November 10, 2005

In This Issue . . .

Tightening the TSH Range: What Should be Considered Normal?
Julie L McDowell

When the National Academy of Clinical Biochemistry (NACB) published thyroid disease diagnosis and monitoring guidelines in 2002, editors Laurence M. Demers, PhD, FACB, and Carole A. Spencer, PhD, FACB, fueled an ongoing debate when they predicted the likelihood of a future downward shift in the upper limit of the normal serum thyroid stimulating hormone (TSH) reference range from approximately 4.5 to 2.5 mIU/L. This forward thinking assertion was based on data from two long term studies—a subgroup of the third National Health and Nutrition Examination Survey (NHANES III) and a 20-year follow-up to the Whickham survey—that found a high proportion of people with a TSH above 2.5 mIU/L have an increased risk of developing overt hypothyroidism. However, many endocrinologists believe that it’s premature to change the parameters of the normal reference range. In this issue, Strategies examines both sides of this debate, and the clinical lab’s role in this midst of these arguments.

Currently, in the appropriate clinical setting, physicians and laboratorians utilize TSH level measurements in the diagnosis of clinical thyroid disease, based on the current normal reference range of 0.45–4.0 or 0.45–4.5 mIU/L, depending on the laboratory's perspective. When patients present TSH levels above this upper normal range, clinicians will often treat them with levothyroxine. But in September, Leonard Wartofsky, MD, chairman of Washington Hospital Center’s Department of Medicine and The Endocrine Society’s (TES) 2006 President, reignited a common argument in the endocrinology community when he published a Clinical Controversies article in the Journal of Clinical Endocrinology & Metabolism saying that the time has come to narrow the TSH reference range by lowering the upper limit. This upper limit shift is based on data from NACB’s laboratory guidelines for thyroid disease that 95% of normal individuals have TSH levels between 0.4 and 2.5 mIU/L.

“Advances over the past two decades in both the sensitivity and precision of assays for TSH and in our understanding and definition of mild thyroid disease (subclinical hypothyroidism and subclinical hyperthyroidism) have helped to fuel a controversy regarding what constitutes the normal range for TSH,” noted Wartofsky and his co-author, Richard A. Dickey, MD, in the JCEM article. “The issue is important because it relates to whether to screen for thyroid disease and what to do when a patient is found to have mild abnormalities in TSH, whether by screening or otherwise. Thus, a more precisely defined reference range will allow the detection of patients with mild thyroid dysfunction who could benefit from therapy or at least closer follow-up.”

Risks of Over and Undertreating

Other data used by those in favor of lowering the upper TSH limit comes from NHANES III. In addition to other studies on thyroid dysfunction, researchers studied the TSH levels of a subgroup of the population made up of 13,334 patients who were determined to have no laboratory evidence of overt hypothyroidism or hyperthyroidism and were clear of other factors that might affect TSH levels, including pregnancy or the presence of antithyroid antibodies. Out of this reference group, a significant number of patients had serum levels lower than 2.5 mIU/L, although the exact figure is open to debate: Wartofsky’s paper indicates the figure is close to 95%, while 86% is the figure identified by Martin I. Surks, MD, Division of Endocrinology and Metabolism, Montefiore Medical Center (Bronx, N.Y.), who wrote an article in the same issue of JCEM arguing in favor of maintaining the current TSH reference range. This figure was provided by one of the senior authors of the NHANES III study, Surks explained to Strategies.

Lowering the upper limit of the TSH reference range should only be done after intense debate and additional critical studies, said Surks. Changes in diagnostic criteria for diabetes, hypertension, and hypercholesterolemia were only instituted after a multitude of critical studies and discussion. “In each of these instances, the guiding principle for change was the firm knowledge that interventions employing these guidelines would have important health benefits that outweigh potential risks,” he wrote in JCEM. While clinicians like Wartofsky argue about the risks of underdiagnosis and subsequent undertreatment of levothyroxine, Surks insists that levothyroxine intervention results can result in overtreating more than 20% of patients and has serious risks. Adverse outcome risks associated with this treatment include an increased rate of bone mineral loss, as well as an increased risk of cardiovascular mortality and atrial fibrillation. In fact, Surks said that approximately 500,000 people would be at an increased risk for these adverse outcomes of levothyroxine treatment if the upper limit were increased to 2.5 mIU/L.

The Dilemna for Lab Directors

Despite disagreeing with Wartofsky’s assertion, Surks recommends that lab directors educate themselves on both sides of the argument, while conceding that it is a tough issue to negotiate given the strength of the arguments on both sides. “It’s very hard for laboratorians, but the clinical data is simply not there to lower the upper limit,” he said.

The key issue in this debate is who should be treated, and no one is suggesting that every patient with a minimally elevated TSH, regardless of the normal range chosen, should automatically be treated with levothyroxine, pointed out one of Wartofsky’s colleagues, Kenneth D. Burman, MD, Director of Endocrinology at the Washington Hospital Center. “The decision to treat with levothyroxine should be a decision mutually agreed upon by the physician and patient and should be based on TSH values as well as the clinical context.”

In January 2004, a consensus panel, which included members from the three primary endocrine societies—the American Association of Clinical Endocrinologists, TES, and the American Thyroid Association—recommended only treating patients with levothyroxine who were pregnant, planning a pregnancy, or had TSH levels below 0.1 mIU/L or above 10 mIU/L (Strategies, Feb. 10, 2005,
). There was dissension among this group, however, and in January 2005, another panel including AACE, ATA, and TES members published clinical guidelines in JCEM recommending routine treatment of subclinical hypothyroidism patients with TSH levels between 4.5 and 10 mIU/L.

“Individual circumstances might, on occasion, suggest that patients with an elevated TSH, but still lower than 10 mIU/L, be treated on a case by case analysis,” said Burman, who also pointed out that the TSH range on most commercial lab assays is about 0.5 to 5.5 mIU/L, which is almost universally discounted by the endocrinology community.

If the reference range is tightened, then it’s likely that testing will identify more people who have subclinical forms of the disease, which are considered mild and present few or no symptoms. In addition, clinicians and laboratorians will probably have to contend with an increased number of false positive results, said Laurence M. Demers, PhD, Professor of Pathology and Medicine at the Pennsylvania State University College of Medicine’s M.S. Hershey Medical Center (Hershey, Pa.) and an editor of the NACB guidelines. Other factors might be pushing the TSH level up—many drugs taken for other disorders can have this effect. Nevertheless, if these patients are considered abnormal, clinicians may begin treating them for thyroid diseases, which puts them at risk for adverse outcomes.

“Hypothyroidism is a disorder of the elderly, so you tend to see age related differences in TSH levels—the older you are, the higher it seems to be that the baseline goes, so you end up with many more elderly patients in a population of those false-positives, and if you treat them, you could be causing harm,” he added.

For these reason, Demers cautioned about hastening the reference range shift that he mentioned in the NACB guideline. “The studies that have to be done need to be longitudinal,” he said. “The definitive studies to really define the reference range in somebody who is absolutely clean of thyroid disease would have to be done, and these longitudinal studies have not been done to that extent.”

For more information:

  • Surks and Wartofsky’s articles can be found online on JCEM’s Web site:
  • To read more about the debate over subclinical thyroid disease diagnostic guidelines, see the July 2005 issue of Clinical Laboratory News online at

    Julie McDowell is the Editor of Strategies. She can be reached by email.

  • Page Access: