In a preview of his on-demand CEU 2020 session, “When It’s Not Your Thyroid,” James V. Hennessey, MD, presents a series of questions about thyroid patients, what their symptoms mean, as well as what they don’t mean.
Sir Arthur Conan Doyle once said, “It is a capital mistake to theorize before one has data.” This is perhaps nowhere truer than in the field of medicine, as an upcoming CEU program will demonstrate.
The Endocrine Society’s 2020 Clinical Endocrinology Update (CEU) meeting being delivered virtually September 10 – 12, 2020, will feature a presentation by James V. Hennessey, MD, FACP, ECNU, associate professor of medicine at Harvard Medical School and director of clinical endocrinology at Beth Israel Deaconess Medical Center in Boston, Mass. His session, called “When It’s Not Your Thyroid,” will be live-streamed (his on-demand program “Drug Effects on the Thyroid” was written about in the special CEU issue of Endocrine News). Hennessey is a master of presenting in a patient case format, dropping clues as the case unfolds and weaving in evidence and supporting data from relevant papers as he goes.
The effect is Holmesian, with Hennessey playing lead detective and the audience his de facto Watson.
“When It’s Not Your Thyroid” will be an exercise in what Hennessey calls “endocriminology,” or solving the mystery of one condition masquerading as another. Attendees will be asked to engage with the case, as Hennessey will stop along the way to ask pointed questions and get them thinking. This Meet the Detective session is going to be a lot of fun!
Three objectives provide the backbone of this talk: to contemplate the evidence underlying the association of symptoms consistent with the presence of hypothyroidism and our ability to biochemically confirm that diagnosis; to understand what is happening in primary care offices when it comes to ordering thyroid function tests (TFTs); and to realize who is being started on thyroid hormone by their primary care doctor before seeing an endocrinologist.
Regarding the first objective, “all guidelines, including those of the Endocrine Society,” Hennessey explains, “say don’t treat symptoms, only treat biochemically confirmed hypothyroidism. That’s where sometimes that first step is skipped and we’re off to the races.” This can happen, he explains, because laboratory results can be misunderstood. A thyroid-stimulating hormone (TSH) value can appear suspicious to a primary care physician simply because the lab tech deemed it high and marked it conspicuously. Or some holistic practitioners have been known to simply talk to patients rather than relying on biochemical confirmation before treating. In either scenario, the problem then compounds — the patient gets labeled as “hypothyroid” and instantly feels worse.
“It turns out that patients being labeled hypothyroid tend to complain of symptoms consistent with the presence of hypothyroidism, perhaps because they read things on websites that are not Endocrine Society supported about how they should feel,” Hennessey says. Their quality of life deteriorates from there because being aware of the diagnosis meant that they focus on their symptoms — the labeling effect. “In a broad epidemiological study, labeling patients hypothyroid by giving them hypothyroid medication automatically led to a decrement in the way they felt,” he says. “In comparing those who are actually hypothyroid with those labeled as such, it was better to be actually hypothyroid but unaware of it than to have normal thyroid function but be treated as hypothyroid.”
“We have to make sure the diagnosis is correct even if it takes some time. Then maybe we’re doing something good for our patients. At every turn in this case, the diagnosis of hypothyroidism could be made — should it be?” – James V. Hennessey, MD, FACP, ECNU, associate professor of medicine, Harvard Medical School; director, clinical endocrinology. Beth Israel Deaconess Medical Center, Boston, Mass.
Beyond the effect of labeling, the elephant in the room here is that treatment for hypothyroidism in a euthyroid patient is not going to make the patient better. “That,” says Hennessey, is the basic principle of ‘nonthyroidal hypothyroidism’ — if a patient was not hypothyroid to begin with, why would a thyroid remedy provide symptom relief? Even an endocrinologist can’t make those go away if the diagnosis was incorrect.”
The Case of Nonthyroidal Hypothyroidism
Hennessey will then get to the case, that of a 71-year-old woman who sees her primary care doctor for a three-month history of being tired and feeling depressed. “Everyone in the audience will recognize tiredness and depression as being symptoms that will lead a primary care doctor to do TFTs,” he says.
