- 304 Downloads
Biliary atresia remains one of the most challenging conditions in pediatric surgery. The incidence ranges from 1 in 5,000 in Asia to about 1 in 16,000 live births in Europe with a slight female preponderance. The cause is still unknown with a number of possible causes giving rise to a range of different presentations. It is usually an isolated condition but can be associated with other anomalies.
Currently the management is almost entirely surgical with an initial attempt to restore bile flow and preserve the native liver using a Kasai portoenterostomy. Liver transplantation is an option if this fails or for those infants who present late with obvious end-stage cirrhosis. Beyond effective surgical technique, the role of adjuvant medical therapy is unclear and evidence of benefit is lacking. Despite this the use of postoperative steroids, prophylactic antibiotics, and choleretic agents such as ursodeoxycholic acid is common.
Experience from high-volume centers suggests clearance of jaundice can be achieved in 50–60% of infants with 10-year native liver and real survival rates of 40% and 90%, respectively.
KeywordsBiliary atresia Kasai portoenterostomy Surgical jaundice Cirrhosis BASM Cystic biliary atresia
Biliary atresia (BA) is a rare but important cause of neonatal jaundice. BA remains a somewhat elusive disease, confined as it is to infancy but yet with its origin essentially unknown. If untreated it results in progressive cholestasis which leads to hepatic fibrosis and cirrhosis. This leads to liver failure complicated by portal hypertension and ascites and ultimately death.
It can however be effectively treated in a high proportion by surgery. For the remainder, liver transplantation is an option and indeed BA remains the single most common cause for this in the pediatric age group.
John Thomson, a physician working in Edinburgh, reported an infant he felt had “congenital obstruction of the bile ducts” in 1891. This child who became deeply jaundiced had clay-colored stool and dark urine from the time of birth and ultimately died from liver failure or sepsis at a few months of age. The illustrations from the postmortem showed a normally formed but empty gallbladder with two small bile-filled cysts in the porta hepatis indicating an absence of the common hepatic duct (Thompson 1891).
The influential American pediatric surgeon William Ladd published a series in 1928 of 11 cases of surgical jaundice he had operated upon (Ladd 1928). He describes using restoration of bile flow using reconstruction techniques such as hepaticojejunostomy. Though not all of these cases were truly biliary atresia, some appearing to be choledochal cysts and luminal blockages with inspissated bile, it presaged a surgical approach to such infants. Nonetheless, with increasing surgical experience in infants with biliary atresia, it was quickly realized that most actually appeared to have an entirely solid extrahepatic biliary tree and therefore were “uncorrectable” by the conventional surgical techniques available at the time.
In the 1950s, Morio Kasai, working in Sendai, Japan, developed a much more radical approach to the biliary dissection, advocating complete excision of the extrahepatic biliary tree and anastomosis of the Roux loop to the denuded, transected, albeit “solid” porta hepatis (Kasai and Suzuki 1959). He recognized that actually most retained some communication with the intrahepatic bile ducts via microscopic biliary channels and exposure of these could restore bile flow in a proportion. This operation, now known as a Kasai portoenterostomy (KPE) , still fails in a disappointingly high proportion but still remains the only realistic option in most cases as an attempt to salvage the native liver.
The first liver transplant anyway was performed in March 1963 by Thomas Starzl in Denver, Colorado, on a child who had been born with biliary atresia (Starzl et al. 1963). Though unsuccessful as she died on-table from uncontrollable bleeding, it did mark a historical milestone. Liver transplant programs then emerged throughout the world in the 1960s, but in the face of ineffective immunosuppression and invariable recipient demise, they were inevitably closed down. By the end of the 1970s, increasing use of the effective immunosuppressive cyclosporine allowed a resumption of liver transplantation, and it then became a viable proposition for initially older children who had had some response to a KPE. Our own program began as a joint venture in the mid-1980s as the King’s College Hospital – Cambridge collaboration led by Sir Roy Calne and Alex Mowat.
In types 1 and 2, there is usually some degree of preservation of structure of the intrahepatic bile ducts though these are clearly still highly abnormal with blunting, irregularity, and pruning (and importantly absence of dilatation, even when obstructed). Type 3 (aka “uncorrectable” BA) is the commonest type seen where the intrahepatic bile ducts are grossly abnormal with myriad microscopic ductules coalescing at the porta hepatis. There are then two further distinct subtypes often dependent on the appearance of the gallbladder. Most have an atrophic gallbladder, and these can be associated with an intact, yet solid biliary tree or complete absence of the common bile duct. This latter appearance is characteristic of the extrahepatic ducts of the syndromic form (see later). Alternatively the gallbladder can look entirely normal and be filled with a clear mucus (it is these that are misdiagnosed on ultrasound). These typically have a normal and often patent common bile duct but an absence of the common hepatic duct. Finally, in some infants there is an obvious inflammatory reaction with edema and hypertrophy of the adjacent lymphoid tissue; in others this appears entirely absent with translucency of the peribiliary tissues.
