I have some concerns about RCT by Wang et al., (1). First, there are several significant baseline characteristics. I re-calculated p-value using (Interactive Statistics — One-way ANOVA from Summary Data) to check the integrity of RCT from summary data (2). Inconsistencies and significant differences exist after recalculations of p-values of total thyroxine, free thyroxine, testosterone, and antithyroperoxidase antibody levels. This raises a concern of potential data falsification (2), as well raises a concern about validity of randomization process (3). Although the authors reported significant difference in only TSH levels, they described it as “a little higher median value” in their discussion. However, it is not “little higher” difference, it is a wide gap with high level of significance (P < 0.0001). This raises a concern about the validity of study findings as well raises doubts about the validity of randomization process (3).
Second, major discrepancies in data exist between the abstract presented in European Society of Human Reproduction and Embryology (ESHRE) conference-2017 (4) and JAMA-RCT. Discrepancies exist in the patient enrollment period. In the ESHRE-abstract, patient enrollment was reported “Between July 1st, 2012, and Nov 30th, 2016” while in the JAMA-RCT patients were enrolled “September 6,2012, and June15, 2016”. Discrepancies also exist in clinical outcomes as well flow of participants.
Third, this RCT was retrospectively registered 15 months after patient enrollment. For RCTs, pre-registration is an essential requirement (5). Retrospective registry of RCT after start is a cause of concern (6).
Fourth, the authors estimated, according to their pilot study, up to 150-180 participants could be screened out yearly. To reach the sample of 1035 screened patients, they may need two more years period of patient enrollment. This may raise doubts about the number of patients reported in this study.
More details of data discrepancies and my concerns are attached. I also attached the abstract book of ESHRE conference where the study was presented with completely different data (P-708, Page 459‐460).
I declare no competing interests.
Many thanks
References
1- Wang H, Gao H, Chi H, et al. Effect of Levothyroxine on Miscarriage Among Women With Normal Thyroid Function and Thyroid Autoimmunity Undergoing In Vitro Fertilization and Embryo Transfer: A Randomized Clinical Trial. JAMA. 2017;318(22):2190-2198.
2- Nielsen J, Bordewijk EM, Gurrin LC, Shivantha S, Flanagan M, Liu S, Linn MM, Zhou KX, van Eekelen R, Brown NJL, Thornton J, Mol BW. Assessing the scientific integrity of the collected work of one author or author group. J Clin Epidemiol. 2025 Apr;180:111603.
3- Nielsen J, Mol BW. Trustworthiness of studies investigating umbilical cord clamping and milking. Lancet. 2025 Jan 11;405(10473):125-126.
4- Wang H, Chi H, Gao H, Li R, Zeng LI, Mol BW, Hong T, Qiao J. Pregnancy Outcomes Study in euthyroid women with Thyroid Autoimmunity after Levothyroxine intervention-an open labelled RCT study (The POSTAL study). Human reproduction (Oxford, England), 2017, 32, i459‐i460.
5- Khan KS; Cairo Consensus Group on Research Integrity. International multi-stakeholder consensus statement on clinical trial integrity. BJOG. 2023 Aug;130(9):1096-1111.
6- Mousa A, Flanagan M, Tay CH, Norman RJ, Costello M, Li W, Wang R, Teede H, Mol BW. Research Integrity in Guidelines and evIDence synthesis (RIGID): a framework for assessing research integrity in guideline development and evidence synthesis. eClinicalMedicine, Volume 74, 102717.