Euthyroid diabetes study finds no independent thyroid-nodule link

In a hospital-based cross-sectional study of 934 adults with type 2 diabetes whose thyroid function tests were normal, a lower free triiodothyronine-to-free thyroxine ratio was initially associated with thyroid nodules, but the association disappeared after adjustment for age, sex and other clinical factors. A nonlinear statistical analysis suggested that estimated nodule odds may be higher at the low end of the ratio distribution, although confidence intervals were wide. Researchers therefore found no independent association under the conventional linear model and characterized the nonlinear pattern as hypothesis-generating rather than conclusive.

Thyroid nodules are often discovered through ultrasonography, even when they cause no symptoms, and most are benign. Their development can involve genetic susceptibility, iodine status, thyroid hormone biology, insulin resistance and chronic inflammation, several of which are common in type 2 diabetes. The study focused on the ratio of free triiodothyronine, or FT3, to free thyroxine, or FT4, both circulating forms of thyroid hormone. The ratio is used as an indirect marker of peripheral conversion of T4 into the more active hormone T3, although its physiological meaning can vary with obesity, inflammation and systemic illness.

Researchers enrolled patients admitted to Tianjin Medical University General Hospital between 2021 and 2023. Participants were between 18 and 80 years old and had type 2 diabetes confirmed using World Health Organization criteria. Each underwent anthropometric measurements, physical examination, laboratory testing and thyroid ultrasonography. The analysis excluded people with type 1 or other specific forms of diabetes, acute diabetic complications, severe hypoglycemia, several cancers and blood disorders, prior thyroid disease or surgery, thyroid-related antibodies, and use of medicines known to affect thyroid function.

Laboratory reference ranges were 2.43–6.01 pmol/L for FT3, 9.01–19.05 pmol/L for FT4 and 0.35–4.96 μIU/mL for thyroid-stimulating hormone, or TSH. Patients with values within all three ranges were considered euthyroid. The FT3/FT4 ratio was calculated directly from FT3 and FT4 concentrations, while experienced thyroid specialists used a 10-MHz linear-array transducer to assess nodule number, size, location, echogenicity, margins, cystic components and Thyroid Imaging Reporting and Data System category. Because cytology or histopathology was not routinely available, the presence of a thyroid nodule was defined by ultrasonography rather than pathological examination.

Thyroid nodules were detected in 730 participants, or 78.2% of the cohort. The prevalence was 84.2% among women and 73.9% among men. This high proportion describes only the selected hospital cohort and should not be interpreted as the prevalence among all people with type 2 diabetes. A population-based study cited by the researchers reported a prevalence of 36.9% in China, but the two estimates are not directly comparable because the present sample consisted of hospitalized patients who may have been older, had more metabolic illness or underwent systematic ultrasound that could reveal small, asymptomatic nodules.

Patients with nodules were older than those without nodules and had lower FT3/FT4 ratios, estimated glomerular filtration rates, fasting plasma glucose, HbA1c, cholesterol measures, liver enzymes and 24-hour urinary uric acid excretion. Hypertension and carotid artery stenosis were also more frequent in the nodule group, whereas lower-extremity atherosclerotic disease was less frequent. Insulin and metformin use was less common among patients with nodules. These differences illustrate why an unadjusted statistical relationship can be misleading: age, sex, metabolic health, kidney function and medication use differed substantially between the groups.

In an unadjusted binary logistic regression, each 0.1-unit increase in the FT3/FT4 ratio was associated with lower odds of thyroid nodules, with an odds ratio of 0.65 and a 95% confidence interval of 0.45–0.95. That difference was statistically significant at P = 0.027. Once age and sex were included, however, the estimate was attenuated and no longer significant, yielding an odds ratio of 0.78 with a 95% confidence interval of 0.52–1.17. A fully adjusted model added body mass index, diabetes duration, smoking, HbA1c, insulin use, metformin use and TSH; the association remained nonsignificant at an odds ratio of 0.77, with a 95% confidence interval of 0.49–1.20.

Restricted cubic spline analysis produced a different statistical picture. After adjustment for body mass index, diabetes duration, smoking, HbA1c, insulin use, metformin use and TSH, the analysis indicated an overall association between the FT3/FT4 ratio and nodules at P = 0.036 and evidence of nonlinearity at P = 0.022. Using the cohort median ratio of 0.296 as the reference, estimated nodule odds declined sharply as the ratio rose through its lower range. Above the reference, estimated odds increased modestly before returning toward the null at higher values. Yet the 95% confidence intervals included 1.0 across most intermediate and higher values, signaling substantial uncertainty.

The ratio also varied with several metabolic and demographic measures, though all reported correlations were weak. It was positively correlated with estimated glomerular filtration rate, triglycerides, alanine aminotransferase, gamma-glutamyl transferase, urinary uric acid excretion, diastolic blood pressure and body mass index. It was inversely correlated with HbA1c and age. Women had a lower ratio than men, and participants with HbA1c below 7% had a higher ratio than those with HbA1c of 7% or more. Nodule prevalence tended to decrease across ratio quartiles, from 82.3% in the lowest quartile to 75.9% in the highest, but the differences among quartiles were not statistically significant.

Several biological interpretations remain possible, but the study did not measure the mechanisms directly. In metabolic states involving insulin resistance or energy excess, increased peripheral T4-to-T3 conversion may be adaptive; by contrast, chronic inflammation and systemic illness may suppress deiodinase activity and reduce peripheral T3 production. Thyroid nodules themselves do not necessarily indicate impaired conversion. Differences in age, sex, comorbidity, obesity, inflammation or other characteristics could generate an apparent relationship. The authors did not measure inflammatory biomarkers or deiodinase activity, so the proposed links among inflammation, thyroid hormone metabolism and nodule formation remain speculative.

The findings are also constrained by the retrospective, single-center, cross-sectional design. Measurements and nodule detection occurred at one point in time, preventing assessment of which came first and excluding causal interpretation. Reverse causation cannot be ruled out. Information on medication duration, dose, adherence and some concomitant treatments was incomplete, and the hospitalized sample limits generalization to community populations, people without diabetes and other regions. The study therefore does not show that changing the FT3/FT4 ratio affects nodule development or that the ratio can diagnose or predict nodules independently. Its clearest conclusion is that the crude inverse association disappeared after multivariable adjustment, while a possible nonlinear pattern at low ratio values requires confirmation in prospective studies with direct measures of thyroid hormone conversion, inflammation and nodule development.

Subject of Research: Association of peripheral thyroid hormone conversion indices with ultrasound-detected thyroid nodules in euthyroid adults with type 2 diabetes.

Article Title: Association between the FT3/FT4 ratio and thyroid nodules in euthyroid patients with type 2 diabetes: a cross-sectional study

Article References: Huang, B., Li, X., Xie, D., Zhang, X.-X., Li, S.-W., Tian, Y.-X., Li, C.-Y., & Cui, J.-Q. (2026). Association between the FT3/FT4 ratio and thyroid nodules in euthyroid patients with type 2 diabetes: a cross-sectional study. Frontiers in Medicine, 13, Article 1908557. https://doi.org/10.3389/fmed.2026.1908557

Image Credits: AI Generated

DOI: 10.3389/fmed.2026.1908557

Keywords: type 2 diabetes, thyroid nodules, FT3/FT4 ratio, euthyroidism, thyroid ultrasonography, peripheral thyroid hormone conversion, cross-sectional study, metabolic disease, restricted cubic splines, Association, between, ratio