Development of trimester-specific reference intervals for thyroid hormones based on real-world data using a maximum likelihood method

Abstract

Introduction: We sought to establish trimester-specific reference intervals of thyroid-stimulating hormone (TSH) and free thyroxine (FT4) using real-world data. Methods: Deidentified data for FT4, TSH, and anti–thyroid peroxidase antibody (TPO-Ab) associated with pregnancies from July 1, 2014, to December 31, 2019, were extracted from the institutions’ medical records. After data cleaning, trimester-specific reference intervals were established using the maximum likelihood method in TPO-Ab–negative pregnancies. Results: We included 55 323 records after data cleaning. Reference intervals for TSH and FT4 built using the maximum likelihood method in the first, second, and third trimesters were 0.40 to 4.09, 0.57 to 4.04, 0.73 to 4.07 mIU/L and 12.2 to 20.5, 10.2 to 18.2, and 9.0 to 15.5 pmol/L, respectively. Compared with reference intervals from the Guidelines of the American Thyroid Association for the Diagnosis and Management of Thyroid Disease During Pregnancy and the Postpartum (2nd edition), the maximum likelihood method–calculated reference intervals of FT4 were comparable with guideline-suggested reference intervals, with the first trimester reference intervals slightly lower and the second trimester reference intervals showing narrower limits. For TSH, the maximum likelihood method–calculated reference intervals were narrower than guideline-suggested reference intervals. Discussion: Trimester-specific reference intervals of TSH and FT4 for pregnancies were established using the maximum likelihood method. Compared with guideline-suggested reference intervals, no clinically significant discrepancies in FT4 and narrower limits in TSH were observed.