Objective: We assessed the association between thyroid autoimmunity and thyroid cancer in a retrospective series of unselected thyroid nodules submitted to fine-needle aspiration cytology (FNAC) to avoid the selection bias of surgical series. Subjects and methods: Ultrasound (US)-guided FNACs were obtained from 590 unselected consecutive patients with single thyroid nodules and positive (ATA + , n = 197) or negative (ATA − , n = 393) serum anti-thyroid antibody (ATA). Cytological results were classified in three classes of increased risk of malignancy: low risk or benign (class II); indeterminate risk (class III); and suspect or malignant (class IV). Results: A higher prevalence of class III (28.9% vs 21.4%, P < 0.05) and class IV (18.8% vs 9.2%, P < 0.001) and lower prevalence of class II (52.3% vs 69.5%, P < 0.001) were found in ATA + vs ATA − nodules respectively. By multivariate logistic regression analysis ATA + conferred a significant risk (odds ratio (OR): 2.29 (95% confidence interval (CI): 1.39–3.76)) for class IV cytology independently from age and sex. In 106 patients where thyroidectomy was carried out, thyroid cancer was found in 54/61 (88.5%) patients with class IV nodules (with similar positive predictive value for cancer in ATA + (96.4%) and ATA– (81.8%) nodules), in 6/31 (19.3%) of class III nodules (all ATA − ) and in none of 14 class II nodules. Non-specific cytological atypias from hyperplastic nodules in lymphocytic thyroiditis probably accounted for the different prevalence of cancer in class III ATA + and ATA − nodules. Histologically proven thyroid cancer (mostly papillary) was then observed in a higher proportion (27/197 = 13.7%) of ATA + , when compared with ATA − nodules (33/393 = 8.4%, P = 0.044), but the significance of this finding is limited by the low number of class II nodules operated on. Conclusions: The presence of ATA + confers an increased risk of suspicious or malignant cytology in unselected thyroid nodules. Since ATA + is not responsible for increased false- positive class IV FNAC, our study provides indirect evidence supporting a significant association between thyroid carcinoma and thyroid autoimmunity, although further studies with a different design are needed for a definitive histological proof.