Selective neck dissection (SND) is known to be a valid procedure to stage the clinically N0 neck but its reliability to control metastatic neck disease remains controversial. This study analysed if selective neck dissection is a reliable procedure to prevent regional metastatic disease in head and neck squamous cell carcinoma (HNSCC). We retrospectively analysed the medical records of 163 previously untreated patients with squamous cell carcinoma of the oral cavity, oropharynx, larynx and hypopharynx treated initially in our department from January 1990 to December 2002. All patients had unilateral or bilateral SND, in combination with surgical resection of the primary tumour. SND was performed in 281 necks. Finally, 146 patients who underwent 249 SND (39 I-III, I-IV, 210 II-IV, II-V) had adequate follow-up and were assessed for the regional control. The median follow-up was 37 months (1-180 months). The end points of the study were neck control following SND and overall survival. Twenty-five percent (30/119) of patients staged cN0 had lymph node (LN) metastasis. Overall, regional recurrence was observed in 2.8% of the necks (7/249): 1.6% (4/249) in dissected field and 1.2% (3/249) in undissected field. Seventy-eight percent (194/249) of the necks were staged pN0 with a subsequent failure rate of 1.5% (3/194); 16% (39/249) were staged pN1 and postoperative radiotherapy (PORT) was proposed in 21 of these patients. The failure rate with PORT was 9.5% and 5.5% without PORT. Six percent (16/249) of the necks were staged pN2b and all had PORT with one subsequent recurrence. Extracapsular spread (ECS) was reported in 16.5% of positive SND specimens (9/55); all by one were treated by PORT with a subsequent failure rate of 22% (2/9). At 3 years, overall survival for the whole population was 70% and statistically highly correlated with pN stage (p<0.001). These results support the reliability of SND to stage the clinically N0 neck. SND is a definitive operation not only in pN0 but also in most pN1 and pN2b necks. PORT is not justified in pN1 neck without ECS. In pN2b necks, the low rate of recurrence supports adjuvant PORT. The presence of ECS, despite adjuvant PORT, remains associated with a higher risk of recurrence.