Routine laboratory test records from a paediatric hospital can be used to establish age-specific and sex-specific thyroid reference intervals without the ethical and logistical difficulties of recruiting healthy children, according to a study conducted by researchers in Poland. Clinical interpretation of thyroid hormone concentrations relies heavily on accurate reference boundaries, but thyroid physiology undergoes substantial maturation from infancy through adolescence. Traditional direct sampling methods require recruiting large cohorts of verified healthy donors, creating substantial hurdles for clinical laboratories seeking representative paediatric reference ranges.
To overcome these challenges, investigators evaluated indirect statistical methods using routine diagnostic results gathered between April 2023 and December 2024 at the University Children’s Hospital in Lublin, Poland. The study analysed laboratory measurements from patients aged one day to 18 years, focusing on three core endocrine markers: thyroid-stimulating hormone, free thyroxine, and free triiodothyronine. By deriving intervals retrospectively from electronic health and laboratory information systems, clinical scientists aim to provide low-cost tools to verify thyroid thresholds when direct reference participant recruitment remains impractical.
The study protocol applied strict selection criteria to the routine data pool, extracting exactly one median value per patient for each analyte to prevent individual over-representation. This sampling strategy yielded 13,563 patient observations for thyroid-stimulating hormone, 11,143 patient observations for free thyroxine, and 7,346 patient observations for free triiodothyronine. All hormonal measurements were performed using electrochemiluminescence immunoassays with Elecsys reagents from Roche Diagnostics across cobas e 411 and cobas e 601 analytical platforms.
Separating physiological measurements from values altered by underlying pathology required specialised statistical algorithms. The researchers employed refineR, an indirect estimation software package designed to isolate underlying non-pathological distributions from mixed hospital cohorts using Box-Cox transformations and robust mathematical modelling.
Reference intervals representing the 2.5th, 50th, and 97.5th percentiles were calculated for each analyte, with statistical uncertainty quantified through 20 bootstrap resamples. To determine whether partitions were medically and statistically justified, the team assessed age and sex influences using generalised additive models combined with robust regression analysis. Final decisions to partition patient sub-cohorts were established against Haeckel-Wosniok equivalence limits, confirming that age significantly altered the distribution of all three tested hormones.
For thyroid-stimulating hormone, age-dependent intervals were defined as 1.02 to 7.04 micro international units per millilitre for infants aged one month to one year, decreasing to 0.787 to 5.03 micro international units per millilitre for children aged one to six years. Among older groups, reference intervals shifted to 0.915 to 5.26 micro international units per millilitre between six and 11 years, before declining to 0.670 to 4.45 micro international units per millilitre among adolescents aged 11 to 18 years.
Concentrations of free thyroxine also demonstrated marked age-related shifts across developmental stages. Reference limits for free thyroxine were determined to be 1.06 to 1.63 nanograms per decilitre in infants aged one month to one year, 1.02 to 1.63 nanograms per decilitre for children aged one to 10 years, and 0.924 to 1.59 nanograms per decilitre for participants older than 10 years. Unlike free triiodothyronine, sex differences observed in free thyroxine concentrations were insufficient to meet the equivalence criteria required for separate male and female partitioning.
From one month to 10 years of age, a single combined reference interval of 3.28 to 5.37 picograms per millilitre was established. However, for patients older than 10 years, physiological divergences between sexes became pronounced enough to justify distinct reference intervals, resulting in estimated ranges of 2.94 to 4.98 picograms per millilitre in boys and 2.51 to 4.59 picograms per millilitre in girls.
The study demonstrates that paediatric thyroid testing requires dedicated age-specific partitions across all three biomarkers and sex-specific boundaries for free triiodothyronine in children older than 10 years. While indirect statistical derivation offers an accessible, low-cost methodology when direct clinical recruitment is unfeasible, the investigators caution that derived boundaries must be evaluated using explicit concordance criteria and thoroughly verified within local laboratory environments before clinical implementation.
Subject of Research: Indirect estimation of paediatric thyroid hormone reference intervals
Article Title: Indirect reference intervals for TSH, fT3 and fT4 in a paediatric population: a low-cost verification tool for medical laboratories
Article References: Chojęta, A., Janiszewska, M., Wawer, J., Chojęta, M., Ćwik, K., Beń-Skowronek, I., Wawer, G. A., Natorska-Chomicka, D., Grywalska, E., & Swadźba, J. (2026). Indirect reference intervals for TSH, fT3 and fT4 in a paediatric population: a low-cost verification tool for medical laboratories. Frontiers in Endocrinology, 17, Article 1809439. https://doi.org/10.3389/fendo.2026.1809439
Image Credits: AI Generated
DOI: 10.3389/fendo.2026.1809439
Keywords: paediatric endocrinology, thyroid reference intervals, thyroid-stimulating hormone, free thyroxine, free triiodothyronine, refineR, indirect estimation, laboratory diagnostics, Box-Cox transformation, Haeckel-Wosniok equivalence limits, clinical chemistry, electrochemiluminescence immunoassay, cobas e 411, cobas e 601, paediatric laboratory medicine, bootstrap resampling
Cite Scienmag News
Harold Sullivan. (November 21, 2026). Researchers Derive Indirect Paediatric Thyroid Reference Intervals Using Routine Laboratory Data. Scienmag. https://scienmag.com/researchers-derive-indirect-paediatric-thyroid-reference-intervals-using-routine-laboratory-data/
Harold Sullivan. "Researchers Derive Indirect Paediatric Thyroid Reference Intervals Using Routine Laboratory Data." Scienmag, 21 November 2026, https://scienmag.com/researchers-derive-indirect-paediatric-thyroid-reference-intervals-using-routine-laboratory-data/. Accessed 21 November 2026.
Harold Sullivan. "Researchers Derive Indirect Paediatric Thyroid Reference Intervals Using Routine Laboratory Data." Scienmag. November 21, 2026. https://scienmag.com/researchers-derive-indirect-paediatric-thyroid-reference-intervals-using-routine-laboratory-data/








