Scoliosis-Specific Exercises Applied During Bouldering Therapy or Standard Physiotherapy Over One Year Have Similar Effects on Deformity in Mild AIS

Abstract

Study Design

Randomized controlled trial over one year.

Objectives

Bouldering as a popular climbing discipline might have a therapeutic potential in adolescent idiopathic scoliosis (AIS). This trial examined the effectiveness of an innovative program compared to standard care.

Methods

Patients suffering from mild AIS (13.1 ± 1.5 yrs, ♀ = 68%) were weekly treated by scoliosis- specific exercises (SSEs) during bouldering therapy (BT: n = 20, Potsdam model) or physiotherapy (PT: n = 21, Schroth-Method). After 12 months, changes in group-blinded max and sum score of Cobb-Angles as well as angle of trunk rotations (ATRs) were mainly analyzed by use of a mixed-ANOVA.

Results

Disregarding group, max Cobb: −1.9° ± 4.9°; sum Cobb: −3.7° ± 8.4° (excl. one outlier/group); max ATR: −1.3° ± 1.5°; sum ATR: −2.6° ± 2.7° were reduced significantly (main time effects: P < 0.001 to 0.021, f = 0.39 to 1.05). Post-hoc tests revealed significance in both groups for max and sum ATR (p_adj < 0.001 to 0.006, d = 0.56 to 1.03). A significant group-by-time interaction (P = 0.039, f = 0.202) was found in sum ATR (BT: −3.48° vs PT: −1.76°). The improvements in ATR but not Cobb-Angle were below measurement error on group level. In each group, 20% improved and 5% deteriorated to a clinically meaningful extent in max Cobb (≥5°). Regarding max ATR, 15% (BT) and 5% (PT) revealed an improvement ≥4°. No clinically relevant deterioration was seen. 85.7% wanted to continue with BT after intervention period. Most (74%) did not interpret it as a kind of treatment.

Conclusions

Both, SSEs during BT and PT over one year can prevent potential progression of mild AIS and partly improve deformity. The BT program is suitable for youth and might improve adherence to specific exercises until or even beyond skeletal maturity.

Introduction

Exercise therapy as a treatment for adolescent idiopathic scoliosis (AIS) is a current issue with increasing publications especially in the last decade. ¹ According to a recent Cochrane review, best type of therapeutic exercising for AIS is still unclear due to limited evidence. ² By citing one long-term randomized controlled trial (RCT), ³ review authors report low-certainty evidence favoring scoliosis-specific exercises (SSEs) over unspecific ones. ¹ International guidelines recommend SSEs to prevent curve progression and bracing. ⁴ The present article evaluates a novel program including SSEs during therapeutic bouldering. Bouldering is a popular climbing discipline performed near ground level. The great potential lies in, i. a., the motivating character of a climbing wall. ⁵ Therapeutic climbing or bouldering as an attractive treatment option, especially for long-standing treatments as often necessary in AIS, could open up a chance to promote adherence – a key factor of therapy success. ⁶

The therapeutic effects of climbing were already discussed at the end of the 19^th century. ⁷ At that time, climbing was mainly reserved as an outdoor activity, more specifically called mountaineering. However, growing amounts of indoor climbing walls in the last decades have opened up further therapeutic possibilities. Thus, climbing is not only practiced by athletes but also by people suffering from different diseases or injuries. A recent meta-analysis demonstrated some evidence of its safe application and effectiveness on physical, mental as well as social health dimensions. ⁸ However, authors claimed a need of more high-quality trials.

In orthopedics, climbing as an exercise therapy has the potential to treat postural weaknesses and instabilities. In patients suffering from back pain, it was found that a climbing therapy regime can significantly improve health perception, core strength and mobility.^9,10 Healthy sedentary adults, ¹¹ children and adolescents, ¹² can benefit from climbing to improve such physical abilities, too. However, studies considering minor patients in orthopedic contexts are rare as previously reviewed.^13-16

Regarding AIS, a climbing training of 24 sessions of 90 min has resulted in significant improvements of strength and mobility similarly for the convex and concave segments of the spine. ¹⁷ Ascending routes or traversing walls are seen as unspecific exercises in terms of a scoliotic spine. In contrast, specific exercises aiming to strengthen core muscles for a 3D-autocorrection are preferred in the treatment of AIS. ⁴ Thus, SSEs have been developed and implemented in a therapeutic bouldering program according to the “Potsdam Model” (PM). ¹⁸ It has been proofed to be feasible and attractive for adolescents. ¹⁹

The PM is a method to design therapeutic exercises on a climbing wall.^20,21 Additionally, load can be controlled for a standardized training regime. It was already used in the treatment of shoulder impingement and stroke.^22-25 The PM is also used as the methodological basis of the intervention in the here presented RCT.

The aim of the trial was to analyze the effectiveness of the bouldering therapy (BT) program on AIS in comparison to standard physiotherapy (PT) in our region (mainly Schroth-Method). Schroth is the most evaluated method for AIS. Various meta-analyses confirmed its effectiveness,^26-28 and superiority to unspecific core exercises with “limited evidence”.^29-31 The primary analysis considered the hypothetical influences of both interventions on the shape of the spine, i.e., Cobb-Angle and angle of trunk rotation (ATR).^32,33

Method

This section was prepared by using CONSORT reporting guidelines. ³⁴

Patients

A priori sample size calculation revealed a minimum of n = 34. Details of the calculation can be found in the supplemental material (P. 1). Due to the long study period (12 months) a drop-out rate of 25% was chosen. ³ Thus, an additional recruitment of up to 12 participants seemed to be reasonable.

