Background

Lumbar spine instrumentation and fusion is a commonly indicated treatment for infection, tumors, trauma, deformity, and degenerative disease. With broad indications and continued improvement in surgical techniques and technology, there has been a tremendous increase in the number of lumbar spine fusions performed in the last decade. Unfortunately, these surgeries can result in unplanned complications such as proximal disease, failure of fusion, and loss of motion with associated adjacent level disease, all of which can be problematic for patients and the treating physician. The purpose of this chapter is to discuss each of these fusion-related complications and provide some insight into how to properly manage them.

Proximal Junctional Kyphosis

With the advent of modern instrumentation and selective fusions, junctional kyphosis at the transition from fused to mobile segments may be a common radiographic finding. After long-segment spinal instrumentation, proximal junctional kyphosis (PJK) can be seen as a postoperative complication ( Fig. 55.2 ). In adult spinal deformity surgery, the reported incidence ranges from 11.0% to 52.9%; however, the description and criteria for defining PJK and its clinical impact vary in the literature. PJK has traditionally been defined by a 10 degree or greater increase in kyphosis at the proximal junction as measured on a sagittal radiograph with a Cobb angle from the caudal end plate of the uppermost instrumented vertebrae (UIV) to the cephalad end plate of the vertebrae two segments cranial to the UIV. This measurement was reported by Glattes et al. and has been validated in several subsequent studies. Although PJK, as defined above, may have a high postoperative occurrence rate, the significance of PJK remains debatable. Some reports indicate that PJK is only a radiographic phenomenon with little clinical significance, whereas others indicate that PJK is associated with significant pain, neurologic deficit, and need for revision surgery. A recent review of causes for hospital readmission after adult deformity surgery reported that PJK was the most common postoperative complication requiring surgical treatment in a cohort of 836 patients.

Despite some benign reports of PJK, investigators have recognized a subset of patients with a more severe version of PJK and an increased need for revision surgery. In addition to increased deformity and pain, these patients are at an increased neurologic risk. The term proximal junctional failure (PJF) is used to define this group and to distinguish it from PJK. PJF is associated not only with an increase in kyphosis but also with structural failure. The structural failure occurs at either the UIV or the vertebrae immediately proximal to the fusion construct (UIV +1). Structural failure is considered a vertebral body fracture, disruption of the posterior osseoligamentous complex, or both. Unlike traditional PJK, PJF has a clear association with increased pain, spinal instability, risk of neurologic injury, and need for revision surgery.

Risk factors for developing PJK/PJF have been identified and include advanced patient age, poor bone quality, posterior spinal ligamentous disruption, instrumentation rigidity, fusion to the sacrum, and postoperative spinal alignment. With respect to adult patients, Kim et al. found age over 55 to be a significant risk factor for the development of PJK. They also found combined anterior/posterior fusions to have higher PJK rates than posterior-alone procedures. Yagi et al. found higher rates of PJK in patients with fusion to the sacrum. Some cases of PJK may develop from structural damage. Studies on adolescent patients have shown an increased incidence of PJK with posterior instrumentation compared with an anterior approach. The conclusion was that disruption of the posterior tension band from surgery and deformity correction forces applied during surgery results in an increased incidence of PJK. Ligamentum flavum damage has also been attributable to PJK progression. Vertebral body fractures have been associated with the mechanical stress generated by pedicle screw instrumentation at the proximal junction. This has been shown to be higher with inclusion of the sacrum. Yagi et al. also associated fusion of the sacrum with higher PJK rates. Both of these reports support the idea that correction forces, and possibly surgical dissection, contribute to PJK and PJF.

The occurrence of PJK is most prevalent in the early postoperative period. Kim et al. reported PJK to be most common in the first 8 weeks after surgery with 59% of patients progressing in this time. In a review of adolescent idiopathic scoliosis patients, Kim et al. found a lower prevalence of PJK (26%) with no significant progression from 2 years on. It appears that the most dramatic progression of PJK occurs in the first 3 to 6 months.

