Expandable Lateral Lumbar Cages With Integrated Fixation: A Viable Option for Rostral Adjacent Segment Disease

Discussion

While MIS fusion has been identified with similar fusion rates, patient satisfaction, patient-reported outcomes, and complications as open techniques, it has the advantages of less blood loss, length of hospital stay, and cost savings.27 However, both clinically symptomatic ASD and radiographic degeneration classifications remain substantial following spinal fusion. Rates of radiographic degeneration adjacent to previous PLIF are 29% to 34% over 5 to 10 years.28 A review of 4206 patients reported incidences of 5.9% and 1.8% per year for adjacent level degeneration and ASD, respectively.28 Sears et al reported 5- and 10-year ASD rates of 9% and 16%, 17% and 31%, and 29% and 40% after 1-, 2-, and 3/4-level PLIF, respectively.2 Accordingly, most patients (75%) in our study had previous 2-level or 3/4-level PLIF.

In this study, the mean duration between primary PLIF and ASD was 7.8 years, which was longer than the 4.7 years observed in a retrospective review of 1000 patients.3 Our mean age at presentation for ASD was 65 years, similar to ages reported previously.3,4 The most common initial ASD pathologies reported previously3 were degenerative spondylolisthesis and canal stenosis, which were similar to our most common pathologies: spondylolisthesis, foraminal stenosis, and canal stenosis.

Sex did not appear to affect the risk of ASD. Women and men were approximately equally represented in this study, which is consistent with previous results.29 Higher BMI is a risk factor for ASD, and our cohort had an average BMI of 29.5 kg/m². Laminectomy for stenosis adjacent to a fusion also increases the risk of ASD; hence, extension of PLIF2 or LLIF with supplemental posterior fixation3 is recommended.

LLIF is an alternative technique for ASD. It enhances interbody fusion and provides better biomechanical stability.30,31 It preserves all posterior stabilizing ligamentous structures, prevents injury to paraspinal muscles, and avoids denervation or injury to the adjacent superior facet joints. Epidural scarring, perineural fibrosis, cerebrospinal fluid leak, and direct nerve root injury are also avoided.5 LLIF also produces lower infection rates and less estimated blood loss. No wound infections occurred in our study.

Intervertebral cages used with LLIF span both apophyseal cortical rims, providing optimal endplate support to minimize subsidence risk.32 Reported subsidence rates after LLIF range from 10% to 62%.33–35 Higher rates of subsidence with stand-alone LLIF cages occur with reduced bone density (T scores less than −1.0),34 but we found no cases in our patient cohort with eLLIFp cages. Cage geometry is important for avoiding subsidence.32 Subsidence rates are higher with 18-mm polyetheretherketone cages than with 22-mm polyetheretherketone cages (14.1% vs 1.9%).36 The eLLIFp cage width of 20 mm is thus advantageous for avoiding subsidence. Effects of cage length are less clear. Previous studies reported that length does not affect subsidence unless it is insufficient to cover the periphery of the endplates.31,36 However, a greater cage height appeared to increase the rate of subsidence.35 Hence, we took great care to expand the cage using tactile feedback to obtain a snug fit and not torque the expansion mechanism.

Frisch et al reported subsidence rates of 16% with static cages and 0% with expandable cages.12 However, patients with static cages underwent unilateral posterior stabilization, whereas most patients with expandable cages underwent bilateral posterior stabilization requiring repositioning to the prone position. Compared with static cages, expandable cages are inserted at a lower profile and expanded in situ, resulting in less trialing and iatrogenic endplate disruption secondary to impaction. Li et al noted significantly lower subsidence rates with expandable cages compared with static cages at 12-month follow-up.13 The absence of ECS or DCS in our study reflects an experienced lateral surgeon considering the following factors: (1) appropriate case selection; (2) choosing the side of approach best for disc space entry; (3) judicious use of anteroposterior and lateral fluoroscopy; (4) Cobb elevator release of the contralateral lateral ligament; (5) careful endplate preparation; (6) minimal endplate trauma with trials; (7) using angled instruments if indicated (especially at L4-L5); (8) impaction of a wide footprint cage orthogonal to the disc space; (9) the fact that a cage of adequate length is critical to span the apophyseal ring appreciative of ipsilateral osteophytes; (10) careful distraction of the cage avoiding torque release (designed to protect the cage); and (11) lateral plate fixation screws equal to or 10 mm less than cage length.

