In Vitro Study of Biomechanical Analysis of Modied Lateral Lumbar Interbody Fusion—crenel Lumbar Interbody Fusion

Background To analyze the biomechanical stability of the modied LLIF, crenel lateral interbody fusion (CLIF), and compare various methods of instrumentation of CLIF in vitro. Methods Three fresh-frozen cadaveric lumbar spines (L1-S1) were used in our study. The modied CLIF interbody cage was inserted into the L3/4 level in each specimen. Every specimen was tested under 5 conditions: intact group; stand-alone CLIF group; CLIF with lateral plate group (CLIF + LP); CLIF with lateral plate and unilateral pedicle screw group (CLIF + LP + UPS); CLIF with bilateral pedicle screw group (CLIF + BPS). Results The ROM of each CLIF group was signicantly reduced when compared with intact group in all directions of loading (p< 0.05). The CLIF + LP + BPS group was the most stable in all directions of loading. CLIF + LP group has less ROM when compared with stand-alone group except for the extension condition. CLIF + BPS group has less ROM than CLIF + LP group in every condition. Conclusions CLIF combine with lateral plate and bilateral pedicel screw is the most stable supplemental xation, and lateral plate could reduce the ROM under rotation and lateral bending conditions. For patients with good bone quality, stand-alone with or without is a alternative method to achieve a good clinical result. CLIF, crenel lumbar interbody fusion; LP, lateral plate; BPS, bilateral pedicel screw; AF, anterior exion; PE, posterior extension; LB, left bending; RB, right bending; LR, left rotation; RR, right rotation.


Background
The lateral lumbar interbody fusion (LLIF) is a minimal invasive technique to treat degenerative spinal diseases, which is rstly introduced by Ozgur [1]. LLIF can preserve posterior structures, minimize soft tissue dissection and decrease blood loss when compare to traditional posterior approach. However, LLIF still has the risk of neurologic, vascular and visceral injury, including the lumbar plexus, genitofemoral nerve, aorta injury [2].
In order to reduce the approach-related complications of LLIF, a new modi ed technique named as crenel lateral interbody fusion (CLIF) was rstly introduced by Li et al [3]. CLIF reached the lateral space of the spine through anterior 1/3 of the psoas muscle. The purpose of this study is to analyze the biomechanical stability of the CLIF technique by in vitro.

Specimen preparation
Three adult lumbar specimens were obtained from the department of anatomy of Zhejiang University.
The test was approved by the medical ethics procedure. The mean donor's age was 57.67 years (range from 52 to 63 years). In all specimens, anterior-posterior and lateral X rays were taken to rule out spinal deformity, tumor or fracture. Dural-energy X ray was also performed to quantify the bone mineral density (BMD) of each specimen. Nonstructural soft tissue was removed for each specimen, and the discs, facet joint capsules, and all ligaments were preserved. All specimens were wrapped and sealed with salinein ltrated gauze and frozen at -20℃ for further test. Before the biomechanical testing, all specimens were thawed at room temperature. Half of the cranial and half of the caudal vertebral body were embedded in polymerymethacrylate cement (PMMA).

Instrumentation and surgical procedures
The modi ed CLIF interbody cage was inserted into the L3/4 level in each specimen by the standard CLIF procedure that was introduced by Li [3]. The entire nucleus, lateral annulus and cartilaginous endplates were removed. The pedicle screws, rods and modi ed lateral plate were provided by Shanghai Sanyou Company (Fig. 1). The experimental groups were divided into as follows: Intact; Stand-alone CLIF group; CLIF with lateral plate (CLIF + LP); CLIF with lateral plate and unilateral pedicle screw (CLIF + LP + UPS); CLIF with bilateral pedicle screw (CLIF + BPS) (Fig. 2).

