Instrumentation Complications

Introduction

The evolution of spinal instrumentation from simple wiring, to segmental, rigid fixation, to motion-preservation and dynamic implants has been powered by an improved understanding of spinal disease and its impact on the biomechanics of the spine. Coincident with that understanding have been improvements in metallurgy, perioperative spinal imaging, and a wealth of clinical outcomes data. Technical advances have also improved surgeons' ability to place implants in a safer and less morbid manner. These advances include imaging modalities from fluoroscopy to navigation to intraoperative axial imaging. Neurologic monitoring and tubular and minimally invasive retractor systems have also affected the incidence of soft tissue complications after spinal implant surgery. Overall, these advances affect the types of problems that can be addressed, procedural morbidity and the potential for medical complications, postoperative bracing or activity restrictions, and duration of muscle and overall surgical recovery.

These new implant categories offer a host of potential benefits to the patient. However, in some cases, our reach for more powerful spinal stabilization has exceeded our grasp and subjected our patients to a host of often unforeseen complications. More frequently, though, the plethora of treatment options has challenged surgeons' in-depth understanding of the specific goals for which the implant was designed. Simply put, even perfect placement of the wrong device will not afford the patient the optimal chance at recovery. Additionally, new tools have allowed us to treat conditions or patient populations previously considered unsuitable for surgery. The combination of higher-risk conditions, older and sicker patients, and a logarithmic increase in the number of implants placed ensures that implant-related complications remain an important topic for spine surgeons and learners.

This chapter highlights changing our evolving understanding of means to prevent and address these complications. For example, previous generations of spine surgeons were taught that “exposure is key.” When difficulties in implant placement were encountered, increased dissection to improve access to and visualization of the surgical field was the most immediate means to reduce the risk of misplacement. While proper exposure and knowledge of spinal anatomy remain key features of a successful surgery, new tools have allowed safe implantation with a more limited exposure using image guidance. Additionally, the surgeon is assumed to have a general understanding of human spinal anatomy; however, optimal implant strategies increasingly require a detailed preoperative assessment of the specific patient's anatomy as gleaned from planar and axial imaging and often incorporating several modalities (e.g., radiographs and magnetic resonance imaging [MRI] or computed tomographic [CT] myelography and angiography). Another evolving critical concept in complication avoidance lies in suitable selection of the operative candidate, followed by preoperative optimization of that patient. Weight loss, smoking cessation, blood sugar control, and osteoporosis management are far more effective means of reducing global complication rates than any single change in intraoperative technique or modality.

Definitions and Classification

Complication can be a loaded term. It has different technical and emotional import for the patient, the surgeon, the medical board, the legal community, and other stakeholders. While a number of definitions are readily available, most are inadequate, from a technical perspective, to demarcate issues along a spectrum from normal sequelae of the procedure to criminal negligence.

For example, some issues, while unfortunate, are a necessary correlate to the implant itself. These include soft tissue dissection and displacement. Even here, however, surgical morbidity may be affected by patient vulnerability, surgical technique, and implant bulk and placement. Most spinal implants span motion segments decreasing overall motion and increasing the loading on adjacent segments. However, proper surgical decision making may minimize the number of fusion levels. Intraoperative technical factors may affect the impact on adjacent segments (e.g., placement of the implant too near the adjacent disc space anteriorly or violation of the superior facet joint posteriorly by a pedicle screw). Interestingly, however, postoperative complaints of stiffness are not common in patients undergoing surgery for degenerative or deformity conditions. In fact, in patients undergoing single-level fusion for degenerative disease, most reported less stiffness postoperatively.

In the pediatric population, spanning a motion segment may impact subsequent growth. In the context of congenital and early-onset deformity correction, arthrodesis and implant placement may impact later growth, including the crankshaft phenomenon, uneven growth of the trunk, and especially restrictive lung disease from inadequate thoracic growth. Recently, increased attention has been focused on growing rods and VEPTR (Vertical Expandable Prosthetic Titanium Rib; DePuy Synthes). One recent study found that addition of a cross-link did not affect spinal canal area. Another recent study found acceptable cervical alignment and growth within the fused areas in 40 children undergoing rigid atlantoaxial or occipitocervical fusion.

