|Home | About | Journals | Submit | Contact Us | Français|
For routine thyroid function testing many laboratories now confine themselves to assay of thyroid stimulating hormone (TSH). This will identify the great majority of affected patients1 though it can miss secondary hypothyroidism.2,3 TSH measurement is also used to monitor thyroxine replacement, and on rare occasions this too can be misleading.
A woman of 66 was seen in the emergency department with a history of recurrent falls, urinary incontinence and new-onset confusion. She walked with a shuffling small-stepped gait and her abbreviated mental test score was 1 out of 10. 10 years earlier she had been found to have primary hypothyroidism, TSH 12 mU/L (reference range 0.3-4.2), and on treatment with thyroxine 100 μg daily the TSH had become normal. A year before the present episode her TSH was noted to have fallen to 0.49 mU/L, which in the absence of a free T4 was interpreted as suggesting that she was taking too high a dose of thyroxine. The dose was reduced to 75 μg daily and three months later the TSH was 1.26 mU/L. However, a further six months later the TSH had fallen to 0.2 mU/L and the thyroxine dose was once again reduced, to 50 μg daily.
CT of her head showed hydrocephalus and a large cystic lesion in the region of the pituitary. MRI confirmed these findings (Figure 1a,b). Her baseline pituitary hormonal profile showed a very high prolactin of 398 500 mU/L (5650). The TSH was now raised at 22.0 mU/L, with a free thyroxine of 9.6 pmol/L (9.0-26.0); she had not taken thyroxine for about two weeks because of her confusion. Further assessment revealed a follicle stimulating hormone of 0.4 IU/mL, undetectable luteinizing hormone, growth hormone 0.3 IU/L with an IGF-1 4.0 nmol/L (6.0-30.0), and cortisol 765 nmol/L. These results indicated partial hypopituitarism. Thyroid peroxidase antibodies were strongly positive. She had a bitemporal hemianopia on perimetry. She was restarted on thyroxine replacement.
Transcranial drainage of the intraventricular cystic lesion did not improve her clinical state or lessen the hydrocephalus (Figure 1c). She was therefore started on cabergoline 500 μg twice weekly, and over the subsequent nine months her tumour shrank (Figure 1d-f), her visual fields improved, and her confusion resolved completely, to the extent that she was able to proof-read this paper.
This patient with primary autoimmune hypothyroidism went on to develop a giant macroprolactinoma. We propose that the continuous decline in her TSH despite reduction of the thyroxine dose was due to partial pituitary failure. Such a phenomenon has been observed by others;4 the low TSH in this patient reflected the development of hypopituitarism, not over-replacement leading to thyrotoxicosis.
The routine measurement of TSH to monitor the adequacy of thyroxine replacement has in this case given a fascinating insight into the changes occurring in the thyroid axis as hypopituitarism developed. The use of TSH as the sole index of function led to erroneous interpretation and inappropriate reductions of thyroxine dosage. Only when the patient was taking half the correct thyroxine dose, and stopped taking thyroxine altogether for a short period, did the TSH rise, by which time the pituitary tumour had caused hydrocephalus severe enough to produce confusion, incontinence and falls. Although the dose of thyroxine required to maintain the euthyroid state has been reported to fall with age,5 this would not have been a plausible explanation for the rapid decline in TSH seen in this patient. Had the free thyroxine also been measured, it might have prompted a search for other causes of a falling TSH during thyroxine replacement and led to diagnosis of the macroprolactinoma before the development of hydrocephalus.
We thank Dr J Kraemer for supplying details of thyroid function tests performed before the acute episode.