Objectives To determine the effect of arthroscopic capsular shift surgery on pain and functional impairment for people with atraumatic shoulder (glenohumeral) joint instability.

Methods We conducted a randomised, placebo-controlled clinical trial in a specialist secondary care facility. Patients aged 18 years and over who reported insecurity (apprehension) in their shoulder joint and had evidence of capsulolabral damage on arthroscopic examination were included. Patients were excluded if their shoulder apprehension symptoms were precipitated by a high velocity shoulder injury, they had bony or neural damage, a rotator cuff or labral tear, or previous surgery on the symptomatic shoulder. Sixty-eight participants were randomised and received diagnostic arthroscopy, followed by arthroscopic capsular shift or diagnostic arthroscopy alone. All participants received the same postoperative clinical care. The primary outcome was pain and functional impairment measured with the Western Ontario Shoulder Instability Index. The prespecified minimum clinically important effect was a reduction in pain and disability of 10.4 points.

Results Mean reductions in pain and functional impairment for both groups were similar. Compared with diagnostic arthroscopy, arthroscopic capsular shift increased pain and functional impairment by means of 5 points (95% CI −6 to 16 points) at 6 months, 1 point (95% CI −11 to 13 points) at 12 months and 2 points (95% CI −12 to 17 points) at 24 months.

Conclusions Compared with diagnostic arthroscopy alone, arthroscopic capsular shift confers, at best, only minimal clinically important benefit in the medium term.

Trial registration number NCT01751490.