Cystic Biliary Atresia
Extrahepatic cyst formation (containing clear mucus or bile) occurs in about 5–10% of cases and can be termed cystic biliary atresia (CBA) (Caponcelli et al. 2008). It is distinguishable clinically and histologically from simple obstruction in a cystic choledochal malformation as there is preservation of an epithelial lining and retention of a normal and often dilated intrahepatic biliary tree in the latter. Where retrograde cholangiography is possible in CBA, the intrahepatic bile ducts are often seen as a primitive “cloud-like” apparition above the porta hepatis.
Biliary atresia splenic malformation (BASM) syndrome
Cystic biliary atresia
Isolated biliary atresia
Other entities are fewer in number, but there does seem to be a relationship with other gastrointestinal anomalies such as esophageal atresia and jejunal atresia in a small proportion (<2 % of large series) and occasional cases with a defined chromosomal abnormality (e.g., cat-eye syndrome and chromosome 22 aneuploidy).
Developmental biliary atresia is a term which we have proposed for those where the onset is almost certainly prenatal and evident by the time of birth. It currently includes BASM, BA arising with other congenital anomalies, chromosomal BA, and cystic BA (Livesey et al. 2009). The onset of occlusion in isolated BA is much more contentious, and it is widely believed that the bile duct can be normally formed and even patent at the time of birth becoming occluded secondarily, perhaps by virally mediated cholangiocyte damage.
Biliary Atresia Splenic Malformation (BASM) Syndrome
Biliary atresia splenic malformation (BASM) syndrome
Polysplenia, double spleen
e.g., ASD, VSD, Fallot’s tetralogy
Preduodenal portal vein
Absence of vena cava
Absence of portal vein
Malrotation, annular pancreas
Biliary atresia is not simply a mechanical extrahepatic biliary obstruction – in which case there would be dilatation of the intrahepatic bile ducts. Early features include portal tract edema, plugging, and then duplication of biliary ductules. Increasingly, fibrosis is evident ultimately leading to overt macronodular cirrhosis. There is early portal tract inflammation and in a proportion of cases an obvious mononuclear cell infiltrate typical of an inflammatory response.
There will be abnormal expression of class II antigens on biliary and vascular epithelium with expression of cellular adhesion molecules (ICAM and VCAM) in about 50% of infants (Davenport et al. 2001). The nature of the infiltrate is still the subject of ongoing research but appears to be largely CD4+ T cells and, within that genera, predominantly Th1 and Th17. Polarized Th1 cells also appear to be oligoclonal suggesting perhaps a proliferative response to a specific antigen (speculatively viral proteins) (Mack et al. 2007). Monocytes/macrophages also have an important role as antigen-presenting cells and also as cellular mediators perpetuating the inflammatory process and initiating fibrosis.
There is a parallel increase in inflammatory mediators (especially sVCAM) and cytokines (especially TNF-α, IFN-γ, IL-2, and IL-18 ) identifiable in the serum, which continues to increase during the early postoperative period and only abates from about 6 months (Narayanaswamy et al. 2007).
There is a marked geographical variation in the incidence of BA across the world. It appears most common in Asian countries such as Japan and China. Taiwan, for instance, has the highest reported incidence at 1 in 5000 live births. By contrast it is much less common in Western Europe and North America with rates of 1 in 15,000–20,000 being reported. Our own study in England and Wales had an overall incidence of 1 in 17,000, but there was also a markedly significant variation within the country with diminishing incidence along a north and west axis (Livesey et al. 2009). There is a female predominance in development BA, not seen in isolated BA.
Seasonality in BA has been sought in a few studies in order to strengthen a possible link with perinatal viral infection though there was no evidence for this in the largest of such studies (Livesey et al. 2009).
Infants who are later diagnosed with BA have a normal range of birth weights and gestational age, even those with the developmental forms of BA. Failure to thrive then becomes evident and their postnatal weights at diagnosis are subnormal.
