Studies, data and dialogue needed to determine TSH range

The gold standard for thyroid dysfunction screening remains debatable.

Subclinical hypothyroidism is a common disorder that ranges in prevalence from 1% to 10% in the U.S. adult population, according to the American Association of Clinical Endocrinologists. For many years, experts have debated the thyroid stimulating hormone reference range.

“Because hypothyroidism is so common, the decision about whether a patient has a new diagnosis of hypothyroidism or whether a hypothyroid patient is on the correct thyroxine dose is a decision made daily by physicians worldwide,” Tom Hamilton, MD, PhD, clinical associate professor of endocrinology and metabolism at the University of Washington in Seattle, told Endocrine Today.

Tom Hamilton

The decision is one that puts endocrinologists and practicing physicians in a quandary over the conflicting opinions and evidence surrounding the upper limit of the TSH reference range. Age-specific ranges and screening pregnant women are among topics of debate.

As a result of inconsistent data and various studies published about the upper limit of the reference range, experts are at a disconnect in terms of what the limit should be, how the limit should be established and which population the limit — or limits — should apply to.

Endocrine Today spoke with Hamilton and other thyroid experts to determine what has caused the controversy, where research stands and what it might take to reach a consensus.

Considering the evidence

According to David S. Cooper, MD, director of the Division of Endocrinology at Sinai Hospital of Baltimore, “the only argument for lowering the upper limit of normal is an epidemiologic study that shows that people who have TSH levels above 2.5 mIU/L seem to have a higher rate of progression to overt hypothyroidism.”

Cooper is referring to results of the Whickham Survey, in which researchers analyzed the prevalence of thyroid disorders in a sample of 2,779 adults randomly selected from Great Britain. The researchers tracked patients for 20 years without performing any interventions or assigning any forms of treatment. Of the 1,877 known survivors, 96% participated. At follow-up, 30% of initial participants had died.

Findings showed that the hazard ratio for the development of hypothyroidism increased with age, but there was no link between hyperthyroidism and age. The odds ratio of developing hypothyroidism with raised serum TSH alone was 8 for women and 44 for men; with positive antithyroid antibodies alone, the OR was 8 for women and 25 for men; and with both raised TSH and positive antibodies, the OR of developing hypothyroidism was 38 for women and 173 for men.

David S. Cooper

The researchers suggested that increasing TSH levels above 2 mIU/L at the initial survey increased the likelihood of developing hypothyroidism.

In contrast, a study published in the Journal of Clinical Endocrinology & Metabolism by Hamilton and colleagues suggested an upper limit of 4.0 mIU/L.

To analyze TSH distribution, the researchers measured the TSH and TPOAb in 1,861 participants of the Hanford Thyroid Disease Study cohort. Of those studied, 766 participants had no evidence of thyroid disease, no positive antibodies and a normal thyroid ultrasound.

A right-skew in the TSH distribution of the thyroid disease-free participants followed an approximate lognormal distribution, according to the researchers. Using third generation immunochemiluminometric assay to measure TSH, they estimated the 97.5th percentile, percentage of participants above 2.5 mIU/L and those above 3.0 mIU/L to be 4.1 mIU/L, 20% and 10.2%.

Based on their findings, they concluded that a narrower upper reference limit of about 4.0 mIU/L would be most appropriate.

Conflicting data

In a study published in the Journal of the American Medical Association in 2004, Jacobijn Gussekloo, MD, PhD, and other researchers from the Netherlands aimed to determine whether elderly people with subclinical thyroid dysfunction should be treated. Their follow-up of the Leiden 85-Plus Study, a prospective, observational, population-based trial, tracked 87% (n=599) of a 2-year birth cohort (1912 to 1914) from age 85 through 89 years.

No adverse events were reported, despite abnormally high levels of thyrotropin. The researchers did not find a link between plasma levels of thyrotropin and free thyroxine and disability in daily life, depressive symptoms or cognitive impairment either at baseline or follow-up. In fact, the increasing levels of thyrotropin were associated with lower mortality. Based on their findings, they concluded that individuals with hypothyroidism may have a prolonged lifespan.

“If this notion is true, this lends some validity to the idea that maybe older people normally do have higher TSH levels that are not a problem for them. Perhaps it is just normal for that age group,” Cooper said.

Argument for an age-specific range

Gregory A. Brent, MD, professor of medicine and physiology, David Geffen School of Medicine at UCLA. Photo by Paula Berger

In another study published in the Journal of Clinical Endocrinology & Metabolism, Martin Surks, MD,professor in the department of medicine at Albert Einstein College of Medicine, Bronx, and Joseph Hollowell, MD, from the University of Kansas, reported that higher concentrations of serum TSH were associated with age. In accordance with evidence from Gussekloo et al, the pair argues for an age-specific TSH range to avoid overestimating the prevalence of subclinical hypothyroidism.