Biomechanical data for LLIF indicate that spinal segment stability is greater with supplemental fixation than with stand-alone cages.30,37 Indications for supplemental “gold standard” posterior pedicle screw-rod fixation follow a published treatment algorithm and include reduced bone density, facet arthropathy, instability, deformity, pars defect, and at least 3-level pathology.8 The benefits of lateral fixation are a shorter surgical time and no patient repositioning for posterior fixation. Lateral plates reduce lateral bending significantly and also reduce axial rotation, flexion, and extension compared with LLIF cages alone.30,37 However, coronal vertebral body fractures have occurred in osteoporotic patients with a separate 2-hole lateral plate and unilateral pedicle screws10 and in nonosteoporotic patients with 1- and 2-level separate 2-hole lateral plates.11 Separate 4-screw plates are more rigid in all motion planes38 but are technically difficult to insert without causing psoas trauma and are associated with higher risk of lumbar plexus injury. Static lateral cage with integrated 2-screw fixation was initially utilized to limit cage migration after planned or inadvertent anterior longitudinal ligament disruption.39 In our experience, the eLLIFp device was robust with screw lengths equal to or 10 mm less than the cage length to optimize fixation and did not require supplemental posterior pedicle screw-rod fixation.

In our eLLIFp technique, we used a drill, rather than an awl, to create holes for the integrated lateral plate screws and to navigate the superior screw and the inferior screw more accurately. We observed no misplaced screws, cage migration, screw pullout, screw fracture, or vertebral body fracture during follow-up. Likewise, no patient required secondary lateral surgery, posterior decompression, or supplemental posterior fixation at the index or adjacent levels.

The transpsoas approach is associated with anterior thigh pain, weakness, or sensory loss in 27.5% to 38.5% of patients, with the highest risk at the L4-L5 level.40 A rising psoas sign or “Mickey Mouse”-shaped psoas muscle rising away from the vertebral column41 with a more anterior position of the lumbar plexus is associated with a higher incidence of iatrogenic plexus injury.17 In a prospective, multicenter study of over 300 LLIF procedures at L4-L5, mean retraction time was 32 minutes in patients with new postoperative symptomatic lumbar plexus neuropraxia and 23 minutes in those without.42 Despite longer mean retraction times for eLLIFp than for impaction of lateral static cages,43 we observed low rates of anterior thigh pain and sensory changes. The majority of our eLLIFp cages was at the L2-L3 and L3-L4 levels, with only 20% at L4-L5 level. One 79-year-old female patient developed a new motor deficit postoperatively. She had a 12-month history of worsening low back pain and bilateral L4 radicular pain unresponsive to conservative treatment. Magnetic resonance imaging showed moderate central canal and lateral recess stenosis above the L4-S1 PLIF performed 4 years before. L3-L4 eLLIFp was performed, with normal IONM results pre-, intra-, and postoperatively. The patient awoke with severe right L4 radiculopathy, which improved 40% by 6 weeks and 70% by 12 months. The presumed etiology was ipsilateral lumbar plexus traction injury, given the unilateral improving weakness and sensory loss, rather than L4 nerve root tethering from scar in the lateral recesses, which would cause bilateral deficits. The low rate of anterior thigh symptoms and lumbar plexus injury using the eLLIFp in our series is multifactorial. First, an experienced lateral surgeon and operating team took great care in patient positioning to minimize the table break, thus reducing the stretch on the psoas muscle. Second, multimodal IONM detected nerve compromise earlier (compared with electromyography alone), enabling protective adjustments such as reducing table break and repositioning the initial probe and K-wire to ensure the tubular retractor is anterior to the plexus. Critically, the surgeon must directly visualize the target disc space and use triggered EMG to ensure that no nerves, particularly the femoral nerve, cross the operative field. Third, sequential opening of the tubular retractor, initially to a 22-mm diameter for discectomy and then ultimately trialing to a 25-mm diameter, accommodated the eLLIFp prosthesis. Moreover, the eLLIFp is low profile, reducing the risk of psoas muscle and nerve entrapment underneath the insertion. The integrated 2-hole plate is fixed to the cage to prevent posterior plate migration and requires less exposure compared with larger separate lateral plates.

