Percutaneous iliosacral screw placement has become widely popular and is now part of the surgeon’s armamentarium to treat unstable pelvic and sacral fractures. The technique has been refined over the past three decades with improved anatomic and fluoroscopic understanding, which has resulted in decreased morbidity.15,19,20 More recently, evidence has supported sacroiliac arthrodesis for treating SIJ dysfunction in carefully selected patients.4–8 Irrespective of the surgical indication, there are no prior studies that have clearly evaluated the biomechanics of lag screws across the SIJ pertaining specifically to compressive forces and relaxation rates.
In this study we examined the biomechanical properties of standard trauma lag screws and iFuse-TORQ implants with a porous structure. Mechanical properties of screw fixation is dictated by outer and core diameters, length, thread design, and pitch, which all determine the pull-out strength of the construct and can improve construct stability.21,22 The insertion and removal torques were significantly higher for the TORQ implant placed at S1 than the standard trauma lag screw implant placed at S1. Similarly, the insertion torque was significantly higher for the TORQ implant placed at S2 than the trauma lag screw placed at S2. Comparatively, the results suggest that the TORQ implants are less likely to back out especially at S1, due to increased removal resistance. The lack of a significant difference between the implants placed at S2 may be due in part to the lack of a roughened porous layer of the 10 mm implants where it interfaces with the SI joint.
This rotation resistance may be further enhanced by the ongrowth, ingrowth, and through-growth that will take place over time in the implant’s porous and fenestrated features. Due to the larger diameters of implants, there is a theoretical increased risk of sacral foramina or pelvic cortex breakthrough. All implants however were successfully placed within the bony corridors above the S1 foramen for the S1 implants and between the S1 and S2 foramina for the S2 implants; this included the larger 11.5mm implants at S1 and 10.0 mm implants at S2. This was confirmed using C-arm fluoroscopy imaging for all instrumentations and direct visualization of each specimen.
While not included in the current study, it is commonly understood that 3D printed, additively manufactured implants are not as strong or as fatigue resistant as forged and machined implants of a similar geometry.23 One advantage of additive manufacturing, however, is the ability to design and produce complex shapes and surfaces such as those described in the current study. The mechanical testing required during implant development established that the strength and endurance limit of the current study’s additively manufactured implant was greater than that of commercially available 6.5 mm and 7.3 mm titanium alloy trauma screws (SI-BONE data on file). Also, the 5.5 mm pitch of the additively manufactured implants is greater than that of the 2.75 mm pitch of the 7.3 mm trauma screws. The increased pitch has the advantage of faster implant insertion, but the increased advancement speed might come with less tactile feel, which requires the surgeon to be aware of the fine difference between the final position and implant stripping. At the same time, increased pitch might reduce the initial amount of compression achievable across the SIJ.
Although not included in the current study, the additively manufactured implants were designed for long-term bony ongrowth, ingrowth, and through-growth via the implant’s roughened surface, porous layer and fenestrations. This bony interaction is designed to reduce implant loosening and backout, and similar porous implants have demonstrated the ability to allow for bony ongrowth, ingrowth, and through-growth.24
Stress relaxation models were previously shown to be more reflective of physiological conditions compared to a traditional pull-out protocol.22,25 Peak load, time to 50% peak load and percent load drop at steady state were comparable between both implants, highlighting a similar performance over time in terms of mechanical and viscoelastic properties of the bone-screw interface. When assessing compressive forces, there were no significant differences between the TORQ and the trauma lag screw at both sacral levels tested. About 50% of the load relaxation took place in the first two hours (average of approximately 67 minutes) while the load dropped to 30% within approximately 15 hours after implant insertion. This load drop may still provide enough compressive load to allow for fracture reduction, stabilization, and healing, but the percentage of remaining compression may be lower than expected by some.
The current study placed implants to the midline within the S1 and S2 sacral bodies. It is well understood that bone density within the sacrum varies greatly from the lateral cortices, through the low density ala, and into the higher density sacral bodies.26–29 The midline placement of the current study increased the likelihood of implant bony engagement and comparable measurements between treatment groups. This is the first study to report load relaxation characteristics in iliosacral pelvic fixation. The load relaxation demonstrated in the current study is not unexpected, however, the magnitude of ~ 70% may be a bit surprising. In other bony anatomy, Beadle et al. and Gruszka et al. demonstrated that load relaxes considerably in scaphoid fracture repair, and Cantwell et al. and Migliorati et al. demonstrated force and torque relaxation of dental implants.18, 30–32 In a foam model, Wähnert et al. reported on force relaxation of 6.5 mm cannulated lag screws and Inceoğlu et al demonstrated cyclic load relaxation of pedicle screws.22,25,33 None of the studies, however, reported load relaxation of up to 70%. This difference is due to the cancellous nature of the pelvis and the likelihood of cancellous bone stress relaxing more than the cortico-cancellous bone used in other studies. Future studies focused on the load relaxation of trans-iliac, trans-sacral screws may bear different results due to the termination of the implants in the cortico-cancellous bone of the ilium.
This study includes some limitations. Although a pilot study was performed on seven specimens, no set limit on tightening torques was found. Stripping torques were found to be mostly dependent on feel based on our cadaveric pilot but also from operating room experience. It is therefore likely that the peak load measured from our experiments was not necessarily the maximal compressive force that can theoretically be applied to the screws for fear of stripping them. It is also worth noting that screw stripping occurred with both types of implants at the S2 level. Additionally, only one surgeon inserted the implants and therefore this study did not account for surgeon variability. The load cells were also larger than surgical washers and were more likely to resist washer penetration of the lateral iliac wall than the surgical washers. Similarly, the polymeric wedge washers were quite stiff and are not believed to contribute meaningfully to the load relaxation findings. In addition, the study incorporated different implant diameters and thread profiles, which may confound the interpretation of the results. Ideally we would also like to repeat some of these experiments for bilateral trans-iliac fixation with longer implants. Lastly, the current study focused on compression relaxation loads at t = 0 and did not account for further changes to the loading environment following a patient’s initial movement while rising from a bed or the first assisted steps.
Achieving compression along the axis of a screw is a central surgical principle in many applications and anatomic regions of the skeleton. In the posterior pelvis, compression is desirable for several specific pathologies, including for SIJ fusion, stabilization of a traumatically disrupted SIJ, and fracture compression of vertical sacral fractures. In this study, we found that a novel posterior pelvic implant with a larger diameter, roughened surface, and dual pitch design, achieved improved insertion and removal torques compared to a standard screw and therefore less likely to back-out. This may improve outcomes when treating posterior pelvic pathology, and our results highlight the importance of addressing this question in a clinical trial.