INTRODUCTION
Dysfunction of the sacroiliac joint is painful and debilitating with effective conservative treatments for a subset of the patient population, and may arise as a result of infection, degeneration, or injury [1-5]. Fusion is required when conservative treatments fail; however, it is only effective in one-third of surgical patients [6-8]. The biomechanics of this joint are both unique (including nutation/counternutation motion and coupled motions) and diverse (including sexual differentiation and left/right side differentiation) [9,10]. The first descriptions of sacroiliac fusion approached the joint through an osseous iliac window [11,12]. The transosseous approach accesses the joint without disrupting the stabilizing ligaments surrounding the joint. It is a technically challenging procedure, risking a breach of the anterior cortex and damage to the external iliac artery and L5 nerve root. Alternatively, direct posterior approaches were described [13,14]. The approach accesses the joint directly, dissecting the posterior ligaments and preparing the joint surfaces for fusion. The approach is not without technical challenges. Mis-targeting the serpiginous joint and destabilizing the joint through dissection of the ligaments are substantial risks. Both approaches would be iterated and improved over the decades after their first introductions to improve safety risks and effectiveness [15-21].
Current methods for the evaluation of sacroiliac joint fusion techniques rely heavily on existing spinal biomechanical test methods, utilizing pure-moment loading, fatigue loading, and rotational relative motion qualities [19,20,22-28]. One major common conclusion from these studies is the implant’s ability to reduce motion (predominantly in the nutation-counternutation plane). However, stabilization has been shown to be highly variable and somewhat dependent on bone quality, implant placement, and design. While studies such as Whang et al. [29], Cross et al. [24], and Sayed et al. [27], have attempted to compare posterolateral threaded implants, lateral compression screws, lateral triangular rods, and posterior distraction cage systems in clinical and cadaveric biomechanical models, each of these comparisons only compared results from the subject device to the existing literature, or between literature, significantly limiting the conclusions, due to differences in cohort demographics, sample size, and interlaboratory variations in biomechanical methodology. For instance, the systematic review by Whang et al. [29] presents the posterolateral approach, albeit having the smallest cohort, as being safer in comparison to the posterior and lateral approaches, with respective cohort sizes being approximately 2 and 10 times larger, and with an inverse trend in ODI improvement. However, its conclusions contrast with those from the clinical trial by Claus et al. [30], which reports equivalent safety and efficacy when comparing the posterolateral and lateral approaches in an equivalent cohort.
At this time, no data exists comparing biomechanical effectiveness between posterior, posterolateral, and lateral techniques in the same cadaver population or comparing the posterior approach to either the lateral or posterolateral approaches in a clinical cohort. None of the existing comparative studies have evaluated the durability of fixation, and the few standalone studies that have assessed fixation durability through fatigue loads have only done so within 1,000 and 5,000 cycles, equivalent to 20 and 100 days, which do not represent clinically reported time to fusion [22,31-33]. Furthermore, while previous research has computationally compared the volume and density of bone along the trajectory of fixation, it neither accounted for the actual implant geometries nor evaluated the surface prepared for fusion [34].
Thus, the objective of our study was to evaluate the biomechanical fixation characteristics of the posterior interposition technique and compare it to the lateral transosseous technique as a concurrent control in the same cadaver specimens. We aimed to compare the amount and durability of sacroiliac fixation achieved between the 2 techniques. We hypothesized that posterior interposition requires less bone removal, facilitates an implant design with a larger area available for fusion, fixes the joint more effectively, and has fewer instances of subsidence and migration.
MATERIALS AND METHODS
1. Computer Modeling of the Sacroiliac Joint
The 0.625 mm contiguous axial computed tomography (CT) scan slices of a male sacrum with pelvis were converted and imported into modeling software (OnShape, Cambridge, MA, USA). Each Food and Drug Administration (FDA) cleared implant was assembled in the joint according to each manufacturer’s instructions, displayed in Fig. 1. A cylindrical screw 12 mm in diameter and 60 mm long was modeled as described previously (Rialto, Medtronic, Minneapolis, MN, USA). A conical implant with ancillary screws was modeled with ventral and dorsal diameters of 16 mm and 8 mm, respectively, and a length of 30 mm (SiLO TFX, Aurora Spine, Carlsbad, CA, USA).
2. Cadaveric Modeling of the Sacroiliac Joint
Twelve paired fresh-frozen human cadaveric sacroiliac joints were utilized. The anatomy included L4 to the pelvis. All tissues were procured from American Association of Tissue Banks accredited donor programs. The cadavers’ medical histories were screened to exclude sacroiliac joint disease. Each joint was prescreened to exclude donors exhibiting bony bridging [35]. L4 dual-energy x-ray absorptiometry scans were used to exclude osteoporotic bone density. The average age was 58 ± 9 years, and the average L4 t-Score was -0.9 ± 1.1. The sex distribution of the cadavers was 3:3 male: female. The specimen was dissected from the torso and cleaned of musculature, with care taken to preserve the pubic symphysis and ligaments around the sacroiliac joint.
