Variability of sensory threshold determination in clinical use

The variability of perception threshold determination for vibration, tactile stimuli and thermal stimuli, with instruments intended for clinical use, was studied in 13 healthy subjects and 27 patients with chronic polyneuropathy. Normal thresholds for tactile and thermal stimuli were determined in 51 healthy subjects. Determinations were made for vibration on hand, lower leg and foot, for touch on pulp of forefinger and great toe and for temperature on hand and foot. Normal thresholds for both tactile and thermal stimuli were age-dependent. Short-term variation, with intervals of some minutes between determinations, remained within 8-18% change from first value. Long-term variation, with intervals of days to some weeks, was pronounced for all types of threshold, with extremes of -90% and +256% change from first determination in 3 or 4 subsequent determinations. Variation was most marked for tactile stimuli and smallest for vibration, but magnitude and pattern of variation was similar for all sensory modalities and for both patients and healthy subjects. Confidence intervals, derived from analysis of variance, showed than as an average a change of less than -60% or greater than +150% from the initial value was needed to ascertain with 95% probability that a subsequent value will reflect a true change of sensory threshold. Basing every threshold value on 2 or more measurements per occasion will reduce the confidence interval. The main cause of variability seems to be central processing mechanisms, i.e. the psychological variability. With proper attention to the variability, sensory threshold determinations should still be a valuable aid in clinical practice and clinical research.