Owing to the expanded knowledge of the molecular mechanisms of cholestasis liver diseases, remarkable progress has been made in the differential diagnosis and management of hereditary cholestatic diseases since the late 1990s (1–7); however, the diagnosis of cholestatic jaundice in infancy continues to be a difficult challenge for pediatricians because of a wide variety of etiologies and overlapping clinical presentation. Genetic diagnosis and specific metabolic tests are costly, time-consuming, and only available in a few academic centers, and they usually take weeks to months. These have limited its use in the early diagnosis and treatment of infantile cholestasis.
To promptly and properly manage clinical patients, efforts are needed to improve the accuracy and efficacy of early diagnosis for infantile cholestatic patients. The disease patterns and frequencies may vary a lot in patients from different ethnic or geographic backgrounds. We have reported the clinical and genetic diagnosis of infantile cholestatic patients from Taiwan, including ATP8B1 (familial intrahepatic cholestasis [FIC1]), ABCB11 (bile salt export pump [BSEP]), ABCB4 (MDR3), and SLC25A13 (citrin) defects in specific groups of patients (8–11). The overall scope and relative frequencies of infantile cholestasis in east Asia, however, remain unclear. This information has a great impact on the differential diagnosis process, the diagnostic algorithm used, and the choice of laboratory tests in a clinical setting.
In the diagnosis of hereditary cholestasis, γ-glutamyl transpeptidase (GGT) has been an important marker for predicting poor prognoses and an important clinical marker for progressive familial intrahepatic cholestasis (PFIC)-1, PFIC-2, as well as inborn errors of bile acid metabolism (12–14). The cutoff values of GGT levels and its clinical application have, however, not been well defined. The major puzzles that clinicians usually face are the following: what are the leading causes (or top lists of differential diagnosis) of inherited intrahepatic cholestasis in our population, which patient should receive genetic diagnosis, and what type of diagnostic approach and candidate gene analysis should they start with?
The first aim of our study was to analyze the etiology and relative frequencies of infantile cholestatic disorders in Taiwan, after the molecular diagnosis of hereditary cholestatic disease was available. Our second aim was to clarify the role and the cutoff level of GGT as an initial diagnostic marker and a predictive factor for prognosis. Based on the rationale that patients with high probability of poor prognosis would need more sophisticated diagnostic modalities such as specific genetic or metabolic analysis, we aim to propose a more efficient and effective management algorithm for clinical applications.
Children of age <1 year who were admitted to our tertiary pediatric center between January 2000 and October 2012 for investigation of cholestatic jaundice were included retrospectively. All of the patients had direct bilirubin levels exceeding 2 mg/dL and >20% of the total bilirubin. We excluded cases of cholestatic liver diseases resulting from preexisting medical problems, such as prematurity, long-term parental alimentation, congenital heart diseases, multiple congenital abnormality, or malignancy. To analyze patient outcomes, we also excluded those patients who were lost to follow-up before age 1 year. The medical records were reviewed under approval from the institutional review board.
A modified 3-day protocol was routinely used to initially exclude extrahepatic cholestasis and common infectious or metabolic causes of intrahepatic cholestasis (15) (Table 1). As for extrahepatic causes, we diagnosed biliary atresia, choledochal cyst, and inspissated bile syndrome through a series of investigations. These included ultrasonography, magnetic resonance cholangiography, liver histology, and operative findings.
After exclusion of extrahepatic causes, advanced investigations for intrahepatic cholestasis were performed based on the presenting symptoms, signs, and biochemical data. The investigations included surveys for infectious pathogens, metabolic diseases (tandem mass, serum bile acid, lactic acid, ammonia, blood gas, and plasma amino acids), genetic analysis for targeted genes, liver histology, and immunohistochemical stains (for BSEP), as well as urinary bile acid analysis. Urine isolation positive for cytomegalovirus did not exclude the possibility of other causes for cholestasis. The complete diagnostic investigations were individualized on a clinical basis.
GGT as Biochemical Marker
GGT level was measured by an autoanalyzer (Beckman Au5800, Brea, CA) in every patient at the time of admission (Table 1). GGT level has been considered an important clinical marker for PFIC and inborn errors of bile acid metabolism defects. The initial levels of GGT may largely affect the diagnostic strategies in patients with intrahepatic cholestasis. Therefore, initial GGT levels were used in this study as a key biochemical marker. Serum GGT levels below the normal value for infancy (≤100 U/L) were classified as low-GGT group and those with levels >100 U/L were classified as high-GGT group in further analysis (13).
