To establish and verify the biological reference intervals for third generation thyroid stimulating hormone (TSH) assay using the Beckman Coulter Unicel DxI800 immunoassay platform, and assesse the effects of age, seasonal climate, and temperature variation on TSH.

Referring to the Clinical and Laboratory Standards Institute (CLSI) EP 28-A3c guidelines, 47 502 subjects with test results of TSH were initially obtained from the database of Computer Center of Shenzhen Traditional Chinese Medicine Hospital using big data technology from April 1, 2020 to March 31, 2021. The subjects had an average age of (42±15) years old. In total, 1834 cases with missing values, 1568 with outliers, and 1960 with positive thyroid peroxidase antibody and/or thyroglobulin antibody were excluded. Finally, 42 140 cases meeting the inclusion criteria were considered as the overall population (group A). Group A was divided into male, female, adult and non-adult groups. After data conversion again, outliers were removed according to the above methods. Finally, patients who fully met the inclusion criteria, 14 566 were male (group B) and 27 394 were female (group C); 41 575 were adults (group D) and 571 were juvenile (group E). Furthermore, the individuals aged 31-96 years old in the overall population were divided by age into six groups to analyze the correlation between TSH and age. In addition, the data of the overall population were divided into by season to evaluate the difference of TSH values in different seasons. The biological reference intervals for TSH were determined statistically using SPSS 22.0 and Minitab 18.0 statistical software. One-way ANOVA was used for comparison among groups, and Mann-Whitney U test was used for pairwise comparison. Pearson correlation test was used for correlation analysis.

The biological reference intervals for TSH of group A-E were 0.40-6.29 mIU/L, 0.45-5.80 mIU/L, 0.43-6.52mIU/L, 0.40-6.28 mIU/L, and 0.46-6.65 mIU/L, respectively. Mann-Whitney U test showed that there were significant differences between group B and C (P＜0.001), and between group D and E (P＜0.05). There were significant differences in TSH in different age groups and among different seasons (F=9.26, P＜0.01; F=4.10, P＜0.01). There was a positive correlation between TSH and age (r=0.740, P＜0.01). Conversely, there was a negative correlation between TSH and temperature variation (r=-0.658, P＜0.05).

There are differences in biological reference intervals for TSH with regard to gender and age. Establishment of gender- and age-specific reference intervals properly will be helpful to clinical diagnosis and treatment of thyroid dysfunction. Seasonal and temperature variations should also be taken into account.