C.V. Guy Rutledge Fogel M.D. As of 5/30/2008 Current As of 5/30/2008

C.V. Guy Rutledge Fogel M.D. as of 5/30/2008

RECENT PUBLICATIONS 2003-2008:

1. Coccygodynia Evaluation and Treatment G Fogel MD, S. Esses MD, JAAOS 2003 2. Hip Spine Syndrome G Fogel MD, S. Esses MD, The Spine Journal 2003 3. Cervical Plates Comparison of Physical Characteristics and In Vitro Push Out Strength G Fogel MD,C. Reitman MD, W. Liu Phd., S. Esses MD, The Spine Journal 2003 4. Physical Characteristics of Polyaxial-Headed Pedicle Screws and Biomechanical Comparison of Load With Their Failure, G Fogel MD,C. Reitman MD, W. Liu Phd., S. Esses MD, Spine March 03. 5. Management of Chronic Limb Pain with Spinal Cord Stimulation G Fogel MD,O. Calvillo MD Phd., S. Esses MD Pain Practice-03 6. Biomechanical Evaluation of Relationship of Pullout Strength, Insertional Torque and Bone Mineral Density in the Cervical Vertebral Body Journal of Spinal Disorders 2003, C. Reitman, G. Fogel MD, Lyndon Nguyen, MS 7. Spinal Epidural Lipomatosis 3 case reports and Literature Metanalysis, The Spine Journal 2005 Fogel, G. R., P. Y. Cunningham, 3rd, SI Esses 8. Meningeal Cyst Evaluation and Management, American Journal of Orthopedics 2004. Fogel, G. R., P. Y. Cunningham, 3rd, SI Esses 9. Anterior Hardware Removal Improves Dysphagia Following Anterior Cervical Discectomy and Fusion, The Spine Journal March 2005, G Fogel M.D., M McDonnell M.D. 10. Complications, Pitfalls, and their management in a clinical practice setting. Book Chapter in “Lumbar Interbody Fusion Cage” John W. Brantigan MD, Guy R. Fogel MD 2006 11. Outcomes Of L1-2 Posterior Lumbar Interbody Fusion With the Lumbar I/F Cage Reporting Unexpected Poor Fusion Results at L1-2, Guy R Fogel MD, John S. Toohey, Arvo Neidre, John Brantigan. Spine Journal July 2006 12. Is One Cage Enough in Lumbar Interbody Fusion: A Comparison of Unilateral Single Cage Interbody fusion to Bilateral Cages. Guy R. Fogel M.D., John S. Toohey M.D., Arvo Neidre M.D., John W. Brantigan M.D. Journal of Spinal Disorders and Techniques 2006 13. Repair of Pars Interarticularis Defect with a Modified Cable-Screw Construct- Case Reports Gordon R. Bozarth MD Guy R. Fogel MD John S. Toohey MD Arvo Neidre MD. Journal of Surgical Orthopedic Advances 2006 14. Fusion Assessment in Posterior Lumbar Interbody Fusion Using Radiolucent Cages: X-rays and Helical CT scan Compared with Surgical Exploration of Fusion. Guy R. Fogel M.D., John S. Toohey M.D., Arvo Neidre M.D., John W. Brantigan M.D. Accepted The Spine Journal 2007 15. Outcomes Of 9mm Width Posterior Lumbar Interbody Fusion With the Lumbar I/F Cage, Guy R Fogel MD, John S. Toohey, Arvo Neidre, John Brantigan. Accepted Journal of Surgical Orthopedic Advances 2007

SUBMITTED FOR PUBLICATION 2007 16. Improved Fusion Rates with Anterior and Posterior Fusion in Tobacco Abuse Patients Treated for Dominant Axial-Mechanical Cervical Spine Pain, G Fogel M.D., M McDonnell M.D., In Press 17. Biomechanical Comparison Of Translational and Constrained Anterior Cervical Plate With Two-level Corpectomy Of The Cervical Spine, Guy R Fogel MD, Weiqiang Liu PhD, In Press SUBMITTED FOR PUBLICATION 2008 18. Biomechanical comparison of Medicrea Cervical Staple and Cervical Plate fixation, Guy R Fogel MD, Weiqiang Liu PhD, In Press 19. Biomechanical Comparison Facetbolt to Pedicle Screw Fixation, Guy R Fogel MD, Weiqiang Liu PhD, In Press

Current as of 5/30/2008 C.V. Guy Rutledge Fogel M.D. as of 5/30/2008

20. In Vivo Study of Lumbar Fusion with Bone Marrow and Adipose Stem Cells, Guy R Fogel MD, Daria Neidre MS, In Press 21. Biomechanical Study of Edge Erosion Wear in Vertebron Pedicle Screw Tulip and Rod, Guy R Fogel MD, Weiqiang Liu PhD, In Press 22. Biomechanical Evaluation of Vertebron Cervical Cage versus Cervical Plate Fixation, Guy R Fogel MD, Weiqiang Liu PhD, In Press

RECENT PRESENTATIONS PAPERS/POSTERS

2005 1. Cervical Plates Comparison of Physical Characteristics and In Vitro Push Out Strength Texas Spine Society Austin Texas May 05 2. Anterior Hardware Removal Improves Dysphagia Following Anterior Cervical Discectomy and Fusion Cervical Spine Research Society Scottsdale Az 12/2004 3. Improved Fusion Rates with Anterior and Posterior Fusion in Tobacco Abuse Patients Treated for Dominant Axial-Mechanical Cervical Spine Pain Texas Orthopedic Association San Antonio 04 4. Total Disc Arthroplasty San Antonio Rheumatology Group January 2005 5. Stingers and Burners Spine and Sports Orthopedic Residents February 2005 6. Osteoporosis and Vertebroplasty Orthopedic Residents April 2005 2006 7. Spondylolisthesis Orthopedic Grand Rounds Methodist March 2006 8. Spinal Cord Injury Orthopedic Residents March 2006 9. Cervical Spine Trauma Orthopedic Residents April 2006 10. Spine Surgery overview for Physical Therapy Students April 2006 11. Is One Cage Enough in Lumbar Interbody Fusion: A Comparison of Unilateral Single Cage Interbody fusion to Bilateral Cages. Paper Mid American Orthopedic Association San Antonio April 2006 12. Outcomes Of L1-2 Posterior Lumbar Interbody Fusion With the Lumbar I/F Cage Reporting Unexpected Poor Fusion Results at L1-2 Paper Mid American Orthopedic Association San Antonio April 2006 13. Outcomes of Posterior Lumbar Interbody Fusion with the 9mm Width Lumbar I/F Cage and the Variable Screw Placement System. Poster Mid American Orthopedic Association San Antonio April 2006 14. Repair of Pars Interarticularis Defect with a Modified Cable-Screw Construct- Case Reports. Poster Mid American Orthopedic Association San Antonio April 2006 15. Is One Cage Enough in Lumbar Interbody Fusion: A Comparison of Unilateral Single Cage Interbody fusion to Bilateral Cages. Presented paper Mid American Orthopedic Association San Antonio April 2006 16. Biomechanical Comparison Of Translational and Constrained Anterior Cervical Plate With Two-level Corpectomy Of The Cervical Spine Paper Texas Orthopedic Association Houston May 2006 17. Fusion Assessment in Posterior Lumbar Interbody Fusion Using Radiolucent Cages: X-rays and Helical CT scan Compared with Surgical Exploration of Fusion. Paper Spine Across the Sea Maui Hi July 2006 18. Outcomes of Posterior Lumbar Interbody Fusion with the 9mm Width Lumbar I/F Cage and the Variable Screw Placement System. Poster Spine Across the Sea Maui Hi July 2006 19. Is One Cage Enough in Lumbar Interbody Fusion: A Comparison of Unilateral Single Cage Interbody fusion to Bilateral Cages. Poster Spine Across the Sea Maui Hi July 2006 2007 20. Orthopedic Spine Surgery Overview for Physical Therapists March 07

Current as of 5/30/2008 C.V. Guy Rutledge Fogel M.D. as of 5/30/2008

21. Metanalysis of Lumbar Fusion. Grand Rounds UTSA Orthopedic Department. April 2007 2008 22. Orthopedic Spine Surgery Overview for Physical Therapists March 08 23. Methodist Hospital Grand Rounds Stem Cell Therapy in Orthopedics. March 08

Current as of 5/30/2008 Coccygodynia: Evaluation and Management

Guy R. Fogel, MD, Paul Y. Cunningham III, MD, and Stephen I. Esses, MD

Abstract

Coccygodynia is pain in the region of the coccyx. In most cases, abnormal mobility incidence of fusions between seg- is seen on dynamic standing and seated radiographs, although the cause of pain is ments.4 Symptomatic patients have a unknown in other patients. Bone scans and magnetic resonance imaging may show slightly higher incidence of sacral coc- inflammation and edema, but neither technique is as accurate as dynamic radiog- cygeal fusion and a more angular sag- raphy. Treatment for patients with severe pain should begin with injection of local ittal alignment of the coccyx than anesthetic and corticosteroid into the painful segment. Coccygeal massage and stretch- asymptomatic patients.4 No pathologic ing of the levator ani muscle can help. Coccygectomy is done only when nonsur- findings have been noted in the id- gical treatment fails, which is infrequent. Coccygectomy usually is successful in care- iopathic or hypermobile segments in fully selected patients, with the best results in those with radiographically demonstrated patients with coccygodynia. The ab- abnormalities of coccygeal mobility. normal subluxation and hypermobil- J Am Acad Orthop Surg 2004;12:49-54 ity appear to cause pain, but coccygodynia can occur with an immobile coccyx. Only occasionally have posttraumatic nonunion, arthritis, or Coccygodynia, pain in the region of Anatomy malunion been found in surgical spec- the coccyx, typically is triggered by imens. or occurs while sitting. The intensity Five fused sacral and three or four of the pain varies and sometimes is fused coccygeal vertebrae form the ter- aggravated by arising from a seated minal end of the spinal column. Pri- Etiology position. Less severe symptoms may mary ossification centers are evident be managed by changing the sitting in the 9th to 10th week of gestation The clinical factors associated with position or injecting the painful area in the axial skeleton, but the coccyx coccygeal abnormalities are obesity, with local anesthetic and corticoste- does not begin to ossify until after birth. roids. Orthopaedic surgeons often In the sagittal plane, the sacrum is ky- see patients with more severe and photic and connects the lumbosacral Dr. Fogel is Spine Fellow, Department of Ortho- disabling symptoms, which in se- junction to the coccyx. The inferior sa- pedic Surgery, Baylor College of Medicine, Hous- lected cases can be managed suc- cral apex at S5, the sacral cornu, ar- ton, TX. Dr. Cunningham is Medical Research cessfully with surgery. ticulates with the coccygeal cornu, a Fellow, Department of Orthopedic Surgery, Bay- Coccygodynia is five times more facet-disk complex on the dorsal ar- lor College of Medicine. Dr. Esses is Professor of Orthopedics and Brodsky Chair of Spinal Surgery, prevalent in women than men. Al- ticulating surface of the coccyx. This Department of Orthopedic Surgery, Baylor Col- though it can occur over a wide age articulation can be a symphysis or a lege of Medicine. range, mean age of onset is 40 years. synovial joint. The coccyx is a trian- Coccygodynia has many causes, but gular structure comprising three or four None of the following authors or the departments it may be posttraumatic, beginning coccygeal units that usually are fused, with which they are affiliated has received anything of value from or owns stock in a commercial com- after a fracture or contusion or after although the first coccygeal segment pany or institution related directly or indirectly a difficult vaginal delivery.1-3 In the may not fuse with the second. The sa- to the subject of this article: Dr. Fogel, Dr. Cun- largest number of cases, the tip of crococcygeal joint also may be fused. ningham, and Dr. Esses. the coccyx is subluxated or hyper- The coccyx provides attachments for mobile, which can be seen on dy- the gluteus maximus muscle, coccygeal Reprint requests: Dr. Esses, Baylor College of Med- icine, Suite 1900, 6560 Fannin, Houston, TX namic radiographs taken with the muscle, and anococcygeal ligament. 77030. patient standing and seated1-3 (Fig. There are no significant differenc- 1). The cause of pain is unknown in es between asymptomatic patients and Copyright 2004 by the American Academy of patients with normal coccygeal mo- those with coccygodynia in the num- Orthopaedic Surgeons. bility.2 ber of coccygeal segments or in the

Vol 12, No 1, January/February 2004 49 Coccygodynia: Evaluation and Management

ecation. Women with a history of vaginitis, discharge, or associated pelvic pain should be referred for gyne- cologic consultation. Concomitant constipation should be managed ap- propriately. If the patient has blood in the stool, a tumor or metastasis should be considered.

Physical Examination The surrounding skin and soft tissue should be inspected for evidence of pilonidal cysts or fistulas. External Figure 1 Maigne’s technique1 for comparing positions of sacral and coccygeal vertebrae palpation or rectal examination may from lateral standing and seated radiographs. A, Standing view. B, Seated view. C, Super- reveal bone spicules, local swelling, imposed views with the sacrum aligned by rotating the seated view through an angle of sag- or coccygeal masses. The coccyx ittal pelvic rotation (I) shows coccygeal angulation and subluxation (II) and the angle at which the coccyx strikes the seat surface (III). should be palpated externally,and the distal segment should be manipulated rectally to detect pain generated antecedent trauma or childbirth, and and sinuses, and perirectal abscess- by motion of the coccygeal segments. exacerbation of pain with arising es.4 Low back pain has occurred con- Local tenderness may occur on the su- from sitting. Obesity, which decreas- currently with coccygodynia. In one perficial coccygeal surface or only on es pelvic rotation when the patient series, 24 of 50 patients (48%) also had manipulation of the coccygeal tip by sits, is three times more common in lumbar disk herniation or bulge with- rectal examination. Tenderness may patients with coccygodynia than in out sciatica or low back pain,5 so al- be greatest at the sacrococcygeal joint the normal population.2 However, the though low back pain is common, it rather than at the coccygeal tip.2 A incidence of radiographically demon- appears to be separate from coccygo- palpable internal mass (eg, chordo- strated coccygeal instability is the dynia. In the series of Postacchini and ma) on the anterior surface of the coc- same regardless of a history of remote Massobrio,4 87% of patients with low cyx or sacrum may be evident on rec- trauma, which suggests that only re- back pain at the time of coccygecto- tal examination. Every examination cent trauma (within 3 months) caus- my had an excellent or good result. should include a stool guaiac to de- es coccygodynia. Patients with nor- The only poor results were in patients tect occult blood. mal coccygeal mobility have the with low back pain and a configura- idiopathic type, which may be asso- tion of the coccyx wherein the first ciated with pelvic floor spasticity or coccygeal segment is partially fused Imaging Evaluation other anomalies of the midpelvic to the sacrum. muscles. Coccygodynia occurring Single-Position Radiographs with an immobile coccyx frequently Because orthogonal anteroposteri- is associated with bursitis of the ad- Clinical Evaluation or and lateral radiographs of the coc- ventitia at the coccygeal tip. Other cyx seldom show differences between suggested etiologies include post- Symptoms normal individuals and those with traumatic arthritis of the sacrococ- Onset of pain may be insidious, re- coccygodynia, these views are usual- cygeal joint and ununited fractures sulting in a possibly long delay from ly not diagnostic.5 Postacchini and or dislocations of the coccyx. Abnor- onset to diagnosis. Patients usually Massobrio4 reported that, in both mal psychological states as a cause of present with pain in and around the healthy and symptomatic patients, coccygodynia have been largely dis- coccyx without significant low back 83% to 95% had two or three coc- proved as behavioral testing of these pain or radiating or referred pain. The cygeal segments. The sacrococcygeal patients reveals personality profiles pain is localized to the sacrococcygeal joint was fused in 51% of the patients similar to other patient groups.5 Oth- joint or mobile segment of the coccyx with idiopathic coccygodynia (26/ er causes of pain in the region of the and may be relieved by sitting on the 51), usually the first intercoccygeal sacrum and coccyx are lesions of the legs or on either buttock. Chronic joint was mobile, and the second in- lumbar disks, arachnoiditis of the pain is that which persists >2 months. tercoccygeal joint was fused in 49% lower sacral nerve roots, tumors of Patients may feel a frequent need to (25/51).4 The curvature of the sacrum the coccyx or sacrum, pilonidal cysts defecate or may have pain with def- was flexed or anterior in 68% of

50 Journal of the American Academy of Orthopaedic Surgeons Guy R. Fogel, MD, et al asymptomatic patients and in 31% of patients with coccygodynia. The tip curved forward sharply in 23% and subluxated posteriorly in 22% of patients with coccygodynia.4

Dynamic Radiographs Maigne and colleagues1-3,6 compared standing and sitting lateral radiographic views of the coccyx in a total of 582 patients with coccygodynia and reported abnormalities in 70%. The normal coccyx pivots slightly (5° to 25°) either posteriorly or an- teriorly with sitting and returns to its Figure 2 The anatomic signs of coccygodynia.1 A, Normal standing appearance of the coc- original position with standing (Fig. cyx. B, Increased flexion mobility of the coccyx when patient is seated. C, Posterior sublux- 2, A). Abnormalities of the coccygeal ation of the coccyx when patient is seated. D, Coccygeal spicule (arrow) arising from the dorsal surface of coccygeal segment. segments in the seated views have anterior hypermobility >25° (Figs. 2, B and 3). Subluxation or posterior displacement of the mobile segment of ma. Physical therapy consisting of injections and manipulations were re- the coccyx is seen when the patient diathermy and ultrasound may pro- peated. If the patient did not respond, is seated (Figs. 2, C and 4, A). A spic- vide temporary relief. a coccygectomy was done 6 weeks ule of the distal tip (Fig. 2, D) is seen Wray et al5 used a stepwise treat- later. The cure rate with injection most commonly with an immobile ment, with each step somewhat more alone was 59% but was 85% with ma- coccyx (<5° of motion with sitting).2 invasive than the previous. First, nipulation and injection.5 Although methylprednisolone (40 mg) and relapses occurred in the injection Advanced Imaging Modalities bupivacaine (10 mL 0.25%) were in- group (21%) and the manipulation Technetium Tc 99m bone scans jected around the side and tip of the group (28%), repeat treatment in each may show inflammation in the area coccyx. With persistent coccygodynia, group achieved good success. Coc- of subluxation or hypermobility. the coccyx was reinjected and manip- cygectomy was done in 20% of pa- Magnetic resonance images also may ulated under general anesthesia by tients and had a success rate of 91%. demonstrate edema in such areas of repeatedly flexing and extending it Maigne and Chatellier6 prospec- inflammation (Fig. 4, B). Neither im- for a minute. If the treatment was suc- tively compared levator ani muscle aging modality can definitively diag- cessful initially but pain recurred, the massage, joint mobilization, and mild nose coccygodynia, nor are they as accurate as the compared standing and seated dynamic radiographs. Provoc- ative tests, such as needling of the painful coccygeal joint to produce pain, and relief with injection of lo- Figure 3 Lateral radiographic seated view of a 45-year-old woman with increased- cal anesthetic are helpful in diagnos- flexion coccygodynia (arrow). She fell on her ing all patients with subluxation or buttocks 2.5 years before presentation and had hypermobility and nearly half of progressive symptoms, including painful 3 bowel movements and inability to sit on a those with normal mobility. chair. Physical examination revealed a tender and nonmobile coccyx with an exquisitely tender distal coccygeal segment on rectal examination. At 1 year postcoccygectomy, she had Nonsurgical Management complete relief from her symptoms and was able to sit normally. (Courtesy of Paul A. Nonsurgical management options in- Anderson, MD, Madison, WI.) clude nonsteroidal anti-inflammatory and analgesic medications, rest, hot baths, and a cushion to protect the coccygeal region from repetitive trau-

Vol 12, No 1, January/February 2004 51 Coccygodynia: Evaluation and Management

sisting of complete coccygectomy or simply excision of the mobile segment, should be done only after nonsurgical management fails.4 Absence of physical ﬁndings or signiﬁcant abnormal psychologic evaluation is a contraindication to surgery. In carefully selected patients, especially those with dynamic radiographic instability or hypermobility, success rates range from 60% to 91%4,5,7-11 (Table 1). In a patient with normal coccygeal mobility who fails nonsurgical care, coccygectomy may be recommended. However, the surgical result is less predictably favorable than in patients with dynamic radiographic coccygeal instability or hypermobility.

Technique The patient should take an oral mechanical bowel preparation, such as saline and polyethylene glycol so- Figure 4 A 19-year-old man fell on his buttocks 2 years before presentation and had im- lution (4 L), the day before surgery. mediate onset of coccygodynia. His symptoms were chronic and disabling. A, Lateral ra- Oral neomycin, erythromycin, or diograph demonstrates posterior subluxation of the coccygeal mobile segment. B, Sagittal metronidazole are given three times T2-weighted spin-echo magnetic resonance image shows edema of the distal coccygeal segments, especially the subluxated coccygeal segment. the day before surgery. Appropriate prophylactic antibiotics for bowel surgery are given 1 hour before sur- stretching of the levator ani, without ration), 8 weeks of rest with use of a gery. Postoperatively, most authori- the addition of injections. At 6 stool softener, adjustable seating, and ties recommend two or three more months, successful treatment was nonsteroidal anti-inflammatory med- doses.12 29.2% with massage, 16% with mo- ication are prescribed. If this fails to Coccygectomy is done with the pa- bilization, and 32% with stretching, relieve the symptomatic coccygo- tient in the prone position with the for a total of 25.7% overall success. dynia or the patient presents with hips and knees flexed. A vertical in- When a patient had a satisfactory re- chronic symptoms (>2 months’ dura- cision is made over the coccyx, ex- sult, it was invariably achieved with- tion), a workup is done, including tending from just above the sacrococ- in a week. Good results tended to re- standing and seated radiographs of cygeal joint into the buttock crease main stable. Patients with normal the coccyx in addition to MRI to eval- without extending it to the perianal mobility of the coccyx fared the best uate for injury edema, tumor, or oth- skin. The incision is carried down (43% success at 6 months). Those with er pathology. Stretching, massage, through the fascia and gluteus max- an immobile coccyx fared the worst and injection are usually initiated at imus muscle with meticulous dissec- (16%). The outcomes for those with this time. If these treatments are un- tion directly to the bone. Subperi- instability subluxation (22.2%) and successful or if pain recurs, coccygec- osteal dissection should be done by hypermobility (25%) were modestly tomy may be offered. gradually working side to side. The successful. Massage and stretching coccyx then should be elevated either were more effective than manipula- by working from the tip proximally tion. When therapy failed, patients Surgical Management or from the side, under direct vision. went on to injections or surgery. All segments must be removed. The Based on these studies, we have The indication for coccygectomy is sig- tip of the coccyx, which is most like- developed an evaluation and treat- nificant, disabling coccygodynia with ly to be left inadvertently, should be ment algorithm for coccygodynia5-7 radiographic subluxation; instability; separated sharply from the underly- (Fig. 5). When a patient presents with or a spicule, particularly on the tip of ing rectum and dense fascia. With acute coccygodynia (≤2 months’ du- an immobile coccyx.2,7 Surgery, con- blunt dissection, the rectum and

52 Journal of the American Academy of Orthopaedic Surgeons Guy R. Fogel, MD, et al

History of coccygodynia confined to coccyx, no associated back pain

≤8 weeks >8 weeks

Rest; avoid sitting on coccyx; use Standing and seated radiographs stool softener, NSAIDS, analgesics as needed for 8 weeks

Normal Hypermobility, subluxation, mobility spicule on tip of coccyx Not improved

MRI coccyx for edema, tumor, other etiology Symptomatic relief

May offer stretching/massage of levator ani muscle and/or corticosteroid/anesthetic injection in areas of radiographic abnormality

If successful, Unsuccessful may repeat

Recurrence Offer coccygectomy

Unsuccessful; Successful; continued complete relief coccygodynia in 3-6 months

Figure 5 Evaluation and management of coccygodynia.

dense fascia then are freed to the lev- tation of the coccygeal specimen and tention, the result may be successful. el of the sacrococcygeal joint. This limit damage to the rectum and its Wound infection and delayed wound joint then may be transected and the venous drainage, which can decrease healing rates range from 2% to coccyx removed. postoperative drainage and increase 22%.4,5,7-10 In a very thin patient with The wound bed should be closely the risk of infection. a kyphotic sacrum, the remaining end examined and palpated for any re- of the sacrum may be prominent and maining segments of bone. Radio- Complications be a source of continued pain that is graphs generally are not necessary. The primary complication is not easily managed. The end of the sacrum may be perineal contamination of the wound. smoothed by rasp, rongeur, or burr. Usually a wound infection is super- Bleeding from the hemorrhoidal ﬁcial or is a simple wound dehiscence Summary venous complex of the rectum may and will heal with intravenous anti- require ligation. Meticulous hemosta- biotics and local wound care. Deep Dynamic radiographs can help iden- sis and a tight layered closure to oblit- infection must be débrided and tify causative abnormalities in most erate dead space should be done. This drained. Antibiotics should be given. patients with coccygodynia, and technique should decrease fragmen- Even with healing by secondary in- those with normal coccygeal mobil-

Vol 12, No 1, January/February 2004 53 Coccygodynia: Evaluation and Management

Table 1 Outcomes From Coccygectomy

Study No. Patients Follow-up (yrs) Outcome

Postacchini and 36 7.8 (mean) 12 excellent, 20 good, 2 fair, 2 poor Massobrio4 Wray et al5 23 2.75 (mean) 21 excellent Maigne et al7 37 2 (minimum) 23 excellent, 11 good, 3 poor Hellberg and 55 15 (mean) 32 cured, 13 improved, 5 slightly improved, 5 dissatisﬁed Strange-Vognsen8 Grosso and van Dam9 9 4.7 (mean) 3 complete, 5 partial, 1 slight relief Eng et al10 27 1.5 (minimum) 9 cured, 9 improved, 6 slightly improved, 3 not improved Bayne et al11 48 7 (mean) 29 acceptable ity also may be successfully treated. nonsurgical management fails, coc- normal coccygeal mobility who fail Corticosteroid and anesthetic injec- cygectomy is usually successful in nonsurgical care, but results are less tions combined with massage or carefully selected patients with dy- predictably favorable than in patients stretching of the levator ani muscle namic instability. Coccygectomy may with dynamic radiographic coccygeal are successful in most patients. When be recommended in patients with instability or hypermobility.

References

1. Maigne JY, Tamalet B: Standardized ra- 4. Postacchini F, Massobrio M: Idiopathic coccyx. Acta Orthop Scand 1990;61: diologic protocol for the study of com- coccygodynia: Analysis of ﬁfty-one op- 463-465. mon coccygodynia and characteristics erative cases and a radiographic study 9. Grosso NP, van Dam BE: Total coc- of the lesions observed in the sitting po- of the normal coccyx. J Bone Joint Surg cygectomy for the relief of coccygo- sition: Clinical elements differentiating Am 1983;65:1116-1124. dynia: A retrospective review. J Spinal luxation, hypermobility, and normal 5. Wray CC, Easom S, Hoskinson J: Coc- Disord 1995;8:328-330. mobility. Spine 1996;21:2588-2593. cydynia: Aetiology and treatment. 10. Eng JB, Rymaszewski L, Jepson K: Coc- 2. Maigne JY, Doursounian L, Chatellier J Bone Joint Surg Br 1991;73:335-338. cygectomy. J R Coll Surg Edinb 1988;33: G: Causes and mechanisms of common 6. Maigne JY, Chatellier G: Comparison of 202-203. coccydynia: Role of body mass index three manual coccydynia treatments: A 11. Bayne O, Bateman JE, Cameron HU: and coccygeal trauma. Spine 2000;25: pilot study. Spine 2001;26:E479-E484. The inﬂuence of etiology on the results 3072-3079. 7. Maigne JY, Lagauche D, Doursounian L: of coccygectomy. Clin Orthop 1984;190: 3. Maigne JY, Guedj S, Straus C: Idiopath- Instability of the coccyx in coccydynia. 266-272. ic coccygodynia: Lateral roentgenograms J Bone Joint Surg Br 2000;82:1038-1041. 12. Norden CW: Antibiotic prophylaxis in in the sitting position and coccygeal dis- 8. Hellberg S, Strange-Vognsen HH: Coc- orthopedic surgery. Rev Infect Dis 1991; cography. Spine 1994;19:930-934. cygodynia treated by resection of the 13(suppl 10):S842-S846.

54 Journal of the American Academy of Orthopaedic Surgeons

The Spine Journal 3 (2003) 238–241

Hip spine syndrome: management of coexisting radiculopathy and arthritis of the lower extremity Guy R. Fogel, MDa, Stephen I. Esses, MDb,* aSpine Fellow and bDepartment of Orthopedics, Baylor College of Medicine, 6560 Fannin Suite 1900, Houston, TX 77030, USA Received 20 February 2002; accepted 31 July 2002

Abstract Background context: Significant lumbar spinal stenosis and lower extremity arthritis may coexist in the elderly. This combination of lumbar stenosis with radiculopathy and lower extremity arthritis may lead to diagnostic uncertainty. Purpose: To describe the findings of hip spine syndrome, a constellation of symptoms with extensive overlap of radiculopathy and lower extremity arthritis. Conclusions: Evaluation of the patient with lower extremity pain in consideration for total joint arthroplasty should include functional inquiry of the spinal nerves. Diagnostic tests and injections may allow an informative weighting of the patient’s symptoms, leading to a better understanding of the patient’s pain syndrome. There is a group of patients who have a total hip arthroplasty and then develop or may continue to have pain of groin and buttock, secondary to sciatica of lumbar spinal stenosis. For the patient undergoing total hip arthroplasty with asymptomatic spinal stenosis, there may be increased neurological risk at surgery, related to the stenosis. The patient with both conditions may require surgical decompression of the lumbar stenosis as well as joint arthroplasty of the arthritic joint. © 2003 Elsevier Inc. All right reserved.

Keywords: Aged; Aged, 80 and over; Hip joint; Hip prosthesis/adverse effects; Laminectomy; Myelography; Pain/etiology/ therapy; Spinal stenosis/complications/radiography/surgery

Introduction the patient’s pain syndrome. The patient with both conditions may require surgical decompression of the lumbar Significant lumbar spinal stenosis and lower extremity stenosis as well as joint arthroplasty of the arthritic joint. arthritis may coexist in the elderly. Which diagnosis leads to Osteoarthritis is the most common musculoskeletal dis- the patient’s lower extremity complaints? Clinically, symp- ease of aging. It has been reported radiographically in more toms of lumbar spinal stenosis, radiculopathy and neuro- than 80% of individuals older than 55 years. Arthritis is genic claudication may be similar to the pain of an arthritic present in more than 40 million US citizens. Osteoarthritis hip or knee. Additionally, an asymptomatic lumbar stenosis of the spine, hip and knee may result in significant impair- may become activated after a total joint arthroplasty and ment, disability and loss of function. Radiographic evidence cause unacceptable levels of radicular pain or newfound of osteoarthritis of the knee has been reported in 40% of pa- muscular weakness, that is, foot drop, thigh or hip abductor tients 80 years or older. Similarly, radiographic arthritis of weakness. Evaluation of the patient with lower extremity the hip has been reported in 12% of those older than 80 pain in consideration for total joint arthroplasty should in- years [1]. Likewise, spondylosis of the lumbar spine is prev- clude functional inquiry of the spinal nerves. Diagnostic alent, and lumbar stenosis, a subset of arthritis, is not un- tests and injections may allow an informative weighting of common. Some aspects of stenosis, such as spondylolisthe- the patient’s symptoms leading to a better understanding of sis, may be seen in 6% of men and 9% of women from a Dutch cross-section study. Computed tomography (CT) FDA device/drug status: not applicable. scan may be a better measure of stenosis, with reported inci- Nothing of value received from a commercial entity related to this dence of 3.4% in individuals older than 40 years. Magnetic research. resonance imaging (MRI) in asymptomatic individuals may * Corresponding author. Department of Orthopedics, Baylor College of Medicine, 6560 Fannin Suite 1900, Houston, TX 77030, USA. Tel.: report even higher incidence of stenosis. Borenstein et al. (713) 986-5740; fax: (713) 986-5741. performed MRI studies of volunteers and reported a preva- E-mail address: [email protected] (S.I. Esses) lence of 60% in those older than 60 years [2]. 1529-9430/03/$ – see front matter © 2003 Elsevier Inc. All rights reserved. PII: S1529-9430(02)00453-9

G.R. Fogel and S.I. Esses / The Spine Journal 3 (2003) 238–241 239 The causes of hip arthritis may be multifactorial and in- pattern. The most common symptom of lumbar spinal clude osteoarthropathy, spondyloarthropathy and avascular stenosis is neurogenic claudication. Neurogenic claudica- necrosis. In general, these patients will have buttock and tion, or pseudoclaudication, refers to pain radiating to the groin pain radiating to the knee, but not below, associated lower extremities that begins and worsens as the patient am- with decreased range of motion. As the arthritis progresses, bulates. The pain worsens with the increased lordosis of the gait disturbances may be seen. Total joint arthroplasty has lumbar spine, with standing and walking, which may rela- made a significant positive impact on the treatment of pa- tively pinch close the foramina and narrow the canal with tients with disabling arthritis. Total hip arthroplasty is one infolding of the ligamentum flavum. The pain often resolves of the most successful surgical procedures. More than spontaneously or resolves rapidly as the patient bends for- 120,000 primary arthroplasties are performed in the United ward or sits down. In contrast, muscular claudication of vas- States each year at an estimated cost of over $2.5 billion. cular origin will produce symptoms with walking up hills or Because most hip replacements are done in patients older bicycling. Alternately, absence of pulses below the hips, ru- than 65 years of age, the number of procedures is expected bor and pallor changes with elevation are classic for vascu- to increase as the population ages. Outcome studies have lar claudication but not neurologic claudication. In uncer- documented marked improvements in pain, sleep, range of tain cases vascular Doppler, flow studies or arteriography motion and physical ability after total hip arthroplasty. may be required. Patients who present with concomitant os- Overall, the data have documented that there is a significant teoarthritis of the lower extremities (hip, knee, ankle), caus- improvement in functional status and quality of life. ing restricted and painful joint range of motion will require Beattie et al. [3] looked at lumbar MRI in the symptom- special consideration. In certain individuals, there may be atic lumbar stenosis patient and found that only the most se- isolated areas of extremity pain at the hip, knee, calf, ankle vere compression situations, such as severe foraminal en- or heel, with asymptomatic areas between painful foci [5]. croachment or large disc extrusion, would correlate with the The most important features of the physical examination patient’s symptoms. Mild to moderate compression, disc in stenosis are the strength, reflex and the palpatory exami- degeneration and central canal stenosis were not signifi- nations. Strength deficits, such as partial foot drop and hip cantly associated with the reported patient pain pattern. flexor or quadriceps weakness may be subtle. In patients Lawrence et al. [1] reviewed hospital discharge registry for with more severe chronic lumbar spinal stenosis, there may the state of Washington from 1986 to 1988, reporting ad- be visible atrophy of calf musculature. The patient may be missions for lumbar spine operations in 18,122 patients, of unable to to walk. Reflexes often are diminished at the which 3,380 (18.6%) were for diagnosis of spinal stenosis. ankles in patients with more significant lumbar stenosis. If 18.6% were representative of the United States, then Straight leg raising, a helpful clinical sign in disc herniation, 52,000 operations for stenosis nationwide were done yearly usually is negative in patients with spinal stenosis, except in [1]. Stenosis may begin at an earlier age, related to develop- patients in whom disc herniation is superimposed in a region mental stenosis. Eighty percent of patients will have back of stenosis. However, the femoral stretch test often is posi- and leg pain, 62% have symptoms of pseudoclaudication tive, even when only L5 and S1 roots are involved. This may and 50% will have neurological deficit. Hall et al. [4] re- occur because of the end effect of increasing lordosis and ported a known association of lumbar stenosis and periph- pelvic tilt increasing the stenosis and stretching the femoral eral vascular disease. roots. Areas of sensory deficit usually correlate poorly with The true incidence of the combination of lumbar stenosis patients’ pain and disability. Decreased range of motion of and hip arthritis is unknown. There can be a substantial joints, with pain at extremes of motion, is commonly found overlap in symptoms of stenosis and hip arthrosis. The in arthritis of the hip and knee. There may be an effusion or stenosis may be asymptomatic while the hip is symptom- induration of the knee, associated with arthritis of the knee. atic. Clinical symptoms arising from the hip often prompt The clinician must evaluate leg pain carefully. Leg pain not medical treatment. Or the stenosis may be the more obvious attributable to radiculopathy may result from hip bursitis, problem, demanding more immediate surgical attention. osteoarthritis of the hip or knee or myofascial pain. Lumbar stenosis or spondylolisthesis causing radiculopathy Plain radiographs may demonstrate osteoarthritis of the hip is usually recognized and treated. There is a group of pa- and lumbar spine (Fig. 1). The plain radiographs are poor tients who have a total hip arthroplasty and then develop or measures of lumbar stenosis, and CT myelography or MRI may continue to have pain of groin and buttock, secondary scanning is far better at defining stenosis. A block with local to sciatica of lumbar spinal stenosis. anesthetic at the most symptomatic location may be of diagnostic benefit. A hip injection with local anesthetic may allow the confident diagnosis of hip disease. Kleiner et al. [6] report Diagnosis injection of 10 mm of bupivacaine HCl into the hip with X-ray Although a careful history and physical examination of- localization to discern the origin of hip pain [6]. They reported ten may differentiate pain of radicular versus joint origin, identification of the hip as the correct source of pain in 88% of sometimes the distinction is difficult. Leg pain should be the total cases. Alternately, in the case of a suspected radicul- described in detail to see whether it simulates a dermatomal opathy, a nerve root block with anesthetic will be nondiagnos-

240 G.R. Fogel and S.I. Esses / The Spine Journal 3 (2003) 238–241 tic. However, the addition of a steroid may give therapeutic from component loosening, hematogenous infection, soft-tis- pain relief [7]. If the radiculopathy is within a dermatome, im- sue problems such as tendonitis or bursitis or radiculopathy of plying a single or perhaps two nerve roots, these nerve roots lumbar stenosis. The location of the pain is important in iden- may be anesthetized with a fluoroscopic guided selective tifying its source. Groin pain usually is caused by acetabular block, but there is low specificity with such blocks. Adjacent component loosening, which can also cause buttock pain. Ili- nerve roots are affected by an injection, and in other cases opsoas inflammation is more common with collared implants there may be some effect of distal injection on referred pain. overhanging the medial calcar, or impingement on prominent cementless sockets causing groin pain. Groin pain also may be caused by an L4–L5 stenosis [8]. Elderly patients with L4–L5 Pain after arthroplasty protruding herniation of the annulus fibrosus were most likely Although total hip arthroplasty is one of the most success- to experience groin pain. The sinuvertebral nerve that inner- ful procedures in orthopedic surgery, a small percentage of pa- vates the posterior annulus fibrosus, the posterior longitudinal tients have pain after surgery. A systematic approach to the ligament and the dura was indicated as the afferent nerve caus- evaluation of this pain, with careful attention to the patient’s ing groin pain [9]. Thigh or knee pain may be the result of history, physical examination and laboratory and radiographic femoral component loosening. studies, is necessary to reach a correct diagnosis. The time of The coexistence of lumbar stenosis with hip arthritis may onset of hip pain is important. Pain that occurs early after sur- be an increased risk factor for neurologic injury with total gery and is out of proportion to the usual postoperative pain hip arthroplasty. Pritchett [10] reported 21 patients with may indicate a postoperative infection, an unstable implant, lumbar stenosis who developed foot drop after total hip ar- heterotopic ossification or radiculopathy. Late pain may result throplasty. This implies that less nerve compression is re-

Fig. 1. Lumbar spinal stenosis Anteroposterior lateral radiographs. (Left) Previous total hip arthroplasty on right hip. (Right) Severe arthritis, left hip. Note associated aortic vascular calcification on lateral.

G.R. Fogel and S.I. Esses / The Spine Journal 3 (2003) 238–241 241 quired to produce symptoms in the presence of these coex- now, a lumbar decompression may be required in the future isting conditions. Pritchett concludes that if nerve injury if the stenosis becomes symptomatic. occurs after hip surgery, lumbar spinal stenosis should be Although progressive neurologic deficit or cauda equina considered, and some patients will improve after lumbar syndrome in association with lumbar spinal stenosis are in- stenosis decompression. Bohl and Steffee [11] theorized dications for urgent operative intervention, Bohl and Steffee that increased walking endurance afforded by the new hip [11] suggested that nonprogressive neurologic deficit (pin arthroplasty may unmask neurogenic claudication, in pa- prick, vibration, reflexes, leg muscle power) correlated poorly tients incapable of walking that distance before total hip ar- with pain and physical disability and therefore should not be throplasty [11]. Although clinically apparent nerve injury is a reason for operative intervention. uncommon, detectable neurological loss after total hip arthroplasty is more common than postoperative infection. Summary Schmalzried et al. [9] found a 1.7% incidence of postoperative neuropathy, mostly sciatic, noted clinically. In most pa- Significant lumbar spinal stenosis and lower extremity ar- tients the cause of the neuropathy is unknown. This author thritis may coexist in the elderly. Evaluation of the patient did not comment about associated lumbar stenosis. McNamara et with lower extremity pain in consideration for total joint ar- al. [12] report 14 patients who underwent lumbar stenosis throplasty should include functional inquiry of the spinal release after total hip arthroplasty [12]. Five patients ini- nerves. Diagnostic tests and injections may allow a better un- tially presented with symptoms of both joint disease and derstanding of the patient’s pain syndrome. There is a group spinal stenosis. Nine patients began to have radicular pain of patients who have a total hip arthroplasty and then develop after their total joint replacement. Eight of nine who under- or may continue to have pain of groin and buttock, or neuro- went decompression of lumbar stenosis after the total joint claudication secondary to sciatica of lumbar spinal stenosis. arthroplasty were rated as “well” to “excellent” outcome. The For the patient undergoing total hip arthroplasty with asymp- authors noted the average time to presentation with symptomatic spinal stenosis, there may be increased neurological toms for spinal stenosis after arthroplasty was 9.3 months. risk at surgery, related to the stenosis. The patient with both It is important to determine whether the patient has pain conditions may require surgical decompression of the lumbar with activity, pain at rest or both. Pain caused by component stenosis as well as joint arthroplasty of the arthritic joint. instability usually is activity related. Lumbar spinal stenosis with radiculopathy may be painful at rest or have a pseudoclaudication nature that may be confused with pain of References component instability. Clinically, a history of “start-up” pain [1] Lawrence RC, Helmick CG, Arnett FC, et al. Estimates of the preva- may indicate a loose component. The patient may report that lence of arthritis and selected musculoskeletal disorders in the United after 5 or 10 steps there is less pain in the groin. Pain with a States. Arthritis Rheum 1998;41(5):778–99. loose component often is triphasic in that the first few steps [2] Borenstein DG, O’Mara JW Jr, Boden SD, et al. The value of magnetic resonance imaging of the lumbar spine to predict low-back pain in as- cause acute sharp pain, the pain lessens with more walking ymptomatic subjects: a seven-year follow-up study. J Bone Joint Surg and then with a moderate amount of walking, pain again in- Am 2001;83-A(9):1306-11. creases. With loose components, the pain is either in the groin [3] Beattie PF, Meyers SP, Stratford P, et al. Associations between pa- or thigh, with occasional buttock pain caused by a loose tient report of symptoms and anatomic impairment visible on lumbar socket. Pain that occurs at night, or at rest, should suggest the magnetic resonance imaging. Spine 2000;25(7):819–28. [4] Hall S, Bartleson JD, Onofrio BM, et al. Lumbar spinal stenosis. possibility of spinal stenosis. It also is important to determine Clinical features, diagnostic procedures, and results of surgical treat- the presence of peripheral vascular disease, because activity- ment in 68 patients. Ann Intern Med 1985:103(2):271–5. related pain may also result from vascular claudication. [5] Friis ML, Guliksen GC, Rasmussen P, et al. Pain and spinal root compression. Acta Neurochir [Wien] 1977;39(3–4):241–9. [6] Kleiner JB, Donaldson WF III, Curd JG, et al. Extraspinal causes of Treatment lumbosacral radiculopathy. J Bone Joint Surg Am 1991;73(6):817–21. [7] North RB, Kidd DH, Zahurak M, et al. Specificity of diagnostic nerve Which element should be treated first? McNamara [12] blocks: a prospective, randomized study of sciatica due to lumbosac- recommended total joint arthroplasty first because of the de- ral spine disease. Pain 1996;65(1):77–85. monstrable excellent results with arthroplasty. In patients [8] Yukawa Y, Kato F, Kajino G, et al. Groin pain associated with lower with concomitant stenosis and coxarthrosis, the patient lumbar disc herniation. Spine 1997;22(15):1736–40. [9] Schmalzried TP, Amstutz HC, Dorey FJ. Nerve palsy associated with should be counseled that two operations may be required to total hip replacement. Risk factors and prognosis. J Bone Joint Surg treat the problem, a hip arthroplasty and a lumbar decom- Am 1991;73(7):1074–80. pression. For the patient with asymptomatic spinal stenosis [10] Pritchett JW. Lumbar decompression to treat foot drop after hip ar- undergoing total hip arthroplasty, there may be increased throplasty. Clin Orthop 1994;303:173–7. neurological risk at surgery, related to the stenosis. How- [11] Bohl WR, Steffee AD. Lumbar spinal stenosis. A cause of continued pain and disability in patients after total hip arthroplasty. Spine 1979; ever, the authors would not recommend a decompression of 4(2):168–73. an asymptomatic lumbar stenosis. Again, the patient may be [12] McNamara MJ, Barrett KG, Christie MJ, et al. Lumbar spinal steno- counseled that although the stenosis may be asymptomatic sis and lower extremity arthroplasty. J Arthroplasty 1993;8(3):273–7.

