GLP-1 + Thyroid Labs: TSH Changes and Monitoring Protocol

Two questions about GLP-1 therapy and the thyroid come up in almost every initial consult. The first is the boxed warning for medullary thyroid carcinoma (MTC), which is a rodent finding (Bjerre Knudsen 2010^[2]) whose human relevance has been actively investigated in large registries (Bezin 2023^[1]). The second is what to expect on thyroid lab work during weight loss: TSH commonly falls by roughly 0.5–1.0 mIU/L over the first 6 months, free T4 usually moves very little, and patients already taking levothyroxine often need a small dose reduction. The underlying physiology is well documented — obesity elevates TSH via the leptin-TRH axis, and weight loss reverses that signal (Knudsen 2005^[5], Laurberg 2012^[6], Reinehr 2011^[8]). This article walks through the boxed warning context, the lab changes you should expect, and the quarterly monitoring protocol that keeps levothyroxine patients in range.

The honest summary

• The MTC boxed warning is grounded in rodent biology. Bjerre Knudsen 2010^[2] showed GLP-1 receptor agonists activate rodent thyroid C-cells, which led to the class-wide boxed warning. The human signal is weaker. The Bezin 2023^[1] French national nested case-control study reported an elevated adjusted hazard ratio for any thyroid cancer in long-term users vs non-users, but the methodology was contested in several published comments and the absolute risk remained small. The contraindication for personal or family history of MTC or MEN2 syndrome stays in the label.
• Calcitonin screening is not recommended. The revised ATA medullary thyroid carcinoma guidelines (Wells 2015^[3]) and the 2015 ATA differentiated thyroid cancer guidelines (Haugen 2016^[4]) do not recommend routine baseline or surveillance calcitonin measurement before starting a GLP-1, because the pre-test probability is too low and false positives carry a real downstream-imaging cost.
• TSH falls with weight loss. The Reinehr and Wolters lifestyle-intervention cohorts (Wolters 2013^[7], Reinehr 2011^[8]) and the population biology in Knudsen 2005^[5] show obesity elevates TSH via the leptin-TRH-TSH axis, and weight loss reverses the signal. The typical drop on a GLP-1 at 6 months is 0.5–1.0 mIU/L with free T4 essentially unchanged.
• Levothyroxine patients often need a dose reduction. Two effects compound: weight loss reduces required dose (about 1.6 mcg/kg per day), and tirzepatide-mediated slowed gastric emptying may alter the pH-dependent absorption pattern (Virili 2022^[11]). We monitor TSH four weeks after each GLP-1 dose escalation and adjust the levothyroxine dose accordingly.

The MTC boxed warning, in context

The class-wide boxed warning on liraglutide, semaglutide, and tirzepatide stems from a rodent observation: Bjerre Knudsen 2010^[2] showed long-acting GLP-1 receptor agonists bind GLP-1 receptors on rodent thyroid C-cells, triggering calcitonin release and C-cell proliferation. Rodent thyroid C-cells express GLP-1 receptors at high density; human C-cells express them at far lower density, which is the biological basis for skepticism that the rodent finding translates to humans at therapeutic doses.

The largest human safety signal study to date is Bezin 2023^[1] — a French national registry nested case-control study using SNDS data — which reported an elevated adjusted hazard ratio for thyroid cancer in patients with 1–3 years of cumulative GLP-1 use compared with non-users. The methodology drew several published comments (immortal-time bias, surveillance bias, indication confounding, and the rarity of MTC making the case ascertainment unstable) and the absolute risk in the cohort remained small. Other large analyses have not replicated the signal cleanly. The label contraindication for personal or family history of MTC or Multiple Endocrine Neoplasia type 2 (MEN2) remains in force regardless of how the population-level signal is adjudicated.

