Preoperative Fulcrum Flexibility >80% Is Associated With Clinical Success in Vertebral Body Tethering

Abstract

Study design

Single-centre prospective study.

Objectives

Our study aims to investigate the relationship between preoperative and intraoperative factors with VBT surgical outcomes.

Methods

This is a single-centre prospective study involving consecutive VBT patients. All patients included were followed up for >2 years postoperatively. Primary outcomes were changes in radiographic parameters from pre-operative to immediate post-operative and between subsequent follow-ups. Secondary outcomes were the relationships between preoperative and intraoperative factors with coronal correction rate. Preoperative factors included fulcrum flexibility and skeletal maturity. Intraoperative factors were intraoperative tensioning, screw positioning, screw length and immediate coronal correction. Presence of growth modulation was defined as >5° further curve correction postoperatively, while postoperative clinical success was defined as <30° coronal curve magnitude and avoidance of revision surgery.

Results

Forty-one patients (35 females, 6 males) with 53 curves were included. Preoperative curve magnitude was 49.0 ± 7.4° and improved to 17.4 ± 8.0° immediately post-op (P < 0.001), with mean correction rate at 64.6 ± 15.4%. Coronal correction rate was persistently maintained at 62.1 ± 29.0% at 2-year. Preoperative flexibility correlated with correction rate at immediate post-operatively up to 18-month (correlation coefficient (r) = 0.768 to 0.933; P = 0.002 to P = 0.044). Strength of intraoperative tensioning at upper and apical curve segments was negatively associated with correction rate up till 9-month (r = −0.409 to −0.620) but could be explained by decision to tension more in less flexible curves. Preoperative fulcrum flexibility >79.8% (i.e. 80%) predicted postoperative clinical success (AUC 0.760; P < 0.001).

Conclusion

VBT improved curve magnitude and shoulder imbalance in scoliosis with progressive spinal growth. Flexible curves were associated with persistent coronal correction, while flexibility above 80% increases chance of clinical success with minimal risk of further progression. Growth modulation was associated with skeletal immaturity.

Keywords

vertebral body tethering fulcrum flexibility growth modulation idiopathic scoliosis skeletal maturity

Introduction

Vertebral body tethering (VBT) is an alternative to traditional spinal fusion in treatment for idiopathic scoliosis.¹ Current interest in VBT lies in its fusionless nature and ability to preserve motion and spinal growth.^2,3 In VBT, a flexible tether is instrumented on the convex side of the curve. Compression effect by tensioning helps achieve immediate correction while VBT is expected to allow growth modulation by the Heuter-Volkmann principle.¹ Successful coronal correction and sagittal improvement were documented in recent studies,^4-7 with favourable short-to-medium term results.

However, the behavior of the tether is unpredictable and there is uncertainty on whether growth modulation occurs.^2,6,7 Further, surgeons have different techniques and preferences, such as tensioning methods and screw positioning.^2,4 A recent peliminary study have reported preoperative fulcrum flexibility as a predictor for coronal correction,⁸ but only its relationship with the first postoperative X-ray was assessed without investigating its role in subsequent correction and sagittal plane outcomes. While curve magnitude was also reported as a determinant for postoperative correction in the same study, other intraoperative factors like segmental tensioning or positioning of vertebral screws were not accounted for. There is currently no consensus on how surgeons should modify these factors to achieve the best surgical outcomes. Hence, it is of utmost importance to address this knowledge gap and increase likelihood of favorable surgical outcomes.

Our study aims to investigate radiographic changes in VBT in terms of coronal correction, sagittal changes, spinal growth as well as assessing for any presence of growth modulation (GM). Secondary aims include (1) assessing the relationship between preoperative and intraoperative factors with postoperative coronal correction. Factors of interest include fulcrum flexibility, intraoperative tensioning, screw positioning, screw length and skeletal maturity, (2) performing subgroup analysis for patients achieving clinical success as well as those with postoperative GM, (3) analyzing VBT-specific complications and their associated predictive factors and (4) postoperative functional level assessing for presence of back pain, return-to-sports and truncal imbalance.

Methods

Study Design

This is a single-centre prospective study involving consecutive VBT patients. All patients were followed up for >2-year postoperatively. Fulcrum-bending radiographs were routinely performed in our institution for assessment of preoperative curve flexibility with reference to method described by Cheung et al.⁹ Skeletal maturity is defined as Sanders stage 7b,¹⁰ distal radius and ulnar classification-radius (DRU-R) grade 10 or DRU-ulnar (DRU-U) grade 8,¹¹ or proximal femur maturity index (PFMI) grade 6 (or grade 5 with concomitant Risser stage 5).^12,13 Assessment of skeletal maturity with PFMI was only used when hand-wrist radiographs were not available. Postoperatively, they were followed up at 3-6 monthly intervals depending on clinical needs. Radiographic data was obtained from postoperative X-rays, while intraoperative details were retrieved directly from operative records. Details on patient mobility, presence of back pain at latest follow-up, range of motion (ROM) were collected from physiotherapy and occupational therapy records to assess for postoperative functional level. Ethical approval from the local institutional review board (IRB; Ref 23-620) was obtained before study commencement. Individual informed consent was not required.

