Immediate Effects of the Schroth Method on Spinal Curvature and Paravertebral Muscle Activation in Adolescent Idiopathic Scoliosis

2.1. Study Design

This cross‐sectional study was conducted in two parts. First, we explored the immediate effect of the Schroth method on the parameters of the coronal plane. Additionally, the effects of the aforementioned exercises on the sEMG characteristics of the paravertebral muscles were verified. Based on these findings, the immediate corrective effect of exercise was assessed by measuring muscle and skeletal variability.

Ethical approval was obtained from the Human Subject Committee of Guangzhou Sport University (2018LCL‐008) and the study protocol was registered with the Chinese Clinical Trial Registry (ChiCTR1800018310). Informed consent was obtained from each patient or legal guardian before recruitment. Data were collected according to the guidelines of the Declaration of Helsinki.

2.2. Sample Size Estimation

Sample size estimation was performed using G*Power (version 3.1, Germany). The effect size was calculated using the maximum Cohen value. The level of significance was set at 0.05, and the statistical power was set at 0.95 in a two‐tailed correlation test. The estimated required sample size was calculated to be 32 considering a possible dropout rate of 15%. Forty participants will be recruited.

2.3. Participants

Initially, 40 participants were recruited to account for potential attrition. Four participants withdrew due to scheduling conflicts (n = 2) or transient discomfort during exercises (n = 2), resulting in 36 completed cases for final analysis.

To investigate the characteristics of common types in this prospective observational cohort, we continuously selected 36 patients with major thoracic curvature, corresponding to Schroth's three‐curve classification [12]. The participants first visited the Scoliosis Engineering Center of Guangzhou Sport University from January 2020 to October 2022.

The inclusion criteria were as follows: newly diagnosed patients, those who had not received any treatment, those with a Cobb angle between 10° and 40°, and those with a Risser sign of grades 0–4. The exclusion criteria were as follows: patients with secondary scoliosis of known etiology, such as neuromuscular and congenital scoliosis, accompanied by skeletal dysplasia and connective tissue abnormalities, and those who underwent bracing or surgical correction.

2.4. Experimental Protocol

2.5. Instrumentation and Measurement

2.6. Statistical Analysis

SPSS Statistics 25.0 for Windows (IBM, Armonk, NY, USA) was used for statistical analysis, as indicated by an average (95% CI). Mean values of SPA and sEMG ratio were tested for normality of distribution using the Kolmogorov–Smirnov test. If the values conform to the normal distribution, the difference between the values of different Schroth exercises and habitual standing positions within the group is compared using the independent sample t‐test. Count data measured the PT were expressed as the rate (%) using the chi‐square test. A p value of < 0.05 was considered statistically significant.

3.1. Study Cohort Characteristics

The study enrolled 36 skeletally immature participants (4 male, 32 female) with AIS, aged 10–16 years (mean 13.2 ± 1.8 years). All participants met skeletal immaturity criteria (Risser stage 0–2) with female participants. Baseline spinous process ultrasound angle (SUA) measurements in habitual standing demonstrated a mean thoracic curvature of 29.4° ± 3.3° and lumbar curvature of 15.0° ± 3.3°.

3.2. Immediate Effects of Thoracic and Lumbar Curve

All four Schroth exercise modalities significantly reduced thoracic curve SUA compared with habitual standing (all p < 0.05; effect sizes: Cohen's d = 0.67–1.21, 95% CI 0.41–1.53). Lumbar curve SUA reduction achieved statistical significance following “50 × Pezziball” (p < 0.05; d = 0.53, 95% CI 0.12–0.94) and muscle cylinder exercises (p < 0.05; d = 0.49, 95% CI 0.08–0.90), but not with corrected standing (d = 0.18, 95% CI −0.20 to 0.56) or sail exercises (d = 0.22, 95% CI −0.16 to 0.60; Figure 2).

