Paper # 1 Cervical Range of Motion Following Anterior Cervical Fusion Dan Riew, MD, David Hannallah, MS, Kevin Lee, BA, Kari Goette, RN, BSN (St. Louis, MO) INTRODUCTION: One of the most frequently asked questions by patients about to undergo cervical fusion involves the loss of motion associated with the operation. Despite this, we were unable to find any studies that examined the amount of motion lost following anterior cervical fusion. We undertook this prospective study to determine the amount of neck flexion and extension lost following anterior cervical fusion. METHODS: 33 patients, with no history of previous cervical fusion, who were about to undergo anterior cervical fusion anywhere between C3 to C7 were enrolled in the study. Lateral photographs were obtained with the patient standing upright and maximally flexing and extending their neck. All photographs were obtained by the same person using the same camera and included the same vertical reference (door edge). Pictures were obtained preoperatively and repeated at 6 weeks, 3 months, 6 months, and 12 months postoperatively. The patients had either an ACDF or a corpectomy procedure. All pictures were scanned into a computer and the patient's range of motion was assessed by summing the extent to which they could flex and extend their heads relative to a vertical line. The angles were measured using the NIH Image software by two different observers. The inter- and intra-observer variability was calculated and an average of three different measurements was used in calculating the range of motion for each picture. The change in range of motion was determined by subtracting the pre-operative range of motion from the maximum post-operative value. RESULTS: Of the 33 patients analyzed, 13 (40%) had one level fused, 11 (33%) had two levels fused, 7 (21%) had three levels fused, and 2 (6%) had four levels fused. 14 (42%) of the 33 patients had a herniated disc and the remaining 19 (58%) had cervical spondylosis. 9 (64%) of the 14 patients with a herniated disc had one level fused and the remaining 5 patients had two levels fused. The inter-observer variability for the measurement technique was 0.96 and the intraobserver variability was 0.99. The patients who had one level fused lost an average 11 degrees (the maximum loss: 82 degrees in one patient with the remaining patients losing less than 25 degrees or actually gaining motion), the patients who had two levels fused gained an average of 5 degrees (maximum loss: 20 degrees, maximum gain: 66 degrees), the patients with 3 or 4 levels fused lost an average of 11 degrees (maximum loss: 34 degrees, maximum gain:13 degrees). None of the above differences between pre-operative range of motion and post-operative range of motion reached statistical significance. Among the single level ACDF patients, those with a diagnosis of a herniated disc lost an average of 16 degrees compared to those with cervical spondylosis who gained an average of 6 degrees. The average pre-operative range of motion was 90 degrees (maximum: 132 degrees, minimum: 50 degrees) for patients with a ruptured disc and 83 degrees (maximum 121 degrees, minimum 32 degrees) for patients with cervical spondylosis. The maximum post-operative range of motion was obtained at an average of 21 weeks post-operatively. DISCUSSION/CONCLUSION: To our knowledge, this is the first prospective study that has evaluated the impact of anterior cervical fusion on range of motion. Inter- and intra-observer variability for our measurement technique was minimal. The loss or gain in range of motion was found to be minimal and not statistically significant. Although there was a trend suggesting that those with a herniated disc lost more motion than those with spondylosis, the differences did not reach statistical significance. A possible explanation for the lack of limitation of motion following fusion may be that the preoperative motion was already limited by either pain or degeneration. In cases where pain limits preoperative motion, relief of pain following surgery would be expected to result in increased motion. In cases where spondylosis physically limits the preoperative motion, fusion of that segment would be expected to have minimal effect on the overall motion. Cervical flexion and extension can be accurately and reproducibly measured using the described technique. It appears that anterior cervical fusion involving 1-3 levels has minimal effect on cervical flexion and extension range of motion. Paper # 2 ACDF for Neck Pain: Clinical Outcome Study Timothy A. Garvey, MD (Minneapolis, MN), Ensor E. Transfeldt, MD (Minneapolis, MN), James R. Malcolm, MD, Paul B. Kos, BS (Rochester, MN) INTRODUCTION: Clinical outcome studies with validated patient perceived outcome measurements are few. This study documents an average 4.4 year follow-up of the clinical outcomes for patients undergoing an anterior cervical discectomy and fusion for neck pain as the chief complaint. While some report inferior outcomes when neck pain is the chief complaint, this series documents an eighty-four percent self-perceived satisfactory outcome. Eight-seven patients were identified from our database, who had a minimum of 2 year follow-up, had had an anterior cervical discectomy and fusion for the primary indication of dominant neck pain, and who had completed extensive outcome questionnaires regarding pain and self function at long-term follow-up. Surgical candidates generally met all of the following criteria: greater than twelve months of symptomatology, failure of an active non-operative exercise program, preoperative radiographs including flexion extension views, pre-operative MRI, cervical discography with concordant reproduction of pain, and no apparent psychosocial contraindications. Eight-seven patients, were followed for an average 4.3 years, and were broken down by the number of levels fused, educational status, source of injury, their responses regarding pain, functional outcome (Oswestry Disability Index and Roland and Morris Disability Index), and global satisfaction. DISCUSSION/CONCLUSION: At an average 4.4 year follow-up, eighty-four percent of patients self perceived the results to be good to excellent. Eightyseven percent would repeat the same treatment again if they had the same condition. Ninety-three percent of patients reported improvement of pain with average visual analog scale changing from 8.4 pre-operatively to 3.8 postoperatively. The self rated functional status improved approximately fifty percent on both the Oswestry Disability Index and Modified Roland and Morris Disability Index. The presence or absence of worker’s compensation or litigation did not correlate with clinical outcome. The educational level did not correlate with clinical outcome. Summary: Properly selected patients with chief complaint of chronic neck pain have a high percentage change of self perceived improvement of pain and function when treated with an anterior cervical discectomy and fusion. While their neck is not made normal, this self expressed satisfaction when compared to their pre-operative dissatisfaction because of their neck condition, documents this surgical treatment as a reasonable option. Paper # 3 Outcome Analysis of Cervical Burst Fractures Treated with Corpectomy and Anterior Plating vs Halo Charles G. Fisher, MD, FRCSC, Marcel Dvorak, MD, Jordan Leith, MD, Peter Wing, MD (Vancouver, BC, Canada) INTRODUCTION: The optimal management of cervical burst fractures has historically been controversial. With the evolution of anterior cervical spine locking plates, treatment has shifted from halo immobilization to surgical stabilization. Although outcomes of these treatment alternatives have been analyzed, they all have suffered from the inherent bias of retrospective studies without a standardized health related quality oflife (HRQL) instrument and a control or comparative group. Furthermore, study populations have lacked homogeneity with respect to fracture patterns. Therefore, the purpose of this study was to compare the outcomes of two groups of patients with unstable cervical burst fractures, those treated with halo thoracic vests (HTV) and those treated with anterior corpectomy plating (ACP). METHODS: A spine database generated retrospective cohort analysis was conducted for all sub axial cervical burst fractures treated by the Acute Spinal Cord Injury Unit at Vancouver Hospital from 1992 to 1996. Fractures had to consist of failure anteriorly in compression and posterioly in tension with or without neurology. The primary outcome was radiographic kyphosis at the time of clinical and radiographic union. Secondary outcomes included the gold standard of HRQL instruments, the Short Form 36 (SF-36) and the Cervical Spine Research Society (CSRS) disease specific instrument. Apriori sample size calculation was conducted and the study was appropriately powered for the primary outcome with an alpha of .05 a beta of 0.1. RESULTS: 44 patients met inclusion and exclusion criteria of which 21 were treated with HTV and 23 with ACP group. The two groups were comparable for baseline demographic data, with the exception of the American Spinal Injury Association (ASIA) motor score representing neurologic deficit. 67% of the HTV patients had a neurological deficit, while 96% in the ACP group had deficits. ASIA motor scores were 68.7 in the HTV group and 38.9 in the ACP group. The majority of injuries occurred at the C5 level. All 44 patients had radiographic follow up but only 12 of the 21 patients in the HTV group and 11 of the 23 patients in the ACP group completed HRQL instruments. The mean kyphosis was 12.0 degrees for the HTV group and 3.54 for the ACP group. This was statistically significant (p < 0.001). The difference remained significant with the baseline variables controlled with multivariant regression analysis. In the HTV group there were 5 failures. 4 of these were managed operatively. There were no major intra or post operative complications in the ACP group. The mean of the SF-36 mental and physical component scores and the CSRS satisfaction sub scale scores were calcuated for each of the two groups and there was no significant difference. The results remained consistent after adjusting for motor score discrepancy between groups. DISCUSSION/CONCLUSION: The results of this study indicated that the optimal treatment to restore and maintain sagittal alignment and minimize treatment failures is anterior cervical plating. The relationship of residual kyphosis and functional outcome is yet to be determined in the literature. Despite a difference in residual kyphosis between the two treatment groups, the patient outcomes did not show any signficant difference and indeed there was a trend that the SF-36 physical component score was higher in the HTV group. It is likely this trend was related to the large number of patients with neurologic deficit in the ACP group. The SF-36 mental component scores were similar between groups. Interestingly, the CSRS satisfaction with treatment score was worse in the ACP group. The results regarding the HRQL are hypothesis generating at best. The study is under-powered to look at these variables. The results indicate that fracture morphology at follow up may not accurately predict patient outcome. Further research is warranted in this area. Paper # 4 Graft Displacement in Multilevel Cervical Corpectomies Robert A. Hart, MD (Portland, OR), Jeffrey C. Wang, MD (Los Angeles, CA), Sanford E. Emery, MD (Cleveland, OH), Henry H. Bohlman, MD (Cleveland, OH) INTRODUCTION: Cervical corpectomy is commonly performed in the presence of multi-level cervical disease and often requires the use of long anterior strut grafts. Migration or dislodgment of these grafts is a reported complication, sometimes requiring revision surgery. The purpose of this study was to examine the factors related to graft displacement in cervical corpectomy, and evaluate the outcomes of patients suffering this complication. METHODS: Patients undergoing anterior cervical corpectomies and strut grafting by two senior surgeons were reviewed. No patient had had a prior laminectomy. Medical records and radiographs of all patients with postoperative graft migration or displacement were evaluated. Measurements from preoperative and postoperative radiographs included: (1) cervical lordosis (C2-C7), (2) segmental lordosis of fusion segments, (3) segmental height of fusion segments. Levels of corpectomy and grafting were also recorded. RESULTS: 248 total patients underwent one to five level anterior corpectomies and fusions over a 25-year period (94 one-level, 76 two-level, 71 three-level, 6 four-level, and 1 five-level corpectomy). All were unplated fusions utilizing autogenous bone graft (iliac crest for 1 level corpectomy; fibular struct for 2 or more). Fifteen of the patients (average age 61.4 years, 9 females, 6 males) had grafts that migrated or displaced in the postoperative period. Of the 15 patients with graft migration, 3 were one-level, 4 were two-level, 7 were three-level, and 1 was a four-level corpectomy, giving migration rates of 3.2%, 5.3%, 9.9% and 16.7% for one-two-three-, and four-level corpectomies, respectively. 14 of the 15 patients' procedures (93%) included a corpectomy of the C6 vertebra with a fusion to the C7 vertebrae. No corpectomy of two or more levels that did not involve a distal fusion at C7 suffered migration or dislodgment. No relationship could be determined between the preoperative cervical lordosis, fusion segment lordosis, or segmental height of the fusion segments and graft migration or dislodgment. Five patients (33%) underwent revision procedures acutely. No patient suffered a neurological or respiratory complication due to graft migration. DISCUSSION/CONCLUSION: A direct relationship was demonstrated in our patients between increasing number of vertebral bodies resected and frequency of graft displacement. While significant grant displacement may require revision surgery, no patient suffered a permanent adverse outcome from this complication. Decompression involving the C6 vertebral body and fusion ending at C7 was associated with a higher rate of graft migration and displacement. Multi-level corpectomies extending to C7 should be considered for additional stabilization or postoperative immobilization to protect against graft migration. Paper # 5 Four-Level Anterior Cervical Discectomy and Fusion Report of 18 Cases with One-Year Follow-Up Paul M. Arnold, MD, Donald A. Eckard, MD (Kansas City, KS) INTRODUCTION: Controversy exists regarding optimal treatment for multilevel cervical spondylotic myelopathy. Recent reports have indicated that fusion rates are unacceptably low with this procedure. We present our series of 18 patients who underwent four-level anterior cervical DISCECTOMY with fusion and plate fixation. METHODS: Eighteen patients underwent 4-level anterior cervical DISCECTOMY and fusion (ACF). There were fifteen males and three females with an age range of 33 to 67. All patients had clinical evidence of myelopathy with documented spinal cord compression on MRI and/or CT myelogram. Each patient had 4 disks removed via modified Smith-Robinson technique, burring of the end plates, placement of the autogenous iliac crest graft, and fixation with a cervical spine locking plate (Synthes Spine, Paoli, PA) or Atlantis plate (Sofamor Danek, Memphis TN) with either 8 or 10 screws. Plates were bent to allow normal sagittal alignment. All patients were followed for at least one year with periodic x-rays in an outpatient setting. Patients wore external othosis for 12 weeks following surgery. RESULTS: No patient was made neurologically worse by surgery. Twelve patients had improvement in their myelopathic symptoms, and six patients were unchanged. Nine patients that had radiculopathy all reported improvement. Five patients continued to complain of neck pain following surgery; The rest were improved. Six patients complained of dysphagia following surgery; this eventually resolved over time in each patient. One patient had a deep wound infection which required a second operation for debridement. This patient had transient right upper extremity weakness which completely resolved after six months. One patient had aspiration pneumonia and respiratory failure which required prolonged intubation. He had transitory quadriparesis, but this resolved over the next three months. All patients underwent radiographic follow-up for at least 12 months post-op. 15 of 18 patients had solid fusion at all 4 levels (83%), and 3 patients had non-union at one level. Sagittal alignment has been maintained in all patients. No patient underwent a second procedure due to hardware failure. On patient had back-out of her two screws a C7; She had a solid fusion at this level and we have continued to follow her. Two patients had partial back-out of a screw in C3; they have backed out halfway but have been stable over several months. DISCUSSION/CONCLUSION: We report an 83% fusion rate for four-level ACF, and we continue to perform this procedure for multilevel cervical disk disease. No patient has required a second procedure for hardware failure, and we have been pleased with patient’s outcomes, both clinically and radiographically. We recommend four-level ACF for multilevel cervical spondylotic myelopathy. Paper # 6 Dysphagia, Hoarseness, and Vocal Fold Dysfunction Following the Anterior Cervical Approach to the Spine: A Study of Potential Risk Factors and Management Guidelines William F. Young, MD, Eli. M. Baron, MD, Ahmed M.S. Soliman, MD, John P. Gaughan, PhD, Lisa Simpson, MA, CCC-SLP (Philadelphia, PA) INTRODUCTION: Dysphagia, hoarseness and recurrent laryngeal nerve injury are well known complications of anterior cervical discectomy that are only briefly mentioned in the neurosurgical literature. This study was undertaken to determine potential risk factors for dysphagia and hoarseness after anterior cervical discectomy. This study was also undertaken to determine the incidence of these complications. METHODS: The charts of 100 cases of anterior cervical discectomy with fusion performed at our institution form January 1996 until February 1999 were reviewed. Univariate logistic regression was used to estimate the relationship of individual factors including patient age, sex, race tobacco use, alcohol use, diabetes hypertension, side of approach, indication for surgery, number of levels being operated on, cervical plating, clinical indication (radiculopathy, myelopathy and neck/shoulder pain), pathology (herniated nucleus pulposus, cervical spondylosis, pseudoarthrosis , cervial stenosis) and difference in surgeons with occurrence of hoarseness dysphagia. Also, the literature on anterior cervical discectomy regarding the incidence of these problems was reviewed and an overall incidence of dysphagia, hoarseness and recurrent laryngeal nerve injury was calculated. Selected cases are presented and the management and possible prevention of these complications are discussed. RESULTS: Patient age was found to be a significant predictor of postoperative dysphagia (p<0.006) with and odds ratio of 1.113 (95% C.L 1.04, 1.21) per year of age. Patient age was not found to be a significant predictor of hoarseness. Patient sex, race, tobacco use, alcohol use, diabetes, hypertension, side of approach, indication for surgery, number of levels being operated on, cervical plating, clinical indication, pathology and difference in surgeons were not found to be significant predictors of postoperative dysphagia, hoarseness and recurrent laryngeal nerve injury in the literature was calculated as 12.3%, 4.9% and 1.4% respectively. DISCUSSION/CONCLUSION: Dysphagia, hoarseness and recurrent laryngeal nerve injury continue to remain common complications of anterior cervical discectomy, despite over 40 years of experience with the technique. Age was found to be a significant risk factor for postoperative dysphagia, Dysphagia following anterior cervical discectomy that lasts over 2 days should be studied with modified barium swallow. Current evaluation of persistent hoarseness and vocal cord immobility following anterior cervical discectory includes assessment with swallowing evaluation, fiberoptic laryngoscopy and laryngeal electromyography. Management of the immobilized vocal fold utilizes Gelfoam injection and medialization laryngoplasty as the main corrective surgical techniques. Nerve identification and intraoperative electromyographic nerve monitoring may reduce the risk of recurrent laryngeal nerve injury. Paper # 7 Vocal Cord Paralysis Following Anterior Cervical Spine Surgery Brian A. Jewett, MD, Gregory A. Mencio, MD, Dan M. Spengler, MD, Sean C. Colman, MD, James L. Netterville, MD (Nashville, TN) INTRODUCTION: The anterior approach to the cervical spine is commonly used for surgery of the cervical spine. In most large series of cervical spine fusions performed through the anterior approach, vocal cord paralysis (VCP) is reportedly rare. We review our experience from the Vanderbilt Voice Center to further delineate risk factors associated with vocal cord paralysis following anterior cervical spine surgery. METHODS: A retrospective review of the database at the Vanderbilt Voice Center from 1989 to 1999 revealed 365 patients treated for vocal cord paralysis. Forty-one of these patients sustained unilateral vocal cord paralysis following anterior cervical spine fusions. These charts were reviewed and a database was compiled with regards to patient demographics, surgery, approach, presenting symptoms, treatment and eventual outcome. RESULTS: All patients, except one, had right sided paralysis following a right sided surgical approach. Twelve patients underwent ACDF at C6-7, twenty patients underwent C5-6 ACDF, five had C5-6, C6-7 multilevel ACDF, 2 had C3-4, C4-5 ACDF and 2 had C5 corpectomy and fusion. The majority of patients presented with immediate post-operative hoarseness and voice fatigue, while some experienced recurrent aspirations and dysphagia. All patients were initially diagnosed with VCP by video-laryngoscopy. Twelve patients’ symptoms responded to conservative treatment, fifteen improved after injection of gelfoam, and fourteen eventually required unilateral vocal cord medialization surgery to correct their deficits. DISCUSSION: Anatomic cadaveric studies published from our Otolaryngology laboratories clearly show that the recurrent laryngeal nerve is at higher risk for stretch induced injury from the right side approach due to its shorter recurrent anatomic course. Without knowing the actual number of cases performed in our referral population, the true incidence of VCP following anterior spine surgery remains elusive. However, it is clear that symptomatic vocal cord paralysis following anterior cervical spine surgery is directly related to the right sided approach, which should when possible be abandoned in favor of the left sided approach. Paper # 8 On The Incidence, Cause, And Prevention of Recurrent Laryngeal Nerve Palsies During Anterior Cervical Spine Surgery Ronald I. Apfelbaum, MD, Mark D. Kriskovich, MD (Salt Lake City, UT) INTRODUCTION: This study analyzes the incidence of recurrent laryngeal nerve (RLN) palsy as a complication of anterior cervical spine surgery in a consecutive series of 900 patients. A hypothesis regarding the etiology of this, along with cadaveric verification and a technique to significantly reduce this complication is reported. METHODS AND RESULTS: A contemporaneous computerized database of all patients undergoing anterior cervical spine surgery with plating since 1986 is maintained at the University of Utah. Any patient who had persistent postoperative hoarseness was evaluated by the ENT service with laryngoscopy to establish whether they had a vocal cord paralysis. When we noted a 6.4% incidence of this complication in our first 250 cases in this series, a search for causes led to the suggestion that endolaryngeal injury to the vulnerable endolaryngeal segment of the nerve, where it crosses the thyroid lamina in a submucosal position, might be a possible factor. We postulated that this might occur due to retraction of the larynx against the unyielding shaft of the endotracheal tube (ET) since the tube is fixed proximally by tape at the mouth and distally by the inflated cuff. Cadaveric studies in 4 unfixed human cadavers, imaged with water soluble contrast within the larynx, demonstrated that the larynx is indeed displaced into the ET shaft after placement of the anterior cervical spine retractor system. This study also confirmed that deflating and then reinflating the ET cuff allowed the tube to recenter within the larynx and released the contact with the laryngeal wall. Clinically we began systematically deflating the ET cuff, after placement of the self-retaining retractor beneath the longus coli muscles, and then reinflating the cuff to “just sealed” pressure. This resulted in a dramatic and statically significant decrease in the incidence of RLN palsies from the previously noted 6.4% to 1.7% (P=0.0002). DISCUSSION: It has often been stated that RLN injury in anterior cervical spine surgery is due to direct injury to the nerve in the course of dissection or to stretch injury of an aberrantly located nerve, but no anatomic proof exists for either of these hypotheses. The fact that RLN palsy is reported after surgery that does not involve the neck or thorax indicates that another mechanism exists. Review of the ENT literature suggest such a mechanism involves endolaryngeal injury to the anterior branch of the nerve as it traverses submucosally with the unyielding thyroid lamina behind it. We therefore hypothesized that such might be occurring in our cases and began to routinely deflate the ET cuff, after the self-retaining retractors were placed beneath the longus coli muscles, and then reinflating the cuff to just sealed pressures. This was repeated if the retractor was moved or replaced. This simple maneuver resulted in a dramatic, highly statistically significant, decrease in RLN paralysis, which averaged 1.7% over the last 650 cases in our series. To establish the anatomic basis for this hypothesis, 4 unfixed human cadavers were studied by imaging their larynxes using a water soluble contrast agent. This clearly showed contact between the ET shaft and the lateral wall of the larynx, in the above noted vulnerable region, and movement of the tube away from the wall after deflation and reinflation of the cuff. Based on this anatomic and clinical study we conclude that most cases of RLN palsy are not due to injury to the nerve during neck dissection but rather is due to neuropraxia secondary to endolaryngeal compression of a short but vulnerable segment of the nerve by the shaft of the endotracheal tube. This can often be prevented by the simple technique of deflating the ET cuff after retractor placement and then reinflation. Paper # 9 The Efficacy of Air Bags in Preventing Spinal Injuries Michael A. Catino, MD, Molly T. Vogt, PhD, William Donaldson, III, MD (Pittsburgh, PA) INTRODUCTION: No studies to date have examined the role that air bags have in preventing spinal injury. This paper discusses the role of automobile air bags in preventing spinal injury and describes the distribution of spinal injuries sustained in motor vehicle accidents in patients in four different protective device groups: air bags only, seat belts only, seat belts and air bags, and no protective devices. METHODS: Information on patients sustaining a spinal injury secondary to a motor vehicle accident between 1990 and 1997 was obtained from the Pennsylvania Trauma Systems Foundation. Analysis was performed on 7,170 patients evaluated and admitted at 26 accredited statewide trauma centers. RESULTS: Relative to the air bag and seat belt combined group, the incidence of spinal injury is 37% greater if no protective device is used (p<0.001), 42% greater if only a seat belt is used (p<0.001), and 35% greater if only an air bag is deployed (p<0.03). No statistical variation in the distribution of spinal injuries by region of the spine for the four protective groups was determined. Analysis of the mean ages for the four groups revealed a significantly higher age for patients wearing a seat belt with air bag deployment. DISCUSSION/CONCLUSION: This study of spinal injuries secondary to motor vehicle accidents in Pennsylvania indicates that individuals who are wearing seat belts during deployment of an air bag are at significantly decreased risk of sustaining a spinal injury when compared to air bag only, seat belt only or no protective device groups. Furthermore, individuals in this group are significantly older than those in the other three groups. However, the distribution of spinal injuries by spinal region does not vary (p>0.05) across the four groups. Paper # 10 Is There a Relationship Between Pre-existing Spinal Degeneration and Whiplash Associated Symptoms in Victims of Rear-end Automobile Impacts? Allan F. Tencer, PhD, Sohail K. Mirza, MD (Seattle, WA) INTRODUCTION: Cervical spine “whiplash” symptoms reported by victims of rear-end automobile accidents may include not only pain about the neck and shoulders, but also associated symptoms such as tingling and numbness in the arms or low back pain. While mechanisms have been described for damage to neck tissues, these associated symptoms are more difficult to explain. Possibly, the extent of pre-existing degeneration of the victim’s cervical or lumbar spine contributes to these symptoms. Therefore a retrospective study was performed to determine whether the degree of degeneration of the spine of a “whiplash” victim was a risk factor in development of these associated symptoms. METHODS: A database of 237 actual rear end motor vehicle impacts with full documentation of vehicle damage, police collision reports, accident circumstances, statements of crash victims, self reported initial symptoms, and a medical history of the victim's own medical provider, was constructed. Accidents in which the maximum change in velocity of the struck vehicle was calculated to be less than 7 mph (11.3 kph, peak vehicle acceleration = 4.2 g) were selected. This level of impact encompasses most collisions with little vehicle damage. Three groups were considered: Group N victims (n = 63) reported only neck pain, Group NB victims (n = 109), reported neck and low back pain, and Group NA victims (n = 50), reported neck pain and arm problems. Symptoms considered were only those reported at the victim’s first visit to a medical provider after the accident. No independent medical exams were performed by the investigators. Exclusion criteria included lack of a head restraint in the vehicle, being unrestrained, impacting the interior of the vehicle, and impacts with a struck vehicle speed change greater than 7 mph. Comparisons were made between Group N and Group NB with respect to severity of pre-existing lumbar degeneration, and between Group N and Group NA with respect to severity of cervical spinal degeneration. Additional comparisons were made between groups for awareness of the impact, and position of the head just prior to impact. Spinal degeneration was rated from radiographic reports (x-rays and usually CT or MRI studies) as: none - normal spine, mild - minimal radiographically apparent changes such as some loss in disc height, moderate - disc bulge with loss of disc height, severe - disc herniation or protrusion, significant osteophyte formation, facet arthritis. RESULTS: (i) Demographics: The sample consisted of 237 subjects, 93 males (39%, avg. age 40 years, weight 190 lb., height 70 in) and 142 females (61%, avg. age 35 years, weight 150 lb., height 64 in). There were no significant differences in demographics between the whole group and any of the three subgroups. (ii) Group N v Group NB (low back pain): There were no significant differences in demographics between these two groups. There was a moderate but not significant trend for those victims with low back pain to have a greater degree of lumbar spinal degeneration (for example Group N had 85% with no degeneration and 4% severe vs. Group NB had 57% with no degeneration and 16% severe). (iii) Group N vs NA (arm symptoms): There were no significant differences in demographics between these two groups. There was a trend toward a difference in the percentage of females to males (Group N had 57% female, Group NA had 70% female, [Z = 1.3, p = 0.097]) and a statistical difference with respect to awareness of the impact (Group N had 7% aware and Group NA had 40% aware, [Z = 3.58, p < 0.01]. The differences in pre-existing cervical spine degeneration and head position between these groups was not significant. DISCUSSION: Using the findings of the victim's own health provider removed potential bias that the investigators might have had when interpreting the extent of symptoms of these difficult to quantify injuries. In this study, selection of vehicle speed change eliminated the effect of a very significant confounding variable, the level of force acting on the occupant at impact. CONCLUSIONS: (i) There was a trend (not significant) for victims who reported low back pain after whiplash injury to have a greater degree of preexisting lumbar spinal degeneration. (ii) Having the head turned, being female, or having pre-existing cervical degeneration did not predispose the victim to whiplash associated arm symptoms. (iii) Being aware of the impending impact did result in a greater likelihood of arm symptoms. Paper # 11 Fluoroscopic Evaluation of the Cervical Spine in the Polytrauma Patient Mitchel B. Harris, MD, Steve Kronlage, MD , Phyliss Carboni, RN, Norman Chutkan, MD (New Orleans, LA) INTRODUCTION: To evaluate the safety and efficacy of a fluoroscopic guided spine examination in the trauma victim unable to be "cleared" by standard measures METHODS: All trauma victims with normal trauma cervical spine x-ray series (AP,lat,dens) that were unable to be deemed free of an unstable injury due to an incomplete clinical exam were prospectively enrolled. Reasons for an incomplete examination included the presence of an altered sensorium secondary to drugs or alcohol, significant 'distracting injuries', or those intubated and narcotized prior to their trauma unit evaluation. Utilizing real time fluoroscopy, with the patient anesthetized, 20-25 lbs of traction is applied through a cervical halter. In-line distraction, i.e. the "stretch test" is performed first. If the stretch test is negative (no instability), passive flexion/extension views are performed with direct fluoroscopic visualization. A positive finding on any examination includes the following: >2mm of aberrant disc space distraction, vertebral body translation>3.5mm, or segmental angulation>11 degrees. Additional positive findings would include splaying of the spinous processes, or subluxation of the facet joints. If the cervical spine is unable to be visualized adequately at either junction (oc.-cx'l, or cx'l- thoracic) during the fluoroscopic evaluation, standard measures for cervical spine clearance are utilized while the collar is maintained. RESULTS: During a 27 month period, 113 pts were prospectively enrolled in this IRB approved protocol; 74 males,39 females. 92% were involved in vehicular trauma. Average ISS=17, average age 32. 109/113 have completed data. 9/113 were unable to be cleared intra-operatively due to poor visualization of the cx'l-thoracic junction. 3 positive findings were recorded and these patients were subsequently surgically stabilized. Intra-operative findings confirmed that all three patients had sustained injuries rendering their spines "unstable".83/113 patients were admitted directly to the OR from the trauma unit. 78/83 were able to have the cervical spines cleared within 8 hours of admission, the remaining 5 patients had poor visualization of their cx'l-thoracic junction and thus could not be cleared through the protocol. DISCUSSION/CONCLUSION: This preliminary study illustrates the safety and utility of intra-operative fluoroscopic evaluation of the cervical spine in the polytrauma victim. A similar protocol has been successfully and safely utilized in the isolated head injured patients(Davis et al. J of Trauma'95, Sees et al. J of Trauma'98).Not only does this method of evaluation of the cervical spine allow for early recognition of occult unstable ligamentous injuries, it also facilitates early removal of the hard collar. These occult soft tissue injuries represent an increased risk for catastrophic sequelae which can occur during routine patient care activities. Early removal of the rigid collar in a polytrauma victim will aid in respiratory mangement, avoidance of skin problems and overall improved patient care. Paper # 12 Pediatric Cervical Spine Injuries: A 10 Year Analysis Richard W. Murphy, MD, Dirk H. Alander, Sarah W. Alander, MD, Laura Fitzmaurice, MD (Kansas City, MO) INTRODUCTION: Cervical spine injuries in the pediatric population are rare. The purpose of this study was to describe injuries of the cervical spine and spinal cord in a large urban pediatric population. This large case series helps to delineate demographics and circumstances of various pediatric cervical spine and spinal cord injuries. METHODS: Medical records of all patients presenting to a regional children’s hospital with cervical spine injuries from January 1988 to December 1998 were reviewed. Cervical spine injuries were categorized as bony, ligamentous or injuries to the spinal cord. The severity of spinal cord injuries was defined by ASIA scores. RESULTS: Data from sixty-one patients were obtained. There were 36 males and 25 females ranging in age from 1 month to 16 years with a mean age of 8.8 years. Children less than 8 year comprised 46% (28). Children older than 8 comprised 54% (33). There were an equal number of injuries in both sexes in the younger age group (14 vs14) while injury predominated in the males in the older age group (22 vs11). Death occurred in 11 patients; all were motor vehicle accident victims. Injuries occurred from motor vehicle accidents (62%), falls (16%), sports (10%) and other injuries (12%). Sixty-six percent of cervical spine and 50% of the cord injured patients were unrestrained passengers in motor vehicle accidents. Motor vehicle accidents were the cause of injury in two-thirds of the younger age group and half of the older group's injuries. Cervical injuries were bony injuries in 56% of the patients; pure ligamentous injuries in 34% and 10% were SCIWORA. In all age and type of injury groups upper cervical spine injuries (occiput to C4) were noted in 62%, lower cervical injuries (C5-T1) in 26%, and mid cervical (C4-C6) in 2%. Ligamentous injuries predominated in the younger age group (19/28) while bony injuries predominated in the older age group (24/33). Cervical spinal cord injuries had a near equal distribution between high (54%) and low (47%) cervical levels in both the younger and older age groups. Spinal Cord injury was documented in 49% (30) of our patients. High cord injuries were seen in 41% and lower cord injuries in 50%. ASIA scores were A’s-18, B’s-0, C’s-4, D’s-8 and E’s-31. Two thirds of those with spinal cord injury were in motor vehicle accidents and were unrestrained. Nine (82%) of the 11 deaths were in the younger child in a motor vehicle accident with a high complete cervical cord lesion. A total of 6 SCIWORA injuries occurred: 3 in sports, 2 in motor vehicle accidents and 1 from a fall. DISCUSSION/CONCLUSION: Serious cervical spine injury is rare in the pediatric population. Cervical cord injuries are usually associated with very high-energy accidents and are not likely to be isolated. Our study aggrees with previous studies that younger children have higher cervical injuries, which has been previously explained by differences in anatomy between age groups. Sports injuries had a low incidence of severe cord injury and had a good chance of cord recovery. In our patient population with cervical spine injuries all patients suffering an early death had high-energy injuries, serious associated injuries and were unrestrained passengers in motor vehicle accidents. Paper # 13 Why Acute Cervical Spine Injuries Are "Missed" in Infants and Children: 12-Year Experience from Level I Adult and Pediatric Trauma Center Sohail K. Mirza, MD (Seattle, WA), Anthony M. Avellino, MD (Seattle, WA), Fred A. Mann, MD (Seattle, WA), M. Sean Grady, MD (Philadelphia, PA), Jens R. Chapman, MD (Seattle, WA) INTRODUCTION: To determine the incidence of "missed" diagnoses of acute cervical spine injuries (ACSI) in infants and children (less than15 years old) initially evaluated at an urban Level I Trauma Center (more than 4000 annual "major" trauma admissions). METHODS: Retrospective, single institution, case series of pediatric cervical spine fractures whose diagnoses were "missed" during initial evaluation from July 1985 to August 1997. RESULTS: The prevalence of ACSI in infants and children was 0.8%. Nineteen percent (7/37) were incorrectly interpreted on initial imaging evaluation. Five percent (2/37) were "missed" fractures (i.e., type I hangman’s fracture and segmental fracture of C1 anterior tubercle). Fourteen percent (5/37) were "normal" and developmental variants read as fractures (i.e., condyle fracture, Jefferson fracture, C4 body fracture, C7 spinous process fracture, and atlanto-occipital dislocation). DISCUSSION/CONCLUSION: The "error rate" for infants and children less than or equal to 8 years old was 24% (4/17) and for children greater than 8 years old was 15% (3/20). The occiput to C2 region was the most common site of diagnostic error. The most common factors predisposing to misdiagnosis were: (1) unfamiliarity with pediatric cervical spine anatomy; (2) sub-optimal conventional films; and (3) failure to aggressively use MRI when neurologic status and conventional imaging findings were divergent. Paper # 14 Cervical Interbody Fusion with Threaded Titanium Fusion Cages: LongTerm Results Scott H. Kitchel, MD (Eugene, OR) • (a –Sulzer Spine Tech – BAK-C) INTRODUCTION: Cervical interbody fusion cages represent a new alternative in anterior cervical diskectomy and fusion surgery. This paper represents the follow-up of patients in an FDA approved IDE study to evaluate one such device. METHODS: Following the FDA approved IDE study protocol, 54 patients were prospectively randomized into one of three groups. The BAK interbody fusion cage (BAK-C), the BAK interbody fusion cage coated with hydroxyapatite (BAK-HA), or traditional Smith Robinson fusion (SRF). with autologous iliac crest bone graft. No supplemental fixation was allowed. Data was collected prospectively and at 3 months, 6 months 12 months, 24 months and 36 months on all patients. Inclusion criteria required neck and/or radicular pain. Parameters evaluated included neck pain, arm pain, neurologic function, and fusion. RESULTS: 21 patients were randomized to the BAK-C group, 18 to the BAKHA group, 15 to the SRF group. Follow-up ranged from 24 to 54 months with an average of 36.2 months. There were no statistically significant differences in the groups for neck pain, arm pain, neurologic recovery or fusion rate at 24 months follow-up. The BAK-C and BAK-HA groups did show development of late kyphosis averaging 2.6 degrees. DISCUSSION/CONCLUSION: Titanium threaded cervical interbody fusion cages do allow safe and effective cervical fusion. They yield results similar to a traditional Smith Robinson fusion with autologous bone graft harvest. The functional significance of the late kyphosis requires further study. • If noted, the author indicates something of value received. The codes are identified as: a – research or institutional support, b – miscellaneous funding, c – royalties, d – stock options, e – consultant or employee. For full information, refer to inside back cover. Paper # 15 Prospective Randomized Multi-Center Clinical Trial of Cervical Fusion Cages Robert Hacker, MD (Eugene, OR), Joseph Cauthen, MD (Gainsville, FL), Thomas Gilbert, MD (St. Louis Park, MN) INTRODUCTION: Threaded cages have been developed for use in anterior cervical interbody fusions to obviate the need for allografts or autogenous bone grafting procedures while providing initial stability during the fusion process. The BAK/C® (Sulzer SpineTech, Minneapolis, MN), has been evaluated in a randomized, multicenter Investigational Device Exemption (IDE) clinical trial that began in December 1994. The purpose of this study is to report on the results from this trial that compared fusions performed with the BAK/C implant to non-instrumented bone only fusions. MATERIALS AND METHODS: The study protocol, conducted at twenty-six investigational centers with 50 surgeons investigators, was prospective with randomization (after obtaining informed consent) of patients to three different study cohorts of anterior cervical fusions: instrumented with HA- coated BAK/C®, instrumented with non-coated BAK/C, and uninstrumented, bonegraft only fusions (anterior cervical decompression with fusion: ACDF). Inclusion criteria were radiographic evidence of symptomatic discogenic radiculopathy from one or two contiguous cervical spine segments (C3 to C7); no moderate to severe myelopathy; no previous fusion at the affected level, cervical trauma, or rheumatoid disease; age 18 to 65; no systemic infection; no significant metabolic, circulatory, cardiac or pulmonary diseases; and no active malignancy. Post-operative follow-up data was collected at 3-, 6-, 12- and 42months. Fusion was assessed on flexion/extension radiographs by an independent orthoradiologists who based his assessment on lack of visible motion or significant radiolucenies; for two-level cases, both had to be fused to be considered a success. Patient-based outcome was assessed by a 10-point visual analog scale measuring neck and radicular pain, and a SF-36 questionnaire to evaluate overall function. Patients were also asked to rate their overall perception of their procedure as excellent, good, fair, or poor. Complications were recorded at each follow-up visit. RESULTS: Total enrollment in this ongoing clinical trial at the time of this report was 488 patients; data analysis included 288 patients at one-year and 140 at two years. There were 79.9% one-level cases and 20.1% two-level cases performed. Overall no difference could be detected between the HA-coated BAK/C group and the non-HA-coated BAK/C group in any of the outcome measures evaluated, therefore these experimental groups were combined and compared to the control group of bone-only fusions. Mean duration of surgery (99.5 vs 94.1 min) and a mean hospital stay (1.5 days) were not significantly different between the two groups (BAK/C vs ACDF, respectively). Mean blood loss was low in both BAK/C (74.6cc) and the ACDF (87.5cc) groups with a statistically larger amount of blood loss in the ACDF group (p<0.05cc). 