In total, 43 patients of Oberlin Clinic or its medical centers in Potsdam, Germany, were consecutively recruited (♀ = 30; ♂ = 13). The recruitment rate was 84%. The study team decided to stop recruitment after 18 months due to very low intermediate drop-out rates. Patients were asked to participate during their doctor’s appointment. The inclusion criteria were a minimum age of ten and maximum of 16 years, an idiopathic scoliosis with a Cobb-Angle between 10° and 25° and a medically certified sports capability. Exclusion criteria were a known cause of scoliosis, cessation of growth (Risser sign: 5), a continuous treatment >12 months and bracing. Figure 1 illustrates inflow and outflow of patients.

Figure 1.

Schematic Sequence of the Study. SSE: Scoliosis-Specific Exercises. Note: Results of the Physical Fitness Tests Will be Addressed in Another Article (in Preparation).

The scoliotic curve types were classified according to the apical vertebral body and dimension of the curve. ³⁵ Major curves were defined as Cobb-Angles ≥10° and minor ones <10° (rounded off to whole numbers). The full Lenke classification system could not be used since only mild forms of scoliosis were considered and side bending radiographs were not taken.

The study was conducted according to the guidelines of the Declaration of Helsinki and approved by the Ethics Committee of the University of Potsdam, Germany (protocol code 51/2018; 11.03.2019). All participants and parents gave their written informed consent to participate.

Intervention

The experimental group took part in a supervised BT program with 4 subgroups of 5 persons each. BT was performed over 12 months as a replacement for the usually prescribed physiotherapy. An exercise therapist (M. Sc.) – additionally qualified in therapeutic bouldering – carried out the BT sessions. These were conducted on 2 pivoted walls (±15°, Figure 2: a primary 5 m × 3 m one (DIY brand) and adjacent 1.25 m × 2.50 m one (Physioclimb Standard).

Figure 2.

Therapeutic Bouldering Walls From Left to Right: Main Wall (DIY Brand), Physioclimb Standard (Miesbach, Germany).

Basically, the same intervention was delivered to all subgroups and performed as planned. It was marginally modified according to the group needs. Each session lasted 60 min once per week. After warm up on the floor (approx. 10 min) including games and coordinative exercises, BT sessions consisted of 2 main parts (each 20 min ± 5 min):

• 1. The bouldering only part, which contained playful teaching of climbing principles and techniques.

• 2. The therapeutic part, which included preparatory floor exercises and scoliosis-specific exercises (SSEs) performed at the bouldering wall. ‘Specific’ means that the exercises were tailored to the individual shape of the spine with an additional 3D-autocorrection.^18,19

As a prerequisite for SSEs, climbing techniques needed to be imparted. This was done in the first main part of the BT session by use of bouldering games and boulder problems (sequence of holds) forcing specific climbing techniques.

In the second part, preparatory floor exercises needed to be performed in the first weeks and later implemented in the circuit training. Regarding SSEs, the PM was applied. The model structures climbing movements to design therapeutic exercises and to control loading.^20,36 Four extremities on separate holds represent start and end position of a movement (square’s position: Figure 3A). Such a position corresponds to a relaxation phase since least physical effort is demanded to stay on the climbing wall. If one extremity releases a hold, the contralateral one is loaded particularly and represents a loading phase (triangle’s position: Figure 3B). In this way, bouldering exercises can be used in a systematic training regime including controllable relaxation and loading phases as well as training parameters.

Figure 3.

(A) Examples of Square’s Position and (B) Triangle’s Position according to the Potsdam Model of Climbing Therapy.

The following training parameters were standardized within a session. Triangle positions were maintained 3 – 8 times for 2 – 5 s depending on exercise type and training progress. Generally, 2 – 3 sets of 1 or 2 SSEs were performed. According to the heterogeneity of the sample, theses bouldering exercises had to be tailored to the curve shape of the spine and trunk rotation of each participant (individualization). After 3 ± 1 sessions, a new exercise was introduced to elicit variety in the program. All exercises with training parameters and individualizations can be found in the supplemental materials (Table S1 and S2). Exact loading intensity was not calculated due to the high complexity of climbing movements and biomechanical differences between participants. However, it can be estimated by counting durations of triangle positions until exhaustion within one set (40 – 60 s = 40 – 60% of maximum strength; 8 – 30 s = 60 – 80%; 1 – 4 s = 90 – 100%). ³⁶ At the beginning, estimated intensities of <30% were used to focus on 3D-autocorrections without exhaustion. Intensity was increased over time according to the progression principle of training. However, correct posture during exercises were a necessary criterion for tailored progression.

The final part (5 – 10 min) included relaxation exercises, reflection of the session with an education of the importance of a permanent 3D-autocorrection, introduction and reminder of daily unsupervised home exercises (Table S1 and S3). Sometimes a little bouldering game rounded off the session. Games were implemented to deliver joyful sessions and to motivate for a continuous exercising.

