Low TSH with normal thyroid hormones: when subclinical hyperthyroidism is treated, and what the evidence shows

A blood test comes back with a thyroid-stimulating hormone (TSH) level below the reference range, while the thyroid hormones themselves, free thyroxine (T4) and triiodothyronine (T3), are normal. This combination is called subclinical hyperthyroidism. "Subclinical" describes the laboratory picture, not how a person feels: some people have symptoms, many have none. The practical questions are what the number means, whether it will stay that way, and when treatment is recommended. The answer depends heavily on how low the TSH is, how old the person is and what else is going on with the heart and bones, and the evidence behind it is weaker than the confident tone of many lab reports suggests.

What the result means

TSH is made by the pituitary gland and tells the thyroid how hard to work. When the thyroid produces a little more hormone than the body needs, the pituitary turns TSH down, often before the hormone levels themselves leave their reference ranges. A low TSH with normal T4 and T3 therefore points to a mild excess of thyroid hormone action.

The 2015 European Thyroid Association (ETA) guideline divides the condition into two grades:

• Grade 1: TSH between 0.1 and 0.39 mIU/L.

• Grade 2: TSH below 0.1 mIU/L.

The guideline defines the diagnosis as a persistently low TSH with normal free hormone levels. A single low result is not enough to apply the label.

A single low TSH is often not the end of the story

A low TSH is not always a lasting thyroid disease. The 2015 ETA guideline lists transient forms, such as the overactive phase of some types of thyroid inflammation (subacute, painless or silent thyroiditis) and the period after treatment of overt hyperthyroidism. It also lists situations in which a low TSH is not called subclinical hyperthyroidism at all: severe illness outside the thyroid, some medicines, pituitary or hypothalamic disease, psychiatric illness, and late in the first trimester of pregnancy. Taking more thyroid hormone tablets than needed is listed separately, as an external cause. If TSH is low and free T4 and T3 are normal, the guideline recommends repeating the tests within 2 to 3 months before drawing conclusions.

How often the result changes on repeat has been measured in large record-based studies:

• In an Israeli primary-care database of 422,242 adults without known thyroid disease, 1.2% had a TSH below the reference range on the first test. Among people not taking thyroid-related medication, 51.2% of those with a low initial TSH had a normal value on a later test within five years.

• In the Scottish TEARS study, which used linked health records to follow 2,024 adults whose TSH was low on at least two measurements at least four months apart, most untreated people still had subclinical hyperthyroidism two, five and seven years later (81.8%, 67.5% and 63.0%). About 0.5–0.7% progressed to overt hyperthyroidism, while 17.2%, 31.5% and 35.6% returned to normal at those time points. Returning to normal was more common when the starting TSH was between 0.1 and 0.4 mU/L than when it was lower.

In other words, in this population persistent subclinical hyperthyroidism rarely turned into overt disease, and a substantial share of people drifted back to normal, especially those with a milder grade.

Finding the cause

The ETA guideline names three main causes of the persistent, endogenous form (that is, not caused by thyroid medication): Graves' disease, a single autonomously functioning nodule, and multinodular goitre. Telling them apart changes what follows. The same guideline names three tests, each with its place. TSH-receptor antibodies can confirm an autoimmune cause such as Graves' disease. Ultrasound can be helpful when the gland is nodular. A radionuclide scan, sometimes with a 24-hour radioiodine uptake test, is recommended for nodular goitre when TSH is below 0.1 mIU/L, to show which parts of the gland are overactive.

What large studies link it to

Most of what is known about risk comes from pooled analyses of prospective cohort studies by the Thyroid Studies Collaboration. These are observational: they show what travels together with a low TSH, not what a low TSH causes.

• Heart and survival. Individual data on 52,674 adults from 10 cohorts: 2,188 (4.2%) had endogenous subclinical hyperthyroidism. Compared with people with normal thyroid function, they had higher total mortality (hazard ratio 1.24), higher coronary heart disease mortality (1.29) and more atrial fibrillation (1.68; analysed in 8,711 participants from 5 cohorts). The risks of coronary death and atrial fibrillation were higher when TSH was below 0.10 mIU/L than when it was between 0.10 and 0.44 mIU/L.

• Atrial fibrillation in older people. In the Framingham Heart Study, 2,007 people aged 60 and over without atrial fibrillation were followed for ten years; the group included people taking thyroid hormone and people with past hyperthyroidism, so it was not limited to the endogenous form. Among those with a TSH of 0.1 mU/L or lower, the ten-year cumulative incidence of atrial fibrillation was 28%, compared with 11% in those with normal TSH; after adjustment, the relative risk was 3.1. People with a slightly low TSH (above 0.1 to 0.4) did not differ significantly from the normal group.

• Fractures. Individual data on 70,298 adults from 13 cohorts: subclinical hyperthyroidism was associated with more hip fractures (hazard ratio 1.36) and more fractures of any kind (1.28). The associations were stronger with a TSH below 0.10 mIU/L and in the endogenous form, excluding people on thyroid medication.