The Spine Journal 3 (2003) 118–124

Cervical plates: comparison of physical characteristics and in vitro pushout strength Guy R. Fogel, MD, Weiqiang Liu, PhD, Charles A. Reitman, MD, Stephen I. Esses, MD* Department of Orthopedic Surgery, Baylor College of Medicine, 6560 Fannin Suite 1900, Houston, TX 77030, USA Received 27 February 2002; accepted 22 May 2002

Abstract Background context: There are many cervical plates available to the spine surgeon today. A single plate design may not be appropriate for every clinical situation. It is important for the surgeon to understand the differences of these plating systems. Plate systems are known to fail by screw pullout from the bone, screw and plate breakage and a less frequent but clinically observed screw pushout from the plate. Pushout testing of the screws from the plate have not previously been subjected to study. Purpose: This compares the features of cervical plating systems and the strength of the locking mechanisms to allow the surgeon to make a knowledgeable choice of plating system. Study design: This is a review of descriptive geometric characteristics of cervical plate systems and a biomechanical evaluation of locking mechanism screw pushout strength. Methods: Physical characteristics of each plate were determined. Features of plates and screws were cataloged. Each of the test plate systems had a different locking mechanism. Biomechanical testing of the locking mechanism–screw-plate constructs was performed to determine the pushout strength of the fixation screw from the plate-locking mechanism. Results: Physical characteristics of the plating systems, including lengths, widths, shortest screw lengths and distance from edge of plate to nearest screw, were determined. Biomechanical testing showed significant differences in pushout strength, in part explained by the type of locking mechanism. Conclusions: Biomechanical screw pushout data demonstrate that a significant range of pushout strengths exist across the available cervical plate systems today. Knowing the physical characteristics of the cervical plating systems available may allow the selection of a plate best suited for a given clinical situation. © 2003 Elsevier Science Inc. All rights reserved.

Keywords: Biomechanics; Bone plates; Bone screws; Cervical vertebrae; Surgery; Equipment design; Spinal fusion/instrumentation

Introduction ders has been more controversial but recent studies support the use of plating, especially for multiple-level fusions [3–8]. Anterior plating systems have become increasingly pop- The plate systems have been recommended to decrease the ular for fixation of the subaxial cervical spine. Internal fixa- pseudarthrosis rate in cervical surgery [9–15]. Other advan- tion of the anterior cervical spine using instrumentation in tages of the plate include prevention of graft extrusion and patients with fracture instability is well recognized [1,2]. decreased need for external postoperative immobilization The role of instrumentation for cervical degenerative disor- [7,8,14–19]. The disadvantages may include the cost, stress shielding of a bone graft, fixation failure from the bone with FDA device/drug status: not applicable. loosening of the screws and the risk of neurological or vas- Nothing of value received from a commercial entity related to this cular injury [12,15,17,18,20–24]. Currently, cervical plating research. systems incorporate a bone fixation screw secured by a * Corresponding author. Department of Orthopedic Surgery, Baylor College of Medicine, 6560 Fannin Suite 1900, Houston, TX 77030, USA. locking mechanism. The locking mechanism makes the Tel.: (713) 986-5740; fax: (713) 986-5741. screw-plate construct more rigid without using bicortical E-mail address: [email protected] (S.I. Esses) fixation [25,26] and prevents pushout of the bone screw. 1529-9430/03/$ – see front matter © 2003 Elsevier Science Inc. All rights reserved. PII: S1529-9430(02)00464-3

G.R. Fogel et al. / The Spine Journal 3 (2003) 118–124 119 This screw-plate construct thus acts mechanically as a single-piece implant. Without the need for posterior cortical penetration in the vertebral body, there is a decreased risk of screw misplacement and resultant spinal cord injury [27]. These plating systems are subject to failure. These are not common complications but may include screw pullout from the bone or from the plate and hardware breakage. There has been a proliferation of plating systems with variable locking mechanisms in recent years, all with their own purported advantages. No information exists in the spine surgery literature about the physical characteristics of these plating systems. The first purpose of this study was to consol- idate physical information for 10 commonly used plates. Sec- ondly, six of the systems were evaluated for screw pushout strength. The goal was to expand the spine surgeon’s understanding of the versatility of each plating system.

Materials and methods Test systems included cervical spine locking plate (CSLP) (Synthes, Paoli, PA); Orion, Premier, Zephr and Atlantis (Sofamor-Danek, Memphis, TN); Vuelock (EBI, Parsip- pany, NJ); PEAK and DOC (Depuy-AcroMed, Cleveland, Fig. 1. Potting technique used for each plate. PEAK plate demonstrated. Pit OH); Aline (Surgical Dynamics, Memphis, TN) and the beneath each of the upside down screws allows pushout of the individual screw. Blackstone (Blackstone Medical, Inc., Springfield, MA). These 10 plate systems are not inclusive of all plating systems available today. Additionally, only six of the systems were made available for the biomechanical pushout testing. Statistical analysis These systems were CSLP, Orion, Vuelock, PEAK, Aline A single-factor analysis of variance (ANOVA) was per- and the Blackstone. formed to compare the pushout strengths of the screw-plate Using a caliper accurate to 0.1 mm, repeated physical constructs for the different plate systems measurements were made for each plate variable. Charac- teristics studied for each system included the height or profile, the maximal width, the distance between the nearest Results screw hole to the end of the plate and the shortest normal- sized and rescue screw. Details were obtained to demon- Table 1 summarizes height, width, screw length and dis- strate the range of lengths for plates designed for one, two, tance from the end of the plate to the nearest screw hole. three and four levels of fusion, as well as the incremental Heights ranged from 1.8 to 3.2 mm, widths form 14.9 to 20.7 length increases for each of these plate designs. mm, and end of the plate to near screw measurement ranged Biomechanical testing was carried out with six different from 1.4 to 3.2 mm. Of note also, four systems had rescue plate styles. Each plating system tested was prepared for screws for which the minimum length was longer than the screw pushout testing by potting the plate in a polyvinyl minimum length standard screw. Table 2 summarizes the chloride cup with lead epoxy with a pit beneath the screw to locking mechanism and angulation position of the screw to allow the pushout of the single screw being tested (Fig. 1). the plate. Some are fixed angle and some are variable. The plate was stabilized with additional screws to the potted Table 3 details the variation in plate length. The Orion surface to prevent motion of the plate. Each plate was tested has a continuous slot and therefore has a non–level-specific with the locking mechanism in place at the manufacturers’ design for any number of fusion levels. A single screw may recommendations, including appropriate torque settings for be placed at any point between the two ends of the plate, as the locking mechanism screws. The fixation screw was the surgeon desires. The multilevel style of plate assumes tested at manufacturers’ optimal position to the plate, either that the surgeon is trying to place screws in separate verte- 90 degrees or slight angulation away from the center of the bral bodies after individual anterior discectomies. These de- plate. All constructs were maintained at body temperature signs are specific for a given number of fusion levels de- (37 degrees Centigrade) for the testing. Testing was per- pending on the number of levels of screw holes between the formed with a materials testing system (model 1321; Instron ends of the plate. The variability is in the overall length of Corporation, Canton, MA). The pushout strength value was the plate, the increment in plate length, and the number of the force at which the screw uncoupled from the plate. intermediate holes.

120 G.R. Fogel et al. / The Spine Journal 3 (2003) 118–124

Table 1 Physical measurements of cervical plates Height Distance edge of Name of plate (profile) Width plate to near screw Shortest normal Screw rescue Aline* 2.4 19.0 1.7 10 12 Blackstone† 2.6 17.8 1.4 12 12 Premier‡ 2.6 17.9 3.1 10 13 Orion‡ 2.6 17.9 3.2 10 11 Atlantis‡ 2.7 17.8 2.6 10 13 Zephyr‡ 1.8 14.9 1.7 13 13 PEAK§ 3.2 18.1 2.9 10 10 DOC§ 2.4 17.7 2.0 12 12 CSLP 2.0 20.7 2.2 12 12 CSLP SS 2.1 16.4 2.1 12 12 CSLP VA 2.6 18.1 2.3 12 12 Vuelock¶ 2.4 18.0 1.8 12 12 SS short stature; VA variable angle. *Surgical Dynamics, Memphis, TN. †Blackstone Medical, Inc., Springfield, MA. ‡Sofamor-Danek, Memphis, TN. §Depuy-AcroMed, Cleveland, OH. Synthes, Paoli, PA. ¶EBI, Parsippany, NJ.

Each of the plate systems reviewed had some differences ing ring; 3) blocking plate, set screw or flange and 4) a sep- in locking mechanism. However, the plates could be loosely arate threaded bushing within the plate. The expansion grouped into general categories by type of locking mecha- screw locking mechanism (Type 1) is characterized by a nism (Table 4). The types were 1) expansion screw; 2) lock- hollow head of the bone fixation screw head, with longitudi-

Table 2 Locking mechanisms and screw angulation to plate Name Locking mechanism Allowable angulation of screws to plate Aline* Inner expansion screw compresses head of Screw to plate variation screw to plate. Blackstone† Top-locking cover plate Variable screw alignment 0–30 degrees Premier‡ Fixed angle screw Orion‡ Locking screw covers head of plate screws. Fixed at one end but may angle 0–12 degrees away. A variable slot at other end, ideal 12 degrees up to 20 degrees away Atlantis‡ Locking screw covers head of plate screws. Variable angulation middle screws. Single screw placed through a slot that allows placement anywhere top to bottom of plate. End screws fixed at 15 degrees away from middle of plate Zephyr‡ Locking flange blocks screw from backing out. Fixed 12 degrees away, variable 2 degrees to center, 0–22 degrees away PEAK§ Plate bushing with separate threads locks screw Cephalic and caudal screws 0–16 degrees. Middle screws 0–7 degrees arc. Two screws to plate. at each end, but only one for middle segment DOC§ Up to 25 degrees angulation in a conical arc CSLP Inner expansion screw compresses head of Fixed at 0 degrees screw to plate. CSLP SS Inner expansion screw compresses head of Fixed 0 degrees caudal and 12 degrees cephalic screw to plate. CSLP VA Inner expansion screw compresses head of Fixed 0 degrees caudal and 6 degrees cephalic screw to plate. Vuelock¶ Preattached expansive ring design locks screw 20 degrees variable in conical arc below the ring. SS short stature; VA variable angle. *Surgical Dynamics, Memphis, TN. †Blackstone Medical, Inc., Springfield, MA. ‡Sofamor-Danek, Memphis, TN. §Depuy-AcroMed, Cleveland, OH. Synthes, Paoli, PA. ¶EBI, Parsippany, NJ.

G.R. Fogel et al. / The Spine Journal 3 (2003) 118–124 121

Table 3 Length of plates designed for one- to four-segment coverage One-segment Two-segment Three-segment Four-segment Name of plate Length IncrementLength Increment Length Increment Length Increment Aline* 22–28 2 28–48 4 48–108 6 Blackstone† 23–36 2 38–54 2 56–70 2 70–90 2.5 Premier‡ 23–25 2 25–90 2.5 90–110 5 Orion‡ 21.5–23 1.5 23–25 2 25–90 2.5 90–110 5 Atlantis‡ 14–25 2 35–47.5 2.5 50–72.5 2.5 62.5–90 2.5 25–32.5 2.5 90–110 5.0 Zephyr‡ 22.5–25 2 27.5–47.5 2.5 50–55 2.5 98–110 4 PEAK§ 24–30 2 39–63 3 57–90 3 59–87 4 30–36 3 90–94 4 74–109 5 DOC§ 16–28 2 30–42 3 44–64 4 68–92 4 CSLP 22–34 2 39–54 3 53–77 3 CSLP SS 20–34 2 34–54 3 47–77 3 CSLP VA 23–35 2 37–55 3 54–78 3 Vuelock¶ 12–24 2 26–42 2 44–66 2 60–92 4 SS short stature; VA variable angle. *Surgical Dynamics, Memphis, TN. †Blackstone Medical, Inc., Springfield, MA. ‡Sofamor-Danek, Memphis, TN. §Depuy-AcroMed, Cleveland, OH. Synthes, Paoli, PA. ¶EBI, Parsippany, NJ.

nal slots in the head to allow expansion of the head of the pushout of the locking ring from the plate, carrying the screw. There is an inner screw applied to expand the head of screw with it. The Orion failed by pushing out the central the fixation screw, which increases the rigidity of the cou- locking screw. The Blackstone failed by the screw pushing pling of the fixation screw to the plate. The CSLP was the the cover plate up and away from the main plate. The PEAK initial design of the expansion screw locking design, and the failed by pushing out the bushing from the plate carrying Aline plate has a modification on the expansion screw the screw with it. Comparison of pushout strengths showed mechanism with a more bulbar head and a longer inner a significant difference between all the plates, with p values screw. The inner screw can apply more torque slowly ranging from 1.5 109 to .007. Of the six systems, the ex- through its longer threaded length. The DOC has a similar pansion screw locking mechanism demonstrated the weak- expansion screw locking mechanism. The locking ring type est pushout values. of locking mechanism (Type 2) is characterized by a ring device pressed into the plate. The Vuelock has a snap-ring Discussion that the fixation screw passes through and is locked beneath. This ring locks the screw beneath the ring and the surface of As the usage of cervical plates has increased, so too has the plate but does allow a degree of conical screw position- the number of cervical plating systems. There is variability ing for optimal screw placement, without locking the screw in these plates. Technical problems arise during surgery, rigidly to the plate. The blocking type of mechanism (Type and certain features within the plating systems more easily 3) is characterized by a second screw, flange or blocking address these. For example, some patients have a smaller cover that is screwed down onto the heads of the fixation bone surface area of the anterior cervical spine and do not screws to block the fixation screw from pushing out. The accommodate wide plates and longer screws. Lordosis is blocking type includes the plates of Sofamor-Danek (Atlan- variable also. Therefore, optimal screw to plate angles vary tis, Orion, Prestige and the Zephyr) and the Blackstone as well. There is generally an optimal plate length, one that plate. Type 4, the PEAK, has a movable polyaxial bushing, clearly spans the fusion levels but does not abut or overhang which secures the screw to the plate by a separate threaded adjacent levels. Sometimes there is a wide range of accept- area mating the screw to the plate. able lengths, but other times, particularly in the presence of Pushout strengths were tested for six systems and ranged a partial corpectomy, there is a narrow range of optimal from 215 neutons (N) to 2662 N. Fig. 2 tabulates the push- lengths. Features of the optimal plate include not only the out strengths of the six plate systems with statistical com- incremental distance between holes, but also the distance parison (ANOVA). The CSLP and the Aline failed with a from the end of the plate to the near hole. Lastly, particu- pushout of the screw from the plate The Vuelock failed by a larly with three- or four-level fusions, it has been our expe-

122 G.R. Fogel et al. / The Spine Journal 3 (2003) 118–124

Fig. 2. Pushout strengths of six plate systems with statistical comparison (analysis of variance).

rience that there is a much greater variation in usual length Lowery and McDonough [28] found hardware failure in of plates. Sometimes, the ideal length plate does not corre- the cervical spine is more common the longer the patient is spond to the number of fusion levels for which it is in- observed after surgery. There was a 35% failure rate overall tended. In this situation the surgeon is unable to use most if but only 18% failure of constrained cervical plates [28]. not all of the intermediate fixation screws. Overlap in Most hardware failures are inconsequential when the patient lengths of plates designed for three- and four-level fusions is not symptomatic. Long-term observation is suggested, helps prevent this problem. and immediate removal of hardware is rarely necessary. Familiarity with plating system features should allow the They reported no patient had tracheal-esophageal erosion or surgeon to individualize and optimize plate selection. Ide- neurovascular compromise as a result of the instrumentation ally, plates should have narrow and wider choices, small in- failure [28,29]. crements in plate length, with short end of plate to near hole Although not common, cervical plate bone screws do fail distances, overlap in lengths from two- to three- and from with pushout from the plate. Geyer and Foy [30] described three- to four-level implants, low profiles and good visibil- erosion through the esophagus with subsequent oral extru- ity of the grafted levels. Screws should ideally come in vari- sion of a locked expansion type screw originating from a able lengths starting at 10 mm, have variable placement an- cervical plate. In this case initial failure was pullout of the gulation capability, have rescue screws of the same length screw from the bone, followed by later pushout failure of as the corresponding standard screw and be easily placed the screw from the plate [30]. The authors have had two ad- with a reliable locking mechanism. ditional cases that required revision because of symptomatic The contemporary cervical plating system is designed for screw pushout (Figs. 3 and 4). unicortical placement to prevent posterior bicortical pene- This pushout testing may be criticized because the forces tration of the cervical vertebra and injury to neurologic may be higher than are physiologically possible, and clini- structures. The locking mechanism has evolved for two cally, failure may not be purely a pushout mechanism. Each functions; one is to increase plate -screw rigidity while al- plate was tested to failure of the screw from the plate, and lowing unicortical fixation of the plate to the vertebra [25]. some plates required large pushout forces. In vivo, theoreti- The second is to prevent pushout failure of the screw from cally one might not require such large forces. Perhaps the the plate [25]. failure of a locking mechanism may occur more slowly, first

G.R. Fogel et al. / The Spine Journal 3 (2003) 118–124 123 plate has several unique physical characteristics, which permits the knowledgeable surgeon to choose the best plating system for each individual patient.

References

[1] Vaccaro AR, Balderston RA. Anterior plate instrumentation for disorders of the subaxial cervical spine. Clin Orthop 1997 Feb;(335):112–21. [2] Abraham DJ, Herkowitz HN. Indications and trends in use in cervical spinal fusions. Orthop Clin North Am 1998;29(4):731–44. [3] Connolly PJ, Esses SI, Kostuik JP. Anterior cervical fusion: outcome analysis of patients fused with and without anterior cervical plates. J Spinal Disord 1996;9(3):202–6. [4] Isomi T, Panjabi MM, Wang JL, et al. Stabilizing potential of anterior cervical plates in multilevel corpectomies. Spine 1999;24(21):2219–2223. [5] McLaughlin MR, Purighalla V, Pizzi FJ. Cost advantages of two- level anterior cervical fusion with rigid internal fixation for radicul- Fig. 3. Proximal screw failure from plate and locking mechanism. opathy and degenerative disease. Surg Neurol 1997:48(6):560–5. [6] Vaccaro AR, Falatyn SP, Scuderi GJ, et al. Early failure of long segment anterior cervical plate fixation. J Spinal Disord 1998;11(5):410–5. after failure of the fixation to the bone. The authors acknowl- [7] Wang JC, McDonough PW, Endow KK, et al. Increased fusion rates edge there are differences in failure from a single pushout with cervical plating for two-level anterior cervical discectomy and fusion. Spine 2000;25(1):41–5. force and failure from physiological cyclical stress, as might [8] Wang JC, McDonough PW, Kanim LE, et al. Increased fusion rates occur from nonunion or after failure of the screw from the with cervical plating for three-level anterior cervical discectomy and bone. This mode of testing of a locking mechanism is a new fusion. Spine 2001;26(6):643–7. method to demonstrate differences of characteristics of the [9] Zdeblick TA, Hughes SS, Riew KD, et al. Failed anterior cervical cervical plate locking mechanisms. It does not prove that one discectomy and arthrodesis. Analysis and treatment of thirty-five patients. J Bone Joint Surg Am, 1997;79(4):523–32. pushout strength is necessarily better, only that there are dif- [10] Coric D, Branch CL, Jenkins JD. Revision of anterior cervical ferences between the locking mechanisms. The clinical sig- pseudoarthrosis with anterior allograft fusion and plating. J Neuro- nificance of this difference requires further investigation. surg 1997;86(6):969–74. [11] Geisler FH, Caspar W, Pitzen T, et al. Reoperation in patients after anterior cervical plate stabilization in degenerative disease. Spine Conclusions 1998;23(8):911–20. [12] Zoega B, Karrholm J, Lind B. Plate fixation adds stability to two- All locking mechanisms significantly increased the push- level anterior fusion in the cervical spine: a randomized study using out strength of the tested screw-plate systems. The expan- radiostereometry. Eur Spine J 1998;7(4):302–7. sion screw had the lowest pushout values. These data dem- [13] Geer CP, Papadopoulos SM. The argument for single-level anterior onstrate a range of pushout strengths exists across the cervical discectomy and fusion with anterior plate fixation. Clin Neu- rosurg 1999;45:21–9. selections of available cervical plate systems today. Further [14] Zaveri GR, Ford M. Cervical spondylosis: the role of anterior instru- study is needed to understand the optimal or minimal push- mentation after decompression and fusion. J Spinal Disord 2001; out strength to avoid this mode of failure. In addition, each 14(1):10–6. [15] Grob D, Peyer JV, Dvorak J. The use of plate fixation in anterior surgery of the degenerative cervical spine: a comparative prospective clinical study. Eur Spine J 2001;10(5):408–13. [16] Bose B. Anterior cervical instrumentation enhances fusion rates in multilevel reconstruction in smokers. J Spinal Disord 2001;14(1):3–9. [17] van Limbeek J, Jacobs WC, Anderson PG, et al. A systematic literature review to identify the best method for a single level anterior cervical interbody fusion. Eur Spine J 2000;9(2):129–36. [18] Alvarez JA, Hardy RW. Anterior cervical discectomy for one- and two- level cervical disc disease: the controversy surrounding the question of whether to fuse, plate, or both. Crit Rev Neurosurg 1999;9(4):234–51. [19] Katsuura A, Hukuda S, Imanaka T, et al. Anterior cervical plate used in degenerative disease can maintain cervical lordosis. J Spinal Dis- ord 1996;9(6):470–6. [20] Sampath P, Bendebba M, Davis JD, et al. Outcome of patients treated for cervical myelopathy. A prospective, multicenter study with independent clinical review. Spine 2000;25(6):670–6. [21] Sampath P, Bendebba M, Davis JD, et al. Outcome in patients with cervical radiculopathy. Prospective, multicenter study with independent clinical review. Spine 1999;24(6):591–7. Fig. 4. Loosening of cover plate locking mechanism and bone screw back- [22] Branch CL. Anterior cervical fusion: the case for fusion without plating out of cervical plate. ing. Clin Neurosurg 1999;45:21–4.

124 G.R. Fogel et al. / The Spine Journal 3 (2003) 118–124

[23] Zoega B, Karrholm J, Lind B. One-level cervical spine fusion. A ran- [27] Ebraheim NA, Fow J, Xu R, et al. The vertebral body depths of the domized study, with or without plate fixation, using radiostereometry cervical spine and its relation to anterior plate-screw fixation. Spine in 27 patients. Acta Orthop Scand 1998;69(4):363–8. 1998;23(21):2299–302. [24] Huckell CB. Clinical outcomes after cervical spine fusion. Orthop [28] Lowery GL, McDonough RF. The significance of hardware failure in Clin North Am 1998;29(4):787–99. anterior cervical plate fixation. Patients with 2- to 7-year follow-up. [25] Spivak JM, Chen D, Kummer FJ. The effect of locking fixation Spine 1998;23(2):181–7. screws on the stability of anterior cervical plating. Spine 1999;24(4): [29] Lowery GL, Swank ML, McDonough RF. Surgical revision for failed 334–8. anterior cervical fusions. Articular pillar plating or anterior revision? [26] Pitzen T, Wilke HJ, Caspar W, et al. Evaluation of a new monocorti- Spine 1995;20(22):2436–41. cal screw for anterior cervical fusion and plating by a combined bio- [30] Geyer TE, Foy MA. Oral extrusion of a screw after anterior cervical mechanical and clinical study. Eur Spine J 1999;8(5):382–7. spine plating. Spine 2001;26(16):1814–6.

Two tive of Scotland, was a physician in the Royal Navy. Various reports of the use of citrus to treat or prevent Hundred scurvy among sailors preceded Lind’s work, but his organized trial and thorough report convinced the Ad- Fifty miralty and led to measures that eradicated scurvy from Years Ago the Navy. in Spine . . .

The prototype of the controlled thera- References peutic trial was the experimental work done on scurvy [1] Lind J. A treatise of the scurvy. Edinburgh: Sands, Murry and by James Lind, which he reported in 1753. Lind, a na- Cochran,1753. Reprinted with notes, Edinburgh, 1953. SPINE Volume 28, Number 5, pp 470–473 ©2003, Lippincott Williams & Wilkins, Inc. Physical Characteristics of Polyaxial-Headed Pedicle Screws and Biomechanical Comparison of Load With Their Failure

Guy R. Fogel, MD, Charles A. Reitman, MD, Weiqiang Liu, PhD, and Stephen I. Esses, MD

Clinically, the senior author has observed that the incidence Study Design. Pedicle screw strength or load to failure of broken pedicle screws has diminished over the past sev- was biomechanically evaluated, and the geometric char- eral years coincidentally with the usage of polyaxial screws. acteristics of pedicle screw instrumentation systems were A hypothesis was developed to explain the decreased compared. Objectives. To compare the features of pedicle screw breakage of pedicle screws. It may be that there is a subtle systems, and to demonstrate the failure point of the loosening or failure of the polyaxial head that removes polyaxial pedicle screw head. some of the stress from the pedicle screw. By decreasing the Summary of Background Data. Many pedicle screw stiffness in the coupling of the polyaxial head to the pedicle instrumentation systems are currently available to the screw, the bending stresses on the pedicle screw would be spine surgeon. Each system has its unique characteristics. It is important for the surgeon to understand the lessened. A MEDLINE search showed no biomechanical differences in these pedicle screw systems. Pedicle screw evidence in the English spine literature regarding the testing load to failure has not been subjected to a comparison of polyaxial screws. In addition, polyaxial-headed pedicle study. screw load-to-failure testing has not been subjected to a Methods. The physical characteristics of each pedicle comparison study. screw instrumentation system were determined. Features of rods, instruments, and pedicle screws were cataloged. Methods Biomechanical testing of the pedicle screw construct was performed to determine the site and force of the load to Nine pedicle screw systems were evaluated: the Silhouette (Sul- failure. Nine pedicle screw systems were evaluated. Test- zer Spine-Tech, Minneapolis, MN), Blackstone (Blackstone ing was performed with a pneumatic testing system un- Medical, Springfield, MA), Click-X (Synthes, Paoli, PA), Xia der load control. Three polyaxial screws were used for (Stryker-Howmedica, Warsaw, IN), M8 (Sofamor-Danek, each test at a load rate of 100 N/second. The load failure Memphis, TN), Miami-MOSS, Monarch, and Magnum value was the force at which the pedicle screw or polyaxial head–screw interface initially deflected. (Depuy-AcroMed, Cleveland, OH), and the SD-90 (Surgical Results. Biomechanical testing demonstrated in all in- Dynamics, Memphis, TN). Not all of the commercially avail- stances that the polyaxial head coupling to the screw was able pedicle screws were available for testing. The screw the first failure point. Although there have been subtle lengths were standardized at 45 mm, and the diameters varied design differences in the instruments over time, the fea- from 6 to 7.5 mm. tures of the pedicle screw instrument sets have become Each tested screw was mounted perpendicularly on the ap- remarkably similar. propriate rod provided by the vendor at the manufacturer’s Conclusions. Biomechanical pedicle screw load-to-fail- recommended torque settings. The distal half of the screw body ure data demonstrated that the polyaxial head coupling to was potted in the shape of a ceramic cylinder to enhance con- the screw is the first to fail and may be a protective tact with the MTS machine. Testing was performed with a feature of the pedicle screw, preventing pedicle screw breakage. Knowing the physical characteristics of the materials testing system (Model 1321; Instron, Canton, MA) available pedicle screw instrumentation systems may al- (Figure 1). The MTS force was applied at a point 30 mm from low the choice of pedicle screw best suited for a given the rod perpendicular to the long axis of the screw. The MTS clinical situation. [Key words: biomechanics, equipment compressed the screw at a load rate of 100 N/second until design, pedicle screws, spinal fusion/instrumentation, failure occurred. The load failure value was the force at which surgery] Spine 2003;28:470–473 the pedicle screw initially deflected or uncoupled from the polyaxial head. Polyaxial heads have made the pedicle screw more versatile, Results particularly improving ease of connecting rod application. The geometry of the coupling between the screw and From the *Department of Orthopedic Surgery, Baylor College of Med- head shows a conforming hemispherical interface that icine, Houston, Texas. allows for polyaxial motion of the head on the screw. Acknowledgment date: June 5, 2002. First revision date: July 26, 2002. Acceptance date: July 30, 2002. The fixation of the polyaxial head to the rod is with an Device status/drug statement: The device(s)/drug(s) is/are FDA ap- internal screw, external nut, or both, pushing the rod proved or approved by a corresponding national agency for this into the slot of the head. The Silhouette has an external indication. Conflict of interest: No funds were received in support of this work. No nut securing the rod into the head of the screw, whereas benefits in any form have been or will be received from a commercial the Miami MOSS and the Magnum have an inner screw party related directly or indirectly to the subject of this manuscript. head and an external nut both securing the rod into the Address reprint requests to Stephen I. Esses, MD, Brodsky Chair of Spinal Surgery, Baylor College of Medicine, 6560 Fannin St., Suite screw head. The other five systems have an internal screw 1900, Houston Texas 77030. E-mail: [email protected]. device securing the rod to the pedicle screw head. To gain

470 Polyaxial-Headed Pedicle Screws • Fogel et al 471

Figure 1. Setup of the Instron MTS. The pedicle screw is en- cased in a ceramic cylinder, and the rod construct is held in a large chuck in a vise. The screw then is compressed at a load rate of 100 N/second until failure occurs.

final fixation, all the screws and nuts require torque to themselves. The crosslink systems have been updated, specified levels except the Surgical Dynamics SD-90. The and there currently is a wide selection of screw diame- SD-90 has a helical wedge that requires only a 90° turn of ters, lengths, and sizes of cross-links. Two of the systems the inner screw to gain the final fixation. The Monarch do offer more than one rod diameter. also has a dovetail top with a center screw that locks the The results from the load-to-failure testing of the head to the rod. polyaxial pedicle screws demonstrates that the weakest The Silhouette had the lowest mean failure load of point of the construct is the head-to-screw coupling (Ta- 213 N, whereas the Magnum was the highest at 486 N. ble 1 and Figure 2). This failure of the polyaxial head The full results from load-to-failure testing of the pedicle may be a protective factor for the pedicle screw shaft, screws are shown in Table 1. Figure 2 shows that the preventing early breakage. statistical differences ranged from a P of 0.13 to a P of The polyaxial head has three tasks: 1) to secure the 0.0009. The statistical power was limited in some cases rod to the head, 2) to prevent the head from deforming in by the small sample size. diameter, and 3) to secure the polyaxial head to the pedi- Discussion cle screw. The outside nut, pin-nut, helical or dovetail wedges, and thicker walled polyaxial head are designed Pedicle screw systems have undergone continual modifi- to prevent the head from deforming. The inner screw cations over the past several years. As recently as 1 to 2 locks the rod in the head and the head to the screw. The years ago, there were substantial differences in the design helical wedge actually can do all three tasks with one features of these sets. However, the systems have evolved locking device. All pedicle screw heads have some to match the strengths of each other such that there cur- method to stabilize the diameter of the head and a rently are very few differences between instrument sets method to hold the rod in the head. The single outside (Table 2). There have been design changes in the screws nut–locking mechanism was statistically weaker than any other design. Biomechanical tests of pedicle screw constructs have Table 1. Load to Failure of Polyaxial Head of demonstrated the fundamental importance of the bone Pedicle Screw implant interface, bone density, and screw pullout 4,5,7,9,11,12,15,16 Screw Screw strength. The essential need for fit and fill System Mean (N) SD Diameter (mm) of the screw in the isthmus of the pedicle has been proved.7,15 The direct relation between pullout strength Silhouette 213.21 25.71 6.5 5,6 BMI 268.88 38.06 6.5 and insertional torque has been well demonstrated, Moss Miami 280.02 13.31 6 and the fundamental improvement in pullout strength M8 340.57 31.18 6.5 14 Click’X 349.54 42.53 7 obtained by cross-linking has been documented. The SD90 357.99 25.44 6.75 stabilizing influence of using converging screws has been Xia 397.72 25.5 7.5 shown.1 The major diameter of a pedicle screw has been Monarch 447.67 28.38 7 2,3 Magnum 486.24 24.05 7 shown to control pullout strength. Bicortical purchase increased pullout strength fundamentally both in indi- ϭ SD standard deviation. vidual vertebrae and in the sacrum. However, bicortical 472 Spine • Volume 28 • Number 5 • 2003

Figure 2. Comparison of load-to- failure data for nine pedicle screw systems.

Table 2. Pedicle Screw Sets Characteristics

Screw Diameter (mm) Rod Diameter (mm) Screw-rod Locking Mechanism

Danek CDH 4.5, 5.0, 5.5, 6.5, 7.5, 8.5 4.5, 5.5, 6.35 Inner set screw, Buttress thread Surgical Dynamics SD-90 4.75, 5.75, 6.75, 7.75, 8.75 5.5 Twisting saddle nut Depuy Miami-MOSS 4.35, 5.0, 6.0, 7.0, 8.0 5.0, 5.5, 6.35 Inner pin nut in 5.5 mm diameter rod.* Inner and outer nut for 5.0 and 6.35 mm rod Monarch 4.75, 5.5, 6.25, 7, 7.75, 8.5 5.5 Dove tail cap with prethreaded set screw Magnum 6-7-8 6.35 Inner set screw and Outer nut Blackstone 4.5, 5.5, 6.5, 7.5, 8.5 5.5 Inner set screw Spine-Tech Silhouette 4.5, 5.5, 6.5, 7.5, 8.5 5.5 Outer nut Synthes Click-X 5.2, 6.2, 7.0, 8.0, 9.0 6.0 Inner set screw Stryker XIA 4.5, 5.5, 6.5, 7.5, 8.5 6.0 Inner set screw, Buttress thread

* Currently the Miami Moss 5.5 Titanium system uses the pin nut, while 5.0 and 6.35 mm rods in both steel and Titanium use the inner screw and outer nut locking mechanism.

purchase, except at the sacrum, has not been widely Conclusions adopted by surgeons because of the risk for vascular Biomechanical pedicle screw load-to-failure data demon- 8,10,13 injury. Nevertheless, bicortical sacral purchase has strated that the polyaxial head coupling to the screw was the been proved extremely safe and has gained widespread first failure point and may be a protective feature of the pedicle 8,10,13 acceptance. screw and rod, preventing pedicle screw or rod breakage. Biomechanical testing of the load characteristics of Knowing the physical characteristics of the available pedicle the polyaxial pedicle screw was not reported in the liter- screw instrumentation systems may allow the choice of pedicle ature reviewed for this study. This testing has further screw best suited for a given clinical situation. detailed the biomechanical characteristics of the polyaxial pedicle screw. The results of this testing demonstrate a range of load tolerances for the various sys- Key Points tems. Although the testing does show that some ● The polyaxial head coupling of the pedicle screw polyaxial pedicle screws are stronger, it is important to is the first feature to fail. note that it is not clear whether a stronger coupling of ● This may protect the pedicle screw from pedicle screw to head is better. There may be instances in breaking. which more or less rigid fixation is preferable. Further ● There is a wide range of pedicle screw construct study is needed to investigate the physical properties and failure loads. their clinical application. Polyaxial-Headed Pedicle Screws • Fogel et al 473

References 8. Lu WW, Zhu Q, Holmes AD, et al. Loosening of sacral screw fixation under in vitro fatigue loading. J Orthop Res 2000;18:808–14. 1. Barber JW, Boden SD, Ganey T, et al. Biomechanical study of lumbar pedicle 9. McCormack T, Karaikovic E, Gaines RW. The load sharing classification of screws: Does convergence affect axial pullout strength? J Spinal Disord 1998; spine fractures. Spine 1994;19:1741–4. 11:215–20. 10. Mirkovic S, Abitbol JJ, Steinman J, et al. Anatomic consideration for sacral 2. Daftari TK, Horton WC, Hutton WC. Correlations between screw hole screw placement. Spine 1991;16:S289–94. preparation, torque of insertion, and pullout strength for spinal screws. J Spi- 11. Oktenoglu BT, Ferrara LA, Andalkar N, et al. Effects of hole preparation on nal Disord 1994;7:139–45. screw pullout resistance and insertional torque: A biomechanical study. 3. Hirano T, Hasegawa K, Takahashi HE, et al. Structural characteristics of the J Neurosurg 2001;94:91–6. pedicle and its role in screw stability. Spine 1997;22:2504–9, discussion 10. 12. Ono A, Brown, MD, Latta LL, et al. Triangulated pedicle screw construct 4. Kostuik JP, Munting E, Valdevit A. Biomechanical analysis of screw load technique and pullout strength of conical and cylindrical screws. J Spinal sharing in pedicle fixation of the lumbar spine. J Spinal Disord 1994;7:394– Disord 2001;14:323–9. 401. 5. Liljenqvist U, Hackenberg L, Link T, et al. Pullout strength of pedicle screws 13. Robertson PA, Plank LD. Pedicle screw placement at the sacrum: Anatomical versus pedicle and laminar hooks in the thoracic spine. Acta Orthop Belg characterization and limitations at S1. J Spinal Disord 1999;12:227–33. 2001;67:157–63. 14. Suzuki T, Abe E, Okuyama K, et al. Improving the pullout strength of pedicle 6. Lill CA, Schlegel U, Wahl D, et al. Comparison of the in vitro holding screws by screw coupling. J Spinal Disord 2001;14:399–403. strengths of conical and cylindrical pedicle screws in a fully inserted setting 15. Tencer AF, Hampton D, Eddy S. Biomechanical properties of threaded in- and backed out 180 degrees. J Spinal Disord 2000;13:259–66. serts for lumbar interbody spinal fusion. Spine 1995;20:2408–14. 7. Lim TH, Eck JC, An HS, et al. Biomechanics of transfixation in pedicle screw 16. Vaccaro AR, Garfin SR. Pedicle screw fixation in the lumbar spine. J Am instrumentation. Spine 1996;21:2224–9. Acad Orthop Surg 1995;3:263–74. PPR3021 4/3/2003 10:51 AM Page 1

CASE CONFERENCE

Management of Chronic Limb Pain with Spinal Cord Stimulation

Guy R. Fogel MD*; Stephen I. Esses, MD‡; Octavio Calvillo, MD, Phd†,‡ *Spine Fellow, †Department of Anesthesiology, and ‡Professor of Orthopedics, Brodsky Chair of Spinal Surgery, Baylor College of Medicine, Houston, Texas

Abstract: syndromes. The treatment modalities of the pain man- Background: Spinal Cord Stimulation (SCS) is a treatment agement practitioner include all of the modalities and option for chronic pain patients. The most common indica- therapies, conservative, pharmacologic, and invasive, tion for SCS is the failed back syndrome with leg pain. In the used to treat chronic pain syndromes. The purpose of last decade, advances in our understanding of appropriate this paper is to describe the basis for the rational use of stimulation programming, lead placement and the physiol- 1 implantable modality, Spinal Cord Stimulation (SCS). ogy of SCS, have led to changes in multi-site stimulation, and Electrical stimulation was first considered for treat- stimulation with differing programs. In the past, low back, ment of intractable pain based on the publication of the axial neuropathic type pain was not responsive to SCS. With dual electrode arrays, and dual stimulation with alternating gate control theory of pain by Melzack and Wall in 1 programs of stimulation, steering of stimulation paresthesia, 1965. Shealy et al, in 1967, first introduced electrical and versatile programmable stimulation parameters, SCS has stimulation of the spinal cord and peripheral nerves for become a more versatile form of analgesia. chronic limb pain.2 In the decade following the studies, Purpose: To describe the current treatment rational for SCS several thousand stimulators were implanted. The initial and the results of that treatment. enthusiasm was dampened by reports of high compli- Results: The SCS is most efficient in patients with neuro- cation and failure rates. At issue, were increased break- pathic pain of the extremities and less efficacious in patients with axial pain. age and displacement of electrical leads, and failure of Conclusion: SCS is the most effective treatment for limb the implanted receiver, the high cost of the implant, and pain not amenable to surgical decompression. The success of difficulty defining the patient population, which could SCS in this chronic pain group is 80% successful in treatment respond with a reasonable pain relief percentage.3–5 of leg pain, and much less effective in treatment of axial SCS has been applied to a variety of diagnoses, pain. including tumors, brachial plexus injuries, spinal cord injury, multiple sclerosis, peripheral vascular disease INTRODUCTION and ischemic limb pain, ischemic cardiac angina, arach- The rational use of implantable technologies for pain noiditis, and pain after failed back surgery.3 However, control should be founded on the knowledge of the neu- even among those with intractable lower extremity pain, robiology of pain, and the clinical presentations of pain the outcome results have shown wide variability.6–10 In the last decade, advances in our understanding of Address correspondence and reprint requests to: Stephen I. Esses, MD, appropriate stimulation programming, lead placement Professor of Orthopedics, Brodsky Chair of Spinal Surgery, Baylor College and the physiology of SCS, have led to changes of mul- of Medicine, 6560 Fannin Suite 1900, Houston Texas 77030. Tel: (713) 986- 5740; Fax: (713) 986-5741; E-mail: [email protected]. tisite stimulation, and stimulation with differing programs. In the past, low back, axial neuropathic type

© 2003 World Institute of Pain, 1530-7085/02/$15.00 pain was not treatable, but with dual electrode arrays, Pain Practice, Volume 3, Issue 2, 2003 ••–•• and dual stimulation with alternating programs of stim- PPR3021 4/3/2003 10:51 AM Page 2

2•fogel et al.

ulation, steering of stimulation paresthesia, and versa- limbs. In terms of pain type, SCS is most effective for tile programmable stimulation parameters, SCS has treating peripheral neuropathic pain which results from become a more versatile form of analgesia. However, actual damage to the peripheral nerves. Causalgia is an based on results, the SCS is most efficient in patients example of peripheral neuropathic pain. SCS is gener- with neuropathic pain of the extremities and less effica- ally not effective for treating 2 other types of pain: (1) cious in patients with axial pain.11,12 nociceptive pain, which results from nerve irritation (not damage) caused by noxious stimuli such as heat, CHARACTERISTICS OF CHRONIC PAIN pressure, or chemicals (burn pain, muscle injury pain, Chronic physical pain may be divided into 2 types: and cancer pain are examples of nociceptive pain and nociceptive and neuropathic pain. The differences are (2) central pain which is caused by Central Nervous important in predicting the efficacy of SCS to treat the System (CNS) damage from a stroke or spinal cord chronic pain. Nociceptors are the nerves which sense injury. and respond to injury of the body. They signal tissue irritation, impending injury, or actual injury. When HISTORICAL REVIEW activated, they transmit pain signals (via the peripheral Application of electrical current through the skin began nerves as well as the spinal cord) to the brain. The pain in the mid 1790 and by the early 20th century, different is typically well localized, constant, and often with an types of electrical devices were available. However, with aching or throbbing quality. This is normal pain in the increased availability of different types of analgesic response to injury of the body. drugs and ablative pain relieving procedures such as Neuropathic pain is the result of an injury or rhizotomy, cordotomy, and thalamotomy, the electrical malfunction in the peripheral or central nervous system. devices had little use and were abandoned. Modern pain Neuropathic pain is not caused by nociceptors; however, control with electricity started with the report of the pain is often triggered by an injury. The pain Melsack and Wall who proposed the gate theory of frequently has burning, lancinating, or electric shock pain.1 Within the year, Wall and Sweat demonstrated qualities. Hyperpathia and allodynia are symptoms of peripheral nerve stimulation could bring pain relief.13 neuropathic pain. Hyperpathia is an increased pain Shealy devised an implantable type of stimulator for the from a stimulus which would be painful normally. Allo- spinal cord known as a dorsal column stimulator.2 In dynia is pain from stimuli, which are not normally the seventies, Long and Sweat independently reported painful, or pain that occurs other than in the area stim- implantable peripheral nerve stimulators.3,4 Soon ulated. Persistent allodynia is also a common charac- afterward, an implantable deep brain stimulator was teristic of neuropathic pain. The pain may persist for reported by Hosobuchi and Adams.14 The initial theory months or years beyond the apparent healing of any of the function of the stimulator was to inhibit the damaged tissues. Examples include post herpetic neu- C-fibers by stimulating the larger myelinated fibers.2 ralgia, reflex sympathetic dystrophy, causalgia, compo- Even today, the mechanism of pain relief is poorly nents of cancer pain, phantom limb pain, entrapment understood. neuropathy, and peripheral neuropathy. Neuropathic Initially, stimulation was difficult to promote widely pain is frequently chronic, and tends to have a less for several reasons. There was the general lack of expe- robust response to treatment with opioids. rience with treatment of this difficult group of pain In some conditions the pain appears to be caused by patients. Also a lack of understanding of the co- a complex mixture of nociceptive and neuropathic morbidities of the patient’s chronic pain syndrome, and factors. An initial nervous system dysfunction or injury inappropriate patient selection, all lead to poor wide- may trigger the neural release of inflammatory media- spread acceptance. With the advent of electrodes tors and subsequent neurogenic inflammation. For implanted through a needle under local anesthesia and example, Failed back syndrome probably represent a a decrease in complications such as breakage and migra- mixture of neuropathic and nociceptive pain. A second tion of leads, and the infection rate, the usage of the example, myofascial pain is probably secondary to noci- spinal cord stimulators has risen again. ceptive input from the muscles, but the abnormal muscle activity may be the result of neuropathic factors. INDICATIONS Spinal Cord Stimulation is indicated for patients The patient with intractable pain who is most likely to suffering from chronic intractable pain of the trunk or be helped by a spinal cord stimulator would have leg PPR3021 4/3/2003 10:51 AM Page 3

Chronic Limb Pain and Spinal Cord Stimulation •3

pain of greater intensity than back pain, chronic pain of THE STIMULATOR more than 6 months duration that failed to respond to The stimulator is commercially available. The basic conservative measures, or was not a candidate for con- principles have remained the same, since the early ventional surgical treatment. Many have had multiple pioneering work of Shealy. The electrode is designed to back surgeries. Some had reflex sympathetic dystrophy. fit the field of stimulation, spinal cord, peripheral nerve, In addition to the long history of back and leg com- or deep brain. The electrode may be monopolar or plaints, many patients have additional health problems, quadripolar. A bipolar lead was evaluated and dis- including diabetes, hypertension, arthritis, renal and carded. There is no particular difference in the electrode cardiac disease, multiple nonspinal operations, history performance. However, the need for subsequent revision of Hodgkin’s lymphoma, other cancers, gastrointestinal of the electrode is 24.5% in the quadripolar versus maladies and Paget’s disease. These patients are of any 68.4% in the monopolar. There is less “lead migration,” age 20–85 years. They have all been treated with nar- breakage, and a lower infection rate. One factor in the cotic medications for pain relief before the spinal cord difference of revision rates between the 2 electrode types stimulator. The patient is found to be refractory to con- may be that the monopolar is always used in trial stim- servative modalities of pain relief. All patients should ulation and may be incorporated in the permanent have a benign form of intractable pain with an organic placement of an implant. The quadripolar is only used basis. The pain may be radicular or nonspecific but in permanent implant group. should be worse in the extremities, usually legs. Some The electrode is connected to a passive receiver, or a axial pain is acceptable. There should not be a major battery powered stimulator. The battery powered units psychiatric condition. Testing may include visual analog are self powered and activated by external control. The scales to assess leg and back pain intensity. Pain Draw- passive receivers are activated by a radio signal trans- ings identify the extent of the painful areas of the back mitted by antennae. The patient has complete control and extremities. of the duration and strength of stimulation. Incorpora- SCS has been most effective in treating neurogenic tion of the trial stimulation period had improved the pain of peripheral origin. The SCS will relieve pain only selection of long term implant patients.7,8,9,10 if paresthesias are induced in the area of the patient’s pain, and some say, if the pain returns after the stimulation is stopped, although there is no scientific THE TRIAL PROCEDURE substantiation. There should be a separate trial stimulation procedure The most common diagnosis treated by SCS, is failed and a permanent implantation in those patients who back syndrome. Failed back syndrome (FBS) may be have a favorable pain relief response with paresthesias defined as chronic pain associated with degenerative produced in the area of the patient’s pain. The implan- spondylosis usually lumbar, associated with a history of tation of trial electrodes may be done with a local anes- previous surgery and absence of current surgical indi- thetic in a day surgery, or as an over-night hospital stay. cations. The FBS patients have lower extremity pain and Most trial and permanent electrodes are inserted dysthesias associated with axial back pain. The limb through an epidural Tuohy needle. Sometimes a pain may be in a particular dermatome, however many laminectomy is required in the lower thoracic spine to FBS will not have an anatomic radiculopathy. Most place the electrode in the epidural space under direct series do not differentiate in the types of diagnosis or vision. The laminotomy is usually required in the pres- leg pain.10,11,15 Many authors report less satisfactory ence of previous surgery, technical difficulty, or epidural outcome in treatment of axial pain. This may be related scarring. After implantation, the electrode is connected to the presence of less receptive nociceptive pain or to a handheld programmer that allows various levels of during the trial setting, that paresthesias may not be stimulation to be tested during the trial period. Patients reproduced in the axial locations of the pain.11,12,16–20 are encouraged to increase their activities to near Relief of cancer pain is also reported as poor.21,22 normal during the trial. This should give some indica- Usually SCS is not effective in central pain such as stroke tion of the extent to which the permanent stimulator or spinal cord injury. However, SCS may be effective, if could be expected to control pain levels. After demon- paresthias are reproduced in the patient’s painful area. strating greater than 50% reduction of pain, the patient SCS has been effective in angina and claudication of will undergo the permanent implantation of the stimu- peripheral vascular disease.23–25 lating system. The pulse generator is implanted in a sur- PPR3021 4/3/2003 10:51 AM Page 4

4•fogel et al.

gically prepared pocket on the abdomen. An extension SCS increases the concentration of glycine in spinal wire is tunneled to the lead for the SCS system. Further cord tissue, and proposes glycine as a reasonable can- parameter adjustment may be undertaken as an didate to explain SCS-induced analgesia.29,30 Another outpatient. The versatility of the current programming neuro-transmitter, gama-amino butyric acid(GABA), is means there are literally thousands of combinations of increased with SCS.31,32 Linderoth demonstrated SCS is stimulation parameters to test. Computerized testing capable of inducing significant increases of GABA in the systems have been developed to maximize the patient’s dorsal horn of the rat.33 Further investigations by response. Linderoth demonstrated increases of serotonin with SCS Some of the patients fail the trial. The failure of the in decerebrate cats. Serotonin is an important mediator trial in cord central pain situations may be from several of analgesia at the spinal level;34 therefore, it is possi- sources. One is difficulty accessing the epidural space ble that serotonin could be a mediator in SCS-induced secondary to trauma or surgery. Second is difficulty pro- analgesia. Catecholamines in SCS induced analgesia ducing paresthesias over a large enough area of the may be increased. Noradrenergic mechanisms may patient’s pain. Third, is the die-back of dorsal column modulate the nociception at the spinal level.26 above a severe cord lesion seen in spinal cord injury SCS is applied with low intensity and generally in the patient. Risks are low and neurologic injury is virtually 50–70Hz and pulse width of 0.2–0.5msec. Stimulation zero. However, a canal stenosis at the proposed implan- paresthesiae must cover the entire painful area. SCS tation site may cause increased risk to the underlying must be applied continuously for periods of at least spinal cord. Pre-operative evaluation of the canal space 20–30 minutes. Effective pain relief with SCS is most at the proposed implantation site is important. Infec- likely to occur in cases with neuropathic pain. There is tions may occur in 5%. Also mechanical and electrical little evidence that purely nociceptive forms of pain, failures become more common with the passage of time. whether acute or chronic, will respond.28,33,35 SCS is There remains a fairly constant revision rate for non- most effective for on-going spontaneous pain and less functioning components. for pain evoked by load and posture.

