1 Introduction

The clinical and radiological outcomes in scoliosis surgery are assessed on coronal curve correction, rib hump reduction, overall truncal balance, coronal balance, sagittal balance and shoulder balance (SB).¹ The cosmetic implications of shoulder imbalance (SI) are significant and is often a major determinant to assess the final outcome following scoliosis surgery. SI may affect patient satisfaction and functional self-image assessment scores.^1–3

Factors postulated to be associated with post-operative SI include structural proximal thoracic curves, selection of upper instrumented vertebra (UIV), magnitude and correction of the middle and distal curves.^4–7 Literature has reported on SI primarily with regards to adolescent idiopathic scoliosis with few papers assessing SB in other aetiologies such as congenital, neuromuscular and syndromic scoliosis.^1,5,6,8 SI assessment in idiopathic scoliosis has largely focused on type 2 Lenke idiopathic curves.^4,7,9,10 Lenke type 2 curves have a proximal thoracic curve which is structural and can present with postoperative SI especially if the magnitude and flexibility of the proximal curve are underestimated.^5,6

Numerous radiological criteria have been used to assess SI which includes coracoid height difference, clavicle angle, T1 tilt, soft tissue shoulder height difference, clavicle chest cage angle difference and clavicle rib intersection difference.^1,11–13 However, whether these radiological indices assessing shoulder symmetry correlate with the patient functional outcomes score for self-image have not been studied adequately.¹⁴

We conducted a retrospective study to assess clinical and radiological factors associated with SI in all aetiologies of scoliosis including idiopathic, congenital, neuromuscular and syndromic scoliosis. The impact of radiological shoulder symmetry was assessed for patients with SI and SB to compare the functional self-image outcomes scores between the two groups.

2 Methods

The medical record database of the institute was searched for all patients undergoing spinal scoliosis deformity correction and fusion surgery during the period 2016–2018. Exclusion criteria was any patient undergoing revision deformity surgery and growing rod procedures. A retrospective analysis was performed for 43 patients with a minimum of 2 years available follow-up. The study was approved by the ethics committee review board (IRB approval no.052) and all applicable institutional and governmental regulations concerning the ethical use of human volunteers and informed consent were followed during the course of this research.

Hospital medical records, radiographs, and clinical photographs of the patients were assessed. SI was measured on the whole spine standing radiograph using the coracoid height difference. A line was drawn to join the upper margin of each coracoid process and the distance to a horizontal was measured as illustrated in Fig. 1. Radiological SI was considered to be present if a difference of 2 cm was noted. Measurements were performed on the preoperative, 12-week (early) postoperative and 2-year (late) follow up radiographs.

Close

Fig. 1 (a)The standing whole spine radiograph, the coracoid process is identified on both sides and a line is drawn connecting the two points. Another horizontal line is drawn and the vertical height distance between the coracoid process and horizontal line was measured. (b) the clinical photo of the same patient.

Export to PPT

Patients demographics were noted and relevant imaging was reviewed from PACS (Imagebyte software version 4.9.3.23.2). Imaging and records were reviewed to identify the aetiology of scoliosis including congenital, idiopathic, neuromuscular and syndromic scoliosis. Patient radiographs were assessed to identify upper and lower end vertebrae (UEV and LEV) of primary and secondary curves. Postoperative radiographs were reviewed to identify the upper and lower instrumented vertebrae (UIV and LIV). Mean preoperative Cobb angle, mean postoperative corrected Cobb angle and an average correction was recorded. Adolescent idiopathic scoliosis was classified using Lenke classification¹⁵ and proximal, middle and distal curves were recorded. All radiographic measurements were assessed by two authors (SS) and (PB) and independently validated by the senior author (AS).

Functional outcomes were assessed based on the SRS-30 questionnaire. Preoperative and postoperative 2-year SRS-30 scores were assessed and subgroup analysis for outcomes for patients with SB and SI was performed. An unpaired ‘t’ test was used for statistical analysis. All numerical data were reported with mean ± standard deviation and statistical significance were noted at p < 0.05.

3 Results

The retrospective review of 43 patients undergoing surgical deformity correction treatment documented 19 males and 24 females. The mean age was 14.5 ± 4.7 years. Among the 43 patients, there were 16 adolescent idiopathic scoliosis (Lenke 1 N = 4, Lenke 2 N = 0 Lenke 3 N = 7, Lenke 4 N = 2, Lenke 5 N = 0 Lenke 6 N = 3), 20 congenital scoliosis, 1 neuromuscular (Cerebral palsy) and 6 syndromic scoliosis. Mean preoperative Cobb angle for the primary curve was 70.88° ±25.66°, the mean postoperative Cobb angle was 26.11° ±18.85° with a mean cobb angle correction of 45° ±21.73°.

