Removal of infected bone in-growth pelvic fixation implants and microcomputed tomography analysis of osseous integration: illustrative case

Abstract

BACKGROUND

A novel pelvic fixation screw fusion device (iFuse Bedrock Granite implant) was designed to promote bony fusion through self-harvesting fenestrations throughout the outer screw shank. Bone on-growth and in-growth using this design have been demonstrated in a sheep model, but data from human subjects have not been reported. A 66-year-old medically complex female with 2 prior spine fusions developed spondylodiscitis cephalad to a prior fusion, requiring the removal of instrumentation, which included screw fusion devices.

OBSERVATIONS

Within this case report, the authors present a novel bone-preserving technique for the removal of well-fixed screw fusion devices. One screw backed out with a driver, and the other 3 had to be trephined out. One screw sheared off the T30 torque driver, implying a removal torque greater than 270 inch-lbs (30.5 Nm). In addition, the removed implants were analyzed using microcomputed tomography (micro-CT) and demonstrated bone on-growth, in-growth, and through-growth in all 4 screws.

LESSONS

Micro-CT clearly demonstrated osseous integration of the screw fusion devices and provided significant support for the use of these implants for lumbopelvic fixation. The authors’ hope is that the novel bone-preserving technique will help other surgeons when faced with the difficult removal of well-fixed screw fusion devices.

https://thejns.org/doi/10.3171/CASE24274

Pelvic fixation strengthens the base of spinal fusion constructs. While various techniques and instrumentation have been employed, screws are most commonly placed in an iliac or sacral-alar-iliac (SAI) trajectory and connected to the spinal rods. Conventional metal screws are solid and do not promote bony fusion of the sacroiliac (SI) joint when placed in an SAI trajectory, which can lead to new-onset SI joint pain postoperatively. A recent systematic review found that this occurs in 24% of cases according to the weighted average.¹ Additionally, there is a 16% prevalence of SI joint pain preoperatively in patients with adult spinal deformity.²

A novel pelvic screw fusion device (iFuse Bedrock Granite implant, SI Bone) was designed to promote bony fusion through self-harvesting micro- and macrofenestrations throughout the outer screw shank. Bone on-growth, in-growth, and through-growth using this design have been demonstrated in a sheep model,³ but data from human subjects have not been reported. Deep spine infection often necessitates implant removal, which presents an opportunity to analyze the explanted instrumentation. Here, we present a patient who developed spondylodiscitis and required implant removal, including screw fusion devices. Microcomputed tomography (micro-CT) scans of the screw fusion devices were subsequently obtained to evaluate bone on-growth, in-growth, and through-growth.

Illustrative Case

A 66-year-old female with a complex medical history and multiple prior spine surgeries underwent removal of well-fixed screw fusion devices due to biopsy-confirmed methicillin-sensitive Staphylococcus aureus (MSSA) spondylodiscitis cephalad to a prior fusion. In early winter 2019, the patient underwent minimally invasive oblique lateral lumbar interbody fusions (OLIFs) at L3–4, L4–5, and L5–S1 with posterior percutaneous instrumentation at an outside institution because of low-back pain and radiculopathy secondary to degenerative scoliosis (Figs. 1A, 1B, 2A, and 2B).

FIG. 1.

A: Preoperative anterioposterior lumbar radiograph demonstrating degenerative lumbar scoliosis with changes at the L3–4, L4–5, and L5–S1 levels. Disregard arrow. B: Radiograph obtained after the L3–4, L4–5, and L5–S1 OLIFs with posterior instrumentation. C: Coronal CT scan demonstrating haloing around the pedicle screws (white circles). D: Final intraoperative image of the revision construct. E: Six-week postrevision follow-up radiograph demonstrating intact instrumentation with slight L2–3 proximal junction kyphosis (PJK). F: Coronal CT scan demonstrating haloing of pedicle screws (white circle) and destructive endplate changes at L2–3. G: Radiograph after spinopelvic instrumentation removal and L2–3 TLIF with posterior instrumentation.

FIG. 2.

