Metabolic syndrome-like phenotype, BMI-based metabolic burden, and size-threshold thyroid nodules: a cross-sectional study of 50,023 health-examination attendees

Metabolic syndrome (MetS) has been linked to thyroid nodules, but most evidence relies on any-size detection dominated by clinically trivial lesions and adjusts for few confounders. Using a body-mass-index (BMI)-based operational definition (a ‘MetS-like’ phenotype, distinct from formally defined MetS), we examined its association, and that of its components and overall metabolic burden, with size-threshold thyroid nodules (≥ 1.0 cm; primary outcome) and, as an exploratory outcome, sonographically suspicious nodules (TI-RADS ≥ 4). In this cross-sectional study following STROBE, the first thyroid ultrasound of each adult attending a Beijing health-management centre (2014–2026) was analysed, with metabolic laboratory data linked by examination date (± 60 days). After excluding participants without ultrasound or with incomplete metabolic data, 50,023 formed the complete-case sample; included and excluded participants were compared. A metabolic-burden score (0–5) counted BMI-defined obesity (≥ 28), elevated blood pressure, elevated glucose, high triglycerides and low HDL-cholesterol. Odds ratios (OR; logistic) and prevalence ratios (PR; robust-variance modified Poisson) were adjusted for age and sex; components were also entered simultaneously. Robustness was assessed by the waist-based Chinese Diabetes Society (CDS) definition, additional adjustment for uric acid, C-reactive protein and thyroid peroxidase antibody in available subsets, period-stratified analysis, multiple imputation, restricted cubic splines and E-values. Nodules ≥ 1.0 cm occurred in 3,854 participants (7.7%) and TI-RADS ≥ 4 nodules in 619 (1.2%). The MetS-like phenotype was associated with nodule ≥ 1.0 cm (OR 1.52 (1.42–1.63); PR 1.45 (1.36–1.54)). BMI-defined obesity, elevated blood pressure, elevated glucose and low HDL-cholesterol remained associated when entered simultaneously, whereas high triglycerides did not. Prevalence rose from 4.4% to 12.2% across 0–5 components (trend P < 0.001). Estimates were stable across the CDS definition (OR 1.61), additional confounder adjustment, examination period, and multiple imputation that included the outcome (OR 1.50); the E-value was 2.30. TI-RADS ≥ 4 nodules were analysed only from 2020 onward, when structured grading became available. No association was seen (MetS-like OR 0.97, 0.82–1.16), but the estimate was imprecise and a small-to-moderate association could not be excluded. A higher BMI-based metabolic burden showed a graded cross-sectional association with size-threshold, but not sonographically suspicious, thyroid nodules. These associations may inform metabolic-health counselling; given the cross-sectional design they do not establish causation and do not support changing thyroid-ultrasound follow-up intensity.