Question #1. “The first question is, how likely is it that she’s actually hypothyroid?” The answer to this question, supported by data from three pivotal studies, is, basically, unknowable without checking TFTs.
In The Colorado Thyroid Disease Prevalence Study (Canaris GJ, et al. Arch Intern Med. 2000;160:526–534), 26% of euthyroid patients reported symptoms like drier skin, poorer memory, slower thinking, weaker muscles or muscle cramps, fatigue, feeling cold, puffy eyes, a deeper or hoarse voice, and constipation as compared to 28% of patients with elevated TSH levels. Although the difference in values was statistically significant, the rates of complaints are proportionally quite similar.
In another study by the same group (Canaris GJ, et al. J Gen Intern Med. 1997;12:544–550), 60% of patients with normal thyroid function had symptoms consistent with the presence of hypothyroidism, 63% with subclinical hypothyroidism had such symptoms, and 73% with overt hypothyroidism had them. “The symptoms don’t tell you the difference if they’re absent,” Hennessey says.
In the third study cited, “Hypothyroid Symptoms Fail to Predict Thyroid Insufficiency in Old People: A Population-Based Case-Control Study” (Carle A, et al. Am J Med. 2016;129:1082–1092), even as many as eight or more symptoms consistent with the presence of hypothyroidism did not predict that patients older than age 60 years were actually hypothyroid because aging people tend to have such symptoms.
“Symptoms don’t reliably predict the presence of hypothyroidism, so don’t treat your patients with thyroid medications unless they are biochemically confirmed,” Hennessey says. (Or, to put a spin on Conan Doyle’s quote: “It is a capital mistake to treat before one has biochemical data.)
Question #2. “After I’ve convinced the audience that symptoms need to be confirmed, I will ask, how likely is it that she’s going to have TFTs ordered?” Although this one seems like an obvious one based on the fact that her symptoms alone are inconclusive, Hennessey cites a British study (Bould H, et al. Family Practice. 2012;29:163–167) that looked at the frequency with which patients presenting with depressive symptoms were referred for TFTs and found that those referred for TFTs had high rates of psychological distress, but there was no correlation of symptoms with TFTs.
Furthermore, mild TSH elevations sometimes resulted in thyroxine treatment assuming that the patient’s symptoms were due to hypothyroidism.
“The problem with this is, if it’s not hypothyroidism, the search for alternative explanations immediately stops,” Hennessey says. “If a patient is depressed or has lupus or something else that is causing these symptoms, they will not be evaluated further. The diagnostic search ends, and maybe the patients are being shortchanged.”
Back to the case, the patient is tested one afternoon and is found to have a TSH of 4.6, which is considered mildly elevated over the presumed upper limit of normal (4.12).
Question #3. “How likely is it that she’ll get treated with thyroid hormone based on one mildly elevated TSH?” According to Hennessey, “the underlying admonition is, who’s going to label their patient as hypothyroid and get them started on the hypothyroid trail?” “Falling Threshold for Treatment of Borderline Elevated Thyrotropin Levels—Balancing Benefits and Risks
Evidence from a Large Community-Based Study” (Taylor PN, et al. JAMA. 2014;174:32–39) revealed that vast majority of patients receiving treatment had TSH levels 4–10 (subclinical hypothyroidism). “This raises the questions of whether it’s appropriate to treat them at all and whether they respond,” Hennessey says. Moreover, some being treated were biochemically euthyroid and were treated based on symptoms, thus labeling them as hypothyroid.
One downside to this inappropriate treatment is the mounting cost for the patient and the healthcare system. “That is unacceptable if the treatment is for the wrong diagnosis, and if the patient wasn’t likely to respond in the first place,” Hennessey says. “Going back to evidence-based medicine, I like to prescribe things for my patients that I’m convinced are in their best interest, and they’ll feel better.”
Question #4. “Should the patient be treated now to relieve her symptoms or for inevitable overt hypothyroidism?” Here, Hennessey cites some prospective studies of subclinical hypothyroidism and what happens when it’s untreated (Vanderpump, MPJ. Clin Endocrinol. 1995;43:55–68; Huber, et al. JCEM. 2002;87:3221–3226; and Somwaru LL, et al. JCEM. 2012; 97:1962–1969).
