Changes in the incidence of thyroid cancer between 1991 and 2005 in Italy: a geographical analysis

The incidence of thyroid cancer (TC) has been increasing over the last 30 years in several countries, with some of the worldwide highest TC incidence rates (IRs) reported in Italy. The objectives of this study were to evaluate by histological subtypes the geographical heterogeneity of the incidence of TC in Italy and to analyze recent time trends for papillary thyroid carcinoma (PTC) in different cancer registries (CRs).The study included cases of TC (85 years of age) reported to 25 Italian CRs between 1991 and 2005. Age-standardized IRs were computed for all histological subtypes of TC according to CRs. Estimated annual percent change and joinpoint regression analysis were used for analysis of PTC.In women, IRs of PTC ranged between 3.5/100,000 in Latina and 8.5/100,000 in Sassari for the period 1991-1995 (a 2.4-fold difference) and between 7.3/100,000 in Alto Adige and 37.5/100,000 in Ferrara for 2001-2005 (a 5.1-fold difference). In men, IRs ranged between 0.7/100,000 in Latina and 3.4/100,000 in Sassari for the period 1991-1995 (a 4.9-fold difference) and between 2.0/100,000 (Alto Adige, Trento) and 10.6/100,000 in Ferrara for 2001-2005 (a 5.3-fold difference). In both sexes, IRs significantly higher than the pooled estimates emerged for the most recent period in the majority of CRs located within the Po River plain and in Latina, but they were lower in the Alpine belt. For women, CRs reported higher IRs than pool estimates showed, between 1991 to 2005, a significantly more marked annual percent change (+12%) than other CRs (+7%). For men the corresponding estimates were +11% and +8%.The distribution of PTC does not lend support to a role of environmental radiation exposure due to the Chernobyl fallout, iodine deficiency, or (volcanic) soils. Between 1991 and 2005, wide geographic variations in the incidence of PTC and heterogeneous upward trends emerged, suggesting that the heterogeneity was a relatively recent phenomenon; this appeared to be mainly explained by variations, at a local level, in medical surveillance.