Introduction

The surgical treatment of severe rigid adolescent idiopathic scoliosis (AIS) is still highly debated [1,2,3]. One-stage Posterior-only spinal fusion (PSF) has historically been the mainstay of AIS treatment, however, achieving an adequate triplanar correction of this kind of curves may be challenging. Therefore, combined antero-posterior spinal fusions [1, 2], have been proposed to gain a more powerful correction, but the high rate of pulmonary complications remains cause of concern [3]. Staged corrections have therefore gained wide spread, using pre-operative halo traction, either standalone [4], or after anterior [5,6,7,8] or posterior [9] releases. These techniques, although effective, are really demanding for the patients and lead to significant health-related costs, considering the prolonged period of hospitalization and the multiple surgical procedures required. To offer a better tolerated alternative to these patients, a temporary internal distraction with magnetically controlled growing rods (MCGR) [10, 11] has been proposed; however, increased concerns about its cost-effectiveness may arise due to the addition of high hardware costs. One-staged correction using aggressive tricolumnar osteotomies (3COs) such as vertebral column resection (VCR [12,13,14] represents an effective alternative, but at a cost of high technical complexity, alongside an high complications rate.

All the above-mentioned techniques proved to be precious and effective, but each particular procedure should be applied according to the pathological anatomy of the deformity. The Authors’ hypothesis is that severe and stiff AIS with deformity spread over more than five vertebral levels should be effectively treated by a technique named Hi-PoAD (High Density Pedicle screws, Ponte Osteotomies, Asymmetric Rods Contouring, Direct Vertebral Rotation), already described for the management of rigid Adult AIS [15]. The rationale is to perform a safe and effective one-stage posterior procedure, exerting strong corrective forces by a combination of different maneuvers, via an aggressive release and high-density constructs. The aim of this article is to present the results of Hi-PoAD technique for the treatment of these curves, at minimum 2 years of follow-up, focusing on its efficacy and safety.

Materials and methods

Study sample

A retrospective review of AIS patients with a major thoracic curve (Lenke 1, 2, 3 patterns) above 90°, with a side bending (SB) reduction < 25% and with major curve spread over more than 5 vertebral bodies, who underwent surgery with Hi-PoAD technique, was undertaken. Non-idiopathic scoliosis, infantile or juvenile idiopathic scoliosis diagnosis were considered exclusion criteria. Follow-up evaluations were performed post-operatively, at 1 and 2 years and up to final follow-up. All patients signed an informed consent on the use of their clinical documentation for scientific purposes. The local Ethics Committee approved this retrospective study.

Data collection

Operative time, blood loss [both as an absolute volume (EBL) and as a percentage of total blood volume (%EBV, estimated as 75 ml multiplied by the body weight)], length of stay, intra- and post-operative complications were recorded. Coronal Cobb angle of each curve, coronal flexibility index (difference in Cobb angle of the main curve between the pre-operative full-length standing and lateral supine side-bending X-rays, expressed as a percentage), T5–T12 thoracic kyphosis (TK) and L1–S1 lumbar lordosis (LL) were measured. Apical vertebral rotation (AVR) was measured according to Nash-Moe. Total Deformity Angular Ratio (T-DAR; the sum of Coronal and Sagittal DAR, which are calculated as the Cobb angle divided by the number of vertebrae involved in the curve) was evaluated. C7 plumb line (C7PL)/central sacral vertical line (CSVL) and sagittal vertical axis (SVA) were used to assess coronal and sagittal balance. Trunk Height (TH) was measured as the distance between the midpoint of the upper endplate of T1 and the midpoint of the lower endplate of the LIV, to evaluate the spinal height. SRS-22 questionnaire was administered pre-operatively and at 1 year of follow-up.

Surgical technique

All surgeries were performed by the first author, according to the Hi-PoAD technique [15]. The fusion area was defined according to the criteria by Lenke et al. [16]. All pedicle screws were placed with the free-hand, power-assisted technique [17]. To optimize the transmission of forces from the tulip to the screw, and therefore to the vertebral body, all-uniplanar-screws constructs were preferred in all cases (Fig. 1). Differently from the original technique [15] the first Ponte osteotomy (Fig. 2) was performed at the apex of the main curve and at the four periapical levels (two above and two below the apex) in all cases, to maximize the corrective potential. In Lenke 3 cases, 5 Ponte or Smith-Petersen osteotomies were performed at thoracolumbar-lumbar curve apex if curve magnitude exceeded 90°. Two 5.5-mm cobalt-chrome rods were asymmetrically molded, to achieve a synergic action on sagittal profile restoration, with a lifting effect at the physiological apex of thoracic kyphosis and a pushing effect in the lumbar spine to restore lumbar lordosis; at the thoracic scoliosis apex, the concavity rod was higher (over-bended) than the convexity rod, to obtain an indirect derotational effect. The rods were passed through each anchor point and simultaneously rotated with the curves consensually oriented on the sagittal plane. Rotation tubes were applied at apical screws and at the LIV to exert derotational forces and counter-torque forces respectively, both in convexity and concavity. Keeping derotational forces applied, the translation maneuver was progressively performed. Before the definitive engage of the rods in the tulips, selective compression and distraction were performed at apical and periapical levels to close gaps at the Ponte osteotomy sites. Final engagement of the rods was then performed. In situ coronal or sagittal bending was then performed if needed. An extensive decortication of the posterior elements of the instrumented vertebrae was performed to obtain a solid fusion bed, followed by apposition of autologous and homologous bone. Subfascial drain was placed, and the wound was closed in multiple layers, with running intradermal suture for the skin. Exemplificative cases are presented (Figs. 3, 4, 5, 6, 7, 8). 

