Surgical treatment of scoliosis in neurofibromatosis type I: A retrospective study on posterior-only correction with third-generation instrumentation.

BACKGROUND: Scoliosis is the most common orthopedic complication of neurofibromatosis type I. Scoliosis can be occurred with two patterns: dystrophic or idiopathic-like. In adolescence, in consideration of bone dystrophy, osteopenia, and often associated hyperkyphosis, most of the authors recommend an anterior-posterior approach. According to other authors, modern instrumentations could be sufficient to sustain a solid posterior arthrodesis.
MATERIALS AND METHODS: Ten patients were diagnosed with scoliosis in neurofibromatosis type I aged between 8 and 25 years, Cobb angle of the thoracic curve >45°, and minimum follow-up (FU) of 1 year and treated with posterior-only approach with third-generation high-density instrumentations. Radiographic measurements were performed on the coronal and sagittal planes. Nonparametric tests (Friedman test and Wilcoxon test) were applied to evaluate the reducibility of the preoperative curve (T0), the postoperative surgical correction (T1), and its maintenance on FU.
RESULTS: Statistics showed results compared to those evaluated in the literature with a combined approach regarding surgical correction and its maintenance on FU. On T1, a median correction of 53.5% of the scoliotic curve and of 33.7% of the thoracic hyperkyphosis was observed. On FU, the correction was maintained. A global improvement in balance was appreciated. The curves, despite rigid, showed a relative reducibility to bending tests and traction. No significant complications occurred.
CONCLUSIONS: The posterior-only approach produces a satisfactory correction of the dystrophic neurofibromatosis scoliosis if associated with the use of high-density third-generation instrumentations. We are confident in recommending posterior-only approach in dystrophic neurofibromatosis scoliosis with coronal curves till 110° and coexisting thoracic kyphosis till 80°. Copyright:

INTRODUCTION

Neurofibromatosis type I (NF-I) is an autosomal dominant genetic disorder (mutation of the NF-1 encoding the neurofibromin protein).[1234] Scoliosis is the most common orthopedic complication of NF-I which is present in 10%–55% of the cases who occurred with two patterns, dystrophic or idiopathic-like, often progressively evolving.[567] Idiopathic-like curves can be treated with decision-making criteria similar to idiopathic scoliosis and needed close observation.[8] Dystrophic curves need aggressive treatment because of their inexorable evolution; the treatment with brace resulted unsuccessful.[3] This monocentric retrospective study aims to evaluate the correction of dystrophic scoliosis in NF-I obtained with posterior-only approach performing with high-density hybrid instrumentations.

MATERIALS AND METHODS

From January 2008 to December 2018, 195 consecutive scoliotic patients who underwent a surgical treatment by the same spine surgical team were retrospectively analyzed.

Ten patientes, 6 males and 4 females, were selected according to the following inclusions criteria: diagnosis of distrophic scoliosi in NF-1, age between 8 and 25 years, thoracic coronal curve Cobb angle ≥ 45°, surgical treatment with posterior-only approach with hybrid instrumentations (pedicle screws, sublaminar bands and hooks), implant density ≥ 0.70 and FU period of at least 1 year.[910] General data of cases are reported in Table 1. The median age at surgical time is 13.5 years (range: 11–23 years). The median of FU timing is 4 years (range: 2–5 years).

Table 1

General information of the 10 cases treated

Variables	Values
n	10
Median age (years), range	13.5, 11-23
Gender (male:female)	6:4
Follow-up (years), range	4, 2-5
Pattern of the curve	Dystrophic
Treatment	PSF Apical asymmetric Ponte osteotomy
Median of implant density, range	0.73, 0.70-0.86
Hypokyphosis (n)	1
Hyperkyphosis (n)	9
Normokyphosis (n)	0

