Patient characteristics
A total of 72 patients (37 men) underwent LMD and ACD implantation between January 2013 and October 2016 at a single institution. Twenty-nine patients had surgery and ACD implantation within the randomized controlled trial and 43 were “commercial cases”. The median age at time of surgery was 47 ± 11.63 years. Overall median follow-up was 14.67 ± 4.77 months. Forty-six patients (64%) had smoked at some point in their lives; 33 (46%) of them were still active smokers and 14 (19%) had quit the habit. Forty-seven (67%) patients had a job requiring heavy lifting. The remaining 25 (33%) had an occupation with lighter physical demands and 5 (20%) were already retired.
Intraoperative Parameters initial surgery
There were no surgical complications. The mean annular defect size was 7.77 ± 1.27 mm in width and 1.35 ± 0.77 cc of disc material was removed on average. The difficulty of surgical procedure with respect to ACD implantation was generally rated as easy or acceptable.
Endplate changes
Endplate changes were found in 71 (99%) patients included in this study. These EPC volumes were significantly larger in the lower endplate (EP) of the vertebra, with a median of 230 mm3, compared with those in the upper EP, with a median of 105 mm3 (p=0.006) (Figure 3). The largest EPC was 4.2 cm3. A multivariate regression analysis of risk factors for development of EPC at the upper or lower endplate only produced one significant predictor: the ACD anchor localization (p = 0.038) (Table 1). Subsequent univariate regression analysis showed that multivariate contribution of anchor localization is based on a significant effect of lower EPC after anchor implantation in the superior EP (p= 0.025) (Table 2). In the current study group, 46 (64%) ACD titanium anchors were implanted in the superior EP with the polymer mesh linked to the lower EP of the adjacent vertebra. We analyzed median upper and lower EPC volumes in association to anchor localization (Figure 4). A significant difference between the inferior and superior anchor localization was only present at the lower EP (p=0.022).
Descriptive score
This grading system was based on the fact that patients with ACD in inferior endplate had, on average, smaller EPC compared to superior EP fixation. Table 3 presents the descriptive criteria for the ACD - EPC scoring system. The scoring system was validated in our patient cohort and Table 4 presents the distribution in our study cohort.
Case report
Figure 5 illustrates a case of a 52 year-old women who underwent LMD and ACD implantation at L4/5 level due to disc herniation and persistent L5 radicular pain. Preoperative CT confirmed absence of any EPC. LMD and ACD implantation in the superior EP of L5 was successful, no complications occurred intraoperative and radicular pain was resolved postoperatively. At one-year follow-up, EPC (size = 0.57 cm3) had appeared around the mesh at the lower EP of L4. Only minimal changes were noticeable below the titanium anchor at upper EP of L5. After 2 years, EPC (size = 1.15 cm3) progressed in the lower EP of L4 and at the 3 years follow-up, EPC (2.46 cm3) increased to a grade 3, with polymer mesh subsidence into the lower EP of L4. Instability caused increasing facet joint-associated low back pain and the patient underwent subsequent fusion and removal of the implant.
Reoperation due to reherniation or device failure
At the time of study analysis in 2018, 65 patients (90%) were still living with the implant and in 7 (10%) patients the whole implant (n=4), or only the polymer mesh (n=3) had been removed. A total of 17 (24%) patients revealed reherniation in the postoperative follow-up MRI, 10 patients (13.8%) were symptomatic and 7 (10%) asymptomatic. In addition to conservative treatments (n=4), reoperation was necessary in 6 (8.3%) patients in the symptomatic cohort. The procedures included recurrent LMD only (n=3), LMD with ACD removal (n=1) and LMD with mesh removal (n=2). A device failure occurred in 19 (26.4%) patients. "Failure" was defined as: dislocation of the whole device >2mm (n=5, 6.9%), device anchor-head breakage (n=1, 1.3%), or posterior dislocation of the mesh into spinal canal (n=13, 18%). Mesh subsidence into EPC was documented in 15 (20.8%) patients (Table 5 and Figure 6). Overall, 7 patients (9.7%) underwent reoperation due to device failure within the follow-up period. In terms of reoperation techniques, the following surgical approaches were used in a total of 13 reoperations (18.1%): Re-LMD only (n=3, 4%), removal of polymer mesh and LMD (n=2, 3%), LMD with mesh and anchor-head removal (n=1, 1%), explant of the whole ACD and LMD (n=1, 1%), transpedicular fusion with ACD in place (n=3, 4%) and fusion with ACD removal (n=3, 4%) (Table 6). All reoperations occurred on average 18 ± 11 months after LMD and ACD implantation. According to the surgeons, reoperation and ACD removal was more difficult than re-LMD only. An unintended durotomy during reoperation occurred in 31% of cases. During all revision surgeries, portions of the mesh, bone and disc material were sent for bacteriological growth testing, and any ongoing infection was excluded.
Clinical Outcome
In summary, the surgery achieved significant improvements with good clinical outcome. These findings are most likely independent from the ACD implantation and a normal consequence of LMD surgery. Mean preoperative ODI of 57.27 decreased significantly postoperative to a mean of 17.58 (95% CI [35.71, 46.99] p < 0.001). With regard to low back pain VAS, a preoperative mean of 63.53 showed significant postoperative reduction to 19.80 (95% CI [35.32, 57.26] p < 0.001). Left leg pain VAS, decreased significant from a mean of 45.66 preoperative to 11.46 postoperative (95% CI [23.72, 49.26] p < 0.001). Similarly, right leg pain VAS reduced from a preoperative mean of 42.97 to postoperative 10.79 (95% CI [21.45, 49.72] p < 0.001) (Figure 7). We also investigated if the magnitude of EPC had any influence on postoperative outcome. In particular, lower EPC showed significant correlation with postoperative ODI at the last follow-up only (p = 0.01) and a trend was present with lower EPC and last postoperative low back pain VAS (p = 0.6). In fact, EPC do not seem to affect postoperative clinical outcome in a significant manner.
Disc degeneration MRI findings prior and after surgery were compared with a mean postoperative follow-up of 22.7 ± 9.7 months to assess DDD. Preoperative mean disc height measured 6.2 ± 1 mm and decreased postoperatively to a mean of 5.5 mm ± 1 mm. This 0.7 ± 0.6 mm reduction in disc height was non-significant. Disc degeneration at index level was Pfirrmann grade III in 49% preoperative and 66% postoperative (Figure 8). A third of patients (31%) had no Modic changes before surgery, and type 1 increased from 14% to 44% (Figure 9).