Thyroid hormone replacement has been used for more than 100 years in the treatment of hypothyroidism, and there is no doubt about its overall efficacy. Desiccated thyroid contains both thyroxine (T<sub>4</sub>) and triiodothyronine (T<sub>3</sub>); serum T<sub>3</sub> frequently rises to supranormal values in the absorption phase, associated with palpitations. Liothyronine (T<sub>3</sub>) has the same drawback and requires twice-daily administration in view of its short half-life. Synthetic levothyroxine (L-T<sub>4</sub>) has many advantages: in view of its long half-life, once-daily administration suffices, the occasional missing of a tablet causes no harm, and the extrathyroidal conversion of T<sub>4</sub> into T<sub>3</sub> (normally providing 80% of the daily T<sub>3</sub> production rate) remains fully operative, which may have some protective value during illness. Consequently, L-T<sub>4</sub> is nowadays preferred, and its long-term use is not associated with excess mortality. The mean T<sub>4</sub> dose required to normalize serum thyroid stimulating hormone (TSH) is 1.6 µg/kg per day, giving rise to serum free T<sub>4</sub> (fT<sub>4</sub>) concentrations that are slightly elevated or in the upper half of the normal reference range. The higher fT<sub>4</sub> values are probably due to the need to generate from T<sub>4</sub> the 20% of the daily T<sub>3</sub> production rate that otherwise is derived from the thyroid gland itself. The daily maintenance dose of T<sub>4</sub> varies widely between 75 and 250 µg. Assessment of the appropriate T<sub>4</sub> dose is by assay of TSH and fT<sub>4</sub>, preferably in a blood sample taken before ingestion of the subsequent T<sub>4</sub> tablet. Dose adjustments can be necessary in pregnancy and when medications are used that are known to interfere with the absorption or metabolism of T<sub>4</sub>. A new equilibrium is reached after approximately 6 weeks, implying that laboratory tests should not be done earlier. With a stable maintenance dose, an annual check-up usually suffices. Accumulated experience with L-T<sub>4</sub> replacement has identified some areas of concern. First, the bioequivalence sometimes differs among generics and brand names. Second, many patients on T<sub>4</sub> replacement have a subnormal TSH. TSH values of ≤0.1 mU/l carry a risk of development of atrial fibrillation and are associated with bone loss although not with a higher fracture rate. It is thus advisable not to allow TSH to fall below – arbitrarily – 0.2 mU/l. Third, recent animal experiments indicate that only the combination of T<sub>4</sub> and T<sub>3</sub> replacement, and not T<sub>4</sub> alone, ensures euthyroidism in all tissues of thyroidectomized rats. It is indeed the experience of many physicians that there exists a small subset of hypothyroid patients who, despite biochemical euthyroidism, continue to complain of tiredness, lack of energy, discrete cognitive disorders and mood disturbances. As organs vary in the extent to which their T<sub>3</sub> content is derived from serum T<sub>3</sub> or locally produced T<sub>3</sub> from T<sub>4</sub>, these complaints may have a biologic substrate; for example, brain T<sub>3</sub> content is largely determined by local deiodinase type II activity. Against this background it is of interest that a number of psychometric scores improved significantly in hypothyroid patients upon substitution of 50 µg of their T<sub>4</sub> replacement dose by 12.5 µg T<sub>3</sub>. Confirmatory studies on this issue are urgently awaited. It could well be that a slow-release preparation containing both T<sub>4</sub> and T<sub>3</sub> might improve the quality of life, compared with T<sub>4</sub> replacement alone, in some hypothyroid patients.