Transforaminal Lumbar Interbody Fusion Using LOOP® PEEK Cage Implants: Safety, Feasibility, Radiographic and Clinical Outcome

Objective: A variety of newly designed grafts for transforaminal lumbar interbody fusion (TLIF) have been introduced for clinical application. Biomechanical properties of the LOOP® PEEK cage (Medtronic GmbH, Meerbusch, Germany) have been shown in cadaver laboratory investigations, but not in clinical studies so far. In this study we analyze the safety, clinical and radiological outcome of the LOOP® PEEK cage implant in a clinical setting. Methods: Forty one consecutive patients undergoing fluoroscopic-guided posterior pedicle screw fixation combined with TLIF using the LOOP® PEEK cage for degenerative spine diseases between January 2010 and December 2011 were included. Time intervals for follow-up, clinical and radiological outcome data collection were at 1, 3 and 12 months. Visual analog pain scales (VAS), neurological exam, patient-reported SF-12®, CTscans and plain x-rays of the lumbar spine were used as clinical and radiologic outcome measures. Following data were recorded for safety evaluation: procedure duration, intraoperative blood loss, number of levels fused, intraoperative complications, hospitalization time, and postoperative complications. Results: A total of 49 cages were implanted during 41 procedures with an average procedure time of 225.25 minutes. Four patients (9.8%) experienced a dural tear, While new sensory and motor deficits were seen in 2 (4.9%) and 1 (2.4%) patients respectively. complications were not associated with implant insertion. Significantly reduced pain scores (p<0.05, paired t-test) were reported by 29 patients (70%) at 1, 3 and 12 months. SF-12® results showed PCS and MCS scores below the healthy population average, one year post-op. Cage dislocation was observed in 2 (4.9%) patients, one required late revision. Implant fracture did not occur. Inchoate fusion of the vertebra was seen in 39 patients (95.1%) at one year. Conclusion: TLIF procedure combined with lumbar fusion using LOOP®-PEEK cage, provides a safe and feasible intraoperative alternative as well as good clinical and radiologic outcome, without increasing the overall complication rate of TLIF procedures.


Introduction
Various surgical techniques for lumbar interbody fusion combined with posterior pedicle screw fixation have been proven being reasonable for treatment of degenerative spinal disorders. Posterior lumbar interbody fusion (PLIF) [1,2], transforaminal lumbar interbody fusion (TLIF) [3,4], and anterior lumbar interbody fusion (ALIF) [5] are the most frequently performed whereby all three columns of the spine are stabilized, resulting in a circumferential fusion [6,7]. The advantage of TLIF and PLIF procedures over ALIF is that they only require a single approach. In addition, ALIF procedures are associated with a risk of retrograde ejaculation, injury of abdominal vessels and greater blood loss due to the trans-or retroperitoneal approach [3,[8][9][10]. PLIF procedures show increased risk of epidural bleeding, arachnopathy and peridural fibrosis. Furthermore, they are limited to segments L3-S1 due to the risk of conus medullaris damage [3,11,12]. The TLIF technique -a modification of PLIF by Harms -seems to be simpler than and as safe as PLIF [4,6]. The advantage of TLIF is its unilateral approach, less arachnoiditis, and avoidance of excessive nerve root retraction and coagulation of the epidural vessels, resulting in less epidural scaring [3,6,11,13,14].
Many studies have proven the efficiency of TLIF procedures, yet many interbody graft products are introduced to the market without primary evaluation in a clinical setting. The aim of our study was to analyze the safety, as well as the clinical and radiological outcome of patients undergoing TLIF, using the LOOP ® PEEK cage implant combined with posterior pedicle screw fixation in a clinical setting.

Patient characteristics
Data were collected and analyzed between January 2010 and December 2011 for all patients undergoing TLIF using the LOOP ® PEEK cage implant combined with fluoroscopy-guided free hand posterior pedicle screw fixation. The study was approved by the IRB of the Kantonsspital Aarau, Aarau, Switzerland. Inclusion criteria were: failed period of conservative therapy for at least 3 months combined with clinical and/or radiologic lumbar instability, recurrent herniated disc, symptomatic spinal-or foraminal stenosis, spondylolisthesis Meyerding grade I or II [25], or degenerative osteochondrosis. Patients with arachnoiditis, infection, severe osteoporosis, tumor, or life expectancy under 3 months were excluded from the study. Demographic data were retrospectively collected from patients' charts in addition to following parameters: patients' history (BMI, smoking status, occupation, working disability, previous spine operations, preoperative pain, duration of ailment and neurologic deficits), conservative therapy (analgesics, physiotherapy, infiltration), pathologies on preoperative radiographic scans (spondylolisthesis grade, osteochondrosis grade, foraminal stenosis, instability), intraoperative findings (operation time, intraoperative blood loss, operated level, number of levels fused, operative technique, intraoperative complications), hospitalization time, postoperative complications, clinical outcome and radiographic outcome.

