Fracture of robust fusion mass after sequential long-construct instrumentation removal: illustrative case

Abstract

BACKGROUND

Adult idiopathic scoliosis presenting after the age of 18 years affects 2%–4% of the adult population younger than 45 years. In the cases of multilevel spinal fusion, downstream complications, particularly pain and prominence, may warrant instrumentation removal. Even with proper fusion, deformity recurrence and loss of sagittal correction are common concerns. However, the literature on fusion mass fractures is minimal.

OBSERVATIONS

A 52-year-old woman with a history of adult spinal deformity and posterior spinal fusion presented with discernable prominence at the proximal end of her instrumentation. Imaging confirmed a solid fusion mass, and sequential partial removal of her instrumentation was performed. Nine years after index removal, she returned with immediate pain following a low-energy injury. A subsequent CT scan showed a linear fracture line in her L3–4 fusion mass. Posterior fusion was performed with pedicle screw fixation of two levels above and below.

LESSONS

Removal of thoracolumbar instrumentation poses potential risks, including fusion mass fracture, and requires careful risk-benefit analyses.

https://thejns.org/doi/10.3171/CASE25582

Adult idiopathic scoliosis (AdIS) is a chronological progression of adolescent idiopathic scoliosis (AIS) presenting after the age of 18 years and affects 2%–4% of the adult population younger than 45 years.^1–3 Compared to AIS, AdIS curves are often less flexible and present with lumbosacral degeneration.^3,4 Bracing is often an early treatment strategy for AIS curves that exceed 20°–25°, while multilevel spinal fusion offers a durable long-term treatment option to mitigate progression of curves that exceed 45°–50°.^5–9 Conversely, bracing has a limited role for AdIS, and surgery is generally indicated only for progressing or symptomatic curves.¹ Recently published data have shown that at an 8-year follow-up, surgery provides better outcomes for adult scoliosis than nonoperative treatment.¹⁰ Furthermore, a large propensity score–matched study showed no significant difference in perioperative outcomes between AdIS and AIS after posterior spinal fusion.¹¹

Despite indications for surgical intervention, downstream complications may arise such as infection, pain, screw loosening, pseudarthrosis, proximal junction kyphosis, structural implant failure, and prominence.^12–15 Although such complications, particularly pain and prominence, may warrant instrumentation removal after adequate fusion, removal presents further risks.^16–19 Even with complete spinal fusion and a robust fusion mass, deformity recurrence and loss of sagittal correction are common concerns.^20–27 Defects and fractures of the fusion mass have been attributed to severe curve progression after instrumentation removal; however, the literature on fusion mass fractures is minimal.^20,28 Early correction with reinstrumentation of the affected region is crucial to mitigate rapid curve progression and further complications.

Here we present the case of a patient with AdIS who underwent instrumentation removal in a staged fashion to allow bony remodeling and then sustained a low-energy fusion mass fracture, which required surgery.

Illustrative Case

A 52-year-old woman, with a history of idiopathic scoliosis, presented with proximal thoracic pain and tenderness. The patient had received early nonsurgical management before several thoracolumbar fusion attempts were initially performed. After referral to a second institution, an anterior-posterior fusion extending to the pelvis was performed and provided symptomatic relief. She was able to return to daily activity and exercise before later developing adjacent segment degeneration following an injury. Her instrumentation was thus extended cephalad.

When the patient was seen at our institution, she presented with discernable prominence in addition to pain and exquisite tenderness at the proximal end of her left-sided instrumentation. It was thought that the transverse process hooks, cross-link, and domino rod connectors near T6–7 were likely the primary pain generators. Prior selective nerve root blocks of the left-sided proximal instrumentation provided substantial pain relief. A review of current radiographs and prior CT scans ruled out pseudarthrosis and confirmed a solid fusion mass (Fig. 1). The bone mineral density (BMD) at T3 was 256 Hounsfield units (HU); however, opportunistic BMD had not been described at this time. Instrumentation removal and associated risks including deformity recurrence were thoroughly discussed along with alternative treatment options. The patient strongly desired instrumentation removal, and a sequential partial removal was agreed on to allow bony remodeling.

FIG. 1.

Preoperative radiographs of the robust fusion mass. Posteroanterior (PA) (A) and lateral (B) biplanar images obtained 3 weeks before index instrumentation removal, demonstrating a solid fusion mass. Sagittal measurements were obtained (PI 60.7°, lumbar lordosis 54.6°, T1–12 thoracic kyphosis 56.9°, pelvic tilt 30.2°, T1PA 18.7°, L1PA 12.9°, T4 pelvic angle [T4PA] 16.7°). The solid fusion mass was confirmed with a prior CT scan obtained 19 months before index removal (C), where a lateral cut of the lumbar fusion mass is shown.