In their study, Surks and Hollowell analyzed data from the National Health and Nutrition Examination Survey (NHANES) III. They aimed to determine the cause of the age-related increase in 97.5 percentile for high TSH concentration (4.5 mIU/L) demonstrated in the NHANES III trial. They proposed two possibilities for the increase: a larger number of patients with subclinical hypothyroidism in groups with normal TSH distribution or age-specific shifts toward higher serum TSH levels.

They found that 10.6% of 20 to 29- year olds without thyroid disease had TSH levels higher than 2.5 mIU/L. For those aged older than 80 years, 40% had TSH greater than 2.5 mIU/L. Furthermore, Surks and Hollowell found that frequency distribution curves for TSH in the 80-or-older group with or without antibodies were displaced to higher TSH, including peak frequency.

For the 20 to 29 year olds and those 80 years and older, the 97.5 percentiles for TSH concentration were 3.56 mIU/L and 7.49 mIU/L. According to the researchers, 70% of those older patients with TSH higher than 4.5 mIU/L were within their age-specific reference range.

According to Gregory A. Brent, MD, “some of these recent data, in terms of age changes, make the issue even more difficult. Certainly a large fraction of the patients that fall into this upper TSH range are the older patients, especially older women.”

Current guidelines

At present, the typical statistical reference range used in U.S. laboratories is between 0.5 mIU/L and 5.0 mIU/L, according to Hossein Gharib, MD, professor of medicine at, the Mayo Clinic College of Medicine. However, AACE guidelines suggest a TSH reference range of 0.3 mIU/L to 3.0 mIU/L.

According to Cooper, “If the upper limit of the TSH reference range is lowered, you are going to call a lot of people who are probably normal ‘subclinically hypothyroid,’ and you are going to treat them. You can see from the data that at least in older people, there is no evidence that treatment is going to help them. In fact, it may even hurt those people in their 80s.”

According to AACE, rates of progression to overt hypothyroidism are highest among patients with TSH levels above 10 mIU/L and those with levels between 5 mIU/L and 10 mIU/L with goiter and/or positive TPOAb. Therefore, these patients should be treated.

“If 3 mIU/L is the upper limit of normal, then a TSH of 6 mIU/L is definitely abnormal. However, if you consider the upper limit of normal to be 5 mIU/L, then a TSH of 6 mIU/L is only marginally abnormal. So, there is some controversy as to what to do with a mildly abnormal TSH,” Gharib said.

TSH and pregnant woman

The added stress placed on a thyroid during pregnancy is well established, and several studies have addressed the issue in hopes of determining an appropriate upper limit for women both before and during pregnancy.

In a study published in the Journal of Clinical Endocrinology & Metabolism, Roberto Negro, MD, and researchers from Italy and Brunei Darussalam, examined 984 euthyroid pregnant women. The researchers aimed to determine the effects of autoimmune thyroid disease on obstetrical complications, along with the efficacy of levothyroxine treatment in these women, 11.7% of whom were thyroid peroxidase antibody positive. They divided the TPOAb+ women into two groups: those treated with LT4 (n=57) and those who were not (n=58). The remaining population of those who were not TPOAb+ (n=869) served as the control group.

Compared with the treated and negative groups, the untreated group had higher TSH values through gestation and had lower values of free T4 after 30 weeks and after parturition. The untreated population had a higher rate of miscarriage, compared with those in the treated and TPOAb- groups (13.8% in the untreated group vs. 3.5% in the treated and 2.4% in the TPOAb- groups). The rate of premature deliveries was also higher in the untreated group, compared with those in the treated and TPOAb- groups (22.4% in the untreated group vs. 7% in the treated and 8.2% in the TPOAb- groups).

The researchers determined that euthyroid pregnant women with positive thyroid antibodies developed impaired thyroid function and had an increased rate of miscarriage and premature delivery.

“It is a very important idea that women with even the mildest degree of thyroid disease should probably be treated. This gets into the whole issue of whether women who are pregnant, or those who wish to be, should be screened for thyroid disease. This paper by Negro et al is so powerful that many physicians are screening women,” Cooper said.

The Endocrine Society, along with some experts, suggests the implementation of a trimester-specific range. Both preconception and during the first trimester, the upper limit should be <2.5 mIU/L. During the second and third trimesters, the limit should be raised to 3 mIU/L.

Deciding when to treat

Due to a lack of data, specifically data regarding who to treat and what effects treatment may have on certain populations, experts and practicing physicians remain torn about the limit and deciding when to treat for hypothyroidism.