Surgical details

All surgeries were performed by two orthopaedic surgeons of the same institution, and were done either by thoracoscopic or mini-open approach. Implant systems used were REFLECT^TM (Globus Medical, USA) and The Tether^TM (Zimvie/Highridge Medical, USA). Length of screw used was determined intraoperatively according to surgical needs (5.0 to 6.5 mm). By dividing the vertebral body into 3 equal parts on lateral view, the positioning of vertebral screw was determined as located in the anterior one-third, posterior one-third or at the centre third of vertebra. After screw insertion, the flexible tether was anchored and sequentially tightened. One common tensioning technique is to maximally tension at every level, while other surgeons may prefer only tensioning maximally at apex region, or applying minimum tensioning throughout, depending on individual patients and curve characteristics. Strength of tensioning was measured in newtons (N). Immediate intraoperative correction of coronal curve was achieved through direct compression by screw-tensioning.²

Study Outcomes

Primary outcomes were interval changes in radiographic parameters from pre-op to immediate post-op and subsequent follow-ups at 3-6 monthly. Parameters measured included coronal curve magnitude (Cobb angle), T1-T12 kyphosis for thoracic curves,⁴ L1-L5 lordosis¹⁴ for thoracolumbar/lumbar curves, shoulder height difference, pelvic obliquity and T1-S1 spinal height. Curve correction rate (%) and the fulcrum bending correction index (FBCI) were subsequently obtained.¹⁵ Clinical success was defined as coronal curve magnitude <30° and the avoidance of revision surgery^7,16, while GM was defined as further curve correction (>5 degree (°) improvement) compared to immediate post-op. Further correction of >5° was selected as cut-off for GM given the potential 5° measurement error¹⁷. Postoperative clinical success and GM were each assessed separately at postoperative 2-year. Shoulder height difference was obtained by first drawing vertical lines through the posterior axillary folds, then measuring the height difference between the intersection of horizontal lines on shoulders and the vertical lines.¹⁸ Pelvic obliquity was measured with the Osebold technique by first drawing a straight line between bilateral superior aspects of iliac crests and another line parallel to lower end of radiograph, the angle between the intersection of the two lines was obtained.¹⁹ Absolute values for shoulder height and pelvic obliquity were used to indicate severity of imbalance.

Secondary outcomes were relationships between preoperative and intraoperative factors with coronal correction rate. Preoperative factors included curve flexibility and skeletal maturity in terms of Sanders and DRU classification. Intraoperative factors were strength of tensioning at upper/apical/lower segments, tensioning technique (maximal tensioning or not), screw positioning, screw length and immediate coronal correction. Screw positioning was defined as screw location relative to centre of individual vertebra on lateral spine radiograph (anterior, middle and posterior 1/3 of vertebral body). Subgroup analysis of patients achieving postoperative clinical success and GM were performed. The incidences of VBT-specific complications, including tether breakage, overcorrection and revision procedures were recorded. Incidences of early perioperative complications were investigated in a recent study by our institution²⁰ and it would be interesting to report on mid-to-long term complications specific to the procedure. Lastly, functional level was assessed by functional status (return-to-sports), any presence of back pain at latest follow-up, aesthetic concerns (rib hump/loin hump) and any truncal imbalance.

Statistical Analysis

Patient demographics, preoperative characteristics and surgical details were presented. Data normality was tested. Comparisons between radiographic parameters at multiple time-points were done with the Friedman test, while the Mann-Whitney U test was utilised to compare between two groups of continuous variables. Subgroup analysis for thoracic and thoracolumbar/lumbar curves were carried out. For sagittal profiles, patients were further subdivided according to degree of kyphosis/lordosis. Normokyphosis was defined as T1-T12 kyphosis of 15° to 40°, and hyperkyphosis and hypokyphosis as >40° and <15° respectively.²¹ Normolordosis was defined as having L1-L5 lordosis of 10° to 40°, and hyperlordosis as >40°.²² Fulcrum flexibility (%) was assessed by: (Preoperative curve magnitude –preoperative fulcrum-bending curve magnitude)/(Preoperative curve magnitude) x 100%, while correction rate (%) was assessed by: (Preoperative curve magnitude – curve magnitude at follow-up)/(Preoperative curve magnitude) x 100%. FBCI (%) was assessed by: (correction rate/fulcrum flexibility) x 100%.

Correlations of parameters were tested at each individual time-point, at 3-6 month intervals, up till post-op 2-year. We recognize that both tethered curves in a double VBT may not be completely statistically independent and any clustering effect from patients with double VBTs were controlled by implementation of partial correlation analysis whenever feasible. In such circumstances where further controlling for single vs double VBTs is not feasible, conventional correlation tests like Pearson and Spearman correlation were utilized for analysis.

For significant factors associated with postoperative clinical success and GM (eg, curve flexibility), further receiver operating characteristic (ROC) analysis was performed. The area under the curve (AUC) was calculated and optimal cut-off values were determined with the highest sensitivity and specificity. Binary logistic regression with bootstrapping was done after controlling for single/double VBTs, in order to further investigate relationship between flexibility and clinical success, and between GM and flexibility respectively. Odds ratio (OR) and 95% confidence intervals (CI) were presented.

For complication anaylsis, any potential correlations between preoperative/intraoperative factors with tether breakage and overcorrection were tested. Tether survival was assessed by Kaplan-Meier survival analysis. Despite the number of relationships assessed, the authors decided to report unadjusted P-values to avoid missing potential significant relationships, owning to the exploratory nature of our study.²³ Statistical significance was set at P < 0.05. All statistics were generated with SPSS statistics (v29.0; IBM, USA).

Results

Patient Characteristics

A total of 41 patients (35 females, 6 males) with 53 curves (33 thoracic, 20 thoracolumbar/lumbar) were included, with mean follow-up of 40.6 ± 13.7 months (range: 24 to 68 months). All curves were skeletally mature at final follow-up. All curves completed >2 years follow-up all had scheduled follow-ups at the 2-year time point (Table 1). Twenty patients had major thoracic curves (Lenke 1), 3 had double thoracic (Lenke 2), 7 had double major (Lenke 3), 8 had major thoracolumbar/lumbar (Lenke 5) and 3 had Lenke 6 curves. In patients (n = 13) with more than one structural curve (Lenke type 2, 3, 4 and 6), 4 patients had single VBT with only the major thoracic curve tethered.