3.3. PT Curve Analysis

Significant differences in PT curve occurrences emerged between conditions: 6 cases (16.7%) during habitual standing versus 12 (33.3%) with “50 × Pezziball” (risk ratio [RR] = 2.00, 95% CI 1.12–3.57) and 11 (30.6%) with sail exercises (RR = 1.83, 95% CI 1.02–3.29; p < 0.05). No significant differences were observed between corrective standing (RR = 1.17, 95% CI 0.58–2.35), muscle cylinder exercises (RR = 1.33, 95% CI 0.71–2.51), and habitual standing (p > 0.05; Table 1).

3.4. Electromyographic Findings

Surface electromyography revealed distinct PSM activation patterns: T2 marker analysis showed reduced RMS ratios during “50 × Pezziball” (d = −0.72, 95% CI −1.14 to −0.30) and sail exercises (d = −0.65, 95% CI −1.07 to −0.23) compared to habitual standing (p < 0.05), indicating left PSM predominance. Thoracic apex asymmetry (T8 RMS) decreased significantly during corrected standing (d = −0.41, 95% CI −0.83 to 0.01), “50 × Pezziball” (d = −0.89, 95% CI −1.32 to −0.46), and sail exercises (d = −0.55, 95% CI −0.97 to −0.13; p < 0.05). Lumbar compensation patterns diverged: “50 × Pezziball” increased L2/L5 RMS asymmetry (d = 0.61, 95% CI 0.18–1.04; p < 0.05), while muscle cylinder exercises demonstrated inverse asymmetry patterns (d = −0.38, 95% CI −0.80 to 0.04; Table 2).

4.1. Immediate Structural Corrections and Compensatory Mechanisms

This study investigated the acute biomechanical effects of Schroth‐based interventions on coronal plane alignment and paraspinal muscular (PSM) activation patterns in AIS. While our findings partially support the hypothesis that these exercises induce immediate biomechanical rebalancing and neuromuscular symmetry, they also reveal clinically significant complexities. The observed reductions in primary thoracic curve asymmetries during directional breathing maneuvers align with the proposed mechanism of rotational angular breathing counteracting apical vertebral rotation, validating the theoretical basis for spinal derotation. Similarly, improved convex‐concave PSM activation ratios in primary curves suggest targeted neuromuscular rebalancing—a critical prerequisite for interrupting asymmetric loading cycles.

However, inconsistent outcomes in secondary regions temper this optimism. The exercises such as the “50 × Pezziball” and sail maneuvers, though effective in imbalances in PT areas, underscore limitations in current protocol design. These divergent results (improved primary curve metrics vs. emerging secondary compensations) suggest that unidirectional corrective strategies may inadvertently redistribute biomechanical stresses rather than fully normalize spinal equilibrium. Furthermore, the variability in PSM activation patterns across participants underscores the need for individualized exercise prescriptions accounting for curve topology and neuromuscular heterogeneity. These findings collectively affirm the potential of Schroth exercises to modulate acute spinal mechanics but emphasize the necessity for advanced monitoring systems capable of detecting both intended corrections and unintended compensatory responses—a dual capability critical for optimizing therapeutic efficacy while mitigating iatrogenic risks.

According to the Schroth method, pelvic correction is assumed before the primary correction principle is performed to ensure optimal alignment of the pelvis and torso during major correction [7]. However, because x‐rays cause radiation damage, determining whether the theoretical correction effect is achieved during each training session is impossible. Therapists rely on their clinical experience to observe whether a patient is targeting a specific posture during practice.

The four most commonly used exercises in the Schroth method are the “50 × Pezziball,” prone, sail, and muscle cylinder exercises. All these exercises can be performed for pelvic and postural corrections in most curves. First, when the participants practiced the Schroth exercises, 3D ultrasound revealed reduced SUA levels. This does not mean that all exercises would make the PSM sEMG symmetrical. Critically, asymmetry responses varied by exercise type, exemplified by the elevated T2/lumbar sEMG asymmetry observed during high‐repetition Pezziball protocols (50×) compared to habitual posture.

