Author’s conclusion
	Subclinical hypothyroidism is a controversial subject, and there is no consensus on how it should be dealt with. However, increasingly more specialists consider that it should be treated. The evidence provided under "tissue effects" supports this view. Furthermore, as discussed in the preceding section, the author considers that persons with a TSH between 2.01–4.00 mU/l may not be entirely normal, and that at least part of them have an early degree of thyroid failure. So, what should the practicing physician do? Since there is no universally accepted practice (12,34), the author presents here his personal conclusions, taking of course into consideration all the various reports discussed in this review. These recommendations may be summarized as follows: Screen with a sensitive TSH assay all males 50 years and all females 35 years. If symptoms, signs or laboratory findings (such as lipid abnormalities) possibly due to hypothyroidism are present, screening should be performed irrespective of age. If TSH 5.00 mU/l, treat with thyroxine in order to normalize the serum level. If TSH 0.41–2.00 mU/l, repeat screening after 5 years. If TSH ? 0.40, the levels of serum fT4 and fT3 should be measured, in order to diagnose possible subclinical or overt hyperthyroidism. If TSH is 2.01–5.00 mU/l, check serum antithyroid autoantobodies (AAB ) and fT4 . If TSH 2.01–5.00 and AAB (+), treat those with a TSH 3.00 mU/l, and follow up those with a TSH 2.01–3.00 mU/l every one or two years. In the persons with a TSH 2.01–5.00 mU/l and negative AAB, repeat after six months or one year serum fT4, TSH and AAB estimations. If serum TSH 4.00 mU/l in more that two occasions, treat for hypothyroidism, if TSH 2.01–4.00 mU/l, follow-up every one or two years. These conclusions are also shown in Table 1. Of course, several other factors should be taken into account: The serum fT4 value may tip the balance for or against treatment in patients with marginal TSH alterations, the physician should be more conservative in debilitated persons with cardiac arrhythmias etc, but more willing to prescribe thyroxine in patients with serum lipid abnormalities, e.t.c. Such an individualized attitude is also recommended by Samuels (34). In other words, the physician must, in some way, practice "holistic" medicine, i.e. decide what to do not on the basis of a single laboratory abnormality, but after considering the patient as a whole, applying his knowledge and experience to the best of the individual patient who has asked for his advice. The Hippocratic dogma "benefit, or (at least) don’t harm" is always true.