3.2% of BAK/C patients required an iliac crest bone harvest compared to 69.0% of ACDF (p<0.001, Chi-square test). Neck pain was significantly reduced immediately post-surgery (e.g., 6.1± 2.7 vs 3.2 ± 2.6; pre-operative vs postoperative for the BAK/C group) with no statistical difference in the change detected between the BAK/C or ACDF group (p>0.05). Similarly, radicular pain in the affected limb was improved for both BAK/C and ACDF groups (e.g., 4.2 ± 3.4 vs 2.0 ± 2.5; pre-operative vs post-operative for the BAK/C group). These symptom improvements were maintained at the two-year followup. SF-36 scores demonstrated that pre-operatively, the patient cohort was, indeed, not a sampling from normal healthy individuals. Furthermore, significant improvements were noted similarly for both groups 6 months after surgery in both the Physical Component Subscale (PCS) and the Mental Component Subscale (MCS) with the improvements maintained at two-years. At 12 months post-surgery 76.5% of patients in the BAK/C group perceived their surgery outcome to be excellent or good and 74.7% of patients in the ACDF groups rated their procedure as excellent or good. Fusion outcome is given in Table 1. Table 1 BAK/C ACDF Follow-up 1-Level 2-Level 1-Level 2-Level 6 months 98.8% 79.6% 80.6% 78.6% 12 months 98.7% 80.0% 86.4% 80.0% 24 months 100% 91.7% 96.4% 77.8% Overall complication rate, defined as the number of patients experiencing one or more events, in the BAK/C group of 346 patients was 10.1% with the three most frequent events noted being dysphagia/hoarseness or wound-related problems or pseudoarthrosis (1.7% for each of these three). Continued or new symptoms were seen in 1.4% of the cohort. By comparison, the ACDF group of 142 patients showed an overall complication incidence of 16.2% with woundrelated problems seen in 6.3% of patients, dysphagia/hoarseness in 2.1% of patients, and graft collapse/migration in 4.8% of patients. Complications requiring a second operative procedure occurred in 4.0% of the BAK/C group and 4.2% of the ACDF (NS, Chi-square test). DISCUSSION/CONCLUSIONS: The use of an interbody bone graft for anterior cervical fusion after discectomy is a well-accepted, well-characterized procedure with predicable outcomes even in light of current controversy regarding the use of internal plating. The BAK/C threaded fusion cage represents an alternative that obviates the need for allograft and its inherent potential problems of reduced incorporation, as well as reduces the necessity for autograft harvesting and the associated potential morbidity. This large multi-center trial demonstrated that outcomes after a cervical fusion procedure with the BAK/C device are comparable to those of a conventional uninstrumented, bone-only anterior discectomy and fusion. The low risk (i.e., low incidence of complications) and the positive clinical outcomes demonstrated in this study suggest that threaded cervical interbody cages are a viable alternative to add the armamentarium of the spine surgeon. Paper # 16 Prognostic Value of Intraoperative Ultrasonography in Cervical Spondylotic Myelopathy Katsuhiro Fujioka, MD (Kani, Gifu, Japan), Kouichi Hashimoto, MD (Gifu, Japan), Akihiko Kurachi, MD (Gifu, Japan), Yukihiro Matsuyama, MD (Aichi, Japan) INTRODUCTION: It is very difficult to predict postoperative prognosis in cervical spondylotic myelopathy. To evaluate the reliability of three examinations in this aspect, We reviewed clinical records, MRI, CT-myelo and intraoperative ultrasonogram of 35 patients operated by laminoplasty. MATERIAL AND METHODS: 35 patients with cervical spondylotic myelopathy treated by laminoplasty from May,1996 to July,1998. 69% were male, 31% were female, and the average age was 60 years. The severity of myelopathy was evaluated according to the Japanese Orthopaedic Association's scoring system and surgical outcomes were evaluated by the formula of Hirabayashi for determining recovery rate. MRI were reviewed pre and post operatively, examining high intensity area in T2-weighted images and atrophy or enlargement of spinal cord. CT-myelo were reviewed pre and post operatively, measuring compression rate of both spinal cord and dural canal. Intraoperative ultrasonograph were also reviewed, measuring compression rate of spinal cord after decompressive procedure. RESULTS: The average recovery rate was 50% (range 0-100%). The high intensity area in T2-weighted images were seen in 60% of preoperative MRI and 80% postoperatively. Atrophy of spinal cord were seen in 54% of postoperative MRI and enlargement were 14%. The average Compression rate of spinal cord were 58% and 80% in pre and postoperative CT-myelo respectively. Those of dural canal were 63% and 89%. The average compression rate of spinal cord after decompressive procedure in intraoperative ultrasonograph were 85%. In these four factors (recovery rate, MRI, CT-myelo and intraoperative ultrasonograph), recovery rate were correlated only with intraoperative ultrasonograph with statistically significant differences (p<0.001). DISCUSSION/CONCLUSION: The plasticity of spinal cord just after the decompressive procedure can be evaluated only by Intraoperative ultrasonography. In this study we demonstrated that recovery rate and intraoperative ultrasongraph were correlated mutually. We concluded that Intraoperative ultrasonography is very valuable examination to predict the prognosis of postoperative cervical spondylotic myelopathy. Paper # 17 Inter- and Intra-observer Reliability of the Japanese Orthopaedic Association Scoring System for Evaluation of Cervical Myelopathy Kazuo Yonenobu, MD (Suita, Japan), Kuniyoshi Abumi (Sapporo, Japan), Kensei Nagata (Kurume, Japan), Eiji Taketomi (Kagoshima, Japan), Kazumasa Ueyama, (Hirosaki, Japan) INTRODUCTION: Several assessment scales for clinical outcome of treatment of cervical myelopathy have been proposed. Most of them include items evaluated by observers. However, no system has been validated in terms of inter- and intra-observer reliability. The Japanese Orthopaedic Association has proposed a scoring system for cervical compression myelopathy (revised JOA Scoring system(17-2 points)), which consists of 7 items: motor function of fingers, shoulder and elbow, and lower extremity, sensory function of upper extremity, trunk and lower extremity, and bladder function, and evaluates the severity of myelopathy by rating each item. This scoring system has been widely used in Japan, and employment of it in countries other than Japan has recently increased. We studied the inter- and intra-observer reliability of the JOA system. METHODS: The observer was 10 experienced spine surgeons, 10 orthopedic surgeons eligible for the Japanese Board of Orthopedic Surgery and 13 residents in the first or second year of orthopedic residency programs in 5 different university hospitals. Twenty-nine patients with myelopathy secondary to ossification of the posterior longitudinal ligament were the subjects of this study. Inter- and intra-observer reliability of total score for each item was evaluated with the intra-class correlation coefficient, and extension of the kappa coefficient (Kappa) for each item was also calculated to assess reliability of multivariate categorical data based on the results of observer reliability study. RESULTS: The inter-observer reliability of the total score was 0.813 ( 95% confidential interval (C.I.) was 0.704 - 0.888). Inter-observer and intra-observer reliability was 0.826 ( 95% C.I. 0.729 - 0.893). Regarding level of experience, the inter-observer reliabilities of the experienced spine surgeons and the orthopedic surgeons eligible for the board were high (0.808 and 0.916 respectively), while that of the residents was 0.649. The Kappa for inter-observer data were relatively low for motor function of shoulder-elbow and sensory function of lower extremity. The Kappa for intra-observer data tended to be high for each item. CONCLUSIONS: The inter- and intra-observer reliability of the JOA scoring system were high, suggesting that the JOA system is useful for assessment of cervical compression myelopathy. Paper # 18 Atrophic Myelopathy: Late Sequelae of Cervical Spondylotic Myelopathy Laurence D. Rhines, MD (Baltimore, MD), Seth M. Zeidman, MD (Rochester, NY), Thomas B. Ducker, MD (Annapolis, MD) INTRODUCTION: The optimal management of patients with cervical spondylotic myelopathy remains incompletely defined. Part of the reason for this is a failure to understand the natural history of this process and the longterm results of surgical intervention. We present a retrospective analysis of 134 patients with between 5- and 10-year follow-up after cervical laminectomy. METHODS: Between 1977 and 1991 we operated on 328 patients for progressive cervical spondylotic myelopathy. Three hundred patients had at least 2 years of sufficient follow-up to permit detailed analysis. Extended follow-up (5-10 years) was available for 134 of the 300 patients. Patients were graded both pre- and postoperatively using our modification of the Nurick grading scale. RESULTS: Of the 134 patients, 66 remained neurologically stable, 22 had minor deterioration but did not fall a complete grade, and 46 patients suffered deterioration by one or more Nurick grades. Definition of the etiology of the deterioration in this last group of patients is most important. Of the 46 patients that deteriorated, 13 had surgically correctible causes, including insufficient laminectomies, postlaminectomy membrane, or anterior compression necessitating anterior decompression and were operated on with favorable outcome. Thirty-two patients had demonstrable marked cord atrophy on follow-up neuroimaging. DISCUSSION/CONCLUSION: Our data suggest that approximately one quarter of patients with cervical spondylotic myelopathy who undergo decompressive surgical procedures will, nonetheless, develop spinal cord atrophy. Progressive atrophy and associated myelopathy was greatest in those patients with preoperative deficits of long duration. This may have profound implications regarding the role of early intervention in preventing irreversible spinal cord changes in myelopathic patients. Paper # 19 Clinical Significance of the Spinal Cord Atrophy on the Late Long Tract Dysfunction after Decompression Surgery for Cervical Spondylotic Myelopathy Hironari Takaishi, Morio Matsumoto, Masahiko Watanabe, Hirofumi Maruiwa, Kazuhiro Chiba, Yoshikazu Fujimura, Yoshiaki Toyama (Tokyo, Japan) INTRODUCTION: Cervical spondylotic myelopathy (CSM) is a common disease caused by chronic segmental compression of the spinal cord. Although the central gray matter and lateral tracts are vulnerable in cases of severe compression, the anterior and dorsal columns were remarkably resistant to degeneration. It still remains unclear whether postoperative late neurological deterioration in CSM is associated with the preoperative severity of cord damages and age-related changes. The purpose of the current study was to investigate the effectiveness and limitations of anterior decompression for CSM by analyzing the changes in the spinal cord with magnetic resonance images (MRI) in patients who were followed for more than 10 years after surgery. MATERIALS AND METHODS: Forty-nine patients with CSM (spondylosis in 30 cases and disc herniation in 19 cases) treated by anterior spinal decompression and fusion from 1964 to 1986 at our hospital were examined in this study. The postoperative follow-up period ranged from 10 to 27 years with a mean of 15.7 years. We analyzed the correlation between the neurological recovery and the imaging parameters. Postoperative changes in the spinal cord were investigated at the fused levels and the adjacent segments with 1.5 tesla fast spin echo MRI by measuring signal intensity ratios in the gray matter and the transverse area of the spinal cord with a digitizer linked to computer. The spinal cord atrophy was evaluated by comparing asymptomatic normal patients divided into age and gender categories, previously reported. Clinical records on the duration of diseases, age at the time of surgery and neurological status were retrospectively reviewed. RESULTS: The transverse area of the spinal cord after decompression correlated significantly with the neurological recovery rate positively (R=0.57) and preoperative severity of myelopathy negatively (R=0.39). Of 19 patients with a transverse area of less than 40 mm2, 8 (42.1%) had a deterioration of trunk and lower extremity function with spasticity. Nine of 49 patients (18.4%) showed marked cord atrophy extended cephalad and caudal over the entire cervical spine with T2-high intensity area on MRI. Focal or multi-segmental high intensity of T2-weighted images was noted in 21 patients (42.9%), low intensity of T1weighted images was observed in 8 patients (16.3%). The presence of these signal intensity changes on MRI tended to have poor surgical results associated with upper extremity dysfunction but without the long tract deterioration. On the other hand, duration of diseases and age at surgery was not predictors for surgical outcomes. DISCUSSION/CONCLUSIONS: In general, patients with great degree of cord damages have a lesser chance of recovery after decompression. It has been suggested that the presence of abnormal signal intensity on MRI should not be taken as a contraindication for surgery, because decompression and stabilization of a progressively deteriorating patient may still give satisfactory result. This study revealed that the cord atrophy, which may be caused by the promotion of age-related degenerative changes based on irreversible cord damages, was a significant factor of severe long tract dysfunction, such as demyelination in the white columns. In addition, marked cord atrophy with abnormal signal intensity may represent terminal degenerative changes of extensive infarction of both gray matter and lateral corticospinal tracts. Therefore adequate criteria for selection of operative indication should be considered to maintain a favorable long-term result. Paper # 20 Treatment of Complex Cervical Myelopathy Thomas B. Ducker, MD (Annapolis, MD), Laurence D. Rhines, MD (Baltimore, MD) INTRODUCTION AND METHODS: In our clinically severe myelopathic patients, 6.1% over the last seven years have presented with nonambulatory myelopathy (Nurick Grade IV or V), diffuse congenital spinal stenosis, superimposed multilevel spondylotic stenosis, various degrees of kyphosis (2 to 28 degrees), and diagnostic MRIs with edematous areas within the spinal cord. For this patient population (33 cases) our treatment has been multilevel anterior cervical discectomies and segmental decompressions, followed by oversized freeze dried fibular ring allograft segmental reconstruction and restoration of alignment (4 to 5 levels) without anterior plating. Immediately, the patient is turned on the operative wedge frame from supine to prone (tong traction at 15 pounds). Then, a posterior cervical decompressive laminectomy (usually C3 to C7) is completed. The neck is further extended and a posterolateral plate fusion is performed. Pedicle screws are used in C2 and sometimes in C7 or T1 and local bone autograft is placed posterolaterally. Postoperatively, the patient is placed in a 2-poster brace during the daytime and in a Philadelphia collar at nighttime for 6 to 8 weeks. RESULTS: Total operative time for this procedure is 5.7 hours (2.0 hrs anteriorly, 0.6 hrs for the turn, and 3.1 hrs posteriorly). Total blood loss averaged only 422 ml and no patients required transfusion. There were no wound infections and no new postoperative deficits. No patients required feeding tubes or parenteral nutrition for swallowing difficulties, although 6 patients were on liquids primarily for the first 3.8 days after surgery. There were no vocal cord problems. At one year, all patients had fused all segments. Corrective lordosis was maintained. Two patients had early subluxation at the bottom of the fused segment (C7 on T1) of 3 and 4 mm, respectively, but this was without sequelae and required no further surgery. No grafts dislodged anteriorly and no screws pulled out posteriorly. Neurologic recovery was greater in the upper extremities than the lower extremities. Arm/hand power (normally 50) changed from 33 +/- 7 to 46 +/- 8. Leg function improved from 32 +/- 9 to 39 +/- 10. Reflexes changed little. Thirty-one of the 33 patients were ambulatory, but Canadian crutches carried much of the weight in 8 of these patients. The two remaining patients improved from Nurick grade V to IV and follow-up MRI showed severe cord atrophy. DISCUSSION: The combined anterior/posterior approach on average did not take longer than the more extensive corpectomies with anterior plate reconstruction. More importantly, corrective lordosis was maintained and the complications were considerably less. The combined operation was well tolerated by the patients (hospital stay averaging 3.2 days). Severely myelopathic patients were routinely transferred to a rehabilitation facility on the third postoperative day. Paper # 21 Minimum 10 Years Outcomes of Operative Treatment of Cervical Myelopathy Due To Ossification of the Posterior Longitudinal Ligament Kosei Ijiri, Shunji Matsunaga, Hiroki Koga, Shinji Nakahara, Kazunori Yone (Kagoshima, Japan) INTRODUCTION: Surgery for ossification of the posterior longitudinal ligament (OPLL) is indicated for patients who have moderate or severe myelopathy. Few reports are available comparing the surgical outcomes of anterior and posterior procedures. The purpose of this study was to determine the long-term results of operative treatment (anterior decompression and fusion, laminectomy, laminoplasty) for cervical myelopathy due to OPLL and to determine those factors influencing the postoperative clinical course. METHODS: A retrospective study was performed for patients who had undergone operation for cervical OPLL before 1988 in our institute. Eighty-six patients were followed for an average of 13.9 years, with a minimum of 10 years. Thirty-two patients underwent anterior decompression and fusion, 13 patients laminectomy, and 41 laminoplasty. There were no statistically significant in age at operation, or follow-up period among the three operation groups. All patients were examined for clinical course of neurological symptoms and functional abilities by Japanese Orthopaedic Association (JOA) score. Factors possibly affecting the course of postoperative changes are discussed. Statistical analysis was performed by means of Student’s t-test for parametric variables and the _2 test for independent samples analysis. RESULTS: Postoperative outcomes are shown for three years after operation and at final follow-up (Fig. 1.). Age at operation, history of trauma, severity of myelopathy before operation and period of disease before operation were factors affecting short-term results. Adjacent segmental instabilities in the anterior decompression and fusion group, and spinal canal stenosis at the thoracic or lumbar level in the laminoplasty group were factors associated with worsening of the clinical symptoms during long-term follow-up. Progression of ossification after the operation had not effect on deterioration. Seven (9%) patients underwent additional surgery of the thoracic and lumbar spine, and all exhibited functional improvement after additional surgery. DISCUSSION: Our indications for anterior surgery is segmental type, involvement of less than 3 segments, and severe kyphosis. Posterior decompression is performed for widely extended type lesions, combined developmental stenosis and multisegmental type lesions of more than 3 segments. This study showed that our indications are reasonable and acceptable from the point of view of long-term results. However, clinical function deteriorated slightly over long-term observation with increasing age. It is known that thoracic and lumbar canal stenosis due to ligament ossification is present in 60% of cervical OPLL patients. Not only cervical lesions but thoracic or lumbar stenosis should be treated to prevent deterioration of symptoms in surgicallytreated cervical OPLL patients. CONCLUSION: Long-term follow-up of patients with surgically-treated cervical OPLL revealed the effectiveness of surgery. The adjacent segmental instability in the anterior decompression and fusion group, and canal stenosis in the laminoplasty group were factors related to postoperative deterioration. Paper # 22 Influence Of Minor Trauma To Neck In Cervical OPLL And Cervical Spondylotic Myelopathy Kyoung-Suok Cho, MD, PhD, Chun-Kun Park, MD, PhD, Choon-Keun Park, MD, PhD, Sung-Chan Park, MD, PhD, Do-Sung Yoo, MD, PhD, Pil-Woo Huh, MD, PhD, Dal-Soo Kim, MD, PhD, Joon-Ki Kang, MD, PhD (Seoul, Korea) INTRODUCTION: Degenerative disease of the cervical spine (including cervical OPLL and cervical spondylotic myelopathy) is the most common causes of the cervical spinal stenosis. The influence of minor trauma to the neck on the neurological outcome in patients with cervical OPLL and cervical spondylotic myelopathy was evaluated retrospectively. MATERIALS AND METHODS: Out of 147 patients treated in our department for cervical OPLL and cervical spondylotic myelopathy between 1992 and 1997, 45 (30.6%) had sustained minor trauma to the cervical spine. Of these 45 patients, 13 developed myelopathy, 9 showed deterioration of preexisting myelopathy, and no neurological change was observed in 23 patients. Minor trauma included motor vehicle accident, fall-down, slip down, being struck by an objects and sports activity. RESULTS: Regarding the relationship between the diameter of the residual spinal canal and the neurological outcome in these 45 patients, 16 out of the 18 patients with a narrow spinal spinal canal (< 10 mm) developed neurological deterioration, whereas that occurred in 6 of the 27 patients with a wider spinal canal (> or = l0 mm)(P<0.05). The surgical outcome is poorer in narrow spinal canal group (33.3% improvement) than in wider spinal canal group (81.4% improvement)(P<0.05). DISCUSSION/CONCLUSION: These results indicate that even indirect minor trauma to the neck can cause irreversible changes in the spinal cord if there is marked stenosis of the cervical spinal canal; such patients who are at risk, must be educated. Paper # 23 Quantitative Assessment of Cervical Spondylitic Myelopathy by a Simple Walking Test Alan Crockard, FRCS, Anoushka Singh, MSc (London, England) INTRODUCTION: One of the great problems in assessing outcome is to define exactly what the disability was before and after surgery. The Japanese Orthopaedic Association Score (with European modifications) do provide some help but it is essentially questionnaire rather then a measurement. The object of this research has been to establish if a simple reliable test of motor function could be used in patients presenting with cervical spondylitic myelopathy. METHODS: The present study explores the use of a timed walking test, 15 metres turning and walking 15 metres back to see if there are differences between patients which cervical myelopathy and the normal population and if there is a difference before and after surgery in the same patient. Time taken to do the walk, turn and walk back was noted as well as the number of steps. Forty-one patients with CSM consecutively referred to four neurosurgeons independent of study team were assessed prior to surgery and at three, six months and nine months after surgery for cervical myelopathy. Each patient group was compared with an equal number of controls and was matched for age and sex. Each trial was performed three times, the patient was requested to walk at his maximal comfortable speed. A large control population had been previously measured. It was shown that there was good inter and intra observed correlation for the tests in controls and on the patients. RESULTS: There were 26 males with a mean age of 59.4 ± 11.9 years and 15 females with a mean age of 62.9 ± 18.2. The mean preoperative walking time for the control group was 24.3 seconds ± 0.8 seconds and the mean number of steps taken to walk 30 metres was 46.9 ± 1.2 steps. In those patients with a cervical myelopathy the mean preoperative walking time was 85.4 ± 11.2 and this was significantly worse than for healthy age and sex matched controls (2 tailed T-tests unpaired unequal variants p = 2.4 x 10–6). The mean postoperative patient walking time was 64.7 ± 8.4 this was a significant improvement in the patients following surgery but still slower than the control population (2 tailed paired T-test p = 0.0018). The mean number of steps taken to walk the 30 metres was 74.8 preoperatively and 63.5 postoperatively. For control, the mean number of steps to walk the same distance was 46.9. The surgery carried out for these patients was entirely within the practice of each surgeon, most were one or two level anterior cervical discectomies with and without fusion. A small number had laminoplasties. CONCLUSIONS: The study indicates that the easily performed walking test may be a suitable measure of a degree of myelopathy associated with cervical spondylosis. It is reproducible and reliable with a low inter trial variability. The distance of 30 metres is long enough to measure the time relatively accurately, the turn at the 15 metre mark is also considered to be important. If these tests can be substantiated in multicentre trial it would provide another method to quantify disability and thus open up the possibilities of objective assessment. Paper # 24 Cervical Cord Compression and the Hoffmann's Sign John A. Glaser, MD, Joel Cure, MD, Kelly Bailey, PA-C, David Morrow, MD (Charleston, SC) INTRODUCTION: The Hoffman’s sign is used as a clinical indicator of dysfunction of the cervical spinal cord. This purpose of this study was to correlate the Hoffman’s sign with the presence or absence of cervical spinal cord compression when imaging of the spinal canal was employed. METHODS: All new patients with complaints related to their cervical spine seen by a single spine surgeon were evaluated. The study period was from May of 1997 through February of 1999. To evaluate for the presence of a Hoffman’s sign the nail on the middle finger of each hand was flicked with the cervical spine in the neutral and forward flexed position. Any flexion of the ipsilateral thumb and/or index finger was recorded as a positive sign. All radiographic images of the spinal canal, MRI and /or CT, were reviewed for evidence of cord compression by the treating physician. All MRI scans done at our institution with a standardized technique were read by a neuroradiologist with no knowledge of the patient’s clinical condition. Cord compression was defined as; obliteration of CSF in the area of compression and any deviation of the shape of the cervical cord when compared to normal levels. Sensitivity, specificity and predictive value of the Hoffman’s sign for compression were then calculated. RESULTS: 165 patients were evaluated by the treating physician. 104(63%) were female and 61 male with a mean age of 48.9 years. 49 patients had a positive Hoffman’s sign, 39(79.6%) of these were female. 124 patients had undergone imaging of the spinal canal. For these patients the sensitivity of the Hoffman’s sign relative to cord compression was 58% and the specificity was 78%. The positive predictive value, the percentage of patients with a positive Hoffman’s that had compression, was 62% and the negative predictive value 75%. 49 of these patients had MRI studies read by a blinded neuroradiologist. For these patients the sensitivity was 33%, specificity 59%, positive predictive value 26% and negative predictive value 67%. DISCUSSION: Although attractive as a simple method of screening for cervical cord compression, the Hoffman’s sign, in the absence of other clinical findings, is not in our experience a reliable test for doing this. Paper # 25 Long Term Follow-up Study of Anterior Decompression and Fusion in Patients with Cervical Myelopathy over Ten Years Yoshiyasu Arai, MD, PhD (Bunkyo-ku, Tokyo, Japan), Hiromichi Komori, MD, PhD (Bunkyo-ku, Tokyo), Kiyoshi Mochida, MD, PhD (Mishima-City, Shizuoka), Kenichi Shinomiya MD, PhD (Bunkyo-ku, Tokyo, Japan) INTRODUCTION: We have prospectively performed cervical anterior surgery for cervical myelopathy patients. The purpose of this paper is to analyze long term follow-up results of cervical anterior surgery, and to report the usefulness and pitfalls of this procedure. MATERIALS AND METHODS: From 1980 to 1988, 127 cervical surgeries were performed in our hospital. Ninety-nine cases (78%) were operated on with anterior decompression and fusion. Another 28 cases underwent laminoplasty because of being elderly over 75 or OPLL extension to the C2 level. We had a chance to examine 79 anterior surgery cases (follow-up rate 80%). Among them, 16 cases were excluded because 10 cases were already dead within 10 years after the operation, and six cases had cerebral palsy. Consequently, 63 cases were followed-up for more than 10 years (cervical spondylotic myelopathy (CSM): 26, OPLL myelopathy: 27, disc herniation (CDH): 10). The patients' average age at the time of the operation was 54.3 years old and the average follow-up term was 13 years (10-18 years). One-level fusion was performed in 18 cases, two-level in 14, three-level in 18, four-level in 8 and five-level in 2. We studied operative results at the time of pre and post operation and the final follow-up to find important factors that affected on final outcome. RESULTS: Averaged JOA scores (Japanese Orthopaedic Score for cervical myelopathy) were 9.2 before operation, 14.9 after the operation (maximum improvement rate 70%), 13.6 final (final improvement rate 55%). Satisfactory JOA score appeared to be maintained until the final follow-up. However, seventeen cases showed deterioration in which nine patients underwent reoperation subsequently. Short term follow up: The averaged maximum recovery rate was the highest in one or two level fusion (1-level; 77%, 2-level: 78%, 3-level: 70%, 4 or 5-level: 67%). The cervical disc herniation group showed the highest recovery rate (CDH: 87%, CSM: 59%, OPLL: 73%). Long term follow up: With regard to fusion number, the final recovery rate was the highest in one level fusion (1-level: 63%, 2-level: 53%, 3-level: 50%, 4or5level: 51%). In each group, recovery rate showed deterioration compared with short-term follow up, but it was maintained over 50%. As to pathology, the final recovery rate didn't show any difference between each group (CDH: 57%, CSM: 57%, OPLL: 53%). Deterioration occurred within 5 years following first operation in 15 out of 17 deteriorating cases (CDH: 2 cases, CSM: 3 cases, OPLL: 12 cases). Among them, six OPLL patients had development of ossification and the other 11 deteriorated due to instability of adjacent discs. Six patients also had lumbar disorders other than cervical problems (In 4 cases, lumbar operations were done while under treatment.). Nine cases (CDH: 1 case, CSM: 2 cases, OPLL: 6 cases) underwent reoperation from two to 15 years (average 7.9 years) following first operation. Only one case underwent additional anterior decompression and the other patients were operated on with laminoplasty. The details of averaged JOA score in these cases as follows; 8.7 before first operation; 14.1 highest score after the first operation (first improvement rate 66%); 8.7 just before reoperation; and 11.9 final. The final recovery rate was above 50% in each disorder. DISCUSSION / CONCLUSION: At the time of the final follow-up, the averaged age was 67.3 years old and almost 40% were aged above 70 years. Although they were old enough to show functional deterioration caused by aging, 73% of the patients maintained satisfactory function even after more than 10 years. We make a diagnosis for responsible lesions by using evoked spinal cord potentials to localize the operative levels. Patients with disc herniation showed higher postoperative recovery rate because they had a single level lesion and underwent one or two level fusion. However, recovery rate of the OPLL group was lower because they had multilevel fusion due to multiple neurological lesions. Although nine patients (14%) needed reoperation, the result of second surgery was satisfactory. There are two factors for deterioration. One is the adjacent disc problem, and the other is development of OPLL. To avoid it, we can not decide operative levels based only on neurological responsible lesions but also on morphological findings such as disc degeneration. We also believe that the entire length of OPLL should be included for decompression. Paper # 26 Postoperative Alignment of the Cervical Spine After Expansive Open Door Laminoplasty With or Without Lamina Spacers for Cervical Stenotic Myelopathy Kazuhiko Satomi, MD, Jun Ogawa, MD, Masato Takahashi, MD, Masaichi Hasegawa, MD (Tokyo, Japan) INTRODUCTION: Expansive open-door laminoplasty is widely performed for the treatment of cervical stenotic myelopathy. We started to use hydroxyapatite lamina spacers for prevention of reclosing of the opened laminae since 1993. However postoperative curvature of the cervical spine by this method is unknown. The purpose of this study was to determine the postoperative curvature of the cervical spine after the procedure comparing with the methods with out without lamina spacers. METHODS: The subject of these studies were 81 patients. Thirty-nine patients underwent laminoplasty with lamina spacers (Group A) , and another 42 patients underwent laminoplasty without lamina spacers (Group B). A mean age of the patients is 59 years in Group A, and 55 years in Group B. Preoperative severity of the clinical symptoms was evaluated by a scoring system proposed by the Japanese Orthopaedic Association which had 17 points in fill (JOA scores). Postoperative recovery of the clinical symptoms was evaluated by this score and recovery rates were calculated using this score. Radiographically, a sagittal diameter of the spinal canal, curve index by Ishihara’s method and range of motion between C2 and C7 preoperatively and postoperatively. Ishihara’s curve index shows state of lordosis of the cervical spine, and an increase of the index means cervical spine becoming lordotic than before. A mean follow-up period were two years in Group A and five years in Group B. RESULTS: The mean JOA scores increased to 14.3 from 9.1 in Group A and to 13.7 from 9.1 in Group B, postoperatively. Recovery rates were 65.1% in the former and 58.1% in the latter. Mean increase of the sagittal spinal canal was 3.9mm in Group A and 4.1mm in Group B. Curve indexes increased 1.5 in Group A, however it decreased 2.4 in Group B. Mean reduction rates of the range of motion of the cervical spine were 32.9% in the former and 61.7% in the latter. DISCUSSION/CONCLUSION: The open-door laminoplasty for the cervical stenotic myelopathy was known as an easier method among various laminoplasties. Reclosure of the opened laminae was reported with deterioration of the neurological symptoms, because opened laminae were tied by threads to the capsule of the facets in an original method. We started to use lamina spacers between the opened laminae in 1933. Present study showed that the modified method developed similar good clinical results and widening of the spinal canal comparing with original method. The increase lordosis and a smaller decrease of the range of motion were achieved after the laminoplasty with lamina spacers than that without lamina spacer. As a conclusion the open-door laminoplasty with lamina spacers was acceptable for the treatment of the cervical stenotic myelopathy. Paper # 27 ‘ T-Saw’ Laminoplasty for the Management of Cervical Spondylotic Myelopathy: Clinical and Radiographic Outcome Charles C. Edwards, MD, D. Hal Silcox, John G. Heller (Atlanta, GA) INTRODUCTION: Spinous process splitting laminoplasty for the management of cervical spondylotic myelopathy has been well accepted in Japan. The results in non-Japanese patients are unknown. METHODS: Eighteen patients underwent expansive, midline,‘T-Saw’ laminoplasty (C3-C7) for multi-level cervical spondylotic myelopathy at a single institution. Independent clinical and radiographic evaluations at latest follow-up (mean 24 months, range 18-36 months) were performed by a single physician. Objective measures included a patient self-assessment questionnaire, physical examination, Pavlov’s ratio, sagital canal diameter (CT), cord compression index, cervical lordosis, ROM and complications. RESULTS: Progression of myelopathy was arrested in all patients. Patients reported improvement in subjective symptoms: strength (78%), dexterity (67%), numbness (83%), pain (83%), and gait (67%). Bowel/bladder compromise resolved in 5/6 patients. The mean Nurick score improved from 2.7 to 0.9 (p<0.001). Mean score on the Robinson pain scale improved from 2.0 to 0.89 (p=0.002). Narcotic use decreased by 87%. Objectively, 68% of patients with motor weakness regained normal strength (p=0.001), while 50% regained normal sensation (p=0.003). Radiographic canal expansion was verified by a statistically significant increase in the mean Pavlov ratio (0.63 to 1.08) and the mean osseous sagital CT measurement (10mm pre-operatively to 18mm after laminoplasty). The mean cord compression index improved from 0.49 to 0.61 (p=0.01). There was no significant change in mean cervical lordosis. Cervical F/E motion (C2 to C7) decreased from 37 to 23 degrees (p=0.05). Graft dislodgment or segmental instability did not occur. Complications were infection (1) and persistent post-operative motor root lesion (C5) (1). CONCLUSIONS: 'T-Saw' laminoplasty appears to be a safe and effective method of arresting the progression of myelopathy and allowing marked functional improvement in a majority of patients with multilevel cervical spondylotic myelopathy. Paper # 28 Expansive Cervical Laminoplasty: Follow-Up Evaluation Of Residual Flexibility, Pain, And Outcome Julie A. Long, MD, Joseph H. Perra, MD, Timothy A. Garvey, MD, Ensor E. Transfeldt, MD, Micheal D. Smith, MD (Minneapolis, MN) OBJECTIVES: To evaluate longer term results of cervical laminoplasty, utilizing the mini titanium plate technique, with emphasis on residual flexibility and pain. METHODS: We reviewed patients with multilevel cervical spondylosis with myelopathy that underwent an expansile open door cervical laminoplasty. We identified 37 patients that underwent cervical laminoplasty from 1993-1998. All but six patients had frank signs of cord compression All had either congenitally or acquired narrow spinal canals as defined by a pavlov's ratio of 0.8 or less. The laminoplasty was performed from two to six levels, internal fixation with titanium mini-plates was used in all cases. We evaluated plain radiographs, an SF-36, a supplementary questionnaire, charts, and when possible a current physical exam. RESULTS: Average follow-up was 19 months (3 months - 5 years). Ten patients had mild to moderate residual deficits. Five patients were worse or showed no change at recent follow-up. Average follow-up cervical lordosis was -14 degrees (+25 to -60). Five patients had follow-up cervical kyphosis that averaged 12 degrees (+1 to +25). Three patients developed instability postoperative. The average radiographic ROM of flexion to extension from C2-C7 was 45 degrees pre- operatively, and 33 degrees at follow-up. ROM from within the laminoplasty, pre-operatively was 29 degrees, follow-up was 25 degrees. Clinical ROM was better than radiographic ROM with average degrees of flexion 37 extension 15, left rotation 43, right rotation 44, left side bend 28, and right side bend 28 degrees. Eighteen patients had dominant neck pain in addition to neurologic changes pre-operatively. Post-operatively, eleven patients still had dominant neck pain, despite improvements neurologically. Greater than 60% of patients did not feel their neck had less motion than before surgery. There were no fixation failures. There was an average of 6.6mm of spinal canal opening as measured from plain radiographs, with 1mm or less of closure at follow-up. Most of the patients over age 65 were retired, and most of the remaining patients were working, some with jobs as strenuous as orthopedic and neurosurgery. The overall general health of the patients as determined by the SF36 was improved. CONCLUSIONS: Posterior cervical laminoplasty is a viable alternative for patients with small spinal canals and neurologic deficits. It appears to allow some preservation of cervical motion. It does not appear to help patients that have had chronic dominant neck pain. Paper # 29 Postoperative Management Without Immobilization of the Cervical Spine in Patients Undergoing Cervical Laminoplasty: Effects on Clinical and Radiological Outcomes Satoshi Asano, MD, Yutaka Nohara, MD, Tetsuro Kiya, MD, Tomohiro Takemoto, MD (Koshigaya, Saitama, Japan) INTRODUCTION: In the treatment of cervical myelopathy due to spondylosis and ossification of posterior longitudinal ligament (OPLL), laminoplasty is common and can obtain good neurological recovery. In general, the patient's cervical spine is immobilized with a collar for several months after this surgery. However, nobody knows how long the patient's cervical spine should be immobilized after cervical laminoplasty. Recently, some authors mentioned the relationship between postoperative neck pain and the duration of cervical immobilization. The objectives of this study were, 1) to compare the results of two groups with different postoperative management, and 2) to evaluate the necessity of collar fixation after cervical laminoplasty. METHODS: From April 1994 to June 1997, forty-two patients with cervical myelopathy were treated by spinous process-splitting laminoplasty using bioactive ceramic lamina spacers. These patients were prospectively and randomly divided into two groups with different postoperative management. In group A, there were 21 patients whose cervical spine was immobilized with a collar for two months after cervical laminoplasty. Their mean age at surgery was 58. Group B included 21 patients, too. However, they had no postoperative immobilization and they were encouraged to get ROM exercise of cervical spine and isometric paraspinal muscle exercise as early as possible. The mean age at surgery was 59 in this group. Diagnoses in group A were cervical spondylotic myelopathy (CSM) in 10 patients, and OPLL in 11. The average number of enlarged laminae was 6. In group B, there were 11 CSM and 10 OPLL. The average number of enlarged laminae was 5.3. At 2 years after the surgery, each patient was examined with Japanese Orthopaedic Association (JOA) score (full marks = 17 points) and its recovery rate. Neck pain was also evaluated. In radiological examination, cervical lordosis and ROM of the cervical spine were measured periodically after the surgery. RESULTS: The postoperative follow-up period ranged from 24 to 60 months (mean, 40 months) in group A, and from 24 to 48 months (mean, 31 months) in group B. The mean hospital stay were 27.1 days in group A and 21.9 days in group B, respectively. In group A, mean preoperative JOA score was 7.3, and it increased to 14.0 points after the surgery. In group B, mean JOA score was 9.0 before the surgery and 14.8 after the surgery. It was no statistically significant difference in the recovery rate of JOA score between two groups. At the followup, severe neck pain was seen in 4 patients of group A. However, no patient complained with severe neck pain in group B. In the patients of group A, the decrease of cervical lordosis was greater than that of group B patient. Statistically significant difference was observed in the decrease of cervical lordosis between the two groups immediately after the surgery and at 3 months after the surgery. Moreover, at two years after the surgery, both the decrease of cervical ROM and the decreasing rate of cervical ROM in group A were significantly greater than those of group B patient. DISCUSSION/CONCLUSION: In general, cervical laminoplasty can obtain satisfactory neurological recovery. However, as for disadvantages of this surgery, it had been reported that cervical ROM usually decreased and several patients complained with neck pain for a long time. In order to prevent these problems, it may be necessary to minimize the damage of the facet joint and muscle, and to mobilize the cervical spine as early as possible. Recently, a few authors reported postoperative management with shorter duration of immobilization after cervical laminoplasty. However, there has been no report about postoperative management without any immobilization after the surgery. In this study, the patients without postoperative immobilization obtained good neurological recovery and had no postoperative neck pain. Moreover, their ROM of the cervical spine was significantly greater than that of patients with postoperative collar fixation. The hospital stay of patients without postoperative immobilization was shorter than that of patients with postoperative immobilization. Therefore, in conclusion, postoperative immobilization of the cervical spine is not necessary after cervical laminoplasty, and early ROM exercise of the cervical spine and isometric neck muscle exercise can lead to better clinical and radiological outcomes. Paper # 30 Impact of Longitudinal Distance of the Cervical Spine on the Results of Expansive Open-door Laminoplasty Kazuhiro Chiba, MD, Masahiko Watanabe, MD, Hirofumi Maruiwa, MD, Morio Matsumoto, MD, Yoshikazu Fujimura, MD, Kiyoshi Hirabayashi, MD, Yoshiaki Toyama, MD (Tokyo, Japan) INTRODUCTION: Cervical kyphosis is considered to be one of the deteriorating factors of expansive open-door laminoplasty (ELAP). However, we experienced a number of patients with postoperative cervical kyphosis who had favorable results. Preliminary analysis on this patient population revealed that patients with cervical spondylotic myelopathy (CSM) tended to have better clinical results than those with ossification of the posterior longitudinal ligament (OPLL). We hypothesized that the shortening in the longitudinal distance of the cervical spine caused by multiple disc space narrowing may have a certain impact on the postoperative results by inducing so called redundancy in the spinal cord. Purpose of the present study is to test this hypothesis by analyzing the postoperative results of patients with postoperative malalignment who underwent ELAP with respect to the longitudinal distance of the cervical spine. MATERIAL AND METHODS: The study group comprised 70 patients, 56 males and 14 females who underwent ELAP and whose postoperative cervical alignment was judged as non-lordotic (kyphosis, straight and sigmoid curves) using modified Ishihara's classification on lateral X-ray films at the final followup. Their average age at the time of surgery was 58 years. There were 38 patients with CSM and 32 patients with OPLL. All patients were followed for at least 2 years postoperatively with an average of 3.7 years. Postoperative results were assessed using Japanese Orthopaedic Association scoring system for the treatment of cervical myelopathy (JOA score). Length of vertical line drawn between the postero-inferior edge of the C2 body and postero-superior edge of the C7 body was divided by antero-posterior length of the C4 body to give "Longitudinal distance index (LDI)". The effects of LDI on JOA scores were analyzed statistically using either unpaired t-test or Mann-Whitney U-test. RESULTS: Average recovery rate calculated using pre and postoperative JOA scores was higher in patients with CSM than those with OPLL (61.9% vs 51.8%; p=0.19) in this population although the difference was not statistically significant. However, JOA score at the final follow-up was significantly higher in the CSM group than in the OPLL group (14.3 vs 13.1; p=0.02). LDI was significantly smaller in the CSM than in the OPLL group (4.9 vs 5.1; p=0.02). When we divide whole patients into to two groups by drawing a line at LDI = 5.0, patients whose LDI was below 5.0 had significantly higher recovery rates than those over 5.0 (47.8 vs 67.3%; p=0.02). DISCUSSION: The present study demonstrated for the first time that the longitudinal distance of the cervical spine may have a significant impact on the surgical results in patients with postoperative malalignment. In patients with CSM whose X-ray films demonstrate multiple disc space narrowing, the redundancy induced in the spinal cord may dissipate the focal compression on the cord even in the cases of postoperative kyphosis. On the other hand, patients with OPLL is less likely to have such disc space narrowing due to ossification maintaining the vertical tensions of the spinal cord giving more static pressure on the spinal cord when the alignment is non-lordotic. Cervical kyphosis may not always be a contraindication of ELAP for patients with CSM and further study is ongoing to test the feasibility of using LDI as a determining factor for selecting appropriate surgical approaches in patients with preoperative kyphosis. Paper # 31 Preoperative Cervical Instability Does Not Affect Clinical Outcomes In Patients With Cervical Spondylotic Myelopathy Treated With Laminoplasty Mamoru Kawakami, Tetsuya Tamaki, Munehito Yoshida, Muneharu Ando (Wakayama City, Wakayama, Japan) INTRODUCTION: Cervical spinal fusion has been used to treat instability associated with cervical spondylotic myelopathy (CSM). Although there are many reports concerning the results of laminoplasty for patients with CSM, it is still unclear whether cervical instability influences the clinical outcome of laminoplasty. In addition, there are some patients with CSM in whom magnitude and direction of instability on standing and lying differ. However, there is no report concerning differences on instability on standing and lying positions and the direction of instability in patients treated with laminoplasty. The purposes of this study were to observe preoperative instability in the standing and lying positions in patients with CSM treated with laminoplasty, to evaluate the clinical significance of instability, and finally to determine if instability influences the clinical outcome of CSM treated with laminoplasty. METHODS: From June 1992 to March 1997, sixty-seven patients undergoing laminoplasty without spinal fusion at our institute for cervical myelopathy due to cervical spondylosis were considered for inclusion in this trail. There were 55 men and 12 women, ranging in age 35-85 years, with a mean age of 63.0. The mean period of symptoms of myelopathy prior to surgery was 2 years, and the mean follow-up period was 3.5 years, with a range of 2-6.5 years. Axial symptoms, neurological function, the range of motion of the cervical spine (ROM), and cervical sagittal alignment and instability were evaluated. For neurological evaluation, recovery rates were calculated at follow-up examination using Hirabayashi’s method with the criteria proposed by the Japanese Orthopaedic Association (JOA score). ROMs were measured from C2 to C7 by Cobb’s method. Instability was considered present if more than 3 mm of slippage was observed on dynamic X-ray studies. In addition, the facet joint angles and the heights of the intervertebral discs were also measured. Patients were divided into three groups. In Group A, patients had no instability while either standing or lying. Patients with instability in the standing position comprised Group B. In Group C, instability was increased or appeared in the lying position. In addition, based on the direction of instability, patients were divided into the anterior and posterior instability groups. Differences between groups in these measurements were tested by _2 analysis and Student’s t-test. The minimum level of significance was p < 0.05 for two-tailed tests. RESULTS: The mean ages at surgery were 57.7, 64.5, and 67.5 years in Groups A (n=24), B (n=21) and C (n=22), respectively. The age of patients with instability was higher than that of those without instability, and the age in Group C was significantly older than that in Group A (p<0.05). The mean periods of symptoms prior to surgery were 3.0, 1.4, and 1.5 years in Groups A, B, and C, respectively. There was a tendency for this period to be longer in Group A than in other groups (p=0.06). There were no significant differences in gender, prevalence of axial symptoms, or JOA scores among Groups A, B, and C. Mean recovery rates were 64.5, 65.9, and 60.2% in Groups A, B, and C, respectively (p>0.05). There were no differences in cervical sagittal alignment at preoperative or follow-up examination among the three groups, and no instability was observed in any patients at follow-up. ROM at follow-up was 43.5% of preoperative ROM (p<0.05), and limitation of flexion in the cervical spine was significant in all groups. However, there were no significant differences in the ROMs, facet joint angles of heights of intervertebral discs among the three groups. Preoperative JOA score in the anterior instability group (n=9) was significantly lower than that in the posterior instability group (n=34) (5.8 vs. 9.0 points, p,0.05). Preoperative lordosis in the anterior instability group was less than that in the posterior instability (8.8º vs. 19.2º, p<0.05). DISCUSSION/CONCLUSION: Patients with instability were old, but did not differ in bone and joint morphology such as cervical sagittal alignment, facet joint angle, and disc height loss from those without instability. It is thus possible that the mechanism of instability is related to dysfunction of the soft tissues around the cervical spine that develops with age. The period symptoms prior to surgery was long in patients without instability, suggesting that cervical instability may rapidly induce myelopathy due to spondylosis. Patients in the anterior instability group had low preoperative JOA scores and decreased lordosis, compared with those in the posterior instability group. Anterior instability with decreased lordosis may produce local kyphosis of the spinal cord at motion segments and may cause severe neurological deficits preoperatively. Preoperative instability influenced neither axial symptoms nor neurological recovery. This may result from posterior decompression of the spinal cord, maintenance of the cervical spine, and reduction of ROM after laminoplasty. Cervical instability is a preoperative factor which does not influence clinical outcome and can be neglected if laminoplasty is indicated for patients with CSM. Paper # 32 Morphological Evaluation of the Extensor Musculature of the Cervical Spine By Means of Coronal View of MRI Haku Iizuka, MD, Takachika Shimizu, MD, Hideo Edakuni, MD, Keisuke Hueki, MD (Gunma, Japan) INTRODUCTION: Laminoplasty is widely accepted as a treatment of multisegmental cervical myelopathy. Posterior procedure, however, involves the extensor musculature of the cervical spine. We have been observing the cervical extensor musculature following surgery by means of coronal view of magnetic resonance imaging (MRI). The purpose of this study is to clarify the relationship between post-operative cervical alignment and morphological evaluation of the cervical extensor musculature, especially of “semispinalis cervics”, by means of coronal view of MRI in patients undergoing laminoplasty. MATERIALS: From 1988, in posterior cervical spine surgery we have attempted to reapproximate the extensor musculature inserting C2 spinous process, after decompression in the neck position in extension. Twenty patients with cervical myelopathy who were treated by this procedure were examined and followed up for an average of 12 months (R group). Also 6 patients who underwent laminoplasty at another hospital, without anatomical repair of extensors, were examined (NR group). METHODS: In R group, semispinalis cervics were evaluated morphologically by means of MRI coronal view 1 month, 6 months and 1 year post operatively. Cervical alignment at one year post operatively was compared with the preoperative alignment using a lateral view of cervical radiographs. We mostly imaged the coronal view of MRI along the plane containing the transverse process of the upper thoracic spine and C2 spinous process so as to define the shape of the semispinalis cervics clearly. The curvature of the cervical spine was determined by measuring the angle formed by two lines extending from the inferior border of C2 vertebral body and the superior border of C7 vertebral body on the lateral radiograph in the neutral position. RESULTS: In R group, we found morphological repair of semispinalis cervics in all cases on coronal view of MRI (Fig.) Pre and postoperative cervical angulation was 13 degrees and 16 degrees on average, respectively. Good cervical alignment was obtained and maintained postoperatively. In NR group, we found caudal retraction of semispinalis cervics in all cases on this MRI examination (Fig.) Pre and postoperative angulation was 14 degrees and –6 degrees on average, respectively. The average loss of angulation was 20 degrees. DISCUSSION/CONCLUSION: Laminoplasty is a useful treatment for cervical myelopathy, but posterior procedures usually involve stripping of the extensor musculature and their insertions on the spinous processes. Especially, semispinalis cervics inserting C2 spinous process is easy to be disrupted. In R group, we could recognize the morphological repair of semispinalis cervics following surgery and in fact, the postoperative cervical alignment was good. In contrast, in NR group, caudal retraction of semispinalis cervics was revealed by MRI, and the cervical lordosis was lost post operatively. In conclusion, this MRI investigation suggests that semispinalis cervics plays an important role in maintaining the cervical alignment. Paper # 33 The Natural History of Destructive Spondyloarthropathy of the Cervical Spine In Long-Term Hemodialysis Patients Akihiro Nagamachi, MD, Toru Endo, MD, Mamoru Hirohata, MD, Shinji Komatubara, MD (Kagawa, Japan) INTRODUCTION: Incidence of destructive spondyloarthropathy (DSA) of the cervical spine increases the number of hemodialysis patients increases and their life span prolongs. Patients with DSA are suffering from severe neck pain and sometimes have neurologic deficit. However, specific management for these patients is not established. Although it is important to know the natural history of DSA for the treatment of the patients, there are few previous longitudinal studies. The purpose of this longitudinal study is to elucidate the natural history of DSA and the factors that associate with the destructive changes in patients with long-term hemodialysis. SUBJECT AND METHODS: Between 1989 and 1998, consecutive 42 patients (24 men and 18 women) being managed with hemodialysis were followed by plain radiograph of the cervical spine. The mean age was 58.1 years (range 38 to 80), and the mean duration of hemodialysis was 15.3 years (range 9 to 25) as of 1998. Gradings of radiological feature from lateral view were as follows; grade 0: normal, grade 1: erosive changes in anterior rim of the vertebrae and endplate, grade 2: disc space narrowing with minimal osteophyte formation, grade 3: spinal fusion and/or subluxation of the vertebrae. All patients were graded each year. The relationship between the duration of hemodialysis and the grade, and between the age at which hemodialysis started and the grades were evaluated. The radiological changes of each disc level were noted. Statistical analysis was performed using Chi-square test, ANOVA and Fisher’s PLSD for post hoc test. RESULTS: In 1989, nine patients (21.4%) were classified into grade 0, thirty patients (71.4%) were grade 1, three patients (7.1%) were grade 2 and no patient was found in grade 3. In 1998, only one patient (2.4%) was classified into grade 0, twenty-six patients (61.9%) were grade 1, seven patients (16.7%) were grade 2 and eight patients (19.0%) were grade 3. The mean duration of hemodialysis of grade 1, grade 2, grade 3 were 39.2 years, 48.2 years and 49.6, respectively, and the mean age at which hemodialysis started of grade 2 and grade 3 were significantly older than that of grade 1. Radiological changes of grade 2 and grade 3 were most commonly observed in lower cervical spine (fig 1). The average numbers of the involved disc level were 1.6 in grade 2 and 1.1 in grade 3 in 1998. Of the eight patients in grade 3, three patients had grade 2 changes in an adjacent disc level. DISCUSSION: In the present study, erosive changes (grade 1) of the cervical spine have progressed in almost all patients in nine years. However, the patients with destructive changes (grade 2 and/or grade 3) were fifteen (35.7%) in 1998. The mean age at which hemodialysis started of these patients were older than that of grade 1 patients. Although the cause of DSA is thought to be the deposition of advanced glycation end products (AGEs) ß2 – microglobulin deposition is the main cause of DSA, destructive changes should have occurred in larger numbers of the patients. Therefore, degenerative changes of the spine due to aging is thought to be on of the most important factors of DSA. DSA developed commonly in lower cervical spine of which the range of spinal motion was large, and these destructive changes were limited in one or sometimes two disc levels. It was though that once the destructive changes occurred, axial and shearing loads concentrated to the involved level and the destruction progressed in limited disc levels. The mechanical stress to the spine is associated with the pathogenesis of DSA. From these point of view, anterior body fusion of the disc of grade 3 should be avoided and posterior decompression and long fusion of the cervical spine should be considered when adjacent disc presented grade 2 destructive changes. CONCLUSIONS: Of the 42 patients managed with fifteen years of hemodialysis in average, destructive changes (grade 2 and/or grade 3) were found in 15 patients (35.7%). Degenerative changes of the spine due to aging and the mechanical stress to the spine were thought to be one of the most important factors in DSA. From the natural history, anterior body fusion of the disc of grade 3 should be avoided and posterior decompression and long fusion of the cervical spine should be considered when adjacent disc presented grade 2 destructive changes. Figure 1: Radiological changes of each disc level in 1989 and 1998 Grade 2 and grade 3 were most commonly observed in lower cervical spine. Paper # 34 Surgical Treatment of Cervical Spinal Disorders Associated with LongTerm Hemodialysis Kuniyoshi Abumi, MD, Manabu Ito, MD, Kiyoshi Kaneda, MD (Sapporo, Japan) INTRODUCTION: The number of long-term hemodialysis survivors has been increasing with the development of medical management. Recent clinical and pathological studies have revealed that patients on long-term hemodialysis show various changes in their bone and joints including destructive spondyloarthropathy (DSA). The cervical spine is the most commonly involved spinal region. The purposes of this report are to review the surgical results of 15 patients with cervical spinal disorders associated with long-term hemodialysis, to investigate the pathological features by histological examination, and to propose the optimum surgical procedure for this disorder. METHODS: Between October 1991 and June 1997, 15 patients underwent surgical treatment for cervical spinal disorders related to hemodialysis at our affiliated hospitals. The patient population consisted of 11 men and 4 women with an average of 56.6 years (range 44-72 years). Duration of hemodialysis ranged from 8 to 27 years with an average of 17.3 years. The compromised spinal levels were limited in the middle and lower cervical spine below C3/4 in 13 patients, and expanded caudad from C1/2 in two. Thirteen patients showed marked cervical myelopathy preoperatively and the remaining two patients showed intolerable radiating pain due to cervical radiculopathy. Eight of 13 patients with myelopathy were unable to walk. For surgical treatment, nine patients with marked destructive changes underwent circumferential reconstructive surgery which consisted of posterior fixation using the pedicle screw fixation, anterior strut bone grafting, and posterior and/or anterior decompression. The number of the fixed spinal segments in the nine patients ranged from one to five (average 2.8). For longitudinal members connecting the pedicle screws, plates were used in 14 patients and rods in one patient. Secure fixation of the plate or rod to the pedicle screw was obtained by a constrained connection mechanism. Two patients with radiculopathy underwent posterior nerve root decompression by foraminotomy and fusion by using pedicle screw or spinous process wiring. Remaining four patients without marked instability underwent posterior spinal cord decompression by laminoplasty or laminectomy alone. Pre and postoperative neurologic status was assessed using a cervical myelopathy scoring method of Japanese Orthopaedic Association reducing 3 points of the bladder function (full score: 14 points). Histological examination was conducted by light and scanning electron microscopy to determine the distribution pattern of amyloid deposits in the spinal components resected during surgery. RESULTS: Average follow-up period in the surviving 13 patients was 53 months (range 25- 92 mos.). One patient who underwent laminoplasty died for sepsis at six months postoperatively. One patient who underwent circumferential reconstruction died after cardiac surgery at 46 months postoperatively. Marked neurologic recovery was obtained in all patients after surgery. Preoperative average JOA score of 6.1 (range: 2–11) improved to 11.0 (range: 6–14) at the final follow up. There were no patients with neurovascular complications directly attributable to pedicle screw fixation. Bony union was obtained in all patients except one who underwent posterior fusion by spinous process wiring. Progressive destructive changes with significant instability at the adjacent mobile segments were observed in two patients who underwent circumferential fusion. These two patients required extension of spinal fusion using pedicle screw/rod fixation system at 26 and 31 months after the initial surgery. Histology showed intensive amyloid depositions at the facet joints, intervertebral discs, and other soft tissues. Affected vertebral bodies were also replaced by amyloid granulation and showed enhanced activity of osteoclast. DISCUSSION/CONCLUSION: Pathological conditions of the cervical spine affected by long-term hemodialysis varies extensively according to the duration of the hemodialysis, predisposing degenerative changes of the spinal segment, sites and extent of amyloid deposition. Surgical procedures must be selected according to the pathological conditions. For the patients with myelopathy or radiculopathy caused by spinal canal stenosis due to amyloid deposition at the capsular fibers, the ligamentum fravum and others without marked segmental instability, posterior decompression alone is the sufficient surgical intervention. However, the patients with marked destructive changes with severe instability require reconstructive surgery. Since destructive changes expand to the threejoint complex of the spinal column in such condition, circumferential spinal fusion which consist of posterior fusion and decompression and replacement of the destructed vertebral bodies and discs is the recommended surgical procedure. Conventional internal fixation procedures such as wiring do not provide sufficient stability for this complicated condition with significant instability. The pedicle of the cervical spine is a strong structural element of the vertebrae, as in the thoraco-lumbar spine. By the results of this series, pedicle screw fixation provides sufficient initial stability to the reconstructed spinal construct and excellent fusion rate. Particularly because this procedure does not require the lamina as the stabilizing anchor, it is advantageous for simultaneous posterior decompression and fusion. We conclude that the pedicle screw fixation is the optimum internal fixation procedure for reconstruction of the cervical spine with marked destructive changes in long-term hemodialysis patients. Paper # 35 Anomalous Vertebral Artery: A Cadaveric and Clinical Case Study Lukasz J. Curylo, MD, Harold Mason, MS, Henry Bohlman, MD, Jung Yoo (Cleveland, OH) INTRODUCTION: Anterior cervical decompression by subtotal corpectomy combined with fusion is a well-accepted treatment for cervical spondylotic myelopathy or radioculopathy. Injury to the vertebral artery during the phase of decompression is a rare complication with potentially catastrophic consequences. Well-accepted surgical intraoperative landmarks are often used to define the lateral extent of corpectomy. However, the presence of an abnormally tortuous vertebral artery with migration towards midline has been well documented in the literature. Lack of recognition preoperatively of this anomaly can lead to laceration of the vessel even by decompression within generally accepted safe limits. The incidence of this anomaly within the general population as well as its characteristics are not known. MATERIALS AND METHODS: 3 patients with an ectopic vertebral artery undergoing anterior cervical corpectomy within the last 2 years were identified at our institution. This prompted us to perform a study on 123 consecutive cadaveric adult subaxial cervical specimens of C3-C6 (total 492 vertebral levels) from the Todd-Hamman Collection at the Cleveland Museum of Natural History. The sex and age of each specimen was recorded. A digital high precision caliper with customized tips was used to measure the distances between the uncovertebral joints (UVJ) at each level. Also the distance between the UVJ and the medial aspect of the vertebral foramen on each side was measured. Next an additional 99 (giving a total of 222) adult specimens (total 888 vertebral levels) were subjected to visual inspection only. From the entire group specimens with frank ectasia of the vertebral artery into the vertebral body were selected. The selected specimens were evaluated by axial CT scanning. A radiographic marker at the level of the vertebral artery marked the location of the UVJ. The distance from the UVJ (visualized by a marker) to the most medial aspect of the vertebral artery foramen was then measured directly off the axial CT scan. The distance (recorded in mm) was given a negative value if the medial aspect of this foramen was medial to the radiographic marker, and a positive value if it was lateral. RESULTS: 7 vertebrae (6 out of 222 cadaveric specimens or 2.7%) were identified as having an ectopic vertebral artery. There were three abnormal C3 vertebrae, three C4 vertebrae and one anomalous C6 (one cadaver had an abnormal C3 and C6). All anomalies were unilateral. The mean distance from the UVJ to medial aspect of vertebral foramen as measured by CT in the abnormal vertebrae was -0.14mm +/- 1.19 (range: -2.0 to +1.5). The mean UVJ to vertebral foramen distance was on the right +5.4mm +/-2.7 (range: -0.75 to +14.0) and on the left +5.7mm +/-2.2 (range: -2.0 to +14.9). This was not statistically significant (paired T-test). The mean inter-UVJ distance was 26.8mm +/-2.6 (range: 20.2 to 35.8). The mean UVJ to vertebral foramen distance for each level were as follows - C3: +5.3mm, C4: +5.1mm, C5: +5.3mm, C6: +5.15mm. The differences were not statistically significant (ANOVA p=0.73). The first case of ectasia was identified intraoperatively when a laceration of the vertebral artery occurred during the initial stage of a motorized corpectomy. The laceration was repaired and the 3 level anterior cervical corpectomy and fusion (ACCF) was completed in a standard fashion. The patient healed without neurologic sequelae. The next two cases were identified preoperatively. Both patients instead of undergoing a standard 3 level ACCF were managed by a 2 level ACCF combined with a wide cervical discectomy and fusion (ACDF) adjacent to the anomalous segment. The intraoperative course, as well as postoperative healing was uneventful in both cases. DISCUSSION: Tortuosity of the vertebral artery is a well-recognized phenomenon. Lack of preoperative recognition of this problem can result in serious complications. The UVJ or joint of Luschka provides usually a safe margin for the lateral extent of decompression with an average of 5 mm of bone remaining to the vertebral foramen. This 5mm average distance between the medial aspect of UVJ and the medial aspect of the vertebral foramen in our study is in concordance with the literature. However, in 2.7% of patients in our study the vertebral artery was located medial to UVJ or less than 1.5mm lateral to UVJ. The UVJ is usually located superior to the ectopic vertebral foramen. Therefore the two structures are not visible on a single axial CT cut and their relation can be difficult to assess preoperatively.Therefore the UVJ is even less reliable during preoperative planning. In the face of an abnormally tortuous vertebral artery, with ectasia into the vertebral body, other levels of the cervical spine must be carefully inspected since this anomaly can be present at multiple levels (1/6 anomalous cases in our study) with special attention to C3 and C4, since most anomalies were located at these levels. The treatment protocol must be appropriately altered. In 2 of our patients the cervical pathology was successfully managed by anterior decompression by combining a wide ACDF at the abnormal level with standard decompression at other normal levels, but a posterior decompression with laminoplasty has been reportedly used with success in the literature. CONCLUSIONS: The incidence of abnormally ectopic vertebral artery in the general population is low but significant. Most anomalies are located in the upper cervical subaxial vertebrae. Even though rare, vertebral art. ectasia can occur at multiple levels simultaneously. In patients with ectasia the UVJ is not a safe anatomic landmark of the lateral extent of the decompression. Preoperative recognition of the anomaly is key to avoiding vascular complications. Treatment by wide discectomy and fusion above and below the anomalous vertebrae, instead of a corpectomy, and combining this with a standard ACCF can give an excellent clinical result without risking significant complications. REFERENCES: 1. Pait T., et al Neurosurgery 39(4); 1996: pp 796-776 2. Oh S., et al Neurosurgery 38(6); 1996: pp 1139-1144 3. Ebraheim N., et al Surgical Radiological Anatomy Vol. 20; 1998: pp 389392 4. Heary R., et al Spine 21(18); 1996: pp 2074-2080 5. Oga M, et al Spine 21(9); 1996: pp 1085-1089 6. Smith M, et al JBJS 75-B(3); 1993: pp 411-415 Paper # 36 Anterior Spine Fusion Using Minimally Invasive Gene Therapy Technique Dan Riew, MD, Jueren Lou, MD, Neill M. Wright, MD, Su-Li Cheng, PhD, Louis Avioli, MD (St. Louis, MO) • (a - Washington University Institute for Minimally Invasive Surgery and Ethicon Endosurgery) INTRODUCTION: We previously demonstrated posterior spinal fusion in rabbits using adenovirus-mediated BMP-2 gene transfer into mesenchymal stem cells (MSC). The purpose of the present study is to determine if this genetherapy technique can be utilized to achieve anterior intradiscal fusion in pigs using minimally invasive techniques. METHODS: Rib marrow derived mesenchymal stem cells (MSC) were isolated from each pig and expanded in culture in vitro. Adenoviruses carrying the gene for BMP-2 (Adv-BMP2) or b-galactosidase (Adv-bgal) were produced in our lab as previously reported. The MSC were transduced with either Adv-BMP2 or the control virus Adv-bgal at 50pfu/per cell. In Vitro Study: BMP-2 protein expression in conditioned media was tested by immunoprecipitation and Western blot with a specific monoclonal antibody h4b2/5.10.24. Alkaline phosphatase activity and matrix mineralization (Von Kossa’s stain) were determined. Protein expression of Type I collagen, osteopontin, and bone sialoprotein were analyzed by Western blot. In Vivo Study: Three pigs underwent general anesthesia and thoracoscopic visualization of the anterior thoracic spine was obtained. Three thoracic disc spaces in each of 3 pigs were punctured and their endplates disrupted with a small curette under thoracoscopic visualization. Each of these 3 disc spaces were separated by one intervening disc space that was left untouched. The discs were then injected with autologous MSC transduced with either Adv-BMP2 or Adv-bgal (control 1) or nontransduced MSC (control 2). Pigs were sacrificed 6 weeks post implantation. CT scans with 2-dimensional reconstructed images were obtained of all specimens and blindly interpreted by a radiologist for the presence of fusion. Histologic examination was performed of all specimens in a blinded fashion. RESULTS: In Vitro Study: A human BMP-2 protein band at molecular weight 20 KD was detected by specific antibody in conditioned medium of Adv-BMP2 transduced pig MSC. No such band was seen in conditioned media of Adv-bgal transduced MSC or untreated MSC. Compared to Adv-bgal transduced or untreated MSC, Alkaline phosphatase activity of Adv-BMP2 transduced MSC increased 5 fold and matrix mineralization was induced. Consistent with the expression of functional BMP-2 protein, Adv-BMP2 transduced MSC increased protein expression of Type I collagen, osteopontin, and bone sialoprotein. In Vivo Study: Anterior spine fusion was demonstrated by radiographic examination in all discs implanted with Adv-BMP2 transduced MSC. The control discs implanted with either Adv-bgal or untreated MSC had little or no intervening bone. Histologic examination demonstrated bridging bone from endplate to endplate for all discs implanted with Adv-BMP2 transduced MSC. The control discs implanted with either Adv-bgal or untreated MSC had no bridging bone. DISCUSSION/CONCLUSION: We demonstrated that recombinant adenoviral vector mediated BMP-2 gene transfer into pig MSC induces the cell to produce BMP-2 protein. The transduced MSC differentiated into osteoblasts in vitro, demonstrated by marked increase in Alkaline phosphatase activity and characteristic matrix mineralization. Furthermore, autologously implanted AdvBMP2 transduced MSC into the disk space induced anterior spine fusion. Our results suggest that it is possible to induce anterior spinal fusion using a minimally invasive recombinant DNA technique. • If noted, the author indicates something of value received. The codes are identified as: a - research or institutional support, b - miscellaneous funding, c - royalties, d - stock options, e - consultant or employee. For full information, refer to inside back cover. Paper # 37 Spinal Cord Regeneration in Infant Rat - Morphological, Electrophysiological and Molecular Biological StudiesYoshiaki Wakabayashi, MD, Toshiyuki Kawa-uchi, MD, Kenichi Shinomiya, MD (Tokyo, Japan) INTRODUCTION: Spinal cord injury (SCI) has been believed to be incurable, but recent studies have suggested that regeneration and functional recovery can be obtained after SCI with specific conditions such as peripheral nerve transplantation with fibrin glue containing aFGF (Cheng H. et al., Science 273:510-513, 1996). Prior to these reports, it is widely accepted that an infant rat recovers its hind limb function after spinal cord transection even without any specific treatments. However, mechanism of this recovery is still controversial, because trained spinal cord reflex may bring about functional restoration as well as axonal regeneration. In order to challenge to repair complete spinal cord injury, this mechanism including required conditions for regeneration must be revealed. The purpose of this study is 1) to show functional recovery from SCI in infancy is not only due to trained spinal cord reflex, but axonal regeneration of some descending pathway(s), 2) to evaluate which descending tract(s) are reconnected after SCI, and 3) to analyze whether or not mRNA expressions of endogenous aFGF and bFGF in infant rat are enhanced after SCI. METHODS: 1) Laminectomy and transection of spinal cord were performed at Th10 in 27 infant rats (two weeks of age). As sham operation, only laminectomy was done in 17 rats. Six weeks after operation, muscle-evoked potentials (MEPs) of hind limbs by transcranial electrical stimulation to the brain were recorded. Re-transection of spinal cord was performed in some recovered rats, and then MEP and locomotion were observed. 