Serum TSH
	Modern assay techniques permit to reliably distinguish between normal TSH values in euthyroid subjects and suppressed TSH levels in hyperthyroid patients, obviating the former need for TRH testing. Analytical assay sensitivity has dramatically improved over the past years down to 0.001 mU/l. However, since analytical sensitivity per se has little clinical relevance, sensitivity of a TSH assay should be given in terms of its functional sensitivity currently defined as the point of the interassay profile where the coefficient of interassay variability (CV) is 20%. Functional sensitivity should be below 0.10 mU/l. Tests not fulfilling this criterion should no longer be used. Table 1 summarizes criteria for quality assessment of TSH assays to be employed for routine diagnostic purpose. Experience in our own laboratory with more than 12 000 samples per year has shown that functional assay sensitivity is the most important performance criterion and should be assessed independently in each laboratory. In addition, control samples with TSH concentrations of about 0.1 mU/l, 0.3 mU/l and 4.0 mU/l should be included in every run in order to early recognize imprecision of TSH determinations. It is the manufacturer|s and the laboratory|s responsibility, to correctly define analytical and clinical performance of a TSH assay. Table 1: Criteria of quality assessment for routine TSH assays Functional sensitivity <0.1 mU/l. Cross reactivity with other glycoprotein hormones (LH, FSH, HCG) <0.01%. Parallel course of sample dilutions and the standard curve. Full recovery of a TSH standard added to patient's sample. Estimates of WHO- or Medical Research Council (MRC) standards ±5% of the expected values. Clear-cut separation of euthyroid and hyperthyroid patients (with an overlap of less than 1%). Lack of interference with heterophile antibodies. Absense of a high dose "hook" effect at TSH concentrations of <300 mU/l. When these quality measures are operative retesting of an inappropriate TSH value has been a rare necessity in our experience. Nonspecific interference will best be detected in a situation where examination of a patient is done by a thyroid expert who should notice apparent discrepancies between the presentation of the patient and results of in vitro tests. In case of discrepant laboratory results we recommend to retest the sample and to check for parallelity with standard curve in diluted samples. If necessary, a new specimen from the patient should be collected and TRH stimulation should be performed. When using this approach, retesting the sample in a different type of TSH assay and the need for checking for the presence of heterophile antibodies is hardly ever necessary. In our setting it is important to note that low normal or suppressed TSH values may be frequently found in patients with secondary hypothyroidism due to pituitary abnormalities. In such situations clinical information is particularly helpful to distinguish primary hyperthyroidism with suppressed TSH from secondary hypothyroidism with inappropriately low TSH values. Suspicion of secondary hypothyroidism is an important condition where screening for thyroid dysfunction by basal TSH alone may be inappropriate or misleading. In a considerable portion of patients TSH values may be found within the normal range despite manifest hypothyroidism as evidenced by decreased FT4 values. The condition can only reliably be diagnosed by interpreting TSH in relation to free thyroid hormone concentrations (FT4). Since determination of TSH alone is unable to distinguish between subclinical and overt thyroid dysfunction, each abnormal TSH value has to be followed or accompanied by assessment of peripheral hormone status. TRH testing has become unneccessary in most thyroid disorders except for hypothalamic or pituitary diseases.