Interbody cage support improves reconstruction of sagittal balance after anterior selective fusion in Lenke type 5 idiopathic scoliosis patients

Abstract

Objective:  To investigate the effect of interbody cage support on reconstruction of the sagittal profile after anterior selective correction in Lenke type 5 adolescent idiopathic scoliosis (AIS) patients.

Methods:  In this retrospective study, a total of 40 AIS patients with a Lenke type 5 curve who underwent anterior selective fusion (ASF) using a single rod were studied. They were divided into two groups: Group A, bone grafting with an interbody titanium mesh cage (TMC); and Group B, bone grafting without a TMC. There were 19 female and 1 male patient aged 12–18 years in Group A, and 17 female and 3 male patients aged 13–20 years in Group B. All patients were followed up for more than 18 months. The coronal correction of the scoliosis and reconstruction of the sagittal profile were evaluated retrospectively.

Results:  By final follow‐up, more than 70% correction had been achieved for the primary curve and a spontaneous correction for the secondary curve in both groups. Satisfactory reconstructions on the sagittal plane were also observed. Similar changes were found in Group A and Group B, including increases in thoracic kyphosis, slight changes in thoracolumbar junctional kyphosis, no significant losses of lumbar lordosis, mild increases in proximal junctional measurement, and remarkable lordosis losses in the instrumented segments. With lessening of the sagittal balance and L₁ offset, a significant anterior shift of the C₇ plumb line was noted during follow‐up in Group A, indicating a more balanced spine in this group.

Conclusion:  Anterior selective single rod instrumentation and fusion is a recommended method for Lenke type 5C AIS. A structural interbody cage does not appear to improve the regional profile, nor the profiles in the instrumented area and the adjacent proximal and distal segments; yet could result in a better total balance in the long‐term.

Introduction

Despite the continual evolution of correction surgery, the goals of surgical treatment for adolescent idiopathic scoliosis (AIS) remain the same: aiming to obtain solid and safe correction, and to provide coronal and sagittal balance. In comparison with traditional posterior spinal fusion, anterior fusion has the advantages of achieving fusion more quickly and better derotation results ¹ , ² , ³ , ⁴ .

As part of the surgical procedure of anterior fusion for thoracolumbar/lumbar scoliosis (Lenke type 5C), the intervertebral discs localized within the fused segments have to be removed. This procedure can be accompanied by height loss of the anterior column, which can, in turn, result in thoracolumbar kyphosis, kyphotic deformity of the fused segments or even sagittal imbalance post‐operatively ⁵ , ⁶ . With the aim of preserving a balanced sagittal profile post‐operatively after anterior single rod instrumentation, Sweet et al. recommended placing a TMC as an interbody structural support to preserve the intervertebral height ³ . In a previous study of TMC with short‐term postoperative follow‐up, we found that a balanced sagittal profile could be obtained in AIS patients with Lenke type 5C and 6C ⁷ . However, there are few reports regarding the sagittal profile in Lenke type 5C patients at a long‐term follow‐up. The current study investigates the relatively long‐term effect of TMC support on reconstruction of the sagittal profile after ASF using TMC implantation.

Materials and methods

Patients

A series of 82 consecutive idiopathic thoracolumbar/lumbar scoliosis patients who had undergone ASF at our hospital from January 2002 to December 2005 was retrospectively reviewed. The inclusion criteria for this study were as follows: curve pattern type 5C according to the Lenke classification system ⁸ ; a minimum follow‐up of 1.5 years; and full records of pre‐ and post‐operative standing‐position X‐ray films of the whole spine. Fifty‐one patients met these criteria, among whom 11 were excluded because the poor quality of their lateral X‐ray films of the spine did not allow adequate assessment. Thus a total of 40 patients were included in this study, and these were divided into two groups according to the intervertebral bone grafting techniques employed: Group A, bone autograft with TMC implantation (n= 20); Group B, bone autograft without TMC implantation (n= 20). There were 19 female and 1 male patient in Group A aged from 12–18 years (mean, 14.8 years); and 17 female and 3 male patients in Group B aged from 12–18 years (mean, 15.8 years). The Risser Grade ranged from two to five in Group A and three to five in Group B. Sixteen patients were classified as Lenke type 5CN and 4 as Lenke type 5C‐ in Group A, and 12 as Lenke type 5CN and 8 as Lenke type 5C‐ in Group B.

