Diagnostic value of routine biochemical markers for biliary atresia in neonates with cholestasis: a retrospective study

Introduction

Biliary atresia (BA) is a serious hepatic disorder in infancy characterized by progressive inflammation and obliteration of the intrahepatic and extrahepatic bile ducts. The incidence of BA is approximately 1 in 19,000 to 1 in 15,000 live births in Western countries, but it is significantly higher in Asian countries, reaching up to 1 in 5,000 live births (1). Kasai portoenterostomy remains the first-line treatment aimed at restoring bile drainage. However, despite undergoing this procedure, nearly 50% of infants develop progressive hepatic inflammation and fibrosis, eventually requiring liver transplantation within 2 years after surgery. BA continues to be the leading indication for pediatric liver transplantation (2). Evidence suggests that early intervention, particularly within 30 days after birth, is critical for improving long-term native liver survival in BA patients (3,4).

Nonetheless, differentiating BA from other causes of cholestasis during the early stages remains challenging. Cholangiography remains the gold standard for diagnosis (5). Noninvasive diagnostic tools include abdominal ultrasound (6) and a variety of biochemical markers, such as the ratio of direct to total bilirubin (TBIL) (4), elevated gamma-glutamyl transferase (GGT) (7), serum bile acid profiling (8), and alkaline phosphatase (ALP) levels (9). In recent years, emerging serological markers such as matrix metalloproteinase-7 (MMP-7) have shown diagnostic potential (10). However, due to the lack of clearly established cutoff values and limited validation, MMP-7 is not yet recommended for routine use. Given their simplicity, low cost, and accessibility, routine biochemical tests remain the most practical noninvasive approach for diagnosing BA. Nevertheless, few studies have focused specifically on diagnostic indicators in the neonatal period.

This study aims to identify potential biochemical markers for BA in neonates and to establish optimal cutoff values for early differentiation of BA from other causes of cholestasis. We present this article in accordance with the STARD reporting checklist (available at https://tp.amegroups.com/article/view/10.21037/tp-2025-aw-755/rc).

Methods

Patient population

A retrospective study was conducted on neonates with cholestasis who were treated at Beijing Children’s Hospital between 2018 and 2025. The diagnostic criteria for neonatal cholestasis (NC) were as follows: direct or conjugated bilirubin concentration >17.1 µmol/L when the serum TBIL concentration was ≤85.5 µmol/L, or a direct bilirubin (DBIL) concentration accounting for more than 20% of the TBIL concentration when TBIL was >85.5 µmol/L (11). The diagnosis of BA was confirmed by intraoperative cholangiography (IOC). The diagnosis of non-BA cholestasis was confirmed by: (I) IOC showing intact bile ducts; (II) liver biopsy consistent with non-BA pathology (neonatal hepatitis and metabolic liver disease); (III) genetic testing confirming pathogenic variants; or (IV) clinical improvement after conservative treatment (jaundice resolution, or DBIL ≤17.1 µmol/L) with a minimum follow-up duration of 6 months. Infants with non-BA cholestasis were excluded if cholestasis persisted beyond 6 months without a confirmed etiology.

The criteria for inclusion in the study were as follows: (I) age ≤30 days at the time of initial biochemical testing; and (II) biochemical results meeting the diagnostic criteria for NC. The exclusion criteria were as follows: (I) patients who did not receive further diagnosis or treatment at our institution; and (II) patients in whom the etiology of cholestasis could not be clearly identified.

This study was approved by the Medical Ethics Committee of Beijing Children’s Hospital (No. 2019-k-386), and informed consent requirements were waived for this retrospective study. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

Data collection

Biochemical test results from the first outpatient visit were collected, including the following parameters: potassium (K), sodium (Na), chloride (Cl), calcium (Ca), phosphorus (P), magnesium (Mg), osmolality (OSM), total protein (TP), albumin (ALB), prealbumin (PA), globulin (GLO), blood urea nitrogen (BUN), creatinine (Cr), ALP, aspartate aminotransferase (AST), alanine aminotransferase (ALT), TBIL, DBIL, indirect bilirubin (IBIL), total bile acid (TBA), GGT, lactate dehydrogenase (LDH), creatine kinase (CK), total cholesterol (Chol), triglycerides (TG), high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol (LDL-C), and very low-density lipoprotein cholesterol (VLDL-C). At our institution, DBIL refers to conjugated bilirubin measured using the enzymatic method (glycerol-3-phosphate oxidase-peroxidase aminophenazone method). These laboratory indicators were analyzed to assess their diagnostic value in BA from other causes of NC. Among the 226 included neonates, missing data were observed for the following indicators: PA (n=8, 3.5%), HDL-C (n=5, 2.2%), and VLDL-C (n=5, 2.2%). No missing data were found for the key markers (GGT, TBA, and DBIL) or baseline characteristics. Given the low proportion of missing data (<5%), we adopted a complete case analysis approach for statistical testing, which is unlikely to introduce significant bias.

