ascending cholangitis and portal hypertension. In addition, all patients should be routinely monitored for fat-soluble vitamin deficiencies, as described above.

Ascending cholangitis — Cholangitis is a common complication in patients with BA who have undergone a successful Kasai HPE. The incidence of cholangitis in these patients is between 40 and 90 percent, with the majority of patients experiencing at least one episode prior to two years of age [91,92]. These patients are at risk for cholangitis because of the abnormal anatomy and bacterial stasis in the region of the Roux limb. Clinicians should have a high level of suspicion for cholangitis in children presenting with fever without a clear source of infection, especially if the fever is accompanied by acholic stools, irritability, and laboratory abnormalities. By contrast, patients with an unsuccessful Kasai (ie, in which bile drainage is not achieved) have a low risk for ascending cholangitis but are still at risk for other infections and sepsis.

Because cholangitis can be life-threatening and may impact long- and short-term outcomes [93,94], most clinicians prescribe prophylactic antibiotics for at least one year after Kasai HPE. Small nonrandomized trials suggest that the benefits of antibiotic prophylaxis outweigh the risks of antibiotic resistance [95,96]. As an example, in one series, infants treated with prophylactic antibiotics had one-half the rate of cholangitis as compared with infants in a historical control group [95]. Either trimethoprim-sulfamethoxazole (4 mg/kg/day trimethoprim and 20 mg/kg/day sulfamethoxazole) or neomycin (25 mg/kg/day divided four times daily) appear to be equally effective in decreasing the incidence of cholangitis [95]. There may be evidence to suggest that treatment with probiotics can afford a prophylactic effect that is similar to that of antibiotics, but further studies are needed [97].

Recurrent cholangitis may predict the need for liver transplantation as it can lead to progressive cirrhosis [93]; however, one episode of cholangitis does not predict early transplantation [8]. Bile lakes develop in a significant minority of patients, and these patients have an increased risk for recurrent cholangitis [98,99].

Portal hypertension and variceal bleeding — The chronic hepatobiliary inflammation characteristic of BA leads to progressive biliary cirrhosis. Biliary cirrhosis causes portal hypertension, which can lead to esophageal varices and ascites. Obliterative portal venopathy has been reported as a pathologic finding in liver explants of children with BA undergoing liver transplantation and could theoretically contribute to portal hypertension [100].

●Clinical presentation – Development of splenomegaly and/or declining platelet counts after Kasai HPE suggest the evolution of portal hypertension. In some patients, portal hypertension is first noted when they present with upper gastrointestinal bleeding from esophageal varices.

●Prevalence – In a registry study of more than 850 children with BA, approximately one-third of children had features of portal hypertension (splenomegaly and/or thrombocytopenia) and 9 percent developed variceal bleeding by five years of age [101]. Among those who developed variceal bleeding, 2.5 percent died, which is considerably lower than historical reports, and 12.5 percent proceeded to liver transplant within six weeks. Recurrent variceal bleeding and refractory ascites are indications for liver transplantation [102,103]. (See 'Liver transplantation' below.)

●Monitoring

•Physical examination and laboratory tests – Assessment for development of splenomegaly and monitoring of platelet count are important to identify the development of portal hypertension. These are not reliable indicators in patients with asplenia or polysplenia.

•Endoscopy – Some centers perform surveillance endoscopy to gauge the size and status of developing varices in patients who demonstrate clinical and ultrasonographic signs of portal hypertension following the Kasai procedure. In a series of 47 children with BA managed with endoscopic surveillance, varices developed in approximately one-half of the children at a mean of 19 months (range 4 to 165 months) after successful Kasai HPE [104]. Other centers begin routine surveillance endoscopies only after the patient has experienced the first variceal bleed. There is no evidence that beta blockers prevent or delay the development of esophageal varices [105].

•Noninvasive tests – Studies have reported the use of noninvasive measurements of liver stiffness (FibroScan; transient elastography) as an index of fibrosis in patients with BA. In a multicenter study, nonfasted liver stiffness correlated with presence of clinically evident portal hypertension (splenomegaly and thrombocytopenia). In another report, transient elastography predicted the presence of esophageal and gastric varices [106,107].

●Management – Endoscopic ablation of varices is generally recommended after the first variceal bleed (secondary prophylaxis), as reflected in a consensus statement [105]. Patients presenting with acute bleeding from varices should receive volume resuscitation as needed to restore hemodynamic stability. In addition, therapy with vasoactive drugs (eg, octreotide) may be started prior to endoscopy [105]. The varices are ablated by band ligation or sclerotherapy, and this therapy is repeated in subsequent sessions as needed with a goal of complete variceal obliteration.