Infants with BA are often o therwise healthy and usually present soon after birth with persistent jaundice, acholic stools, and dark urine. Around 40% of infants with cystic BA are detected antenatally with the cyst seen on the maternal ultrasound scan, typically at around 20 weeks of gestation (Caponcelli et al. 2008). It is important to have prompt differentiation postnatally between BA and a choledochal malformation to ensure prompt appropriate management.
Infants usually demonstrate a degree of failure to thrive by the time of diagnosis which is thought to be due to fat malabsorption. Subsequent deficiency of the fat-soluble vitamins A, D, E, and most importantly K can mean some infants will present with a bleeding tendency, and in severe cases, this can be as catastrophic as an intracranial hemorrhage.
Liver fibrosis and are time-dependent features which seem to begin perinatally even in those infants with developmental BA (Makin et al. 2009). Features such as obvious ascites, marked hepatosplenomegaly, etc. should therefore be considered a late sign and are not seen before 80–90 days.
Differential diagnosis of conjugated hyperbilirubinemia
“TORCH” infections toxoplasma, cytomegalovirus, hepatitis, syphilis
US and MRCP
Inspissated bile syndrome
Percutaneous transhepatic cholangiogram
Clinical features (e.g., abnormal facies, echocardiography)
Spontaneous perforation of bile duct
US and radioisotope scan
US and CT scan
Parenteral nutrition-associated cholestasis
Progressive familial intrahepatic cholestasis
? low GGT
Metabolic causes, e.g., galactosemia
Hemoglobin and white cell counts are normal although the platelet count may be raised. This latter observation appears to be specific for BA. Liver biochemistry will show a conjugated hyperbilirubinemia, slightly raised transaminases (AST and ALT), and significantly raised γ-glutamyltranspeptidase (GGT). Protein and albumin levels are usually normal. The AST-to-platelet ratio index (APRi) can also be calculated and has been used as a surrogate marker of liver fibrosis and perhaps as a prognostic index (Grieve et al. 2013).
An abdominal ultrasound is a key part of the diagnostic process in order to exclude other possible surgical diagnoses. The signs in BA however are less specific but might include an atrophic gallbladder with no evidence of filling between feeds and a “triangular cord sign” representing the sonographic appearance of the solid proximal biliary remnant in front of the bifurcation of the portal vein (Park et al. 1997).
Radioisotope (technetium (Tc)-labeled iminodiacetic acid derivatives) hepatobiliary imaging was formerly quite popular in showing absence of biliary excretion. However, it is rarely specific, and there is considerable overlap with neonatal hepatitis (Shanmugam et al. 2009).
Percutaneous liver biopsy , looking for histological features of “large-duct obstruction” as against those of “neonatal hepatitis,” is safe and well-tolerated but needs an experienced pathologist to interpret (Russo et al. 2016). Currently it is the diagnostic method of choice in two out of the three English BA centers.
Direct cholangiography is certainly possible, either using ERCP (Shanmugam et al. 2009) or at laparoscopy (Nose et al. 2005). ERCP is technically challenging even with the right equipment but can avoid laparotomy in the larger infants. Laparoscopy and direct puncture of the gallbladder is also relatively straightforward as an access point for a cholangiogram.
Screening for Biliary Atresia
Screening programs have been initiated in some countries in order to reduce the time to definitive surgery. The most well-developed has been that in Taiwan (Hsiao et al. 2008) but other European countries are also pioneering it. Mothers are issued with color-coded cards and asked to compare with their infant’s stool. Recognition of pale stool then prompts further investigation and referral. This has lowered the time to surgery in Taiwan, but still it is no quicker than in the UK even though there is no such coordinated program. They do avoid the problem of the “missed” infant presenting late (>100 days old) with BA and overt cirrhosis which is still seen in England and Wales. A recent review of our program suggested that this had fallen from 7.8% in the period 1999–2004 to 4.8% in the period 2009–2013 (personal observation).
Surgery: Kasai Portoenterostomy (KPE)
Confirmation of diagnosis: A limited right-upper quadrant muscle-cutting incision allows access to the gallbladder and a cholangiogram (if possible). If there is no lumen, then that, in itself, is evidence for BA. The cholangiogram should show the complete biliary tree and drain into the duodenum to exclude BA.
Mobilization of the liver: Our practice suggests that complete porta hepatis dissection requires mobilization of the liver allowing it to be everted outside of the abdominal cavity. The anesthetist needs to be aware of this maneuver as it reduces venous return and requires judicious intravenous volume support. Alternatively, some surgeons leave the liver in situ but sling the right and left vascular pedicles and use traction to expose the porta hepatis. There is a risk of portal vein thrombosis with this technique.
