SPINE Volume 39, Number 13, pp E800-E810 ©2014, Lippincott Williams & Wilkins
DEFORMITY
Pedicle Screw Versus Hybrid Construct Instrumentation in Adolescent Idiopathic Scoliosis Meta-analysis of Thoracic Kyphosis Yun Cao, MD, Wei Xiong, MD, and Feng Li, MD
Study Design. A quantitative meta-analysis was conducted on publishing studies reporting results of spinal surgery in adolescent idiopathic scoliosis with instrumentation of pedicle screw or hybrid construct. Objective. The primary objective is to get overview of the power to restore thoracic kyphosis using 2 different instrumentations. Summary of Background Data. During the past years, surgery has mainly focused on correcting the coronal curve, but nowadays, the aims of posterior fusion and instrumentation are to achieve a stable, well-balanced spine in the coronal as well as in the sagittal planes. Methods. A PubMed and EMBASE search was conducted using combinations of the key words “hybrid construct” or “pedicle screw” with “adolescent idiopathic scoliosis” up to October 2013. A hand search of reference lists of obtained articles was also performed. Results. A total of 24 studies were identified and included in this meta-analysis. Of this, 12 included pedicle screw group only, 5 included hybrid construct group only, and 7 included both groups. The total number of patients was 1615. Age ranged from 9 to 26 years, with a mean of 15. For pedicle screw group, we got standard mean difference (SMD) as 0.40 (95% confidence interval, 0.31–0.50); I2= 91.4%. For hybrid construct group, we got SMD as 0.15 (95% confidence interval, 0.04–0.26). Both intervals of SMD lay in positive value side. Overall SMD was 0.30 (95% confidence interval, 0.23– 0.37). The positive value of SMD means thoracic kyphosis improves after surgery, whereas negative value means kyphosis loss after surgery. Conclusion. There is overall tendency for both instrumentations to restore thoracic kyphosis. Hybrid construct seems to be more powerful in restoring kyphosis than pedicle screw. Preoperative From the Department of Orthopaedics, Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, Hubei, People’s Republic of China. Acknowledgment date: December 9, 2013. First revision date: February 24, 2014. Second revision date: March 8, 2014. Acceptance date: March 13, 2014. The manuscript submitted does not contain information about medical device(s)/drug(s). No funds were received in support of this work. No relevant financial activities outside the submitted work. Address correspondence and reprint requests to Yun Cao, MD, Department of Orthopaedics, Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, Jiefang Road 1095, Wuhan, Hubei, People’s Republic of China; E-mail:
[email protected] DOI: 10.1097/BRS.0000000000000342
E800
important factor was found to be reasonably distributed or balanced. For operative factor, rod stiffness, fashion of hybrid construct, in situ bending, and rod rotation maneuver could have influence on results of kyphosis in different ways. Loss of correction and limited restoration of loss may do exist after correction. Key words: adolescent idiopathic scoliosis, pedicle screw, hybrid construct, thoracic kyphosis, meta-analysis. Level of Evidence: 3 Spine 2014;39:E800–E810
A
dolescent idiopathic scoliosis is a complex 3-dimensional musculoskeletal disorder characterized by lateral spine curvature in the coronal plane, thoracic hypokyphosis in the sagittal plane, and vertebral rotation in the transverse plane. During the past years, surgery has mainly focused on correcting the coronal curve, but nowadays, the aims of posterior fusion and instrumentation are to achieve a stable, well-balanced spine in the coronal as well as in the sagittal planes, prevent curve progression, and also achieve solid fusion, while keeping the fusion as short as possible.1 In the sagittal plane, results are differing in studies concerning scoliosis populations with thoracic hypokyphosis. Some authors have found an increase of kyphosis with hook, screw, or hybrid constructs whereas others have found no change or decrease of kyphosis with the same type of constructs particularly in recent studies.2–7 Most of the authors found an important number of cases remaining in hypokyphosis after surgery and a postoperative angle of kyphosis close to the inferior value. It seems that there is no ideal instrumentation in regard to sagittal plane correction.8 The purpose of the study is (1) to get an overview of the power to restore thoracic kyphosis using 2 different instrumentations, (2) to compare the efficacy of kyphosis restoration, and (3) to investigate the source of heterogeneity among the studies.
MATERIALS AND METHODS Search Strategy A PubMed and EMBASE search was conducted using combinations of the key words “hybrid construct” or “pedicle
www.spinejournal.com
June 2014
Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. SPINE131442_LR E800
08/05/14 3:26 PM
DEFORMITY
Pedicle Screw Versus Hybrid Construct Instrumentation in Adolescent Idiopathic Scoliosis • Cao et al
screw” with “adolescent idiopathic scoliosis” up to October 2013. A hand search of reference lists of obtained articles was also performed.
(if available); (23) preoperative sagittal balance; and (24) implant density.
Inclusion/Exclusion Criteria
Thoracic kyphosis from the included studies was grouped according to whether all-pedicle screws (>80%) were implanted or hybrid construct. A random-effects model acknowledges that the effect of treatment may not be identical from study to study because of heterogeneity. To study the source of heterogeneity, we conducted metaregression with some candidate potential factors, such as anchors type, Risser grade, flexibility of thoracic curve, and sagittal balance. Furthermore, we identified studies that included both pedicle screw group and hybrid construct group and applied metaanalysis of these studies alone. Most of the statistics underwent with the assistance of software Stata 12.0 (StataCorp LP, College Station, TX).