Fig. 1  
figure 1

Exemplificative surgical image of a high-density pedicle screw construct: T3–L2 fusion with screws placed at all pedicles, except for left T10 pedicle

Full size image
Fig. 2
figure 2

Ponte osteotomies: exemplificative drawing

Full size image
Fig. 3 
figure 3

Case 1, a 17 years old female with a Lenke 2 curve, treated by T4–L4 fusion; pre and post-operative X-rays

Full size image
Fig. 4 
figure 4

Case 1, a 17 years old female with a Lenke 2 curve, treated by T4–L4 fusion; pre and post-operative clinical appearance

Full size image
Fig. 5
figure 5

Case 3, a 13 years old female with a Lenke 3 curve, treated with a T3–L5 fusion; pre- and post-operative X-rays

Full size image
Fig. 6
figure 6

Case 3, a 13 years old female with a Lenke 3 curve, treated with a T3–L5 fusion; pre- and post-operative clinical appearance

Full size image
Fig. 7
figure 7

Case 4, a 18 years old female with a Lenke 2 curve, treated with a T3–L3 fusion; pre- and post-operative X-rays

Full size image
Fig. 8
figure 8

Case 4, a 18 years old female with a Lenke 2 curve, treated with a T3–L3 fusion; pre- and post-operative clinical appearance

Full size image

Early mobilization started on post-operative day 1. For the first 4 weeks after surgery, a thoracic lumbar sacral orthosis (TLSO) to restrict spinal movements and facilitate initial bone graft fusion was prescribed.

Patients characteristics

19 (16 females and 3 males) were included, with a follow-up of 29.2 ± 3.1 (range 28–37). The average age was 13.7 ± 1.6 (range 12–18). Patient’s characteristics are summarized in Table 1.

Table 1 Patients' characteristics
Full size table

Statistical analysis

Parametric test was used to compare samples in case of continuous variables, normal distribution and appropriate numerousness. The Shapiro–Wilk test was used to verify normal distribution. As parametric test, two-tailed student t test was used to compare the average of the variables for homoscedastic paired groups. Pearson's chi-squared test was applied to compare categorical data. Continuity correction was applied in case of discrete distribution. p values < 0.05 were considered to be significant. Furthermore, Intraclass Correlation Coefficient (ICC) and Cohen’s Kappa were used to assess intra- and inter-observer agreement in the measurement of pre-operative and last follow-up values of TK and AVR, respectively. Jamovi statistical analysis software (The jamovi project (2021), jamovi Version 1.6) was used to perform statistical analysis.

Results

Average surgical time was 263.1 ± 39.1 min, and average blood loss was 892.7 ± 137.3 (21.8 ± 3.7% EBV) (Table 1). Patients were discharged after an average length of stay of 7.9 ± 4.1 (range 5–19).

An effective triplanar correction was achieved (Table 2): the average post-operative Cobb angle decreased by 65.0% (from 101.9° ± 14.2° to 35.7° ± 15.9°, p < 0.001). On the sagittal plane, both post-operative TK and LL remained stable, passing from 24.7° ± 21.8° to 26.3° ± 15.3° (+ 6.5%; p = 0.454) and from 51.9° ± 15.3° to 51.0° ± 11.6° (− 1.7%; p = 0.710) respectively. The average AVR significantly reduced from 3.3 ± 0.7 to 1.3 ± 0.5 (− 60.6%; p < 0.001). The average C7PL/CSVL reduced with statistical significance from 1.5 cm ± 0.8 to 0.9 cm ± 0.6 (− 40.0%; p = 0.013). TH significantly increased from 31.1 cm ± 3.5 to 37.0 cm ± 3.4 (+ 19.0%; p < 0.001). At the final follow-up no significant differences were noted in the radiological results compared to post-operative values, except from a significative improvement in C7PL/CSVL (from 0.9 cm ± 0.6 to 0.6 cm ± 0.3; − 33.3% p = 0.017). Intra- and inter-observer agreement showed good agreement for TK both in pre-operative and last follow-up measurements, while AVR did record substantial agreement in the pre-operative measurements and only moderate agreement in the last FU measurements (Table 3).