Primary prophylaxis (endoscopic ablation of esophageal varices before the first variceal bleed) is generally not recommended, as no mortality benefit has been demonstrated. In extenuating circumstances, such as when the child is not in reasonable proximity to emergency care, primary prophylaxis can be considered.

Ascites is common in infants with BA awaiting transplant and is typically managed with diuretics [108]. Paracentesis may be required in refractory cases.

LIVER TRANSPLANTATION — The majority of individuals with BA eventually require liver transplantation; BA is the most common indication for liver transplantation in infants and children. In the current era, at least 60 to 80 percent of patients with BA will require liver transplantation, even with optimal management. (See 'Prognosis' below.)

The indications for liver transplantation for patients with BA include [79,109-111]:

●Primary failure (lack of bile drainage) of the Kasai hepatoportoenterostomy (HPE)

•Prompt referral for liver transplantation evaluation is recommended if total bilirubin is >6 mg/dL (100 micromol/L) three months or more beyond HPE

•Referral for liver transplantation evaluation also should be considered if total bilirubin is persistently 2 to 6 mg/dL (34 to 100 micromol/L) three months or more beyond HPE

●Refractory growth failure – Referral for transplantation should be considered for patients with moderate or severe growth failure that does not respond to intensive nutritional support

●Complications of portal hypertension (if these cannot be managed with other measures)

•Repeated variceal bleeding

•Refractory ascites that compromises respiratory, bowel, or renal function

•Hepatopulmonary syndrome

•Portopulmonary hypertension

●Progressive liver dysfunction

•Intractable pruritus

•Refractory coagulopathy

●Recurrent cholangitis requiring prolonged intravenous antibiotic therapy and need for a central line or peripherally inserted central catheter

●Hepatocellular carcinoma without extrahepatic involvement (very rare)

Among patients transplanted before two years of age, persistent cholestasis (often complicated by growth failure) and recurrent or resistant cholangitis are the most common indications for liver transplantation.

Although it is clear that the vast majority of BA patients will ultimately require liver transplantation, preemptive transplants should be deferred as there are advantages to transplanting older and larger patients. Because the outcomes of liver transplantation improve for infants with weights >10 kg as compared with smaller infants, transplantation of infants with persistent cholestasis may be postponed if growth can be achieved and the infants are otherwise stable. Supplemental nasogastric tube feedings are often necessary to attain adequate growth. The cholestasis may cause chronic pruritus, which often responds to medical therapy. (See "Pruritus associated with cholestasis".)

For those children who require liver transplantation, prognosis is generally good. In the United States, the one-year patient and liver-graft survival rates for pediatric patients undergoing primary liver transplant for BA are 96 and 89 percent, respectively [112]. At five years postoperatively, patient and graft survival are still above 95 and 80 percent, respectively [113]. In international series, long-term survival is approximately 70 to 85 percent at 5 years, and these rates continue to improve in more recent case series [70,114,115].

BA patients can receive whole or segmental deceased donor grafts, as well as segments from a living donor [114]. The liver transplant surgery in BA patients is complicated by presence of intraabdominal adhesions attributed to previous HPE. Postoperative outcomes appear to be less favorable in the subset of BA patients with BASM, due to the enhanced difficulty associated with liver dissection and vascular reconstruction in these patients [116]. Poor nutrition is associated with increased mortality either while awaiting transplantation or after the procedure [117]. Thus, vigorous nutritional support, including nasogastric feeds, is essential in the pre- and postoperative care of these patients. (See 'Nutrition' above.)

PROGNOSIS

Overall survival — Although long-term prognosis for BA patients is variable, the complementary and sequential approach of hepatoportoenterostomy (HPE) and liver transplantation affords long-term survival, with upwards of 90 percent of BA patients surviving into adulthood [118,119]. Importantly, waitlist and post-transplant mortality is higher in patients undergoing transplant earlier in life. In a large series, patients undergoing transplant when they were less than two years of age had five-year survival of 93.8 percent, compared with 97.1 percent for those transplanted after two years of age [112]. This observation underscores the importance of achieving biliary drainage with HPE and delaying transplantation when possible.

Survival without transplantation — Overall survival with native liver (ie, without transplantation) ranges from 30 to 55 percent at 5 years, 30 to 40 percent at 10 years, and 20 to 40 percent at 20 years [48,119-127]. The long-term outcomes are illustrated by the following series:

●Among 1429 patients diagnosed with BA between 1986 and 2015 in France, 94 percent underwent the Kasai HPE, leading to initial clearance of jaundice in 38 percent [128]. Survival with the native liver was 41 percent at 5 years, 35 percent at 10 years, 26 percent at 20 years, and 22 percent at 30 years of age.

