Efficacy of Biphasic Calcium Phosphate Ceramic With a Needle-Shaped Surface Topography Versus Autograft in Instrumented Posterolateral Spinal Fusion

Study Design. A multicenter randomized controlled noninferiority trial with intrapatient comparisons. Objective. The aim of this study was to determine noninferiority of a slowly resorbable biphasic calcium phosphate with submicron microporosity (BCP<μm, MagnetOs Granules) as an alternative for autograft in instrumented posterolateral fusion (PLF). Summary of Background Data. Successful spinal fusion with a solid bone bridge between the vertebrae is traditionally achieved by grafting with autologous iliac bone. However, the disadvantages of autografts and unsatisfactory fusion rates have prompted the exploration of alternatives, including ceramics. Nevertheless, clinical evidence for the standalone use of these materials is limited. Methods. Adults indicated for instrumented PLF (1 to 6 levels) were enrolled at 5 participating centers. After bilateral instrumentation and fusion-bed preparation, the randomized allocation side (left or right) was disclosed. Per segment 10 cc of BCP<μm granules (1 to 2 mm) were placed in the posterolateral gutter on one side and 10 cc autograft on the contralateral side. Fusion was systematically scored on 1-year follow-up CT scans. The study was powered to detect >15% inferiority with binomial paired comparisons of the fusion performance score per treatment side. Results. Of the 100 patients (57 ± 12.9 y, 62% female), 91 subjects and 128 segments were analyzed. The overall posterolateral fusion rate per segment (left and/or right) was 83%. For the BCP<μm side only the fusion rate was 79% versus 47% for the autograft side (difference of 32 percentage points, 95% CI, 23-41). Analysis of the primary outcome confirmed the noninferiority of BCP<μm with an absolute difference in paired proportions of 39.6% (95% CI, 26.8-51.2; p < 0.001). Conclusion. This clinical trial demonstrates noninferiority and indicates superiority of MagnetOs Granules as a standalone ceramic when compared to autograft for posterolateral spinal fusion. These results challange the belief that autologous bone is the most optimal graft material.

][11] Much less controversial are allograft or ceramics, although their superiority as standalone alternative has not been demonstrated. 12Even noninferiority is not generally accepted and has only recently been demonstrated for specific ceramics. 13eramics offer numerous advantages as bone graft substitutes, including minimal disease transmission risk, excellent biocompatibility, long shelf life, and cost-effective manufacturing.As a consequence, ceramics have been used in spinal fusion procedures for decades and intensive research to improve their performance has continued. 14][21][22] Our research group previously investigated such a commercialized microporous biphasic calcium phosphate in a randomized clinical trial and demonstrated that the standalone use of this AttraX Putty (NuVasive Inc., CA) in PLF was noninferior to autograft. 13In that study, we also recognized rapid resorption of both the iliac crest autograft and ceramic within the first year, leaving only half of the intended fusions successful.We therefore seized the opportunity to investigate a modified version of this ceramic (MagnetOs Granules, Kuros Biosciences B.V., Bilthoven, The Netherlands; referred to as BCP < μm), designed with a slower resorption rate and a surface topography consisting of submicron needles instead of micrograins.

Study Design
This study is a multicenter, randomized, intrapatient controlled noninferiority trial (ClinicalTrials.govNCT03 625544).The study design and protocol are similar to a previous trial recently published by our research group and discussed elsewhere. 13,23fter obtaining approval by the medical research ethics committee of the University Medical Center Utrecht and local institutional review boards, the study was conducted in 5 Dutch hospitals, in accordance with international legislation and Dutch law.Based on computerized simple randomization, each subject got one side of their spinal fusion trajectory grafted with the BCP < μm ceramic and the contralateral side treated with autograft.At 1-year follow-up, the primary efficacy outcome was  assessed on CT scans to evaluate posterolateral fusion.Fusion performance of the BCP < μm was tested with a noninferiority margin of 15%.Safety was evaluated by analysis of (serious) adverse events.

Subjects
Patients between 18 and 80 years of age undergoing primary instrumented posterolateral spinal fusion of 1 to 6 levels in the thoracolumbar region were considered eligible for this study.The complete list of inclusion and exclusion criteria can be found in Table 1.

