Serum TSH measurement in children with thyroid disorders.

A number of thyroid function tests have recently been developed in which components of the hypothalamic-pituitary-thyroid axis are directly measured; these include estimation of the total concentrations of the thyroid hormones, thyroxine (T4) and tri-iodothyronine (T3), their free (unbound) fractions, the thyroid hormone-binding proteins, and the serum TSH (thyroid-stimulating hormone, thyrotrophin). Thus, though the peripheral effects ofthe thyroid hormones are difficult to assess quantitatively, derangement of thethyroid axis can be precisely defined. The theoretical advantages of such direct measurements are reflected in greatly increased diagnostic sensitivity. Of the tests currently available the serum T4 and TSH have become the cornerstones of thyroid investigation, and the papers in this issue by Jackson, Vanderschueren-Lodeweyckx, and Grant (page 522) and Hamilton and Hutchison (page 567) emphasize the value of serum TSH measurement in the assessment of children with thyroid disorders. The laboratory diagnosis of thyroid disease in children was well reviewed by Fisher (1973). TSH is a glycoprotein hormone synthesized by the thyrotroph cells of the anterior pituitary. TSH secretion is regulated directly by the circulating level of thyroid hormones and indirectly by the tripeptide hypothalamic thyrotrophin-releasing hormone TRH. By stimulation of thyroid epithelial cell adenyl cyclase, TSH enhances iodide uptake, hormone synthesis, and release. Serum TSH concentration shows little diurnal variation and is constant in both sexes throughout life from day 4 onward. Cord blood levels are higher than maternal levels and a marked TSH surge occurs immediately after birth, reaching a peak at 30 minutes and falling to normal by day 3 (Fisher and Odell, 1969). Early bioassays for serum TSH have been superseded by radioimmunoassays which are now routine procedures in many laboratories and in Britain are widely available through the supraregional assay service. The laboratory and clinical aspects of TSH immunoassay were reviewed by Hall (1972). TSH assays require only small quantities of serum and are reliable and specific. The sensitivity of most current assays is such that basal levels are undetectable in about 10-20% of normal subjects and decreased levels cannot reliably be detected. Normal values should be established for each assay and each laboratory, but basal concentrations in normal subjects usually fall within the range 0-5 ,uU/ml of the MRC standard 68/38. Evered et al. (1975) found that less than 2 5% of normal adults had serum TSH values greater than 5tLU/ml, and Jackson et al., using a similar assay, now report identical findings in normal children. Serum TSH may be measured under basal conditions, after stimulation by TRH, and after suppression by administration of thyroid hormones. TRH is active when given orally, but the intravenous stimulation test of Ormston et al. (1971), which can conveniently be combined with insulin and gonadotrophin-releasing hormone stimulation in a comprehensive test of anterior pituitary function (Mortimer et al., 1973), has been most thoroughly evaluated and widely used. After a base-line sample for serum TSH has been obtained, intravenous TRH is given (a dose of 200 rig, given as a bolus, is suitable for all ages; transient nausea occurs frequently but no serious side-effects have been recorded) and the serum TSH is measured 20 and 60 minutes later. Normal subjects show a sharp rise in serum TSH to a peak at 20 minutes and a fall toward the basal value at 60 minutes. The response in children is similar to that in adults (Job et al., 1971; Foley et al., 1972; Malvaux and Beckers, 1973). The value of the test may be enhanced by measurement of the thyroid hormone response, the T4 showing a small, and the T3 a larger, response, both reaching a peak by 4 hours (Uller, Van Herle, and Chopra, 1973). The normal limits of response to TRH stimulation in children have not yet been adequately defined and are clearly subject to the same variability between assays as the basal values. Fortunately, however,