1 Introduction

The degenerative lumbar disease consists of a group of anatomical and morphological changes that influence sagittal parameters and sagittal balance. It is characterized by disc degeneration, thickening and calcification of the ligamentum flavum, facet joint arthritis and loss of lumbar lordosis.¹ Patients often complain of low back and/or leg pain, and may present with clinical objective spinal imbalance.¹ When conservative treatment fails, decompression and fusion surgery are an option. However, fusion surgery increases spine stiffness and can modify biomechanics and sagittal parameters.¹ With degenerative changes and consequent loss of lumbar lordosis, there are some compensatory mechanisms: the pelvis compensates with retroversion (increasing the pelvic tilt), knee bending, ankle dorsiflexion and gait pattern changes.^2,3 Thus, whenever a spinal fusion is being performed, attaining spinal parameters and achieving appropriate postoperative sagittal balance should be a goal of treatment.

Rod contouring is often used to achieve adequate spinal alignment when performing spinal fusion. However, while the use of polyaxial screws facilitates rod alignment and contouring, it also allows possible mismatches between the rod lordosis and the final lordosis of the fused segment.⁴ In 2019 Moufid and Gille published a study investigating the mismatch between rod bending and post-operative lordosis and proposed three radiological factors involved in obtaining the planned lordosis in short-segment fusions using polyaxial screws. In their study, they suggested that to achieve a match between rod lordosis and segmental lordosis using polyaxial screws, screws should be positioned parallel to the superior endplate, perpendicular to the rod and to a homogenous depth.⁴ The objective of this work was to reproduce the measurements performed by Moufid and Gille⁴ in a different cohort of patients, in an attempt, to validate these findings.

2 Material and Methods

2.2

2.2 Radiological parameters

Radiological parameters were measured on standing lateral view X-rays performed 6 weeks after surgery: pelvic incidence (PI); lumbar lordosis (LL); fused segment lordosis (SL); rod lordosis (RL). The following additional parameters were analysed, similarly to those described by Moufid and Gille⁴: the distance between the posterior vertebral wall and the rod for each vertebra (the standard deviation of the three distances was calculated and called Alpha) (Fig. 1); the angle between the screw and the rod for each screw (mean of the three was calculated and called Theta) (Fig. 2); the angle between screws and superior endplate for each vertebra (mean of the three was calculated and called Lambda) (Fig. 3). We also calculated the DiffL value which is the difference between postoperative segmental lordosis and the lordosis of the rod (Fig. 4). All the parameters were independently measured by two authors and the correlation between them was assessed by the intraclass correlation coefficient, which was >0.8 for all the parameters. A mean of these values was reached for final value analysis.

Close

Fig. 1 Measurement of the distance between the rod and the posterior wall of the vertebral body (Alpha).

Export to PPT

Close

Fig. 2 Angle between the screw and the rod (Theta).

Export to PPT

Close

Fig. 3 Angle between superior endplate and the screw (Lambda).

Export to PPT

Close

Fig. 4 Difference between lordosis of the fused segment and lordosis of the rod (DiffL).

Export to PPT

3 Results

After analysis of a total of 62 X-rays (4 scoliosis excluded) 58 cases were included. The mean age was 65.7 (SD 8.6) years and women were predominant (67.2%). The most frequent fusion surgery was posterolateral fusion (77.6%); in 6 cases a one-level TLIF was performed and in 7 a 2-level TLIF was performed. Descriptive analyses are discriminated in Table 1.

DiffL didn't associate with type of fusion (p = 0.89 if comparing APL, APL+1 level TLIF and 2 level TLIF; p = 0.62 comparing APL and 2 level TLIF). DiffL varied between 0.1° (practically no mismatch between the rod and lumbar lordosis) and 30.5° of mismatch. DiffL didn't correlate with gender, age, PI and Alpha, Theta or Lambda (Table 2). There was a significant positive correlation between lumbar lordosis and DiffL (ρ = 0.28; p = 0.03) (Table 2). No correlation was found between the radiological parameters for the cut-off point proposed by Moufid and Gille (Alpha 4.7 mm, Theta 86°, Lambda 2.8°) and the DiffL value.

4 Discussion

Over the last decade, achieving postoperative sagittal balance has gained an increased focus in the management of patients submitted to spine surgery and is a current debatable issue. An adequate surgical planning in lumbar degenerative spine, to restore and correct sagittal balance is key to achieve good clinical results, better outcomes and higher patient satisfaction. Spinopelvic parameters can be divided into those linked to the curvature of the spine (LL, which depends on PI, and SL) and those linked to the sacrum and pelvis (PI, which depends of pelvic tilt (PT) and sacral slop (SS)).⁵ They are interdependent, so a correction of a single parameter may influence the others and help improve global balance.^2,5 The primary methods for correcting sagittal imbalance are pedicle subtraction osteotomy (PSO) and Smith-Petersen osteotomy (SPO).^6–8 To help to restore lumbar lordosis other gestures emerged: interbody fusion techniques, monoaxial screws with compression/distraction application, and rod bending.⁴ Polyaxial screw use, while allowing for an easier placement of the rod may also allow placement of a rod in a incorrectly implanted screw, and therefore, their usefulness in achieving a desired spine curvature (as defined by the pre-contouring of the rod) is debatable. Hand contouring of rods before, during, and after (in situ bending) implantation is a routine procedure, particularly in deformity cases.⁹ However, the role of posterior instrumentation remains to be elucidated to just maintaining the angle, or to correct and change the lumbar curvature. Even when adequate rod bending is achieved, a mismatch may occur between the rod and the actual lordosis of the fused segment due to other influencing factors.^4,10 As an example in complex deformity cases, surgeons may be confronted with difficulties in shaping and reducing rods into pedicle screw heads, which may result in forceful reduction maneuvers, potential screw loosening or pull-out, and longer surgery time.¹¹ Moufid and Gille⁴ proposed an ideal screw and rod placement to allow for a better match between rod lordosis and lumbar lordosis, therefore allowing polyaxial screw use while still achieving the desired lordosis.

