A history of benign thyroid diseases has been associated with the risk of thyroid cancer. We have analyzed this issue using data from a case-control study conducted in northern Italy between 1986 and 1992 on 399 incident, histologically confirmed thyroid cancer cases and 617 controls admitted to the hospital for acute, nonneoplastic, non-hormone-related diseases. The overall multivariate relative risk (RR) estimates were 2.8 [95% confidence interval (CI), 0.6-12.4] for previous episodes of thyroiditis, 27.1 (95% CI, 6.5-111.9) for adenoma, 8.2 (95% CI, 3.5-19.1) for goiter, 3.8 (95% CI, 1.4-10.9) for hyperthyroidism, and 1.5 (95% CI, 0.4-5.1) for hypothyroidism when all histotypes were analyzed. The RR for any thyroid disease was 7.7 (95% CI, 4.6-12.8). A family history of thyroid disease was significantly related to thyroid cancer with an RR of 1.6. The RR for having resided in endemic goiter areas was 1.3 for < 20 years of residence and 1.6 for 20 or more years. These associations were somewhat stronger when only papillary, follicular, and mixed papillary/follicular cancers were considered. Analyses of data in separate strata of sex and age suggested that several benign conditions play a more important role in females and in subjects younger than 50 years. Results were similar to the overall ones when papillary and follicular carcinomas were considered separately. The population-attributable risk for any previous thyroid disease was approximately 20% in this Italian population. These results confirm that history of thyroid disease is a relevant indicator of subsequent thyroid cancer risk also in areas at relatively low prevalence of goiter and other thyroid diseases.