CHARACTERISTICS OF RESULTS THERAPEUTIC STIMULATION Some patients (20% to 40%) will fail the trial stimula- Recently, some insight has been gained in the physio- tion. There are several reasons reported. There may be logic mechanisms underlying the pain relief of SCS. Ini- a failure to get paresthesias in the area of pain. There tially, SCS evolved as a direct clinical application of the may be no response to less than 50% relief of pain to well known gate-control theory of Melzack and Wall. stimulation. Patients who receive 50% relief are usually However their gate theory would imply all acute and implanted with permanent stimulator. Kim et al noted chronic nociceptive pain could be suppressed. This has 20% immediate failures after permanent implantation.8 not turned out to be accurate. The SCS is almost exclu- Kim et al noted the electrode was the same for the trial sively beneficial for neuropathic pain and not helpful for and permanent, allowing some placebo effect. An addi- nociceptive pain. In the patients with ischemia of angina tional 25% to 40% may work well long term and then and peripheral vascular disease, which would be lose pain control, usually for technical reasons. Most of described as nociceptive pain, the SCS may actually have the technical late failures will improve with revision of its beneficial effect because of improved ischemia rather the stimulator or electrodes. However, there is a small than any nociceptive pain relief. In addition to the group that may lose efficiency and not regain it with electrophysiologic mechanism of Melzack and Wall, revision. These are termed “late failures” and current neurochemical mediators may mediate the SCS induced thought is that this may be a manifestation of “toler- analgesia. There is no evidence that endogenous opiod ance” to the stimulator.36 Tolerance is described as the release is caused by SCS. No increase in opiod peptides gradual loss of pain relief while stimulated, without have been demonstrated.26 The effects of SCS are mechanical problems. This tolerance has been reported unchanged by Naloxone, an opiod antagonist.27,28 with deep brain stimulation also.9 Kumar notes Various aminoacid neurotransmitters have been shown Amytriptylline and L-Tryptophan have no benefit in to participate in nociception. Inhibitory transmitters like tolerance cases. taurine and glycine have been studied in relation to SCS- See Table 1 for a summary of results from the liter- induced analgesia, using microdyalisis has demonstrated ature of the last 10 years. PPR3021 4/3/2003 10:51 AM Page 5

Chronic Limb Pain and Spinal Cord Stimulation •5

Table 1. A Decade of SCS Clinical Series

Reference Study Design Size Type Pain Follow-up Results

Devulder J et al 199047 Retrospective 45 Mixed: failed back 5 yr 77% Very good pain relief Kumar K 199115 Retrospective 121 Mixed: most failed back 6 mo to 10 yr ave 40 mo 40% Pain control by SCS

North RB 199112 Retrospective 50 Failed back 2.2 yr (5 yr max) 53% > 50% pain relief; 5.0 yr: 47% > 50% pain relief Simpson BA 199148 Retrospective 60 Mixed trauma and 2–9 yr (ave 20 mo) Modest 23.3%; signiﬁcant 46.7% failed back Spiegelmann R 199149 Retrospective 43 Mixed, RSD, Failed surgery 3–33 mo (ave 13 mo) 63% Pain relief

Tasker RR 199250 Retrospective 35 Mixed, Iatrogenic, 30 yr 50% success. 25 pts had >than inﬂammation, vascular experience 50% pain relief at 1 yr.

De La Porte C 199351 Retrospective 64 Failed back surgery 1–7 yr (ave 4 yr) 55% had >50% pain relief North RB 199352 Retrospective 205 Mixed 2–20 yr (ave 7.1 yr) 52% HAD >50% relief pain. LeDoux 199311 Retrospective 32 Failed back syndrome 2 yr 76% Good at 1 yr, 37% 5 yr. Ohnmeiss 199610 Prospective 40 Failed back syndrome 1–2 yr 53% at 6 wks, 26% at 1–2 yr Burchiel 19967 Prospective 219 Failed back and leg pain 1 yr 55% success in 70 pts at 1 yr. Kumar et al 988 Retrospective 66 mo ave Kim 20018 Retrospective 122 Mixed Non-speciﬁc limb 3.9 yr (0.3–9 yr) 60.7% received pain evoked pain implant 79% success.

Pain, narcotic usage, return to work, and activities of only pain relief available. In a systematic literature daily living are deemed measures of success at last review of patients treated for FBS with SCS, Turner and follow-up. Pain criteria for success of the SCS have been associates found an average of 50% to 60% of patients difficult to decide. Most authors consider SCS success- with FBS report >50% pain relief with the use of the ful if there continues to be greater than 50% relief of SCS at follow-up.20 the index pain level at 1 year after implantation. Nar- SCS has been very effective in treating Reflex Sym- cotic usage is a highly variable follow-up parameter. pathetic Dystrophy (RSD), in well-defined indications Some series report reduction in drug usage, and changes including a history of limb trauma; pain of more than to less potent analgesics. Other studies find no reduc- 1 peripheral nerve distribution; physical findings of tion. Many authors avoid treating patients who use dystrophy; and a positive response to a sympathetic narcotics excessively. Improved return to work data blockade.40–44 Kemler and associates conducted a ran- again is equivocal. Some authors report a higher domized, controlled trial of SCS for reflex sympathetic incidence of return to work. However, a high percent- dystrophy.40 The results show that SCS reduces the age of the patients are disabled, or retired, skewing the intensity of pain caused by this disorder in patients in data.7,8,10,37 whom all conventional treatments have failed. The success rate was 56% at 6 months in permanently RESULTS IN DIFFERENT DIAGNOSES implanted SCS. Kumar etal, reported 12/12 patients Non-specific limb pain may be treated as effectively as with RSD experienced pain relief with trial stimulation neuropathic pain if a peripheral nerve is the root cause and had permanent stimulators implanted.9 At follow- of the pain. Central pain is amenable if paresthesias are up, 8 patients reported excellent and 4 good results. obtained in area of pain.8 SCS is the most important Kumar concluded SCS is an effective treatment for the pain technique available for treatment of intractable pain of RSD and that SCS is superior to ablative sym- pain of benign origin. SCS may be most valuable in pathectomy in the management of RSD. failed back syndrome (FBS).6,11,18–20,37,38,39 SCS has SCS with peripheral neuropathy of diabetes and clearly been more effective in relieving intractable leg causalgia may have up to 75% success rate.6,7,8,9 pain, than in treating axial back pain. SCS is at least as Kumar reported only moderate success with peripheral good if not better than re-operation in patients with neuropathy. Cases showing the best results were causal- FBS. For those who have failed surgery, it may be the gia (100%) and diabetic neuropathy (75%). Kumar PPR3021 4/3/2003 10:51 AM Page 6

6•fogel et al.

reported less responsiveness in post-herpetic neuralgia, In conclusion, SCS is an effective and safe therapy that other inter-costal neuralgias, spinal cord injury with improves the quality of life and activities of daily living, dysthesias and pain.9,45 in patients disabled by their chronic pain. The mode of SCS is less effective in treating amputation stump and action of SCS is not completely known, but is better phantom limb pain, cauda equina syndrome, and bone defined now, and may be related to controlling neuro- and joint pain syndrome. These cases responded initially transmitters, such as glycine and GABA. With dual with early pain relief, but all subsequently lost pain stimulation further inroads in the treatment of FBS are relief. In a small number of cases of amputation stump made. SCS continues to be most effective for neuro- and phantom limb pain, none were internalized.9,45 pathic pain and less effective for nociceptive mediated Multiple Sclerosis showed high early pain relief pro- pain. ducing a 92% internalization rate. However, perhaps due to the development of tolerance, there have been REFERENCES good short term then progressive worsening of pain 1. Melzack R, Wall PD. Pain mechanisms: a new theory. relief in multiple sclerosis.7,8 Science. 1965;150(699):971–979. More recently, excellent results have been obtained 2. Shealy CN, Mortimer JT, Reswick JB. Electrical inhi- in vascular claudication, Kumar reported 69% suc- bition of pain by stimulation of the dorsal columns: prelimi- cess.23,28,37,45 There is an interesting side effect in the nary clinical report. Anesth Analg. 1967;46(4):489–491. vascular patient, with the blockade of sympathetic 3. Long DM, Erickson DE. Stimulation of the posterior control of the vessels there is a relative increase in columns of the spinal cord for relief of intractable pain. Surg Neurol. 1975;4(1):134–141. perfusion. 4. Sweet WH. Control of pain by direct electrical In spinal cord injury, cases of complete paraplegia did stimulation of peripheral nerves. Clin Neurosurg. 1976;23: not respond. All cases with satisfactory responses had 103–111. incomplete paralysis, with the majority of their pain 5. Shealy CN. Dorsal column stimulation. Surg Neurol. below the level of the spinal cord lesion. Cauda equina 1977;7(4):192. syndrome had early success with erosion to poor pain 6. Burchiel KJ, et al. Prognostic factors of spinal cord control.6,7,8,10,46 stimulation for chronic back and leg pain. Neurosurgery. Average short-term success is 66% to 75% while 1995;36(6):1101–1110, discussion 1110–1111. long-term success is about 50%. North points out in 7. Burchiel KJ, et al. Prospective, multicenter study of 1993 that the late success rate is better if one excludes spinal cord stimulation for relief of chronic back and extrem- the implants needing revision of electrode or receiver.36 ity pain. Spine. 1996;21(23):2786–2794. These equipment failures should have pain relief 8. Kim SH, Tasker RR, Oh MY. Spinal cord stimulation for nonspecific limb pain versus neuropathic pain and restored once the implant is re-activated. spontaneous versus evoked pain. Neurosurgery. 2001;48(5): 1056–1064, discussion 1064–1065. SUMMARY 9. Kumar K, et al. Epidural spinal cord stimulation Chronic pain remains one of the most debilitating of all for treatment of chronic pain—some predictors of success. A medical disorders. Chronic pain may lead to loss of 15-year experience. Surg Neurol. 1998;50(2):110–120, discussion 120–121. employment, destruction of interpersonal relationships 10. Ohnmeiss DD, Rashbaum RF, Bogdanffy GM. and drug addiction. Reasonable long-term success with Prospective outcome evaluation of spinal cord stimulation in SCS may be achieved by adhering to these principles. patients with intractable leg pain. Spine. 1996;21(11): 1. Modest selection criteria. All patients should 1344–1250, discussion 1351. have exhausted all conservative treatments. 11. LeDoux MS, Langford KH. Spinal cord stimulation Narcotic usage alone should not be a non- for the failed back syndrome. Spine. 1993;18(2):191–194. 12. North RB, et al. Failed back surgery syndrome: exclusionary finding. Many patients will 5-year follow-up after spinal cord stimulator implantation. decrease drug usage after stimulator. Neurosurgery. 1991;28(5):692–699. 2. State of the art SCS equipment should be used. 13. Wall PD, Sweet WH. Temporary abolition of pain in 3. A trial period should be undertaken in all cases. man. Science. 1967;155(758):108–109. Some patients (15% to 55%) will be excluded 14. Hosobuchi Y, Adams JE, Rutkin B. Chronic thala- from the expense and possible risk of permanent mic stimulation for the control of facial anesthesia dolorosa. implantation, by failing a trial period. Arch Neurol. 1973;29(3):158–161. PPR3021 4/3/2003 10:51 AM Page 7

Chronic Limb Pain and Spinal Cord Stimulation •7

15. Kumar K, Nath R, Wyant GM. Treatment of chronic 32. Duggan AW, Foong FW. Bicuculline and spinal inhi- pain by epidural spinal cord stimulation: a 10-year experience. bition produced by dorsal column stimulation in the cat. Pain. J Neurosurg. 1991;75(3):402–407. 1985;22(3):249–259. 16. North RB, et al. Failed back surgery syndrome: 5- 33. Linderoth B, et al. Dorsal column stimulation year follow-up in 102 patients undergoing repeated operation. induces release of serotonin and substance P in the cat dorsal Neurosurgery. 1991;28(5):685–690, discussion 690–691. horn. Neurosurgery. 1992;31(2):289–296, discussion 296– 17. North RB, et al. Spinal cord stimulation for chronic, 297. intractable pain: superiority of “multi-channel” devices. Pain. 34. Basbaum AI, Fields HL. Endogenous pain control 1991;44(2):119–130. systems: brainstem spinal pathways and endorphin circuitry. 18. North RB, et al. A prospective, randomized study of Annu Rev Neurosci. 1984;7:309–338. spinal cord stimulation versus reoperation for failed back 35. Meyerson BA, et al. An experimental animal model surgery syndrome: initial results. Stereotact Funct Neurosurg. of spinal cord stimulation for pain. Stereotact Funct Neuro- 1994;62(1–4):267–272. surg, 1994;62(1–4):256–262. 19. North RB, Kidd DH, Piantadosi S. Spinal cord stim- 36. North RB, et al. Spinal cord stimulation for chronic, ulation versus reoperation for failed back surgery syndrome: intractable pain: experience over two decades. Neurosurgery, a prospective, randomized study design. Acta Neurochir Suppl 1993;32(3):384–394, discussion 394–395. (Wien). 1995;64:106–108. 37. Lang P. The treatment of chronic pain by epidural 20. Turner JA, Loeser JD, Bell KG. Spinal cord stimula- spinal cord stimulation—a 15 year follow up; present status. tion for chronic low back pain: a systematic literature Axone. 1997;18(4):71–73. synthesis. Neurosurgery. 1995;37(6):1088–1095, discussion 38. Long DM. The current status of electrical stimula- 1095–1096. tion of the nervous system for the relief of chronic pain. Surg 21. Long DM. Fifteen years of transcutaneous electrical Neurol. 1998;49(2):142–144. stimulation for pain control. Stereotact Funct Neurosurgery. 39. Bell GK, Kidd D, North RB. Cost-effectiveness 1991;56(1):2–19. analysis of spinal cord stimulation in treatment of failed back 22. Long DM. Conquering pain. Neurosurgery. surgery syndrome. J Pain Symptom Manage. 1997;13(5):286– 2000;46(2):257–259. 295. 23. Augustinsson LE, et al. Spinal cord stimulation in 40. Kemler MA, et al. Electrical spinal cord stimulation cardiovascular disease. Neurosurg Clin N Am. 1995;6(1): in reflex sympathetic dystrophy: retrospective analysis of 23 157–165. patients. J Neurosurg. 1999;90(1 Suppl):79–83. 24. Grill WM, et al. Emerging clinical applications of 41. Kemler MA, de Vet HC. Health-related quality of life electrical stimulation: opportunities for restoration of func- in chronic refractory reflex sympathetic dystrophy (complex tion. J Rehabil Res Dev. 2001;38(6):641–653. regional pain syndrome type I). J Pain Symptom Manage. 25. Moore R, Chester M. Neuromodulation for chronic 2000;20(1):68–76. refractory angina. Br Med Bull. 2001;59:269–278. 42. Kemler MA, et al. Impact of spinal cord stimulation 26. Linderoth B, et al. Release of neurotransmitters in on sensory characteristics in complex regional pain syndrome the CNS by spinal cord stimulation: survey of present state of type I: a randomized trial. Anesthesiology. 2001;95(1):72–80. knowledge and recent experimental studies. Stereotact Funct 43. Rho RH, et al. Complex regional pain syndrome. Neurosurg. 1993;61(4):157–170. Mayo Clin Proc. 2002;77(2):174–180. 27. Meyerson BA, Brodin E, Linderoth B. Possible 44. van De Beek WJ, et al. Neurophysiologic aspects of neurohumoral mechanisms in CNS stimulation for pain patients with generalized or multifocal tonic dystonia of reflex suppression. Appl Neurophysiol. 1985;48(1–6):175–180. sympathetic dystrophy. J Clin Neurophysiol. 2002;19(1):77– 28. Meyerson BA, Linderoth B. Mechanisms of spinal 83. cord stimulation in neuropathic pain. Neurol Res. 2000;22(3): 45. Kumar K, et al. Improvement of limb circulation in 285–292. peripheral vascular disease using epidural spinal cord stimu- 29. Simpson RK, Robertson CS, Goodman JC. Segmen- lation: a prospective study. J Neurosurg. 1997;86(4):662–669. tal recovery of amino acid neurotransmitters during posterior 46. Alo KM, Holsheimer J. New trends in neuromodu- epidural stimulation after spinal cord injury. J Am Paraplegia lation for the management of neuropathic pain. Neurosurgery. Soc. 1993;16(1):34–41. 2002;50(4):690–704. 30. Simpson RK Jr., Robertson CS, Goodman JC. 47. Devulder J, et al. Spinal cord stimulation: a valuable Release of segmental amino acid neurotransmitters in response treatment for chronic failed back surgery patients. J Pain to peripheral afferent and motor cortex stimulation: a pilot Symptom Manage. 1997;13(5):296–301. study. Life Sci. 1991;49(17):PL113–PL118. 48. Simpson BA. Spinal cord stimulation in 60 cases of 31. Duggan AW. Release of neuropeptides in the spinal intractable pain. J Neurol Neurosurg Psychiatry. 1991;54(3): cord. Prog Brain Res. 1995;104:197–223. 196–199. PPR3021 4/3/2003 10:51 AM Page 8

8•fogel et al.

49. Spiegelmann R, Friedman WA. Spinal cord stimula- 51. De La Porte C, Van de Kelft E. Spinal cord stim- tion: a contemporary series. Neurosurgery. 1991;28(1):65–70, ulation in failed back surgery syndrome. Pain. 1993;52(1): discussion 70–71. 55–61. 50. Tasker RR, DeCarvalho GT, Dolan EJ. Intractable 52. North RB. Spinal cord stimulation for chronic, pain of spinal cord origin: clinical features and implications intractable pain. Adv Neurol. 1993;63:289–301. for surgery. J Neurosurg. 1992;77(3):373–378. ORIGINAL ARTICLE

Biomechanical Evaluation of Relationship of Screw Pullout Strength, Insertional Torque, and Bone Mineral Density in the Cervical Spine Charles Alan Reitman, MD, Lyndon Nguyen, and Guy R. Fogel

help prevent implant loosening and add rigidity to the plate- Background: Understanding of implant failure mechanisms is im- screw construct. Some screws actually lock to the plate, while portant in the successful utilization of anterior cervical plates. Many in most cases, there is some kind of blocking plate or screw variables influence screw purchase, including the quality of the bone. head expansion to secure the screw to the plate. In these cases, The purpose of this study was to assess the relationship of screw pullout and screw insertional torque across a wide range of bone mineral the screws are initially placed securely per the surgeon’s own densities (BMDs). perception, in most instances without specific torque control. Screw pullout and stripping (exceeding maximal inser- Methods: A total of 54 cervical vertebrae in 12 cervical spines were tional torque) are possible modes of failure. Some factors af- evaluated for BMD using dual-energy x-ray absorptiometry scan- fecting the pullout strength of a cancellous bone screw are spe- ning. Actual and perceived peak torques of 3.5-mm anterior cervical cific to the screw design and include the major diameter of the screws were determined at each level followed by screw pullout screw, the length of engagement of the thread, and screw strength testing. thread depth and pitch.4 Furthermore, tapping was found to Results: A high correlation was observed between screw pullout reduce pullout force by an average of 8% compared with non- strength and BMD. However, there was a low correlation of peak tapped holes. In addition to these characteristics of the instru- insertional torque to pullout strength. mentation, the quality of the bone probably plays an important Conclusion: These findings suggest the quality of the bone is more role in the strength of fixation. As a result, there is particular instrumental in the success or failure of anterior cervical screws than concern regarding implant failure when used in the osteopo- is the insertional torque with which the screws are placed. rotic spine. Since bone mineral density (BMD) is one clinical measure of bone quality, there is interest in determining the Key Words: cervical spine, bone mineral density, screw pullout relationships between BMD and internal fixation. Although strength, insertional torque, biomechanics studies in the cervical spine in particular are limited, these data (J Spinal Disord Tech 2004;17:306–311) combined with more extensive reports in the thoracic and lumbar spine would suggest that there is a linear relationship of insertional torque and BMD to pullout force of vertebral body nterior cervical discectomy and fusion is frequently em- fixation screws. Aployed in the surgical management of the cervical spine. The purpose of this study was to define the relationships Recently, the benefit of the addition of anterior cervical plates of insertional torque and pullout strength of anterior cervical 1,2 has been established. There are several plates available, each plate locking screws to a wide range of BMDs. In addition, the with their own characteristics, with variation in features such difference in maximum torque and the torque applied by the 3 as implant thickness, rigidity, and dimensions. Success of feel of the surgeon was evaluated. We hypothesized that screw these implants is in part dependent on secure screw anchoring pullout strength would be dependent on both BMD and peak of the plate to the vertebral column. insertional torque. Additionally, it was felt that a linear rela- These systems initially used bicortical fixation screws tionship of torque to BMD could be established such that op- but have now evolved to safer unicortical cancellous screws. timal forces of screw application could be predicted based on All of these screws are locked to the plate in some fashion to BMD. Furthermore, it was suspected that a minimal BMD existed below which screw purchase would be unreliable. Received for publication April 4, 2003; accepted July 29, 2003. MATERIALS AND METHODS From the Baylor College of Medicine, Houston, TX, USA. A total of 54 fresh-frozen human vertebral bodies from Reprints: Charles Alan Reitman, MD, Baylor College of Medicine, 6560 Fan- nin, Suite 1900, Houston, TX 77030, USA (e-mail: creitman@bcm. 12 different cadaveric cervical spines (3 female, age 59–88 tmc.edu). years; 8 male, age 47–88 years) procured from ScienceCare Copyright © 2004 by Lippincott Williams & Wilkins Anatomic (Phoenix, AZ, USA) were used for this experiment.

306 J Spinal Disord Tech • Volume 17, Number 4, August 2004 J Spinal Disord Tech • Volume 17, Number 4, August 2004 Screw Pullout Strength and Insertional Torque

All 12 specimens contained C3–C7 vertebral bodies with parts of C2 and T1 and were kept frozen until the time of testing. Five screws failed to provide data owing to poor placement or a weak area of the bone consistent with a fracture, and one pullout test failed owing to failure of the coupling mechanism. Prior to testing, each cervical spine was vacuum sealed, frozen, and submitted for BMD testing (Dexascan; River Oaks Imaging, Houston, TX, USA). In addition, each specimen was subjected to radiographic evaluation to rule out any tumors or significant destructive changes. All specimens were allowed to reach room temperature and kept moist throughout the testing period. Two bone screws from the PEAK Polyaxial Anterior Cervical Plate System (DePuy AcroMed, Raynham, MA, USA) were inserted into each cervical vertebral body, C3–C7. All screws were inserted by a single orthopedic spine surgeon. These were cancellous screws specifically designed for unicortical placement in the cervical spine. All screws were 14 mm in length. At this length, all screws resulted in unicortical bone fixation in this sample of spines. Each screw was passed through an aluminum custom sleeve adapter prior to implantation to allow for mechanical pullout testing. The thickness of the adapter simulated the thickness of an average cervical plate. A special screwdriver with a torque transducer (Trans- ducer Techniques, Temecula, CA, USA) mounted to the handle enabled continuous acquisition of torque measurements as the surgeon was placing the screws. Predrilled holes using a 2.5-mm bit were made in the vertebral bodies to accommodate each screw. One bone screw was inserted until it stripped the bone (actual peak insertional torque). A second screw was threaded until the physician felt that a sufficient amount of torque was achieved to maximize holding force without stripping the bone (perceived peak insertional torque). Torque values were documented and analyzed in Microsoft Excel (Microsoft, Redmond, WA, USA). During insertion of all screws, the surgeon was blinded to the BMD of the specimen as well as the torque being generated with each screw. FIGURE 1. Screw pullout testing apparatus. Destructive mechanical pullout testing was performed by means of a custom pneumatic testing apparatus (Fig. 1). sitioned directly parallel to the configuration of the screw. One pneumatic cylinder (Parker Hannifin Corp., Cleveland, High tolerances during the machining process required that the OH, USA) applied a uniaxial tensile load at a constant rate of 5 sleeve be within approximately 0.2° of vertical to couple with N/s. The pneumatic cylinder was powered by an individual the load cell. This ensured that the pullout force was directed pneumatic valve (Proportion-Air, McCordsville, IN, USA) parallel to the alignment of the screw. The sleeve adapter was and controlled by LabView (National Instruments, Austin, threaded to the force transducer and loaded to failure. Only TX, USA). A universal “S” load cell fixed between the pneu- screws placed by feel were tested for pullout. Stripped screws matic cylinder and sleeve adapter noted the tension force at a were not tested in this manner and were placed only to deter- rate of 30 Hz throughout the loading process. Each specimen mine peak insertional torque. Maximum failure load was veri- was securely mounted to a multiaxial adjustable table vise and fied by graphing the force data over time and visually identi- floating x-y table to permit self-alignment during the testing fying the critical load. procedure. Use of the multiaxial vise allowed us to carefully Average failure loads and statistical comparisons were adjust the alignment. Owing to the intentional and careful de- calculated using a statistical analysis package to determine lin- sign of the sleeve and various adapters, the sleeve could be ear regression and square correlation coefficient as well as connected to load cell and testing apparatus only if it was po- analysis of variance (ANOVA).

RESULTS These relationships imply that the critical variable for screw A total of 54 vertebra from 12 cervical spines were purchase was the density of the bone into which it was placed, tested. This provided data for a broad range of BMDs from and not the force with which it was applied. Clearly, there was 0.256 to 1.273 g/cm2. a correlation of torque to BMD, which was also expected. The relationships of BMD, torque, and pullout strength However, despite the fact that denser bone could accommo- were evaluated by linear regression analysis. Figure 2 demon- date more torque, the amount of torque did not appear to be strates a high correlation of average BMD to pullout strength critical in preventing screw pullout. (R2 = 0.710). There was also a strong relationship of average These findings are consistent with those of Lim et al.5 BMD to maximum torque (R2 = 0.707) (Fig. 3). However, They were based on biomechanical evaluation of 6.5-mm an- there was much less correlation of torque to pullout strength terior lumbar screws, but results were similar in that pullout (R2 = 0.422) (Fig. 4). ANOVA was also used to assess inter- strength was strongly related to BMD but not to the peak actions of BMD, torque, and pullout strength (Figs. 5 and 6). insertional torque of the screws. Several other human and Relationships of torque to average BMD were significant (P = animal studies have been done in the thoracic and lumbar spine 0.000087), while pullout strength to average BMD showed a that support the high correlation of BMD to screw pullout trend (P = 0.093). force.5–12 Maximum torque compared with the perceived peak Fewer studies have been conducted in the cervical spine. torque results were assessed. On average, the surgeon’s per- Zink et al13 evaluated several variables including relationships ceived peak torque was about 85% of the actual peak torque. of BMD and screw torque and axial forces for anterior cervical Average maximum torque was found to be 2.56 ± 1.47 Nm, fixation of Caspar screws. They did not measure pullout while average perceived torque was 2.16 ± 1.28 Nm (P = strength directly. They found that torque was related to BMD. 0.003) (paired t test). They also established that larger rescue screws generated higher torques than standard screws. They concluded that for DISCUSSION BMD under 150 mg/mL, unicortical fixation with 3.5-mm Screw pullout strength is a critical factor in the success screws was inadequate fixation without additional reinforce- of an anterior cervical implant. These data showed that screw ment and recommended that all patients get BMDs prior to pullout strength was substantially related to BMD as expected surgery to help determine fixation and need for postoperative but had a much lower correlation with peak insertional torque. immobilization.

FIGURE 2. Relationship for BMD and pullout strength: y = 64.049x + 1.5126 (R2 = 0.710).

.(R2 = 0.707) 3.8488 מ FIGURE 3. Relationship of BMD to torque: y = 31.343x

FIGURE 4. Relationship of pullout strength to torque: y = 1.8503x + 14.363 (R2 = 0.422).

FIGURE 5. Relationship of maximum torque to average BMD (P = 0.000087).

FIGURE 6. Relationship of pullout strength to average BMD (P = 0.093).

Ryken et al14 also looked at properties of Caspar screws onstrated that the surgeon was able to place the screws satis- in the cervical spine. Contrary to our findings, they found a factorily without the need for feedback, had we been able to significant correlation of torque to pullout strength as well as standardize torque, there would have been further direct mea- BMD to pullout. These findings were more significant for bi- sure of pullout versus BMD, which would have provided ad- cortical purchase versus unicortical placement, and range of ditional support to our conclusions. bone densities was fairly narrow (0.787 ± 0.154 g/cm2). A recent report also found positive correlations with pullout CONCLUSIONS strength and BMD as well as pullout strength and insertional BMD significantly influences pullout strength of ante- torque of anterior unicortical cervical screws.15 However, in rior cervical screws. Peak insertional torque is related to BMD this study, midrange insertional torque was evaluated rather but as an independent variable does not demonstrate a strong than peak insertional torque. They additionally observed that influence on pullout strength. It is important for the clinician to longer screws (14 and 16 mm) had greater pullout strength than acknowledge that the critical factor contributing to pullout shorter screws (12 mm), and in general, their pullout strength strength is the density of the bone and not the force with which findings were much more significant in the presence of longer the screw is placed into the bone. The surgeon is capable of screws. BMD has also been shown to be important in implant placing a screw by feel that has adequate torque. Specific fixation in the posterior cervical spine.16 torque screwdrivers should not be necessary to ensure optimal It appears to be clear that pullout strength is affected by bone purchase, and the surgeon then does not risk exceeding BMD. This has been demonstrated in multiple studies. How- the peak insertional torque capacity of the bone. BMD should ever, effects of insertional torque have been less well estab- be considered when choosing a cervical implant. lished for anterior cervical screws. While a recent study suggests a high correlation of insertional torque to pullout REFERENCES strength, they did not measure peak insertional torque, and the 1. Wang JC, McDonough PW, Endow KK, et al. Increased fusion rates with torque in the midranges of screw placement was probably cervical plating for two level anterior cervical discectomy and fusion. Spine. 2000;25:41–45. more directly dependent on BMD. This is the first study to look 2. Wang JC, McDonough PW, Kanim LEA, et al. Increased fusion rates with at relationships of peak insertional torque to pullout strength of cervical plating for three level anterior cervical discectomy and fusion. unicortical anterior cervical screws over a large range of BMD. Spine. 2001;26:643–644. 3. Fogel GR, Reitman CA, Nguyen L, et al. Cervical plates: comparison of Results of this study imply that it is not as important to attain physical characteristics and in vitro push out strength. Spine J. (in press). near peak insertional torque values, and, in fact, care should be 4. Chapman JR, Harrington RM, Lee KM, et al. Factors affecting the pullout taken not to overtighten and strip the screw as this would more strength of cancellous bone screws. J Biomech Eng. 1996;118:391–398. 5. Lim TH, An HS, Evanich C, et al. Strength of anterior vertebral screw likely alter and weaken pullout properties. Of interest, we did fixation in relationship to bone mineral density. J Spinal Disord. 1995;8: look at the comparison of the surgeons’ perceived peak torque 121–125. versus the actual torque to failure. The average perceived 6. Coe JD, Warden KE, Herzig MA, et al. Influence of bone mineral density on the fixation of thoracolumbar implants. Spine. 1990;15:902–907. torque was 85%. We initially hypothesized that peak inser- 7. Hadjipavlou AG, Nicodemus CI, Al-Hamdan FA, et al. Correlation of tional torque would be important and were set to establish op- bone equivalent mineral density to pullout resistance of triangulated ped- timal torque values for screw placement depending on the icle screw construct. J Spinal Disord. 1997;10:12–19. 8. Oduyama K, Sato K, Abe E, et al. Stability of transpedicle screwing for BMD. The study invalidated this hypothesis and supports the osteoporotic spine. Spine. 1993;18:2240–2245. placement of screws by feel without the need to reach a mea- 9. Shimamoto N, Kotani Y, Shono Y, et al. Biomechanical evaluation of sured optimal torque. anterior spinal instrumentation systems for scoliosis: in vitro fatigue simulation. Spine. 2001;26:2701–2708. We did observe a linear relationship between pullout 10. Snyder BD, Zaltz I, Hall JE, et al. Predicting the integrity of vertebral bone strength and BMD down to about 0.4 g/mL. Below this den- screw fixation in anterior spinal instrumentation. Spine. 1995;20:1568– sity, pullout strength seems to reach a minimum of 106.8 N. 1574. 11. Wittenberg RH, Shea M, Swartz DE, et al. Importance of bone mineral Although we feel that the knowledge of BMD preoperatively is density in instrumented spine fusions. Spine. 1991;16:647–652. important to help determine the relative stability of internal 12. Zdeblick TA, Kunz DN, Cooke ME, et al. Pedicle screw pullout strength: fixation, this analysis does not establish a threshold safety correlation with insertional torque. Spine. 1993;18:1673–1676. 13. Zink PM. Performance of ventral spondylodesis screws in cervical verte- value above which implant failure in pullout would reliably be brae of varying bone mineral density. Spine. 1996;21:45–52. decreased. This question would have to be answered with fur- 14. Ryken TC, Clausen JD, Traynelis VC, et al. Biomechanical analysis of ther study. Furthermore, we did not standardize torque values bone mineral density, insertion technique, screw torque, and holding strength of anterior cervical plate screws. J Neurosurg. 1995;83:325–329. prior to pullout testing. One of the goals was to mimic the in- 15. Hitchon PW, Brenton MD, Coopes JK, et al. Factors affecting the pullout sertional torque that would be present during surgery and to strength of self-drilling and self-tapping anterior cervical screws. Spine. assess the need for a torque feedback device in placing screws. 2003;28:9–13. 16. Lovick DS, Ryken TC, Traynelis VC, et al. Assessment of primary and Therefore, all pullout testing was done on screws placed by salvage lateral mass screw insertion torque in a cadaveric model. J Spinal feel and not to a designated torque value. Although it was dem- Disord. 1997;10:431–435.

Case Studies Spinal epidural lipomatosis: case reports, literature review and meta-analysis Guy R. Fogel, MDa, Paul Y. Cunningham, III, MD, JDb, Stephen I. Esses, MDb,* aHouston Spine Surgery, 5225 Katy Freeway, Suite 600, Houston, TX 77007, USA bDepartment of Orthopedics, Baylor College of Medicine, 6560 Fannin, Suite 1900, Houston, TX 77030, USA Received 19 September 2003; accepted 13 May 2004

Abstract BACKGROUND CONTEXT: Symptomatic spinal epidural lipomatosis (SEL), a rare cause of spinal cord compression, has most often been associated with exogenous steroid use. PURPOSE: Identify four associations with SEL, correlate the associated groups with level of disease and compare treatment with outcome data in these groups. STUDY DESIGN/SETTING: Case reports of three patients and analysis of 104 cases from the literature. PATIENT SAMPLE: Three patients from the senior author’s practice. OUTCOME MEASURES: Not applicable. METHODS: The authors report three new cases of SEL not associated with steroid use. They review all available English literature and present a table of all 104 reported cases. RESULTS: The clinical course of three new patients is reported. CONCLUSIONS: Associated conditions are exogenous steroid use, obesity, endogenous steroid excess, and some remain idiopathic. Although SEL is a rare condition, our review of the literature reveals many more reported cases than previously thought. With increased awareness of this condition and improved imaging techniques, further studies of this disease should be undertaken. Ć 2005 Elsevier Inc. All rights reserved. Keywords: Dura mater/pathology; Lipoma/complications/pathology/radiography/surgery; Spinal cord compression/etiology

Introduction disease, it was often underdiagnosed. Symptomatic epidural lipomatosis was first described in 1975 [1]. Since that time, Spinal epidural lipomatosis (SEL) is a disease consisting a number of other cases have been diagnosed with the use of of an excessive deposition of normal adipose tissue in the imaging in combination with clinical symptoms, history, spinal canal, compressing the spinal cord. Symptomatic SEL surgical findings and the absence of other identifiable causes. is exceedingly rare and often associated with exogenous Patients may present with progressive and longstanding steroid use. Although less common, obesity and Cushing complaints of pain, weakness, numbness, incontinence, syndrome/disease (hypercortisolism) have also played a role ataxia, abnormal reflexes and even rarely paralysis. We in SEL. Epidural lipomatosis becomes symptomatic in report three new cases of SEL. rare occasions by causing compression of the spinal cord or nerve roots. Because SEL can mimic other common spine conditions, such as spinal stenosis and degenerative joint Case report 1 A 47-year-old man presented with low back pain of 2 FDA device/drug status: not applicable. years’ duration that began after a work-related injury. The Nothing of value received from a commercial entity related to this pain was situated in his lower back with radiation to the but- research. tocks and thighs bilaterally. It was associated with decreased * Corresponding author. Department of Orthopedics, Baylor College of Medicine, 6560 Fannin, Suite 1900, Houston, TX 77030, USA. Tel.: sensation and weakness in both lower extremities. This pain (713) 986-5740. was not associated with any bowel or bladder incontinence. E-mail address: [email protected] (S.I. Esses) The patient also had a history of neck pain with associated

1529-9430/05/$ – see front matter Ć 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.spinee.2004.05.252 G.R. Fogel et al. / The Spine Journal 5 (2005) 202–211 203 numbness, weakness and loss of sensation in the right upper extremity. His past medical history was significant for renal cancer that was treated by nephrectomy 6 years prior. The patient had no history of steroid use. On physical examination, the patient was 5 feet 6 inches tall and weighed 140 pounds, with a body mass index (BMI) of 23.4. He had no structural spinal deformity. His range of motion was limited on both flexion and extension. Some tenderness was noted over his lower back bilaterally. Neuro- logical examination revealed that the patient was intact with Fig. 2. First patient: axial computed tomography myelogram views of L4– no nerve root tension signs. A computed tomography (CT) L5 shown in Fig. 1 demonstrate the extrinsic compression of the thecal myelogram revealed prominent epidural fat from L2 to S1 sac just above the L4–L5 disc space. The compression is symmetrical and with thecal sac compression. Complete cutoff of the intrathe- is close to water density, certainly less dense than ligamentum flavum, cal dye occurred at the L4–L5 level. CT scan cuts through capsule or disc. There is an abnormal dural sac shape changed by the the L4–L5 and L5–S1 levels demonstrated thecal sac com- extrinsic pressures of the epidural lipoma. pression by an extrinsic circumferential mass consistent with epidural lipomatosis (Figs. 1 and 2). A laminectomy and decompression was performed from him severe low back pain. He denied leg pain after this L2 to L5. Operative findings included marked adipose tissue injury. The patient participated in physical therapy and took in the spinal canal and stenosis. The patient lacked any leave of work. Two days later, he was unable to continue neurological abnormalities postoperatively, although he did the physical therapy because of pain. still complain of lower back pain. Magnetic resonance imaging (MRI) sagittal spin-echo T1-weighted images demonstrated widening of fat tissue in the epidural spinal canal rounding and compressing the thecal Case report 2 sac and nerve roots. Hypertrophic spondylosis of L2–S1 in the lumbar spine was noted. Axial T1-weighted MRI of the A 48-year-old man developed an acute onset of low back lumbar spine just cephalic to L5–S1 shows circumferential pain that he attributed to a work-related fall. The patient hyperintense and homogenous tissue elevating the dural sac continued to work, and his pain eventually resolved. Subse- and the nerve root. A sagittal T2-weighted image showed a quently, the patient sustained a lifting injury at work, causing high contrast between adipose tissue, and the dural sac on T2-weighted image permitted an accurate evaluation of the extent of pathologic overgrowth of epidural fat in the spinal canal. Hypertrophy of facet joints and end plates with discal bulging resulted in mild canal stenosis at multiple levels and increased epidural fat posterior to the dural sac that appeared compressed and indented below L2 (Figs. 3 and 4). The patient received an epidural steroid shot. He reported that the pain became worse and began to radiate to both buttocks after the steroid injection. He was unable to walk after the steroid injection. He reported that his symptoms were occasionally accompanied by left leg numbness that extended to the bottom of the foot. His past medical history was negative. His past surgical history was significant for a gunshot wound to the stomach sustained during the Gulf War. He denies taking steroids. His only medication was tramadol. On physical examination, the patient was 6 feet tall and 270 pounds, with a BMI of 36.6. The patient had no spinal deformities. He stood with his lumbar spine flexed to 20 degrees, because of pain. He was unable to lie flat. Paraver- tebral muscle spasm was noted. He had no tenderness to palpation. His neurological examination showed that the Fig. 1. First patient: anteroposterior and lateral views of the lumbar myelogram demonstrate some extrinsic compression of the thecal contents at L3– patient was intact. A laminectomy and decompression was L4 with complete blockage of the intrathecal dye above the L4–L5 disc performed from L3-S1 with bilateral foraminotomy. Direct space. Hypertrophic spondylosis is seen radiographically only at L5–S1. removal of the epidural lipomatosis was not reported. 204 G.R. Fogel et al. / The Spine Journal 5 (2005) 202–211 pathologic overgrowth of epidural fat in the spinal canal. On the sagittal T1-weighted image, the posterior epidural stripe of hyperintense lipomatosis is 8 to 10 mm in width throughout the thoracic spine. Axial views of T5–T6 showed extradural compression from epidural fat. Axial cuts from T10– T11 showed restoration of the symmetry of the dural contents and cerebrospinal fluid without the extradural compression (Figs. 5 and 6). The patient was diagnosed with thoracic SEL and cervical spondylosis. A laminectomy and decompression was performed from C5 to C7 and T1 to T7. Although the patient made some improvement, at time of discharge he was unable to stand or walk and lacked bowel and bladder continence.

