A new principle for assessing vibrotactile sense in vibration-induced neuropathy.

The vibrotactile perception threshold has been evaluated in 20 manual workers not exposed to occupational vibration and in 27 workers using hand-held vibrating tools. Threshold values were evaluated with a Békesy type of vibrometer at frequencies ranging from 8 to 500 Hz. The earliest sign of vibration lesion was a reduction of sensation at frequencies of 125 to 250 Hz, indicating dysfunction in receptors of the fast adapting type II (stage 1 curve), followed in more severe cases by a prominent loss of sensation at all frequencies higher than 65 Hz (stage 2 curve). The stage 3 curve indicated the most severe loss of sensation in which the function of slowly adapting type I receptors, as well as fast adapting type I receptors, was deteriorated; consequently the vibrotactile thresholds at low-, median-, and high-frequency ranges were impaired. The vibrotactile changes corresponded well to numbness and the presence of vibration-induced white finger.

increasing interest in exploring neurological di sturbances oc curring in the upper extremities a ft er longterm use of hand-held vibra tin g tools (2 , 5, 7). The symptoms of peripheral nerve involvement may vary from int ermittent tingling to permanent numb ness and, in extreme cas es, pain and mu scle was ting (19). Us ually either the median or the ulnar nerve is involved (12), but so me authors report sim ulta neo us involveme nt o f both nerv es (17).
The early diagnosis of vibra tion-induced neuropathy is associated with se veral difficulties. Although nerve conduction velocity may be impaired in advanced stag es (3,5), neurophysiological test s a nd two-point di scrimination tests may be perfectly normal (15). It ha s , ho wever, been reported tha t the impa irm ent o f vib ro tac tile sense m ay be one of the firs t ch anges ob ser ved after occupational exposure to vibration (4,II ). Thus cha nge s in vibro tactile threshold may predict the o nse t of vibratio n injury. We present a method for determining th e vibrotactile sense a t several frequencies. The purpose was to develop a diagnostic method for th e early detection of vibra tio n-ind uc ed neuropa th y.

Materials and methods
The vibrometer An audiometer of the Bekesy type was modified in accordance with a previous description (14). To evoke stable vibration without compression circuitry, a powerful vibrator (Bruel & Kjaer, type 4809)was used (figure 1). The area of the vibrating head was 5 mm-, Th e vibration exciter consisted of an audiometer (Briiel & Kjaer , type 1800) modified to operate at low frequencies. Th e signal was amplified with a power amplifier (Bruel & Kjaer, type 2706). The vibrating head of the exciter was placed into the center of a small opening in a testing table. The system provided a sinusoidal signal at controlled frequencies and ampli tudes, and it could be applied to a finger pulp placed on top of the table (figure 2). Vibration at seven fixed frequencies ranging from 8 to 500 Hz were automatically delivered to the probe on the top of the exciter.
The signal was recorded on an X-V record er with a builtin test signal generator. The test frequency was presented on the X axis, the frequencies being automatically changed by the instrument itself. Th e Y axis deflection represents the perception threshold for vibration. The level of vibratio n is controlled by an automatic attenuator, operated by the subject by means of a hand-h eld switch (figures 3 and 4). Instructions were given to press the switch when the vibration of the probe could be felt, thus causing the automat ic attenu ator to decrease the intensity (amplitude) . When vibration could no longer be felt, the switch was released. In this way the subjects could regulate the intensity of the vibration and thereby track their threshold levels. The perception threshold was recorded on a preprinted chart at frequen cies of 8, 16, 33, 65, 125, 250, and 500 Hz, the frequencies being aut omatically changed through this frequency range. The level was recorded in decibels as acceleration re 10-6 m/ 2 • In all the subjects the third and fifth fingers of both hands, innervated by the median and ulnar nerves, respectively, were tested . The testing time for one finger was 5-6 min. Since the test is psychophysical and requires cooperation from the patient, a preliminary test was done for trainin g purpo ses, before the actual testing procedure was carried out. The testing was usually performed by an occupational therapi st.
The equipment was mobile and could be brough t to any desired place. The testings were perfo rmed in various local health care centers associated with the workplaces.

Workers exposed to occupational vibration
Twe nty-seven workers who had been using hand -held vibrating tools for various period s of time forme d the exposed grou p. The mean age was 35 (ra nge 18-60) years, the exposure time averaging 9 years (range 9 mon ths -27 years). The subjects tested included car mechani cs and factory workers exposed to vibration from hand-held vibrat ing tools. In a qu estion nai re each patie nt indicated the occurrence of white finger s a nd/or n umbness of the hand. Th e investigation did not include objecti ve assessments of vasospastic disease. Vibrograms were obta ined from the third an d fift h fingers of both hands no less than 2 h a fter the cessati on of exposure to vibra tion .

Ref erence subjects
Vibrograms obtained from the reference group showed a typical shape with a mainly ho rizon tal co urse in the lower frequencies, followed by a peak within the range of 125-250 Hz, ind icating a bett er percept ion threshold fo r fast adapt ing type II (FA II) recepto rs (pacinian corpuscles) than for fast adapt ing type I (FA 1) or slowly ada pting type I (SA I) or 11 (SA 11) recept or s (figure 5). At 500 Hz ther e was an increase in th e perception thre shold s, indicated by a slope of the cur ve.
The level and shape of the curve was identical to curves obtained from 55 referents from the hospital staff, who were not involved in heavy manual work (14).

