Reply: A New Age for Recombinant Human Thyroid-Stimulating Hormone?

REPLY: We thank the reader for the constructive criticism of our recent article describing similar short-term clinical outcomes using either recombinant human thyroid-stimulating hormone (rhTSH) or traditional thyroid hormone withdrawal preparation for radioactive iodine remnant ablation (1). Like the reader, we agree that a short period of thyroid hormone withdrawal followed by moderate administered activities of radioactive iodine is successful in achieving high rates of remnant ablation (in fact, in our study this was the gold standard to which patients prepared with rhTSH were compared). Furthermore, by minimizing the period of thyroid hormone withdrawal, the symptoms of hypothyroid withdrawal (although still present) are certainly less clinically significant than are prolonged periods of thyroid hormone withdrawal before therapy.

Rather than concluding that rhTSH preparation was superior to thyroid hormone withdrawal based on nonrandomized, retrospective data with short-term follow-up, we were careful to conclude that rhTSH-assisted ablation is simply a “safe and effective option” to thyroid hormone withdrawal. We continued by stating that the “choice of preparation for [radioactive iodine remnant ablation] ([thyroid hormone withdrawal] vs. rhTSH) should be made on the basis of patient's quality of life and the ‘cost’ of hypothyroidism versus the cost of rhTSH” (1).

As with any retrospective study there are significant limitations, which we carefully pointed out in the discussion. Even though the differences in age (44 ± 15 y for thyroid hormone withdrawal vs. 48 ± 15 y for rhTSH, P = 0.03) and administered activity (3,811-MBq [103-mCi] median for thyroid hormone withdrawal vs. 4,033-MBq [109-mCi] median for rhTSH, P = 0.01) were statistically significant, we do not consider these to be clinically significant differences. It seems unlikely that these small differences, although statistically significant, would have a meaningful impact on the clinical outcomes. Oftentimes in clinical studies we detect differences that are statistically significant that have little if any clinical meaning. Therefore, it is important to put statistically significant differences in a proper clinical context.

Following standard practice, we reported the overall P values for the contingency tables presented in the article (Tables 5 and 6). Although it is certainly possible to calculate specific P values for each category, these individual values still need to be interpreted in light of the overall contingency analysis. For example, in Table 5, although there were no significant differences overall between rhTSH and thyroid hormone withdrawal with respect to the clinical outcomes, analysis of the individual category described as “no clinical evidence of disease” (vs. all other outcomes) does demonstrate a statistical significance (P = 0.02) between thyroid hormone withdrawal and rhTSH preparation, whereas each of the other individual categories demonstrates no significant individual differences (vs. all other outcomes). In our opinion, one must be careful in attributing significance to individual categories when the overall contingency analysis does not find significant differences. Therefore, we chose not to emphasize the individual category analysis and simply reported the overall P value for the contingency table.

We have similar concerns about reporting specific categoric P values for Table 6, although in this case the overall P value for the contingency analysis is significant and therefore individual category analysis seems more reasonable. To that end, comparing the category of “no clinical evidence of disease” with all other outcomes demonstrates a P value of 0.002, and comparing the category of “persistent disease” with all other outcomes reveals a P value of 0.02, indicating a statistically significant difference within each of those categories when thyroid hormone withdrawal is compared with rhTSH. Therefore, based on both the initial contingency table analysis and this additional individual-category analysis, we continue to conclude that “when the definition of no clinical evidence of disease included a suppressed thyroglobulin level of less than 1 ng/mL and a stimulated thyroglobulin level of less than 2 ng/mL, rhTSH-assisted [radioiodine remnant ablation] was associated with significantly higher rates of no clinical evidence of disease…and significantly lower rates of persistent disease…than was [radioiodine remnant ablation] after [thyroid hormone withdrawal]” (1).

We thank the reader for pointing out the typographic error in Table 7: the total excluding distant metastases at diagnosis should be 371 (rather than 394). The remainder of the data in this table are correct.

In both Table 5 and Table 6, we included as a specific category “thyroid bed uptake only,” defined as persistent uptake in the thyroid bed with no structural evidence of persistent disease and stimulated thyroglobulin values less than 10 ng/mL. This is always a difficult group to categorize. Some of these patients probably have persistent disease, whereas others likely have just normal thyroid remnants. Therefore, we could not with confidence classify them as either no clinical evidence of disease or persistent disease. Since it did not seem reasonable to exclude this group from analysis, we included them as a separate clinical endpoint as we have done in our previous studies (2,3). As with any retrospective study, not all patients received identical follow-up (neck ultrasound, CT scans, MRI scans, ¹⁸F-FDG PET scans, or radioactive iodine scans); therefore, we included all patients regardless of the extent of follow-up studies to provide our best disease status classification for each individual patient.

In conclusion, we view rhTSH stimulation as a safe and effective alternative to traditional thyroid hormone withdrawal preparation for routine radioactive iodine remnant ablation. Additional studies are needed to define the minimal administered activity of radioactive iodine that can achieve both successful remnant ablation and acceptable long-term clinical outcomes.