Discussion Fig. 3. Second patient: magnetic resonance imaging (MRI) sagittal spin- We conducted a complete review of the available English echo T1-weighted images demonstrate widening of fat tissue in the spinal literature, charting all reported cases of SEL, noting the most canal rounding and compressing the thecal sac and nerve roots. This abnormal fat-density compression is thicker between the disc levels and thinner at common associations of SEL, correlating these associations the areas of degenerative discal bulging. MRI sagittal image also shows to the location of the disease process and examining the hypertrophic degenerative changes at L2–S1. treatment to outcome data. We found 104 cases of SEL reported in the literature, including our three cases (Table 1). Postoperatively, the patient improved, although he still complained of low back pain and the inability to stand completely Pathogenesis erect. The patient started a supervised physical therapy pro- The underlying pathological mechanism of SEL is un- gram with some improvement of his back pain. known. A review of the literature and the data reported in the chart below reveal four categorical associations with SEL: exogenous steroid use, obesity, endogenous steroid excess Case report 3 or Cushing syndrome and an idiopathic group. The most common association is exogenous steroid use. SEL has been A 43-year-old man developed an acute onset of back pain after sustaining a lifting injury. The pain was situated in his mid-thoracic spine and radiated into his neck. Several days later, the pain became much worse and he was unable to walk or stand. He was immediately brought to the emergency room and admitted. His past medical history was negative. The patient had had three prior operations on his back for work-related injuries. The patient had no history of steroid use. On physical examination the patient weighed 270 pounds and measured 6 feet 4 inches, with a BMI in excess of 30. On palpation, the patient complained of tenderness throughout his spine. His range of motion was severely limited by pain. His motor examination revealed flexor paralysis of both lower extremities. There was also weakness in handgrip bilaterally. Deep tendon reflexes were absent in the lower extremities and for the triceps tendon bilaterally. Plantar stimulation elicited no response. Sensory examination Fig. 4. Axial T1-weighted magnetic resonance imaging of the lumbar spine revealed decreased sensation to pinprick from and including just cephalic to L5–S1 shows circumferential hyperintense and homogenous the T1 level and down. tissue elevating the dural sac and the nerve roots toward the posterior MRI of the thoracic spine with axial and sagittal T1- lamina, distorting and compressing the dural contents. Calculations of the dural sac to epidural fat anterior and posterior to the dural sac revealed a weighted images showed a marked high contrast between ratio of 1.05, and the epidural fat to spinal canal anteroposterior distance adipose tissue and the dural sac on T1-weighted images. was 67%. This would be a moderate compression Grade II according to The contrast permits an accurate evaluation of the extent of the classification of Borre el al [18]. G.R. Fogel et al. / The Spine Journal 5 (2005) 202–211 205

Fig. 6. Third patient: axial views of T5–T6 show displacement of the dural sac cerebrospinal ﬂuid without extradural compression from epidural fat.

Fig. 5. Third patient: the sagittal T1-weighted image of the thoracic spine reveals that the posterior epidural stripe of hyperintense lipomatosis is 8 to 10 mm in width throughout the thoracic spine. patient class may be caused by a pseudo-Cushing state exhib- iting elevated urinary free cortisol levels. Unfortunately, no studies of obese patients with SEL have reported cortisol documented in association with steroid use for many condi- levels to date. Furthermore, some investigators question tions, including transplantation, systemic lupus erythemato- whether obesity plays a causal role in SEL or is merely a sus, rheumatoid arthritis, Graves disease, chronic hepatitis, predisposing factor [6]. Borre et al reported 53 severe cases dermatomyositis, nephritic syndrome, glomerulonephritis, of SEL with 39 obese patients. Additionally, 4 were on sarcoidosis, Crohn’s disease, multiple sclerosis, chronic obstructive pulmonary disease, atopic dermatitis, diabetes mel- steroids, 2 were hypothyroid and 1 was obese, had hypothy- litus, prostatic cancer, lichen ruber planus, pineoblastoma, roidism and was taking steroids [18]. Our case review re- cerebral lymphoma, polyarthritis, asthma and polyarteritis vealed24.5% ofreported cases attributableto obesity alone(It nodosum. It is well established that hypercortisolism leads to is important to note that this percentage includes only those an accumulation of adipose tissue in a typical distribution, patients who did not take steroids and were classified as on the face, neck, trunk and mediastinum [2]. Hypertrophy of obese by either the case reporter or BMI.) adipose tissue already present in the spinal canal is theorized Cushing syndrome/disease from endogenous sources is to be the cause of SEL in certain cases of exogenous the third associated category in SEL. To date, only three steroid use [3]. Based on our review, 55.3% of all reported cases of Cushing syndrome associated with SEL have been cases in the English literature were associated with exoge- reported in the English literature accounting for 3.2% of nous steroid use. Although the majority of these cases SEL cases [7–9]. were associated with long-term steroid use, three of the We found no identifiable association with SEL in 17% cases arose from multiple epidural steroid injections: cases of the cases. This patient group includes those patients who 13, 43 and 47. Patient 13 received a total of 103 injections did not take exogenous steroids, were not obese and did not over 12 years, eventually resulting in an abrupt onset of have an underlying Cushing syndrome/disease to account for neurologic deficits at the end of this period. Patient 43 re- the SEL. Unfortunately, some of these patients also lacked ceived a series of injections over a period of 3 years totaling certain of the data criteria. Of the 16 cases, 2 of these 1,200 mg of methylprednisolone. Her physical examination lacked data on the patient’s BMI. However, after consider- revealed sequelae of Cushing syndrome, including moon ation, we opted to include these cases in our analysis to be facies and buffalo hump. Patient 47 received a total of five complete in analysis of all known cases in the English injections; the last three injections reportedly exacerbated his literature. symptoms. In fact, one of these patients developed Cushing Hypothyroidism in previous papers has been associated syndrome sequelae from these local injections [3–5]. with SEL [4,6,10,11]. Upon review, we disagree with this Obesity is the second most common associated category conclusion. There has been only one reported case of hypo- of SEL. Koch et al. [2] hypothesized that obesity in this thyroidism associated with SEL [11]. This patient was obese, 206 G.R. Fogel et al. / The Spine Journal 5 (2005) 202–211

Table 1 One hundred and four cases of spinal epidural lipomatosis from English literature review with diagnosis, presence of steroid usage or obesity, type and outcome of treatment Case Reference Pathology Levels Steroids Obesity Treatment Outcome Glomerulonephritis 1 [1] Transplantation C7–L4 Yes NR Laminectomy Improved 2 [20] Asthma T2–T10 Yes Yes Laminectomy Unchanged 3 [21] Arteritis nodosum T1–T9 Yes NR Laminectomy Improved 4 [22] Thyroiditis L4–S1 Yes Yes Laminectomy CR Chronic hepatitis 5 [23] Cushing syndrome T7–T9 Yes Yes Laminectomy CR Polycystic kidney 6 [23] Cushing syndrome T2–T10 Yes Yes Laminectomy Improved Polyarthritis 7 [23] Cushing syndrome T2–T10 Yes NR Laminectomy CR Glomerulonephritis 8 [24] Transplantation S1–S3 Yes Yes Laminectomy Improved 9 [25] Asthma L4–S1 Yes NR Laminectomy CR 10 [26] Urethral stenosis T1–L5 Yes NR Steroid taper CR 11 [27] Asthma T4–T8, L4–L5 Yes NR Laminectomy Unchanged 12 [28] Polyarthritis, asthma L4–S1 Yes No Laminectomy CR Multiple sclerosis 13 [28] Multiple epidural steroid injections T10–L5 Yes No Laminectomy Improved 14 [29] Heart transplant L3–L4 Yes Yes Laminectomy CR 15 [30] Asthma T4–T8 Yes Yes Laminectomy Unchanged 16 [30] Asthma T5 Yes Yes High-dose steriods NR 17 [31] Rheumatoid arthritis T4–T8 Yes NR Laminectomy Improved 18 [32] Heart transplant T4–T10 Yes Yes Laminectomy LE paretic 19 [32] Heart transplant T2–T10 Yes NR Steroid taper CR 20 [17] Asthma T4–T9 Yes NR Laminectomy Improved 21 [17] Sarcoidosis T4–T9, L5–S1 Yes NR Laminectomy Improved 22 [17] Crohn disease T2–T8 Yes NR Observation Lost of FU 23 [17] Prostatic cancer T4–T8 Yes NR Medical Improved 24 [17] Pineoblastoma T3–T9 Yes NR Laminectomy Worse 25 [13] Heart-lung transplant T6–T8 Yes No Laminectomy CR 26 [33] Heart transplant L3–S1 Yes NR Laminectomy CR Asthma Weight reduction 27 [33] Treated with steroid inhaler L3–S1 Yes Yes and steroid taper Improved 28 [33] Chronic obstructive pulmonary disease T5–T12 Yes NR Laminectomy Unchanged 29 [33] Lichen ruber planus L3–S1 Yes NR Steroid taper CR 30 [33] Renal transplantation T4–T7 Yes NR Laminectomy Improved 31 [34] Renal transplantation L5–S2 Yes Yes Laminectomy CR 32 [34] Cerebral lymphoma L4–L5 Yes NR None NR Systemic lupus 33 [35] erythematosis L1–S1 Yes Yes Laminectomy CR Systemic lupus 34 [36] erythematosis T3–T9 Yes Yes Laminectomy Improved Radiation pneumonitis 35 [37] For adenocarcinoma T8–T10 Yes NR Laminectomy Improved 36 [38] Nephrotic syndrome T5–T10 Yes NR Steroid taper Improved Atopic dermatitis 37 [39] Cushing syndrome T6–T9 Yes NR Laminectomy Paraplegic 38 [40] Renal transplantation T3–T9 Yes NR Laminectomy CR 39 [41] Renal transplantation T1–T12 Yes Yes Weight reduction CR and steroid taper 40 [42] Dermatomyositis T4–T7 Yes NR Laminectomy and CR steroid taper 41 [43] T-cell lymphoma T10–T11 Yes NR Laminectomy Unchanged 42 [44] Polymyalgia rheumatica L4–L5 Yes Yes Laminectomy CR Cushingoid sequelae from multiple epidural steroid 43 [4] injections L3–S1 Yes NR Steroid taper CR 44 [45] Juvenile rheumatoid arthritis T4 Yes Yes Laminectomy and CR steroid taper (Continued) G.R. Fogel et al. / The Spine Journal 5 (2005) 202–211 207

Table 1 Continued Case Reference Pathology Levels Steroids Obesity Treatment Outcome 45 [46] Ulcerative colitis T1–T11 Yes Yes Laminectomy Improved 46 [47] Temporal arteritis T1–T10 Yes NR Laminectomy Improved None Multiple epidural steroid 47 [5] injections L2–L5 Yes No Laminectomy CR Systemic lupus erythematosis Weight reduction 48 [48] Cushing syndrome Thoracolumbar Yes Yes and steroid taper Improved Juvenile rheumatoid arthritis 49 [49] Cushing syndrome T6–T7 Yes NR Laminectomy CR 50 [49] Juvenile rheumatoid arthritis L5–S1 Yes NR Laminectomy Improved Cushing syndrome None; previous anabolic 51 [50] steroid use for body building Lumbar Yes No Bed rest, analgesics Improved Thrombocytopenia purpura 52 [51] Human immunodeﬁciency virus L5–S1 Yes No Laminectomy Improved treated with ritonavir 53 [35] None T1–T10 No Yes Weight reduction CR 54 [35] None T6–T8 No Yes Laminectomy CR 55 [35] None L5–S1 No Yes Laminectomy Improved 56 [15] None T5–T8 No Yes Laminectomy Improved 57 [52] None L4–L5 No Yes Laminectomy NR 58 [53] None L3–L5 No Yes Weight reduction CR 59 [16] None L5–S1 No Yes Weight management CR 60 [6] Diabetes mellitus T3–T9 No Yes Laminectomy Unchanged 61 [11] Hypothyroidism T1–T12 No Yes Laminectomy Unchanged 62 [54] None L4–S1 No Yes Laminectomy Almost CR 63 [55] None L3–S1 No Yes Weight reduction CR 64 [56] None L3–S1 No Yes Weight reduction Improved 65 [57] Prostatic cancer treated L2–L4 No Yes Weight reduction and CR with cyproterone acetate cyproterone cessation 66 [58] None L5–S1 No Yes Weight reduction CR 67 [59] None L4–L5 No Yes Laminectomy CR 68 [59] None L4–S1 No Yes Weight reduction CR 69 [60] None L5–S1 No Yes Weight reduction Unchanged 70 [60] None L4–S1 No Yes Weight reduction Improved 71 Our case None L3–L5 No Yes Laminectomy Improved 72 [61] Spermatocytic seminoma L4–S1 No Yes Laminectomy CR 73 [19] None T1–T10 No Yes Laminectomy Improved 74 [62] None L4–S1 No Yes Physiotherapy NR 75 Our case Cervical spondylosis T1–T7 No Yes Laminectomy Paraplegic 76 [63] None T3–T7 No No Laminectomy CR 77 Our case None L2–L5 No No Laminectomy CR 78 [64] None T1–T10 No No Laminectomy Improved 79 [12] None L4–S1 No No Laminectomy CR 80 [65] None L3–S2 No No Laminectomy Improved 81 [66] None L4–S1 No No Laminectomy and Improved endoscopic decompression Cerebral palsy Intrathecal baclofen 82 [67] Pump placement Thoracolumbar No No Laminectomy Unknown 83 [17] None T4–T8 No NR Laminectomy Improved 84 [68] None L3–L4 No No Laminectomy Improved 85 [68] None L3–L5 No No Laminectomy CR 86 [68] None L3–S1 No No Laminectomy Improved 87 [10] None T3–T10 No No Laminectomy CR 88 [69] None C6–L3 No No Laminectomy Improved 89 [70] None T4–T9 No No Laminectomy CR 90 [71] None L5–S1 No NR Laminectomy Improved 91 [19] None T6–T8 No No Laminectomy Improved (Continued) 208 G.R. Fogel et al. / The Spine Journal 5 (2005) 202–211

Table 1 Continued Case Reference Pathology Levels Steroids Obesity Treatment Outcome Cushing disease Surgical removal 92 [9] Pituitary microadenoma C7–L1 No NR of pituitary tumor Improved Adrenal tumor 93 [8] Cushing syndrome C5–T2 No No Tumor removal Improved Bronchial carcinoid Medical: metyrapone, 94 [3] Cushing syndrome T5–T10 No NR ketaconazole Improved 95 [72] Diabetic, paraplegic T1–T9 No Yes Laminectomy Improved Klippel-Trenaunay-Weber 96 [73] syndrome Thoracic No NR Laminectomy Improved Obese posttraumatic cauda 97 [74] equina syndrome Lumbar No Yes Laminectomy Improved 98 [75] Cushing syndrome Thoracic No NR Medical management Improved Cushing syndrome with pituitary tumor Surgical removal of 99 [76] (macroprolactinoma) Thoracic No NR pituitary tumor Improved 100 [77] Becker nevus associated T4–T7 No NR Laminectomy Improved 101 [78] None Lumbar No 102 [78] None Lumbar No Yes Laminectomy Improved 103 [78] None Lumbar No Yes Laminectomy Improved 104 [78] None Lumbar No Yes Laminectomy Improved CRϭcomplete recovery; FUϭfollow-up; LEϭlower extremity; NRϭnot reported.

the second leading association with SEL. There is the persua- dural sac. Grade II SEL was symptomatic in 14% of cases. sive argument that hypothyroidism is associated with gener- Grade III was defined as the epidural fat more than 75% of alized fat deposition resulting from decreased lipolysis, canal width and the dural 30% the width of the fat. All However, the data are insufficient to link hypothyroidism Grade III SEL cases were symptomatic. In Grade III cases, and SEL [3,4,10,11]. This linkage would require cases of there was a 42% rate of associated substantial pathology, nonobese patients with hypothyroidism and SEL to be con- such as disc herniation or stenosis. Also in the Grade III sidered an independent risk factor. The case of an obese SEL, the epidural fat produces centripetal pressures on the patient with hypothyroidism and SEL can only be considered thecal sac changing the morphology or shape. Most com- a case of obesity associated with SEL. monly a trifid or Y shape may be seen. Borre et al. [18] report six other patterns seen in axial magnetic imaging. Imaging Level of disease Most case reports relied heavily on CT imaging and myelography [12,13]. However, MRI is now recognized as Examining all available case reports, we were able to the most sensitive and specific modality for evaluating fatty determine incidences of spinal level involvement. From the tissue [14–16]. T1-weighted images differentiate epidural reported cases, 45.8% had thoracic involvement only, 43.6% fat from dural content with a high degree of specificity and had lumbosacral involvement only and 10.6% had involve- allow for measurement of adipose thickness. Quint et al. ment in both thoracic and the lumbosacral area. We cor- [17] conducted a study in which 28 normal patients were related location of disease to the associated categories to imaged and their epidural adipose measured. The mean sagit- ascertain whether any trends existed. Of the 52 patients with a tal thickness of their epidural fat was 4.6 mm with a normal history of steroid use, 55.8% were found to have thoracic range of 3 to 6 mm. In contrast, imaging in 6 patients with involvement only, 32.7% with lumbosacral involvement SEL revealed a mean thickness of 8 mm [17]. Borre et al. only and 11.5% with both thoracic and lumbosacral involve- [18], on the lumbar MRIs of 2,258 patients, measured the ment. Data from the patients in the obese category reveal a anterior posterior diameters of the dural sac and spinal stronger trend: 69.6% had lumbosacral involvement only, canal and the thickness of the epidural fat. Borre et al. de- and 30.4% had thoracic involvement only. The idiopathic veloped an MRI classification based on these ratios. Grade 0 group consisted of 16 patients. Of the 16 patients, 37.5% or normal was defined as epidural fat less than 40% of canal had thoracic spine involvement only, 50% had lumbosacral width, and dural 150% width of epidural fat. Grade I was involvement only and 12.5% had involvement of both the defined as epidural fat less than 50% of the canal width lumbosacral and thoracic spine. Of the three patients with and less than 50% of the dural width. Grade I was not Cushing syndrome/disease from endogenous sources, symptomatic. Grade II was defined as the epidural fat 50% 66.6%, had both thoracic and lumbosacral involvement and to 75% of the canal and 100% to 150% the width of the 33.3% had thoracic involvement only. Although four of the G.R. Fogel et al. / The Spine Journal 5 (2005) 202–211 209 cases reported cervical involvement, these cases had SEL by physiotherapy and nonsteroidal anti-inflammatory medi- throughout the spine and were included in the percentages cation, not weight loss. The other patient failed to lose of thoracic and lumbosacral involvement. We have found weight. Because conservative care was not successful, they no cases of SEL isolated to the cervical region. were not included in that group. Based on these data, weight Investigators initially reported a significantly higher inci- loss as a treatment modality appears to be very successful dence of thoracic SEL. This was theorized to be secondary and should be considered the first line of treatment in this to the fact that the thoracic region has the largest amount patient group, with surgical correction reserved for those of epidural fat [10,17,19]. Contrary to this popular belief, our patients who fail to respond clinically to a weight loss plan case review indicates that SEL is found in approximately or are unable to lose weight. the same number of cases in both the thoracic and the lumbo- One of the patients in the idiopathic group lacked outcome sacral region. The previous causal relationship is no longer data and was therefore considered to have a poor outcome. All present with this new information. It is clear that further stud- of the patients in this group were treated by laminectomy, ies into the etiology of SEL are required. with a success rate of 93.75%. The only patient without improvement lacked outcome data. Surgical treatment of patients with no discernable cause for SEL appears to be Treatment the treatment of choice.The three patients who were diag- Treatment of SEL ranges from conservative management nosed with SEL associated with endogenous steroid excess to surgical excision. The success rates were calculated as- or Cushing disease/syndrome were treated differently. Be- suming that none of the patients in the surgical group were cause all three of these patients had an underlying disorder first treated conservatively. In the event that the reviewed causing endogenous steroid excess, the underlying disorder was treated. Two of the patients had surgical removal of data revealed a failed conservative treatment that then pro- the tumor causing the steroid excess, and the remaining ceeded to surgery, these patients were included in both patient was treated with ketaconazole, an inhibitor of ste- groups or were specifically discussed. In the steroid patient roidogenesis. All three of the patients improved. group, three of the cases lacked outcome data and are therefore considered to have no improvement for purposes of this analysis. Thirty-nine patients received a laminectomy and Conclusions debulking, accounting for 75% of those patients in the Symptomatic SEL is a rare condition consisting of excess steroid group (39 of 52). The success rate of this modality adipose tissue in the spinal canal causing compression of was 77% (31 of 39 patients receiving a laminectomy had the spinal cord and resulting neurologic symptoms. Four results ranging from improved symptoms to complete recov- categories have been identified as associated with SEL: ex- ery). The remaining 13 patients (25%) received a combina- ogenous steroid use, obesity, endogenous steroid excess and tion of different medical treatments, including weight loss, idiopathic. Thoracic and lumbosacral levels are usually af- steroid taper, analgesics, bed rest and observation. One of fected by SEL, with the incidence between the two roughly the 13 received high-dose steroids for symptoms. The suc- equal. No case of isolated cervical involvement was found. cess rate of medical management of these patients was 77%. A new MRI grading scale may help define those patients Borre et al. [18] reported that 26 patients were treated medi- who will improve expectantly (Grade I) from those who cally and 18 surgically. In the surgical group 16 of 18 had may require surgical decompression (Grade III). Surgical associated spinal pathology treated, such as disc herniation treatment in the Grade III SEL may also treat commonly and stenosis decompression. Two patients were decom- associated degenerative stenosis, and facet pathology. Obese pressed surgically with such a high ratio of epidural fat to patients tend to develop SEL in the lumbosacral region canal and dural width and presented with neurologic com- three times more often than in the thoracic, whereas steroid promise and cauda equina syndrome [18]. Although both use tends to cause SEL in the thoracic region slightly less modalities appeared to treat SEL successfully in approxi- than twice as often as in the lumbosacral region. Obese mately 75% of the cases, these data do not reveal criteria patients should be managed by diet alone initially, with for selecting a treatment modality. surgery reserved for those without a significant clinical The obese patient group was split between surgical cor- response. Although SEL is a rare condition, our review of the rection and weight loss treatment modalities: 52.2% were literature reveals many more reported cases than previously treated by laminectomy and debulking, 47.8% by weight thought. With increased awareness of this condition and loss. The success rate of the surgically managed group was improved imaging techniques, further studies of this disease 66.7%. The patients managed conservatively by weight should be undertaken. loss improved in 81.8% of the cases, with the one patient lacking outcome data considered having no improvement. Two patients not included in the conservative group pre- References viously failed conservative treatment before surgical treat- [1] Lee M, Lekias J, Gubbay SS, Hurst PE. Spinal cord compression by ment was administered. One of the patients was managed extradural fat after renal transplantation. Med J Aust 1975;1(7):201–3. 210 G.R. Fogel et al. / The Spine Journal 5 (2005) 202–211

[2] Koch CA, Doppman JL, Patronas NJ, Nieman LK, Chrousos GP. Do [27] Russell NA, Belanger G, Benoit BG, et al. Spinal epidural lipomatosis: glucocorticoids cause spinal epidural lipomatosis? When endocrinol- a complication of glucocorticoid therapy. Can J Neurol Sci 1984;11(3): ogy and spinal surgery meet. Trends Endocrinol Metab 2000;11(3): 383–6. 86–90. [28] Roy-Camille R, Mazel C, Husson JL, Saillant G. Symptomatic spinal [3] Koch CA, Doppman JL, Watson JC, Patronas NJ, Nieman LK. Spinal epidural lipomatosis induced by a long-term steroid treatment. Review epidural lipomatosis in a patient with the ectopic corticotropin syn- of the literature and report of two additional cases. Spine 1991;16(12): drome. N Engl J Med 1999;341(18):1399–400. 1365–71. [4] McCullen GM, Spurling GR, Webster JS. Epidural lipomatosis com- [29] Randall BC, Muraki AS, Osborn RE, Brown F. Epidural lipomatosis plicating lumbar steroid injections. J Spinal Disord 1999;12(6):526–9. with lumbar radiculopathy: CT appearance. J Comput Assist Tomogr [5] Sandberg DI, Lavyne MH. Symptomatic spinal epidural lipomatosis 1986;10(6):1039–41. after local epidural corticosteroid injections: case report. Neurosurgery [30] Jungreis CA, Cohen WA. Spinal cord compression induced by steroid 1999;45(1):162–5. therapy: CT ﬁndings. J Comput Assist Tomogr 1987;11(2):245–7. [6] Badami JP, Hinck VC. Symptomatic deposition of epidural fat in a [31] Haid RW, Kaufman HH, Schochet SS, Jr, Marano GD. Epidural lipo- morbidly obese woman. Am J Neuroradiol 1982;3(6):664–5. matosis simulating an epidural abscess: case report and literature [7] Doppman JL. Epidural lipomatosis. Radiology 1989;171(2):581–2. review. Neurosurgery 1987;21(5):744–7. [8] Noel P, Pepersack T, Vanbinst A, Alle JL. Spinal epidural lipomatosis [32] Zampella EJ, Duvall ER, Sekar BC, et al. Symptomatic spinal epidural in Cushing’s syndrome secondary to an adrenal tumor. Neurology lipomatosis as a complication of steroid immunosuppression in cardiac 1992;42(6):1250–1. transplant patients. Report of two cases. J Neurosurg 1987;67(5): [9] Sivakumar K, Sheinart K, Lidov M, Cohen B. Symptomatic spinal 760–4. epidural lipomatosis in a patient with Cushing’s disease. Neurology [33] Fessler RG, Johnson DL, Brown FD, et al. Epidural lipomatosis in 1995;45(12):2281–3. steroid-treated patients. Spine 1992;17(2):183–8. [10] Selmi F, Davies KG, Sharma RR, Redfern RM. Idiopathic spinal [34] Peeters FL, Koster PA. Epidural lipomatosis simulating spinal canal extradural lipomatosis in a non-obese otherwise healthy man. Br J stenosis or herniated nucleus pulposus. Rofo Fortschr Geb Rontgenstr Neurosurg 1994;8(3):355–8. Neuen Bildgeb Verfahr 1991;154(3):342–3. [11] Toshniwal PK, Glick RP. Spinal epidural lipomatosis: report of a [35] Haddad SF, Hitchon PW, Godersky JC. Idiopathic and glucocorticoid- case secondary to hypothyroidism and review of literature. J Neurol induced spinal epidural lipomatosis. J Neurosurg 1991;74(1):38–42. 1987;234(3):172–6. [36] Healy ME, Hesselink JR, Ostrup RC, Alksne JF. Demonstration by [12] Bednar DA, Esses SI, Kucharczyk W. Symptomatic lumbar epidural magnetic resonance of symptomatic spinal epidural lipomatosis. Neu- lipomatosis in a normal male. A unique case report. Spine 1990;15 rosurgery 1987;21(3):414–5. (1):52–3. [37] Soloniuk DS, Pecoraro SR, Munschauer FE. Myelopathy secondary to [13] Hierholzer J, Benndorf G, Lehmann T, et al. Epidural lipomatosis: spinal epidural lipomatosis. A case report. Spine 1989;14(1):119–22. case report and literature review. Neuroradiology 1996;38(4):343–8. [38] Maehara T, Tanohata K, Noda M, Nakayama D. Medically treated [14] Hierholzer J, Vogl T, Hosten N, Lanksch W, Felix R. Imaging in steroid-induced epidural lipomatosis. Neuroradiology 1988;30(3):281. epidural lipomatosis. Crit Rev Neurosurg 1998;8(5):279–81. [39] Kaplan JG, Barasch E, Hirschfeld A, et al. Spinal epidural lipomatosis: [15] Kumar K, Nath RK, Nair CP, Tchang SP. Symptomatic epidural lipo- a serious complication of iatrogenic Cushing’s syndrome. Neurology matosis secondary to obesity. Case report. J Neurosurg 1996;85(2): 1989;39(8):1031–4. 348–50. [40] Tobler WD, Weil S. Epidural lipomatosis and renal transplantation. [16] Qasho R, Ramundo OE, Maraglino C, Lunardi P, Ricci G. Epidural Surg Neurol 1988;29(2):141–4. lipomatosis with lumbar radiculopathy in one obese patient. Case [41] Vazquez L, Ellis A, Saint-Genez D, Patino J, Nogues M. Epidural report and review of the literature. Neurosurg Rev 1997;20(3):206–9. lipomatosis after renal transplantation—complete recovery without [17] Quint DJ, Boulos RS, Sanders WP, et al. Epidural lipomatosis. Radiol- surgery. Transplantation 1988;46(5):773–4. ogy 1988;169(2):485–90. [42] Buthiau D, Piette JC, Ducerveau MN, et al. Steroid-induced spinal [18] Borre DG, Borre GE, Aude F, Palmieri GN. Lumbosacral epidural epidural lipomatosis: CT survey. J Comput Assist Tomogr 1988;12(3): lipomatosis: MRI grading. Eur Radiol 2003;13(7):1709–21. 501–3. [19] Stern JD, Quint DJ, Sweasey TA, Hoff JT. Spinal epidural lipomatosis: [43] Pennisi AK, Meisler WJ, Dina TS. Lymphomatous meningitis and two new idiopathic cases and a review of the literature. J Spinal steroid-induced epidural lipomatosis: CT evaluation. J Comput Assist Disord 1994;7(4):343–9. Tomogr 1985;9(3):595–8. [20] Butcher DL, Sahn SA. Epidural lipomatosis: a complication of cortico- [44] Taborn J. Epidural lipomatosis as a cause of spinal cord compression steroid therapy. Ann Intern Med 1979;90(1):60. in polymyalgia rheumatica. J Rheumatol 1991;18(2):286–8. [21] Godeau P, Brunet P, Wechsler B, Fohanno D. [Compressive extradural [45] Perling LH, Laurent JP, Cheek WR. Epidural hibernoma as a complica- lipomatosis as an unusual complication of corticosteroid therapy: one tion of corticosteroid treatment. Case report. J Neurosurg 1988;69 case (author’s translation)]. Nouv Presse Med 1979;8(47):3889–91. (4):613–6. [22] Lipson SJ, Naheedy MH, Kaplan MM, Bienfang DC. Spinal stenosis [46] Andress HJ, Schurmann M, Heuck A, Schmand J, Lob G. A rare case caused by epidural lipomatosis in cushing’s syndrome. N Engl J of osteoporotic spine fracture associated with epidural lipomatosis Med 1980;302(1):36. causing paraplegia following long-term cortisone therapy. Arch [23] Chapman PH, Martuza RL, Poletti CE, Karchmer AW. Symptomatic Orthop Trauma Surg 2000;120(7-8):484–6. spinal epidural lipomatosis associated with Cushing’s syndrome. Neu- [47] Pinsker MO, Kinzel D, Lumenta CB. Epidural thoracic lipomatosis rosurgery 1981;8(6):724–7. induced by long-term steroid treatment case illustration. Acta Neu- [24] Guegan Y, Fardoun R, Launois B, Pecker J. Spinal cord compression rochir (Wien) 1998;140(9):991–2. by extradural fat after prolonged corticosteroid therapy. J Neurosurg [48] Crayton HE, Partington CR, Bell CL. Spinal cord compression by 1982;56(2):267–9. epidural lipomatosis in a patient with systemic lupus erythematosus. [25] Archer CR, Smith KR. Extradural lipomatosis simulating an acute Arthritis Rheum 1992;35(4):482–4. herniated nucleus pulposus. Case report. J Neurosurg 1982;57(4): [49] Arroyo IL, Barron KS, Brewer EJ. Spinal cord compression by epi- 559–62. dural lipomatosis in juvenile rheumatoid arthritis. Arthritis Rheum [26] George WE, Wilmot M, Greenhouse A, Hammeke M. Medical man- 1988;31(3):447–51. agement of steroid-induced epidural lipomatosis. N Engl J Med 1983; [50] Fiirgaard B, Madsen FH. Spinal epidural lipomatosis. Case report and 308(6):316–9. review of the literature. Scand J Med Sci Sports 1997;7(6):354–7. G.R. Fogel et al. / The Spine Journal 5 (2005) 202–211 211

[51] Ebright JR, Stellini MA, Tselis AC. Spinal epidural lipomatosis in a [65] Kurt E, Bakker-Niezen SH. Neurogenic claudication by epidural lipo- human immunodeﬁciency virus-positive patient receiving steroids and matosis: a case report and review of literature. Clin Neurol Neurosurg protease inhibitor therapy. Clin Infect Dis 2001;32(5):E90–1. 1995;97(4):354–7. [52] Gero BT, Chynn KY. Symptomatic spinal epidural lipomatosis without [66] Frank E. Endoscopic suction decompression of idiopathic epidural exogenous steroid intake. Report of a case with magnetic resonance lipomatosis. Surg Neurol 1998;50(4):333–5. imaging. Neuroradiology 1989;31(2):190–2. [67] Huraibi HA, Phillips J, Rose RJ, et al. Intrathecal baclofen pump [53] Beges C, Rousselin B, Chevrot A, et al. Epidural lipomatosis. Interest implantation complicated by epidural lipomatosis. Anesth Analg of magnetic resonance imaging in a weight-reduction treated case. 2000;91(2):429–31. Spine 1994;19(2):251–4. [68] Lisai P, Doria C, Crissantu L, et al. Cauda equina syndrome secondary [54] Stambough JL, Cheeks ML, Keiper GL. Nonglucocorticoid-induced to idiopathic spinal epidural lipomatosis. Spine 2001;26(3):307–9. lumbar epidural lipomatosis: a case report and review of literature. [69] Hogg JP, Shank T, Gingold M, et al. Childhood presentation of idio- J Spinal Disord 1989;2(3):201–7. pathic epidural lipomatosis: a case report with magnetic resonance [55] Berenbaum F, Revel M, Deshays C, Rousselin B, Amor B. [Lumbo- imaging and pathologic conﬁrmation. J Child Neurol 1996;11(3): radicular pain caused by epidural lipomatosis in an obese patient: 236–40. recovery after hypocaloric diet]. Rev Rhum Mal Osteoartic 1992; [70] Hughes JP, Jones EW. Case of the month: inside a thin man—an 59(3):225–7. engrossing problem. Br J Radiol 1995;68(806):213–4. [56] Pouchot J, Si-Hassen C, Damade R, et al. Cauda equina compression [71] Supik LF, Broom MJ. Epidural lipoma causing a myelographic block by epidural lipomatosis in obesity. Effectiveness of weight reduction. in a patient who had sciatica and lumbosacral spondylolisthesis. A J Rheumatol 1995;22(9):1771–5. case report. J Bone Joint Surg Am 1991;73(7):1104–7. [57] Grayling M, Jardine DL, McClintock AD, Spar J, Wilton GN. Symp- [72] Draganescu M, Daka M, Manon-Espaillat R. Epidural lipomatosis. tomatic epidural lipomatosis following cyproterone acetate therapy. Neurology 2002;58(9):1387. [73] Ashkan K, Moore AJ. Spinal cord compression caused by an extradural Aust N Z J Surg 2000;70(3):233–5. lipoma in Klippel-Trenaunay-Weber syndrome. Case illustration. [58] Boutsen Y, Donckier J. Epidural lipomatosis. Postgrad Med J 2000; J Neurosurg 2002;97(2 suppl):269. 76(891):60–1. [74] Schizas C, Ballesteros C, Roy P. Cauda equina compression after [59] van Rooij WJ, Borstlap AC, Canta LR, Tijssen CC. Lumbar epidural trauma: an unusual presentation of spinal epidural lipoma. Spine lipomatosis causing neurogenic claudication in two obese patients. 2003;28(8):E148–51. Clin Neurol Neurosurg 1994;96(2):181–4. [75] Dumont-Fischer D, Rat AC, Saidenberg-Kermanac’h N, et al. Spinal [60] Borstlap AC, van Rooij WJ, Sluzewski M, Leyten AC, Beute G. epidural lipomatosis revealing endogenous Cushing’s syndrome. Joint Reversibility of lumbar epidural lipomatosis in obese patients after Bone Spine 2002;69(2):222–5. weight-reduction diet. Neuroradiology 1995;37(8):670–3. [76] Fujisawa H, Hasegawa M, Tachibana O, Yamashita J. Spinal epidural [61] Millwater CJ, Jacobson I, Howard GC. Idiopathic epidural lipomatosis lipomatosis associated with pituitary macroprolactinoma. Acta Neu- as a cause of pain and neurological symptoms attributed initially to rochir (Wien) 2002;144(2):213–4. radiation damage. Clin Oncol (R Coll Radiol) 1992;4(5):333–4. [77] Kawai M, Udaka F, Nishioka K, et al. A case of idiopathic spinal [62] Sundram SR. CT of epidural lipomatosis: a case report. Radiogr epidural lipomatosis presented with radicular pain caused by compres- Today 1990;56(632):26–8. sion with enlarged veins surrounding nerve roots. Acta Neurol Scand [63] Iplikcioglu AC, Berkman MZ, Sengoz A. Idiopathic spinal epidural 2002;105(4):322–5. lipomatosis. Acta Neurochir (Wien) 1998;140(4):405–6. [78] Payer M, Van Schaeybroeck P, Reverdin A, May D. Idiopathic symp- [64] Citow JS, Kranzler L. Thoracic epidural lipomatosis with associated tomatic epidural lipomatosis of the lumbar spine. Acta Neurochir syrinx: case report. Surg Neurol 2000;53(6):589–91. (Wien) 2003;145(4):315–21.

A REVIEW PAPER

Surgical Evaluation and Management of Symptomatic Lumbosacral Meningeal Cysts Guy R . Fogel, MD, Paul Y. Cunningham III, MD, and Stephen I . Esses, M D

cysts are dural diverticula that are of congenital or Abstract acquired onset and that may develop a pedicle with an ostium that works in an all-valve fashion to collec t Sacral meningeal cysts are a fairly common finding i n the workup of sciatica . In most instances, a cyst cerebrospinal fluid (CSF) . Normal fluctuations i n causes no symptoms . Occasionally, a symptomati c CSF pressure may account for growth of the cyst and sacral cyst may present with chronic low back pai n erosion of adjacent bone surfaces . (radiculopathy), sensory loss in sacral dermatomes , Type I extradural meningeal cysts are dural diver- perinea) pain, or bowel or bladder dysfunction . Com- ticula that may arise anywhere along the thecal sac . pared with computed tomography, magnetic reso- Type I sacral cysts often have a pedicle at the caudal nance imaging shows meningeal cysts more ofte n tip that connects to the thecal sac adjacent to dorsa l and allows better localization of sacral cysts. nerve roots . In this article, we present clinical guidelines tha t Type II cysts are dilations of the nerve root sleeve . may be used to distinguish symptomatic cysts fro m The dura may have tears, which can be microscopic t o asymptomatic cysts . We conclude that surgica l fairly large. The wall of type II cysts contains nerve roo t treatment of a symptomatic cyst may includ e laminectomy with fenestration and imbrication of the fibers. Bone erosion (eg, canal-widening pedicular ero- cystor percutaneous treatment methods . Surgery sion, foraminal enlargement, scalloping of vertebral bod - for sacral meningeal cysts can lead to successfu l ies or sacrum) is usual in type II cysts. Figure 1 shows improvement of pain and function in activities of dail y types of extradural cysts that may be surgically treated. living in more than 80% of cases . Extradural Meningeal Cysts acral cysts are a fairly common finding in th e workup of sciatica . On myelography for investi- Sagittal View Corona) View S gation of leg and back pain, the incidence of sacral cysts is 17%. On magnetic resonance imaging (MRI) of 500 asymptomatic patients, the incidence i s 4.6%. ) These cysts are usually considered incidental findings on MRI investigations, but occasionally the y Large pedicle may cause sciatica and other symptoms such as bowe l and bladder problems and perinea) pain . In this arti- Large rent at cle, we describe the clinical presentation, the MRI and base of cyst computed tomography (CT) findings, and the surgical treatment options for sacral meningeal cysts .

Types of SRacra! Cysts In 1988, Nabors and colleagues2 classified sacral cysts into 3 types—extradural cysts (type I) ; extradural cysts with nerve roots included within the cyst , including the Tarlov perineural cyst (type II) ; and SmaPt rent Narrow pedicle intradural cysts (type III) . Most sacral meningeal or pedicl e

Dr. Fogel is Spine Fellow, Dr. Cunningham is Medical Researc h Figure 1 . Variation of connection of extradural cyst to the dura . Fellow, and Dr. Esses is Professor of Orthopedics and Brodsk y A narrow pedicle can be ligated if there is no nerve root involve- Chair of Spinal Surgery, Baylor College of Medicine, Houston , ment in the cyst wall of the pedicle . The large and small rents Texas . may be oversewn to seal the cyst from the dura . 278 JUNE 2004 G . R . Fogel et al

TABLE . LITERATURE REVIEW FOR MENINGEAL CYSTS Measure Extradural Type I Extradural Type II Intradural Type II I

No. patients 67 20 7 Cyst location Throughout spine Lumbosacral Throughout spin e Cyst size 2-5 cm Unknown Unknow n Surgery Laminectomy cyst excision Cyst wall excision, Cyst excision oversewing

Identified from the 1960-2001 English-language literature . 1 -3,6-23

Type III cysts are intradural and may occur any - localization of sacral cysts . MRI determine s where along the posterior subarachnoid space . I n whether cysts are filled with fluid and thereb y compression, symptomatic type III cysts act as an y excludes solid tumors . When the signal intensity of intradural masses may . The Table lists the meningea l the mass is the same as that of the theca] sac, a diag- cyst cases found in the literature and classifies the m nosis of CSF-containing structure may be confi- by type . dently made . MRI shows a higher intensity on T z weighting compared with CSF, probably related t o Clinical Presentatio n increased protein and solute content and absence o f Symptomatic sacral cysts may present in various forms . CSF motion effects . Tsuchiya and colleagues 4 state d Typically, a patient complains of low back pain for sev - that MRI with myelography may be especially sen- eral years . The patient has minimal neurologic deficits , sitive with postoperative patients because myelo- absent deep tendon reflexes, and bowel and bladde r graphic material may highlight cysts despite th e abnormalities, including constipation, incontinence, o r bony changes related to surgery or the scallopin g recurring urinary tract infections . Usually, the lowe r associated with the cyst itself . Figure 2 show s extremity sensation is spared . Radicular pain often i s meningeal cyst changes in the lumbosacral are a relieved or disappears when the patient is recumbent , on MRI . and it is aggravated often by the Valsalva maneuver. Routine radiographs of the sacrum and lumbar Perineal pain may be present in approximately 50% o f spine may show bone erosions, but this result i s patients ., Sacral meningeal cysts have been associate d uncommon . CT myelogram is important in deter - with childbirth, papilledema, sacral fracture, neurofibro - mining whether a cyst communicates with th e matosis, and dysraphism of the caudal spine . subarachnoid space. If intrathecal dye does no t fill the cyst even with some delay, then simpl e MRI and CT Finding s oversewing of the posterior cyst wall is all that i s MRI is the best single test . Compared with CT, MR I required . If the cyst communicates with the sub- shows meningeal cysts more often and allows bette r arachnoid space, the pedicle or communicatio n must be found and ligated to prevent recurrenc e of the cyst . CT myelogram may show the presence or absence of the free communication of th e cyst with the spinal subarachnoid space . Delaye d myelographic images are important and may sho w an extradural arachnoid cyst that fills slowl y through a small pedicle connecting the cyst wit h the thecal sac . Myelography may produce a false - negative result if the pedicle of the cyst is to o narrow to allow entry of intrathecal contras t material . Another problem is that CT seldo m shows sacral roots, except S1, because CT is limited to pedicle-to-pedicle scanning in most cases , and lumbar CT seldom is ordered below L5-S1 . CT of the lumbar and sacrum, sacral tomography , and epidurography and intraoperative ultrasoun d Figure 2. Sagittal magnetic resonance imaging views of 2 lurn- bosacral meningeal cysts . The L4-L5 cyst, which extends int o may be helpful but remain unproved in any larg e the left foramina, was causing radicular pain in the left leg . number of cases .

THE AMERICAN JOURNAL OF ORTHOPEDICS ® 279

Symptomatic Lumbosacral Meningeal Cyst s

Authors Clinical Guidelines symptoms should correspond anatomically with th e Clinical guidelines for distinguishing a symptomatic cyst; they should correspond to the same side as th e cyst from an asymptomatic cyst are useful. cyst and to symptoms affecting the same level o f First, it is important to rule out any other possible nerve roots as the cyst location suggests . causes of back and leg pain . It is absolutely neces- Third, it is useful to see whether pulsatile change s sary not to ascribe complaints of back and leg pain t o in CSF pressure increase symptoms . Postural a cyst unless it is reasonably certain that the cyst is changes in CSF pressure may affect complaints . causing the pain . Unfortunately, this is not alway s Radicular pain often is relieved or disappears when possible, and there is not a reliable test for determin - the patient is lying down . Increased pain with Val- ing the clinical relevance of a sacral cyst . It is impor- salva maneuver, cough, or sneeze tends toward a tant to rule out disc herniation, spinal stenosis, an d diagnosis of symptomatic cyst. spondylolisthesis as causing the patients pain . Fourth, aspiration of the meningeal cyst may b e Second, it is important to determine that the cys t diagnostic of a symptomatic cyst if the symptoms corresponds anatomically to the patients complaints. decrease at least temporarily. This may be done with A symptom-producing sacral cyst should cause CT or fluoroscopic guidance. Aspiration may be sacral radiculopathic symptoms, with radiculopathy repeated if initially successful and symptoms recur. to the buttock and S1 distribution in the leg associ- If all 4 criteria are met, then it is reasonable to ated perhaps with bowel or bladder symptoms . The suggest that a cyst is symptomatic (Figure 3) .