A: Preoperative lateral lumbar radiograph demonstrating degenerative lumbar scoliosis with changes at the L3–4, L4–5, and L5–S1 levels. Disregard arrow. B: Radiograph obtained after L3–4, L4–5, and L5–S1 OLIFs with posterior instrumentation. C: Sagittal CT scan demonstrating haloing around the pedicle screws (white circles). D: Final intraoperative radiograph of the revision construct. E: Six-week postrevision follow-up radiograph demonstrating intact instrumentation with slight L2–3 PJK. F: Sagittal CT scan demonstrating haloing of pedicle screws (white circle) and destructive endplate changes at L2–3. G: Radiograph after spinopelvic instrumentation removal and L2–3 TLIF with posterior instrumentation.

The patient’s postoperative course was complicated by ongoing pain and radiculopathy despite treatment with medications, physical therapy, and a spinal cord stimulator. Ultimately, the patient was diagnosed with SI joint pain and received bilateral SI joint injections several times, which provided 4–6 weeks of ≥70% pain relief. Following this diagnosis, the patient was referred to our clinic for definitive management of sacroiliitis.

Upon presentation, the patient complained of signs and symptoms consistent with SI joint pain and presumed pseudarthrosis of the previous instrumented vertebral segments. Her Oswestry Disability Index (ODI) was 51%, and her visual analog scale (VAS) back pain score was 8. The diagnosis was supported by evidence of S1 screw loosening, degenerative changes within the SI joint, and a lack of fusion mass on radiographs and advanced imaging (Figs. 1C and 2C). The diagnosis was confirmed with the use of a discogram and computed tomography (CT)–guided SI joint injections, which reproduced low-back pain at L3–4 and L5–S1 and resulted in 50% pain reduction.

Initial interventions included weight loss counseling and therapeutic SI joint injections. Despite these interventions, the patient presented with ongoing symptoms in the summer of 2022 and was ultimately indicated for revision posterior instrumented spinal fusion with pseudarthrosis repair (L3–S1 with L3–4, L4–5, and L5–S1 Smith-Petersen osteotomies, L4–5 and L5–S1 transforaminal lumbar interbody fusion [TLIF]) and bilateral pelvic fixation (bilateral stacked SAI trajectory screw fusion devices; Figs. 1D and 2D). The patient developed a cardiac dysrhythmia after the L4–5 and L5–S1 TLIFs and L3–4, L4–5, and L5–S1 Smith-Petersen osteotomies were performed, for which anesthesia appropriately intervened, and the patient stabilized. The patient’s sagittal and coronal alignment had been significantly corrected at this point, and due to the cardiac dysrhythmia, TLIF at L3–4 was not performed. However, the fusion was augmented posteriorly.

Following revision surgery, on postoperative day 64, the patient presented to the emergency department with fever, hematuria, and pneumaturia. A CT scan revealed a colovesical fistula and acute diverticulitis, for which she was treated medically and subsequently discharged. In late winter 2023, she underwent laparoscopic robot-assisted colovesical fistula takedown, sigmoidectomy with end-descending colostomy, and ureteral stent placement.

Routine follow-up was performed at 6 weeks (Figs. 1E and 2E), 6 months, and 1 year postoperatively. During the immediate postoperative time frame, the patient was managed with ongoing CT-guided SI injections for SI-related joint pain. However, at her 1-year postoperative visit, the patient presented with new and worsening back pain. CT and magnetic resonance imaging (MRI) were performed due to the new presentation of back pain. Both CT (Figs. 1F and 2F) and MRI (Fig. 3) findings raised concerns for osteodiscitis at the level immediately cephalad to her prior surgery. C-reactive protein and erythrocyte sedimentation rates were elevated at 126 mg/L and 39 mm/hr, respectively. Due to the concern for osteodiscitis, interventional radiology obtained a biopsy from the L2–3 level, confirming the diagnosis of MSSA-positive osteodiscitis.

FIG. 3.

MRI demonstrating endplate changes at L2–3 concerning for spondylodiscitis. FRFSE = fast recovery fast spin echo; STIR = short tau inversion recovery.

Removal of the spinal instrumentation with L3–S1 debridement and an L2–3 TLIF due to segmental instability was performed in winter 2024 (Figs. 1G and 2G). A complex revision was performed, which included the removal of multiple well-fixed screw fusion devices. Below, we report our novel removal technique for screw fusion devices, which included a prototype head-splitting device for tulip head removal.

