Thyroid over-expression of type 1 and type 2 deiodinase may account for the syndrome of low thyroxine and increasing triiodothyronine during propylthiouracil treatment.

Although propylthiouracil inhibits type 1 deiodinase, leading to a more rapid fall in triiodothyronine (T(3)) than thyroxine (T(4)) levels in patients treated for hyperthyroidism, we report a patient with Graves' disease whose free T(3) paradoxically rose during such treatment, despite low free T(4) levels and increasing doses of propylthiouracil. A similar response has previously been associated with high levels of thyroid stimulating antibodies, but it has been unclear why there should be a dichotomy in the circulating thyroid hormone profile. Thyroid tIssue from our patient contained very high levels of type 1 and, especially, type 2 deiodinase, in contrast to other patients treated with Graves' disease, which were most likely secondary to high levels of thyroid stimulating antibodies. This unusual response to propylthiouracil is important to recognise therapeutically, and represents a further situation in which abnormal expression of deiodinase enzymes has clinical significance.