PSF - Posterior spinal fusion

Dystrophic features were studied: all ten patients had vertebral scalloping. According to the classification system proposed by Li et al.,[11] on a total of 280 pedicles of the vertebrae constituting the thoracic curve, evaluated on computed tomography (CT) images with reconstruction according to the axial plane of each vertebra, the results were as follows: A: 22; B: 96; C: 127; D: 35; and E: 0. The details are shown in Table 2 and Figure 1. On the sagittal plane, wedging of vertebral bodies is present in three patients, and on the coronal plane, it is present in all ten patients. Rotation was assessed on CT images with reconstruction according to the axial plane of the apical vertebra.[12] The median of the rotation angle is equal to 40.5° (range: 28°–45°). Nine patients had a scoliotic curve with short-range angular deviation. Five patients had rib penciling. A single patient had rib dislocation with dislocation of the head of the rib in the vertebral canal. The patient was asymptomatic without neurological deficits. Resection of the head of the rib was not performed in the operative site [Figure 2]. No patient had paraspinal or intraspinal soft-tissue masses. Dystrophic elements on simple are recapped in Table 3.[3]

Table 2

Frequency distribution of vertebral pedicles

Patient	A	B	C	D	E	Total
1	1	11	10	10	0	32
2	3	12	12	3	0	30
3	2	13	7	0	0	22
4	10	10	10	0	0	30
5	3	13	13	5	0	34
6	0	4	17	7	0	28
7	1	11	12	2	0	26
8	2	8	15	5	0	30
9	0	10	10	2	0	22
10	0	4	21	1	0	26
Total	22	96	127	35	0	280

Figure 1

Bar chart on the frequency of vertebral pedicles

Figure 2

Rib dislocation with dislocation of the head of the rib in the vertebral canal on computed tomography image. (a) Axial view; (b) coronal view; (c) sagittal view

Table 3

Dystrophic elements on cases treated

Dystrophic features	Simple
Vertebral scalloping (n)	10
Rib penciling (n)	5
Vertebral pedicles
A	22
B	96
C	127
D	35
E	0
Coronal wedging of vertebral bodies (n)	10
Sagittal wedging of vertebral bodies (n)	3
Median apical rotation (°), range	40.5, 28-45
Short curve (n)	9
Rib head dislocation into the spinal canal (n)	1

Radiographic data collection was performed by two independent operators. All radiographic measurements were made on the preoperative time (T0), on immediate postoperative time (T1), and on FU.[13] All patients underwent magnetic resonance imaging and CT scan that were performed before surgery in order to evaluate the presence of spinal cord malformation.[14151617181920]

Cobb angle measurements of coronal thoracic curve and coronal balance were achieved on standing posteroanterior X-rays. Kyphosis and sagittal balance were determined on standing lateral X-rays. Curve flexibility was evaluated on the preoperative side-bending and traction anteroposterior radiographs.[21]

One patient had thoracic hypokyphosis (<25°), and nine patients had hyperkyphosis (>50°).

The patients underwent posterior high-density arthrodesis with third-generation instrumentations. Asymmetric apical Ponte osteotomies were performed in patients who had hyperkyphosis. For spinal fusion, autologous decorticated bone was used with a homologous supplement from the bone bank.

No patients were lost during FU.

Collected data are expressed as median (range). The Friedman test and Wilcoxon test were performed to show differences in angular values of coronal and sagittal thoracic curves, in values of coronal and sagittal imbalance, and in values of flexibility. The level of significance was set as P < 0.05. Statistical analysis was performed using MedCalc® software (omicX, 72 rue de la rèpublique 76140 Le-Petit-Quevilly France).

RESULTS

Radiographic results on preoperative (T0), postoperative (T1), and follow-up

Data collection are reported in Table 4.