Pre-and postoperative assessment
Pre-and postoperative clinical outcome was assessed prior to the procedure, then at 6 weeks, 3 months, and 1 year post-op. Clinical outcome was evaluated using the visual analogue pain scale (VAS), neurological exams and SF-12 ® self-reporting surveys at one year. Preoperative radiological assessment included plain and functional x-rays, in addition to magnetic resonance imaging (MRI) of the lumbar spine. Radiographic outcome measures included computer tomography (CT) scan on first postoperative day and plain x-rays at 1, 3, and 12 months.

Operative techniques
All patients were operated using an open surgical technique. The patient is placed in a prone position under general anesthesia. The posterior elements of the spine are exposed to the base of the transverse processes. After fluoroscopic guided pedicle screw insertion, the superior and inferior articular processes of one facet joint are resected and the disc is exposed in the neuroforamen. Laminectomy, interlaminotomy, recessotomy or foraminotomy procedures are completed depending on the patient's clinical and specific symptoms. The epidural veins are coagulated before incising the disc. The disc is subtotally resected using rongeurs and curettes. After scraping the endplates, the LOOP ® PEEK cage implant is filled with bone chips taken from resected lamina. It is then inserted under fluoroscopic guidance into the anterior-central region of the disc space. The radius between cage and introducer can be adjusted in each phase of the implantation. Finally, both rods are mounted under light compression. The muscle fascia and skin are then sutured in the customary manner. All patients were operated by the same team using the described standardized surgical technique.

Data analysis
The SF-12 ® surveys were analyzed using QualityMetric's QM Certified Scoring Software (QualityMetric, Lincoln RI, USA). Analysis of the final data set was performed using JMP software (SAS Institute Inc., Cary NC, USA) and Superior Performing Software System (SPSS; IBM SPSS Inc., USA). For statistical data analysis, the paired t-, Mann-Whitney U, Fisher's extract, and Spearman's rho test were used. Values were expressed as mean ± SD. A p value <0.05 was considered significant. The authors did not receive any funding and there is no conflict of interest.
48.8% (n=20) had undergone a prior lumbar operation, 14 (34.1%) at the same level, 2 (4.9%) at an adjacent level, and 4 (9.8%) at the same and an adjacent level. Patients' clinical characteristics are presented in Table 1, and the most common secondary diagnoses are listed in Table 2.
In total, 49 cages were implanted. Cage implantation was possible without total facet joint removal in 29 patients (71%). Implant handling and placement was performed easily according to the manufacturer's instructions and without incident in all cases. The non-radiolucent implant markers guided implantations reliably in all cases.
Dural tear occurred in 4 (9.8%) patients during decompression and wasn't associated with implant insertion. In 3 (60%) cases of dural tear, the patients had undergone prior lumbar surgery at the same level.
The average length of surgery was 225.25 minutes (min; ± 48.83 minutes; range 90-335 min), whereby mono-segmental (219.36 min ± 48.01 min; range 90-290 min) operation time was not significantly (p>0.05) longer than multi-segmental procedures (245.56 min ± 48.83 min; range 180-335 minutes). Age had no influence on the duration, although a higher BMI, and female sex was associated with increased surgery time (p<0.05).
Average blood loss was 552 ml (±463.77 ml; range 100-2500 ml). Although age, sex and number of fused segments did not influence blood loss significantly (p>0.05); a higher BMI, and longer operating time resulted in significantly more intraoperative blood loss (p<0.05).

Postoperative early clinical and radiologic findings
The mean hospitalization time was 9.3 days (±5.3 days; range 5-38 days). Sex, age, BMI, number of fused segments and intraoperative complications (e.g. dural tear) did not influence the length of stay (p>0.05).
Of 4 (9.8%) patients with intraoperative dural tear, one patient developed a diffuse swelling in the operation field causing caudal symptomatic and neurologic deficits, requiring reoperation. New sensory and motor deficits were seen in 2 (4.9%) and 1 (2.4%) patients,    respectively. Three (7.3%) patients required a postoperative blood transfusion. One (2.4%) patient developed a deep venous thrombosis in the lower extremities and was treated with oral blood thinners. Postoperative infections or bleedings did not occur.

