Should we really determine a reference population for the definition of thyroid-stimulating hormone reference interval?