Table 1.

Patient Demographics and Operation Details

Parameters
Gender, N(%)
Males	6 (41.6)
Females	35 (85.4)
Mean age at surgery, years	12.1 ± 1.2
Mean follow-up, months	40.6 ± 13.7
Standing height, cm ± SD	153.5 ± 6.8
Torso height, cm ± SD	83.5 ± 13.9
Weight, kg ± SD	38.7 ± 5.8
Arm span, cm ± SD	154.0 ± 8.7
BMI, kg/m² ± SD	16.4 ± 2.0
Skeletal maturity
Risser stage	Stage 0: 56.1%
	Stage 1: 19.5%
	Stage 2: 9.8%
	Stage 3: 12.2%
Sanders stage	Stage 1: 2.9%
	Stage 2: 8.6%
	Stage 3: 48.6%
	Stage 4: 31.4%
	Stage 5: 2.9%
	Stage 6: 5.7%
DRU-R	Stage 6: 17.9%
	Stage 7: 25.0%
	Stage 8: 53.6%
DRU-U	Stage 4: 3.6%
	Stage 5: 28.6%
	Stage 6: 39.3%
	Stage 7: 25.0%
Preoperative radiographic parameters
Mean Cobb angle (°)	49.0±6.9
Mean Fulcrum-bending Cobb angle (°)	10.2 ± 7.3
Mean fulcrum flexibility (%)	79.5 ± 13.0
Mean shoulder height/shoulder imbalance, mm ± SD	15.3 ± 11.0
Mean thoracic kyphosis (°)	22.7 ± 14.6
Mean lumbar lordosis (°)	39.3 ± 10.3
Pelvic obliquity (°)	2.7 ± 2.6
Operation details
VBT type, N (%)
Single VBT	29 (70.7)
Double VBT	12 (29.3)
Instrumented vertebral segment, N (%)
Thoracic	33 (62.3)
Thoracolumbar/lumbar	20 (37.7)
Mean instrumented levels	7.0 ± 1.6

VBT, vertebral body tethering; SD, standard deviation; BMI, body mass index.

Trends of radiographic Parameters

Preoperative curve magnitude was 49.0 ± 6.9° and improved to 17.4 ± 8.0° immediately post-op, and was maintained at 21.1 ± 14.0° at 2-year (P < 0.001). Mean correction rate and FBCI at immediate post-op were 64.6 ± 15.4% and 82.7 ± 19.1% respectively. A mild but non-significant increase in kyphosis was seen in hypokyphotic curves (n = 8) from 5.6 ± 3.6° to 12.7 ± 10.3° immediate post-op (P = 0.244). In contrast, a slight decrease in lordosis was observed in both normolordotic and hyperlordotic thoracolumbar/lumbar curves (n = 20) but 2-year lordosis was similar to pre-op (P > 0.05). Imbalance in shoulder height improved from 15.3 ± 11.0 mm to 11.1 ± 7.7 mm (P = 0.062). However, minimal changes were seen with pelvic obliquity (P = 0.809). Progressive increase in spinal height was noted with most significant increases after index operation (365.3 ± 21.5 to 384.9 ± 20.8 mm), ultimately progressing to mean 398.8 ± 21.4 mm at 2-year (P = 0.045). (Table 2).

Table 2.

