P30: Establishing Representative Pediatric Liver Enzymes Reference Ranges Utilizing United States Optum® Real World Data
Sinead Maria Nally
Clinical Development Director
We aimed to establish reference ranges for liver enzymes for ALT, AST, Alkaline Phosphatase, and Total Bilirubin in a representative pediatric population with low risk for subclinical liver disease using Optum® claims database, and compared these ranges to other existing datasets.
We identified a cohort of male/female subjects aged<=18 years in Optum® from 2017-2019. We excluded records with diagnostic codes indicative of conditions that may directly impact liver enzymes based on ICD10 class ±15 days of lab record. We excluded outliers if labs were >5x the Harriet Lane ULN.
We aimed to develop standards for liver enzymes reference ranges in a representative pediatric population with low risk for subclinical liver disease using Optum® claims database to (1) identify upper limit thresholds for ALT, AST, Alkaline Phosphatase (ALP), Total Bilirubin (TB), and (2) compare ranges to other existing datasets (Canadian Laboratory Initiative on Paediatric Reference Intervals (CALIPER), Harriet Lane (HL), LabCorp).
Data in Optum® were evaluable in over 145,000 patients for each test. For neonates, only TBIL was analyzed due to low sample sizes for other tests. Age was imputed using year of birth or, when available in infants, date of the first medical record or admission. Age- and gender-specific 95th percentiles were estimated with the skewness-median-coefficient of variation (LMS) method to determine the upper reference range. Only the patient’s first record was used in the analysis. The upper reference ranges were compared to existing ranges reported in HL, LabCorp, and CALIPER. The 95th percentile curves followed age-related trends that corresponded to known biological changes during childhood. The strongest differences between the various sources were found for ALT in infants, and ALP and TBIL in late adolescence. Upper ranges for ALT were estimated to be higher in Optum® than the other sources in the first 6 months of life and showed a trend as opposed to being static as reported by the other sources. Optum® TBIL demonstrated a sharper peak in the first two weeks of life followed by a steeper decrease toward a steady level throughout infancy. Age-related trends seen in Optum® TBIL were concordant with CALIPER from childhood to adolescence, whereas TBIL ranges were reported to be static by HL and Labcorp. Optum® and Labcorp ALP ranges overlap closely and trend down after 15 years of age, while HL has a static trend and CALIPER showed an increase in estimated ranges for older males.
Hepatic function monitoring is critical for identification of drug induced liver injury (DILI), and routinely conducted in clinical studies. This highlights the importance of utilizing pediatric-appropriate reference ranges for hepatic function monitoring and safety. Despite the widespread use of liver enzymes in pediatrics, thresholds for detecting liver disease in children are unknown and the proper interpretation of liver enzyme assays performed in children is unclear. We aimed to establish liver enzymes reference ranges in representative pediatric population with low risk for subclinical liver disease using Optum® claims database, and compared to other existing datasets (CALIPER, Harriet Lane, LabCorp).
Utilizing real world data, this large claims dataset provides robust age- and gender- specific liver enzyme ranges in a United States pediatric population presumed to be free of confounding comorbidities as subjects with suspected hepatotoxicity were excluded based on reported ICD10 codes. The 95th percentile curves followed age-related trends that corresponded to known biological changes during childhood. The strongest differences between the different ranges were found for ALT in infants, and ALP and TBIL in late adolescence. The differences could be partially due to variation in the estimation methods, and not necessarily differences in the data sources.
This real world data provide context to existing lab ranges and additional assessment of lab abnormalities. The limitations of claims data include potential ascertainment bias in patients with available liver enzyme testing, and only commercially insured patients since Optum® does not include people who are uninsured or covered by government plans. The utility and applicability of this real world data source of standard pediatric liver reference ranges for industry, regulators, and clinicians needs to be further validated.