Are thyroid function tests too frequently and inappropriately requested?

In spite of data supporting the use of the serum thyrotropin (TSH) concentr ation as the best test to detect abnormal thyroid function, measurement of circulating thyroid hormones with or without a serum TSH continues to be fr equently requested to evaluate thyroid function. We have analyzed how combi nations of thyroid function tests were ordered by referring physicians and the results of the tests in order to offer some suggestions as to how to us e thyroid function tests in a cost effective manner. During 1995, 19,181 in patient and outpatient requests (45,865 different tests) for thyroid functi on tests were received by the laboratory of a 1600 bed University Hospital in Parma, Italy. The following tests were carried out: T4, free T4, T3, fre e T3 and TSH. Serum TSH values below and above the normal range were consid ered to reflect abnormal thyroid function i.e. hyperthyroidism, or hypothyr oidism including subclinical disease independent of the results of the othe r tests. Combinations of ordered tests and the percent of the total for eac h combination were: TSH+T4+T3 (56%), TSH+FT4+FT3 (14%), TSH (12%), TSH+FT4 (9%), TSH+T4 (1%), TSH+T4+T3+FT4+FT3 (5%), others (3%). The T4+T3+TSH panel (10,780 requests) had normal serum TSH values in 80.6% and the FT4+ FT3+TS H panel (2,590 requests) had normal TSH values in 73.2%. Elevated serum TSH concentrations were observed more frequently in hospitalized than in ambul atory patients (9.7% vs 7.4% p<0.001). T3 (elevated serum T3, normal T4 and low TSH concentrations) and T4 (elevated serum T4, normal T3 and low TSH c oncentrations) toxicosis were observed in 8.1% and 9.4%, respectively, of t he requested test (NS). FT3 and FT4 toxicosis, defined as for T3 and T4 tox icosis, were observed in 7.5% and 4.9%, respectively (NS). The low T3 and l ow FT3 syndrome in hospitalized patients was present in 1.6% and 2.3% of th e requests, respectively (NS). The low T4+low T3 and low FT4+low FT3 syndro me was present in only 0.3% and 0.2%, respectively, of the requests. Our st udy shows that a) in hospitalized patients thyroid function tests were requ ested in 20% of the patients and only one in 14 of these patients at the hi ghest could have abnormal thyroid function, as indicated by abnormal TSH va lue b) FT4 (or T4) is as useful as FT3 (or T3) in the diagnosis of hyperthy roidism, c) in hospitalized patients the low T3 syndrome was far less commo n than that reported in the literature, probably due to the lower severity of illness, d) panels which include T3 and FT3 are not justified, and e) se rum TSH alone is the most appropriate initial thyroid function test. (C) 19 99, Editrice Kurtis.