Do Cerebral Potentials to Magnetic Stimulation of Paraspinal Muscles Reﬂect Changes in Palpable Muscle Spasm, Low Back Pain, and Activity Scores?

Objective: Previous studies have shown that cortical-evoked potentials on magnetic stimulation of muscles are influenced by muscle con- traction, vibration, and muscle spasm. This study was carried out to determine whether these potentials correlate with palpatory muscle spasm, patient symptoms, and disability in patients with low back pain. Methods: A prospective observational study was performed on 13 subjects with a history of low back pain visiting an orthopedic hospital-based clinic. Patients were screened for serious pathologic conditions by an orthopedic surgeon. The patients were then evaluated for the pres- ence of muscle spasm by one of the investigators who was blinded to the results of the evoked potential studies. Patients were asked to complete a low back pain visual analogue scale (VAS) and a Roland-Morris Activity Scale (RMAS). Cortical-evoked potentials were recorded with a magnetic stimulator placed over the lumbar paraspinal muscles with the patient in the prone position. The palpatory examination, VAS, RMAS, and the cortical potentials were repeated after 2 weeks of therapy commonly used to reduce muscle spasm. Results: The patients demonstrated a signiﬁcant decrease in low back pain VAS and RMAS scores after treatment compared with before treatment. There was a reduction in the amount of palpatory muscle spasm in 11 of 13 cases. The cortical potentials before treatment were attenuated compared with previously reported controls and showed a signiﬁcant increase before and after treatment in the amplitude of these potentials with multivariate analysis of variance. There was signiﬁcant correlation between the changes in cortical poten- tials after treatment and the changes noted in paraspinal muscle spasm and VAS and RMAS scores. Conclusions: This study confirms the previous report that the amplitude of cerebral-evoked potentials on magnetic stimulation of paraspinal muscles is depressed in the presence of palpa- ble muscle spasm. The close correlation among these potentials, paraspinal muscle spasm, and clinical symptoms suggests that the measurement of muscle activity may be more important in the assessment of low back pain than is commonly accepted. (J Manipulative Physiol Ther 2000;23:458-64)


IINTRODUCTION
The role of muscle in low back pain and even the existence of muscle spasm has been the subject of controversy, [1][2][3] primarily because of the inadequate measurement techniques that correlate to patient symptoms and other clinical findings. Despite the skepticism, physicians commonly attribute spinal pain to muscle spasm, sprains, strains, or soft-tissue trauma. 4,5 Patients often point to muscles as the source of their pain, and many physicians describe muscle firmness or spasm as a palpable finding in the clinical examination. 6 The American Medical Association Guidelines for the Evaluation of Permanent Impairment considers vertebral muscle guarding or spasm an objective finding for differentiating categories of impairment. 7 Muscle firmness, spasm, or so-called trigger points are perceived by many physicians as palpable and by many patients as specifically tender. 8,9 Needle electromyography of these tender muscles, however, has demonstrated that they are electrically silent. 2 As yet, there are no reliable electrodiagnostic methods to confirm the presence of clinically palpable muscle firmness and subjectively tender muscles. The ability to record cerebral potentials after magnetic stimulation of muscle has provided a new and noninvasive method for examining muscle afferent activity in human beings. 10 Cortical potentials can be evoked on magnetic stimulation of muscles in both the upper and lower extremities and the paraspinal muscles. These magnetically evoked potentials can be attenuated by vibration and voluntary muscle contraction, which suggests that they are the result of 1aafferent fiber stimulation. The observation that these potentials are also attenuated by involuntary paraspinal muscle spasm raises the possibility that they may be of value in monitoring muscle spasm in patients with low back pain. 11 We are interested in defining the relation between cerebral potentials evoked by magnetic stimulation of paraspinal muscles and clinical findings such as palpable muscle spasm, pain, and activity scores and determining whether evoked potentials change in concert with patient symptoms.

METHODS
This paper was designed as a prospective observational study to determine the correlation between magnetically induced, cortical-evoked responses from paraspinal muscles, clinically determined muscle spasm, Visual Analogue Scales (VAS) for pain, and Roland-Morris Activity Scales (RMAS).
Thirteen patients presenting to a hospital-based back pain clinic were screened by an orthopedic surgeon not affiliated with the project to exclude patients with serious pathologic conditions. Patients were then evaluated for inclusion criteria and clinical assessment by one of the investigators (CYH). Inclusion criteria were low back pain with or without leg pain, the presence of lumbar muscle spasm on palpation (the working definition for palpable spasm in this study was the presence of palpable hardness, bands, and localized tenderness of paraspinal muscles that could be differentiated from adjacent muscles), no prior history of lumbar surgery, and no contraindications for lumbar spinal manipulation. Table 1 presents the demographic data and working diag-noses of these patients. The patients signed a consent form to participate in the project after an explanation about the benefits and risks of the research and treatment procedures.
The patients were given a questionnaire to complete, including an RMAS and low back pain VAS. A crude assessment of the degree of muscle spasm was made from the clinical examination for each side of the spine. This ranged from no spasm (-) to marked spasm (+++), based on the clinical experience of the assessing clinician. The clinician was blinded to the results of the cortical-evoked potentials and VAS and RMAS scores but not to patient symptoms or the remainder of the clinical examination.

