Prevalence and Distribution of Goiter Among Commercially Insured,Working-Age Adults in the United States

Background

Goiter has been linked to iodine deficiency, female sex, older age, environmental exposure, autoimmune disease, genetic susceptibility, and socioeconomic disparities.^1–5 Iodine deficiency is the most important risk factor for goiter and was historically endemic in landlocked regions of the United States, where rates declined sharply following iodization of dietary salt in the 1920s.^6,7 The most recent national estimate of goiter prevalence in the U.S. dates to the Ten-State Nutrition Survey (1968–1970), which identified a prevalence of 31 per 1000 adults and children by physical examination of 36,000 participants. ⁸ Although researchers have continued to monitor goiter in other regions, particularly in low- and middle-income countries where iodine deficiency remains endemic,^9,10 no comparable estimates have been reported in the U.S. To begin to address this gap, we report a cross-sectional analysis of goiter prevalence and distribution among working-age adults in a large commercially insured U.S. cohort.

Methods

Data were derived from the Merative™ MarketScan^® Commercial Claims and Encounters Database, a commercially available database aggregating employer-sponsored insurance claims from participating insurers across all U.S. states, obtained through the Center for Health and the Social Sciences at the University of Chicago. The cohort was restricted to adults ages 18–64.

Cases were identified using an exhaustive set of ICD-10 codes for iodine-deficiency goiter (E01), non-toxic goiter (E04), and toxic goiter (E05) over a two-year lookback period (January 1, 2022 through December 31, 2023). Codes for solitary thyroid nodule were excluded from the case definition but permitted as co-occurring diagnoses. Individuals with CPT codes for thyroidectomy on or before December 31, 2023, were excluded as treated disease. Point prevalence was calculated as the number of individuals meeting the case definition on December 31, 2023, divided by total enrollment on that date and estimated under three nested case definitions (≥1, ≥2, or ≥2 qualifying codes ≥30 days apart). The strictest definition was used for further analysis. Regional prevalences were age-standardized to the overall MarketScan population. Multivariable negative binomial regression modeled associations with age, sex, U.S. Census region, toxic status, and ultrasound utilization. All analyses were performed in R version 4.4.0.

Results

Overall goiter prevalence

On December 31, 2023, 19,606,273 individuals were enrolled in commercial insurance plans captured in MarketScan. 15,207,303 were ages 18–64 and comprised the denominator. Within the two-year lookback period, 385,490 individuals had ≥1 qualifying ICD-10 code for goiter, 255,888 had ≥2 qualifying codes, and 203,291 had ≥2 qualifying codes on service dates separated by ≥30 days. Excluding individuals with a CPT code for thyroidectomy on or before the index date, the estimated prevalence of goiter was 24.6 per 1000 persons based on a case definition requiring ≥1 qualifying code, 16.1 per 1000 persons based on a case definition of ≥2 qualifying codes, and 12.7 per 1000 persons based on a case definition of ≥2 qualifying codes ≥30 days apart (Fig. 1).

OPEN IN VIEWER

FIG. 1.

Case definition.

Using the strictest case definition as the primary estimate (≥2 codes ≥30 days apart), females accounted for 85.1% of prevalent cases. The median age was 51 (IQR 41–58), and prevalence was noted to increase steadily with age, peaking at 36.0 per 1000 females ages 60–64 and 8.9 per 1000 males ages 60–64. 59.6% of individuals with goiter (n = 114,872) also had at least one ICD-10 code for thyroid nodule during the lookback period.

Regional variation in overall goiter prevalence

Age-adjusted regional prevalence of any goiter was highest in the U.S. Northeast (17.5 per 1000), followed by the South (13.1), Midwest (10.8), and West (9.7) (Fig. 2). To assess whether regional differences in goiter prevalence reflected differences in clinical practice, ultrasound and thyroidectomy utilization were examined as markers of diagnostic and treatment intensity across regions. Neck ultrasound was performed in 72.1% of individuals during the lookback period (n = 139,066), with similar proportions across regions. Among individuals excluded from prevalence calculations due to thyroidectomy during the lookback period (n = 10,442), rates of surgery ranged from 3.9% of goiter cases in the Northeast to 6.2% in the Midwest. Consistent with only modest regional variation in clinical practice, after multivariable negative binomial regression adjusting for age, sex, toxic status, and ultrasound utilization, region remained a significant predictor of goiter prevalence (Fig. 2).