Controls

The active control group got standard PT according to the Schroth-Method – a type of SSE – in an outpatient setting. In one case, Schroth was not possible due to missing local Schroth-therapists. Thus, other physiotherapeutic SSEs were performed. In 2 other cases PT consisted of a mixture of Schroth and other SSEs. These cases were not excluded from intention to treat analysis. Treatment frequency was documented (generally 1 – 2 times/week for 20 – 30 min).

Outcomes – Data Acquisition and Processing

• 1. Radiography:

Radiological AP and LAT studies were taken while standing 1.5 ± 1.8 months before (min: <1; max. <7 months) and 0.7 ± 0.9 months (min: <1; max. <4 months) after the end of the interventional phase. A microdose biplanar linear radiography (EOS^TM imaging system GmbH, Frankfurt, Germany) was used without loss of image quality.^37,38 Due to the consecutive recruitment strategy, time period between examinations and start of intervention could not be fully standardized. BT group usually received physiotherapy in between (3.1 ± 4.6 sessions). The exact amount of therapeutic exercise sessions per patient was documented until the second radiological examination (Table S4).

In one case at t0 and 2 cases at t2, a digital radiography was already available. Thus, no new EOS image was taken to reasons minimize radiation exposure. Due to high inter-method reliability and excellent between-method agreement of digital radiography and EOS imaging, both can be used interchangeably. ³⁹

The main outcome – Cobb-Angle – was calculated semi-automatically by use of the EMR-manager (Version: 7.2, Vepro AG, Pfungstadt, Germany). Especially in small curves, the detection might be difficult. To minimize the measurement error in detecting the highest measurable Cobb-Angle, 2 specialists in orthopedics analyzed all EOS images. Both raters were blinded with respect to the group. One rater was additionally blinded to the measurement time point. The other one was the attending physician during the study period and, therefore, could not be blinded regarding time points.

Raters were instructed to take every convexity of ≥10° in thoracic and lumbar spine (coronal plane) into account. Of these resulting Cobb-Angles, the highest (max Cobb) was used for further analyzation. If 2 or up to 4 Cobb-Angles ≥10° were present, the data set with the highest sum score was used (sum Cobb). ⁴⁰ For an intraindividual comparison of sum scores between t0 and t2, the identical number of Cobb-Angles was a prerequisite. The intraclass correlation coefficients (two-way mixed, absolute agreement, single values) between both raters were good to excellent ranging from 0.90 to 0.93 for max Cobb and 0.79 to 0.83 for sum Cobb. The average measurement error of Cobb-Angles is about 3° between and 1° to 2° within raters. ⁴¹

Furthermore, pelvic obliquity (bilateral height difference of the most prominent point of iliac crest in mm) ⁴² and Risser sign (scores: 0 – 5) ⁴³ as a general measure of skeletal maturity was determined by use of the already given radiological images. Side of pelvic obliquity did not change from t0 to t2.

• 2. Clinical examination:

Patients were examined by the attending physician clinically 0.5 ± 0.8 months before and 0.1 ± 0.4 months after the end of the interventional phase (each min: <1; each max. <3 months).

Main outcomes were the ATR measured by a scoliometer during Adam’s forward bend test. ⁴⁴ It has revealed good to high intra- and interrater reliability (r = 0.86 – 0.97).^32,45,46 Maximum (max ATR) and sum scores (sum ATR) were used as statistical parameters.

Secondarily, fingertip-to-floor distance (FFD) in cm was measured, as well as Ott’s- and Schober’s sign. ⁴⁷

During clinical examinations, group blinding could not be maintained completely at t2. This must be considered in the data interpretation.

• 3. Pain:

The actual sensation of pain and pain over the last 2 months was questioned before and at the end of the interventional phase. Questions came from the ‘German Questionnaire of Pain for Children and Adolescents’. ⁴⁸ The outcome was the score on the pain scale reaching from 0 to 10.

• 4. Adverse events, adherence and interest in continuing BT:

Adverse events, except for trivial ones like muscle soreness or skin irritations, were documented. Participation in therapy sessions and reasons of absence were used to calculate the adherence rate in relation to the number of offered sessions: $\frac{number\hspace{0.17em}of\hspace{0.17em}attended\hspace{0.17em}therapeutic\hspace{0.17em}sessions}{total\hspace{0.17em}number\hspace{0.17em}of\hspace{0.17em}offered\hspace{0.17em}therapeutic\hspace{0.17em}sessions}$ .

After the intervention phase, interest in continuing BT was anonymously questioned by form.

Study Design and Procedure

The phase III interventional study characteristics were randomized, controlled and partly single blinded. The random allocation sequence was generated by drawing lots (balanced grouping) for ten consecutively recruited patients who fulfilled all pretests. The simple randomization technique was conducted by the study manager and audited by an independent person. This resulted in 4 subgroups with 5 study participants per type of treatment as mentioned above (parallel assignment). The study manager assigned participants to the groups and was the only one who had access to this list. Blinding of patients was not possible due to the active intervention. In contrast, examiners during data acquisition were blinded regarding group with one exception: As mentioned above, clinical examinations could not be blinded completely at t2. The sequence of the study is shown schematically in Figure 1.

