 |
 |
 |
|
| home | CME | help | feedback | subscriptions | archive |
|
© 2000 The Journal of Bone and Joint Surgery, Inc.
The Use of Pedicle-Screw Internal Fixation for the Operative Treatment of Spinal Disorders*
Investigation performed at the Columbia Orthopaedic Group,
Columbia Spine Center, and the Department of Orthopaedic Surgery,
University of Missouri Health Sciences Center, Columbia, Missouri
Pedicle screws have dramatically improved the outcomes of spinal reconstruction requiring spinal fusion. Short-segment surgical treatments based on the use of pedicle screws for the treatment of neoplastic, developmental, congenital, traumatic, and degenerative conditions have been proved to be practical, safe, and effective. The Funnel Technique provides a straightforward, direct, and inexpensive way to very safely apply pedicle screws in the cervical, thoracic, or lumbar spine. Carefully applied pedicle-screw fixation does not produce severe or frequent complications. Pedicle-screw fixation can be effectively and safely used wherever a vertebral pedicle can accommodate a pedicle screw - that is, in the cervical, thoracic, or lumbar spine. Training in pedicle-screw application should be standard in orthopaedic training programs since pedicle-screw fixation represents the so-called gold standard of spinal internal fixation. Pedicle screws have revolutionized the surgical treatment of spinal disorders, although their introduction and widespread adoption by spinal surgeons has created one of the most difficult regulatory problems ever seen in orthopaedics.
Hadra65 first used silver-wire internal fixation for the treatment of a cervical fracture-dislocation and tuberculous spondylitis in 1891. Later, King84 introduced facet screws for the treatment of degenerative lumbar conditions. Although Boucher17 has been widely credited with the first use of pedicle screws in North America, his report suggests that his innovation was a longer facet screw that occasionally obtained oblique purchase across the pedicle. His screws were not aimed down the long axis of the pedicle. Thus, it seems that Harrington and Tullos68 deserve credit for the first deliberate attempt to put pedicle screws through the isthmus of the pedicle. Their report, published in 1969, described the attempted reduction of two cases of high-grade spondylolisthesis. Pioneering the use of pedicle-screw internal fixation proceeded in France and Switzerland during the 1970s. Clinical success with the screws was reported during the 1980s by investigators including Cotrel and Dubousset31, Dick40, Roy-Camille et al.137,138, and Louis99. A strong stimulus to North American use of the pedicle-screw technique was the presentation by Roy-Camille at the 1979 American Academy of Orthopaedic Surgeons meeting in San Francisco. Subsequently, a number of American surgeons began to use pedicle screws in the United States. Arthur Steffee used them most creatively. Steffee et al.152 developed the variable-screw-placement (VSP) plate, which permitted pedicle screws to be placed according to individual patient anatomy. This device provided much more clinical latitude than the Roy-Camille plate137,138, which had fixed screw-hole distances for application of the screws (Fig. 1).
Recognizing the clinical utility of pedicle-screw fixation, the North American Spine Society and the Scoliosis Research Society collaborated on a series of introductory meetings to educate spinal surgeons about this new technique. The first of these meetings was held in 1984. These two societies subsequently mutually sponsored the International Meeting on Advanced Spine Techniques, and this annual meeting continues to be held to the present day. Since the introduction of pedicle screws, engineers, surgeons, radiologists, neurophysiologists, anatomists, epidemiologists, and statisticians have made fundamental efforts to improve what was recognized universally as a clinical liability in surgical spine care: the lack of truly high-quality spinal internal fixation similar to what was available for internal fixation of long bones. The basic and clinical science of pedicle screws developed from their introduction and infancy in the early 1970s to widespread acceptance, as indicated in "State-of-the-Art Treatment" statements by the North American Spine Society in 1993 and 1996, which endorsed the use of pedicle screws by experienced surgeons. Review articles described the evolution of treatment methods as surgical experience accumulated33,48,62,114,158.
The anatomy of the human pedicle has been studied exhaustively in different races6,25,26,73,83, in children50,108,140, and in adults123,146. Patterns of pedicle anatomy unique to the cervical spine44,81,82, the thoracic and the lumbar spine45,46,109,126,132,146,148,159,160, and the sacrum47,94,120,172,173 have been clearly identified. Measurements of the outer and inner diameters of the pedicle have been performed extensively71,121. The inner diameter of the pedicle - the critical surgical dimension - has been shown to be more directly related to the height of the patient than to the gender81. However, wide individual variations within common patterns of anatomy are the rule. This forces the surgeon to understand the individual anatomy of the patient, in order to achieve clinical success, and to appreciate the patterns of pedicle anatomy that are common to the human race in general. There do not seem to be more than modest differences in pedicle anatomy from race to race6,25,26,73,83.