Table 2 Average values of the examined variables pre-operatively, post-operatively, at 1-year, 2-years and last follow-up
Full size table
Table 3 Intra and interobserver agreement for AVR and TK in the pre-operative and last follow-up measurements
Full size table

SRS-22 questionnaire score increased in all patients, from a pre-operative value of 2.1 ± 0.5 to a mean of 3.9 ± 0.4 at 1 year of follow-up (p < 0.001).

3 patients had an intraoperative SEP and MEP reduction during the corrective maneuver. In these cases, a counter-maneuver was performed and the mean arterial pressure was raised, to enhance the recovery of the evoked potentials. Once the recovery happened, a temporary rod was placed in slight distraction on the concavity of the main curve (and on the concavity of the lumbar curve in Lenke 3 cases) to stabilize the spine in a partial corrected position and surgery was interrupted. The definitive surgery was performed 5 days later, to allow the spinal cord to get used to the new corrected spinal alignment (Figs. 9, 10). 2 of these 3 cases (66.7%) had a T-DAR > 25; conversely, among patients who had a one-stage procedure, only 1 (6.2%) had a T-DAR > 25 (p = 0.008) (Table 4). No mechanical complications (screws pull-out, pedicle breakage, necessity to extend the fusion area) were registered intraoperatively.

Fig. 9
figure 9

Case 2, a 15 years old female with a Lenke 3 curve, treated by T3–L4 fusion. The patient had a transient SEP and MEP drop during corrective maneuver, so the patient was temporarily stabilized with two titanium rods and managed with a second stage surgery after 5 days. Pre-operative and post-operative (first and second stage surgery) X-rays

Full size image
Fig. 10 
figure 10

Case 2, a 15 years old female with a Lenke 3 curve, treated by T3–L4 fusion; pre and Postoperative clinical appearance

Full size image
Table 4 Relationship between T-DAR and surgical stages required for correction
Full size table

At the last follow-up, no mechanical nor infective complications were registered; no cases of correction loss, adding on or junctional pathology were observed.

Discussion

The described technique proved to be really efficient, allowing a powerful coronal correction rate of 65.0%, which is superior compared to the results reported by others all-posterior one-stage correction studies without 3COS [18,19,20] (50.9–55.9%). This difference can be explained by the combination of corrective strategies that the described technique provides. First, an extremely high screw density with screws placed at every level permits to dissipate corrective forces, thus limiting pull-out risks. Secondly, the aggressive release based upon periapical Ponte Osteotomies enforces the translation and derotational maneuverers allowing in addition segmental correction with convex compression and concave distraction. Thirdly, the simultaneous application of the asymmetrically molded CoCr rods allows to apply a translational and derotational force over both rods at the same time, improving the overall corrective capacity. Finally, for the coupling principle, DVR and translation maneuvers act simultaneously: the DVR has a translation component that acts in synchrony with the translation performed over the rods; on the other side, the asymmetric shape of the rods exerts an indirect derotational force that combines with DVR. This combination of strategies, that come at the cost of an increased technical complexity, allowed also a good triplanar correction of the deformity, respecting a physiological sagittal profile (no significant changes in TK and LL), reducing the vertebral body rotation (significant AVR reduction) and restoring a correct coronal balance (significant C7PL-CSVL reduction). In particular, when looking at TK, our results are in line with a growing body of literature that reports a better control of kyphosis when combining pedicle screws and Ponte osteotomies [21, 22]. The addition of Asymmetric Rods Contouring to high-density screws systems, may further make possible the respect of TK even when large amounts of coronal and axial plane corrections are needed.

It may be underlined that despite the complexity of the technique average surgical time and blood loss resulted in line with the majority of the case series dealing with all posterior techniques without 3COS (263.1 min vs 180.5–300 min and 892.7 ml vs 850–1752 ml respectively [18,19,20]).

Some authors [23, 24] reported a correction rate as high as 77% for severe scoliosis curves, adopting a staged technique based upon anterior release. In fact, anterior release allows to achieve a direct control over the anterior disk restraint to DVR. On the contrary, the presented technique, as it is based on an all-posterior approach, allows to exert powerful corrective forces which may overcome the anterior disk restraints only indirectly and/or partially. Despite that, the morbidity related to the anterior approach may outweigh the benefit of the additional correction.

