Persistent problems with the specificity of immunometric TSH assays

Sensitive immunometric TSH assays have led to improved diagnosis of thyroid function abnormalities, and serum TSH is often used as the initial screening method for evaluation of thyroid function. Shortly after the marketing of the first of these assays, it was discovered that some sera contained heterophilic antibodies leading to spuriously high TSH values. This problem is considered to be solved by modifications of the assays. We have, however, found that these assays may still give nonspecific results. A panel of sera from 63 patients with untreated hyperthyroidism (51 Graves' disease, 12 Plummer's disease) were used for evaluation of 6 different TSH assays (BeriLux, Spectria, London, Amerlite, Delfia, Delfia Ultra). Several problems were encountered. Most importantly, although all 63 sera gave totally suppressed TSH values in at least one of the assays, many sera gave clearly and reproducibly measurable values in one or more assays. Eleven sera gave values of more than 0.10 mU/L (6 more than 0.40 mU/L) in at least one assay. The pattern of nonspecific values was different in all 6 assays. TRH stimulation did not alter the measured values. Addition of 10% mouse serum to the patient sera gave unpredictable alterations with lower TSH measurements in one assay, but unaffected values in other assays. Hence a measurable serum TSH alone does not always exclude primary hyperfunction of the thyroid. Nonspecific TSH values are the most common cause of nonsuppressed TSH in hyperthyroidism and should be considered before other causes, such as pituitary TSH producing neoplasms or pituitary resistance to thyroid hormone.