Sensitivity and specificity of different conduction block criteria

Objectives: To resolve the discrepancy between conduction block criteria derived from healthy controls and stricter criteria suggested by computer simulation of interphase cancellation through altered motor units. Methods: An EMG database provided control nerves from patients with amyotrophic lateral sclerosis (ALS) or neural muscular atrophy (CMT1) (disease controls) and from subjects without neuromuscular diseases (healthy controls). We estimated normal limits from the healthy controls (criterion A) and from the pooled sample of healthy and disease controls (criterion B). We compared their sensitivity with that of an arbitrary limit of 0.5 (criterion C) in acute (AIDP) and chronic inflammatory demyelinating neuropathy (CIDP) and in multifocal motor neuropathy (MMNP). Specificity was assessed in ALS and CMT1. Results: Limits estimated from healthy controls (criterion A: amplitude ratio of <0.7 in median and peroneal nerves and <0.8 in the ulnar nerve) gave false positive results in 17.3% of the ALS nerves. High scatter of the amplitude ratio of the nerves with distal response amplitudes below 1 mV required amplitude-dependent limits (0.36 for distal responses below 1 mV, 0.56 between 1 and 2 mV, and between 0.67 and 0.73 for higher response amplitudes) for criterion B. It was false positive in 4.3% of the ALS nerves and in 28.3% of the CMT1 nerves. A limit of 0.5 for nerves with distal responses above 1 mV and a limit of 0.36 for smaller responses (criterion D) avoided false positive results in ALS without further impairing sensitivity per patient in MMNP. Sensitivity in AIDP was 34.9% for criterion A, 19.5% for criterion B, and 10.2% for criterion D. Amplitude ratios were more sensitive than area ratios in CIDP and MMNP, but less specific in CMT1. Conclusions: Limits derived from healthy controls are unspecific in chronic neuromuscular diseases and in nerves with low response amplitudes. Criterion D should be used if motor unit restructuring or conduction delays have to be taken into account. Criterion A may be applicable in early AIDP if the distal response amplitude is above 1 mV. (C) 2000 Elsevier Science Ireland Ltd. All rights reserved.