Abstract
In order to investigate the mechanism of nerve irritation in thoracic outlet syndrome (TOS), we studied 150 patients who presented with symptoms of neurologic TOS between 1985 and 1999. They first performed various provocative physical manoeuvres and then underwent injection of contrast medium into the supraclavicular part of the brachial plexus. Several of the provocative manoeuvres were then repeated and radiographs were again obtained. Based on the neuroradiographs, we identified three subsets of patients; those with only compression (type 1 TOS, n=27, 18%), those with combined compression and stretching (type 2 TOS, n=111, 74%), and those with only stretching (type 3 TOS, n=12, 8%). We were able to correlate the neuroradiological subsets with symptoms elicited by pre-radiographic provocative manoeuvres; in 92 patients (61%) these were elicited by traction manoeuvres. We conclude that stretching is an important factor of nerve irritation in TOS.
INTRODUCTION
Thoracic outlet syndrome (TOS) was first described by Peet et al. (1956) as a syndrome caused by compression or stretching of the brachial plexus and/or the subclavian artery and vein as they transverse the thoracic outlet. No specific diagnostic tests for TOS currently exist. TOS is primarily a clinical diagnosis based on the presence of a combination of arm pain, color change of the hand and a radicular pattern of sensory loss and weakness. As tests evaluating positional changes on the radial pulse have proved unreliable, examination of neurologic symptoms elicited by provocative manoeuvres is important for establishing the diagnosis of neurologic TOS (Roos, 1979). More than 95% of TOS is attributed to entrapment of the brachial plexus at the thoracic outlet and vascular TOS is thought to be present in only about 5% of patients (Schwartzman, 1991). Wilbourn (1990) divided TOS into four distinct subgroups; arterial vascular, venous vascular, true neurologic, and disputed neurologic TOS. While the first three are uncontroversial and their incidence is low, disputed neurologic TOS exhibits unique features that are unparalleled in the field of peripheral neurology and its etiology is unclear.
Swift and Nichols (1984) reported that some patients with TOS had droopy shoulder syndrome and suggested that their symptoms resulted from stretching of the brachial plexus. Nakatsuchi et al. (1995) proposed that the symptoms of TOS might be related to increased tension of the brachial plexus and surrounding vasculature due to muscular imbalance and the resultant downward traction.
Since 1985, we have studied the pathology of TOS using neuroradiography of the brachial plexus (Ide et al., 1994). One of the provocative tests for this condition is performed by pulling the arms down and back, and should produce pinching of the brachial plexus between the first rib and the clavicle. However, on the neuroradiographs we observed that, in some TOS patients, this manoeuvre resulted in stretching between the proximal and distal fixed points of the brachial plexus rather than brachial plexus compression. Their symptoms were aggravated by downward traction of the arm (traction on the brachial plexus) and relieved by pulling the arms upward (relaxing the brachial plexus) (Ide et al., 2001). This observation led us to renewed attempts to understand the mechanism of nerve irritation in TOS. This study investigates the correlation between the neuroradiographic findings in patients with a diagnosis of neurologic TOS and the symptoms and signs produced by provocative manoeuvres.
PATIENTS AND METHODS
We studied the brachial plexus neuroradiographs of 150 patients who presented with symptoms of neurologic TOS between 1985 and 1999. There were 93 women and 57 men with mean ages of 30 (range, 13–56) and 35 (range, 16–60) years, respectively. As shown in Table 1, the patients presented with combinations of symptoms and signs; none had gangrene.
The patients had the following criteria: (1) persistent diffuse pain and/or paresthesia in the upper limb aggravated by specific arm movements and positions, (2) a positive Tinel’s sign over the brachial plexus, (3) elicitation of reproducible pain and/or paresthesia by at least one provocative manoeuvre (Adson, 1947; Eden, 1939; Roos, 1979; Wright, 1945) or induction and/or aggravation of symptoms upon pulling downward on the arm and their improvement or elimination upon pulling the arms upward, (4) exclusion of diseases of the cervical spine and a peripheral neuropathy.