Changes in Radiological Parameters Across Follow-Ups

Radiographic parameters (±SD)	Pre-op (n = 53)	Immediately post-op (n = 53)	3-month (n = 46)	6-month (n = 45)	9-month (n = 28)	12-month (n = 37)	18-month (n = 30)	24-month (n = 53)	P-value
Cobb angle (°)	49.0 ± 6.9	17.4 ± 8.0	19.3 ± 8.7	19.1 ± 9.5	16.5 ± 11.6	18.5 ± 9.6	19.4 ± 14.1	19.5 ± 14.0	<0.001^a
Thoracic (n = 33)	50.2 ± 7.0	20.2 ± 7.8	22.9 ± 8.0	21.6 ± 9.7	18.7 ± 12.0	21.5 ± 9.8	19.5 ± 16.1	21.1 ± 15.9	<0.001^a
Thoracolumbar/lumbar (n = 20)	47.2 ± 6.5	12.7 ± 6.1	13.7 ± 6.7	15.0 ± 7.6	11.1 ± 8.8	14.2 ± 7.9	19.3 ± 7.5	16.9 ± 17.6	<0.001^a
Correction rate (%)	/	64.6 ± 15.4	60.9 ± 16.2	61.6 ± 18.7	67.0 ± 23.6	62.7 ± 18.9	62.4 ± 29.0	62.1 ± 29.1	0.021^a
Thoracic	/	59.7 ± 14.7	54.4 ± 15.0	57.2 ± 19.5	63.1 ± 25.5	57.5 ± 18.8	63.8 ± 32.3	58.7 ± 33.3	0.074
Thoracolumbar/lumbar	/	72.6 ± 13.1	70.8 ± 12.8	68.7 ± 15.3	76.2 ± 6.3	70.2 ± 17.0	58.6 ± 18.9	67.6 ± 34.2	0.665
FBCI	/	82.7 ± 19.1	77.2 ± 18.5	78.2 ± 22.6	85.9 ± 26.0	78.3 ± 22.5	81.0 ± 33.2	77.7 ± 33.7	0.021^a
Thoracic	/	79.5 ± 18.4	71.8 ± 17.2	74.6 ± 23.2	83.2 ± 29.5	77.2 ± 25.6	83.3 ± 35.7	75.5 ± 36.9	0.074
Thoracolumbar/lumbar	/	87.9 ± 9.6	85.5 ± 17.9	84.1 ± 20.8	92.4 ± 13.9	79.9 ± 17.8	75.0 ± 26.7	81.3 ± 37.0	0.933
T1-T12 kyphosis
Hypokyphosis (n = 8)	5.6 ± 3.6	12.7 ± 10.3	12.3 ± 4.2	11.2 ± 6.0	11.3 ± 3.9	13.8 ± 6.2	19.5 ± 9.2	15.3 ± 9.0	0.244
Normokyphosis (n = 23)	26.3 ± 5.7	33.2 ± 12.1	32.6 ± 8.3	30.3 ± 10.7	29.8 ± 9.3	32.8 ± 8.5	28.4 ± 9.0	29.0 ± 8.1	0.891
Hyperkyphosis (n = 2)	46.5 ± 6.4	41.5 ± 2.1	45.0 ± 2.8	46.0 ± 1.4	47.0 ± 2.8	50.5 ± 2.1	45.5 ± 9.1	37.0 ± 9.9	/
L1-L5 lordosis
Normolordosis (n = 8)	29.8 ± 5.4	23.6 ± 5.7	29.0 ± 6.1	30.7 ± 6.9	25.7 ± 0.6	29.7 ± 5.3	26.5 ± 5.8	29.0 ± 3.9	0.305
Hyperlordosis (n = 12)	47.4 ± 5.6	42.4 ± 10.0	48.6 ± 6.7	48.7 ± 8.6	47.5 ± 5.3	49.1 ± 7.7	47.8 ± 7.4	47.0 ± 9.2	0.127
Radiographic shoulder height (mm)	15.3 ± 11.0	11.1 ± 7.7	8.4 ± 5.8	9.4 ± 6.3	9.6 ± 6.3	9.2 ± 8.3	11.9 ± 8.7	9.1 ± 6.5	0.062
Pelvic obliquity (°)	2.7 ± 2.6	2.1 ± 2.0	2.4 ± 2.4	2.3 ± 1.9	3.2 ± 2.4	3.2 ± 3.3	1.9 ± 0.9	2.8 ± 1.7	0.809
Spinal height (mm)	365.3 ± 21.5	384.9 ± 20.8	379 ± 17.0	381.4 ± 17.4	385.4 ± 18.2	382.9 ± 18.6	386.7 ± 18.1	398.8 ± 21.4	0.045^a

^acomparisons between parameters at follow-up were done with the Friedman test (P < 0.05).

Relationship Between Factors and Coronal Correction

Coronal curve flexibility correlated significantly with correction rate from immediate post-op up to 18-month (correlation coefficient (r) = 0.768 to 0.933; P = 0.002 to P = 0.044), with stronger associations in thoracic curves up till 2-year (r = 0.367 to 0.718; P < 0.001 to P = 0.041). Strength of intraoperative tensioning at upper and apical curve segments was negatively associated with correction rate up till 9-month (r = −0.409 to −0.620). Interestingly, curve flexibility in our cohort was inversely correlated to upper segment tensioning (r = −0.424; P = 0.020). Hence, the observed correlation between tensioning and correction might be a result of surgeon’s preference to increase tensioning in less flexible curves. This implied that more flexible curves do not require aggressive tensioning to achieve desirable correction, while satisfactory correction in rigid curves might not be achievable even if surgeons decided to tension more intraoperatively. Method of tensioning at different segments did not contribute much to correction rate (P > 0.05). Placement of screws did not correlate with correction rate. More correction rate was anticipated at subsequent follow-ups when surgeons aimed for more intraoperative correction by direct compression (r = 0.824 to 0.940; P < 0.001). (Table 3 and table 4).

Table 3.

Relationship Between Preoperative Factors With Coronal Correction Rate

Correlation coefficient (r), [P-value]	Immediate correction (n = 53)	3-month (n = 46)	6-month (n = 45)	9-month (n = 28)	12-month (n = 37)	18-month (n = 30)	24-month (n = 53)
Preoperative factors
Curve flexibility
Fulcrum flexibility (coronal)^b	0.825 [0.022]^*	0.932 [0.002]^*	0.933 [0.002]^*	0.855 [0.014]^*	0.788 [0.035]^*	0.768 [0.044]^*	0.737 [0.059]
Thoracic (n = 33)^a	0.367 [0.036]^*	0.321 [0.096]	0.452 [0.016]^*	0.718 [<0.001]^*	0.283 [0.202]	0.449 [0.041]^*	0.555 [<0.001]^*
Thoracolumbar/lumbar (n = 20)^a	0.171 [0.470]	0.183 [0.468]	0.032 [0.903]	0.548 [0.160]	0.404 [0.136]	0.595 [0.120]	0.390 [0.089]
Skeletal maturity
Sanders stage^b	0.443 [0.445]	0.046 [0.941]	0.230 [0.710]	0.134 [0.829]	0.291 [0.635]	0.318 [0.602]	0.391 [0.515]
DRU-R^b	0.230 [0.710]	−0.015 [0.981]	0.221 [0.721]	0.143 [0.819]	0.262 [0.671]	0.275 [0.654]	0.550 [0.337]
DRU-U^b	0.993 [<0.001]^*	0.813 [0.094]	0.877 [0.051]	0.759 [0.137]	0.846 [0.071]	0.935 [0.020]^*	0.863 [0.060]

*statistical significance (P<0.05)

^aRelationships were assessed by Spearman correlation test for nominal variables and Pearman correlation test for test involving categorical variables

^brelationships assessed by partial correlation controlled for single versus double VBT.

