Variability of quantitative sensory testing: implications for clinical practice

The conventional estimation of motor, sensory, and mixed nerve conduction velocities reflects activity in the fastest conducting, heavily myelinated nerve fibers that are only a small proportion of the total. Unmyelinated and thinly myelinated fibers are not evaluated by this technique and numerically represent the largest group of fibers in human cutaneous nerves. The availability of new quantitative techniques to study this aspect of sensory function is an important addition to standard electrodiagnosis. Patient understanding and cooperation is essential because subjective responses are evaluated. We evaluated a reference range for 20 healthy subjects as well as variability on repeated testing. Vibration and thermal thresholds were measured bilaterally at several sites. Measurements were repeated at intervals ranging from two days to three months. There was no side difference but substantial site differences were noted for all measurements. Intraindividual variation was substantial but within the expected range for a psychophysiologic test. Close correlation was noted between various measurements at same and different sites, indicating a great degree of interindividual variation. The Marstock method is recommended for routine clinical use.