2) To determine possible pathways involved in the functional recovery, we conducted tract-tracing studies using Fast Blue as a retrograde tracer and WGAHRP as an anterograde tracer, following the electrophysiological study. 3) Infant (11 days after birth) and adult (10 weeks of age) rats underwent spinal cord transection or sham operation as described above. Twelve or twenty-four hours after surgery, total RNAs were extracted with TRIZOL reagent from distal end of the transected spinal cord. cDNAs of aFGF, bFGF and beta-actin were amplified by RT-PCR with specific primers. After electrophoresis, amount of each product was determined by computerized analysis, and normalized with beta-actin. Product ratios of SCI to sham were examined. RESULTS: 1) After transection, 13 rats recovered to walk, 8 rats were not able to walk, and the other 6 rats were died. Fourteen animals survived after sham operation without any dysfunction. MEPs were recorded in sham operated and recovered rats, but not in unrecovered rats. Moreover, MEPs of recovered rats disappeared after spinal cord re-transection that caused complete paralysis. 2) Histological studies showed continuous re-connection of rubrospinal, vestibulospinal and reticulospinal tracts, while regeneration of corticospinal tract was not observed. 3) Twelve hours after transection, mRNA expressions of aFGF and bFGF in adult rat spinal cord were equivalent to sham, however the expressions were decreased after 24 hours. mRNA levels of FGFs in infant rat after transection showed a similar tendency. DISCUSSION/CONCLUSION: Electrophysiological evaluation suggested regeneration of some descending pathways might play a pivotal role in functional recovery from SCI in infancy as well as trained spinal cord reflex. Histological study revealed this functional recovery is attributed to axonal regeneration of rubrospinal, vestibulospinal and reticulospinal tracts, but not corticospinal tract. Neurotrophic factors such as FGFs are considered to be important for nerve regeneration, however expression patterns of endogenous FGFs in infant rat after transection showed no significant difference from adult rat within 24 hours after operation. We conclude that regenerative ability of infant rat spinal cord is not attributed to activation of endogenous FGFs, at least within 24 hours after SCI. Paper # 38 Reduction in the Metabolic Rate of Schwann Cells Induced by Intervertebral Disc Cells Nahshon Rand, MD (Nashville, TN), Saul F. Juliao, BS (Nashville, TN), John M. Dawson (Nashville, TN), Yizhar Floman (Jerusalem, Israel), Dan M. Spengler (Nashville, TN) • (a - Cervical Spine Research Society) BACKGROUND: The etiology and pathophysiology of radiculopathy in disc herniation is not fully understood. It is not fully explained by mechanical compression. Inflammation and its mediators are believed to play a role in radiculopathy, through a yet unknown mechanism. Disc cells and their secreted products may affect the adjacent Schwann cell (SC) lining of the spinal nerve roots, leading eventually to clinical signs of radiculopathy. There are however, very little data addressing such interaction. The objective of this study was to describe the effects of compounds secreted by disc cells on the metabolic rate (MR) of cultured SC. METHODS: Intact lumbar and coccygeal intervertebral discs and sciatic nerves were harvested from inbred mice. Discs were separated to annulus fibrosus (AF) and nucleus pulposus (NP) and cultured for 5 days. Sciatic nerves were enzymatically digested. SC were isolated and cultured in medium (50,000 cells/well) in a total volume of 200 microliters. The baseline MR of cultured SC was studied using radiolabeled thymidine incorporation assay (RLTI). The effects of administering cultured medium of AF and NP to SC culture on the metabolic rate of the cultured SC were assayed using RLTI. Conditioned media of AF or NP (0, 25, 50, 100, and 175 microliters in 96 well plates (corning NY)) were added to a SC culture for 48 hours. During the last 18 hours, tritiated thymidine (0.5 mCi/ well) was added and the amount of thymidine incorporation was measured on a beta plate liquid scintillation counter (Wallac, Turku, Findland). Results were calculated as percent of the baseline MR. RESULTS: The table and figures 1 and 2 describe the results. There was a considerable reduction in MR in response to both NP and AF cultured medium. The reduction was directly related to the amount of cultured medium added. Table. Metabolic rate of Schwann cells after exposure to different volumes of disc cell cultured medium (expressed as % of baseline metabolic rate) Tissue Type Nucleus Pulposus Annulus Fibrosus Volume of Added Medium Sample # Mean Value 1 2 3 4 5 0 100 100 100 100 100 100 25 68 91 100 75 65 79.8 50 56 77 91.5 62 53 67.9 100 38 65 75 40 35 50.6 175 14 31 30 25 18 23.6 0 100 100 100 100 100 100 25 52 60 56 59 75 60.4 50 52 41 46 38 87 52.8 100 20 24.7 29 35 84 35.84 175 20 16.2 19 15 49 23.84 DISCUSSION: Recent findings attribute a role to disc cells in the pathophysiology of inflammation in disc herniation. Disc cells and the inflammation found in disc herniation, are likely to be involved in the pathophysiology of radiculopathy. Only little data is available concerning the interactions of disc cells and their secreted products, and SC. The current study demonstrated a reduction in MR of SC in response to their exposure to cultured medium of disc cells. The reduction in SC MR may correlate with the clinical findings in radiculopathy. Other cellular functions of SC may also be impaired as a result of the effects of herniated disc cells and their secreted products. Such potential effects of disc cells and their products on different cellular functions of SC, and identifying active compounds in the disc cell cultured medium, require further research. CONCLUSION: Cultured disc cells secrete compounds which induce a reduction in MR of cultured SC. Further study of the specific compounds which induce the reduction, and the nature of the interaction between SC and disc cells is still required. • If noted, the author indicates something of value received. The codes are identified as: a - research or institutional support, b - miscellaneous funding, c - royalties, d - stock options, e – consultant or employee. For full information, refer to inside back cover. Paper # 39 Molecular Mechanisms Underlying the Process of Experimental Murine Spondylosis Takanobu Nakase, MD, PhD, Kenta Ariga, MD, Shimpei Miyamoto, MD, PhD, Motoki Iwasaki, MD, PhD, Eiji Wada, MD, PhD, Shinya Okuda, MD, Wenxiang Meng, MD, Kazuo Yonenobu, MD, PhD (Osaka, Japan) [INTRODUCTION: Little is known about the molecular mechanisms of spondylosis. Thus, the initial purpose of this study was to analyze the distribution of genes for bone and cartilage extracellular matrix (ECM) proteins using our previously established experimental model in mice. Additionally, although genetic expressions of these ECM genes have been reported to be regulated by several bone-related growth factors, little is known about the involvement of such growth factors in the process of spondylosis. Thus, the second purpose was to investigate the expression of several bone-related growth factors, such as bone morphogenetic protein (BMP)-4, 6, Indian hedgehog (Ihh) and PTHrP in the process of spondylosis using our mice model. In this study, an in situ hybridization technique was used to detect mRNAs for specific molecules on histological sections. METHODS: A total of 36 five-week-old ICR mice (male) were used. Of these, 18 underwent spinous process resection (sp-mouse) as described previously (Spine, 1991). Three sp-mice each were fully anesthetized and the cervical and thoracic spine were removed at 1, 2, 3, 4, 6 and 12 months post surgery (E1, 2, 3, 4, 6 and 12) together with three age-matched controls for each group (C1, 2, 3 ,4, 6 and 12). The tissue samples were fixed in 4% PFA, decalcified in EDTA, embedded in paraffin, and then prepared for histological sectioning. The sections were stained with Hematoxylin-eosin and safranin-O-fast green. In situ hybridization using DIG system was performed as described previously (J. Bone Miner. Res. ,1994) to evaluate genetic expression of ECM proteins and growth factors utilizing cRNA probes for collagen types II, IX, X, XI (Col I,II,IX,X,XI), aggrecan (Ag), osteopontin (OP), BMP-4, -6, Ihh and PTHrP. The number of positive cells in in situ hybridization was analized using an image analyzing computer system. RESULTS: Histological events were assessed as follows: Stage I: Metaplasia of fibroblastic cells in the annulus fibrosus into chondrocytes. (Mainly observed in E1 and E2). Stage II: Proliferation of chondrocytes. (Mainly observed in E3 and E4). Stage III: Endochondral ossification leading to apparent osteo-chondrophyte formation (Mainly observed in E6 and E12). In control groups, no obvious degenerative changes as described above were found. i i i Results of ECM: In stage I, mRNAs for early chondrogenic genes, such as Cols II, IX, XI and Ag were expressed by cells in the annulus fibrosus and in the attachment of the anterior spinal ligament. In stage II, the majority of proliferating chondrocytes expressed these genes. A few proliferating chondrocytes expressed mRNAs for Col X and OP. In stage III, the number of cells positive for early chondrogenic genes (Cols II, IX, XI and Ag) decreased, whereas the number of cells expressing marker genes for endochondral ossification, such as Col X and OP increased. Results of growth factors: In stage I, only mRNA for BMP-4 was expressed by cells in the annulus fibrosus and in the attachment of the anterior spinal ligament to disc. In stage II, the majority of proliferating chondrocytes expressed PTHrP. Several hypertrophic chondrocytes expressed mRNAs for Ihh and BMP-6. No BMP-4 signals were observed. In stage III, the number of cells positive for PTHrP mRNA decreased, whereas the number of cells expressing BMP-6 and Ihh increased. In the control groups, cells expressing the genes for ECM proteins and growth factors were scarcely found. DISCUSSION/CONCLUSIONS: A A ECM: i) The distribution of early chondrogenic genes in stage I suggested that formation of osteochondrophytes originates from both annulus fibrosus and attachment of the spinal ligament at the molecular level. ii) The findings also demonstrated that the pattern of genetic distribution of ECM proteins altered along with time-course histologic changes in murine experimental spondylosis. Growth factors : Genes for BMP-4, 6, Ihh and PTHrP were differentially distributed in the present murine spondylosis model. i) BMP-4 is known as a bone/cartilage inducing substance in vivo. Expression of BMP-4 mRNA in stage I in sp-mice suggested that BMP-4 gene was induced by mechanical load and initiated osteochondrophyte formation. ii) PTHrP has been reported to promote chondrocyte growth/differentiation in in vitro culture systems. The dominant expression of PTHrP in stage II suggested that PTHrP is involved in chondrocyte proliferation/differentiation in spondylosis. iii) Ihh-BMP6-PTHrP signalling is known to be important for embryogenic skeletogenesis. Expression of BMP-6, PTHrP and Ihh genes in stage III indicated that osteochondrophyte formation is regulated by Ihh-BMP6PTHrP signalling in spondylosis. Finally, co-localization of mRNAs for ECM proteins and growth factors indicated that site and stage specific regulation of the genetic expression of these ECM genes by several growth factors (BMP-4,6,Ihh and PTHrP) is involved in the process of spondylosis. Paper # 40 Degeneration of Intervertebral Disc by Smoking-Experimental Study in Rat Smoking Model Hiroshi Oda, Hiromi Matsuzaki, Yasuaki Tokuhashi, Hidehiko Hagiwara, Yosinao Uematsu, Masaki Iwahasi (Toyko, Japan) INTRODUCTION: Epidemiological surveys have confirmed that a high proportion of patients suffering from low back pain are smokers. In this study, we established a rat smoking model in which the animals were exposed to smoke passively in order to investigate the effects on the intervertebral disc histopathologically and immunohistochemically. METHODS: Twenty SD female rats weighing 200-300g were used. Using a newly developed smoking box which enabled intermittent passive smoking, smoke from a short piece of tobacco, was fed into the atmosphere for about 5 minutes hourly and then ventilated for a subsequent period of 5 minutes. A total of 20 cigarette were smoked daily at one-hour intervals. The intervertebral discs were extracted from the models with smoking for 1-,2-,4-, and 8-weeks. Control models were kept in similar smoking box for same periods and their discs were extracted in the same way. The intervertebral discs were fixed with 4 % paraformaldehyde solution, decalcified with 20% EDTA solution, embedded with paraffin, and sliced at 5 _ m. The tissues sections were stained with HE, alcian blue + PAS, and safranin O to observe the tissues. The sections were also observed for immunohistochemical localization by the AMC method using IL1ß and TNF- _ as primary antibodies. A part of the tissues of the intervertebral discs were homogenized with 3ml of added PBS. The homogenate was centrifuged at 3000 rpm for 10 minutes and the IL-1ß and TNF-_ levels in the supernatant were measured. RESULTS: There were no histopathologic or immunostaining changes in the control, 1-week and 2-week models. In the 4-week model, (1) the nucleus pulpous was stained unevenly by alcian blue staining and cells producing mucus in the nucleus pulpous reduced. (2) myxopolysaccaride was unevenly distributed in the annulus fibrosus. (3) the chondrodisc had matrix presenting slight metachromasia, and the hyperplasia of the chondrocytes was found. There was no invasion into the blood vessels in the intervertebral discs. In the 8-week model, the distribution of chondrocytes become more uneven, apparently resulting degeneration of the intervertebral disc. Immunohistochemically, in the staining of chondrocytes increased and the production of inflammatory cytokines was confirmed in the 4- and 8-week models. The IL-1ß level in the supernatant increased significantly in all smoking models that continued for at least 2 weeks. DISSCUSSION/CONCLUSION: Microscopically an uneven distribution of cells producing mucus appeared in the 4-week model, and marked reductions and increases of mxyosaccarides were observed in the annulus fibrosus. It is reported that such distribution conditions inhibit the arrangement of annulus fibrosus and consequently result in the formation of small cracks. Using immunohistologic staining, the localization of each cell type producing cytokines was identified. The findings indicate that IL-1ß and TNF- _ were present in chondrocytes in the early stage of the degeneration of the intervertebral disc by smoking, and these antibodies were assumed to be involved. Paper # 41 Static Hydrostatic Loading Induces In-Vitro Apoptosis in Human Intervertebral Disc Cells Nahshon Rand, MD, Saul F. Juliao, BS, Dan M. Spengler, MD, John M. Dawson, PhD (Nashville, TN) INTRODUCTION: Apoptosis, a biological homeostatic mechanism, is involved in processes of cellular aging and degeneration. Apoptosis has been shown to occur in intervertebral disc (IVD) cells. It is unknown whether mechanical loading (at levels to which IVD cells are physiologically exposed) affects the rate of apoptosis in IVD cells. The purpose of this study was to determine the effect of static hydrostatic loading of cultured human IVD cells on the rate of apoptosis. METHODS: Culture Protocol: Intact healthy human discs were obtained from patients undergoing anterior release surgery for correction of deformity. Annulus fibrosus and nucleus pulposus were separated. The tissues underwent enzymatic digestion to release the cells from their matrix. The cellular pellet was used to establish monolayer cell cultures, using Dulbecco's modified Eagle solution (DMEM) supplemented with fetal calf serum (FCS)(20%), 2mM glutamine, 1% streptomycine and an antifungal drug. Cells were cultured in well-plates (150,000 cells/well) and in chamber slides (6,000 cells/chamber). Acidity of the control (nonpressurized) cultures was adjusted to a pH of 6.8 with 0.1 N HCl, to equal the acidity in the pressurized cultures, which was measured to be 6.8 (probably due to the relatively high concentration of CO2 in the pressurized chamber). Hyberbaric Chamber: A thick-walled stainless steel chamber was built to hold cells/tissue cultures. A computer-controlled servo-control valve regulates pressures in the chamber using pre-mixed air and CO2 (5%), allows gas exchange, and controls pressure frequency and duration. Pressure output is proportional to an electrical signal input. The chamber sits in water bath in an incubating oven at 37°C. Load Protocols: We have studied apoptosis in IVD cells at a pressure of 200 psi under three time protocols: Acute (20 seconds), Intermediate Term (6-36 hours) and Long Term (36 hours). For Intermediate Term loading, pressure was held at 200 psi and released at 6 hour intervals so samples could be gathered at 6, 12, 18, 24, 30, and 36 hours. Control samples were maintained at ambient pressures. After loading, samples were returned to an incubator (37° C, 5% CO2) for 72 hours. To establish statistical significance, control and loaded quadruple samples of cultured IVD cells from 6 different patients were studied at 24 hours. Assay protocol: After a 72 hour equilibration, cells of the experimental and control samples were released from the dishes by trypsinization, washed and centrifuged, and cell smears were prepared and fixed in cold acetone. Apoptosis was detected using a commercial apoptotic detection system (Promega) based on the TUNEL technique. Apoptotic cells were visualized by flourescent microscopy, and an apoptotic cell index was calculated. RESULTS: The apoptotic cell index in human disc cell cultures in response to static loading to 200 psi for acute, intermediate, and long-term loading are presented in Table 1 and Figure 1. The results were normalized to the number of cells per well and expressed as a percentage. Assaying quadruples of control and loaded samples from six different patients at 24 hours showed a highly significant difference in ACI, as presented in Figure 2 (p=0.0001). DISCUSSION/CONCLUSION: Static hydrostatic loading of cultured human IVD cells to 200 psi increases the apoptotic cell index in these cell cultures. The detected change in the rate of apoptosis between the loaded and the control group was found highly significant statistically (p=0.0001) at 24 hours. The results indicate that mechanical compression may induce apoptosis in IVD cells. This process in turn may play a role in the pathophysiology of degenerative disc disease. Paper # 42 Rescue of Human Disc Cells from Apoptosis Helen E. Gruber, PhD, Edward N. Hanley, Jr., MD (Charlotte, NC) • (a - North American Spine Society / Cervical Spine Research Society) INTRODUCTION: Previous work from our laboratory has suggested that apoptosis (programmed cell death) may be an important event which contributes to the diminished cell population in the aging and degenerating human intervertebral disc. In the present study, we have asked the question whether disc cells grown in tissue cultures can be rescued from this type of death when exposed to specific cytokines. METHODS: Studies were approved by out institutional Research Review Committee and human subjects Institutional Review Board. Disc specimens were obtained from surgery or from control donor discs obtained from the Cooperative Human Tissue Network. Cells from the annulus were established in monolayer culture using modified Minimal Essential Medium and 20% fetal bovine serum (FBS). Cells were plated onto 8 well NUNC Lab-Tek Chamber Slides at a concentration of 5,000-8,000/well. Treatment groups included: usual growth conditions (20% FBS), apoptosis positive control conditions (1% FBS), insulin (100 ng/ml), insulin-like growth factor-1 (IGF-1) at selected doses up to 500 ng/ml, and platelet-derived growth factor (PDGF) at doses up to 100 ng/ml. IGF-1, PDGF and insulin were added to cells grown in 1% FBS only. Cells were grown for 10 days, fed every 3 days, and fixed on day 10. Apoptotic cells were localized using the Genzyme TUNEL insitu Apoptosis Detection Kit. Data were derived from 2 replicates/subject/assay. At least 600 cells/treatment were scored. Data are expressed as mean ± s.e. m for the mean values from cells from 13 disc from 11 patients (9 adults, mean age 48 years and two pediatric subjects aged newborn and 2 months). RESULTS: As shown in the figure below, cells exposed to 500 ng/ml IGF-1 showed a significantly lowered incidence of apoptotic cells compared to that seen in the 1% positive control group (p=0.007). Cells exposed to 100 ng/ml PDGF showed significanly fewer apoptotic cells compared to the 1% serum positive control group (p=0.006). Neither 0.005, 5 or 50 ng/ml IGF-1, 100 ng/ml insulin, or l or 10 ng/ml PDGF significantly lowered the % of apoptopic cells (data not shown for these IGF-1 and PDGF doses). DISCUSSION/CONCLUSION: Results demonstrate significant reduction of the % of apoptotic cells in tissue culture following exposure to 500 ng/ml IGF1 and 100 ng/ml PDGF. These two cytokines have been shown by others to reduce apoptosis in other cell types (fibroblasts, EMBO J 13:3286; oliogodendrocytes, Development 118:283). Our findings expand the understanding of apoptosis in human cells from the annulus in culture, and suggest that selected cytokines can sucessfully reverse or retard apoptosis in vitro. This avenue of research has clinical therapeutic potential in the treatment of disc degeneration. • If noted, the author indicates something of value received. The codes are identified as: a - research or institutional support, b - miscellaneous funding, c - royalties, d - stock options, e – consultant or employee. For full information, refer to inside back cover. Paper # 43 Map Kinase Pathways Through Ask1 in the Transmission of Apoptosis Signals After Spinal Cord Injury in Rats Shinji Nakahara, Shunji Matsunaga, Hiroki Koga, Kazunori Yone, Kosei Ijiri, Yasuhiro Ishidou (Kagoshima, Japan) INTRODUCTION: Numerous experimental studies have been performed to clarify the complex pathophysiology of spinal cord injury. These studies have suggested a two-step mechanism of neuronal impairment consisting of primary damage due to external mechanical force and secondary damage due to additional damage initiated by the primary damage. Recently, apoptosis has been focused on as one of important factor to affect the secondary injury. It has recently been reported that the mitogen-activated protein kinase (MAPK) signaling cascades may participate in signaling pathways leading to neuronal apoptosis in culture cells. The MAPK signaling cascade, a signal transduction pathway, consists of three distinct members of the protein kinase family; MAPK, MAPK kinase (MAPKK), and MAPKK kinase (MAPKKK). Apoptosis Signal-regulating Kinase 1 (ASK1), a recently identified MAPKKK which activates c-Jun Nterminal kinase (JNK) and p38 MAP kinase, plays important roles in apoptosis signaling. The aims of this study are to clarify the mechanisms of apoptosis after spinal cord injury, and to examine the role of the MAPK cascade in the transmission of apoptosis signals. MATERIALS AND METHODS: Male Wistar rats aged 16 weeks were used. Following laminectomy of T8 and T9, the spinal cord was injured by extradural static weight-compression. In this fashion, a model of reproducible spinal cord injury can be created, with a complete transverse spinal cord lesion and paralysis of the lower extremities. The thoracic spinal cord was removed en at several times and embedded in paraffin after fixation in neutral-buffered formalin. The rat spinal cord was studied by hemotoxylin-eosin staining, Nissl-staining, terminal deoxynucleotidyl transferase (TdT) mediated dUTP nick-end labeling (TUNEL) staining and immunostaining using polyclonal antibodies against ASK1, JNK, and p38. Electron microscopic analysis was performed to reveal the occurrence of apoptosis. To examine the type of apoptotic grail cells, sections were doubled-stained by the TUNEL method and immunohistochemical staining with oligodendrocyte-specific monoclonal antibody RIP or with the polyclonal antibody GFAP. RESULTS: TUNEL-positive cells were maximal at 3 days after injury (Fig. 1). Electron microscopic analysis revealed the occurrence of apoptosis in both neuronal cells and glial cells. The oligodendrocyte-specific maker stained TUNEL-positive cells, but not by the astrocyte-specific maker. Expression of ASK1 was maximal at 24 hours after injury (Fig. 1). Following the expression of ASK1, activated forms of JNK and p38 were observed in apoptotic cells detected by the TUNEL method. Co-localization of ASK1 and activated JNK or activated p38 was observed in the same cell in mirror section. Figure 1: Apoptosis and MAPK Family DISCUSSION/CONCLUSION: In the present study, the occurrence of apoptosis in both neuronal cells and oligodendrocyte was demonstrated. Loss of oligodendrocytes due to apoptosis might have been responsible for the demyelination of axons, since oligodendrocytes myelinate multiple axons in the central nervous system. Following the expression of ASK1, activation and nuclear translocation of both JNK and p38 in neuronal and glial cells were observed after injury. Furthermore, expression of ASK1 and the activated forms of JNK and p38 were observed in the same cells in mirror sections. These finding suggest that induction of ASK1 may contribute to the activation of JNK and p38. That also suggests the possible involvement of the MAPK pathways through ASK1 in the transmission of apoptosis signals after spinal cord injury. Paper # 44 Pro-Inflammatory Cytokines Gene Expression in Human Intervertebral Disc Cells Nahshon Rand, MD (Nashville, TN), Saul F. Juliao, BS (Nashville, TN), Dugan Schwalm, BS (Nashville, TN), John M. Dawson, PhD (Nashville, TN), Yizhar Floman, MD (Jerusalem, Israel), Dan M. Spengler, MD (Nashville, TN) • (a - Cervical Spine Research Society) BACKGROUND: Intervertebral disc cells have a role in the modulation of inflammation in disc herniation. Cultured disc cells have previously been shown to secrete pro-inflammatory cytokines. Recent studies demonstrated the expression of constitutive genes for interleukin-6, TGF-beta and phospholipase A-2 in disc cells. The objective of this study was to examine the patterns of proinflammatory cytokine messenger RNA (mRNA) expression in human intervertebral disc cells. METHODS: Intact human thoracic and lumbar intervertebral discs were obtained from 6 patients undergoing anterior release surgery for correction of deformity. Mean patients' age was 14.6 years (range 9-30). Discs were processed immediately after excision. Nucleus pulposus (NP) was separated and enzymatically digested, and cell cultures were established from the cell pellet. Because several cytokines mRNAs exist at low levels, we extracted total RNA from NP cell cultures with a mixture of guanidine thiocyante and phenol (Tri Reagent, Sigma) according to manufacturer's instructions. The polyadenilated mRNAs were reverse transcribed using a first strand cDNA synthesis kit (Pharmacia Biotech) according to manufacturer's instructions. The cDNAs were detected with high sensitivity by the polymerase chain reaction (PCR) with specific primers for interleukin-1 (IL-1), interleukin-6 (IL-6) and tumor necrosis factor(TNF-alpha). RESULTS: All NP cell lines expressed mRNA sequences for the inflamatory cytokines; IL-1, IL-6 and Beta-actin. In contrast, only four of six samples yielded a strong signal for TNF-alpha, one yielded a moderate signal, and one yielded no signal (figure). DISCUSSION: We have previously reported on the inherent inflammatory properties of intervertebral disc cells, including secretion of pro-inflammatory cytokines. The results of this study support our previous findings demonstrating this capacity of NP cells to synthesize and secrete pro-inflammatory cytokines. CONCLUSION: Disc cells inherently possess the capability to function as inflammatory mediators during disc herniation. • If noted, the author indicates something of value received. The codes are identified as: a - research or institutional support, b - miscellaneous funding, c - royalties, d - stock options, e – consultant or employee. For full information, refer to inside back cover. Paper # 45 Localization of Cathepsins B, D, K and L in Degenerated Intervertebral Discs Kenta Ariga, MD, Shimpei Miyamoto, MD, PhD, Takanobu Nakase, MD, PhD, Shin'ya Okuda, MD, Wenxiang Meng, MD, Eiji Wada, MD, PhD, Takashi Matsuoka, MD, Kazuo Yonenobu, MD, PhD (Suita, Osaka, Japan) INTRODUCTION: Cathepsins B, D, K and L have been investigated as the enzymes contributing to matrix destruction in the articular catilage in patients of osteoarthritis and rheumatoid arthritis. Then these cathepsins degrade collagen type I, II and aggrecan which are the major components of the intervertebral disc matrix. Although these cathepsins are considered as the candidates of the factors contributing to destruction of the intervertebral disc matrix, little is known about the relationship between intervertebral disc degeneration and these cathepsins. The purpose of this study is to examine the involvement of cathepsins B, D, K and L in the pathomechanism of intervertebral disc degeneration by monitoring the expression of these cathepsins in the degenerated intervertebral disc tissues. METHODS: Eight human intervertebral discs were obtained after informed consent from seven patients at the time of the surgery. The surgeries were anterior spinal fusion for cervical spondylosis (four discs) or PLIF for lumbar spondylolisthesis (four discs). The tissue samples were fixed in 4% PFA, decalcified in EDTA, embedded in paraffin, and then prepared for histological sectioning. These sections were evaluated histologically and were examined for immunolocalization using antibodies for cathepsins B, D, K and L by the streptoavidin-peroxydase technique. For further characterization of the stained cells, immunohistochemical staining with antibody for human CD68 and TRAP staining were added. RESULTS: H&E staining indicated significant degenerative findings in all specimens. The findings of degeneration were as follows, 1) fissure formation, 2) disorganization of the lamellar structure of the annulus fibrosus, 3) the cloning of chondrocyte-like cells, 4) cystic or myxomatous change and 5) vascular invasions. Immunohistochemical study showed the expression of cathepsins D and L by disc fibrochondrocytes at the site of degenerative changes, and also showed the expression of cathepsins D and K by TRAPpositive multinucleated chondroclast-like cells localizing at the separation site of the cartilagenous end plate from the vertebral body. Whereas, at annulus fibrosus which remained normal lamellar structure of collagen fibers, there was no positive cell for any cathepsins. DISCUSSION/CONCLUSIONS: Remarkable expression of cathepsins D and L was observed at the various degeneration sites, and expression of cathepsins D and K was observed in the chondroclast-like cells localized at the separation site of the cartilagenous end plate from the vertebral body. These cathepsins are capable of degrading major matrix proteins, then our results suggest that cathepsins D, K and L are involved in the process of intervertebral disc degeneration. Paper # 46 Classification and the Prognosis of Cervical Spondylotic Amyotrophy Diagnosed by Electrophysiological Examination Takanori Saito, MD, Shigeo Akagi, MD, Kunihiko Sasai, MD, Isashi Kato, MD, Ryokei Ogawa, MD (Moriguchi, Osaka, Japan) INTRODUCTION: Cervical spondylotic amyotrophy, known as Keegan type cervical radiculopathy, is a type of cervical spondylosis causing a predominant motor neuron palsy. Anterior horn of the spinal cord and/or anterior nerve root may be damaged but the precise origin is not clear. There is no established treatment for this disorder and the prognosis differs in various reports. The purpose of this study is classifying the cases with cervical spondylotic amyotrophy diagnosed by physical and radiological findings into 4 groups by electrophysiological examinations, clarifying the prognosis of each group and determine the proper treatment of cervical spondylotic amyotrophy. PATIENTS/METHODS: We examined 27 patients (23 men and 4 women, aged 28-72y.o. average,56y.o.) who were suspected, by physical and radiological examinations, having cervical spondylotic amyotrophy between 1993 and 1998 in our hospital. The follow up period ranged from a year and a month to 5 years and 9 months( average, 2years and 11 month). Electrophysiological examinations included needle electrommyogram(nEMG), segmental sensory nerve action potential(seg SNAP), Erbs point stimulated compound muscle action potential(CMAP) recorded in deltoid muscle, motor evoked potential(MEP) in abductor pollicis brevis by transcranial magnetic stimulation, and median or ulnar nerve stimulated somatosensory evoked potentials(SEP). Values deviated more than 3SD were considered to be abnormal in these examinations. RESULTS: Compression of the spinal cord was found in all 27 patients by MRI and the levels were C3/4 in 11 cases, C4/5 in 27 and C6/7 in 5 (some patients had more than two lesins). Muscle strength of shoulder abductor muscles on their first visit was MMT 3 in 4 patients, MMT2 in 16 and MMT01 in 7. Slight sensory disturbance was found in 4 patients and numbness in 7. Compression on the spinal cord was electrophysiologically confirmed in 11 of the 27 patients. Among the 11 patients, prolonged CMCT(more than 12.5msec) of MEP measured by Kimura's method was observed in 10 (average 14.7msec) and prolonged Erb-P14 latency (more than 5.7msec) of ulnar nerve stimulated SEP was found in 7 (average 6.1msec). Results of examinations for peripheral nerves were as fllows. nEMG revealed neurological changes of only C5 innervated muscle in 6, C5/6 innervated muscles in 9, more than 3 segments innervated muscle in 12, and paravertebral muscles in 11. CMAP showed that the amplitude was less than50% of the contralateral side (average 0.9mv) in all cases and the latency was also prolonged (average 6,1msec). The amplitude and /or latency of anterograde segSEP stimulathy at 1st,3rd ans 5th fingers were abnormal. These patients were classified into 4 groups according to the above result of electrophysiological examination. 1)Brachial neuritis(BN)group in 12 patients, 2) motor neuron disese (MNF) group in 2, 3)Keegan type radiculopathy (KR) group in 6, and 4)Keegan type myeloradiculopathy in 7. BN group was diagnosed by normal MEP and SEP (i.e. non-compression of cervical spinal cord), extensive abnormalities beyond C5,6 in nEMG and segSNAP but normal nEMG in paravertebral muscles. MND group was diagnosed by normal SEP and segSNAP, prolonged MEP and extensive abnormalities beyond C5,6 including paravertebral muscles in nEMG. KR group was diagnosed by non-compression of cervical spinal cord(i.e. normal MEP and SEP), C5,6 limited abnormalitiees in nEMG and segSNAP, and neeurological chages of nEMG in paravertebral muscles. KM group was diagnosed by compression of cervical spinal cord disclosed by MEP and SEP, C5,6 limited abnormalities in nEMG and segSNAP, and neurological chages of nEMG and in paravertebral muscles. All patients in BN group were conservatively treated and shoulder elevation became possible in all cases. Microcervical foraminotomy was performed in 3 of 6 patients in KR group, and difficulty in shoulder elevation persisted in one conservatively treated patients and one patients having surgeery. In KM group, 4 patients were operated and 4 were conservatively treated. Three of the 4 patients who were conservatively treated did not restore muscle strength, but 3 patients who were operatd restored it and shoulder elevation became possible. Patients in KR and KM group who were conservatively treated showed a tendency to recover in muscle strength within 3 mmths. Nobody who had CMAP less than 0.1mv could restore shoulder abductor muscle strength greater than MMT 3. DISCUSSION: Cervical spondylotic amyotrophy was finally diagnosed by the effectiveness of decompression surgery so far, while there were several reports describing effectiveness of conservative therapy also. It is speculated from the present study that good response to conservative therapy was attributed to the fact that disease with good prognosis such as BN or focal type Guillain Barre syndrome might be involved in this disorder. The present findings suggest that cervical spondylotic amyotrophy is not a disease with good prognosis if BN or focal type Guillain-Barre syndrome are excluded and that surgical intervention should be considered in cases with low amplitude(less than o.1mv) of CMAP or without any signs of restoration within 3 months from the onset. Paper # 47 Muscle Motor Evoked Potential Detects Segmental Gray Matter Dysfunction During Surgery for Cervical Spinal Cord Tumor Hiroshi Nakamura, MD, Hideki Kitagawa, MD, Yoshiharu Kawaguchi, MD, Ryuichi Gejo, MD (Toyama, Japan) INTRODUCTION: Both the gray and white matter are at risk during the surgery for cervical spinal cord. White matter dysfunction mainly results in gait disturbance and gray matter dysfunction results in upper limb disturbance. However, it has been difficult to monitor the gray matter dysfunction using previous monitoring method. Theoretically, recently developed muscle MEP (evoked muscle potentials after transcranial electrical stimulation) recorded from upper limb can detect segmental gray matter dysfunction. Using muscle MEPs, we could detect the segmental neurological deterioration during the surgery for cervical spinal cord tumors. The purpose for this paper is to report the effectiveness of muscle MEPs in detecting segmental motor deterioration that has not been detected by previous methods. METHODS: Ten patients (4 males, 6 females, aged 23-72 years) who suffered from spinal cord tumor (Eight intramedullary tumors, two extramedullary tumors) were monitored. Four types of evoked potentials were recorded simultaneously; Muscle MEPs, Spinal MEPs (Evoked spinal cord potentials after transcranial electrical stimulation), ESCPs (Evoked spinal cord potentials after spinal cord stimulation) and SEPs (Evoked cortical potentials after peripheral nerve stimulation). All of the patients received propofol-based anesthesia supplemented with low dose fentanyl and nitrous oxide. The neuromuscular block was monitored using DATEX Relaxograph. Transcranial electrical stimulation was applied using a Digitimeter D185 stimulator connected with surface electrodes affixed to the scalp. Each electrode was located 7 cm lateral and 2 cm anterior to the vertex. Trains of 1-9 stimuli 50 ms in duration were delivered with a voltage range of 300-1000 V and an interstimulus interval (ISI) of 1-5 ms. Spinal cord stimulation was applied by epidural catheter electrode. Compound muscle action potentials were recorded with a pair of needle electrodes inserted into the limb muscles in a belly-tendon fashion. Muscle MEPs of diaphragm and sphincter ani were also recorded. Recording muscles were determined in each cases based on patients clinical status and surgical procedure. Spinal cord potentials were recorded from epidural catheter electrodes. Evoked responses were amplified 30Hz to 3kHz and recorded. RESULTS: Muscle MEPs were disappeared in 4 of 10 cases. All of these patients suffered from, at least, transient motor. In two patients, hemilateral segmental muscle response was disappeared which corresponded to postoperative neurological deterioration. In one of these two patients, response from left intrinsic hand muscle disappeared although the other muscle responses were stable. This patient showed left-hand clumsiness and weakness after the operation. In the other patient, left arm response disappeared that corresponded to postoperative deterioration of left arm. The postoperative deterioration in these two patients could no be detected precisely by other evoked potentials. DISCUSSION/CONCLUSION: Transcranial repetitive electrical stimulation and propofol based anesthesia enabled stable recording of muscle MEPs that reflected postoperative neurological status of the individual myotomes. Even a hemilateral segmental deterioration could be detected by this method. Muscle MEPs become monitoring of choice for the surgery of spinal cord tumors. Paper # 48 Internal Morphology of Human Cervical Pedicles Manohar M. Panjabi, PhD, Eon K. Shin, Neal C. Chen, Jaw-Lin Wang, PhD (New Haven, CT) INTRODUCTION: Transpedicular screw fixation is one of the most sophisticated procedures currently being used to stabilize the cervical spine. Anatomic studies have documented the external dimensions and angular parameters of the cervical pedicle, thereby providing the spinal surgeon with relevant information for accurate pedicle screw insertion. Data is lacking, however, regarding its internal architecture and cortical shell thickness along the pedicle axis. It has been shown that screw stability and pull-out strength depend largely on the internal characteristics of the pedicle, not its external dimensions. The purpose