JCM | Free Full-Text | Using Electric Stimulation of the Spinal Muscles and Electromyography during Motor Tasks for Evaluation of the Role in Development and Progression of Adolescent Idiopathic Scoliosis

In this study, we selected two isokinetic tasks of left and right lateral bending and rotation when standing, since asymmetric muscle activity was expected [44]. As early as 1955, Riddle and Roaf [16] suggested investigating the EMG activity for scoliosis and Cheung et al. [48] suggested evaluating EMG as a predictor of risk of progression. In this study, we anticipated differences in muscle activity across the deformity, since previous studies have demonstrated asymmetry of the convex and concave muscles in AIS when compared to normal spines in lateral bending. It is especially interesting if this would be predictive for the development and determining risk of progression of AIS [21,28,44]. When evaluating local mean and peak EMG signals at the three levels, we were unable to find consistent differences in general. In some instances, convex activation was higher than concave—in particular at the apex level of the primary curve and for subjects with curve progression—as demonstrated in previous studies. However, this is an inconsistent finding [28]. When evaluating for differences in the mean and peak EMG signal ratios, we found differences between the concave and convex sides for a specific motor task of a specific subject but again without a consistent pattern emerging across the subjects, similar to previous studies [28]. However, following the path of EMG activity of progressive and non-progressive AIS, we had two consistent findings. One of the five subjects had rapid progression and one regressed to a straight spine. When evaluating the subjects by adding their absolute EMG ratios for all four motor tasks, the subject with progression had an almost 10-fold less summed mean and peak EMG ratio (X + Y) when compared to the stationary curves and the subject with regression. The latter had a more than 3-fold higher summed mean and peak EMG ratio (X/Y). We interpret the higher ratios as heightened asymmetric spinal muscle activity trying to rebalance the spine to maintain or correct the deformity. When progression occurs, this is preceded or accompanied by a decreased summated ratio across the deformity. Notice that our two mentioned subjects were diagnosed with cerebral palsy, which can affect muscle activity registered by EMG. However, due to their typical AIS curve and chronological age over 14, they may be characterized as AIS cases.

As mentioned, asymmetry in recorded EMG has been an inconsistent finding when utilized to detect the risk of curve progression in the literature [28]. Our findings suggest that rather than detecting asymmetric EMG patterns, we should identify a ‘silencing’ of summed mean and peak asymmetry. When monitored longitudinally, and when curve progression occurs, or preceding the curve progression, we should investigate a decrease in EMG asymmetry (or maybe less normalized EMG activity). Farahpour et al. [44] found that trunk (spine) muscle activities in AIS are asymmetrical and higher than those in normal subjects in AIS. The asymmetry in muscle response was interpreted as perturbations as a compensatory strategy in trying to rebalance the spine and not an inherent characteristic [9,44]. Muscle activity varies with the motor tasks in AIS [23] and asymmetrical muscle activity is induced during anticipatory postural perturbations and symmetrical muscle activity to sudden balance threats [49]. The asymmetrical muscle activity is suggested as a neuromuscular adaptation to the altered mechanical situation of AIS [44]. Based on these findings and our results, we suggest that asymmetric muscle activation is a postural neuromuscular adaptation when the spine rebalances itself towards the midline. Previous studies found that symmetric muscle activity is present in smaller AIS where less mechanical correction is needed [50,51,52], whereas in larger curve AIS, more corrective, asymmetric muscle hyper-activation is required [53,54]. Muscle endurance in AIS is similar to normal subjects initially, but with increased muscle activity and fatigue and there is an increase in convex rotational mobility without additional active postural stability in AIS [55]. Moreover, there seems to be a higher metabolic cost in AIS, thus a higher risk of fatigue [56]. This is indicative of the EMG ratios being affected or diminished as seen in our study. We suggest that asymmetric EMG activity will be detected as a compensatory correcting mechanism as a consequence of neuromuscular efficacy [24,57,58,59]. Approaching fatigue (and initially) when muscles are weakened, the (asymmetric) muscle activity will increase [54], but when the mechanical demands are superseded and following fatigue, the asymmetrical EMG decreases, rotational instability increases, and opens the path for a vicious cycle progression [10,14,55,60,61]. Interestingly, when examining differences in muscle strength for AIS, there is a consistent asymmetric rotation strength deficit and less relative trunk strength with a rotational weakness on the concave side of the primary curve in AIS [20,21,24,25].