Follow-Up and Prognosis
All of the patients included in the study were studied until total clinical and biochemical recovery, death, or persistent/progressive disease at 1 year of age. A good prognosis was defined as clinical and biochemical recovery of the diseases before 1 year of age. If the patient had a liver transplant, died before 1 year of age, or had persistent disease beyond 1 year of age, the patient was defined as having a poor prognosis.
The diagnosis of all of the patients was reviewed after collecting all of the available clinical/laboratory data from the time of admission and until the last follow-up. In this study, the final diagnosis was verified based on a uniform set of criteria, which are described below.
The diagnosis of Alagille syndrome was based on the clinical diagnostic criteria established by Alagille et al (16) and on liver histopathology. The genetic diagnosis of the Jag1 mutation was performed in 6 patients after parental consent.
The phenotype diagnosis of PFIC was based on clinical, serologic, bile acid level, histological features (8–10,17), and a history of infantile onset of cholestasis persisting or progressing when studied up to 1 year of age. We divided these patients into 2 groups according to their serum GGT levels: those having levels that were ≤100 U/L were classified as low-GGT PFIC; and those with levels of >100 U/L were classified as high-GGT PFIC. Genetic analysis of the ATP8B1 or ABCB11 mutations were performed in 15 patients who were suspected of having low-GGT PFIC as previously reported (8,10).
The diagnosis for neonatal intrahepatic cholestasis caused by citrin deficiency (NICCD) was based on clinical suspicions, supported by laboratory evidence that included aminoacidemia profiles, mutational analysis of the gene SLC25A13, and/or liver histology (6,11,18). The inborn error of bile acid synthesis was suspected if the total serum bile acid was normal or low (19). Urine bile acid analysis was performed by mass spectrometry analysis as previously reported (20–22).
The diagnosis of neonatal hepatitis was based on clinical and biochemical features: cholestatic liver disease onset within 3 months of age; spontaneously recovered within 1 year of age; had negative metabolic, genetic screening, or other organ abnormalities. A histopathological diagnosis of hepatocellular injury, primarily giant cell hepatitis while excluding other metabolic or storage diseases, was noted in this group of patients. The category includes patients with positive viral isolations, with detectable infectious pathogens, as well as those with unidentified causes.
The diagnosis of other metabolic/developmental/endocrine abnormalities was based on associated extrahepatic organ involvements, disease-specific laboratory assays, and the confirmation of diagnosis by a pediatric geneticist, neurologist, or endocrinologist.
For between-group comparisons, the χ2 test was performed using SAS version 9.3 (SPSS Inc, Chicago, IL). A P value <0.05 was considered statistically significant.
Diagnosis and Disease Frequencies
In a total of 256 cholestatic cases, 143 boys and 113 girls, the percentage of patients diagnosed with extrahepatic cholestasis was 47.3%, and biliary atresia comprised 76% of all of the cases of extrahepatic cholestasis. Among the 135 patients with intrahepatic cholestasis, neonatal hepatitis (49.6%) was the most common cause, followed by phenotypic PFIC, including the low- and high-GGT type (21.5%), NICCD (10.4%), and Alagille syndrome (7.4%) (Table 2). A wide variety of diagnoses were noted. In the intrahepatic cholestasis group, 86 patients (63.7%) were defined as having a good prognosis, including 67 patients with neonatal hepatitis, 13 with NICCD, 2 with Alagille syndrome, 2 with hypopituitarism, and the other 2 were idiopathic.
Genetic Diagnosis of PFIC
Fifteen of the 19 patients who had a phenotypic diagnosis of low-GGT intrahepatic cholestasis underwent genetic diagnosis for the ATP8B1 or ABCB11 mutation analysis after parental consent, as previously described (8,10). Disease-causing mutations were confirmed in 7 among 15 patients who underwent genetic analysis. In the 8 patients without identifiable mutations, immunohistochemical staining for BSEP was performed; all of the patients exhibited a positive canalicular BSEP stain in liver sections (Fig. 1).
GGT Levels in Specific Diseases
The initial GGT levels in specific diseases are shown in Figure 2. In addition to PFIC and inborn errors of bile acid synthesis, initial GGT levels <100 U/L were noted in many cases with neonatal hepatitis and in 2 cases of NICCD. Initial GGT levels <100 U/L were also noted in 3 of the 4 cases of mitochondrial liver diseases (16, 39, and 53 U/L).