Portal dissection: The gallbladder is mobilized from its bed and the distal CBD divided and then dissected back toward the porta hepatis (Figs. 2 and 3). Division of small veins from the back of the portal confluence to the porta plate facilitates downward traction of the portal vein and exposes the caudate lobe. On the left side, there is often an isthmus of liver parenchyma (from segments III to IV) which may need division by coagulation diathermy to open up the recessus of Rex (where the umbilical vein joins the left portal vein). On the right side, the division of right vascular pedicle into anterior and posterior should be visualized. The “width” of the transected portal plate should extend from this bifurcation and a small innominate fossa on the extreme right to the point where the umbilical vein joins the left portal vein.
Porta hepatis transection: Excise remnants flush with the liver capsule by developing a plane between solid white biliary remnant and the underlying liver starting at the gallbladder fossa. Excising liver parenchyma itself does not seem to improve bile drainage and the so-called deep coring adds nothing. All of the denuded area now needs to be incorporated into the Roux loop.
- (v)Roux loop and portoenterostomy: A standard retrocolic Roux loop measuring 40–45 cm should be constructed. The jejunojejunostomy lies about 10 cm from the ligament of Treitz and can be stapled or sutured. The proximal anastomosis must be wide (~ 2 cm) and end-to-side is appropriate. Fine precise suturing (e.g., 6/0 PDS®) at the edge of the portal plate is satisfactory (Fig. 4).
Options and Alternatives
There is of course an option not to proceed with KPE following confirmation of the diagnosis though this is uncommon. In our own national program, only 2.8% of 620 infants were directed toward primary liver transplantation. The usual reasoning is that an attempt at KPE is felt to be futile by reason typically of advanced cirrhosis. These are usually infants who present late, perhaps at > 100 days of age. Though even in this cohort we have shown a not insignificant 5-year native liver survival of about 40% (Hadzic et al. 2003).
When there is patency of the native gallbladder and common bile duct, some, at least in France, would consider a portocholecystostomy, as it does have the advantage of abolishing the risk of postoperative cholangitis. However, the anastomosis is not as flexible as a standard Roux loop, and revisions for repeated biliary obstruction have been described (Zhao et al. 2011).
Laparoscopic KPE have been reported (Dutta et al. 2007), but this has not been taken up by the larger centers performing the standard KPE on a regular basis. It has become apparent that laparoscopic Kasai surgery doesn’t offer anything advantageous to the child beyond the cosmesis of a better scar and perhaps an adhesion-free abdominal cavity for the transplant surgeon, though even this is arguable (Hussain et al. 2017).
Key outcomes such as clearance of jaundice results are certainly not better and are rarely comparable to the open approach. Some initial advocates in Hong Kong and Hannover have since returned to the standard open approach (Wong et al. 2008; Ure et al. 2011). The problems seen are likely to be due to the difficulties with portal plate dissection using currently available laparoscopic instruments. Radical resection of all extrahepatic biliary remnants from all biliary sectors and a wide portoenterostomy encompassing all the margins of that resection are the key features to maximize results, and it can be a difficult, delicate dissection in the open operation without the constraints of laparoscopic surgery.
Sometimes, the anatomical configuration of the less common type 1 and type 2 BAs, particularly cystic biliary atresia, will allow a hepaticojejunostomy to be performed as there is still a bile duct to join to. However, given that these groups do have a better long-term outcome (Davenport et al. 1997; Superina et al. 2011), it is probably more sensible to dissect it higher and clear the portal plate as in a standard KPE than risk this approach.
Adjuvant Therapy for Biliary Atresia
A number of drugs have been highlighted to have the potential to improve the outcome of KPE, but there has been little hard scientific data published to provide real evidence for their use.
There have now been two prospective, double-blind, randomized, placebo-controlled trials. The first one used a low dose of prednisolone (2 mg/kg/day) in two English high-volume centers (Davenport et al. 2007) in 73 infants. This showed a statistically significant improvement in early bilirubin levels (especially in the “younger” liver) in the steroid group but did not translate to a reduced need for transplant or improved overall survival. The other recently published study is the START Trial (Bezerra et al. 2014). This randomized 70 infants from 14 North American centers to a steroid arm using initially IV methylprednisolone 4 mg/kg/day for first 3 days followed by oral prednisolone (4 mg/kg/day till second week, 2 mg/kg × 2 week, followed by a tapering protocol over the next 9-week period). Although there was a difference in the main primary endpoint (clearance of jaundice) from 48.6%, in the placebo group, to 58.6% in the steroid group, this did not attain statistical significance. As the median time to KPE was relatively high in this study, they also did a subgroup analysis of infants <70 days at KPE (n = 76). This showed that 72% (28/39) in the steroid group cleared their jaundice compared to 57% (21/37) in the placebo group, still not statistically different (P = 0.36), but noticeably the same magnitude as was found in the UK study. There is therefore a clear possibility of a type 2 error (i.e., accepting the null hypothesis when there actually is a true difference).
