2.1 Patient population

Our institutional review board approved the study protocol. This study is a retrospective study conducted at an academic teaching facility. Because this study is retrospective and non-invasive, informed consent was obtained through opt-out, and our institutional ethics committee waived the need to obtain informed consent from individual patients.

This is a retrospective study consisting of records of 46 AIS patients consecutively retrieved from our database from August 2011 to December 2020. The inclusion criteria for this study were (1) Lenke type 1 or 2, (2) posterior corrective surgery, and (3) follow-up at least 24 months after surgery. The exclusion criteria were Chiari malformation and other symptomatic scoliosis types. Twenty-eight patients who met these criteria were recruited for the study. Four were men and 24 were women, with a mean age of 14.5 years at the time of surgery. Patient parameters are provided in Table 1.

2.2 Surgical technique

All patients underwent posterior corrective surgery. Preoperative computed tomography (CT) was performed, surgery was planned, and pedicle screws were inserted as far as possible. When the pedicle was thin and screw insertion was difficult, a hook or sublaminar tape fixation was used. The surgeon determined the extent of fixation, following the report by Lenke et al..^19,20 After facetectomy at the apical lesion, two titanium alloy rods were bent identically, 6.0 mm in diameter. The two rods were connected to screw heads and then rotated simultaneously. Compression was performed on the convex side and distraction was performed on the concave side. When residual vertebral rotation or additional correction was needed, direct vertebral rotation or additional bending was used. A soft corset was worn for 6 months postoperatively.

2.3 Measurement parameters

Full-spine posteroanterior radiographs were obtained. Each radiograph was obtained in the natural standing position with the eyes straight ahead. The patient was in a standing position so that the line connecting the toes of both feet was parallel to the x-ray cassette. The hands naturally droop in the anteroposterior position. Using posteroanterior radiographs at preoperative and at 1 week, 3 months, and 2 years postoperatively, the following parameters were measured. Clavicle angle (CA), coracoid height difference (CHD), and T1 tilt were measured to evaluate the shoulder balance. For the pelvis, the left-right ratio of the iliac width (L/R ratio) and pelvic obliquity (PO) of the right and left pelvis were measured. The CA was defined as the angle between the horizontal plane and the line connecting the highest point of each clavicle. The CHD was defined as the height difference in millimetres between the coracoid processes by drawing a horizontal line at the upper margin of each process. T1 tilt was defined as the angle between the upper endplate of T1 and the horizontal plane (Fig. 1A). The PO was defined as the angle between the horizontal plane and the line connecting the upper edges of both iliac ridges. The radiographic measurement of pelvic rotation was based on the report of Lucas et al.,²¹ in which the best-positioned and most reliable landmarks were the inferior iliac crest of the sacroiliac joint on the medial side and the anterior superior iliac spine on the lateral side; this distance was measured to determine the L/R ratio (Fig. 1B). The CA, CHD, T1 tilt, and PO were defined as positive values that increased toward the left. The extent of change in CA, CHD, and L/R ratio from 1 week postoperative to 2 years postoperative were calculated as ΔCA, ΔCHD, ΔT1 tilt, ΔPO and ΔL/R ratio. Each absolute value was described as |CA|, |CHD|, |ΔCA|, |ΔCHD|, |ΔT1 tilt|, |ΔPO|, and |ΔL/R ratio|. A CA≧2° and a CHD≧9 mm were defined as shoulder imbalance.²² Pelvic rotation was defined as a change in the L/R ratio of 0.1 or more.²³ Patients were divided into two groups according to the presence of preoperative pelvic rotation or not (rotation group [pre L/R ratio≧0.1] and non-rotation group [pre ΔL/R ratio<0.1]). These two groups were examined to determine 1) whether they had balanced shoulders at 2 years postoperative and 2) whether their shoulder balance worsened during the postoperative course. Additionally, a comparative analysis was performed between the two groups. The correlation between the measured parameters and the extent of change in each parameter was also examined. Patients fixed to L2 or lower were defined as the lumbar spine fixation group, and those fixed up to L1 were defined as the non-fixation group. The effect on the relationship between PAR and shoulder balance was examined using partial correlation analysis.

Close

Fig. 1 Measurement of radiographic parameters A. Clavicle angle (CA) was defined as the angle between a line connecting the highest points of both clavicles and the horizontal plane. The coracoid height difference (CHD) was defined as the height difference in millimetres between the coracoid processes by drawing a horizontal line at the upper margin of each. T1 tilt was defined as the angle between the horizontal line and the line through the upper endplate of T1 B. Pelvic obliquity (PO) was defined as the angle between the line connecting the upper edges of both iliac ridges and the horizontal line. The medial side was the inferior iliac crest of the sacroiliac joint and the lateral side was the anterior superior iliac spine, and this distance was measured to obtain the left-right ratio of iliac width (L/R ratio).

Export to PPT

Each parameter was measured by three independent observers from images created using the DICOM viewer on a DICOM workstation. An electronic caliper was used for the measurements. All measurements were taken by three independent observers and averages were used. Inter- and intra-rater reliability was good (κ > 0.80).