After the performance of provocative physical manoeuvres (Tinel’s sign over the brachial plexus at the level of the scalene muscles or at the supraclavicular fossa; aggravation of pain or paresthesia with the arm in 90° abduction and external rotation or with arm traction; relief of pain or paresthesia by pulling the arms upward), neuroradiographs were taken. Several of the manoeuvres were then repeated and standardized radiographs were obtained. Based on the neuroradiographic findings, we defined three subsets of patients and compared the physical signs and characteristics of each of the three types of TOS. For evaluation, we used data obtained on the worse side.
Brachial plexus neuroradiography technique
We used the percutaneous supraclavicular or inter-scalene muscle approach under fluorographic guidance (Takeshita et al., 1991). Contrast medium (10 ml iopamidol, 15 ml saline, 5 ml of 1% lidocaine) was injected slowly over a 60-s period to fill the plexus from the scalenus anterior proximally to the subcoracoid region distally.
Classification of TOS based on brachial plexus neuroradiographs
Radiographs were taken with the patient in standing position. To evaluate pinching of the brachial plexus between the first rib and the clavicle, anteroposterior and oblique lateral radiographs were obtained with the arm at the side and elevated. For study of elongation of the brachial plexus resulting in stretching between its proximal and distal fixed points downward traction (3 kg load stress) was applied to the arm and anteroposterior radiographs were obtained. The radiographers were blind with respect to the patients’ symptomatology and the results of the pre-injection provocation manoeuvres.
Brachial plexus compression was recorded when the contrast medium around the brachial plexus in the costoclavicular space between clavicle and first rib was narrowed by 50% by elevation of the arm (compared with when the arm was at the side) (Fig 1). This was assessed by measuring the distance between the inferior surface of the clavicle and the superior surface of the first rib on oblique lateral radiographs obtained. With the arm elevated and hanging by the side: the compression ratio of this distance was calculated. Brachial plexus stretching was recorded when the contrast medium of the brachial plexus was more elongated and taut with the arm in downward traction than with the arm hanging by the side (Fig 2). The distance between first thoracic spinal process and coracoid process on anteroposterior radiographs obtained with the arm hanging by the side was compared with that when downward traction was applied and the elongation ratio was calculated. Radiographic findings led to the identification of three patients subsets; those who had evidence of only compression, those with only elongation or stretching, and those with both compression and stretching. Based on these findings, we established three types of TOS: compression- (type 1), combined- (type 2), and stretching type TOS (type 3).
Statistical analysis
For analysis of differences between groups, we used the unpaired t-test, for multivariate analysis, the chi-squared test.
RESULTS
Of the 150 patients, 138 (92%) had evidence of brachial plexus compression and 123 (82%) of brachial plexus stretching. Twenty-seven (18%) had type 1, 111 (74%) had type 2, and 12 (8%) had type 3 TOS. Table 2 shows the gender and age of patients in each group. The ratio of women to men was significantly higher in type 2 and type 3 TOS than in type 1 (P<0.001) and the mean age of type 1 patients was significantly higher than that of type 3 patients (P<0.05). The mean compression ratio between clavicle and first rib and elongation ratio of the distance between first thoracic spinal process and coracoid process for each group are shown in Table 3. The mean compression rate in type 1 or 2 TOS was significantly higher than that in type 3 TOS (P<0.0001). The mean elongation rate in type 3 TOS was significantly higher than that in type 1 TOS (P=0.01).
Table 4 indicates the correlation between the signs and symptoms elicited by the pre-radiographic provocation manoeuvres and TOS type based on brachial plexus neuroradiographs. Tinel’s sign in the supraclavicular fossa was common to all TOS types. Significantly, more types 1 and 2 than type 3 TOS patients manifested aggravation of pain with the arm in 90° abduction and external rotation (P<0.0001). However, Tinel’s sign on the scalene muscles, aggravation of discomfort by the traction manoeuvre and symptom relief upon pulling the arms upward were elicited in significantly more patients with type 2 or type 3 TOS than in those with type 1 (P<0.01).
