“Yellow stools in neonatal cholestasis exclude biliary atresia.” This conventional wisdom led to the development of the infant stool color card, which alerts parents to seek medical referral when pale stools are observed, a strategy that has been shown to improve survival in infants with biliary atresia (BA). Here, we present a case of a newborn with significant direct hyperbilirubinemia (direct bilirubin level of up to 9.2 mg/dL on day of life 10) who continued to produce colored stools. Whole-genome sequencing results were negative for genetic causes of cholestasis. Hepatobiliary scintigraphy findings were nonexcretory. A liver biopsy specimen revealed cholestasis, ductular hyperplasia, giant cell formation, minimal inflammation, minimal portal or periportal fibrosis, and no evidence of viral changes. On day of life 38, during the exploratory laparotomy, the patient was found to have complete absence of the extrahepatic biliary tree, or biliary aplasia, possibly a rare, severe form of BA. This report aims to increase our vigilance and help prevent diagnostic error in patients with signs and symptoms of BA who may produce pigmented stools. Primary care physicians should hence refer an infant (early and urgently) to a pediatric gastroenterologist for further workup for a direct bilirubin level >1.0 mg/dL with any total bilirubin level, irrespective of the color of the infant’s stools.
Cholestasis refers to decreased bile formation or flow with hepatic retention of biliary substances normally excreted into the intestinal lumen.1 It is evidenced by a direct bilirubin level >20% of the total serum bilirubin level,2 or a direct bilirubin level >1.0 mg/dL if the total serum bilirubin level is <5.0 mg/dL, and usually presents with jaundice and, if there is obstruction to the bile flow (such as in biliary atresia [BA]), with acholic stools.1,3 Cholestatic jaundice occurs in 1 of 2500 term infants.4 It has a multitude of possible etiologies1 ; the most common during the first months of life is BA, occurring in 25% to 40% of the cases.1 Because BA is a progressive destructive cholangiopathy, the vast majority of patients with BA develop acholic stools at some point during their first 60 days of life,5 as a result of the expanding destruction of the extrahepatic biliary system, after having excreted pigmented stools earlier in life.
A Kasai hepatic portoenterostomy as early as possible within 45 days of life is the preferred primary treatment of BA,6 an oftentimes temporizing procedure because at least 80% of patients with BA will eventually require a liver transplant.7 In a Kasai procedure, the entire (variably obliterated) extrahepatic biliary tree is excised, and a jejunal Roux loop is anastomosed to the cut surface to drain the bile from the nonobliterated part of the biliary ductules.8 Importantly, patient survival with the native liver successively decreases with increasing age at the time of the Kasai procedure9 ; the procedure is less likely to be successful if performed after 60 days of age.
A term, large for gestational age male infant weighing 4440 g was born via cesarean delivery to a 47-year-old G6P5 mother with diabetes. Apgar scores were 8 at both 1 and 5 minutes. After delivery, the infant required nasal continuous positive airway pressure for respiratory distress and was admitted to the NICU. He was found to be hypoglycemic, with a glucose level of 17 mg/dL, and thrombocytopenic, with 38 000 platelets per mm3, and was treated for suspected sepsis. An echocardiogram was performed because of a murmur and worsening hypoxic respiratory failure and revealed total anomalous pulmonary venous return to the coronary sinus. The total anomalous pulmonary venous return was repaired on day of life (DOL) 6.
The infant passed meconium within 24 hours and continued to have green-brown-yellow stools (for more details, see Fig 1). A direct bilirubin level of 3.4 mg/dL and a total bilirubin level of 14.3 mg/dL on DOL 5 (Fig 2) prompted an abdominal ultrasound. A 1.5 × 1.0 × 1.4-cm complex cyst was identified within the right lobe of the liver, which, on repeat ultrasound with Doppler, was found to represent a thrombus within the umbilical vein. The liver had normal echogenicity; however, no gallbladder or common bile duct could be visualized, raising concern for BA. The Doppler revealed significant hepatic arterial flow with possible superimposition of portal venous flow, suggestive of a hepatic arterioportal fistula. Computed tomography angiography confirmed the hepatic arterioportal fistula (Fig 3A) and also revealed a hepatic cyst versus a rudimentary intrahepatic gallbladder in segment IV of the liver. Whole-genome sequencing was performed on DOL 11, and results were negative for genetic causes of cholestasis.