Surgical technique

All patients were treated by anterior‐only selective fusion with a single rigid rod. All procedures were performed through traditional thoracoabdominal approaches to allow exposure of the convexity of the major thoracolumbar curve after removal of the 10th or the 11th rib. In general, the plan was to fuse the vertebral levels localized within the primary thoracolumbar/lumbar curve, as measured by the Cobb method, from the upper to the lower end vertebrae, as described in our previous study ⁹ . The segmental vessels on the convex side of the vertebral region to be fused were isolated and ligated. Thorough discectomies were performed within the region. The annulus fibrosis and nucleus pulposus were removed back to, but not through, the posterior longitudinal ligament. The cartilage vertebral endplates were abraded while the integrity of the bony endplates was preserved with care. Screws (CD‐Horizon, Medtronic Sofamor‐Danek, Minneapolis, MN, USA; and ISOLA, Depuy Spine, Raynham, MA, USA) were then inserted into all vertebrae within the fusion region. The rotational alignment of the screws was adjusted according to the rotational status of each vertebra, and care taken to ensure coronal plane placement of the screws parallel to the lower endplate of each vertebra. For patients in Group A, a TMC filled with morselized rib and iliac bone was placed into the prepared intervertebral space; and in Group B, grafting with morselized rib and iliac bone was performed. Both the TMC and the morselized autografts were mainly placed in the anterior part of the intervertebral space with the aim of preventing anterior column collapse and segmental kyphosis. A single rod was cut and contoured to the appropriate coronal and sagittal alignment of the instrumented segments. The rod was engaged with all screws, and then rotated until the appropriate coronal and sagittal alignment had been achieved. After rod rotation had been completed, intersegmental compression was applied and the screws immediately tightened. Plugs were tightly locked to the screws once a C‐arm radiograph showed satisfactory correction of the scoliosis. Chest tube drains were placed in all patients before routine wound closure.

Radiographic analysis

Radiographic analyses were carried out on both side‐bending and standing‐position whole‐spine X‐ray films which had been taken preoperatively, 1–2 weeks postoperatively and at the most recent follow‐up. The flexibility of the primary curve was evaluated through measurements on side‐bending X‐rays. Sagittal balance, recorded as the sagittal vertical axis (SVA), was assessed by measuring the horizontal distance between the C₇ plumb line (C₇PL) and the posterior superior corner of the first sacral vertebra ¹⁰ . Additionally, to assist with the analysis of sagittal balance, the distance from the C₇PL to the center of the first lumbar vertebra (L₁ offset) was also measured ¹¹ . The SVA and L₁ offset were considered positive if the C₇PL was anterior, and negative if posterior, to the posterior superior corner of the first sacral vertebra and L₁ center, respectively. Angulation measurement on the sagittal planes included thoracic kyphosis (TK, from T₅ to T₁₂), thoracolumbar junctional kyphosis (TJK, from T₁₀ to L₂), lumbar lordosis (LL, from T₁₂ to S₁) and sacral slope (SS). In addition, the instrumented segments angle (ISA) was measured from the upper instrumented vertebra (UIV) to the lowest instrumented vertebra (LIV), and proximal junctional measurement (PJM), distal junctional measurement (DJM) and distal junctional kyphosis (DJK) were analyzed as described in previous studies ¹² . For the measured angles, positive values were used to denote kyphosis and negative to indicate lordosis. These means of measurement are shown in 1, 2.

Figure 1.

Depiction of the means of measurement used in this study. The measurements of the sagittal regional angulation and the balance include TK (from T₅ to T₁₂), the TJK (from T₁₀ to L₂), LL (from T₁₂ lower end plate to S₁ superior plate), SS (the angle of sacral slope), the SVA (the horizontal distance between C₇PL and the posterior superior corner of S₁), and the L₁ offset (the horizontal distance between C₇PL and the center of L₁). ll, line of the lower endplate; ul, line of the upper endplate.

