Normal serum free thyroid hormone concentrations in patients treated with phenytoin or carbamazepine: a paradox resolved

The finding of decreased thyroid hormone levels in people taking anticonvulsants appears to be an artifact of the test used to measure the hormone levels. Researchers took blood samples from 9 people taking phenytoin and 10 people taking carbamazepine. All 19 had no history of thyroid disease. Adding therapeutic concentrations of the drugs to the blood samples caused total T4 thyroid hormone to drop significantly, but free T4 hormone increased substantially when measured by an assay that uses undiluted blood. This increase in free T4 was not seen when the hormone was measured using an assay that dilutes the blood. In fact, free T4 levels dropped. These drugs displace T4 from T4-binding proteins. This would tend to raise free T4 levels. However, when blood samples are diluted, the drug concentration drops substantially and no longer displaces T4. Physicians should monitor TSH levels in these patients, since TSH levels are usually normal.
Objective. - To address the paradox that phenytoin- and carbamazepine-treated patients have decreased serum free thyroxine ([T.sub.4]) and triiodothyronine ([T.sub.3]) concentrations but appear clinically euthyroid and have normal serum thyroid-stimulating hormone (TSH) concentrations. Design. - In vitro studies comparing measurements of total arid free [T.sub.4] and [T.sub.3] by ultrafiltration assay (undiluted serum) and a commercial free [T.sub.4] estimate kit in control serum samples or serum samples containing added therapeutic levels of phenytoin or carbamazepine. These measurements were made in serum samples diluted 1:5 with either identical serum or phosphate buffer, pH 7.4, and in serum samples from patients with seizure disorders who were treated with phenytoin or carbamazepine. Setting. - A 650-bed teaching hospital. Patients. - Selected patents (n=19) who were in good health except for seizure disorder, with stable anticonvulsant drug levels in the upper had of the therapeutic range, and were not taking any other drugs that could affect thyroid parameters. Main Outcome Measure. - Serum concentrations of free [T.sub.4] and free [T.sub.3] in patients taking phenytoin or carbamazepine vs normal controls. Results. - Addition of phenytoin or carbamazepine to normal human serum in vitro resulted in a significant increase in free [T.sub.4] fraction and free [T.sub.4] (P < .001). In patients taking phenytoin or carbamazepine, serum total [T.sub.4] decreased significantly to 60% and 74%, respectively, of the control serum concentration (P < .001 for both phenytoin and carbamazepine); free [T.sub.4] fraction (by ultrafiltration assay) increased 65% and 44%, respectively (P < .001 for phenytoin, P < .01 for carbamazepine); and free [T.sub.4] remained unchanged. Free [T.sub.4] concentration measured by a commercial kit (1:5 serum dilution) was significantly lower than the control concentration in both phenytoin- and carbamazepine-treated patents. Serum free [T.sub.3] and serum TSH were also normal in phenytoin- and carbamazepine-treated patients. Conclusions. - Therapeutic levels of phenytoin and carbamazepine displace [T.sub.4] and [T.sub.3] from serum binding proteins. When added to serum, the drugs effect an increase in free hormone fractions and free [T.sub.4] and [T.sub.3]. In drug-treated patients, increased free [T.sub.4] and free [T.sub.3] fractions offset the significant decrease in serum [T.sub.4] and [T.sub.3], resulting in normal free [T.sub.4] and free [T.sub.3] concentrations. Since currently available clinical tests will continue to show decreased free [T.sub.3] concentrations in patients taking phenytoin or carbamazepine, clinicians should rely on serum TSH measurements to confirm the euthyroid status of these patients.