The Pitfalls of Using Research Grade Immunoassays for Thyroid Health Measurements

• Published in: Journal of the Endocrine Society Vol. 5; no. Supplement_1; p. A834
• Main Authors: Boggs, Ashley S P, Galligan, Thomas M, Kemp, Jennifer V, Ravinder, Singh J, Grebe, Stefan Karl Gunther
• Format: Journal Article
• Language: English
• Published: 03.05.2021
• ISSN: 2472-1972; 2472-1972
• DOI: 10.1210/jendso/bvab048.1701

Abstract Introduction: Immunoassay technology is subject to matrix interferences that can produce inaccurate results and incorrect conclusions when using samples not previously validated. While many commercially available research grade (RG) immunoassay kits are available, caution should be applied when using RG kits for thyroid health assessments, particularly on samples from pregnant individuals whose blood chemistry is unique to non-pregnant individuals. Question: Do RG immunoassay kits reliably provide precise and accurate measurements of thyroid health biomarkers in serum Standard Reference Materials (SRMs) from pregnant and non-pregnant donors? Methods: T4, T3, rT3, Tg and TSH measurements were conducted on SRMs 971, Hormones in Frozen Human Serum, and SRM 1949, Frozen Prenatal Serum, using RG immunoassay kits. When available, performance was assessed against validated FDA approved immunoassays or mass spectrometric (MS) methods. Results: RG kits were variable, inaccurate, or imprecise for four of the six biomarkers assessed. RG kit total thyroid hormone measurements overall performed comparably to MS methods, except rT3 measurements, which were twofold greater than mass spectrometric measurements (971M RG mean = 0.51 ng/mL, MS mean = 0.20 ng/mL, p < 0.0001; 971F RG = 0.48 ng/mL, MS = 0.18 ng/mL, p < 0.0001) and had CVs over 30 %. RG kit Tg measurements varied sometimes by as much as tenfold (971M means of 6.50 ng/mL up to 63.3 ng/mL, p < 0.0001; 971F means of 0.350 ng/mL up to 14.5 mg/mL, p < 0.0001). TSH values differed by RG kit manufacturer (971M means of 1.27 µIU/mL up to 1.82 µIU/mL, p < 0.0001; 971F means of 1.36 µIU/mL up to 2.27 µIU/mL, p < 0.0001) and by dilution scheme using the same manufacturer with one case indicating a diagnosis of hypothyroid versus normal TSH levels (1949 non-pregnant undiluted mean = 2325 pg/mL, half dilution mean = 1631 pg/mL, p < 0.0001). Conclusions: RG immunoassays are often used for research projects because they do not require expensive equipment and are simple to conducted. However, we demonstrate here that not all kits are accurate for all patient samples. By utilizing a matrix matched SRM with well-defined quantities of thyroid health biomarkers, one can assess method accuracy, making measurements from different methods comparable. Thereby, data can be harmonized to contribute reliable data on thyroid biomarkers to advance the field of thyroid health.