Christopher G. Paramore, M.D., Curtis A. Dickman, M.D., and Volker K. H. Sonntag, M.D.
Division of Neurological Surgery, Barrow Neurological Institute, St. Joseph's Hospital and Medical Center, Phoenix, Arizona
Although they are excellent clinical tools, Caspar anterior cervical plates have not been studied closely with regard to their mechanisms of failure. As more extensive operations are contemplated on older, sicker patients, it is imperative to know when a plating system might be prone to failure and what the mechanism of that failure might be. Therefore, the authors reviewed 49 patients undergoing Caspar plate placement in whom sufficient radiographs were obtained to determine if the fate of the hardware was related to the patient's age, type of operation, and the length of construct. Eleven of 49 patients suffered hardware failure, defined as any amount of screw backout or breakage, plate pullout, or pseudarthrosis. Four patients underwent hardware removal; one underwent posterior fusion for pseudarthrosis. Only two required treatment in a halo brace. There was an eventual fusion rate of 100%, including one fibrous union, and one of the patients who underwent repeat surgery was lost to follow-up review. No graft extrusions or new neurological deficits were incurred as a result of hardware failure. Plate length predicted plate failure in a statistically significant manner. Increasing age and reoperation correlated with plate failure but were not statistically significant in this small number of patients. Telescoping of the bone graft and vertebral bodies, with concomitant migration of the plate and slippage of the screws, was common. However, telescoping was more profound in the group in which the plates failed. The authors conclude that Caspar plate failures are more likely to occur in the elderly and in patients who need longer constructs. Bone fusion can be expected even when the hardware loosens.
Key Words * cervical corpectomy * cervical spine instrumentation * anterior cervical screw plate * Caspar plate
Caspar trapezoidal plates (Aesculap, San Francisco, CA) are safe and effective in stabilizing the spine and in promoting fusion after decompression and grafting of the anterior cervical spine.[4,8,12,14,15] Despite biomechanical studies that have suggested that anterior plating is insufficient to stabilize three-level cervical injuries,[6,9,10,16] there is ample clinical evidence to the contrary.[1-4,8,12-15] Caspar anterior cervical plates are also effective in promoting fusion after decompression for cervical spondylotic myelopathy, osteomyelitis, and tumors involving the vertebral bodies.[11,14]
The biomechanical basis of Caspar plate failure has not been evaluated systematically. Because the Caspar system is nonconstrained (that is, screws are not locked into the plate), bicortical screw purchase is mandatory to secure the system. Several reports have documented screw "backout" or "breakage,"[4,8,11,12] but screw-plate failures have not been analyzed. The biomechanical literature suggests that repetitive cyclical loading causes loosening at the screw-bone interface (toggling), leading eventually to screw backout.[5,7] However, ex vivo models cannot reproduce the load-bearing and load-sharing responsibilities of the graft and plate, respectively. Settling of the graft-plate construct in the vertical axis, or "telescoping," is a potential reason for failure. Telescoping might be expected when there is a mismatch in the bone densities of the graft and native bone, as when banked bone is used in an osteoporotic patient. However, telescoping has not been reported in association with Caspar plate failure in the clinical literature.
The purpose of this retrospective clinical review was to analyze critically the failure mechanisms of Caspar plates from a biomechanical perspective. Additionally, several clinical variables including age, diagnosis, and the length of a construct were analyzed statistically to determine if there was an association between clinical factors and the likelihood of plate failure.
CLINICAL MATERIAL AND METHODS
Clinical records and preoperative, immediate postoperative, and 1-month postoperative radiographs were reviewed in 49 patients undergoing Caspar anterior cervical plating at our institution. Patients were excluded from the study if less than 1 month of follow-up evaluation was available. The purpose of the study was not to evaluate the long-term outcome of patients treated with Caspar plating but rather to correlate the radiographic appearance of the construct with the patient's clinical course during the first weeks after surgery. Our previous experience had indicated that plate failures usually occurred within 1 month of operation. All patients underwent operation via the standard anterior cervical approach. Autologous iliac bone crest was used as the graft material in all one-level fusions and in all corpectomies up to two levels. Allograft fibula bone was used when decompressions longer than two vertebral bodies were performed.
A construct was considered a mechanical failure if follow-up radiographs revealed screw backout, screw breakage, plate pullout, or pseudarthrosis, whether or not the patient experienced clinical symptoms. After assigning a "failed" or "nonfailed" grade to each construct, the independent variables of age, diagnosis, length of Caspar plate, and previous operation were evaluated for a predictive association with plate failure.
The radiographic appearance of the construct immediately postoperatively and at follow-up evaluation was compared in each patient for whom the appropriate radiographs were available. Each screw was examined for evidence of failure, as noted. Additionally, the distance between the cephalad cortex of the upper body and the caudal cortex of the lower body of the construct was measured on radiographic films obtained immediately postoperatively and at follow-up review. The length of the plate was measured and all distances normalized to the actual plate length, as recorded in the operative notes. The amount of construct telescoping could thus be determined for all cases, whether failed or nonfailed.