Figure 2.

Depiction of the means of measurement used in this study. The measurements of the sagittal segmental and junctional angles include the ISA (from the upper plate of the UIV to the lower end plate of the LIV), the PJM (between the upper plate of the second vertebra cephalad to the UIV and the lower end plate of the UIV), the DJM (between the upper plate of the second vertebra caudal to the LIV and the lower end plate of the LIV) and the DJK (between the upper plate of the first vertebra caudal to the LIV and the lower end plate of the LIV).

Statistical analysis

Data was analyzed using SPSS 13.0 (SPSS, Chicago, IL, USA). The sagittal postoperative outcomes of patients in both groups were described. One‐way ANOVA was used to determine the difference between patients in Group A and those in Group B regarding sagittal balance and angulations on the preoperative, postoperative, and the latest follow‐up sagittal X‐ray films. A paired t‐test was used to determine the significance of changes in sagittal balance and angulations in each group. In this study, statistical significance was set at P < 0.05.

Results

The preoperative Cobb angles of the primary and secondary curves, ratio of the primary to the secondary curve and curve flexibility were similar in the two groups. In Group A, the primary curve averaged 45.3° (range, 40°–60°); the secondary curve averaged 24.9° (12°–34°); the mean flexibility of the primary curve was 81.7%; and the mean ratio of the primary to the secondary curve was 1.97 (range, 1.47–3.71). In Group B, the primary curve averaged 47.5°± 6.1° (range, 40°–62°); the secondary curve averaged 29.9° (15°–40°); the mean flexibility of the primary curve was 81.4%; and the mean ratio of the primary to the secondary curve was 1.76 (range, 1.18–2.91). Fusion levels were similarly distributed in both groups: in Group A, 12 cases were fused from T₁₁ to L₃, 3 from T₁₁ to L₄, 2 from T₁₂ to L₄, 2 from T₁₀ to L₃, and 1 from T₁₀ to L₂; and in Group B, 7 from T₁₁ to L₃, 5 from T₁₂ to L₄, 4 from T₁₂ to L₃, 2 from T₁₁ to L₄, 1 from T₉ to L₂, and 1 from T₉ to L₃. After the surgery, patients in Group A and Group B were followed up for a mean duration of 21.7 months (range, 18–33) and 25.6 months (range, 19–38), respectively. By the most recent follow‐up, bone grafting with or without TMC was found to have fused in both groups, with trabecular bone integrating through the intervertebral space. In both groups there was no deep infection, neurological or visceral injury, or instrumentation related complication, and pseudoarthrosis occurred in no case.

Coronal plane correction

The correction results for both groups are shown in Table 1. The mean preoperative flexibility of the primary curve was similar for both groups (81.7% in Group A vs. 81.4% in Group B). Satisfactory correction of the primary curve was obtained in both groups. The correction rate for Group A was slightly higher than that for Group B (postoperative, 79.9% vs. 76.2%; and at the most recent follow‐up, 76.3% vs. 72.7%) but these differences are not statistically significant (P > 0.05). At the most recent follow‐up, eight cases in Group A displayed a mean of 3.5° loss of correction, and six in Group B a mean of 5.9° loss of correction, while one case in Group B had a correction loss of 18°. In the unfused thoracic curve, spontaneous correction of 12.0° and 9.6° was achieved in Group A and B, respectively.

Table 1.

Coronal plane correction of the primary curve in both groups

Group	n	The primary curve magnitude
		Preoperative (flexibility)	Postoperative (correction rate)	Most recent follow‐up (correction rate)
A	20	45.3°± 5.6° (81.7%)	9.0°± 5.4° (79.9%)	10.7°± 6.5° (76.3%)
B	20	47.6°± 6.1° (81.4%)	11.4°± 9.7° (76.2%)	12.9°± 7.9° (72.7%)

The primary curve magnitude, flexibility and correction rate show no significant difference between the groups.