In 1986, we52,53 clearly demonstrated the improved quality of spinal internal fixation provided by a pedicle-screw implant with a fixed link to a plate, in the first nonfailure stability testing done for spinal fractures. We also showed the fundamental importance of load transfer across the spinal column itself along with the implant - a concept known as load-sharing53, which was subsequently widely adopted by the spinal surgery community. Biomechanical tests of pedicle-screw constructs have demonstrated the fundamental importance of the bone-implant interface, bone density, and screw pullout strength11,28,63,66,87,179. The essential need for fit and fill of the screw in the isthmus of the pedicle has been proved19,88,106. The direct relationship between pullout strength and insertional torque has been well demonstrated178, and the fundamental improvement in pullout strength obtained by cross-linking39,64,100,130 (that is, attachment of the two screws in an individual vertebra with use of a metal cross-link) has been documented. The stabilizing influence of using converging screws (a high pedicle angle [Fig. 2]), if possible, also was emphasized9,23,139.
The internal forces and moments of bilevel constructs have been measured exhaustively with strain gauges on all components of rod and plate-based constructs, and the fundamental importance of load-sharing through the vertebral column has been reemphasized. Without load-sharing, an unstable four-bar mechanism was clearly demonstrated when all four pedicle screws were parallel (Fig. 3)23. In addition to in vitro testing, finite-element modeling has been used extensively59,96 to analyze stress transfer through implants as well as through implant-based constructs.
The bone-mineral density of vertebrae has been evaluated exhaustively with densitometry. Japanese investigators demonstrated conclusively that patients with multiple spontaneous compression fractures were very poor candidates for pedicle-screw-based internal fixation because their bone-mineral density was poor (one-fifth of normal bone density) preoperatively (Fig. 4)151. Other authors have strongly suggested the routine use of bone-mineral-density testing before pedicle-screw-based surgery, although mild osteoporosis (bone-mineral density of less than 100 milligrams per milliliter) did not seem to affect screw purchase or clinical outcome adversely87.
The success of hook-based constructs compared with that of screw-based constructs was shown to be based on bone-mineral density28,112. In very osteoporotic specimens, hook-based claw constructs performed as well as screw-based constructs28. A few authors have suggested use of both a screw and a hook at the same level111,174. The enthusiasm for pedicle-screw fixation among spinal surgeons created a group of engineering collaborators who quickly became extremely knowledgeable regarding the engineering issues related to structural and clinical success. A model for fatigue testing that is now widely used was adopted by the American Society for Testing and Materials (Fig. 5)34,60. A corpectomy model is created with two polyethylene blocks to mimic vertebrae, which are attached to a standard Instron tester (Instron, Canton, Massachusetts). This model can be used for subcomponent or construct testing.
The major diameter of a pedicle screw has been shown to control pullout strength (Fig. 6)170,179. Bicortical purchase both in individual vertebrae and in the sacrum increased pullout strength fundamentally152, although bicortical purchase, except at the sacrum, has not been widely adopted by surgeons because of the risk of injury to the aorta or vena cava. Bicortical sacral purchase, however, has been proved to be extremely safe and has gained widespread acceptance. The first sacral segment has been analyzed exhaustively, and several safe and secure fixation sites have been demonstrated within it47,94,120,172,173. The anatomical liabilities of the second sacral segment also have been clearly shown.
The presence of one or more pedicle fractures has been shown to reduce, although not to completely eliminate, the integrity of pedicle-screw-based constructs85. Many different styles of mechanisms have been developed by manufacturers to attach a pedicle screw to a longitudinal member (a plate or a rod); some are quite rigid and some, less rigid. All of these designs have shown some clinical utility, although there were dramatic differences in their biomechanical performance in vitro (Fig. 7)34. The patient's body weight has been shown to be the main determinant affecting structural survival of rods used for scoliosis correction75.
Nitrite has been added to the surface of the screw to improve surface hardness97, and both titanium and steel implants have been widely investigated, manufactured, and used. Titanium implants are used for patients who are allergic to nickel, chromium, or cadmium or for those who may require frequent postoperative magnetic resonance imaging or computed tomographic scans43,128,163.
Roy-Camille et al.137,138 suggested that a pedicle screw should be introduced by drilling the path and then applying the screw. Most American surgeons realized the danger of this approach; they adopted a blunt technique to identify the pedicle and routinely used biplane image intensification during the placement of pedicle screws152. The use of taps of gradually increasing diameter to assess the quality of cortical purchase through the isthmus of the pedicle, and the use of image intensification to assess the length of the screw necessary to obtain purchase in the vertebral body but not through the anterior vertebral cortex, have become standard procedures for safe screw application, resulting in strong fixation. This technique, called the Funnel Technique, is now used widely (Fig. 8).