Statistical Analysis

SPSS Statistics 27 (IBM, Armonk, New York, USA) was used for statistical comparisons. For unpaired t-tests, normal distribution was checked with the Shapiro-Wilk test for metric data. In case normal distribution was not fulfilled for all comparative variables, non-parametric tests were used. If heterogeneity of variances was present (Levene test), Welch-correction was applied. Ordinally scaled data like “climbing experience” were analyzed by non-parametric tests (Mann-Whitney U-Test).

For main analysis, mixed ANOVA for within and between subjects was performed. Residuals were not always normally distributed (Shapiro-Wilk test: P > 0.05). However, mixed ANOVA seem to be sufficiently robust against violation of the assumption of normal distribution.^50-52 Error variances and covariances were homogeneous (Levene test, Box test: P > 0.05, respectively). The effect size partial eta squared (η²) was converted into Cohen’s f by = $\sqrt{\frac{{\eta }^2}{1-{\eta }^2}}$ . ⁵³ Significance level for main (time) and interaction (group-by-time) was set at α = 0.05. If significant results were present, Bonferroni corrected pairwise comparisons were performed. Additionally, mean differences (t2−t0) and 95%-confidence intervals (CI) of all parameters measured twice are presented. One patient in PT group was lost to follow up of radiological parameters (missed examination). Generally missing data were removed from analysis.

Due to mixed ANOVA’s less robustness against outliers, those were removed in a second step. For that, cut-off z-standardized value was set > |2.68| representing 1.5-fold interquartile range, ie, 99.632 % of all values were still included. Results of mixed ANOVA are presented with and without outliers.

Results

The 41 study participants were 13.1 ± 1.5 years of age, weighted 52.0 ± 11.1 kg and were 164.9 ± 11.0 cm tall. Table 1 shows patient characteristics separated per group and measurement time point (before = t0 and after = t2 intervention period). Between group analyzes revealed some differences highlighted in bold. The scoliotic curve types were very heterogeneous (Table 1). In 39% of the sample, an atypical ATR was present, ie, rib hump and/or lumbar bulge was atypically located on the side of the concavity.

Table 1.

Patient Characteristics (Means ± Standard Deviations) at Start and End of the Study

	Bouldering therapy group f = 13; m = 7		Physiotherapy group f = 15; m = 6
	t0	t2	t0	t2
Age (in years)	13.3 ± 1.5	14.3 ± 1.5	12.9 ± 1.6	13.9 ± 1.6
Height (in cm)	164.7 ± 11.1	168.3 ± 10.0	165.0 ± 11.1	168.7 ± 9.9
Weight (in kg)	50.8 ± 9.4	54.2 ± 9.8	53.2 ± 12.7	56.7 ± 12.1
Physical activity a (in min/week)	242.1± 210.4	242.2 (n = 19)± 90.8	251.6± 142.1	327.6 (n = 20)± 193.1
Current treatment sessions (no./week)	0.8 ± 0.5	1.0 ± 0.2	0.8 ± 0.7	1.0 (n = 20) ± 0.4
Home-based exercise sessions (no./week)	2.3 ± 2.2	1.4 ± 1.8	1.9 ± 2.8	1.5 (n = 20) ± 1.7
Climbing experience (0 = never; 1 = once; 2 = few times; 3 = several times; 4 = regularly)	1.5 ± 1.0	4.0 ± 0.2	2.2 ± 1.0	2.4 (n = 20) ± 0.9
Instructed climbing courses (no.)	0.4 ± 1.1	1.4 ± 1.1	0.9 ± 1.9	1.1 (n = 20) ± 2.3
Pain spine last 2 month (NAS 0-10)	2.3 ± 3.0	1.6 ± 2.3	1.6 ± 2.5	1.9 (n = 20) ± 2.6
Ott’s test (rel. increase in cm)	4.5 ± 1.8	4.0 ± 0.8	3.8 (n = 17) ± 1.3	3.6 ± 1.0
Schober’s test (rel. increase in cm)	4.6 ± 0.9	4.3 ± 1.0	5.2 (n = 18) ± 0.7	4.9 ± 1.0
Fingertip-to-floor distance (in cm)	8.8 ± 9.2	7.8 ± 9.6	6.0 (n = 19) ± 7.5	3.4 ± 5.7
Risser-sign (grading scale 1-5)	1.9 ± 1.7	3.0 ± 2.0	1.7 ± 1.8	2.6 (n = 20) ± 1.9
Pelvic obliquity (in mm)	3.9 ± 4.5	4.2 ± 3.9	2.6 ± 3.0	4.8 (n = 20) ± 4.0
Max cobb-angle (in °)	15.5 ± 3.5	13.7 ± 4.8	14.5 ± 3.4	12.5 (n = 20) ± 5.1
Sum score cobb-angle (in °)	25.4 ± 10.3	23.0 ± 14.6	23.7 ± 11.5	21.4 (n = 20) ± 17.3
Max ATR in ° (max angle of trunk rotation)	6.5 ± 1.9	4.9 ± 1.9	5.7 ± 1.4	4.5 ± 2.0
Sum score ATR (in °)	10.6 ± 3.5	7.1 ± 3.2	8.3 ± 3.0	6.5 ± 3.0
Atypical ATR (no.) (ATR on side of concavity)	8/20	6/17	8/21	6/13
Type of scoliosis (no.) (acc. apex vertebra)
No scoliosis	0	3	0	7
Main thoracic	1	1	5	3
Double thoracic	4	1	0	1
Double major	7	3	11	4
Triple major	2	4	2	2
Thoracolumbar/ lumbar	6	8	3	3
Total	n = 20	n = 20	n = 21	n = 20

^aexcl. treatment, incl. physical education in school (according to questionnaire of Schmidt el. al 2016) ⁴⁹ .