Why we do not screen calcitonin routinely. The revised ATA MTC guidelines (Wells 2015^[3]) and the ATA differentiated thyroid cancer guidelines (Haugen 2016^[4]) explicitly do not endorse calcitonin screening in the general population because the false-positive rate is high and the pre-test probability is low. Adding a GLP-1 prescription does not change that calculus for a patient without personal or family history of MTC or MEN2. The Endocrine Society has taken a similar position. We document the family history on intake, screen for the contraindicated history, and do not order baseline or surveillance calcitonin in low-risk patients.

What TSH actually does on a GLP-1

Obesity is associated with mildly elevated TSH at the population level, with free T4 in the normal range — the picture often labeled “isolated hyperthyrotropinemia” or “subclinical hypothyroidism of obesity.” Knudsen 2005^[5] showed that small differences in thyroid function track with BMI in a population of 4,649 Danish adults, with higher TSH and lower free T4 in the higher-BMI groups even within the reference range. Laurberg 2012^[6] reviewed the mechanism: adipose-derived leptin signals to the hypothalamus to upregulate thyrotropin-releasing hormone (TRH); higher TRH drives higher TSH; the change is partially reversed by weight loss.

The pediatric weight-loss data is the cleanest demonstration of the reversibility. Wolters 2013^[7] followed obese children in a lifestyle intervention and reported significant TSH and free T3 declines in children who lost weight, with regression toward baseline values in those who regained. Reinehr 2011^[8] reviewed the broader pediatric literature and reached the same conclusion: the TSH elevation seen in obesity is a downstream consequence, not a primary cause, and reverses with weight loss in most children.

The same physiology applies to adult GLP-1 patients. In the STEP-1 program (Wilding 2021^[10]) and SURMOUNT-1 (Jastreboff 2022^[9]) the published adverse-event tables do not flag incident hyperthyroidism, and the observed pattern in clinical practice is a TSH decline of roughly 0.5–1.0 mIU/L by week 24 on weight-loss-dose GLP-1 therapy, with free T4 essentially unchanged. The decline tracks total body-weight loss; patients who lose more weight typically see a larger TSH drop.

Magnitude: how much does TSH actually drop?

Levothyroxine on a GLP-1: the dose-reduction problem

Levothyroxine dose tracks body weight at roughly 1.6 mcg per kg per day in adults with primary hypothyroidism. A patient who loses 20% of body weight on tirzepatide is mechanically over-dosed by the same fraction by week 36 unless the prescription is re-titrated. That alone produces falling TSH on follow-up labs — sometimes pushing patients into the suppressed range, which carries cardiovascular and bone consequences in older adults.

A second issue is absorption. Levothyroxine absorption is pH-dependent — gastric acidity dissolves the tablet coating and enables intestinal uptake. Virili 2022^[11] demonstrated experimentally that a higher gastric pH reduces levothyroxine absorption in humans, which is the same mechanism that drives the well-known levothyroxine-PPI interaction. GLP-1 therapy slows gastric emptying, which can shift the timing and possibly the magnitude of levothyroxine absorption, particularly for tablet (not liquid or soft-gel) formulations. The published magnitude of the GLP-1-levothyroxine pharmacokinetic interaction is modest, but the combination of dose-reduction from weight loss plus absorption-shift from delayed gastric emptying explains why TSH often moves more than expected on follow-up.

The practical solution is unchanged from standard levothyroxine pharmacology: take the dose on an empty stomach with water, wait at least 60 minutes before food, coffee, or other oral medications, and ideally separate it from the weekly GLP-1 injection by several hours. We routinely recommend taking levothyroxine on waking and giving the weekly GLP-1 injection later in the day.

