Variability of three standard neurophysiological techniques in established symptomatic diabetic polyneuropathy

The reproducibility of three standard neurophysiological techniques (motor and sensory nerve conduction velocities, vibration perception threshold and mean expiratory: inspiratory electrocardiographic R-R ratio) was determined from duplicate measurements approximately four weeks apart in 50 patients with chronic symptomatic diabetic polyneuropathy. There was no change in average glycaemic control (mean +/- SD) between the measurements (HbA1 10.7 +/- 2.0 vs. 10.4 +/- 2.1% at zero and four weeks respectively). While there were no significant differences in the mean values at zero and four weeks for any technique considerable intra-individual variability was observed in several measurements. Mean coefficients of variation (CV) for motor nerve conduction velocity ranged from 9.8% (ulnar nerve) to 10.7% (median) to 12.2% (peroneal nerve). Variability in sensory nerve conduction velocities (CVs 8.6, 8.7 and 14.7% for radial, ulnar and sural nerves respectively) was complicated by a high proportion of unrecordable action potentials. Duplicate action potentials were recordable from the sural nerve in only 15 (30%) patients. The highest intra-individual variability was observed in measurements of vibration perception threshold with CVs of 21.0% (thumb) and 18.0% (hallux). The apparently satisfactory CV of 3.6% for the cardiac E:I ratio test may represent an artefactually low degree of variability. Neurophysiological techniques in current use are characterized by technical limitations and intra-individual variability in patients with established diabetic neuropathy.