Thyroid Hormone Shifts in Pregnancy May Signal Rising Insulin Resistance in PCOS Mothers

Pregnancy rewires almost every hormonal system in the maternal body, but few of those rewirings are as quietly consequential as the ones involving the thyroid. A new analysis of 284 pregnancies in women with polyendocrine metabolic ovarian syndrome (PMOS), the condition long known as polycystic ovary syndrome or PCOS, suggests that the trajectory of thyroid hormones across gestation tracks closely with the trajectory of insulin resistance. The study, published in Frontiers in Endocrinology, found that women whose thyroid-stimulating hormone rose only modestly during pregnancy, and whose free thyroxine levels sat at the lower end of the spectrum, went on to show higher glucose, insulin and HOMA-IR values later in gestation. The findings hint that the thyroid axis may act as an early warning system for metabolic deterioration in a group of women already at elevated risk of gestational diabetes.

The research team, led by Anastasia Trouva of Karolinska Institutet together with colleagues in Sweden and Norway, did not set out to run a new trial. Instead, they mined data from two earlier randomized, double-blind, placebo-controlled studies, the Pilot study and PregMet, in which pregnant women with PMOS had been assigned to metformin or placebo from the first trimester until delivery. Blood samples had been drawn at inclusion, between gestational weeks 5 and 12, and again at weeks 19, 32 and 36. That unusually dense sampling schedule allowed the investigators to reconstruct how thyroid function evolved over time and to test whether that evolution predicted subsequent metabolic markers, rather than simply correlating two snapshots taken at the same moment.

The metabolic readouts in the analysis were the standard currency of insulin resistance research. Fasting and post-load glucose were measured during 75-gram oral glucose tolerance tests at baseline, week 19 and week 32, with an additional fasting glucose at week 36. Fasting insulin was quantified centrally by ELISA, and the two values were combined into the homeostatic model assessment of insulin resistance, HOMA-IR, a proxy index calculated as fasting glucose multiplied by fasting insulin and divided by 22.5. Thyroid function was characterized through serum TSH and free thyroxine, with thyroid peroxidase antibodies assessed to flag autoimmune thyroid disease. None of the participants had overt hypothyroidism, and TSH values stayed within the manufacturer’s reference range throughout.

At entry into the trials, the women carried a mean body mass index of 28.9, at the upper edge of the overweight range, and a mean HOMA-IR of 3.2, a value generally interpreted as insulin resistance in non-pregnant populations. That baseline metabolic burden is not incidental. Insulin resistance is a pathophysiological cornerstone of PMOS, present in most affected women independently of body weight, and it compounds the natural insulin resistance of pregnancy, which intensifies sharply after mid-gestation as placental hormones rise. Women with PMOS face roughly a threefold increased risk of gestational diabetes compared with their peers, making any signal that tracks metabolic decline in this group clinically interesting.

The central result of the new analysis is an inverse one: the smaller the rise in TSH across pregnancy, the higher the glucose levels that followed. A one milliunit-per-liter smaller increase in TSH between baseline and week 19 corresponded to post-load glucose at week 19 that was 0.24 millimoles per liter higher. The same pattern held later in pregnancy, with smaller TSH increases from baseline to week 32 and to week 36 predicting higher fasting glucose at week 36. Free thyroxine told a parallel story. Lower fT4 at baseline predicted higher fasting insulin at weeks 19 and 36, and lower fT4 at weeks 19 and 32 predicted higher insulin and HOMA-IR at week 36. A smaller TSH rise was also linked to higher HOMA-IR at week 32.

Crucially, these associations emerged independently of whether the women were taking metformin or placebo. The statistical models included treatment group as a covariate, and interaction tests found no evidence that the drug modified the thyroid-metabolism links. Metformin did leave its own fingerprint: the treated group showed significant early-pregnancy reductions in insulin and HOMA-IR that the placebo group did not. But by the later trimesters the two arms converged, echoing the consistent finding from prior trials that metformin, despite its insulin-sensitizing credentials, has not reduced the incidence or severity of gestational diabetes in women with PMOS. The thyroid signal, in other words, persisted beneath and beyond the pharmacological noise.