Only articles that reported all of the following information were included: (1) the patient population was adolescent idiopathic scoliosis only; (2) the intervention groups included pedicle screw group, hybrid construct group, or both; and (3) thoracic kyphosis result was reported. Pedicle screw group must satisfy that pedicle screws consisted more than 80% of anchors implanted. Hybrid construct group must satisfy that proximal hooks, wires, or clamps combined with distal pedicle screws were implanted. Articles were excluded if they matched 1 of these conditions: (1) patient population was mixed with other scoliosis, such as adult adolescent scoliosis, congenital scoliosis; (2) only anterior spinal fusion was used; (3) curve type of patients was mostly Lenke 5; and (4) article was case report, or review. Furthermore, taken the following reason into consideration, that is, the same institution was more likely to report results observed from the same population, we chose the most suitable one from articles reported by the same institution.
Review Procedure After being selected, studies were screened independently by authors (Yun Cao and Wei Xiong). Reviewers were not blinded to authors and journal. Rather, a study was included for analysis as long as both screeners agreed that it met inclusion criteria. No authors were contacted.
Data Extraction A meta-analysis database was created from included studies with the following categories: (1) study ID (author, journal, year published, institution); (2) reference; (3) study type (e.g., prospective, randomized, controlled, retrospective series); (4) data evidence class. Prospective randomized controlled trials were considered class I evidence. Good quality prospective cohort and retrospective case-control studies were considered class II evidence. Retrospective series and prospective observational studies were considered class III evidence; (5) study patient inclusion/exclusion criteria; (6) description of surgical procedure; (7) total number of patients; (8) patient age statistics including mean, standard deviation, median, and range (if available); (9) number of male and female patients; (10) final follow-up time; (11) number of patients undergoing pedicle screw instrumentation; (12) number of patients undergoing hybrid construct instrumentation; (13) use of monoaxial or polyaxial pedicle screw; (14) type of hybrid construct; (15) special surgical techniques applied (such as simultaneous translation on 2 rods, direct vertebral derotation); (16) distribution of curve types; (17) Cobb angle of thoracic curve; (18) flexibility of thoracic curve; (19) number of fused levels; (20) Risser grade; (21) preoperative thoracic kyphosis statistics including mean, standard deviation, median, and range (if available); (22) postoperative thoracic kyphosis statistics including mean, standard deviation, median, and range Spine
Statistical Analysis
RESULT Description of Study The search identified 117 articles. Of these, 73 were excluded. Excluded studies included case reports, reviews, non-English articles, and articles that did not match inclusion criteria. Of the remaining 44 articles, 4 studies reported mainly Lenke 5 curve, 1 did not include either pedicle screw group or hybrid construct group, and 1 was duplicate. Fourteen articles were excluded because the same institutions reported other articles included. The 24 remaining studies fit criteria and were included in the meta-analysis. Twelve studies included pedicle screw group only,9–20 5 included hybrid construct group only,1,7,21–23 and 7 included both groups.3,6,24–28 Descriptive information for each included study is given in Table 1. The total number of patients was 1615. Age ranged from 9 to 26 years, with a mean number of 15. Of these 24 remaining studies, 12 included pedicle screw group only. The density of screws in each study did vary. In some study, nearly 100% screws were implanted on both concave and convex sides,9 and in some other studies, only concave side needed 100% screws, whereas convex side got 33% to 50% screws.11 Some studies specifies the screw types into monoaxial and polyaxial.923 Furthermore, 3 studies contained more than 1 pedicle screw group because of different surgical techniques.16,17,19 Five studies included hybrid construct group only. Hybrid construct consisted of proximal hooks (can also be wires or clamps) and distal pedicle screws. One study included 2 types of hybrid construct groups because of different anchors,23 in which one was hook and the other was universal clamp. In total, there were 7 studies that included paired-compare groups between pedicle screw group and hybrid construct group. In these studies, main characteristics of preoperative demographics showed no significant difference between pedicle screw group and hybrid construct group, such as age, sex, curve type, and Cobb angle. Preoperative and postoperative thoracic kyphosis is shown in Table 2. For correction index, it equals to (mean of postoperative kyphosis/mean of preoperative kyphosis − 1) × 100% and was www.spinejournal.com
E801
Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. SPINE131442_LR E801
08/05/14 3:26 PM
TABLE 1. Characteristics of Included Study
SPINE131442_LR E802
25 24 24 20
PS PS PS PS PS PS (VCA) PS (SR) PS (C-T)
Retro.
Retro.
Prosp.
Retro.
13
14
Retro.
Retro.
Retro.
Retro.
Retro.
Retro.
Retro.
Mladenov et al19
Liljenqvist et al20
Blondel et al21
Sucato et al7
Chaiyamongkol et al22
Krallis et al1
Ilharreborde et al23
18
Retro.
Retro.
Chunguang et al17
Basu et al
Prosp.
Qiu et al16
15
Fu et al
Clement et al
Vallespir et al
Potter et al
12
Retro.
Suk et al
15
PS (SR)
75
75
HC (UC) HC (hook)
43
52
86
60
HC
HC
HC
HC
18
15
PS (DVD)
PS
14
26
PS (CSLRR) PS
24
PS (N-T)
62
25
20
203
116
PS
89
Retro.