Evaluation and Treatment Algorithm : Symptomatic Sacral Cys t

Patient with a sacral cyst an d Consider Diagnose s Chronic low back pai n Low back sprai n Normal lower extremity sensatio n Osteoarthriti s Absent deep tendon reflexe s Herniated dis c Bowel and bladder abnormalitie s Spinal stenosi s Spondylolisthesis YES Instabilit y Does the cyst cause back and leg pain ? NOT the cy s Rule out non-sacral etiology of pai n MRI, CT-myelograph y Electromyograph y Activity modification Postural changes : recumbent and Valsalva according to severity of symptoms YES Pharmacologic pain contro l NO Pain Relief ? Bracing, modalities, exercise s Diagostic aspiration relieves pain Epidural steroid injections YES Repeat aspiratio n Percutaneous drainage of the cyst Closure of the cyst with fibrin glu e Failure of treatment may require additional testin g Not Successfu l and treatment consistent with greater level o f Surgical treatmentConsideration s severit y 1. Delay of 2-6 months to allow conservative treatmen t 2. Documented failure to respond to treatment ; physical signs an d radiographic findings of a surgically treatable lesio n 3. Goal is correction of pathological condition, to attai n functional recovery .

Surgical treatmen t Decompression laminectomy wit h Oversewing of cyst neck and partial or total excision of the cys t MRI = magnetic resonance imaging Shunting of cerebrospinal fluid from the cyst to the subarachnoid spac e CT = computed tomography

Figure 3 . Evaluation and treatment algorithm for symptomatic sacral cysts .

280 JUNE 2004 G. R . Fogel et a l

Figure 4 . Clinical example of a lumbosacral meningeal cys t treated surgically . The patient had symptomatic radiculopath y and no other discernable cause for the leg pain other than th e cyst . (A) Multiloculated extradural cyst is exposed with a wid e laminectomy. (B) The base of the cyst is oversewn, and th e cyst is decompressed with syringe and needle . (C) The base i s oversewn, and the cyst is collapsed and is not refilling .

suggested that fibrin glue may be definitive therapy fo r sacral meningeal cysts . They extended their method for sealing dural tears to a CT-guided percutaneou s procedure for delivering fibrin glue to sacral cysts. Aseptic arachnoiditis has been a worrisome complication of this percutaneous glue technique . The most popular treatment may be a combination Surgical Treatment of laminectomy, fenestration, and drainage of the cyst ; Surgical management usually includes extensive bon y blockage repair of the communication with the arach- decompression with laminectomy alone or combined noid space; and suture use, perhaps supplemente d with either partial resection and oversewing of the cys t with as-needed placement of fibrin glue to sea l or total cyst excision, which may include sacrifice o f drainage . Figure 4 (A—C) shows a symptomati c the involved sacrococcygeal nerve roots . Additional extradural cyst that was ligated at its base and aspi- treatments may include cyst drainage (percutaneou s rated. The surgical outcome was good, but the patient aspiration or external drain placement) ; incision , reported having mild radiculopathy-type pain without drainage, and plication of the cyst wall ; CSF shunting weakness intermittently at the 6-month interval ; at th e to the peritoneum or the subarachnoid space ; and clo- time of his report, this pain was improving . sure of the connection between the cyst and the theca l sac. Type I meningeal cysts are treated by closing th e Summary pedicle between the cyst and the subarachnoid space . Sacral meningeal cysts have fascinated spine surgeon s With type II meningeal cysts, because there is no pedi- for decades . In most instances, these cysts are asympto - cle to block off, the aim of surgery is to obliterate the matic. It is necessary to relate complaints of back an d cyst by partial resection and oversewing of the cys t leg pain to a cyst only when it is reasonably certain tha t wall . One must move or protect the nerve roots if they the cyst is causing the pain . Unfortunately, althoug h are within the wall of the cyst, as in the Tarlov variety MRI and CT can be used to determine the presence of a of type II cysts . Type III intradural cysts should be meningeal cyst, there is no reliable test for determinin g excised by marsupialization, opening to the surround- the clinical relevance of this cyst . Clinical guidelines are ing intradural fluid . These cysts are likely to recur . used to distinguish a symptomatic cyst from an asymp- Recently, cyst drainage, either percutaneous or open , tomatic cyst. Treatment options for a symptomati c has been recommended to decrease symptoms, bu t sacral meningeal cyst range from aspiration and shunt- cysts treated this way are likely to recur . , In 2001 , ing of the cyst to laminectomy and closure of the com- Morio and colleagues o reported on a case successfull y munication of the cyst with the thecal sac . Percutaneous managed with cyst subarachnoid shunting (as a varia- treatment with fibrin glue is interesting but may have tion). In a preliminary report, Patel and colleagues ? residual aseptic arachnoiditis as a complication .

THE AMERICAN JOURNAI, OF ORTHOPEDICS ® 281

Symptomatic Lumbosacral Meningeal Cysts Editoria l

References (continued from page 272) 1. Paulsen RD, Call GA . Murtagh FR . Prevalence and percutaneou s this is what all patients perceive and dread. It seemed drainage of cysts of the sacral nerve root sheath (Tarlov cysts) . AJNR Am J Neuroradiol. 1994;15 :293–297 : discussion 298–299 . so long and I never really slept . Each time a medica- 2. Nabors MW, Pait TG, Byrd EB, et al . Updated assessment and curren t tion was due or a nursing procedure needed, th e classification of spinal meningeal cysts . J Neurosurg . 1988;68 :366-377 . nurse would say that she'd be there as soon as sh e 3. Van de Kelft E, Van Vyve M . Chronic perineal pain related to sacra l meningeal cysts . Neurosurgery. 1991 :29 :223–226 . could, but it seemed forever before she returned with 4. Tsuchiya K . Katase S . Hachiya J . MR myelography of sacral meningea l the medication or whatever. There was something cysts . Arta Radio/ . 1999 :40 :95–99 . 5. Bartels RH, van Overbeeke JJ . Lumbar cerebrospinal fluid drainage for about inefficiency in the pharmacy ; that the nurses symptomatic sacral nerve root cysts : an adjuvant diagnostic procedur e couldn't just get the ordered medications themselve s and/or alternative treatment? technical case report . Neurasurgen . 1997 ;40 :861–864 ; discussion 864–865 . when they are due . I loved the old days when I was a 6. Morio Y. Nanjo Y. Nagashima H . et al . Sacral cyst managed with cyst- resident . The nurse and I would round with a car t subarachnoid shunt : a technical case report . Spine . 21)0126 :451-453 . with bottles of pills and the charts . If the patient 7. Patel MR . Louie W, Rachlin J . Percutaneous fibrin glue therapy of meningeal cysts of the sacral spine . AJR Am J Roentgenal . wanted or needed something, it was dispensed an d 1997 :168 :367–370 . charted on the spot. Efficient patient care, but "some - 8. Carrillo R . Lumbar cerebrospinal fluid drainage for symptomatic sacral nerve root cysts : an adjuvant diagnostic procedure and/or alternativ e one's" concern about potential drug abuse by doctor s treatment? technical case report. Neurostugerv . 1998 :42 :952–953 . and nurses, the possibility of dosage mistakes, and , 9. Chavda SV. Davies AM, Cassar-Pullicino VN. Enterogenous cysts of th e l central nervous system : a report of eight cases . Clin Radio/ . obviously more important, the failure to properly bil 1985 ;36:245–251 . the patient, have led to the computerized infallible , 10. Cilluffo JM, Gomez MR . Reese DF, et al . Idiopathic ("congenital" ) but possibly less efficient system . spinal arachnoid diverticula: clinical diagnosis and surgical results . Mayo Clin Prot.. 1981 ;56 :93–101 . Finally, the day came, and, eager to take the firs t 11. Cole GP. Flannery AM, Gulati AK . Intrasacral meningocele : case report train out of Dodge . I thanked my surgeon, fel t and review of the literature . Spine . 1989 :14 :1418–1420 . 12. Crellin RQ . Jones ER . Sacral extradural cysts : a rare cause of low back- immensely better than before the surgery, an d ache and sciatica . J Bone Joint Surg Br. 1973 :55 :20–31 . received my discharge diploma . I don't know whether 13. Davis DH, Wilkinson JT, Teaford AK, et al . Sciatica produced by a sacral I learned anything new, but I certainly experienced th e perineural cyst . Tex Med . 1987:83 :55–56 . 14. Ehara S, Rosenberg AE, el-Khoury GY. Sacral cysts with exophytic com - ebb and flow of emotions that many of my more per- ponents : a report of two cases . Skeletal Radio/ . 1990;19 :117–119 . ceptive patients feel . Certainly the inpatient medical 15. Evans JA . Lougheed LM . Intradural cysts of the cervical spine : report o f three cases . .1 Bone Joint Surg Am . 1978 ;60 :123–125 . care has improved over the years . The respirator y 16. Fardon DF. Intrasacral meningocele complicated by transverse fracture : a therapist and his instructions and equipment hav e case report. J Bone Joint Surg Am . 1980 ;62 :839–841 . improved . The OTs and PTs were there and efficiently 17. Freidberg SR . Fellows T . Thomas CB, et al . Experience with symptomatic spinal epidural cysts . Neurosurgery. 1994 ;34:989–993 ; discussio n rendered their instructions and advice . The preopera- 993 . tive, operative, and recovery room processes are pro- 18. Gortvai P. el-Gindi S . Spinal extradural cyst: case report. J Neurosurg. 1967 ;26 :432--435 . fessionally rendered and appropriate, although the 19. Kim CH . Bak KH. Kim JM. et al . Symptomatic sacral extradural arach- recent well-intended JCAHO requirements result in a noid cyst associated with lumbar intradural arachnoid cyst . Clin Neural Neurosurg . 1999 ;101 :148–152. three-stooges comedy routine, with everyone indis- 20. Mummaneni PV, Pitts LH, McCormack BM . et al . Microsurgical treat- criminately going through the same questions ove r ment of symptomatic sacral Tarlov cysts . Neurosurgery. 2000;47 :74–78 ; and over again . Possible nursing shortages and per- discussion 78–79 . 21. Raftopoulos C . Pierard GE. Retif C, et al . Endoscopic cure of a giant haps inpatient pharmacy issues may exist, but percep- sacral meningocele associated with Marfan's syndrome : case report . tion and reality may not coincide . Nonetheless, I no w Neurosurgery- . 1992;30 :765–768 . 22. Su CC, Shirane R . Okubo T, et al . Surgical treatment of a sacral nerv e know what my patients can expect and can respon d root cyst with intermittent claudication in an 85-year-old patient : case with some level of understanding . My fellow sur- report . Surg Neuro/ . 1996 ;45 :283–286 . e 23. Voyadzis JM. Bhargava P. Henderson FC . Tarlov cysts : a study of 1 0 geons, if you have to have surgery, you will appreciat cases with review of the literature . J Neurosurg . 2001 ;95 :25–32 . the "education"

282 JUNE 2004 The Spine Journal 5 (2005) 202–211

Fig. 6. Third patient: axial views of T5–T6 show displacement of the dural sac cerebrospinal ﬂuid without extradural compression from epidural fat.

Surgical treatment of dysphagia after anterior cervical interbody fusion Guy R. Fogel, MD*, Mark F. McDonnell, MD Houston Spine Surgery, 5225 Katy Freeway, Suite 600, Houston, TX 77007, USA Received 22 September 2003; accepted 30 June 2004

Abstract BACKGROUND CONTEXT: Dysphagia is a frequent complication after anterior cervical interbody fusion (ACIF). Although dysphagia usually improves over 6 months, it remains a significant and persistent problem for some patients. The etiology is poorly understood but has been reported to be associated with vocal cord paralysis, dislodgement of instrumentation and unidentified causes, such as hematoma, adhesion formation and denervation of the pharyngeal plexus. A surgical treatment of dysphagia after ACIF has not been reported. PURPOSE: We report the surgical treatment of persistent dysphagia occurring after ACIF with instrumentation. STUDY DESIGN/SETTING: A retrospective review of cervical discectomy and interbody fusion patients identified a subset of patients with symptomatic dysphagia who chose surgical treatment of the dysphagia. The hypothesis is that removal of the anterior cervical plate will release mechanical adhesions of the esophagus to the anterior spine around the plate. Outcome was graded by examination and a final telephonic interview with a dysphagia questionnaire. METHODS: Thirty-one patients who elected surgical treatment for persistent dysphagia were assessed at clinic visits after surgery at 3, 6 and 12 months for symptomatic dysphagia, and with a final telephonic questionnaire. The average time from initial surgery to time of surgical treatment for dysphagia was 18 months. Final follow-up was an average 11 months (range, 6 to 25 months) with a dysphagia questionnaire using the Bazaz-Yoo dysphagia score. Thirty-one patients responded to a phone questionnaire with the Bazaz-Yoo dysphagia score. RESULTS: The primary operative finding was extensive adhesions attaching the esophagus to the prevertebral fascia and anterior cervical spine around the periphery of the cervical plate. Seventeen patients (55%) were significantly improved to no dysphagia of solids and liquids (pՅ.0001). Ten patients (32%) reported mild dysphagia occasionally with specific foods. Three patients had persistent moderate occasional dysphagia with solid food. Two patients had persistent severe dysphagia of solids and liquids. Previous cervical surgery, particularly with pre-existing dysphagia, and unexpectedly extreme amounts of adhesions at surgery were contributing factors to the cases with persistent severe dysphagia. CONCLUSIONS: Surgical treatment of dysphagia after ACIF has not been reported. Removal of the cervical instrumentation in patients will improve the dysphagia. This improvement with surgical management, as compared with the dissatisfaction before surgical treatment, documents that this surgical treatment is a reasonable option. Ć 2005 Elsevier Inc. All rights reserved. Keywords: Dysphagia; Complications, cervical; Cervical plate

Introduction

FDA device/drug status: approved for this indication (DePuy cervical Anterior cervical discectomy and interbody fusion is plate). frequently performed for disc herniation and spondylosis. Nothing of value received from a commercial entity related to this Although the procedure is associated with low morbidity, research. dysphagia is a common postoperative complaint [1–17]. * Corresponding author. South Texas Orthopaedics and Spinal Surgery Associates, 9150 Huebner Road, #350, San Antonio, TX 78240, USA. Tel.: Dysphagia is deﬁned as difﬁculty with swallowing both (210) 561-7234; fax: (210) 561-7240. solids and liquids and includes the inability to protect the E-mail address: [email protected] (G.R. Fogel) airway from aspiration. Patients with dysphagia report

1529-9430/05/$ – see front matter Ć 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.spinee.2004.06.022 G. R. Fogel and M. F. McDonnell / The Spine Journal 5 (2005) 140–144 141 difficulty with solid and liquid food. Some report painful Methods swallowing (odynophagia), difficulty with saliva, food stick- Forty-four patients were identified as undergoing plate ing in the throat and coughing with eating. Common symptoms are listed in Table 1. removal and lysis of adhesions for the indication of dysphagia in the period January 1999 to December 2001. Attempts Postoperative dysphagia usually improves with time. by mail and phone were made to contact all patients, and However, there are some patients with persistent symptoms 31 of the 44 responded to a final telephone questionnaire that do not improve spontaneously. Winslow et al. [5] in using the Bazaz-Yoo dysphagia score giving a last follow- a questionnaire study of noninstrumented anterior cervical up an average 11 months (range, 5 to 25 months) after interbody fusion (ACIF) patients found early dysphagia in instrumentation removal in 31 patients. In the 31 patients 60% and long-lasting dysphagia in 23% [5]. In a prospective with complete follow-up, the average time to surgical treat- consecutive series of both instrumented and noninstru- ment for dysphagia was 18 months. mented ACIF, Bazaz et al. [2] described a significant rate The 31 patients underwent anterior exploration of the of dysphagia that decreased over time and stabilized at a fusion, with removal of the anterior cervical instrumentation 12.5% rate at 12 months. Yue et al. [14] evaluated 74 patients and lysis of adhesions between the esophagus and the ante- at an average 7.2 years after instrumented ACIF and found a rior cervical spine. The right-side anterior approach was 35% incidence of persistent dysphagia with 17.6% moderate made excising the previous incisional scar. The anterior and severe symptoms. instrumentation was the PEAK plate (DePuy, Warsaw, IN), in Early in our experience, several patients were referred all cases. The surgeon used self-retraction systems (Trimline; for otolaryngology evaluation for persistent dysphagia after Sofamor Danek, Memphis, TN). Anesthesia included the instrumented anterior cervical interbody fusion. The patients use of an esophageal stethoscope, and the endotracheal cuff returned without meaningful abnormalities detected on was kept at low pressure throughout the procedure. Iliac barium swallow, laryngeal endoscopy or esophageal manom- crest autograft and additional PEAK plate instrumentation etry. The otolaryngology specialists suggested no remedy was used in revision of nonunion or treatment of adjacent for the dysphagia. In the same time period, the senior author segment deterioration cases. Postoperative immobilization had noted a clinical improvement in dysphagia occurring consisted of a hard collar in revision cases, and no immobili- after cervical instrumentation removal in two patients with zation was required otherwise. severe preoperative symptoms of dysphagia. The hypothesis Patients were evaluated at clinic visit by the author is that removal of the anterior cervical plate will release (MFM) 1, 6 and 12 months after the procedure. Questions mechanical adhesions of the esophagus to the anterior about dysphagia have been standard intake questions at spine around the plate. follow-up in the practice. We use the dysphagia score pub- With clinical success in two patients, cervical instrumen- lished by Bazaz et al. [2]. These authors published a vali- tation removal was offered to patients with significant persis- dated grading system for postoperative dysphagia (Table 2). tent dysphagia after ACIF. Our series of anterior cervical The patients are graded as having none, mild, moderate or discectomy and interbody fusion (ACIF) for treatment of disc severe dysphagia. Patients with no episodes of swallowing herniation and spondylosis was reviewed. Patients were difficulty were graded as “none.” Patients who experienced identified who chose to have exploration of the anterior rare episodes of dysphagia were graded “mild.” These pa- fusion with removal of anterior cervical instrumentation. We tients thought their dysphagia was insignificant. “Moderate” report the surgical treatment results in this group of ACIF dysphagia was defined as occasional swallowing difficulty patients with persistent dysphagia. with specific foods (ie, steak or bread). “Severe” dysphagia was defined as frequent difficulty swallowing the majority of foods. Using the dysphagia score retrospectively, the patients Table 1 were graded as having none, mild, moderate or severe Symptoms of dysphagia dysphagia of liquids and solids based on follow-up visit More obvious symptoms clinical evaluation and a telephonic questionnaire in those Pain with swallowing patients unavailable for a final clinical visit. Records were Difficulty swallowing Heartburn Coughing or choking with swallowing Regurgitation of old foodstuffs Table 2 Nasal regurgitation Dysphagia score symptoms Feeling of blockage Weight loss Severity Liquid Solid Less obvious symptoms 0.None None None Food avoidance 1.Mild None Rare Frequent throat clearing 2.Moderate None or rare Occasionally with specific food Changes in breathing after swallowing 3.Severe None or rare Frequent (majority of solids) Wet voice quality From Bazaz et al. [2]. 142 G. R. Fogel and M. F. McDonnell / The Spine Journal 5 (2005) 140–144 reviewed for patient’s age, gender, surgical procedure, tobacco use and number and location of surgical levels addressed. The average age was 43 years with a range of 25 to 66 years. Twenty-eight (64%) were men and 18 (41%) women. The number of levels in the initial anterior cervical interbody fusion in the 31 patients ranged from one to four levels with a mean of three levels. Seventy percent of the fusions were at three or four levels. All patients had indication of symptomatic anterior hardware and dysphagia categorized as moderate or severe. Addi- tional indications at the time of surgical treatment of dysphagia were exploration for pseudarthrosis (five; 16%) or adjacent segment deterioration (six; 19%). Three patients had undergone previous surgery, and two of these had pre- existing dysphagia and dysphonia before the index ACIF.

Results Fig. 1. Scar tissue forms with the incision from skin to spine, sweeping over the plate and attaching to the anterior cervical spine along the edges of The Bazaz-Yoo dysphagia score before surgical treatment the plate. Adhesions may bind the trachea and the esophagus to the scar overlying the plate. Removal of the plate will release the adhesions and of dysphagia in the 31 patients was moderate dysphagia in free the esophagus from the spine surface. 15 (48%) or severe dysphagia in 16 (52%). The patients had more severe dysphagia with solids than liquids (Table 3). The Bazaz-Yoo dysphagia score after instrumentation removal screw loosening observed. In three cases, there was some in the 31 patients at last examination showed significant tissue staining beneath the plate, typical of titanium plates improvement in dysphagia scores of liquids and solids and screws. There were no cases with visible metallic wear Յ (p .0001). One (3%) had continued severe dysphagia of debris reported. solids and liquids. Three patients (10%) had persistent mod- Approaching the cervical spine through the previous inci- erate dysphagia. In 17 patients (55%) there was no dysphagia sion was not difficult. However, twice there was unusually at all. Ten (32%) reported mild occasional dysphagia with dense and prolific adhesions found at the time of surgery. In certain solids, such as steak and bread (Table 3). these two cases, the dissection was more difficult. In both At the time of cervical instrumentation removal, the pri- of the patients there was recurrent moderate to severe dys- mary operative finding was extensive adhesions attaching phagia. In one patient, there was a history of three additional the esophagus-trachea midline structures to the prevertebral previous surgeries, with pre-existing dysphonia and dyspha- fascia and anterior cervical spine around the periphery of gia rated as moderate. An otolaryngologist performed preop- the cervical instrumentation and through any open aperture erative evaluation of the vocal cords and the approach at the in the plate. Releasing the adhesions between the esophagus index surgery and reported intense scarring but no damage to and the anterior cervical spine was the key to restoring the vocal folds or visceral structures. At final examination, mobility to the esophagus. Fig. 1 illustrates the plane of there was barium swallow evidence of a pharyngeal divertic- surgical release of adhesions accomplished with removal ulum. His dysphagia was rated severe. His dysphonia was of the cervical plate. The status and condition of the plates mild. In the other patient with extreme scarring at surgery, at the time of surgery was uniformly benign. The plates were there was an extraordinary odynophagia, choking of solids not fractured or corroded. There was no screw failure or and liquids, and difficulty breathing before surgery rated as severe dysphagia and dysphonia. For 2 months after sur- Table 3 gery, the dysphagia was improving; thereafter the symptoms Dysphagia of liquids and solids before surgery and last examination recurred and worsened. At final examination there was Liquid Solid severe dysphagia and a vocal fold paralysis that responded finally to a gelfoam implant. Complications of the surgical Preoperative/ Preoperative/ Severity last score (%) p Value last score (%) p Value treatment of dysphagia were continued symptoms of dysphagia as severe in 2 (5%), moderate in 3 (7%) and mild in 16 0.None 0/81 Ͻ.0001 0/55 Ͻ.0001 1.Mild 26/16 Ͻ.0001 0/32 Ͻ.0001 (36%). Two were initially much improved in their dysphagia, 2.Moderate 65/3 Ͻ.0001 48/10 Ͻ.0001 but symptoms recurred 6 to 12 weeks after surgical treat- 3.Severe 10/0 .0377 52/3 Ͻ.0001 ment. One patient had a persistent vocal fold paralysis and Median 2/0 3/0 3 more had dysphonia without known vocal fold paralysis. SD 1/0 1/1 One developed a symptomatic pharyngeal diverticulum G. R. Fogel and M. F. McDonnell / The Spine Journal 5 (2005) 140–144 143

Table 4 Complications Levels Complications LFU Symptoms C3–C7 Dysphonia Dysphonia mild with singing C4–C6 None Snoring increased C3–C7 Dysphagia and dysphonia before surgery Mild occasional choking on large solids, mild dysphonia, continued neck pain, diabetic C3–C6 None Continued dysphagia improved 1 grade C3–C6 Most severe scarring, 1 inch thick Severe residual dysphagia, pharyngeal diverticulum, severe neck pain and headaches C3–C6 None Initially good but recurrent dysphagia C3–C7 Dysphonia Continued dysphonia C4–C6 Horner syndrome for 3 months Horner syndrome resolved C3–C6 Most intense scarring, VFP VFP with implant; dysphagia improved 2 months, then worsened; normal barium swallow C4–T1 Dysphagia and dysphonia before surgery Good relief of dysphagia, mild dysphonia LFUϭlast follow-up; VFPϭvocal fold paralysis. associated with severe dysphagia. One Horner syndrome with vocal cord paralysis [1,15,16,24] or prevertebral edema resolved at 3 months after surgery (Table 4). [10]. Frempong-Boadu et al. [1] reported 48% of ACIF patients had preoperative evidence of radiographic swallowing abnormalities, and Doran et al. [25] found 50% of Discussion patients with prior cervical surgery had postoperative barium swallow abnormalities. Revision surgery with additional The literature on the incidence of dysphagia is largely scarring, the risks of reoperation through a previous inci- retrospective. The incidence varies widely from the litera- sional scar and previous dysphagia and dysphonia may in- ture, with reports varying from 2% to 60% [2,5,15,18,19]. crease the risk of recurrent and persistent dysphagia. From reports in the literature, persistent dysphagia occurs Increased prominence of the instrumentation, graft or screws whether or not the anterior cervical spine is instrumented could cause esophageal irritation, even erosion and death [1,2,10,17]. The incidence in the instrumented versus nonin- [26]. All current anterior cervical plating systems are de- strumented cases in the literature may be slightly higher signed with minimal profile and the screw lock to the plate by in the instrumented group. Albert et al. [7] in the multicentric various methods to prevent screw back out, which could cervical plate study sponsored by the Cervical Spine Re- cause esophageal irritation and dysphagia [27]. The PEAK search Society are currently reporting incidence of 7.9% Plate system of Depuy-Acromed used in this study is a dysphagia in instrumented ACIF patients at 24 months and 5.3% in nonplated ACIF patients. Edwards et al. [8] com- contemporary system with a profile height of 3.2 mm and pared the outpatient records of four spine surgeons from two a width of 18.1 mm. Dislodgement of the graft or instrumen- different academic medical centers with the results of a brief tation is a known cause of dysphagia [18,28]. No case of survey mailed to patients. Dysphagia was recorded as present instrumentation dislodgement or bone graft displacement in surgeon records 25 times and 104 times by patient survey. was identified in this series. Anterior instrumentation, screw Edwards et al. concluded that dysphagia was underreported heads and grafts should be kept flush with the surface of by 76% in surgeon records with very poor correlation to the the spine to minimize local esophageal irritation. patients’ responses in the survey. Intense adhesions were found in each of our cases at Bazaz et al. [2] found a 50% incidence of dysphagia at instrumentation removal surgery. These adhesions seem to 1 month after anterior cervical surgery that improved to flow smoothly around the surfaces of the plate along all 12.6% at 12 months. Yue et al. [14] invited 74 patients edges, without actually attaching to it. Adhesions extended to return for examination by two independent reviewers at an from the esophagus through the holes in the plate down to average 7.2 years (range, 5.4 to 11.1 years). Persistent the bone surface. The adhesions attach the esophagus to the dysphagia of mild to severe was present in 36% and moderate anterior vertebral surfaces around the borders of the plate. or severe dysphagia was found in 17.6% [14]. The persistent While removing the plate, adhesions attaching the esophagus dysphagia rate in the current study is very similar to that de- to the vertebral surface adjacent to the plate are removed. termined by Yue et al. The dysphagia incidence in the current The esophagus is freed from the cervical surface below. study may appear higher than expected, but it is clear from With the initial ACIF, the magnitude of the adhesions is the Yue et al. and the Edwards et al. reports that the true increased by the presence of raw and oozing bone exposed incidence of persistent dysphagia is higher than that during the subperiosteal dissection, drilling and burring nec- previously stated in the literature and is often underreported. essary for the adequate preparation of the site. The adhesion Risk factors for dysphagia, such as age, gender, primary formation after plate removal may be less intense because versus revision surgery, usage of cervical plate and number there is no exposed bleeding bone surface to produce a of levels treated, have been previously described [1,5,10– significant hematoma. The patient has a better chance of not 12,20–23]. Dysphagia has been reported to be associated adhering the esophagus to the bone again. 144 G. R. Fogel and M. F. McDonnell / The Spine Journal 5 (2005) 140–144 One shortcoming of this study is the inability to answer evaluation. Paper presented at: Cervical Spine Research Society the question of whether relief of dysphagia with lysis of Annual Meeting, Monterey, CA, November 29, 2001. [7] Albert TJ, Riley L, Skolasky RL. Cervical plate study. Paper presented adhesions in other situations, such as noninstrumented cases at: Cervical Spine Research Society, Scottsdale, AZ, December 11, or instrumented cases treated without plate removal, would 2003. be as effective in treatment of dysphagia. Unfortunately, [8] Edwards CC, Heller JG, Murakami H. Corpectomy versus lami- this study, without a noninstrumented control group, cannot noplasty for multilevel cervical myelopathy: an independent matched- answer that question. In literature review, there is a similar cohort analysis. Spine 2002;27(11):1168–75. [9] Edwards CC, Karpitskaya Y, Cha C, et al. Accurate identification of incidence of dysphagia in the instrumented and noninstru- adverse outcomes after spine surgery. J Bone Joint Surg Am 2004; mented cases. However, it is interesting that dysphagia may 86-A:251–6. be improved in cases with revision for pseudarthrosis or [10] Martin RE, Neary MA, Diamant NE. Dysphagia following anterior extension for adjacent segment deterioration with removal cervical spine surgery. Dysphagia 1997;12(1):2–10. and insertion of a new plate. Another shortcoming of the [11] Stewart M, Johnston RA, Stewart I, Wilson JA. Swallowing performance following anterior cervical spine surgery. Br J Neurosurg present study is the lack of pre- and postoperative barium 1995;9(5):605–9. swallow and otolaryngology evaluations. It is our opinion [12] VanderveldtHS, Young MF. The evaluation of dysphagia after anterior from the literature and our experience that the testing does cervical spine surgery: a case report. Dysphagia 2003;18(4):301–4. not reflect the severity of the dysphagia, and it does not [13] Welsh LW, Welsh JJ, Chinnici JC. Dysphagia due to cervical spine suggest any remedy for the severe symptoms. A last surgery. Ann Otol Rhinol Laryngol 1987;96(1 Pt 1):112–5. [14] Yue WM, Brodner W, Highland TR. Persistent swallowing and voice shortcoming is a proper assessment of the effect of the instru- problems after anterior cervical discectomy and fusion with allograft mentation’s contribution to dysphagia; for example, the andplating:a5-to11-year follow-upstudy.Paperpresentedat: Cervical thickness of the plate and such properties as corrosion and Spine Research Society Annual Meeting, Scottsdale, AZ, December tissue staining may have some effect. This study did not 11, 2003. find any direct contribution of the instrumentation to [15] Morpeth JF, Williams MF. Vocal fold paralysis after anterior cervical diskectomy and fusion. Laryngoscope 2000;110(1):43–6. dysphagia. [16] Netterville JL, Koriwchak MJ, Winkle M, Courey MS, Ossoff RH. Vocal fold paralysis following the anterior approach to the cervical spine. Ann Otol Rhinol Laryngol 1996;105(2):85–91. Conclusions [17] Cauthen JC, Kinard RE, Vogler JB, et al. Outcome analysis of noninstrumented anterior cervical discectomy and interbody fusion in Surgical treatment of dysphagia after ACIF has not been 348 patients. Spine 1998;23(2):188–92. reported. Removal of the anterior cervical instrumentation [18] Bose B. Anterior cervical fusion using Caspar plating: analysis of results and review of the literature. Surg Neurol 1998;49(1):25–31. in patients with residual significant dysphagia will improve [19] Clements DH, O’Leary PF. Anterior cervical discectomy and fusion. the dysphagia. This improvement with surgical management, Spine 1990;15(10):1023–5. as compared with the dissatisfaction before instrumentation [20] Mathers-Schmidt BA, Kurlinski M. Dysphagia evaluation practices: removal surgery because of dysphagia, documents that this inconsistencies in clinical assessment and instrumental examination surgical treatment is a reasonable option. decision-making. Dysphagia 2003;18(2):114–25. [21] Lind CD. Dysphagia: evaluation and treatment. Gastroenterol Clin North Am 2003;32(2):553–75. [22] Leslie P, Carding PN, Wilson JA. Investigation and management of References chronic dysphagia. Br Med J 2003;326(7386):433–6. [23] Langmore SE. Evaluation of oropharyngeal dysphagia: which diag- [1] Frempong-Boadu A, Houten JK, Osborn B, et al. Swallowing and nostic tool is superior? Curr Opin Otolaryngol Head Neck Surg 2003; speech dysfunction in patients undergoing anterior cervical discec- 11(6):485–9. tomy and fusion: a prospective, objective preoperative and postopera- [24] Beutler WJ, Sweeney CA, Connolly PJ. Recurrent laryngeal nerve tive assessment. J Spinal Disord Tech 2002;15(5):362–8. injury with anterior cervical spine surgery risk with laterality of surgi- [2] Bazaz R, Lee MJ, Yoo JU. Incidence of dysphagia after anterior cal approach. Spine 2001;26(12):1337–42. cervical spine surgery: a prospective study. Spine 2002;27(22): [25] Doran SE, Walsh J, Kinkaid A, Cutler D. Prospective analysis of 2453–8. dysphagia following anterior cervical spine fusion. Paper presented [3] Flynn TB. Neurologic complications of anterior cervical interbody at: Cervical Spine Research Society Annual Meeting, Scottsdale, AZ, fusion. Spine 1982;7(6):536–9. December 11, 2003. [4] Riley LH, Jr, Robinson RA, Johnson KA, Walker AE. The results of [26] Gaudinez RF, English GM, Gebhard JS, Brugman JL, Donaldson DH, anterior interbody fusion of the cervical spine. Review of ninety-three Brown CW. Esophageal perforations after anterior cervical surgery. consecutive cases. J Neurosurg 1969;30(2):127–33. J Spinal Disord 2000;13(1):77–84. [5] Winslow CP, Winslow TJ, Wax MK. Dysphonia and dysphagia follow- [27] Fogel GR, Liu W, Reitman CA, Esses SI. Cervical plates: comparison ing the anterior approach to the cervical spine. Arch Otolaryngol of physical characteristics and in vitro pushout strength. Spine J Head Neck Surg 2001;127(1):51–5. 2003;3(2):118–24. [6] Donnelly RE, O’Brien M, Dart D, Lowe TG, Dwyer TF. Dysphagia [28] Brown JA, Havel P, Ebraheim N, Greenblatt SH, Jackson WT. Cervical after multilevel cervical arthrodesis: a clinical and cineradiographic stabilization by plate and bone fusion. Spine 1988;13(3):236–40. 92S Proceedings of the NASS 18th Annual Meeting / The Spine Journal 3 (2003) 67S–171S

DISCUSSION: There are multiple clinical studies correlating a previous disc space, there is a markedly increased risk of moderate to severe ALOD. fusion with adjacent lumbar degeneration. Our study examines long-term Based upon our data, we now strive to keep the plate as far away from survivorship of segments adjacent to lumbosacral fusions. The adjacent the adjacent disc space as possible. segment degeneration in our study is in agreement with previous findings. CONCLUSIONS: To our knowledge, the alarming rates of ALOD follow- The rate and type of repeat surgeries, however, differ from previous data. ing anterior cervical plates, and its relationship to the plate-to-disc distance CONCLUSIONS: Adjacent segment disease appears to be a real concern have not been previously reported. We recommend that anterior cervical following lumbosacral posterior spinal fusion. At an average follow-up of plates be placed at least 5 mm away from the adjacent disc space in order 6.9 years, 26% of patients required reoperation, with 91% of these being to decrease the likelihood of ALOD. A greater initial distance would be proximal lumbar fusions. Although the majority of the adjacent segment necessary when using subsidence plates. disease occurred at the immediately adjacent segment, levels more proximal DISCLOSURES: Device or drug: anterior cervical plates. Status: ap- were affected as well. There was radiographic progression of the adjacent proved. segment disease, but preoperative grading did not correlate with the need CONFLICT OF INTEREST: No conflicts. for further surgery. DISCLOSURES: No disclosures. doi: 10.1016/S1529-9430(03)00234-1 CONFLICT OF INTEREST: No conflicts. doi: 10.1016/S1529-9430(03)00233-X

4:46 4:42 55. Anterior hardware removal improves dysphagia following 54. Adjacent level ossification disease (ALOD) secondary anterior cervical discectomy and fusion to anterior cervical plates Guy Fogel, MD1, Mark McDonnell, MD1; 1Houston Spine Surgery, 1 1 1 K. Daniel Riew, MD , Jong-Beom Park, MD , Yong-Sun Cho, MD ; Houston, TX, USA 1Washington University in St. Louis School of Medicine, St. Louis, MO, USA HYPOTHESIS: Removal of the anterior cervical plate in patients with symptomatic dysphagia after anterior instrumented fusion of multiple cervi- HYPOTHESIS: There are few studies concerning the association between cal levels will improve dysphagia. osteophyte formation and anterior cervical plates. It has been the senior METHODS: Two hundred and ninety-one patients treated with instru- author’s empiric experience that peri-plate ossification of the adjacent seg- mented anterior discectomy and interbody fusion (ACIF) were assessed at ment occurs commonly. Our hypothesis is that the closer the plate to the 3 months, 6 months, and 12 months for symptomatic dysphagia. Eighty- adjacent disc space, the greater the ossification. two patients had residual complaints of dysphagia of solids or liquids at METHODS: One hundred and eigtheen patients, who had no previous 6 months. Forty-nine patients had moderate or severe symptoms. Ninety- cervical surgery and achieved a solid fusion following anterior cervical eight patients had removal of the anterior cervical plate. Forty-six patients arthrodesis with a plate for degenerative cervical conditions, with a mini- had the indication of dysphagia. 26 patients had moderate or severe mum of 1 year follow-up were identified from a database for inclusion in symptoms. The average time follow-up from the index ACIF was 30 months. this study. All of the procedures were performed by the senior author The averagetime to hardwareremoval was 16months. These 46patients were and all of the radiographic analyses were independently performed by 2 assessed at average 13 months after plate removal for dysphagia, with an experienced spine surgeons who were uninvolved in the patients’ care. The extensive outcome questionnaire using the Bazaz dysphagia score(Bazaz, mean age was 51.8 years. The mean duration of post-operative follow-up Lee et al. 2002), visual analog scale, the North American Spine Society was 25.7 months (range: 12–76). Thirty-six patients underwent a 1-level Satisfaction Questionnaire, and a modified Oswestry Disability Index. arthrodesis; 47 a 2-level; 32 a 3-level; and 3 had a 4-level arthrodesis. The RESULTS: The primary operative finding was extensive adhesions at- distance between the tip of the plate and the caudal as well as the rostral taching the esophagus-trachea midline structures to the pre-vertebral fascia adjacent disc was measured (plate-to-disc distance). This distance was used and anterior cervical spine around the periphery of the cervical plate. All Ͻ to divide the patients into two groups: group A, plate-to-disc distance 5 patients reported improvement in the level of dysphagia after hardware Ͼ mm and group B, 5 mm. The severity of ALOD (adjacent level ossification removal. Forty-one patients were improved to only rare dysphagia of solids disease) at the 2 adjacent disc spaces were classified using the following and no problems with liquids. Five patients reported dysphagia occasionally grading system: 0: none; 1 (minimal): extends across Ͻ50% of the disc with specific foods. No patients had persistent frequent dysphagia with space; 2 (moderate): Ͼ50% of the disc space; 3 (severe): complete bridging solid food. of the disc space. Each observer determined the grade twice and the DISCUSSION: Dysphagia is a frequent complication at one month after average of the four measurements was used as the final grade. Eighteen ACIF. Although dysphagia usually improves over 6 months, it remained patients were excluded from the measurement of distal plate-to-disc distance a significant and persistent problem at 6 month follow up in 18% of our due to bony overlapping of the shoulder and cervicothoracic junction. ACIF patients The etiology is poorly understood, but has been reported to The incidence and severity of ALOD were compared between the groups be associated with vocal cord paralysis, dislodgement of hardware, and by the Chi-square test and Mann-Whitney test. unidentified causes such as hematoma, adhesion formation, and denervation RESULTS: The inter-observer variability for the measurement technique of the pharyngeal plexus(Frempong-Boadu, Houten et al. 2002). Intense was 0.96 and the intra-observer variability was 0.98. An alarming 59.5% adhesions were found in each of our cases at surgery. All patients improved (70/118) developed adjacent level ossification at the rostral disc space while with removal of instrumentation and lysis of adhesions. Adhesion formation 29% (29/100) developed it at the caudal disc space. The incidence of rostral is the primary cause of dysphagia after anterior cervical surgery. Surgical ALOD was 67.0% (65/97) in patients with a plate-to-disc distance Ͻ5mm treatment of dysphagia after ACIF has not been reported. Removal of the vs. 23.8% (5/21) for Ͼ35mm(pϽ0.001). The severity grade of rostral cervical hardware in patients with residual significant dysphagia will ALOD was 0.96Ϯ0.88 in patients with a plate-to-disc distance Ͻ5mm vs. improve the dysphagia. 0.24Ϯ0.44 for Ͼ35mm(pϽ0.001). The incidence of caudal ALOD was 45% CONCLUSIONS: Adhesion formation is the primary cause of dysphagia (27/60) in patients with a plate-to-disc distance Ͻ5mm vs. 5% (2/40) for after anterior cervical surgery. Surgical treatment of dysphagia after ACIF Ͼ35mm(pϽ0.001). The severity grade of caudal ALOD was 0.67Ϯ0.90 has not been reported. Removal of the cervical hardware in patients with in patients with a plate-to-disc distance Ͻ5mm and 0.10Ϯ0.44 for Ͼ3 residual significant dysphagia will improve the dysphagia.This improve- 5mm (pϽ0.001). ment with surgical management, as compared with the dissatisfaction DISCUSSION: We believe that the current findings definitively demon- before hardware removal surgery because of the dysphagia, documents that strate that if the plate is placed Ͻ5 mm from either the rostral or caudal this surgical treatment is a reasonable option. Proceedings of the NASS 18th Annual Meeting / The Spine Journal 3 (2003) 67S–171S 93S DISCLOSURES: No disclosures. Friday, October 24, 2003 CONFLICT OF INTEREST: No conflicts. 9:31–10:20 AM doi: 10.1016/S1529-9430(03)00235-3 Outstanding Paper Presentations

4:50 56. Vascular injury during anterior lumbar surgery 9:31 Salvador Brau, MD1, Rick Delamarter, MD2, Michael Schiffman3, Lytton OP1. Transplantation of oligodendrocyte precursors and sonic Williams, MD4, Robert G. Watkins, MD4; 1Sapine Access Surgery hedgehog results in improved function and white matter sparing in Associates, Los Angeles, CA, USA; 2Saint John’s Health Center, Santa the spinal cords of adult rats after contusion 1 1 1 Monica, CA, USA; 3OCCLA, Los Angeles, CA, USA; 4Los Angeles Spine Nicholas C. Bambakidis, MD , Robert H. Miller, PhD ; Departments of Surgery Institute, Los Angeles, CA, USA Neurological Surgery and Neurosciences, Case Western Reserve University, Cleveland, OH, USA HYPOTHESIS: Vascular injury is a known complication of the approach for anterior lumbar surgery. Case reports and reviews of complications BACKGROUND CONTEXT: A substantial cause of neurological disabil- have appeared in the literature, but the true incidence of both arterial ity in spinal cord injury (SCI) is oligodendrocyte death leading to demyelin- and venous injuries is not well established. Our large study tries to more ation and axonal degeneration. Rescuing oligodendrocytes and preserving accurately establish this incidence by evaluating 1315 consecutive cases myelin is expected to result in significant improvement in functional outcome undergoing anterior lumbar surgery at various levels from L2 to S1. after SCI. While previous investigators have utilized cellular transplantation METHODS: One of the authors (SAB) performed 1315 approaches on of xenografted pluripotent embryonic stem (ES) cells and observed improved 1310 patients between August 1997 and December 2002 using a mini- functional outcome, these transplants have required steroid administration open approach previously described by him. A concurrent database was and only a minority of these cells develop into oligodendrocytes. kept on these patients to track the approach complications with particular PURPOSE: The objective of the present study was to determine whether emphasis on vascular injury. Calcification of the vessels alone was not allografts of oligodendrocyte precursors transplanted into an area of incom- considered a contraindication to surgery. All patients had pedal pulses plete spinal cord contusion would improve behavioral and electrophysiolog- evaluated pre-operatively, but no vascular studies were performed before ical measures of spinal cord function. Additional treatment incorporated surgery. Pulse oxymetry was used to measure oxygen saturation (SaO2) the use of the glycoprotein molecule Sonic hedgehog (Shh) which has been in the left foot during and for 4 hr. after surgery in our last 629 cases. shown to play a critical role in oligodendroglial development and induce RESULTS: There were 643 males and 667 females. Age ranged from 19 proliferation of endogenous neural precursors after SCI. to 84 years and weight from 94 to 337 lbs. Six of the 1315 cases (0.45%) SETTING: Laboratory study. (5 female 1 male) had left iliac artery thrombosis (LIAT). All six had METHODS: Moderate spinal cord contusion injury was produced in 39 exposure at L4-5. Five were diagnosed at surgery and one in the PAR after adult rats at T9-10. Ten animals died during the course of the study. Nine posterior surgery. All patients had 2+ pedal pulses pre-op except for 1 patient rats served as contusion controls (group 1). Six rats were treated with with a 1+ left pedal pulse who ended up having LIAT. Pulse oxymetry oligodendrocyte precursor transplantation 5 days after injury (group 2). confirmed the diagnosis in the the last two patients when LIAT was The transplanted cells were isolated from newborn rat pups using immuno- not clinically obvious. Four had thrombectomy (1 with intimal tear repair) panning techniques. Another eight rats received an injection of recombinant 1 a fem-fem bypass and 1 an axillo-femoral bypass. Four had no sequelae Shh along with the oligodendrocyte precursors (group 3), while six more and 2 had compartment syndromes. There were 19 venous injuries (1.4%) rats were treated with Shh alone (group 4). Eight additional rats received to the left common iliac vein and all but 4 were at L4-5. All were repaired only T-9 laminectomies to serve as non-injured controls (group 0). Animals with blood loss ranging from 100cc to 3000cc (mean 650cc). In 2 patients were followed for 28 days. the procedure was aborted following repair in one and ligation of the vena RESULTS: Following an initial complete hindlimb paralysis, rats of all cava and both iliac veins in the other. groups receiving a contusive injury recovered substantial function within DISCUSSION: Our study shows that although the incidence of vascular 1 week. By 28 days, rats in groups 2 and 3 scored 4.7 and 5.8 points better injury is low (0.45% for LIAT and 1.4% for venous injury), the sequelae can on the BBB open field locomotor score than rats in group 1 (groups 2 and be quite significant. Two patients with LIAT required bypass surgery and 3=18.2 and 19.4 points respectively after 28 days vs. group 1=13.6 points; two that developed compartment syndrome had significant residual damage p=0.015). Rats in group 4 scored no better than those in group 1 (BBB= to the left leg. The two patients with venous injury that had to be aborted 16.4). Motor evoked potential (MEP) recordings revealed significant im- recovered without incident, but later required posterior instrumentation. provement in latency measurements in all treatment groups compared to CONCLUSIONS: A prior study by one of the authors (SAB) has shown controls at 28 days, although 3 animals in group 1 and two animals in group that the majority of patients undergoing anterior surgery at L4-5 have 3 were not recordable. Histological examination demonstrated significantly complete occlussion of the left iliac artery during retraction. Venous return more spared white matter in the same groups that correlated with the is also obstructed at the same time. With the amount of exposure required it improvements in BBB scores and MEP latencies. Immunohistochemical is seems likely that vascular injuries may occur. Ligation of the ileo-lumbar analysis showed the survival, proliferation, and migration of the transplanted vein in exposing L4-5 is mandatory to avoid tearing it while mobilizing the cells, as well as the induction of proliferating endogenous neural precursor left iliac vein to the right. If small venous lacerations occur during cells in animals treated with Shh. exposure and they can be repaired easily then the procedure can continue. CONCLUSIONS: These findings suggest that the transplantation of oligo- The sequelae of LIAT can be minimized by monitoring SaO2 in the left dendrocyte precursors may improve axonal conduction and spinal cord leg with pulse oxymetry. Failure of the SaO2 to return to normal after function in the injured spinal cord. The benefits seem more pronounced with removal of the retractors is diagnostic of LIAT and treatment should be the addition of Shh, and the addition of Shh alone results in the proliferation instituted immediately. Since only 5 of 1315 patients suffered significant of an endogenous population of neural precursor cells. sequelae from vascular complications it appears that a mini-open, anterior DISCLOSURES: No disclosures. approach is safe, though it must be carried out with great respect for the CONFLICT OF INTEREST: No conflicts. vessels to avoid possible catastrophic outcomes. DISCLOSURES: No disclosures. CONFLICT OF INTEREST: No conflicts. doi: 10.1016/S1529-9430(03)00236-5 doi: 10.1016/S1529-9430(03)00237-7 : = g=2

CAE

Cpltn, tfll, nd hr Mnnt

hn W. rntn, G . l

Te uma Ieoy usio Cage (euy Sie, ayam, MA use o oseio uma ieoy usio (I a goo esus i is iiia cusom use a was a oe y e A i ase o aoae esus i a muicee osecie suy i wic o sugeo eome moe a a ew oe cases. Aoug saisica e sus ae es eemie oug osecie suies, oe essos may e es eae ae a ew sugeos ae a a oouiy o eom seea ue cases. e goa o is suy is o eemie esus, comicaios, ias, a guieies a ca e escie we e uma Ieoy usio Cage wi eice scew iaio is use ouiey i a ciica iae acice seig o eame o aies wi isaig ack ai secoay o egeeaie isc isease.

e cue suy is a eosecie eauaio o e is 42 aies wo ae a su gey y a sige gou a oigiay icue ie ooeic sie sugeos ewee euay a Ocoe 200, we e gou icue ee sugeos. e suy was iiiae i Ocoe 2004 o esue a eac aie a e oouiy o a yea oowu. Icusio a ecusio cieia ae ise i o . Aoug yea is ge eay isuicie o eemie sugica oucome o a saisica asis, i sou e sui cie o ieiy ias a comicaios.