With eLLIFp, back and leg pain improved over 90% by the time of solid fusion. These results were somewhat better than the 80% improvement reported for expandable lateral cages14 and clearly superior to the 49% to 60% improvement with static cages.4,8,13,19 Functional status and disability improved 47% and 69% in our cohort, which were similar to improvements reported for static cages but lower than that reported for expandable lateral cages.13

Our results suggest that indirect neural decompression with eLLIFp was successful, as none of our patients required subsequent posterior direct decompression for either central or foraminal stenoses. In a prospective study of 122 consecutive patients who underwent lateral interbody fusion with impacted 10° lordotic cages, 9% required subsequent unplanned direct posterior decompression for missed unstable spondylolisthesis, bony lateral recess stenosis, or misaligned cages, resulting in contralateral foraminal encroachment.43 The authors concluded that patients with neurogenic claudication and radiculopathy from spinal canal or foraminal stenosis could be successfully treated with lateral interbody fusion, but those with leg pain secondary to bony lateral recess stenosis or unstable degenerative spondylolisthesis may benefit from second-stage direct decompression and fixation.

Solid interbody fusion was achieved for 94% of eLLIFp cages on CT evaluation at 12 months. Our fusion rate was superior to fusion rates reported for stand-alone LLIF cages33 and similar to those for LLIF cages with supplemental posterior fixation.32 A recent study proposed evaluating segmental immobilization in LLIF by both interbody and zygapophyseal joint fusion rather than anterior fusion alone.44 We observed no significant difference in eLLIFp interbody fusion rates between 1 and 4 segment previous fusions, although the 2 nonunions in our study were both found at the L4-L5 level. We suspect the eLLIFp nonunion rate is more dependent on the specific level (L4-L5), rather than on the number of levels (1-, 2-, 3-, or 4-level) of previous posterior fusions, due to the higher biomechanical stress at the L4-L5 segment. The 2 cages in our study without bridging intertrabecular bone showed titanium endplates fused to the vertebrae and no radiolucent lines, subsidence, or screw loosening, consistent with stable locked pseudarthrosis.32 These patients also had no increased mechanical pain indicative of nonunion.

SL and PDH improved significantly from preoperatively to postoperatively and were maintained until fusion, without the need for posterior fixation. The 8 to 17 mm × 5° to 20° eLLIFp cages were expanded to an average 10 mm and 8°, which added 2.2° of SL and 2.7 mm of PDH. These were similar to the changed with static 10 mm × 10° lordotic cages, which increased SL by 2.4° and PDH by 2.6 to 2.7 mm from preoperatively to fusion.19,45 Expandable lateral cages with posterior fixation provide superior SL improvement and similar PDH improvement, compared with eLLIFp. The SL difference may reflect differences in operative levels: 80% of eLLIFp cages were used at less lordotic upper lumbar levels,46 whereas 76% of expandable lateral cages were inserted at L3-L4 and L4-L5.13

We found a comparative mean total cost analysis of eLLIFp (AU$19,715) to single-level PLF, ALIF, LLIF with posterior fixation, TLIF/PLIF (range AU$21,966–30,393) was consistent with median total US costs by the corresponding surgical approach type (range US$18,038–29,640) for single-level lumbar fusions,47 and the mean total cost of lumbar fusions by US national Medicare claims of US$31,716 ± 18,124.48.48 The proportion of our mean total costs for the 6 procedures for lumbar ASD comprising implant charges, OR charges, and hospital charges was similar to the implant, OR, and room charges reported in a cost analysis of single-level lumbar fusions at the University of California.47 The total implant charges were lower for eLLIFp than for alternative procedures (Table 1). Additionally, eLLIFp has lower direct hospital costs with a shorter operation time and hospital length of stay than ALIF, PLF, and TLIF/PLIF.25 Both eLLIFp and LLIF require IONM but no vascular access surgeon,26 cell saver, or perfusionist. None of our eLLIFp patients required intensive care unit admission. These costs have external validity for surgeons in other countries to compare surgical treatment options for symptomatic ASD.

The strengths of this study include the consistent surgical technique performed by a single senior surgeon well experienced in lateral fusion procedures. Furthermore, IONM facilitated a low rate of neurological injury. CT was used for serial radiologic assessments of subsidence, SL, and PDH, with measurements performed using digital software. Additionally, thin-section CT assessments of fusion are superior to evaluating plain static and flexion-extension radiographs.49 It is also the first study to assess efficacy and cost-effectiveness of the eLLIFp device as a stand-alone treatment in appropriately selected patients with 1 to 2 levels of ASD with minimal to no deformity.

Limitations of this study include the relatively small cohort size and intermediate duration of follow-up. Although a comparison cohort was not included, our results are consistent with those reported in the literature. Long-term follow-up of eLLIFp is planned to evaluate the incidence and timing of further ASD. In patients with significant global deformities (sagittal and/or coronal imbalance) more extensive fusion incorporating posterior fixation and osteotomies would be indicated.