Hossein Gharib

According to Brent, long-term outcome studies that demonstrate the implications of lowering the normal range are lacking. Studies showing the benefits of treating such patients, especially those in the 3.5 mIU/L to 5 mIU/L range, are also needed.

“As endocrinologists, we are aware of target ranges for patients who are not diagnosed and those who are on levothyroxin, which most people believe to be below 2 mIU/L or 2.5 mIU/L. However, confusion arises as to whether everyone with a TSH above 3 mIU/L should be diagnosed as hypothyroid,” Brent told Endocrine Today.

The experts interviewed by Endocrine Today agree that additional data, discourse and evidence are needed in order to reach a more widely accepted range, and some do not believe a consensus will be reached at all.

“A consensus may not be reached; the groups are polarized on this subject,” Jerome Hershman, MD, distinguished professor of medicine at the David Geffen School of Medicine at the University of California at Los Angeles, told Endocrine Today.

Cooper proposed a large, randomized controlled trial to analyze the use of thyroid hormone vs. placebo in 10,000 patients with TSH levels between 2.5 mIU/L and 4 mIU/L. If such a study could demonstrate the benefits of thyroid treatment — the patients felt better, their cholesterol levels improved, their mood was better or they lived longer — it may help build a case for treatment.

“There is certainly a need for more data, more studies with larger populations and larger data that come from prospective analyses to shed light on this subject. But, until that happens, we need to look at the information currently available and use expert opinions. Though there is disagreement, we need to look at the guidelines and consider physicians’ personal experience and preference. In any decision making, you have to look at all of those factors,” Gharib suggested.

In practice

“A key message to remember is that while endocrinologists will continue to debate the most appropriate upper reference limit for TSH and whether thyroid hormone replacement produces benefit in subclinical hypothyroidism, they make up a minority of the physicians who make daily decisions about whether a patient is hypothyroid,” Hamilton said.

Despite the ongoing debate, some experts agree that until there is more evidence to support either argument, physicians should remember to treat patients individually and not relative to a reference range.

“If you follow the patient carefully it is, in itself, an appropriate management schedule. If you decide to treat a patient, follow them and make sure you do not over- or under-treat that individual. Essentially, following your patient is paramount,” Gharib said.

In terms of treating patients based on their individual signs, symptoms and test results, the experts interviewed by Endocrine Today agreed that the decision is one that should be made by the physician based on personal preference and what they believe is best for their practice and their patient.

“Physicians need to do what they think is best for their patients, that is the bottom line,” Cooper said.

“The evidence is very strong for treating everyone with a TSH above 10 mIU/L, and it comes down to an individual physician’s judgment in collaboration with their patient. It is just like treating patients for high cholesterol: if it is elevated, you treat them. It is all preventive medicine,” said Hershman, who is also a member of the Endocrine Today Editorial Board.

“While there are many factors other than TSH that should influence the decision to start thyroid hormone treatment or change a person’s thyroid dose, it is the upper reference limit that many physicians in practice rely on. Consequently, it is important that we try to determine the most appropriate upper reference limit for TSH and also to educate health care professionals about the many clinical factors that should be considered, in addition to TSH, before making a diagnosis or initiating therapy,” said Hamilton. – by Stacey L. Adams

How should the TSH reference range be determined?

For more information:

• AACE thyroid guidelines. Endocr Pract. 2002;8:457-469.
• Abalovich M, Amino N, Barbour LA, et al. Clinical practice guideline: management of thyroid dysfunction during pregnancy and postpartum: an Endocrine Society clinical practice guideline. J Clin Endocrinol Metab. 2007;92: s1-s47.
• Gussekloo J, van Exel E, de Craen AJM et al. Thyroid status, disability and cognitive function and survival in old age. JAMA. 2004;292:2591-2599.
• Hamilton TE, Davis S, Onstad L et al. Thyrotropin levels in a population with no clinical, autoantibody, or ultrasonographic evidence of thyroid disease: implications for the diagnosis of subclinical hypothroidism. J Clin Endocrinol Metab. 2008;93:1224-1230.
• Negro R, Formoso G, Mangieri T, et al. Levothyroxine treatment in euthyroid pregnant women with autoimmune thyroid disease: effects on obstetrical complications. J Clin Endocrinol Metab. 2006;91:2587-2591.
• Surks MI, Hollowell JG. Age-specific distribution of serum thyrotropin and antithyroid antibodies in the U.S. population: implications for the prevalence of subclinical hypothyroidism. J Clin Endocrinol Metab. 2007;92:4575-4582.
• Vanderpump MP, Tunbridge WM, French JM et al. The incidence of thyroid disorders in the community: a twenty-year follow-up of the Whickham Survey. Clin Endocrinol (Oxf).1995;43:55-68.