Table 4

Radiographic results on preoperative (T0), postoperative (T1) and follow-up

	Median
	Thoracic curve (Cobb°)	Coronal imbalance (mm)	Kyphosis (Cobb°)	Sagittal imbalance (mm)
Before surgery (T0)	93 (60-111)	15 (0-38)	59.5 (24-83)	23.5 (0-72)
Bending	77.5 (57-90)	-	-	-
Traction	75.5 (40-97)	-	-	-
After surgery (T1)	45 (25-55)	19 (0-39)	40.5 (29-51)	42 (0-67)
FU	43 (27-56)	4.5 (0-21)	46 (32-57)	12.5 (0-39)

FU - Follow-up

Statistical results concerning reducibility of thoracic curves

The Wilcoxon test showed significative differences (P = 0.0020) between thoracic curve angles measured on standing posteroanterior X-rays on T0 (median: 93°, range: 60°–111°) and thoracic curve angles measured on preoperative side-bending anteroposterior radiographs (median: 77.5°, range: 57°–90°). The median percentage of bending test reducibility is 17.1% (range: 3.3%–30.5%).

The same procedure showed significative differences (P = 0.0020) between thoracic curve angles measured on standing posteroanterior X-rays on T0 and thoracic curve angles measured on preoperative traction anteroposterior radiographs (median: 75.5°, range: 40°–97°). The median percentage of the correction in the traction test is 15.9% (range: 8.3%–51.2%). Data are summarized in Table 5.

Table 5

Statistical results concerning flexibility of thoracic curves

	Thoracic curve
Wilcoxon test T0 versus bending (P)	0.0020
Wilcoxon test T0 versus traction (P)	0.0020
Bending correction (%), range	17.1, 3.3-30.5
Traction correction (%), range	15.9, 8.3-51.2

Statistical results on coronal plane

On the coronal plane, significative differences (P < 0.00001) were found, using the Friedman test, as regards Cobb angle measured on the three different times T0, T1, and FU. The Wilcoxon test was applied between T0 and T1 and between T1 and FU. A significative difference between Cobb angle on T0 and Cobb angle on T1 (median: 45°, range: 25°–55°) is demonstrated (P = 0.0020). The median correction was obtained as 53.5% (range: 21.7%–71.6%). ID on postoperative time was also calculated: the median was 0.73 (range: 0.70–0.86). There were no significative differences (P = 0.2969) between Cobb angle on time T1 and Cobb angle on time FU (median: 43°, range: 27°–56°). The median value of the correction loss at FU was 2% (range: 8%–16.3%).

The coronal imbalance was assessed using the Friedman test, which showed a significative difference (P = 0.0433) among the measurements taken on T0, T1, and FU. The Wilcoxon test was applied to evaluate the imbalance between T0 and T1, between T1 and FU, and then between T0 and FU. Statistics showed no significance (P = 0.9102) between the time T0 (median: 15 mm, range: 0–38 mm) and T1 (median: 19 mm, range: 0–39 mm). The median value concerning the variation of the imbalance between T0 and T1 was 9.5 mm (range: 0–29 mm). A significative difference (P = 0.0156) between T1 and FU (median: 4.5 mm, range: 0–21 mm) was demonstrated. The median variation of the imbalance between T1 and FU is 10 mm (range: 0–29 mm). Significative differences (P = 0.0488) between the median of the coronal imbalance on T0 and on FU were verified. The median variation of the imbalance between T0 and FU is 13.5 mm (range: 2–24 mm) [Figure 3].

Figure 3

Patient 1 (female, 14 years old). X-ray examinations on coronal plane: (a) before surgery (T0); (b) immediate postoperative time (T1); (c) 4-year follow-up

Statistical results on sagittal plane

On sagittal plane, hypokyphosis was excluded from the sample. The Friedman test was applied: there was a significative difference (P = 0.0003) in the comparison between Cobb angle, concerning thoracic kyphosis, measured in the three different times T0, T1, and FU. The Wilcoxon test was applied between T0 and T1 and between T1 and FU. The results showed significative differences (P = 0.0090) between kyphosis angle on T0 (median: 60°, range: 51°–83°) and the value on T1 (median: 41°, range: 31°–51°). The median percentage correction of thoracic hyperkyphosis on the postoperative time was 33.7% (range: 14.8%–46.1%). There were no significative differences (P = 0.3528) between Cobb angle measured on T1 and the value measured on FU (median: 47°, range: 35°–57°). The median variation on FU compared to T1 was 11.7% (range: 2%–22.5%). Concerning the case with hypokyphosis (24°), the percentage correction between T0 and T1 was 16%. The loss of correction between T1 and FU was 8%.