Study ID
11
Study Type
Abul-Kasim et al10
PS
Group
Retro.
Number
E802
Hamzaoglu et al9
Age, yr (Range) 15 (13–17)
16 (15–17)
14 (10–18)
15 (13–17)
15 (12–18)
15 (13–18)
15 (16–18)
15 (13–17)
14 (9–23)
15 (10–17)
16 (11–21)
15 (10–20)
14 (13–16)
15 (12–22)
12 (10–24)
14 (12–15)
14 (9–18)
15 (12–24)
16 (13–19)
Sex (Female/ Male) 59/16
64/11
29/14
44/8
52/8
11/3
58/12
54/8
94/22
77/12
Risser Sign 3
3.8
3.6
3
Flexibility 0.42
0.47
0.48
0.624
0.623
0.55
0.6
0.49
0.54
0.51
Sagittal balance 15.48
1.7
1.9
30.5
24.5
16.5
16.1
16.6
Fused Level 12
14
10.5
10
12
7.7
Lenke 1 0.48
0.51
0.74
0.44
0.53
1
1
0.91
0.91
0.91
1
1
0.64
0.63
1
1
0.52
0
Lenke 3 0.08
0.03
0.09
0.06
0.46
0
0
0
0
0
0
0
0.12
0.02
0
0
0.16
0.52
Follow-up ≥2 yr
≥2 yr
≥2 yr
≥2 yr
2y
3m
≥2 yr
≥2 yr
≥2 yr
1 (1–2) yr
≥2 yr
≥2 yr
≥2 yr
IP
IP
2y
IP
IP
4.6 (2–7) yr
≥5 yr
6 wk
5.4(2–8) yr
Category of Hybrid Construct H+ W
UC
H+ W
www.spinejournal.com
Y
Y
Y
Y
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
Y
N
In situ Bending (Continued )
N
N
Y
N
6.35SS
H+ W
N
N
Y
Y
N
Y
Y
Y
Y
N
Y
N
N
5.5Ti
5.5Ti
5.5Ti
5.5Ti
5.5Ti
5.5Ti
6.0Ti
N
Y
Y
Y
H+ W
UC
Rod Stiffness 5.5Ti
Rod Rotation N
DEFORMITY Pedicle Screw Versus Hybrid Construct Instrumentation in Adolescent Idiopathic Scoliosis • Cao et al
Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
June 2014
08/05/14 3:26 PM
SPINE131442_LR E803
Retro.
Retro.
Retro.
Retro.
Retro.
Lowenstein et al6
Lonner et al25
Silvestre et al26
Crawford et al27
Yilmaz et al28
33 33 25
PS (POLY) HC PS
29 35 35
HC PS HC
34
34
PS (MONO)
PS
17
HC
27
17
PS
HC
29
29
PS HC
71
HC
56
PS
Age, yr (Range) 15(11–18)
14 (13–17)
15 (13–16)
15 (11–20)
16 (11–20)
14 (12–16)
14 (12–16)
15 (12–18)
13 (11–18)
14 (10–16)
15 (10–18)
15 (12–21)
14 (12–16)
Sex (Female/Male) 28/7
28/7
26/3
27/7
23/4
16/9
15/2
14/3
25/4
20/9
104/23
Risser Sign 3
3
4
3
2.34
3.04
2.4
2.8
Flexibility 0.42
0.47
0.51
0.52
0.27
0.26
0.5
0.51
0.53
0.47
0.43
0.51
Sagittal balance 42.2
30.5
64
52.5
1.5
3.9
3.9
21.24
28.82
20
19
1.09
0.91
Fused Level 10.1
10.6
12.89
11.84
12
10
11
13
11.7
11.3
10.3
Lenke 1 0.69
0.77
1
1
0.40
0.36
0.61
0.76
0.70
0.48
0.48
0.61
0.68
Lenke 3 0.09
0.09
0
0
0.11
0.12
0.09
0.09
0.12
0.28
0.28
0.07
0.07
Follow-up ≥2 yr
≥2 yr
≥2 yr
≥2 yr
6.7 (4.5–8.5) y
6.7 (4.5–8.5) yr
≥2 yr
≥2 yr
≥2 yr
2m
2m
2 yr
2 yr
≥5 yr
≥5 yr
Category of Hybrid Construct H
H+ W
H
H+ W
H
H+ W
H+ W
Rod Stiffness 5.5Ti
5.5Ti
5.5Ti
5.5Ti
Rod Rotation Y
Y
N
N
Y
Y
N
Y
Y
Y
Y
Y
Y
N
N
Y
Y
N
N
N
N
N
N
N
N
N
N
N
N
N
In situ Bending
Retro. indicates retrospective study; PS, pedicle screw; Prosp. prospective study; VCA, vertebral coplanar alignment; SR, simple rotation; C-T, conventional thoracoplasty; IP, immediate postoperative follow-up; N-T, no thoracoplasty; CSLRR, convex short length rib resection; DVD: direct vertebral derotation; HC, hybrid construct group; UC, universal clamp; MONO, monoaxial screw; POLY, polyaxial screw.
Retro.
Kim et al3
Study ID
Retro.