Patients who undergo greater sagittal realignment are also at higher risk of PJK/PJF. In a study by Hart et al., patients who experienced PJF had a greater number of pedicle subtraction osteotomies, an increase in lumbar lordosis, pelvic-incidence and lumbar lordosis mismatch reduction, and sagittal vertical axis (SVA) correction. More importantly, from a clinical perspective there is evidence that patients with PJF do worse clinically than patients without this complication. Hostin et al. reported on their experience with 1218 adult deformity surgeries. There were 68 cases of PJF (5.6%) with 28 of those patients undergoing revision surgery within 6 months of the operation. Patients undergoing revision surgery were identified with PJF on average 9 weeks after the initial surgery compared with 13 weeks for patients who did not have revision surgery. This again demonstrates the impact of mechanical failure as opposed to recurrent deformity. There is a clear clinical significance with the occurrence of PJF given the frequent need for extension of instrumentation.

Several strategies have been proposed to mitigate the occurrence of PJK/PJF, including preservation of the posterior ligament complex and adjacent facet joints, use of vertebral augmentation with polymethyl methacrylate (PMMA) to prevent vertebral compression fractures, and use of less rigid fixation at the proximal terminal junction of the construct with transitional rods, transverse process hooks, and dynamic stabilization techniques. Cadaveric models have shown a reduced incidence of junctional fractures with vertebroplasty. Kebaish et al. demonstrated this with 18 cadaveric spine specimens in an axial loading model showing 5 out of 6 fractures in the control group, 6/6 with cement at UIV, and only one fracture in the UIV +1 cement group. This has also been demonstrated in clinical practice with a report of 8% PJK rates and 5% PJF rates in 41 patients treated with prophylactic UIV +1 vertebroplasty. Additional data is needed, but this treatment may prove to be advantageous in reducing PJK rates. There has been no good data to support transitional rods or the use of hooks over pedicle screws at the upper levels of instrumentation. Percutaneous instrumentation has also not produced a decrease in PJK over open screw placement.

One technique that does seem to offer the potential to reduce the occurrence of PJK/PJF is the use of junctional tethers. Using a polyester tether such as woven polyethylene Mersilene tape (Ethicon, Somerville, NJ), the spinous processes are interwoven in an attempt to dissipate forces at the proximal junction. Although their clinical effectiveness remains under investigation, Bess et al. recently provided a finite element analysis that demonstrated that posterior tethers created a more gradual transition in range-of-motion and adjacent-segment stress from the instrumented to noninstrumented spine. There are numerous techniques for the application of these tethers. Most simply, the spinous processes of the UIV +1 and UIV −1 are tethered together by passing the Mersilene tape through each and tying it together. Recently at our institution, we have begun tying the Mersilene tape from the UIV +1 to a crosslink and creating distraction to increase tension on the Mersilene tape ( Fig. 55.1 ). In a retrospective review, our data shows a 40.8% PJK rate with no tether, 34.3% with standard tether technique, and 19% with use of a crosslink. This is pending publication.

(A) Illustration of Mersilene tether tied from the UIV +1 spinous process to a crosslink device. The tape is tied tight with the crosslink in a cephalad position. (B) With the Mersilene tied tightly, the crosslink is distracted, putting increased tension on the tether and adding support to the transition from instrumented to noninstrumented segments.

(A) Preoperative standing lateral radiograph showing loss of sagittal balance, lordosis, and proximal junctional kyphosis (PJK) with a pelvic incidence (PI) 49 degrees, pelvic tilt (PT) 33 degrees, lumbar lordosis (LL) 2 degrees, and thoracic kyphosis (TK) 15 degrees and an 11-cm SVA. (B) T10 to iliac with L2 PSO postoperative standing lateral radiograph showing PJK. Measurements include PI 49 degrees, PT 35 degrees, LL 50 degrees, and TK 59 degrees and a −1-cm SVA. Kyphosis is 50 degrees across the PJK segment. (C) T4 to iliac extension postoperative standing lateral radiograph. Measurements include PI 49 degrees, PT 30 degrees, LL 49 degrees, and TK 46 degrees and a −1-cm SVA. Restored sagittal alignment.