We recently reported a follow-up study to the original trial which now examined the use of a high-dose prednisolone cohort (starting at 5 mg/kg/day) (Davenport et al. 2013). This showed the same beneficial biochemical effects (now including a reduction in AST and APRi levels) with an increased proportion of those who cleared their jaundice in the steroid groups [67% (41/62) versus 52% (47/91); P = 0.04].
Dexamethasone has also been recommended by one UK center (Leeds) commencing orally (0.3 mg/kg twice daily for 5 days, 0.2 mg/kg twice daily for 5 days, and 0.1 mg/kg twice daily for 5 days), beginning on postoperative day 5 (Stringer et al. 2007).
Despite the lack of hard research evidence, many centers continue to use steroids in a variety of regimens. Its use is particularly high in Japan with over 90% of the institutions using some form of steroids after KPE. Its use in the USA is a little more restricted but still 46% of the infants were being prescribed steroids in one survey (Lao et al. 2010). All three of the English specialist centers use a variety of post-KPE steroid regimens.
Ursodeoxycholic Acid (UDCA)
This is widely thought to be beneficial, but only if surgery has already restored bile flow to a real degree. UDCA “enriches” bile and has a choleretic effect, increasing hepatic clearance of supposedly toxic endogenous bile acids, and may confer a cytoprotective effect on hepatocytes. A study assessed the effect of UDCA on liver function in children >1 year post-KPE in a crossover study in 16 children with BA who had undergone “successful” surgery defined by resolution of jaundice 6 months after surgery. These patients were all treated with UDCA (25 mg/kg/day in three divided doses) for 18 months at which point treatment was stopped and their clinical and biochemical status monitored. Six months later only one had worsened clinically with recurrence of jaundice; however, all but two had sustained significant worsening in liver enzymes. These were all then restarted on UDCA, and 6 months later, all of these had significant diminution in their liver enzymes (Willot et al. 2008).
Both Japanese and Chinese centers routinely prescribe the Chinese herb “inchinko-to” to infants post-KPE, and one of the claimed benefits includes inhibition of apoptosis and inhibition of liver fibrosis (Inuma et al. 2003). For instance, Tamura et al. (2007) reported a prospective study of 21 children post-KPE who had cleared their jaundice but who had persistent elevated liver enzymes and GGT. Inchinko-to was given to 12 for up to 3 years, while the remainder persisted in their standard regimen without any herb. Liver enzymes, bile acids, and markers of liver fibrosis were measured sequentially. There were no side effects of treatment. In the inchinko-to group, markers of liver fibrosis (e.g., hyaluronic acid) were significantly decreased at 1 and 3 years without change in liver enzymes, bile acids, or bilirubin.
Continuation of the natural history of BA and ineffective KPE are the most common problems postoperatively and lead invariably to end-stage liver disease and cirrhosis. Jaundice deepens and is accompanied by abdominal distension and ascites with failure to thrive and malnutrition. These infants require urgent consideration of liver transplantation. Other problems are related to the KPE procedure itself, and there are some specific complications which can occur independently of this process though.
The direct anastomosis of bowel to the liver creates a bilio-intestinal link which predisposes to ascending cholangitis and is seen in up to 60% of large series (Ecoffey et al. 1987; Rothenberg et al. 1989; Ernest van Heurn et al. 2003). This is much more likely to occur in those with BA compared to those with choledochal malformations who have the same reconstruction as the latter’s bile flow is so much better than even the best KPE. The risk is apparent in the first 2 years postsurgery although the reason for the diminution in risk is obscure. Presumably there is some time-dependent change in local immunological defenses. The CHiLDREN consortium reviewed 219 long-term survivors with a median age of 9 years and reported an incidence of cholangitis of 17% in the preceding 12 months (Ng et al. 2014).
