TSH: the thyroid number the reference range argues about

TSH is the most-ordered hormone test in medicine, and it comes with a strange property: the experts don’t fully agree on where “normal” ends. Most labs print a reference range topping out around 4.0–4.5. A serious body of endocrinologists has argued for two decades that a truly healthy upper limit sits closer to 2.5. That’s not a rounding error — millions of people live in the gap, many of them tired, cold, and told their thyroid is fine.

This page takes the argument seriously instead of picking a side for you: what TSH actually measures, what the range fight is really about, what “optimal” defensibly means, why you can be normal-on-paper and exhausted in real life — and what a careful physician actually does with a value like 3.8.

What TSH actually measures

Here’s the twist most people miss: TSH is not a thyroid hormone. It’s the signal sent to your thyroid — thyroid-stimulating hormone, produced by the pituitary gland at the base of your brain. The pituitary continuously reads the level of actual thyroid hormone in your blood (free T4 and T3) and adjusts its shouting accordingly: hormone drifts low, TSH rises to demand more; hormone runs high, TSH falls toward silence.

So TSH is a thermostat reading, and it’s exquisitely sensitive by design — the relationship is log-linear, meaning a small fall in free T4 produces a large rise in TSH. That’s why TSH is the standard first-line test: it amplifies trouble before the hormone levels themselves look clearly abnormal. High TSH generally means the pituitary is compensating for an underperforming thyroid (hypothyroid direction). Suppressed TSH generally means it’s backing off from an overactive one (hyperthyroid direction).

The sensitivity is also the limitation. A thermostat tells you the system is working hard; it doesn’t tell you why, and on rare occasions the thermostat itself is the problem. Which is why TSH is where thyroid assessment starts, not where it ends.

The reference-range argument, honestly told

Both sides of this fight have real evidence. You deserve to hear both.

The case for a lower limit (~2.5): standard reference ranges were built from population samples that quietly included people with early, undiagnosed thyroid disease — inflating the upper end. When the large U.S. NHANES III analysis restricted the sample to a rigorously screened, disease-free reference population (no thyroid history, no antibodies, no goiter, no interfering medications), most healthy adults clustered low: the population’s median TSH sat around 1.4, and roughly 95% of the truly disease-free group fell below about 2.5. People with TSH between 2.5 and 4.5 are also more likely to carry thyroid antibodies — for some of them, upper-normal is early autoimmune thyroid disease in slow motion. This is the basis on which the National Academy of Clinical Biochemistry suggested, back in 2003, that the upper limit would eventually narrow.

The case against (~keep 4.0–4.5): TSH rises naturally with age — follow-up analyses of the same NHANES data showed the TSH distribution shifts upward in older adults without thyroid disease, meaning a 3.8 in a healthy 75-year-old may be that person’s normal, not pathology. Lowering the cutoff to 2.5 would label tens of millions of people “abnormal” overnight, most of whom would never develop thyroid disease. And crucially, treating mildly elevated TSH doesn’t reliably help: in the largest randomized trial of levothyroxine for subclinical hypothyroidism in older adults (TRUST), treatment normalized the number but produced no improvement in tiredness or other symptoms versus placebo.

The honest synthesis: a TSH between about 2.5 and 4.5 is neither “fine, forget it” nor “diseased, treat it.” It’s a watch value — most meaningful in context: your symptoms, your antibodies, your free T4, your age, and your own history. Which brings us to what a good physician actually does with one.

What “optimal” defensibly means

Caveats that change the number without your thyroid changing: TSH runs a daily rhythm (highest overnight and early morning — an afternoon draw can read ~20–30% lower than a dawn one); acute illness and recovery bend it in both directions; pregnancy uses entirely different trimester-specific ranges; several drugs (steroids, dopamine agonists, lithium, amiodarone) push it around; and high-dose biotin — beloved of hair-and-nails supplements — can corrupt the assay itself, mimicking hyperthyroidism on paper. The FDA has formally warned about biotin interference; skip biotin for at least two days before a thyroid draw.