An additional Doppler documented abnormal drainage of a left hepatic vein into the left atrium (Fig 3B). Hepatobiliary scintigraphy on DOL 19 revealed normal hepatic uptake of the contrast but no visualized gallbladder, biliary tree, or bowel and no excretion into the small bowel 24 hours post injection, which was suggestive of BA. A liver biopsy specimen on DOL 23 revealed cholestasis, ductular hyperplasia, few portal tracts, giant cell formation, minimal intralobular inflammation, minimal portal triad inflammation, minimal portal or periportal fibrosis, and no evidence of viral changes. The matrix metalloproteinase-7 level, a highly sensitive and specific serum biomarker for BA,10 was obtained on DOL 26 and was markedly elevated to 234.1 ng/mL (median of normal is 2.86 ng/mL, with an interquartile range of 1.32–5.32 ng/mL), further supporting the diagnosis of BA.10
Despite the complicated vascular anatomy (hepatic arterioportal fistula and left hepatic vein draining into the left atrium), the decision was made to proceed with an intraoperative cholangiogram and possible hepaticojejunostomy on DOL 38. The unusual intraoperative findings are described in the surgical report:
Intraoperative findings included a slightly medialized porta hepatis, found more towards the midline and the right side of the abdomen. The portal vein was identified as well as the bifurcation. The arterial inflow of the liver was identified as well. There was no evidence of any atretic biliary tree, or gallbladder. There was no trace of a strand of tissue along the porta hepatis climbing up towards the bifurcation of the portal vein, which would represent the scarred area for a possible portoenterostomy. The parenchyma of that area was no different than the surrounding parenchyma from any other identifiable location on the liver. I asked [another surgery attending] to perform an intraoperative consultation. We judged that the patient would not benefit from a portoenterostomy as there was no identifiable target to sew to, rather there was fairly homogeneous looking Glisson’s capsule in hepatic parenchyma.
Despite an absent biliary tree and no Kasai procedure, the patient’s stools remained pigmented for the majority of his hospitalization. Still moderately jaundiced, he was discharged from the hospital on DOL 44 on ursodeoxycholic acid, expressed breast milk or hydrolyzed formula, and vitamin supplementation. A genetic cholestasis panel obtained after his hospitalization revealed heterozygosity for genes for autosomal-recessive disorders, including the genes LIPA (in Wolman disease), MSK1 (in Bardet-Biedl syndrome and Joubert syndrome), and UGT1A1 (in Crigler-Najjar syndrome and Gilbert syndrome), none of which would explain his direct hyperbilirubinemia or absent extrahepatic biliary tree. He is being followed-up in a liver transplant clinic and continues to pass pigmented stools (Fig 1). At 11 months of age, he has done remarkably well clinically, although he is being referred for a liver transplant evaluation given a decline in ponderal growth and severe deficiency of vitamin D despite supplementation.
Here, we present an unusual case of what would most closely correspond to biliary aplasia given the complete absence of any (atretic) biliary tree, gallbladder, fibrous strands, or identifiable target to perform a cholangiogram or Kasai procedure.