Surgical Technique
All subjects underwent a single or multilevel PLF with pedicle screw instrumentation through a midline approach.When deemed necessary, decompression and/ or an additional interbody fusion procedure with local bone graft were performed.After placement of instrumentation and bilateral fusion-bed preparation via decortication, the randomized allocation side (left/right) of the BCP < μm condition was revealed by opening a sealed envelope.
For each segment, 10 cc of the BCP < μm granules were prepared in a surgical steel bowl by soaking them in 10 mL venous blood that was allowed to clot.The resulting slurry was then positioned onto the graft bed with a 20 cc syringe.
For autograft, corticocancellous bone was harvested from the posterior iliac crest on the autograft allocation side, through the initial skin incision.Both local decompression bone and iliac crest bone were morselized into 2 to 4 mm pieces.To match the contralateral use of 10 cc of BCP < μm, a volume of 8 to 10 cc autograft per fusion level was intended.The contribution of iliac crest bone to the autograft condition had to be at least 50%.Graft volumes were assessed by slight compression in a 20 cc syringe, that was then used to position the graft.
Both grafts were placed at the allocated side around the posterior instrumentation in the decorticated lateral gutters, bridging the dorsal surfaces of the transverse processes, facets, and laminae.The wound was then closed in layers, followed by standard postoperative care.

Outcome Measures
Clinical and radiographic assessments were conducted preoperatively, and at 6 weeks, 3 months, and 1 year postoperatively.Patient-reported outcomes mea- sures (PROMs) included a Visual Analogue Scale (VAS) for back and leg pain, the Oswestry Disability Index (ODI) and the EQ-5D-5L.The condition-specific ODI ranges from 0% to 100%, with higher scores indicating greater functional disability related to low back pain. 246][27] Generic health status was measured with the EQ-5D-5L and converted into a single index value ranging from -0.446 (worst health state) to 1.000 (full health). 28

Fusion Assessment
For the primary efficacy outcome, thin-sliced ( ≤ 1 mm) CT scans with multiplanar reconstructions were obtained at the 1-year follow-up.Posterolateral fusion was evaluated independently by 2 spine surgeons blinded to the treatment sides using the previously developed assessment method based on Christensen (2001) and Carreon (2007). 8,29,30Interobserver reliability of this method is moderate (Kappa = 0.45) and comparable to other radiological studies. 13Both sides of each instrumented segment were evaluated in 3 reconstructed planes.To discriminate ceramic remnants from bone, the scatter reduction was switched of.The intertransverse area and the area around the rod, including the facet joint, were scored separately as fusion, doubtful fusion or nonunion.Additional interbody fusion was assessed similarly in the sagittal and coronal planes.CT scans with disagreements were re-examined to reach consensus.For statistical analyses, the posterolateral fusion scores of each segment and side, as well as the scores for interbody fusion, were dichotomized into "fused" (fusion) and "not fused" (doubtful fusion or nonunion).

Safety Evaluation
To assess safety, adverse events potentially associated with the (surgical) procedure were recorded until last follow-up and examined for any potential relation with BCP < μm.Adverse events were defined as any unexpected, undesirable medical experience occurring to a subject during the study.Events were classified as serious when they resulted in death, were life-threating, required hospitalization or prolongation of existing hospitalization, and/or resulted in persistent or significant disability or incapacity.

Statistical Methods
2][33] Weighing the disadvantages of autografting against the consequences of less successful fusions at the BCP < μm side, the noninferiority margin was set at an absolute difference of 15%.With a desired power of 80% and one-sided significance level of 0.05, a minimum sample size of 84 patients was calculated.Assuming that ~15% of the subjects would not be evaluable for primary efficacy analysis (e.g., because of revision surgery with graft removal or lost to follow-up), the total number was set at 100.
Study data were processed in an electronic data capture system (Castor EDC, Ciwit BV, Amsterdam, The Netherlands) and analyzed using SPSS Statistics, version 29.0.1 (IBM Corp., New York, NY).Baseline characteristics, surgical details, PROMs and fusion rates on segment level were summarized using descriptive statistics.The VAS for back and leg pain and ODI at baseline and 1-year follow-up were compared with paired samples t test when applicable (p < 0.05).
To examine fusion at segment level while accounting for clustering of fusion scores within segments and within patients, a 3-level Generalized Estimating Equations (GEE) model with an independent correlation structure and treatment condition as predictor was used.The relation between successful interbody fusion and posterolateral fusion on either or both sides was analyzed using a similar 2-level GEE model with spinal level and interbody fusion as predictors.For both models, the significance  Performance score of 1 means more or an equal number of segments fused on that side compared with the contralateral side, 0 is less or none of the segments fused.The absolute difference in paired proportions of successful fusion performance was 39.6% with a 95% CI (26.8-51.2),p < 0.001.level was p = 0.05 and the odds ratio (OR) along with the 95% confidence interval (CI) is reported.
For the primary outcome analyses, a posterolateral fusion performance score per treatment condition was calculated to correct for multilevel fusions.This score was based on a higher, equal or lower number of fused segments on one side compared to the contralateral side.That way each subject had a single outcome for each condition (1 = more or equal number of segments fused, 0 = less or none of the segments fused).Noninferiority of BCP < μm versus autograft was tested against the upper limit of the 2-sided 95% CI around the difference in paired proportions for successful posterolateral fusion performance, corresponding to a one-sided significance level of 0.025.