Note: sig. differences (P < 0.05) between groups are highlighted in bold.

Drop-Outs, Adherence and Adverse Events

Average number of offered therapeutic sessions over 12 months were similar between BT (39.6 ± 4.4) and PT (39.5 ± 15.2). In both groups, one girl decided to terminate therapy after 8 months each. Those data result in drop-out rates of 1/20 and 1/21 during the intervention phase. Drop-outs were not excluded from intention to treat analysis.

In BT, adherence rate was 84.2% ± 13.3% including the mentioned drop-out. Reasons of absence were illness / injury sustained in school sports or leisure time / corona quarantine (n = 58); doctor’s appointment (n = 14); school matters (n = 15), not specified (n = 7). In PT, the drop-out and one other girl defaulted together 10 prescribed sessions leading to an adherence rate of 94.3% ± 20.0%. Absolute number of attended therapeutic sessions between radiological and clinical examinations differed not significantly between groups (Table S4).

No adverse events were detected during the intervention phase. Average number of unsupervised home-based exercise sessions per week can be found in Table 1.

Cobb-Angle

Table 2 shows mean differences and corresponding standard deviations of all parameters. Regarding the main analysis, Cobb-Angles decreased on average in both, BT and PT. Max Cobb improved significantly −1.9° ± 4.9° over all participants (main time effect): F(1; 38) = 5.77; P = 0.021; f = 0.39; 1-β = 0.65). The interaction effect (group-by-time) was not significant, ie, no type of intervention improved better than the other one, illustrated by parallel dashed lines in Figure 4A. Sum Cobb decreased −2.2° ± 10.4° considering all participants. No significant main or interaction effects were found, ie, the improvements found in the main curves (max Cobb) were not to the detriment of secondary curves expressed by sum Cobb. It should be noticed, that the high SD can be explained by one outlier in each group (Figure 4B). By excluding these outliers, main effect of sum Cobb (−3.7° ± 8.4°) would be significant (F(1; 36) = 6.96; P = 0.012; f = 0.44; 1-β = 0.73).

Table 2.

Mean Differences Before and After Intervention (Δ t2−t0) [95%-Confidence Intervals]

		Bouldering therapy n = 20	Physiotherapy n = 21
Main	Max cobb-angles in °	−1.9 [−4.2 to 0.5]	−1.9 [−4.2 to 0.4] (n = 20)
	Sum cobb-angles in °	−2.4 [−2.3 to 2.2]	−2.0 [−7.2 to 3.1] (n = 20)
	Max ATR in °	−1.6 [−2.3 to −0.9], P adj < 0.001, d = 0.81	−1.2 [−1.9 to −0.3], P adj < 0.001, d = 0.66
	Sum ATR in °	−3.5 [−4.8 to −2.12], P adj < 0.001, d = 1.03	−1.8 [−2.8 to −0.7], P adj = 0.003, d = 0.56
secondary	Pelvic obliquity difference in mm	0.3 [−0.8 to 1.5]	2.0 [0.1 to 0.9] (n = 20)
	FFD in cm	−1.0 [−3.6 to 1.6]	−2.4 [−4.8 to −0.1] (n = 19)
	Ott’s test in cm	−0.1 [−0.7 to 0.5]	−0.4 [−1.1 to 0.4] (n = 17)
	Schober’s test in cm	−0.3 [−0.7 to 0.1]	−0.3 [−0.6 to 0.0] (n = 18)
	Pain in spine of last 2 months (NAS 0-10)	−0.7 [−2.1 to 0.7]	−0.3 [−0.8 to 1.4] (n = 20)

Note: Sig. differences of Bonferroni corrected post-hoc test are highlighted in bold.

Figure 4.

(A) Max and (B) Sum of Cobb-Angles in Bouldering Therapy (BT Red) and Physiotherapy (PT Blue) Group. Individual Values Expressed by Dots, Mean Group Values by Triangles and Overall Means (With Outliers) by Crosses With Respective Dashed Trend Lines. P-values of main time effects are Presented. Bonferroni Corrected post-hoc Tests Revealed No Significance. *Cut-Off Z-Standardized Value > |2.68| (1.5-fold Interquartile Range).

Bonferroni corrected post-hoc tests revealed no significant improvements in Cobb-Angles within groups (Table 2. Figure 4).

Angle of Trunk Rotation

In regard to the main clinical parameter, the max and sum ATR decreased over all participants highly significantly with main time effects of F(1; 39) = 32.18; P < 0.001; f = 0.91; 1-β = 1.00 and F(1; 39) = 42.71; P < 0.001; f = 1.05; 1-β = 1.00, respectively. Max ATR decreased −1.6° ± 1.5° in BT and −1.2° ± 1.6° in PT.

Post-hoc tests with Bonferroni correction revealed significant improvements in both groups over time (Table 2, Figure 4). The interaction effect (group by-time) was not significant for max ATR but for sum ATR with F(1; 39) = 4.57; P = 0.039; f = 0.202; 1-β = 0.56, ie, BT improved better (−3.5° ± 2.8°) than the PT (−1.8° ± 2.33°). The results of the ATR are illustrated in Figure 5, depicting steeper dashed trend lines in BT.