of this study is two-fold: 1) To quantify the internal dimensions and cortical shell thicknesses of middle and lower cervical pedicles, and 2) To provide comparative graphical data which would document the variability in pedicle morphology. METHODS: Twenty-six human cervical vertebrae (C3-C7) were secured to a thin sectioning machine to produce three 0.7 mm-thick pedicle slices along its axis. The first slice was made at the isthmus of the pedicle. To document variation along the pedicle axis, two additional slices were obtained: one anteriorly and another posteriorly to the first slice. Plain film radiographs of the pedicle slices were scanned and digitized to facilitate measurement of the internal dimensions. Computer software was specifically developed to determine the external dimensions (i.e. pedicle height and width) and the internal dimensions (i.e. cortical shell thicknesses of the superior, inferior, lateral, and medial walls, and the cancellous core height and width) of cervical pedicles (Fig. 1A). RESULTS: Superior and inferior wall cortical thickness were found to be similar. The medial cortical shell (mean value range: 1.2-2.0 mm) was measured to be 1.4 to 3.6 times as thick as the lateral cortical shell (mean value range: 0.4 –1.1 mm) across all pedicle thin slices and vertebral levels (Fig. 1B). Graphical comparisons of thin slices from C3- to C5 demonstrated that the cervical pedicle was semicircular. , rectangular, or even triangular on cross section. Large variability was observed not only between individual spines and vertebral levels, but also within the pedicle axis. The cancellous core in such specimens ranged from being finely trabeculated to nearly absent. These variabilities were reflected in the large standard deviations for our measurements. Thin slices from C6 and C7 exhibited greater uniformity in size and ovoid shape, with the slices becoming narrower as one moved antero-posteriorly along the pedicle axis. CONLCUSIONS: The lateral cortex was significantly thinner than the medial cortex in all thin slices and at all vertebral levels. The lateral cortex of cervical pedicles is responsible for protecting the vertebral artery, with the exception of C7 pedicles. Thus, extreme caution must be exercised when inserting pedicle screws so as not to perforate the thin lateral cortex of the cervical pedicle. The extensive variability in pedicle morphometry must also be recognized as the standard deviations for measured cortical values were found to be large. Graphical representations of the pedicle demonstrated its complexity as a threedimensional structure whose features, with respect to cortical thickness, orientation, and overall shape, change as one moved antero-posteriorly. Characteristics of the cervical pedicle at different spinal levels must be carefully noted prior to transpedicular screw fixation to avoid cortical wall violations. Paper # 49 Cervical Pedicle Screws: Comparative Accuracy of the Abumi Method versus Stealth Station™ Steven C. Ludwig, MD, Joseph M. Kowalski, MD, Charles Edwards, II, MD, John G. Heller, MD (Decatur, GA) INTRODUCTION: Authorities disagree on the relative safety and accuracy of different cervical pedicle screw insertion techniques. The reported clinical experience for both traumatic and nontraumatic cervical spine reconstruction uses a fluoroscopic assisted method (Abumi). The most accurate in vitro technique uses a computer assisted image guided surgical system (Stealth Station™, Sofamor Danek). The purpose of our investigation was to assess the safest method of transpedicular screw placement in the human cervical spine using these two separate surgical techniques. METHODS: Twelve human cadaveric cervical spines were instrumented from C3-C7 according to two techniques. In Group I, fifty screws were placed using the Stealth Station™, (Sofamor Danek). In Group II, sixty-seven screws were placed by the Abumi/fluoroscopy-assisted method. Postoperative CT scans were obtained. A cortical breach was considered critical if the screw encroached upon any vital structure. Findings were confirmed by open dissection by an independent observer. RESULTS: In Group I (Stealth Station™) 82% of screws were placed in the pedicle, and 18% had a critical breach. In Group II (Abumi method) 88% of screws were placed in the pedicle, and 12% had a critical breach. No statistically significant differences were demonstrated between the groups. The most common structure injured in each group was the vertebral artery. Further analysis based upon minimum external pedicle diameters revealed a significantly higher rate of critical breaches with pedicle diameters less than 4.5mm. DISCUSSION/CONCLUSION: This investigation compared the anatomic safety of the two most accurate methods for inserting cervical pedicle screws. Computer assisted image guidance did not appear to enhance the safety or accuracy of insertion. Independent of the method used, there remains a noteworthy risk of a critical breach between the C3-C6 levels, especially with pedicle diameters less than 4.5mm. In those unusual circumstances which call for cervical pedicle screw fixation, we recommend that surgeons select their method of insertion with a full understanding of the rate and type of potential complications that may occur. Paper # 50 Cervical Pedicle Screws: A Biomechanical Comparison of Two Insertion Techniques Joseph M. Kowalski, MD, Steven C. Ludwig, MD, William C. Hutton, DSc, John G. Heller, MD (Decatur, GA) INTRODUCTION: Wiring techniques remain the gold standard for posterior cervical fixation. However, absent or deficient posterior elements may dictate the use of alternative fixation techniques. Lateral mass screws are most commonly used in posterior cervical plating. However, local bone deficiencies due to fracture, tumors, inflammatory lesions or neural decompression may dictate the use of an alternative fixation point, such as the pedicle. Though they have inherently greater anatomic risk, cervical pedicle screws have been shown to have significantly higher pullout strength than lateral mass screws. Different insertion methods have been recommended, with varying degrees of accuracy demonstrated in vitro. The purpose of this study was to determine if there is a significant difference in screw purchase for two distinct methods of cervical pedicle screw insertion. METHODS: Fifty fresh disarticulated human vertebrae (C3-C7) were screened with computed tomography for anatomical pathology and pedicle morphometry. The right and left pedicles were randomly assigned to one of two insertion methods. The Standard method requires drilling through the lateral mass and the pedicle along the intended screw path into the vertebral body, prior to tapping and inserting the screw. With the Abumi technique the cortex and cancellous bone of lateral mass are removed with a high-speed burr, which provides a direct view of the pedicle introitus. The pedicle is then probed bluntly, tapped and the screw inserted. Thus screws inserted with the Abumi method do not purchase either the cortex or cancellous bone of the lateral masses. After insertion of the 3.5mm cortical screws, any pedicle wall violations were documented prior to subjecting each screw to a uniaxial load-to-failure. RESULTS: There was no significant difference in the mean pullout resistance between the Abumi (696 N) and Standard (636.5 N) insertion techniques (p=0.412). There were no differences in pull-out resistance between vertebral levels or within vertebral levels. Two (2%) minor pedicle wall violations were observed. DISCUSSION/CONCLUSION: In the event that pedicle screw purchase is deemed appropriate for a particular cervical lesion, a surgeon must select a method of insertion. Differences in the safety and accuracy of various methods of cervical pedicle screw insertion have been quantified in vitro. Whereas the reported clinical safety of the Abumi method compares favorably with the in vitro evaluation of frameless stereotactic guidance technology, doubt has arisen over the mechanical effect of sacrificing the cortical and cancellous bone of the lateral mass in preparing the screw hole. We conclude that surgeons need not be concerned about reduced screw purchase when deciding between the Abumi methods and its alternatives. Further clinical studies are required to clarify the issues of relative safety in vivo. Paper # 51 Cervical Spine Pedicle Screws - Clinical and Radiographic Evaluation in 30 Consecutive Patients Claes Olerud, MD, PhD (Uppsala, Sweden), Bengt Lind (Gothenburg, Sweden), Bo Sahlstedt (Uppsala, Sweden) • (d, e – company manufacturing implants owned by family) INTRODUCTION: In biomechanical experiments cervical spine pedicle screws give improved purchase compared to traditional fixation techniques but cadaver experiments have shown a high incidence of misplaced screws with risk of neuro-vascular injury. This has, however, not been the experience in the few clinical reports on the subject. The reason may simply be differences in insertion technique. We use a combination of anatomical landmarks to identify the correct entrance point on the lamina, lateral projection fluoroscopy, and gentle probing with a blunt "pedicle feeler" to prepare the screw hole in the pedicle. METHODS: To evaluate the safety and efficacy of this technique a follow-up study was done on 30 consecutive patients (14 women) with a mean age of 68 (37-85) years undergoing posterior cervical spine fixation. Indications for surgery were rheumatoid arthritis in 10, spinal stenosis in six, trauma in six, metastases in four, miscellaneous in four. Twenty-three transarticular C2-C1 screws, 50 pedicle screws in C2-C7, and 57 pedicle screws in the upper thoracic spine were used. The patients were followed clinically and radiologically at 6, 12, and 26 weeks. To evaluate the screw position a special radiographic technique was used: plain radiographs in four standard projections were obtained under fluoroscopic guidance by an experienced radiologist (BS). An independent follow-up investigation was conducted at 12 months by an experienced spine surgeon (BL) from an other university institution who reviewed the records and radiographs and evaluated the clinical outcome in 19 of the 20 still alive patients with special attention to neurological function. RESULTS: All pedicle screws in the cervical spine, and all but one screw in the upper thoracic spine were in the correct position. The misplaced screw was lateral to the pedicle of Th2 but without any consequence to the patient. There were no peroperative neuro-vascular injury that could be referred to the use of pedicle screws. One patient had experienced neurological deterioration of the function of one nerve root. This was non-instrument-related and could be attributed to insufficient decompression at time of surgery. No pedicle screw had loosened. There were no disturbances of fusion healing related to the use of pedicle screws. Two patients experienced fixation failure of instrument extension to the occiput. Both healed their fusions uneventfully after revision. CONCLUSION: The applied technique for placing pedicle screws in the cervical and upper thoracic spine seems both safe and efficient. • If noted, the author indicates something of value received. The codes are identified as: a - research or institutional support, b - miscellaneous funding, c - royalties, d - stock options, e - consultant or employee. For full information, refer to inside back cover. Paper # 52 Posterior Reduction and Fusion of Cervical Spine Injuries with Traumatic Disc Herniation - Indirect Posterior Reduction of the Herniated Disc Using Cervical Pedicle Screw System Yasuhiro Shono, MD, Kuniyoshi Abumi, MD, Kiyoshi Kaneda, MD (Sapporo, Japan) INTRODUCTION: This study retrospectively reviewed the sixteen patients of middle and lower cervical spine injuries with traumatic disc herniation who were surgically treated by single posterior reduction and fusion procedure using cervical pedicle screw system. The study was undertaken to evaluate whether the procedure effectively reduces the disc herniation, and whether it can be safely conducted without performing anterior decompression. METHODS: Total 73 patients with middle and lower cervical spine injuries were identified. In 50 patients, pre- and post-operative MR images were obtained, and disc herniation was defined as the presence of an extruded disc that deformed the thecal sac or nerve roots. Sixteen patients (32%) had traumatic disc herniation and single posterior reduction-fusion procedure was performed using cervical pedicle screw system. Average follow-up period was 4 years and 3 months (2 years - 6 years and 3 months). RESULTS: All patients had kyphotic deformity of 18 degrees in average, and this was corrected to 0.7 degrees of lordosis. Anterior translation was reduced from 8mm to 0.7mm. Preoperative disc height ratio of 63% (normal=100%) was improved to 104%. Preoperative MR images showed traumatic disc herniation in all 16 patients, and, after surgery, reduction or reversal of a disc herniation was observed in all. Thecal sac and/or spinal cord compression had vanished after indirect decompression by posterior procedure. No additional decompressive procedures were required to remove residual herniated mass. Before surgery, four patients presented with cervical radiculopathy, 10 with myelopathy (eight incomplete and two complete), and two without neurologic symptoms. At final follow-up, all four patients with radiculopathy showed complete recovery and six patients (60%) with myelopathy improved at least one Frankel grade. No patients showed any neurologic deterioration immediately after surgery and at the course of the final follow-up. All patients obtained solid bone union and there were no implant related complications. DISCUSSION/CONCLUSION: Traumatic disc herniation may occur frequently in association with the cervical spine injury. In our series, 16 patients out of 50 (32%) showed traumatic disc herniation, and most commonly occurred in distractive flexion injury (50%). Cervical spine injury associated with disc extrusion should be treated with a great caution. Extruded disc material could potentially increase cord compression following closed reduction, particularly if the facet dislocation holds the spinal canal open in an angulated position. Open reduction and fusion using posterior wiring may also cause neurologic deterioration due to extruded hernia mass compressing the spinal cord. Conventional instrumentations including posterior wiring technique primarily rely on compression force to reduce dislocation or subluxation of the injured spinal segment. Consequently, reduction of the anterior translation with local kyphotic deformity by compression force application causes narrowing of the injured disc. At the same time, dislocated caudal vertebra drags extruded disc material into the spinal canal. This further displaced disc material causes increased neural compression. Therefore, it has been contraindicated to perform posterior reduction surgery alone in treatment of cervical spine injuries with traumatic disc herniation. Circumferential fusion consists of anterior decompression and fusion followed by posterior reduction and stabilization has been a recommended method in this type of injury pathology. On the other hand, The pedicle screw fixation system allowed three dimensional reduction of the injured cervical segment with the restoration of the disc height and lordosis alignment. Postoperative MR images showed reduction or reversal of a disc herniation after surgery and the thecal sac and/or spinal cord compression had vanished after indirect decompression by posterior pedicle screw procedure. Moreover, no patients showed any neurologic deterioration after surgery. The cervical pedicle screw system provides effective reduction-fixation of the cervical spine injuries with traumatic disc herniation, and the surgery can be performed safely in a single posterior procedure without need of additional anterior decompression. Paper # 53 Revision Surgery of the Previously Operated Cervical Spines: Application of Pedicle Screw Fixation Sytems Kuniyoshi Abumi, MD, Yasuhiro Shono, MD, Kiyoshi Kaneda (Sapporo, Japan) INTRODUCTION: Revision surgery of the previously operated cervical spine holds many difficult reconstruction problems. The purpose of this report is to investigate the clinical results of reconstructive surgery using pedicle screw fixation for previously operated cervical spine. METHODS: Between 1991 and 1997, 45 patients with a previously operated cervical spine underwent revision surgery using pedicle screw fixation systems. There were 24 men and 21 women, and their average age was 58.1 years (range, 22-75 years). Cervical disorders for primary surgery were cervical spondylotic myelopathy in 20 patients, rheumatoid arthritis in 11, spinal injuries in five, spinal cord tumor in three and others in six. The main etiologies those required revision surgery were progression of degenerative change at the adjacent fusion level causing myelopathy in 13 patients, failure of anterior fusion in 11, postlaminectomy kyphosis or instability in ten, pseudarthrosis of posterior or circumferential fusion in five, insufficient posterior decompression in four, and recurrence or residual spinal cord tumor in two. Among the 11 patients with failure of anterior fusion, four patients had been managed by using anterior plate fixation. Before the revision surgery, 16 patients had marked cervical kyphosis with an average angle of 29 degrees (range 15-46 degrees). For revision surgery, 32 patients underwent posterior surgery alone using pedicle screw fixation, and the remaining 13 patients required supplemental anterior surgery. Twenty-five patients underwent simultaneous posterior decompression including laminectomy, laminoplasty and foraminotomy. Pedicle screw-plates system designed for the cervical spine (CPS) was used in 37 patients. Combination of CPS screw and Isola rod was used in four patients. VSP for thoracolumbar spine was utilized for one patient in the early phase of this series. Occipitocervical fixation using occipitocervical rods and CPS screws was performed in three patients. The number of the spinal segments in which fixation was performed ranged from one to 11 (average 3.5). Screws were inserted after probing and tapping under lateral X-ray image control. If pedicle prove could not pass the hard portion of the pedicles, holes by which to initiate probing were created by using Kirschner wire. After surgery, short soft neck color or Philadelphia color was used in all patients, and worm for varied time depending on extent of osteoporosis and range of fusion. However, no patients required halo-vest immobilization. All patients were permitted to walk or sit up in bed one day postsurgery unless contraindicated by their general condition. RESULTS: All patients underwent more than 2 years follow-up review. Bony union was obtained in all patients. Kyphosis was corrected to 5.0 degrees in average (range: -8 to 32 degrees) at the latest follow-up. Loss of correction during bony union was within 3 degrees in all patients. There were no instrumentation failures. Of the 38 patients who had myelopathy preoperatively, 31 patients improved after surgery to some extent. There was no postoperative neurological deterioration. There were two cases of deep infection healed by continuous irrigation without metal removal. There were three patients with cerebrospinal fluid leakage related to posterior decompression. According to postoperative screw evaluation by CT and plain x-ray films, 18 of 244 screws (7.4%) inserted into the cervical pedicles showed perforation of the screw from the pedicle wall in various extents. However, there were no patients with neurovascular complications directly attributable to screw insertion into the cervical pedicles. Iatrogenic foraminal stenosis which caused C5 radiculopathy was observed in one, and the radiculopathy healed by foraminotomy and application of distraction force without removing the screws. DISCUSSION/CONCLUSION: There are many etiologies those require revision surgery in the cervical spine including pseudarthrosis, postlaminectomy instability or kyphosis, insufficient decompression in previous surgery, development of degenerative change in the adjacent fused segment, etc. Revision surgery of the previously operated cervical spine holds many difficult reconstruction problems including a difficulty in exposing the operated field,redecompression, insufficient bed for bone grafting, kyphotic and/or translational deformity, etc. Repeated anterior approach in revision of failed anterior cervical fusion requires difficult dissection with risks of the vascular and visceral complications, and sometimes require extend of the fusion level to obtain the optimum bone graft bed. Furthermore, many patients who require one-stage posterior decompression and stabilization in revision surgery. Therefore, posterior reconstructive procedure is invariably required regardless of the initial surgery. Among the various posterior cervical instrumentation, facet wiring, lateral mass plate-screw and pedicle screw fixation does not require the lamina for fixation. However, the pedicle screw fixation has many biomechanical advantages. According to the results of this series, pedicle screw fixation provided high fusion rate and sufficient correction of kyphosis in the reconstructive surgery of previously operated cervical spine with complicated condition. We sometimes encountered the difficulties in pedicle screw in the patients who had undergone posterior surgery. As the results, rate of screw malposition by postoperative radiological evaluation was 7.4%; however, no neurovascular complications directly attributable to screw insertion were observed. Probability of neurovascular injury by perforated screw from the pedicle may be low. Nevertheless, use of pedicle probe and X-ray image, and use of K-wire in some patients raised the reliability of the cervical pedicle screw insertion. In conclusion, pedicle screw fixation which does not require the lamina for stabilization is the useful procedure in revision surgery for the previously operated cervical spines. In addition, using this procedure, rigid fixation provides a high fusion rate as well as corrective capability of kyphosis. Paper # 54 Standing Cervical Sagittal Alignment In Solid Uninstrumented Anterior Cervical Fusions James W. Hardacker, MD, Peggy Pattern, RT(R), Philip W. Pryor, MD (Carmel, IN) INTRODUCTION: Anterior cervical discectomy and fusion is a widely accepted reconstructive procedure. However, little information on standing preoperative vs. postoperative cervical sagittal alignment measures in uninstrumented fusions is available. MATERIALS/METHODS: We reviewed a consecutive series of 45 patients undergoing primary one or two level anterior cervical discectomy and fusion with iliac crest autograft for radicular symptoms. 32 patients had one level reconstruction while 13 patients had two level reconstruction. The majority of operative sites were C5-C6 and C6-C7 but not exclusively. Attempts at maximizing intraoperative and postoperative sagittal plane alignment was stressed. All patients were braced until fused with mean follow-up 22 months, range 12-52 months. All patients had preoperative and postoperative radiographs in AP and standing lateral planes. Mean age was 46 years (range 29-69 years) with 23 patients female and 22 male. Total and segmental measures were completed via that Cobb method from the occiput to C7 by a single observer with a digital level. The odontoid lateral plumb line was referenced to the C7 vertebral body. Preoperative and postoperative measures were assessed with a t test or Mann-Whitney rank sum test as appropriate. Intraobserver and interobserver reliabilities were assessed with Pearson correlations by randomly remeasuring all lordosis values in 33% of patients. Preoperative and postoperative work function, analog pain scales, and comorbidities were documented. RESULTS: Mean preoperative total cervical lordosis was (minus=lordosis/positive=kyphosis) – 41.3 degrees (-6.9- - 67.0 degrees) while mean postoperative total cervical lordosis was – 44.1 degrees (-14.7 - -74.2 degrees) and was not statistically significant (p=.30). Preoperative vs. postoperative segmental lordosis values and p values were: occiput-C1 (-3.1 degrees vs. –2.1 degrees, p=.38), C1-C2 (-32.7 degrees vs. –33.3 degrees, p=.65), C2-C3 (-1.6 degrees vs. –1.8 degrees, p=.86), C3-C4 (-1.2 degrees vs. –3.7 degrees, p=.05), C4-C5 (.83 degrees vs. –1.4 degrees, p=.64), C5-C6 (-.33 degrees vs. .11 degrees, p=.71), C6-C7 (-3.9 degrees vs. –3.6 degrees, p=.83). Most segmental lordosis existed at C1-2 and was not statistically different preoperatively vs. postoperatively at any level except C3-4. Preoperative and postoperative odontoid plumb line were 17.1mm vs. 17.7mm anterior to the midpoint of the C& vertebral body and not statistically different (p=.83). Pearson correlation coefficients for intraobserver reliability of randomly remeasured lordosis values ranged from .88-.99. Interobserver reliability of remeasured cervical lordosis values revealed correlation coefficients of .85-.99. Mean preoperative vs., postoperative maximum pain level on a 10 point scale was 8.4 vs. 2.8 (p=.001). Of the 58 fused levels, 62% maintained their lordosis postoperatively, and an additional 7% increased their operative level segmental lordosis a mean of 8.8 degrees while 18 levels lost greater than 5 degrees of segmental lordosis (mean 8.5 degrees). Of the operative sites, 8 levels had 5 degrees or more of preoperative segmental kyphosis while postoperatively 15 of 58 fused levels had greater than 5 degrees of segmental kyphosis. Pain levels diminished in this subpopulation with segmental kyphosis from preoperative mean 7.7 to 3.4 postoperatively. Multilevel cervical segmental kyphosis including nonoperative sites existed in 20% of patients preoperatively, but in only 6% postoperatively. The most common levels for preoperative segmental kyphosis were C4-C5 and C5-C6, while most common lever of postoperative segmental kyphosis was C5-C6. There was no preoperative or postoperative total cervical kyphosis present. Work status was maintained or increased in 60% of patients while 35% remained retired, not working or in domestic roles. Only 2 patients (4%) stopped working after surgery. DISCUSSION/CONCLUSION: Anterior cervical fusion has excellent pain reduction and statistically maintained preoperative vs. postoperative cervical total and segmental alignment. The majority of segmental cervical lordosis existed at C1-C2 before and after surgical intervention. Although the majority of operative levels maintained or improved lordosis segmentally, 31% of operative levels lost a mean of 8.5 degrees segmental lordosis. Although this lordosis loss is not statistically significant, it should be considered when preoperative planning is conducted for anterior cervical fusions without instrumentation. Paper # 55 In Vivo Determination of Cervical Spine Kinematics Using Fluoroscopy A. Alexander M. Jones, MD, Richard D. Komistek, PhD, James T. Brumley, II (Denver, CO) • (a – Osteonics, Stryker Spine) INTRODUCTION: Cervical disc degeneration can be debilitating, significantly reducing a patient’s range of motion and their ability to perform normal neck motions. Cervical arthrodesis, by definition, decreases neck motion and disc degeneration adjacent to fused levels is well recognized clinically. Previous kinematic studies on the human spine have been conducted under in vitro conditions, but recently, fluoroscopy has been used on other joints to determine in vivo kinematics. This study focuses on the determination of the in vivo kinematics during active flexion and extension of normal, degenerated, and fused cervical spines. METHODS: Fifteen adult subjects (five normal, five degenerative, and five 2level fused) were analyzed under fluoroscopic surveillance. The subjects having a fused spine were fused at the C5-C6 and C6-C7 levels (deemed clinically successful). Fluoroscopic images were downloaded directly to a workstation computer and were analyzed at ten various increments of spinal flexion/extension. During the computer analysis, constant points on each of the seven cervical spine vertebrae were tracked throughout the range of motion (Figure 1). Fig. 1: Schematic of digitization process used to analyze the in vivo fluroscopy images. The points were analyzed to determine the rotation angle of each individual vertebra. The rotation of each vertebra relative to the subsequent vertebra was plotted with respect to time and the data was curve-fit to obtain a temporal function that represented the motion pattern. Using a mathematical model, the relative velocities were obtained and used to determine the partial angular and linear velocities. Kinematic comparisons were then made for each subject within each group and with subjects from the other two groups. RESULTS The kinematic patterns for the three groups varied considerably. Normal cervical spines showed a smooth, arc like motion, whereas, the degenerative spines demonstrated inconsistent motion patterns especially at the C5, C6, and C7 vertebrae (location of the degeneration) (Figures 2 & 3). The two-level fusion patients demonstrated no motion at the fused C5-C6 and C6C7 levels and marked abnormalities in the motion pattern at the C4-C5 level (level above fusion) (Figure 4). There was also a distinct difference in the relative angular velocities for the three groups. The subjects having a fused spine experienced a significant change in the relative angular velocities above the fused joint. Fig. 2: Kinematic patterns for a normal cervical spine throughout the full range-of-motion cycle. Fig. 3: Kinematic patterns for a degenerative cervical spine throughout the full range-of-motion cycle. Fig. 4: Kinematic patterns for a fused cervical spine throughout the full range-of-motion cycle. DISCUSSIONS: This is the first study to document the in vivo motions of the cervical spine. This analysis determined that there is a significant difference in the in vivo kinematics of the normal cervical spine compared to a degenerative or fused cervical spine. The abrupt change in motion detected above the fused levels may lead to further complications, possibly including accelerated disc degeneration due to the abnormal loading conditions adjacent to the fused cervical segments. • If noted, the author indicates something of value received. The codes are identified as: a - research or institutional support, b - miscellaneous funding, c - royalties, d - stock options, e - consultant or employee. For full information, refer to inside back cover. Paper # 56 The Role of Cervical Spine Radiography in Asymptomatic Trauma Patients: A Meta-Analysis of Literature Sloane R. Blair, MD (Houston, TX), Debi Mujherjee, MD (Shreveport, LA), Ronald W. Lindsey, MD (Houston, TX) INTRODUCTION: Routine cervical spine radiography (CSR) of trauma victims in the emergency setting is known to be highly inefficient and costly. However, the validity of forgoing CSR in asymptomatic patients (ASPS) is highly controversial. The objective of this study was to perform a meta-analysis on to the existing literature pertaining to this issue. METHODS: A study group of trauma CSR was established from twenty articles and five abstracts which included the patients clinical symptoms. ASPS had no alteration of consciousness, neck pain, tenderness, neurologic deficit, and/or severe polytrauma. The study group papers were categorized according to the quality of each study: Group A (8 papers): prospective articles with complete data on the number of ASPS and false-negative diagnoses, Group B (5): prospective study with incomplete asymptomatic and/or false-negative data, Group C (7): retrospective studies, Group D (5): abstracts, Group E (8): significant data omission or misclassification . The data was subjected to chisquare tests as normally used for the categorical models and contingency tables were constructed. The computations were done using statistical software (JMP software, SAS Institute Inc.) in a Macintosh Centris 650 computer. RESULTS: Only Group A presented sufficient data on the ASPS and the number/type of false-negative diagnosis. When these papers were combined, A totaled 5,168 patients, 1,923 ASPS and only 1 false-negative diagnosis (a nonthreatening transverse process fracture). The statistical analysis demonstrated prob>ChiSquare value of 0.0083 (Likelihood ratio) for the data in Group A. This analysis was significant in suggesting that CSR was not warranted on these ASPS. The data in Groups B,C, and D although not significant, demonstrated the same trend. DISCUSSION/CONCLUSION: The literature on the emergency clearance of the cervical spine although voluminous, is controversial as to the merits of a more clinical based criteria for determining which patients warrant CSR. Our study suggests that CSR in trauma victims based on clinical indicators appears feasible. Paper # 57 Diagnosing Cervical Fusion After Anterior Cervical Surgery: A Comprehensive and Critical Literature Review Navin Sethi, MD, Dan Riew, MD, James Devney, DO (St. Louis, MO) INTRODUCTION: Since Smith and Robinson first described anterior cervical fusion surgery, there have been numerous studies examining fusion rates. A cursory review of the literature revealed a wide disparity in fusion percentages. One possibility for this discrepancy may be due to the manner in which fusion was determined. We undertook this study to review all existing English language literature on anterior cervical fusion procedures over the last thirty years to ascertain the methodology by which fusion was assessed. We sought to determine if there was a consensus on what constituted a solid arthrodesis. METHODS: A Medline literature search (1966 to January, 1999) was performed for articles related to anterior cervical fusion. Eighty-nine English language articles were found in which the authors presented their results following anterior cervical fusions. These studies were then critically analyzed as to patient number, age, diagnosis, number of levels fused, type of graft, instrumentation, follow-up period and fusion rate. The definition of a solid fusion was also noted for each of the studies. RESULTS: There were a total of 7001 patients in the combined studies. Most studies included a breakdown of the number of levels that were fused; however, for 2049 patients the number of levels that were fused could not be determined. There were 2723 one-level fusions, 1586 two-level fusions, 594 three-level fusions and 49 four-level or higher fusions. The follow-up period ranged from six months to nine years. Fusion rates in the studies ranged from 45% to 100%. The definition of fusion was highly variable among the studies. The most stringent definition of a solid fusion was spanning trabeculae across the graft/host surface AND the absence of motion on flexion/extension lateral cervical spine radiographs. Such criteria were noted in only 18% of the studies (16/89). One third of the studies considered the presence of EITHER spanning trabeculae across the graft/host surface OR the absence of motion on flexion/extension x-rays as adequate for determining successful fusion (29/89). Fifteen studies based their definition of a solid fusion solely on the presence of spanning trabeculae (17%) while seven studies based their criteria on the absence of motion on flexion/extension lateral radiographs (8%). The vast majority of studies gave no criteria for their diagnosis of fusion (44/89). Four of the studies used CT scans along with plain radiographs and one study used MRI to diagnose fusion. Some studies based their solid arthrodesis rate solely on clinical grounds. DISCUSSION/CONCLUSION: To our knowledge, this is the first comprehensive review of the literature looking at the criteria for a solid arthrodesis following anterior cervical surgery. It is clear from a review of the literature that no universal criteria exist for a solid arthrodesis. In order to perform prospective studies to determine outcome following anterior cervical surgery and to assess whether fusion status has a bearing on outcome, a universal definition of solid arthrodesis must be present. Although the indications, levels of fusion, type of graft, and use of instrumentation varied among these studies, there was a wide range of fusion rates even for similar procedures. One reason for such a discrepancy is the modality by which fusion was ascertained since this differed amongst the various studies. Only 18% of the studies used the presence of both bone bridging and the absence of motion on flexion/extension films as their criteria for a solid arthrodesis. Such criteria were used in the original Smith and Robinson study; however, our data show that multiple studies since that time have wavered from this definition. As the fusion rates in these studies are based on different criteria, meaningful comparisons among these studies are not valid. There seemed to be no universal standard for assessing solid fusion after anterior cervical spine surgery. Until common criteria are adopted for solid fusion following anterior cervical spine surgery, the actual rate of fusion for most clinical studies is both unknown and suspect. Studies need to be performed determining the components of a solid arthrodesis so that a universal definition of fusion can be adopted. Paper # 58 Plain Radiographic Assessment of Anterior Cervical Fusion K. Daniel Riew, Alan Hilibrand, Jeff Wang, James Kang, Lawrence G. Lenke, Keith Bridwell PURPOSE: Numerous articles have appeared in the scientific literature describing fusion rates following anterior cervical arthrodesis, without any uniform standard for radiographic assessment of fusion. Although flexionextension views have been widely utilized, no study has critically assessed the importance of obtaining these views in addition to neutral lateral radiographs. Before newer bone graft substitutes can be prospectively studied for their ability to produce successful anterior fusion, a gold standard is needed for the plain radiographic assessment of fusion. The purpose of the present study was to determine 1) how often the assessment of fusion based on neutral lateral radiographs agrees with the findings of flexion-extension views, 2) whether anterior cervical plating prevents detection of motion on flexion-extension views, and 3) how accurate is the interpretation of fusion status by a treating surgeon at the time of the patient's office visit? MATERIALS AND METHODS: All patients who had undergone anterior decompression and arthrodeses without posterior surgery from August 1995 to December 1997 by a single surgeon at a single institution with at least one year follow-up with A/P, lateral and flexion/extension lateral radiographs. We excluded the patient if the anterior and the posterior elements of the fused segments were not easily visible on any of the radiographs, if they had previous or subsequent posterior surgery at the involved segments, or if cages were placed in the anterior cervical spine. A total of 41 patients qualified for the study. The radiographs were then presented to each of five spine surgeons who had no involvement in the care of any of the patients. For the first reading, each surgeon was independently presented with the AP and lateral views only. They were asked to rate each level as a solid arthrodesis, pseudarthrosis, questionable or unable to determine due to poor radiographs. Any patient who was felt to have poor radiographs by any of the five surgeons was excluded from the rest of the study. Criteria for fusion were: bridging bony trabeculae, with no pseudarthrotic line. For multiple level procedures, they were asked to comment about each level. For corpectomies greater than one level, they were only asked to comment about the cephalad and caudad ends of the construct. For the second readings, each surgeon was presented with flexion-extension laterals and asked to rate films as a pseudarthrosis if there was any motion. Finally, we compared the assessments of fusion made by the 5 treating surgeons with the assessments made by the treating surgeon at the time of the patient office visit. RESULTS: Of the 41 patients included in the study, 34 were felt to have adequate lateral and flexion/extension radiographs by all 5 readers. A total of 60 fused levels were evaluated by each of the 5 surgeons for a total of 300 assessed levels. Based upon neutral lateral radiographs, 242 of these levels were diagnosed as solid, 30 as a pseudarthroses, and 28 as inconclusive. Of the 242 "solid fusions" on neutral radiographs, 35 (14%) had motion consistent with a pseudarthrosis on dynamic lateral radiographs. Of the 30 levels diagnosed as pseudarthrotic on neutral views, 24 were confirmed on the dynamic views and 6 were found to be solid. Of the 28 levels read as "inconclusive" on neutral films, all were able to be interpreted as solid (15) or pseudarthrotic (13) on the dynamic views. Anterior cervical plating did not appear to prevent the detection of a pseudarthrosis, as 11 of the 26 patients treated with a plate had detectable motion on flexion-extension views. In addition, 9 patients who were plated across multiple levels had motion identified at one level only, suggesting that individual levels behind a plate can be assessed. At the office visits, the operating surgeon diagnosed pseudarthroses in 9 of the 34 patients. The five surgeons agreed with this diagnosis for all nine but found an additional 4 out of 34 (12%) patients who had detectable motion on dynamic views. DISCUSSION: The rates of successful anterior cervical fusion reported in the scientific literature vary widely, with a variety of plain radiographic criteria used to assess fusion status. The present study demonstrated that even a confident assessment of fusion status based upon a neutral lateral radiograph is likely to be incorrect in 10 -- 20% of cases. In addition, in this study 10% of assessments based upon a neutral lateral