Irrigation with saline solution down the pedicle5, and visualization of the pedicle by an endoscope51, have both been used to assess proper screw placement. Routine monitoring of somatosensory135 and dermatomal somatosensory29 evoked potentials, as well as electromyography21,27,36,37,56,72,135,136,166, also have been used for this purpose. Unpublished data have suggested the need for routine laminectomy in patients with pedicle-screw fixation, although this approach has not been adopted widely. Hertlein et al.70 described the application of pedicle screws from the anterior approach. The use of robotics1 and of computer-based guidance technology also has been investigated14,22,58,79,115,116,124,125.
Methylmethacrylate has been used in vitro and in vivo to improve the fixation of pedicle screws in vertebrae129,170,179 (Fig. 9). The improvement in pullout strength is obvious. However, the occasional spread of the cement into the spinal canal limits the successful application of this technique. The use of calcium phosphate122 has been reported, but there has been very limited acceptance of this option. The use of carbonated apatite also has been reported98.
Early clinical reports of pedicle-screw fixation demonstrated that most pedicle screws did not obtain cortical purchase (only cancellous purchase was obtained), whereas recent reports have emphasized the essential nature of cortical purchase and the direct relationship between pullout strength, insertional torque, and bone-mineral density121,178. In our practice, the use of the direct Funnel Technique to identify the pedicle isthmus permits tapping the isthmus to obtain cortical purchase for each inserted screw. Bicortical purchase is routinely used at the first sacral level but never at any other level.
The improvements in fixation at individual vertebral levels stimulated the need for improvements in spinopelvic fixation to manage spinopelvic disorders with greater security. Iliosacral screws, iliac-wing screws, intrasacral fixation, and lumbosacral end-plate fixation all have been recommended, researched, and adopted in limited applications for specific indications102,119.
With improved spinal internal fixation, surgeons have attempted to limit their use of extra autogenous bone graft to avoid morbidity at the donor site. Bone stimulators134, allograft materials18, synthetic hydroxyapatite, calcium phosphate, and coral all have been used with pedicle-screw-based constructs, with greater or lesser success depending on the pathology and the host-related variables4. Smoking has been demonstrated to be a risk factor for pseudarthrosis in some series but not others8,55.
Many surgeons have abandoned formerly routine techniques of spinal fusion such as decortication of the posterior elements or facet fusion when pedicle-screw-based implants are used. The time to healing of a scoliosis fusion was reduced from six to twelve months in the 1960s to two to four months in the 1990s because of the secure internal fixation provided by pedicle-screw-based implants67. There has been a decrease in the need for postoperative bracing of patients with constructs based on internal fixation with pedicle screws. This has been particularly notable in patients with scoliosis10,67,95,110,154,155.
Stress-shielding of bypassed vertebrae under implants that were deliberately disproportionately large has been well documented in dogs32. However, loosening of these huge implants from the host spine occurred when the bypassed spinal segment was osteoporotic. After screw-loosening, the stress-shielding effect of the disproportionately large implants was alleviated and the bone density of the bypassed segment of the spinal column returned to normal.
The limitations of plain radiographs and computed tomography in demonstrating proper screw position have been well documented54,117,165. Surgical exploration is the reference standard for determining union. Only the radiographic demonstration of trabeculation across the intertransverse (lateral) or interbody area comes close to surgical exploration with regard to accuracy in determining the presence or absence of solid union of a spinal fusion15.
Static kinematic testing has been used to examine and identify the increased motion that occurs at motion segments adjacent to spinal constructs immobilized by pedicle-screw-based implants175. Finite-element models also have demonstrated this increased motion96. Case reports30,93 of spinal fractures occurring adjacent to spinal instrumentation constructs have documented clinically the vulnerability of the spine after a violent injury. Disc pressures are substantially changed adjacent to a spinal instrumentation construct as well35,164. Other case reports have demonstrated the necessity of protecting the supraspinous and interspinous ligaments, the ligamentum flavum, and the capsular ligaments at segments adjacent to a spinal instrumentation construct, and they have shown the importance of restoring balance in the sagittal plane to avoid degeneration of adjacent segments49.