RESULTS
The pertinent clinical data for the patients included in this study are presented in Table 1. Bicortical screw purchase, verified by postoperative radiographs, was achieved in all patients. Several patients who did not meet the criteria for plate failure were still noted to have construct telescoping on their follow-up radiographs. Eleven patients met the criteria for plate failure (Fig. 1). There were five cases of screw backout, three of screw breakage, two of plate pullout, and one case of pseudarthrosis. Of these patients six were asymptomatic and were treated with a hard collar until fusion was documented radiographically, whereas four required reoperation for plate removal because of dysphagia. The two patients whose plates pulled out required treatment with halo vests. The patient with a symptomatic pseudarthrosis underwent posterior fusion. Nine patients developed a bone fusion, one developed a fibrous union, and one was lost to follow-up review after 30 days.
Fig. 1. Radiographic studies showing Caspar plate failures. A and B: This 78-year-old woman underwent a one-level corpectomy and Caspar plating for spondylotic myelopathy (A). At 2-month follow-up examination, she was noted to have telescoping of the plate across the lower disc space and an inferior screw backout (B). She was treated for an additional month in a hard collar and developed a radiographic bone fusion with no adverse symptoms. C and D: This 74-year-old man underwent a two-level corpectomy and Caspar plating for spondylotic myelopathy (C). Four months postoperatively, he was noted to have profound telescoping, lower screw breakage, and upper screw backout (D). Additionally, he had developed dysphagia. The instrumentation was removed and solid fusion was noted. His dysphagia resolved. E-G: This 79-year-old woman underwent a four-level corpectomy and Caspar plating for postlaminectomy kyphosis (E). At the initial operation her bone quality was considered extremely poor, so she was placed in a halo vest immediately postoperatively. A follow-up radiographic film obtained 2 weeks later demonstrated a plate pullout (F). She suffered dysphagia and the instrumentation was removed. The patient was treated in the halo vest for several months and bone fusion was noted at a 1-year follow-up examination (G).
The likelihood of construct failure was associated with increasing age (Fig. 2) and increasing plate length (Fig. 3). Similarly, hardware failure was found in 42% of patients undergoing reoperation. Using logistic regression modeling, a probability equation was generated that described the likelihood of plate outcome given a certain variable. Plate length correctly predicted the outcome in 81.4% of cases, and this relationship was statistically significant (chi-square test, p = 0.0024). The other variables only marginally increased the power of the equation and were thus discarded. However, it is likely that age and reoperation would prove statistically significant if the patient sample were larger.
Fig. 2. Bar graph showing the relationship of patient age to plate failure. The percentage of patients experiencing hardware failure increased with increasing age. For patients older than 70 years, the likelihood of plate failure was 50%.
Fig. 3. Bar graph showing the relationship of plate length to plate failure. Caspar plates 60 mm and longer had a failure rate of 64%. By logistic regression analysis, plate length was shown to be a statistically significant predictor of plate outcome (the longer the plate, the higher the probability of plate failure).
Twenty-three patients had radiographs that allowed evaluation of construct telescoping. More telescoping occurred in patients with longer constructs, and it is also this group that is more prone to hardware failure. The mean amount of telescoping in the 16 patients in the nonfailed group was 2 ± 1.6 mm and the mean decrease in height was 4.3% ± 3.6%. In contrast, the mean telescoping in the seven patients in the failed group was 6 ± 3.1 mm and the mean decrease in height was 9% ± 3.1%.
DISCUSSION
The clinical effectiveness of the Caspar anterior cervical plating system has been well established in reports on the treatment of fractures of the cervical spine.[4,8,12,14,15] These authors documented screw loosening or breakage in 0% to 7.5% of cases but did not comment on construct telescoping. Our experience verifies these findings; none of the 13 patients with cervical fractures in this series showed clinical or radiographic evidence of failure. However, in general, trauma patients tend to be relatively young and healthy compared to the overall population. Advances in medical technology have resulted in many older and sicker patients becoming candidates for anterior cervical plating. Many of these patients suffer from multilevel disease, necessitating long constructs. Often these patients have already had one or more operations to the anterior or posterior cervical spine. The fate of cervical instrumentation in these complicated cases has not been thoroughly addressed.
This study reviewed the effectiveness of Caspar plates in maintaining construct integrity in the presence of many types of cervical pathology. Some type of shortcoming was noted in 11 of 49 patients, but clinical symptoms, usually dysphagia, occurred in only five patients. A consistent finding in all patients was telescoping, which occurred as the graft and vertebral bodies settled in the early phase of the healing process. Usually, the Caspar instrumentation compensated by small movements of the screws along the slots and migration of the plate along the anterior surfaces of the vertebral bodies. However, if the amount of telescoping became excessive, screw breakage or plate pullout resulted. Screw backout rarely occurred without telescoping. Telescoping might be conjectured to occur more frequently in the Caspar system, and others like it, in which the screws are not locked into the plate. However, such a system also allows the vertebral bodies to load-share more effectively than a locking screw system, conceivably improving the rate of bone fusion. Such improvements remain to be determined.
