TSH and Fertility — What Your Thyroid Result Actually Means (India 2026)

TSH is thyroid stimulating hormone, released by the pituitary to tell the thyroid how hard to work. It is checked before fertility treatment because thyroid function affects ovulation and early pregnancy — and because it is cheap and easily corrected.

The counter-intuitive part is the direction. TSH rises when thyroid output falls, so a high TSH suggests an underactive thyroid, not an overactive one. People frequently read the number the wrong way round.

• The 2.5 mIU/L TSH figure is from ATA 2011; the 2017 guidelines revised the reference upper limit to 4.0 mIU/L in healthy pregnant women.
• TABLET randomised 952 euthyroid antibody-positive women to levothyroxine or placebo and found no significant difference in live births.
• T4LIFE reached the same conclusion in antibody-positive women with recurrent pregnancy loss.
• Evidence differs where thyroid function is genuinely abnormal: a 2018 meta-analysis found lower miscarriage rates with levothyroxine in subclinical hypothyroidism during IVF/ICSI.

It is almost always read alongside free T4, and often alongside thyroid peroxidase antibodies. A TSH result on its own answers less than it appears to.

The reason all three are needed is that they answer different questions. TSH says how hard the pituitary is pushing. Free T4 says whether the thyroid is actually keeping up despite the pushing. Antibodies say whether the immune system is involved.

A raised TSH with normal free T4 describes a very different situation from a raised TSH with low free T4, and only the second is unambiguous hypothyroidism.

Thyroid function also fluctuates. A single reading is a snapshot of a system that moves, and results drift with time of day, intercurrent illness and the laboratory's own assay. That is a reason to repeat a borderline result before acting on it, rather than to dismiss it. Our overview of how the thyroid affects fertility covers the wider picture.

Not according to current guidance. The 2.5 mIU/L figure comes from the American Thyroid Association's 2011 guidance for early pregnancy. The ATA revised it upward in 2017 to a reference upper limit of 4.0 mIU/L in healthy pregnant women.

This matters because 2.5 is still quoted widely — in clinic handouts, on fertility forums, and in a great deal of online content that has not been updated in a decade. A woman with a TSH of 3.2 can be told she has a thyroid problem on the strength of a threshold that was superseded.

The ATA published further guidelines for preconception, pregnancy and postpartum in May 2026. Check the current edition rather than any figure quoted second-hand, including this one — and note that the 2026 guidance itself records that much of the underlying evidence remains of low-to-moderate quality.

A raised TSH is still worth investigating. The point is narrower: 2.5 is not a line between healthy and unhealthy, and a result between 2.5 and 4.0 is not by itself abnormal under the guidance that replaced it.

It is worth understanding why a threshold moves at all, because it tends to be read as the experts changing their minds arbitrarily. Reference ranges are derived from populations, and a range built from one population does not necessarily transfer to another.

When larger and more varied data became available, the upper limit that described a healthy pregnant population turned out to be higher than the earlier figure suggested. The number changed because the evidence underneath it did.

The practical consequence for a patient is uncomfortable but simple: the threshold you are measured against depends on which edition your clinic is using. That is a reasonable thing to ask outright, and a clinic that cannot say is worth pressing.

Subclinical hypothyroidism means TSH is above the reference range while free T4 remains normal. The thyroid is being pushed harder than usual but is still producing enough hormone, so there is often no symptom at all.

It is a laboratory description rather than an illness in the ordinary sense, and that is precisely why treating it is contested. If the thyroid is keeping up, the question of whether intervention helps is genuinely open rather than obvious.

Its prevalence also depends entirely on where the upper limit is drawn. Move the threshold from 4.0 to 2.5 and a large group of women acquire a diagnosis without anything having changed in their bodies.

That is not a quibble about definitions — it determines who is offered treatment, and it is the clearest illustration of why the threshold question matters more than it sounds.

A diagnosis of subclinical hypothyroidism should therefore prompt two questions before any treatment question: which reference range was applied, and has the result been repeated. Both can change the answer. We look at the range of thyroid problems that affect fertility separately, including the overt disease this is distinguished from.

For euthyroid women with thyroid antibodies, the best available trials say no. TABLET randomised 952 such women across 49 UK centres to levothyroxine 50 micrograms or placebo before conception and through pregnancy, and found no significant difference in live births or any other outcome.

The T4LIFE trial reached the same conclusion in euthyroid women with thyroid antibodies and recurrent pregnancy loss: levothyroxine did not raise live birth rates against placebo.

Two well-conducted randomised trials pointing the same way is about as clear as fertility evidence gets. It does not support offering levothyroxine to a woman whose thyroid function is normal simply because antibodies were found.

The picture is less settled where thyroid function is genuinely abnormal. A 2018 meta-analysis of randomised trials in women with subclinical hypothyroidism and/or thyroid autoimmunity undergoing IVF or ICSI reported a significantly lower miscarriage rate with levothyroxine. That is a different population from the one TABLET studied, and the distinction is the whole argument.

These findings are often quoted against each other, as though one must be wrong. They are not in conflict. TABLET asked whether giving thyroid hormone helps women whose thyroid hormone is already adequate; the answer was no.

The meta-analysis asked whether correcting a genuine deficiency helps; the answer was more encouraging. What separates them is whether there was anything to correct in the first place.

This is why the threshold question and the treatment question are the same question. If a woman is labelled subclinically hypothyroid only because a 2.5 cut-off was applied, she is closer to the TABLET population than to the one where treatment showed benefit — and the evidence for treating her is correspondingly weaker.

Thyroid peroxidase antibodies indicate thyroid autoimmunity. They are associated with pregnancy loss, but the trials above show that correcting thyroid hormone in women whose levels are already normal does not reverse that association.

One reading of that result, offered by the trialists, is that the risk in these women is driven by the immune environment rather than by any shortfall of thyroid hormone. If that is right, giving thyroid hormone would not be expected to help — which is what the trials found.

A positive antibody result is still worth knowing, because it raises the chance of thyroid function deteriorating later and is a reason to monitor. It is not, on current evidence, a reason to start treatment while function is normal.

This is a genuinely difficult result to receive, and the difficulty deserves naming. Being told you have an antibody associated with pregnancy loss, and in the same breath that the obvious treatment does not help, is not a satisfying place to be left.

It is nonetheless what two well-conducted trials found, and being told otherwise would be worse — a treatment that does not work still has a cost, in money, in monitoring, and in the belief that something is being done.

What the finding does not mean is that nothing can be done about recurrent loss. It means this particular intervention, in this particular group, did not change the outcome. Other causes of recurrent loss are investigated on their own terms, and a positive antibody result does not close that inquiry.

Ask three questions: what the reference range on your report is, whether free T4 was measured alongside TSH, and whether antibodies were checked. A single TSH figure without those cannot support a treatment decision.

• A TSH between 2.5 and 4.0 is not abnormal under the guidance that replaced the 2.5 threshold.
• Normal function with positive antibodies is not, on trial evidence, an indication for levothyroxine.
• Genuinely raised TSH with low free T4 is a different situation and is treated.
• Guidance is being revised — the ATA issued new guidelines in May 2026.

If you are already taking levothyroxine and pregnant or planning pregnancy, none of this is a reason to stop. Dose requirements change in pregnancy, and that is a conversation with the clinician who prescribed it — not something to adjust on the strength of an article. For how thyroid hormone sits alongside the rest of the endocrine picture, see the role of hormones in female reproduction.