Multidirectional bending was conducted in a custom-designed fixture previously reported (Fig. 2) [3,25,27]. The details of this fixture include alignment of anatomical axes, adaptation of the cadaver tissue to the test frame, biomechanical pure-moment loading of the sacroiliac joint, calibration of sacral motion relative to the ilium, biomechanical loading profiles, and range of motion calculation. Each joint was cyclically loaded from 0 Nm to 5.625 Nm at 1-Hz extension for 18,500 cycles. Extension loading was chosen as the predominant motion direction in the nutation-counternutation plane. 18,500 cycles were selected to represent 12 months of motion [31,36,37]. 1 Hz was chosen to complete the cycle count within an 8-hour timeframe, and 5.625-Nm maximum moment was the highest loading allowed by the system controller to produce stable loading control and is in line with previous cyclic loading of the joint [22] and ASTM guidelines (F2077-22: Standard Test Methods for Intervertebral Body Fusion Devices) [38].
The pubic symphysis ligaments were sharply dissected with a scalpel knife to separate the left and right sacroiliac joints for independent analysis of each [19]. Each joint was destabilized after intact data collection and before fixation. All posterior sacroiliac ligaments were sharply dissection with a scalpel knife [19,22].
The posterior interposition technique was performed under an anterior-posterior-rotated fluoroscopic view, and the central third of the sacroiliac joint was targeted with a guide pin. The pin was aimed at the apex of the sacral ala using a lateral fluoroscopic view. A joint finder was placed, and a working channel was docked into the joint space. The joint space was reamed 30 mm deep. The implant was placed 30 mm deep. Two screws were put through the implant and into the sacral and iliac bones. The screws were locked into the cone, and the instruments were removed. Representative CT scans of placement are displayed in Fig. 1.
The posterolateral transosseous technique was performed under an AP-rotated fluoroscopic view; a guide pin was placed between the S1 and S2 foramen. The pin was advanced under lateral view perpendicular to the joint in the direction of the sacral promontory. The pin was advanced to 10 mm before the anterior sacral cortex. The implant length was chosen from this pin placement, the channel was drilled over the pin, and the implant was advanced to the end of the pin. The instruments were removed. Representative CT scans of placement are displayed in Fig. 1.
3. Data and Statistical Analysis
The computer models were analyzed for bone volume removed to place the implant and bony surface area available for fusion after placement. Bone volume removed was calculated as the volumes of each implant lodged within the sacrum and ilium. The surface area available for fusion was calculated as the implant external and internal surfaces exposed within the joint space in addition to the adjacent cartilaginous and fibrous portions of the joint. All measurements were performed using the aforementioned modeling software.
The range of motion was calculated for each plane of loading. Fixation and postfatigue were calculated as the percentage of destabilized motion. The stability of the bone-implant interface in terms of implant migration and joint collapse was measured with CT scans taken after fixation and after fatigue loading. Outcome measures were compared statistically within each group in Excel (Microsoft, Redmond, Washington, USA) using repeated-measures analysis of variance with post hoc comparisons using the Holm method. Comparisons between the 2 implant systems were performed using unequal variances t-tests. For all comparisons, a 95% confidence interval (α=0.05) was used to determine statistical significance.
RESULTS
Computer modeling results of bone volume and surface area available for fusion are displayed in Table 1 and Fig. 3. The posterolateral transosseous implant required 3,522 mm3 of bone removed for placement. The posterior interposition implant required 1,212 mm3 of bone removed, 66% less than the posterolateral implant. The joint surface available for fusion with the posterolateral transosseous implant was 1,659 mm2, 12% more than the 1,466 mm2 for the posterior interposition. The implant windows of the posterior interposition implant facilitate a surface of 1,067 mm2 for bony fusion, which is 76% more area than the posterolateral transosseous implant’s 606-mm2 windows.
The results of the multidirectional bending tests are displayed in Table 2 and Fig. 4. Fixation resulted in a 42% ± 8% reduction (p=0.035) in nutation/counternutation motion for the posterior interposition technique and 14% ± 4% (p=0.013) for the posterolateral transosseous technique. This difference in fixation was statistically significant (p=0.018). Reductions in motion during lateral bending resulted in 28% ± 12% (p=0.022) for posterior and 27% ± 3% (p=0.030) for posterolateral but was not statistically significantly different between the fixation groups (p=0.939). Reduction during axial rotation resulted in 21% ± 10% (p=0.112) for posterior and 6% ± 6% (p=0.341) for posterolateral, which was not statistically significantly different (p=0.243). Postfatigue fixation decreased in all directions for both techniques. Nutation/counternutation reduction in motion resulted in 28% ± 11% for posterior and 0% ± 10% for posterolateral; however, these reductions and differences were not statistically significant (p>0.084). Lateral bending reduction in motion resulted in 16% ± 19% for posterior and 12% ± 19% for posterolateral; however, these reductions and differences were not statistically significant (p>0.135). Axial rotation reduction in motion resulted in -8% ± 16% for posterior and -16% ± 13% for posterolateral; however, these reductions and differences were not statistically significant (p>0.451).