Secondary Diagnosis Number of Patients
CT scans on the first postoperative day showed 2 (4.9%) pedicle screw misplacements; reoperation was necessary in both cases. No cage dislocation was detected.

Clinical and radiologic findings at 1, 3, and 12 months postoperative
Clinical follow-up at 1, 3 and 12 months showed significant reduction in VAS pain scores (p<0.05, paired t-test; Figure 2).  (Table 3). Throughout the entire follow-up time, 37 patients (90.2%) showed no new neurologic deficits. While postoperatively and at 1 month follow up new postoperative neurologic deficits occurred in 4 patients (9.8%), after 12 months these deficits recovered completely in 3 of these patients and were probably due to postoperative swelling caused by manipulation of the nerve root. However, in the patient showing persistent neurologic deficits after 12 months, a postoperative spondylitis, spondylodiscitis and arachnoiditis occurred, causing these permanent postoperative deficits. New postoperative neurologic deficits were never caused directly by the LOOP ® -PEEK cage.
Radiology films obtained after 1 month (n=33) showed 1 (3%) cage dislocation with no further migration at 1 year post-op. After 3 months (n=27), further cage dislocation (3.7%) occurred in one patient together with late infection, which required surgical implant removal and antibiotic therapy. After 12 months (n=21), 1 patient (4.7%) showed radiographic signs of pedicle screw loosening and 2 patients (9.5%) showed signs of adjacent segment disease. Throughout the whole follow-up time none of the patients showed a progression of spondylolisthesis or cage subsidence into the vertebral endplates. Conclusive evaluation of the fusion rate after 1 year was not possible however, 39 patients (95.1%) showed an inchoate fusion of the vertebra (Table 4).

Cage migration into endplates none
High fusion rate 39 (95.2%)   images of a 46 year old male suffering from back pain and pain to both his legs. In addition to the degenerative changes to the lumbar spine, the patient presented with lumbar disc herniation of the segment L4/5 on the left side and L5/S1on the right side. We indicated fluoroscopic guided pedicle screw insertion in the segments L4 -S1 and microscopic dorsal decompression, discectomy and cage implantation in L4/5 and L5/S1.

Discussion
Our study shows that TLIF using a LOOP ® PEEK cage implant combined with posterior pedicle screw fixation provides a surgically safe, feasible and effective alternative resulting in good clinical and radiological outcomes. The main findings of this study show a significant reduction of the VAS pain score at 1, 3 and 12 months postoperative compared to preoperative. Reduction of lumbar and radicular pain was shown in 90.2% and 82.9% of the patients respectively. The majority (90.2%) showed no new postoperative neurologic deficits, while 4 (9.8%) presented with new neurologic deficits. Neither these, nor dural tears were associated with LOOP ® -PEEK cage handling and implantation.

Transforaminal interbody fusion
Multiple studies have demonstrated equally good clinical and radiographic results of the TLIF procedure compared to PLIF and ALIF [3,4,6,11,13]. The most important advantage over ALIF is the elimination of an additional ventral approach and its complications. Unlike PLIF, nerve root retraction and associated complications are also prevented. Hee et al. claims TLIF as the preferred technique because it is associated with shorter operating time, less blood loss, shorter hospital stay, and lower incidence of complications [9]. TLIF was found to be a safe procedure with good clinical outcome in several recent studies [3,13,[26][27][28]. Our study confirms these results with an overall good clinical outcome of 90.2%, significant reduction of pain score after one year, high surgical safety and no cage-associated intraoperative complications.

The use of PEEK cages in the literature
PEEK biomaterials in the human body were first employed in the field of spinal implants in the early 1990's and are now widely accepted as the material of choice for fusion cages [29]. PEEK was tested in vitro for implant related complications and appears to be as safe as titanium or other metal-based cages [29,30]. Clinical data on neat PEEK implants in lumbar spine fusion has not yet been published in the literature. The first study evaluating PEEK cages in spinal surgery was conducted by Brantigan et al., starting in May 1989 [31]. This two-year    [32]. However, carbon (68%) reinforced PEEK cages were used in these studies. Neat PEEK cages for anterior cervical fusion have also been investigated and findings published [33][34][35]. For lumbar spine fusion, however, the literature is generally limited to in vitro biomechanical, animal or cadaver studies [36][37][38]. Similarly, the LOOP ® PEEK cage has only been analyzed in cadaver laboratory investigations [24]; our study is the first to evaluate a PEEK cage implant usage for lumbar spine fusion in a clinical context.

