We performed the subclavian perivascular approach to the brachial plexus using contrast medium to confirm the location of the tip of the needle and the spread of the injected solution to obtain a high success rate and to minimize the risk of pneumothorax. Review of the cases led to the hypothesis that the solution injected inside the costal attachment of the middle scalene muscle spreads into the interscalene space. Because of the difference in the placement of the tip of the needle using our technique and the supraclavicular approach, including the subclavian perivascular approach, we termed our technique the supracostal approach. We conducted the present study to establish the supracostal approach by proving this hypothesis. A total of 173 blocks in 149 adult patients were studied. Eighty-four blocks in 74 patients were achieved by using the supracostal approach with contrast medium. The needle was inserted 1 cm lateral to the palpated subclavian artery and 1-2 cm above to the clavicle to touch a specific part of the first rib, which we believed to correspond to the inside of the costal attachment of the middle scalene muscle. After injecting the anesthetic solution with contrast medium, radiographs were obtained for each block, while computed tomographic (CT) studies were performed for five blocks. Five blocks in five patients were achieved by using the subclavian perivascular approach with contrast medium and both radiographs and CT studies. In addition, the anatomical difference between the two approaches was evaluated in five adult cadavers. Based on these studies, we determined the proper part of the first rib that corresponded to the inside of the costal attachment of the middle scalene muscle. Eighty-four blocks in the remaining 70 patients were performed with the supracostal approach without contrast medium. Of the 84 blocks with contrast medium, 80 (95%) produced successful blockade defined by sensory and motor examination. The radiological studies showed that, with the supracostal approach, the injected solution, which spread from the middle scalene muscle into the interscalene space, did not spread below the first rib. However, with the subclavian perivascular approach, the solution was confined within the perineural sheath and spread below the first rib to the axilla. The anatomical studies could explain this difference, revealing that the perineural space of the brachial plexus is not identical to the interscalene space. There was no failure in the 84 blocks performed with the supracostal approach without contrast medium after we determined the proper part of the first rib. We conclude that the supracostal approach to the brachial plexus is reliable, easy to perform, and associated with a low complication rate. CONCLUSIONS A new fluoroscopically guided approach for brachial plexus block has been established on the basis of anatomical and radiological studies to be reliable, easy to perform, and associated with a low complication rate.