In this cross-sectional study of a community-based population in China showing more than adequate iodine intake, we were unable to confirm a significant association between high BMI and risk of thyroid nodules, except for the subgroup of women. In contrast, we found an independent, positive correlation between waist circumference – treated as a categorical or continuous variable – and risk of thyroid nodules. Participants with central obesity were at 1.62-fold higher risk of thyroid nodules than those with normal waist circumference, and this relationship was also observed in nearly all subgroup analyses. Our findings suggest that in Chinese individuals with more than adequate iodine intake, higher waist circumference is more strongly associated than higher BMI with elevated risk of thyroid nodules. This may mean that adipose tissue in the waist area may influence risk of thyroid nodules differently from adipose tissue elsewhere in the body.
Previous studies in different populations have reached conflicting conclusions about the association between BMI and risk of thyroid nodules. Two community-based studies in China reported that overweight and general obesity (as measured using BMI) were associated with higher risk of thyroid nodules only in women [5, 13], and we found the same result in our population when we treated BMI as a categorical variable. However, another Chinese study did not detect this association, perhaps because of insufficient sample size . A previous community-based study in China found that BMI defined as a continuous variable correlated positively with risk of thyroid nodules , but two other studies involving healthy individuals undergoing physical exams failed to detect this association [6, 7], similar to our negative result in the present study. To make things more complicated, two studies outside Asia reported a negative relationship between BMI and risk of thyroid nodules [12, 14]. In light of the literature, we speculate that BMI may not correlate linearly with thyroid nodule risk, and so it may be unsuitable for assessing the influence of adiposity on the presence of thyroid nodules. Whether this is true only for Chinese populations or more broadly requires further study.
The poor performance of BMI as an indicator of thyroid nodules may relate to the fact that it is a quite nonspecific measure of adiposity, aggregating measures of muscle mass, peripheral and abdominal adipose tissue, and bone mass . Waist circumference, in contrast, specifically reflects abdominal adipose distribution, which mainly consists of subcutaneous and visceral adipose tissue [25, 26]. This specificity may help to explain why waist circumference appears to be a better indicator of thyroid nodule risk. Central obesity has already been linked to greater likelihood of adverse metabolic health conditions [18, 27], including hyperglycemia, hypertension and dyslipidemia, which reflect a cluster of components in metabolic syndrome. Only approximately 20% of obese individuals (based on BMI) have metabolic disorders because of their smaller proportion of visceral adipose tissue . In other words, larger waist circumference appears to be a stronger risk factor than BMI for metabolic syndrome . Our finding of a strong correlation between waist circumference and thyroid nodule risk may therefore reflect the well-established correlation between metabolic syndrome and thyroid nodule risk [2, 4, 8, 9, 19]. Indeed, several of those previous studies have reported correlations between central obesity and thyroid nodule risk [4, 8, 9]. Central obesity lies at the core of metabolic syndrome , so it may not be surprising that our results show central obesity to be more closely associated with thyroid nodules than overweight and general obesity.
In our association analyses, we took into account age, gender, TSH and UIC as potential confounders. Thyroid nodules are known to be more prevalent in women [2,3,4] and older individuals [2,3,4,5,6,7,8,9], while waist circumference tends to larger in males and in the elderly. Thyroid nodule formation has been associated with TSH  and UIC [10, 11], and TSH has been associated with waist circumference . Our observation that the increased risk of thyroid nodules was markedly attenuated after adjustment for these potential confounders suggests that these covariates also contribute to overall risk. Subgroup analyses showed that central obesity was significantly and independently associated with higher risk of thyroid nodules in nearly all subgroups, while BMI ≥24 kg/m2 significantly correlated with increased risk of thyroid nodules only in women. This provides further evidence that, overall, central obesity is more strongly associated with risk of thyroid nodules than overweight and general obesity. The present study shows no evidence that gender, age, TSH, or UIC affects the observed relationship between risk of nodules and either BMI ≥ 24 kg/m2 or central obesity. Certainly, according to American Diabetes Association definition for overweight (≥23 kg/m2) and obesity (≥25 kg/m2) for Asians , higher BMI was also significantly associated with higher thyroid nodule risk among individuals with TSH > 4.2 mIU/L, and TSH played an interactive role in the association between BMI and TNs, which might deserve further researches.
We found central obesity to be significantly related to higher thyroid nodule risk in men but not in women. While further work is needed to confirm that this is not merely a sample size effect, we suggest that it may reflect sex differences in the proportion of abdominal fat components. Women tend to have larger stores of subcutaneous fat than visceral fat, while men tend to have more visceral fat than subcutaneous fat for any given waist circumference [25, 26]. This implies that increasing waist circumference represents, in men, primarily accumulation of visceral fat. Obese individuals with greater proportion of subcutaneous fat than visceral fat are at lower risk of metabolic syndrome than those with more visceral than subcutaneous fat . These findings, when taken together with our present results, suggest that waist circumference-associated visceral fat may play a key role in the development of thyroid nodules.
These considerations may point to a key role of insulin resistance in formation of thyroid nodules. Visceral fat is the strongest predictor of insulin resistance [22, 25], which is a central contributor to metabolic syndrome . One study in Italy found that while waist circumference was significantly associated with the presence of thyroid nodules, this association was no longer significant after adjusting for insulin resistance . These results suggest that insulin resistance may be associated with thyroid nodule formation more strongly than even waist circumference. Indeed, our previous study of a large population suggested that insulin resistance is associated with the distribution, construction, and density of blood vessels in thyroid nodules . Differences in such vascularization may help determine nodule growth and progression. It is possible that insulin resistance may cause changes in proliferative pathways activated directly by insulin or insulin-like growth factor-1 (IGF-1), which helps regulate thyroid gene expression and may be important in thyrocyte proliferation and differentiation [4, 30, 31]. Previous studies and the present work argue for focusing future research on the potential role of waist circumference-associated insulin resistance in the formation of thyroid nodules. The available evidence further suggests that effective diagnosis and treatment of insulin resistance may help prevent such nodules.
The present study extends previous work on associations of thyroid nodules with BMI or waist circumference to the case of a population with more than adequate iodine intake (UIC = 239 µg/L). In addition, our study was able to show that waist circumference was associated with thyroid nodules independently of TSH and UIC. These two thyroid nodule risk factors are usually ignored as potential confounders in the literature. Finally, our study systematically compared two adiposity measures, whereas most previous studies have focused on one or the other.
Nevertheless, the results of the present work should be interpreted with caution in light of several limitations. First, the cross-sectional study design does not allow us to establish causal links between obesity measures and thyroid nodules. Second, although the multivariate model adjusted for as many confounders as possible, we did not control for other components of metabolic syndrome, including hyperglycemia, hypertension and dyslipidemia. This reflects the fact that the original purpose of our study was to investigate relationships between iodine nutrition and thyroid diseases, so we did not collect information about history of hypertension, diabetes or dyslipidemia. Third, we did not examine whether the relationship of thyroid nodule risk to BMI or waist circumference depends on the specific nodule subtype.