radiograph were inconclusive, and over half of these were reliably determined to be a pseudarthrosis based upon dynamic lateral radiographs. It should also be noted that in this study the treating surgeon underestimated the rate of pseudarthrosis by 10% despite the presence of flexion-extension views. Consequently, we believe that studies reporting fusion rates based upon the treating surgeon's assessment of neutral lateral radiographs should be viewed with caution. Paper # 59 Pseudoarthrosis of the Cervical Spine: A Comparison of Radiographic Diagnostic Measures Sanford E. Emery, MD (Cleveland, OH), Lisa K. Cannada, MD (Cleveland, OH), Steve Scherping, MD (Great Falls, VA), Paul K. Jones, PhD (Cleveland, OH) INTRODUCTION: The accuracy of diagnostic methods for detecting pseudoarthrosis has been poorly documented. Radiographic criteria mentioned in the literature include a lack of bridging trabeculae between involved vertebrae, the presence of a radiolucent gap between the graft and the vertebral body, motion between the involved segments of flexion-extension views and the Cobb angle. The purpose of this study is to compare accuracy of two objective radiographic techniques in identifying nonunion following anterior cervical discectomy and fusion (ACDF). METHODS: Twenty-four patients with 26 cervical fusions ranging from 1-3 levels were included in this study. There were 14 females and 10 males with an average age of 48. All patients had a known outcome. The charts were reviewed for original diagnosis, dates of procedure and follow-ups, operative notes past medical and social histories. The mean follow-up was 25 months. Two measurements were obtained from lateral flexion-extension radiographs: 1.) Cobb angles and 2.) Distance between the tips of the spinous process of the operated levels. Data was obtained independently by three physicians: a fulltime spine attending, a spine fellow and a fourth-year orthopaedic resident. A #1 fine-tip lead pencil and the same goniometer were used for all measurements. The operated disc spaces were covered so the measurements were performed in a blinded fashion. RESULTS: The reliability among the observers as measured by Cronbach’s alpha was 0.94 for the spinous process method and 0.57 for the Cobb angle method. A measurement of 2 mm or more between spinous processes was noted in patients with pseudoarthrosis and using the spinous method was 0.43. DISCUSSION/CONCLUSION: The measurement of the distances between the tips of the spinous processes provides an objective method to detect pseudoarthrosis radiographically. Measuring the distances between the tops of the spinous processes is simpler to perform than the Cobb angle method. Those patients with pseudoarthrosis consistently had measurements of the tips of the spinous process greater than a 2mm difference between flexion and extension views. We feel the measurement of distance between spinous processes on lateral flexion-extension radiographs is an objective, accurate method in radiographic diagnosis of pseudoarthrosis. Paper # 60 Inter-observer Agreement in the Assessment of Cervical Spine Fusions Using Flexion-extension X-rays Lane D. Spero, MD, John Hipp, PhD, William Watters, III, MD, Stephen I. Esses, MD (Houston, TX) INTRODUCTION: Flexion-extension x-rays are commonly used in the clinical assessment of cervical spine fusions, and are also used to compare or assess fusion methods in research. The purposes of this study were to measure the inter-observer agreement in assessing motion at cervical spine fusion sites, and to test a technique to improve the agreement. METHODS: Flexion-extension radiographs of 20 anteriorly plated cervical spine fusions were assessed by 9 surgeons. Observers completed a standardized questionnaire to record whether they observed motion at the fusion site. RESULTS: The kappa value was 0.19 indicating poor inter-observer agreement. The same 20 flexion-extension radiographs were digitized and aligned so that the plate remained in the same position. The images were shown to the same 9 observers, and the images were rapidly cycled between the flexion and extension views to simulate motion. The inter-observer agreement was only slightly better (kappa=0.21) with the aligned and animated images. However, despite the low kappa values, there was 100 percent agreement for 7 of the 20 cases using the aligned and animated images, wheras there was 100 percent agreement for only 3 of the 20 cases using the conventional method. Motion at the fuson site was reported in 140 of the 180 recorded observations using the aligned and animated images, whereas motion was reported in only 54 of the 180 observations made using the conventional flexion-extension x-rays. To determine whether the agreement between observers was better when there was a lot of motion, quantitative measurements of intervertebral motion were obtained. There was no correlation between the amount of intervertebral motion and the percent of interobserver agreement. CONCLUSIONS: The results of this investigation raise concerns about the daily clinical use of fexion-extension radiographs to assess cervical spine fusions, and raise concerns about the use of flexion-extension x-rays as a primary outcome variable in research investigations of cervical fusions. Paper # 61 Radiation Exposure Using Intraoperative Fluoroscopy for Anterior Cervical Diskectomy and Fusion Avi J. Bernstein, MD, David L. Spencer, MD (Park Ridge, IL) INTRODUCTION: The use of intraoperative fluoroscopy, in place of standard radiographs in anterior cervical decompression and fusion surgery, has received little attention and appears to be used by a minority of surgeons on a consistent basis. Identification of the appropriate level is not only necessary for a positive outcome, but it is the standard of care. From personal experience, I found that the use of intraoperative fluoroscopy increased surgeon confidence and reduced operating room time. I undertook this study in an effort to determine the risks of radiation exposure using fluoroscopy to the operating surgeon and the surgical team, as well as its impact on efficiency, cost, and surgeon confidence. METHODS: Thirty consecutive patients undergoing anterior cervical diskectomy and fusion for neck pain, radiculopathy, and/or myelopathy were studied. All patients received autologous iliac crest bone graft. All cases employed anterior plate fixation. Intraoperative fluoroscopy was used in all cases. Fluoroscopic images were obtained: (1) after initial patient positioning, (2) for determination of placement for skin incision, (3) for determination of appropriate level of resection, (4) for assessment of provisional plate fixation, (5) and for final reconstruction. Dosimeter badges were worn during all fluoroscopic maneuvers. The test badge was worn by the operating surgeon on the collar, in an unprotected fashion. The second control badge was placed on the waist behind a leaded apron. The badges wee evaluated on a monthly basis and the reading were used to determine radiation exposure. Also recorded were the number of fluoroscopic images per case, seconds of fluoroscopic exposure, the age and gender of the patient, the operative time, the number of levels, and any postoperative complications. RESULTS: The test dosimeter badge demonstrated no significant radiation exposure to the operating surgeon beyond background routine background radiation, relevant to medical personnel. There was no significant difference between the test badge and control badge. The total second of exposure per case averaged approximately 3 seconds. The additional cost for intraoperative fluoroscopy relative to the cost of two plain crosstable radiographs was negligible, and less expensive when the relative time savings is considered. The operative times averaged 80 minutes for a single case, and 105 minutes for double level cases. There were no complications related to the procedure including infection, neurologic compromise, or revision surgery. Fusion rate was 100%. DISCUSSION/CONCLUSION: 1 The risk of radiation exposure to the operating surgeon during anterior cervical diskectomy and fusion surgery, is exceedingly small. 1 The use of intraoperative leaded aprons during fluoroscopy can be avoided at the surgeon’s discretion, and still maintain safe dosimeter rates even when performing large numbers of cases. 1 When used properly, intraoperative fluoroscopy will reduce operative time, increase the surgeon’s confidence and accuracy, and decrease the cost of surgery. An appropriate technique for use of intraoperative fluoroscopy will be demonstrated. Paper # 62 Virtual Fluoroscopy for Cervical Spine Surgery Kevin T. Foley, MD (Memphis, TN), Y. Raja Rampersaud, MD (Memphis, TN), David A. Simon, PhD (Broomfield, CO), Thomas H. Jansen, MSc (Memphis, TN) • (a - Medtronic/Sofamor Danek) INTRODUCTION: The utility of fluoroscopy for cervical spine surgery is variable. As a localizing and navigational tool, fluoroscopy is essential in certain procedures, such as odontoid screw placement. In other procedures, such as anterior cervical plating, its disadvantages typically outweigh its benefits. The purpose of this study was to evaluate the utility of computer-assisted fluoroscopic navigation (virtual fluoroscopy) in the cervical spine. METHODS: Using three fresh cadavers, the following cervical spine procedures were performed with the FluoroNav virtual fluoroscopy system (Medtronic Sofamor Danek, Memphis, TN): 1) odontoid screw placement, 2) lateral mass screw placement bilaterally from C3 to C6, and 3) anterior cervical corpectomy (three-level), with anterior plating. All procedures were done using standard surgical exposures and techniques. Procedural guidance was provided by a computer-enhanced fluoroscopic technique as follows. The head was rigidly secured in a Mayfield apparatus. A dynamic reference arc was rigidly attached to the head holder. True AP and lateral fluoroscopic images of the C-spine were obtained using a standard C-arm (OEC Model 9600, OEC, Salt Lake City, UT) fitted with a calibration target. Each image was automatically calibrated and the C-arm was then removed from the surgical field. The locations of various optically-tracked tools (e.g. probe, drill guide, high-speed drill) were graphically projected onto the previously acquired fluoroscopic images in real-time. Anatomic correlation of the virtual tool position with the actual tool tip was verified by touching various anatomic landmarks. Odontoid screws were placed in each cadaver by simultaneously utilizing virtual (pre-acquired) open mouth and lateral views, providing typical biplanar guidance with a single fluoroscope. The second and third procedures were enhanced by software that quantified the linear and angular relationships of a given tool to a defined plane. Using trajectories similar to those described by Magerl (30º lateral, parallel to facet joints), lateral mass screw placement was carried out bilaterally from C3 to C6 in each cadaver. The pre-acquired fluoroscopic AP and lateral views were superimposed with a real-time graphical overlay of the sagittal drill trajectory relative to the facet joints and the axial drill trajectory in degrees. For the third procedure, an image-guided high-speed drill was used to perform three-level corpectomies (16 mm wide) on each specimen (C3-C6) followed by placement of a spanning anterior plate. Using pre-acquired AP and lateral fluoroscopic images and a virtual tool overlay, simultaneous feedback was provided with respect to width of the corpectomy trough (i.e. tool tip position from the midline), depth of the drill tip relative to the posterior vertebral body line, and the angulation of the drill relative to the mid-sagittal plane. The same previously acquired images were utilized to ensure proper AP and lateral placement of the plate and screws. Subjective assessment of the virtual fluoroscopic system in performing the given tasks was noted. Post-procedure, the cervical spines were assessed with radiographs (AP and lateral images) and CT scans. The CT scans were loaded into an image-guided surgery workstation. Using a digital measuring tool as well as a standard goniometer, linear and angular measurements were made. RESULTS: The virtual fluoroscopy system provided multiplanar procedural guidance with previously acquired AP and lateral fluoroscopic views, allowing for removal of the C-arm unit from the operative field during surgery. Anatomic correlation of the real and virtual tool tips was excellent. All three odontoid screws were placed properly on the post-procedure AP and lateral films as well as on the axial CT images. The average lateral angular deviation of the 24 lateral mass screws relative to the mid-sagittal plane was 30.5º (range: 27º-35º). The average corpectomy trough width was 16.8 mm. The troughs were symmetric with regard to the midline and their sidewalls were parallel to the mid-sagittal plane. All three anterior cervical plates were midline. DISCUSSION/CONCLUSIONS: Although conventional fluoroscopy can be useful in cervical spine surgery, it has several limitations. The difficulties inherent in odontoid screw fixation, where two C-arm units must be positioned for simultaneous AP and lateral fluoroscopic guidance, illustrate this point. Virtual fluoroscopy offers several advantages. Multiplanar guidance is possible using a single C-arm. Tool position can be updated in real-time without acquiring new fluoroscopic images, minimizing radiation exposure and the ergonomic challenges of C-arm use. Real-time quantitative information can be provided, such as tool tip distance from the posterior vertebral body line or anatomic midline, and drill guide angulation in the axial plane relative to the mid-sagittal plane. This type of quantitative information may improve upon inter-surgeon variability in the performance of a geometric surgical task. We conclude that computer-assisted fluoroscopic navigation is feasible in the cervical spine, has several advantages over conventional fluoroscopy, and seems to have acceptable accuracy. • If noted, the author indicates something of value received. The codes are identified as: a - research or institutional support, b - miscellaneous funding, c - royalties, d - stock options, e - consultant or employee. For full information, refer to inside back cover. Paper # 63 Clinical and Radiological Evaluation of the Codman Anterior Cervical Plate: Results of a Prospective, Multicenter Study with Two Year Followup Michael G. Fehlings, MD, PhD (Toronto, ON, Canada), Dieter-Karsten Boker, MD (Giessen, Germany), Charles Branch, MD (Winston-Salem, NC), Fred Geisler, MD (Chicago, Ill), Regis Haid, MD (Atlanta, GA), Robin Johnston, MD (Glasgow UK), Iain Kalfas, MD (Cleveland, OH), Christopher Paramore, MD (Birmingham, Al), Gerald Rodts, MD (Atlanta, GA), Volker Sonntag, MD (Phoenix, AZ), William Taylor, MD (Glasgow UK), Vincent Traynelis, MD (Iowa City, IA), Michael Winking, MD (Giessen, Germany) • (a – Codman/Johnson and Johnson Professional, Inc.) INTRODUCTION: There is increasing interest in the application of anterior cervical fixation systems which are semi-constrained and allow greater load sharing between the cervical spine and the bone graft. Such systems theoretically should promote a higher rate of bone fusion, but have the potential disadvantage of providing less rigid fixation. In order to address these issues, we conducted a multicenter, prospective clinical and radiological assessment of the Codman anterior cervical plate, an example of a semi-constrained system, in 194 patients, with independent, blinded evaluation at a follow-up of 2 years. METHODS: A total of 194 patients (118 M, 76F; age range: 16-85) undergoing anterior cervical reconstruction at 10 centers were enrolled in the study from 10/95-09/96. Clinical and radiological data were collected at defined intervals up to 2 years postoperatively according to a specified protocol. All radiographic studies were quantitatively assessed by an independent, blinded observer at the coordinating site (University of Toronto). RESULTS: Two patients were excluded from the study due to insufficient data. Indications for anterior cervical reconstruction included degenerative disease (82%), trauma (15%), neoplasia (2%) and rheumatoid arthritis (1%). Autograft was used in 60% of cases with allograft being used in the remainder (40%). Most cases involved multilevel reconstruction (66%) with 32% of cases involving 3 or more segments of the cervical spine. Satisfactory construct stability occurred in 95.7% of cases. Definite radiographic fusion occurred in 90.5% of cases. A total of 8 cases (4.3%) required revision with the most common indication being graft extrusion or screw pullout (n=4). Hardware failure (screw fracture, partial or complete screw pullout) occurred in 11 cases (5.7%). No significant change in the cephalad screw plate angle occurred from the time of insertion up to 2 years of follow up. The caudal screw plate angle changed from -3.4 (initial) to 4.5 (6 months) degrees with no further change occurring between 6 months and 2 years. DISCUSSION/CONCLUSION: The Codman anterior cervical plate was associated with successful construct stability in 95.7% of cases. Changes in screw plate angle occurred at the caudal site over the first 6 months of followup and remained stable after that. These data suggest that the semi-constrained design of the Codman anterior cervical plate provides excellent clinical results with the potential of greater load sharing between the bone graft and the reconstructed anterior cervical spine. • If noted, the author indicates something of value received. The codes are identified as: a - research or institutional support, b - miscellaneous funding, c - royalties, d - stock options, e - consultant or employee. For full information, refer to inside back cover. Paper # 64 Clinical Experience With A New Load-Sharing Anterior Cervical Plate Ronald I. Apfelbaum, MD (Salt Lake City, UT), Andrew T. Dailey, MD (Salt Lake City, UT), Jose Barbera, MD (Valencia, Spain) • (e - Aesculap Instrument Corp, Drs. Apfelbaum & Barbera) INTRODUCTION: Anterior plate fixation is widely accepted as an adjunct to surgical correction of traumatic or degenerative conditions of the cervical spine. The various systems have evolved from nonconstrained constructs to rigid plate screw constructs. With the popularity of the latter, concerns are rising on both the biomechanical and clinical levels about the adverse effects of graft stress shielding. A new plating system (ABC plate) has been developed which seeks to avoid stress shielding of the graft by allowing vertical migration of the fixation screws within the plate while preventing screw back-out and still effectively stabilizing the spine to restore or preserve lordosis. METHODS: This study reports on our experience with 149 patients at two different centers utilizing the ABC system. Indications for surgery were herniated disks in 43%, degenerative disease in 38%, trauma in 13%, deformity and failed fusion in 2% each, and tumors and OPLL in 1% each. Patients ranged in age from 18 to 76 years (average 46 yrs.). Unicortical screw placement was used in 55% and bicortical in 45% according to the surgeon's preference and judgment. Single level plates were used in 43%, two level plates in 44%, and three or more levels were plated in 13%. Corpectomies were used in 24% with the remainder being interbody fusions. Allograft was used in 56%, autograft in 44%. Patients are seen and follow-up radiographs, which include lateral flexion/extension views, are taken at the 1, 3, 6 and 12 month intervals after surgery. Settling was measured and corrected for magnification. Cobb angles were measured to determine changes in angulation from immediately preoperatively to subsequent interval films. Fusion was determined by bridging trabecular bone and the absence of motion at the tips of the spinous processes on flexion and extension films. RESULTS: Minimum follow-up of 3 months was available on 93 patients, involving 174 instrumented levels. SETTLING: Settling occurred in most patients. It was more pronounced with the use of allograft, but averaged 1 to 2 mm/level overall with a maximum of 6 mm in one patient with a three level fusion. Ninety-five percent of the settling occurred in the first month, at which time the grafts often were often indistinguishable from the three month grafts in rigidly plated patients. No settling occurred after 3 months. FUSION: Virtually all levels showed impressive early graft incorporation. Using the strict criteria defined above, 122 of 174 levels (70%) were fused at 3 months and 103 of 120 levels (86%) at 6 months. No translational instability was observed. KYPHOTIC ANGULATION: Lordosis was preserved unchanged in 141 patients, while there was a 5° loss of lordosis in 2 patients. Kyphosis in 6 patients was maintained within 5° of the postoperative correction. There were no hardware-related complications and no patients had a painful nonunion. CONCLUSIONS: This constrained true load sharing system appears to attain the goals of stabilizing the spine after surgery or trauma while promoting earlier and more substantial graft incorporation. Since no negative downside has been identified, the use of this type of anterior cervical plating system would seem to confer a significant benefit to the patient. • If noted, the author indicates something of value received. The codes are identified as: a - research or institutional support, b - miscellaneous funding, c - royalties, d - stock options, e - consultant or employee. For full information, refer to inside back cover. Paper # 65 Dynamic Cervical Plates: Do They Load-Share at the Expense of Stability? Darrel Brodke, MD, Sohrab Gollogly, MD, Alexander Mohr, BS, Andrew T. Dailey, MD, Kent N. Bachus, PhD (Salt Lake City, UT) INTRODUCTION: Single or multi-level corpectomy with anterior strut grafting is a common surgical technique for correcting degenerative and traumatic conditions of the cervical spine. Many different styles of anterior plates have been designed to augment the initial stability and improve the fusion rate of corpectomy reconstructions. Recently, dynamic cervical plates have been introduced that allow for axial settling in order to increase load sharing with the graft and minimize device related osteopenia, however, there is some concern regarding the initial stability of these plates. We evaluated the load sharing properties and stiffness in flexion and extension, lateral bending, and axial torsion, of the Depuy-Acromed DOC( and Aesculap ABC( dynamic cervical plates and compared them with the Synthes CSLP( and Sofamor-Danek Orion( locked cervical plates in a corpectomy reconstruction model. This model was intended to simulate initial surgical conditions, followed by graft subsidence or resorption with a 10% shorter graft, and complete graft failure or expulsion when the graft was removed. MATERIALS AND METHODS: A simulated corpectomy reconstruction model using UHMWPE bodies was inserted into a custom 4-axis spine simulator, designed to apply compressive axial loading to the specimen, as well as pure flexion/extension, lateral bending, and axial torsion moments. Six constructs of each type were first precision drilled by a machinist in accordance with the plate manufacture’s recommendations for screw geometry and spacing, the plates were implanted on the blocks, and the constructs were tested with a 30mm graft in place. The load sharing properties of the plates were calculated by measuring the load borne by the graft as a percentage of the applied axial load of 15 to 120N, using the Tekscan( system. The stiffness of each construct was then calculated in flexion/extension, lateral bending, and axial torsion by applying ±2.5Nm moments in one plane, while maintaining the others at zero moment, and measuring displacement with the OptoTrak( system. These measurements were then repeated using a 27mm graft to simulate 10% graft subsidence or resorption. The two dynamic plates were then compressed until they reached their limits for axial settling and the stiffness of all the plates in all three axes was then measured with no graft in place in order to simulate graft failure or expulsion. RESULTS: With a 30mm graft in place, there were no significant differences between the Orion, CSLP, and ABC plates with respect to flexion/extension stiffness, while the DOC was significantly less stiff. When the graft was shortened by 10% to 27mm, the ABC plate retained its initial stiffness significantly more than the other three plates, which became essentially equal. With removal of the graft the stiffness of all constructs was diminished (figure 1). Fig. 1: Flexion/Extension Stiffness In lateral bending with a 30mm graft, the Orion, CSLP, and ABC plates were not significantly different, while the DOC plate was significantly less stiff. With a 27mm graft the Orion and CSLP retained the majority of their initial rigidity, while the ABC plate lost some initial rigidity due to the ability of the screw heads to slide in the elongated screw holes when lateral loads were applied. Again the DOC plate was significantly less stiff than the other three. This pattern was repeated when the graft was removed. In axial torsion, the four plates were significantly different, the Orion being the stiffest, followed by ABC, CSLP, and DOC plates. This pattern was essentially unchanged when testing was repeated with a 27mm graft and without a graft. Analysis of load sharing data reveals that all four plates shared applied axial loads with the 30mm graft with a range of 51-95%. In general, the stiffer plates carried a greater proportion of the applied axial loads and the dynamic plates load shared more effectively. When the graft was shortened by 10%, only the two dynamic cervical plates shared a significant proportion of the applied axial load to the 27mm graft over the entire 15 to 120N range. The locked cervical plates did not allow the graft to share any portion of the axial load until at least a 90N load was applied, and they allowed a maximum of approximately 20% of the applied load to be transmitted through the graft at 120N. Fig. 2: Lateral Bending Stiffness Fig. 3: Axial Torsion Stiffness CONCLUSIONS: There is a significant debate in the orthopaedic and neurosurgical literature about the use of instrumentation to augment anterior cervical discectomy and corpectomy reconstruction and fusions. Those who argue in favor of the use of anterior plates cite greater initial stability, increased fusion rates, lesser requirements for additional orthotic use, and an earlier return to work with decreased indirect costs to the patient. However, there are reports of hardware associated complications, increased operative times and direct costs with the use of instrumentation, and concerns about the stress shielding properties of locked cervical plates and the risk of device related osteopenia and pseudoarthrosis. There are scant published data on the comparative biomechanical properties of the various styles of plates that are currently available and the load sharing properties of dynamic plates as compared to locked plates has not been evaluated. By using a simulated corpectomy reconstruction model made of UHMWPE, we were able to eliminate the inconsistencies associated with cadaveric spines and isolate the differences between four types of anterior plates. Notably, the two dynamic plates tested were able to effectively share applied physiologic axial loads with a graft under conditions simulating the initial reconstruction, and also with a subsequent 10% subsidence or resorption of the graft. In contrast, the locked cervical plates did not share any portion of the applied axial load with the shorter graft until at least 90N was applied to the model. With respect to rigidity, the locked cervical plates and the ABC dynamic plate were relatively similar in terms of their initial stiffness in flexion and extension, lateral bending and axial rotation. The DOC plate was consistently the least stiff of all the constructs in all modes of testing. Simulated graft subsidence or loss changed the flexion and extension and lateral bending properties of the all plates with little effect on axial torsion stiffness. In summary, the ABC plate was able to dynamically settle along the vertical axis and load share with a graft more effectively than locked cervical plates and without a significant difference in initial stability. In contrast, the DOC dynamic plate was also able to effectively load share, but it did so at the expense of stability. Basic Science Award Human Cervical Intervertebral Disc Cells are Susceptible to AdenovirusMediated Gene Therapy. Seong-Hwan Moon, MD, James D. Kang, MD, Kotaro Nishida, MD, Lars G. Gilbertson, PhD, Christopher Niyibizi, PhD, Patrick N. Smith, MD, Mark A. Knaub, MD, Paul D. Robbins, PhD, Christopher H. Evans, PhD (Pittsburgh, PA) INTRODUCTION: Intervertebral disc degeneration of the cervical spine and its related clinical disorders (i.e., degenerative spondylosis, cervical radiculopathy, and myelopathy) can result in substantial patient morbidity and is associated with increased health care costs. The available technology for the treatment of intervertebral disc degeneration is generally invasive and usually requires surgery in which all or a portion of the affected intervertebral disc is removed necessitating further procedures to re-stabilize the segment (i.e., fusion). Disc degeneration is characterized in part by a progressive decrease in proteoglycan content leading to dehydration of the nucleus pulposus. Although growth factors (such as TGF-beta1, IGF-1, and OP-1) appear to have promising therapeutic properties such as stimulating proteoglycan synthesis1, 2, 3, there is no currently practical method for sustained delivery of exogenous growth factors to the disc for chronic types of disease such as disc degeneration. An alternative possibility is genetic modification of disc cells through gene transfer such that the cells manufacture the desired growth factors endogenously, on a continuous basis. In previous studies, we successfully used an in-vivo adenovirus-mediated gene transfer technique to deliver exogenous marker genes and therapeutic genes to rabbit lumbar intervertebral discs4, 5. Before contemplating extending these approaches to a clinical setting, it is necessary to determine whether human cervical disc cells are susceptible to adenovirus-mediated gene transfer and biologically modifiable by therapeutic gene transfer. Accordingly, the objectives of this in-vitro study were (1) to test the efficacy of the adenovirus-mediated gene transfer technique for transferring exogenous marker genes (lacZ, luciferase) to the cultured human disc cells, and (2) to determine whether the disc cells can be biologically modulated by gene transfer of therapeutic gene (TGF-beta1). METHODS: All of the experimental protocols were approved by the human subjects Institutional Review Board at the University of Pittsburgh. Cervical intervertebral disc tissue was obtained from twelve patients undergoing anterior decompression and fusion surgeries for cervical radioculopathy or myelopathy. Degenerative status of the intervertebral disc of each patient was assessed based on magnetic resonance imaging. An attempt was made by the operating surgeon (JDK) to carefully obtain tissue from the central aspect of the disc to optimize harvest of only the nucleus pulposus and the inner annulus fibrosus. The sequential enzymatic isolation of disc cells was performed as described elsewhere6. In Phase I Marker gene studies, two different adenoviral constructs were prepared: Ad/LacZ (encoding E coli beta-galactosidase) and Ad/luciferase (encoding firefly luciferase). At full confluence of the cultures, the cultures were organized into three groups: (1) lacZ gene group, (2) luciferase gene group, (3) control group. In Group 1, 50 microliter of normal saline with six different doses of adenovirus (Ad/lacZ) were added to the culture wells to determine optimal multiplicity of infection (MOI) giving a 100% transduction rate. The six doses were 7.5x106 plaque forming unit (PFU), MOI = 15; 1.5x107 PFU, MOI = 30; 3x107 PFU, MOI = 60, 5x107 PFU, MOI = 100; 7.5x107 PFU, MOI = 150; and 1.5x108 PFU, MOI = 300. In Group 2, 50 microliter of the Ad/luciferase at an MOI of 150 were added to the culture wells, enabling quantification of transduced gene expression. In Group 3, 50 microliter of normal saline without adenovirus was added to the culture wells. All cultures were incubated at 37°C for one hour with gentle agitation. F12/DMEM (950 microliter) with 10% fetal bovine serum was added to each well, the cells were further incubated at 37°C. Gene expression was assessed at one and three weeks, by 5-bromo-4-chloro-3-indolyl-beta-galactosidase (X-Gal) staining technique for lacZ gene expression and enzyme activity assay for luciferase gene expression. Phase II Therapeutic gene (TGF-beta1) study: We used Ad/TGF-beta1 with MOI of 75. The cultures were organized into four groups, Group 1: as a saline control, Group 2: same as Group 1 except administration of exogenous hTGFbeta1 at 2ng/ml of medium, Group 3: as a viral control, transduction with Ad/luciferase in monolayer culture, Group 4: same as Group 3 except replacement of virus with Ad/hTGF-beta1. All cultures were incubated for 2 day after transduction in serum free condition. TGF-beta1 concentrations in each group were measured by enzymatic-linked immunosorbent assay (ELISA). Newly synthesized proteoglycan was assessed using S35-sulfate incorporation using chromatography on Sephadex G-25 in PD-10 columns and scintillation count was normalized by cell number. Collagen and non collagen synthesis were measured with H3-proline incorporation followed by collagenase digestion. Oneway analysis of variance with Fisher's protected LSD post-hoc test was conducted to compare newly synthesized proteoglycan. RESULTS: Phase I: The adenovirus-lacZ construct at an MOI of 15, 30, 60, and 100 transduced approximately 5%, 20%, 40%, and 70% of the cells respectively. The adenovirus-lacZ construct at an MOI of 150 rendered nearly 100% of the cultured cells of Group 1 (lacZ gene group) lacZ+ without evidence of cytotoxicity (Figure 1). Reporter gene expression persisted up to at least 3 weeks after transduction. There was no observable decrease in expression at three weeks compared to one week. The other groups exhibited negative X-Gal staining. There were no differences in transduction rates between the cells harvested from different grades of degeneration. In Group 2 (luciferase gene group), the adenovirus-luciferase construct at an MOI of 150 resulted in a significant increase in the activity of luciferase over that of the other groups (p<.05). The luciferase activity at three weeks showed no statistically significant decrease from that of one week. Phase II: The concentration of TGF-beta1 in exogenous TGF-beta1 applied group (Group 2) was 0.56ng/ml, and in Ad/TGFbeta1(Group 4) was 3.12ng/ml on day 2. There was statistically significant three-fold increase in newly synthesized proteoglycan in Ad/TGF-beta1 treated cells comparing control (p<.05) The exogenous TGF-beta1 treated cells showed significant increase in newly synthesized proteoglycan (p<.05). Both Ad/TGFbeta1 treated cells and exogenous TGF-beta1 treated cells showed 350% increase in collagen synthesis and 250% increase in non collagen synthesis, compared with control. DISCUSSION: The many steps important for successful gene therapy include a clear understanding of the pathogenesis of disease, an appropriate target tissue for gene therapy, effective therapeutic gene(s), and an animal model that closely simulates disease for preclinical testing. The rate-limiting step for successful gene therapy, however, is the ability to transfer efficiently the appropriate therapeutic gene to the target tissue. Previously, we have shown that adenovirus-mediated gene transfer to the rabbit intervertebral disc resulted not only in efficient and prolonged gene expression4 but also in modulation of biologic activity of the intervertebral disc5. In the current study, we achieved 100% transduction rate of human cervical disc cells using an adenovirus-mediated approach-even in cells harvested from clinically degenerated discs. Furthermore, we showed more effective upregulation of proteoglycan synthesis by gene therapy than by application of exogenous TGFbeta1. These preliminary results, together with those of our previous rabbit studies, add growing support to the notion that intradiscal gene therapy may have potentially clinically useful applications in the prevention and treatment of degenerative disc disease of the spine. REFERENCES: 1. Thompson JP et al. Spine 16(3): 253-260, 1991 2. Osada R et al. JOR 14:690-699, 1996 3. Takegami K et al. 45th Annual Meeting, Trans Orthop Res Soc, 201, 1999 4. Nishida K et al. Spine 23(22): 2437-2442, 1998 5. Nishida K et al. Spine (In press), 1999 6. Chelberg MK et al. J Anat 186:43-53, 1995 Figure 1. Cultured human cervical disc cells 3 weeks after transduction by AdlacZ in vitro. Adenovirus construct at an MOI of 150 rendered nearly 100% of the cells lacZ+ (original magnification: x100). Figure 2. Significant three-fold increase of newly synthesized proteoglycan in Ad/TGF-beta1 Group over control groups (*: p< .05), significant two-fold increase in exogenous TGF-beta1 Group from that of control (p< .05). ACKNOWLEDGEMENT: The authors thank Dr. Savio L-Y Woo and Dr. Freddie H. Fu, and the Musculoskeletal Research Center for generous guidance and support of this study. J. William Fielding Resident-Fellow Award Oligodendroglial Apoptosis Occurs along Degenerating Axons and is Associated with Fas and P75 Expression following Spinal Cord Injury Steven Casha, MD, Wen Ru Yu, MD, Michael G. Fehlings, MD (Toronto, Ontario, Canada) INTRODUCTION: Recent studies have implicated apoptotic cell death in the pathophysiology of spinal cord injury (SCI). The association of apoptosis with a cell death program presents many new potential therapeutic targets. This study aims to characterize the temporal and spatial distribution of apoptosis following cervical SCI using the clip compression model; to identify the cell populations undergoing apoptosis; to examine their relationship to degenerating axons (as indicated by the interruption of axonal beta-amyloid precursor protein (BAPP) transport); and to examine the potential role of FAS and P75 in initiating post traumatic apoptosis in the spinal cord. METHODS: SCI was performed in a rat model using the calibrated clip compression model at C7-T1. Tissue was examined by agarose electrophoresis to demonstrate DNA laddering, in situ terminal-deoxy-transferase mediated dUTP nick end labeling (TUNEL), and electron microscopy to accumulate several lines of evidence that apoptosis occurs following SCI. TUNEL histology was quantified to examine the temporal and spatial distribution of apoptotic cells in the spinal cord following injury. Double labeling immunohistochemistry for cell specific markers (NF200 for neurons, GFAP for astrocytes, OX42 for microglia and CNPase for oligodendrocytes) with TUNEL was performed to identify the cells undergoing post-traumatic apoptosis. Double labeling immunohistochemistry using anti-BAPP antiserum and TUNEL was used to examine the relationship of apoptotic cells to degenerating axons. Double labeling immunohistochemistry using anti-Fas and anti-p75 antisera with TUNEL was used to determine the potential role of these "death receptors" (TNF receptor gene family members) in initiating post-traumatic apoptosis. Changes in expression of these receptors were also examined using western blotting. RESULTS: DNA agarose electrophoresis (DNA laddering), TUNEL and electron microscopy all provided evidence that apoptotic cell death occurs following SCI. TUNEL revealed apoptotic glia in grey and white matter following SCI. These were largely oligodendrocytes as identified by cell specific markers (CNPase). Double labeling also identified apoptosis in microglia (OX42 positive) and few astrocytes (GFAP positive). No apoptotic neurons (NF200 positive) were identified. Counting of TUNEL cells revealed that they were uniformly distributed in the injured and adjacent segments, increasing in numbers during the first week and declining beyond seven days. BAPP has been shown to accumulate in degenerating axons due to interruption of axonal transport, and has been used as a molecular marker of axonal degeneration. Degeneration of axons, as determined by BAPP positivity, was temporally and spatially colocalized with TUNEL-positive glia on double labeling immunohistochemistry. FAS and P75 protein expression was seen in astrocytes, oligodendrocytes and microglia. Fas expression increased during the first week following injury and declined beyond. P75 expression increased beyond the first week. Fas and P75 positive TUNEL positive cells were identified. Fas positive apoptotic cells were more likely found at day 3 and day 7 post injury while P75 positive apoptotic cells were more likely found at day 14. DISCUSSION: We conclude that delayed axonal degeneration after traumatic SCI is associated with glial, in particular oligodendroglial, apoptosis and that the death receptors FAS and P75 may be involved in initiating this process. Molecularly targeted therapies directed at minimizing glial apoptosis may be a clinically useful strategy to prevent white matter degeneration after traumatic CNS injury. Our observation that apoptosis occurs up to 2 weeks after SCI implies that neuroprotective strategies may be useful for several days after injury in contrast to current therapies which are limited to the first 48hrs. Death receptors are capable of activating specific caspase cascades, which result in the apoptotic demise of a cell. These specific proteases may present specific targets to which some semi-selective antagonists are already available. Our data implicates death receptors and thus the caspases they activate in post-traumatic apoptosis following SCI. Clinical Award Subtotal Corpectomy Versus