TITLE:
Urinary Iodine Concentration, Glycated Hemoglobin, and Thyroid Autoantibodies in US Adults: A Complex Survey-Weighted Cross-Sectional Analysis of NHANES 2007-2012
AUTHORS:
Chuanzao Chen, Tongmin Ji, Jiayi Pan, Chanhong Shi
KEYWORDS:
Urinary Iodine Concentration, HbA1c, Thyroid Autoantibodies, NHANES, Complex Survey, Sensitivity Analysis
JOURNAL NAME:
Journal of Biosciences and Medicines,
Vol.14 No.10,
October
30,
2026
ABSTRACT: Objective: To examine cross-sectional associations of spot urinary iodine concentration (UIC) with glycated hemoglobin (HbA1c) and thyroid autoantibodies in US adults, and to assess the robustness of the estimates across analytic specifications. Methods: We combined three National Health and Nutrition Examination Survey (NHANES) cycles (2007-2012). The principal common sample comprised 8,278 adults aged ≥20 years, excluding known pregnancies, with required measurements. Survey weights, strata, and primary sampling units were incorporated. Log2-transformed UIC was related to HbA1c, thyroid peroxidase antibodies (TPOAb), thyroglobulin antibodies (TgAb), and positivity for either antibody. Weighted quartiles, natural splines, interaction tests, jackknife replicate weights, and sensitivity analyses were used. Selected secondary tests were corrected for false discovery rate. Results: Weighted median UIC was 141.5 μg/L; 15.26% were positive for at least one antibody. After full adjustment, each doubling of UIC corresponded to an HbA1c difference of −0.0127 percentage points (95% CI −0.0316 to 0.0061; P = 0.180). Geometric mean ratios were 0.972 (95% CI 0.926 - 1.020) for TPOAb and 0.982 (0.959 - 1.005) for TgAb; the odds ratio for either antibody positivity was 0.972 (0.895 - 1.055). For the three UIC-antibody outcomes, all Benjamini-Hochberg-adjusted P values were ≥0.368. The four-degree-of-freedom spline yielded an overall P = 0.0218 and a nonlinearity P = 0.2967; corresponding values in the stricter diabetes- and glucose-lowering-medication-exclusion sample (n = 7,084) were 0.1389 and 0.2025. Adjustment for urinary creatinine changed the magnitude or direction of the linear coefficient. Conclusions: Spot UIC showed no clearly supported adjusted linear association with HbA1c or thyroid autoantibodies under the prespecified cross-sectional models. Exploratory signals were not consistent across analytic specifications. Interpretation should therefore remain limited to cross-sectional associations rather than clinical equivalence, long-term iodine effects, or causal mediation.