Some issues are not ubiquitous but are extremely common and difficult to avoid, such as occipital numbness after placement of C1 lateral mass screws. Many are less frequent occurrences over which the treating physician may have limited control. A common example here is deep vein thrombosis after implant-related spine surgery in general, but also more specifically after placement of anterior lumbar cages or disc replacements. Careful exposure and handling of the iliac veins reduces, but does not eliminate, this risk. Wound infections and pseudarthrosis also fall into this category.

No spine surgeon likely has a 100% accuracy rate with transpedicular fixation in the lumbar spine. At some point, however, the “miss rate” exceeds an acceptable level. That level has changed as newer technologies, such as intraoperative navigation, have become available to assist the surgeon in difficult cases with limited landmarks. What is “acceptable” also varies by venue. In a medicolegal setting, a community standard is often used. But do we really have good data about these miss rates in a community setting?

Technical misadventures may also occur with inadvertent injury to the surrounding neurologic or vascular structures during implant placement. Some issues exceed even the term “complication.” Major failures in surgical planning or technique, operating outside the scope of one's training, recommending surgeries for financial gain when little patient benefit can be expected, and an absence of informed consent may all reach the level of assault. Certainly, attorneys have been using pathways other than malpractice claims at an increasing rate.

To avoid this type of conflation, Sokol and Wilson describe a four part definition of “surgical complication” :

• 1.

  A surgical complication is any undesirable and unexpected result of an operation. A scar need not be a complication. This will depend on whether it was expected. Our revised definition, however, does not specify the subject of the unpleasant result; hence a surgeon's needlestick injury would constitute a surgical complication. The solution is to specify the recipient of the surgical complication.

• 2.

  A surgical complication is any undesirable and unexpected result of an operation affecting the patient.

• 3.

  A surgical complication is any undesirable and unexpected result of an operation affecting the patient that occurs as a direct result of the operation.

• 4.

  A surgical complication is any undesirable, unintended, and direct result of an operation affecting the patient that would not have occurred had the operation gone as well as could reasonably be hoped.

Most spinal implant-related complications arise from failures in biology or biomechanics or from errors in surgical strategy or technique. Often, the most severe complications arise from a combination of these issues. Some complications require a failure in more than one area. In others, a deficit in one area can increase the risk of a complication in another.

Ultimately, implant failure rates are tied to the patient's underlying diagnosis. In one series of 289 patients treated for spinal metastatic disease, a 10.7% reoperation rate was reported. Of these, instrumentation failure accounted for 26%. Another recent study found that durotomy was associated with a 2.2 times higher risk for pseudarthrosis.

Implant-related complications vary with different patient demographics, including osteoporosis, spinal deformity, and previous surgery. Fujimori and others compared the safety and complication rates of pedicle screw placement between patients in three age-based cohorts: 0 to 5 years, 5 to 10 years, and 10 to 15 years of age. In total, 5054 screws were placed, although far more were placed in the older age group (4219) than in the two younger cohorts. Although there were no neurologic complications associated with screw placement in this study, the overall complication rate ranged from 0.1% to 0.6% per screw, with the rates increasing as patient age decreased. Here, too, risk factors may span several of the groups in the failure taxonomy above. For example, osteoporosis or other bone quality issues, which are increasingly common in the aging population, confer risks across the spectrum from the biology of bone healing to the mechanics of spine fixation to technical issues encountered intraoperatively (poor bone quality limits tactile feedback and challenges intraoperative imaging, thereby increasing the risk of implant misplacement).

Biologic Failure

Biologic failure implies a disruption in the physiologic processes associated with healing from surgery. This disruption affects the ultimate success of the procedure and subjects the patient to possible revision surgeries or abject surgical failure. Common biologic issues complicating spinal implant surgeries include infection and other inflammatory issues, osteoporosis, and vascular issues. These issues may, singly or in combination, increase the risk of several adverse outcomes, including pseudarthrosis, implant loosening, bone failure and fracture, and adjacent-segment problems. These biologic risk factors are often driven by the individual patient's medical history. However, choices in spinal implants also have an effect on some of these phenomena ( Fig. 98.1 ).