Most children will present with pyrexia, worsening jaundice, and a change in liver biochemistry and should be treated aggressively with broad-spectrum intravenous antibiotics effective against Gram-positive organisms (e.g., gentamicin, meropenem, Tazocin™ (piperacillin/tazobactam)) (Wong et al. 2004).There is some evidence to suggest synergistic action of third-generation cephalosporin with aminoglycosides (Luo and Zheng 2008).
Many protocols also advocate long-term prophylactic antibiotic use to reduce the incidence of cholangitis (Bu et al. 2003; Mones et al. 1994) though again the evidence is thin. Bu et al. (2003) reported a prospective randomized study of 19 children who had had at least one previous episode of cholangitis to evaluate the efficacy of trimethoprim/sulfamethoxazole or neomycin as oral prophylactic agents to prevent cholangitis. Though there was no difference between the two treated groups, they did have a lower frequency of cholangitis and a higher survival rate compared to historical controls.
A number of modifications to the Roux loop have been suggested to reduce the risk of cholangitis. Early surgeons suggested bringing the Roux loop out as a stoma (Ohya et al. 1990), separating it from the intestinal stream, while latterly others have advocated creation of a “valve” in the loop by intussuscepting mucosa (Endo et al. 1999) though none have really stood the test of time (Ogasawara et al. 2003).
There is one surgical modification which does significantly reduce the risk of cholangitis altogether by avoiding a Roux loop and draining the denuded portal plate into the opened out native gallbladder as a portocholecystostomy. This has been a particular favorite of the French group in Bictre, Paris, and is anatomically possible in about 10–20% of KPE where the gallbladder is a mucocele and the common bile duct is patent. The subsequent cholangitis rate approaches zero, but the revision rate is higher as the drainage is more tenuous (Matsuo et al. 2001 and personal communication Prof. F Gautier).
Portal Hypertension and Esophageal Varices
Abnormally raised portal venous pressure (i.e., portal hypertension) is seen at the time of KPE in about 70% of BA infants (Shalaby et al. 2012). It is caused by increasing liver fibrosis and correlates with age at KPE, bilirubin level, and ultrasound measured spleen size. Overall however it is a poor predictor of outcome either in terms of response to KPE or more surprisingly even in those who will go on to develop varices. The result of the KPE in terms of clearance of jaundice and more importantly the abbreviation and perhaps attenuation of the hepatic fibrotic process determines variceal formation.
Using endoscopic surveillance about 60% of children who survived beyond 2 years will have definite varices and of these about 20–30% will bleed (Stringer et al. 1989; Duché et al. 2010). In the recently reported North American consortium study based on long-term survivors with BA, 43 (26%) of the cohort of 163 subjects (median age 9 years) had had some form of complication of portal hypertension (variceal bleed, ascites), while a further 37 (23%) subjects met their definition of portal hypertension [i.e., spleen palpable> 2 cm below the costal margin and thrombocytopenia (platelet count < 150,000/ml)] (Shneider et al. 2012b).
Esophagogastric varices take time to develop, and bleeding is unusual before 9 months of age and more usually occurring from 2 to 3 years. Emergency treatment of bleeding varices specifically includes the use of vasopressin (e.g., terlipressin) or somatostatin analogues (e.g., octreotide) sometimes even a Sengstaken-Blakemore tube (Eroglu et al. 2004). Most are treated endoscopically with banding or in the very young injection sclerotherapy. In those with reasonable restoration of liver function, this should be all that is necessary; however, those who are still jaundiced may require transplant assessment. The role of propranolol in BA children with portal hypertension particularly those with cirrhosis has not been formalized.
This is related to and caused by portal hypertension in part, but other contributory factors include hypoalbuminemia and hyponatremia. It predisposes to spontaneous bacterial peritonitis. Conventional treatment includes a low-salt diet, fluid restriction, and the use of diuretics particularly spironolactone. Often seen in settings of malnutrition, consideration should be given to nasogastric feeding to try and increase calorie and protein intake.
Nutrition and Growth Failure
It should be apparent that maintenance of good nutrition and restoration of normal growth are vital to an infant well-being. This is particularly so (and harder to achieve) in those who have failed to clear their jaundice and should be heading down the transplant pathway. There is a clear correlation between improvements in nutrition and better outcome following transplantation (De Russo et al. 2007, Carter-Kent et al. 2007). Indeed malnutrition is one of the variables used in calculation of the pediatric end-stage liver disease (PELD) score used in the USA to prioritize children for liver transplantation (McDiarmid et al. 2002).
