The aim of this retrospective study was to analyze the incidence of further nonsentinel node metastases at completion lymphadenectomy of the groin after a positive sentinel node biopsy to evaluate whether radical dissection remains the treatment of choice for these patients. Patients treated at the National Cancer Institute of Milan between January 1999 and December 2006 were reviewed retrospectively. The analysis included patients with a diagnosis of positive sentinel node biopsy of the groin (clinically negative) who underwent completion groin, iliac, and obturatory dissections. The primary melanoma was located on the lower extremities and trunk in 82.5 and 17.5%, respectively. The median follow-up was more than 30 months. The number of positive sentinel nodes was considered, as well as the size and location of the metastases (micro vs. macro). After radical dissection, the number and the location (groin, iliac, or groin+iliac nodes) of further nonsentinel node metastases were analyzed. The frequency of further nonsentinel node metastases at completion of groin dissection was correlated to Breslow's thickness and to the characteristics of the positive sentinel node biopsy. A total of 1581 patients with primary melanoma (>1 mm, or Clark's IV-V) underwent lymphatic mapping and sentinel node biopsy: 752 patients had sentinel node biopsy at the groin basin; among these, 150 (20%) patients presented positive sentinel node biopsy and underwent completion radical dissection (groin, obturatory, and external iliac+obturatory radical node dissections). We found further positive nonsentinel node metastases in 36 of 150 (24%) patients, 69% (25 of 36) of which were located in the iliac-obturator area and 31% in the groin area only: 16 patients (44.4%) had one additional metastatic node and seven patients (19.4%) had two, whereas 13 (36.1%) had three or more. In 22 cases (61.1%), the sentinel node showed a macrometastasis (>2 mm deposit in the node) and in 14 cases (38.9%) a micrometastasis (<2 mm deposit). In conclusion, there is clear evidence that patients with a positive sentinel node biopsy could have further positive nonsentinel node metastases (in our series, 24%). Although it is well known that the impact of sentinel node biopsy on survival of melanoma patients has yet to be defined, to obtain a clear nodal basin and regional control a groin+iliac-obturatory radical node dissection remains an appropriate procedure in the presence of a positive sentinel node biopsy at the groin level. This could be considered a standard treatment until new data, provided by ongoing studies, indicate new parameters for selecting patients for completion lymph node dissection.