Bone Quality and Osteoporosis

As with infection, poor bone quality can increase the risk of implant-related complications in several ways. Most importantly, poor bone quality requires careful consideration of operative strategy. Osteoporotic patients, for example, are often more fragile and not good candidates for larger, multisegmental spinal reconstructions. Unfortunately, from a mechanical point of view, these patients often require multipoint fixation and more aggressive correction of spinal alignment problems to avoid kyphotic collapse, adjacent-segment fracture, or implant failure.

With the aging of the population, type 2 osteoporosis is becoming more common. However, changes in hormone replacement practices have also increased the incidence of postmenopausal (type 1) osteoporosis. Often ignored are secondary causes of bone loss from other medications or medical problems (steroids, hyperparathyroidism). Such causes of type 3 osteoporosis are less responsive to medication management.

Although newer medications offer patients better bone-building ability, these agents have among the highest noncompliance rates of any medication class. Even after the “bone decade,” a sizable proportion of at-risk patients still make it to surgery without a preoperative assessment of bone mineralization.

At a minimum, at-risk patients should have a dual-energy absorptiometry scan, although surgeons should carefully consider the trabecular architecture and cortical thickness on a preoperative CT scan as well. Surgeries can often be delayed to allow bone building. Perioperative administration of bisphosphonates, denosumab, and teriparatide have been studied in spinal fusion patients. Teriparatide has been found to be most effective and can offer clinically relevant improvements in bone quality over 6 to 12 months of use.

With a sense of the patient's bone quality in mind, potential changes in surgical strategy should be undertaken. In some elective conditions, the decision to move forward must be tempered by a realistic appraisal of the impact of fragility on the intended procedure's risk/benefit balance. In others, a delay to address bone quality is recommended. Such a delay is not likely to provide net gains for patients markedly incapacitated or nonambulatory due to their spinal pathology. In these patients, modifications to the surgical strategy must be considered. These considerations include types of bone anchors, number of anchors used, polymethylmethacrylate augmentation of screws, and avoidance of anterior/interbody procedures.

Bone quality will limit the surgeon's ability to achieve distraction intraoperatively; in general, in the face of significant deformity, a spinal column–shortening approach is favored. Significant preload on bone anchors must be avoided ( Fig. 98.2 ). Selection of motion-sparing approaches is compromised by early failure of the bone-implant junction.

Intraoperatively, radiographic visualization may be compromised as well. The surgeon must be ready to abandon percutaneous procedures if the proper landmarks cannot be visualized. Osteoporotic bone is vascular. Late decortication, bone wax, blood product support, Cell Savers, and other techniques have been recommended to support these patients.

Postoperative concerns in osteoporotic patients include activity restriction and bracing to reduce implant stress. Osteoporotic bone likely progresses to union more slowly than healthy bone, but, even when fused, may be subject to higher bending forces. Implant removal is therefore discouraged, as late deformity progression and compression fractures have been described even within solid arthrodeses.

Dysvascular Change

Vascular injury is a concern with implant misplacement, but subtler vascular embarrassment may lead to biologic failure of a spinal construct. Implant placement typically requires greater exposure and devascularization than decompression- only procedures. This devascularization may be particularly important in smokers and in surgeries in a previously irradiated bed ( Fig. 98.3 ).

Biomechanical Failure

As with biologic failure, biomechanical failure exists at the nexus between the patient's pathophysiology and the goals of the reconstructive procedure. Thus, this failure can arise from errors in surgical planning or from patient comorbidities. Biomechanical failure is typically expressed through one of several modes of failure, such as cage subsidence, rod fracture, or screw loosening or fracture.

Cage subsidence occurs when the loads through the cage exceed the carrying capacity of the bone on which it rests. Frequently, this failure is seen in patients with poor cancellous bone quality or when a hard cage subjected the host bone to a marked modulus of elasticity mismatch. Preoperatively, prevention may include selection of a wider cage supported by a larger portion of the vertebral endplate, especially the cortical ring. However, subsidence is seen even with wide lateral interbody fusions. In one recent study, 30% of patients had either grade I or II subsidence.

Intraoperatively, careful endplate protection during preparation and placement will prevent the cage from sitting on softer subchondral bone. Postoperative loading should also be considered at each stage. Greater activity restrictions or bracing may be considered in larger patients. Cage failure is also more likely when there is ongoing instability in planes other than axial loading. For example, posterior tension band failure leads to unloading of the posterior load-bearing columns and overloading of the endplate. This could be addressed with separate, posterior instrumentation.