Figure 5.

(A) Max and (B) Sum of Angle of Trunk Rotation (ATR) in Bouldering Therapy (BT Red) and Physiotherapy (PT Blue) Groups. Individual Values Expressed by Dots, Mean Group Values by Triangles and Overall Means by Crosses With Respective Dashed Trend Lines. P-values of main time effects and Bonferroni Corrected Ones per Group (P_adj) are Presented.

Pelvic Obliquity, Ott’s-, Schober’s Test, Fingertip-To-Floor Distance and Sensation of Pain

Mixed ANOVA of secondary parameters revealed no significant interaction effects but some clinically neglectable main effects were found: an increase (1.17 mm ± 3.46 mm) in pelvic obliquity (F (1;38) = 4.82; P < 0.034; f = 0.36; 1-β = 0.57), minimal decrease (−1.69 cm ± 5.22 cm) in FFD (F (1;37) = 4.16; P < 0.049; f = 0.34; 1-β = 0.51), minimal decrease (−0.32 cm ± 0.78 cm) in Schober’s test (F (1;36) = 6.14; P < 0.018; f = 0.41; 1-β = 0.67). Table 2 shows that post-hoc t-test results were not significant. There was no significant difference in pain reduction between groups.

Interest in Continuing Bouldering Therapy

After intervention phase, 85.7 % of patients treated by BT wanted to keep on. 15 of 20 actually participated in a follow-up program. Two could not participate due to time overlap with other appointments and one due to financial reasons. The drop-out and one more had no interest in continuing BT. In PT, all but 2 (drop out and one more) continued physiotherapy after study intervention period. A flow chart is given in Figure 6.

Figure 6.

Flow Chart Post Intervention.

Interestingly, 14 of 19 (1 drop-out) patients of the BT group ticked “no” in the form referring to the question if they are currently in therapy. In the PT group, everyone (19/19, 1 drop-out and 1 blanked questionnaire) answered “yes”.

Discussion

Climbing by itself can strengthen core muscles,^11,17 but in scoliotic spine, an asymmetric training of muscles is indicated. ⁴ The here presented RCT examined the effects of such a training by use of SSEs performed at a bouldering wall. This intervention was applied as sole treatment for mild AIS. It was compared to the common standard care in our region (Schroth) – the best evaluated SSE-intervention.^26-31 The supervised intervention period over 12 month was relatively long compared to other SSE-trials.

Cobb-Angle Changes in Comparison to Previous Studies

Regarding the main radiological outcome, max Cobb decreased about 1.9° (±5°) on average in BT as well as PT group. In addition to max Cobb changes, the influence of the intervention on other relevant curves is also important. This can be expressed sum Cobb, which showed similar reductions between groups (2.4° ± 10° and 2.0° ± 11°, respectively).

The result of max Cobb is in line with the finding of a recent meta-analysis including 7 RCTs with SSE-interventions. Frequency weighted mean Cobb-Angles have improved by 2.0° ± 4.2°. ³¹ In contrast, control conditions (bracing only, unspecific physical exercises) have showed less improvement 0.1° ± 3.2°. ³¹ Stratified by SSEs in terms of Schroth (n = 4 RCTs), improvements have been found to be higher 3.9° ± 2.9° compared to the results of control group (PT) in the present study. ³¹ However, none of the included studies has used intervention periods >6 months.

In the following, studies which examined the influence of SSEs with long intervention periods ≥12 months as exclusive treatment are considered.^3,54-57 Nevertheless, those differ by other criteria, e.g., home-based or supervised programs, exercise volume, patient characteristics, analytical methods.

An elaborated RCT of Monticone et al (2014) has compared traditional spinal exercises (controls) to task-oriented and specific self-correction exercises (SSE). ³ Max Cobb has decreased 5.3° ± 0.6°. In contrast, controls have worsened 1.7° (±0.3°) with a significant interaction effect. Compared to the presented study, the weekly duration of SSEs has been considerably higher (100 min vs 20 – 30 min). Furthermore, the mean treatment duration until skeletal maturity has been about 42 months – more than 3 times longer. The resulting higher exercise volume might explain the higher improvement.

However, a higher exercise volume does not always lead to better results. Two prospective controlled studies found max Cobb changes of −0.33° and +1.7°, respectively,^54,57 by use of the Scientific Exercises Approach to Scoliosis (SEAS). Changes have not been statistically significant. Supervised exercise volumes were about 80 min per week over 12 months ⁵⁴ and 91.84 min ± 25.68 min over 25 months. ⁵⁷

Another prospective study by Karavidas et al. (2024) mainly used video-based Schroth exercises (5 times for 30 min over 12 to 56 months). ⁵⁵ Lumbar (−1.6°) but not thoracic (−0.5°) Cobb-Angles have improved significantly on average.

Significantly highest thoracic Cobb-Angle reduction of 8.25° was stated Otman and colleagues (2005) in their uncontrolled cohort study. ⁵⁶ Even if the study excluded non-compliant participants, it shows what could be possible with extremely high exercise volumes (4 hours per day on 5 days per week over 12 months) in combination with high compliance.