Study Type
Hwang et al24
Group
Spine
Number
TABLE 1. (Continued )
DEFORMITY Pedicle Screw Versus Hybrid Construct Instrumentation in Adolescent Idiopathic Scoliosis • Cao et al
Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
www.spinejournal.com
E803
08/05/14 3:26 PM
DEFORMITY
Pedicle Screw Versus Hybrid Construct Instrumentation in Adolescent Idiopathic Scoliosis • Cao et al
listed in Table 2. Most studies measured thoracic kyphosis as T5–T12 Cobb angle in the lateral radiograph. Only 2 studies were claimed as prospectively designed series and the rest 22 were retrospectively series. So, all of the studies were considered class III evidence. We chose the following factors as candidate potential sources of heterogeneity according to clinical and practical consideration: (1) Risser grade; (2) flexibility of thoracic curve; (3) fused level; (4) curve type, which we calculated as the percentage of Lenke 1 and Lenke 3 curve; and (5) sagittal balance, which was usually measured as C7 plumb line’s translation from S1 vertical line in the lateral radiograph and was expressed in millimeter. The results showed in Table 1.
RESULT OF META-ANALYSIS Meta-analysis of all Groups For pedicle screw group, we got standard mean difference (SMD) as 0.40 (95% confidence interval [CI], 0.31–0.50). SMD equals to mean difference of postoperative and preoperative thoracic kyphosis divided by standard deviation. I2= 91.4%. For hybrid construct group, we got SMD as 0.15 (95% CI, 0.04–0.26). Both intervals of SMD lay in positive
Study ID
value side. Overall SMD was 0.30 (95% CI, 0.23–0.37). The positive value of SMD means that thoracic kyphosis improves after surgery, whereas negative value means kyphosis loss after surgery. Details for each study are shown in Figure 1.
Meta-analysis of Paired Groups We chose postoperative thoracic kyphosis of hybrid construct group as control and compared it with pedicle screw group. In this meta-analysis, we got SMD as −0.46 (95% CI, −0.64 to −0.27). The negative values mean that hybrid construct has more power to restore thoracic kyphosis than pedicle screw. This result is shown in Figure 2.
Publication Bias For publication bias, we used Begg’s funnel plot to show. For pedicle screw group, there were 10 of 23 out of the funnel. For hybrid construct group, there were 6 of 13 out of the funnel. Figure 3 shows the result.
RESULT OF METAREGRESSION To explore the source of heterogeneity, we underwent metaregression of several chosen candidate potential sources. Flexibility of thoracic curve was measured by Cobb angle in
%
Pedicle screw group Hamzaogln et al9 Kazi 2011 Suk et al11 potter et al12 Vallespir et al13 Clement et al14 Fu et al15 Qiu et al16 VCA zhou 2011 C-T Basu et al18 Mladenov et al19 DVR Liljenqvist et al20 Hwang et al24 screw Kim et al3screw Lowerstein et al8 screw Lonner et al25 MONO Silvestre et al26 screw Crawford et al27 screw Yilmaz et al28 screw Qui et al16 SR zhou 2011 N-T zhou 2011 CSLRR Mladenon et al19 SR Lonner et al20 POLY Subtotal (I 2 = 91.4%, P = 0.000) Hybrid construct group Blondel et al21 Sucato et al7 Chayamong et al22 Krallis et al1 Ilharreburde et al23 clamp Ilharreburde et al23 hook Hwang et al24 hybrid Kim et al3 hybrid Lowerstein et al6 hybrid Lonner et al25 hybrid Silvestre et al26 hybrid Crawford et al27 hybrid Yilmaz et al28 hybrid Subtotal (I 2 = 80.7%, P = 0.000)
SMD (95% CI)
Weight
1.39 (1.06–1.72)
4.88
1.34 (1.05–1.62)
6.45
0.52 (0.32–0.72)
13.38
–1.66 (-2.38 to 0.93)
1.00
0.06 (-0.50–0.61)
1.70
1.24 (0.86–1.63)
3.53
–0.15 (-0.70–0.41)
1.70
0.79 (0.20–1.38)
1.51
0.64 (0.00–1.27)
1.29
–1.25 (-2.07 to 0.44)
0.79
–1.02 (-1.79 to 0.26)
0.90
0.04 (-0.61–0.70)
1.23
–0.45 (-1.26–0.36)
0.80
–0.70 (-1.23 to 0.17)
1.86
–0.90 (-1.61 to 0.19)
1.05
–0.57 (-1.06 to 0.09)
2.22
–0.57 (-1.13 to 0.00)
1.64
0.95 (0.45–1.46)
2.07
–0.97 (-1.46 to 0.47)
2.13
0.30 (-0.27–0.87)
1.62
0.60 (0.02–1.17)
1.56
1.02 (0.44–1.60)
1.56
0.31 (-0.41–1.03)
1.01
0.00 (-0.48–0.48)
2.25
0.40 (0.31–0.50)
58.13
0.86 (0.49–1.24)
3.74
0.04 (-0.26–0.33)
5.86
-0.09 (-0.48–0.29)
3.54
0.64 (0.21–1.07)
2.79
0.75 (0.42–1.08)
4.77
-0.12 (-0.44–0.20)
5.10
0.35 (0.01–0.68)
4.77
-0.16 (-0.67–0.36)
1.97
-0.35 (-1.03–0.33)
1.14
-0.57 (-1.07 to 0.08)
2.16
-0.46 (-1.00–0.08)
1.79
-0.57 (-1.10 to 0.04)
1.90
0.32 (-0.15–0.79)
2.35
0.15 (0.04–0.26)
41.87
0.30 (0.23–0.37)
100.00
Heterogeneity between groups: P = 0.001
Overall (I 2 = 89.4%, P = 0.000) -1
Kyphosis loss E804
0
1
Kyphosis restore
Figure 1. Meta-analysis of thoracic kyphosis in adolescent idiopathic scoliosis with pedicle screw or hybrid construct instrumentation. The positive values of SMD represent kyphosis restoration after surgery and negative values of SMD represent kyphosis loss after surgery. SMD indicates standard mean difference; VCA, vertebral coplanar alignment; C-T, conventional thoracoplasty; DVR, direct vertebral rotation; MONO, monoaxial screw; SR, simple rotation; N-T, no thoracoplasty; CSLRR, convex short length rib resection; POLY, polyaxial screw.