A posterior midline incision was made, and exposure of the prior instrumentation was performed. Pus was present along the implants, and granulation tissue was encountered throughout the surgical wound consistent with low-grade infection. Due to the concern for biofilm presence, the decision was made to explant the preexisting instrumentation. The rods and screws from L3 to S1 were removed, and sharp debridement was performed along the approximately 17-cm wound, followed by irrigation. The fusion mass from L3 to S1 appeared solid and thus was not revised.

A total of 4 SAI trajectory screw fusion devices were removed during the complex revision. Three of the 4 implants were well fixed, 1 of which sheared off the T30 torque driver, implying a removal torque greater than 270 inch-lbs (30.5 Nm). They subsequently had to be trephined out. The fourth implant was loose due to suspected screw tract infection and was backed out with a driver.

Removal of the well-fixed implants was performed as follows (Video 1): 1) following set plug and rod removal, a wheeled, diamond-cutting, high-speed burr was used to score the tulip head (13.5-mm outer diameter; Fig. 4A). 2) Using the prototype head-splitting device, the tulip head was removed (Fig. 4B). 3) Trephinated (11-mm outer diameter) reamers were then passed over the headless screw (10.5-mm outer diameters) and advanced until the screw was captured within the reamer (Fig. 4C).

FIG. 4.

A: A wheeled, diamond-cutting, high-speed burr was used to score the tulip head. B: Tulip head removal using the prototype head-splitting device. C: Trephinated reaming over the headless screw until the screw was captured.

VIDEO 1. Clip showing the removal technique. Click here to view.

A curette was used to debride the tracts, followed by irrigation with 2 L of antibiotic fluid and 3 L of saline. The tracts were then grafted with antibiotic-soaked, crushed cancellous bone allograft, and bone morphogenetic protein sponges were placed at approximately the level of the SI joint. Vancomycin and tobramycin antibiotic beads were placed throughout the deep aspect of the surgical wound, and vancomycin powder was applied superficially to the fascia before skin closure.

The removed screw fusion devices were subsequently analyzed with micro-CT scans to determine bone on-growth (Fig. 5A), in-growth, and through-growth (Fig. 5B). The patient recovered well in the immediate postoperative period and successfully completed 6 weeks of intravenous antibiotics with no recurrence of infection. Her ODI has decreased to 33%, and her VAS back pain score has decreased to 3. At 6 months postoperatively, a CT scan of the lumbar spine and pelvis will be obtained to assess healing, and she will return to the clinic for evaluation.

FIG. 5.

A: Photographs and micro-CT images of the removed granite screws demonstrating bone on-growth. B: Micro-CT sections of the granite screws demonstrating bone in-growth and through-growth. White indicates metal and gray indicates bone.

Patient Informed Consent

The necessary patient informed consent was obtained in this study.

Discussion

Observations

This is the first report of retrieved bone on-growth, in-growth, and through-growth screw fusion devices in a human subject. The micro-CT scan results demonstrated the effectiveness of the screw fusion device design in allowing bone growth throughout the porous surface. Due to the nature of the bone on-growth and in-growth, the removal of well-fixed implants can leave a significant bony void, creating reconstructive challenges. By removing the tulip head prior to screw removal using the novel head-splitting device, we were able to use smaller trephinated reamers than would otherwise have been required and thereby preserve bone. In doing so, we were able to exchange the infected devices with bone grafts and restore pelvic stability.

Spondylodiscitis is a rare complication, with an estimated 4–24 million cases per year globally.^4–8 In this patient, we reasoned that she developed spondylodiscitis at the level cephalad to her prior lumbar fusion due to an unknown source, which then communicated with the prior fusion down to the pelvis. Given her colovesical fistula, she presumably had several episodes of bacteremia since the first revision. Biopsy grew MSSA, which has been reported to be the most common pathogen in several clinical studies.^9–15

Surgical management of spondylodiscitis is recommended when conservative treatment fails after several weeks of antibiotics.¹⁶ Whether to retain or remove implants during revision spine surgery is debated,¹⁴ but removal without major morbidity can be challenging. The novel technique used in our case minimized bone loss due to the removal of the screw fusion devices. Previously, we used trephined reamers over the tulip head of the screws; however, by removing the polyaxial tulip head with the novel head-splitting device, we were able to use a smaller-diameter trephined reamer, thus preserving bone. In preserving bone, we could exchange the screws for bone graft and restore lumbopelvic stability. Alternatively, we could have used a burr to cut off the tulip heads completely, but that would have introduced significantly more metal debris into the surgical wound compared to the novel head-splitting device.

