Aim The upper reference limit of thyroid-stimulating hormone (TSH) is critical for defining patients with subclinical hypothyroidism, a condition which carries a higher risk of progression to overt hypothyroidism and adverse cardiovascular events. Yet, there is a lack of consensus on its absolute value, and data in non-pregnant adult Chinese are lacking.

Methods Apparently healthy and drug-free local adult Chinese were recruited by completing health questionnaires. Their serum samples were tested for TSH, free thyroxine (FT4), thyroglobulin antibody and thyroid peroxidase antibody levels. After excluding subjects with thyroid antibodies, the TSH level was log-transformed, and the reference limits were defined as mean±1.96SD. The 2.5th and 97.5th percentiles of FT4 were also calculated.

Results Serum samples from 212 subjects were used in this study. 51 subjects were seropositive to thyroglobulin antibody, 31 were seropositive to thyroid peroxidase antibody, and 27 were seropositive to both. The reference intervals after excluding subjects seropositive to thyroid antibodies were: TSH: 0.68–3.70 mIU/l; FT4: 13.5–21.3 pmol/l (male) and 12.6–19.7 pmol/l (female). Including subjects with thyroid antibodies only minimally changed the reference intervals of these hormones.

Conclusion The authors have set up the reference interval of TSH for the local population, and their findings also suggest that the importance of excluding subjects with thyroid antibodies in the reference population should not be overemphasised. Moreover, the international authorities should consider recommending percentile-equivalent action limits instead of an absolute cut-off on TSH for categorisation of different types of thyroid dysfunction.