Henry Ford Hospital Medical Journal


We conducted a pilot study to evaluate an algorithm for thyroid function testing consisting of initial serum thyrotropin values, measured by a sensitive immunoradiometric assay (TSH-IRMA) followed by a computer-directed decision to order further studies. We divided 216 outpatients according to their serum TSH-IRMA values as follows: suppressed (< 0.1 mU/L, group I); low (0.1 to 0.4 mU/L, group II); normal (0.5 to 5.0 mU/L, group III); and high (> 5.0 mU/L, group IV). Thyroxine (T4), resin uptake (RU). and free thyroxine index (FTI) tests on groups I, Il, and IV revealed that T4 and RU were normal for most patients in all groups and FTI was normal in 80% of group 1, 93.4 % of group ll, and 93.3% of group IV. All patients in group I were designated hyperthyroid from either an exogenous or endogenous source. All patients in group ll were clinically euthyroid except one; 50% were taking either L-thyroxine or propylthiouracil and 50% had no identifiable thyroid disease. Patients in group IV were hypothyroid. Overall, TSH was more effective in detecting both hypothyroidism and hyperthyroidism than either serum T4, RU ratio, or both combined in FTI since results of these measures fell in the normal range for most patients in all groups. We conclude that a computer-directed algorithm with TSH-IRMA as the initial step is useful in the evaluation of suspected thyroid dysfunction, that T4 and RU may be helpful when TSH is abnormal or borderline, and that suppressed TSH-IRMA values (<0.1 mU/L) but not low values (0.1 to 0.4 mU/L) are consistently associated with hyperthyroidism. Results obtained by use of the algorithm may be misleading in patients with hypothalamic pituitary dysfunction, but its use should reduce the number of redundant and unnecessary T4 and RU tests.