GGT Levels in Relation to Prognosis
We further analyzed the distribution of initial GGT levels in relation to prognosis as defined above. In the intrahepatic cholestasis group, 86 patients (63.7%) were defined as having a good prognosis. The distribution of their initial GGT levels was normal distribution, with the highest frequency of patients (40%) having levels of approximately 101 to 200 U/L (Fig. 3A). Forty-nine patients (36.3%) were categorized into the poor prognosis group (Fig. 3B). We found that patients with GGT levels in the low or high extremes tended to have a poor prognosis. Overall, patients with initial GGT levels between 75 and 300 U/L tended to have a good prognosis (61/74, 82.4%), whereas patients with initial GGT levels <75 U/L or >300 U/L have a much lower rate of having a good prognosis (25/61, 41%, P < 0.0001).
In the low-GGT group (initial GGT ≤ 100 U/L), 47.4% (27/57) of the patients had poor prognosis, that is, 52.6% (30/57) of the patients recovered by 1 year of age. Patients with lower levels of initial GGT had a higher probability of a poor prognosis. In the present study, we aimed to select an appropriate cutoff level for defining low-GGT levels as a predictor for poor prognosis. For more convenient clinical application, we set several different cutoff levels of GGT and test the predictive value of poor prognosis. In patients with initial GGT levels below each cutoff value, the probability of poor prognosis was calculated (Table 3). When the cutoff value for GGT was decreased, the sensitivity of screening poor prognosis patients decreased but that the specificity and positive predictive value for poor prognosis increased. For a cutoff level set at ≤75 U/L, the sensitivity, specificity, positive predictive value, and negative predictive value were 100%, 43.3%, 61.4%, and 100%, respectively.
Similarly, we compared the sensitivity, specificity, and positive predictive values of different cutoff GGT levels in predicting poor prognosis in the high-GGT group. In patients with initial GGT levels above that certain level, the probability of having a poor prognosis was calculated. We found that as the GGT cutoff level increased, the positive predictive value of poor prognosis increased (Table 3).
In an era of expanded knowledge on genetic cholestasis, the diagnosis of infantile cholestasis has also dramatically improved. This is the first study to show the various etiologies and relative frequencies of infantile cholestasis in east Asian patients. The most common causes of intrahepatic cholestasis in Taiwan were neonatal hepatitis, PFIC, and NICCD. The prevalence in this study was different from previous studies, which were mainly conducted in Western countries (23–25). For example, α-1-antitrypsin deficiency, cystic fibrosis, and tyrosinemia, although searched for, were not found in this study. The diagnosis of neonatal hepatitis was based on clinical and biochemical parameters. This category may overlap with the diagnosis of transient neonatal cholestasis in some other studies (26).
In this study, the incidence of biliary atresia (36% among all cases of infantile cholestasis) was high. In addition to the higher incidence of biliary atresia reported in Taiwan (27), some patients in this study were referred to our hospital for a liver transplant, possibly overestimating the frequency of biliary atresia.
It is well known that patients with PFIC-1 and PFIC-2 (FIC1 and BSEP-deficient patients) present with low or normal GGT levels. Clinically, it is therefore assumed that patients presenting with low or normal GGT levels in infantile cholestasis should be suspected of having PFIC and possibly a poor prognosis. Our study demonstrated that although PFIC patients have low/normal GGT levels, the reverse interpretation, that is, the prediction of poor prognosis by low/normal GGT levels, is not necessarily true. Although GGT levels have been an important clinical marker for the diagnosis of PFIC-1 and PFIC-2, the cutoff value and its application in selecting patients for genetic analysis are still not well established. The normal range of the GGT levels was age dependent. In most of the previous studies, the upper normal limit of GGT levels was defined to be approximately 100 U/L, based on the normal infant values and on studies of patients with idiopathic neonatal hepatitis (28); however, we found that 52.6% of the patients with initial GGT ≤100 U/L recovered by 1 year of age. Thus, the definition of low GGT ≤100 U/L is not a good indicator for predicting PFIC or poor prognosis in paitents with infantile cholestasis. We found that with the lower level of initial GGT, there is a higher chance the patient may have a poor prognosis. Those patients with GGT levels lower than 75, 50, and 25 U/L had a 61.4%, 82.6%, and 100% chance of having a poor prognosis, respectively.