Current management for PJK involves both conservative and surgical interventions. When diagnosed radiographically, PJK may be followed with routine imaging and close observation. Some patients will remain asymptomatic. Other patients may be managed by pain medications, physical therapy to strengthen the back, or epidural injections. We have found aquatics-based physical therapy programs to be most effective because they unload the spine and allow patients to increase their activity level with reduced pain. Surgical management when needed involves extension of the fusion superiorly to correct the deformity. Regardless of where the fusion ends, PJK will remain a risk factor.

Nonunion

Pseudarthrosis is a well-reported complication of lumbar spine surgery. Its diagnosis is based on appropriate clinical history and imaging findings or implant failure, loss of fixation, deformity, or radiolucencies. However, the presentation is unpredictable, and it can occur up to a decade postoperatively despite the presence of solid bone formation at earlier time points. Surgery for spinal deformity correction relies heavily on arthrodesis across several levels of the spinal column. To achieve a solid fusion, locally harvested autologous bone graft is used along with supplemented allograft. Occasionally osteoinductive and osteoconductive agents are used as well. Fusion in children and adolescents is rarely a concern due to high bone quality. In adult patients nonunion is a real concern and must be monitored closely. Rates of pseudarthrosis after lumbar spine fusion have ranged from 5% to 35% with a higher incidence in surgeries spanning three or more levels. Pseudarthrosis should be suspected when a patient presents with recurrent pain and/or neurologic symptoms during long-term follow-up from fusion or in the presence of instrumentation failure. A mechanical exacerbation of symptoms may suggest instability at the surgical site. Diagnosis can be difficult because symptoms are not specific to pseudarthrosis but may be attributable to other causes, such as infection or adjacent segment disease (ASD). A pain-free interval in the postoperative period is a useful clue into the history. Patients with no symptom relief postoperatively should be studied further to rule out additional causes.

It is often difficult to predict when or whether a pseudarthrosis will become symptomatic for patients. DePalma and Rothman retrospectively reviewed outcomes in patients with radiographic evidence of pseudarthrosis compared with those in matched controls with a successful lumbar fusion and found no significant difference between the two groups in terms of subjective satisfaction, symptom relief, or return to activity. More recent studies, however, suggest that a solid fusion correlates with improved long-term outcomes and decreased symptom severity. Kornblum et al. reported on patients with symptomatic spinal stenosis and spondylolisthesis treated with posterolateral arthrodesis. Eighty-six percent of patients with a solid fusion had “excellent” or “good” long-term outcomes compared with only 56% of patients with a pseudarthrosis. It remains unclear why some patients can tolerate a nonunion with good long-term clinical outcomes, whereas others require surgical treatment. Local factors that can lead to nonunions include poor preparation or decortication of the fusion surface, insufficient viable graft material, vascular insufficiency, smoking, poor nutrition, or metabolic problems. Meticulous surgical preparation and adequate-quality bone graft will minimize the risk for fusion failure. Global parameters may also contribute to pseudarthrosis. Malalignment of the spine with poor sagittal balance, insufficient compression forces, and inadequate stability at the fusion site all contribute. These mechanical concerns become increasingly significant when fusions extend across transition zones such as the lumbosacral junction. In the adult population, a correlation between poor postoperative sagittal alignment and pseudarthrosis has been noted.