www.spinejournal.com
June 2014
Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. SPINE131442_LR E804
08/05/14 3:26 PM
DEFORMITY
Pedicle Screw Versus Hybrid Construct Instrumentation in Adolescent Idiopathic Scoliosis • Cao et al
Study
%
ID
SM D ( 95 %
Hwang et al24 3
Kim et al
6
Lowenstein et al Lonner et al
25 26
Silvestre et al
27
Crawford et al
CI)
28
Yilmaz et al
Overall (I = 72.7%, P = 0.001) 2
-1
Favors hybrid construct
0
Weight
-0.41 (-0.77 to 0.06)
27.53
-1.04 (-1.59 to 0.49)
11.42
-0.32 (-1.00–0.35)
7.53
-0.14 (-0.62–0.34)
15.00
-0.29 (-0.84–0.26)
11.53
0.14 (-0.35–0.64)
14.02
-1.28 (-1.80 to 0.77)
12.97
-0.46 (-0.64 to 0.27)
100.00
- 0 . 4 1 ( - 0. 77 ,
- 0 . 0 6)
- 1 . 0 4 ( - 1. 59 ,
- 0 . 4 9)
- 0 . 3 2 ( - 1. 00 ,
0. 35 )
- 0 . 1 4 ( - 0. 62 ,
0. 34 )
- 0 . 2 9 ( - 0. 84 ,
0. 26 )
0. 14 ( - 0 . 3 5,
%
SMD (95% CI)
0 . 6 4)
- 1 . 2 8 ( - 1. 80 ,
- 0 . 7 7)
- 0 . 4 6 ( - 0. 64 ,
- 0 . 2 7)
1
Favors pedicle screw
Figure 2. Meta-analysis of thoracic kyphosis in studies including both PS group and hybrid group. The positive value of SMD representing pedicle screw has more power to restore kyphosis, and negative value of SMD representing hybrid construct has more power to restore kyphosis. SMD indicates standard mean difference.
middle and side-bending anterior-posterior radiograph. Sagittal balance was usually measured as distance between C7 plumb line and vertical line of S1 superior posterior edge. Risser grade was defined as 0 to 5 grades according to the
SUBGROUP ANALYSIS Category of Hybrid Construct There were 3 main categories of hybrid construct, that is, hook only fashion, hook plus wires fashion, and universal clamps fashion. According to this situation, we stratified studies to 3 subgroups, respectively. Charts for subgroup analysis on category of hybrid construct are shown in Figure 4A.
Follow-up In included studies, length of follow-up did vary because of different purposes of studies and various realities. To investigate the maintenance of sagittal correction, we stratified studies to 3 subgroups: less than 2 years, a minimum followup of 2 years, and a minimum follow-up of 5 years. Charts for subgroup analysis on length of follow-up are shown in Figure 4B (hybrid construct) and Figure 4C (pedicle screw).
Begg's funnel plot with pseudo 95% confidence limits 2
1
SMD
percentage of bilateral iliac epiphyseal closed surface to assess the maturity of skeletal. Percentage of curve type specified by Lenke classification as type1 (main thoracic curve) and type 3(double main curve) was recorded. We took the SMD of thoracic kyphosis as effect variable and candidate potential sources of heterogeneity (CPSofHG) as assumed dependent variables. Using the software Stata 12.0 for Windows (StataCorp LP, College Station, TX), we got the coefficient of these CPSofHG. As shown in Table 3, all of the coefficient 95% CIs lined on both side of 0 values, P value for test was greater than 0.05 each. According to statistical knowledge, only when the whole interval lines on 1 side of 0 value and P value is less than 0.05, we can say that the factor contributes to the heterogeneity of effect.
0
In situ Bending In situ bending was supposed to directly improve the sagittal curve. Some surgical procedures preferred to use this technique when they considered necessary whereas other surgical procedures found no interest in this technique and chose not to apply it in practice. According to this situation, we stratified studies to 2 subgroups: In situ bending used and in situ bending not used groups. Charts for subgroup analysis on length of follow-up are shown in Figure 4D (hybrid construct) and Figure 4E (pedicle screw).
-1
-2 0
.1
.2 s.e. of: SMD
.3
.4
Begg's funnel plot with pseudo 95% confidence limits 1
Rod Rotation Maneuver
SMD
.5
0
-.5 0
.2 s.e. of: SMD
.4
Figure 3. Begg’s funnel plot of report bias. Top graph represents studies including pedicle screw group, and bottom graph represents studies including hybrid construct group. SMD indicates standard mean difference. Spine
Rod rotation maneuver is one of the basic curve correction methods. But in some studies, they did not use this maneuver. Instead, they used other maneuvers, such as translation technique, vertebral coplanar alignment, and so on. According to this situation, we stratified studies to 2 subgroups: Rod rotation maneuver used and rod rotation maneuver not used groups. Charts for subgroup analysis on lord rotation maneuver are shown in Figure 4F (hybrid construct) and Figure 4G (pedicle screw).
