Neonatal jaundice lasting longer than 14 days after birth is unusual. In some of these infants it is associated with breast feeding (1), but cholestasis must always be excluded by measurement of the conjugated bilirubin in addition to the total bilirubin. Neonatal cholestasis is a serious condition requiring immediate further investigation because it may be due to life-threatening disorders that must be addressed early in order to rapidly initiate specific treatment (2).

The incidence of neonatal cholestasis is ~1 in 2500 live births (3, 4). It results from impairment in bile excretion, caused by various disorders such as defects of intrahepatic production or transmembrane transport of bile, or a mechanical ductal obstruction preventing bile flow. Of the various conditions that can present with neonatal cholestasis, biliary atresia is the most common cause throughout the world and has been reported to account for one-third of the cases (5–7). Early diagnosis is particularly important in biliary atresia because timely surgical intervention with Kasai portoenterostomy correlates with better long-term outcome and patient survival (8–10). However, the diagnosis of cholestatic jaundice in neonates is often delayed and this may result in a poor patient outcome.

Here, we report the retrospective analysis of 82 infants with neonatal cholestasis who were presented to a single tertiary center during a 4-year period. This report provides information on the prevalence and mortality of the various causes of neonatal cholestasis, the diagnostic algorithm, and the timing of specific medical and surgical intervention in infants with biliary atresia in this cohort.

A variety of disorders can present with cholestasis during the neonatal period with biliary atresia as the most frequent cause and idiopathic cases as the second common entity. A previous study reported that 35% of infants with neonatal cholestasis had biliary atresia and 30% were considered as idiopathic cases having neonatal hepatitis (5). Other causes included α1AT deficiency (17%), congenital viral infections (9%), Alagille syndrome (6%), and choledochal cysts (3%). Almost 25 years later, this retrospective study confirms these data in part since biliary atresia accounted for about 41% of all infants with neonatal cholestasis that were referred to our hospital (Figure 1; Table 2).

In contrast to Mieli-Vergani (5), however, the percentage of infants diagnosed with idiopathic cases is lower in the current cohort (13 vs. 30%). This is likely due to the introduction of new diagnostic tools that enable the detection of previously unrecognized conditions. Rates of PFIC and TPN-induced cholestasis have been underestimated in recent years as these disorders were not described in previous demographic studies (Table 2). PFIC was identified in 10% of the cases in this cohort, which is in line with a previous reported empirical evidence of 10–15% of children with neonatal cholestasis (20). The true incidence is not precisely known but is estimated to be 1/50000–1/100000 births (20). Subtypes 1 and 2 represent two-thirds and PFIC3 the remaining third of cases with PFIC (21). Although to date three types of PFIC have been identified (20), only the subtypes PFIC2 (n = 7) and PFIC3 (n = 1) were diagnosed in this cohort. In addition, the prevalence of α1AT deficiency in the current cohort is low compared to previous reports identifying α1AT deficiency as the third common cause among Caucasians (7).

Recently, molecular genetic tests became available to diagnose the Alagille syndrome, PFIC I-III, and α1AT deficiency (22). These tests have been reported to efficiently identify disease-causing mutations and thus represent an adjunct alternative to timely clear up the cause of cholestasis. Moreover, a more frequent genetic testing for ATP8B1, ABCB11, and ABCB4 would generally identify more cases of PFIC because the immunostaining may namely recognize the relevant proteins but does not detect a loss function mutation – a reason why PFIC may be underdiagnosed in this and current other studies. It thus can be expected that genetic testing will imply the potential to expand the PFIC group among cases of neonatal cholestasis in future studies. In preterm infants, neonatal cholestasis was regressive in all cases. The cause for the development of neonatal cholestasis in these children remains speculative. The majority of preterm neonates developed cholestasis early after birth (63%), a point of time when total parenteral nutrition was just initiated. Alternatively, prematurity in combination with perinatal stress and/or asphyxia as the worst case may have resulted in hepatic dysfunction leading to the perinatal onset of neonatal cholestasis and may represent a potential causal factor for a generally benign and reversible cholestasis in preterm and also at term neonates (23).

In this study, biliary atresia represents by far the most frequent disorder causing neonatal cholestasis (Figure 1; Table 2). However, this study has a limitation. The relative incidences reported here but also in other previous studies (6) need to be handled carefully, since they represent retrospective analyses of single tertiary pediatric transplant centers. The subanalysis of our cohort demonstrates that the study group includes two different groups – one part is the local region’s infants with cholestasis undifferentiated and for further investigation at initial referral, and the second a selected group containing already diagnosed children referred from a wider area for further treatment and/or transplantation (Figures 1 and 2). The latter subset of the group does bias the incidence and mortality of disorders of the whole population to a certain extent because children referred from other centers may represent a grouping of more severe cases whereas some of the cases with more benign cases of neonatal cholestasis would not have been referred. Therefore, the severity of diseases may be skewed toward a more severe spectrum of possible causes of neonatal cholestasis (e.g., biliary atresia) that may generate the differences in the incidences of other population-based studies.

On the other hand is the fact that the Children’s Hospital Essen is located in one of the biggest congested urban area of Germany containing over 5 million habitants. The catchment area is small and the majority of the children (87%) that were referred to our hospital with neonatal cholestasis were from community hospitals from cities within an area of 250 km. Although the consideration of a catchment area may not completely abrogate a possible skewing toward more severe disorders, these children with neonatal cholestasis represent a homogenous population within a well-defined area of West-Germany. The differences with other published series, namely the decrease of idiopathic cases of neonatal cholestasis and the increase of PFIC, may thus be caused at least in part by an increased knowledge of new diagnoses and improvements of the diagnostic work-up.

