The aim of this study was to determine the value, advantages and limitations of ultrasound-guided fine-needle aspiration biopsy (US-FNAB) in an endemic goitre area. US-FNAB was performed on all outpatients who presented with hypoechoic and/or hypofunctional and/or growing nodules. A total of 4518 US-FNABs were performed and 718 patients from this series underwent surgery. Cytological results of the primarily performed US-FNAB of these patients were compared retrospectively with the histological results. US-FNAB results were grouped as (1) non-malignant (n = 303), (2) non-malignant follicular proliferation (n = 177), (3) malignancy cannot be ruled out (n = 133), (4) malignant (n = 61), (5) inadequate (n = 34), and (6) sampling error; biopsy of a non-malignant nodule (n = 10). Nodules as small as 5 mm in diameter could be biopsied, gaining representative material. US-FNAB found a malignant or suspicious cytology in 65 out of 87 cases with malignant histology (74.71%). Diagnosis of early tumour stages was often possible: 12 of 18 thyroid carcinomas biopsied and smaller than 10 mm in diameter had malignant or suspicious cytology (groups 3 and 4). US-FNAB was performed incorrectly within non-malignant nodules in ten patients (1.39%) with multinodular goitre (ten papillary carcinomas, nine smaller than 10 mm). Regarding the cytology of groups 1 and 2 as benign and those of groups 3 and 4 as malignant, US-FNAB performance was as follows: sensitivity 87.84%, specificity 78.50%, negative predictive values 98.13%, positive predictive values 33.51% and accuracy 79.53%. Biopsies with inadequate material were obtained in 4.73% of all biopsies. No major adverse effects occurred. Re-biopsies in 61 cases did not alter the cytological outcome in those cases where adequate material was obtained. US-FNAB is a valuable method in the pre-operative assessment of thyroid nodules in order to select patients for surgery, as malignancy can often be detected even in early tumour stages. However, even with ultrasonographic guidance, the minimal tumour size detectable by US-FNAB is around 5 mm. The cytological interpretation in cases with regression and microfollicular proliferation also sets limits on the method. However, patients with non-malignant cytologies can be followed up safely by sonography due to the high NPV of US-FNAB as long as thyroid nodules do not become larger. Re-biopsies seem to be of limited value as long as adequate material was obtained by US-FNAB.