2.4 Statistical analysis

All analyses were performed using SPSS Statistics for Windows version 13 (SPSS Inc.). Values are presented as means and standard deviations (SD). Statistical significance was set at P < 0.05. Normality was checked and Pearson's product-rate correlation and Spearman's rank correlation were used for correlation comparisons as appropriate. Each independent variable was compared between the two groups using independent t-tests for continuous variables and chi-square tests for categorical variables. Correlation analysis was evaluated using Spearman's rank correlation coefficient. Patients fixed to L2 or lower were defined as the lumbar spine fixation group, and those fixed up to L1 were defined as the non-fixation group. Partial correlation analysis was performed to examine the relationship between PAR and shoulder balance, controlling for potential confounding variables, while stratifying by the presence or absence of lumbar fusion. Lumbar fixation is recorded as a binary variable (0 = no fixation, 1 = fixation). This method allowed for the evaluation of the association between PAR and shoulder balance independent of other influencing factors, such as lumbar fusion.

3.1 Participants

The mean Cobb angle for the 28 patients was 54.1° ± 2.28. The mean postoperative Cobb angle was 11.7° ± 1.12, and the correction rate was 76.3 % ± 1.92. The mean L/R ratio was 0.94 ± 0.01, and preoperative PAR was observed in 11 of 28 cases (39.3 %). The most extensive fixation was from T1 to L3, with an average fixation level of 9.75 vertebrae. Twelve patients were fixed to the lumbar spine below L2. Fig. 2 shows pre-to postoperative changes in the L/R ratio. There was no significant change in the PAR between the preoperative and immediate postoperative periods (1 week); however, a significant change was observed from 1 week to 3 months postoperatively, which remained stable until 2 years. During the 2-year postoperative course, pelvic rotation was observed in 6 cases (|ΔL/R ratio|≧0.1). Shoulder imbalance persisted for 2 years postoperatively in 16 of the 28 cases. There were 8 cases of worsening shoulder balance during the 2-year postoperative course.

Close

Fig. 2 Pre-to postoperative change in the left-right ratio of iliac width (L/R ratio) There was no significant change in pelvic axis rotation (PAR) between the preoperative and immediate postoperative periods; however, significant changes were observed at 1 week and 3 months postoperatively and remained stable until 2 years. NS: no significant; S: significant.

Export to PPT

3.2 Comparison analysis

The rotation group had significantly more instances of balanced shoulders at 2 years postoperatively than the non-rotation group (Table 2a). In one patient, where the CHD was balanced but the CA was imbalanced, the comparison in CA did not show a significant difference between the two groups (Table 2b).

The rotation group had significantly greater ΔCHD and ΔCA than the non-rotation group. However, there was no significant difference in the extent of change in pelvic rotation between the two groups (Table 3).Most patients in the rotation group showed no deterioration of shoulder balance during the postoperative period (Table 4a, Table 4b). The ratio of shoulder balance to shoulder imbalance was the same for both CA and CHD. Of the six cases in whose pelvis rotated in the postoperative course, five had preoperative PAR.

3.3 Correlation analysis

The ΔL/R ratio was negatively correlated with ΔCHD, ΔCA, and ΔT1 tilt (r = −0.425, P = 0.024.; r = −0.457, P = −0.015.; r = −0.520, P = 0.005) (Table 5a, Table 6a). |ΔL/R ratio| was positively correlated with |ΔCHD| and |ΔCA| (r = 0.375, P = 0.049.; 0.388, P = 0.041.) (Table 5b). No correlation was observed between the values at 2 years postoperatively. The ΔPO did not correlate with the extent of change in shoulder parameters. PS density was not significantly correlated with any of the parameters. Partial correlation analysis was performed controlling for the presence or absence of lumbar fusion, but the correlation rate did not change significantly (Table 6a, Table 6b).

3.4 Representative case

A 15-year-old Lenke type 2 AIS male patient, a main thoracic spine Cobb angle of 70°, and preoperative and postoperative changes are shown. Preoperative pelvic axis rotation (PAR) was an L/R ratio of 0.85 (Fig. 3A). There was no change in PAR immediately postoperatively (L/R ratio = 0.83) (Fig. 3B); however, PAR changed during the postoperative course (L/R ratio = 1.16) (Fig. 3C). Shoulder balance also improved during the postoperative course (CHD: 22.83 mm to −2.38, CA: 3.9° to −1.6°, CA: 5.3°–0.8°).

Close

Fig. 3 Representative case A 15-year-old male patient, Lenke type 2 AIS, main thoracic spine Cobb 70°, showing preoperative and postoperative changes. A. Preoperative pelvic axis rotation (PAR) was defined as a left-right ratio of the iliac width (L/R ratio) = 0.85. B. Immediately postoperatively, there was no change in PAR (L/R ratio = 0.83). C. The PAR changed during the postoperative course (L/R ratio = 1.16). Shoulder balance also improved during the postoperative course (CHD: 22.83 mm to −2.38, CA: 3.9° to −1.6°, CA: 5.3°–0.8°). CA: Clavicle angle; CHD: Coracoid height difference.

Export to PPT