Table 4.

Relationship Between Intraoperative Factors With Coronal Correction Rate

Correlation coefficient (r), [P-value]	Time-points
	Immediate correction (n = 53)	3-month (n = 46)	6-month (n = 45)	9-month (n = 28)	12-month (n = 37)	18-month (n = 30)	24-month (n = 53)
Tensioning strength
Upper segment tensioning^a	−0.312 [0.094]	−0.250 [0.228]	−0.455 [0.029]^*	−0.620 [0.008]^*	−0.197 [0.450]	−0.267 [0.378]	−0.376 [0.041]^*
Apical tensioning^a	−0.409 [0.028]^*	−0.461 [0.023]^*	−0.449 [0.036]^*	−0.446 [0.071]^*	−0.499 [0.041]^*	−0.095 [0.758]	−0.196 [0.309]
Lower segment tensioning^a	−0.254 [0.175]	−0.128 [0.540]	−0.364 [0.087]	−0.634 [0.006]^*	−0.413 [0.100]	−0.365 [0.219]	−0.296 [0.113]
Tensioning method
Maximal tensioning throughout all segments^a	−0.072 [0.688]	−0.199 [0.311]	−0.148 [0.462]	−0.125 [0.609]	−0.219 [0.353]	0.239 [0.373]	0.076 [0.668]
Screw position (anterior = 0/Central = 1/Posterior = 2)
Cranial screws^b	0.026 [0.956]	0.201 [0.666]	0.101 [0.829]	0.019 [0.968]	0.099 [0.833]	0.258 [0.576]	0.191 [0.681]
Apical screw^b	0.082 [0.862]	0.463 [0.295]	0.239 [0.607]	0.198 [0.671]	0.139 [0.766]	0.242 [0.601]	0.068 [0.885]
Caudal screws^b	0.371 [0.413]	−0.035 [0.941]	−0.041 [0.931]	0.157 [0.737]	0.301 [0.512]	0.373 [0.409]	0.205 [0.659]
Screw length^b
Immediate intraoperative correction^b	/	0.824 [0.023]^*	0.877 [0.009]^*	0.869 [0.011]^*	0.908 [0.005]^*	0.940 [0.002]^*	0.886 [0.008]^*

*statistical significance (P<0.05)

^aRelationships were assessed by Spearman correlation test for nominal variables and Pearman correlation test for test involving categorical variables

^brelationships assessed by partial correlation controlled for single vs double VBT.

Subgroup Analysis

Clinical success: Thirty-five curves (66.0%) had magnitude <30° and did not require revision surgery at 2-year postoperatively (clinical success). Patients achieving clinical success had higher preoperative fulcrum flexibility compared to those without (71.0% vs 83.9%; P < 0.001). Presence of clinical success was associated with higher pre-op flexibility (r = 0.494; P < 0.001). With ROC analysis (Figure 1), utilising pre-op flexibility to predict presence of clinical success with area-under-curve (AUC) of 0.760 (P < 0.001), yielding optimal cut-off of 79.8% (sensitivity 0.78, specificity 0.69). Binary logistic regression showed that higher preoperative fulcrum flexibility predicted postoperative clinical success (OR 1.097; 95% CI 1.045 to 1.213; P = 0.003) with significance level P < 0.001 after further bootstrapping (1000 samples). (Table 5).

Figure 1.

ROC curve of preoperative fulcrum flexibility predicting postoperative clinical success at 2-year

Table 5.

Binary Logistic Regression Using Preoperative Fulcrum Flexibility Predicting 2-Year Clinical Success

Model: Using preoperative fulcrum flexibility predicting 2-year clinical success
Results before bootstrapping
B	S.E	Wald	Exp (B)	P-value
0.093	0.031	9.054	1.097	0.003
Results after bootstrapping (1000 samples)
B	Bias	Standard error	95% confidence interval	P-value
0.093	0.010	0.037	0.044 to 0.193	<0.001

B, unstandardised Beta coefficient; S.E., standard error.

Growth modulation: Twelve curves (22.6%) exhibited GM, with 4 curves showing progressive curve improvement at subsequent follow-ups. Other curves generally exhibited pattern of initial curve progression over initial 3-6 months followed by progressive curve improvement. Higher fulcrum flexibility trended towards GM but not statistically significant (r = 0.270, P = 0.076). GM was associated with preoperative Risser stage ≤2 (r = 0.301; P = 0.047), and was also associated with postoperative clinical success (r = 0.343; P = 0.023).

VBT-specific Complications

Overall incidence for tether breakage was 32.1% (17/53; 9 thoracic and 8 thoracolumbar/lumbar curves). By Kaplan-Meier analysis, breakage-free survival at 12-month, 24-month and 36-month were 98.1% (95% confidence interval (CI) 94.4 to 100.0%), 84.5% (95% CI 74.7 to 94.3%) and 57.3% (95% CI 40.7 to 73.9%) respectively (Figure 2A). Thoracolumbar/lumbar curves were at greater risk of tether breakage (2-year survival 78.5% vs 87.9%; log-rank test P = 0.020) with hazard ratio of 2.69 (95% CI 1.00 to 7.22) (Figure 2B). Regarding curves with tether breakage (n = 17), there was a slight but statistically insignificant curve progression of 3.7 ± 3.4° (P = 0.197) and 6.8 ± 1.5° (P = 0.104) at 6-month and 12-month post-breakage respectively. Most tether breakages (11/17; 64.7%) occurred caudal to the curve apex (lower segments). 6 breakages (35.3%) occurred at the curve apex, while one lumbar curve had breakages both cranial and caudal to curve apex (both upper and lower segments). We did not find any significant correlations between fulcrum flexibility with incidence of tether breakage.