Implant migration and joint collapse after fatigue loading are displayed in Table 3 and Fig. 5. Zero precent of the posterior interposition implants migrated during fatigue loading, and 0% of the joints collapsed. 50% of the posterolateral transosseous implants migrated, and 20% of the joints collapsed.
DISCUSSION
The evidence presented is a biomechanical rationale for sacroiliac fusion using a posterior interposition technique. The investigation was conducted using computer-based and cadaver-based models. The posterior interposition technique removed less bone, left more surface area for fusion, immobilized the sacroiliac joint more effectively, and was more durable in resisting migration and subsidence into the sacral and iliac bones.
This study’s results are corroborated by previous sacroiliac joint biomechanical investigations [19,22,26,27]. Best case fixation as a percent of the destabilized range of motion for these previous studies ranged from 38% to 71% in nutation/counternutation, 31% to 65% in lateral bending, and 29% to 65% in axial rotation. While the posterior system investigated utilizes up to 2 implants, and the posterolateral system utilizes up to 3 implants, our study investigated the least invasive scenario using a single implant for both study groups. Our results for the posterior interposition implant were similar to those of previous studies, with 42% in nutation/counternutation, 28% in lateral bending, and 21% in axial rotation. However, our results for the posterolateral transosseous implant were better than those reported in a previous cadaver investigation [26], which analyzed nutation/counternutation motion before and after implant placement. The investigators were unable to observe any statistically significant decreases in motion using a single posterolateral implant. The percent reduction in motions ranged from 2% to 16%, which is slightly worse than the current study (14% ± 4%). The previous investigators failed to destabilize the joint before fixation, which may have yielded less favorable results. The same investigators have reported improved performance using 2 posterolateral implants, with no added benefit of a third implant [39]. Future studies may look into comparing both systems when 2 implants are used. However, the nutation/counternutation motion reduction with a single posterolateral implant in our study (14%) is comparable to that previously reported using 2 posterolateral implants (19%). Both motion reductions are less than that observed in our study using a single posterior implant (42%). We observed a trend between fixation and bone quality in the posterolateral technique that was not present in the posterolateral samples, which is displayed in Fig. 6. Posterior interposition exhibited a stronger relationship with bone quality, with correlation coefficients an order of magnitude larger. Also, the slopes of these trends were an order of magnitude larger. While both devices performed equivalently in lateral bending, the superior fixation in nutation-counternutation, and stronger fixation relationship with bone quality exhibited by the posterior interposition, in comparison to the posterolateral implant, may be due to the differences in fixation principles. The posterolateral implant pierces the joint perpendicular to its surface and thus is oriented along a similar trajectory to the axis of rotation of the joint during nutation-counternutation which is the predominant motion of the joint. In contrast, the posterior implant lies perpendicular to the axis of rotation and should, therefore, be more effective at obstructing nutation-counternutation motions. In addition, the posterolateral implant only serves to bridge the joint, which is a means of form closure; in contrast, while the posterior implant bridges the joint as well, it also exerts force closure on the joint by means of distraction. An increase in forces exerted on the joint surface, increases the friction on the joint, as well as the tension and subsequently the stiffness of the supporting ligaments, thus contributing to added motion resistance in shear motions such as nutation-counternutation.
The study's strengths include the concurrent control design, which controls variability from sex, age, height, bone quality, and weight. We screened our cadavers for sacroiliac joint fusion/bridging. We utilized the multidirectional bending flexibility model, a standard test method used by many laboratories for consistency and reliability [40,41]. 18,500 is the most extended fatigue cycle count performed on sacroiliac fixation to date. Optical markers were placed close to the joint to minimize any effects of bone deformation [42].
Limitations of our study include pure-moment loading. Isolated loading directions do not accurately represent the combined multidirectional in vivo loading of standing up, walking, and daily activity. As with all cadaveric biomechanical evaluations, our sample size was low. Our results only indicate the investigated devices' performance when implanted according to the manufacturer’s specifications, as shown in Fig. 1. Modified placements may influence the performance of both devices. Biological changes over time are not included in our cadaver model. While our model includes soft tissues involved in passive stabilization, such as ligaments and cartilage, our model does not simulate the effects of active stabilization via musculature. Finally, while our study design included specimens with osteopenia and normal bone density, we excluded osteoporotic specimens, as patients with osteoporosis are currently contraindicated by the FDA for sacroiliac joint fusion. The majority (66%) of sacroiliac joint-mediated chronic low back pain patients are females aged 58 ± 5 years at diagnosis [43-45]. According to World Health Organization, over half of these are osteopenic, and just under one-third are osteoporotic [46]. This demographic will, therefore, benefit from future studies investigating treatment performance in osteoporotic specimens. These limitations should be factored in when transferring the findings to a clinical setting.
CONCLUSION
Fixation of the sacroiliac joint with the posterior interposition technique removes less bone, provides more bony surface area for fusion, and effectively reduces joint motion in a cadaver model. The amount of fixation in nutation-counternutation and lateral bending are respectively superior and comparable to posterolateral transosseous techniques. The posterior interposition technique may provide more durable fixation; however, further clinical evidence is necessary.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation
Print