Here, 58 cases of lumbar fusion where polyaxial screws were utilized were analysed to understand if Moufid and Gille⁴ results were reproducible. To our knowledge, the framework of the current study is identical to the one in the study by Moufid and Gille,⁴ despite the lack of information about the rod curve before surgery as well as corrective maneuvers (compression, distraction, facetectomies), which as we know, are influencing factors. This study was innovative, as it was the first one to try to analyse the relation between rod contouring and final lordosis in short-segment fusion. If validated, information from such study is important as it may help spinal surgeons plan and adequately perform fusions and to achieve the pretended segmental and lumbar lordosis.

Interbody fusion devices are aimed at increasing graft surface area and to increase lordosis. In this study, however, no differences were found between the type of fusion and DiffL. The lack of correlation with fusion type may be explained by insufficient sample size.

The depth of the screws was represented by Alpha, the perpendicularity with the rod was represented by Theta and the parallelism between the screws and the superior endplate was represented by Lambda. However, no correlation was found between these parameters and DiffL not even for the cut-off point proposed by Moufid and Gille (Alpha 4.7 mm, Theta 86°, Lambda 2.8°) and the DiffL value.

A significant positive low correlation was found, however, between LL and DiffL. The SL mean was 28.4° and RL mean was 20.5° (in Moufid study it was 25.9°). This means, that in the cohort of patients analysed here a great lumbar lordosis was achieved with lower rod lordosis. A possible explanation for this is that the middle screw was introduced to a lesser degree in these patients, therefore achieving a higher lordosis with less contoured rods. This is plausible, as in very degenerate spines, facet hypertrophy in the intermediate level may lead the surgeon to introduce the screw less than desired. Another possible explanation for the differing results is that insufficient bending was applied to the rods. Here, the mean RL was 20.5°, probably the initial curvature of the rod (20°) with no additional bending, whereas in the study by Moufid it was 25.9°.

Since no correlation was found between DiffL and the parameters defined by Moufid et al., optimal cut-off values using ROC curves as well as multivariate regression were not performed in the present study.

This study has limitations. First, the small size of our study which can compromise our results. Second, x-rays were evaluated at 6 weeks after surgery, and screw loosening and plastic or elastic screw or rod deformity may already have happened at this stage. Third, no clinical evaluation was performed and, therefore, it is not known if patients with different DiffL have different clinical outcomes. Fourth, this was a monocentric study and the surgeries were made by different surgeons. Finally, pre-operative x-rays were not evaluated.

While the results in this study do not contradict the results found by Moufid and Gille, they highlight a common difficulty in spine surgery, which is related to achieving the planned lordosis, despite adequate rod bending.

The main reason for the lack of statistically significant correlation in your study may be the small sample size. However, other factors may also have influenced these results. While several surgical gestures and techniques such as osteotomies, facetectomies, discectomies,^4,7,12 interbody fusion,¹³ screw compression or distraction¹⁴ and patient positioning^12,15 may aid in making the spine more lordotic and less stiff, therefore allowing better sagittal correction, a mismatch may still occur when using polyaxial screws, since they will allow for a 30° variation on each end of the rod. Further studies, with larger cohorts, may help clarify these results.

In addition to the obtained results, it is the author's opinion, that monoaxial screws, despite adding difficulty when reducing the rod to the screw (as in the cases of degenerative and stiff spines) are ideal to achieve the planned spine curvature. We propose that, in adult degenerative spines, in short segment (1- or 2- level) fusions monoaxial screws are used (in combination with other techniques, such as interbody devices and osteotomies) to achieve adequate spinal curvature.

5 Conclusion

No significant factors were identified in this study to aid in achieving an ideal match between rod and segmental spine lordosis, therefore not validating the study by Moufid and Gille. While the measurements proposed in that study may still be important, it is the author's opinion that, to achieve the desired spinal curvature, monoaxial screws (for short segments) or a combination of mono and polyaxial screws should be used. More studies are needed to clarify all the factors related to this mismatch.

Funding/sponsorship

This research did not receive any specific grant from funding agencies in the public, commercial or not-for-profit sectors.

Authors contribution

Sara Elisa Diniz: Conceptualization, data curation, formal analysis, writing – original draft. Filipa Cordeiro: Data curation, writing – original draft. Ana Ribau: Data curation, writing – original draft. João Vale: Data curation, writing – original draft. Ricardo Rodrigues-Pinto: Conceptualization, data curation, formal analysis, supervision, Writing – review & editing. All authors have approved the final article.