Laminoplasty For Multilevel Cervical Spondylotic Myelopathy: A Long-Term Follow-Up Study Over Ten Years Eiji Wada, MD (Suita, Osaka, Japan), Shozo Suzuki, MD, Kazuo Yonenobu, MD, Keiro Ono, MD, (Osaka , Japan) INTRODUCTION: There has been controversy about which surgical procedure should be indicated for multilevel cervical spondylotic myelopathy, subtotal corpectomy or laminoplasty. To answer this question, we had changed our bias on therapeutic strategy from subtotal corpectomy to laminoplasty in 1984. Patients treated with subtotal corpectomy or laminoplasty are comparable with each other regarding clinical backgrounds which affect surgical outcomes and surgical techniques. The purpose of this study is to clarify the long-term outcomes of subtotal corpectomy and laminoplasty and offer basic data in surgical planning for multilevel cervical spondylotic myelopathy. PATIENTS AND METHOD: Twenty-three patients treated with subtotal corpectomy and 24 with laminoplasty were followed up at least ten years after surgery. Prognostic factors including age at surgery, duration of symptoms, number of blocks on myelogram, the severity of myelopathy before surgery, anteroposterior canal diameter and transverse area of the spinal cord were identical between the corpectomy and the laminoplasty groups. The severity of myelopathy was assessed with the Japanese Orthopaedic Association(JOA) score. The JOA score was evaluated before and after surgery(at one-year, 5-year, and final visit), Lateral radiographs in neutral, flexion and extension were taken periodically. Axial pain was graded by modifying the criteria of Robinson(none, mild, moderate or severe). Neurological recovery, late deterioration, axial pain, radiographic results(degenerative changes at adjacent levels, alignment and range of motion of the cervical spine) and surgical complications were compared between the two groups. RESULTS: The averaged JOA score was 7.9 points before surgery, 13.3 points at one year follow-up, 13.9 points at five year's and 13.4 points at the final visit in the corpectomy group. The averaged JOA score in the laminoplasty group was 7.4 points before surgery, 13.1 points at one year follow-up, 12.9 points at five year's and 12.2 points at the final visit . There was no statistically significant difference between the two groups at each follow-up visit(Table 1). One patients in the corpectomy group needed an additional laminoplasty for the treatment of relapsed myelopathy secondary to the adjacent degeneration. One patient in the laminoplasty group deteriorated neurologically after hyperextension injury. Severe or moderate axial pain was observed in 15 % of the corpectomy group and in 40 % of the laminoplasty group at the latest follow-up. Patients treated with laminoplasty manifested axial neck pain more frequently than those with corpectomy(The Mann-Whitney U test, p<0.05)(Table2).In the corpectomy group, listhesis over 2mm developed in 38% of the upper adjacent levels and disc space narrowing with osteophytes was found in 54 % of the lower adjacent levels. In the laminoplasty group, listhesis over 2 mm was seen in sixteen patients(twenty-two discs) before surgery. At the final follow-up, it was unchanged in eight patients(twelve discs) and improved in the other eight patients(ten discs). No segmental instability developed after laminoplasty. From the radiographic evaluation, the sagittal alignment was deteriorated in two patients(straight:1 and kyphosis:1) of the corpectomy group and in three patients(straight:2 and kyphosis:1) of the laminoplasty. There was no statistically significant difference between the two groups in the prevalence of kyphotic deformity after surgery. The averaged range of motion was reduced to 49 % of the preoperative value in the corpectomy group and to 29% in the laminoplasty group, respectively. Lifted lanimae were frequently fused in the laminoplasty group with an average of 2.5 segments. Spontaneous fusion at C2/3 level was also observed in 40 % of the laminoplasty group. Patients with laminoplasty had lost their range of cervical motion more frequently than those with corpectomy(The Mann-Whitney U test, p<0.05). Eight patients of the corpectomy group encountered surgical complications, including seven graft complications(two dislodgement, one fracture and six pseudoarthrosis) and one esophageal fistula. The patients with pseudoarthrosis needed posterior interspinous wiring. There was a significance correlation between pseudoarthrosis and number of fused segments(Chi-squared test:p=.0015).Two of these graft complications were associated with neurologic symptoms(one; myelopathy and one C5 root palsy). In the laminoplasty group, four patients suffered from C5 root palsy which recovered in the course of time. Averaged operative time and blood loss in the corpectomy group were 264 minutes and 986 grams, respectively. These values in the laminoplasty group were 182 minutes and 608 grams, respectively. The corpectomy group needed longer operative time with a statistical significance(Mann-Whitney U test;p<0.001) and tended to have more blood loss(Mann-Whitney U test;p=0.24). DISCUSSION/CONCLUSION: Subtotal corpectomy and laminoplasty revealed an identical effect as a surgical treatment for multilevel cervical spondylotic myelopathy. These neurological recoveries usually last over ten years. In the subtotal corpectomy group, the disadvantages were longer surgical time, more blood loss and pseudoarthrosis. In the laminoplasty group, axial pain manifested frequently and range of motion reduced severely. Table 1. Preoperative and Postoperative JOA Scores Table 2. Summary of Disadvantages Paper # 66 Cervicomedullary Angle in Pre- and Postoperative MR Imaging: Evaluation of the Angular Deformity at the Craniocervical Junction Takashige Takada, MD, Kuniyoshi Abumi, MD, Yasuhiro Shono, MD, Kiyoshi Kaneda, MD (Sapporo, Japan) INTRODUCTION: The purpose of this report is to evaluate the angular deformity at the craniocervical junction in the pre- and postoperative MR imaging using a cervicomedullary angle in occipitocervical disorders. METHODS: 25 patients with lesions at the occipitocervical junction were treated by occipitocervical fixation using cervical pedicle screws and O-C platerod system. Occipitocervical lesions included in this series were RA in 21 patients, os odontoideum in two and others in two. MR imaging was performed in all patients at before and after surgery. The angle between the lines on the ventral side of the cervical spinal cord and the medulla oblongata on MR imaging was determined as the cervicomedullary angle. Also lateral cervical radiographs were obtained for all patients, and atlantoaxial angle, O-C2 angle and McRae value were measured. On the other hand, MR imaging was performed in 50 adults (male: 26, female: 24. Avg. age 49.3 yrs.) without cervical disorders to serve as normal controls. RESULTS: The normal value of cervicomedullary angle was 163.0+/-5.4 degrees. In patients who had occipitocervical lesions, preoperative value was 136.7+/-8.7 degrees. After surgery, malalignment of the occipito-atlanto-axial region was corrected and compression of the medulla oblongata was disappeared or reduced. Postoperative value was improved to 157.2+/-5.5 degrees. In 15 of 25 patients, the atlantoaxial angle, O-C2 angle and McRae value were difficult to determine, due to gross bone erosion of the dens, dens abnormalities and atlantoaxial spontaneous fusion. However, bone abnormalities and erosion did not affect measurement of the cervicomedullary angle using MR imagings. DISCUSSION/CONCLUSION: Lateral cervical radiographs were used to evaluate the instability and subluxation of craniocervical junction using atlantodental distance, McRae value, McGregor value, Chamberlain value and so on. However, in cases of bone erosion of the dens, abnormalities and atlantoaxial spontaneous fusion, it is often difficult to evaluate exact radiographic pathologies of the upper cervical spine. Although the posterior atlantoaxial interval has been a most sensitive indicator to predict paralysis, soft tissue lesions including rheumatoid pannus and inflammatory tissues can not be evaluated by this method. MR imaging is very useful not only for depicting the bony abnormalities in the cervical spine, but also can directly show the compression of the spinal cord and brain stem. Soft tissue changes are clearly demonstrated by MR imaging including distortion of normal ligaments and bursae around the dens, particularly in rheumatoid arthritis. The cervicomedullary angle is a useful method to evaluate and diagnose the patients with compressive cervical myelopathy. Angular deformity at the craniocervical junction and upward migration of the odontoid process are the main causes of neurologic symptoms in patients with upper cervical disorders. Therefore, realignment of the craniocervical junction by application of extension and distraction force on O-C2 using pedicle screw fixation improves cervicomedullary angle and reduces the anterior compression of the spinal cord. Thus,O-C2 fusion using pedicle screw fixation allows decompression at the craniocervical junction by posterior realignment which enables to obviate the anterior decompression. Paper # 67 Occipitocervical Fusion for the Cervical Spine Lesions in Rheumatoid Arthritis Atsunori Kanazawa, MD, Kazuo Yonenobu, MD, Eiji Wada, MD, Takashi Matsuoka, MD, Keiju Fujiwara, MD, Takahiro Ochi, MD (Osaka, Japan) INTRODUCTION: The natural course of rheumatoid arthritis (RA) is known to uniform. The natural course, however, can be classified into three types according to number of joints involved at ten-year affection of RA (Ochi, 1988). The severity of cervical lesions was reported to parallel the severity of the rheumatoid arthritis itself, and cervical spine lesions were found to pursue different natural courses depend on the type of RA. Therefore, the severity of arthritis should be taken into consideration when surgical results are evaluated. No study on surgical results from this point of view has been done. The purpose of this study is to evaluate the results of occipitocervical fusion (OCF) in rheumatoid arthritis in terms of the disease severity. METHODS: This is a retrospective study of twenty-nine patients (3 men and 26 women) with classic or definite rheumatoid arthritis, who underwent the OCF for the cervical lesions. The OCF consisted of rectangular rod with sublaminar wiring and bone grafting from the occiput down to C2, sometimes C3. Indications of OCF were progressive neurological symptoms and/or intractable pain. The mean age at the operation was 60.0 years (range, 47-60 ys). The average period of follow-up was 89 months (range, 1-195 mos). The mean duration of the disease at the operation was 173.8 months (range, 34-362 mos). Surgical results were assessed by Ranawat's classification of neurological defecit. Radiological results were evaluated on flexion and extension lateral cervical radiographs. Regarding the disease severity of rheumatoid arthritis, patients were classified into three subsets based on the Ochi's criteria of rheumatoid arthritis 1). Three subsets were created based on the number of joints with erosion or destruction on radiographs as follows: those with least erosive disease (LES), those with more erosive disease (MES), and those with mutilating disease (MUD). In terms of subsets, the patients were evaluated with respect to clinical symptoms, and the cumulative survival rate according to the Kaplan-Meier method. The radiological findings were also evaluated in patients who could be followed more than 3 years after surgery. There were 14 with MES, and 15 with MUD. There were no patients with LES. With respect to the cervical subluxations, seven of 14 patients with MES had irreducible atlantoaxial subluxation (AAS), 4 had vertical subluxation of dense of axis (VS), 2 had VS combined with subaxial subluxation (SS), and 1 had SS before surgery. Six of 15 patients with MUD had irreducible AAS, 7 had VS, and 2 had VS combined with SS before surgery. RESULTS: Results in MES: Occipitocervical pain improved in 11 (78.6%) of 14 patients. More than one-level improvement in Ranawat's classification of neurological deficit was seen in 4 (50.5 %) of 8 patients who had neurological symptoms preoperatively. However, 2 (50.0 %) of these 4 patients lost surgical improvement during the follow-up period, because of development of subaxial lesion in one patient and of progressive joint destruction of lower extremities in another. Pain relief was maintained in 11 (78.6 %). In twelve patients out of 14 who were able to be followed up more than 3 years after operation, no pseudarthrosis was found in all patients. However subaxial subluxation adjacent to the fusion segments developed in 3 (25.0 %) patients. The cumulative survival rate in 5 years after surgery was 70.7 %. Results in MUD: Occipitocervical pain improved in all patients except one. More than one-level improvement in Ranawat's classification was achieved in 3 (30.0 %) of 10 patients who had neurological deficit preoperatively. However, 2 (66.6 %) of these 3 patients lost surgical improvement during the follow-up period, because of development or progression of subaxial lesion in one patients and of thoracic spine lesion in another. As to pain relief, 10 (66.7 %) of 15 patients were satisfied. In eleven out of 15 who were able be followed more than 3 years after surgery, subaxial lesions adjacent to the fusion segments developed in 5 (45.5 %) patients. There was 1 case of pseudoarthrosis due to infection. The cumulative survival rate in 5 years after surgery was 66.7 %. DISCUSSION: Incidence of subaxial subluxation adjacent to the previous fusion has been reported from 9 % to 40 % of patients with OCF or posterior atlantoaxial fusion (PAA). However no predictor for SS after OCF or PAA has been reported. The results of this study indicate that the severity of rheumatoid arthritis affects long-term results of OCF. The patients of MUD subset, the most aggressive subset, had higher incidence of regression of neurological symptoms due to subaxial subluxation after OCF than those of MES subset and their life prognosis was shorter than that of the MES patients as well. The results suggest that for a patients of the MUD subset posterior total cervical spinal fusion including the occiput is indicated to avoid to avoid neurological regression due to SS after OCF. Paper # 68 Translaminar Screws of the Atlas Timothy Floyd, MD, FACS (Boise, ID), Dieter Grob, MD (Zurich, Switzerland) INTRODUCTION: Spondylodesis of the atlantoaxial complex with translaminar screws is indicated in certain cases of rheumatoid arthritis, osteoarthritis, posttraumatic instability and congenital anomaly. Screw fixation can be supplemented with posterior fixation techniques, such as sublaminar wiring, which also can be used for fixation of graft. However, sublaminar wiring is not possible in cases where the C1 lamina is deficient. In these rare cases, extension of the fusion to the occiput has been described. We describe a more conservative technique for posterior fixation of graft in such cases. METHODS: Over a three year period of time four patients who required atlantoaxial spondylodesis and who had deficiency of the posterior arch of C1 presented to the clinic. One had rheumatoid arthritis with C1C2 instability and C1 spina bifida, one had atlantoaxial osteoarthritis with C1 spina bifida, and one had painful pseudarthrosis of a Type II odontoid fracture. She had anterior subluxation with an atlantodens interval of 15mm and 5mm of instability on flexion/extension films. The subluxation created stenosis which required a C1 decompressive laminectomy. The fourth patient had painful posttraumatic arthritis of C1C2. Technique: After the C1C2 transarticular screws have been placed, the superior fourth of the spinous process of the axis is osteotomized with a power saw, and either a C1 laminectomy is performed (for stenosis) or the blunt ends of the incomplete arch of the atlas are osteotomized with rongeurs. A 2.0mm drill is used to drill down the medullary canal of the arch on both sides. The depth is usually 10 to 15mm. A rectangular unicortical iliac crest graft is harvested and shaped to fit over the C1 arch and a notch is created in the graft to fit into the osteotomized spinous process of C2. A 2.5mm drill is used to created two gliding holes in the graft for compression between with the atlas. The graft is then wedged into position such that the notch fits tightly against the C2 spinous process. Two 2.7mm screws are placed through the graft obliquely into the hemi-arches of the atlas (See Figure). Other cancellous graft can be placed around the site. Postoperatively, the patient is placed in a soft collar for six weeks. RESULTS: Three of the patients were followed until radiographic union occured, while the fourth is still early in the postoperative period. No patient developed a new neurological deficit postoperatively. All patients had resolution of their preoperative pain symptoms. CONCLUSIONS: Translaminar wiring of graft to the atlas is a safe method for supplementation of fixation when transarticular screws are used and when the posterior arch of the atlas is deficient. The technique provides solid fixation of the graft to the atlas, and with press fit into the spinous process of the axis could enhance three dimensional stability of the construct. This technique is a conservative alternative to extension of the spondylodesis to the occiput. The use of translaminar screws of the atlas preserves atlantooccipital motion and avoids the morbidity and complications of occipitocervical arthrodesis. Paper # 69 The Treatment of Cervical Instability in Children with Skeletal Dysplasia Rajiv V. Taliwal, MD (Philadelphia, PA), Suken A. Shah, MD (Wilmington, DE), William G. MacKenzie, MD (Wilmington, DE) INTRODUCTION: Children with a skeletal dysplasia manifest many orthopaedic abnormalities, but few are as life threatening or disabling as those that involve the cervical spine. Guidelines have been established for the diagnosis and management of cervical instability, with or without overt myelopathy. Surgical stabilization with posterior cervical fusion is the mainstay of operative intervention, but little has been written about the outcome of these children managed operatively. The purpose of this study is to retrospectively review our experience with the surgical management of cervical instability in children with skeletal dysplasia. We aim to describe the physician-based outcome, define the fusion rate, and report the complication rate in both the perioperative and postoperative periods. METHODS: Twenty-nine children previously diagnosed with a skeletal dysplasia underwent a posterior cervical fusion for instability at our institution from 1980 to 1998. A retrospective review was performed in order to determine the type of skeletal dysplasia, level of instability, preoperative neurologic findings, levels of decompression and/or fusion, technique, type and duration of postoperative immobilization, and complications. RESULTS: Of the 29 children with skeletal dysplasia that underwent a posterior cervical fusion for instability, 5 were excluded from the study because they had prior procedures done elsewhere. Of the remaining 24 patients, 13 were male, and 11 were female. Six patients had congenital spondyloepiphyseal dysplasia, 4 had metatropic dysplasia, 3 had diastrophic dysplasia, 3 were unknown, 2 had spondyloepimetaphyseal dysplasia, and 1 each had spondylometaphyseal dysplasia, camptomelic dysplasia, achondroplasia, metaphyseal chondrodysplasia, pseudoachondroplastic dysplasia, and DyggveMelchior-Clausen syndrome. The instability diagnoses in 18 children was atlantoaxial instability and 6 had subaxial instability and deformity; additionally, 6 of the patients exhibited myelopathy of varying severity. At a mean age of 6 yrs. + 8 mo., 17 children underwent an occipitocervical fusion, 4 had a posterior subaxial cervical fusion, and 3 had an atlantoaxial arthrodesis. Postoperatively, 20 patients were immobilized in a halo brace or cast, 2 in a Minerva cast, and 2 in a hard collar; the immobilization period averaged 12 weeks. The mean follow up period was 49 months. Three children had recurrent instability due to pseudarthrosis and underwent a revision posterior cervical fusion. All of these children had attempted occipitocervical fusions and were immobilized in a halo vest after the index procedure; 2 had allograft bone graft and none had instrumentation placed. One child developed junctional degeneration and instability necessitating an extension of her fusion mass subaxially. Two of the 6 myelopathic children improved neurologically. With the exception of one patient who had a transient CN IV palsy, there were no neurologic complications. Fourteen of the 20 children immobilized in halos developed pin tract complications. Additionally, there were 3 children with pulmonary problems, 2 falls necessitating early halo removal, and one superficial wound infection. DISCUSSION/CONCLUSION: To our knowledge, this is the largest review of the surgical treatment of cervical instability in children with skeletal dysplasia. Twenty of the 24 children (83%) in our study had a successful arthrodesis with minimal postoperative morbidity. However, neurologic improvement in the myelopathic children was limited. Additionally, this patient population is fraught with complications; the complication rate (major and minor) was 67%. The 3 children with pseudarthroses postoperatively had varied types of dysplasia, but all underwent an attempted 4 level occipitocervical fusion. The use of the halo brace in patients with skeletal dysplasia is usually required for postoperative immobilization, but comes with a high rate (70%) of pin site problems. To enhance chances of a successful result in this difficult group of patients, we recommend a meticulous fusion of only the involved cervical levels with autograft bone and halo brace immobilization with vigilant pin care. Paper # 70 Non-Rigid Immobilization of Odontoid Factures Ingo Schwinnen, MD, E.J. Mueller, MD, M. Wick, MD, O. Russe, MD, G. Muhr, MD (Ruhruniversity Bochum, Germany) INTRODUCTION: The appropriate treatment of odontoid fractures is still controversial. Operative as well as non operative treatment is recommended for the same type of fracture. For non operative treatment rigid immobilisation (Halo device) is the method of choice for the majotity of cases. In a retrospective analysis we reviewed 26 patients with type II and III fractures of the odontoid treated with a cervical orthesis only. The results and complications of this group are discussed and a treatment rationale is presented. METHODS: Between 1985 and 1999 26 patients with a fracture of the odontoid were treated with a Philadelphia collar only at our institution.There were 13 female and 13 male patients. The average age was 59.1 (range 15-86 yrs). 8 injuries were due to road traffic accidents, in 16 cases the underlying cause was a minor fall, 1 injury resulted from a fall from a significant height and in one case the mechanism of injury could not be evaluated. The fractures were classified according to Anderson and d'Alonzo. There were 19 type-II and 7 type-III injuries. Ten (38%) fractures were displaced on initial x-rays ( 3.5 mm translation and/or 10° angulation), 16 (62%) fractures were undisplaced. Flexion-/extension views did not reveal any significant instability ( 2mm translation) in all of the fractures. After application of the orthesis all patients were mobilised immediately. External stabilisation of the fractured odontoid with a Philadelphia-collar was maintained for 3 months. All patients could be followed up at 25.4 months on average (range 6-75 months). RESULTS: There were no posttraumatic or treatment related neurological deficits. Solid union of the fracture could be obtained in 20 (77.5%) patients. In 4 (15%) patients a stable and clinically asymptomatic pseudarthrosis was documented. 7 (35%) of the 20 fractures showed significant displacement ( 4.0 mm anterior/posterior translocation and/or than 10° angulation) of the consolidated odontoid. 9 (35%) patients complained about persistent neck pain and a restriction of cervical spine motion. The clinical symptoms did not correlate with the radiological results. Secondary procedures because of persistant instability or failure of fixation had to be performed in 2 (7.5%) patients. Both were initially undisplaced type II injuries. Another multiple injured patient suffered from pulmonary embolism. The overall complication rate was 11.5 %. No patient died during the observation period. DISCUSSION/CONCLUSION: The ideal treatment of fractures of the odontoid process, particulary type II fractures, has been controversial. For non-operative treatment the best results have been reported with halo-thoracic bracing. This study demonstrates, that stable type II and III fractures of the odontoid, even if they are displaced, can be successfully treated with non-rigid immobilisation. Evaluation of the stability of the fractures with flexion-/extension views is mandatory for decision making. Elderly patients seem to have a rate of pseudarthrosis. Paper # 71 Cervical Myelopathy Associated with Os Odontoideum and Atlantoaxial Instability: Magnetic Resonance Imaging and Its Clinical Significance Jong-Beom Park, MD, Woo-Sung Choi, MD, Han Chang, MD (Uijongbu, Korea) INTRODUCTION: Os odontoideum may result in an atlantoaxial instability and cervical myelopathy. Previous studies have shown that the development of cervical myelopathy, which is the most serious complication associated with Os odontoideum, is thought to be related to the degree of atlantoaxial instability, decrease in the space available for cord (SAC), or direction of atlantoaxial instability. However, these indirect radiographic parameters using plain radiographs can not provide direct information concerning the causes of cord compression in patients with Os odontoideum. In the current study, the authors measured four radiographic parameters including the direction of instability, the degree of instability, and SAC in flexion and extension to assess the relationships with the development of cervical myelopathy. In addition, the pathologic structures compressing the spinal cord were evaluated with magnetic resonance imaging (MRI). METHODS: 13 patients who had been treated with posterior atlantoaxial fusion were classified into two groups according to clinical manifestations with special regard to the occurrence of cervical myelopathy. There were four patients without myelopathy and nine with myelopathy. The mean age was 41.7 years (range, 564 years) and the mean duration of symptoms was 23 months (range, 2-83 months). Four radiographic parameters were measured with flexion and extension lateral radiographs: the main direction of instability of Os odontoideum (anterior or posterior), the maximum distance that Os odontoideum moved (the degree of instability), and SAC in flexion and extension. MRI was taken in nine patients with cervical myelopathy. MannWhitney U test was used to analyze the statistical differences for the degree of instability and SAC in flexion and extension between the patients with and without myelopathy. Kappa statistics was used to test the agreement between the presence of cervical myelopathy and main direction of instability. P < 0.05 was considerd significant. RESULTS: There were no significant statistical differences in the degree of instability (6.83 versus 7.38, p = 0.816), SAC in flexion (6.94 versus 7.13, p = 0.938), and SAC in extension (7.56 versus 5.75, p = 0.434) between the patients with and without myelopathy. Kappa statistics showed a poor agreement between the main direction of instability and the presence of cervical myelopathy (kappa = 0.268, p = 0.308). MRI showed, however, cord compression by retrodental reactive lesions in nine patients with cervical myelopathy: fibrocartilaginous masses were seen in seven patients, and cystic masses, in two. DISCUSSION/CONCLUSION: The current study did not support the results of previous studies which suggested a strong relationship between the development of myelopathy and the direction of instability, the degree of instability, and decrease in SAC. However, MRI provided a direct visualization of cord compression by showing fibrocartilaginous or cystic masses, which were not detected on plain radiographs. The current study suggests that the radiographic parameters using plain radiographs should be reevaluated and retrodental reactive lesions should be considered as the potential cause of cervical myelopathy in patients with Os odontoideum and atlantoaxial instability. Paper # 72 Fracture Etiology of the Odontoid Process Christian M Puttlitz, PhD (San Francisco, CA), Vijay Goel (Iowa City, IA), Charles Clark (Iowa City, IA), Vincent Traynelis (Iowa City, IA) • (a - Surgical Dynamics, Inc.) INTRODUCTION: Fractures of the odontoid process of the second cervical vertebra comprise 7-13% of all cervical spine fractures. Most published reports involving odontoid fracture use the classification system detailed by Anderson and D’Alonzo. There is considerable controversy as to the major load path that causes odontoid fractures [1,2]. A review of the literature fails to designate a consensus on this issue. The large combination of exogenous and/or muscle loading on the upper cervical spine can produce complex loads that lead to injury or tissue injury. It is impractical to study such a wide range of load combinations in a cadaver model. In an attempt to study the changes in stress distribution and load transfer due to these types of loads, a validated finite element (FE) model was subjected to combinations of compression or tension, and horizontal shear forces alone or force loading coincident with flexion, extension, lateral bending, and axial rotation loads. METHODS: A ligamentous three-dimensional (FE) model of the C0-C1-C2 complex was generated from 0.5 mm thick serial computed tomography scans. Validation of the model has been described previously . Force loads were applied at the posterior margin of the occiput. These loads were applied as lone entities or in concert with flexion, extension, or lateral bending moments. Applied rotation moments (1.5 Nm) were used to induce full range-of-motion. The posterior occipital forces simulate gross loading to the back of the head. These exogenous loads induce not only force loading but add some component of rotational moment generation due to its offset from the center of the upper cervical spine. The loads were varied in the sagittal, coronal, and transverse planes (Table 1). Force vectors imposed on the model produced a resultant 100N load. The location (Type I, II, or III fracture region) and magnitude of the maximum stresses were reported. RESULTS: The data indicate that pure rotational loading, in the absence of additional force vectors, produces relatively minor maximal von Mises stresses (to be less than or equal to 44 MPa) in the axis, with extension producing the highest stress. Conversely, shear loading, in the absence of coincident rotational moment application, resulted in maximal axis stresses in excess of 100 MPa for numerous (6) force vector scenarios. Inferior (compression) and lateral forces, or combinations thereof, were integral in the production of these high stresses. All of these maximal stresses were experienced in the Type II fracture region. The coupling of rotation moment and force vector application demonstrated some interesting results. Flexion mitigated the stress response due to compressive loading. Extension loading produced a very high maximum von Mises stress (226 MPa) when it was applied with a coincident inferoposterior load. In fact, extension loading with coincident force application resulted in ten scenarios wherein the maximum axis von Mises stress exceeded 100 MPa. Most of the stress locations for these maximal stresses fell within the Type I or Type II fracture areas. Lateral bending results support the findings of posterolateral loading in heightening the magnitude of the maximal stress. Inferior loading (compression) with lateral bending also indicated an ability to increase the maximum stress as compared to just compressive loading. In all cases, axial rotation failed to appreciably (greater than 4 MPa) increase the maximum stress of the axis as compared to force application only. CONCLUSIONS: Overall, this investigation has indicated that extension and the application of extension via force vector application causes the greatest risk of superior odontoid failure. This finding is in disagreement with those found by Althoff , who dismissed the contributions of sagittal plane rotation to odontoid fracture. The hypothesis of extension as a causal mechanism of odontoid fracture from the findings of this study includes coupling of this motion to other rotations. Flexion seems to provide a protective mechanism against force application that would otherwise cause a higher risk of odontoid failure. To our knowledge this finding has not previously been described. REFERENCES: 1)Althoff, Acta Orthop Scand 177S:1-95, 1979 2) Doherty et al., Spine 18:178-84,1993 3)Puttlitz et al., ORS Transactions:471, 1999. • If noted, the author indicates something of value received. The codes are identified as: a - research or institutional support, b - miscellaneous funding, c - royalties, d - stock options, e - consultant or employee. For full information, refer to inside back cover. Paper # 73 Anterior Odontoid Screw Fixation Julie E. York, MD (Salt Lake City, UT), Russell Lonser, MD (Salt Lake City, UT), Robert Veres, MD (Budapest, Hungary), Adrian Casey, MD (London, UK), Ronald I. Apfelbaum, MD (Salt Lake City, UT) • (e - Aesculap Instrument Corp, Drs. Apfelbaum & Barbera) INTRODUCTION: The use of anterior odontoid screw fixation for the management of type II odontoid fractures has gained acceptance because it confers immediate stability, preserves C1-2 rotatory motion, and has a fusion rate that compares favorably with alternative treatment methods. In an attempt to refine the indications for this procedure, we retrospectively reviewed the surgical outcome of patients who underwent this procedure at two institutions. METHODS: Between 1986 and 1998, one hundred thirty-eight patients underwent direct anterior screw fixation for type II odontoid fractures at either The University of Utah (n=93) or National Institute of Traumatology in Budapest, Hungary (n=45). There were 92 males and 46 females, with a mean age of 50 years (range, 15-92 years). One hundred twenty-five patients had recent fractures (less than 6 months old) and 13 patients had remote fractures (greater than 6 months old). The follow-up ranged from 3 to 60 months, with a mean of 18.2 months. RESULTS: Patients with recent fractures had an overall bony stabilization rate of 88%, whereas patients with remote fractures had a significantly lower rate of fusion (25%, p<0.05). Anatomic bony fusion of recent fractures was significantly higher in fractures oriented in the horizontal and posterior oblique direction relative to anterior oblique fractures (p<0.05). Factors that did not have a statistically significant impact on fusion rate included patient age, sex, direction of displacement, and amount of displacement. Complications related to hardware failure occurred in 14 patients, with a disproportionately higher rate seen in patients with remote fractures. There was one mortality. DISCUSSION: Our results suggest that direct anterior screw fixation is effective for treating recent type II odontoid fractures. The most important predictor of successful fusion was the length of time between injury and surgery. Patients treated with in 6 months of injury had a significantly higher rate of fusion. Furthermore, patients with horizontal or posterior oblique fractures had a significantly higher fusion rate relative to patients with anterior oblique fractures. • If noted, the author indicates something of value received. The codes are identified as: a - research or institutional support, b - miscellaneous funding, c - royalties, d - stock options, e - consultant or employee. For full information, refer to inside back cover. Paper # 74 3D-CT Based, Personalized Drill Guide for Posterior Cervical Stabilization at C1-C2 Jan Goffin, MD, PhD (Leuven, Belgium), Karel Van Brussel, MSc, Kirsten Martens, MSc, Jos Vander Sloten, PhD, Remi Van Audekercke, PhD, Maria-Helena Smet, MD, PhD, Guy Marchal, MD, PhD (Leuven, Belgium), Wilfried Van Craen, MSc (Haasrode, Belgium) • (a - Brite-Euram Project PISA) INTRODUCTION: Posterior fusions with wiring of the cervical spine at C1-C2 according to Gallie, Brooks-Jenkins and Dickman et al. are associated with failure rates of fusion up to 25 %, primarily in cases with rotational instability. The addition of transarticular screw fixation according to Magerl and Seemann offers a better biomechanical stability of the C1-C2 complex, providing much higher fusion rates up to 95 % or more. Nevertheless the Magerl technique is not without risk, particularly for the vertebral artery. Even the help from frameless stereotactic guidance does not provide ideal screw placements in all situations. Material and methods. In the context of the Brite-Euram project PISA, a cadaver study and clinical trial with a medical image based, personalized drill guide were carried out. A high precision spiral CT scan (slice thickness 1mm, table speed 1mm/sec) of the inferior occiput-C1-C2 complex of the cadaver or patient is obtained. After processing and segmenting of the digital CT data (MIMICS software by Materialise N.V., Belgium) the CT data are transferred to a 3D solids CAD environment (Unigraphics by EDS) to design a template for the posterior course of C2 as well as to allow preoperative navigation for screw trajectories: the exact entry points for the screws at C2 as well as the orientation of the screws inside the left and right isthmus C2 are defined by the surgeon preoperatively. This information is then introduced into the CAD program. The template should assure a stable, unique and correct position upon C2. To this end the template features a number of clamps to interface with the posterior course of C2. The template is produced in medical grade (USP Class VI) acrylate resin (Stereocol by Zeneca) using the stereolithography rapid prototyping technique. Stainless steel cylinders with an inner diameter of 2.4 mm and a length of 20 mm are produced and fixed in the template to serve as a drill guide and interface between the drill and the polymer template. The 2.4 mm diameter corresponds exactly to the diameter of the drill. Two series of cadaver studies were carried out. For the first series of 5 cadavers a template with clamps connecting only to the lamina C2, excluding the spinous process from the interface, was tried out. For the second series of 3 cadavers the template was connected not only to the lamina, but also the spinous process of C2. Both cadaver studies were performed without any fluoroscopic control at surgery. Eventually two patients were operated upon using this new technology. RESULTS: The results of the first and second cadaver studies are presented in tables 1 and 2. As the results concerning rotational stability of the template towards the lamina C2 were insufficient during the first series, reflected in a sometimes unacceptable discrepancy between navigated and surgically obtained sagittal orientation of a number of screws, it was decided to include also spinous process C2 into the interface for the second series of cadavers. Both the entry points and screw trajectories were very satisfactory for this second series as well as for the two clinical cases, eventually operated upon. Table 1. Comparison of planned screws vs. inserted screws (set 1). The first part of the table lists the absolute differences (in mm) between the (x,y) coordinates of the origin. The second part of the table lists the angles (in degrees). The xaxis runs from left to right, the y-axis antero-posterior. (L = left, R = right, m = mean, s = standard deviation; ++ = optimal, + = acceptable, - = unacceptable, -= failure). Cadaver 2 was excluded because of total failure (probably due to exchange of cadavers). Table 2. Comparison of planned screws vs. inserted screws (set 2). DISCUSSION: Transarticular C1-C2 fixation according to Magerl is a procedure not without risk, particularly for the vertebral artery. Frameless stereotactic guidance might lower these risks, but even this technique does not provide adequate screw positionings in every patient. Although the actual experience is very limited, the idea of using a template with drill guide, as presented here, simplifies the surgical act and at the same time might enhance the accuracy of C1-C2 transarticular screw positioning. The planning of the screw trajectory is completely done in the presurgical period. Intraoperative planning, based on the preopertively stored data, as is done in actual stereotactic guidance systems, can be avoided. This fact shortens the total operation time and, while avoiding possible problems with peroperative redefining of anatomical landmarks, might increase the accuracy of the screw trajectories. • If noted, the author indicates something of value received. The codes are identified as: a - research or institutional support, b - miscellaneous funding, c - royalties, d - stock options, e - consultant or employee. For full information, refer to inside back cover. Paper # 75 Comparison of Atlantoaxial Stabilization with Posterior Cervical Wiring or Posterior Transarticular Screw Fixation Thomas M. Reilly, MD, Rick C. Sasso, MD (Indianapolis, IN) INTRODUCTION: The upper cervical spine has a very unique architecture which enables multidirectional mobility, with the atlantoaxial (AA) complex allowing significant axial rotation. However, this distinctive anatomy results in a significantly higher failure rate after AA fusions, as compared to arthrodesis in the lower cervical spine. Conventional management of AA instability includes