Altered bioavailability due to changes in the formulation of a commercial preparation of levothyroxine in patients with differentiated thyroid carcinoma

Abstract

OBJECTIVES: In July 1995 we began noticing an unusually high rate of elevated TSH levels in patients with differentiated thyroid cancer treated with levothyroxine-specifically the brand Levothroid-becoming more obvious from September 1995. Faced with the possibility that these findings had some relationship to the drug taken, we carried out a prospective study, changing this brand of levothyroxine for another. DESIGN AND PATIENTS: We studied 58 patients with differentiated thyroid carcinoma (50 women and eight men; aged 22-75 years) who were being treated with levothyroxine and who had previously had adequate TSH suppression. Their Levothroid tablets were changed to the same dose of Dexnon tablets, and their clinical and analytical response was evaluated 2 months later. The patients were divided into two groups according to their TSH level at the start of the study: group 1, 42 patients with TSH > 0.2 mU/l (not suppressed) and group 2, 16 patients with TSH < or = 0.2 mU/l (suppressed). RESULTS: After 2 months with Dexnon the TSH levels in group 1 fell significantly (P < 0.0001) also decreasing in group (P < 0.09). The free T4 and free T3 rose significantly in both groups. After the change to Dexnon, 17 patients (40%) in group 1 had suppressed TSH and 26 (62%) had free T4 levels above the upper limit of normal vs none at baseline (P < 0.001). The group 2 patients maintained their inhibited TSH values after treatment with Dexnon, and the free T4 was above the upper limit of normal in 15 (94%) vs 3 (19%) at baseline (P < 0.001). The Levothroid tablets collected from the patients in both groups formed part of those which the manufacturer later withdraw from the market. These batches possessed the correct dosage, but they had been made from 'non-micronized' raw materials from another supplier. CONCLUSIONS: The most probable cause of the inadequate TSH suppression in our patients was the reduction in bioavailability in certain batches of Levothroid, although we are unable to rule out the possibility that the results obtained after the changeover to Dexnon were due to its greater bioavailability. Simple changes in the manufacture of levothyroxine tablets may produce important variations in their bioavailability, having an adverse effect on the clinical control of the patients, and causing extra expense by the need for repeated patient visits and thyroid function tests.