The assessment of fusion can be difficult. Plain radiographs are often the initial assessment for pseudarthrosis and other diagnoses given their availability and relatively low cost, but the radiographic presentation of nonunions can vary. In a study using plain radiographs, Kim et al. found an average of 3.5 years (range: 12–131 months) before fusion failure could be detected. In a similar study, Dickson et al. showed that out of 18 patients with known pseudarthrosis, only 13 (72%) were detected by radiographs 2 years postoperatively. This data along with other findings suggests that annual radiographic follow-up should be implemented for multilevel fusions even if bony union is apparent at early time points. Computed tomography (CT) imaging has the strongest correlation of fusion assessment and should be obtained if a nonunion is in question. Although there are no universally accepted criteria for pseudarthrosis with an interbody fusion, most studies use the following to identify a nonunion: motion on dynamic films, absence of continuous trabecular bone between adjacent vertebrae, gas in the disc space, and periimplant radiolucency. Current radiographic guidelines for successful lumbar fusion include less than 3 mm of translation motion and less than 5 degrees of angular motion on flexion and extension radiographs. As CT technology advances, so does our ability to detect pseudarthrosis. Shah et al. reported bridging trabecular bone to be appreciated on 95% of thin-section CT scans compared with 4% of plain films 6 months postoperatively. These authors suggest that thin-section CT should be the modality of choice for the detection of pseudarthrosis.

Additional methods used to detect pseudarthrosis include bone scintigraphy and positron emission tomography (PET) scans. Bone scintigraphy uses radiographic tracer to localize tissue with high metabolic activity (indicating active tissue changes or repair). This is more commonly used to detect infections, neoplasm, and occult fractures. It currently remains a poor choice for the detection of pseudarthrosis due to poor sensitivity. Similarly, PET scans detect gamma ray emission from positron-emitting radioactive tracers, which have an affinity for metabolically active cells. It has recently been suggested that tracers, which are used more commonly for detection of infections and neoplasm, can also measure bone graft healing by correlating increased uptake at the fusion site. Although studies have shown this to be a modality for monitoring active bone formation, little data exists on clinical applications and correlation with rates of nonunion.

The treatment of pseudarthrosis varies but is almost always surgical. In cases of asymptomatic patients, they may be observed and followed closely with radiographs and routine evaluation. When symptomatic, patients will often experience pain in the axial spine with occasional radicular symptoms. A delayed fusion with no evidence of instrumentation loosening may be treated with bracing, activity limitation, and observation. Primary principles of surgery include stabilization of the existing posterior fixation and regrafting. Treatments may require a circumferential fusion with anterior lumbar interbody fusion (ALIF) or lateral lumbar or posterior lumbar interbody fusion. Interbody devices allow for increased surface area under compressive forces, creating an ideal environment for fusion. Advances in biologics, allograft materials, and growth factor augmentation have all improved arthrodesis. The ideal material demonstrates osteoinductive, osteoconductive, and/or osteogenic properties. Iliac crest bone graft has been the gold standard autograft bone material; however, complications with harvest and limitations in supply have led to the development of additional agents. A review of posterolateral lumbar fusion rates reported iliac crest fusion rate of 79%, allograft bone 52%, ceramics 87%, demineralized bone matrices 89%, autologous bone marrow 74%, and bone morphogenetic properties 94%. Implants must be solidly anchored and screws increased in size as needed. Osteoporotic bone requires segmental fixation, and when a fusion extends to the sacrum, supplemental instrumentation with iliac screws is often required.

The best treatment for pseudarthrosis is to prevent it from occurring at the initial operation. Improvements in bone graft materials, instrumentation, and techniques have all led to decreased rates of nonunion. Treatment of pseudarthrosis has also benefited from these advancements. Another critical aspect of pseudarthrosis prevention is the assessment of the patient’s preoperative condition. Risk factors such as alcoholism, osteoporosis, advanced age, malnutrition, and tobacco use have all been attributed to decreased fusion rates. During surgery for pseudarthrosis, emphasis must be placed on bone preparation and arthrodesis. This involves aggressive removal of fibrous tissue, renewed grafting with autologous bone, and revision instrumentation when necessary. Poor alignment must be addressed with evaluation of sagittal parameters. This may involve osteotomies or corrective maneuvers. The importance of proper alignment is increased with the length of fusion and extension across junctional zones.