The TSH Reference Range Debate: Is a TSH of 4 really healthy?

Joseph Hollowell and colleagues at the US Centers for Disease Control and Prevention (CDC) analyzed blood samples from 17,353 people in NHANES III, representative of the US population. In the strictly selected reference group without antibodies and without thyroid disease, mean TSH was 1.40 mU/l. Around 11 percent of the apparently healthy people had TPO antibodies, and these were associated with thyroid dysfunction.

What this means for you: The typical TSH of a person with a healthy thyroid is closer to 1 to 2 than to 4. And among apparently healthy people, a noticeable share has silent autoimmune activity.

Leonard Wartofsky and Richard Dickey from Washington argued that earlier reference groups were mixed with people in early stages of thyroid disease. They point to the laboratory guidelines of the National Academy of Clinical Biochemistry, NACB for short: more than 95 percent of healthy people have a TSH below 2.5 mU/l. They interpret the mean TSH of 1.18 in Black Americans, who have Hashimoto's less often, as the true normal value.

What this means for you: From this point of view, a TSH above 2.5 is an outlier that may often reflect early Hashimoto's thyroiditis.

Martin Surks, Gayotri Goswami and Gilbert Daniels analyzed the same NHANES III reference group of 14,333 people without antibodies. 85 percent were below 2.5, another 9.7 percent between 2.5 and 4.5. Lowering the limit would have classified roughly 20.6 million additional people in the United States alone as having subclinical hypothyroidism, many of them without thyroid disease.

What this means for you: A TSH between 2.5 and 4.5 can be a signal. But it can also simply be your normal value, and no benefit from treatment has been shown for this range.

Mark Vanderpump and the team in Newcastle followed the Whickham cohort for two decades. In women, an elevated TSH together with antibodies was associated with an odds ratio of 38 for later hypothyroidism. Even rising values above 2 mU/l increased the likelihood.

What this means for you: A TSH in the upper normal range is not a diagnosis. Together with antibodies, however, it is a good reason to keep an eye on the trend.

Martin Surks and Joseph Hollowell divided the thyroid-healthy people from NHANES III by decade of life. In 20 to 29 year olds the 97.5th percentile was 3.56 mU/l, in people over 80 it was 7.49 mU/l. Seventy percent of older people with a TSH above 4.5 were therefore within their age-appropriate range.

What this means for you: For young adults a limit of around 4 may be rather generous. For very old people it may be rather too strict.

Lei Zhao, Weiping Teng and colleagues from Shenyang analyzed the nationwide Chinese TIDE study with 78,470 participants. TSH rose with higher iodine intake, and the upper limit increased by about 0.53 mU/l per decade of life. By the test manufacturer's limits, 13.6 percent would have had subclinical hypothyroidism, by the study's own reference range only 3.0 percent.

What this means for you: How many people are considered ill depends heavily on which limit is used. That is an argument for locally and individually appropriate ranges, and at the same time a warning against overdiagnosis.

Stig Andersen and the team in Aalborg measured 16 healthy men monthly for one year. The individual range of variation was about half as wide as the group's reference range, and a single TSH described the set point only to within about plus or minus 50 percent. The authors conclude that a value within the lab limits is not necessarily normal for an individual person.

What this means for you: Old lab results from times when you felt well may be more valuable than any table.

Pia Skov Hansen and colleagues from Odense studied 690 healthy Danish twin pairs. Around 64 percent of the differences in TSH could be explained by heredity, and a similar share for fT4 and fT3. Iodine intake played a small but measurable role.

What this means for you: Your set point is partly something you were born with. Whether it should be taken into account in treatment is, according to the authors, still an open question.

David Stott and a European consortium gave 737 people aged 65 and over with mildly elevated TSH between 4.6 and 19.99 mU/l either levothyroxine or a placebo. TSH dropped clearly under therapy. After one year, however, there was no difference in hypothyroid symptoms or tiredness, and the other outcome measures also showed no benefit.

What this means for you: According to this trial, hormone therapy brings no noticeable gain for older people with mildly elevated TSH. That is an important argument against rapid treatment. For young people or when you want to conceive, the trial says little.

Bernadette Biondi, Anne Cappola and David Cooper summarized the state of knowledge on subclinical hypothyroidism in 2019. It affects up to 10 percent of adults, is probably overestimated in people over 70 because of the age-related rise in TSH, and most people could be monitored without treatment. They see therapy as an option mainly from a TSH of 10, or in young and middle-aged people with matching symptoms.

What this means for you: The other side does not dismiss symptoms either. But it calls for careful weighing instead of an automatic response.

Roberto Negro and colleagues from Lecce studied 4,123 pregnant women without TPO antibodies in the first trimester. With a TSH between 2.5 and 5.0 mU/l, miscarriage occurred more often than with values below 2.5, namely in 6.1 versus 3.6 percent. There was no difference in preterm births.

What this means for you: A TSH that is considered normal outside pregnancy may be assessed differently in early pregnancy. This is an observation, not proof of cause and effect.

An expert panel led by Erik Alexander revised the ATA guideline on thyroid disease in pregnancy and the postpartum period in 2017. It covers the interpretation of thyroid values in pregnancy, iodine, antibodies and pregnancy complications, as well as the thyroid in infertility. The recommendations are graded by strength of evidence.

What this means for you: When you want to conceive, your TSH belongs in a different frame of assessment than at a routine check.

Myrthe van Dijk, Mariëtte Goddijn and colleagues from the Netherlands, Belgium and Denmark studied 187 women with recurrent miscarriages, TPO antibodies and a normal TSH. Before conception they received levothyroxine or a placebo. The live birth rate was 50 versus 48 percent, with no meaningful difference.

What this means for you: According to this trial, antibodies alone with normal function are no reason for routine treatment. The assessment belongs in the hands of a specialist.