Clinical and laboratory features of biliary atresia and patterns of management practices: Saudi national study (2000–2018)

Abstract

Background:

We utilized the data from the Saudi national biliary atresia (BA) study (2000–2018) to describe the clinical, biochemical, imaging, and histopathological features of BA and the perioperative clinical practices among local pediatric gastroenterologists.

Methods:

This is a retrospective, multicenter, nationwide study that included 10 tertiary care governmental hospitals including the four liver transplant (LT) centers in different regions across Saudi Arabia.

Results:

BA was diagnosed in 204 infants (106 females; 10% preterm). The median age at referral was 65 days. Congenital anomalies were present in 68 patients (33%); 22 were splenic malformation (10.8%). The medians of laboratory investigations were total bilirubin (189 μmol/l), direct bilirubin (139 μmol/l), ALT (164 u/l), and GGT (472 u/l). The level of serum GGT was normal in 26 cases (12.7%). The ultrasound findings included hypoplastic or atrophic gall bladder (GB) (65%), normal GB (30%), and cord sign (5%). A HIDA scan was performed in 99 cases (48.52%). Magnetic resonance cholangiopancreatography (MRCP) was performed in 27 cases (13%). A total of 179 liver biopsies (88%) were obtained. The most common histopathologic findings were bile duct proliferation (92%), canalicular cholestasis (96%), bile plugs (84%), and portal fibrosis (95%). Cholangiography was performed in 139 cases (68%): operative in 122 (60%) and percutaneous in 17 (8%). A total of 143 children (70%) underwent Kasai portoenterostomy (KPE) at a median age of 70 days. After KPE, steroid was used in 37% of the cases and 100 cases (70%) were prescribed prophylactic antibiotics for variable duration (ranging between 3 and 12 months).

Conclusion:

Our data show marked variation in the diagnostic evaluation and perioperative management of BA cases among the different tertiary centers. There is a need to establish a national BA registry in Saudi Arabia aiming to standardize pre- and postoperative clinical practices. Additionally, normal serum GGT level, normal GB size on ultrasound, and being a premature baby should not preclude the diagnostic workup for BA.

INTRODUCTION

Biliary atresia (BA) is an idiopathic obstructive cholangiopathy characterized by a progressive fibrosclerosing obliteration of the large bile ducts that presents during the neonatal period with cholestasis. Neonatal cholestasis represents a common end result of several pathogenetic mechanisms that might affect the highly vulnerable liver in young infants. As a result, the differential diagnosis of neonatal cholestasis is very broad and identification of etiology is challenging for clinicians because the list includes many entities with overlapping clinical, biochemical, and histological features. Furthermore, the diagnostic workup must be completed in a short time, since some treatable conditions need prompt medical or timely surgical intervention. The age at Kasai portoenterostomy (KPE) procedure has been a significant predictor of outcome of BA, with 70% to 80% of patients clearing the jaundice if KPE is performed before 60 days of age, in contrast to the success rate of 40% to 50% if performed between 60 and 90 days of life and up to 25% if performed after 90 days of life.[1,2,3,4,5] In view of these challenges, clinicians need to promptly recognize the clinical, laboratory, radiologic, and histologic clues that assist in the early diagnosis of BA.

Previous reports from different countries showed marked variation in clinical practices and little consensus on the perioperative management of BA. There is a lack of data on the pattern of practices in the management of BA among pediatric gastroenterologists in Saudi Arabia. The clinical practices could reflect on BA outcomes; hence, understanding the current clinical practices helps the efforts to unify the practice to provide the best care to BA patients. Here, we utilized the data from the Saudi national BA study (2000–2018) to address two main issues: first, to describe the clinical, biochemical, imaging, and histopathological features of BA, and second, to characterize the diagnostic approach and pattern of perioperative clinical practices among local pediatric gastroenterologists.

PATIENTS AND Methods

Study setting and design

This is a retrospective, multicenter, nationwide study that included 10 tertiary care governmental hospitals (including four liver transplant [LT] centers) in different regions across Saudi Arabia. Saudi Arabia consists of 13 provinces; the 10 hospitals are located in the three most populated provinces: Central (five hospitals), Eastern (three hospitals), and Western (two hospitals). In Saudi Arabia, the national health system guarantees free-of-charge services and access to health care for all citizens in the 13 provinces. The healthcare system is designed so that tertiary care is highly centralized in these three main provinces that receive referrals for complex cases from the remaining 10 provinces.