Statistical analysis

All data were analyzed using SPSS version 26.0 (IBM Corp., Armonk, NY, USA). Continuous variables were tested for normality using the Shapiro-Wilk test. Variables following a normal distribution were expressed as mean ± standard deviation (SD) and compared using the independent-samples t-test; variables with non-normal distribution were expressed as median and interquartile range (IQR) and compared using the Mann-Whitney U test. Categorical variables were described as percentages and analyzed using the Chi-squared test or Fisher’s exact test, as appropriate. A two-sided P value of <0.05 was considered statistically significant.

Univariate analysis was first performed to compare the BA group with the non-BA group. Variables with statistically significant differences in the univariate analysis were entered into a multivariate logistic regression model to identify independent predictors for the diagnosis of BA in neonates. The odds ratio (OR) and 95% confidence interval (CI) were calculated to quantify the association strength of each independent predictor. Collinearity among variables in the multivariate model was evaluated using variance inflation factors (VIFs), with VIF <2.0 indicating no significant collinearity. The Hosmer-Lemeshow test was performed to assess the model fit, with a P value >0.05 suggesting a good fit between the model and the observed data.

We used receiver operating characteristic (ROC) curve analysis to determine the optimal cutoff values of continuous variables (based on the Youden index). Then, we calculated the corresponding diagnostic performance indicators. These included the area under the curve (AUC), sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy.

Results

Patient characteristics

A total of 262 neonates with cholestasis were screened at Beijing Children’s Hospital between 2018 and 2025. Among them, 36 neonates were excluded: 17 were lost to follow-up and 19 had an unclear etiology of cholestasis. Finally, 226 eligible neonates were included in the study, of whom 53 were diagnosed with BA and 173 were classified as non-BA cholestasis (Figure 1). The non-BA group consisted of distinct etiologies with clear diagnostic evidence: neonatal hepatitis syndrome (n=92, 53.2%), cytomegalovirus hepatitis (n=32, 18.5%), total parenteral nutrition induced cholestasis (n=16, 9.2%), sepsis-related cholestasis (n=14, 8.1%), metabolic diseases (n=12, 6.9%), and Alagille syndrome (n=7, 4.0%). Baseline characteristics of the two groups are summarized in Table 1. There were no significant differences in gender, age at initial biochemical testing, birth weight, or onset time (age at jaundice appearance) between the BA and non-BA groups (all P>0.05).

Figure 1 Participant flow diagram of the study. A total of 262 neonates (age ≤30 days) with cholestasis were screened at Beijing Children’s Hospital between 2018 and 2025. Among them, 36 neonates were excluded due to loss to follow-up (n=17) or unclear etiology of cholestasis (n=19). Finally, 226 neonates were enrolled in the study, including 53 cases diagnosed with BA and 173 cases classified as non-BA cholestasis. The diagnosis of BA was confirmed by IOC. The non-BA group was diagnosed based on IOC findings (intact bile ducts), liver biopsy results, genetic testing for pathogenic variants, or clinical improvement (e.g., jaundice resolution, DBIL ≤17.1 µmol/L) with a minimum 6-month follow-up. BA, biliary atresia; DBIL, direct bilirubin; IOC, intraoperative cholangiography.

Comparison of laboratory indicators between the BA group and non-BA groups

As shown in Table 1, significant differences were observed in multiple biochemical parameters between the BA and non-BA groups. Neonates with BA had significantly higher levels of K, Ca, P, TP, ALB, and ALB/GLO ratio. Regarding liver function indicators, the BA group exhibited markedly elevated levels of TBIL, DBIL, IBIL, TBA, and GGT (all P<0.05). Notably, GGT levels were more than three times higher in the BA group (667.0±374.8 U/L) compared with the non-BA group (206.2±175.9 U/L). Conversely, the BA group showed significantly lower levels of blood BUN, Cr, OSM, and LDH compared to the non-BA group (all P<0.05). Lipid metabolism indicators such as LDL-C and Chol were also significantly higher in the BA group (P<0.001 and P=0.001, respectively). Other indicators, including Na, Cl, ALT, AST, PA, ALP, TG, HDL-C, and CK, showed no statistically significant differences between the two groups (all P>0.05).