Reconstruction of the sagittal profile

Changes in sagittal balance and angulations are shown in Table 2. Preoperatively, the values of TK, TJK, LL and SS were similar for both groups. TK increased by 1.0° and 1.3° in Group A and Group B postoperatively, while displaying another increase of 1.5° and 1.3° by the most recent follow‐up, respectively. Variations in TJK were not significantly different between the groups from preoperatively to the most recent follow‐up. Despite a slight decrease between pre‐operation and post‐operation in Group A, an tendency for TJK to increase by 2° was observed in both groups during the post‐surgical follow‐ups. By the most recent follow‐up, three cases in each group had a TJK greater than 10°. The values of, and the changes in, LL and SS were similar in both groups between pre‐operation and the most recent follow‐up.

Table 2.

Changes in sagittal balance and angulations in both groups

	Group	TK (°)	TJK (°)	LL (°)	SS (°)	SVA (mm)	L1 offset (mm)
Pre‐operation	A	15.2 ± 7.2	3.2 ± 6.4	−48.1 ± 9.8	35.6 ± 5.5	−22 ± 24	−23 ± 16
	B	12.2 ± 7.3	1.2 ± 5.7	−46.8 ± 8.0	35.7 ± 6.0	−17 ± 24	−23 ± 16
Post‐operation	A	16.2 ± 5.8	2.0 ± 5.4	−48.3 ± 8.5	34.2 ± 5.2	−24 ± 19	−24 ± 15
	B	13.5 ± 7.6	1.6 ± 5.7	−46.6 ± 6.6	35.3 ± 4.7	−22 ± 22	−24 ± 15
Latest follow‐up	A	17.7 ± 6.6	4.7 ± 5.4	−48.4 ± 8.3	35.5 ± 5.8	−15 ± 21	−12 ± 16†‡
	B	14.8 ± 7.3	3.5 ± 6.9	−48.9 ± 10.4	36.6 ± 7.4	−21 ± 22	−19 ± 15

^{† and ‡} indicate a significant difference (P < 0.05) between that at most recent follow‐up and preoperatively, and between that at most recent follow‐up and postoperatively, respectively.

The preoperative SVA and L₁ offset in both groups averaged about −20 mm, indicating a slightly negative sagittal balance, and were not significantly different between the two groups. After the correction surgery, no marked change in either index was found. However, the negative value of the SVA in Group A displayed a decreasing trend, from −24 mm postoperatively to −15 mm at the latest follow‐up, though this difference was not significant (P= 0.066). Moreover, the L₁ offset turned out to be −12 mm by the most recent follow‐up, significantly less than the pre‐ and post‐operative measurements (P= 0.015 and 0.001, respectively). However, no obvious changes in the SVA and L₁ offset were found during follow‐up in Group B.

Table 3 shows the changes in segmental and junctional angles from pre‐operation to the most recent follow‐up. Prior to surgery, the ISA averaged 6.0° and 8.6° lordosis in Group A and Group B, respectively. The lordosis of ISA showed a decrease of 1.1° and 3.8° after surgery, gradually decreased by 5.5° and 7.3° during follow‐up, and turned out to be 0.5° and 1.3° lordosis by the most recent follow‐up in Group A and Group B, respectively. The changes in the SVA in both groups were statistically significant (P < 0.05). For one case in Group A and three cases in Group B, an ISA of greater than 10° kyphosis was recorded at the most recent follow‐up. The PJM displayed an increase of 0.5° and 1.0° during follow‐up in Group A and Group B respectively, which was found to represent a significant change between pre‐operation and the most recent follow‐up (P= 0.030). A similar trend in the DJM was observed in both groups, that is, decreasing between pre‐ and post‐operation and then increasing during follow‐up. The DJK was observed to be rapidly reduced postoperatively, but there were only trivial variations during the follow‐up.

Table 3.

