# Fully Automated Serum LC-MS/MS Platform and Pediatric Reference Intervals for Organic Acids, Amino Acids, and Acylcarnitines in Children (Ages 0–6 Years): Toward Quantitative Diagnosis of Inborn Errors of Metabolism

## TL;DR

A new automated blood test platform can detect metabolic disorders in children by measuring multiple metabolites at once, improving diagnosis speed and accuracy.

## Contribution

A fully automated serum-based LC–MS/MS platform with pediatric reference intervals for integrated metabolic profiling in children.

## Key findings

- The platform simultaneously quantifies 54 metabolites with high precision and accuracy.
- Pediatric reference intervals enabled effective interpretation of metabolic abnormalities in IEM patients.
- Z-score models successfully distinguished major IEM categories like organic acidemias and fatty acid oxidation disorders.

## Abstract

Background/Objectives: Conventional diagnosis of inborn errors of metabolism (IEMs) requires multiple specimen types—urine organic acids, plasma amino acids, and serum acylcarnitines—analyzed on distinct analytical platforms. This multi-assay approach is labor-intensive and limits timely clinical decision making. We aimed to develop a fully automated serum-based LC–MS/MS platform for integrated quantitative metabolite profiling and to establish pediatric reference intervals (RIs) to support diagnostic interpretation. Methods: A fully automated LC–MS/MS system integrated with the CLAM-2030 automated pretreatment module was developed to enable simultaneous quantification of 25 organic acids, 8 amino acids, and 21 acylcarnitines. Analytical performance was assessed for linearity, limits of detection and quantification, precision and accuracy. Serum samples from 296 non-IEM children aged 0–6 years were analyzed to establish pediatric RIs using Box–Cox transformation and Gaussian modeling. Clinical utility was evaluated in sera from 89 patients diagnosed with IEM using z-score-based logistic regression models. Results: The method demonstrated excellent performance, with linearity (r2 > 0.99) across calibration ranges, limits of detection and quantification defined by signal-to-noise ratios > 3 and >10, and intra- and inter-assay precision < 15% CV for all 54 analytes. Twenty-one analytes met the acceptance criterion of ±20% accuracy at all quality-control levels. Pediatric RIs provided a quantitative framework for interpreting the metabolic abnormalities. In IEM patients, disease-specific metabolites were consistently outside the established ranges, and z-score-based logistic regression models successfully distinguished major IEM categories, including organic acidemias and long-chain fatty acid oxidation disorders. Conclusions: This fully automated, serum-based LC–MS/MS platform provides a clinically practical and quantitative framework for integrated metabolic profiling using pediatric RIs, supporting diagnosis and monitoring of IEMs in pediatric settings.