Predictors for the diagnosis of neonatal BA

Based on Table 1, variables with statistically significant differences in the univariate analysis were entered into a multivariate logistic regression model to identify independent predictors for the diagnosis of BA in neonates. The final model showed that GGT, TBA, and DBIL were independent predictors for the diagnosis of BA (Table 2). Among them, GGT was the most powerful predictor, with an OR of 1.007 (95% CI: 1.004–1.009; P<0.001), followed by TBA (OR =1.019; 95% CI: 1.009–1.029; P<0.001) and DBIL (OR =1.010; 95% CI: 1.004–1.016; P=0.001). Collinearity analysis using VIFs showed that all VIF values were <2.0 (GGT: 1.32, TBA: 1.28, DBIL: 1.19), indicating no significant collinearity. The Hosmer-Lemeshow test confirmed good model fit (χ²=10.463, P=0.25).

Diagnostic performance of key markers

ROC curve analysis was performed for the three independent predictors (GGT, TBA, and DBIL) to determine optimal cutoffs for clinical application. The optimal cutoff values were: GGT 274.5 U/L (AUC =0.893; 95% CI: 0.843–0.943), TBA 67.4 µmol/L (AUC =0.857; 95% CI: 0.802–0.912), and DBIL 46.2 µmol/L (AUC =0.803; 95% CI: 0.742–0.864). To further evaluate clinical applicability, the diagnostic performance of individual markers at optimal cutoffs was summarized (Table 3). For GGT >274.5 U/L, the sensitivity was 90.6%, specificity 75.7%, PPV 53.3%, NPV 96.3%, and accuracy 79.2%. For TBA >67.4 µmol/L, the sensitivity was 83.0%, specificity 75.7%, and accuracy 77.4%. For DBIL >46.2 µmol/L, the sensitivity was 86.8%, specificity 68.8%, and accuracy 73.0%.

Discussion

Timely identification of BA remains crucial to improving outcomes in neonates with cholestasis, yet differentiating BA from other causes remains difficult due to overlapping clinical and biochemical features in the early stage. As noninvasive diagnostic tools become increasingly important in neonatal hepatology, there remains a need to define neonate-specific and clinically validated thresholds for commonly used biochemical markers. In this study, we analyzed a large cohort of neonates with cholestasis and identified GGT, TBA, and DBIL as significant independent predictors for the diagnosis of BA. By determining optimal neonate-specific cutoff values for these markers, we provide preliminary practical tools that may assist in improving early diagnostic efficiency and supporting clinical decision-making before invasive examinations or surgical interventions are performed.

GGT is a well-established marker of biliary epithelial injury and is located in the bile ducts near the portal areas of the liver. When bile ducts are damaged, GGT is released into the circulation (12). Its diagnostic value in BA has been widely validated in previous studies (13-15). Reported diagnostic thresholds for GGT in BA have generally ranged from 184 to 368 U/L; however, most of these values were derived from studies involving older infants or mixed-age pediatric populations (16,17). In addition, prior research has demonstrated a significant correlation between GGT levels and age in patients with BA, and the diagnostic thresholds established in older infants may not be directly applicable to neonates (18). In the present neonatal cohort, GGT showed relatively robust diagnostic performance for BA, with a high AUC and acceptable sensitivity and specificity at the identified cutoff value. Notably, the high NPV of GGT suggests that this marker may have potential utility for early exclusion of BA in neonates with cholestasis. However, the magnitude of the observed effects should be interpreted with caution, considering the relatively limited sample size of the BA group and the inherent variability of liver enzyme expression during the neonatal period. Collectively, these findings support the potential clinical utility of GGT as a neonate-specific biomarker, and its reliability and applicability require further verification in multicenter studies.

Bile acids are synthesized by hepatocytes, and disturbances in their synthesis and metabolism often result in liver dysfunction. Serum TBA levels are recognized as sensitive indicators of hepatic injury (19,20). Previous studies have reported significantly elevated TBA levels in patients with BA (21), but their application in the early diagnosis of BA during the neonatal period remains limited. One study developed a diagnostic scoring system incorporating the glycochenodeoxycholic acid/chenodeoxycholic acid ratio, acholic stools, and GGT, which achieved an AUC of 0.87, with 85.3% sensitivity and 81.3% specificity at a score cutoff of 15 (22). In our study, a serum TBA level exceeding 67.4 µmol/L was significantly associated with the diagnosis of BA. We therefore recommend that TBA be considered a valuable adjunctive biomarker in the early screening of BA, particularly in scenarios where imaging findings are inconclusive or cannot be obtained in a timely manner.