One more underappreciated fact: your personal TSH setpoint is far narrower than the population range. Month to month, an individual’s TSH varies within a tight band unique to them — so a 3.9 in someone who ran 1.2 for a decade is a bigger signal than the population range implies, even though both values print “normal.” This is the strongest argument for having a baseline and a trend rather than a single lifetime data point.

“My TSH is normal — so why am I still exhausted?”

Three honest possibilities, in descending order of frequency:

1. It genuinely isn’t your thyroid. Fatigue is the most nonspecific symptom in medicine. The most common lab-findable culprit in menstruating women isn’t thyroid at all — it’s iron: ferritin can be deeply depleted, with very real fatigue and hair shedding, while TSH and even a full blood count stay normal. Sleep debt, depression, and early insulin resistance (fasting insulin high while glucose still looks fine) are the other big three. A TSH-only fatigue workup rules out one cause and silently skips the rest — this is the single most common way “your labs are normal” goes wrong.
2. The TSH is normal but the picture isn’t complete. TSH-only testing assumes the pituitary is telling the whole story. Checking free T4 alongside catches the uncommon discordant cases, and TPO antibodies identify autoimmune thyroid disease years before TSH leaves the range — relevant if your TSH is drifting up through the 2.5–4.5 band with symptoms.
3. You’re at the edge of your own setpoint. The within-person argument above. One value can’t show this; a trend can.

The pattern across all three: the answer isn’t a better thyroid test, it’s the neighboring tests drawn at the same time — iron, metabolic, inflammatory, and the thyroid panel proper — read together by someone qualified to tell the stories apart.

What actually moves TSH

Honest answer: less than the wellness internet suggests. TSH is a signal to investigate, not a dial to turn.

• Iodine matters in both directions — deficiency raises TSH, but excess iodine (kelp supplements are the classic offender) can disrupt thyroid function too. In iodine-sufficient countries, supplementing blind is more likely to hurt than help.
• Selenium has modest evidence for lowering TPO antibody levels in autoimmune thyroiditis; evidence it improves how you feel is thin.
• Recovery and repeat testing “fix” more mild TSH elevations than any supplement: transient thyroiditis, recovery from illness, and simple assay variation resolve on the second draw about half the time.
• A warning that earns its bold: some “thyroid support” supplements have been found to contain actual thyroid hormone, unlabeled and undosed. That’s not support; that’s unsupervised medication. Nothing over-the-counter should be “treating” your TSH.
• When medication is right, it’s very right. Genuine hypothyroidism is one of the most treatable conditions in medicine — a decision that belongs to a licensed clinician looking at your whole picture, not to a webpage, ours included.

Terve Health measures and tracks TSH in context; diagnosis and treatment decisions are always made with a licensed clinician.

How to get your TSH tested

Three realistic routes:

1. Your physical may already include it — TSH is the one hormone that sometimes makes the default panel. The catch is what comes with it: usually nothing. TSH alone, no free T4, no antibodies, no iron studies for the fatigue differential — and a binary in-range/out-of-range read.
2. Order it a la carte. Direct-to-consumer routes run roughly $29–55 all-in — every route priced in what a TSH test costs, including the $150-panel upsell to skip. Cheap and accessible — but a solo TSH inherits every limitation on this page, and nobody reviews it with you.
3. Measure it inside a full panel a physician actually reads. Terve Health is a whole-body panel — drawn at any of 2,000+ Quest locations, ordered by a licensed clinician, and returned as a written plain-language review: your TSH next to your free T4, antibodies, ferritin, fasting insulin, and inflammation markers, with the fatigue differential actually worked instead of waved at. You don’t choose a panel — a physician orders the right one for you. And if a result needs real follow-up, the medical practice behind your panel can actually see you.