Changes in segmental and junctional angulations in both groups

	Group	ISA (°)	PJM (°)	DJM (°)*	DJK (°)
Pre‐operation	A	−6.0 ± 7.9	2.5 ± 3.0	−32.7 ± 8.0	−17.7 ± 6.9
	B	−8.6 ± 9.8	2.4 ± 3.8	−35.1 ± 5.9	−18.8 ± 5.1
Post‐operation	A	−4.9 ± 8.2	4.3 ± 2.9	−28.6 ± 8.8	−13.6 ± 5.4
	B	−4.8 ± 6.8	2.4 ± 4.3	−29.0 ± 8.3†	−15.5 ± 5.5
Latest follow‐up	A	−0.5 ± 7.0†‡	4.8 ± 4.0†	−32.6 ± 8.2	−13.9 ± 5.8
	B	−1.3 ± 9.6†‡	3.4 ± 4.0	−33.3 ± 5.7	−15.6 ± 6.9

The preoperative parameters were measured on the preoperative standing X‐ray film according to the fusion strategy.

^* Measurements of the DJM were completed in only 15 cases from Group A and 13 from Group B, due to the fact that the DJM equals the DJK if the LIV is localized at L₄.

^{† and ‡} indicate a significant difference (P < 0.05) between that at most recent follow‐up and pre‐operation, and between that at most recent follow‐up and post‐operation, respectively.

Discussion

The clinical outcome of ASF for type 5 AIS

AIS Lenke type 5C refers to a primary thoracolumbar/lumbar curve with or without a secondary nonstructural thoracic curve ⁸ . ASF has been proven to be an effective treatment for scoliosis of this type. However this approach is not indicated in younger patients with Risser Grade 0 or 1 because of the risk of coronal decompression and aggravation of the secondary curves. As recommended by Lenke, anterior selective thoracolumbar/lumbar fusion is indicated in patients who are Risser Grade 2 or more, have a ratio of the primary thoracolumbar/lumbar to the secondary thoracic curve of not less than 1.25, and a secondary thoracic curve of less than 50° ¹³ . Sanders et al. summarized that if patients have a closed triradiate cartilage, a ratio of the thoracolumbar/lumbar to the thoracic curve of more than 1.25, and the thoracic curve bending out to 20° or less, ASF could result in a satisfactory outcome ¹⁴ . And Ogilvie stated that the cosmetic appearance should be taken into account, and recommended supplementary indications which included a thoracic curve of less than 40° which was supple and with no obvious rib hump deformity ¹⁵ . In our series, all patients were classified as Lenke type 5C, had Risser Grade no less than 2 and closed triradiate cartilages. In only one patient was the ratio of the thoracolumbar/lumbar to the thoracic curve less than 1.25 (being 1.18), in this case no obvious rib hump deformity was observed. There have been some studies regarding the clinical outcome of anterior selective thoracolumbar/lumbar fusion, and it is generally agreed that this surgical approach leads to a good correction rate (up to 60%–81%) ³ , ⁷ , ¹⁶ . In this study, satisfactory correction of the primary curve was obtained in both groups up to the most recent follow‐up (correction rate greater than 70%), with spontaneous correction of the unfused thoracic curve, apart from one case in Group B who had a correction loss of the primary curve greater than 15°. There were no instrumentation related complications nor pseudoarthrosis. The satisfactory results of this study might be attributed to the fact that the indications for ASF were strictly controlled.

Regional angulation and balance reconstruction in the sagittal plane

Presently, evaluation of the sagittal spine mainly focuses on the regional angulations and total balance. The parameters used in the current study include TK, TJK, LL, SS, SVA and L₁ offset, which reflect all of the following: thoracic, thoracolumbar, lumbosacral angulations and the sagittal alignment of the vertebrae of C7, L1 and S1. With thorough measurement of these parameters, we believe the sagittal angulations and balance of Lenke type 5C patients after ASF can be comprehensively assessed.