In this study, we performed unilateral or bilateral electric stimulation on subjects with small-curve AIS. At the initial visual evaluation, the frontal plane spatial changes in the vertebral bodies when stimulated only showed discrete changes without any apparent pattern. However, when analyzing the spatial changes with calibrated measurements of displacement and rotation (Cobb angle), we detected an overall average general compression of the vertebral bodies of 4.4 mm compression and 0.9° increase in CA, thus a 4% (Equation (3)) and 70% (Equation (4)) increase in CA, respectively, by paravertebral muscular stimulation. In the small deformity AIS, subject number 5 with a small curve (~13°) demonstrated a relatively large compressive effect in the relative ratio of CA/displacement (Equation (5)) with a ratio less than 1 (=0.61) when the spinal muscles were stimulated. In the deformity, with more than 20° of CA, subject 1 with a larger curve (~24°) demonstrated a relatively large increase in the relative ratio of CA/displacement (Equation (5)) exceeding four times (=4.33) when the spinal muscles were stimulated. In deformities less than 7° of CA, we found a 4.8 mm compression or expansion and 1.8° increase in CA, thus a 1% (Equation (3)) and 20% (Equation (4)) increase in CA. Based on this finding, we suggest that in small-curve AIS (CA 10–20°), the spinal muscles do not exert a scoliogenic response but rather have a stabilizing function to maintain the spine in the midline without a major effect/change in CA. Subsequently, the role of the muscles converts from counteracting and stabilizing the spine to being scoliogenic for curves with a CA of more than 20°. Interestingly, there was a vivid and diverse response in both displacement and angle for straight spines, which we interpret as spinal muscle alertness and muscle fatigue in AIS as discussed in the following section. For a deformity exceeding 20° CA, we induced a muscle response with moderate and unilateral electric stimulation (~55) on the concave side of the primary curve in a (the only) female subject at age 14.7. For a small deformity (10–20° CA), we induced a muscle response with a small and unilateral electric stimulation (~34) on the concave side but below the primary curve in a male subject at age 16.5. For a straight spine (CA 57] found that the spinal muscles maintained static and dynamic stability of the spine, thus achieving a corrective effect of a lumbar scoliotic curve. This corrective effect in small-curve AIS was affirmed by inducing a temporary paralysis of the iliopsoas leading to curve correction and derotation [32]. We were unable to identify studies using a similar experimental muscular stimulation as ours; thus, we are unable to relate our findings directly, but lateral electric surface stimulation (LESS) mediates correction after short-duration stimulation for initial scoliosis angles less than 25 ° [34], whereas for curves more than 30–40 °, it generates a shorter term correction [36,37]. This corroborates our findings indirectly. In experimental animal models [34,62], unilateral LESS led to a scoliosis-like deformity with the convexity located on the opposite side to the stimulated side. The extent of the deformity was ascribed to fatigue of stimulated spinal muscles by long-term serial stimulation, which is indirectly consistent with our EMG findings. LESS also accommodates changes in muscle fibres with a conversion to type 1 (fast) in the stimulated muscles [34,63]. This conversion follows histological findings of a significantly lower proportion of type I (slow-twitch oxidative) fibres in the spinal muscles on the concave side of the scoliotic curves in AIS, where the muscles adopt a faster or more glycolytic profile to accommodate reduced low-level tonic activity following curve progression or due to general disuse of the spinal muscles associated with trying to maintain or balance the spinal deformity [27,63]. Muscle fatigue characteristics are similar to normal patients; thus, this fibre conversion is reactive and a physiological response to the increased mechanical demand [50,55]. Long-term LESS seems to mediate compensatory changes in deformity, histology, number of motor endplates and muscle fibre type and long-term LESS seems to lead to exhaustion with a reduced number of motor endplates and histological atrophy [62,64,65]. We speculate if such gradual changes occur in the natural history of AIS with a decline in force level generation as a physiological response dependent on the duration and extent of central and peripheral fatigue and changing from being reversible to irreversible over time [66,67].

When evaluating the EMG data, we also excluded a large number of subjects due to not performing all the selected motor tasks, not performing the five repetitions or the quality of movements and/or the EMG recordings being inadequate. In general, we included subjects as a convenience sample with variance in curve morphology and level, and varied Cobb angles (but was small-curve AIS). Moreover, the subjects were tested at different times from diagnosis and when reaching maturity. All radiological examinations were performed as part of the subject’s clinical follow-up, justifying the induced radiation and some subjects had regressed to an almost straight spine when examined. Since the focus was on small-curve AIS, the risk of progression was small for our population [15]; thus, using EMG for prediction should be considered circumstantial for our population, and it would be unlikely to be able to establish a predictive relationship for asymmetry as expressed in EMG ratios for our subjects. However, this might also be obscured by the other above-mentioned factors and the small number of subjects. Still, we believe that our findings are important to communicate, since electrostimulation is unique and probably will not be performed in the future due to the tenacity of the method.

A pioneering experimental animal study on muscle imbalance by Schwartzmann and Miles [68] produced scoliosis by the unilateral release of the sacrospinal muscles, but release without imbalance maintained a straight spine [68]. Rotational muscle imbalance was already considered an ‘old and forgotten idea’ in 1955 when the unopposed activity of the spinal muscles was hypothesized to cause vertebral rotation and initiate AIS deformity [16,69]. These studies indicate that spinal muscular imbalance plays a role in scoliogency; with our study, we would like to reintroduce the spinal muscles as an important factor in scoliogency by suggesting two mechanisms. Firstly, spinal muscles mediate stability and counteract scoliosis in the straight spine and small-curve AIS but are scoliogenic in further curve progression (>~20 ° of CA). Specific muscle exercises that seem to reduce the risk of AIS progression are indicative of this role of the spinal muscles [52,70,71]. Secondly, in the process of maintaining spinal balance, fatigue and morphological changes can occur. If the spinal muscles are unable to maintain balance, then curve progression occurs, and this is preceded or accompanied by the subsidence of EMG asymmetry as a heralding sign. We suggest that rather than looking for asymmetry in EMG responses, the detection of reduced EMG asymmetry could be taken as a sign that the spinal muscles are fatigued and failing to maintain stability for the spinal deformity. When the initial stabilizing spinal muscle actions are exhausted, they will be unable to counteract the deformity and change their role to becoming scoliogenic. Proof of these two mechanisms requires larger prospective studies with longitudinal EMG monitoring, showing the suggested decrease in EMG ratios in deformity progression, and tests of interventions with a focus on specific rotational muscle strengthening for influence on the natural development of AIS [10].