The monitoring protocol

1. Baseline labs before the first GLP-1 injection. TSH and free T4 are the minimum. Anti-thyroid peroxidase (anti-TPO) antibodies if not measured in the prior 12 months and the patient has any family history of autoimmune thyroid disease, miscarriage, or symptoms consistent with Hashimoto thyroiditis. Free T3 is not routinely necessary at baseline unless symptoms suggest hyperthyroidism. Cost at Costco lab pricing: TSH and free T4 are roughly $25 each.
2. Repeat TSH at week 12, week 24, and week 52. The week 12 check catches early movers; the week 24 check captures the largest dose-driven shift; the week 52 check confirms steady-state at the maintenance dose. Free T4 is added at week 24 and week 52, or any time TSH is outside the reference range.
3. For levothyroxine patients: TSH four weeks after each GLP-1 dose escalation. Each step up the GLP-1 dose ladder produces a fresh wave of weight loss plus a small change in gastric-emptying kinetics. Check TSH four weeks after each titration step and reduce the levothyroxine dose by 12.5–25 mcg per day when TSH drops below the lower reference limit.
4. Separate levothyroxine from the GLP-1 injection. Take levothyroxine on waking, on an empty stomach, with water; wait at least 60 minutes before food, coffee, calcium, iron, or other oral medications. Inject the weekly GLP-1 later in the day. If GI side effects make morning dosing intolerable, switch to bedtime levothyroxine with the same 4-hour fast.
5. Subclinical hypothyroidism (TSH 4.5–10, normal free T4) before starting: if the working diagnosis is obesity-driven (anti-TPO negative, no goiter, no family history), a reasonable approach is to defer levothyroxine initiation, recheck TSH at week 24, and confirm reversal with weight loss. If anti-TPO positive or symptoms suggest Hashimoto, start levothyroxine per standard guidelines and monitor as above.
6. Endocrinology referral triggers: TSH unstable across two escalation steps, free T4 outside range, new thyroid nodule on examination, neck mass, or any patient with confirmed personal or family history of MTC or MEN2 (these patients should not be on a GLP-1 in the first place).

Hashimoto thyroiditis and the autoimmune-inflammation angle

Patients with Hashimoto thyroiditis often ask whether GLP-1 therapy will worsen the autoimmune process. There is no published evidence that GLP-1 receptor agonists exacerbate Hashimoto. Weight loss itself reduces several markers of systemic inflammation, and the inflammatory milieu of obesity plausibly contributes to autoimmune-thyroid disease activity in subsets of patients. We treat Hashimoto patients the same way we treat any levothyroxine patient on a GLP-1 — monitor TSH carefully and adjust dose as weight comes down — with anti-TPO antibodies as a baseline anchor and repeat measurement only if TSH stability is unexpectedly difficult.

Related research and tools

• Does GLP-1 cause cancer? MTC and thyroid evidence — the full primary-source review of the boxed warning, the Bezin 2023 registry, and the Bjerre Knudsen rodent mechanism
• GLP-1 + levothyroxine (Synthroid) interaction — the absorption pharmacokinetics and the dose- adjustment protocol in depth
• T3 and thyroid hormone for weight loss — why supraphysiologic T3 is not a weight-loss strategy, and what the evidence actually shows
• What the GLP-1 trials actually showed for side effects — STEP-1, STEP-2, STEP-3, SURMOUNT-1 adverse-event tables in detail
• GLP-1 muscle-loss prevention protocol — the protein and resistance-training half of the on-GLP-1 monitoring picture
• GLP-1 side-effect timeline — the canonical onset and resolution windows for common adverse events

Important disclaimer. This article is educational and does not constitute medical advice. Patients with a personal or family history of medullary thyroid carcinoma or Multiple Endocrine Neoplasia type 2 (MEN2) should not take a GLP-1 receptor agonist. Levothyroxine dosing changes should be coordinated with the prescribing clinician and confirmed by repeat TSH; do not self-adjust. Thyroid lab interpretation in pregnancy uses different trimester-specific reference ranges and is outside the scope of this article. PMIDs were verified live against the PubMed E-utilities API on 2026-05-28.

Last verified: 2026-05-28. Next review: every 12 months, or sooner if new prospective trial data on GLP-1 thyroid safety endpoints (post-marketing registry updates, SURMOUNT or STEP long-term extensions) is published.