Why would a blunted TSH rise or lower thyroxine foreshadow worsening insulin resistance? The physiology is plausible even if causality remains unproven. Thyroid hormones are master regulators of metabolic rate, stimulating hepatic glycogenolysis and gluconeogenesis, promoting adipose lipolysis, and modulating pancreatic secretion of insulin and glucagon. They also influence hepatic glucose production through a sympathetic neural pathway running from the hypothalamic paraventricular nucleus to the liver. In a healthy pregnancy, TSH drifts slightly upward from the first trimester toward term while free thyroid hormones dip minimally. A failure to mount that normal TSH increase, or a relatively low free thyroxine level, could reflect a thyroid axis that is not keeping pace with the metabolic demands of gestation, leaving glucose regulation to strain against a rising tide of placental insulin resistance.

The authors are careful to frame the work as exploratory and hypothesis-generating rather than practice-changing. Because the analyses are observational in nature, with no randomization to thyroid function, residual confounding cannot be excluded, and body mass index absorbed a substantial share of the associations in adjusted models, a reminder that adiposity shapes both thyroid physiology and insulin action. The researchers also note methodological caveats: glucose was measured locally at participating hospitals while insulin and thyroid markers were assayed centrally, trimester-specific reference ranges for thyroid hormones were unavailable, and immunoassay-based free thyroxine measurements can be influenced by the pregnancy-induced changes in thyroid-binding proteins. Whether thyroid parameters add predictive power beyond established risk factors such as BMI and baseline HOMA-IR remains an open question for future studies.

Even with those caveats, the study adds a distinctive piece to a growing puzzle. Previous work in non-PMOS populations has linked declining or low-normal free thyroxine to less favorable metabolic profiles and to gestational diabetes, and hypothyroxinemia in the second trimester has been associated with the condition. What is new here is the longitudinal, within-pregnancy dimension in a high-risk group, and the observation that the change in TSH, not merely its absolute value, carries information. A prior study in non-pregnant women with PMOS found that TSH at or above 2 milliunits per liter was associated with insulin resistance, yet the women in the current cohort mostly sat below that threshold, suggesting that even subtle shifts within the normal range may matter when metabolic vulnerability is already high.

The practical implications are likely to unfold slowly. If confirmed in prospective cohorts designed to test mechanism and prediction, thyroid trajectories could eventually complement existing metabolic markers in identifying pregnant women with PMOS who are drifting toward glucose intolerance, opening a window for earlier intervention in a condition where prevention has proven stubbornly difficult. For now, the message is one of biological insight rather than clinical prescription: in pregnancy, the thyroid and the pancreas appear to be in constant conversation, and in women with PMOS, a quieter thyroid conversation may foreshadow a louder metabolic one.

Subject of Research: Associations between maternal thyroid function and insulin resistance during pregnancy in women with polyendocrine metabolic ovarian syndrome

Article Title: Maternal thyroid function and subsequent markers of insulin resistance in pregnant women with polyendocrine metabolic ovarian syndrome

Article References: Trouva, A., Vanky, E., Alvarsson, M., Calissendorff, J., Åsvold, B. O., Ujvari, D., & Hirschberg, A. L. (2026). Maternal thyroid function and subsequent markers of insulin resistance in pregnant women with polyendocrine metabolic ovarian syndrome. Frontiers in Endocrinology, 17, Article 1884684. https://doi.org/10.3389/fendo.2026.1884684

Image Credits: AI Generated

DOI: 10.3389/fendo.2026.1884684

Keywords: thyroid function, insulin resistance, pregnancy, PCOS, PMOS, gestational diabetes, TSH, free thyroxine, HOMA-IR, metformin, endocrinology, glucose metabolism