The Thyroid Test Nobody Reordered

Levothyroxine dose requirement tracks body weight, so a drug that removes a lot of weight predictably changes the dose someone needs. A 2026 study looked at whether anyone actually rechecked. Among 5,370 matched patients on stable levothyroxine, thyroid testing after starting a GLP-1 was indistinguishable from after starting a drug with far smaller weight effects — same proportion tested, same timing.^[1]

What was compared

The study emulated a trial using Medicare claims: adults aged 65 and over with type 2 diabetes, on a stable levothyroxine dose, starting either a GLP-1 or an SGLT2 inhibitor, matched one to one on propensity scores. The outcome was simply whether a TSH test was ordered during follow-up.^[1]

The comparator is well chosen. SGLT2 inhibitors also reduce weight, but considerably less than GLP-1 drugs do — so this is not a comparison against no treatment. Two groups whose thyroid-monitoring needs genuinely differ were monitored identically.

Eighty-three percent in both arms. One hundred and thirty days in both arms. Nothing moved.

Why it should have moved

Levothyroxine is dosed by body weight, and the therapeutic window is narrow enough that people notice when it is wrong — fatigue and cold intolerance at one end, palpitations and anxiety at the other. Lose a substantial amount of weight and the dose that was correct becomes too much.

There is a second, separate reason specific to the oral GLP-1: taking it alongside levothyroxine raised total thyroxine exposure by 33% in a dedicated trial, which we covered in interactions. That is a pharmacokinetic effect on top of the weight effect, and it applies from the first dose rather than after months of loss.

What the study cannot tell you

• Whether anyone came to harm. It measured testing, not outcomes. Under-monitoring is a risk, not an injury.
• The population is narrow: Medicare, aged 65 and over, 71.6% women, 87.1% White, with type 2 diabetes.
• The period was 2011 to 2020 — largely before the modern high-dose weight-loss era. The weight changes involved were smaller than someone on a current dose would experience, which makes the finding conservative rather than weaker.
• 83% is not bad testing in absolute terms. The finding is not that monitoring was poor; it is that it did not respond to a change in need.

What to do with it

If you take levothyroxine and are starting or already on a GLP-1, this is a reasonable thing to raise: whether your thyroid levels should be rechecked as your weight changes, and when. It is a question rather than a demand, and the answer belongs to whoever manages your thyroid.

It is also worth knowing what an over-replaced thyroid feels like, because the symptoms overlap with things people attribute to the GLP-1 itself — palpitations, anxiety, heat intolerance, trouble sleeping. Attributing everything to the new drug is the easy mistake when two things changed at once.