A cas wee eiewe a aa ace i a comueie aaase y wo eei ece sugeos, oe o wom a o aiciae i e sugey (G.... ecause e eoi o e suy was e as ecoe oowu, aies wee suie o aiae ieas om o 4 yeas. May u o a aiogas wee eiewe i is suy. a iogas wee eiewe o eey aie wo a ess a a goo ciica oucome. I wo , e sugeos ntrprttn o fn was ecassiie as fn ai ue i e aaase. ,, 2:2 C:,

O I CUSIO ECUSIO CIEIA

Icusio Ecusio isaig ack ai esuig om egeeaie Iecio o e sie isc isease a oe o moe uma ees Sigiica oseooosis imay egeeaie isc isease Meaoic oe isease aie iscecomy ogem seoi eay Soyoisesis Sia maigacy aie usio egeeaie soyoisesis

emogaic aa wee ecoe icuig age, se, comesaio issues, euoogic ucio, ume o io ack sugeies, oeae ees, oeaie iicuies, comica ios, eoeaios, o eisios.

Ciica eauaios icue aigs o ai, ucio, ecoomic saus, a meicaio us age accoig o oi ike scaes use i e eious IE suy. e oi ike scaes wee ae o ceae a 4 o 20oi moiie oo scoe o eie ecee ( o 20 ois, goo ( o 6 ois, ai ( o 2 ois, a oo (4 o 8 ois, cosise wi eious ieaue a escie uy i Cae .

usio success was eie as aiogaic eiece o oe igig e usio aea wi o ucecies. I ay ucecy was see eeig acoss e isc iesace, e ee was cosiee o use. I muieee sugeies, a ees a o e use o e a ie o e cosiee a usio success. ecause cao ie eioce oyme (C cages ae aiouce, igig oe ca e accuaey assesse o ai aiogas.

Ae e oigia eiew o 42 cas, yea aa wee aaiae o 2 aies (6.. ecause 8 aies a ee os o oowu, G... sueise e oice sa o coac ese aies a oai ue iomaio y eeoe ieiew (oes 2 a . oysee aies wee coace a oie ew ioma io. O ese 4 aies, 26 (. wee oig we a e o ee o eu, a see (4. aie o eu ecause ei sugey was usuccessu a ey wee cosuig oe ocos. e comue aaase was uae o icue e ciica i omaio oie y eeoe ieiew. O e aies se o coecio agecies, may a ecee eay sugica esus we as see. I ese aies wee as see ess a a yea osoeaiey, ei saus was cosiee os o oowu. We ae u awae o ay oe suy eemiig e easos a aies ae os o oowu.

COMICAIOS, IAS, A EI MAAGEME 4

O 2 SUM MAY O 8 AIES IIIAY OS O OOWU

See i oice: 4 Coac wi ew iomaio: 4 ecease: 2 Se o coecio agecy: o kow oe ume: 4

x EASOS O 4 AIES O EUIG

oig we o ee o oowu: 26 Cage o os isuace cou o eu: ie oo a away: Sugey usuccessu cosuig oe ocos: Oe mao ea oems moe imoa:

Some aies oie muie aswes.

Oe ime aiioa iomaio was oaie o seea oe aies, esuig i ci ica oowu o o 42 o 8.8 o aies. I is gou, ciica success was aciee i 22 o (8.8 a usio success was aciee i o 0 a ies (.. oe a some aies a oy ciica o usio aa u o o. ese success saisics ae cosise wi umes eiousy eoe. o saisi cay sigiica ieeces i oucomes eise amog e aiciaig sugeos.

e sugica ecique i Cae 2 ecommes aceme o aiioa oe ga ewee a esie e cages weee ossie. Some sugeos ecomme iig e aeio isc sace wi oe ga eoe aceme o soe cages, a some ae use oy oca oe ga om e ecomessio. . 8 I e cue seies, a aies a sugey y a seio sugeo assise y a seco seio sugeo. e oeaios wee eome o a cosugeo asis i a eac sugeo i e wok o ei esec ie sie. A cages wee ie wi caceous oe aese oug a sma wiow i e iiac ces. e oy ieece i ecique amog e aiciaig sugeos was a oe sugeo i o wis o ace aiioa oe i e isc sace, saig a

eeeces , , 4, 20, 22, 26. aceme o ea oe was "giig e iy." I aayig e aa o e suy, we ie— u wee uae—o oe a is sugeo a a owe usio ae.

I a ey comee iscecomy is oe, e sace is ie wi oo ae sugey. e sie oeigs o e cages oie coac wi oe ga, a e eiousy emy sace e comes esseiay a acue emaoma. We ae see cases i wic o ea oe was ae, u aiogas ake 2 yeas osoeaiey sow a e oe as e ee o i e eie isc sace. We si eiee a sugica icies equie iig e eie isc sace wi as muc oe as ca e ace. owee, i ee is isuicie sace, isuicie oe, o isuicie ime aaiae, e usio ae is si saisacoy. We cauio agais muie comomises o ecique.

Comicaios i e 42 aies ae summaie i o 4 a ca e comae wi ose eoe i e IE suy i ae . ee wee 26 eiceeae comica ios, ioig e cage a ioig e eice scews. ee wee ewe oke eice scews i e cue suy (0.4 a i e IE suy (6. a ewe oose scews i e cue suy ( .8 a i e IE suy (2.. A scews i e IE suy wee mae o saiess see a scews i e cue suy wee mae o iaium. Oea, e ume o scew comicaios ee owe i e cue suy. e cue suy coime e ow ae o ese comicaios eoe i e oigia IE suies.

Aiioa cage aceme ias wee osee a esee meio. I oe aie, cages wee ou o e uesie. ig. sows a aea aioga o is aie. We eiee a is oem was cause y icoec ieiicaio o e isc sace. eae o is oem, ee wee ou cases eee o as "ase cae" o "mao siioe oac." ig. 2 sows a case i wic e oac cu eey io e ee

o 4 COMICAIOS O CIICA SUY GOU

eiceeae Comicaios: oeiceeae Comicaios: 26 o 42 (6.02 o 42 (4.2 eoeaio o eai CS eak: (.82 C Cpltn Iecio: 6 (.6 Cage migaio: (0. Eiua emaoma: (0.24 oke cage: (0.2 oo o: 2 (0.48 ase cae: 4 (.00 a ack: (0.24 Aeio aceme: (0. Mao meica: (0.24 Uesie cage: (0.24 Seoma: (0.24 Aeio eea acue: (0.24

dl Sr Cpltn oke scew: 2 (0.4 oose scew: 0 (.8 Maosiioe scew: (0.24 COMICAIOS, IAS, A EI MAAGEME 4

COMICAIOS COMAE WI IE SUY Cage IE ( Cue Suy ( eoeaio o ua ea .80 .82 ee iecio 2. .6 Mao meica comicaio 0.4 0.24 S .6 0 eice comicaio .00 6.02 Sugica ea 0.0 0 ee oo iuy . 0.4 Eiua emaoma 0 0.26

ig. I is case, cages wee uesie ecause o aiue o accuaey ocae e isc sace.

ig. 2 oaces cu eey io e a S oe ceaig a "ase cae" a a imoe i.

a eaes, esuig i seuaosis. e aoiae ecique o use o aoi is oem is iusae i Cae 2. A i eame o seae is ace io e isc sace a sea. I ee is ay quesio o e accuacy o e aceme o is isu me, a aioga sou e ake o eiy is osiio. e seae is use o is ac e isc sace o oma isc eig. "Wokig aes" ae igee ow o o e isacio. A ou eame eaes a ou oe. e goa o eamig a oac ig is meey o ae e eae, o o cu a ee cae. eeoe o aceme o a mm cage, e isc sace sou e isace o mm. A mm eame sou e use, oowe y mm oaces.

e "guie oac" oows e ou cae a makes i squae. e "ia oac" o ows e eious cae a squaes e oommos oio o e cae. e ia oac sou e e wi igei oce oy a aowe o oow e eious cae. A ase cae is ceae i e ia oac is ace i a iecio iee om e guie oac, us ceaig a ecessiey age cae a wi eisose o a oose iig cage a seuaosis. We ecomme a i ee is ay quesio o e co ec iecio o e ia oac, e ia oac ca meey e omie.

ee wee ee cases o aeio aceme o e cage. ig. sows a aea a ioga o oe case. I mos cases e cage comes o a im so agais e aeio auus iosus. I ig. , e cage as eeae oug e auus. Oe e e ec i e auus is cause y secig equie o eucio o a soyoisesis. A ee o ou aeio aceme oems occue a S a cause o aese e ec. I aeio cage isaceme occus a 4 o aoe, e sugeo sou eosiio e cage, oe a iicu ask a may equie a seaae aeio aoac o e sie.

Make ea o aeioy ace cage

ig. Oe cage was ace oug e aeio auus. I is occus a ige e es, i sou e emoe oug a aeio aoac, i ecessay, o aoi amage o e gea esses. COMICAIOS, IAS, A EI MAAGEME 4

I oe case, a aeio acue o e eea oy was osee (ig. 4. is acue occue ecause e eae cae was o suiciey ee o coai e cage. We e sugeo ammee e cage eee, i acue e coe o e e ea oy. o aese eec was oe uig e eaig ocess. We o o eiee a ay secia ecauios ee o e ake osoeaiey we is occus.

I ee cases, e cages wee cassiie as aig migae. ig. sows oseio migaio o oe cage, iicae y e osiio o e make ea. is aie a ou io uma sugeies a eeoe aicua ai aoimaey mos ae cage usio. e cage was emoe oug a oseio eosue. ecause ee was ey i e moio, e eice scews wee aso emoe. ig. 6 sows aiogas 6 mos ae. ecause eio a eesio ims sowe o moio, e aies su

ig. 4 e aeio i o was acue as e esu o a aem o ie e cage eee a e eae cae.

ig. Aeooseio (A a aea ( iews o a case i wic e cage eouse. geo cassiie im as a usio success. We ae ecassiie is case as a ocke seuaosis a a usio aiue. O ia oowu, e aie a ey ie ack ai a o aicua ai, a was cosiee a ciica success.

ig. sows e ia iaoeaie aioga o a aie eae o egeeaie soyoisesis. e cages aea o e coecy osiioe. ee mos ae su gey, ee was a ecue 4 mm si (ig. 8. y mos ae sugey, is si a icease o 6 mm (ig. . Aoug e aie a o euoogic symoms a oy mi ack ai, i was oug a e cosuc sou e eise. e eisio sugey was eome eiey oseioy. e eice scews a 4 wee ou o e oose as a esu o oseooosis. e cages wee emoe, aigme was esoe, age cages wee ace, age scews wee ace a 4, a a seco oi o o i ma iaio was gaie a . A ia aioga is sow i ig. 0. e aie a

ig. 6 Aeooseio (A a aea ( iews ae e eouse cage was e moe. ecause e segme was sae, e eice scews a aes wee aso e moe. is esu is oug o e a "ocke seuaosis."

ig. ia iaoeaie iew o aie wi oseooosis. COMICAIOS, IAS, A EI MAAGEME 4

E=2

. 8 ee mos ae sugey, ee was a 4 mm ecue owa si.

. A mos ae sugey, e si a icease o 6 mm.

. 0 Ae a oseio eisio, aigme was esoe.

COMICAIOS, IAS, A EI MAAGEME 4

I ecame immeiaey aae a sugica imes a oo oss wee cosieay e cease i is ciica suy we comae wi e IE suy. I e IE suy, e a eage aie a wo io aie sugeies a wo ees. e aeage oo oss i e IE gou was m. I e is 60 aies i is ciica acice gou," e ae age oo oss o imay oceues was 26 m o oeee a 6 m o woee usios. I aies wi io sugey, e aeage oo oss was 8 m o oeee a 40 m o woee oceues. e aeage sugica ime i e IE suy was 2 miues. I is ciica gou, e aeage sugica ime o imay cases was 202 miues o oeee a 2 miues o woee oceues. I aies wi io sugey, e aeage sugica ime was 208 miues o oeee a 2 miues o woee oeaios. I ou cases we aways a wo seio sugeos aiciae i eac sugey, a we eiee ee is cosieae ime sae we comae wi oe se io sugeo wokig wi a esie o eow. I moe ece cases, ou sugica ime o oeee imay cases is ouiey ue 20 miues.

I e IE suy, ee was o eo o imi aies use o oseoia aiiamma oy ugs (SAIs eoe sugey. I is seies, we equie aies o aoi ai ioa SAIs o 4 weeks eoe sugey. We esic SAI use oge a . Sow e ecommes i Cae 8. CO2 iiios cause o iicuies i is ega. Aoug i may e aecoa oiio, we ae ou a aies wo a ueecey ig sugica oo oss wee ou o e akig ueoe SAIs. I aiio, we ae mae a coscious ecica eo o coo oo oss a eey se o e sugey. We eiee a ime se o emosasis owes oo oss, aows moe ecise ima aceme, a soes e oa sugica ime.

I e IE suy, oe sugeo ouiey a oo oss igues oue e aeage o e oe sugeos. I wo o is sugeos cases, e oo oss eceee 0,000 m. wo o is sugeos aies ie iaoeaiey. We a ig ee o oo oss is oeae o a egy eio, e eeig may suey acceeae o a aamig ae as a esu o iaascua coaguoay. Sugeos sou e kowegeae o e icies o asusio maageme summaie i Cae 8.

rdt Ud t Cntrl ldn

A ume o oucs ae aaiae o use i cooig eeig uig sie sugey. Aoug mos sugeos ae wokig eeiece wi ese oucs, may ae u awae o imoa eais i ei ackage ises.

e Geoam soge is iee o aicaio o eeig suaces as a emosaic age. I is a waeisoue, owie, iae ouc eae om uiie ok ski, geai, a wae. Geoam owe is eae y miig asoae Geoam soge. e moe o acio o Geoam i cooig eeig is mosy mecaica. Geoam is usuay comeey asoe wii 4 o 6 weeks, a ca e use y o sauae wi seie saie souio. Aoug mos sugeos sauae Geoam wi oica omi, e eicacy o is comiaio as o ee ese y e mauacue, a e ack age ise saes a is use "cao e ecommee." Geoam aso may swe o a soig uis a eeoe sou e emoe ae use i amiecomy oceues. e mauacue saes: "We Geoam was use i amiecomy oeaios, muie euoogic ees wee eoe, icuig u o imie o caua equia syome, sia seosis, meigiis, aacoiiis, eaace, aesesia, ai, ae a owe ysucio, a imoece." Aoug may sugeos use a see o Geoam o o o e ua as a oecie aie uig wou cosue, i is oay imoa o aoi eaig is ouc i ig, oy saces suc as e sia caa.

oe wa is comose o 80 eie wie eeswa a 20 isooy amiae. oe wa aciees oca emosasis o oe y acig as a mecaica aie. Aoug i is a asoae maeia, i is oug o icease e isk o iecio a o esis a e oy sie o yeas. I sou e use saigy, a ecess maeia sou e emoe fr th rl t.

Aiee is a micoiia coage emosa use o coo oo oss we emosasis y igaue o oe coeioa oceues is ieecie o imacica. Aiee is o ie i a owe om, i sees, a as a coage soge cae Uaoam. I a suy o eeig om kie icisios i ig sees, e mauacue emosae a Ua oam wiou omi eome as we as Geoam wi omi. ey caime a e aaage was a oe ca aoi use o oie omi a e aea ossii iy o omig aioies o oie omi, wic cou eciiae aco eiciecy a aomaiies o emosasis o eeosue.

Sugice is a oiie, egeeae ceuose i a see weae. I is a ase a as o aima o uma comoes. Sugice cooms o u oes o aee o iegua suaces. We sauae wi oo, i aiy swes io a geaious mass. Sugice aciees emosasis ecause is ow eaues aumi a goui. I is usuay a soe wiou issue eacio. e mauacue ecommes a Sugice e emoe oowig use i e sie. ee ae ee a eas si eos o aaegia esuig om Sugiceeae oacic co comessio.

osea is a asoae emosaic age ackage as a ki coaiig a egieee oie coage—eie geai mai comoe a a omi comoe. A oi eay isesio eee is oie o mi e comoes uig use. e mie mai cooms o iegua eeig suaces a swes aoimaey 20 o coac wi oo o uis. osea is yoiic a aees o we issue. e aicuae aue a ows ecess maeia o icooae io e co o e emoe y gee iigaio wiou isuig e emosaic sea. Sugeos sou e awae a ecause o e sweig o e geai mai, osea is o iicae o use i euosugica oceues. I is, owee, ey eecie i soig eeig om eiua esses mae ouesome y aiioa SAIs. = =, $4 f= K .2= C"2

o ue isaceme oe ime. e imay sugeo ieee ia aiogas as sowig usio owee, we eiewe is aies ims a ise e esu i e aa ase as seuaosis.

O e aies wi cage oems, 0 we o o oma usio success. ee wee coue as usio aiues. Oe case esue i a eiie seuaosis, oe was co siee a ocke seuaosis, a i oe e usio was uceai a eeoe coue as a usio aiue.

e aie wi a maosiioe scew a egeeaie isc isease a 4 a S wi esie aicuoay. o eious sugey a ee eome. is aie awoke wi seee igsie aicuoay wi o moo weakess a o e eg ai. ee was o esose o seoi eay. ee mos ae sugey, a C sca (ig. sowe a e scew o e ig sie a eeae e aea wa o e ee a oy i a aea cose o a ee oo aeioy. Iecio o oca aeseic wi co as meium (ig. 2 eiee e aicua ai. ue sugey was oe immei aey o emoe e oeig scew. e ia ciica esu was usaisacoy. e aie cou si a wok mos o e ay wiou uue iscomo, u a make aic ua ai wi a aciiies. We eiee a a MI a C sca sou e oe immei aey osoeaiey we a aie eeieces ew aicua ai ae sugey. is was oe o oy wo ee oo iuies a we ecouee amog e 42 cases i is suy. e oe ee oo iuy is escie i Case o Cae 6.

ig. e scew eeae e aea wa o e eea oy a cause iiaio o a ee oo aeioy.

,..? , 1-,119,5 wi coas a uiacaie, e ai was emoaiy ocke. ecause o is e scew was emoe. COMICAIOS, IAS, A EI MAAGEME = = •

Esse ii Seaa is mae y miig wo comoes, eiee oug a oi

E:: , uaisesio eee: a coceae iioge wi aoii, a omi wi i coie. issee oms a soi coaguum wii o miues o eiey a e

: :aces e as sages o aua emosasis, coeig iioge o ii sas, _ waeig seas. issee comoes ae eie om uma asma a eeie aoii. issee is wiey use ae uoomy eai i sie sugey o a waeig sea. Mos sugeos eiee a ae ua eai wi issee, a

E .S ca e amuae o a oma sceue a o o ee ays o e es, wic :e ecommee ae ua eai wiou issee.

A ouiey iies ue sugey io ee caegoies: eoeaio, eisio, emoa sugeies (o . prtn is eie as a sugica oceue i : e eices wee o aee o ecoigue. vn is eie as sugey i

: ie eices wee ecoigue, eee, o eace, u some oigia eemes

E — si. vl iicaes a e cages o eice scew eices ae emoe. Oea, aies (0.6 a 4 ue oeaios o ei uma sie.

prtn EIUA EMAOMA

aie eeoe a eiua emaoma. euoogic ucio was oma immei E=Ei ae sugey owee, ogessie owe eemiy weakess eeoe wii e

s: 24 ous osoeaiey. A MI sca oaie e moig ae sugey ocu

e ce e emaoma. e emoac ai aeae o e ucioig omay. Sugey ES eome a ay o ai e emaoma. e comessio was ou i a aea , E e o om e ai. euoogic ucio imoe quicky, a e aie egaie ..eg wii a ew weeks o e iiia oceue. e imoace o oig imme : magig suies i e esece o ay ew aicua ai o euoogic aoma

=o e oesesse.

o UE SUGEY4 OEAIOS I 0 AIES

eoeaios eisios eoai CS eak: 8 (.8 eise eice scews: 4 (0. ea iecio: 4 (0. eise isace cage: 2 (0.4 AI o seuaosis: 2 (0.4 Ee usio: 40 (.4 eai seuaosis: (. eisumeaio: (0.2 _ue ecomessio: (0.6 AI (emoe aie AK: (0.2 Eiua emaoma: (0.2 emoas emoe eice scews: 8 (. 2 COMICAIOS

CS EAK

ome o eai CS eaks. Ou gou woe a eo o si ,..4,. ,c■k.e,..c.„,e, ,.,_

cases o osoeaie CS eak assoiaa l o, u a a e E eece o uic a i,o ass ies o ou cases was uise, — ao iue a eome ewee Augus s: e. cases assoae6 - = -1 Si 7 G. • "L. -•."" "3- h rrrr6 rr .

}?° c °&° 24, 2 4 ?& Q?? , G 622c ? 2 &•=c 2_ • a 0222 a sie sugey a geeay ae o ogem eeeious eec. Aoug eoea io is icoeie o eeyoe ioe, oe o ese aies a ay ogem a ese eec. ese ua eas a occue uig issecio o sca, o uig oacig o cage aceme. Scauo e a 2 eoe i 200 a aceme o cyii ca cages oug a I aoac a a 2 ae o ua eas, oe a e oo a ia, wi a 20 ae o osoeaie weakess o aicuoay. ese comicaios i o occu i is seies, ecause ess ee oo eacio is equie o aceme o ecagua C cages a ecause u ee oo moiiy was aciee eoe ay I ses wee iiiae.

EAME O IECIO

ie o si cases o iecio occue wii a mo eio i 2002. ese iecios ioe ee iee osias a iee ogaisms. esie iesie iesigaio, a cause was ee ou, a e iecios cease. e oea iecio ae o .6 is owe a a eeiece i e iiia IE suy a is wii a acceae age. We ouiey use iaeous isgeeaio ceaosoi eoeaiey a o 24 ous ae sugey. acos associae wi icease iecio icue ooge sugi ca ime, muie eious sugeies, a aiig wou, a coic iecios i e a e, ee, o oe aeas.

We ecomme a a emoacye ai e use i a cases. Aoug osoeaie eeig ca usuay escae oug e eges o e icisio, e ack o a ucioig ai is associae wi icease isk o eiua emaoma, coic aiage, a i cease isk o iecio.

I e esece o a coic aiig wou o susece iecio, e aie sou e eue o e oeaig oom, e wou oee a iigae oougy, a aii oic eay iiiae ase o iiia emiic seecio, moiie ae y cuue esus. I e imas ae sae, ey sou e e i ace. We ecomme a e wou e cose oe emoac ais a e aie e eue o sugey eey ays ui e wou is cea a cuues ae seie. I e wou is iiiay acke oe, i may o e ossie o oai imay cosue we e iecio is cooe. I e iecio e CS EAK

Si eoeaios wee eome o eai CS eaks. Ou gou woe a eo o si cases o osoeaie CS eak associae wi e o ACO (Giaec, Ic., Cee a, O, a aiaesie age. is eo was eece o uicaio, u a ae se ies o ou cases was uise, 8 a ACO was susequey wiaw om e make. e cases associae wi ACO wee a eome ewee Augus a Oc oe 8. e si cases eoe ee wee o associae wi e use o ACO. Sice e gou ega e ouie use o issee o ay ua eais, is comicaio as occue i oe case.

We agee wi Wag e a 2 a Cammisa e a a ua eas ae a ouie a o um a sie sugey a geeay ae o ogem eeeious eec. Aoug eoea io is icoeie o eeyoe ioe, oe o ese aies a ay ogem a ese eec. ese ua eas a occue uig issecio o sca, o uig oacig o cage aceme. Scauo e a 2 eoe i 200 a aceme o cyii ca cages oug a I aoac a a 2% ae o ua eas, oe a e oo a ia, wi a 20% ae o osoeaie weakess o aicuoay. ese comicaios i o occu i is seies, ecause ess ee oo eacio is equie o aceme o ecagua C cages a ecause u ee oo moiiy was aciee eoe ay I ses wee iiiae.

EAME O IECIO

ie o si cases o iecio occue wii a mo eio i 2002. ese iecios ioe ee iee osias a iee ogaisms. esie iesie iesigaio, a cause was ee ou, a e iecios cease. e oea iecio ae o .6% is owe a a eeiece i e iiia IE suy a is wii a acceae age. We ouiey use iaeous isgeeaio ceaosoi eoeaiey a o 24 ous ae sugey. acos associae wi icease iecio icue ooge sugi ca ime, muie eious sugeies, a aiig wou, a coic iecios i e a e, ee, o oe aeas.

We ecomme a a emoacye ai e use i a cases. Aoug osoeaie eeig ca usuay escae oug e eges o e icisio, e ack o a ucioig ai is associae wi icease isk o eiua emaoma, coic aiage, a i cease isk o iecio.

I e esece o a coic aiig wou o susece iecio, e aie sou e eue o e oeaig r, e nd pnd nd rrtd oougy, a aii oic eay iiiae ase o iiia emiic seecio, moiie ae y cuue esus. If th plnt r tbl, ey sou e e i ace. We ecomme a e wou e

wou is cea a cuues ae seie. I e wou is iiiay acke oe, i may c: ossie o oai imay cosue we e iecio is cooe. I e iec: COMICAIOS, IAS, A EI MAAGEME

g?, . 4:2, ,42 : , W CI, quies oe ackig, i is ossie o oai secoay cosue usig a iecoece aissimus osi—gueus maimus muscuocuaeous a."

vn CAGES O EICE SCEWS

Cage eisio cases wee escie eaie. I a aiioa case, a aie wo a a a 4 AK usio a uegoe eisio wi iaea eice scews a oseoa ea usio om oug S wi usio usig cao ie eioce oyme (C cages a S. e aie ae a a aeio oceue o emoe e AK cages a eace em wi a sackae C cage. is case is iscusse i Cae 4 (see Case .

EE USIO

ue sugey was eome o ee e usio i 40 aies wo a aace segme egeeaio. A ume o auos ae aesse e oem o aace segme egeeaio. I 88 ee eice a aicaio o a igi iea ia io sysem ouce geae sess coceaio i e aace segme. Kuma e a 6 suie 28 aies wo uewe usio sugey a a mace gou o 28 a ies wo uewe ousio uma sugey. ei suy ase 20 yeas. e ici ece o aiogaic cages i e aace segme was wice as ig i e usio gou as i e ousio gou owee, ee wee o saisicay sigiica ie eces ewee e wo gous i oucome measues. ai e a suie 0 aies wo a uegoe usio wi eice scews a ou a aies wi esee oseio come iegiy ewee e use segme a e aace segme ee oe isaiiy i 6.% esus 24.% wiou esee oseio come iegiy. Scege e a 24 suie 8 aies wo a uegoe uma usio a wo a a a eage symomee iea o . yeas. A e oi o suy, ese aies eeoe ecue seee symoms. Sceges gou ou a segmes aace o e aa ce segme wee as ikey o eak ow as e aace segmes. amy e a agee a egeeaie cages occue i e seco ee aoe e use ees wi equecy equa o ose occuig i e is ee. Gisei e a suie 2 a ies wo a uegoe oseio uma usio a aeage o 6. yeas eaie a e oe a e ae o symomaic aacesegme egeeaio waaig ue sugey aeae o e 6.% a yeas a 6.% a 0 yeas. Ou eious eo o e 0yea esus o aies eae wi e uma eoy usio Cage i e IE suy iicae a 6% o aies a aiogaic cages i e aace seg me, u e cages wee ciicay sigiica i oy 20%. O ese, % a ue goe aace oceues a e oe aie was coemaig sugey.

I e cue gou, .4% o e gou as a woe uewe aacesegme usio oceues wi aiae ieas o oowu om o yeas. We ecomme a e ue 0% o e amia e esee aog wi e igameous sucues weee

eeeces 2, , 6, , 24. ossie, wi e eceio a a u amiecomy sou e eome i i is ecessay o ecomess o oec e eiig ee oo. We a oio o e amia a e oseio igameous come ae esee, a suseque usio sugey ca oe e eome oug a iesace aig oma issue wi ie iicuy om eious sca. I is ou imessio a aies wo ae a goo iiia esus ae cage usio ca agai aciee goo esus ae aacesegme oceues. oo ew o ou a ies ae suicie oowu ae e seco cage oceue o eauae is coce saisicay.

EMOAS

eice scews a aes wee emoe i 8 o 42 aies (.%. is is a owe ae a eeiece i e IE suy, i wic eice scews wee emoe i .2% o aies. Iicaios o emoa o eice scews ae ue eiewe i Cae 4 i a iscussio o eauaio a eame o aie cages. We eiee a i is iicu o eemie we eice scews ae a souce o ai. Ceaiy, i a scew is oose, i may e aiu. Aems o amiise a iagosic iecio o ock e ai may ai o eie e aeseic age o e aiu scew ac eeoe a egaie iecio suy may o e meaigu. We eice scews ae emoe, we ecomme a e mi i e icisio e oee o e ascia ee a ski as eeoe o o sies. We ecomme a a iaea ascia icisio e mae aou 4 cm aea o e miie a Wiseye muscesiig aoaces e use o emoe e scews a aes. is aoac is ess aumaic a oeig e miie eosue, wic may e e quie i a cosscoeco was use. e oceue ouiey akes ess a a ou, e aie equies oy a oeig say, a ca usuay eu o u aciiies i a week o wo. e oy oem wi is aoac is a i ceaes a oeia ocke o ea sace aoe e ascia a ca accumuae ui. e oe aie wo a a eoeaio o seoma a coiuous eaccumuaio o ui o mos om is souce. issee is useu o seaig e ocke sou is oem occu.

ECOMESSIO A UUSE EES

Seee eeaey wae a sugeos sou ee e a usio a a ecomesse eea ee. Weee a ecomessio is eee aoe a usio, a ee sou aso e icue i e isumeaio a usio. Ceaiy is ecommeaio makes iuiie sese. I emoig e oseio igame come aoe e usio esus i ige aacesegme aiue, as ai e a ae osee, e moe eesie esa iiaio sou aso esu i moe eque mecaica aiue o e segme.

wo o e sugeos eece Seees aice a eome ecomessios aoe usios. O ese aies, 26 wee aaiae o oowu. Ciica success i is su se was aciee i o 26 (%. We ese esus ae comae wi ose o e suy gou as a woe usig e ci squae saisic, ciica success aes wee o saisicay iee (ci squae . p = 0.2. ue sugey was oe i 2 o th 26 ptnt th djnt dprn (46.%. Whn th rlt r rn COMICAIOS, IAS, A THEIR MAAGEME

ae wi ose o e suy gou as a woe usig e ci squae saisic, a saisi ca ieece is aoace (ci squae 2.60 0.0. Sugey o ee e usio o e ecomesse ee was eome i si aies (2%. We ese umes ae comae wi ose o e suy gou as a woe, saisica sigiicace is eace (ci squae .8 0.48. eeoe Seees ecommeaio is coime y is suy gou: o o e a usio a a ee o ecomessio.

2 n

e aiga Ieoy usio Cage was aoe y e A i euay o ea egeeaie isc isease a oe o wo ees ewee 2 a e sacum we wo cages ae accomaie y iaea eice scew saiiaio. 2 was ecue om e IE suy ecause usio a is ee is aey equie, a saisica comaisos ae o aaiae. owee, occasioay uma usio may e iicae a e 2 ee ecause o aacesegme egeeaio ae usio o 2, mao isc eiaio wi egeeaio, aie eious amiecomy o iscecomy, egeeaie uma soyo sis, o oe causes.

wee aies wee ieiie i e suy gou wo a cage usio a 2. 8 I see cases, e iicaio was egeeaio aoe a og usio. I ie cases, e aace segmes a o ee eiousy use. ecause o icease isk o euoogic iuy o e cous meuais, e sugica oceue was moiie. e eie as ieaicuais a ieio aicua ocesses wee emoe iaeay, a cages wee ace oug a asoamia uma ieoy usio (I aoac wiou eacio o e ua. See aies (% aciee ciica success. Ciica esus icue eo ecee, wo goo, ie ai, a ie oo as measue y e moiie oo scae eoe e iousy i Cae . usio was successu i see cases a aie i ie.

e uma Ieoy usio Cage IE suy eoe usio success i 6 o 8

(8.°4 I e age ciica seies om ou gou, usio success was aciee i 2 o 26 o—2 aies (8%. We e usio esus o ou age seies ae comae wi e IE esus usig e ci squae saisic, e usio aes wee o saisicay iee (ci squae 0.44 0.06. We e usio esus o e 2 gou ae comae wi e age ciica seies om ou gou usig e ci squae saisic, 2 usio aes ae sigiicay owe (ci squae 64.2 <0.000.

e uma Ieoy usio Cages IE suy eoe ciica success i o (86.8%. I e age ciica seies om ou gou, ciica success was aciee i 28 o 4 (82% o o—2 aies. We e ciica esus o ou age ciica seies ae comae wi e IE esus usig e ci squae saisic, ciica success aes wee o saisicay iee (ci squae 0. 0.46. We e cii ca esus o e 2 gou ae comae wi e age ciica seies om ou gou usig e ci squae saisic, ciica success aes a 2 ae sigiicay owe a eeiece a e owe uma ees o I (ci squae 4.88 0.022. ig. sows aiogas o a aie wo a uegoe a 2 cage usio mo eaie. e osiio o e imas aeas saisacoy. ee is egee o measue kyosis. ig. 4 sows a aea im ake 2 mos ae sugey. e kyosis as icease o 0 egees. e kyosis icease o 2 egees 2 mos ae (ig. . A 6 mos ae sugey, e kyosis a icease o 4 egees (ig. 6. A 8 mos ae sugey, e kyosis seeme o ae saiie a 4 e gees wi e use o a SO ace (ig. . A 0 mos ae sugey, e kyo sis a icease o egees (ig. 8 esie e use o e SO ace. e a ie eee eisio sugey i e oe a e usio wou saiie. eisio sugey was ue eaye y e wokes comesaio caie.

ig. Aeooseio (A a aea ( iews ae sugey a 2. ee is egee o kyosis.

ig. 4 wo mos ae sugey, e kyosis a icease o 0 egees. COMICAIOS, IAS, A EI MAAGEME

c ou mos ae sugey, e kyosis _ we o 2 egees.

. 6 Si mos ae sugey, e kyosis a icease o 4 egees.

Eig mos ae sugey, e kyosis

: saiie a 4 egees wi e :ace.

. 8 e mos ae sugey, e kyosis a icease o egees. is aie equie eisio sugey. eomiy a auma sugeos ae og ecommee a a usio sou o e a ecause o a ecessiey ig ae o mecaica sess a aiue a e oa couma ucio. Isea, wo oacic ees sou e icue i e isumeaio a usio. We a eece a e ea saiiy aoe y aeio cage aceme wi eice scew iaio wou comesae o e ae sesses a is ee. is e ecaio was o eaie. e aiues wee o imie o aies aig a og usio eig a . O e ie aies wo a 2 use wi a ouse aace 2 seg me, wo aies a usio aiue a ee wee usuccessu ciicay, wi oo aia scoes. eisio as ee ecommee i o usio aiues.

asimoo a coeagues eo i Cae 6 ei aoae eeiece wi aeio oeee usio a e oacouma ucio, wic aeas o e muc moe success u. e aeio oceue may e moe successu ecause i aois emoig e sa iiig oseio eemes. I ay case, i mus e oe a C cage usio wi ei ce scew iaio eome oug a aiioa oseio aoac as a ige a eece ae o ciica a usio aiue.

U f SnlC I

A mecaica esig o e C cages assume use o iaea cage aceme a iaea eice scew iaio. e mauacue is equie o sae a aa o o e is o usiy e use o a sige cage comoe o uiaea eice scew iaio. ig. sows a aeooseio aioga o a case eome esewee i wic e sugeo ace oy a sige cage a S ecause e ee oo scaig om io sugey mae sae aceme o a seco cage quesioae. ig. 20 sows a C i mage o e same aie. O couse, is is e coec iaoeaie ecisio ecause a seco cage ca e ace oug a aeio aoac ae i equie. ee is cea aiue o usio oug is sige cage. e sugeo aie o ack ea oe aou e cage, a is was o a coec ecisio.

Moiai e a" eoe a suy i wic acieuy miiay esoe wee eae wi wo C cages a 6 wee eae wi a sige cage. We oy a sige cage was use, oe ga was isee om a iaea aoac. Moiai e a eoe a e esus wee geeay goo a a aies aig a sige cage a usio a ciica success equa o ose aig wo cages.

Some aies may equie a sigecage usio ecause o cicumsaces ou a e i me o sugey. e iicaios may icue o eoug sace o a seco cage (e cause a is cage was ace i e miie, aie eious iscecomy wi ee oo scaig, a cooie ee oos o oe oo aomaies eeig sae oo eac io. I aiio, we a ee o ouee ieoy usio is eome, e sugeo may ee o use sige cages a some ees o sae ime a miimie oo oss. I ig. 2, e is cage was ace i e miie a ee was isuicie sace o a seco cage. is case a coious ea oe gaig a esue i a soi usio a ciica success. COMICAIOS, IAS, A EI MAAGEME

ig. Use o a sige cage wiou ea oe ga esue i usio aiue.

ig. 20 Cooa C sca emosaes aie usio a S.

2 Aeooseio (A a aea ( iews o a sige cage case i wic e

.. EE .. ace. is esue i soi usio. I e cue suy gou, aies wee ieiie wo a I wi a sige um a Ieoy usio Cage a aiae scew aceme i a eas oe usio ee. weysi o ese aies a ciica success. usio was successu i 0 aies (%. e esus o usio a ciica success usig a sige cage wee saisicay o iee om e esus eoe i e IE suy ( 0.06 o usio success 0.46 o ciica success. e sige aie usio occue a 2, a ee a as ee ecogie as a ieee isk aco o aie usio. We emasie a we a usio is eome wi a sige cage, i is aicuay imoa o e sugeo o ack aiioa oe i e isc sace aou e cage.

U f C

We e uma Ieoy usio Cage was is esige, e oima imesios wee ukow. I aeae a e aeage wi a wou aow iaea cages o e ace was mm. ossie aiue moes icue susiece io e eea oies (wic ague o maima wi o e sus a usio aiue (wic ague o maima wi o e oeig o oe ga suace aea. I e esig o e mm cage ooyes ese mecaicay i caae sies e wo coces wee gie equa weig: e oeig was . mm wie a e su wi was . mm (o 2. mm eac. is e sig eome we i caae sies a oe ysioogic oas wiou susiece.

I e esig o age a smae cage sies, e wi o e sus was eaie a 2. mm ecause o e coce o susiece i e suo aea a ee euce. us e cage a is mm i wi as a oeig o oe ga a is . mm wie, ooioaey 8% aowe oea a e mm cage, u 6% smae i suace aea o oe ga. e mm cage as a oe suace aea 4% ess a e mm cage. ecause oe eaig was oeiay comomise wi e mm wie cages y e ooioay geae eucio i oe ga suace aea, i is ciicay imoa o suy e ciica a usio success o ese aow cages.

aies may equie a mm wie cage ecause o cicumsaces ou a e ime o sugey. e iicaios may icue o eoug sace o a age wi o eig cage i e isc sace, ee oo aomaies a may imi access o e isc sace, a im ie aiiy o eac ee oos ecause o eiua scaig.

I e cue suy, 0 aies wee ieiie wo a I a oe o moe ees us ig oe o wo mm wie cages. ° e mao saisica ieece ewee e emo gaics o e mm gou a e IE suy is aie gee. I e IE suy 4% o aies wee wome, a i e mm gou 66% wee wome, a saisica ieece ( 0.00. is ieece is o suisig ecause wome yicay equie smae im as.

I e mm gou, ciica success was aciee i 4 aies (82% a usio suc cess i 82 (.%. I e mm cases ae emoe om e age suy gou, usio COMICAIOS, IAS, A EI MAAGEME 6

success was aciee i 6.% o cases. We e usio esus o e age seies wee comae wi IE esus usig e ci squae saisic, e usio aes wee o saisicay iee (ci squae 0.44, p = 0.06. We e usio esus o e mm gou ae comae wi e age suy gou, e esus wi e mm cages ae tttll lr (h r = 0.04, p = 0.00. eeoe we cocue a use o mm cages esus i owe usio success o aou %.

