Subclinical hypothyroidism is associated with aspecific complaints such as tiredness, cognitive and depressive complaints, subtle disturbances in lipid values, an increased risk of cardiovascular disease, ovulatory dysfunction and a negative effect on foetal psychomotor development and pregnancy outcome. Subclinical hyperthyroidism is associated with atrial fibrillation, osteoporosis and dementia. Not enough prospective randomised studies with hard outcomes are available to provide evidence-based general recommendations. Therefore, the decision as to whether or not a patient should be treated needs to be made on an individual basis. For subclinical hypothyroidism it is advisable to consider treatment in the case of positive thyroid peroxidase antibody tests, a TSH concentration higher than 10 mU/l, the presence of one or more risk factors for cardiovascular disease, infertility on the basis of ovulatory dysfunction, and pregnancy. In the case of complaints of tiredness and certainly in the case of depression or cognitive dysfunction, a 3-month trial treatment can be considered. This leads to a decrease of the complaints in about 25% of cases. As negative effects are associated with the treatment, we advise an expectant approach in all other cases with a yearly monitoring of the TSH concentration. For subclinical hyperthyroidism it is advisable to consider treatment in the case of a nodular goitre, and especially in the case of atrial fibrillations. If subclinical hyperthyroidism persists in the absence of nodular thyroid disease, an expectant approach appears to be justified.