A total of 168 primary supraclavicular decompressions were performed on 146 patients with neurogenic thoracic outlet syndrome. This report compares the results of rib resection (supraclavicular anterior and middle scalenectomy and first rib resection) with rib-sparing (supraclavicular anterior and middle scalenectomy alone) operations. All patients with cervical ribs were excluded. In total, 125 rib resections and 43 rib-sparing procedures were performed between 1983 and 1992 by a single surgeon. The patients were otherwise comparable in symptoms and physical signs. During surgery there was a significantly higher proportion of pleural injury associated with rib resection (59%) than with rib-sparing (40%) procedures. The mean hospital stay was also prolonged by 1 day in patients undergoing rib resection (p = 0.005). There was no significant difference in early success between the two groups (83% for rib resection, 91% for rib sparing) and no difference in those resuming employment (52% and 63% respectively). Life-table analysis showed that the two groups have similar long-term results (69% and 76% at 2 years). The only important factor determining clinical outcome in primary supraclavicular thoracic outlet syndrome decompression was the duration of symptoms before operation. Some 83% of patients with symptoms less that 2 years had a successful result compared with only 68% in those with symptoms longer than 2 years (p < 0.05). Spontaneous or post-traumatic neurogenic symptoms responded to operation identically. The theoretical benefit of first rib resection to relieve mechanical compression of the brachial plexus is not evident from this review. Thorough removal of the scalene musculature and other myofascial anomalies, preferably through the supraclavicular approach, leads to less patient morbidity, shortens hospitalization, and is recommended for patients with neurogenic thoracic outlet syndrome requiring operative intervention.