We documented preoperative SI in 34 patients including 17 congenital scoliosis, 13 adolescent idiopathic scoliosis, 3 syndromic and 1 neuromuscular scoliosis patient. There were 9 patients where the preoperative radiograph showed well-balanced shoulder symmetry and all 9 patients continued to have good SB in the early and 2 year follow up radiographs. The adolescent idiopathic scoliosis with preoperative SI showed the curve types as follows Lenke 1 (N = 4), Lenke 3 (N = 5), Lenke 4 (N = 2) and Lenke 6 (N = 2).

Postoperatively 12-week radiographs showed that 19 out of 34 patients with preoperative SI had improved after surgery and had well balanced shoulder symmetry, which was well maintained at the final follow up. There were 15 patients where SI was present at postoperative 12 weeks radiographs. At 2-year follow up spontaneous correction and improvement in the SI was noted in 8 out of the 15 patients leaving 7 patients with persistent SI at the final follow up. The demographic, clinical and radiological details of the patients with postoperative SI at 12-week radiographs are highlighted in Table 1.

3.2

3.2 Idiopathic scoliosis with shoulder imbalance

The shoulder balance outcomes for patients with idiopathic scoliosis classified as per Lenke classification are listed in Table 3. There were five patients with adolescent idiopathic scoliosis that had a SI at 12 weeks and 2 patients had residual SI at 2 year follow up. SI was seen in one patient with a Lenke type 1 curve (Fig. 2) and one Lenke type 3 curve. In both patients, a right-sided main thoracic curve with a preoperative high right shoulder was seen. Postoperatively both patients showed a contralateral left shoulder being higher.

Table 3 Distribution of curve patterns as per Lenke classification in idiopathic scoliosis patients showing preoperative, 12 week and at 2 year follow up for shoulder balance.

Sr No	Lenke curve type	Proximal thoracic curve	Middle Curve	Distal curve	Preoperative SI	12 weeks shoulder symmetry	Final follow up shoulder symmetry
1	Lenke 1	8	65	25	Right shoulder elevated	Balanced	Balanced
2	Lenke 3	23	70	35	Right shoulder elevated	Balanced	Balanced
3	Lenke 4	58	110	45	Right shoulder elevated	Balanced	Balanced
4	Lenke 4	30	110	70	Right shoulder elevated	Right shoulder higher	Balanced
5	Lenke 1	20	50	30	Balanced	Balanced	Balanced
6	Lenke 6	16	60	73	Right shoulder elevated	Balanced	Balanced
7	Lenke 3	16	52	48	Right shoulder elevated	Balanced	Balanced
8	Lenke 3	12	100	43	Right shoulder elevated	Right shoulder elevated	Balanced
9	Lenke 3	22	95	48	Right shoulder elevated	Balanced	Balanced
10	Lenke 3	21	60	38	Balanced	Balanced	Balanced
11	Lenke 6	18	67	95	Right shoulder elevated	Balanced	Balanced
12	Lenke 6	15	55	65	Right shoulder elevated	Balanced	Balanced
13	Lenke 1	10	65	30	Right shoulder elevated	Left shoulder elevated	Balanced
14	Lenke 3	20	60	40	Right shoulder elevated	Balanced	Balanced
15	Lenke 3	20	90	40	Right shoulder elevated	Left shoulder elevated	Left shoulder elevated
16	Lenke 1	30	80	36	Right shoulder elevated	Left shoulder elevated	Left shoulder elevated

Close

Fig. 2 (a) Adolescent idiopathic scoliosis Lenke 1 curve with clinical photo shows the right sided curve. (b) Standing radiographs show a right shoulder at a higher level preoperatively (c) Final follow up radiograph at 30 months showed the left shoulder at a higher level. UEV/LEV were T5/L1 respectively while UIV/LIV were T3/L3 respectively.

Export to PPT

Spontaneous improvement in SI was seen in some patients as illustrated in Fig. 3 where the patient with Lenke 1 type curve showed satisfactory shoulder balance at 2 years however, there was evidence of SI at the early 12 weeks follow up. There was one patient that had a syndromic large thoracolumbar curve (UEV = T8 to Pelvis), the apex at L2 with a pelvic obliquity and right shoulder at a higher level. She was treated with a T4 to S1 instrumented fusion however postoperatively she developed an opposite left shoulder asymmetry due to a stiff proximal thoracic curve T1- T8 of 46° Fig. 4.

Close

Fig. 3 (a) Whole spine standing radiograph show a right shoulder at a higher level with a Lenke 1 type curve preoperatively (b) clinical photos shows an elevated right shoulder (c) postoperative radiographs at 12 weeks shows a contralateral left elevated shoulder (d) final follow up radiographs shows spontaneous correction with good shoulder balance.

