Defining age partitions for thyrotropin (TSH) reference intervals in infancy: a robust statistical approach

Area of Science:

• Pediatric Endocrinology
• Clinical Chemistry
• Neonatal Medicine

Background:

• Thyroid hormones are vital for infant growth and neurodevelopment.
• Accurate interpretation of thyroid-stimulating hormone (TSH) requires age-specific reference intervals (RIs), which are currently ill-defined for the first year of life.
• Diagnostic uncertainty and unnecessary testing arise from the lack of established pediatric TSH RIs.

Purpose of the Study:

• To establish statistically derived age partitions for TSH RIs reflecting thyroid physiology in infants up to one year old.
• To improve diagnostic accuracy for thyroid disorders in the pediatric population.
• To provide a reproducible method for defining age-dependent analyte partitions.

Main Methods:

• Retrospective analysis of de-identified TSH results from 11,045 infants at a tertiary hospital.
• Exclusion of data from infants with repeat tests, extreme TSH values, or conditions affecting thyroid function.
• Sequential application of the Mann-Whitney U test to identify age partitions, followed by validation using Lahti et al. criteria.

Main Results:

• Five age partitions were identified: days 0-1, 2, 3, 4-89, and 90-365.
• TSH levels peaked on days 0-1 (RI: 1.9-40.1 mIU/L), decreased sharply by day 2, and reached near-adult levels by days 90-365 (RI: 0.7-5.7 mIU/L).
• No significant sex-based differences in TSH levels were observed.

Conclusions:

• A robust, data-driven framework for age-specific TSH RIs in infants aged 0-365 days has been established.
• The defined TSH RIs enhance diagnostic accuracy for infant thyroid disorders.
• The methodology offers a reproducible approach for defining age-specific reference intervals for other analytes.