Aside from PFIC, initial low/normal GGT levels may also have been caused by neonatal hepatitis, NICCD, inborn errors of bile acid synthesis, or mitochondrial disorders. To find a possible cutoff value for GGT levels that would prompt a genetic diagnosis investigation, we have identified that GGT levels <75 U/L may be more suitable.
In this regard, we suggest an algorithm using different GGT levels to select patients for further sophisticated diagnostic approaches or referral (Fig. 4). For patients with GGT levels in the range of 75 to 300 U/L, there was a high possibility (82%) that the patients may have a good prognosis. Liver biopsy can be considered at initial evaluation or any time when indicated. Alternatively, close follow-up for 3 months may be appropriate if there have been no signs of liver cirrhosis, hepatic failure, progressive disease, positive family history, or specific clues for inherited diseases.
For patients with low GGT levels that are <75 U/L, inborn errors of bile acid metabolic defects should be surveyed, especially when low or normal serum bile acid level is detected. In patients with elevated serum bile acid levels, a genetic analysis of ATP8B1 (FIC1, for PFIC-1) or ABCB11 (BSEP, for PFIC-2), may be initiated. The lower the GGT levels (eg, ≤50 U/L or ≤25 U/L), the higher is the probability the patient may have a poor prognosis. Consequently, a more rapid investigation and treatment, such as biliary diversion surgery, medical, and nutritional treatments, or even evaluation for liver transplantation, should be considered (20).
For patients with high GGT levels ≥300 U/L, the likelihood of a poor prognosis is highly increased. An overall investigation to rule out Alagille syndrome, PFIC-3, or genetic/metabolic studies for inherited disorders based on ethnic/geographic background should be arranged. An aggressive management of the disease may be planned.
Advanced diagnostic methods such as genetic and metabolic analyses, liver immunohistochemical staining, electron microscopy, and bile analysis were only available in a few medical centers, and the molecular diagnosis was costly and time-consuming. For patients who need advanced diagnostic procedures, preparation for liver transplantation, or genetic consultation, early referral to a tertiary center for advanced survey and management is suggested.
The limitation of this study is that it was not a prospective study. The disease investigation was based on the availability of diagnostic modalities in our own and collaborator's centers. Owing to the highly diverse disease etiologies and the cost and capacity in diagnosing these rare diseases, there has been a gap between the findings in research and the patient diagnosis applied in daily clinical practice. Our results reflect the application of recently developed knowledge and the advanced diagnostic procedures to bedside daily practice in real-world clinical settings.
The diagnosis of PFIC in our study was primarily a clinical phenotypic diagnosis. Molecular and immunohistochemical investigations revealed that in the low-GGT–level group, only a portion of the patients had either ATP8B1 or ABCB11 mutations (10). The lower yield of confirmed cases may suggest that there are PFIC types with unknown genetic basis. Alternatively, this lower yield could have occurred because the investigations of the patients were not extensive enough because this study was retrospective and was on a clinical instead of a research basis. A novel genetic disorder, mutations in tight junction protein 2 gene (TJP2), was found to cause low-GGT PFIC (29). It is expected that in the future there will be easier, cheaper, and more efficient diagnostic methods for PFIC to be developed, as well as further elucidation of novel etiologies of infantile cholestasis.
Highly diverse etiologies of infantile cholestasis are noted in Taiwanese patients. Patients with low/normal GGT levels of ≤100 U/L were not necessarily cases of PFIC. Initial GGT levels that are <75 U/L or >300 U/L may be a quick index for predicting poor prognosis and should prompt further molecular/metabolic investigation. Further research is required for increasing the clinically available diagnostic efficiencies of rare genetic diseases and for investigating novel etiologies of infantile cholestasis.
