The shoulder block may impair ventilatory function and diaphragmatic movement less than the interscalene brachial plexus block. We randomly allocated 30 adults who underwent shoulder arthroscopy under general anaesthesia to ultrasound-guided shoulder block or interscalene block with 20 ml bupivacaine 0.5%. The primary outcome, rate of ultrasound-measured hemidiaphragmatic excursion < 25% of baseline 30 min after blockade, was reduced from 12/15 with brachial plexus block to 2/15 with shoulder block, a difference (95%CI) of 67% (40-93%), p < 0.001. The mean (SD) numeric rating scale pain scores at rest after shoulder block were higher than after interscalene block at two postoperative hours, 1.4 (1.2) vs. 0.3 (0.7), p = 0.02, but lower at 24 postoperative hours, 1.3 (1.3) vs. 3.4 (2.3), p = 0.008. Mean (SD) pain scores on movement in the shoulder and interscalene blocks were similar, with respective values of 1.9 (1.9) vs. 0.7 (1.2), p = 0.13 at two postoperative hours and 3.7 (2.3) vs. 5.3 (2.5), p = 0.41, at 24 postoperative hours.