Abstract
Recently, Livadas et al. presented the results of their clinical study on the potential overuse of levothyroxine (LT4) supplementation in this journal (1). They concluded that there is overuse of LT4 therapy based on the finding that 60.8% of subjects on long-term LT4 treatment remained euthyroid six to eight weeks after treatment cessation. A single thyrotropin (TSH) value of ≥4.5 IU/L post LT4 discontinuation was considered as underlying hypothyroidism. They suggested a six- to eight-week period of LT4 discontinuation trial in patients without a solid diagnosis of hypothyroidism so as to evaluate the necessity of LT4 therapy continuation.
In our opinion, the interval of six to eight weeks off LT4 for estimating the thyroid function status in patients under long-term LT4 treatment is relatively short. It is known that in euthyroid subjects, withdrawal of LT4 treatment is a cause of transient hypothyroidism (2,3). Indeed, upon abrupt withdrawal of LT4 treatment in euthyroid subjects, the restoration of pituitary and thyroid function could take up to 12 weeks, especially in patients under long-term LT4 treatment (3). Briefly, within the first week off LT4, thyroxine (T4) and triiodothyronine (T3) drop below the lower limit of normal range, reach a nadir by the end of the second week, and reach low normal values five weeks later. Secretion of TSH is transiently impaired for two to five weeks, as indicated initially by low serum TSH levels and subsequently by normal levels, while serum T4 and T3 remains low. Thereafter, in some patients, a rebound of TSH can be observed, probably due to the recovery of the hypothalamus–pituitary–thyroid (HPT) axis, and thus serum TSH can be found slightly elevated within the next weeks (3). Therefore, Livadas et al. may have characterized some subjects as hypothyroid after LT4 discontinuation who were actually euthyroid due to this potential TSH rebound. In addition, a phenomenon of impaired TSH secretion has been described after restoration of long-standing hyperthyroidism—so-called hysteresis of the HPT axis—that could last for weeks or even months (4). Inspecting the data of Livadas et al., we observe that the standard deviations of serum TSH in both groups were relatively high, namely 0.79 and 0.83, taking into consideration that the mean TSH is approximately 1.5 IU/L. This observation indicates that some subjects in both groups had serum TSH levels just above the lower limit of reference range. So, they may have underlying iatrogenic subclinical hyperthyroidism that could lead, to some extent, to a delay in TSH increase (“hysteresis”) after LT4 discontinuation. Hence, the hypothyroidism in these subjects could have been masked.
It is also noteworthy that there is a discrepancy between the TSH values in table 1 and the ones mentioned in the text (Results section, fourth paragraph). In conclusion, although the authors recognize the limitations of their study, we believe that it is important to stress the necessity of extending the follow-up post LT4 cessation beyond eight weeks and having a second TSH measurement at least 12–16 weeks post LT4 cessation before characterizing a subject as euthyroid or hypothyroid.