Growth failure may be defined as height or weight <2 standard deviations below the population mean and is a major risk factor for poor outcome, and a recent report from the Biliary Atresia Research Consortium (BARC) identified this as a key measure of outcome associated with transplantation or death by 24 months of age (Sokol et al. 2007). Similarly, Studies in Pediatric Liver Transplantation (SPLIT) looked at 775 children with BA awaiting transplantation and again identified growth failure as an independent risk factor for pre-and posttransplant mortality and graft failure (Utterson et al. 2005).
The main emphasis to correct nutritional deficiencies has been on increasing the supply of calories. Sometimes this may simply be by adopting a more reliable mode of delivery. Chin et al. (1990) and Holt et al. (2000) achieved higher growth rates simply with supplemental feedings via a nasogastric tube. Overall calorie prescription should aim for 110–160% of normal caloric intake and may need to consist of semi-elemental formula, medium-chain triglycerides (MCT) (by up to 60% of fat provided), and may be supplements of branched chain amino acids. In infants our standard prescribed formula feed has included Heparon© Junior and Caprilon© (higher proportion of MCT) (UK manufacturer, SHS International) providing approximately 120–150 kcal/kg/day. Short-term PN should be considered in those where NG feeding is not being tolerated then and can improve nutrition of children on the transplant waiting list (Sullivan et al. 2012).
Virtually all infants are deficient in fat-soluble vitamins (D, A, K, and E) at presentation and all require active correction of this. This may forestall complications such as rickets and pathological bone fractures (vitamin D), coagulopathy and spontaneous bleeding (vitamin K), and even cerebellar ataxia and impaired vision (vitamin E). Supplementation and monitoring of fat-soluble vitamins should be regarded as absolutely mandatory and again is particularly important in those failing to clear their jaundice (Sokal et al. 1994; Schneider et al. 2012).
Developmental Delay and Neurocognitive Function
Many of these infants and children are severely affected by chronic liver disease requiring extended periods of hospital stay and as such have the potential to miss out on normal educational opportunities. Much is directly proportional to the severity and duration of illness and malnutrition though there may also be more specific impairment of normal neurocognitive function by vitamin and mineral deficiencies. Howard et al. (2001) published a comparative study of developmental outcome from adolescents with their native livers in the UK and Japan. This showed actually an overall high quality of life and comparable attainment of educational standards in both countries. By contrast, there is some evidence that those coming to transplant may not be as fortunate. Neurocognitive function was seen to be significantly deficient in 10–15%, 26% had learning disabilities, and almost 40% had special educational needs in one long-term American study of 51 children living with a liver transplant (Midgley et al. 2000).
Outcome and Results
Though the results of surgery of BA are largely unpredictable in individual cases, a number of factors can be identified as important (Nightingale et al. 2017). These include:
Age at Kasai Portoenterostomy
The effect of age on outcome following KPE is still complex and incompletely understood. If a bile-obstructed liver is left untreated, then fibrosis and cirrhosis are invariable. But, the effect is neither simple nor linear. We examined the concept with respect to infants treated in our institution as an example of a relatively homogenous, uniform scheme of management with experienced surgeons in a high-volume center. We divided infants coming to KPE in the 1980s and 1990s into age quartiles and calculated their subsequent native liver survival. Only the oldest quartile (<100 days) had a demonstrably worse outcome (Davenport et al. 1997), even then not reaching statistical significance. We then looked at 225 infants from the 1990s and 2000s and divided them again into specific age cohorts and also by their putative etiological group. This time a significant effect on outcome was shown but only for those where the BA was considered developmental (BASM and cystic BA). For all those remaining infants with isolated BA, there was barely any discernable effect again up until about 100 days of age (Davenport et al. 2008). Throughout this period we have never found any evidence of a significant cutoff (e.g., < 60 days, < 6 weeks, < 8 weeks, etc.), and such an arbitrary simplistic approach to prognosis needs to be consigned to the dustbin of history.
It has previously been shown in the UK that the more KPEs a surgeon does (as a center, arbitrarily >5/year), the better the outcome (McClement et al. 1985; McKiernan et al. 2000). This has led to superspecialization in England and Wales, Denmark (Kvist and Davenport 2011), Finland (Lampela et al. 2012), and the Netherlands. Certainly each country must have a duty of care to even its smallest citizens and make provisions to improve collaboration and communication between centers and allow the best possible outcome.
