New Study Suggests Thyroid Testing and Treatment Are Overdone

Minor changes may spur overtreatment due to a new definition of clinical levels

With James Hennessey MD and Christopher Symonds MD

There is debate in the medical community about how often thyroid tests should be given.

If you change the lower range of what constitutes hypothyroidism, the number of tests done and prescriptions for levothyroxine written will rise. A new study has found that the rates of those changes are not equal, and this is probably a good thing.

A changing definition of hypothyroidism

The study, done through the University of Calgary, looked at rates of prescriptions among the population, and the number of tests for thyroid secreting hormone (TSH) that were conducted before and after the lower range for what was considered hypothyroidism was change from 6 mlU/L to 4 mlU/L. The change was in line with recommendations in the United States designed to corral more subclinical cases of thyroid disorder, especially where there have been decreases in TSH but not enough to push the person into clinical definitions of hypothyroidism.

The rationale for the changes come from some studies that suggested doctors are failing some patients by not treating them when they were off by small amounts. However, the results of these studies are mixed and many have not been replicated. Further, some of those who have initial testing within the new limits eventually normalize without any medication. Some end up over-treated to the point of hyperthyroidism, which has its own dangers.

The changes in definition led to a large number of questions and referrals to the endocrinology clinic at the university, says author Christopher Symonds, MD.


  • There were as many as 5000 additional tests per month.
  • In the three months prior to the change, there were nearly 200,000 tests done, just over 6600 of which had results between 6 and 10 mlU/L – a prevalence of 3.3%. Afterwards, the prevalence went up to 9.1%.
  • Levothyroixine prescriptions increased by 25.3%, from 3.24 to 4.06 per 1000 population.
  • The rates of new prescriptions changed little over the first 3 years of the study, but jumped in the last year – the year of changing recommendations. 

Symonds notes that there was already a high volume of thyroid testing being done, but there was a lack of good literature that more testing and treatment would lead to better outcomes.

Expert discussion

The results of the study are hardly surprising, says James Hennessey, MD, an endocrinologist and professor at Harvard College. “They noticed that by changing the lower limit, they saw an increase in diagnoses of hypothyroidism. More patients with newly abnormal TSH tests see repeat tests, as well, so the number of tests went up threefold.”

But this is an emphasis on laboratory results, he says, not whether a patient has symptoms of hypothyroidism.

It was gratifying to see that the increased number of patients identified as having thyroid disorders didn’t result in an equal number of new prescriptions or changed prescriptions for levothyroxine, says Hennessey. “That means some clinicians looked at more than a lab report, but it is still concerning that some of these treatments undoubtedly occurred in patients with no outward symptoms of a thyroid disorder” such as being cold, gaining weight, being constipated or having new hair loss.

There is an ongoing question of why doctors are doing a thyroid test to begin with. “There is some frequency that is based on the idea that one is taking blood annually, so check it,” Hennessey says. “But there are no major endocrinology organizations that endorse the approach of a yearly test for everyone.”

Unless a patient has a risk factor – such as a family history of thyroid disease – there is little reason to test so often. “It isn’t like a cholesterol or other routine test, or it shouldn’t be,” Hennessey says. “But it’s being treated like that now for people who are asymptomatic. Risk factors and symptoms should be what drives testing and prescriptions.”

Even if the lab tests show a patient in the new range, unless the patient has symptoms, Hennessey is reluctant to suggest treatment. Part of the reason relates to the above-mentioned studies that show some resolution without medication of those abnormal TSH levels. The changes in clinical norms, by themselves, should not result in a prescription. “That would mean a patient who last week was considered normal and felt fine, but this week isn’t any longer in the normal range, and can lead to patients wanting treatment they do not need.”

There are also patients who have some symptoms, but because levothyroxine is seen as an easy fix, they get a prescription and yet see no resolution to their symptoms. Hennessey says he would like to see physicians look for other potential reasons for symptoms. 

Age-based evaluation of hypothyroidism

One thing Hennessey noted that wasn’t mentioned in the paper is the age based relationship of hypothyroidism. For older people – over 60 – a cutoff of 6 mlU/L is probably appropriate. For even older people it’s possible that it is 7.5. And treatment should be carefully considered for older patients before their TSH reaches 10 mlU/L, he says.

He sees the lesson from this as the same as what happened with vitamin D a few years ago. “Researchers found that a level of over 30 was associated with optimal bone health, so they changed the limit from 20 to 30. All of a sudden we got all sorts of inquiries about vitamin D. We created a new disease of vitamin D deficiency. But studies didn’t find evidence that treating this condition with extra vitamin D changed outcomes.”

Symonds says that it’s possible to watch TSH levels and not treat in patients with no symptoms, and their levels may spontaneously normalize.

close X