Figure 2.

2A, Graph showing tether breakage-free survival in VBT curves; 2B, Graph showing tether breakage-free survival of thoracic vs thoracolumbar/lumbar curves

Six patients (14.6%) required revision surgery at mean 29.2 ± 12.2 months (range: 24 to 54). Amongst them, two required conversion PSF due to primary thoracic curve deterioration and one other had PSF for T10-L3 deterioration. Another patient had upper untethered thoracic (T2-5) deterioration, ultimately requiring T2-L3 PSF. A total of 4 instrumented curves in 3 patients had overcorrection (excessive growth modulation from −3.4 to −36.8°), with 2 patients ultimately requiring removal of implants after 2-year.

Postoperative Functional Level

Only 2 patients complaining of mild back pain at final follow-up, with others all asymptomatic. Twenty-six patients (63.4%) had retrievable physiotherapy/occupational therapy assessments. Five patients were frequent sports players, with one returning to competitive sports (rugby). Ten patients has mild residual thoracic humps while only two had mild loin humps. Six patients had mild truncal shifts. None had problems with walking, balancing and ADL.

Discussion

Our study demonstrated that VBT improved curve magnitude and shoulder imbalance in scoliosis, while simultaneously allowing for progressive spinal growth. Flexible curves were associated with persistent coronal correction, while flexibility above 79.8% (i.e. 80%) increases chance of subsequent clinical success. GM occurred in 22% of patients and was associated with skeletal immaturity. Less flexible curves warranted more intraoperative tensioning but did not directly contribute to postoperative correction.

Majority of curve correction was achieved intraoperatively in VBT,^5,6,24,25 while a proportion of patients achieved further curve correction post-index operation (growth modulation). It was reported by Catanzano et al that progressive thoracic curve improvement occurred in half of the patients.²⁶ Most flexibility of the curve was taken up during intraoperative correction, as reflected from the immediately post-op FBCI (>80%).^8,9,15 Thoracic kyphosis was mildly increased in hypokyphotic patients. This echoed with findings from Baroncini et al, which also reported restoration of normokyphosis in preoperative hypohyphotic patients.⁴ Direct compression by instrumentation on anterior column might explain this kyphotic effect. The most significant advantage of VBT undoubtedly relied on its ability to preserve growth and range-of-motion. Progressive increase in spinal height was demonstrated throughout and previous literature also reported spinal growth was allowed to occur at an expected rate with VBT.²⁷ Overall imbalance was improved with significant decrease in postoperative shoulder height difference, and was reported to be comparable to PSF.²⁸ Previous studies documented a 30% decrease in shoulder height, implying the potential of spontaneous proximal thoracic curve correction after VBT.²⁹ Pelvic obliquity was more or less unchanged throughout our study, it should be noted most pelvic obliquity in a healthy AIS cohort results from leg-length discrepancy and was not corrected by scoliosis surgery.

Previous literature suggested that more flexible curves was associated with more curve correction in the first erect post-op X-ray.⁸ Our study further consolidated this relationship by demonstrating flexibility was also associated with subsequent correction rate up till 2-year. Consensus on intraoperative tensioning was not achieved but previous techniques described ideal tensioning should be able to obtain intraoperative correction <20° and parallel apical disks.² Generally, surgeons tend to increase segmental tensioning when more rigid curves are encountered. Compared to rigid curves, desirable correction in more flexible curves could be sufficiently obtained even with less tensioning per segment, possibly explaining the inverse relationship we observed between tensioning and correction. Hence, the observed relationships could merely be a reflection of common surgeon practice. Although our results did not favor a particular tensioning technique, tensioning more at the apex and less at peripherals was shown to decrease risk of screw pull-out in addition to optimizing correction in previous studies.³⁰ Hence, over-tensioning for more correction might not be necessary in more flexible curves, given the risks of more aggressive tensioning on increasing stress on screw-tether junction and potential for tether breakage.³¹ Previous studies have shown that peripheral screws positioned posterior relative to apex could help achieve cord-tensioning effect, allowing for correction and derotation4. In contrast, screw placement did not seem to impact correction in our study.

Our study reported strong correlation between higher preoperative fulcrum flexibility and achieving clinical success, which was not demonstrated previously. We hypothesize that flexible curves are more amendable to tensioning forces exerted on the concave side of the curve during growth, hence are more prone to resisting further curve deterioration. Further GM is enabled by asymmetrical spinal growth, associated with concave vertebral height gain as well as spinal growth.³² Differenital growth was shown to be more pronounced in instrumented major thoracic curves, while overall total spinal growth remained unchanged despite asymmetrical growth.²⁷ Bernard et al recommended smaller initial intraoperative correction in patients with Risser stages 0-2 and smaller preoperative curve magnitude, allowing for more growth modulation postoperatively.³³ GM was more frequently reported in skeletally immature patients,^17,34 which resonated with our findings. In patients with preoperative flexibility >80%, there is greater chance of postoperative clinical success, which should be conveyed during preoperative counselling. It should also be noted that clinical success is not guaranteed despite having higher rates of treatment success in patients with more flexible curves (AUC 0.760), these should be conveyed in preoperative counselling to adjust patient expectations.

