Objectives: Familial clustering of a disease is an indicator of a possible heritable cause, provided that environmental sharing can be excluded. Thus, data on familial risks are important for genetic studies and for clinical genetic counseling. Design: We carried out a nationwide family study on nontoxic and toxic nodular goiters, and Graves' disease in order to search for familial clustering of these diseases at the population level. Methods: The Swedish Multigeneration Register on 0-75 year old subjects was linked to the Hospital Discharge Register from years 1987 to 2007. Standardized incidence ratios (SIRs) were calculated for offspring of affected parents and for siblings by comparing to those whose relatives had no hospitalization for thyroid disease. Results: The number of hospitalized patients in the offspring generations was 11 659 for nontoxic goiter, 9514 for Graves' disease, and 1728 for toxic nodular goiter. Familial cases accounted for 8.2, 5.2, and 2.1% of all patients respectively. The highest familial risk for offspring of affected parents was noted for Graves' disease (SIR 3.87), followed by toxic nodular goiter (3.37) and nontoxic goiter (3.15). Familial risks were higher for affected siblings: toxic nodular goiter (11.66), Graves' disease (5.51), and nontoxic goiter (5.40). Weaker familial associations were observed between the three diseases. Conclusions: To our knowledge this is a first population-based family study on these thyroid diseases. The observed high familial aggregation for defined thyroid diseases cannot be explained by the known genetic basis, calling for further studies into genetic and environmental etiology of thyroid diseases.