Table 5 shows the diagnostic value of each of the five provocative tests between type 1 and type 3 TOS. For type 3 TOS, the accuracy of Tinel’s sign on the scalene muscles, the arm traction test and arms upward test was high, while that of Tinel’s sign in supraclavicular fossa and the arm abduction in external rotation test was low.
Patients with type 1 TOS manifested several characteristics. Of the 27 type 1 patients, 18 were muscular, square-shouldered men who worked or played sports with the arm in an abducted and externally rotated position. Eleven of the 12 type 3 TOS patients were young, slender women with drooping shoulders and poor posture.
DISCUSSION
We were able to identify three types of TOS with characteristic responses to provocative manoeuvres and establish that the symptoms and signs attributable to nerve irritation by traction occur in a significant proportion of patients with neurologic TOS: 92 (61%) of our 150 patients, traction manoeuvres elicited symptoms. We suggest that traction rather than compression of the brachial plexus is the major mechanism of nerve irritation in type 3 and some type 2 patients with TOS.
Characteristically, patients with type 3 TOS were young, slender women with drooping shoulders and poor posture. The high prevalence of women among our patients with type 2 and type 3 TOS may be due to a combination of factors, including differences in peri-scapular muscle strength and development, and postural variations. The results for the elongation ratio, the distance between first thoracic spinal process and coracoid process suggest that the strength of scapular retractors (trapezius, levator scapulae, the rhomboids, serratus anterior) may be weak in the stretching types of TOS. Tinel’s sign on scalene muscles and the arm traction and arms upward test have a high diagnostic value in these cases. Tinel’s sign over the brachial plexus may reflect the advance of a wave of regeneration or diminution of the threshold of nerve fibres to mechanical stimulation. While patients should undergo thorough neurophysiological investigation, the findings will usually be normal, indicating that nerve conduction studies reflect events in the largest myelinated fibres rather than changes in the behaviour of non-myelinated fibres.
Thoracic outlet decompression surgery may not be effective in patients whose major symptoms are due to stretching of the brachial plexus (Ide et al., 1994; Takagi et al., 1987). Conservative treatment with physical therapy is often successful in patients whose symptoms were aggravated by downward traction and relieved by pulling the arms upward (Kataoka, 1994). Our patients performed shoulder strengthening exercises with an orthosis stabilizing the scapula, which provides shoulder girdle support and corrects poor posture to relax the brachial plexus (Ide et al., 1994; Fig 3). Nakatsuchi et al. (1995) reported excellent or good results in 66 of 86 (77%) conservatively treated TOS patients who used an orthosis to counterbalance the downward traction on the brachial plexus.
We found that 138 (92%) of our 150 patients had neuroradiographic evidence of compression in the costoclavicular space and 123 (82%) had neuroradiographic evidence of stretching of the brachial plexus. Takeshita et al. (1991) noted abnormalities in 153 of 180 (85%) neuroradiographs of TOS patients; narrowing in the scalenus triangle, the costoclavicular space, and at the subcoracoid region was noted in 30%, 75%, and 6%, respectively. They found no suggestion of compression in any arm position in 30 control individuals, but some compression of the brachial plexus may occur in most of the patients with abduction of the shoulder (Wright, 1945). In our study, brachial plexus compression was recorded when the contrast medium around the brachial plexus in the costoclavicular space between clavicle and first rib was more than 50% narrowed with the arm elevated. Takeshita et al. (1991) did not investigate stretching of the brachial plexus. In our study type 1 TOS patients, who had compression but no stretching of the brachial plexus, the mean elongation rate was 5.5%, suggesting that this may be the normal level of the elongation. A comparative study of healthy adults and patients with TOS is needed to clarify the normal range of the neurography.
Our study documents the existence of different mechanism of nerve irritation in TOS: compression and stretching of the brachial plexus. This differentiation is useful as it determines the treatment of TOS.