Rod Stiffness Different rod materials and diameters were supposed to have influence on correction of curves. In some studies, they www.spinejournal.com
E805
Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. SPINE131442_LR E805
08/05/14 3:26 PM
DEFORMITY
Pedicle Screw Versus Hybrid Construct Instrumentation in Adolescent Idiopathic Scoliosis • Cao et al
TABLE 2. Pre/Postoperative TK of Studies TK Preoperative
TK Postoperative
Group
Mean
SD
Mean
SD
Correction Index (%)
Hamzaoglu et al9
PS
18
5.5
26
6
44.4
Abul-Kasim et al
PS
15.1
8.1
24.8
6.3
64.2
Suk et al
PS
18
11
23
8
27.8
Study ID 10
11
Potter et al
PS
29.4
2.8
25
2.5
−15.0
13
Vallespir et al
PS
18
13.5
18.6
7.3
3.33
14
Clement et al
PS
24
12
24
11
0
Fu et al
PS
21
13.7
34.4
6.7
63.8
PS (VCA)
15.8
10.7
14.5
6.5
−8.2
PS (SR)
18.3
10.5
27.7
13.1
51.4
PS (C-T)
15.5
12.6
19.2
12.4
23.9
PS (N-T)
21.2
8
26.3
8
24.1
PS (CSLRR)
20.1
7.6
24.6
7.5
22.4
PS
20.4
7.8
28.8
8.6
41.2
PS (DVD)
24.38
12.12
12.17
6.55
−50.1
PS (SR)
23.4
9
15.2
6.9
−35.0
Liljenqvist et al
PS
13.9
9.7
16.7
8.3
20.1
Blondel et al
HC
29.6
40
31
23
4.7
HC
18.7
8
24.8
6
32.6
Chaiyamongkol et al
HC
19.3
12.7
19.7
9.5
2.1
Krallis et al
HC
1.7
34.6
−1.3
29.1
−176.5
HC (UC)
24
13
30.7
7.1
27.9
HC (hook)
23.8
14.2
32.3
7.3
35.7
PS
23.6
13.8
22.1
10.7
−6.4
HC
21.7
12.9
16.1
11.8
−25.8
PS
26.3
14.1
30.4
9.1
15.6
HC
26
14
18
8
−30.8
PS
30
14
28
11
−6.7
HC
29.65
13.79
19.35
8.45
−34.7
PS (MONO)
26.24
11.22
22.42
10.47
−14.6
PS (POLY)
27
13
21
7
−22.2
HC
27
13
21
7
−22.2
PS
35.32
19.5
26
12.5
−26.4
HC
35.22
15.25
29.3
10.25
−16.8
PS
17.5
3.5
20.5
2.75
17.1
HC
22
3.75
20
3.25
−9.1
PS
22.1
13.2
11.5
8.1
−48.0
HC
18.4
13.2
21.9
8.1
19.0
12
15
Qiu et al16
Chunguang et al
17
Basu et al
18
Mladenov et al19 20
21
Sucato et al
7 22
1
Ilharreborde et al23 Hwang et al24 Kim et al3 Lowenstein et al6
Lonner et al25
Silvestre et al26 Crawford et al27 Yilmaz et al28
TK indicates thoracic kyphosis; PS, pedicle screw group; VCA, vertebral coplanar alignment; SR, simple rotation; C-T, conventional thoracoplasty; N-T, no thoracoplasty; CSLRR, convex short length rib resection; DVD, direct vertebral derotation; HC, hybrid construct group; UC, universal clamp; MONO, monoaxial screw; POLY, polyaxial screw.
E806
www.spinejournal.com
June 2014
Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. SPINE131442_LR E806
08/05/14 3:26 PM
DEFORMITY
Pedicle Screw Versus Hybrid Construct Instrumentation in Adolescent Idiopathic Scoliosis • Cao et al
TABLE 3. Metaregression of CPSofHG* Coefficient (95% CI)
P⬎t
Flex.
1.71 (−1.72 to 5.15)
0.309
SB
0.22 (−0.02 to 0.02)
0.897
RG
0.22 (−0.78 to 1.23)
0.631
FL
0.10 (−0.15 to 0.35)
0.406
Lenke 1
−0.44 (−2.15 to 1.26)
0.597
Lenke 3
−0.16 (−3.14 to 3.08)
0.919
CPSofHG
*P < 0.05 is considered as the factor that contributes to the heterogeneity of effect. If coefficient interval goes across both negative and positive values, it means that no evidence supports that the factor contributes significantly to the heterogeneity of effect. CPSofHG indicates candidate potential sources of heterogeneity; CI, confidence interval; Flex., flexibility of thoracic curve; SB, sagittal balance; RG, Risser grade; FL, fused level.
did claim clearly that all included patients got same kind of rod, mainly 5.5-mm titanium alloy rod. We noticed that other kinds of rod were also used in included studies, such as 6.0-mm titanium alloy rod and 6.35-mm stainless steel. In included studies, no studies claimed that they used CoCr alloy rod alone in included patients’ operation. Only 1 latest PS
published article reported its result of CoCr used in vivo.29 Because it was not published when we did our data search earlier, we simply used it as a reference for sensitive test. A number of studies did not verify the rod material or diameters information. According to this situation, we stratified studies to 4 subgroups: 5.5-mm titanium alloy rod, 6.0-mm titanium alloy rod, 6.35-mm stainless steel, and unidentified groups. Charts for subgroup analysis on rod stiffness are shown in Figure 4H (hybrid construct) and Figure 4I (pedicle screw). All the results and details are shown in Table 4.