The cited studies reflect great differences in the main outcome. An optimal exercise dose-response relationship still needs to be explored in respect of practicability and compliance.

In summary, the presented 12-month BT program according to the PM revealed partly better absolute Cobb-Angle reductions compared to other SSE-interventions despite of a lower exercise volume. In direct comparison to PT group, average sum Cobb reduction was slightly greater (Δ = 0.36°) and max Cobb reduction was negligibly lower (Δ = −0.08°). The 95%-CIs (−3.1 to 3.3) of the letter were lower than the minimal important difference (see next section), giving evidence for equivalence of SSEs performed during BT and PT.

Clinical Relevance of Cobb-Angle Reductions

Rates of MIDs could give additional information of therapy success. For that, MID of max Cobb is mostly defined as ≥ 5°,^{2,39,54,55,58} a value which lies slightly over the 95% upper and lower limits of agreement. ⁴¹ In the presented study, 20% in each group improved ≥5° in max Cobb and only 5% deteriorated despite low volume of SSEs. Similar improvement rates have been found by the other SEE-studies mentioned above (23.5%, ⁵⁴ 23.9%, ⁵⁵ respectively). These studies have shown higher deterioration rates with 11.8%, ⁵⁴ and 12.9%. ⁵⁵ A re-analyzation of the study of Monticone et al (2014), ³ which had used a cut-off value of 3° originally, has revealed a deterioration ≥5° in 7% of patients treated with SSE and 38% in untreated control group. ² Zapata et al (2019) have found even higher rates of deterioration (50%) in untreated or braced mild AIS (Risser 0) after 12 months. ⁵⁸

The high progression rates in untreated patients demonstrate the need of an effective therapy. For that, success can also be defined as curve stabilization (no clinically meaningful curve progression). Instead of a progression threshold (MID), rates of braces prescription can also be used as an outcome in mild AIS. ⁵⁷ Negrini et al. (2019) have found rates of 33%, 46% and 50% in SEAS, usual physiotherapy and controls, respectively. ⁵⁷ In the presented study, only one of 20 patients in the BT group (statistical outlier) had to be braced after study intervention period. The outlier in PT group matured to Risser 5 at t2. Thus, bracing was not indicated.

However, progression rates highly depend, i.a., on the initial curve magnitude, which complicates data interpretation of rates. The 2 outliers showed initial Cobb-Angles (>20°) and might explain the increase due to a higher risk of progression. However, the only other girl (BT group) with an initial Cobb-Angle of >20° improved by 8.9°. In contrast, she had a C-shaped thoracolumbar scoliosis. Thus, the type of scoliosis could be an effect modifying factor. The 2 outliers were under the 4 patients with an initial triple major scoliosis. According to descriptive data, this type was the only one associated with an average increase in max and sum Cobb. One patient with this form of scoliosis treated by Schroth deteriorated and the other one improved Cobb-Angles. Schroth explicitly address triple major scolioses, even though the 3D-autocorrection is much more complex than in other types. In contrast, SSEs of the bouldering program have not addressed this form explicitly, yet. It is suggested to adapt bouldering exercises for triple major scolioses.

Progression rates also depend on skeletal maturity. ⁵⁹ Until now, no ideal maturity index is available. ⁵⁹ The widely used Risser sign reflects a basic index, but is not suggested as a major criterion for guiding therapy. ⁵⁹ However, using a further invasive radiological assessment like Sanders Maturity Scale was rejected for ethical reasons (further exposure to radiation). We have included immature subjects with a mean Risser sign score of <2 in both groups. Interestingly, the girl with the highest Risser sign of 4.5 at t0 had a considerably decrease in max Cobb of 8.8°. In the largest subgroup of our sample (Risser 0, n = 15), max. Cobb improved 2.8° ± 6.0° on average (range: −14.4° to 4.50°). Thus, none of the curves in the most immature participants at t0 worsened to a clinically meaningful extent, despite of the highest risk of progression. ⁵⁹

Changes of Angle of Trunk Rotation in Comparison to Previous Studies

The main clinical parameters – max and sum ATRs – improved significantly in both groups. The reductions in the BT group were greater (0.4° and 1.7°, respectively) than in the PT group with statistical significance for sum ATR. However, this result needs to be interpreted with caution because the attending physician could not be fully blinded at t2. Even if data have been collected with the necessary medical care, bias could not be ruled out. Furthermore, sum ATR was significantly higher in BT group at start compared to PT group. This examination was fully blinded because group randomization followed afterwards.

On individual level, 15% in BT group and 5% in PT group reduced the max ATR to a clinically relevant extent of ≥3.74° (MID). ⁴⁶ Nobody increased in max ATR more than 1°. Stop of potential regression could also be seen as therapy success. Interestingly, the mentioned outlier in BT group which progressed meaningfully in Cobb-Angles, showed reductions in max and sum ATRs (2° and 4°, respectively).