In effect, the presence of subclinical hypothyroidism had no impact on longevity in patients over age 65 years, and 33% normalized without treatment while most remained subclinically hypothyroid. “The worry is of thyroid failure to come,” Hennessey says, “but, it turns out, if you don’t treat people with thyroid hormone nothing really much happens, and you can always repeat the TFT and start treatment then if it’s abnormal. It’s not a high-risk situation.” Interestingly, one of the biggest risk factors for developing overt hypothyroidism was having been treated with levothyroxine, thus being labeled hypothyroid.
TSH values higher than 10 did increase the risk of persistent subclinical hypothyroidism and levothyroxine administration and slightly increased the risk to progress to overt hypothyroidism (somewhat). Also of note is that using a single TSH value led to a 40% misdiagnosis rate. These findings gave rise to guidelines that establish that at least two TFTs must be done before intervention is warranted (Gencer B, et al. Am J Med. 2020, doi.org/10.1016/j.amjmed.2020/01.018).
Accordingly, a second TFT is done on the case patient about a year later at 8:00 am and comes back with a TSH of 5.6.
Question #5. “Are these TSH values actually abnormal in a 71-year-old woman?” The studies Hennessey cites here make clear that TSH ranges must be stratified by age bracket (Surks MI and Hollowell JG. JCEM. 2007;92:4575–4582; Bremmer AP, et al. JCEM. 2012;97:1554–1562; and Samuels MH, et al. Thyroid. 2017;26:1185–1194). Among those age 20–29 years, the upper limit of normal is about 3.56, whereas in those older than age 80 years, it’s 7.5 — a 4-unit difference!
Pertinent to the case, the patient’s two TSH levels are actually normal because in the 70 – 79-year age bracket, 5.9 is the upper limit of normal. (The “presumed” upper limit of normal, 4.12, was an average of all subjects from the NHANES study who had TFT levels taken.) “Stratification makes it clear that you have to take age into account, and you have to develop an age-adjusted TSH,” Hennessey says.
Question #6. “Is there good indication of deterioration of function here?” Probably not, says Hennessey: “Morning TSH is a unit and a half higher than in late afternoon, so this is just sampling variation and a good example of diurnal rhythm. So, if you thought this patient who went from 4.6 to 5.6 was just circling the drain and about to overtly fail, take a look and see when the TFTs were drawn,” (according to Ehrenkranz J, et al. Thyroid. 2015;25(8):954–961).
So, when it’s not your thyroid, what else could it be? Rather than give away any big reveals from Hennessey’s presentation here, guidelines put out by the British Thyroid Association Executive Committee offer clues in their “Management of primary hypothyroidism” statement. Possible causes of persistent symptoms in euthyroid patients range from lifestyle and nutritional factors to drugs to other diseases and conditions.
“In a broad epidemiological study, labeling patients hypothyroid by giving them hypothyroid medication automatically led to a decrement in the way they felt. In comparing those who are actually hypothyroid with those labeled as such, it was better to be actually hypothyroid but unaware of it than to have normal thyroid function but be treated as hypothyroid.”- James V. Hennessey, MD, FACP, ECNU, associate professor of medicine, Harvard Medical School; director, clinical endocrinology. Beth Israel Deaconess Medical Center, Boston, Mass.
Finally, when is it actually your thyroid? Hennessey has a solution for that problem. He proposes a “hypothyroidism triad” to establish a diagnosis that bears similarities to the Whipple triad for hypoglycemia: 1. symptoms consistent with the presence of hypothyroidism (many of which are also on the list in the Whipple triad); 2. symptoms in the presence of unequivocal evidence of biochemical hypothyroidism; and 3. resolution of those symptoms with the administration of levothyroxine. “If we skip one of those steps, maybe the diagnosis is wrong,” Hennessey says. “We have to make sure the diagnosis is correct even if it takes some time. Then maybe we’re doing something good for our patients. At every turn in this case, the diagnosis of hypothyroidism could be made — should it be?”
— Horvath is a freelance writer based in Baltimore, Md. In the July issue, she wrote about a podcast by two neuroendocrinologists that detailed treating COVID-19 patients who also have underlying pituitary conditions.