Endocrine Feedback Loop – Episode 70

Dosing Strategies for Remission in Graves Disease
February 19, 2026
Host: Dr. Chase Hendrickson (Vanderbilt University Medical Center)
Guests: Dr. Andy Crawford (Dartmouth Hitchcock Medical Center), Dr. Giuseppe Barbacino (Massachusetts General Hospital/Harvard Medical School)

Episode Overview

This episode of the Endocrine Feedback Loop explores a recent observational study from Japan investigating whether the minimal pre-discontinuation dose of methimazole (an antithyroid drug) impacts the risk of relapse in patients with Graves disease upon stopping therapy. The discussion examines methodology, data analysis, and practical implications for endocrinologists managing Graves disease, with a focus on international approaches and lingering questions around optimal maintenance dosing before discontinuation.

Key Discussion Points & Insights

1. The Shift to Antithyroid Drugs as First-Line Treatment

[02:36–05:14]

• Thionamides (notably methimazole) are increasingly first-line for Graves disease, with long-term safety established.
• Historical Context: The U.S. favored radioactive iodine due to accessibility, but increased awareness of associated risks (notably thyroid eye disease and hypothyroidism) has shifted practice toward antithyroid drugs.
• Notable Quote:

  “Radioactive iodine was very appealing... but then it became clear that radioactive iodine increases the risk of the dreaded complication of Graves disease, which is thyroid eye disease.” – Dr. Barbacino [03:36]

2. Understanding Remission and Risk of Relapse

[05:14–07:53]

• Goal: Achieve durable remission, but relapse is common (~30–40% post-discontinuation).

• Patients want to maximize the chance of being medication-free long-term.

• Key Risk Factor: Persistently elevated TSH-receptor antibody titers (TRAb) predict higher relapse.

• Distinction between endocrinological and immunological remission explained.

  “People with persistently high levels of TSH receptor antibodies are clearly unlikely to experience a remission once they stop the methimazole.” – Dr. Barbacino [06:25]

3. Study Design and Methodology

[07:53–12:46]

• Design: Retrospective cohort study from a single Japanese institution (Kuma Hospital, Kobe), 2008–2024.

• Inclusion Criteria: Graves disease patients on stable methimazole dose (≤2.5 mg/day) for >3 months before discontinuation.

• Patient Identification: Based on ICD-10 codes, validated for accuracy.

• Primary Outcome: Relapse (resumption of methimazole within 1 year).

• Potential Confounders: Biggest concern—why clinicians choose lower doses prior to discontinuation (“confounding by indication”).

  “It’s possible... you could actually see what the outcome was going to be, and then that changed your prescribing behavior, and it flips that cause and effect relationship around a bit.” – Dr. Hendrickson [14:27]

4. Methimazole Dose Groups Compared

[12:46–15:49]

• Four groups:
  1. <1.25 mg/day
  2. 1.25 mg/day

  3. 1.25–<2.5 mg/day

  4. 2.5 mg/day (reference)
• Other Issues: Discussion about differences between TRAb and TSI (thyroid stimulating immunoglobulin) assays; both sufficient for diagnosis, slight differences in sensitivity.

5. Addressing Confounders & Statistical Rigor

[15:49–18:46]

• Notable omitted variable: Thyroid peroxidase (TPO) antibodies – patients with elevated TPO more likely to remit.
• Propensity score matching used to simulate randomization, but limited by available variables.
• Strength: Multiple sensitivity analyses (e.g., Poisson regression, alternative follow-up definitions) support findings’ robustness.
• Quote:

  "There is always a little bit of concern about the accuracy of these studies, but you know, can't get any better than that with retrospective studies.” – Dr. Barbacino [16:55]

6. Results Overview

[18:46–23:38]

• Sample Size: 4,352 patients analyzed.
• Key Finding: Lower pre-discontinuation doses (<1.25 mg/day and 1.25 mg/day) associated with significantly lower 1-year relapse risk versus 2.5 mg/day group.
• Statistical Rigor: Both crude and adjusted models yielded similar results, increasing confidence.
  • Relative risk of relapse for <1.25 mg/day group ~0.2 compared to 2.5 mg/day.
• Poisson regression confirmed association with low relapse events rate (13% at 1 year).
• Sensitivity Analyses: Findings persisted regardless of follow-up length or exclusion of levothyroxine use.
• Quote:

  “Whenever you make these adjustments, there's always going to be residual confounding. And if you make an adjustment and it shifts your crude number quite a bit to when you get to the adjusted number, then you get really worried… that's exactly what we see here.” – Dr. Hendrickson [21:10]

7. Potential Mechanisms & Biases

[23:38–27:13]

• Authors' Hypothesis: Lower maintenance dose may biologically "stabilize" immune system.
• Panel Skepticism: More likely explanation is that clinicians observe signs of remission and lower dose accordingly (“confounding by indication” and survivor bias).
• Dr. Barbacino’s Insight:

  “There is indeed a dose-dependent effect, which is also linear reverse dependent… One hypothesis is that methemazole affects the immune response—finding a sweet spot that doesn’t result in excessive TSH production and therefore excessive production of thyroid antigens.” [25:37]

8. Limitations (As Discussed by Authors and Hosts)

[27:14–29:44]

• Possible misclassification of relapse (used re-initiation of methimazole rather than direct evidence of hyperthyroidism).
• Residual confounding and survivor bias.
• Referral bias and generalizability: single-center, Japanese population.
• Limited tablet strengths in different locations (e.g., 2.5 mg tabs not available in U.S.).

9. Implications for Clinical Practice

[29:44–34:22]

• Both guests and host agree: study is high quality for retrospective design; large dataset, careful statistical handling.
• Practical Application Questions:
  • Unclear if observed dose relationship is truly causal.
  • Tapering below 5 mg/day is difficult in U.S. due to tablet limitations (2.5 mg tabs not commonly available).
  • Alternating daily dosing is cumbersome; pill splitting is impractical.
• Quotes:

  “I have to do things like every other day dosing... I’m not sure that would have the same biologic effect as taking 2.5 mg every day would.” – Dr. Crawford [32:00]
  “It's a little too soon with one single retrospective study to change practice in a way that would require, at least in the United States, creative management system.” – Dr. Barbacino [33:27]

Notable Quotes & Memorable Moments

• On the appeal of radioactive iodine:

  “It's like telling a person, you have hypertension, you can take a pill for the rest of your life, or we can fix it permanently. That’s very appealing.” – Dr. Barbacino [03:43]

• On measuring TSH receptor antibodies:

  “TSI is a real biological assay… but in the setting of the patient with hyperthyroid TSH receptor antibodies, no matter how measured, they’re almost always diagnostic…” – Dr. Barbacino [11:36]

• On the central clinical question:

  “Should this change our practice? ... I think it’s a little too soon… but I do put patients on 2.5 milligrams sometimes asking them to break the pill… occasionally.” – Dr. Barbacino [33:22]

Timestamps for Key Segments

• [02:36] – Historical shift to thionamide use
• [05:14] – Remission and relapse rates in Graves disease
• [07:53] – Study design and cohort
• [12:46] – Methimazole dosing groups and TRAb vs. TSI
• [15:49] – Confounders and statistical methods
• [18:46] – Results and statistical validation
• [23:38] – Mechanistic speculation and survivor bias
• [27:14] – Limitations of the study
• [29:44] – Discussion on quality of evidence
• [31:48] – Practical considerations and applicability

Takeaway & Clinical Pearls

• Key Finding: Lower pre-discontinuation methimazole doses (<1.25 mg/day) were associated with reduced 1-year relapse rates in Graves disease after drug withdrawal.
• Uncertainty Remains: Causality is not established—lower dose may be a marker of likely remission rather than a modifiable driver.
• Practice Change? Not yet warranted by these results; study highlights need for prospective randomized data and points to practical barriers (tablet strengths).
• Potential Policy Impact: Findings may encourage future development of lower-strength tablets and further research into optimal tapering strategies.

Summary in a Sentence:
While this robust retrospective study suggests that tapering to very low methimazole doses before discontinuation is linked to a lower relapse risk in Graves disease, the evidence stops short of supporting a practice change due to confounding factors and practical limitations in medication formulations in the U.S.

For questions or feedback on the podcast or this episode, reach out to the Endocrine Society at podcastrine.org.