1. Bull LN, van Eijk MJ, Pawlikowska L, et al. A gene encoding a P-type ATPase mutated in two forms of hereditary cholestasis. Nat Genet
2. Bull LN, Carlton VE, Stricker NL, et al. Genetic and morphological findings in progressive familial intrahepatic cholestasis
(Byler disease [PFIC-1] and Byler syndrome): evidence for heterogeneity. Hepatology
3. Strautnieks SS, Bull LN, Knisely AS, et al. A gene encoding a liver-specific ABC transporter is mutated in progressive familial intrahepatic cholestasis
. Nat Genet
4. Deleuze JF, Jacquemin E, Dubuisson C, et al. Defect of multidrug-resistance 3 gene expression in a subtype of progressive familial intrahepatic cholestasis
5. Alagille D, Estrada A, Hadchouel M, et al. Syndromic paucity of interlobular bile ducts (Alagille syndrome or arteriohepatic dysplasia): review of 80 cases. J Pediatr
6. Saheki T, Kobayashi K. Mitochondrial aspartate glutamate carrier (citrin) deficiency as the cause of adult-onset type II citrullinemia (CTLN2) and idiopathic neonatal hepatitis (NICCD). J Hum Genet
7. Balistreri WF. Inborn errors of bile acid biosynthesis and transport—novel forms of metabolic liver disease. Gastroenterol Clin North Am
8. Chen HL, Chang PS, Hsu HC, et al. FIC1 and BSEP defects in Taiwanese patients with chronic intrahepatic cholestasis with low gamma-glutamyltranspeptidase levels. J Pediatr
9. Chen HL, Chang PS, Hsu HC, et al. Progressive familial intrahepatic cholestasis
with high gamma-glutamyltranspeptidase levels in Taiwanese infants: role of MDR3 gene defect? Pediatr Res
10. Chen HL, Liu YJ, Su YN, et al. Diagnosis of BSEP/ABCB11 mutations in Asian patients with cholestasis using denaturing high performance liquid chromatography. J Pediatr
11. Chen ST, Su YN, Ni YH, et al. Diagnosis of neonatal intrahepatic cholestasis caused by citrin deficiency using high-resolution melting analysis and a clinical scoring system. J Pediatr
12. Wang JS, Tan N, Dhawan A. Significance of low or normal serum gamma glutamyl transferase level in infants with idiopathic neonatal hepatitis. Eur J Pediatr
13. Maggiore G, Bernard O, Riely CA, et al. Normal serum γ-glutamyl-transpeptidase activity identifies groups of infants with idiopathic cholestasis with poor prognosis. J Pediatr
14. Davit-Spraul A, Gonzales E, Baussan C, et al. Progressive familial intrahepatic cholestasis
. Orphanet J Rare Dis
15. Chang MH. Neonatal hepatitis: a follow-up study. J Pediatr Gastroenterol Nutr
16. Alagille D, Odièvre M, Gautier M, et al. Hepatic ductular hypoplasia associated with characteristic facies, vertebral malformations, retarded physical, mental, and sexual development, and cardiac murmur. J Pediatr
17. Pawlikowska L, Strautnieks S, Jankowska I, et al. Differences in presentation and progression between severe FIC1 and BSEP deficiencies. J Hepatol
18. Yeh JN, Jeng YM, Chen HL, et al. Hepaticsteatosis and neonatal intrahepatic cholestasis caused by citrin deficiency (NICCD) in Taiwanese Infants. J Pediatr
19. Heubi JE, Setchell KD, Bove KE. Inborn errors of bile acid metabolism. Semin Liver Dis
20. Balistreri WF, Bezerra JA. Whatever happened to “neonatal hepatitis”? Clin Liver Dis
21. Ueki I, Kimura A, Nishiyori A, et al. Neonatal cholestatic liver disease in an Asian patient with a homozygous mutation in the oxysterol 7alpha-hydroxylase gene. J Pediatr Gastroenterol Nutr
22. Ueki I, Kimura A, Chen HL, et al. SRD5B1 gene analysis needed for the accurate diagnosis of primary 3-oxo-Delta4-steroid 5beta-reductase deficiency. J Gastroenterol Hepatol
23. Moyer V, Freese DK, Whitington PF, et al. Guideline for the evaluation of cholestatic jaundice
in infants: recommendations of the North American Society for Pediatric Gastroenterology, Hepatology and Nutrition. J Pediatr Gastroenterol Nutr
24. Liu C, Aronow BJ, Jegga AG, et al. Novel resequencing chip customized to diagnose mutations in patients with inherited syndromes of intrahepatic cholestasis. Gastroenterology
25. Matte U, Mourya R, Miethke A, et al. Analysis of gene mutations in children with cholestasis of undefined etiology. J Pediatr Gastroenterol Nutr
26. Jacquemin E, Lykavieris P, Chaoui N, et al. Transient neonatal cholestasis: origin and outcome. J Pediatr
27. Hsiao CH, Chang MH, Chen HL, et al. Universal screening for biliary atresia using an infant stool color card in Taiwan. Hepatology
28. Cabrera-Abreu JC, Green A. Gamma-glutamyltransferase: value of its measurement in paediatrics. Ann Clin Biochem
29. Sambrotta M, Strautnieks S, Papouli E, et al. Mutations in TJP2 cause progressive cholestatic liver disease. Nat Genet