In infertile women with subclinical hypothyroidism, with or without thyroid peroxidase antibodies, serum TSH during pregnancy follows preconception values and thyroid hormones remain stable.

Study question: How does subclinical hypothyroidism, defined in infertile women during preconception by thyroid-stimulating hormone (TSH) >2.5 or >4.5 mIU/l, with or without thyroid peroxidase antibodies (anti-TPO) >100 IU/ml, impact thyroid hormone levels during pregnancy and after birth?

Summary answer: During pregnancy, TSH levels remain similar to those in preconception, even with supplementary thyroxine, whereas the serum levels of anti-TPO progressively decline.

What is known already: Overt hypothyroidism impacts both pregnancy and offspring but randomized clinical trials and cohort studies failed to detect the benefit of treatment with thyroxine in cases with low-threshold TSH or with anti-TPO during pregnancy.

Study design size duration: First, the prevalence and reproducibility of two candidate cut-off levels of subclinical hypothyroidism in a cohort of 177 infertile women was compared with 171 women not aiming for pregnancy. Second, the impact of distinct setpoints of TSH in preconception (with or without anti-TPO) was monitored during pregnancy in 87 previously infertile women by high-frequency monitoring of thyroid function. Both studies were carried out from 2007 to 2019.

Participants/materials setting methods: Reproducibility and prevalence of subclinical hypothyroidism were examined in infertile women presenting in the fertility care unit of an academic institution. Women not aiming for pregnancy participated as controls. In both groups, TSH and anti-TPO were measured two times on different occasions. In addition, a group of previously infertile women with known preconception setpoints of TSH (with or without anti-TPO) were followed up prospectively throughout pregnancy and after birth. During pregnancy, serum was sampled weekly until Week 12, then monthly until delivery, and once after birth. Only cases with preconception TSH >4.5 mIU/l were supplemented with thyroxine. After collection of all samples, the serum levels of anti-TPO and the major thyroid hormones were measured. Prolactin with known fluctuations during pregnancy was used as reference.

Main results and the role of chance: Measures of both TSH and anti-TPO at two different time points were accurate and reproducible. The odds of subclinical hypothyroidism in infertile women and controls were similar. During pregnancy, TSH closely followed preconception TSH levels, whereas serum levels of the thyroid hormones predominantly remained within or above (not below) the reference. Treatment of infertile women with preconception TSH >4.5 mIU/l with thyroxine resulted in higher free thyroxine (fT4) serum levels. The serum levels of anti-TPO declined as pregnancies evolved.

Limitations reasons for caution: The numbers of participants both in the prevalence study and in pregnancy did not reach the a priori estimated numbers. For ethical reasons, the patients with preconception TSH >4.5 mIU/l were treated with thyroxine. The findings apply to infertile women only.

Wider implications of the findings: We propose to use >4.5 mIU/l as the serum TSH threshold for supplementing women with thyroxine before pregnancy. During pregnancy, fT4 may be the better marker to monitor thyroid function. The consistent decrease of anti-TPO antibody levels during ongoing pregnancies must be considered a protective element.

Study funding/competing interests: The prevalence part of this study was supported by Merck-Serono, Geneva (TH006/EMR200007-603). The hormone measurements of the serum samples collected during the follow-up pregnancies were made possible by financial support of Roche Diagnostica (November 1721, 2017, Rotkreuz, Switzerland). I.D.G. was supported by a grant of the Repronatal Foundation, Basel, Switzerland. All authors declare no conflict of interest.

Trial registration number: Research Database of UniBasel, project no. 576691 (2007).