The controversy in the management of regional lymph nodes in patients with differentiated thyroid cancer is discussed on the basis of a review of the literature. Since no prospective studies have yet compared limited dissections ('node picking') with more extensive dissections [(modified) radical neck dissection], a retrospective analysis was performed using two patient groups in which patients were managed differently with regard to the preoperative diagnosis and treatment of regional lymph node metastases. Only patients with proven lymph node metastases were included in the study. Because of selection methods necessary to create comparable patient groups, only 83 patients could be included in the analysis. There was no difference in survival or recurrence rate in either group, although recurrences occurred less frequently in the explored side of the neck after MRND (3.9% vs. 6.3%). More postoperative morbidity was found in the patients who had been subjected to a more extensive search for and treatment of lymph node metastases. Because of the relatively small number of patients only the difference in occurrence of accessory nerve palsies reached statistical significance (P = 0.05). It is advocated that only in the case of papillary carcinoma with limited lymph node involvement node picking is the procedure of choice. In all other cases a modified radical neck dissection should be standard treatment.