Similar to the above case, there have been several historical citations of post hoc explantation of orthopedic implants. Engh et al. analyzed the fixation of porous-coated anatomical medullary locking femoral prostheses (DePuy Synthes) used in uncemented total hip arthroplasty that were removed during autopsy of 12 subjects at 12–93 months postoperatively.¹⁷ They found the more extensively coated prostheses had less micromotion at the tip than the proximally coated prostheses and concluded that micromotion increases as the extent of porous coating decreases. While unfortunate, the removal of the well-fixed screw fusion devices provided important clinical data regarding osseous integration.

Randers et al. reported the osseointegration of 6 triangular titanium rods (TTRs) that were retrieved from 6 patients due to recurrent pain after minimally invasive SI joint fusion.¹⁸ These included 1 first-generation TTR (iFuse implant, SI Bone) and 5 second-generation TTRs (iFuse 3D, SI Bone). Suspicion of TTR loosening was determined by radiolucency on CT scans around 5 of the implants, and 1 extruded 1.5 cm from the iliac bone. After ruling out other potential causes for the recurrent pain and confirming relief after an SI joint block, patients were eligible for revision surgery. Three TTRs (1 first generation and 2 second generation) were clinically loose during surgery. Micro-CT scans demonstrated minimal or no osseointegration, and no bacterial growth was detected upon further testing. The 3 other TTRs (all second generation) were well fixed clinically and found to have near full-length osseointegration on micro-CT. This demonstrated that porous implants spanning the SI joint enable fusion by providing a framework for bony fusion. Current imaging used in clinical practice is unable to truly demonstrate bone in-growth, on-growth, and through-growth due to the Mach effect at the bone-implant interface. However, micro-CT can overcome this, although it is not clinically utilized.

We acknowledge that this report of the osseous integration of screw fusion devices is based on samples from a single patient. Although bone was observed on and in the fenestrations throughout the devices, this outcome may not be consistent in every case. However, our results support the implant’s ability to harvest bone for fusion. For screw fusion devices that are well fixed, the novel technique we describe offers a bone-preserving method of extraction. In the future, multicenter studies could be performed to analyze a larger sample of these devices that are removed from human patients to provide further evidence of the bone in-growth and on-growth capabilities.

Lessons

Complications after surgery provide unique opportunities for in vivo evaluation of osseous integration of spine implants. Here, we present a novel technique of screw fusion device removal and a postexplantation micro-CT evaluation. The micro-CT clearly demonstrated the osseous integration of the devices and provided significant support for their use in lumbopelvic fixation. Our aim is that the above technique will help other surgeons when faced with the challenging removal of well-fixed screw fusion devices.

Acknowledgments

This research was supported in part by the Minnesota Dental Research Center for Biomaterials and Biomechanics. We thank Bonita VanHeel for her assistance with the micro-CT analysis.

Disclosures

Dr. Jones reported personal fees for consulting from SI Bone and Medtronic outside the submitted work. Dr. Polly Jr. reported royalties and consulting fees from SI Bone and personal fees for consulting from Globus during the conduct of the study, as well as textbook royalties from Springer outside the submitted work. In addition, Dr. Polly Jr. had a patent for SI Bone with royalties paid from the University of Minnesota.

Author Contributions

Conception and design: Polly, Spence, Jones. Acquisition of data: Polly, Spence, Haselhuhn, Jones. Analysis and interpretation of data: Polly, Spence, Godlewski, Haselhuhn. Drafting the article: Polly, Spence, Haselhuhn. Critically revising the article: Polly, Spence, Haselhuhn, Jones. Reviewed submitted version of manuscript: Polly, Spence, Haselhuhn, Jones. Approved the final version of the manuscript on behalf of all authors: Polly. Administrative/technical/material support: Spence. Study supervision: Polly, Spence.

Supplemental Information

Videos

  Video 1. https://vimeo.com/957126470.

Correspondence

David W. Polly Jr.: University of Minnesota, Minneapolis, MN. pollydw@umn.edu.