Evoked Potential Testing
The subjects were tested while lying on an examination table in the prone position and were awake throughout the procedure.
Magnetic stimulation was performed with a MagPro magnetic stimulator (Dantec Medical, Copenhagen, Denmark) (Fig 1). A stimulation coil with a circular outer diameter of 12.5 cm (Dantec Medical, MC-125) was placed tangentially to the skin overlying the paraspinal muscles, 2 to 3 cm lateral to the mid-line at the L2 to L5 levels. The stimulation wave form was monophasic, with a pulse width of 0.16 ms. The magnetic field at the center of the maximum stimulation output was 2.0 T, and the stimulation rate was 1 Hz. Magnetic stimulation was applied unilaterally, resulting in 2 recordings per patient. The strength of stimulus was  30% of maximum output. No contraction of leg muscles was produced at this stimulation intensity.
Recording electrodes were 8-mm diameter silver/silverchloride disks attached with electrode cream to the skin. Electrode impedance was maintained below 2 kο. Recording electrodes were placed on the scalp 2 cm posterior to the Cz position of the international 10-20 System referenced to Fpz. A ground electrode was placed on the scalp between the pair of recording electrodes. The bandpath filters of the amplifier were set at 5 and 1000 Hz. The analysis time was 100 ms. A total of 128 single sweeps were averaged in each trace, and the test was completed 2 times on each occasion to ensure consistency. Paraspinal, muscle-evoked cerebral potentials on magnetic stimulation were recorded at a rate of 1 Hz in each patient before and after the course of treatment. All positive and negative components within 100 ms from the time of stimulation were isolated for the measurement of latency and amplitude.

Treatment Methods
Patients were asked to attend the clinic 3 to 5 times a week for a period of 2 to 3 weeks. Treatment consisted of soft-tissue massage and manipulation. The soft-tissue treatment included ischemic compression applied to the affected muscles in the back and lower extremity, followed by distraction of the lumbar spine. The ischemic compression was delivered through a thumb-like massage hand tool for 1 to 2 minutes until the pain subsided or disappeared. Distraction was carried out with the patient lying prone on a motorized table with the ankles strapped to the table and the lumbar spine placed in flexion (Leander, Inc). Flexion occurred at 6 to 10 seconds per cycle, and the clinician further increased the stretching by spreading the spinous processes in a caudal-cephalic fashion while the table was moving downward. These distraction maneuvers were also performed in extension and lateral flexion. After this procedure, the patients received high-velocity, low-amplitude lumbar spine adjustments in the side posture position, with the affected side up and with direct contact with the mammillary process at the direction of the clinician. The patients had little or no pain during these procedures. General advice on proper lifting and simple home exercise was given to the patients. The total number of treatments ranged from 3 to 15, with a mean of 8.3 (standard deviation 2.8).

Data Analysis
Evoked potentials, clinical assessment of muscle spasm, and VAS and RMAS scores were recorded at the beginning and end of the treatment period, irrespective of the degree of recovery of the patient. These data were all sequestered until the time of analysis.
The data were entered into the SPSS-PC+ statistical program (SPSS, Chicago, Ill). Mean and standard deviations of each variable were calculated. Multivariate analysis of variance, one-way analysis of variance, and paired Student t tests were used for testing. The level of significance (α) was set at .05.

RESULTS
The data from the before-and after-treatment clinical evaluations for muscle spasm, paraspinal muscle corticalevoked potentials, VAS, and RMAS are noted in Tables 2  and 3. Fig 2 is a representation of the VAS scores before and after treatment. The VAS score before treatment (mean 5.6 ± 1.7 cm) compared with the VAS score after treatment (mean 1.9 ± 2.2 cm) showed a significant difference (P < .001) with a paired Student t test. Improvement was noted in 11 of 13 patients, with 2 patients showing no improvement. No patient showed increased symptoms. Nine of 13 patients showed >50% improvement in the VAS score. Fig 3 is a representation of the RMAS scores before and after treatment. The RMAS was administered 2 times on the day before treatment and 1 time on completion of the treatment. The interclass correlation between the testing on the same day was 0.95. To improve the reliability of this measurement, the results from the 2 separate testings before treatment were averaged. The results before and after treatment were then analyzed for differences. The mean score before treatment was 52.7% ± 19.9%. The mean score after treatment was 24.8% ± 20.6%. A significant difference was found with the paired Student t tests (P = .002). Only 1 patient reported a decrease in activity levels after treatment. Seven of 13 patients showed a >50% improvement in their activity scores. Fig 4 is a representation of the paraspinal muscle spasm findings before and after treatment. To enter the trial, the patients had to have been diagnosed clinically with low back pain with palpable unilateral or bilateral palpable muscle spasm in the lumbar paraspinal muscles. Eleven of 13 patients had palpable muscle spasm bilaterally; the remaining 2 patients were diagnosed with unilateral muscle spasm. After treatment, 11 patients were believed to have had a reduction in muscle spasm, with 2 patients (patients no. 2 and 13) perceived clinically as not showing any appreciable change in muscle spasm after treatment. Patients with bilateral muscle spasm who initially showed improvement tended to show improvement bilaterally.

