Thyroid Stimulating Hormone Stability in Patients Prescribed Synthetic or Desiccated Thyroid Products: A Retrospective Study

I am concerned that this study, publicized in Science, and discussed in physician forums online, may be misleading, either accidentally or intentionally.

A large healthcare system such as Kaiser Permanente has thousands of patients taking thyroid preparations prescribed by many types of physicians. In this population are patients with complete absence of endogenous thyroid function, and those with marginal or even questionable need for a thyroid pill. An example of the former group is a 50 year old woman who became hypothyroid from autoimmune thyroiditis, with pretreatment TSH of 60 and a current full replacement dose of 137 mcg. An example of the latter group, whose requirement was more marginal is a 50 year old woman who complained of slow metabolism to her primary doctor, who found a TSH of 6 and prescribed 25 or 50 mcg of levothyroxine. Neither type of patient is rare, but the second woman may well stop taking her thyroid pills in the future and her TSH will have returned to normal; her "TSH stability" will be excellent. I looked in vain for average current dose and pre-treatment TSH in the 25 characteristics carefully matched and reported by these authors. Almost every one of the 25 listed characteristics is less important to TSH stability than the degree of hypothyroidism. Why would the authors not include the current average doses and the pre-treatment TSH levels? It is difficult to imagine an honest motive for concealing a substantial difference in one of these two parameters when so many marginally relevant parameters have been listed.

I was even more astonished to learn that they deliberately excluded nearly all patients who carried diagnoses of all the common forms of severe hypothyroidism (post-thyroidectomy, post-Graves, post-radioiodine, Hashimoto's, hypopituitarism). Why would they do this? Why would they want to compare stability in a patient population who needed only partial replacement and may have included many who needed nothing at all (at least measured by TSH)? Non-endocrinologists, as I presume your editorial writers were, may not immediately recognize what a suboptimal population this would be for comparing two thyroid replacement treatments. I suspect the authors did not discuss their study design with an endocrinologist, nor did the journal editor ask an endocrinologist to review it for publication.

I keep trying to understand what i am missing. Would the authors be willing to respond as to why they excluded most of the patients in their health system who truly needed complete thyroid replacement? It is hard for me to believe them so naïve as to not understand that a real comparative test of treatment equivalence would be among patients with a need for full replacement, not a population that deliberately excluded most of those patients. If my assessment is correct, this paper has been misrepresented in the accompanying editorial and in media reports like the one in Science.

In other words, if the patient groups had substantially different pretreatment TSH levels, omission of that fact is dishonest and you cannot claim stability is comparable. Alternatively, if the patient groups had similar pretreatment TSH levels that were only mildly elevated and are on low doses, they may be stable because of continuing endogenous thyroid function, but such a study cannot be generalized to a population with real, full-replacement hypothyroidism. Who can object to "patient centered care"? Certainly not me. Like every single other endocrinologist who I have ever heard express a public opinion, I put most of my hypothyroid patients on levothyroxine and most do well, but if someone asks for desiccated thyroid, most of us prescribe it.

The letter is already too long, but if the pharmacist authors and the writers of the accompanying editorial really don't in good faith understand how choosing the Hypothyroid code but excluding those also coded for the major causes of full hypothyroidism selects a population of people who are only partly hypothyroid, if at all, I would be happy to explain it. If I have horribly misunderstood the patients in this study, your authors can provide the missing data and I will retract this letter (or they can embarrass me publicly and I will publicly apologize). However, if my surmises are correct, you owe your readers a retraction with apology, or at the very minimum, an explanation that the study may not be relevant to adult patients with real absence of thyroid function. And Drs Scheiderhan and Zick may want to find another study on which to lecture us on "patient-centered care".