Study population

The details of the study population, inclusion and exclusion criteria, and overall outcomes (including clearance of jaundice post-KPE, survival with native liver (SNL), and overall survival) were published elsewhere.[6] In brief, 204 infants were diagnosed with BA between 2000 and 2018 (age at referral: median 65 days, range 7–245 days; 106 females). One hundred and forty-six patients (71.5%) underwent KPE at a median age of 70 days (range, 12–186 days). The definitive diagnosis was made when cholangiography failed to show a patent biliary tree. In cases where cholangiography was not undertaken, the diagnosis of BA was confirmed by operative findings at subsequent LT and histology of the excised extrahepatic biliary remnant typical of BA. Three of the 146 KPE procedures were performed outside Saudi Arabia and excluded from the analysis. The average age of the study patients at the last follow-up was 7.2 years ± 6 years.

Data collection

Medical records were reviewed to collect gestational age and clinical characteristics, associated congenital anomalies, and laboratory investigations at presentation: total and direct bilirubin, alanine transaminase (ALT), aspartate transaminase (AST), international normalized ratio (INR), gamma-glutamyl transferase (GGT), total serum bile acid levels, imaging findings, histopathological findings, and treatment provided postoperatively. The normal values of GGT were as follows: <1 month old, GGT <200 IU/L; 1–2 months old, GGT <150 IU/L; 2–3 months old, GGT <100 IU/L; and >6 months, GGT <60 IU/L.[7,8] Not all variables were available for each participant. For each variable, the number of subjects (“N”) from whom data were available is indicated.

Ethical considerations

The study was approved by the local institutional review board (log number 16-001) at King Fahad Medical City and the ethics committees in all hospitals.

Statistical analysis

All categorical variables such as gender, gestational age, congenital anomaly, and consanguinity were presented as numbers and percentages. Continuous variables such as age were presented as mean ± standard deviation (SD), while total serum bilirubin (TSB), direct bilirubin, alanine transaminase, aspartate aminotransferase, GGT, alkaline phosphatase, and INR were expressed as the median interquartile range (IQR). Nonparametric tests were used when the data were skewed. The Kolmogorov–Smirnov test was used to check the assumption of normal distribution. The Chi-square or Fisher’s exact test was used according to whether the expected cell frequency was smaller than 5, and was applied to determine the significant association between categorical variables. The Mann–Whitney U-test was applied to determine the significant difference between isolated BA and BA with congenital anomalies with respect to ALT, AST, and GGT. All data were entered and analyzed through Statistical Package for the Social Sciences (SPSS) version 25 (SPSS Inc., Chicago, IL, USA) and MedCalc version 18.11.6 (Acacialaan 22, 8400 Ostend, Belgium).

RESULTS

The clinical and laboratory characteristics

Table 1 summarizes the clinical and laboratory characteristics of the 204 BA patients. Preterm babies constituted 10% of the study cohort. The age at KPE was later in preterm babies than in full-term infants, but the difference did not reach statistical significance (70 days [IQR 35-85] vs. 60 days [IQR 26-75]; P = 0.11). The level of serum GGT was normal at 14%. Congenital anomalies were present in 68 patients (33%), including 22 cases of splenic malformation (10.8%). Isolated BA cases were referred earlier and underwent KPE at a significantly earlier age than BA cases associated with congenital anomalies [Table 2].

Table 1.

Clinical and laboratory features of the study cohort (n=204)

Variable	Number (%)
Gender: male/female	98/106
Birth weight	2.83±0.55
Gestational age (n=157)
- Term	118 (75)
- Preterm	16 (10)
- Post-term	1 (0.5)
- IUGR	22 (14)
Median age when first seen in a tertiary care center (days)	65 (7 – 245)
Consanguinity	54.6
Congenital anomalies (n=204)	68 (33)
- BASM (biliary atresia splenic malformation)	9 (4.4)
- BASM + other congenital anomalies	13 (6.4)
- Congenital anomalies (other than BASM)	46 (22.5)
⁰ Heart	32 (69.56)
⁰ Situs inversus	9 (19.56)
⁰ Vascular	8 (17.39)
⁰ CNS	4 (8.69)
⁰ GIT	6 (13.04)
⁰ Renal	12 (26.08)
⁰ Skeletal	4 (8.69)
TSB µmol/L	189 (142-237.2)
Direct bilirubin µmol/L	139 (104.5-179)
ALT U/L	164 (104-224)
AST U/L	222 (150-365.7)
GGT IU/L (n=189)	472 (189-858.2)
- High GGT	163/189 (86)
- Normal GGT	26/189 (14)
Alkaline phosphatase U/L	644 (470-917)
INR	1.1 (1-1.2)
Albumin gm/L	34 (30-38.3)
Serum bile acids (µmol/L)	129.6 (27.6-197.5)
Serum cholesterol (mmol/L)	4.5 (2.8-7)
Alpha-fetoprotein (µg/L)	999 (10.8-1210)

Table 2.

