Shrug radiographs for the diagnosis of long thoracic nerve palsy in traumatic brachial plexus injury

Abstract [Background] Preoperative diagnosis of long thoracic nerve (LTN) palsy is important for shoulder reconstruction after a traumatic brachial plexus injury (BPI). In the present study, we developed an objective diagnostic method for LTN palsy for patients with traumatic BPI. [Methods] This is a retrospective review of 56 patients with traumatic BPI who had been receiving treatment at a single institution for over eight years. The patients were divided into two groups: an LTN palsy group (n=30) and a no palsy control group (n=30). The LTN palsy group had 21 different palsy types with four and five C5-7 and C5-8, whereas the no palsy group had 18 different palsy types with five and three C5-6 and C5-8, respectively. Preoperative plain anteroposterior radiographs were taken in shoulder adduction and shrug positions. Scapulothoracic (ST) upward rotation and clavicle lateral (CL) rotation angles were measured on X-rays. The differences between the adduction and shrug positions for the respective angles were calculated and defined as ΦST and ΦCL, respectively. The differences in the ΦST and ΦCL values due to the presence or absence of LTN palsy were examined, the cut-off values of ΦST and ΦCL for the diagnosis of LTN palsy were determined, and further sensitivity and specificity were calculated. [Results] Both ΦST and ΦCL were significantly decreased in the LTN palsy group compared with the no palsy control group. The sensitivity and specificity for LTN palsy were 0.833 and 1.000 for ΦST and 0.833 and 0.840 for ΦCL, respectively, when the cut-off value was set as ΦST = 15° and ΦCL ≤ 24°. [Conclusion] Dynamic shrug radiographs provide a useful objective diagnosis of LTN palsy after traumatic BPI. Level of evidence Level IV; Case Control Design; Diagnostic Study