The short answer: subclinical hypothyroidism means your TSH is above the reference range while your free T4 is still normal. It is common, affecting roughly 4% to 8.5% of adults and up to 15% of women over 60. For most adults, the current guideline advice is not to treat it. A 2019 BMJ Rapid Recommendation based on 21 randomized trials issued a strong recommendation against thyroid hormone in adults with subclinical hypothyroidism, because levothyroxine did not improve symptoms, quality of life, fatigue, or body weight. The exceptions are real but narrow: pregnancy or planning pregnancy, a TSH above 20, and severe symptoms.
The uncomfortable part is that "your TSH is a bit high" often arrives alongside genuine fatigue, and the evidence says the pill usually will not fix the fatigue. Both things are true.
What is the actual definition?
Two numbers define it. TSH, produced by the pituitary, is above the laboratory upper limit, typically around 4.0 to 4.5 mIU/L. Free T4, the hormone the thyroid actually makes, is inside the normal range.
The mechanism is a compensating system. As the thyroid starts to falter, the pituitary raises TSH to push harder, and for a while that pressure keeps T4 output normal. The elevated TSH is the signal that extra effort is required, not evidence that the output has failed yet. That is why "subclinical" is the right word: the biochemistry has shifted, the hormone level has not.
One practical point before anything else. TSH is a noisy measurement. It varies through the day, rises after a poor night of sleep, and shifts transiently after any acute illness. A single high reading is not a diagnosis. Repeat it, ideally 6 to 12 weeks later, before drawing conclusions. Our guide to thyroid blood tests covers what each value on the panel measures.
How common is it?
NHANES III put the US prevalence at 4.3%. Other population studies report 7.5% to 8.5% of women and 2.8% to 4.4% of men. In women over 60 the figure rises to about 15%.
Prevalence at that level is exactly why over-treatment is a real concern. A finding present in one in twenty adults, screened for widely, will generate a large number of prescriptions if the threshold for treating is low.
Does it progress to real hypothyroidism?
Sometimes, at a rate that can be estimated. Overall progression to overt hypothyroidism runs about 2% to 6% per year. Thyroid peroxidase (TPO) antibodies separate the risk groups: roughly 2.6% per year when antibodies are absent, and about 4.3% per year when they are present.
That makes a TPO antibody test genuinely useful here, not because it changes today's treatment but because it changes the monitoring interval. Positive antibodies plus a rising TSH is a trajectory. Negative antibodies with a TSH of 5.2 that returns to 3.8 on a repeat draw is usually nothing.
Our guide to Hashimoto's thyroiditis covers what those antibodies mean in more detail.
What do the guidelines actually say?
In May 2019, a BMJ Rapid Recommendation panel led by Bekkering and colleagues reviewed a systematic review of 21 randomized trials and issued a strong recommendation against thyroid hormone treatment in adults with subclinical hypothyroidism (BMJ 2019;365:l2006). The panel found moderate-to-high quality evidence that levothyroxine produced no clinically relevant benefit for quality of life, thyroid-related symptoms, depressive symptoms, fatigue, or body mass index.
The American Family Physician summary of that recommendation is explicit about who it does not cover:
- Pregnant patients. Excluded. The benefit is unknown and the risk calculus is different.
- TSH above 20 mIU/L. Excluded, because that generally indicates overt rather than subclinical disease.
- People with severe symptoms. Excluded, because they were not in the trials.
- Adults aged 30 or younger. Very few participants, so the evidence does not extend to them.
The older convention of treating everyone above a TSH of 10 predates this evidence and, as the AFP review notes, had little supporting data behind it.
What did the TRUST trial find?
TRUST is the trial that changed practice most. It randomized 737 adults aged 65 and older with TSH between 4.6 and 19.99 mIU/L and normal free T4 to levothyroxine, at a median dose of 50 micrograms, or placebo.
The treatment worked biochemically. Mean TSH fell from 6.40 to 3.63 mIU/L over 12 months. It did not work clinically. There was no difference between groups in the hypothyroid symptoms score or the tiredness score, the two co-primary outcomes.
That result is the core of the argument. The number moved. The person did not feel different.
When should it be treated?
Treatment remains reasonable, and in some cases important, in these situations:
- Pregnancy, or actively trying to conceive. Thyroid hormone requirements rise in pregnancy and the fetus depends on maternal supply in the first trimester. This population sits outside the recommendation against treatment and warrants specialist input.
- TSH above 20 mIU/L, which is usually treated as overt hypothyroidism rather than subclinical.
- TSH persistently above 10 with positive TPO antibodies, where progression risk is highest. This is a judgment call between patient and clinician, not an automatic indication.
- A goiter, or symptoms severe enough that the trial evidence does not describe the patient.
- Younger adults, particularly under 30, where the evidence base is thin.
If a trial of levothyroxine is started for symptoms, the honest framing is that it is a trial. Set a defined endpoint, typically three to six months, agree in advance which symptom you are watching, and stop the drug if that symptom does not change. Our levothyroxine guide covers dosing, timing, and interactions.
What if the TSH is only slightly high and you still feel terrible?
This is the most common version of the question, and it deserves a direct answer rather than reassurance.
Fatigue, weight change, low mood, and brain fog are non-specific. They are also common. When a mildly elevated TSH turns up during a workup for those symptoms, the elevated TSH is frequently a bystander rather than a cause, which is precisely what the trials demonstrated when treating it changed nothing.
The alternatives worth ruling out are unglamorous and often more productive: iron deficiency with or without anemia, vitamin B12 deficiency, obstructive sleep apnea, depression, anxiety, medication effects, and simple sleep debt. Thyroid symptoms overlap heavily with anxiety in particular, which our sister site anxiety.md covers in detail.
If your TSH is normal and symptoms persist, our guide to why symptoms persist with a normal TSH addresses that situation specifically.
The bottom line
Subclinical hypothyroidism is a high TSH with a normal free T4, found in roughly 4% to 8.5% of adults. Confirm it with a repeat test before acting on it, and add a TPO antibody test to set the follow-up interval.
For most non-pregnant adults, the evidence supports monitoring rather than treating. Twenty-one randomized trials, including TRUST in 737 older adults, found that levothyroxine lowers the number without improving how people feel. Treatment still belongs in pregnancy, at a TSH above 20, and in selected higher-risk or younger patients. If symptoms are the reason you are considering it, look hard at the other causes first.
Last updated: August 2026. This article is for informational purposes only and does not constitute medical advice. Do not start, stop, or change a thyroid medication without speaking to the prescribing clinician, and seek specialist advice if you are pregnant or planning a pregnancy.