Although the initial clinical use of pedicle-screw-based implants was based on their theoretical benefits, their effectiveness in facilitating fusion now has been demonstrated conclusively55,144,145,149,177. Collaborative and individual studies8,55,167,176,177 have demonstrated statistically higher fusion rates associated with use of these implants. The functional benefits for patients also have been demonstrated conclusively with use of preoperative and postoperative testing167.
Mainly on the basis of the results of treatment of scoliosis and the desire to ensure load-sharing in the anterior column, many low-back surgeons in the United States routinely began to use circumferential fusion for all cases of degenerative spinal disease after anterior and middle-column load-sharing was demonstrated to be beneficial biomechanically55,152,156. Anterior grafting was performed either with an anterior open procedure, a laparoscopically based anterior procedure, or posterior lumbar interbody fusion at the same time as the pedicle-screw-based surgery. Some authors have suggested that the routine use of circumferential fusion improves the healing rate in smokers55 and that the union rate is better than that obtained with pedicle-screw-based intertransverse fusion alone8. The development of prosthetic devices that contain autograft or allograft has added another variable to this clinical interface that is still under investigation. An unpublished review of the results in our patients demonstrated a 92 percent union rate with pedicle-screw-based posterolateral fusion alone, without cages, using only the autograft obtained from laminectomy for the majority of patients. None of the patients had anterior surgery. The union rate in smokers and nonsmokers was the same.
Complications have accompanied the use of pedicle screws. The rate of screw misplacement has ranged from 0 percent10 to 2 percent20 to 25 percent95 in patients with scoliosis and to nearly 4.2 percent8,16,49,101,160,168 in those with degenerative diseases. A learning curve has been demonstrated for surgeons in general54. A variable prevalence of nerve-root and/or cauda equina injury has been reported to be associated with pain and sensory deficit in some patients20,55,113,167,168. The prevalence has been reported to be as high as 11 percent114 and as low as 1 percent150 to 2 percent20,113. However, most misplaced screws do not create nerve-root injury20,54,145,154. Safe removal of screws that were drilled into the spinal canal has been reported41. A high prevalence of screw misplacement associated with drilling160 rather than with the blunt technique of screw application has been clearly demonstrated20. Computer-based navigation techniques have been developed and have been suggested for clinical application14,22,58,79,115,116,124,125. Some authors have suggested that the routine use of computer navigation improves screw placement. Our own experience, however, has been that the increased operating time does not improve the quality of screw placement but increases tremendously the expense and potential danger. A dural injury was reported in seven of 124 patients168 and in two of eighty-nine patients145 in two series of pedicle-screw fixation for the treatment of degenerative spine disease. Deep-infection rates have ranged from 1.1 percent (five of 470 patients)8 to 1.2 percent55 to 2.3 percent (two of eighty-five patients)57 to 4.2 percent (four of ninety-six patients)149. Prompt wound débridement and administration of antibiotics, with preservation of the implant and subsequent delayed primary closure, have been adopted routinely, with acceptable intermediate-term clinical outcomes57. The use of pedicle-screw-supplemented fusion has been reported to enhance the results of primary operative treatment of hematogenous discitis and osteomyelitis131. The frequency of screw breakage has ranged from 2.6 percent8 to 4.9 percent55 to 9 percent6 to 36 percent114 to as high as 60 percent105. However, survivorship analysis of a series of patients who had degenerative disease indicated that the clinical survival of pedicle-screw-based constructs was similar to that of total hip and total knee reconstructions101. The two alarmingly high screw-breakage rates105,114 were reported when highly comminuted spinal fractures were internally fixed with posterior short-segment pedicle-screw-based instruments. The load-sharing classification103,127 provides a simple way to differentiate fractures according to comminution. It suggests that anterior vertebrectomy, strut-grafting, and instrumentation be performed in patients with severely comminuted injury to avoid the high screw-breakage rates that occur when short-segment posterior pedicle-screw-based instrumentation is used for highly comminuted injuries (Fig. 10).
Screw pullout and screw-connector disengagement have been reported both in in vitro testing and in patients8,16,55. Rod breakage has been reported only rarely. Most authors have described problems with the screw or the bone-screw interface but not with the longitudinal component8,16,55. Screw-thread fracture has never been reported, to my knowledge. There are established methods for removing broken screws42,104.
The union rate associated with pedicle-screw-based constructs generally has been reported to be nearly 90 percent8,20,55,167,176,177. The benefits provided by very stiff implants were demonstrated statistically in a prospective, randomized trial comparing the rate of union associated with these implants with the rates obtained without use of an implant and with use of a semirigid implant177.
Implant-related pain has led to the need to remove the implant from some patients. The clinical benefits of implant removal have not been uniformly demonstrated74.
|
Top
Introduction
History of Spinal Internal...