Neonatal cholestasis is a severe disorder in early childhood with a pretransplant mortality that is not negligible. Especially for children with biliary atresia it was about 12%. Biliary atresia was the most frequent disorder in our and previously reported cohorts (6, 24) leading to liver transplantation. Although only infants with biliary atresia died after transplantation in our cohort the intraoperative complications could have also affected infants with other causes of neonatal cholestasis. The majority of those infants that died before liver transplantation had disorders with an unfavorable prognosis (three with syndromic variants of biliary atresia, one with complex malformation syndrome, and one with Zellweger’s syndrome), only one child had a non-syndromic biliary atresia. However, the fact that syndromic variants seem to carry a higher complication rate may depend on the small series as previous multi-center studies did not find a difference in between syndromic and non-syndromic variants of biliary atresia (25, 26).

It is the late referral of children with biliary atresia that is widely suggested to be accountable for an increased mortality (26–29) since the success rate of the Kasai procedure is closely associated with the age at the time of surgery (9, 10, 30). Besides age at hepatoportoenterostomy several other risk factors for a failure of Kasai surgery have been described and include histologic features (small or absent bile ducts at the portal plate) and recurrent episodes of cholangitis (13). It is reported that children with biliary atresia referred for surgery before 60 days of age do markedly better with a reestablishment of bile flow in more than 80% than those older than 90 days at the time of operation (5, 27, 31–33). This had been shown both in the short-term and in the long-term course (9, 34, 35). The earlier surgery was performed, the later a liver transplantation was usually needed (8). In the short-term, infants that earlier received a Kasai procedure less frequently developed intrapulmonary vascular dilatations, a condition preceding the development of a HPS (14). In the long-term, every second infant with biliary atresia who underwent surgery <30 days of age was still living with his/her own liver 10 years later compared to 15% of those treated after 90 days of age (34). In our cohort, infants with biliary atresia were initially presented to a secondary health care provider with a mean age of 37 days, the mean age at diagnosis was 62 days. This is in line with the average age at diagnosis of biliary atresia, e.g., in the USA (36) and Germany (37) (~60 days). The Kasai procedure was performed soon after diagnosis at a mean age of 66 days demonstrating that the late diagnosis of biliary atresia caused the delay in an appropriate treatment. The majority of the children in this cohort (~70%) underwent surgery before 10 weeks of age and only 27% before 60 days. Hence, corrective surgery within the first 2 months of age is still by far not the routine although substantial efforts have been undertaken to overcome this problem (38, 39).

Among these it is a current endeavor to centralize the services and treat children with biliary atresia in few experienced centers in order to increase surgical success and survival with the native liver (26). A recent large German multi-center study confirmed a relation between outcome and experience of the surgical center and reported a 2-year survival of 20% with their native liver (40). The 2-year survival with the native liver was even less (12%) in our cohort. However, it is important to emphasize that the numbers of this study are biased by the fact that a majority of children with biliary atresia had a history of a failed Kasai procedure and were referred to our center for the purpose of liver transplantation.

The initial and timely detection of biliary atresia is currently and in future will still be a domain of the primary health care provider and depends on the recognition of prolonged jaundice and abnormal stool and/or urine color. It is often challenging as a substantial group of breast fed newborns (ca. 15%) develop benign neonatal jaundice (1) and infants with biliary atresia often appear well during the first 3 weeks after birth. In Germany, at 4–6 weeks of age a visit to a pediatrician is generally recommended and supported by the governmental health insurance system. A routine dipstick test for urinary bilirubin excretion and/or the use of a stool color card at this visit would be an additional non-invasive and simple check to identify cholestatic infants. Furthermore, both types of biliary atresia, the fetal type with early onset and the perinatal type with later onset during the second to fourth weeks of life (6), would be detected at this time point. However, within the last three decades (5) no essential progress had been achieved in early identification of biliary atresia. There is still an urgent need to increase the awareness of prolonged jaundice caused by neonatal cholestasis and of its severity.

When biliary atresia is suspected, ERCP may serve as a reliable, accurate and safe additional beneficial diagnostic tool in unclear cases of children with non-pigmented stool (37, 41, 42). Here, ERCP clarified the diagnosis in 73% of the cases where biliary atresia was suspected (16/22 infants). Both the low complication rate as well as the average procedure time of ERCP was in line with previous studies (37, 43). Hence, ERCP (if conclusive) may represent a safe method expanding the diagnostic repertoire especially 4–6 weeks after birth when histological findings are sometimes not unequivocal (2, 44). Although it is not generally recommended yet we would therefore encourage performing an ERCP and liver biopsy 4–6 weeks after birth at latest if acholic stools are persisting and biliary atresia is suspected. The ERCP procedure however should be performed by operators specifically skilled in pediatric patients in an expert tertiary center in order to guarantee a safe and reliable procedure that then enables to detect biliary tract abnormalities (e.g., biliary atresia or choledochal cysts) with high sensitivity and specificity.

In conclusion, a variety of hepatobiliary disorders with biliary atresia as the major cause can present with cholestasis during the neonatal period. A systemic approach is the key to reliably achieve a rapid diagnosis in order to initiate disease-specific and in many cases life saving therapy. Therefore, any infant presenting with prolonged jaundice at 14 days of age with or without acholic stool must be evaluated by determination of the fractionated bilirubin and if elevated referred to a subspecialist for pediatric hepatology as soon as possible in order to confirm or rule out biliary atresia and other treatable disorders.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.