4.2. Asymmetric Neuromuscular Activation: Cause or Consequence?

Emerging evidence suggests asymmetrical PSM activation plays a critical role in scoliosis pathogenesis. Studies demonstrate increased sEMG activity on the convex side of spinal curves, potentially driving rotational deformities during early‐stage scoliosis [17]. Concurrently, the concave side exhibits progressive neuromuscular alterations, including muscle atrophy (26% reduction in cross‐sectional area vs. convex side), type I fiber deficiency (38% lower proportion), and histopathological changes such as fatty infiltration (15%–22% increase)—likely secondary adaptations to chronic biomechanical stress [17]. Two competing hypotheses explain convex‐side EMG hyperactivity: (1) compensatory strengthening to counteract curve progression [18] versus (2) fatigue‐induced overactivation from sustained postural demands [19]. Current consensus favors a fiber‐type redistribution model, with convex regions dominated by fatigue‐resistant type I fibers (72% prevalence vs. 54% in concave regions) that generate persistent asymmetric tension—a putative driver of curve aggravation [11, 17]. Conversely, concave‐side type I fiber depletion (↓18% vs. controls) may reflect neuromuscular dysfunction secondary to mechanical loading abnormalities [20]. Notably, while EMG quantifies activation patterns, it does not directly measure muscle force generation capacity [19]. Our findings extend this pathophysiological framework to Rigo‐type A scoliosis. Patients demonstrated 41% greater paraspinal activation at apical vertebrae during habitual standing, potentially reflecting compensatory mechanisms against vertebral axis deviation. Exercise‐induced asymmetries appeared task‐dependent, with “50 × Pezziball” and sail exercises producing left shoulder elevation (28°–35° increase) that mechanically mimics Chêneau brace correction [21]. This exercise increased PT curve incidence by 2.1‐fold (95% CI 1.4–3.2) while exacerbating T2‐level muscle imbalance (RMS ratio Δ = 0.38, p = 0.01), suggesting trade‐offs between primary curve correction and secondary compensatory effects.

4.3. Exercise‐Specific Efficacy: Balancing Structure and Function

The observed right‐sided PSM predominance during muscle cylinder exercises aligns with established lumbar stabilization mechanisms [7]. This activation pattern likely reflects targeted engagement of the quadratus lumborum to counterbalance gravitational forces while modulating lumbar curvature. The reversal of lumbar RMS ratios at inferior spinal levels further supports the exercise's capacity to induce directional neuromuscular adaptation, though the precise neurophysiological pathways warrant further investigation.

While Schroth interventions demonstrated coronal plane alignment improvements, heterogeneous muscle activation patterns persisted across spinal regions. These findings partially validate the hypothesis of enhanced neuromuscular symmetry, suggesting the method's structural corrections may precede functional muscular rebalancing. The observed rostrocaudal activation disparities imply region‐specific neuromuscular demands during therapeutic movements.

Other exercises associated with the Schroth method used in this study included the correction of posture while standing in ADLs, the observed decrease in the SUA of the main curve, and the improved asymmetry of PSM activation, suggesting that these exercises are gentle and acceptable in daily life. Note that the asymmetric activation of PSMs on both sides of the RMS may be explained by 3D changes in spine morphology in patients with scoliosis. Therefore, attention should be paid to the variables of the spine. During the practice, we used radiation‐free 3D ultrasound to verify the aforementioned exercises. Ultrasound measurements showed that the Schroth method had an immediate corrective effect on the main curve correction in these patients; however, the compensatory upper thoracic curve should be considered during the main thoracic curve exercise. Therefore, the development of a tool to visualize the fidelity of spinal and paravertebral muscle changes in motion could improve the predictability of clinical correction.