Comparison of BA cases associated with congenital anomalies versus isolated BA

Variables	Isolated BA (n=131)	BA with congenital anomalies (n=68)	P
Male: female	65:66 (1:1.01)	30:38 (1:1.26)	0.461
Median birth weight (Kg)	3	2.9	0.91
Preterm (%)	19/131 (14.5)	8/68 (9)	0.593
1st- and 2nd-degree consanguinity (%)	52/131 (40)	28/68 (38.2)	0.840
Median age at referral to Ped. GI specialist (days)	60	71	0.021
Median age at Kasai surgery (days)	66	74	0.04
Median total serum bilirubin (µmol/L)	190	189	0.85
Median direct bilirubin (µmol/L)	137.8	139.5	0.67
Median ALT U/L	175	155	0.15
Median AST U/L	246	222	0.21
Median GGT IU/L	418	531	0.19

Data presented as n (%), mean ± standard deviation or median (interquartile range), as appropriate

Imaging findings

Ultrasound of the abdomen was performed in all 204 patients; the findings included hypoplastic or atrophic gall bladder (GB) in 65%, normal GB in 30%, and cord sign in 5%, and the common bile duct was nonvisualized in 53% and visualized in 47%. Hepatobiliary iminodiacetic acid scintigraphy (HIDA) scan was performed in 99 cases (48.52%), all non-excretory. Table 3 compares the BA cases that underwent KPE and had an HIDA scan performed against the cases that did not have an HIDA scan (where data are available). Patients that underwent HIDA scan had a higher median day to KPE as compared to patients who did not (11 vs. 9.5 days, respectively, P = 0.034).

Table 3.

Comparison of patients who underwent HIDA scan versus patients that did not undergo HIDA scan

	HIDA done (n=79)	HIDA not done (n=50)	P
Median (IQR)
Age at referral	49.0 (33.5-80)	60.0 (48.3-74.8)	0.18
Age at KPE	65.0 (49-94.5)	74.5 (60-84.8)	0.53
Days to KPE	11.0 (8-20)	9.5 (5-14)	0.034

Magnetic resonance cholangiopancreatography (MRCP) was performed in 27 cases (13%); GB was hypoplastic, atrophic, or not visualized in all the cases, and the common bile duct was not visualized in 23 cases (85%) and visualized in 4 cases (15%). Preoperative cholangiography was performed in 139 cases (68%): operative in 122 (60%) and percutaneous in 17 (8%).

Liver histopathology

A total of 179 patients had liver biopsy performed (88%) (needle biopsy in 116 cases and wedge biopsy in 100 cases, and both biopsies were obtained in 35 cases). The most common histopathologic findings in both types of biopsies were as follows: bile duct proliferation (92.5%, 92%), canalicular cholestasis (92%, 96%), bile plugs (84%, 83%), portal fibrosis (95%, 92%), and giant cell transformation (57%, 59%). In cases where both needle and wedge biopsies were performed (n = 35 cases), a comparison of the frequency of the histopathologic findings showed no significant differences (P>0.05) between the two types of liver biopsies [Table 4].

Table 4.

Comparison of the liver histopathology findings in needle and wedge biopsies (n=35)

Pathological findings	Needle biopsy (n=35)	Wedge biopsy (n=35)	P
Bile duct proliferation
a) Yes	93.5%	96.2%	0.637
Canalicular cholestasis
a) Yes	100%	96%	0.259
Bile plugs
a) Yes	96.3%	91.3%	0.459
Giant cell transformation
a) Yes	57%	59%	0.897
Portal fibrosis
a) Yes	100%	100%
i) Grade 1–2 fibrosis (no septae)	22%	21%	0.904
ii) Grade 3–4 fibrosis (++ septae)	77%	79%	0.62

Post-KPE management

A total of 143 children (70%) underwent KPE at a median age of 70 days. Clearance of jaundice was achieved in 66 of the 143 (46%) children. After KPE, steroids were used in 37% of the cases; they received intravenous methylprednisolone and/or oral prednisone in different protocols for variable duration and different regimens; the median duration was 30 days (range 7–90 days) post-KPE. All cases that underwent KPE received intravenous antibiotics, and 100 cases (70%) were prescribed prophylactic antibiotics (mostly trimethoprim or sulfamethoxazole) on discharge for variable duration (ranging between 3 and 12 months). Ursodeoxycholic acid was prescribed for 88% of the BA patients post-KPE; two-thirds received ursodeoxycholic acid until the time of LT or last follow-up visit and one-third continued for a duration that ranged between 1 and 12 months.

