Neonatal jaundice is routinely screened transcutaneously or by heel-stick capillary sampling, which assesses only total bilirubin 1, while direct bilirubin requires venous sampling. Recent studies have shown that infants with biliary atresia have elevated direct bilirubin early in life, which could be used for screening 2; however, routine implementation of venous sampling is limited. In adults, a positive urine bilirubin has shown 70%–77% sensitivity and 80%–82% specificity for identifying raised serum direct bilirubin 2; however, data in pediatrics are lacking. In children, immature glomerular filtration rates 3 and hepatic transporters (OATP1B1/B3, MRP2) 4 may modify bilirubin excretion, potentially limiting the applicability of adult data to children. Therefore, we aimed to evaluate the diagnostic performance of urine bilirubin for detecting elevated serum direct bilirubin in the paediatric population. This retrospective cross-sectional study analysed the computerised database of Clalit Health Services, Israel's largest health maintenance organization, which provides care to approximately 54% of the national population across diverse geographic and socioeconomic groups. Data from 2000 to 2020 were reviewed for patients aged 0–18 years who had urine bilirubin results available within 1 week of serum direct bilirubin measurement. The median time difference was 0 days (IQR 0–0.012). When multiple serum measurements were available, the value closest in time to the urine test was used. Collected data included sex, age, medical diagnosis and liver enzymes: alanine aminotransferase (ALT), aspartate aminotransferase (AST), gamma-glutamyl transferase (GGT) and alkaline phosphatase (ALP). Urine bilirubin was considered positive at ≥ + 1. Serum direct bilirubin was evaluated using a threshold of ≥ 1.0 mg/dL. Data are presented as numbers (%) for categorical variables and medians (IQR) for continuous variables. Group comparisons were performed using Chi-square and Mann–Whitney U tests, as appropriate. Sensitivity, specificity, PPV and NPV with 95% confidence intervals were calculated for the overall cohort and for infants younger than 12 months. Analyses were conducted using R software, version 4.2.3. A total of 75 964 children contributed 116 725 urine bilirubin tests. The cohort was 51% male with a mean age of 9.3 ± 5.5 years; 1.9% had a diagnosis of liver disease. Children with a positive urine bilirubin had significantly higher serum direct bilirubin and cholestatic enzymes compared with those with negative urine bilirubin (Table 1). Interestingly, of the 34 children with a BA diagnosis, only 7 (21%) had elevated urine bilirubin at the time of first presentation with cholestasis. There was no statistically significant difference between the serum bilirubin of those BA patients with negative urine bilirubin (4.3 ± 2.66 mg/dL) and those with elevated urine bilirubin (5.91 ± 2.15 mg/dL, p = 0.2). Overall, urine bilirubin demonstrated low sensitivity but very high specificity for detecting cholestasis. For a direct bilirubin cutoff of ≥ 1 mg/dL, sensitivity was 31.4% (95% CI: 29.9%, 33.0%), specificity 99.7% (95% CI: 99.6%, 99.7%), PPV 74.2% (95% CI: 71.9%, 76.4%) and NPV 97.9% (95% CI: 97.8%, 98.0%). Performance was similar in infants < 12 months: sensitivity 27.1% (95% CI: 25.3%, 29.0%), specificity 98.4% (95% CI: 98.2%, 98.7%), PPV 81.2% (95% CI: 78.5%, 84.0%), NPV 84.2% (95% CI: 83.6%, 84.9%). (Supplementary Tables S1 and 2). These findings indicate that urine bilirubin is not a suitable screening test for cholestasis. Importantly, 79% of children with biliary atresia in our dataset had negative urine bilirubin results, underscoring the limited utility of urine bilirubin for BA screening. Adult studies have reported moderate sensitivity and specificity for urine bilirubin in detecting elevated serum bilirubin 2. In contrast, our paediatric cohort showed very high specificity but low sensitivity, likely reflecting age-specific physiology. Several approaches to neonatal cholestasis screening exist, including stool colour cards, targeted bile acid assays and venous direct bilirubin sampling, each with its own limitations 5. There are some limitations to the study. Its retrospective design and reliance on routinely collected laboratory data may introduce selection bias toward the clinically tested population. The use of a large database restricts access to detailed clinical indications for testing. In addition, although neonatal cholestasis was a key clinical motivation, most participants were older children. However, a substantial subgroup of infants younger than 12 months (n = 11 775) demonstrated similar diagnostic performance. Despite this, the results may not be generalizable to neonates. In summary, urine bilirubin should not be considered a reliable diagnostic or screening test in paediatric patients. Its main value lies in its high specificity, which warrants prompt serum-based evaluation for a positive result, but its low sensitivity precludes excluding disease when negative. Miri Dotan: conceptualization, methodology, formal analysis, data curation, writing – original draft. During manuscript preparation, ChatGPT was used to improve language clarity. The authors have nothing to report. The study was approved by the local institutional ethics committee (RMC-0046-21) in keeping with the principles of the Declaration of Helsinki. Informed consent was waived since data were collected anonymously from the computerised medical files, with no active participation of patients. The authors declare no conflicts of interest. Clalit Health Services H.M.O. granted permission to use this dataset for the current analysis. Due to ethical restrictions, data for the current study cannot be shared. Table S1: Diagnostic performance of urine bilirubin for the detection of elevated serum direct bilirubin (≥ 1 mg/dL) in the overall paediatric population. Table S2: Diagnostic performance of urine bilirubin for the detection of elevated serum direct bilirubin (≥ 1 mg/dL) in infants younger than 12 months. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.
Dotan et al. (Sun,) studied this question.