Nocturnal sleep duration, clinical activity score, and thyroid functional remission after radioactive iodine therapy in Graves' disease with thyroid-associated ophthalmopathy.

No patient sleeping less than 6 hours at baseline achieved thyroid functional remission after 131I therapy.

• 0 of 31 patients sleeping <6h achieved remission (95% CI 0%-9.2%).
• The cohort comprised 153 patients (53 in remission, 100 not in remission), followed for at least 12 months after 131I therapy.
• The authors describe the absence of remission below 6h as an exploratory post hoc observation, not a validated threshold.

After FDR adjustment, GD duration, TAO duration, sleep duration, and CAS remained associated with remission in univariate screening.

• Univariate screening used false discovery rate (FDR) control.
• Missing continuous values were multiply imputed (20 chains).

Cross-validated LASSO retained GD duration, sleep duration, and CAS as predictors of remission.

• LASSO used the one-standard-error rule (C = 0.142).
• TAO duration was not retained by LASSO despite passing FDR screening.

Each additional hour of pretreatment nocturnal sleep was associated with substantially higher odds of thyroid functional remission.

• Post-selection logistic regression OR per sleep hour was 6.13 (95% CI 3.16-11.91; P < 0.001).
• Firth sensitivity analysis gave a consistent estimate: OR 5.64 (95% CI 2.99-10.63).

Higher Clinical Activity Score was associated with lower odds of remission.

• OR per CAS point was 0.64 (95% CI 0.48-0.86; P = 0.003) in the post-selection model.
• Mutually adjusted restricted cubic splines supported a non-linear relationship for CAS (P < 0.001).

GD duration was not significantly associated with remission in the post-selection model.

• OR per month of GD duration was 0.991 (95% CI 0.979-1.003; P = 0.152).
• GD duration had been retained by LASSO.

No significant CAS × sleep interaction was detected, and sleep showed no evidence of a non-linear relationship with remission.

• CAS × sleep interaction: Wald P = 0.096; likelihood-ratio P = 0.086.
• Mutually adjusted RCS non-linearity for sleep: P = 0.585.

The study design limits inference: it was retrospective, single-center, with a thyroid-function (not ocular) primary endpoint, and findings require prospective validation.

• 20 patients received prednisone before 131I therapy (discontinued ≥3 days prior) and 133 received none; no thyroid-directed or immunosuppressive treatment was given during follow-up.
• Remission was coded using the 2021 Chinese Society of Nuclear Medicine guideline categories (euthyroidism or hypothyroidism vs. partial or ineffective response).
• The authors state that associations are post-selection and that the fitted threshold effects are exploratory, not validated thresholds.