After ASF, TK was found to increase from pre‐ to post‐operation in this study, with a increment of 2.5° and 2.6° in Group A and Group B respectively, tending to reconstruct a normal thoracic kyphosis. Rhee et al. reported that an increment of about 6° in TK was observed in 27 AIS patients at a minimum of 2 years follow‐up after anterior thoracolumbar fusion ¹² . Though the increment of TK in our study was less than that in the study by Rhee et al. ¹² , the trend in TK change is similar. We think that the trend for TK to increase may result from spontaneous correction of the unfused thoracic curve. Besides, the contribution of spinal growth should not be excluded, as Mac‐Thiong et al. have observed that total thoracic kyphosis (from T₁ to T₁₂) increases with age ¹⁷ . The TJK displayed a trivial variation postoperatively and increased slightly during the follow‐up in both groups, these changes in TJK were statistically insignificant, and thus in accordance with previous studies. In the study by Rhee et al., no obvious change in TJK was observed during a minimum of two years follow‐up after anterior thoracolumbar fusion with TMC ¹² . And, in another study, no remarkable change was found in TK during a greater than two year follow‐up after anterior thoracolumbar fusion with TMC, as reported by Sweet et al. ³ As stated in some previous studies, a decrement in LL represents reconstruction loss of the sagittal profile after anterior thoracolumbar fusion for type 5 AIS ⁵ , ¹⁸ . In our study, we found no obvious decrement in LL from pre‐operation to the most recent follow‐up in both groups, which is comparable to the results reported by Rhee et al. ¹² Whereas, in the study performed by Sweet et al., the loss of LL was approximately 4° during postoperative follow‐up ³ . Watkins et al. reported that structural grafts in the intervertebral space were able to maintain the lordosis of the lumbar region, while morselized rib autografts resulted in approximately 5° loss of LL ¹⁹ . Yet, Lowe et al. compared the outcomes of intervertebral grafting with TMC and morselized rib bones in anterior thoracolumbar fusion, and found slight but not insignificant decrements in LL regardless of the grafting technique ¹⁶ .

SVA, the parameter denoting the sagittal alignment of C₇ and S₁, has been widely used to evaluate the total sagittal balance of the spine ¹⁰ . The L₁ offset, the sagittal translation of L₁ to C₇PL, is another important parameter of sagittal balance and is found to be highly correlated with the SVA ¹¹ . In this study, both the SVA and L₁ offset showed no significant changes from pre‐operation to the most recent follow‐up in Group B, while in Group A, an average of 7 mm decrease in the SVA and 11 mm decrease in the L₁ offset was observed by the most recent follow‐up, implying a less negative balance due to anterior displacement of C₇PL. The results of our study are in accordance with the study of Rhee et al., in which the SVA decreased slightly, from −25 mm to −17 mm, during follow‐up after anterior thoracolumbar fusion with TMC grafts ¹² . However, the outcome regarding sagittal balance after anterior thoracolumbar fusion remains controversial. Watkins et al. reported the SVA changed remarkably, from −8 mm preoperatively to +7 mm at follow‐up, in patients with morselized rib grafts, while it remained at −4 mm in patients with structural grafts ¹⁹ . And Lowe et al. reported that there were no significant changes in the SVA in patients with TMC grafts or morselized rib grafts ¹⁶ .

In our study, the increments in TK and small or slight changes in the TJK and LL were found to be of no statistically significant difference between Group A and Group B, implying that intervertebral grafting techniques using either TMC or morselized rib bones do not significantly influence the sagittal regional angulations after anterior thoracolumbar fusion. However, changes in the SVA and L₁ offset were found to be remarkable, and significant between groups. It has been reported that the SVA averages −3.8 mm in healthy adults ¹⁰ , and ranges from −0.8 mm to −8.6 mm in healthy adolescents 10–15 years of age ¹¹ . Obviously, a normal sagittal spine balance is represented by a small negative value of the SVA. As compared to the reported SVA range in healthy populations ¹⁰ , ¹¹ , the sagittal balance till the most recent follow‐up remained less in both groups of our study. However, in Group A the negative value of the SVA decreased because of the anterior migration of the C₇PL during follow‐up, and the sagittal balance tended to be reconstructed close to normal. This finding indicates that long‐term sagittal balance may be able to improve despite insignificant changes in regional angulations.