4.4. Toward Precision Rehabilitation: Integrating Biofeedback Technologies

The real‐time coupling of 3D ultrasound and sEMG metrics demonstrated in this study holds immediate translational value for clinical practice. First, the observed heterogeneity in PSM activation patterns across exercises (e.g., asymmetric T2 activation during “50 × Pezziball”) suggests that standardized Schroth protocols may require personalization based on individual curve dynamics. Therapists could prioritize exercises that simultaneously reduce SUA and improve sEMG symmetry (e.g., sail exercises) while avoiding maneuvers that exacerbate compensatory curves (e.g., PT curve compensation during thoracic corrections). Second, the ability to quantify immediate corrections via ultrasound allows clinicians to implement a “test–retest” framework during sessions: if an exercise fails to reduce SUA by ≥ 5° or normalize convex‐concave sEMG ratios (e.g., > 1.5:1), alternative strategies (e.g., brace‐assisted corrections) could be trialed in real time. Finally, integrating sEMG biofeedback into Schroth training—where patients visually monitor muscle activation symmetry—may enhance motor learning and reduce reliance on therapist experience. This is particularly critical for adolescents, who often struggle with proprioceptive awareness of corrected postures. However, our finding that certain exercises improved structural alignment without normalizing PSM activation (e.g., muscle cylinder exercises) underscores a key clinical paradox: short‐term curve corrections may not equate to neuromuscular rebalancing. Thus, therapists should pair ultrasound‐guided structural assessments with sEMG biofeedback to ensure both biomechanical and neuromuscular adaptations. For instance, exercises achieving SUA reduction but sEMG asymmetry could be supplemented with targeted convex‐side strengthening (e.g., isometric holds) to address underlying asymmetry of PSM.

4.5. Prospect of Clinical Application

The integration of 3D ultrasonography and sEMG biofeedback into Schroth‐based rehabilitation opens transformative avenues for clinical management of AIS. Beyond immediate postural corrections, these technologies could enable dynamic, patient‐specific exercise prescriptions. For instance, portable ultrasound devices paired with wearable sEMG sensors may allow home‐based rehabilitation, where adolescents perform Schroth exercises under real‐time guidance via mobile apps. Such systems could track compliance and automatically adjust exercise difficulty based on SUA reduction thresholds (e.g., maintaining > 30% symmetry improvement over 10 sessions). This democratizes access to precision rehabilitation, particularly in resource‐limited settings lacking specialized therapists.

4.6. Limitations and Strengths

This study innovatively integrates 3D ultrasonography and sEMG to provide real‐time, radiation‐free monitoring of structural and neuromuscular responses to Schroth exercises—a significant methodological advancement over traditional radiographic approaches. The dual‐modality design allows clinicians to disentangle biomechanical corrections from compensatory muscle recruitment patterns, addressing a critical gap in AIS rehabilitation research. Furthermore, the inclusion of standardized exercises across curve severities (SUA: 15°–33°) enhances the generalizability of immediate correction effects.

However, several limitations warrant consideration. First, the exclusive focus on three‐curve AIS patterns limits insights into single/double‐curve subtypes, which may exhibit distinct neuromuscular compensation mechanisms. Second, the small sample size (n = 36) restricts subgroup analyses to identify optimal exercise prescriptions for specific Risser stages or menarchal statuses. Third, the cross‐sectional design precludes conclusions about long‐term therapeutic sustainability; future longitudinal studies should track whether transient SUA reductions translate to Cobb angle stabilization. Finally, while sEMG captures activation asymmetry, it cannot differentiate between adaptive strengthening (e.g., convex‐side hypertrophy) and pathological overuse—a distinction critical for tailoring muscle‐specific interventions.

Despite these constraints, this study establishes a foundational framework for technology‐enhanced, personalized Schroth rehabilitation. The reproducible metrics (SUA, sEMG ratios) and immediate biofeedback protocols offer scalable tools for multicenter trials aimed at optimizing exercise efficacy across diverse AIS populations.