DISCUSSION

Our initiative to develop nationwide BA database served to collect large amounts of data to characterize the clinical practices in approaching the diagnosis and treatment of BA among pediatric gastroenterologists in Saudi Arabia. The study highlights several important observations. First, the data showed that there was no standard approach to the diagnosis of BA and that postoperative management varied considerably among the 10 participating tertiary care centers. On reviewing the diagnostic tests done for the 204 patients, only liver biochemistry tests and liver ultrasound were uniformly performed in all patients, while the performance of HIDA scan and MRCP was less consistent. Also, there was no uniformity in post-KPE practice; although there was fairly consistent use of ursodeoxycholic acid (88%) and oral prophylactic antibiotics (100%) across the centers, the use of steroids (37%) post-KPE was inconsistent. This non-uniformity in practice has been reported in other countries,[2] indicating that there is little consensus on the perioperative management of BA. In contrast, in Japan, there are standard national postoperative treatment protocols that adopt the use of prophylactic antibiotics, ursodeoxycholic acid, and steroids post-KPE for extended periods of time, as previously described.[9,10] Whether these differences in clinical practices are responsible for the generally improved prognosis of BA in Japan as compared to outcomes elsewhere is not clear and needs large, well-designed, multicenter studies to address the question.

The traditional teaching describes a typical BA case as a full-term, well-looking baby with high GGT cholestasis and pale stool. However, our data and other recent studies showed that infants with BA could be preterm and could manifest normal GGT cholestasis in 12–14% of the cases.[11,12,13,14] The clinical phenotype and biochemical profile of BA in a preterm baby overlap with more common causes of neonatal cholestasis in a neonatal intensive care setting (probably due to multifactorial origin, e.g., prematurity of bile acid metabolism, sepsis, and prolonged total parenteral nutrition), which makes the identification of BA challenging to clinicians, which might result in the delay of the diagnosis of BA and performance of KPE, as observed in our study when KPE was performed at a later date in preterm than full-term infants [median 70 days (IQR 35–85) vs. 60 days (IQR 26–75), respectively]. Therefore, clinicians should exercise vigilance and promptly exclude BA when encountering an infant with cholestasis and pale stool, regardless of the gestational age and level of GGT, as proposed in Figure 1. Nonetheless, commoner causes of normal or low GGT cholestasis (e.g., progressive familial intrahepatic cholestasis types 1 and 2, an inborn error of bile acid synthesis, and panhypopituitarism) should be strongly considered and appropriately investigated.[15]

Figure 1.

An algorithm to diagnose biliary atresia in a neonate with cholestasis

Although late referral of BA cases is a problem all over the world, it is more pronounced in Saudi Arabia. In Saudi Arabia, after birth, a newborn is seen in the well-baby clinic at two months of age for vaccination and checkup; therefore, family physicians miss the opportunity of early detection of prolonged neonatal jaundice within the first month of life. In addition, pediatricians and family physicians may not recognize the importance of investigating prolonged jaundice in neonates greater than 2 weeks of age thinking that it is breast milk jaundice, which is very common at this age. Therefore, efforts are needed to raise awareness of parents, general practitioners, and pediatricians about the necessity of prompt investigations of any case with prolonged jaundice (jaundice persisting for >2 weeks of life) and early referral of cases with cholestasis to a pediatric gastroenterologist.

In our study, we report a higher rate of congenital anomalies (33%), as compared to other national studies, which impacted negatively on the BA outcomes.[16] Several studies showed that mutations in various genes that might cause defects in bile duct development and control cilia development were found in BA patients with major laterality defects (such as situs inversus, polysplenia or asplenia, preduodenal portal vein, and interrupted vena cava).[17,18,19,20,21] The overall rate of consanguinity in our study cohort is high (55%) as compared to all national registries, which might suggest a role for some yet undetermined genetic variants in the higher development of “congenital” or “embryonic” forms of BA in our population.