Serum bilirubin, particularly DBIL, remains a fundamental parameter in the evaluation of cholestasis. Although previous studies have suggested the use of the DBIL/TBIL ratio as a diagnostic indicator (14), our findings indicate that the absolute value of DBIL provides greater diagnostic discrimination in neonates. Despite group differences in TBIL, IBIL, and related ratios, DBIL was the only bilirubin-related parameter retained as an independent predictor in multivariate analysis. A retrospective cross-sectional study of 145 infants with cholestasis also reported that combining GGT and DBIL yielded strong predictive value for BA, which aligns with our findings (15). In our cohort, DBIL levels were significantly higher in the BA group than in the non-BA group, suggesting that BA may cause more severe cholestasis in the neonatal period. Moreover, studies have shown that early DBIL-based two-step screening programs can reduce the age at Kasai surgery from 56 to 36 days (4). In our study, a DBIL threshold of 46.2 µmol/L was significantly associated with BA, but its diagnostic performance requires further validation in larger multicenter cohorts before it can be recommended for routine clinical use.

Although our study provides valuable insights into the use of biochemical markers for the diagnosis of BA during the neonatal period, several limitations should be noted. First, this was a retrospective, single-center study, which may introduce selection bias. Due to the study design and limited sample size of the BA group, we did not perform bootstrapping or cross-validation to estimate optimism bias. In addition, there was a notable imbalance in sample size between the BA and non-BA groups, with the non-BA group being approximately three times larger. This imbalance largely reflects the real-world epidemiology of BA, which is considerably less prevalent than other causes of NC; however, it may still introduce spectrum bias and influence effect size estimates and CI widths. Second, to focus on early differentiation between BA and other causes of NC, we included only neonates within 30 days of age, which may limit the generalizability of our findings to older infants. Finally, although optimal cutoff values were determined using ROC curve analysis, further validation in large-scale, prospective, multicenter cohorts is necessary to confirm their clinical applicability and correct for potential optimism bias.

Conclusions

In conclusion, our study identifies GGT, TBA, and DBIL as potential significant predictors for the diagnosis of BA in neonates with cholestasis. These noninvasive biochemical markers may provide auxiliary tools for the early differentiation of BA from other cholestasis etiologies in neonates, thereby supporting timely clinical referral and intervention. However, given the limitations of this study, the cutoff values reported herein are preliminary and cannot be directly applied to routine clinical practice. Future prospective multicenter studies with larger sample sizes are needed to validate these findings and establish standardized diagnostic thresholds for neonatal BA.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the STARD reporting checklist. Available at https://tp.amegroups.com/article/view/10.21037/tp-2025-aw-755/rc

Data Sharing Statement: Available at https://tp.amegroups.com/article/view/10.21037/tp-2025-aw-755/dss

Peer Review File: Available at https://tp.amegroups.com/article/view/10.21037/tp-2025-aw-755/prf

