Endocrine Abstracts

JOINT2676

Background: Thyroid function tests (TFTs) are among the most frequently performed immunoassays in clinical laboratories. However, the lack of standardized reference materials and methods has led to significant variability in results among different manufacturers, making standardization challenging. Additionally, reference intervals (RIs), which are essential for accurate interpretation, vary across analytical methods. This study aims to establish a common RI applicable across multiple automated immunoassay platforms using percentile transformation.

Methods: We analyzed randomly selected samples covering a wide concentration range (TSH: 283 samples, T3: 265 samples, fT4: 283 samples) from Kyung Hee University Hospital at Gangdong. These samples were tested on the three most widely used automated immunoassay analyzers in South Korea (Abbott Alinity i, Beckman Coulter DxI 800, and Roche cobas e801) to measure TFT values simultaneously. Using percentile transformation, regression analysis was conducted to derive recalibration equations. These equations were applied to TFT measurements from 120 healthy individuals, each tested on all three platforms. Healthy individuals were selected based on normal blood test results from a health screening center, absence of suspected thyroid disease on ultrasound, and negative anti-thyroglobulin antibody (Anti-TG) test and anti-thyroid peroxidase antibody (Anti-TPO) tests. RIs were determined using the nonparametric percentile method (2.5th–97.5th percentiles). The final common RI was derived by integrating the recalibrated results from all three platforms, ensuring a common RI across different methods.

Results: Recalibration effectively minimized method-dependent discrepancies in RI values. The common TSH RI (0.51–4.08 μIU/ml) reduced inter-method variability, narrowing the differences observed across method-specific RIs: Abbott (0.52–4.08), Beckman Coulter (0.49–4.27), and Roche (0.50–3.99). For triiodothyronine (T3), the common RI (0.71–1.41 ng/ml) significantly reduced variations among Abbott (0.84–1.40), Beckman Coulter (0.78–1.47), and Roche (0.65–1.27), demonstrating the most pronounced improvement. Similarly, for free thyroxine (fT4), the common RI (0.84–1.42 ng/dl) narrowed the gap between platform-specific values: Abbott (0.83–1.39), Beckman Coulter (0.83–1.32), and Roche (0.97–1.46). Among the three markers, T3 exhibited the greatest reduction in variability after recalibration, while TSH and fT4 also showed notable improvements in cross-platform consistency.

Conclusions: Harmonization of TFT results using recalibration equations successfully minimized inter-method variability, enabling the establishment of a common RI. This approach improves result interpretation across immunoassay platforms and ultimately enhances clinical decision-making. Future research should validate the RI in independent populations, assess its clinical applicability, and evaluate its stability across different platforms to ensure reliability and facilitate widespread adoption.