DISCUSSION In total, 24 articles were included in the analysis. Seven were retrospectively cohort and 2 were prospective series. Overall, the evidence class of article was class III, which is not robust. To avoid introducing more reporting bias, for articles reported by the same institution, we chose only one of them by taking the publishing time and integrity of data into consideration. For quantitative analysis, for both groups, overall effect had positive intervals that showed kyphosis restoration after surgery. For the 7 group-paired studies, we underwent directly comparing the result of postoperative thoracic kyphosis and got a result of comparison between 2 subgroups, which was
SMD (95% CI)
PS
1.39 (1.06 - 1.72)
SMD (95% CI)
-1.66 (-2.38 to -0.93)
1.39 (1.06 - 1.72) 0.52 (0.32 - 0.72) -1.66 (-2.38 to -0.93) 0.06 (-0.50 to 0.61) 1.24 (0.86 - 1.63) -0.15 (-0.70 to 0.41) 0.79 (0.20 - 1.38) 0.64 (0.00 - 1.27) -1.25 (-2.07 to -0.44) -1.02 (-1.79 to -0.26) 0.04 (-0.61 to 0.70) -0.45 (-1.26 to 0.36) -0.70 (-1.23 to -0.17) -0.90 (-1.61 to -0.19) -0.57 (-1.06 to -0.09) -0.57 (-1.13 to -0.00) 0.95 (0.45 - 1.46) 0.30 (-0.27 to 0.87) 0.60 (0.02 - 1.17) 1.02 (0.44 - 1.60) 0.31 (-0.41 to 1.03) 0.00 (-0.48 to 0.48) 0.34 (0.24 - 0.45)
-0.15 (-0.70 to 0.41) 0.64 (0.00 - 1.27)
HC
HC
SMD (95% CI)
-1.02 (-1.79 to -0.26)
SMD (95% CI)
0.04 (-0.61 to 0.70)
HC
-0.70 (-1.23 to -0.17) -0.57 (-1.06 to -0.09)
0.04 (-0.26 to 0.33) -0.09 (-0.48 to 0.29) 0.64 (0.21 - 1.07) 0.75 (0.42 - 1.08) -0.12 (-0.44 to 0.20) -0.16 (-0.67 to 0.36) -0.57 (-1.07 to -0.08) -0.57 (-1.10 to -0.04) 0.32 (-0.15 to 0.79) 0.09 (-0.04 to 0.22)
0.86 (0.49 - 1.24) 0.75 (0.42 - 1.08) 0.80 (0.55 - 1.05)
UC group 0.04 (-0.26 to 0.33) -0.09 (-0.48 to 0.29) 0.64 (0.21 - 1.07) 0.35 (0.01 - 0.68) -0.16 (-0.67 to 0.36) -0.57 (-1.07 to -0.08) -0.57 (-1.10 to -0.04) 0.03 (-0.12 to 0.18)
0.60 (0.02 - 1.17) 1.02 (0.44 - 1.60) 0.31 (-0.41 to 1.03) 0.00 (-0.48 to 0.48) 0.20 (0.06 - 0.34)
a minimun of 2 yr 1.34 (1.05 - 1.62) 0.06 (-0.50 to 0.61) 1.24 (0.86 - 1.63) -1.25 (-2.07 to -0.44) 0.82 (0.64 - 1.01)
less than 2 yr 0.52 (0.32 - 0.72)
0.35 (0.01 - 0.68) -0.46 (-1.00 to 0.08) 0.13 (-0.16 to 0.41)
B
0.15 (0.04 - 0.26)
-0.09 (-0.48 to 0.29) 0.64 (0.21 - 1.07) 0.75 (0.42 - 1.08) -0.12 (-0.44 to 0.20) 0.32 (-0.15 to 0.79) 0.28 (0.11 - 0.45)
-0.90 (-1.61 to -0.19)
-0.45 (-1.26 to 0.36) -0.57 (-1.13 to -0.00) 0.36 (0.18 - 0.54)
a minimun of 5 yr
0.15 (0.04 - 0.26)
In situ bending (-)
0.79 (0.20 - 1.38)
less than 2 yr
Hook only
0.86 (0.49 - 1.24) 0.04 (-0.26 to 0.33) 0.35 (0.01 - 0.68) -0.16 (-0.67 to 0.36) -0.35 (-1.03 to 0.33) -0.57 (-1.07 to -0.08) -0.46 (-1.00 to 0.08) -0.57 (-1.10 to -0.04) 0.05 (-0.10 to 0.20)
0.30 (-0.27 to 0.87)
0.86 (0.49 - 1.24) -0.35 (-1.03 to 0.33) 0.58 (0.25 - 0.91)
-0.12 (-0.44 to 0.20) -0.35 (-1.03 to 0.33) -0.46 (-1.00 to 0.08) 0.32 (-0.15 to 0.79) -0.10 (-0.33 to 0.12)