The correction rate achieved via the Hi-PoAD technique was comparable to that of staged procedures based on traction (both internal and external) (65.0% vs 51.3–67.4% [4,5,6,7,8, 10, 11]), but without the need of prolonged challenging traction periods, multiple procedures and high health- and hardware-related costs. The rationale of staged procedures is to obtain a gradual distraction resulting in adequate correction minimizing stresses over the spinal cord. Looking in detail at the safety profile of the techniques, Hi-PoAD reported a 15.8% rate of transient intraoperative MEP/SEPs deterioration, without any sequalae at the wakening. However, it must be acknowledged that even staged procedures are not exempt from neurologic risks, with a reported rate of 0–14.2% % [4,5,6,7,8, 10, 11], ranging from transient MEPs reduction [10], to transient [4, 6, 7] and permanent [5] motor deficit. This seem to imply that the spinal cord risks during correction may be dependent upon other factors besides the correction techniques, such as curve intrinsic characteristics. In our cohort, a T-DAR > 25 significantly correlated with the need to perform a 2-stage procedure due to SCM changes during the first correction attempt. On one hand, this seems to support the higher risks of SCM events reported by Wang et al. [25] for patients with a T-DAR > 25. On the other hand, there are certainly other factors acting that still need to be identified, since the patient with the most severe DAR (33.7) of our series still had a one-stage procedure, while one patient with a 22.3 DAR had a two-stage procedure. So, given the unpredictability and the generally transitory nature of SCM events during a cautious and progressive corrective manouver, in addition to the relatively high success rate of the procedure (84.2%), in our experience a one-stage Hi-PoAD procedure may still be suitable for these curves. Moreover, this strategy remained cost-effective even in the 3 cases where a staged procedure was required, given the fact that relatively economic hardware was used (simple titanium temporary rod vs MCGR) and that hospital stay was relatively short (13–19 days) in these 3 cases; certainly shorter compared to internal or external traction techniques [4, 6,7,8,9,10,11].

The correction rate of the presented technique was similar to that achieved by VCR studies (65.0% vs 61–67% [12,13,14]). However, the Hi-PoAD technique allowed significantly lower mean EBL (892.7 ml vs 1610–7034 ml [12,13,14]), general complications rate (0% vs 25–59% [12,13,14]) and neurological complications rate (15.8%, all transient, vs 0–27% [12,13,14]), compared to VCR. Moreover, an often-neglected issue with these patients is that they usually have a mismatch between trunk height and leg length due to the severity of the spinal deformity. The use of VCR, which is a shortening procedure, may therefore result in failure to restore a proper trunk height/leg length ratio compared the increase of TH achieved by the presented technique. For all these reasons, in the Authors’ opinion, VCR remains a fundamental weapon, but with limited indications such as: extremely sharp deformities, often with a kyphotic component and associated myelopathy. In these situations, the powerful translation between the proximal and distal vertebral sections, and the circumferential spinal cord decompression that it offers, are still unmatched by any other technique while it may be too aggressive for severe AIS curves involving more than 5 vertebral bodies.

The results of the present study must be seen in light of some important limitations. First, the definition of severe scoliosis is still unclear, with studies using as cutoff a Cobb angle > 80° [1, 5, 12], others > 90° [2, 11, 14, 23] and others > 100° [10, 24]. The definition of rigid curve is even more confusing, both in terms of cutoff (flexibility < 20% [14], 25% [12], 30% [2, 5, 19, 23], 40% [1]) and of method of assessment (standing SB [5], supine SB [5, 14, 19, 23, 24]). Moreover, traction radiographs under general anesthesia may redefine some of the included curves as non-stiff or partially stiff [26]. In this uncertain context, building solid evidence is challenging. Secondly, the retrospective design and the small sample of this study must be acknowledged. Finally, the AVR measurement according to Nash-Moe method has several limitations, especially in post-operative radiographs, in which the presence of pedicle screws can make the measurement difficult and sometimes inconsistent. In fact, while TK measurements showed good intra- and interobserver agreement (probably partially due to software aid in its measurement), the AVR measurement according to Nash-Moe did record only moderate agreement in the last follow-up X-rays. Moreover, if one considers the wide 95% CI, the intra- and interobserver agreement further decreases. The possible strengths of this study, especially considering the rarity of these cases, are the homogeneity of curves in terms of etiology (all adolescent idiopathic), topography (all main thoracic) and characteristics (number of vertebral levels involved). The present study may represent an attempt to discriminate between different severe AIS curves with the focus of developing a treatment algorithm able to highlight the indication for each corrective technique available, basing on the pathological anatomy of the deformity.

Conclusions

The Hi-PoAD technique proved to be a valid alternative for the treatment of rigid AIS curves exceeding 90° of magnitude, with less than 25% reduction on side bending, and involving more than 5 vertebral bodies.