TSH, Thyroid Hormone, and PTC—Response

We thank Drs. Franceschi and Rinaldi for their commentary (1) on our interpretation of the observed different associations between thyroid-stimulating hormone (TSH) levels outside the normal range and risk of papillary thyroid cancer (PTC) by gender in our article (2). In accordance with the findings from Rinaldi and colleagues (3), our study reported an inverse association between prediagnostic serum TSH levels within the normal range and risk of PTC. Moreover, our study found a significantly increased risk of PTC associated with TSH levels below the normal range among women and with TSH levels above the normal range among men. Drs. Franceschi and Rinaldi commented that the difference by gender could be overinterpreting due to the small number of participants with TSH levels outside the normal range and the controversial reference limits of TSH level (4). On the basis of the comments, we reanalyzed data on quartiles of TSH levels. As shown in Table 1, although the risk of PTC was decreasing with an increased TSH level among the overall population (P_trend = 0.0002), there was a slightly J-shaped association. The strongest inverse association was observed in the third quartile of TSH level (OR = 0.55; 95% CI: 0.40–0.74). This result was similar to that reported by Rinaldi and colleagues (3). When the analysis was stratified by gender, the inverse association between TSH level and risk of PTC was stronger among women, while the J-shaped association was only observed among men but not among women. The gender-specific results suggested that women could be more sensitive to the effect of TSH and more susceptible to the lower TSH levels as compared with men. Although chance due to small numbers, as well as influence of different potential thyroid problems and controversial reference limits cannot be ruled out, the observed different patterns of inverse association between TSH level and risk of PTC by gender warrants further investigation.

Drs. Franceschi and Rinaldi also commented that the gender-specific analysis could obscure the subgroup results by histologic subtypes (classical vs. follicular variant PTCs) and by tumor size (≤10 vs. >10 mm). To address this comment, we performed subgroup analysis among overall population. As shown in Table 2, a significantly inverse dose–response relationship was observed between TSH level within the normal range and risk of classical PTC (P_trend = 0.0007) but not follicular variant PTC (P_trend = 0.055). TSH levels within the normal range were inversely associated with PTC >10 mm (P_trend = 0.0001) but not PTC microcarcinoma (P_trend = 0.25). Since there could be a potential different gender effects on the association between TSH level and risk of overall PTC, to report the subgroup results by gender may provide more information regarding to the gender effects on different histologic subtypes and tumor sizes.

We appreciate the insightful comments provided by Drs. Franceschi and Rinaldi. We believe that these additional results support a potential gender difference in the association between TSH levels and PTC risk. Further studies among different populations with larger sample size are needed to confirm the association. In addition, our study together with the study by Rinaldi and colleagues (3) indicate a potential need for reevaluation of TSH reference limits.

See the original Letter to the Editor, p. 227