No large randomised controlled trials in subclinical hyperthyroidism nor subclinical hypothyroidism have used robust clinical endpoints, particularly cardiovascular morbidity and mortality. Despite this absence of evidence, David Cooper and Bernadette Biondi (March 24, p 1142)1 make rather strong recommendations favouring treatment in both situations, particularly for patients with concentrations of thyroid-stimulating hormone below 0·1 mU/L or above 5·0 mU/L.

As mentioned in their Introduction, subclinical thyroid diseases are defined on a biological basis, might but might not always constitute an early stage of “overt” th