Funding: This work was supported by the Beijing Municipal Science & Technology Commission (No. Z211100002921062), the Beijing High Innovation Plan (No. 20250058), the National Natural Science Foundation of China (Nos. 82400592 and 82300574), and the Beijing Municipal Natural Science Foundation (No. 7252043).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tp.amegroups.com/article/view/10.21037/tp-2025-aw-755/coif). All authors report the funding from the National Natural Science Foundation of China grants (Nos. 82400592 and 82300574) and the Beijing Municipal Natural Science Foundation (No. 7252043). The authors have no other conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. This study was approved by the Medical Ethics Committee of Beijing Children’s Hospital (No. 2019-k-386), and informed consent requirements were waived for this retrospective study. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Tam PKH, Wells RG, Tang CSM, et al. Biliary atresia. Nat Rev Dis Primers 2024;10:47. [Crossref] [PubMed]
2. Smith SK, Miloh T. Pediatric Liver Transplantation. Clin Liver Dis 2022;26:521-35. [Crossref] [PubMed]
3. Fanna M, Masson G, Capito C, et al. Management of Biliary Atresia in France 1986 to 2015: Long-term Results. J Pediatr Gastroenterol Nutr 2019;69:416-24. [Crossref] [PubMed]
4. Harpavat S, Garcia-Prats JA, Anaya C, et al. Diagnostic Yield of Newborn Screening for Biliary Atresia Using Direct or Conjugated Bilirubin Measurements. JAMA 2020;323:1141-50. [Crossref] [PubMed]
5. Davenport M. Updates in Biliary Atresia: Aetiology, Diagnosis and Surgery. Children (Basel) 2025;12:95. [Crossref] [PubMed]
6. Chen X, Zhao D, Ji H, et al. Predictive modeling for early detection of biliary atresia in infants with cholestasis: Insights from a machine learning study. Comput Biol Med 2024;174:108439. [Crossref] [PubMed]
7. Gong Z, Lin L, Lu G, et al. Development and validation of a model for early diagnosis of biliary atresia. BMC Pediatr 2023;23:549. [Crossref] [PubMed]
8. Fu X, Xiao Y, Golden J, et al. Serum bile acids profiling by liquid chromatography-tandem mass spectrometry (LC-MS/MS) and its application on pediatric liver and intestinal diseases. Clin Chem Lab Med 2020;58:787-97. [Crossref] [PubMed]
9. Rafeey M, Saboktakin L, Hasani JS, et al. Diagnostic value of anti-smooth muscle antibodies and liver enzymes in differentiation of extrahepatic biliary atresia and idiopathic neonatal hepatitis. Afr J Paediatr Surg 2016;13:63-8. [Crossref] [PubMed]
10. Xie C, Wang P, Wang D, et al. Is performing the Kasai portoenterostomy in the neonatal period associated with a better prognosis? A single-center, retrospective cohort study from China. BMC Pediatr 2025;25:454. [Crossref] [PubMed]
11. De Bruyne R, Van Biervliet S, Vande Velde S, et al. Clinical practice: neonatal cholestasis. Eur J Pediatr 2011;170:279-84. [Crossref] [PubMed]
12. Kong F, Dong R, Chen G, et al. Progress in Biomarkers Related to Biliary Atresia. J Clin Transl Hepatol 2024;12:305-15. [Crossref] [PubMed]
13. Weng Z, Zhou W, Wu Q, et al. Gamma-Glutamyl Transferase Combined With Conventional Ultrasound Features in Diagnosing Biliary Atresia: A Two-Center Retrospective Analysis. J Ultrasound Med 2022;41:2805-17. [Crossref] [PubMed]
14. Toh Q, Chen Y, Lee YY, et al. Simple biliary atresia score-a validated diagnostic aid for infantile cholestasis. Pediatr Surg Int 2024;40:212. [Crossref] [PubMed]
15. Hsiao YH, Hung WL, Yang YJ, et al. Non-invasive simple predictors of biliary atresia in cholestatic infants - A preliminary report. Pediatr Neonatol 2025;66:588-93. [Crossref] [PubMed]
16. Wang G, Zhang N, Zhang X, et al. Ultrasound characteristics combined with gamma-glutamyl transpeptidase for diagnosis of biliary atresia in infants less than 30 days. Pediatr Surg Int 2021;37:1175-82. [Crossref] [PubMed]
17. Zhang Y, Li T, Wang T, et al. Comparison for the diagnostic performance of early diagnostic methods for biliary atresia: a systematic review and network meta-analysis. Pediatr Surg Int 2024;40:146. [Crossref] [PubMed]
18. Chen X, Dong R, Shen Z, et al. Value of Gamma-Glutamyl Transpeptidase for Diagnosis of Biliary Atresia by Correlation With Age. J Pediatr Gastroenterol Nutr 2016;63:370-3. [Crossref] [PubMed]
19. Harpavat S, Hawthorne K, Setchell KDR, et al. Serum bile acids as a prognostic biomarker in biliary atresia following Kasai portoenterostomy. Hepatology 2023;77:862-73. [Crossref] [PubMed]
20. Anderson L, Hukkinen M, Nyholm I, et al. Serum bile acids early after portoenterostomy are predictive for native liver survival and portal hypertension in biliary atresia. J Pediatr Gastroenterol Nutr 2025;80:462-70. [Crossref] [PubMed]
21. Han YJ, Hu SQ, Zhu JH, et al. Accurate prediction of biliary atresia with an integrated model using MMP-7 levels and bile acids. World J Pediatr 2024;20:822-33. [Crossref] [PubMed]
22. Zhao D, Zhou K, Chen Y, et al. Development and validation of bile acid profile-based scoring system for identification of biliary atresia: a prospective study. BMC Pediatr 2020;20:255. [Crossref] [PubMed]