Sequential partial instrumentation removal was performed as clinically indicated with three operations individually spaced by roughly 2 and 3 years, respectively. The first of three operations involved removal of the left-sided cephalad thoracic instrumentation from T2 to T7 (Fig. 2A and B), which was deemed to be of greatest concern. Crushed allograft was used to augment the fusion mass at the upper fusion levels given a concern for healing where the cross-link was removed. Transient symptomatic relief was experienced before complaints of left distal prominence and pain as well as left-sided midthoracic rib syndrome. A second procedure was performed involving removal of the remaining left-sided instrumentation from T8 to the pelvis (Fig. 2B and C). Of note, the patient exhibited a solid fusion mass throughout, and bone graft was placed into the removed instrumentation defects. Prolonged symptomatic improvement was experienced before the patient was involved in a car accident. She developed pain over her remaining right-sided thoracic instrumentation and pain in her left sacroiliac joint. The third operation involved complete removal of the right-sided instrumentation from T3 to the pelvis (Fig. 2C and D). Exposure of the outer set plugs at and below the distal thoracic levels required bone removal. Separate bilateral pedicle screw instrumentation at L5–S1 required further bone unroofing. Crushed bone allograft was applied to the residual defects. Exploration of the fusion again supported the notion of adequate structural support. Paraspinal muscle mobilization of 30 cm was required for wound closure. Thoracolumbar back pain resolved following completion of instrumentation removal, although sacroiliac joint and upper thoracic/cervical neck pain worsened.

FIG. 2.

Sequential radiographs of instrumentation removal. PA images obtained 3 weeks before index surgery (A), 9 months postremoval of left T2–7 instrumentation (B), 6 weeks postremoval of left T8 to pelvis instrumentation (C), and 14 months postremoval of right T3 to pelvis and bilateral L5–S1 instrumentation (D).

The patient returned 5 years after instrumentation removal following an injury during daily activity. While pulling a rug, the patient stated she heard a loud “pop” and felt immediate excruciating pain in her low back. A subsequent CT scan showed a linear fracture line of her robust lumbar fusion mass at the L3–4 level (Fig. 3A and B). Given the long-fused segments proximal and distal to the fracture and anterior/posterior fusion at this level, this was presumed to be a three-column injury. The opportunistic BMD at L1 was 140 HU, however, the robust posterior fusion mass at this segment likely provided stress shielding to the anterior column.²⁹ Posterior spinal fusion was performed with pedicle screw fixation of two levels above and below the fracture (L2–L5) (Fig. 3C and D). As expected, the patient’s fusion mass was quite extensive and altered anatomy substantially. The fracture line was decorticated, and a bone autograft was used. With multiple prior surgeries, wound closure was a challenge requiring extensive paraspinal mobilization. Tenderness and prominence were noted during immediate follow-up, although the patient experienced no neurological deficits.

FIG. 3.

Diagnostic and postoperative images of fusion mass fracture. CT scans indicated a fracture line of the fusion mass at L3–4, depicted by a red circle in both PA (A) and lateral (B) cuts. Posterior pedicle screw fixation two levels above and below was performed, and PA (C) and lateral (D) biplanar radiographs were obtained 6 weeks postoperation.

The now 66-year-old patient returned 4.5 years after reinstrumentation with moderate symptom improvement. The most recent imaging demonstrated stable instrumentation without concern for loosening or fracture and an opportunistic BMD at T3 of 260 HU. Compared to prior imaging, potential kyphotic progression at her cervical-thoracic junction was observed with proximal adjacent segment degeneration, for which she remains on a chronic narcotic prescription. There was no notable deformity recurrence or further sequelae within the fusion mass (Fig. 4).

FIG. 4.

Radiographic analysis of spinal curvature over 14 years. Standard PA (lower row) and lateral (upper row) biplanar images were obtained 3 weeks before index removal (T1PA 18.7°, T4PA 16.7°, L1PA 12.9°) (A), 9 months after index removal (T1PA 21.4°, T4PA 17.7°, L1PA 14.5°) (B), 29 months after second removal (T1PA 24.3°, T4PA, 17.3°, L1PA 11.1°) (C), 14 months after third removal (T1PA 19.5°, T4PA 13.9°, L1PA 9.5°) (D), 12 months after fracture fixation/fusion (T1PA 18.7°, T4PA 13.0°, L1PA 8.2°) (E), and 4.5 years after fracture fixation/fusion (T1PA 20.2°, T4PA 13.5°, L1PA 9.7°) (F), with a total span of 14 years. No notable coronal/sagittal correction loss was observed within the fused segments, although kyphotic progression at the cervical-thoracic junction was questioned based on the most recent radiograph.

Informed Consent

The necessary informed consent was obtained in this study.

Discussion

Observations

Posterior instrumentation removal after spinal fusion has a relatively high prevalence and is often associated with prominence and pain. Yuan et al. reported a 12.5% instrumentation removal rate for 2177 patients with pedicle screw fixation for spondylolisthesis or fractures.³⁰ Schwab et al. reported a 14.6% removal rate for degenerative disc disease.¹⁹ Instrumentation removal may elicit further complications including fracture and/or deformity recurrence.