Biomechanical testing
In all conditions, specimens were studied using standard pure moment exibility tests. For all tests, an apparatus was used in which a system of cables and pulleys imparted nondestructive, nonconstraining torque in conjunction with a standard servohydraulic test system (Shanghai Sanyou Medical Co., China) [4]. Loads of 5 N m maximum were applied in the appropriate anatomical axes to induce 6 different types of motion: exion, extension, left and right lateral bending, left and right rotation. In each loading condition, 5 N m were applied for three cycles, and the last cycle used for data analysis. Threedimensional vertebral kinematic response was measured by an optical infrared camera system. All collected data were recorded with data acquisition software.

Data Analysis
Statistical analyses were performed using SPSS 22.0 software. Data were normalized such that each specimen served as its own control. Comparisons between multiple sets of sample means were performed by one-way analysis of variance (ANOVA). P < 0.05 was considered statistically signi cant.

Results
The L3-L4 ROM in different groups was showed in table 1 and Fig. 3. In all xation conditions, the ROM was signi cantly reduced relative to the intact condition in all directions of loading (p < 0.05). For standalone group, the ROM was reduced to 93% of the intact ROM during exion condition (p = 0.069), 84% during extension, 68% during left bending, 94% during right bending, 91% during left rotation, 89% during right rotation (p < 0.05). For CLIF + LP group, the ROM was reduced to 79% of the intact ROM during exion condition, 83% during extension, 63% during left bending, 65% during right bending, 86% during left rotation, 80% during right rotation (p < 0.05). For CLIF + LP + BPS group, the ROM was reduced to 73% of the intact ROM during exion condition, 64% during extension, 31% during left bending, 50% during right bending, 69% during left rotation, 70% during right rotation (p < 0.05). For CLIF + BPS group, the ROM was reduced to 75% of the intact ROM during exion condition, 81% during extension, 50% during left bending, 60% during right bending, 73% during left rotation, 73% during right rotation (p < 0.05).
Of all the constructs tested, the CLIF + LP + BPS group was the most stable in all directions of loading.
CLIF + LP group has less ROM when compared with stand-alone group except the extension condition. CLIF + BPS group has less ROM than CLIF + LP group.
Our study showed the lateral plate could reduce the ROM under rotation and lateral bending conditions. Bilateral pedicle screws group demonstrated the greatest reduction in ROM.

Discussion
LLIF provides a larger interbody cage with preserving posterior elements when compared to TLIF or PLIF by posterior approach. Several studies showed the biomechanical stability with LLIF much better than other interbody fusion approaches [5,6]. However, LLIF still has the risk of neurologic, vascular and visceral injuries. In order to reduce the approach-related complications of LLIF, Li introduced a modi ed LLF technique, and his study showed CLIF has the less complications when compared to standard LLIF [3]. This is the rst study to evaluate the biomechanical stability of CLIF.
Our study showed the CLIF + LP + BPS group was the most stable in all directions of loading. CLIF + LP group has less ROM when compared with stand-alone group except the extension condition. CLIF + BPS group has less ROM than CLIF + LP group. Lateral plate could reduce the ROM under rotation and lateral bending conditions.
With regard to LP xation, Nayak et al. showed the pedicle screw and rod xation had less ROM when compared to lateral plate xation under exion-extension and lateral bending conditions in a 2-level LLIF [7]. Some studies demonstrated a combination of lateral and spinous process plate xation could achieve a similar biomechanical stability with BPS [8,9]. In our study, BPS has less ROM than LP, but LP could limit the rotation and lateral bending when compare to intact group, which was similar with the Nayak's study. We do not perform the spinous process xation because the advantage of lateral plate xation could be xed with CLIF in the same approach, while spinous process and pedicle screw need extra posterior approach.
There are several limitations of our study. Firstly, it is an in vitro cadaveric study, did not consider the effect of soft tissues, in vivo study is required. Secondly, the small sample of the current study may affect the results and ndings. Moreover, our study only evaluates the immediate postoperative result, which do not show the results of fusion rate, adjacent vertebral disease or subsidence rate. Further study is needed to con rm our ndings.

Conclusion
Our study showed CLIF + LP + BPS group was the most stable in all directions of loading. BPS has less ROM than LP, but LP could limit the rotation and lateral bending when compare to intact group. For patients with good bone quality, stand-alone with or without is a alternative method to achieve a good clinical result.