Data suggesting that (1) sulfation of the phenolic hydroxyl of iodothyronines plays an important role in thyroid hormone metabolism and (2) maternal serum 3,3'-diiodothyronone sulfate (3,3'-T(2)S) may reflect on the status of fetal thyroid function stimulated us to develop a radioimmunoassay (RIA) for measurement of T(2)S. Our T(2)S RIA is highly sensitive, practical, and reproducible. T(4)S, T(3)S, and T(1)S crossreacted 3.1%, 0.81%, and 5.3%, respectively; thyroxine (T(4)), triiodothyronine (T(3)), and reverse (r)T(3), 3,3'-T(2) and 3'-T(1) crossreacted <0.1%. Although rT(3) sulfate (rT(3)S) crossreacted 55% in 3,3'-T(2)S RIA, its serum levels are very low and have little influence on serum T(2)S values reported here. T(2)S was measured in ethanol extracts of serum, amniotic fluid, and urine. Recovery of nonradioactive T(2)S added to serum was 96%. The dose-response curves of inhibition of binding of (125)I-T(2)S to anti-T(2)S by serial dilutions of ethanol extracts of serum or urine were essentially parallel to the standard curve. The detection threshold of the RIA varied between 0.17 and 0.50 nmol/L (or 10 and 30 ng/dL). The coefficient of variation (CV) averaged 9% within an assay and 13% between assays. The serum concentration of T(2)S was [mean +/- SE, nmol/L] 0.86 +/- 0.59 in 36 normal subjects, 2.2 +/- 0.06 in 10 hyperthyroid patients (P <.05), 0.73 +/- 0.10 in 11 hypothyroid patients (not significant [NS]), 6.0 +/- 1.5 in 16 patients with systemic nonthyroidal illness (P <.001), 18 +/- 2.5 in 16 newborn cord blood sera (P <.02), 2.7 +/- 0.32 in 25 pregnant women [15 to 40 weeks gestation, P <.001], 0.94 +/- 0.10 in 10 hypothyroid women receiving T(4) replacement therapy (NS), and 2.0 +/- 0.38 in 11 hypothyroid women treated with T(4) replacement and oral contraceptives (P <.02); serum T(2)S levels in the third trimester of pregnancy were similar to those in the second trimester of pregnancy. T(2)S concentration in amniotic fluid was 12.5 +/- 2.7 nmol/L (n = 7) at 15 to 20 weeks gestation, and it decreased markedly to 3.3 +/- 1.3 nmol/L (n = 3) at 35 to 38 weeks gestation. Urinary excretion of T(2)S in random urine samples of 19 normal subjects was 10.9 +/- 1.3 nmol/g creatinine. (1) T(2)S is a normal component of human serum, urine, and amniotic fluid, and serum T(2)S levels change substantially in several physiologic and pathologic conditions; (2) high serum T(2)S in pregnancy may signify increased transfer of T(2)S from fetal to maternal compartment, estrogen-induced increase in T(2)S production, decreased clearance, or a combination of these factors. The data do not support the notion that fetal thyroid function is the only or the predominant factor responsible for high serum T(2)S in pregnant women.