Patient Characteristics
Between September 2018 and October 2022, 116 patients provided informed consent, of which 100 subjects were operated according to the randomization scheme.For the primary outcome analysis, 9 subjects were excluded for the circumstances outlined in Figure 2. Patient characteristics and surgical details are summarized in Table 2.The average age was 57 ± 12.9 (range 20 to 79) years, with 62% female.A total of 19 patients were active smokers, the rest were either former smokers (n = 35) or had never smoked (n = 46).The majority underwent surgery in the lumbosacral region (n = 54) and had a single-level fusion (n = 69).A total of 153 instrumented segments were involved, with 55 additional interbody procedures performed in 49 subjects.

Patient Reported Outcomes
During the first year after surgery, clinical outcomes improved, with a mean decrease in ODI of 18 ± 16 percentage points and VAS scores of 24 ± 29 points for back pain (p < 0.001).The decrease in leg pain (median 33 points, interquartile range [IQR] 11 to 65) was not normally distributed and therefore not tested.This improvement is shown in Figure 3 and exceeded the MCID for the majority of patients (VAS back pain 62%, VAS leg pain 67% and ODI 59%).Furthermore, at 1-year follow-up 44% of the subjects achieved a satisfactory symptom state ≤ 22% on ODI.Improvement in clinical status is also reflected in the increased EQ-5D-5L index value, from median 0.40 (IQR 0.24 to 0.58) to median 0.77 (IQR 0.59 to 0.85).

Fusion Assessment
Posterolateral and interbody fusion were assessed in 91 CT scans obtained at 1-year follow-up, encompassing 132 instrumented segments.Four segments were excluded; 2 due to noncompliance with the study procedure regarding grafting, and 2 because of a pedicle subtraction osteotomy.
Of the 128 segments assessed for posterolateral fusion, 83% were fused on either one or both sides.For the BCP < μm side, this was 79% vs. 47% for the autograft side (absolute difference of 32 percentage points, 95% CI, 23-41).The estimated odds ratio favored the BCP < um side at 4.2 (95% CI, 2.7-6.8).For 40% of segments with only one side fused, there is a sharp contrast in the number of fusions between BCP < μm (46 cases) and autograft (5 fusions) (Table 3).
Interbody fusions were assessed in 54 segments, of which 24 were fused.The overall segment fusion rate was a little higher (84%), as 2 levels with unsuccessful posterolateral fusion had successful interbody fusion.Segments with an interbody fusion procedure had a slightly lower overall posterolateral fusion rate compared with segments without (78% vs. 87%).Secondary GEE-analyses, however, showed a positive relation between successful interbody fusion and posterolateral fusion (OR = 5.5; 95% CI, 1.2 to 24.4; p = 0.025), which most likely represents patient specific factors.
The primary outcome is the fusion performance score per treatment condition that adjusts for multilevel fusions.This analysis confirmed the noninferiority of BCP < μm with an absolute difference in paired proportions of 39.6% (95% CI, 26.8-51.2;p < 0.001; Table 4), which even indicates the superiority of the BCP < μm. 34

Safety Evaluation
During the first year, there were 24 serious adverse events related to the (surgical) procedure, involving 17 subjects and 14 reoperations (see Table 5).Reasons for reoperation included surgical site infection (n = 7), persistent cerebrospinal fluid leakage (n = 4), neurologic complications arising from a malpositioned screw (n = 1), and need for extension of the instrumentation (n = 2).Only 3 reoperations included graft removal.Overall, 36 adverse events were reported, including 10 cases of dural tears that were repaired before graft placement.None of the (serious) adverse events could be directly related to BCP < μm.