Cerebral Evoked Potentials
These potentials consisted of both positive (P) and negative (N) components that were designated by custom according to latency in milliseconds: P30, N40, P50, N70, and P90 (Fig 5). The latency of the earliest positive potential (P30) is in the range of 26 to 36 ms when stimulation was applied to the L2 through L5 levels.
When comparing the evoked potentials before treatment with previously published normal controls, 11

Comparison of Clinical Outcomes and Cerebral-evoked Potentials
When defining improvement of cerebral-evoked potentials as showing an increase of the P30-N40 component or both the P30-N40 and N40-P50 amplitudes by >50% after treatment, we were able to identify 10 patients who showed improvement in both palpable muscle spasm and cerebralevoked potentials. One patient had normal evoked response before and after manipulation but showed an improvement in muscle spasm. Two patients had normal evoked potentials before and after treatment and no change in palpable muscle spasm. The sensitivity of evoked potentials in predicting improvement in palpable muscle spasm was calculated to be 91% (10 of 11) and the specificity calculated to be 100% (2 of 2).
We defined an improvement in pain as a 50% reduction in the VAS score. With this definition, the sensitivity of evoked potentials in predicting improvement in pain was 89% (8 of 9) and the specificity 50% (2 of 4). By defining an improvement in function as a 50% reduction in the RMAS score, the sensitivity of evoked potentials in predicting improved activity was 71% (5 of 7) and the specificity 17% (1 of 6).
A multivariate analysis of variance with the evoked potentials before treatment as the dependent variables, out- come of the muscle spasm as the between-subject variable, and the side of the lumbar muscles as the within-subject variable resulted in a statistically significant difference of the combined evoked potentials among the 3 groups who showed different outcomes in the spasm (P = .019). The patients who showed improvement in muscle spasm (ie, total disappearance of spasm (n = 6) or less spasm (n = 5) had depressed cortical-evoked potentials before treatment. The patients (n = 2) who showed no improvement in muscle spasm had normal cortical-evoked potentials before treat-

DISCUSSION
This study confirms the previous report that the amplitude of cerebral-evoked potentials on magnetic stimulation of paraspinal muscles is depressed in the presence of palpable muscle spasm. 11 The amplitude of P30-N40, N40-P50, and P50-N70 components of cerebral-evoked potentials are noted to be lower in patients with palpable muscle spasm  than published normal values and tended to return to normal amplitudes when the muscle spasm improved with treatment. This relation suggests that the paraspinal muscleevoked cerebral potentials may serve as an objective means to assess muscle spasm.
The usage of the amplitude of cerebral-evoked potentials to measure neuronal activity has a different connotation than the more commonly used latency measurements for neuronal deficits. Whereas the latency of an evoked potential depends on the conduction velocity, the amplitude of the response depends more on the volume of receptors that are available for stimulation. Although the amplitude of these responses varies greatly between individuals and is therefore not clinically of value in measuring conduction abnormalities, it is relatively stable in the same individual over a number of days. 10 The ability to attenuate muscle contractionevoked cortical potentials by means of vibration and voluntary contraction of muscles has been interpreted as indicating that the muscle-evoked response is caused by direct stimulation of terminal nerve afferents in the muscle. 11 Lotz et al 12 demonstrated that magnetic stimulation of muscles induced muscle contraction by activation of terminal nerve afferents and not by activation of the muscle fibers directly. In addition, evidence favoring an indirect stimulation of muscles comes from the observation that corticalevoked responses from muscles can be elicited in patients paralyzed by succinylcholine chloride. 13 It is assumed that paraspinal muscle spasm has an effect similar to vibration and voluntary muscle contraction and reduces the number of afferents available for magnetic stimulation. If the muscle spasm saturates the number of available afferent fibers, a diminution of the amplitude of the cortical response would be anticipated. When the spasm diminishes, thereby increasing the availability of 1a-afferents to magnetic stimulation, an increase in the amplitude of the cortical-evoked response would be anticipated.
Although the sample size was small in this study, the evoked potentials appeared to predict which patients were