During the diagnostic approach to BA, all pediatric gastroenterologists in Saudi Arabia requested abdominal ultrasound, many ordered HIDA scan (50%), and few did magnetic resonance imaging (MRI) of the hepatobiliary system (13%), and none performed endoscopic retrograde cholangiopancreatography (ERCP). Abdominal ultrasound is an easy and noninvasive first diagnostic imaging investigation to exclude choledochal cyst and assess for signs suggestive of BA (e.g., hypoplastic or atrophic, cord sign, nonvisualization of the common bile duct, subcapsular blood flow, splenic or vascular abnormalities). However, it is important to remember that none of these findings is diagnostic of BA and the absence of them does not rule out BA.[22] However, GB can be normal in 30% of BA cases. Ultrasound of the liver is operator dependent and detecting the tiny CBD in neonates depends on the experience of the technician and radiologist. Typically, CBD is nonvisualized in BA; nevertheless, CBD has been reported as visualized in 47% of our patients. The fibrotic cord-like structure of the obliterated CBD could be interpreted, by a nonexpert operator, as “visualized CBD.” Therefore, the finding of “visualized CBD” on ultrasound should not make clinicians relax about pursuing efforts to exclude BA in suspected cases.

HIDA scan adds little to the routine evaluation of the cholestatic infant; it has a sensitivity and specificity of 83%–100% and 70%–85%, respectively, and occasionally misses BA or misclassifies other non-BA cases, leading to unnecessary exploratory laparotomy.[23,24,25,26] Concerns about significant time delays to KPE were recently addressed with the reported average elapsed time from presentation to KPE of 22.2 days and 10.2 days when HIDA was performed and HIDA was not performed, respectively (P = 0.084).[27] Similarly, in our cohort, patients with a HIDA scan performed, as opposed to HIDA not performed, had higher median days to KPE of 11.0 and 9.5 days, respectively (P = 0.034). Though the two groups in our cohort did not differ regarding the age at KPE, the delay to KPE in the HIDA-performed group may have led to the loss of the benefit of early presentation, therefore impacting negatively on KPE outcomes.

MRCP is a noninvasive modality for visualizing the biliary system in infants with a reported specificity of 36% and sensitivity of 99%.[28] Although ERCP has proven effective with high positive and negative predictive values for BA (sensitivity 86%–100% and specificity 87%–94%),[29,30] ERCP requires an experienced endoscopist, specific infant endoscopy equipment not readily available at many centers, and general anesthesia. The superiority of ERCP compared with other types of cholangiograms has not been demonstrated.[31]

Liver biopsy continues to be the cornerstone in the diagnostic workup of cases with high suspicion of BA.[22] It has a diagnostic accuracy of 86 to 95% for BA when examined by experienced pathologists.[32,33,34,35,36] The liver in neonates and young infants is vulnerable to insult and usually responds to injury by manifesting a histopathological picture of intracanalicular cholestasis and giant cell transformation, with considerable overlap between entities of different mechanisms, making a histological diagnosis for cholestasis in infants has always been considered a challenge for many pathologists. In our study, we and other groups[32,33,34,35,36] demonstrated that bile duct proliferation, canalicular cholestasis, bile plugs, and portal fibrosis were four major histopathological features of BA. In another study, we found that bile duct plugs and marked ductular proliferation were strong independent histologic predictors of BA.[33] When the above histopathologic features are found in liver histology in association with absent or rare hepatocellular giant cell transformation and no lobular disarray or sinusoidal fibrosis, the diagnosis of BA is most likely and should prompt confirmation by intraoperative operative cholangiography [Figure 1]. Other investigators demonstrated additional histological features that, when found in combination with the above-mentioned histological features, favor BA diagnosis. These include portal stromal edema and absent or rare extramedullary hematopoiesis.[32,36] Whether wedge liver biopsy is more accurate in identifying the histopathologic features suggestive of BA than percutaneous liver biopsy is not clear. In a multicenter study, wedge LB demonstrated histopathologic features consistent with BA in 14 cases, which were falsely thought to be “not consistent with BA” on the corresponding percutaneous LB.[36] The authors suggested that wedge biopsies were more representative, as some of the characteristic features are better noted in larger portal tracts not sampled by needle biopsy. In our study, we did not observe any statistically significant differences between the two types of liver biopsy.

A clear limitation of our study is the retrospective design, the amount of missing information, and lack of consistency in practice among the participating centers.

In conclusion, there is little consensus as to the standard postoperative treatment of BA among pediatric gastroenterologists in Saudi Arabia. There is a need to establish a national BA registry in Saudi Arabia aiming to standardize pre- and postoperative clinical practices and facilitate research to study the effect of these clinical practices on BA outcomes. Normal serum GGT level, normal GB size on ultrasound, and being premature baby should not preclude the diagnostic workup for BA in a baby with cholestasis and pale stool.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.