Of the clinical factors studied, plate length was statistically predictive of plate failure. Other variables such as age, reoperation, and extent of procedure showed definite associations with higher failure rates but were not statistically significant. Together these data indicate, as does common sense, that more extensive operations in older people are more likely to fail. However, an encouraging finding is that most of the hardware-related problems encountered were treated with hard collars or with halo vests. No patient experienced graft dislodgment or a new neurological deficit related to plate failure, and the ultimate fusion rate was nine (90%) of 10 in the failure group, with one patient lost to follow-up review. Overall, 47 of the 49 patients achieved solid fusion without revision of their initial bone graft. One patient required a secondary fusion for symptomatic pseudarthrosis and one was lost to follow-up review.
The Caspar system is a nonconstrained screw plate system requiring bicortical purchase of the vertebral bodies. Telescoping of the bone-plate construct occurs to some degree in almost all patients and is for the most part tolerated by a combination of slot-slippage and screw angulation. Bone fusion can be expected in these patients. Excessive telescoping is to be expected in elderly patients with poor bone quality or when extremely long constructs are required. Excessive telescoping can result in screw backout, screw breakage, or frank plate pullout. Approximately half of the patients with radiographic evidence of screw backout or breakage may be asymptomatic; most of the others may be managed successfully with hardware removal. Halo bracing is only required occasionally, and a stable fusion may be expected even in the presence of hardware loosening. Excellent results were routinely obtained in operations for traumatic cervical injury.
Consequently, the Caspar system is a viable option for anterior cervical plating. The failure mechanisms of the newer constrained locking plate systems remain to be elucidated. A valid comparison can be made only after the hardware failure rates and bone fusion rates for all systems in all circumstances can be predicted.
Acknowledgment
The authors thank Dr. Eric Bullfinch for aiding in the statistical analysis of these data.
References
1. Aebi M, Zuber K, Marchesi D: Treatment of cervical spine injuries with anterior plating. Indications, techniques, and results. Spine 16(Suppl 3):S38-S45, 1991
2. Böhler J, Gaudernak T: Anterior plate stabilization for fracture-dislocations of the lower cervical spine. J Trauma 20:203-205, 1980
3. Bremer AM, Nguyen TQ: Internal metal plate fixation combined with anterior interbody fusion in cases of cervical spine injury. Neurosurgery 12:649-653, 1983
4. Caspar W, Barbier DD, Klara PM: Anterior cervical fusion and Caspar plate stabilization for cervical trauma. Neurosurgery 25:491-502, 1989
5. Clausen JD, Goel VK, Traynelis VC, et al: Biomechanics of Caspar and Synthes plates in quasi-static and cyclic loading modes. Bioengineering Conference ASME 24:618-620, 1993
6. Coe JD, Warden KE, Sutterlin CE III, et al: Biomechanical evaluation of cervical spinal stabilization methods in a human cadaveric model. Spine 14:1122-1131, 1989
7. Gallagher MR, Maiman DJ, Reinartz J, et al: Biomechanical evaluation of Caspar cervical screws: comparative stability under cyclical loading. Neurosurgery 33:1045-1051, 1993
8. Garvey TA, Eismont FJ, Roberti LJ: Anterior decompression, structural bone grafting, and Caspar plate stabilization for unstable cervical spine fractures and/or dislocations. Spine 17(Suppl 10):S431-S435, 1992
9. Kalff R, Ulrich C, Claes L, et al: Comparative experimental biomechanical study of different types of stabilization methods of the lower cervical spine. Neurosurg Rev 15:259-264, 1992
10. Montesano PX, Juach EC, Anderson PA, et al: Biomechanics of cervical spine internal fixation. Spine 16(Suppl 3):S10-S16, 1991
11. Naito M, Kurose S, Oyama M, et al: Anterior cervical fusion with the Caspar instrumentation system. Int Orthop 17:73-76, 1993
12. Randle MJ, Wolf A, Levi L, et al: The use of anterior Caspar plate fixation in acute cervical spine injury. Surg Neurol 36:181-189, 1991
13. Ripa DR, Kowall MG, Meyer PR Jr, et al: Series of ninety-two traumatic cervical spine injuries stabilized with anterior ASIF plate fusion technique. Spine 16(Suppl 3):S46-S55, 1991
14. Tippets RH, Apfelbaum RI: Anterior cervical fusion with the Caspar instrumentation system. Neurosurgery 22:1008-1013, 1988
15. Tuite GF, Papadopoulos SM, Sonntag VKH: Caspar plate fixation for the treatment of complex hangman's fractures. Neurosurgery 30:761-765, 1992
16. Ulrich C, Woersdoerfer O, Kalff R, et al: Biomechanics of fixation systems to the cervical spine. Spine 16(Suppl 3):S4-S9, 1991
Manuscript received August 24, 1995.
Accepted in final form January 22, 1996.
Address correspondence to: Volker K. H. Sonntag, M.D., Neuroscience Publications, Barrow Neurological Institute, 350 West Thomas Road, Phoenix, Arizona 85013-4496.