OPEN IN VIEWER

FIG. 2.

Regional variation and age- and sex-specific prevalences of goiter.

Conclusion

This cross-sectional analysis of a large national claims database of 15.2 million commercially insured, working-age adults ages 18–64 identified an overall goiter prevalence of 1.3–2.5%. A majority of these cases were non-toxic multinodular goiters. The study also reaffirmed well-established associations of female sex and advancing age with the development of goiter.

One of the landmark studies of thyroid epidemiology in iodine-sufficient populations was the Whickham survey, which assessed thyroid size by physical exam in 2779 adults from 1972 to 1974 in a mixed urban and rural population in the United Kingdom. The survey reported a goiter prevalence of 8.6%, which declined to 7% in a 1995 follow-up of the cohort.^11,12 The lower prevalence observed here may reflect a cumulative floor effect, as restriction to adults under 65 omits the demographic in which structural thyroid abnormalities peak, while restriction to commercially insured adults omits uninsured or underinsured populations who may carry a disproportionate disease burden. Globally, more recent studies have focused on populations with iodine deficiency and on the effects of iodine supplementation.^13,14 Closest in aim to the present analysis, a 2021 national survey by Liu et al. using ultrasonography of 77,933 adults from 31 provinces in China reported a comparable goiter prevalence of 1.2% in adults, with higher prevalence among women and the elderly and in individuals with low urinary iodine concentrations. ¹⁵

Against this backdrop, two additional findings deserve attention. First, a small but persistent prevalence of ICD-10-coded iodine-deficiency goiter was observed. Whether these codes represent true iodine-deficiency goiter or misclassification is uncertain, particularly given the absence of urinary iodine measurements and the fact that iodine deficiency is rarely biochemically confirmed in routine clinical practice. Nonetheless, the finding is directionally consistent with NHANES data documenting gradually declining urinary iodine concentrations and a rising risk of iodine deficiency, particularly among women and racial and ethnic minorities. ¹⁶ Second, regional variation remained independently associated with goiter prevalence after multivariable adjustment, with the Northeast exhibiting substantially higher rates than other U.S. Census regions. These findings may reflect database artifact, including detection bias or differential insurer enrollment across regions, but they also generate the hypothesis that environmental or social determinants of goiter in the United States remain incompletely understood.

Several additional limitations warrant consideration. ICD-10 goiter codes do not require standardized thyroid volume measurement, such that the same code may capture both a small, incidentally detected multinodular gland and a massively enlarged, symptomatic substernal goiter, precluding stratification by clinical significance. Furthermore, TSH values are unavailable in administrative claims data, limiting more accurate characterization of functional status. Finally, and perhaps most importantly, the MarketScan database provides a cross-sectional snapshot that does not permit assessment of temporal trends, limiting evaluation of changing iodine status and goiter epidemiology over time.

In summary, this analysis provides the first large-scale, contemporary estimate of goiter prevalence in a U.S. population. Future work using public-payer claims datasets and more granular patient cohorts can help to validate these findings, further characterize goiter prevalence and morphology across the full spectrum of the U.S. population, and investigate potential environmental and social determinants contributing to the goiter burden.

Authors’ Contributions

J.T.: Conceptualization (lead), statistical analysis and interpretation of data (lead), writing of article (lead), and critical revision of article (equal). J.K.: Conceptualization (equal), data acquisition (lead), statistical analysis and interpretation of data (equal), and critical revision of article (equal). K.K.: Conceptualization (equal), statistical analysis and interpretation of data (supporting), and critical revision of article (equal). S.A.A.: Writing of article (supporting) and critical revision of article (equal). T.M.: Writing of article (supporting), and critical revision of article (equal). K.M.: Statistical analysis and interpretation of data (supporting), and critical revision of article (equal). X.M.K.: critical revision of article (equal), supervision (supporting). P.A.: Critical revision of article (equal) and supervision (supporting). M.K.A.: Conceptualization (lead), drafting of article (supporting), critical revision of article (lead), administrative and material supportive (lead), and supervision (lead).

Supplemental Material