Tether breakage occurred in 32.1% of our patients, and documented occurrence can be up to 50% at 36-month.³⁵ Although curve deterioration was insignificant post-tether breakage, it was reported that breakage during the first 12-month resulted in consistent loss of coronal correction.³⁶ Dominant failure mode in tether breakage occurs at the screw-tether junction and higher failure rates were associated with higher compression force at initial correction.³¹ However, our results failed to demonstrate the relationship between tensioning and incidence of breakage. Overcorrection only occurred in three patient, with recent studies suggesting lower BMI and skeletal immaturity to be critical predictors.³⁷

Our study provided insight on optimal surgical techniques like screw positioning and tensioning to aid surgeons in achieving improved outcomes. Strengths of our cohort included consistent 2-year postoperative follow-up and all patients attaining skeletal maturity at final follow-up. Study limitations included single-centre nature limiting size of our study cohort. With only 41 patients, secondary analysis might be underpowered (subgroup analysis for thoracic and lumbar curves). Moreover, the incompleteness of intraoperative documentation (only 46%) of tensioning variables limited further factor controlling during correlation analysis. This should be acknowledged and the effect of tensioning should be assessed in future studies. A multi-centre study would be required in the future to obtain desirable curve number for analysis, verifying relationships observed here. Potential advantages of our single-centre study included less variability in patient demographics and surgeon preferences. The inclusion of patient-reported outcome measures like Scoliosis Research Society-22 revised (SRS-22r) questionnaire would provide more comprehensive evaluation of treatment success, given that the aims of VBT surgery is preservation of mobility and function. Lastly, we were unable to accurately assess for spinal ROM as these were not systematically quantified in follow-up assessments. Some patients had private physiotherapy appointments and had no retrievable reports.

Conclusion

VBT allows for persistent curve correction, especially in flexible curves. Additional benefits of VBT included improvement in shoulder balance and preservation of progression spinal growth. Preoperative fulcrum flexibility >80% effectively predicted postoperative clinical success and these curves were unlikely to progress after index operation. Surgical success was highly correlated with presence of GM, which was associated with preoperative Risser stage ≤2.

Footnotes

ORCID iDs

Kai Chun Augustine Chan

Chi Chun Garvin Cheung

Kai Him Ambrose Chan

Long Ki James Lin

Elvis Kai Yeung Tse

Chi Wai Natalie Leung

Prudence Wing Hang Cheung

Jason Pui Yin Cheung

Ethical Considerations

Ethics approval was obtained from the institutional review board of the University of Hong Kong/Hospital Authority Hong Kong West Cluster (IRB reference number: UW23-620).

Author Contributions

CHAN Kai Chun Augustine: Methodology, data curation, formal analysis, investigation, writing-original draft, writing-review and editing.

CHEUNG Chi Chun Garvin: Data curation, investigation.

CHAN Kai Him Ambrose: Data curation, investigation.

LIN Long Ki James: Data curation, investigation.

TSE Elvis Kai Yeung: Data curation, investigation.

LEUNG Chi Wai Natalie: Data curation, investigation.

CHEUNG Prudence Wing Hang^#: Conceptualization, methodology, resources, data curation, project administration, supervision, writing-review and editing.

CHEUNG Jason Pui Yin^#: Conceptualization, methodology, resources, data curation, project administration, supervision, writing-review and editing.

CHEUNG PWH and CHEUNG JPY are co-corresponding authors of this manuscript. CHAN KCA is the first author.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: PWHC received funding for this work from the University of Hong Kong (HKU Seed Fund, #2302101719), JPYC received funding from the Li Shu Fan Medical Foundation Professorship in Orthopaedic Surgery.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

ICMJE COI statement

All authors have no conflicts of interest to disclose.

References

Raitio

Syvanen

Helenius

. Vertebral body tethering: indications, surgical technique, and a systematic review of published results. J Clin Med. 2022;11(9):2576.

Shah

Kraft

Miyanji

. Anterior vertebral body tethering: a review of the available evidence. J Am Acad Orthop Surg. 2024;32(6):247-256.

Wong

DLL

Mong

, et al. Can anterior vertebral body tethering provide superior range of motion outcomes compared to posterior spinal fusion in adolescent idiopathic scoliosis? A systematic review. Eur Spine J. 2023;32(9):3058-3071.

Baroncini

Courvoisier

Berjano

, et al. The effects of vertebral body tethering on sagittal parameters: evaluations from a 2-years follow-up. Eur Spine J. 2022;31(4):1060-1066.

Boulet

Hurry

Skaggs

, et al. Analysis of three-dimensional spine growth for vertebral body tethering patients at 2 and 5 years post operatively. Spine Deform. 2024;12(4):1009-1016.

Samdani

Ames

Kimball

, et al. Anterior vertebral body tethering for idiopathic scoliosis: two-year results. Spine (Phila Pa 1976). 2014;39(20):1688-1693.

Hoernschemeyer

Hawkins

Tweedy

Boeyer

. Anterior vertebral Body tethering: a single-center cohort with 4.3 to 7.4 years of Follow-up. J Bone Joint Surg Am. 2024;106(20):1857-1865.

Silk

Tishelman

Eaker

Lonner

. Reducing uncertainty in anterior vertebral body tethering: predicting postoperative curvature with fulcrum bending radiographs. Glob Spine J. 2024;14(1):219-224.

Cheung

Luk

. Prediction of correction of scoliosis with use of the fulcrum bending radiograph. J Bone Joint Surg Am. 1997;79(8):1144-1150.

10.

Cheung

PWH

Cheung

JPY

. Sanders stage 7b: using the appearance of the ulnar physis improves decision-making for brace weaning in patients with adolescent idiopathic scoliosis. Bone Jt J. 2021;103-B(1):141-147.

11.

Cheung

Samartzis

Cheung

Luk

. The use of the distal radius and ulna classification for the prediction of growth: peak growth spurt and growth cessation. Bone Jt J. 2016;98-B(12):1689-1696.

12.

