In a prospective study of critically ill hypothyroxinemic we assessed the relationship between serum TSH and T4 during the return of serum T4 to normal during recovery. In this longitudinal study of 60 patients with a variety of critical illnesses, including burns, septicemia, and acute renal failure, serum T4 fell to less than 2.7 micrograms/dl (35 nmol/liter) in 24 patients, of whom 14 survived with return of T4 to normal. A rise in total T4 of more than 1.9 microgram/dl (25 nmol/liter) within 96 h occurred 13 times in 10 patients, while 4 patients had slower increases in T4. All 13 episodes of rapid T4 rise [1.7 +/- 0.8 (+/- SD) to 5.6 +/- 2.1 micrograms/dl] were associated with a marked increase in serum TSH (1.1 +/- 0.8 to 7.0 +/- 5.2 mU/liter), and TSH was transiently above normal during 8 episodes of T4 recovery. In the 6 episodes with sampling less than 6 h apart, the TSH rise consistently preceded the T4 rise. In the 4 patients who received dopamine, TSH and T4 remained low until cessation of therapy. During the TSH rise, only minor changes, which could not account for the increase in total T4, occurred in T4-binding globulin (12.9 +/- 3.3 to 14.8 +/- 3.3 mg/liter), prealbumin (208 +/- 73 to 234 +/- 82 mg/liter), and albumin (28.3 +/- 2.9 to 31.9 +/- 2.9 g/liter). Mean free T4 increased (0.60 +/- 0.34 to 1.45 +/- 0.56 ng/dl), as did total T3 (16 +/- 14 to 76 +/- 44 ng/dl), during the phase of TSH rise, suggesting that the increase in TSH was not simply a consequence of diminished negative feedback due to increased plasma binding. The very close and consistent temporal relationship between TSH and T4 during the recovery phase suggests that TSH may have an essential role in the return of T4 to normal during recovery from critical nonthyroidal illness.

The diuretic furosemide inhibits serum protein binding of T4 in equilibrium dialysis, dextran-charcoal, and competitive ligand binding separation systems and displaces [125I]T4 from isolated preparations of T4-binding globulin (TBG), prealbumin, and albumin. Equilibrium dialysis studies of undiluted normal serum showed that about 10 micrograms/ml furosemide increased the free T4 and free T3 fractions. Displacement occurred at lower drug concentrations in sera with subnormal albumin and TBG levels. Binding of [14C]furosemide to TBG was inhibited by unlabeled T4, suggesting that furosemide and T4 share a common binding site. A single oral dose of 500 mg furosemide given to five patients maintained on peritoneal dialysis increased the percentage of charcoal uptake of [125I]T4 (using serum diluted 1:10) from 4.1 +/- 1.0 (+/- SE) to 10.8 +/- 4.3 (P less than 0.01) after 2 h, while decreasing total T3 from 75 +/- 5 to 56 +/- 13 ng/dl (P less than 0.01) and total T4 from 6.7 +/- 0.9 to 4.8 +/- 0.8 micrograms/dl (P less than 0.01) after 5 h. Various ligands inhibited [125I]T4 binding to serum proteins in the following relative molar relationship: T4, 1; furosemide, 1.5 X 10(3); fenclofenac, 2 X 10(4); mefenamic acid. 2.5 X 10(4); diphenylhydantoin, 4 X 10[4); ethacrynic acid, 10(5); heparin 5 X 10(5); 2-hydroxybenzoylglycine, 10(6); and sodium salicylate, 1.5 X 10(6). These studies demonstrate that furosemide competes for T4-binding sites on TBG, prealbumin, and albumin, so that a single high dose can acutely lower total T4 and T3 levels. The drug is much more potent on a molar basis than other drug inhibitors of T4 binding, but at normal therapeutic concentrations, furosemide is unlikely to decrease serum T4 or T3. However, high doses, diminished renal clearance, hypoalbuminemia, and low TBG accentuate its T4- and T3-lowering effect. Hence, furosemide should be considered a possible cause of low thyroid hormone levels in patients with critical illness. The significance of this drug in reports of impaired hormone and drug binding in renal failure requires further assessment.

Methods to identify the plasma T4-binding abnormalities that can cause euthyroid hyperthyroxinemia were evaluated in patients with excess T4-binding globulin, familial dysalbuminemic hyperthyroxinemia, prealbumin-associated hyperthyroxinemia, and autoantibody binding of T4. Familial dysalbuminemic hyperthyroxinemic serum showed a unique persistence of abnormal [125I]T4 binding when diluted 1:100 in phosphate buffer with added 1000-fold excess of unlabeled T4 (10(-6) M T4). Immunoprecipitation of [125I]T4 by antibody to prealbumin, precipitation of [125I]T4 by polyethylene glycol 6000 19%, and in vitro resin uptake of T3 were specific for prealbumin-associated hyperthyroxinemia, autoantibody binding of T4, and T4-binding globulin excess, respectively. These simple methods facilitate investigation of patients with euthyroid hyperthyroxinemia and will identify individuals and families at risk of misdiagnosis by standard methods. Use of these techniques rules out the known binding abnormalities in hyperthyroxinemic patients and may make the diagnosis of generalized hormone resistance more specific.

Low serum total T4 associated with subnormal concentrations of thyroxine binding globulin (TBG) has been reported in up to 40% of euthyroid Australian aborigines. It has been suggested that these subjects show both diminished concentration of TBG and reduced TBG affinity for T4 (Sarne et al., 1985). We have compared 12 euthyroid aborigines with low T4 (total T4 44 +/- 5 nmol/l) and aborigines with normal T4 (T4 99 +/- 9 nmol/l, n = 12) using measurements of free T4 and T3 by equilibrium dialysis. TBG was measured both by RIA (Henning, Berlin, FRG) and a method dependent on T4 binding (Corning Immophase). Aborigines with low T4 showed lower levels of free T4 (12.6 +/- 0.6 cf. 18.7 +/- 1.0 pM), free T4 index (66 +/- 8 cf. 98 +/- 13), total T3 (1.1 +/- 0.2 cf. 1.6 +/- 0.3 nmol/l), TBG RIA (14.0 +/- 0.6 cf. 25.0 +/- 1.2 ng/l), and TBG Immophase (9.0 +/- 0.5 cf. 22.0 +/- 1.2 mg/l) (P less than 0.01), but free T3 (5.3 +/- 0.4 cf. 4.7 +/- 0.4 pM) and TSH (1.9 +/- 0.2 cf. 1.8 +/- 0.2 mU/l) were not significantly different from the values found in aborigines with normal T4. Scatchard analysis of T4 and T3 binding was performed using serum diluted 1 : 20,000 for T4 and 1 : 500 for T3 (barbitone buffer pH 8.6, 4 degrees C, dextran-coated charcoal separation). In euthyroid low T4 aborigines compared to those with normal T4, both T4 capacity (106 +/- 14 cf. 238 +/- 13 nM, P less than 0.01) and affinity (5.05 X 10(10) cf. 8.47 X 10(10) M-1, P less than 0.05) were significantly reduced. Similarly, both T3 capacity (62 +/- 10 cf. 154 +/- 16 nM, P less than 0.01) and affinity (1.67 X 10(9) cf. 2.28 X 10(9) M-1, P less than 0.02) were reduced. A substantial minority of euthyroid Australian aborigines have a TBG variant characterized by both reduced capacity and affinity of T4 and T3. These findings suggest that TBG may be both qualitatively and quantitatively abnormal in these subjects.