Sagittal imbalance was evaluated using the Friedman test, which showed a significative difference (P = 0.0047) between the measurements taken on times T0, T1, and FU. The Wilcoxon test was applied to evaluate the difference in sagittal imbalance between T0 and T1, between T1 and FU, and between T0 and FU. Statistics showed no significative differences (P = 0.6953) between imbalance on T0 (median: 23.5 mm, range: 0–72 mm) and on T1 (median: 42 mm, range 0–67 mm). The median value concerning the variation of the sagittal imbalance between T0 and T1 was 25 mm (range: 1–59 mm). Statistics showed a significative difference (P = 0.0137) between sagittal imbalance on T1 and FU (median: 12.5 mm, range: 0–39 mm). The median variation of the sagittal imbalance between T1 and FU was 31.5 mm (range: 5–87 mm). The difference between sagittal imbalance on T0 and on FU was not significative (P = 0.0547). The median variation of the sagittal imbalance between T0 and FU was equal to 17.5 mm (range: 0–111 mm) [Figure 4].

Figure 4

Patient 1 (female, 14 years old). X-ray examinations on sagittal plane: (a) before surgery (T0); (b) immediate postoperative time (T1); (c) 4-year follow-up

In Table 6, there is a recap of statistical results.

Table 6

Statistical results on coronal and sagittal planes

	Thoracic curve	Coronal imbalance	Kyphosis*	Sagittal imbalance
Friedman test T0, T1, FU (P)	<0.00001	0.0433	0.0003	0.0047
Wilcoxon test T0 versus T1 (P)	0.0020	0.9102	0.0090	0.6953
Wilcoxon test T1 versus FU (P)	0.2969	0.0156	0.3528	0.0137
Wilcoxon test T0 versus FU (P)	-	0.0488	-	0.0547

*Level of significance sets as P<0.05 hypokyphosis was excluded. FU - Follow-up

Clinical and operative results

The rate of early and late intraoperative and postoperative surgical complications was zero. A frontal pressure injury and an allergic reaction to antibiotic therapy with vancomycin occurred. There were no cases of coronal or sagittal decompensation and no neurological and infectious complications. The median of intraoperative blood losses was 1650 ml (range: 600–3974 ml). The median of the duration of the surgery is 412.5 min (range: 275–520 min).

DISCUSSION

This study intends to support the thesis of the effectiveness of posterior-only spinal fusion in dystrophic scoliotic curves.[2223242526]

In adolescence, in consideration of bone dystrophy, osteopenia, and often associated hyperkyphosis, most of the authors recommend to associate an apical anterior arthrodesis release to the posterior arthrodesis.[27282930] According to other authors, modern instrumentations could be sufficient to sustain a solid and effective posterior arthrodesis.[222324252631]

The anterior instrumented arthrodesis allows a good correction saving fusion levels. However, several complications are reported in the anterior approach, such as injury to the large vessels and adjacent organs, and reduced functionality pulmonary.[32]

The posterior instrumentation involves the use of pedicle screws, hooks, and sublaminar bands in hybrid constructions. Nowadays, such equipment are widely used for the effectiveness of the correction and the relatively low complication rate in the treatment of idiopathic scoliosis.[33]

The combined anterior-posterior approach includes the anterior release, followed by posterior instrumentation and fusion, which is in the same surgical session or deferred.