TSH and the rest of the picture

• Free T4 — the hormone TSH is shouting about; the pair distinguishes subclinical from overt, and catches the rare discordant cases.
• TPO antibodies — the why behind a drifting TSH; positive antibodies with upper-normal TSH is a watch-closely pattern.
• Ferritin — the other half of the fatigue differential, and iron is a required cofactor for thyroid peroxidase, the enzyme that makes thyroid hormone — low iron and sluggish thyroid function travel together more often than chance.
• Fasting insulin — the metabolic overlap: fatigue, weight gain, and sluggishness read “thyroid” to patients and “insulin resistance” on lab paper more often than either alone.
• Vitamin D — the third station on the unexplained-fatigue circuit; drawn with TSH and ferritin so one workup answers all three at once.
• Testosterone — thyroid dysfunction shifts SHBG and produces the same energy/mood/libido story; each hormone workup proofreads the other.

Browse the full biomarker library to see everything the panel covers.

Frequently asked questions

What is an optimal TSH level? In rigorously screened healthy adults, TSH clusters between about 0.4 and 2.5 mIU/L with a median near 1.4 — which is why many clinicians treat that zone as functionally optimal for younger and middle-aged adults. But TSH rises naturally with age, and a value between 2.5 and 4.5 is often entirely normal, particularly in older adults. Optimal is a zone read against your age, symptoms, antibodies, and your own baseline — not a single winning number.

Can my TSH be normal and my thyroid still be the problem? Uncommonly, yes: central hypothyroidism (a pituitary signaling problem) produces low thyroid hormone with a deceptively normal TSH, which is why free T4 alongside TSH matters when symptoms are strong. Far more often, though, a normal TSH is telling the truth and the fatigue has a different measurable cause — iron deficiency and early insulin resistance lead that list. The fix is testing the differential, not re-testing the thyroid.

What does a high TSH mean? The pituitary is working harder to get enough thyroid hormone — usually an underactive thyroid, most commonly autoimmune (Hashimoto’s) in the U.S. Between 4.5 and 10 with normal free T4, it’s “subclinical,” and about half of mild elevations normalize on repeat testing — so a confirmed trend matters more than one draw. Above 10, guidelines generally support treatment. Every step of that path belongs with a clinician.

TSH vs. free T4 — what’s the difference? TSH is the pituitary’s request; free T4 is the thyroid’s delivery. TSH moves first and hardest (that log-linear amplification), making it the more sensitive screen; free T4 confirms severity and catches the rare cases where the two disagree. On a comprehensive panel you read them together, which is the point.

Should a TSH of 5 be treated? Not automatically — and this is where the evidence is genuinely humbling: in the largest randomized trial in older adults with subclinical hypothyroidism, levothyroxine normalized TSH but didn’t outperform placebo on fatigue or quality of life. Guidelines individualize: repeat the test first, weigh symptoms, antibodies, age, and cardiovascular context. It’s a conversation with a clinician, not a threshold reflex.

Sources

1. Hollowell JG, et al. Serum TSH, T4, and thyroid antibodies in the United States population (NHANES III). JCEM. 2002. (Disease-free reference population; median ~1.4.)
2. Baloch Z, et al. NACB Laboratory Medicine Practice Guidelines: laboratory support for thyroid disease diagnosis and monitoring. Thyroid. 2003. (The ~2.5 upper-limit argument.)
3. Surks MI, Hollowell JG. Age-specific distribution of serum TSH and antithyroid antibodies in the U.S. population. JCEM. 2007. (TSH rises with age in disease-free adults.)
4. Stott DJ, et al. Thyroid hormone therapy for older adults with subclinical hypothyroidism (TRUST). NEJM. 2017.
5. Andersen S, et al. Narrow individual variations in serum T4 and T3 in normal subjects. JCEM. 2002. (Within-person setpoint.)
6. Jonklaas J, et al. Guidelines for the treatment of hypothyroidism. Thyroid. 2014. (ATA treatment thresholds and individualization.)
7. FDA Safety Communication: biotin interference with lab tests (updated 2019). fda.gov.
8. Meyerovitch J, et al. Serum thyrotropin measurements in the community: five-year follow-up. Arch Intern Med. 2007. (~50%+ of mild elevations normalize on repeat.)

Terve Health measures and helps you understand your biomarkers; it does not diagnose, treat, or prevent disease. Laboratory testing performed by Quest Diagnostics®; testing is ordered and results are reviewed by licensed clinicians at an independent medical practice. Always discuss results and treatment with a licensed clinician.