Liver Histology and Biliary Remnant
There is improved outcome in types 1 and 2 compared to type 3 BA and cystic BA compared to non-cystic BA (Caponcelli et al. 2008; Davenport et al. 1997; Superina et al. 2011). Almost certainly this is due to improved preservation of the connections between the intra- and extrahepatic bile ducts and ductules. Infants with BASM also have a worse outcome and also appear at risk of sudden death in the first year following KPE. The cause of this is unknown and may be due to an intrinsically worse liver disease, a smaller biliary remnant tissue, or the effect of other anomalies (e.g., cardiac).
Prospective evaluation of the macroscopic features of the hepatobiliary elements (hardness of the liver, presence of ascites, etc.) was relatively poorly predictive of outcome with only actual size of resected biliary remnants being really discriminatory (Davenport and Howard 1996). Microscopic examination of the transected bile duct remnant will show a varying amount of residual ductules. Older series suggested that only those showing large ductules ( >300 μm) had a distinctly better outcome (Howard et al. 1982), but a later evaluation showed that minimal or no ductules in the remnant were also predictive of lack of effect of KPE (Tan et al. 1994).
Conclusion and Future Directions
Cirrhosis , present in about 10% of infants at the time of KPE, is the usual histological state in about half of the long-term survivors by the time of adolescence (Hadzic et al. 2003). Many of these however are completely asymptomatic but it should make one wary about suing the term “cured” in this respect. Almost certainly this will limit the natural life span of the Kasai survivor and the specter of transplant will once more emerge. Limitation of this process by effective antifibrotic drugs remains tantalizingly close but none are ready for real clinical application (Czaja et al. 2014). Many pharmacological agents that diminish oxidative stress [e.g., S-adenosylmethionine (SAMe), N-acetylcysteine (NAC), vitamin E] and at a cellular level reduce the pro-inflammatory cytokine cascade (e.g., infliximab, etanercept) and limit hepatic stellate cell activation and minimize myofibroblast proliferation (e.g., cannabinoid antagonists) have been identified.
Still, although the cause of biliary atresia remains elusive, a complementary system of surgical treatment has evolved, which has improved the overall survival to adulthood in affected infants from around 10% to about 90%. Not many surgical diseases can claim such a dramatic turnaround in prognosis.
- Davenport M, Savage M, Mowat AP, Howard ER. The biliary atresia splenic malformation syndrome. Surg. 1993;113:662–8.Google Scholar
- Kasai M, Suzuki S. A new operation for “non-correctable” biliary atresia – portoenterostomy. Shijitsu. 1959;13:733–9.Google Scholar
- Ng VL, Haber BH, Magee JC, et al. Childhood Liver Disease Research and Education Network (CHiLDREN). Medical status of 219 children with biliary atresia surviving long-term with their native livers: results from a North American multicenter consortium. J Pediatr. 2014;165:539–46.Google Scholar
- Park WH, Choi SO, Lee HJ, et al. A new diagnostic approach to biliary atresia with emphasis on the ultrasonographic triangular cord sign: comparison of ultrasonography, hepatobiliary scintigraphy, and liver needle biopsy in the evaluation of infantile cholestasis. J Pediatr Surg. 1997;32:1555–9.CrossRefPubMedGoogle Scholar
- Russo P, Magee JC, Anders RA, et al. Childhood Liver Disease Research Network (ChiLDReN). Key histopathologic features of liver biopsies that distinguish biliary atresia from other causes of infantile cholestasis and their correlation with outcome: a multicenter study. Am J Surg Pathol. 2016;40(12):1601–15.Google Scholar
- Shneider BL, Abel B, Haber B, et al. Childhood Liver Disease Research and Education Network. Cross-sectional multi-center analysis of portal hypertension in children and young adults with biliary atresia. J Pediatr Gastroenterol Nutr. 2012b;55: 567–73.Google Scholar
- Stagg H, Cameron BH, Ahmed N, et al. Canadian Biliary Atresia Registry. Variability of diagnostic approach, surgical technique, and medical management for children with biliary atresia in Canada – is it time for standardization? J Pediatr Surg. 2017;pii: S0022-3468(17)30074–X.Google Scholar
- Superina R, Magee JC, Brandt ML, et al. The anatomic pattern of biliary atresia identified at time of kasai hepatoportoenterostomy and early postoperative clearance of jaundice are significant predictors of transplant-free survival. Ann Surg. 2011;254:577–85.CrossRefPubMedPubMedCentralGoogle Scholar
- Thomson J. On congenital obliteration of the bile ducts. Edinburgh Med J. 1891;37:523–31.Google Scholar