posterior wiring techniques such as the Gallie and Brooks-Jenkins, as well as posterior interlaminar clamping devices, along with supplemental halo-vest immobilization. More recently, Magerl introduced C1-C2 transarticular facet screws as an alternative method for rigid stabilization of the AA complex. The purpose of this study was to review our experience with these fixation methods in the treatment of AA instability. METHODS: Retrospective review of all patients between 1987 and 1998 who had symptomatic AA instability and underwent AA arthrodesis utilizing posterior cervical wiring and halo-vest immobilization (Group I, n=30), or posterior transarticular screw fixation without rigid external immobilization (Group II, n=23). In Group II patients, after preoperative thin-slice CT imaging, all patients were stabilized with AO 3.5 mm cortical screws, along with autograft bone and Gallie-type cabling, without halo-vest immobilization. The causes of AA instability were: 20 (12 Group I & 8 Group II) patients with type II odontoid fractures or fracture nonunion, 16 (7 Gr I & 9 Gr II)with os odontoideum, 13 (8 Gr I & 5 Gr II)with ligamentous instability, and 4 (2 Gr I & 2Gr II) with tumor. The mean patient follow-up was 41.7 months in Group I and 26.4 months in Group II. RESULTS: Four group I patients (13%) developed a pseudoarthrosis, and 29% had complications including pin site or wound infections. All (100%) Group II patients went on to solid fusion, with 8% developing superficial wound infections or suboccipital numbness. There were no cases of screw malposition, neurolgic deficit,or vascular injury, although there was one late unilateral screw fracture that went on to heal uneventfully. DISCUSSION/CONCLUSION: Recent reports indicate that the Magerl transarticualar screw fixation method is biomechanically stronger than conventional posterior wiring techniques, which have been reported to have failure rates as high as 25%. This study demonstrates that, although technically demanding, transarticular screw fixation appears to be superior to traditional posterior wiring or clamping stabilization, resulting in an extremely high fusion rate with few complications, and avoids the necessity and discomfort of prolonged halo-vest immobilization. This technique is a valuable option in the management of patients with AA instability. Paper # 76 Circumferential Approaches for the Correction of Occipitocervical and Subaxial Deformity David W. Cahill, MD (Tampa, FL) INTRODUCTION: The literature available on spinal deformity surgery is heavily weighted toward thoracic and thoracolumbar reconstruction. At our institution, more than 50 cases of multisegmental cervical or occipitocervical deformity corrections have been performed since 1995. Thirty such cases have at least 2 years of followup. METHODS: All subaxial deformity cases involve sagittal plane correction for kyphosis. One quarter of such cases simultaneously involve coronal and/or axial plane correction as well. 18 supra-axial cases involve sagittal plane correction in 12, axial or coronal plane correction in 8. Causative pathology involves congenital, post-surgical, tumor or infection. Two-thirds of all case involved a 360 degree or 540 degree circumferential approach. Anterior (including transoral) reconstruction usually involved titanium cage and buttress plate or rod instrumentation; posterior fixation usually involved non-constrained plates or rods and cables. All used cancellous autograft except for malignant tumor cases. RESULTS: An aggressive approach to anterior and posterior osteotomy followed by anterior distraction and posterior compression allowed correction of kyphotic deformity intraoperatively in all cases. Late loss of lordosis occurred in 1/3 of cases secondary to cage settling or telescoping. Rigid fixation from the skull to the axis or subaxial spine allowed near complete correction of the cervicomedullary junction in all cases with no late loss of correction. No patient suffered a cord injury. There was one superficial wound infection. There were two vertebral artery injuries one of which was associated with a delayed stroke. There were two next segment failures requiring late reoperation for extension of fusion. DISCUSSION: Modern reconstructive techniques allow for clinically significant correction of cervical and occipitocervical deformities associated with chronic pain, neurologic deficit, and cosmetic deformity. An aggressive circumferential approach to decompression and osteotomy allows near anatomic correction of geometric deformity in any plane. Evolving hardware improvements allow improved anterior and posterior fixation. Clinically significant obstacles remain in the development of non-telescoping devices for anterior column reconstruction and in constrained posterior fixation. Paper # 77 360 Degree Surgery for High Cervical Vertebral Tumours Alan Crockard, FRCS, Robert Quiney, FRCS, Benjamin Taylor, FRCS, Jan Lehovsky, FRCS (London, England) INTRODUCTION: For many primary bone tumours the ideal goal is total excision with wide margins, many are radio-resistant and thus this approach has been useful down the years. Vertebral tumours in the high cervical spine present as a particularly challenging problem if these goals are to be achieved. The tumours are rare and a team approach may be the solution for these difficult problems. For tumours in the upper three cervical vertebral our aim has been in two staged procedures to excise all of the tissue involved to preserve the neuraxis and carry out a primary instrumented fixation and fusion. The approach has been used in chordomas, chondrosarcomas, desmoplastic fibromas and osteogenic sarcomas. METHODS: Preoperative evaluation with MRI scan and CT scans are essential as well as plain radiographs. Usually one vertebral artery is related to the tumour and so a planned trial balloon occlusion of the involved vertebral artery as well as testing the competence of the Circle of Willis by MRA is carried out prior to surgery. If there is a competent Circle of Willis and if there are no problems associated with trial balloon occlusion then an elective transection above and below the lesion of the vertebral artery will allow an en bloc removal of the tumour and vertebral body. The surgical approach is usually dorsal first with division of the vertebral artery, mobilising of the involved portion of the pedicle and lamina, insertion of lateral mass screws or other screws to effect a rod and screw instrumented fixation with bone graft. One week later a ventral approach including a midline mandibulotomy, midline glossotomy and ventral extension to the hyoid bone will allow a midline ventral exposure of the top four or five cervical vertebra. The vertebral body and the previously mobilised portion of pedicle with ligated vertebral artery can then be removed. The uninvolved other vertebral artery is carefully preserved. Iliac crest bone graft and vertebral instrumentation either locking titanium plate or Harms type cage is used. Postoperatively a tracheostomy and percutaneous gastrostomy is used until the swelling has settled. A halo body jacket is used for 6 to 8 weeks. RESULTS: Since 1995 we have carried out this procedure in eight patients, five males and two females, aged 19-64 years. Four patients had chordomas two of them were recurrent chordomas referred from elsewhere after previous surgery, the other two chordomas were primary presentations. There was one desmoplastic fibroma, one malignant schwannoma, one chondrosarcoma and one primary melanoma of the dura. There were no postoperative deaths but the most elderly patient who had had a fifth re-operation following proton beam radiation and previous surgery developed a delayed posterior inferior cerebellar artery syndrome and respiratory difficulties which resulted in him requiring respiratory support. He died 4 months later. One young female with a particularly mytotic chordoma had a recurrence in 8 months and this required further lateral operation and radiation therapy all to no avail. All the other patients have shown a good recovery, two patients had C5 root lesions on the side of the vertebral artery excision, in that the C5 root was excised with the tumour as well as the vertebral artery. All the instrumentation has survived, none of the ventral instrumentation has required removal. One patient had posterior pharyngeal wall difficulties which required several attempts at resuturing. The six patients are alive with evidence of sound bony fusion and no recurrence to date. CONCLUSIONS: While the surgery is extensive it is extremely well tolerated with good evidence of fusion and no construct failure. It may be considered for those rare tumours not responsive to radiation therapy or more conventional treatment. Paper # 78 Whiplash Syndrome: Kinematic Factors Influencing Pain Patterns Joseph F. Cusick, MD, Narayan Yoganandan, PhD, Frank A. Pintar, PhD (Milwaukee, WI) • (a - PHS CDC Grant R49CCR515433, Department of Veterans Affairs Medical Research) INTRODUCTION: The whiplash syndrome usually consists of two pain components: headache and neck pain. The proposed mechanism is the induction of load transfer to the head-neck complex by inertial forces as the result of a rear crash. Corresponding kinematics, especially the temporal aspect, of this complex will be influenced by the anatomy and material properties of the various components. Experimental evaluation of such responses in the human head-neck complex will offer insight into various factors that may cause segmental activation of corresponding nociceptors. MATERIALS AND METHODS: Intact human cadaver head neck complexes were fixated at T2-T3. Posterior skin and soft tissues as well as muscle tissues were maintained. Local tissue removal was done to position retroreflective targets (anterior vertebral bodies, mastoid process, facets (C3-T1) with two targets in the inferior and superior regions of each facet). Specimens aligned such that the Frankfort line was horizontal. A load cell and accelerometer were attached to the pendulum. A six-axis load cell and an accelerometer was attached to the base of the complex. Whiplash loading was applied by a mini-sled pendulum. The specimens were tested sequentially as follows. The first run was conducted at a velocity of 1.3 m/s (1 m/s = 2.2 mph). The specimen was evaluated radiologically and macroscopically, and a second run was conducted at a velocity of 2.2 m/s. It was evaluated again and a repeat test was conducted at 1.3 m/s. This procedure was repeated for velocities of 3.3 and 4.6 m/s. The testing was terminated in the event of failure. Overall and close-up high-speed digital photographs of the specimen were taken during whiplash loading at 1000-4500 frames/sec. The change in angular motion of the upper (occiput-C2), middle (C2-C5) and lower (C5-T1) regions were obtained in the sagittal plane, and facet joint compression and sliding motions were computed using the two pair of targets at each facet joint by adopting the principles of continuous motion analysis system. RESULTS: Predominant moments were in the sagittal plane. During the initial loading, the upper segment undergoes local flexion concomitant with the lag of the head as the lower segment is in local extension establishing a reverse curvature. As dynamic loading continues, the inertia of the head catches up with the spine resulting in a single curvature in the extension mode. The extension moment, postero-anterior shear force, and axial dynamic force during the application of the whiplash acceleration pulse increased with increasing velocity, and were non-linear. At 2.2 m/s, the maximum flexion of the upper and middle regions was 0.1 and 0.05 radians, and maximum extensions of the three regions were 0.6 (upper), 0.3 (middle) and 0.4 (lower) radians. At the lower segment, the posterior aspect of the facet joint compresses more than the anterior aspect. Mean peak sliding motion was 2.0 mm with mean peak compression 2.9 mm posteriorly and 2.0 mm anteriorly. DISCUSSION: The present findings offer some insight into the controversy regarding the pathophysiology of the whiplash syndrome. The upper cervical flexion incurred during the initial stage of loading will induce stretching of this region with potential adverse tensile forces acting upon pain-sensitive structures, including the C1 and C2 dorsal root ganglion. Such alteration of nociceptor threshold may result in suboccipital pain and headache. The dissimilar compression at the anterior and posterior parts of the lower segment facet joint resulting in a pinching mechanism may alter local nociceptor function in this region and lead to the occurrence of neck pain. • If noted, the author indicates something of value received. The codes are identified as: a - research or institutional support, b - miscellaneous funding, c - royalties, d - stock options, e - consultant or employee. For full information, refer to inside back cover. Paper # 79 Biomechanical Comparison of Posterior Cervical Fixations Including an Alternative Technique Hisanori Mihara, MD (Madison, WI), Katsuhiro Ohnari, MD (Yokohama, Japan), Norihiko Akiyama, MD (Yokohama, Japan), Boyle C. Cheng, MS (Madison, WI), Stephen M. David, MD (Madison, WI), Thomas A. Zdeblick, MD (Madison, WI) • (b - Best Medical, Inc., Sofamor Danek, Inc.) INTRODUCTION: Fixation of the posterior cervical spine with interspinous wiring is well known as Rogers' or Bohlman's technique. Recently several plate fixation techniques have been used for posterior cervical stabilization. The authors have developed wavy shaped rods as an alternative interspinous fixation technique. This unique technique has been proved to be clinically useful in Japan. However, no biomechanical studies have been performed until now. The purpose of this study is to compare its in Vitro biomechanical stability to standard posterior cervical fixation techniques. METHODS: Nine fresh frozen cervical human spines were tested at the C5-C6 motion segment. Nondestructive static biomechanical testing was performed with axial compression, flexion-extension, lateral bending, and axial rotation after destabilizing and restabilizing consecutively as following: 1) intact spine, 2) creating of a stage 3 distractive-flexion injury followed by fixation with the wavy shaped rods (Wavy Rod, Best Medical, Tokyo, Japan) bound by three multistrand cables (Figure 1), 3) Rogers' wiring with a multistrand cable, 4) Bohlman's triple wiring technique using multistrand cables with a pair of unicortical grafts from the ilium, 5) lateral mass plate fixation with Magerl's technique. Testing was performed on a material testing machine (MTS Bionix 858, MTS Corporation, Minneapolis, MN) and load displacement curves were obtained utilizing four linear extensometers and one rotatory extensometer across the C5-C6 motion segment. Figure 1. Wavy Rod fixation technique RESULTS: In axial compression loading, the reconstructed specimens showed significant differences in range of motion (mm) measured anterior and posterior (one-way analysis of variance). When comparing the four fixation techniques, the construct with the Wavy Rod indicated significantly less motion both anterior and posterior than those with the other fixation techniques. Also in flexion extension loading, all techniques significantly limited the intervertebral motion below the level of the intact motion segment. Particularly, the construct with the Wavy Rod showed the smallest mobility, 49.9% at anterior and 9.3% at posterior of the intact (Figure 2). In lateral bending, the lateral mass plate provided the greatest stability, superior to the intact segment, but the difference was not statistically significant. In axial rotation, all reconstruction techniques limited the angular motion below the intact level (Wavy Rod; 68.0%, Rogers' wiring; 75.2%, Bohlman's triple wiring; 59.8%, lateral mass plate; 71.7% ), but no significant differences were observed at the P< 0.05 level. Figure 2. % ROM under flexion extension load of 3Nm. * indicates statistical difference at p < 0.05 compared with the intact ROM. Wavy = Wavy Rod fixation, Rogers = Rogers' wiring technique, Triple = Bohlman's triple wiring technique, Plate = Lateral mass plate fixation. CONCLUSIONS: All four reconstruction techniques restored the stability to at least the level of the intact motion segment prior to destabilization. A simple alternative cervical posterior fixation technique, the Wavy Rod system, was considered the most effective technique in stabilizing a cervical motion segment, particularly in axial compression and flexion extension loading. • If noted, the author indicates something of value received. The codes are identified as: a - research or institutional support, b - miscellaneous funding, c - royalties, d - stock options, e - consultant or employee. For full information, refer to inside back cover. Paper # 80 Radiographic and Biomechanical Analysis of Multilevel Smith-Robinson vs. Corpectomy Reconstruction of the Cervical Spine: A Human Cadaveric Model Michael F. O'Brien, MD (Wheat Ridge, CO), D.A.B. Smith, MSC (Wheat Ridge, CO), R. Vraney, MD (Wheat Ridge, CO), T.G. Lowe, MD (Wheat Ridge, CO), D.G. Fitzgerald, BSME (Ft. Collins, CO), C.V. Franke, BSME (Ft. Collins, CO), B.D. Loyd, BSME (Ft. Collins, CO), T.R. Haher, MD (New York, NY), A.P. Dwyer, MD (Denver, CO), A. Merola, MD (New York, NY), S.P. James, PhD (Ft. Collins, CO) INTRODUCTION: For symptomatic multilevel cervical spondylosis, decompression may be accomplished with either a multilevel Smith-Robinson technique (MLSR) or a multilevel corpectomy / reconstruction (MLCR). Criticisms of MLSR have included longer operative times and supposedly higher pseudarthrosis rates. However MLCR has been implicated in numerous complications including graft dislodgement, hardware failure, strut failure, subsidence, poor intraoperative restoration of lordosis, and the development of postoperative kyphosis. Purpose: To investigate the radiographic and biomechanical effects on human cadaveric cervical spines of an MLSR vs. MLCR technique. MATERIALS AND METHODS: Six human cadaveric spines (C2-T2) were prepared by careful removal of all nonstructural soft tissue. Specimens were divided into three level Smith-Robinson (MLSR n=3) or three level corpectomy (MLCR n03) groups. The specimens were potted to preserve lordosis. Radiographs of preloaded intact and post-operative specimens were obtained for measurement. Surgical techniques were used to prepare both groups from either C 3 – C 6 or C 4- C 7. Biomechanical testing was performed on an MTS servohydraulic biaxial (axial/torsional) biomechanical testing system. Overall stiffness and end vertebral motion were measured. Specimens were tested intact, immediately “post=op”, and after 1200 cycles. RESULTS: MLSR Group: Lordosis improved by a mean of 24º (range 21º -27º). Overall global stiffness decreased 4.2% from intact to immediately postop. The mean end-segment movement, comparing intact to post-op, decreased 4.8%. No graft dislodgement occurred. MLCR Group: Lordosis improved by a mean of 9º (range 8º -11º). There was a 26% decrease in global stiffness from intact to post-op while there was a 56% decrease in end-segment movement comparing intact to post-op. In one specimen graft dislodgement occurred at the proximal end of the construct during cycling. In spite of being less stiff and less lordotic as a group preoperatively, the MLSR group ultimately had better overall reconstitution of lordosis and lost less stiffness than the MLCR group did comparing intact to post-op. Stiffness decreased by approximately 14% in both groups during cycling. DISCUSSION: This study suggests that MLSR is more likely to achieve segmental lordosis and preserve overall stiffness than is MLCR. Less effective restoration of lordosis and decreased stiffness may be the nidus for the mechanical failure often identified with the MLCR group. This biomechanically unfavorable situation identified in the MLCR is exacerbated by ligamentous forces generated over multiple levels which are then constructed on cancellous bone (“prepared endplate”) by struts of a typically higher modulus of elasticity and smaller contact cross-sectional area than the vertebral endplates. This may result in subsidence and ultimately failure. Paper # 81 Biomechanics of Impact Injuries in the Cervical Region Vijay K. Goel, PhD (Iowa City, IA), Young E. Kim (Seoul, Korea), T-Y Park (Seoul, Korea), Choonki Lee (Seoul, Korea) • (a - KOHTERF-98-03 and US Airforce) INTRODUCTION: Vertical impact of the head in a neutral, extended or flexed posture is one of the leading causes of injury in the cervical region. The characterization of the mechanism of injury under such loads has been done through in vitro tests. There is lack of consensus between the load vector and the types of injuries produced in cadaver tests, for obvious reasons. For a better understanding, it is essential that repeatable studies be undertaken. Such a repeatable experiment is feasible through a mathematical model. This study delineates the use of a finite element model of the entire cervical spine in analyzing its response to impact loads. METHODS: A three-dimensional non-linear finite element model of the ligamentous cervical spine from C1 to C7 was prepared using the methodology described in the literature. The model contained all of the ligaments; facet articulations and other pertinent anatomical features, including disc (annulus and nucleus). The model consisted of 1348 elements and 1245 nodes. The material properties were taken from the literature. The model responses were computed for an impact load of 6400 N alone or coupled with 8 Nm of flexion or extension moment. The impact load was applied using a triangular impulse of 30 msec duration while the moments were quasi-static. In the axial impact alone case, the motion of C1 was constrained along the vertical axis, thus simulating the action of muscles while in the other two cases, C1 was allowed to move freely. Model responses to quasi-static + 600 N axial forces, + 6 Nm of flexion/extension moment and up to 30 Nm of axial torque were also computed. RESULTS: The predicted displacements (translations and rotations) for quasistatic axial force and moments were found to be in close agreement with the published data; validating the model. In axial compression compact loading case, the largest facet contact force of 4500 N was generated at the C5-6 level followed by C4-5 level (4250N). The intradiscal pressure in the C6-7 disc was the highest (2.75 MPa) followed closely by the C5-6 level. For case 2, flexed and vertical impact loading, the PLL across C2-3 experienced a tension force of 213 N, which exceeds its rupture strength (207 N). In case 3 (extension plus axial impact), the tension force in ALL across C2-3 was very high as compared to the other levels. Capsular ligaments across C4-5 levels experienced the highest forces. DISCUSSION: In axial impact mode, the results predict that the structures across C5-6 level (disc and facets) are likely to undergo fracture. In the flexion/extension modes, the fracture region shifts to C2-3 level and the ligaments are likely to rupture. These results are in agreement with the in vitro data reported in the literature. In summary, we have developed an experimentally validated model that is capable of predicting the injury prone regions in the cervical spine subjected to axial impact force in the constrained and unconstrained modes. Further work is in progress to investigate the response of the spine in whiplash loading scenarios. • If noted, the author indicates something of value received. The codes are identified as: a - research or institutional support, b - miscellaneous funding, c - royalties, d - stock options, e - consultant or employee. For full information, refer to inside back cover. Paper # 82 Effect of Endplate Conditions and Bone Mineral Density on the Compressive Strength of the Graft-Endplate Interface in the Cervical Spine Howard S. An, MD (Chicago, IL) INTRODUCTION: The Smith-Robinson anterior cervical fusion is an established procedure for patients with cervical disc disease. The technique involves placing a tricortical iliac graft against endplates. The endplate is burred and holes are made for contouring and to improve vascularity to the graft. Subsidence of the graft into the vertebral body is a well-known complication in anterior cervical fusion, which can result in graft extrusion, kyphotic deformity, pseudarthrosis, and recurrence of symptoms. There is no information in the literature in regards to the compressive strength of the graft-endplate interface as related to the endplate thickness, holes in the endplate and bone mineral density (BMD) of the vertebral body, in which the Smith-Robinson anterior cervical fusion technique is simulated. The purpose of this study was to investigate the effect of endplate thickness, endplate holes, and BMD on the biomechanical strength in an anterior interbody fusion of the cervical spine. METHOD: A total of 7 cervical spines (C3-C7) were obtained from fresh cadavers (age: 69-86 with mean of 79) and dissected using a surgical blade through the intervertebral disc to obtain the isolated vertebrae. After carefully removing the annulus fibrosus tissues with no damage on the endplates, posterior elements of each vertebrae were removed using a bone saw. 8 vertebrae were found to have gross pathology in plain x-rays and excluded from the study. Remaining 27 vertebrae were sealed in double plastic bags and stored at –20 degC until testing. Bone mineral density (BMD) of each vertebral body was measured using a dual-energy x-ray absorptiometry (DEXA) unit (Lunar DPX-L, Madison, WI) and LUNAR small animal software. Frozen vertebral body specimens were placed in anterior-posterior position and scanned in slow speed and high resolution modes. BMD of each vertebral body was measured from the scanned image three times and average values were used for the data analysis. Thickness of endplates was measured using CT images. Coronal section images of each vertebral body were acquired with 120 kV, 70 mA, and 1.5 mm thick contiguous slices and transferred to a PC. All CT scans were performed while keeping the specimens frozen. Endplate thickness was measured using a computerized image processing system (Image-Pro Plus, Media Cybernetics, Silver Spring, MD). Three CT sections representing the center portion of each vertebral body were selected and used for measurement. For each endplate image, endplate thickness was measured at three points, Center point and both lateral midpoints. An average value of these 9 measurements (3 sections x 3 measurement points/section) was determined as the representing endplate thickness. Thickness of both superior and inferior endplates was measured separately. Following the DEXA and CT, each vertebral body was cut into halves through the horizontal plane and a total of 54 specimens consisting of one endplate and half of the vertebral body was obtained for destructive compression tests. The prepared specimens were assigned into one of the following 3 groups so that group mean BMDs became similar. In Group I, the endplates were preserved intact. The endplates in Group II were burred into approximately _ of the intact thickness. In Group III, the endplates were totally removed using a burr. Each specimen underwent the destructive compression test. After completely thawed, the prepared specimen was placed between the self-aligned plate and a custom made indentor on an Instron material testing system (Model 1321). The indentor was made to have circular cross section. The diameter (8 mm) was determined so that the indentor did not compress the cortical bone during the test. The compressive load was applied using a displacement control at a speed of 10 mm/min. The maximum displacement of the indentor was 2 mm sufficiently larger to ensure the endplate fracture. Quasitstatic loading speed was used to minimize the viscoelastic effect of the specimens. While loading, loaddisplacement curves obtained, and the load to failure (LF) was determined as the maximum load in each curve. Parameters measured in this study were BMD of the vertebral body, endplate thickness and LF. Regression analyses were performed to investigate the relationship between the measured parameters in each group, and ANOVA tests were performed to compare the differences in measured parameters between groups as well as between cervical levels. The effect of holes was studied by using a finite element model (FEM). A threedimensional non-linear FEM of an intact C5-6 cervical spine segment was developed, and anterior discectomy and bone graft without any hole was created as control . Four characteristic hole patterns were created on the upper endplate of C6: one larger central hole, 2 lateral holes, 2 anterior and posterior holes, and 4 holes. The contact area removed was about 20% for all cases. The compression loading was simulated at 110N to the geometric center of the upper vertebral body. The area failed and maximum stress for obtained for each case and compared to the control. RESULTS: Means (+/-SDs) of and endplate thickness and BMD for different cervical levels are listed in Table 1. There was no significant effect of the cervical level on endplate thickness and BMD, and the thickness of superior and inferior endplates was similar. In addition, there was no relationship between the BMD and end-pate thickness. The mean BMDs and LFs (+/-SDs) for Groups I, II, and III are listed in Table 2. In each group, there was a significant linear relationship between BMD and LF (R = 0.57 in Gr I; R = 0.647 in Gr II; R = 0.581 in Gr III). Multiple regression analyses were performed on Gr I data to investigate the relationship of LF with BMD and endplate thickness because the Gr I specimens had intact endplates. LF was found to have relationship with BMD but not with the endplate thickness. Multi-factor ANOVA showed that the endplate condition significantly affected LF (p < 0.05) whereas the cervical level and the endplate position (superior and inferior) had no effect on LF. Tukey multiple mean comparison showed that LF of Gr I (intact endplate) was significantly greater than LF of Gr III (no endplate). The FEM results revealed that the hole pattern does not significantly influence graft subsidence or anterior-posterior translation under 110 N loading, but it influences the fraction of the upper endplate which is exposed to fracturing stresses (Table 3). A large central hole is more effective than the other patterns at distributing a compressive load across the remaining area and minimizing the fracture area. DISCUSSION: The effect of the endplate condition, BMD of the vertebral body on the mechanical strength of the graft-endplate interface in the cervical spine was investigated in this study. The results of this study help draw several clinically relevant conclusions. For example, the significant relationship between BMD and LF indicates that preoperative consideration of BMD would be important for patient selection and the choice of a surgical technique. Significantly larger LF in the intact endplate group (I) than no endplate group (III) suggests that contribution of the endplate for maintaining the compressive strength at the fusion site may be more significant than considered. Thus, it may be important to preserve the endplate as much as possible to reduce the incidence of subsidence when performing the interbody fusion particularly in patients with poor bone quality. For making holes in the endplate for vascularity of the bone graft, one central hole is preferable to more smaller holes in minimizing the exposed surfaces for fracturing stresses. Paper # 83 Biomechanical Analysis of Cervical Spine Endplate Failure: A Comparison of Constructs Following Cervical Corpectomy Stephen M. David, MD (Asheville, NC), Boyle C. Cheng, MS (Madison, WI), Hisanori Mihara, MD (Madison, WI), Thomas Zdeblick (Madison, WI) • (a,b – Medtronic) INTRODUCTION: The optimal reconstructive technique after cervical corpectomy has yet to be determined. In general, reconstructive options after cervical corpectomy include an interbody spacer (autogenous/allograft fibula, or iliac crest, metallic device e.g. Harms cage) supplemented by external immobilization and/or internal fixation (anterior and/or posterior instrumentation). Much attention has been given to internal fixation in an effort to diminish graft complications and decrease the time of external immobilization. Despite internal fixation, failure of fixation, graft dislodgment and endplate failure has been reported. The purpose of this study was to compare the initial biomechanical properties, with reference to endplate strength, of various constructs used to reconstruct the cervical spine after corpectomy. MATERIALS AND METHODS: Fresh human cadaveric subaxial cervical spine vertebra (C3 to T1) were dissected free of all soft tissue. End-plate cartilage was removed using a curette. Care was taken not to perforate adjacent end-plate subchondral bone. Prior to biomechanical testing, plain radiographs and bone mineral density scans (DEXA) were obtained. Single cervical vertebra were assigned treatment groups (Table 1) in a random fashion. Constructs were then applied to each vertebra based on treatment group. Spines were tested on an MTS Bionics 858 (MTS, Eden Prairie, MN) test system. Under displacement control of the main actuator at 50 mm/min (no peak torque or moment was maintained), constructs were tested to failure. Load displacement curves for each construct were analyzed to compare linear elastic stiffness values, energy absorption (area under the load displacement curve) and maximum load at the moment of failure. Specimens were inspected to determine the mode of failure. Fischer’s protected least square differences (PLSD) analysis of variance was used to detect differences between treatment groups. P values less than 0.05 were considered statistically significant. All analyses were performed using SAS statistical software (SAS Institute, Cary, NC). RESULTS: (See Figures 1 and 2) No difference was noted for load to failure, stiffness, or energy absorbed between fibular grafts of iliac crest grafts. Fibular graft constructs failed with endplate perforation without fracture of the graft, while iliac crest grafts failed by shortening of the graft without endplate failure. The addition of an inferior junctional plate significantly increased construct stiffness (p<0.05). When comparing the addition of a spanning anterior plate, the fibula plus plate group was generally stiffer than the iliac crest and plate group. The anterior and posterior plated group was the construct least likely to fail at the endplate or bone/implant interface. The fibular graft with the spanning anterior plate demonstrated the greatest load at failure of the anterior only implant constructs, followed closely by the junctional plate construct. The Harms bone/implant interface. The fibular graft with the spanning anterior plate demonstrated the greatest load at failure of the anterior only implant constructs, followed closely by the junctional plate construct. The Harms cage as a stand alone construct failed at lower loads than the fibular graft with the spanning (p<0.04) or junctional (p<0.02) plates of the tricortical iliac crest graft with a spanning anterior plate (p<0.03). The addition of a spanning anterior plate to the Harms cage construct significantly increased load to failure (p<0.007) and energy absorption (p<0.01). Bone mineral density of the cervical vertebra as measured by DEXA scanning did not correlate with endplate maximum failure load. DISCUSSION: Our findings suggest that anterior and posterior plating minimizes failure of the subchondral endplate and the bone/implant interface after corpectomy. If anterior only constructs are considered, the constructs employing a spanning plate failed at higher loads, closely followed by the junctional plate construct. The addition of a junctional plate enhanced construct stiffness compared to the graft alone construct, most likely on the basis of the screws in the junctional plate being approximated to the endplate/graft interface, increasing the stiffness of this interface. Based on these findings, it is recommended that the screws being used in a junctional plate be placed near the subchondral bone adjacent to the graft/host interface to enhance construct stiffness. The Harms cage alone was least able to withstand higher loads compared to other constructs, most likely due to higher loads over less endplate surface area, compared to constructs with greater surface area contact with the endplate. The addition of an anterior plate to the Harms cage construct significantly increased load to failure and energy absorption. No difference was detected between maximum load to failure, stiffness or energy absorbed between fibula or iliac crest grafts, however the modes of failure differed. Based on our findings, we do not feel that bone mineral density as typically measured by DEXA scanning is a good predictor of endplate failure after corpectomy reconstruction. DEXA scanning measures cancellous bone in the central part of the vertebral body and did not correlate with failure loads at the endplate cortical (subchondral bone). Figure 1. Ultimate Strength Figure 2. Linear Elastic Stiffness • If noted, the author indicates something of value received. The codes are identified as: a - research or institutional support, b - miscellaneous funding, c - royalties, d - stock options, e - consultant or employee. For full information, refer to inside back cover. Paper # 84 Biomechanical Comparision of Five Different Atlanto-Axial Posterior Fixation Techniques Thomas Henriques, MD (Uppsala, Sweden), Bryan W. Cunningham, MSC (Baltimore, MD), Norimichi Shimamoto, MD (Baltimore, MD), Claes Olerud, MD, PhD (Uppsala, Sweden), Guy A. Lee, MD (Baltimore, MD), Suue Larsson, MD, PhD (Baltimore, MD), Paul A. McAtec, MD (Baltimore, MD) • (b – Nordopedic AB) INTRODUCTION: Previous investigations have demonstrated that three-point fixation, using bilateral transarticular screws in combination with posterior wiring, provide the most effective resistance to minimize motion about the C1C2. However, transarticular screws are technically demanding and require considerable experience. Posterior wiring techniques affording one point of fixation, such as Brooks and Gallie, have indicated failure rates of ~15%, which are considered secondary to structural bone graft failures. The present study was undertaken to compare five different reconstructions of the atlanto-axial complex. The primary objective in the present study was to determine whether non-bone graft dependent one-point fixation can afford stability levels equivalent to three-point reconstruction. METHODS: Eight human cervical specimens (C0-C4) (age range 70-94 yr, mean 78.1) were utilized in the current study. Non-destructive biomechanically testing was performed, including axial rotation (±1.5Nm, 50N preload), flexion/extension (± 1.5Nm) and lateral bending (±1.5Nm) loading modes. Following intact spine analysis, fracturing the odontoid process destabilized each specimen. Each specimen was then reconstructed in the following order:  Supralaminar hooks over C2 in combination with C1 claw (HC)  Pedicle screws in C2 in combination with C1 claw (PC)  Bilateral C1-C2 transarticular screws (Magerl) (M)  Bilateral C1-C2 transarticular screws combined with a posterior wire fixation according to Gallie using structural bone graft (MG)  Bilateral C1-C2 transarticular screws in combination with the C1 claw (MC). Unconstrained three-dimensional segmental motion was measured using an optional motion analysis system (Optotrak 3020). Range of motion (degree) at C1-C2 and the level above and below, were statistically compare using a one-way analysis of variance and Scheffe’s post hoc test. RESULTS: Under axial rotation all C1-C2 reconstructions indicated significantly higher stiffness levels than the intact spine (p<.0001). The two and three point reconstructions using transarticular screw (M) provided higher stiffness than the one-point reconstructions. PC vs. M, MG, MC and HC vs. MG and MC (p<0.05) (Figure 1). During flexion/extension, higher stiffness levels were observed in one and three-point fixations when compared to intact spine at p<0.05. In lateral bending, no significant differences between the six groups were indicated, although the trend was that reconstructions including transarticular screws provided greater stability than one-point reconstructions. DISCUSSION/CONCLUSIONS: Three-point reconstruction using bilateral transarticular screws in combination with the C1- claw device (MC), proved to markedly reduce motion in all planes compared to one and tow-point fixation. The one-point fixation constructs –HC and PC- which do not rely on structural bone graft, resulted in less stability compared to the three-point reconstructions. An interesting finding was that adjacent level range of motion during rotation seemed to increase the following reconstruction, although no significant differences were observed. This may explain why bone-graft-depending posterior wire fixation fail in ~15% of cases, and provides a strong argument for the use of three-point fixation in surgical management of C1-C2 instability. The current study substantiates the use of three-point fixation as the treatment of choice for C1-C2 instability. When used in combination with transarticular screws, the C1 claw device provides an alternative to posterior wiring reconstructions. Figure 1. Axial Rotation Stiffness: *Indicates significant difference from the intact condition; # from Magerl-Gallie and Magerl-Claw; $ from Magerl, Magerl-Gallie and Magerl-Claw. Error bars represent one standard deviation and significance is indicated at p<0.05. • If noted, the author indicates something of value received. The codes are identified as: a - research or institutional support, b - miscellaneous funding, c - royalties, d - stock options, e - consultant or employee. For full information, refer to inside back cover.
© Copyright 2017