5.67 μU /ml in the whole cohort.
3.73 μU /ml (1448 patients) after having excluded subjects referring to have had thyroid problems or presenting an abnormal clinical examination.
3.37 μU /ml (1186 patients) after further exclusion of the subjects with anti-thyroid antibodies.
3.37 μU /ml (766 subjects) after further exclusion of subjects with thyroid abnormalities at ultrasonography. To finish, TSH levels were measured with ELISA test (which is now no more used). Performing dosages with a new third-generation system, TSH values were obtained averagely higher of 0.73 μU /ml compared to ELISA method. So, in probably normal subjects, the higher TSH cutoff should be of 4.1 μU /ml (3.37 + 0.73).
We can also accept the conclusions of this trial (4 μU /ml is the higher limit of normality for TSH), but the main aspect remains to be cleared up: from a clinical point of view, is it useful to treat subjects presenting a higher TSH but being completely asymptomatic with thyroid supplementations?