Aoug i is ossie o aoi is oem y eomig geeous oe gaig i e isc sace ewee, esie, a i o o e cages, e easo o usig e mm cages is usuay a isuicie sace is aaiae o wie cages o o oe ga. e eoeace o wome i e mm aie gou imies a e smae cages wee aoiaey seece. Use o mm cages i age me simy ecause i is sugicay easie wi ikey esu i geae ecease i usio success.

wo ieesig oseaios mus e mae. is, i equie a suy o 42 aies o osee a % ieece i usio success. Seco, aoug e A aows e sae o umeous eices wiou ciica aa, e A wou o aow a MA sueme o a oe mm cages wose imesios ae eacy ooioa o e mm cages e cause e imesios ae ousie e owe imis o e eices ese ciicay.

n l n Sl n

aiioa eomiy sugey wi aigo o isumeaio oe icue usios om 4 o 4. ese oceues eome ecaes ago equey esu i aack e omiies a egeeaio eow e usio. ig. 22 sows A a aea aio

. 22 Aeooseio (A a aea ( iews o a aie wi a 4 o 4 a igo o, wic esue i symomaic egeeaie cages a 4. ig. 2 Aeooseio (A a aea ( iews o successu eisio usig wege cao cages.

gas o suc a case, wi symomaic egeeaie cage a 4 eow e usio. is case was eae wi a wege uma Ieoy usio Cage as sow i ig. 2. e o o ca e emoe, u is ca e a eesie oceue i e o is comeey coee wi oe. I is case, e o o was ike o e ew cosuc. We a eig o ese oceues i is seies. ee wee os o oowu. ou o ie wee ciicay successu a aciee usio. Aemig o use S we i is e oy uuse uma segme as a owe success ae.

Saisacoy ciica a usio esus wee aciee wi e uma Ieoy usio Cage wi eice scew iaio i a suy o 42 cosecuie aies. Aoug ua eas wee commo, ey a o ogem sigiicace. wo aies eeiece ee oo iuy i sie o 8 aies aig scaig om io uma sugey. Cage usio wi eice scew iaio as a ige a eece aiue ae a 2. aowe mm cages ae % owe usio success comae wi age cages. Ieoy usio usig caceous oe acke isie e cages wiou aiioa oe aeas o ae saisacoy usio success i aoiaey sie cages ae use. Sugeos sou o e a usio a a ee o ecomessio. Ieoy usio wi a sige C ecagua cage aeas o ae saisacoy usio success i aiioa oe ga is acke aou e cage iig e isc sace.

I , ll I COMICAIOS, IAS, A EI MAAGEME 6

Ez,2. C,

EEECES

. aiga W, eie A. e uma I Cage o I wi S eice scew sysem: 0yea esus o a oo a ug Amiisaio ciica ia. Sie 4:68688, 2004. 2. aiga W, eie A. Acieeme o oma sagia ae aigme usig a wege cao ie eioce oyme usio cage i eame o soyoisesis. Sie :866, 200. . aiga W, Seee A. A cao ie ima o ai ieoy uma usio: woyea cii ca esus i e is 26 aies. Sie 8:2062, . 4. aiga W, Seee A, Geige M. A cao ie ima o ai ieoy uma usio. Me caica esig. Sie 6:S2S282, . . aiga W, Seee A, ewis M, e a. uma ieoy usio usig e aiga I Cage o I a e S eice scew sysem: wo yea esus o a oo a ug Amiis aio IE ciica ia. Sie 2:4446, 2000. 6. Cammisa , Giai Sagai K, e a. Iciea uoomy i sie sugey. Sie 2:266 266, 2000. . Ceia A. Case eo: Acue aaegia ue o Sugice eae oacic co comessio. I aio Imag :4, . 8. oge G, ooey S, eie A, e a. Oucomes o 2 oseio uma ieoy usio wi e aiga I Cage a S eice scew iaio (i ess. . oge G, ooey S, eie A, e a. Oucomes o sigecage oseio uma ieoy usio wi e aiga I Cage a S eice scew iaio (i ess. 0. oge G, ooey S, eie A, e a. Oucomes o oseio uma ieoy usio usig mm wi aiga I Cage a S eice scew iaio (i ess. . ak C, aiga W, Coa . iaea iecoece aissimus osigueus maimus muscua cuaeous a o cosue o suascia iecios i uma sia sugey: A ecica oe. Sie 22:646, . 2. Gisei G, Wag C, aia , e a. Aace segme egeeaio i e uma sie. oe oi Sug Am 86:40, 2004. . amy M, Wise , agaa , e a. e asiio oe aoe a umosaca usio. Sie 2:82, 8. 4. asimoo , Sigeou K, Kaayama M, e a. Ciica esus o sigeee oseio uma i eoy usio usig e aiga I Cao cage ie wi a miue o oca moseie oe. Sie 2:28262, 2002. . oso omm . Eecs o oe wa o aceia ceaace. Sugey 8:20620, 8. 6. Kuma M, acquo , a . ogem oowu o ucioa oucomes a aiogaic cages a aace ees oowig uma sie usio o egeeaie isc isease. Eu Sie 0:0, 200. . ai , Ce , iu CC, e a. eaio ewee amiecomy a eeome o aace seg me isaiiy ae uma usio wi eice iaio. Sie 2:2222, 2004. 8. e A, oges E, awso EG, e a. Uecogie uoomy ae uma iscecomy: A eo o ou cases associae wi e use o ACO. Sie 26:8, 200. . ee CK. Acceeae egeeaio o e segme aace o a uma usio. Sie : , 88. 20. Miua Y, Imagama S, Yoa M, e a. Is oca oe iae as a souce o oe ga i oseio um a ieoy usio? Sie 28:28628, 200. 2. Moiai W, Geige ucioa oucomes o isumee oseio uma ieoy usio i acieuy US seiceme: A comaiso wi ooeaie maageme. Sie :2224, 200. 22. Moiai W, Sooa , osoe . Ae 2 cages eee wi isumee I? A coma iso o esus 2 ieoy cages i a miiay ouaio. Am Oo 2:4, 200. 2. Scauo AA, Gama SC, Yu W, e a. eioeaie comicaios o eae cyiica um a ieoy usio eices: Aeio esus oseio aoac. Sia iso ec 6:02 0, 200. 'CM MLE., ELEZ:E tE`M;; •'EETEEP !MEE, EE:E ELI, M: MEEE, -EME ELM ELM FEEM •;.Z.E.E1 CM, EME KELE

24. Scege , Smi A, Sceusee I. uma moio segme aoogy aace o oa couma, uma a umosaca usios. Sie 2:08, 6. 2. Someg , ooey S, eie A, e a. Comicaios a sugica cosieaios i oseio uma ieoy usio wi cao ie cages a Seee eice scews a aes. Oo aeics 26:004, 200. 26. ueg , a , yeg , e a. usio ae ae oseio uma ieoy usio wi ca o ie ima: yea oowu o aies. Eu Sie :882, 6. 2. Wag C, oma , iew K. ua eas secoay o oeaios o e uma sie. oe oi Sug Am 80:282, 8.

Ii • TIFI .1 ii• • i • iiiiiiiiiiiiiiiiii The Spine Journal 6 (2006) 421–427

Outcomes of L1–L2 posterior lumbar interbody fusion with the Lumbar I/F cage and the variable screw placement system: reporting unexpected poor fusion results at L1–L2 Guy R. Fogel, MD*, John S. Toohey, MD, Arvo Neidre, MD, John W. Brantigan, MD South Texas Orthopaedic and Spinal Surgery Associates, 9150 Huebner Road, San Antonio TX 78240, USA Received 1 July 2005; accepted 30 September 2005

Abstract BACKGROUND CONTEXT: Posterior lumbar interbody fusion (PLIF) was introduced 50 years ago. The Lumbar I/F cage (DePuy Spine, Raynham, MA) was designed to enhance PLIF results. PLIF with the Lumbar I/F cage and posterior Variable Screw Placement System (VSP) has increased the success of fusion to nearly 100% at the four lowest lumbar levels, L2–L3 through L5–S1. Less commonly, PLIF is indicated for the L1–L2 level. Clinical-results of Lumbar I/F cage fusion and VSP at L1–L2 have-not been reported. PURPOSE: The purpose of this study is to report the functional outcomes, fusion rate, and complications related to PLIF with Lumbar I/F cage and VSP of L1–L2 STUDY DESIGN/SETTING: The setting is a retrospective, single-arm cohort study of consecutive PLIF surgical patients at a single center. PATIENT SAMPLE: A review of 373 of 425 patients who underwent PLIF with Lumbar I/F cage and VSP from 1999 to 2002 identified 12 patients who had PLIF with Lumbar I/F cage and VSP at L1–L2. Mean follow-up was 31 months (range 12–65 months). OUTCOME MEASURES: Clinical success was determined with a modified Prolo score evaluating pain, function, medication usage and economic status. Fusion success, determined by evaluation of plain radiographs, was defined by continuous bone bridging the fusion area with no lucencies. METHODS: The 12 patients were evaluated for clinical success and/fusion success at last follow- up. These results were compared with the results of the 373 patients reviewed, and historical groups of the original Investigational Device Exemption study and the 10-year follow-up study. RESULTS: Previous surgery was reported by 10 of 12 patients, with an average symptom-free period of 3 years after previous fusion and before presentation with severe symptomatology necessi- tating further surgery at L1–L2. Seven patients had clinical success (59%), and five patients were clinically unsuccessful (41%). This included zero excellent, 2 of 12 (15%) good, 5 of 12 (42%) fair, and 5 of 12 (42%) poor results. Fusion was successful in seven (58%) and failed in five patients (42%). Three failed fusions were associated with L1–L2 subsidence. Two patients required further revision for non-union. CONCLUSIONS: In 12 patients with L1–L2 fusion, we report an unexpected high rate of failed fusion and poor clinical outcome. Ó 2006 Elsevier Inc. All rights reserved.

Keywords: PLIF; Interbody fusion orthosis; Cage; Pedicle screw; Degenerative disc diseases; Spinal fusion/instrumentation/ methods; Treatment outcome

Introduction Device status: The Lumbar I/F cage and the VSP Spinal Fixation Sys- tem are approved by the US Food and Drug Administration. Dr. Brantigan Posterior lumbar interbody fusion (PLIF) was intro- has a ﬁnancial interest in the Lumbar I/F cage. No ﬁnancial assistance was duced by Cloward [1–4] more than 50 years ago. Limita- provided from any source for this study. tions of PLIF included inadequate compression strength * Corresponding author. South Texas Orthopaedic and Spinal Surgery Associates, 9150 Huebner Road, Suite 350, San Antonio, TX 78240. Tel.: of allograft bone [5] and low fusion rates [6]. The Lumbar (210) 561-7234; fax: (210) 561-7240. I/F cage (DePuy Spine, Raynham, MA) was designed to E-mail address: [email protected] (G.R. Fogel) enhance posterior lumbar interbody fusion results by

1529-9430/06/$ – see front matter Ó 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.spinee.2005.09.011 422 G.R. Fogel et al. / The Spine Journal 6 (2006) 421–427 providing the mechanical strength to provide load-sharing ligament structures above the fusion level. At L1–L2, be- during fusion and to allow the best biological healing with cause of increased risk of neurological injury to the conus autologous bone graft [7]. medullaris, a more lateral approach was used, essentially In an Investigational Device Exemption (IDE) study of a transforaminal lumbar interbody fusion. A spreader is the Lumbar I/F cage, fusion success was reported in 176 placed in the disc space and used to distract the disc space of 178 patients (98.9%) [8]. Almost half of the study cases on each side to normal disc height or to the elastic limit of were two-level procedures. Less than 10% were three- or the annulus. A ‘‘working plate’’ is tightened down over the four-level procedures. The high fusion success was undi- pedicle screws on one side to hold the distraction. In con- minished over multiple fusion levels including 100% of trast to PLIF at the lower levels, it is not generally possible the three- and four-level fusions. Based on this study, the to use bilateral working plates. Next, a round reamer is used Lumbar I/F cage was approved by the Food and Drug Ad- to prepare a round hole in the disc space on the opposite ministration for treatment of degenerative disc disease at side. The goal of reaming and broaching is merely to flatten one or two contiguous levels from L2 to the sacrum, ac- the end plate, not to cut a deep channel. Therefore, for companied by posterolateral fusion and pedicle screw fixa- placement of a 9-mm cage, the disc space should be tion using the Variable Screw Placement (VSP) Spinal distracted to 9 mm. The 9-mm reamer should be used, System. followed by the 9-mm broaches. The ‘‘guided broach’’ The L1–L2 level was excluded from the study protocol follows the round channel and makes it square. The ‘‘final and from the Food and Drug Administration approval be- broach’’ follows the previous channel and squares the bot- cause of the low frequency of isolated degenerative disc tom most portion of the channel. The final broach should be disease at this level. The approval was limited to two levels held with fingertip force only and allowed to follow the pre- because of the small number of three- and four-level proce- vious channel. A ‘‘false channel’’ is created if the final dures in the study. Nevertheless, some patients develop broach is placed in a different direction than the guided symptomatic degenerative disc disease at the L1–L2 level. broach, thus creating an excessively large channel that will The purpose of this retrospective review is to report the predispose to a loose-fitting cage and pseudarthrosis. We functional outcomes, fusion rate, and complications related recommend if there is any question of the correct direction to PLIF with Lumbar I/F cage and VSP of L1–L2. of the final broach that the final broach merely be omitted.

Materials and methods Clinical study design Description of the Lumbar I/F cage From February 1999 to October 2003, we performed The Lumbar I/F cage is manufactured from poly-ether- PLIF using the Lumbar I/F cage with VSP pedicle screw ether ketone reinforced with ‘‘chopped’’ or short carbon fi- fixation on 425 patients in our practice. During the summer bers. The cage is filled with autologous bone and placed in of 2004, we conducted a retrospective chart review of all a disc space channel. This allows bone fusion from end patients. We had a minimum of 1 year of full clinical and plate to end plate across the bone graft within the cage. radiographic data on 377 of 425 patients (87%). Although The cage material is one-tenth the stiffness as compared 1 year is not long enough to define an exact rate of clinical with metal and minimizes stress shielding of the bone graft success, it is sufficient follow-up to identify complications [8]. Because the cages are radiolucent, bridging bone can and failed fusions. From this group we identified 12 pa- be accurately assessed by plain radiographs. tients who had received PLIF at the L1–L2 level. The mean follow-up was 31 months (range 12–65 months). All 425 operated patients had disabling back and/or radicular pain Description of the Variable Screw refractory to conservative management with moderate to Placement System (VSP) severe degenerative change, with instability and adjacent The VSP (DePuy Spine, Raynham, MA) includes pedi- segment deterioration based on flexion-extension radio- cle screws and plates. The screw has a cancellous thread graphs, magnetic resonance imaging (MRI), or discogram. ending in an integral nut, above which a machine thread In the subset of 12 L1–L2 patients, 10 patients had 25 prior with tapered and locking nuts allows the screw to be an- spinal operations (average 2.5). In the previously operated chored securely to the plate [8]. patients, there was an average symptom-free interval of 3 years (1–7 years), followed by onset of significant and se- Description of the operation vere symptoms correlating with imaging identifying L1– L2 as the source of the new symptoms. At the L1–L2 level, In the standard PLIF operation, the extent of the required nine had preoperative radiographic adjacent segment de- laminectomy should expose the entire space between the generation with or without stenosis, and three had herniated medial walls of the pedicles. The superior one-third of degenerative discs causing stenosis in the presence of de- the lamina should be preserved to maintain continuity of generative flexion-extension or rotational instability. G.R. Fogel et al. / The Spine Journal 6 (2006) 421–427 423 Of the 12 L1–L2 patients, there were three men and nine treated segments must be fused for the patient to be consid- women. Average age was 44 years (range 34 to 80). Insur- ered a fusion success. ance was workers’ compensation in four and private insurance in eight. Three patients had positive smoking history Sagittal plane alignment with an average 21-pack/year histories. The sagittal alignment is approximately 5 degrees ky- Clinical success was defined according to previous pub- phosis at the thoracolumbar junction [12]. After surgery, lished literature parameters used over many years and mod- progressive kyphosis may impair the final result. The radio- eled after an expanded Prolo Scale [8–11]. The 5-point graphic criteria of proximal junctional kyphosis was adap- Likert scales for pain, function, economic status, and med- ted from Yang and Chen [13] and defined as greater than 10 ication usage are added to a combined 4- and 20-point scale degrees progression in the proximal segment kyphotic an- as used for the IDE study of the Lumbar I/F cage [8].This gle from preoperative to last follow-up radiograph. Proxi- study defined a patient’s result as a clinical success when mal segment kyphosis was measured from T11–T12 to the final rating was excellent (17 to 20 points) or good L1–L2 disc spaces on the current radiographs. This mea- (13 to 16 points), or fair (9 to 12 points) with a minimum surement included two disc spaces and the T12 vertebral improvement of three points or more in the combined 20- body. The angle was measured by the method of Cobb point scale. [14]. Measurements were made on the preoperative and last follow-up lateral radiographs. Additional Cobb angle measurement was made of the L1–L2 disc space on preopera- Fusion status tive and last follow-up radiographs. Fusion status was recorded for each surgically treated segment at each follow-up interval. Fusion healing was defined by previously published literature parameters [8–10]. Results Fusion success was defined as radiographic evidence of Clinical success bone bridging the disc space within the cages, without evidence of lucency across the cage. If lucency was seen to Clinical success was achieved in 7 of 12 patients (59%). extend across the cage, the level was considered not fused. This included 2 (17%) Good and 5 (42%) Fair results. Clin- For patients undergoing multiple-level fusion, all surgically ical success was not achieved in the other 5 (42%) with

Table 1 Success parameters, prior and further surgeries Clinical Fusion # Prior Previous Further Patient success success Comment surgery operated levels surgery 1 Y Y Fused radiographically. 5 L3–S1 PLIF None Continued pain with limitations. 2 N Y Fused radiographically, T12–L1 ASD 2 L3–L5 PLIF None with instability. 3 N Y Fused radiographically at every level 2 L2–L4 PLIF L4–S1 PLIF with continued pain. 4 Y Y Healed fusion 1 L4–L5 PLIF None at time of exploration effusion. 5 Y Y Fused radiographically with mild pain 2 L2–L4 PLIF I.R. L1–L2, LI screw and limitations. loose, fusion healed 6 N N Reinjury 6 months 0 None None post-op. Recurrent pain. 7 N N Fused radiographically 5 L2–S1 PLIF None ASD T12–L1 continued pain. 8 Y N Pseudarthrosis twice at exploration. 3 L3–S1 PLIF Revision PLF BMP and Treated with re-instrumentation screws x1, IR BMP BMP x1 and BMP, then IR with BMP. 9 Y N Pseudarthrosis at exploration. 2 L2–L5 PLIF IR BMP PLF IR with BMP. 10 Y N Pseudarthrosis radiographically 2 L4–L5 PLIF None with L1–L2 subsidence. 11 N Y Fused radiographically with 0 None L2–L4 PLIF continued pain. 12 Y Y Fused radiographically, with 1 L4–L5 PLIF None asymptomatic L2–L3 stenosis. ASD5adjacent sgemental degeneration; BMP5bone morphogenic protein; IR5Implant removal of pedicle screws and plates; N5no; PLF5posterolat- eral fusion; Y5yes. 424 G.R. Fogel et al. / The Spine Journal 6 (2006) 421–427 Poor results. Both patients without previous surgery were loosening in two, and in one 34-year-old woman, a single clinically unsuccessful. Table 1 lists the clinical status of cage retropulsed 5 mm without clinical neurologic symp- these patients at ﬁnal follow-up. toms. This patient had an isolated L1–L2 PLIF, without previous surgery. One of the two cages retropulsed into the spinal canal, causing back pain without neurologic loss. Fusion success The device failure led to an unsuccessful fusion. The other Fusion success was achieved in 7 (59%) (Table 1 and patient without previous PLIF had L1–L2 and L4–S1 PLIF Fig. 1), and was unsuccessful in 5 (42%) patients. Fusion with successful fusion of both levels (Fig. 1). Table 1 corre- was successful in two of the three patients who smoked, lates the clinical status with fusion success. Figure 1 is an ex- and unsuccessful in one. The failures were not limited to ample of a healed fusion. Figure 2 is an example of a failed patients having a long fusion ending at L1. Of the ﬁve pa- fusion with subsidence of the cages causing increased ky- tients having L1–L2 fused with a nonfused adjacent L2–L3 phosis and lateral tilt through the disc space. Figure 3 dem- segment, two patients had fusion failure and three were un- onstrated computed tomographic (CT) evidence of nonunion successful clinically with poor Prolo scores. Failed fusion above a long fusion. Figure 4 shows CT evidence of non- was associated with subsidence in three, pedicle screw union of L1–L2, with the adjacent L2–L3 level never treated.

Fig. 1. (A) Patient 4 is a 37-year-old man with healed L1–L2 fusion anterior posterior radiograph 1 year after posterior lumbar interbody cage fusion L1–L2. Bone density is demonstrated within the fusion cage and spanning the disc space between the end plates. There is increased bone density suggesting com- pletion of interbody fusion healing. There is no lucency between the bone and the vertebral end plates. (B) Lateral radiograph 1 year after posterior lumbar interbody cage fusion L1–L2. Increased bone density is demonstrated across the fusion cage between the end plates, implying complete fusion. (C) Coronal computed tomographic scan through the interbody cages, with arrow demonstrating bridging bone through the cage spanning the disc space and connecting the end plates. (D) Sagittal computed tomographic scan through an interbody cage: arrow demonstrating increased bone density spanning the disc space continuous with the end plates. G.R. Fogel et al. / The Spine Journal 6 (2006) 421–427 425 Cages used in L1–L2 fusion Ten patients had two cages placed at the L1–L2 level. The other two patients had a single cage placed at L1–L2 (9Â9Â25 mm). The reason given for a single cage was the limited space in the disc space for two cages, and in one report, concern about retraction of the dura with the conus medullaris near-by. In the ﬁrst single-cage case, Hea- los (Depuy Spine), an osteoconductive matrix, was used to ﬁll the other side of the disc space. In the second case, a trans-foraminal approach was used to place a single cage obliquely across the disc space. In 11 patients, the cages were 9Â9Â25 mm length. One patient had two 9Â11Â25 mm cages. A total of 22 interbody cages at L1–L2 were used. All levels were supported with bilateral pedicle screws and VSP plates.

Complications In one patient, an epidural hematoma with acute cauda equina syndrome occurred immediately after L1–L2 surgery. The hematoma was drained immediately with complete resolution of all symptoms. Another had loosening of the pedicle screw ﬁxation, with subsidence and kyphosis across the L1–L2 segment. This patient is using an electro- magnetic bone stimulation device to encourage bone fusion healing.

Further surgery Five patients had further surgery after the L1–L2 fusion at an average of 18 months (range 14–26 months), as reported in Table 1.

Sagittal plane alignment There was no preoperative proximal segmental T11–L1 kyphosis above the L1–L2 fusion. There was no proximal segmental T11–L1 kyphosis at last follow-up. Two patients developed L1–L2 segmental kyphosis greater than 8 degrees associated with radiographic subsidence of one or both of the interbody cages through the L2 cephalic end plate. One patient with 3 degrees angulation did not have

= Fig. 2. (A) Patient 10 is a 45-year-old woman at 1 year after L1–L2 PLIF. Anterior-posterior radiograph 1 year after posterior lumbar interbody cage fusion L1–L2 shows a 16-degree lateral tilt deformity through the disc space, with incomplete or blocked areas of bone density within the fusion cage. (B) Lateral radiograph 1 year after posterior lumbar interbody cage fusion L1–L2. The arrow demonstrates subsidence of the disc space with intrusion of the interbody cage through the cephalic end plate of L2. There is a resultant 14-degree kyphosis at L1–L2. There is nonunion of the interbody fusion at L1–L2. (C) Sagittal helical thin-cut computed tomographic scan through an interbody cage at L1–L2. The arrow demonstrates subsidence of the cage into the L2 cephalic end plate. There is incomplete bridging of the bone density across the disc space. A lucency extends across the disc space, implying incomplete healing of the disc space interbody cage fusion. 426 G.R. Fogel et al. / The Spine Journal 6 (2006) 421–427

Fig. 4. (A) Patient 8, a 73-year-old woman with previous L3–S1 healed fusion, has nonunion of L1–L2. Coronal computed tomographic reconstruction reveals lucency surrounding both interbody cages at L1–L2. Fig. 3. (A) Patient 9, a 48-year-old woman with nonunion L1–L2 above (B) Sagittal computed tomographic reconstruction: superior arrow demon- previously healed L2–L5 posterior lumbar interbody fusion. Computed to- strates lucency around the cage, and inferior arrow shows subsidence into mography 12 months after L1–L2 implant removal and bone morphogenic L2 body. Metallic dot in the disc space is the cage position marker. protein augmentation of posterior-lateral fusion; computed tomographic coronal reconstruction of L1–L2: an arrow demonstrates a lucency running horizontally through the fusion mass within the interbody cage. (B) Com- puted tomographic sagittal reconstruction: arrow shows lucency around MRI findings with positive discogram. The rate of previous a cage with 2 mm retropulsion of the cage into the canal. surgery of this group is not significantly different than the original IDE Lumbar I/F cage study wherein there were 326 surgeries in 170 patients reported (average 1.9 surger- identified subsidence. The three patients with increased ies/patient) [8] (chi-square50.2662, p5.6059). Cobb angle measurement across the L1–L2 disc space did have failed fusions. Clinical success The Lumbar I/F cages IDE study reported clinical suc- Discussion cess in 79 of 91 (86.8%) [8]. In the larger clinical series from our group, clinical success was achieved in 285 of Outcome measures 345 (82%) of non-L1–L2 patients. When the clinical results of our larger clinical series are compared with the IDE At the time of the original Brantigan Lumbar I/F cage results using the chi-square statistic, clinical success rates IDE in 1990, there was little agreement regarding optimum are not statistically different (chi-square5.555, p5.4563). outcome measures in a spinal fusion study. To our knowl- When the clinical results of the L1–L2 group are compared edge, a thorough examination of the statistical validity of with the larger clinical series from our group using the chi- the Prolo Scale has not been undertaken. However, the Pro- square statistic, clinical success rates at L1–L2 are one-half lo Score has been widely accepted and is used in this report of the expected success at the lower lumbar levels of PLIF to maintain continuity with the original IDE and the (chi-square54.588, p5.0322). 10-year follow-up of the Lumbar I/F cages. We believe the Prolo Scale produces results that may be compared with Fusion success literature reports of outcomes over many years [8,9]. The Lumbar I/F cage IDE study reported fusion success Previous surgery in 176 of 178 (98.9%) [8]. In the larger clinical series from our group, fusion success was achieved in 321 of 326 non- Ten of the 12 patients had previous PLIF surgery (aver- L1–L2 patients (98%). When the fusion results of our larger age 2.5 surgeries/patient). In these 10 patients, there was series are compared with the IDE results using the chi- a 3-year average symptom-free time before presentation square statistic, the fusion rates are not statistically differ- with severe symptoms requiring surgical treatment at the ent (chi-square50.1414, p5.7069). When the fusion results adjacent L1–L2 level. In patients with prior surgery, diag- of the L1–L2 group are compared with the larger clinical nosis was based on progressive MRI changes over time series from our group using the chi-square statistic, L1– or by positive concordant pain reproduction on discogram. L2 fusion rates are 60% lower (chi-square564.9297, In patients without adjacent fusion, diagnosis was based on p!.0001). G.R. Fogel et al. / The Spine Journal 6 (2006) 421–427 427 Sagittal plane alignment [2] Cloward RB. Lesions of the intervertebral disks and their treatment by interbody fusion methods. The painful disk. Clin Orthop 1963;27: The development of junctional kyphosis above a short 51–77. segmental fusion for thoracolumbar scoliosis or fracture [3] Cloward RB. Spondylolisthesis: treatment by laminectomy and pos- management is not well understood. An increase in kyphosis terior interbody fusion. Clin Orthop 1981;154:74–82. [4] Cloward RB. Posterior lumbar interbody fusion updated. Clin Orthop may give a displeasing cosmetic result, and may result in 1985;193:16–9. revision of the fusion [15–17]. Yang and Chen found in five [5] Brantigan JW, Cunningham BW, Warden K, McAfee PC, Steffee AD. patients with preoperative junctional kyphosis of more than Compression strength of donor bone for posterior lumbar interbody 10 degrees, four of five developed increased junctional fusion. Spine 1993;18:1213–21. kyphosis [13]. In this series, no junctional kyphosis was iden- [6] Brantigan JW. Pseudarthrosis rate after allograft posterior lumbar interbody fusion with pedicle screw and plate fixation. Spine 1994;19: tified preoperatively, and no increase in kyphosis was found 1271–9; discussion 1280. at last follow-up. Isolated L1–L2 angulation was associated [7] Brantigan JW, Steffee AD. A carbon fiber implant to aid interbody with subsidence of the interbody cages in two failed fusions. lumbar fusion: two-year clinical results in the first 26 patients. Spine 1993;18:2106–7. Biomechanics [8] Brantigan JW, Steffee AD, Lewis ML, Quinn LM, Persenaire JM. Lumbar interbody fusion using the Brantigan I/F cage for posterior Deformity and trauma surgeons have long recommended lumbar interbody fusion and the variable pedicle screw placement that a fusion should not end at L1 because of an excessively system: two-year results from a Food and Drug Administration in- high rate of mechanical stress and failure at the thoracolum- vestigational device exemption clinical trial. Spine 2000;25: 1437–46. bar junction. Instead, two thoracic levels should be included [9] Brantigan JW, Neidre A, Toohey JS. The Lumbar I/F Cage for pos- in the instrumentation and fusion. We had expected that the terior lumbar interbody fusion with the Variable Screw Placement extra stability afforded by anterior cage placement with ped- System: 10-year results of a Food and Drug Administration clinical icle screw fixation would compensate for the added stresses trial. Spine J 2004;4:681–8. at this level. This expectation was not realized. The failures [10] Stromberg L, Toohey JS, Neidre A, Ramsey M, Brantigan JW. Com- plications and surgical considerations in posterior lumbar interbody were not limited to patients having a long fusion ending at fusion with carbon fiber interbody cages and Steffee pedicle screws L1. Of the five patients having L1–L2 fused with a nonfused and plates. Orthopedics 2003;26:1039–43. adjacent L2–L3 segment, two patients had fusion failure, [11] Prolo DJ, Oklund SA, Butcher M. Toward uniformity in evaluating and three were unsuccessful clinically with poor Prolo results of lumbar spine operations: a paradigm applied to posterior scores. None of the five has required further surgery, but lumbar interbody fusions. Spine 1986;11:601–6. [12] Bernhardt M, Bridwell KH. Segmental analysis of the sagittal plane revision has been recommended in both fusion failures. alignment of the normal thoracic and lumbar spines and thoracolumbar junction. Spine 1989;14:717–21. Conclusions [13] Yang SH, Chen PQ. Proximal kyphosis after short posterior fusion for thoracolumbar scoliosis. Clin Orthop 2003;411:152–8. In this small group of L1–L2 posterior lumbar interbody [14] Cobb J. Outline for the study of scoliosis. Instructional Course cage fusion with VSP patients, there was a high rate of Lectures, American Academy of Orthopedic Surgeons, 1948; failed fusion, and poor clinical success. The failures were 261–75. not limited to patients having a long fusion ending at L1. [15] Molz FJ, Kirkpatrick JS, Reza Moeini SM, Partin JI, Bidez MW. Effects of kyphosis and lordosis on the remaining lumbar verte- This study reports an unexpectedly high failure rate at bral levels within a thoracolumbar fusion: an experimental study L1–L2 but does not provide a solution for those patients of the multisegmental human spine. J South Orthop Assoc requiring treatment at this level. 1999;8:261–8. [16] Sink EL, Newton PO, Mubarak SJ, Wenger DR. Maintenance References of sagittal plane alignment after surgical correction of spinal deformity in patients with cerebral palsy. Spine 2003;28: [1] Cloward RB. The treatment of ruptured lumbar intervertebral discs by 1396–403. vertebral body fusion. I. Indications, operative technique, after care. [17] Vaccaro AR, Silber JS. Post-traumatic spinal deformity. Spine J Neurosurg 1953;10:154–68. 2001;26(24 Suppl):S111–8. ORIGINAL ARTICLE

Is One Cage Enough in Posterior Lumbar Interbody Fusion: A Comparison of Unilateral Single Cage Interbody Fusion to Bilateral Cages Guy R. Fogel, MD, John S. Toohey, MD, Arvo Neidre, MD, and John W. Brantigan, MD

osterior lumbar interbody fusion (PLIF) was intro- Abstract: Posterior lumbar interbody fusion (PLIF), as recom- Pduced by Cloward1–4 more than 50 years ago. mended with bilateral lumbar interbody cages and pedicle screw Limitations of early PLIF included inadequate compres- fixation, has increased the successful fusion rate to nearly 100%. sion strength of allograft bone and low fusion rates.5 Presently, a unilateral approach to the disc space with a variant Carbon fiber reinforced polymer (CFRP) rectangular of PLIF, the trans-foraminal interbody fusion is often used. cages were introduced to enhance PLIF results by There are few clinical studies of unilateral interbody fusion. The providing the mechanical strength to provide load- clinical and fusion results of unilateral interbody fusion are sharing during fusion and to allow the best biologic important as the usage of trans-foraminal interbody fusion healing with autologous bone graft.6,7 PLIF with CFRP procedure increases. This retrospective study of 26 consecutive cages also restores disc and neuro-foraminal height,8,9 patients treated with a unilateral cage asks whether fusion and segmental alignment including balance and lordo- healing and clinical outcome is comparable with that obtained sis.10–14 with bilateral cages. In this study, there were no pseudarthroses, PLIF with insertion of bilateral interbody cages has instrumentation failures, or significant subsidence at any of the been a standard treatment to predictably restore seg- single cage levels. Disc space height and foraminal height were mental alignment and balance, and obtain a successful restored by the surgery and maintained at last follow-up. Using fusion. However, occasionally only a unilateral cage may Prolo scores, 23/26 patients had clinical success (88%), and 3 be placed because of circumstances found at the time of were unsuccessful. Fusion was successful at all single cage fusion surgery such as limited size of the disc space, nerve root levels and overall in 23/26 (88%) reviewing all levels of fusion. anomalies that may block access to the disc space, and In conclusion, fusion and clinical success rates were not limited ability to retract nerve roots due to epidural diminished by the use of a unilateral interbody cage rather than scarring of previous surgery. Because unilateral interbody the recommended 2 cages. This retrospective comparative study fusion may be necessary in PLIF for the reasons stated is a Level III-2 Therapeutic Study investigating the results of above, the question is raised as to whether this will give unilateral PLIF with a single interbody cage compared with acceptable balance, strength, and fusion results. historical series with interbody cages. Recently a variation of the PLIF approach, the Key Words: PLIF, TLIF, interbody fusion orthosis, cage, trans-foraminal interbody fusion (TLIF) has allowed pedicle screw, degenerative disc disease, intervertebral disk unilateral placement of interbody cages. The TLIF displacement/pathology/radiography/*surgery, middle aged, approach may reduce operative time, blood loss, dural spinal fusion/*instrumentation/*methods, treatment outcome tears, and radiculopathy risks and yet provide comparable mechanical stability to the PLIF. Biomechanical (J Spinal Disord Tech 2007;20:60–65) studies of TLIF suggest a single interbody cage positioned in sagittal, coronal or oblique orientation provides excellent stability and stiffness.15–17 Clinical and fusion results reported with TLIF are comparable with PLIF results.18–22 In the TLIF reports, there is variety of type and number of implants, and type and location of bone 20,23–32 Received for publication January 31, 2006; accepted June 6, 2006. graft material. The rapid evolution and variety of From the South Texas Orthopaedic and Spinal Surgery Associates, 9150 implant and biologic technologies complicates the assess- Huebner Road, San Antonio, TX 78240. ment of the TLIF technique. Device status: The Lumbar interbody carbon fiber Cage and the VSP Spinal Fixation System are approved by the United States Food and This study asks if unilateral PLIF with a single cage Drug Administration. Dr Brantigan has a financial interest in the will give acceptable results comparable with bilateral Lumbar interbody carbon fiber Cage. No Financial assistance was implants and to recent TLIF studies with unilateral provided from any source for this study. interbody fusion. This study reports the clinical and Reprints: Dr Guy R. Fogel, MD, South Texas Orthopaedic and Spinal Surgery Associates, 9150 Huebner Road, San Antonio, TX 78240 fusion results of patients having PLIF with a unilateral (e-mail: [email protected]). approach to the disc space and compares the outcomes to Copyright r 2007 by Lippincott Williams & Wilkins historical bilateral PLIF series and recent TLIF studies.

60 J Spinal Disord Tech Volume 20, Number 1, February 2007 J Spinal Disord Tech Volume 20, Number 1, February 2007 Is One Cage Enough in Lumbar Interbody Fusion?

MATERIALS AND METHODS foraminal stenosis and radiculopathy), 3 symptomatic The Lumbar I/F cage (DePuy Spine, Raynham, spondylolisthesis refractory to conservative care, and 6 MA) is a CFRP implant made of poly-ether-ether ketone pseudarthroses. Two patients had significant degenerative reinforced with ‘‘chopped’’ or short carbon fibers. The scoliosis. Fourteen patients were revision cases with 24 VSP (DePuy Spine, Raynham, MA) includes pedicle prior spinal operations. In the previously operated screws and plates. The CFRP cage and the VSP System patients, there was an average interval between the 2 have been previously described.33 surgeries of 42 months (7 mo to 7 y). There were 10 men From February 1999 to October 2003, we per- and 16 women. The average age was 50 years (range 34 to formed PLIF using the CFRP cages and VSP instrumen- 80 y). The insurance was workers compensation in 5 and tation on 425 patients in our practice. Surgical indications private insurance in 21. Nine patients had a positive included disabling back and/or radicular pain refractory smoking history with an average 19-pack/year history. to conservative management with moderate to severe Clinical success was defined according to previous degenerative change, with instability and adjacent seg- published literature parameters used over many years and ment deterioration based on magnetic resonance imaging modeled after an expanded Prolo scale.33–36 The 5-point or discogram. During the summer of 2004, we conducted Likert scales for pain, function, economic status, and a retrospective chart review of all patients. Follow-up was medication usage are added to a combined 4-point and achieved on 377 of 425 patients (87%) including 1-year 20-point scale.33 This study defined a patient’s result as a clinical and radiographic evaluation. Although 1 year is clinical success when the final rating was excellent (17 to not long enough to define an exact rate of clinical success, 20 points), or good (13 to 16 points), or fair (9 to 12 it is sufficient follow-up to identify complications and points) if the patient achieved a minimum improvement failed fusions. of 3 points or more in the combined 20-point scale. From this large group, 26 consecutive patients were Fusion success within the interbody space was identified who had PLIF at one or more levels with a defined by previous published parameters of radiographic unilateral CFRP cage. Each of these patients had fusion assessment of Brantigan and Steffee.7 As recom- identical treatment to the large group including iliac crest mended by Blount et al,37 2 observers evaluated the autograft, posterolateral fusion, and bilateral pedicle fusion from Ferguson anterior-posterior (parallel to the screw instrumentation. The only surgical variable was vertebral endplates), and lateral radiographs. A level was that a unilateral single CFRP cage was used rather than regarded as fused if there was radiographic evidence of bilateral cages. The minimum follow-up was 24 months bone bridging the disc space without lucency. If lucency (mean 36 mo, range 24 to 65 mo). Demographic data were was seen to extend across the cage, the level was recorded including age, sex, neurologic function, number considered not fused. For patients undergoing multiple- and type of prior back surgeries, and diagnosis at each level fusion, all surgically treated segments must be fused lumbar level. Patients were examined and data recorded for the patient to be considered a fusion success. Figure 1 before surgery, at surgery, and at postoperative intervals shows a series of anterior-posterior Ferguson radiographs after surgery. Evaluations included ratings of pain, demonstrating progressive solid fusion. Figure 2 shows function, economic status, and medication. Patient the features of a pseudarthrosis with discontinuity in the history, sensory and motor function, tension signs, bone graft and lucency adjacent to the cage. reflexes, and radiographic findings were recorded at each Statistical analysis was carried out using the w2 follow-up interval. Additional clinical data were recorded analysis. Demographics, clinical and fusion success were including surgery levels, operative difficulties, complica- compared with the larger reviewed patient group, the IDE tions, fusion status, and reoperation or revision. cage study,33 and the available TLIF studies with a Indications for the surgery were 17 recurrent disc minimum 2-year follow-up20,21,29,38 using w2 analysis. disease (recurrent herniation or disc space collapse with Statistical significance was defined as P<0.05.

