Transference Validation of Clinically Utilized Immunoassay Reference Intervals: Evidence From Thyroid...

Area of Science:

• Clinical Chemistry
• Laboratory Medicine
• Assay Development

Background:

• Interplatform variability in thyroid immunoassays hinders standardization, particularly for assays lacking metrologic traceability.
• Direct method-derived reference intervals (RIs) are the gold standard but are resource-intensive.
• RI transference offers a cost-effective solution for establishing multiplatform RIs but requires careful consideration of statistical relevance, precision, and clinical feasibility.

Purpose of the Study:

• To validate a reference interval (RI) transference framework for harmonizing thyroid function testing RIs across different platforms.
• To address challenges in assays with inconsistent reference standards using thyroid function tests as a model.

Main Methods:

• A dual-platform framework (Roche Cobas e601 vs. Beckman DxI 800) was employed using community midrange specimens and clinical leftovers.
• Weighted Deming regression was applied for assays with constant coefficient of variation (CV), and standard Deming regression for assays with constant standard deviation.
• Precision was assessed via multiday temporal stability and Monte Carlo sample size optimization, benchmarking against direct RIs using Spearman ρ, CV(b), and bias ratio (|BR|).

Main Results:

• Weighted Deming regression enabled accurate thyroid-stimulating hormone (TSH) transfer and acceptable precision for free thyroxine (FT4), total triiodothyronine (TT3), and total thyroxine (TT4).
• Randomized multiday sampling significantly reduced extreme bias ratio outliers.
• Community-derived specimens showed better alignment with direct RIs compared to pathology-enriched samples, although FT3 presented persistent challenges due to methodologic heterogeneity.

Conclusions:

• A comprehensive thyroid testing RI transference framework was established, integrating multiday sampling, adaptive sample size, and concentration-driven data selection.
• The framework provides optimal solutions for cross-platform RI establishment in immunoassays.
• Resource-limited laboratories can utilize this transference method to leverage multicenter direct RIs, balancing cost-effectiveness with clinical validity.