Feasibility of the LOOP ® PEEK cage
In our experience, surgical handling and feasibility of the LOOP ® PEEK cage appears to be safe. Implantation did not increase overall complication rates of TLIF procedures, as compared to other series. Thanks to the LOOP ® hinge, cage angle adjustment is possible in each phase of the implantation. This enables simple and accurate placement in the median intervertebral space, using a unilateral approach. Due to the pointed tip, we also found insertion in narrow intervertebral spaces easier than with other spinal cages. The LOOP ® PEEK cage had its limitations though, in cases where the intervertebral space was very narrow. The titanium x-ray markers and the radiolucency are helpful during the insertion, facilitating optimal cage positioning (Figures 4-6).

Patient characteristics and intraoperative findings compared to the literature
The mean patient age in our study (51.9 years) was higher than usually described in the literature (mean 44.2 years range 37.2-54.2 years) [26][27][28][39][40][41]. This might be due to the indications and heterogeneous nature of our patient group, as the majority was operated due to degenerative lumbar disorders. Mean blood loss (552 ml) was greater than TLIF procedures described in the literature (474 ml, range 320-609 ml) [3,11,13,19,27], however, the SD (±463.77 ml) was very high and the range (100-2500 ml) very broad. As obesity can lead to significantly higher intraoperative blood loss, and over 70% of our patients were obese (BMI>25 kg/m 2 ), this might explain the findings. Mean surgery time in the literature (187.67 min) [3,11,13,19,42] is lower than that in our study (225.25 min). Mean hospital stay in our study (9.3 days) is also higher compared to the literature (5 days); this might be due to the varying health care systems in each country.

Clinical and radiological outcome compared to the literature
Our results show a significant reduction of the postoperative VAS pain score at 1, 3, and 12 months compared to preoperative. In fact, at one year, the SF-12 ® surveys showed PCS and MCS scores below the healthy population average. Reduction of lumbar and radicular pain was shown in 90.2% and 82.9% of patients respectively. The majority (90.2%) showed no new postoperative neurologic deficits. These results correlate with previous TLIF studies published [3,26,27,[39][40][41][42][43].
Although the VAS pain scores were reduced significantly postoperatively, they had risen again at 12 months. This could explain the low PCS and MCS scores in the SF-12 ® surveys completed at one year, compared to the healthy population. Comparison of SF-12 ® scores at admission and during follow-up would provide better information about the outcome of the chronic pain sub-group. The difference between the chronic back pain cohort post-surgery and healthy population is not significant. Nevertheless, the SF-12 ® survey allows patients to describe their actual life quality in an understandable and objective manner, providing surgeons with a clearer picture of the situation.
In this study 4.9% of the patients (n=2) showed pedicle screw misplacements on postoperative CT scans; both needing operative revision. Cage dislocation within 1 year occurred in 2 cases (4.9%); in one case due to late infection, which needed surgical implant removal and antibiotic therapy. The second case was followed up conservatively and showed no further dislocation after a year, eliminating the need for revision surgery. Radiographic signs of pedicle screw loosening and symptoms of adjacent segment disease were seen in 4.9% and 2.4% of all cases respectively. No patients showed progression of spondylolisthesis or cage subsidence into the vertebral endplates throughout the entire follow-up period. 95.1% of patients showed an inchoate fusion of the vertebra. Our radiological outcome results correlate with previously published literature [3,26,27,[39][40][41][42].

Study limitations
The heterogeneous nature of our patient group with mixed causes of instability was a limitation in our study. The age range was very broad (21-78 years), whereby the mean age was 59.1 years. This indicates that most patients suffered from degenerative lumbar conditions, resulting in worse outcomes as compared to young patients with dysplastic and isthmic spondylolisthesis. Approximately 20% of patients were lost to follow-up after 3 months, and 40% after 1 year. Even though all data was prospectively collected, this study is retrospective in nature and the patients were not randomly selected. There was no control group to compare the LOOP ® PEEK cage directly with a different intervertebral cage. However, with the growing number of implants available for degenerative or traumatic spine diseases, clinical evaluation of these products is of utmost importance in order to improve surgical techniques and patients' outcome.

Conclusions
We conclude that the safety and feasibility profile of the LOOP ® -PEEK cage supports the introduction in routine application with good clinical and radiologic outcome without increasing the overall complication rate of TLIF procedures. Careful evaluation and followup is required in larger series from multiple centers.