●Among 100 children who underwent HPE in the Netherlands between 1987 and 2015 and reached the age of two years with their native liver, long-term survival with native liver was 89, 72, 60, and 54 percent at 5, 10, 15, and 18 years of age, respectively [129].

●Among 80 patients who underwent the procedure between 1970 and 1986 in Japan, the 5-, 10-, and 20-year survival rates with native livers were 63, 54, and 44 percent, respectively [126]. One-half of the 20-year survivors had cirrhosis, and 20 percent went on to liver transplant or death within a few years.

●Among 34 patients with BA who underwent Kasai HPE between 1994 and 2011, survival with the native liver was 87.6 percent at 5 years, 76.9 percent at 10 years, and 48.5 percent at 15 years [41]. This represents substantial improvement in outcomes compared with earlier series, possibly due to earlier identification of affected infants due to a national screening program using stool color cards. (See 'Signs and symptoms' above.)

In addition to complications from portal hypertension and late-onset cholangitis, the risk of developing cancer in the native liver remains an important concern. Careful monitoring and treatment of these late complications are essential. Pregnancy in female survivors of BA is not without risk, and the pregnancy must be monitored to ensure the health and safety of the mother and child.

Given the improvements in early identification of infants with BA and advances in medical therapy after Kasai HPE, including prophylaxis against cholangitis and improvements in nutrition and bile flow, it is reasonable to expect that long-term survival from current cases will exceed that of these older series. Experimental areas of therapy such as ileal bile acid transport inhibitors and antifibrotic or antiinflammatory agents may improve outcomes in the future.

Predictors of the need for transplantation — Although performance of the Kasai HPE clearly improves survival overall, the long-term prognosis is difficult to predict. The three most important prognostic factors for surgical outcome are younger age at the time of HPE, the expertise of the surgeon and care center at which the procedure is performed, and the decrease in serum bilirubin in the first few months after HPE.

●Age at HPE – Many case series have documented better outcomes when HPE is performed before 60 days of age and poorer outcomes after 90 days of age [48,120-122]. As an example, in a series of 349 infants diagnosed and treated for BA in Canada, older age at time of HPE was associated with a progressive decrease in patient survival with the native liver [70]. Among infants who underwent HPE at ≤30 days, 31 to 90 days, and ≥90 days, the survival with the native liver at four years of age was 49, 36, and 23 percent, respectively. In Taiwan, after implementation of a protocol using universal screening for infant stool color increased the rate of accomplishing HPE before 60 days of life from 50 to 66 percent, improvements in several clinically important outcomes were seen: the rate of jaundice-free survival with the native liver at three years of age improved from 32 to 57 percent and the rate of overall survival at five years of age improved from 56 to 89 percent [130]. However, the time at which early is early enough and late is too late is controversial [131].

●Surgery quality – Surgical success is dependent on the expertise of the center and surgeon. Although controversial, it appears that a center that performs at least five Kasai procedures per year has a better success rate, as measured by 5- or 10-year long-term survival with the native liver [121,132,133]. Another surgical factor is anatomic pattern of BA identified at time of Kasai HPE. Children born without atresia at the porta hepatis (Ohi classification 1) have the lowest risk of death or transplantation by two years of age [134].

●Biomarkers – Serum bilirubin post-HPE is an important predictive biomarker of outcome. Data show that the serum total bilirubin level measured three months after Kasai HPE predicts native liver survival [8,135]. In a prospective cohort, among patients with total bilirubin <2 mg/dL three months post-Kasai HPE, two-year survival without transplantation was 86 percent [135]. Among those with total bilirubin ≥2 mg/dL three months post-Kasai HPE, two-year survival without transplantation was only 20 percent. Infants with elevated bilirubin were also more likely to develop poor weight gain, hypoalbuminemia, and coagulopathy.

For infants who achieve normal bilirubin levels, serum bile acid concentrations serve as an additional prognostic biomarker. In a prospective cohort study, serum bile acid concentration ≤40 micromol/L six months after HPE was associated with improved clinical outcomes at two years of age, including biomarkers and spleen size [136]. The same measure also predicted 10-year cumulative incidence of liver transplant or death (8.5 versus 43 percent).

