First-trimester thyroid reference intervals in twin pregnancy: establishment, diagnostic utility, and association with adverse outcomes.

Objective: This large retrospective study aims to establish twin pregnancy-specific reference intervals for serum TSH and FT4 in early gestation and evaluate their clinical utility. Methods: This study analyzed TSH and FT4 levels in early pregnancy among 46,474 singleton and 1,183 twin pregnancies at a tertiary hospital. Local gestation-specific reference intervals for TSH and FT4 were established by calculating the 2.5th to 97.5th percentiles. Chi-square tests were used to compare abnormal detection rates, univariate regression analyzed differences in thyroid dysfunction prevalence, and multivariate regression assessed their predictive value for adverse pregnancy outcomes. Stratified, interaction, and marginal analyses were performed to evaluate the consistency and clinical relevance of the association. Results: Twin pregnancies showed lower median TSH (0.62 vs. 1.17 mIU/L) and higher FT4 (15.40 vs. 14.80 pmol/L) than singletons. Twin-specific intervals can reduce abnormal TSH/FT4 detection (P<0.001). Based on the twin-specific intervals, low TSH was significantly associated with preeclampsia, whereas low FT4 was linked to SGA, LGA, emergency cesarean delivery, and fetal distress (P<0.05). Multivariable logistic regression confirmed that reduced FT4 independently predicted risks of SGA and LGA. Interaction analyses showed the FT4-growth abnormality association was consistent across different maternal characteristics. Marginal value analysis showed that FT4 levels between P2.5 and P5 did not significantly increase risks of SGA or LGA, supporting the clinical validity of the current reference range. Conclusion: Twin-specific thyroid reference intervals in early pregnancy can improve the accuracy of high-risk pregnancy identification and show strong potential for clinical application.