An electronic literature search was conducted by using PubMed to identify any similar published cases. The search yielded 6762 titles, from which 71 articles of case presentations or case series of BA or other anatomic hepatobiliary abnormalities were selected. Of these, only one other case was identified in the existing literature with complete absence of the biliary tree and gallbladder, including lack of any trace of fibrous strands.11
BA is usually classified into 3 types according to the most proximal biliary obstruction: type 1 (5% of the cases) has luminal patency to the common bile duct (which can be cystic) with bile-containing gallbladder, type 2 (2%) has luminal patency to the common hepatic duct, and type 3 (>90%) has an entirely solid extrahepatic biliary tract within the porta hepatis, possibly with a solid gallbladder or mucocele.8 Biliary aplasia could then be classified as a very rare, severe, atypical form of BA type 3. On the other hand, it is conceivable that the presented case, and hence biliary aplasia, is not part of any of the 3 BA types given the complete (likely congenital) absence of any (fibrotic) biliary tree remnant, and it could represent an entirely separate entity possibly secondary to nondevelopment of the caudal pars cystica of the hepatic diverticulum (an embryologic precursor) to the extrahepatic biliary tree and gallbladder.12
Additionally, BA can be categorized into a nonsyndromic group without associated major malformations (84%); a second group with at least 1 malformation, such as cardiovascular or genitourinary anomalies, but without laterality defects (6%); and a third syndromic group with gastrointestinal, cardiovascular, or splenic anomalies with laterality defects, including situs inversus (10%).1,13 Given associated hepatic and cardiovascular abnormalities, including a hepatic arterioportal fistula and an abnormal venous return to the heart, without laterality or splenic defects, the presented case (if considered an extreme case of BA) could be subsumed under the second group.13–15
Now to the central question: How is it possible that there are brown-green-yellow stools even in this severe form of BA without any identifiable biliary system? Already >50 years ago, Cameron et al16 injected bilirubin tagged with radioactive carbon (carbon 14) intravenously into 3 children with BA and found that most of the carbon 14–tagged bilirubin was detected in the urine; however, a small percentage was detected in the feces of these patients. This indicates that despite complete biliary blockage, bile, according to the authors, might be “reaching the gastrointestinal tract by way of the gastric and pancreatic juices and the succus entericus [alkaline secretion produced by glands in the wall of the duodenum] and could account for the radioactive isotope recovered in the feces.”16,17 Despite the evidence that there appears to be at least 1 additional pathway for passage of small amounts of bile besides the biliary tree, no further studies have been conducted to examine this more closely.
Given the higher prevalence of BA in Taiwan, with 1 in 2700 to 6600 live births,5,18 Chang and co-workers5,19 implemented a universal stool card screening program in 2002 to promote early diagnosis of BA. This card improved early diagnosis of BA and increased the rate of Kasai operation at <60 days of life from 49.4% to 65.7% after implementation, with associated improvement in patient survival.19 The sensitivity of the stool card screening program is 89.7%,5 which means that 10.3% of the cases of BA did not have acholic or clay-colored stools but pigmented stools in their first 60 days of life. Furthermore, researchers of a study in the United Kingdom found that only 62.8% of medical providers were able to identify acholic stools.20 This indicates the need for more objective tools, such as the smartphone application PoopMD, which, in a pilot study, was able to differentiate acholic from regular stools with a sensitivity of 100% and a specificity of 89%.21 However, because a number of patients with BA do not develop acholic stools within the critical first 1 to 2 months of life,5 all patients with neonatal cholestasis should be urgently referred to a pediatric gastroenterologist. The pediatric gastroenterologist should complete early full laboratory testing (including matrix metalloproteinase-7, if possible, given its excellent sensitivity and specificity for BA),10 imaging studies as indicated, and likely a liver biopsy in the first month of life. There are at least 2 reports (Harpavat et al22,23 ) demonstrating that a direct bilirubin level measured within 60 hours of life is abnormally high in infants subsequently diagnosed with BA, leading the authors to recommend that all newborns be screened for both total and direct bilirubin levels shortly after birth. Our case emphasizes that workup for BA should ideally be completed early, irrespective of stool color. Kasai portoenterostomy for BA is associated with better survival if done before 60 days of life9 and with best survival if done before 45 days of life.9
Primary care physicians should hence refer an infant early and urgently to a pediatric gastroenterologist for a direct bilirubin level >1.0 mg/dL with any total bilirubin level,1 irrespective of the color of the infant’s stools.
We thank Dr Christopher Dory (Department of Radiology, Rady Children’s Hospital-San Diego) for help with the preparation of the imaging studies. We further thank Moses Utomi, MFA, for proofreading the manuscript.
Drs Hartmann, Carter, Keller, Saenz, and Schwarz are responsible for the reported case, participated in the concept, design, analysis and interpretation of data, and drafting and critical review of the manuscript, approved the final manuscript as submitted, and agree to be accountable for all aspects of the work.
FUNDING : Supported in part by National Institutes of Health grant K12 HD 850-36 (to Dr Hartmann). Funded by the National Institutes of Health (NIH).
POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.