In one article from a multicenter study, clinical variables at two years of age in children with BA surviving with native liver were used to predict the need for liver transplantation later in childhood [137]. Variables included in the final predictive models included total bilirubin, albumin, platelet count, gamma-glutamyl transpeptidase (GGTP), and history of cholangitis or ascites. Use of predictive models such as these may identify patients who could benefit from enhanced clinical monitoring.

Neurodevelopmental — Infants and young children with BA are at increased risk for neurodevelopmental delays, particularly if the HPE was not successful [138-140]. The risk is correlated with poor growth and ascites, both of which are markers for more advanced liver disease. The risk appears to be low for patients who survive with their native liver [141].

SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Pediatric liver disease" and "Society guideline links: Portal hypertension and ascites".)

SUMMARY AND RECOMMENDATIONS

●Overview – Biliary atresia (BA) is a progressive, idiopathic, fibro-obliterative disease of the extrahepatic biliary tree that presents with biliary obstruction in the neonatal period and is the most common indication for liver transplantation in children. BA may occur in isolation (70 percent), in association with laterality anomalies such as situs inversus or asplenia (10 to 15 percent), or with other congenital malformations (10 to 15 percent). Those with laterality anomalies have a somewhat worse prognosis. (See 'Types of biliary atresia' above.)

●Pathogenesis – The causes of BA are not well established and are probably multifactorial; genetic factors may play a permissive role in some cases, but infectious, toxic, or immunologic mechanisms are probably involved. (See 'Pathogenesis' above.)

●Clinical presentation – Most infants with BA are born at full term, have a normal birth weight, and initially thrive and seem healthy. Scleral icterus and/or generalized jaundice typically develop by eight weeks of age. Infants may also have acholic (very pale-colored) stools (see "stool color card" for examples), dark urine, and/or a firm liver and splenomegaly. (See 'Clinical features' above.)

Laboratory testing reveals elevation in serum conjugated bilirubin (>2 mg/dL [34 micromol/L]) and mild or moderate elevations in serum aminotransferases, with disproportionately increased gamma-glutamyl transpeptidase (GGTP). Milder elevations of serum conjugated or direct bilirubin in neonates have been observed in patients later diagnosed with BA and warrant further evaluation. (See 'Laboratory studies' above.)

●Evaluation – The diagnosis of BA is made with a series of imaging and laboratory tests and liver biopsy to exclude other causes of cholestasis. Infants should be evaluated as rapidly as possible because the success of the surgical intervention diminishes progressively with older age at surgery. (See 'Evaluation' above.)

The definitive diagnosis of BA is made by a cholangiogram. This is typically performed intraoperatively; if the diagnosis of BA is confirmed, the surgeon performs a hepatoportoenterostomy (HPE; also known as a Kasai procedure) (figure 1). (See 'Cholangiogram' above and 'Diagnosis' above.)

●Initial management

•Surgery – We recommend that all infants with BA undergo a Kasai HPE (Grade 1B). This surgery should be performed as soon as the diagnosis of BA can be made and preferably before 60 days of age. Younger age at the time of the Kasai HPE is associated with better outcomes. (See 'Kasai procedure' above and 'Predictors of the need for transplantation' above.)

•Nutrition and monitoring – Infants with BA who are jaundiced should be treated with supplements of fat-soluble vitamins and monitored for fat-soluble vitamin deficiencies (table 3). In addition, they should be given high-calorie formula or other nutritional supplements as required to sustain normal rates of growth. Vigorous nutritional support is also indicated for non-cholestatic infants with BA. (See 'Nutrition' above.)

•Medications – We suggest that infants and children be treated with ursodeoxycholic acid (UDCA) after Kasai HPE (Grade 2C). We do not recommend treatment with glucocorticoids (Grade 1A). (See 'Choleretics' above and 'Glucocorticoids (unproven benefit)' above.)

We suggest treating all patients with prophylactic antibiotics for at least the first year after Kasai HPE to reduce the risk for ascending cholangitis (Grade 2B). Repeated episodes of cholangitis hasten the progression of liver disease. (See 'Ascending cholangitis' above.)

●Liver transplantation – Patients with persistent jaundice three months after Kasai HPE should be referred for liver transplant evaluation. Other indications for early liver transplantation include failure to thrive despite vigorous nutritional rehabilitation, complications of portal hypertension (recurrent variceal bleeding or intractable ascites), and progressive liver dysfunction. (See 'Liver transplantation' above.)

●Prognosis – At least 60 to 80 percent of patients with BA will eventually require liver transplantation. A minority of patients with BA treated with Kasai HPE survive to 20 years or more without liver transplantation. However, many of these patients have chronic liver disease with cirrhosis and portal hypertension. (See 'Survival without transplantation' above.)