Export to PPT

Close

Fig. 4 (a) Clinical photograph of a syndromic scoliosis with large thoracolumbar curve, pelvic obliquity and elevated right shoulder. (b, c) whole spine lateral and AP films (d) bending films showed good flexibility (e) standing radiograph at last follow up shows an elevated left shoulder.

Export to PPT

4 Discussion

SB is an important concern in the clinical and radiological outcome following scoliosis surgery.^1–3 SI is as a cosmetic concern and can affect the patient's perception towards the final clinical outcome of the surgery.^2,3 However, there are few publications that have analysed shoulder imbalance in congenital scoliosis and a correlation of patient reported outcomes scores to shoulder balance following surgery has not been reported widely in literature.

The limitations of the presented study include the retrospective nature of data analysis which is susceptible to bias and small sample size of the series. The study sample includes different aetiologies for scoliosis making the groups heterogenous. Radiological assessment was performed using coracoid height difference which primarily evaluates lateral shoulder symmetry and medial trapezius prominence can be underestimated.

4.3

4.3 Correlation of clinical cosmetic appearance and radiological shoulder asymmetry

Clinical appearance of the shoulder and radiological SB has been studied by authors to objectively assess the correlation between the entities.^14,16,17 Qui et al. studied double thoracic idiopathic curves and suggested that there is a discrepancy between the radiological parameters used to assess SI and clinical cosmetic appearance of the shoulder.¹⁸ The authors correlated several radiological parameters including coracoid process height, T1 tilt, first rib angle, clavicle angle, clavicle-rib cage intersection with the cosmetic appearance of the shoulder and concluded that these parameters do not completely and accurately reflect the cosmetic shoulder appearance.¹⁸

Sharma et al. studied the correlation between radiological indices used to assess shoulder and trunk balance and objective cosmetic indices following scoliosis surgery in the clinical appearance.¹⁴ The authors noted that radiographic indices for SB did not reliably correlate with the final cosmetic appearance except for the radiological parameter of trunk shift. Our study documented that although the radiological criteria suggested SI; the patient-reported outcomes for self-image did not show significant differences in the preoperative or postoperative assessment at 2 years follow up.

Review of work from Sharma et al. which analysed radiological and cosmetic improvement correlation with patient self-reported outcomes indicated that the correlation was good for the apical vertebra translation correction and posterior trunk symmetry index.¹⁹ However, the correlation was poor for cobb angle correction and overall coronal balance.

The clinical appearance of SI has been reported to be subdivided into medial trapezial and lateral clavicular imbalance by Ono et al. and the authors suggested T1 tilt correlated with medial trapezial prominence more reliably than clavicle angle to lateral clavicular asymmetry.²⁰ Menon et al. reviewed 69 Lenke 1 and 2 type curves and reported that the inter-coracoid process line and T1 tilt could have a concordant or discordant relationship and the SB may be affected by both the main thoracic and proximal thoracic curves.¹⁷Fig. 5 depicts a patient in our series where the coracoid process height difference suggested good radiographic SB however, the medial trapezial prominence was a significant cosmetic concern and was accompanied by a significant T1 tilt. In summary, radiological SB assessed by coracoid height difference does not clearly correlate well with cosmetic appearance and therefore may not correlate with the patient-reported outcome scores which were supported by the findings of this study.

Close

Fig. 5 20-year-old boy with a right proximal thoracic spine scoliosis deformity (a) whole spine standing radiographs show a higher placed right shoulder. Aetiology of deformity was a congenital proximal thoracic curve due to a segmentation defect. (b) postoperative radiograph shows the coracoid process level and good lateral shoulder balance. (c) the medial trapezial prominence is significant which is concordant with a significant T1 tilt.

Export to PPT

5 Conclusions

SI was seen in patients with congenital scoliosis however, review of literature showed few papers evaluating SI in the setting of congenital scoliosis. Segmentation and mixed anomalies of the proximal thoracic spine were associated with residual SI at the final follow up. Radiological SI based on coracoid height was not associated with significant differences in the patient-reported outcomes for self-image scores both in the preoperative and 2-year postoperative assessment.

Author roles

Dr Shailesh Hadgaonkar: Data curation, Funding acquisition, Formal analysis, data acquisition, analysis of data. Dr Shubham Shah: Data curation, Funding acquisition, Formal analysis, data acquisition, analysis of data. Dr Pramod Bhilare: Data curation, Funding acquisition, Formal analysis, Writing - original draft, data acquisition, analysis of data, drafting of the manuscript, critical revision. Dr Ajay Kothari: administrative support. Dr Ashok Shyam: critical revision, Supervision. Dr Parag Sancheti: Supervision. Dr Siddharth N Aiyer: Supervision, Conceptualization, Data curation, Funding acquisition, Formal analysis, Writing - original draft, conception and design, data acquisition, analysis of data, drafting of the manuscript

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.