Initial research on complications following instrumentation removal for long-construct deformity was published by Deckey et al., who reported that nearly 30% (4/14) of their cases had significant sagittal correction loss.²⁰ Each case was felt to experience spinal collapse due to fracture or defect in the fusion mass, and all required posterior fixation and fusion.²⁰ Farshad et al. reported a mean 33.3% (13°) coronal correction loss of the main thoracic curve for 7 patients with AIS 10 years after removal, which was statistically greater than that for the control group.²² Further studies, largely investigating AIS, report both coronal and sagittal correction loss following instrumentation removal (Table 1) but demonstrate a wide variability in literature.^{20,21,23–28}

TABLE 1.

Literature comparison of sagittal (thoracic kyphosis angle) and coronal (main thoracic Cobb angle) correction loss following instrumentation removal for patients primarily with AIS

Authors & Year	Sample Size (n)	Coronal Correction Loss	Sagittal Correction Loss	Notes
Potter et al., 200621	15	10°	5°	T2–12 TK, mean 5.2 yrs postremoval
Farshad et al., 201322	7	13°	10°	T4–12 TK, 10 yrs postremoval
Burgos et al., 202323	36	3.9°	1.3°	T2–12 TK, 2 yrs postremoval
Potaczek et al., 200924	59	3.6°	—	T2–12 TK, mean 2.2 yrs postremoval
Yamauchi et al., 202525	116	1.8°	7.2°	T5–12 TK, mean 4.1 yrs postremoval
Kotani et al., 201328	1	28°	17°	T5–12 TK, 15 mos postremoval
Muschik et al., 200427	35	11°	10°	T4–12 TK, 2 yrs postremoval
Weighted mean		4.2°	6.6°

TK = thoracic kyphosis; — = no value existed.

In our case, no obvious deformity recurrence was observed in the sagittal or coronal plane, despite the patient sustaining a fusion mass fracture (Fig. 4). We focused on sagittal measures as there were minimal coronal plane residual issues. Of note, all typical measurements were done for the first available full-spine radiograph (Fig. 1). However, a prior anterior L5–S1 fusion with a Harms cage made discerning the sacral endplate difficult; hence, the reliability of pelvic incidence (PI) is questionable. It is our belief that the key measurement for this case was the T1 pelvic angle (T1PA), noted for all key radiographs. The T4–L1 pelvic angle (T4-L1PA) relationship also offers important insight but was not known at the time of the surgeries.

The lack of deformity recurrence is likely attributed to the sequential instrumentation removal, which decreased stress shielding and allowed for bony remodeling. A recent in vivo ovine study demonstrated that rod load declines with progressing bony fusion.³¹ Additionally, our early intervention and reinstrumentation of the fractured fusion mass may have mitigated correctional loss. Even with seemingly sufficient fusion of the posterior spinal elements, instrumentation allows for load sharing. This was illustrated by Teles et al. and Sedney et al., who reported multiple cases of fusion mass fractures following instrumentation removal of fused segments and addition of instrumentation to adjacent segments above or below.^32,33 Jeon et al. further illustrated this for posterior instrumentation of burst fractures, reporting a segmental motion increase from 1.6° to 5.8° after instrumentation removal.¹⁷

During upper extremity retraction, flexion is resisted by fused segments, which experience a bending moment greatest at the instantaneous axis of rotation.^34,35 This can result in instrumentation failure such as rod breakage or screw pullout and can predispose patients without instrumentation to the risk of fracture.^36,37 Although lever arms associated with long fusions make them more susceptible to fracture, even multiple cases of short posterior fusion have reported fusion mass and vertebral compression fractures following instrumentation removal.^38–40

Successfully treating a fracture in the setting of prior spinal fusion can be challenging for multiple reasons, including substantial alteration of anatomy. Kim et al. described this challenge but also demonstrated the effectiveness of pedicle screw fixation with only a 2% malposition rate.⁴¹ Furthermore, two long fusion arms above and below a fracture line (Fig. 3) put the patient at a greater risk for displacement and nonunion. Hence, we opted to implement posterior pedicle screw fixation extending two levels above and below to achieve rigid fixation and fusion. This technique is widely supported from both a clinical and a biomechanical perspective.^42,43

Lessons

In select cases, removal of symptomatic posterior instrumentation for patients with idiopathic scoliosis may be warranted. The potential for fusion mass fracture and associated complications exists. Patients should be counseled about the case-specific risks and benefits.

Disclosures

Dr. Polly reported consulting fees from Medtronic, SI-BONE, and Globus; royalties from SI-BONE and Springer; and institutional research support from Medtronic, SI-BONE, Mizuho OSI, and AO Spine.

Author Contributions

Conception and design: Hendricks, Polly. Acquisition of data: Hendricks, Polly. Analysis and interpretation of data: Hendricks, Polly. Drafting the article: Hendricks, Geisness. Critically revising the article: all authors. Reviewed submitted version of manuscript: Hendricks, Polly. Approved the final version of the manuscript on behalf of all authors: Hendricks. Study supervision: Hendricks, Polly.

Correspondence

Cale J. Hendricks: University of Minnesota, Minneapolis, MN. hend0643@d.umn.edu.