We reviewed 101 patients with injuries of the terminal branches of the infraclavicular brachial plexus sustained between 1997 and 2009. Four patterns of injury were identified: 1) anterior glenohumeral dislocation (n = 55), in which the axillary and ulnar nerves were most commonly injured, but the axillary nerve was ruptured in only two patients (3.6%); 2) axillary nerve injury, with or without injury to other nerves, in the absence of dislocation of the shoulder (n = 20): these had a similar pattern of nerve involvement to those with a known dislocation, but the axillary nerve was ruptured in 14 patients (70%); 3) displaced proximal humeral fracture (n = 15), in which nerve injury resulted from medial displacement of the humeral shaft: the fracture was surgically reduced in 13 patients; and 4) hyperextension of the arm (n = 11): these were characterised by disruption of the musculocutaneous nerve. There was variable involvement of the median and radial nerves with the ulnar nerve being least affected.

Surgical intervention is not needed in most cases of infraclavicular injury associated with dislocation of the shoulder. Early exploration of the nerves should be considered in patients with an axillary nerve palsy without dislocation of the shoulder and for musculocutaneous nerve palsy with median and/or radial nerve palsy. Urgent operation is needed in cases of nerve injury resulting from fracture of the humeral neck to relieve pressure on nerves.

The Mangled Extremity Severity Score (MESS) may be used to decide whether to perform amputation in patients with injuries involving a limb. A score of 7 points or higher indicates the need for amputation. We have treated three patients with a MESS of 7 points or higher, in two of which the injured limb was salvaged. This scoring system was originally devised to assess injuries to the lower limb. However, a MESS of 7 points as a justification for amputation does not appear appropriate when assessing injuries to the major vessels in the upper limb.

We describe 20 patients, aged between 43 and 88 years, with delayed nerve palsy or deepening of an initial palsy caused by arterial injury from low-energy injuries to the shoulder. The onset of palsy ranged from immediately after the injury to four months later. There was progression in all the patients with an initial partial nerve palsy. Pain was severe in 18 patients, in 16 of whom it presented as neurostenalgia and in two as causalgia. Dislocation of the shoulder or fracture of the proximal humerus occurred in 16 patients. There was soft-tissue crushing in two and prolonged unconsciousness from alcoholic intoxication in another two. Decompression of the plexus and repair of the arterial injury brought swift relief from pain in all the patients. Nerve recovery was generally good, but less so in neglected cases. The interval from injury to the repair of the vessels ranged from immediately afterwards to 120 days. Delayed onset of nerve palsy or deepening of a nerve lesion is caused by bleeding and/or impending critical ischaemia and is an overwhelming indication for urgent surgery. There is almost always severe neuropathic pain

The outcome of 260 repairs of the radial and posterior interosseous nerves, graded by Seddon’s modification of the Medical Research Council Special Committee’s system, was analysed according to four patterns of injury; open ‘tidy’, open ‘untidy’, closed traction, and those associated with injury to the axillary or brachial artery. We studied the effect on the outcome of delay in effecting repair and of the length of the defect in the nerve trunk. Of the 242 repairs of the radial nerve we found that 30% had good results and 28% fair; 42% of the repairs had failed. The violence of injury was the most important factor in determining the outcome. Of the open ‘tidy’ repairs, 79% achieved a good or fair result, and 36% of cases with arterial injury also reached this level. Most repairs failed when the defect in the nerve trunk exceeded 10 cm. When the repairs were carried out within 14 days of injury, 49% achieved a good result; only 28% of later repairs did so. All repairs undertaken after 12 months failed. Of the 18 repairs of the posterior interosseous nerve, 16 achieved a good result

We have analysed the results of repair of traumatic lesions of the musculocutaneous nerve in 85 patients, which were graded by Seddon’s modification of the Medical Research Council system into three types of injury: open ‘tidy’, open ‘untidy’and closed ‘traction’. They were also correlated with associated arterial injury. There were 57 good, 17 fair and 11 poor results. The type of injury was the most important factor in determining the result; 12 of 13 open-tidy lesions gave good results compared with 30 of 48 closed-traction lesions. The results were better when the nerves were repaired within 14 days of injury and when grafts were less than 10 cm long. They were worse in the presence of associated arterial or bony injury