Sirs: The success of peripheral nerve repair after accidental lesions relies on the surgical techniques of anastomosis [1] or grafting [2], the regeneration potential of the severed nerve [3] and time elapsed from trauma. Wallerian degeneration or muscular dysfunction originating from disuse and/or release of neuromuscular trophic factors may compromise neuromuscular integrity [4]. Electrophysiological examination provides valuable insights into the localisation and pathophysiology of the injury, severity of the dysfunction and the progress of reinnervation [5], whereas 31p magnetic resonance spectroscopy (MRS) probes muscular energetics and function [6]. Alterations of resting metabolism after muscle denervation or chronic disuse have been documented using MRS in man [7] and animal models [8, 9]. We now describe an electromyographic and spectroscopic follow-up of nerve regeneration and muscle reinnervation in man after surgical anastomosis of sectioned median nerve. A 28-year-old healthy male presented with a deep wound at the left wrist (following accidental section with a sharp blade) and total anaesthesia of the hand in the median nerve territory. Inspection revealed a complete section of the median nerve but no damage of tendons or vessels. Direct epineural nerve repair was carried out with a nylon 9-0 suture within 3 h of the accident. Denervation of the opponens pollicis muscle was assessed by needle examination. Orthodromic motor conduction and action potential amplitude of the median nerve were recorded. Antidromic sensory action potential amplitudes were measured for the first four digits with ring electrodes while stimulating at the wrist level. Thenar eminence muscle metabolism was followed by 31p MRS with a SIS 200/330 spectrometer equipped with a 4.7 T horizontal magnet. Spectra were acquired with a 2.5-cm-diameter surface coil (1024 transients, repetition time 0.6 s), yielding a minimum signal to noise ratio of 44 for creatine phosphate (PCr) in the operated hand at day 12 after the accident. Intramuscular pH was determined from the chemical shift difference between inorganic phosphate (Pi) and PCr resonances [6]. Spontaneous activity was abundant and voluntary recruitment limited to single units, indicating severe muscle denervation at day 17 postinjury (Table 1). The follow-up revealed signs of collateral reinnervation, improving over time. Motor and sensory action potential areas and peak-to-peak amplitudes also recovered. Fairly complete restoration of the amplitude of the evoked motor potential developed after 15 months (Table 1), as compared with a period of 2 months required after nerve crushing in a rat model [8]. Two weeks after injury, the patient presented with anaesthesia for all sensory modalities in a territory encompassing the entire cutaneous distribution of the left median nerve. A progressive reduction in the area of anaesthesia was found 2 months after injury, with shrinkage along the borders of the ulnar and radial nerve territories. The area of anaesthesia to warmth remained larger than that of pinprick. At 15 months post-injury the patient reported a satisfactory recovery of all sensory modalities in the entire median nerve territory. He estimated his sensory function to be