Mismatch between rod bending and actual post-operative lordosis in short lumbar arthrodesis with poly axial screws

1

Keywords:

• Rod bending/contouring
• mismatch
• lordosis
• poly axial screw
• mono axial screw

1. Introduction

Restoring sagittal balance is a common goal in modern spine surgery (Le Huec and Roussouly 2011). To do so, the adequate lordosis need to be restored between the fused vertebrae. In degenerative posterior spine surgery, screws with poly axial head are connected to a rod. The poly axiality helps the surgeon to connect every screw head to the rod. In most cases the rod is bended barehand approximatively during surgery. The role of posterior instrumentation during spinal surgery remains to be defined and the question of rod bending is essential.

With the use of poly axial screws and a rod bended with the exact lordosis needed between the vertebrae, two situations can occur: a perfect match with the adequate lordosis restored between the vertebrae or a total mismatch between rod lordosis and lumbar lordosis. There is no study on the link between rod bending and lordosis.

The purpose of this study was to evaluate parameters that explain the mismatch between lumbar lordosis and rod bending in short lumbar surgery using poly axial screws.

2. Methods

This study was monocentric, retrospective, descriptive and analytic. All patients with posterior L3L5 fusion in a university-affiliated hospital in 2017 were included. Patients with past surgical history of anterior fusion on the levels L3L5, Coronal malalignment with a Cobb angle superior to 5°, the use of dynamic fixation systems were excluded.

To define the match or the mismatch we developed a quantitative variable called: diffL. It results on the difference between the rod lordosis planed, bended during surgery barehand in most cases and the actual spine lordosis of the fused vertebrae. The greater the difference(diffL), the greater the mismatch.

We measured on immediate post-operative standing profile x-ray sagittal parameters known to be important to minimise adjacent strain due to bad sagittal balance restauration (Barrey et al 2013): pelvic incidency, lumbar lordosis, lordosis of the instrumented segment.

But those will not be enough to explain the diffL and three yet undescribed parameters seems to have a clinical relevance and an important role (Figure 1):

Figure 1. Three new parameters studied: ThetaMA, LambdaMA, EcarT.

• the standard deviation of the distances between posterior wall and rod (EcarT) which reflects how homogeneously screws are put in depth. Indeed, if all the upper screws are just barely buried and the lowest are completely inside the vertebrae that will allow a straight spine with no lordosis despite a rod with a significant lordosis

• the angle between screw and superior endplate (lambdaMA). This parameter represents also a link between the instrumentation and the vertebrae. The variation of this parameter must affect the diffL.

• the angle between the screw body and rod (thetaMA) which reflect the poly axiality allowed by the head screw.

Univariate and multivariate analysis were conducted to see if there was a link between all those parameters and the mismatch: vertebral lordosis-rod lordosis (diffL).

3. Results and discussion

74 patients were included, mean age was 67. 18 were 360° fusion and 56 were postero-lateral fusions Demographic variables had no statistically significant effects on the difference between the rod and the vertebral lordosis. Lumbar Roussouly classification, pelvic parameters and the lordosis severity did not show a significant correlation with the outcome.

Statistically significant correlation was found between the fact that screws are put with a homogeneous depth (EcarT) and the diffL (R = 0.47 [0.27, 0.63] p < 0.0001)). Similarly, the diffL is correlated with the positioning of the screws parallel to the superior vertebral endplate (Lambda MA) The correlation between ThetaMA and the diffL was also significantly and negatively correlated with the diffL (R=−0.46 [0.26, 0.62] p < 0, 0001).

We integrated these three indices in a multivariate regression model summarized in Table 1, to create the MAI index. ThetaMA, LambdaMA and EcarT are used as explanatory variables while the dependent variable is the diffL. We get the following index:

Table 1. Multivariate regression model with diffL as dependent variable and lambdaMA,TetaMA, EcarT as explanatory variables.

Variables	Coefficient [CI95%]	p-value
Intercept	40.1 [7.88;72.27]	0.02
TetaMA	−0.46 [−0.82; −0.09]	0.01
LambdaMA	0.71 [0.11;1.31	0.02
EcarT	1.16 [0.31;1.99]	0.008

MAI = 40.1–0.46* ThetaMA + 0.71*LambdaMA + 1.16*EcarT

MAI index gives an improvement on the prognostic of the severity of the diffL. MAI index yields to an AUC of 0.84 [0.75, 0.93] for predicting a diffL greater than 5°. The estimated diagnosis cut point of the MAI index for predicting a diffL greater than 5 is 6.

4. Conclusions

Our study is the first on the link between rod bending and lumbar lordosis. Three new radiologic factors were highlighted and are involved in not obtaining the planned lordosis in short lumbar fusion with poly axial screws.

The Figure 2 is an illustration of our results. In the picture a) we can see a wide mismatch with an almost straight spine despite a rod with a significant lordosis. Then if we apply the three parameters we highlighted by mobilising the vertebrae to have screws parallel to superior endplate, homogeneous depth in screws and body screws with an angle of 90 with the rod, we obtain the image b) where the lumbar lordosis is restored and match the rod lordosis.

Figure 2. Illustration case with a wide mismatch in a) and a proper TetaMA, lambraMA, EcarT, to obtain a perfect match in b).

These three factors explain the difference between rod lordosis and lumbar lordosis.

Two factors depend on the way the surgeon positions screw:

• parallel to the superior vertebral endplate(lambdaMA),

• with a homogeneous depth (EcarT).

And the last factor: ThetaMA is depending on the surgical technique (compression on screws, osteotomies, monoaxial screws, use of interbody devices) (Barrey and Darnis 2015).

The aim of the study is not to show any superiority of mono axial (where thetaMA always =90) vs poly axial screws nor to give the only way arthrodesis should be performed, but it gives a better understanding of the link between screw positioning, rod contouring, osteotomies and sagittal curves.