• Dementia. A meta-analysis of prospective cohorts found a pooled risk ratio for dementia of 1.67 in subclinical hyperthyroidism, based on five studies with 6,410 participants (95% confidence interval 1.04 to 2.69). Decline on the Mini-Mental State Examination, a standard cognitive screening test, did not differ significantly. The authors described the data as limited.

The pattern across these studies is consistent: the associations are clearest for grade 2, a TSH below 0.1.

Trials of treatment: one small trial, no answer yet

Subclinical hyperthyroidism can be treated in the same ways as overt hyperthyroidism, depending on the cause: antithyroid drugs, radioactive iodine or surgery. No randomised trial has shown that treating it reduces fractures or deaths. For atrial fibrillation, one randomised trial has been published (France, 2024): 144 people aged 50 and over with TSH below 0.4 mU/L, normal hormone levels and an autonomously functioning thyroid on a scan were assigned to radioiodine or to monitoring. New atrial fibrillation occurred in 1 of 70 treated and 4 of 74 monitored people, a difference that was not statistically significant. The trial struggled to recruit, and the authors concluded it failed to show a significant reduction; a quarter of those treated with radioiodine developed an underactive thyroid during follow-up. When the ETA guideline was written in 2015, its recommendations were made "despite the absence of randomized prospective trials". A 2012 review in The Lancet called large randomised trials urgently needed, and the fracture analysis above left the same question open: whether treatment prevents fractures still has to be studied.

So the recommendations below rest mainly on observed risks and expert judgement about how to balance them, not on proof that treatment changes the outcomes.

When guidelines recommend treatment

The ETA 2015 guideline uses TSH grade, age and symptoms:

• Grade 2 (TSH below 0.1), age over 65: treatment is recommended.

• Grade 1 (TSH 0.1–0.39), age over 65: treatment could be considered, because of the higher risk of atrial fibrillation.

• Grade 2, age under 65: treatment is suggested if the low TSH persists and/or there are symptoms of thyroid hormone excess (a weak recommendation), and recommended when cardiovascular risk factors or other illnesses are present.

• Grade 1, age under 65, without symptoms: the guideline found no data to support treatment and advises follow-up without treatment.

The 2016 American Thyroid Association guideline separates two ranges. When TSH is persistently below 0.1 mU/L, it strongly recommends treatment for everyone aged 65 or older; for people with cardiac risk factors, heart disease or osteoporosis; for postmenopausal women not taking estrogens or bisphosphonates; and for people with symptoms of hyperthyroidism. For people under 65 without those risks, it says treatment should be considered. When TSH is persistently below the lower limit of normal but at least 0.1 mU/L, it says treatment should be considered at age 65 or older and with heart disease, osteoporosis or symptoms; people under 65 without these can be observed.

What this means for reading your own result

• A low TSH with normal T4 and T3 is a reason to repeat the test, not a diagnosis on its own.

• If it persists, the next step is finding the cause, because Graves' disease and nodular disease are managed differently.

• How low the TSH is matters: the observed links with atrial fibrillation, fractures and mortality are strongest below 0.1.

• Age, heart rhythm, bone health and symptoms shape the decision. For younger people with a mildly low TSH and no symptoms, guidelines advise follow-up rather than treatment.

• Whether treatment changes long-term outcomes is not established: the only randomised trial was small, did not show a significant reduction in atrial fibrillation, and did not study fractures or deaths. That uncertainty is worth discussing openly with an endocrinologist.

For the general logic of which tests have been shown to change outcomes, see our article on screening after 40.

This material is for information only and does not replace a consultation with a doctor.

References

• Biondi B, et al. The 2015 European Thyroid Association Guidelines on Diagnosis and Treatment of Endogenous Subclinical Hyperthyroidism. Eur Thyroid J. 2015. PMID 26558232

• Ross DS, et al. 2016 American Thyroid Association Guidelines for Diagnosis and Management of Hyperthyroidism and Other Causes of Thyrotoxicosis. Thyroid. 2016. PMID 27521067

• Meyerovitch J, et al. Serum thyrotropin measurements in the community: five-year follow-up in a large network of primary care physicians. Arch Intern Med. 2007. PMID 17646608

• Vadiveloo T, et al. The Thyroid Epidemiology, Audit, and Research Study (TEARS): the natural history of endogenous subclinical hyperthyroidism. J Clin Endocrinol Metab. 2011. PMID 20926532

• Collet TH, et al. Subclinical hyperthyroidism and the risk of coronary heart disease and mortality. Arch Intern Med. 2012. PMID 22529182

• Sawin CT, et al. Low serum thyrotropin concentrations as a risk factor for atrial fibrillation in older persons. N Engl J Med. 1994. PMID 7935681

• Blum MR, et al. Subclinical thyroid dysfunction and fracture risk: a meta-analysis. JAMA. 2015. PMID 26010634

• Rieben C, et al. Subclinical thyroid dysfunction and the risk of cognitive decline: a meta-analysis of prospective cohort studies. J Clin Endocrinol Metab. 2016. PMID 27689250

• Cooper DS, Biondi B. Subclinical thyroid disease. Lancet. 2012. PMID 22273398

• Goichot B, et al. Treatment of subclinical hyperthyroidism in patients older than 50 years: a randomized controlled study. Eur Thyroid J. 2024. PMID 39159264