The statistical significance concerning the reducibility of the curve at the bending test and at traction suggests a flexibility that is still partially persistent with optimistic prediction when opting for correction and surgical stabilization. Despite this, the percentage of reducibility to the bending test is still to be considered, which is in any case <25%, thus decreeing these curves as rigid. The median age of the sample, equal to 13.5 years, the bone immaturity and the altered quality of the bone and soft tissues, found in the context of NF-I, could explain the relative flexibility of the dystrophic curve. The evaluation in the following study regarding the reducibility of dystrophic scoliotic curves was not taken into consideration in the studies available in the literature on the surgical treatment of deformity in neurofibromatosis. There are some scientific articles that evaluate the “flexibility” of idiopathic scoliosis.[3435]

The duration of the surgical timing and intraoperative blood losses are compatible with the greater complexity of the surgery due to the dystrophy and the stiffness of the curves compared to the surgery of idiopathic scoliosis. In addition, the high ID of the instrumentations partially justifies intraoperative bleeding. It should also be remembered that two patients had coagulation disorders on preventive hematology tests.

Invasiveness can be considered minor, due to the presence of a single access and the need for only one surgical time, which reduces the operative duration compared to a combined approach; however, the high ID could theoretically increase the risk of complications including infections, pseudoarthrosis, and bad positioning of the screws with possible vascular and nervous complications. However, no intraoperative either postoperative complications were detected in the studied sample.

On coronal plane, the effectiveness of surgical correction has been demonstrated statistically with a median percentage (53.5%) that is satisfactory, consistent with the surgical corrections evaluated in the literature. No significative differences between the measurements on the postoperative time and on the FU demonstrate the maintenance of the correction thanks to a third-generation instrumentation that has allowed an optimal stabilization.

On coronal plane, the imbalance does not seem to improve on immediate postoperative time, probably due to the severe impact of the surgical trauma, but it decreases in the FU leading to a good overall balance.

The statistical significance between the measurements of the kyphotic curves on preoperative time and those on postoperative time has demonstrated the effectiveness of the correction of hyperkyphosis in some cases even severe. This result was possible through the good control of the sagittal curves allowed by the posterior access. The data must be compared with the previous scientific literature which proposed, in the case of hyperkyphosis above 50°, the combined anterior-posterior approach.[28] However, this recommendation could not take into consideration the potential of actual instrumentations that allow, through the use of hybrid constructs, good stability thanks also to the possibility of increasing the ID. The high ID (≥0.70), in cases of pedicle dystrophy which does not allow the safe and stable positioning of the pedicle screws, is guaranteed in these curves by the use of the sublaminar bands. The patient with hypokyphosis (24° Cobb) had a flat back. In this case, the correction was performed in the opposite direction also improving in sagittal radiographic and clinical profile.

Regard to the sagittal imbalance, there is an improvement of balancing the immediate postoperative time, probably caused by severe surgical trauma. Even in the absence of statistical significance, in the long run, there is a reduction of imbalance on FU.

CONCLUSIONS

Considering the available case studies and the scientific literature found, the potential of the new generation instrumentations and posterior-only approach was evaluated, providing the real possibilities that technology can offer at the present moment.

It has been concluded that the posterior-only approach produces a satisfactory correction of the dystrophic neurofibromatosis thoracic curves if associated with the use of third-generation high-density implant instrumentations. The effectiveness is also demonstrated in cases where severe thoracic kyphosis is present, also allowing good control, as well as hypokyphosis, with a substantial improvement in all cases of the sagittal profile. The correction is stable over time.

Based on recent scientific literature and the retrospective study conducted, we are confident in recommending the posterior-only approach in dystrophic neurofibromatosis scoliosis with coronal curves till 110° and coexisting thoracic kyphosis till 80° as these high-density third-generation instruments. The implants allow to achieve results that until now were mainly achieved with the combined anterior-posterior approach, ultimately concretizing the possibility for the patient to undergo a single surgical session and therefore avoiding exposure to complications of the anterior release.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.