The outcomes of 261 nerve injuries in 100 patients were graded good in 173 cases (66%), fair in 70 (26.8%) and poor in 18 (6.9%) at the final review (median 28.4 months (1.3 to 64.2)). The initial grades for the 42 sutures and graft were 11 good, 14 fair and 17 poor. After subsequent revision repairs in seven, neurolyses in 11 and free vascularised fasciocutaneous flaps in 11, the final grades were 15 good, 18 fair and nine poor. Pain was relieved in 30 of 36 patients by nerve repair, revision of repair or neurolysis, and flaps when indicated. The difference in outcome between penetrating missile wounds and those caused by explosions was not statistically significant; in the latter group the onset of recovery from focal conduction block was delayed (mean 4.7 months (2.5 to 10.2) vs 3.8 months (0.6 to 6); p = 0.0001). A total of 42 patients (47 lower limbs) presented with an insensate foot. By final review (mean 27.4 months (20 to 36)) plantar sensation was good in 26 limbs (55%), fair in 16 (34%) and poor in five (11%). Nine patients returned to full military duties, 18 to restricted duties, 30 to sedentary work, and 43 were discharged from military service. Effective rehabilitation must be early, integrated and vigorous. The responsible surgeons must be firmly embedded in the process, at times exerting leadership.

We describe 261 peripheral nerve injuries sustained in war by 100 consecutive service men and women injured in Iraq and Afghanistan. Their mean age was 26.5 years (18.1 to 42.6), the median interval between injury and first review was 4.2 months (mean 8.4 months (0.36 to 48.49)) and median follow-up was 28.4 months (mean 20.5 months (1.3 to 64.2)). The nerve lesions were predominantly focal prolonged conduction block/neurapraxia in 116 (45%), axonotmesis in 92 (35%) and neurotmesis in 53 (20%) and were evenly distributed between the upper and the lower limbs. Explosions accounted for 164 (63%): 213 (82%) nerve injuries were associated with open wounds. Two or more main nerves were injured in 70 patients. The ulnar, common peroneal and tibial nerves were most commonly injured. In 69 patients there was a vascular injury, fracture, or both at the level of the nerve lesion. Major tissue loss was present in 50 patients: amputation of at least one limb was needed in 18. A total of 36 patients continued in severe neuropathic pain.

This paper outlines the methods used in the assessment of these injuries and provides information about the depth and distribution of the nerve lesions, their associated injuries and neuropathic pain syndromes.

A 60-year-old man developed severe neuropathic pain and foot-drop in his left leg following resurfacing arthroplasty of the left hip. The pain was refractory to all analgesics for 16 months. At exploration, a PDS suture was found passing through the sciatic nerve at several points over 6 cm and terminating in a large knot. After release of the suture and neurolysis there was dramatic and rapid improvement of the neuropathic pain and of motor function.

This case represents the human equivalent of previously described nerve ligation in an animal model of neuropathic pain. It emphasises that when neuropathic pain is present after an operation, the nerve related to the symptoms must be inspected, and that removal of a suture or irritant may lead to relief of pain, even after many months.

We studied 27 patients with low back pain and unilateral L5 or S1 spinal nerve root pain. Significant radiological changes were restricted to the symptomatic root level, when compared with controls. Low back and leg pain were graded on a visual analogue scale. Dermatomal quantitative sensory tests revealed significant elevations of warm, cool and touch perception thresholds in the affected dermatome, compared with controls. These elevations correlated with root pain (warm v L5 root pain; r = 0.88, p < 0.0001), but not with back pain. Low back pain correlated with restriction of anteroposterior spinal flexion (p = 0.02), but not with leg pain.

A subset of 16 patients underwent decompressive surgery with improvement of pain scores, sensory thresholds and spinal mobility. A further 14 patients with back pain, multilevel nerve root symptoms and radiological changes were also studied. The only correlation found was of low back pain with spinal movement (p < 0.002). We conclude that, in patients with single level disease, dermatomal sensory threshold elevation and restriction of spinal movement are independent correlates of sciatica and low back pain.

We performed intercostal nerve transfer in 19 patients to relieve pain from preganglionic injury to the brachial plexus. The procedure was successful in 16 patients at a mean of 28.6 months (12 to 68) after the injury.