Cheung

PWH

Canavese

Chan

CYW

, et al. The utility of a novel proximal femur maturity index for staging skeletal growth in patients with idiopathic scoliosis. J Bone Joint Surg Am. 2022;104(7):630-640.

13.

Cheung

PWH

Canavese

Wong

JSH

Luk

KDK

Cheung

JPY

. The use of proximal femur maturity index in guiding brace weaning for adolescent idiopathic scoliosis. JB JS Open Access. 2025;10(4):e25.00191.

14.

de Jonge

Dubousset

Illes

. Sagittal plane correction in idiopathic scoliosis. Spine (Phila Pa 1976). 2002;27(7):754-760.

15.

Luk

Cheung

Leong

. Assessment of scoliosis correction in relation to flexibility using the fulcrum bending correction index. Spine (Phila Pa 1976). 1998;23(21):2303-2307.

16.

Zhu

Qiu

Liu

Man-Chee Cheung

. Minimum 3-year experience with vertebral body tethering for treating scoliosis: a systematic review and single-arm meta-analysis. J Orthop Surg. 2022;30(3):10225536221137753.

17.

Hoernschemeyer

Hawkins

Tweedy

Boeyer

. Growth modulation increases clinical success in vertebral body tethering. Spine Deform. 2025;13(5):1421-1429.

18.

Nikouei

Ghandhari

Ameri

Mokarami

. Shoulder imbalance in adolescent idiopathic scoliosis: a systematic review of the current state of the art. Arch Bone Jt Surg. 2022;10(12):992-1003.

19.

Karkenny

Magee

Landrum

Anari

Spiegel

Baldwin

. The variability of pelvic obliquity measurements in patients with neuromuscular scoliosis. JB JS Open Access. 2021;6(1):e20.00143.

20.

Chan

KCA

Cheung

CCG

Tse

KYE

Cheung

JPY

Cheung

PWH

. The perioperative complications of vertebral body tethering: prevalence and predictive factors. Bone Jt J. 2025;107-B(7):736-743.

21.

Koele

Lems

Willems

. The clinical relevance of hyperkyphosis: a narrative review. Front Endocrinol. 2020;11:5.

22.

Shirazi

Haghighi

Alavi

Nezhad

Emami

. Is application of kinesio tape to treat hyperlordosis more effective on abdominal muscles or hamstrings? J Res Med Sci. 2018;23:9.

23.

Feise

. Do multiple outcome measures require p-value adjustment? BMC Med Res Methodol. 2002;2:8.

24.

Vatkar

Najjar

Patel

Quraishi

. Vertebral body tethering in adolescent idiopathic scoliosis with more than 2 years of follow-up- systematic review and meta-analysis. Eur Spine J. 2023;32(9):3047-3057.

25.

Newton

Parent

Miyanji

, et al. Anterior vertebral body tethering compared with posterior spinal fusion for major thoracic curves: a retrospective comparison by the harms study group. J Bone Joint Surg Am. 2022;104(24):2170-2177.

26.

Catanzano

Jr. Newton

Bastrom

, et al. Spontaneous lumbar curve correction following vertebral body tethering of main thoracic curves. J Bone Joint Surg Am. 2022;104(18):1629-1638.

27.

McDonald

Shah

Hargiss

, et al. When successful, anterior vertebral body tethering (VBT) induces differential segmental growth of vertebrae: an in vivo study of 51 patients and 764 vertebrae. Spine Deform. 2022;10(4):791-797.

28.

Meyers

Eaker

Samdani

, et al. Anterior vertebral body tethering shows clinically comparable shoulder balance outcomes to posterior spinal fusion. Spine Deform. 2024;12(4):1033-1042.

29.

Miyanji

Fields

Murphy

, et al. Shoulder balance in patients with lenke type 1 and 2 idiopathic scoliosis appears satisfactory at 2 years following anterior vertebral body tethering of the spine. Spine Deform. 2021;9(6):1591-1599.

30.

Jeandel

Bremond

Christine de Maximin

Lefevre

Courvoisier

. Thoracoscopic anterior vertebral body tethering in lenke Type-1 right adolescent idiopathic scoliosis. JBJS Essent Surg Tech. 2023;13(3):e22.00027.

31.

Guldeniz

Yip

CCH

Nafo

Cheung

KMC

. Biomechanics of the tether breakage: tensile behaviour of a single-unit vertebral body tethering construct. Spine Deform. 2023;11(4):825-831.

32.

Louer

Jr. Upasani

Hurry

, et al. Growth modulation response in vertebral body tethering depends primarily on magnitude of concave vertebral body growth. Spine Deform. 2024;12(6):1689-1698.

33.

Bernard

Bishop

Herzog

, et al. Dual modality of vertebral body tethering: anterior scoliosis correction versus growth modulation with mean follow-up of five years. Bone Jt Open. 2022;3(2):123-129.

34.

Treuheim

Eaker

Markowitz

Shankar

Meyers

Lonner

. Anterior vertebral body tethering for scoliosis patients with and without skeletal growth remaining: a retrospective review with minimum 2-Year Follow-Up. Int J Spine Surg. 2023;17(1):6-16.

35.

Cahill

Miyanji

Lullo

, et al. Incidence of tether breakage in anterior vertebral body tethering. J Pediatr Orthop. 2024;44(4):e323-e328.

36.

Baroncini

Migliorini

Eschweiler

Hildebrand

Trobisch

. The timing of tether breakage influences clinical results after VBT. Eur Spine J. 2022;31(9):2362-2367.

37.

Pahys

Hwang

McGarry

Quinonez

Grewal

Samdani

. Incidence and predictors of growth modulation and overcorrection after anterior vertebral body tethering. Spine (Phila Pa 1976) 2025;21(4):286-293.