Although late implant failure occurs because of failure of fusion or bone healing, early failure typically occurs for implant overloading. As such, early implant failure is typically related to failure to understand the direction or degree of spinal instability or errors in surgical strategy. For example, rod fractures are most common in stiff, adult spinal deformities at the level of spinal osteotomy ( Fig. 98.4 ). In one recent series of 75 adults undergoing fusion for adult spinal deformity, 9.3% of those who had a pedicle subtraction osteotomy had a rod fracture compared with 2.6% of those who had a Smith-Petersen osteotomy. The authors identified the following additional risk factors: pseudarthrosis at 1 year, sagittal rod contour greater than 60 degrees, presence of dominoes and/or parallel connectors, and fusion construct crossing both thoracolumbar and lumbosacral junctions.

Unlike rods, bone anchors may fracture, loosen, or displace from the bone. In addition, polyaxial pedicle screws can exhibit disengagement between the screw and its tulip. Often motion or a 1-mm radiographic halo around the screw is used to define loosening. One recent study found that a change of 1.9 degrees of angulation between the pedicle screw axis and cranial endplate was 75% sensitive and 89% specific in detecting screw loosening on 6-month radiographs ( Fig. 98.5 ).

In late failure, bone healing and fixation failure are typically described in terms of a “race.” Pseudarthrosis is often a multifactorial problem that involves the patient's biology and pathomechanics as well as the mechanical attributes of the fixation methodologies used. In some cases, the issue lies in the patient's biologic ability to heal in a reasonable time frame. The surgeon may choose to address this issue by adding autologous bone graft or bone morphogenetic protein (BMP). BMP may be more effective in some areas compared with others. A higher than expected fusion failure rate of 10.8% was reported using recombinant human bone morphogenetic protein 2 (rhBMP-2) in a pediatric cohort undergoing occipitocervical and C1–C2 fusion surgery. The same author group reported no increase in the risk of cancer at 4-year follow-up in 57 consecutive cases of rhBMP-2 use.

In other patients, the problem lies in adequate stabilization of the motion segment. Depending on the patient population, pseudarthrosis leading to implant failure has become less common as fixation methods have evolved. In a series of 227 AIS patients undergoing posterior spinal fusion, a 2% rate of implant failure was noted. Often, conversion to a more rigid form of fixation, such as transpedicular instrumentation or interbody cage placement, will successfully address the pseudarthrosis ( Fig. 98.6 ). Similarly, after anterior cervical discectomy, a posterior wiring or clamping procedure will often lead to union anteriorly.

Even in challenging clinical scenarios, properly designed implant schemes exhibit low failure rates. In a cohort of 318 patients with spinal metastases undergoing separation surgery (which seeks to restore mechanical stability and remove only that part of the tumor in contact with the spinal cord). In this series, only 2.8% exhibited symptomatic implant failure that required revision. Of those, failure was more common in women, in cases involving more than six contiguous levels, and in patients who required concomitant chest wall resection.

Screw loosening is not necessarily diagnostic for fusion failure. Screws can loosen as part of a “dynamization” process. Additional issues complicate the biology of the bone-implant junction. For example, diabetics carry a higher risk of aseptic implant loosening. These issues are more critical in dynamic stabilization approaches and likely decrease with solid segmental fusion. Loose or fractured implants may migrate into other anatomic regions. Implant migration into the retroperitoneal space has been reported 6 years after instrumentation without fusion. The authors cautioned against implant placement without concomitant spinal fusion.

These studies are complicated by the limitations of radiographic diagnosis of pseudarthrosis. Spinal implants often impede clear visualization of bone healing, especially stainless steel or trabecular metal. A recent review noted the limitations of radiographic studies in identifying pseudarthrosis due to the high rate of asymptomatic pseudarthroses and the number of conditions, such as adjacent-segment degeneration, that may mimic the symptoms of pseudarthrosis. Dickson and colleagues, in a series of 171 consecutive pedicle subtraction osteotomy patients, identified 18 pseudarthroses, with two diagnosed more than 5 years after the index surgery.