The outstanding study of Monticone et al (2014) has found an improvement in max ATR (3.5° ± 0.2°) with significant main time, and group-by-time effects. ³ As mentioned above, exercise volume was extremely high. The study of Karavidas et al (2024) revealed significant ATR reductions at thoracic (1.5°) and lumbar (1°) level by using a daily exercise program. ⁵⁵ The research group around Negrini found non-significant changes in max and sum ATR in SSE and control groups.^54,57 In comparison to other long-term conducted SSEs, the presented bouldering program shows a promising approach to reduce or at least stabilize ATR. To eliminate the probable bias of data collection, a 3D-analyzation of the EOS-images are planned in the next step. ⁶⁰

Acceptance of Scoliosis-Specific Bouldering

According to the interest of 45/51 patients to take part in the study as well as the actual recruitment rate of 84%, BT seem to be an attractive treatment option for adolescents. The recruitment rate was somewhat higher compared to the pilot study (76%). ¹⁹ Furthermore, drop-out rate was lower (5% vs 11%) despite of a much longer intervention period of 12 months. Other studies with an equal study duration reported higher drop-out rates of about 26% (Schroth), ⁵⁶ 14% (SEAS), ⁵⁴ and 8% (usual physiotherapy). ⁵⁴

The adherence rate of 84% to BT sessions is interpreted as high for a group setting with a fixed weekly appointment. Even though rates of PT with one-on-one sessions were naturally higher, both treatments showed similar effects.

After 12 months of BT, 86% wanted to keep on and 75% actually did. Most (74%) participants even did not recognize BT as a kind of treatment. These results emphasize the high acceptance of the innovative therapy program.

Another conspicuous result is that physical activity (excl. treatment) increased significantly from t0 to t2 in PT group. This result might be explained by a resumption of sports activities after COVID-lockdown. In contrast, the activity in BT group did not increase beyond the intervention program. A speculative interpretation might be the sportive character of the therapy within a group setting, which did not necessarily require engagement in another sports.

Strength and Limitations

Therapeutic Climbing was used in the treatment of scoliosis of children and adolescents at least in the 1990s, ⁶¹ and examined scientifically in the early 2000s.^17,62 To our best knowledge, the presented study is the first which examined the long-time effects of a scoliosis-specific BT program on curvature in a randomized-controlled design.

According to the research design, internal validity is limited in favor of an enhanced external validity. This eases generalizability and practical transfer of study results. It is reasonable to train therapists in BT and the specific exercise protocol. However, equipment for a BT is less common than for standard PT. This limits its practicability to institutions with appropriate resources.

As argued by Ma et al 2024, ¹ RCTs examining SSE for mild to moderate AIS patients compared to no treatment are very difficult to conduct due to ethical reasons. Standard care (PT according to the Schroth-Method) was chosen as control condition. Hence, the effects of the BT program compared to no treatment is still unclear. This could weaken future secondary analyses especially because of the effectiveness of Schroth.^26-31

As mentioned in the method section, in one case, Schroth-therapist was not available and in 2 other cases a mixture of Schroth and other SSEs was performed. By excluding these 3 participants, results would not change except of sum ATR. The found interaction effect (group-by time) would not be significant any more (P = 0.071). It was already mentioned above to interpret the found superiority of BT in regard to sum ATR with care.

Another limitation is the incomplete blinding of the rater during clinical assessments at t2 as discussed above. In contrast, a very conservative approach was used to determine group-blinded Cobb-Angles by using the highest values out of 2 provided by 2 surgeons. However, bias during radiographic data collection could not be ruled out completely. In some patients arm positions differed during t0 and t2. This can lead to changes in radiographic markers, but should not have affected Cobb-Angles significantly. ⁶³ Additionally, markers are very difficult to set in curves <10°. Such angles had to be determined to enable a comparison between time points.

The found main effects sizes (time, independent of the group) with f = 0.39, 0.44, 0.91 and 1.05 for max Cobb, sum Cobb (with excluded outliers), max ATR and sum ATR, respectively, could be interpreted as high. ⁵³ Separated by group, improvements did only reveal Bonferroni corrected statistical significance in ATR, but not Cobb-Angles. The only interaction effect (sum ATR) was low in size (f = 0.2). On group level, the found improvements were below measurement error for Cobb-Angles (<3°) but above measurement error for ATR (>1.35°).

Conclusion

According to the results on group as well as individual level, the 12-month bouldering therapy including SSE is at least as effective as other SSEs like Schroth (control group) in mild AIS. Even a low volume of 20 – 30 min SSEs + additional unsupervised home exercises per week stabilize curvature – corresponding to the main therapy objective in mild scoliosis. Thus, scoliosis-specific bouldering could be used in addition, as a variation or as an alternative to standard physiotherapy (Schroth) according to the preference of the patient and availability.

The high acceptance, adherence and interest for a long-term exercising ease the practical transfer of the study protocol. From a psychological and pedagogical perspective, adolescents could have the option to go to sports with peers instead to therapy.

In future, the program could be evaluated in addition to bracing in more severe forms of AIS and adapted to scoliosis therapy of children.

ORCID iDs

Silas Dech https://orcid.org/0000-0002-8914-8276

Tom Hellriegel https://orcid.org/0009-0008-8003-9242

René Kittel https://orcid.org/0009-0003-6492-0519

Ethical Considerations

The study was conducted according to the guidelines of the Declaration of Helsinki and approved by the Ethics Committee of the University of Potsdam, Germany (protocol code 51/2018; 11.03.2019).

Consent to Participate

All participants and parents gave their written informed consent to participate.

Consent for Publication

All participants and parents gave their written informed consent for publication of anonymized data.

Data Availability Statement

The data presented in this study are available in the main article and supplementary materials.*