DISCUSSION
The current study builds upon previous work that investigated a comparable microporous biphasic calcium phosphate (AttraX Putty). 8,13,23These studies established the effectiveness of the intrapatient controlled design to compare bone graft substitutes to the gold standard, i.e. autologous bone.To avoid bone graft quality as a potential confounder when only local graft is used, we decided to use at least 50% iliac crest bone graft.Like other studies, we demonstrated that achieving fusion with autologous bone graft at a single intended graft location like the posterolateral gutter is challenging and does not exceed 55% after 1 year. 4,5,13The observation that most autograft and ceramics are resorbed within a year, but fusion continues thereafter, 8 supports the idea that post- erolateral fusion relies more on facet ankylosis than on graft-related bone bridge formation on the long term.Moreover, it suggests osteoconductive fusion is not optimally facilitated by autografts nor most ceramics.A lower resorption rate is a key difference between the microporous ceramic previously investigated and the BCP < μm investigated in the current trial.
Even though the primary aim of the current study was to demonstrate noninferiority, our findings indicate superiority of the BCP < μm in terms of CT determined posterolateral fusion at 1 year.This superiority became evident with the primary outcome, the fusion performance score, that adjusts for multilevel procedures by comparing one treatment side to the other.With the additional analysis (GEE), examining fusion rates per segment, we similarly observed the superiority of the BCP < μm condition.This was most prominent when looking at the unilaterally fused segments in Table 3, were BCP < μm was responsible for the fusion in 46 of 51 cases.We realize that the observed superiority of a standalone ceramic has not been shown before and definitely needs confirmation by others in future studies.
There are some important limitations to the current study.First, we used an outcome measure that at best only indicates if the intended fusion has been achieved.Even if this leads to an improved clinical outcome after 1 year, the intrapatient model does not allow for comparison of patient reported outcomes.To really demonstrate clinical benefit, thousands of patients are needed, probably with a much longer followup.We have chosen the objective outcome of radiographic fusion as this is the purpose of the grafting procedure.Second, the intrapatient design only assessed unilateral fusion, which underestimates the fusion rate when any fusion (left and/or right) would be regarded as a fusion.Third, the reliability of the thin-slice CT assessment for fusion determination is not fully established, as highlighted in a recent systematic review and reflected by the moderate interobserver reliability. 35Fourth, because of the slower absorption rate of this BCP < μm, we are not completely sure that the fusion observed in this condition always represents bone and is not a remnant of BCP < μm that perfectly mimics bone.However, by adjusting the scatter reduction function, BCP < μm granular remnants could be identified in 1 year CT scans and distinguished from bone relatively easily (Fig. 4).Interestingly, in subsequent 2 year CT scans (not part of the current study) the granules appear to remodel into bone (Fig. 5).This reveals a fifth limitation, that in many cases the observers could not be truly blinded.Given the radiological resemblances between ceramics and bone, it will be very difficult to completely exclude the human factor for this assessment.Conducting fusion assessments after a longer follow-up period would afford the BCP < μm more time to dissolve, but reduces the graft related component of fusion.

CONCLUSIONS
This clinical trial demonstrates the noninferiority and potential superiority of BCP < μm compared to autologous bone graft in terms of CT-assessed posterolateral spinal fusion after 1 year.Therefore, MagnetOs™ Granules could serve as a standalone bone graft substitute for autograft in instrumented thoracolumbar PLF.
The Dutch Clinical Spine Research Group Consists of the Following Investigators:

➢ Key Points
❑ This randomized intrapatient controlled trial investigated the efficacy of a slowly resorbable biphasic calcium phosphate with submicron microporosity (BCP < μm, MagnetOs Granules) as a standalone bone graft substitute in instrumented PLF.❑ At 1 year the fusion rate on the autograft side was 47% versus 79% for the BCP < μm side.The overall posterolateral fusion rate per segment (left and/or right) was 83%.❑ Primary analysis confirmed the noninferiority of BCP < μm in terms of the fusion performance score that adjusts for multilevel fusions.❑ Primary outcome analysis even indicates superiority of the BCP < μm.

Figure 1 .
Figure1.The BCP < µm granules of 1 to 2 mm in size have a macroporosity and microporosity and display a characteristic submicron needle-shaped surface topography when observed at high magnifications with a scanning electron microscope.

Figure 3 .
Figure 3. ODI (0%-100%; dark blue line) and VAS pain (0-100; back pain in light blue line and leg pain in orange dotted line) scores at baseline and each postoperative followup.Median values along with their interquartile range are given as the data are not normally distributed.ODI indicates Oswestry Disability Index; VAS, Visual Analogue Scale.

Figure 4 .
Figure 4. Computed tomography images of a 70-year-old female 1 year after L2-S1 PLF with L2-3 and L5-S1 TLIFs.(A) Coronal reconstruction showing dense ceramic granule remnants around the rod on the right side.(B) Sagittal reconstruction of the left side showing some autograft remnants posterior to the rod and no signs of fusion.(C) Sagittal reconstruction of the right side indicating bony fusion anterior to the rod and a dense mass posterior to the rod.(D) Without scatter reduction this mass clearly contains ceramic remnants.(E) Axial view L2-3 indicating a mass around the right rod, that can be identified as ceramic remnants without the scatter reduction (F).

TABLE 1 .
Inclusion and Exclusion Criteria

TABLE 4 .
Posterolateral Fusion Performance Per Treatment Condition, After Correction for Multilevel Fusion (n = 91)

TABLE 5 .
Number and Nature of Serious Adverse Events (n = 100)