Workers exposed to vibration
There was a striking correlation between the subjective symptoms presented by the patients and the changes in the curve of the vibrograrn. With increasing symptoms, changes in the curve corresponding to the results obtained from material of nerve entrapment patients were observed (14) so that "stage 1" was indicated by a flattening of the peak within the FA II receptor range of 125-250 Hz, " stage 2" was indicated by, in addition, a steeper slope of the curve starting at a frequency of 65 Hz , and "stage 3" was indicated by the dis-appearance of the vibration sense at the highest frequencies and deterioration of sensation also in the very low frequency range corre sponding to the vibrotactile perception o f the SA I recepto r pop ulation . Typical curves are shown in figures 6-8, representing workers having used hand-held vibrating too ls for 6 months, 18 months, and 27 years, respectively. The changes in the vibrogram corresponded remarkably well to the subjective experience of numbness in the hand and to the occurrence of white fingers. Of six workers suffering from white fingers, the vibrogram was abnormal in five cases. Of five worker s suffering from numbness of the hand, the vibrogram was abnormal in four cases. For seven workers with symptoms of both disturbed peripheral circulatio n and neuropathy the vibrogram was abnormal in six cases. Out  of ten workers with no symptoms, three showed abnormal vibrograms (stage I).

Discussion
Vibration-induced neuropathy of the hand represents a puzzling condition, often imitating carpal tunnel syndrome. The nerve conduction velocity is often normal across the wrist level, and these cases are not always cured by decompression of the median nerve . These facts strongly indicate that the nerve lesion may be located more distally than in the carpal canal, and hypotheses have been proposed that intraneural edema in distal median nerve branches may constitute an important etiologic factor in early vibration-induced neuropathy (13).
In compression neuropathies changes in vibrotactile sense represent early signs, always appearing before changes in two-point discrimination and often appearing before impairment in nerve conduction velocity (6). Changes in the vibrotactile sense of the skin have been reported to represent an early objective sign also in vibration-induced neuropathy (4). The physiology of the vibrotactile sensation of the skin is complex in that vibration is perceived by several receptor systems of the skin. Two receptor populations have been defined on the basis of their properties for adaptation to constant pressure, ie, the SA and FA types. Depending on their response to sharp edges, each of these receptor types can be further divided into receptors responding to sharp contours (type I) and into diffuse contours (type II) (9). SA I receptors are anatomically linked to neurite complexes of Merkel's cells and SA II receptors to Ruffini's corpuscles, both of which respond to lower frequencies of vibration (0.1-60 Hz) at psychophysical threshold values. FA I receptors have been anatomically linked to Meissner's corpuscles, and FA II receptors to pacinian corpuscles, the latter perceiving higher frequencies of vibration at psychophysical thresholds (10).
The psychophysical threshold (vibrogram) curve is linked to the activation of SA I receptors at low frequencies (below 16 Hz), to FA I receptors at middle frequencies (32 to 65 Hz), and to FA II receptors at high frequencies (above 65 Hz). The most sensitive frequency range in the evaluation of psychophysical vibrotactile perception is 250-350 Hz, where the FA II receptors in normal subjects can detect displacement of the skin at an amplitude of 0.1 urn (16). In our study this sensitive threshold of FA II receptors has been expressed in the peak of the vibrogram curve.
The earliest sign of decreased vibrotactile sensation was a flattening of this peak (stage I). Abnormalities in the curve within the lower frequencies occurred only at an advanced stage, which -paralleled by an increase in 2 PD -reflects a lesion in the SA I receptor population.

378
It is interesting to note that abnormalities of the curve correlated not only to numbness of the hand, but also to the occurrence of white fingers. This finding emphasizes the difficulties in separating neurological and vascular problems in the vibration syndrome and suggests a common etiology. Intraneural nerve fibers in vascular walls may playa role in the initiation of vasospasm. On the other hand, all peripheral nerve trunks are well vascularized , and disturbances in intraneural vascular function will rapidly interfere with nerve function. The fact that the intraneural microvessels are sympathetically innervated makes the picture even more complicated. A disorder of sympathetic nerve fibers may have consequences for the blood flow in the vesselsof the hand and for the vessels in the peripheral nerves and hereby contribute to an even further impairment of peripheral nerve function.
The vibrotactile sense is dependent on local temperature, as well as on time and space variables such as frequency, temporal and intensity relationships between stimuli, area of stimulation, and surface gradients on the skin (20,21). The significance of the transient threshold shifts for vibration perception is well known (II). These factors have to be considered when vibrotactile sense is used as an indicator of damage to components of peripheral nerves. All these factors can be completely controlled only under welldefined laboratory conditions. Our method, however, was developed with the purpose of using mobile equipment for screening tests in the field. By necessity, all factors considered in the preceding discussion could not be completely controlled. This deficiency has to be judged against other advantages of the method. However, as long as the testing procedure is standardized maximally in each individual case, we feel that the main source s of error have been excluded .
The results of this pilot study, with a striking correlation between subjective symptoms of neurological and circulatory disturbances and pathological vibrograms, as well as between exposure time and pathological vibrograms, indicate that our method is very useful for screening test purposes. Further investigations are needed to prove the statistical significance of the method. In an ongoing prospective field study including 350 manual workers doing heavy work, the method is now being used as a screening test to study the course of the neuropathy as related to factors such as continuous exposure to vibration and modifications of the work situation.