Segmental and junctional sagittal angulations

Owing to the mechanism by which compression and derotation forces are directly applied to the vertebrae within the fusion range, it is thought by some authors that ASF is kyphogenic in regards to the sagittal plane ⁵ , ⁶ , ¹⁶ . It was demonstrated in a previous report that, compared to posterior fusion, anterior thoracic fusion results in better restoration of thoracic kyphosis ¹² . Similar results have also been found in anterior thoracic fusion assisted by thoracoscopy ²⁰ . Accordingly, will anterior thoracolumbar fusion, in the same way, lead to loss of lordosis, or even increments in kyphosis of the fusion segments? In the case of thoracolumbar scoliosis, the above‐described kyphotic phenomenon in the fusion segments was scarcely seen with the use of earlier instrumentation (like Zielke Instrumentation) ⁵ , ²¹ . However, kyphosis in the fusion levels has been less frequently seen with utilization of the new‐generation instrumentation and improvements in intervertebral fusion techniques ³ , ¹⁶ , ²² . As stated by Sweet et al., ASF with rigid single rod and intervertebral TMC grafting for 20 cases of thoracolumbar scoliosis resulted in insignificant decrease in the ISA after a greater than two year duration ³ . Similar findings have been reported by Lowe et al. ¹⁶ . However, a remarkable decrease in the ISA has also been found in a few studies. Ouellet and Johnston observed a decrease of 8° in the angle of the ISA with a minimum 2‐year follow‐up after anterior thoracolumbar fusion ¹⁸ . It was reported in another study that the ISA changed on average from −7° to 9°, and from −1° to 8° with morselized rib grafting and with TMC grafting, respectively ¹⁹ . In our study, a slight decrease in the lordotic curve of the ISA was observed after surgery, and a further decrease during follow‐up in both groups. As compared to pre‐operation, the ISA at the latest follow‐up had decreased by 5.5° and 7.3° in Group A and Group B, respectively.

The postoperative decrease in the ISA is thought to be mainly related to the surgical procedures, including the insertion of screws, grafting of bones and compression between segments. Generally, the screws should be inserted as posteriorly as possible, intervertebral grafts and supports should be placed as anteriorly as possible, and intersegmental compression should be performed gently, so as to avoid causing kyphosis of the fused segment intraoperatively. Additionally, the residual growth potential of the posterior spinal column is believed to play a role in causing a decrease in the ISA. A previous study found that the potential for further growth in the growth plates is adequate in some AIS patients with Risser Grade 4, while it is poor in patients with Risser Grade 5 ²³ . Wang et al. reported similar findings in the iliac growth plates, which could be useful for evaluating the growth potential of the posterior spinal column ²⁴ . In our study, there were 28 patients graded as Risser 4 and less. Theoretically, the posterior spinal columns in some of these 28 patients have some potential for further growth. Further growth of the posterior column together with cessation of growth of the anterior column could thus lead to a decrease in the ISA or even a kyphotic ISA, providing a possible cause for the decrease in the ISA observed during postoperative follow‐up in this study.

As observed in the current study, the PJM increased marginally till the most recent follow‐up. Often seen during follow‐up after posterior fusion of the thoracic, thoracolumbar and lumbar spine, an increase in the PJM is relatively less often reported to occur during follow‐up after anterior thoracolumbar fusion ¹² , ²⁵ . In a study of 27 cases who underwent anterior thoracolumbar fusion, the PJM increased from 6° to 9° during the greater than 2‐year follow‐up period, and an increase of more than 10° was found in 5 cases ¹² . In addition, in patients with TMC and morselized rib grafting respectively, Watkins et al. reported an increase of 4°and 3° in sagittal angulation from T5 to the upper instrumented vertebra during a follow‐up of more than 2 years, as compared with that found pre‐operatively ¹⁹ . Obviously, the findings in these studies are in accordance with the results of our study.

In this study, patients in Group A had slightly smaller decreases in the ISA than in Group B, but the difference was statistically insignificant. Besides, the changes in the PJM, DJM and DJK were also similar in both groups. These findings suggest that, as compared with morselized rib grafting, intervertebral grafting with TMC might not significantly modify the trend towards decrease in the ISA and increase in the PJM during follow‐up.

Disclosure

The device(s)/drug(s) used in this study is/are approved by the corresponding national agency for this indication. No benefits in any form have been or will be received from a commercial party related directly or indirectly to the subject of this manuscript.