A
-0.97 (-1.46 to -0.47)
a minimun of 2 yr
Hook + Wire
SMD (95% CI)
0.95 (0.45 - 1.46)
0.40 (0.31 - 0.50)
D
In situ bending (+)
0.15 (0.04 - 0.26)
PS PS
1.34 (1.05 - 1.62) -0.97 (-1.46 to -0.47) 0.77 (0.52 - 1.01)
In situ bending (+)
a minimun of 5 yr
C
In situ bending (-)
E
0.40 (0.31 - 0.50)
SMD (95% CI)
SMD (95% CI) 1.39 (1.06 - 1.72) 0.52 (0.32 - 0.72)
HC
1.39 (1.06 - 1.72) 0.06 (-0.50 to 0.61) 1.24 (0.86 - 1.63) 0.79 (0.20 - 1.38) -1.25 (-2.07 to -0.44) 0.04 (-0.61 to 0.70) -0.45 (-1.26 to 0.36) 0.95 (0.45 - 1.46) 0.80 (0.62 - 0.97)
SMD (95% CI)
0.86 (0.49 - 1.24) 0.04 (-0.26 to 0.33) -0.09 (-0.48 to 0.29) 0.75 (0.42 - 1.08) -0.12 (-0.44 to 0.20) 0.35 (0.01 - 0.68) -0.57 (-1.07 to -0.08) -0.57 (-1.10 to -0.04) 0.17 (0.04 - 0.30)
Rod rotation (+)
G
0.64 (0.00 - 1.27) 0.04 (-0.61 to 0.70) -0.45 (-1.26 to 0.36) -0.90 (-1.61 to -0.19) -0.57 (-1.06 to -0.09) -0.57 (-1.13 to -0.00) 0.95 (0.45 - 1.46) 0.60 (0.02 - 1.17) 1.02 (0.44, 1.60) 0.00 (-0.48 to 0.48) 0.35 (0.23 - 0.47)
unidentified 1.34 (1.05 - 1.62)
unidentified
-0.15 (-0.70 to 0.41)
0.86 (0.49 - 1.24) -0.16 (-0.67 to 0.36) 0.32 (-0.15 to 0.79) 0.45 (0.20 - 0.71)
-1.02 (-1.79 to -0.26)
0.79 (0.20 - 1.38) -0.70 (-1.23 to -0.17) -0.97 (-1.46 to -0.47) 0.30 (-0.27 to 0.87) 0.31 (-0.40 to 1.03)
5.5 mm Ti alloy rod
0.35 (0.18 - 0.53)
5.5 mm Ti alloy rod
-0.09 (-0.48 to 0.29) -0.09 (-0.48 to 0.29)
1.24 (0.86 - 1.63) 1.24 (0.86 - 1.63)
6.35 mm stainless steel rod
Rod rotation (+) 0.40 (0.31 - 0.50)
0.64 (0.00 to 1.27)
0.04 (-0.26 to 0.33) 0.64 (0.21 - 1.07) 0.75 (0.42 - 1.08) -0.12 (-0.44 to 0.20) 0.35 (0.01 - 0.68) -0.35 (-1.03 to 0.33) -0.57 (-1.07 to -0.08) -0.46 (-1.00 to 0.08) -0.57 (-1.10 to -0.04) 0.10 (-0.03 to 0.23)
1.34 (1.05 - 1.62) 0.52 (0.32 - 0.72) -1.66 (-2.38 to -0.93) -0.15 (-0.70 to 0.41) 0.64 (0.00 - 1.27) -1.02 (-1.79 to -0.26) -0.70 (-1.23 to -0.17) -0.90 (-1.61 to -0.19) -0.57 (-1.06 to -0.09) -0.57 (-1.13 to -0.00) -0.97 (-1.46 to -0.47) 0.30 (-0.27 to 0.87) 0.60 (0.02 - 1.17) 1.02 (0.44 - 1.60) 0.31 (-0.41 to 1.03) 0.00 (-0.48 to 0.48) 0.25 (0.14 to 0.36)
0.64 (0.21 - 1.07) -0.16 (-0.67 to 0.36) -0.35 (-1.03 to 0.33) -0.46 (-1.00 to 0.08) 0.32 (-0.15 to 0.79) 0.10 (-0.13 to 0.33)
F
0.06 (-0.50 to 0.61)
Rod rotation (-)
Rod rotation (-)
0.15 (0.04 - 0.26)
-1.66 (-2.38 to -0.93)
HC
SMD (95% CI)
H
0.15 (0.04 - 0.26)
6.0 mm Ti alloy rod
I
0.40 (0.31 - 0.50)
Figure 4. A–I, Subgroup analysis on category of hybrid construct, length of follow-up, rod rotation, in situ bending, and rod stiffness. PS, pedicle screw; HC, hybrid construct. Spine
www.spinejournal.com
E807
Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. SPINE131442_LR E807
08/05/14 3:26 PM
DEFORMITY
Pedicle Screw Versus Hybrid Construct Instrumentation in Adolescent Idiopathic Scoliosis • Cao et al
TABLE 4. Subgroup Analysis of TK Hybrid Construct
Pedicle Screw
SMD
95% CI
SMD
95% CI