FIGURE 1. Anterior-posterior Ferguson radiographs of L4-5 taken at interval demonstrate increases in bone density within the single cage and around the cage. The contrasting radiographic densities of bone and carbon cage strut (radiodense dot in the strut) allow a comparison of increased bone density with healing of the fusion. There is no lucency between the bone and the vertebral endplates. There was fusion success at 1 year with maintained maturity through 4 years. r 2007 Lippincott Williams & Wilkins 61 Fogel et al J Spinal Disord Tech Volume 20, Number 1, February 2007

TABLE 2. Correlation of Clinical to Fusion Success Clinical Patient Success Fusion No. Prolo Score Success Final Comment 1 E Y Healed fusion, with ASD L2-4 2 F Y Pain significantly improved clinical success 3 G Y Healed fusion, with previous infection 4 E Y Healed fusion 12 mo after IR BMP 5 F N ASD L1-2 6 F Y Healed fusion with clinical success 7 G Y ASD L2-3 FIGURE 2. L2-4 fusion, L3-4 single cage progresses to 8 G Y ASD L1-2 appearance of lucencies implying pseudarthrosis with non- 9 P Y Healed fusion, groin pain not continuous bone across the single cage. This level was found relieved to be healed at time of fusion exploration. 10 P Y ASD L1-2 constant pain 11 G Y Healed Fusion, ASD T12-L1 12 F Y Healed fusion, clinical success RESULTS 13 F Y Healed fusion, with clinical success A unilateral cage was used for the indications of 14 F Y Healed fusion, ASD L3-5 continued pain previous surgery with limited access to one side of the disc 15 G Y ASD L2-3 space in 12 (46%), not enough space in the disc space for 16 F Y Healed fusion, clinical success another cage in 11 (42%), and nerve root anomaly with 17 F Y ASD L3-4 low position of the exiting nerve root in 3 (12%). 18 G Y ASD L4-5, continued pain 19 G Y Healed Fusion, with clinical Unilateral cages were placed at thirty lumbar levels in success. 26 patients. There were 8 single-level fusions with a 20 F Y Healed fusion, clinical success unilateral cage; 10 with a unilateral cage at a single level 21 F Y Healed fusion, with clinical success in a multilevel PLIF, and 8 with multilevel unilateral 22 E Y Healed fusion, with clinical success cages in a multilevel PLIF. Nerve root anomalies 23 P N L3-4 pseudarthrosis in long fusion, chronic pain prevented cage placement at L4-5 and L5-S1 only. Not 24 G Y Healed fusion, with clinical success enough space for a second cage occurred at the higher 25 F N Healed fusion 16 mo after 1 cage lumbar levels but not at L5-S1. Previous surgery with removal and BMP adhesions prohibited cage placement most frequently at 26 G Y Hip groin pain not relieved by L3-4 but was significant at every level. Details of the fusion PLIF surgery are shown in Table 1. ASD indicates adjacent segment deterioration; BMP, bone morphogenic Clinical success was achieved in 23 patients (88%). protein; E, excellent; F, fair; G, good; IR, implant removal; P, poor; N, no; Y, yes. This included 3 excellent (12%), 9 good (35%), and 11 fair (42%) results. Clinical success was not achieved in 3 with poor results (12%). Table 2 correlates the clinical a fusion success at three years follow-up. There were no and fusion success of these patients. other complications such as dural tear, infection, or A retropulsed cage complication occurred in 1 neurologic injury. patient, at the L2-3 level with 2 cages in place. A cage At the final radiographic evaluation, fusion success was found to have retropulsed into the canal in the third was achieved in all 30 unilateral-cage levels. There were month postoperatively. This was identified and watched. no pseudarthroses, instrumentation failures, or significant At 8 months, the patient was complaining of back pain subsidence at any of the unilateral cage levels. Disc space and hip pain on the same side as the mal-positioned cage. height and foraminal height were restored by the surgery The cage was removed and the pedicle screw instrumen- and maintained at last follow-up. However, there were 3 tation removed, and the posterior-lateral fusion was patients with single level, 2 cage radiographic pseudar- augmented with bone morphogenic protein. This case was throsis. Fusion success was then 23/26 (88%) for all

TABLE 1. Details of Current Fusion Levels Lumbar Total Each 2 Cage Levels 1 Cage Levels No Cage Levels Patients Level Level No. (%) No. (%) Levels No. (%) Treated No. (%) L1-2 2 1 (33) 1 (33) 1 (33) 1 Level 6 (23) L2-3 10 5 (45) 5 (45) 1 (10) 2 Level 11 (42) L3-4 14 2 (13) 12 (75) 2 (13) 3 Level 4 (15) L4-5 14 8 (57) 6 (43) 0 (0) 4 Level 4 (15) L5-S1 11 5 (42) 6 (50) 1 (8) 5 Level 1 (4) Total 51 21 30 5 26

62 r 2007 Lippincott Williams & Wilkins J Spinal Disord Tech Volume 20, Number 1, February 2007 Is One Cage Enough in Lumbar Interbody Fusion? patients. The first pseudarthrosis was the L2-3 retro- unilateral cage had equal fusion and clinical success as pulsed cage removed at 8 months and subsequently fused those having 2 cages. The fusion and clinical results of on radiograph and computed tomography (CT) scan. The TLIF with unilateral interbody support and bone grafting second was an L2-3 level persistent pseudarthrosis from were not significantly different by w2 analysis from this an initial L2-S1 pseudarthrosis reconstruction that was unilateral cage series. explored and treated with pedicle screw implant removal During the design of the CFRP cages, mechanical and bone morphogenic protein augmentation of poster- testing assumed use of bilateral cage placement and olateral fusion and subsequently was healed at last bilateral pedicle screw fixation.14,39 Biomechanical fatigue follow-up with x-rays and CT scan. The third patient strength of 2 cages had a 2-fold safety factor over the had a pseudarthrosis level reported at L3-4 on CT in a maximal loads of daily living.40 The single cage fatigue L1-5 fusion for degenerative scoliosis, and did not strength had a small or nonexistent safety factor over agree with fusion exploration, accepting a fair clinical the maximal loads of daily living.40 Closkey et al41 in an result. Table 2 correlated the clinical status with fusion in vitro analysis found >30% of the vertebral endplate success. surface was required for load transmission across structural interbody grafts. Recent published mechanical data related to single interbody cage in TLIF found no DISCUSSION statistical difference in initial stability and stiffness with a The goal of lumbar interbody fusion is to relieve single cage compared with 2 cages.15–17 Ames et al15 in pain caused by neurologic compression and achievement human cadavers, found no significant difference in of a stable surgical construct. Compared with poster- motion between PLIF with 2 allograft bone spacers and olateral instrumented fusion, adding an interbody fusion TLIF with a single spacer as long as there were bilateral produces a significantly stiffer construct that protects the pedicle screws. Chen et al16 in a porcine model found posterior instrumentation from failure, and provides a unilateral cage and unilateral pedicle screws had similar circumferential fusion mass with an increased rate of stability to the unilateral cage with bilateral pedicle successful fusion.7,23,33 Both the PLIF and the TLIF screws. Kettler et al17 in a human cadaver found a approach offer circumferential fusion from a posterior unilateral cage is as stable as bilateral cages. Position of approach. the graft did not change the stiffness or stability.15,42 The patients in this unilateral cage series presented Ames et al15 varied the position of the unilateral graft with complex surgical problems that precluded the usage between the anterior and middle column without of the recommended bilateral interbody fusion cages. diminution of the stability or stiffness of the construct. Comparison of this series with the historical investiga- Harris et al42 in a human cadaver study placed an tional device exemption clinical trial (IDE) of Brantigan oblique single CFRP cage, and found the addition of et al,33 showed no significant difference in the indications bilateral pedicle screws matched the flexibility of an intact for surgery but there was a significantly higher rate motion segment. Heth et al43 compared anterior and of previous surgery in this group (w2 = 6.5202, P transverse placement of threaded cylindrical cages and value = 0.0107). Comparison of the clinical and fusion found no difference in stability with transverse positioned success showed no significant difference between this cages. Wang et al44 in human cadavers compared series and the IDE study. Despite the complexity of sagittal versus oblique placement of cylindrical cages revision surgery, the outcome of the unilateral cage and found any differences between cage position was patients was as successful as the IDE series. normalized with bilateral pedicle screw fixation. Several With the commonality of unilateral interbody key points are made by these studies. A unilateral fusion with bilateral pedicle screws and posterolateral cage is as stable and stiff as bilateral cages. Cage position graft, the comparison to reported TLIF studies is does not decrease the stability of the construct. All appropriate. Lowe et al38 reported 90% fusion rate and constructs are improved by bilateral pedicle screw 80% clinical success in TLIF with 2 titanium-mesh cages. fixation. Hee et al21 reported 96% fusion rate with the same TLIF Bone grafting of the available surface area of the procedure augmented with autologous growth factors. disc space is important for fusion success. Prolo36 found McAfee et al19 in 120 spondylolisthesis patients treated successful fusion filled 77% of the available disc space with TLIF with a unilateral CFRP cage reported a 97.5% with bone. A clinical study using CT scan to demonstrate fusion rate. Potter et al29 reported 93% fusion rate and disc removal showed more than 56% of the cross- 79% clinical success with interbody structural allograft or sectional area of the endplate could be cleared from a bioabsorbable spacers. Hackenberg et al20 reported 89% unilateral TLIF approach.45 The surface area of exposed fusion rate and significant improvement in Oswestry bone graft in a single CFRP cage is 138 mm2 and the disability index in TLIF with a unilateral CFRP. surface area of a typical L5 lumbar endplate is 1259 mm2. Molinari reported 19 active-duty military personnel were A single CFRP cage will fill only 10% of the endplate.40 treated with 2 CFRP cages and 16 with a unilateral cage. Additional bone grafting to fill all available surface area is When only a single cage was used, bone graft was inserted recommended. Placement of additional bone graft around from a bilateral approach.27 Molinari et al27 reported that the single cage may account for the undiminished high the results were generally good, and that patients having a rate of fusion success in this series. r 2007 Lippincott Williams & Wilkins 63 Fogel et al J Spinal Disord Tech Volume 20, Number 1, February 2007

A shortcoming of this study is that a small adult isthmic spondylolisthesis. J Spinal Disord Tech. 2003;16: population will result in poor specificity or underestima- 469–476. tion of the actual pseudarthrosis rate. To improve the 13. Brantigan JW, Neidre A. Achievement of normal sagittal plane alignment using a wedged carbon fiber reinforced polymer statistical power of the study would require a large fusion cage in treatment of spondylolisthesis. Spine J. 2003;3: number of patients that may be impractical. This study 86–196. demonstrates successful fusion and clinical results when 14. Brantigan JW, Steffee AD, Geiger JM. A carbon fiber implant to the patient selection criteria are expanded to include aid interbody lumbar fusion. Mechanical testing. Spine. complex revision surgical problems. The results of this 1991;16(suppl 6):S277–S282. 15. Ames CP, Acosta FL Jr, Chi J, et al. Biomechanical comparison of study support the use of a unilateral interbody cage posterior lumbar interbody fusion and transforaminal lumbar combined with bilateral posteriolateral fusion and pedicle interbody fusion performed at 1 and 2 levels. Spine. 2005;30: screws. A prospective randomized trial comparing uni- E562–E566. lateral with bilateral CFRP cages is warranted. The 16. Chen HH, Cheung HH, Wang WK, et al. Biomechanical analysis of authors emphasize that when a fusion is done with an unilateral fixation with interbody cages. Spine. 2005;30:E92–E96. 17. Kettler A, Schmoelz W, Kast E, et al. In vitro stabilizing effect of a unilateral approach to the disc space, it is particularly transforaminal compared with two posterior lumbar interbody important for the surgeon to pack additional bone in the fusion cages. Spine. 2005;30:E665–E670. disc space around the cage. 18. Lowe TG, Tahernia AD. Unilateral transforaminal posterior lumbar interbody fusion. Clin Orthop Relat Res. 2002;64–72. 19. McAfee PC, Devine JG, Chaput CD, et al. The indications for CONCLUSIONS interbody fusion cages in the treatment of spondylolisthesis: analysis Fusion and clinical success rates are not diminished of 120 cases. Spine. 2005;30:S60–S65. by the use of a unilateral CFRP cage rather than the 20. Hackenberg L, Halm H, Bullmann V, et al. Transforaminal lumbar recommended 2 cages. Clinical success with the unilateral interbody fusion: a safe technique with satisfactory three to five year single cages was not statistically different from the clinical results. Eur Spine J. 2005;14:551–558. 21. Hee HT, Majd ME, Holt RT, et al. Do autologous growth factors success rates of the IDE study. Fusion success was enhance transforaminal lumbar interbody fusion? Eur Spine J. achieved in 100% of 30 single cage levels and 23 of 26 2003;12:400–407. patients (89%) at all fusion levels. Mechanical failure did 22. Zhao J, Wang X, Hou T, et al. One versus two BAK fusion cages in not occur with the single cage. A single cage with bilateral posterior lumbar interbody fusion to L4-L5 degenerative spondylolisthesis: a randomized, controlled prospective study in 25 patients pedicle screws provides adequate alignment, balance, and with minimum two-year follow-up. Spine. 2002;27:2753–2757. mechanical stability, and allows the maximal amount of 23. Wang JC, Mummaneni PV, Haid RW. Current treatment strategies autologous graft to fill the disc space. for the painful lumbar motion segment: posterolateral fusion versus interbody fusion. Spine. 2005;30(suppl 16):S33–S43. REFERENCES 24. Coe JD, Vaccaro AR. Instrumented transforaminal lumbar inter- 1. Cloward RB. The treatment of ruptured lumbar intervertebral discs body fusion with bioresorbable polymer implants and iliac crest by vertebral body fusion. I. Indications, operative technique, after autograft. Spine. 2005;30(suppl 17):S76–S83. care. J Neurosurg. 1953;10:154–168. 25. Harms J, Jeszenszky D, Stolze D. True spondylolisthesis reduction 2. Cloward RB. Lesions of the intervertebral disks and their treatment and more segmental fusion in spondylolisthesis. In: Bridwell by interbody fusion methods. The painful disk. Clin Orthop. KH, DeWald RL, eds. Textbook of Spinal Surgery. Philadelphia: 1963;27:51–77. Lippincott-Raven; 1997:1337–1347. 3. Cloward RB. Spondylolisthesis: treatment by laminectomy and 26. Humphreys SC, Hodges SD, Patwardhan AG, et al. Comparison of posterior interbody fusion. Clin Orthop. 1981;74–82. posterior and transforaminal approaches to lumbar interbody 4. Cloward RB. Posterior lumbar interbody fusion updated. Clin fusion. Spine. 2001;26:567–571. Orthop. 1985;16–19. 27. Molinari RW, Sloboda J, Johnstone FL. Are 2 cages needed with 5. Brantigan JW, Cunningham BW, Warden K, et al. Compression instrumented PLIF? A comparison of 1 versus 2 interbody cages strength of donor bone for posterior lumbar interbody fusion. Spine. in a military population. Am J Orthop. 2003;32:337–343; Discussion 1993;18:1213–1221. 343. 6. Brantigan JW. Pseudarthrosis rate after allograft posterior lumbar 28. Mummaneni PV, Rodts GE Jr. The mini-open transforaminal interbody fusion with pedicle screw and plate fixation. Spine. lumbar interbody fusion. Neurosurgery. 2005;57(suppl 4):256-261; 1994;19:1271–1279; Discussion 1280. Discussion 256–261. 7. Brantigan JW, Steffee AD. A carbon fiber implant to aid interbody 29. Potter BK, Freedman BA, Verwiebe EG, et al. Transforaminal lumbar fusion. Two-year clinical results in the first 26 patients. lumbar interbody fusion: clinical and radiographic results and Spine. 1993;18:2106–2107. complications in 100 consecutive patients. J Spinal Disord Tech. 8. Tullberg T, Brandt B, Rydberg J, et al. Fusion rate after posterior 2005;18:337–346. lumbar interbody fusion with carbon fiber implant: 1-year follow-up 30. Rivet DJ, Jeck D, Brennan J, et al. Clinical outcomes and of 51 patients. Eur Spine J. 1996;5:178–182. complications associated with pedicle screw fixation-augmented 9. Hashimoto T, Shigenobu K, Kanayama M, et al. Clinical results lumbar interbody fusion. J Neurosurg Spine. 2004;1:261–266. of single-level posterior lumbar interbody fusion using the Brantigan 31. Salehi SA, Tawk R, Ganju A, et al. Transforaminal lumbar I/F carbon cage filled with a mixture of local morselized bone interbody fusion: surgical technique and results in 24 patients. and bioactive ceramic granules. Spine. 2002;27:258–262. Neurosurgery. 2004;54:368–374; Discussion 374. 10. Sears W. Posterior lumbar interbody fusion for lytic spondylolisth- 32. Schwender JD, Holly LT, Rouben DP, et al. Minimally invasive esis: restoration of sagittal balance using insert-and-rotate interbody transforaminal lumbar interbody fusion (TLIF): technical feasibility spacers. Spine J. 2005;5:161–169. and initial results. J Spinal Disord Tech. 2005;18(suppl):S1–S6. 11. Molinari RW, Bridwell KH, Lenke LG, et al. Anterior column 33. Brantigan JW, Steffee AD, Lewis ML, et al. Lumbar interbody support in surgery for high-grade, isthmic spondylolisthesis. Clin fusion using the Brantigan I/F cage for posterior lumbar interbody Orthop Relat Res. 2002;109–120. fusion and the variable pedicle screw placement system: two-year 12. Kwon BK, Berta S, Daffner SD, et al. Radiographic analysis of results from a Food and Drug Administration investigational device transforaminal lumbar interbody fusion for the treatment of exemption clinical trial. Spine. 2000;25:1437–1446.

64 r 2007 Lippincott Williams & Wilkins J Spinal Disord Tech Volume 20, Number 1, February 2007 Is One Cage Enough in Lumbar Interbody Fusion?

34. Brantigan JW, Neidre A, Toohey JS. The lumbar I/F cage for 40. Serhan H. Presentation of mechanical testing of the Brantigan cage. posterior lumbar interbody fusion with the Variable Screw Place- In: The Orthopaedics and Rehabilitation Devices Advisory Panel ment System: 10-year results of a food and drug administration Meeting of the Department of Health and Human Services, Public clinical trial. Spine J. 2004;4:681–688. Health Service, Food and Drug Administration. Maryland: Bethesda; 35. Stromberg L, Toohey JS, Neidre A, et al. Complications and 1997. surgical considerations in posterior lumbar interbody fusion with 41. Closkey RF, Parsons JR, Lee CK, et al. Mechanics of interbody carbon fiber interbody cages and Steffee pedicle screws and plates. spinal fusion. Analysis of critical bone graft area. Spine. 1993;18: Orthopedics. 2003;26:1039–1043. 1011–1015. 36. Prolo DJ, Oklund SA, Butcher M. Toward uniformity in evaluating 42. Harris BM, Hilibrand AS, Savas PE, et al. Transforaminal lumbar results of lumbar spine operations. A paradigm applied to posterior interbody fusion: the effect of various instrumentation techniques lumbar interbody fusions. Spine. 1986;11:601–606. on the flexibility of the lumbar spine. Spine. 2004;29: 37. Blount KJ, Krompinger WJ, Maljanian R, et al. Moving toward a E65–E70. standard for spinal fusion outcomes assessment. J Spinal Disord 43. Heth JA, Hitchon PW, Goel VK, et al. A biomechanical comparison Tech. 2002;15:16–23. between anterior and transverse interbody fusion cages. Spine. 38. Lowe TG, Tahernia AD, O’Brien MF, et al. Unilateral trans- 2001;26:E261–E267. foraminal posterior lumbar interbody fusion (TLIF): indications, 44. Wang ST, Goel VK, Fu CY, et al. Posterior instrumentation reduces technique, and 2-year results. J Spinal Disord Tech. 2002;15:31–38. differences in spine stability as a result of different cage orientations: 39. Brantigan JW, McAfee PC, Cunningham BW, et al. Interbody an in vitro study. Spine. 2005;30:62–67. lumbar fusion using a carbon fiber cage implant versus allograft 45. Javernick MA, Kuklo TR, Polly DW Jr. Transforaminal lumbar bone. An investigational study in the Spanish goat. Spine. 1994;19: interbody fusion: unilateral versus bilateral disk removal—an in 1436–1444. vivo study. Am J Orthop. 2003;32:344–348; Discussion 348.

r 2007 Lippincott Williams & Wilkins 65 Journal of Surgical Orthopaedic CASE REPORTS Advances

Repair of Pars Interarticularis Defect With a Modiﬁed Cable-Screw Construct

Gordon R. Bozarth, MD, Guy R. Fogel, MD, John S. Toohey, MD, and Arvo Neidre, MD

Operative treatment of symptomatic spondylolysis is not common. Multiple surgical techniques have been described for direct repairs of the pars defects. Reported success rates are high, although few reports describe successful return to sports in athletes. The purpose of the study was to assess the outcome after bone grafting and fixation of pars interarticularis defects utilizing a modification of the previously described techniques of Scott and of Songer. A retrospective single-arm cohort study was performed at a single center. This article reports on three athletes with symptomatic spondylolysis or grade I spondylolisthesis unresponsive to conservative management who were treated with bone grafting andascrew-cablerepair.Theoutcomemeasurewas the return to sports activities. A retrospective chart and radiographic analysis was conducted on three athletes. Patients were assessed for return to sports, clinical evidence of return to functional activities, and radiographic evidence of healing of the pars defects. All three patients proceeded to radiographic and clinical success. All patients reported resolution of their preoperative pain and return to sports. One patient did require occasional anti-inflammatory drugs for episodic low back pain. The use of this modified cable-screw technique for symptomatic spondylolysis provided excellent clinical, radiographic, and functional results in this small cohort. (Journal of Surgical Orthopaedic Advances 16(2):0–0, 2007)

Key words: pars defect, spondylolisthesis, spondylolysis

the pars defect. Direct repair has the advantage of main- Spondylolysis and grade 1 spondylolisthesis are comm- tenance of the motion segment and compression across on defects found in the general population, with a reported the bone-grafted pars defect. We have modified the Scott incidence of 6% (1). While many remain asymptomatic technique to pass a cable around a pedicle screw on each or heal with bracing or symptomatic management, a side, rather than the transverse process, and modified the small subgroup of these patients will continue to have significant functional limitations related to their sport (2). Songer technique to pass the cable around the base of the Symptomatic spondylolysis unresponsive to conservative spinous process and not beneath the lamina. We report on management may be an indication for operative inter- this modification for direct repair of symptomatic spondy- vention. While lumbosacral fusion has remained the gold lolysis. standard, direct repair of the pars defect has been shown to be effective in this subset (1, 3Ð9). Multiple surgical tech- Surgical Technique niques have been described in an attempt to directly repair The patient was positioned on a flexion frame in the From South Texas Orthopaedic and Spinal Surgery Associates, San prone position. Posterior laminar dissection was perform- Antonio, TX. Address correspondence to: Guy R. Fogel, South Texas Orthopaedic and Spinal Surgery Associates, 9150 Huebner Road, San ed over the involved segment and a localization radio- Antonio TX 78240; e-mail: [email protected]. graph was obtained. Care was taken not to violate the No financial assistance was provided from any source for this study. facet capsule. The defect in the pars was localized bilat- Received for publication October 21, 2006; accepted for publication November 22, 2006. erally. Soft tissue and callus was debrided from each For information on prices and availability of reprints call 410-494- defect until bleeding bone was evident. Each pars defect 4994 X226. 1548-825X/07/1602-0000$22.00/0 was then bone grafted with iliac crest autograft. Pedicle Copyright  2007 by the Southern Orthopaedic Association screws were inserted within the pedicle at the same level

VOLUME 16, NUMBER 2, SUMMER 2007 1 FIGURE 1 Surgical technique of modiﬁed Songer repair.

as the laminar pars defect, utilizing a standard tech- stiffness of the low back. CT scan at 6 months after nique. A 1-mm AcroMed Songer cable was then passed surgery showed one pars to be healed and one incom- around the base of the spinous process and the ipsilateral pletely healed. At 1 year, the patient was asymptomatic pedicle screw from both sides at the involved level. The and performing his usual soccer and conditioning program cables were then simultaneously tensioning and finally without limitations. CT scan at 1 year showed healing of crimped. Figure 1 is a graphic representation of the proce- both pars defects. dure. Patient 2 was a male 16-year-old baseball player with The surgical technique begins with debridement of the bilateral L5 pars defects. He was first diagnosed at age pars defect, then bone grafting of the pars defects with 14. He was treated with a brace and when pain free iliac crest autograft. Fixation is established with pedicle began gradual return to activities with physical therapy. screw placement into each L5 pedicle, then a 1-mm cable He improved with conservative treatment and returned is woven around the spinous process and both pedicle to full activity. He played school football, baseball, and screws are tensioned to compress the pars defect bone other sports. He did return to playing high school baseball grafts. for the entire season. And then several weeks before his return visit, he began to have shooting pain to his left Case Reports buttock following a diving fall to the ground. Physical exam showed mildly restrictive spine motion. Straight Patient 1 was a male 17-year-old soccer player with leg raising was negative. Leg lengths were equal. Plain a 2-year history of back pain. His initial injury occurred x-rays of the lumbar spine showed a right-sided pars while playing soccer. He complained of episodic signif- defect and the left pars appeared intact. CT scan showed icant low back pain that would limit his school atten- bilateral L5 pars defects. A lumbar MRI scan showed dance and he could not play soccer. Running became Schmorls nodes of several lumbar segments. Bone scan more painful, and basically any quick motion could be showed uptake of the left L5 pars defect and not the painful. He was treated with bracing, and had several right. He had bilateral repair of the L5 pars defect. At rounds of physical therapy consisting of stretching and 6 weeks, he was able to swim and jog. At 12 weeks, x- truncal strengthening. Physical exam revealed negative rays showed definite but incomplete healing of the pars straight leg raising and normal neurologic examination of the lower extremities. Plain x-rays, including obliques, defect bilaterally. He increased the amount and type of his showed pars defect of L5 bilaterally. A lumbar MRI scan aerobic activity. At 5 months, he was playing basketball and doing some running. Radiographs showed one side AQ1 revealed prominent pars defects with edema of the adjacent bone. A žtechnetium-99 bone scan showed increased completely healed and the other side filling in. He began uptake at the L5 level. The L5ÐS1 disc showed moderate a full-time running program. At 8 months, he was jogging dehydration. A discogram was done and was negative at 1 mile. He continues to have some aching in his back. X- L4Ð5 and L5ÐS1. Surgical repair of bilateral L5 pars ray showed one pars defect completely healed and the defects was done. The patient was braced with a lumbar opposite shows progressive healing. At 10 months, he corset for 6 weeks. He returned to jogging at 6 weeks and was back to full basketball sports and running without returned to soccer at 3 months. At 6 months, he continued discomfort. Radiograph showed apparent full healing of to play soccer; however, he complained of intermittent both of the pars defects.

2 JOURNAL OF SURGICAL ORTHOPAEDIC ADVANCES A BC

D E

G H

FIGURE 2 A–C, lateral and oblique radiographs of L5 pars fractures with minimal slip. A, Lateral view suggests an L5 pars defect. B and C, Obliques conﬁrm the pars defects bilaterally at L5. D –F, MRI and technetium-99 bone scan imaging of L5 pars defect. D, Sagittal MRI shows desiccation of the L5 –S1 disc. E, Arrow demonstrates the pars defect. F, Lateral SPECT technetium-99 bone scan reveals increased uptake at the L5 level. G and H, AP and lateral x-rays of pars repair surgical technique. L5 pedicle screw ﬁxation bilaterally with a Songer cable wound around the spinous process and tightened securely compressing graft in the pars defects.

Discussion constructs have been described with a high rate of good to excellent results (3Ð5, 7, 9Ð13). These techniques are Several surgical techniques have been developed for technically difﬁcult and some require wire, hook, or cable surgical repair of symptomatic spondylolysis and up passage beneath the posterior lamina or the transverse to grade 1 spondylolisthesis refractory to conservative process (2, 4, 12, 14). The Scott technique describes a management. Multiple wire, screw, screw-wire, pedicle wiring beneath the transverse process on each side around screw-wire, pedicle screw-hook, and pedicle screw-rod the spinous process. Songer reported a special fabricated

VOLUME 16, NUMBER 2, SUMMER 2007 3 ABC

E F

FIGURE 3 A–C, Preoperative AP, lateral, and oblique radiographs. A, AP view reveals abnormality reﬂecting a pars defect of the left L5 pars. B, Lateral view suggests a bilateral pars defect at L5. C, Arrow on the oblique x-ray demonstrates a pars defect. D –F, MRI and technetium-99 bone scan imaging of the L5 pars defect. D, MRI sagittal disc desiccation of L5 –S1 without spondylolisthesis. E,CTaxial view demonstrates a unilateral pars defect. F, Technetium-99 bone scan shows increased uptake of the unilateral pars defect.

A B C

DEF

FIGURE 4 A–C, Six months postoperative CT scan. Coronal (A), axial (B), and sagittal (C) views demonstrating persistent incomplete healing of the pars defects repair. D –F, CT Scan at 1 year following surgical repair. Much improved healing of the pars defect bilaterally is evident. 4 JOURNAL OF SURGICAL ORTHOPAEDIC ADVANCES pedicle screw with a hole in its head to allow passage 3. Gillet, P., Petit, M. Direct repair of spondylolysis without of a 1-mm cable. This cable was passed from the pedicle spondylolisthesis, using a rod-screw construct and bone grafting screws around the spinous process and beneath the lamina. of the pars defect. Spine 24(12):1252Ð 1256, 1999. 4. Askar, Z., Wardlaw, D., Koti, M. Scott wiring for direct repair of In our practice, there has been occasional loss of fixa- lumbar spondylolysis. Spine 28(4):354Ð 357, 2003. tion with broken Scott wires. There is some hypothetical 5. Tokuhashi, Y., Matsuzaki, H. Repair of defects in spondylolysis increased risk of nerve injury during blind passage of by segmental pedicular screw hook fixation. A preliminary report. wires beneath the transverse process. Sublaminar passage Spine 21(17):2041Ð 2045, 1996. of a cable may not be necessary. Our modification of 6. Reitman, C. A., Esses, S. I. Direct repair of spondylolytic defects in young competitive athletes. Spine J. 2(2):142Ð 144, 2002. previous techniques arose from a need for a stronger mate- 7. Wu, S. S., Lee, C. H., Chen, P. Q. Operative repair of symptomatic rial with less breakage, and lower potential for neurologic spondylolysis following a positive response to diagnostic pars complications related to wire passage beneath the trans- injection. J. Spinal Disord. 12(1):10Ð 16, 1999. verse process or lamina. 8. Roca, J., Moretta, D., Fuster, S., Roca, A. Direct repair of The described modification to the Songer cable-pedicle spondylolysis. Clin. Orthop. 246:86Ð 91, 1989. 9. Buck, J. E. Direct repair of the defect in spondylolisthesis. screw technique does not involve the transverse process Preliminary report. J. Bone Joint Surg. 52-B(3):432Ð 437, 1970. and does not require sublaminar cable passage, decreasing 10. Pedersen, A. K., Hagen, R. Spondylolysis and spondylolisthesis. the degree of technical difficulty. Although the number of Treatment by internal fixation and bone-grafting of the defect. cases is small, all patients reported relief of their symp- J. Bone Joint Surg. 70-A(1):15Ð 24, 1988. toms and were able to resume normal athletic activity. 11. Ivanic, G. M., Pink, T. P., Achatz, W., Ward, J. C., Homann, N. C., May, M. Direct stabilization of lumbar spondylolysis with a This suggests that a well-motivated athlete may return to hook screw: mean 11-year follow-up period for 113 patients. Spine his sport after a direct repair of pars defects for symp- 28(3):255Ð 259, 2003. tomatic spondylolysis. 12. Johnson, G. V., Thompson, A. G. The Scott wiring technique for direct repair of lumbar spondylolysis. J. Bone Joint Surg. 74- B(3):426Ð 430, 1992. References 13. Kakiuchi, M. Repair of the defect in spondylolysis. Durable fixation with pedicle screws and laminar hooks. J. Bone Joint Surg. 79- 1. Fredrickson, B. E., Baker, D. McHolick, W. J. Yuan, H. A., Lubicky, A(6):818Ð 825, 1997. J. P. The natural history of spondylolysis and spondylolisthesis. J. 14. Hefti, F., Seelig, W., Morscher, E. Repair of lumbar spondylolysis Bone Joint Surg. 66-A(5):699Ð 707, 1984. with a hook-screw. Int. Orthop. 16(1):81Ð 85, 1992. 2. Songer, M. N., Rovin, R. Repair of the pars interarticularis defect with a cable-screw construct. A preliminary report. Spine 23(2):263Ð 269, 1998.

VOLUME 16, NUMBER 2, SUMMER 2007 5 QUERIES TO BE ANSWERED BY AUTHOR

IMPORTANT NOTE: Please mark your corrections and answers to these queries directly onto the proof at the relevant place. Do NOT mark your corrections on this query sheet.

Queries from the Copyeditor: AQ1 Pls verify spelling change ARTICLE IN PRESS

The Spine Journal - (2007) -

Clinical Study Fusion assessment of posterior lumbar interbody fusion using radiolucent cages: X-ray ﬁlms and helical computed tomography scans compared with surgical exploration of fusion Guy R. Fogel, MD*, John S. Toohey, MD, Arvo Neidre, MD, John W. Brantigan, MD South Texas Orthopedic and Spinal Surgery Associates, 9150 Huebner Road, San Antonio, TX 78240, USA Received 29 October 2006; accepted 19 March 2007

Abstract BACKGROUND: Plain radiographic assessment of posterolateral fusion has been reported as accurate in only two thirds of patients who were found to be healed at surgical exploration. Plain radiographic techniques for fusion assessment of interbody fusion with radiolucent cages are reported to be accurate. A helical computed tomography (CT) scan shows a high sensitivity for pseudarthrosis compared with plain radiography. PURPOSE: To determine the accuracy of fusion assessment with plain X-ray films and helical CT scans by comparison to results of surgical exploration of fusion. STUDY DESIGN/SETTING: The accuracy and interobserver agreement of plain X-ray films and thin-cut helical CT scans were compared with fusion assessment by surgical exploration in patients with posterior lumbar interbody fusion using a radiolucent carbon fiber reinforced polymer cage (CFRP) and iliac crest bone graft. PATIENT SAMPLE: A review of 90 patients who had surgical exploration of the lumbar fusion. OUTCOME MEASURES: All patients had plain X-ray films including Ferguson anteroposterior parallel to the interbody space. Fifty-four patients had thin-section helical CT scans. METHODS: Fusion assessment by exploration was compared with blinded assessment by plain X-ray films and CT scans. RESULTS: Ninety patients had surgical exploration of 172 lumbar interbody and posterolateral fusion levels. At the time of exploration, fusion was determined to be successful in 87 of 90 patients and 168 of 172 (97%) fusion levels. X-ray assessment showed healed interbody fusions in 87% and posterolateral fusion healed in 75%. CT grading of the interbody fusion found healed interbody fusion in 77%, and the posterolateral fusion was fused in 68%. Plain X-ray films and CT scans had a sensitivity of 100% for pseudarthrosis and a negative predictive value of 100% for healed fusion. Specificity was almost 90% and was not significantly different between X-ray films and CT scans. CONCLUSIONS: Fusion assessment with plain X-ray films and helical CT scans showed equal accuracy after posterior lumbar interbody fusion confirmed by surgical exploration. Our results indicate that when plain X-ray films show strong evidence of fusion or pseudarthrosis, the helical CT is unlikely to provide useful new information. Ó 2007 Elsevier Inc. All rights reserved.

Keywords: Spinal fusion; Evaluation methods; Outcome; Lumbar vertebrae; Radiography; Surgery; Tomography; X-ray computed

Introduction

FDA device/drug status: approved for this indication (Lumbar I/F Cage Modern techniques combining posterolateral fusion and the VSP Spinal Fixation System). (PLF) with interbody fusion (PLIF) using radiolucent inter- Dr. Brantigan has a ﬁnancial interest in the Lumbar I/F Cage. body cages have increased the fusion rate to nearly 100% * Corresponding author: South Texas Orthopedic and Spinal Surgery Associates, 9150 Huebner Road, San Antonio, TX 78240. Tel.: (210) by plain radiographic assessment [1–5]. Nevertheless, 561-7234; fax: (210) 561-7240. 10% to 15% of patients may remain symptomatic. Accurate E-mail address: [email protected] (G.R. Fogel) radiographic assessment of fusion success is important to

2 G.R. Fogel et al. / The Spine Journal - (2007) - identify patients who might benefit from further surgery. minimum follow-up after the index fusion operation was Physicians have continued to struggle with the correct 24 months (mean, 38 months [median, 35 months; range, interpretation of radiographic methods. To date, the ‘‘gold 24-68 months]). standard’’ of fusion determination has been surgical exploration [6–8]. Comparison of plain radiographic assessment Radiographic assessment of fusion of posterolateral fusion does not correlate with surgical exploration in a third or more of patients [6,9,10]. Plain All patients underwent X-ray assessment of 172 levels radiographic techniques for fusion assessment of PLIF of fusion before the surgical exploration using plain static with radiolucent cages have been compared with surgical X-ray films including anterior-posterior, lateral, and exploration and have been reported to be accurate [1,2]. individual Ferguson views of each interbody fusion level. Improvement in computed tomography (CT) scanning The Ferguson view is an anterior-posterior X-ray directed has increased accuracy in fusion assessment. Early studies parallel to the end plates of the vertebral body designed used only axial images, and, when compared with surgical to visualize the interbody fusion. Two of the authors exploration, they showed noncorrelation in up to 43% of (GRF and JST), blinded to the names of the patients and cases [6]. More recently, thin-section 1 mm axial helical the results of the surgical exploration, independently graded CT scans with sagittal and coronal reconstructions provide the X-ray films for evidence of interbody and posterolateral high-quality images with exquisite bone detail. Several fusion. helical CT studies have shown a high specificity for Fifty-four patients with 109 levels of fusion had thin- detection of pseudarthrosis compared with plain radiogra- section helical CT scans with sagittal and coronal recon- phy particularly in evaluation of the interbody fusion structions before the exploration of fusion at an average [11–16]. of 30 months (range, 10-60 months) after the index surgery. From a single surgical cohort, we have performed a ret- These patients underwent 1-mm thin-section helical CT rospective chart and radiographic review of 90 consecutive scanning of the involved lumbar segments performed on patients who had surgical exploration of the fusion after a high-speed helical scanner (Sensation 4; Siemens). posterior lumbar interbody fusion at one or more levels. Reconstructed images were generated in sagittal plane and Our purpose was to show the accuracy of plain X-ray films coronal planes exactly perpendicular to the plane of each and helical CT scans in fusion assessment. To our knowl- cage. Each helical CT scan was interpreted by one of three edge, a comparison of plain X-ray films and helical CT radiologists who were blinded with regard to the clinical scans to surgical exploration in evaluation of PLIF with history and the plain radiographic findings. radiolucent cages has not been previously reported. The interbody and posterolateral fusion were graded by two methods. Each level and each side was judged individually. Interbody fusion was graded by the method of Brantigan and Steffee [17] as modified to describe the Fras- Materials and methods er definition of locked pseudarthrosis (BSF scale) [14] Clinical study design outlined in Table 1. The posterolateral fusion was graded by the method of Lenke et al. [18] outlined in Table 2. From February 1999 to October 2003, we performed PLIF using the Lumbar I/F Cage (DePuy Spine, Raynham, Surgical exploration of fusion MA) with pedicle screw and plate fixation in our practice. Iliac crest bone autograft was used for the interbody fusion At the time of exploration, the pedicle screws were re- and local bone for the posterolateral fusion. Inclusion moved. All soft tissue was removed from the fusion mass. criteria included patients with disabling back pain because The fusion mass, facet joints, and intertransverse areas of degenerative disc disease at one or more lumbar levels and included patients with primary degenerative disc disease, failed discectomy, spondylolisthesis, failed fusion, Table 1 and degenerative spondylolisthesis. During the spring of Classification of interbody fusion success: Brantigan, Steffee, Fraser (BSF) 2005, we conducted a retrospective chart review of all BSF-1: Radiographical pseudarthrosis is indicated by collapse of the patients. From this group, 90 patients were identified who construct, loss of disc height, vertebral slip, broken screws, had surgical exploration of the lumbar fusion. There were displacement of the carbon cage, or significant resorption of the bone 48 men and 42 women. The average age was 43 years graft, or lucency visible around the periphery of the graft or cage. (range, 27-70). The average time from the index fusion BSF-2: Radiographical locked pseudarthrosis is indicated by lucency visible in the middle of the cages with solid bone growing into the cage operation to the exploration of the fusion was 27.2 months from each vertebral endplate. (12-65 months). The indications for exploration of the BSF-3: Radigraphical fusion: bone bridges at least half of the fusion area fusion were persistent nonspecific low back pain in 63, with at least the density originally achieved at surgery. Radiographical adjacent-level operation for indications of instability or fusion through one cage (half of the fusion area) is considered to be stenosis in 18, and radiographic pseudarthrosis in 10. The mechanically solid fusion even if there is lucency on the opposite side. ARTICLE IN PRESS

G.R. Fogel et al. / The Spine Journal - (2007) - 3 Table 2 threshold for fusion was set at interbody radiographic Lenke classification of posterolateral fusion success fusion BSF-3 (Table 1) and posterolateral fusion at Lenke-A Grade A: Definitely solid with bilateral trabeculated stout fusion masses (Table 2). The values for the plain X-ray films were com- present pared with those values for the helical CT scan, with use Grade B: Possibly solid with a unilateral large fusion mass and of the normal distribution to approximate the difference a contralateral small fusion mass Grade C: Probably not solid with a small fusion mass bilaterally between any two percentages. Interobserver agreement Grade D: Definitely not solid with bone graft resorption or obvious for X-ray film interpretation was assessed with the chi pseudarthrosis bilaterally square test, Fischer exact test, and the McNemar’s test. Statistical analysis was performed by using SPSS 6.1.3 (SPSS Inc, Chicago, IL). were carefully explored. Motion was evaluated by one or more methods. Where significant portions of the adjacent spinous processes remained, a Kocher clamp was used to apply distraction or compression; direct compression was Results placed on the fusion mass and facets with a large punch. Ninety patients had surgical exploration of 172 lumbar Long pedicle probes were placed in each pedicle screw hole interbody and posterolateral fusion levels (Table 3). At and used to apply compression and distraction. A solid the time of exploration, fusion was determined to be suc- posterolateral fusion was indicated by observation of a solid cessful in 87 of 90 patients and 168 of 172 (97%) fusion bridge of cortical bone bridging the posterolateral area. levels (Fig. 1). Surgical exploration revealed four pseu- Posterolateral pseudarthrosis was indicated by either an darthroses (4/172, 2.3%). Two of the four pseudarthroses observed defect in the bridging bone or by visible motion occurred at L5–S1, one at L4–5, and one at L1–2. Because in the posterolateral fusion area. Each side of the postero- all pseudarthrosis levels were relatively stable, the pedicle lateral fusion was graded separately. Solid interbody fusion screws were left out and augmentation of the posterolateral was indicated by total rigidity and lack of motion between fusion was performed with a combination of supplemental the manipulated pedicle probes. Interbody pseudarthrosis local bone chips, demineralized bone graft, and/or bone was indicated by any relative motion between the manipu- morphogenic protein. The patient with L1–2 pseudarthrosis lated pedicle probes. had two revision procedures. The first included removal of pedicle screws and augmentation of the posterior-lateral Statistical methods grafting. The second included reinstrumentation with pedicle screw fixation and augmentation of the posterior-lateral The sensitivity, specificity, positive predictive value, and bone grafting (Fig. 2). negative predictive value for detecting pseudarthrosis were X-ray assessment showed healed interbody fusions calculated for all plain X-ray films and the helical CT scans in 87%, and the posterolateral fusion was healed bilater- by comparing the results of the independent interpretations ally in 75% (Table 3). Interobserver agreement for X-ray of those studies with the results of surgical exploration. The interpretation was not significantly different comparing

Table 3 Fusion assessment of interbody fusion (BSF) and posterolateral fusion (Lenke) L5–S1 # (%) L4–5 # (%) L3–4 # (%) L2–3 # (%) L1–2 # (%) Total levels (%) Matched XR to CT (%) Explored 69 (40) 69 (40) 23 (13) 9 (5) 2 (1) 172 Pseudarthrosis 2 1 0 0 1 4 (2) Levels with CT 38 45 16 8 2 109 BSF-3 XR 58 (84) 61 (88) 21 (91) 8 (89) 1 (50) 149 (87) CT 28 (74) 39 (87) 15 (94) 7 (88) 1 (50) 90 (77) 86 (96) BSF-2 XR 11 (16) 8 (12) 1 (4) 1 (11) 0 21 (12) CT 9 (24) 6 (13) 1 (6) 1 (13) 0 17 (16) 9 (53) BSF-1 XR 0 0 1 (4) 0 1 (50) 2 (1) CT 1 (3) 0 0 0 1 (50) 2 (2) 1 (50) Lenke-A XR 57 (83) 48 (70) 17 (74) 6 (67) 1 (50) 129 (75) CT 28 (74) 26 (58) 13 (81) 6 (75) 1 (50) 74 (68) 63 (85) Lenke-B XR 7 (10) 14 (20) 3 (13) 2 (22) 0 26 (16) CT 1 (3) 10 (22) 2 (13) 0 0 13 (12) 6 (46) Lenke-C/D XR 5 (7) 7 (10) 3 (13) 1 (11) 1 (50) 17 (9) CT 9 (24) 9 (20) 1 (6) 2 (25) 1 (50) 22 (25) 2 (40) There were 172 levels explored. Four of 172 were pseudarthroses at surgical exploration. All levels had X-ray films, and 109 levels had helical CT scans. The image grading is reported as the number of levels with that grade and a percentage of levels with X-ray film or CT scan. Matching image grading is reported as number of X-ray films to CT scan matches as a percentage of levels with CT scans. ARTICLE IN PRESS

4 G.R. Fogel et al. / The Spine Journal - (2007) - ARTICLE IN PRESS

G.R. Fogel et al. / The Spine Journal - (2007) - 5 interbody and posterolateral fusion level by level and is based on observations made over many years. Numerous overall (chi-square test50.5595, p5.455, Fischer exact authors have noted that a solid posterolateral fusion has test5.046, and McNemar’s55.400, p5.02). The proportion flexibility of motion up to 4 or 5 [19–22]. Unlike a pos- of overall agreement between observers was 98.6%. terolateral fusion, a successful interbody fusion creates CT grading of the interbody fusion found solidly healed a totally rigid motion segment [23–25]. Luk et al. [24] eval- interbody fusion in 77%, and the posterolateral fusion was uated 52 cases of anterior lumbar interbody fusion 5 years solidly fused bilaterally in 68% (Table 3). When the CT after surgery and noted there was an average of 1 of mo- scan showed a healed interbody or posterolateral fusion, tion measured in the fused segment and 12 to 15 in the the X-ray film was also healed in 96%. If the CT scan adjacent nonfused segments. The authors attributed the 1 showed a healed posterolateral fusion, the X-ray film was of motion to artifact and rotational and measurement also healed in 85% (Table 3). variations. Table 4 presents the data to calculate the sensitivity, The possible errors in each surgical determination specificity, positive and negative predictive values for plain should be considered individually. Failed PLIF and failed X-ray films, and CT fusion assessment confirmed at explo- PLF should be the easiest and most accurate category to ration. The incidence of pseudarthrosis at the time of surgi- evaluate surgically because there is a definite gap in the cal exploration was 4/172 levels or 2.3%. Sensitivity refers posterolateral fusion mass. Axial compression and distrac- to how good a radiographic test is at correctly identifying tion shows definite motion in the posterolateral area, and fusion levels with pseudarthrosis. When calculating sensi- flexion forces on the pedicle probes show motion in the in- tivity, we are therefore interested in only the fusion levels terbody area. With healed PLIF and failed PLF, the pedicle with pseudarthrosis. The sensitivity of both CT and plain probes show no motion on flexion or extension forces; X-ray films was 100% in correctly identifying pseudarthro- however, there is a visible discontinuity in the posterolat- sis. Specificity, on the other hand, is concerned with how eral fusion mass. With failed PLIF and healed PLF, there good the radiographic test is at correctly identifying fusion will be a solid bridge of bone bridging the posterolateral levels that are healed and do not have a pseudarthrosis. fusion area. Axial distraction or compression produces no Specificity was almost 9 of 10 and was not significantly dif- motion in the posterolateral area, but flexion forces on ferent between X-ray film and CT scan at the chosen the pedicle probes produces a slight motion or ‘‘springi- threshold for fusion. The positive predictive value refers ness.’’ Admittedly, this category provides the greatest pos- to the chance that a positive radiographic test for pseudarth- sibility of misinterpretation. With experience, healed PLIF rosis will be correct (ie, the chance was less than one in and healed PLF category can be accurately determined four that there was a pseudarthrosis at exploration when because there is solid bridging posterolateral bone with the X-ray films were interpreted as below the threshold no motion on either axial or flexion forces. A locked pseu- for fusion). The negative predictive value is concerned only darthrosis should appear to be a healed PLIF because, by with the radiographic tests negative for pseudarthrosis. definition, this type of segment is stable. Although this re- When the radiographic test was solidly healed (negative sult is correctly described as ‘‘pseudarthrosis,’’ it should not for pseudarthrosis), 100% of the levels were healed at be considered a noncorrelation in the two-by-two analysis exploration. or surgical exploration. The plain X-ray findings of the CT group were com- In this study, both the helical CT scans and plain radiog- pared with that of those patients who did not have a CT raphy were found to be highly accurate with high sensitivity scan. The incidence of solid fusion was not significantly for pseudarthrosis and high negative predictive value for different in the two groups. However, the incidence of pseu- healed fusion. Helical CT scans did not miss a single darthrosis was significantly greater in the CT group. All of pseudarthrosis. Plain X-ray films missed one pseudarthrosis the patients with obvious plain radiographic pseudarthrosis at L5–S1 that was diagnosed with a helical CT scan. The did have adjunctive CT scanning. All cases of surgical negative predictive value was the most important measure pseudarthrosis had both plain X-ray and helical CT studies. of accuracy. For both X-ray films and CT scans, the negative predictive value was 100% when either the interbody fusion or the posterolateral fusion was solidly healed. The Discussion only difference between X-ray films and CT scans was The accuracy of distinguishing posterolateral fusion the number of false-positive readings for pseudarthrosis, re- success from PLIF fusion success by surgical exploration flected by the positive predictive value. The X-ray films

Fig. 1. True-negative healed fusion at L4–S1 graded 3A on plain X-ray ﬁlms and 3B on CT scan and conﬁrmed to be healed by exploration. (A) An anteroposterior view parallel to L4–5 end plates shows a single interbody cage with bridging bone at L4–5 (BSF-3) and bridging bone bilateral in posterolateral fusion (Lenke ‘‘A’’). (B) A lateral view with bridging bone across the L4–5 interspace. (C) A CT coronal view through posterolateral graft bridging bone across transverse processes bilatrally (Lenke ‘‘A’’). (D) A CT coronal view through interspace (BSF-3). (E) A CT sagittal view through interbody spaces showing bridging bone (BSF-3’’). ARTICLE IN PRESS

6 G.R. Fogel et al. / The Spine Journal - (2007) - ARTICLE IN PRESS

G.R. Fogel et al. / The Spine Journal - (2007) - 7 Table 4 negatives for X-ray films (Table 4) found no significant True-and false positive and-negative findings of surgical exploration difference between the imaging modalities (chi-square to X-ray films and CT scan fusion assessment test5 1.172, p5.2789). The accuracy of plain x-ray films X-ray TP FN FP TN and helical thin-cut CT scans methods was not statistically Interbody fusion 4 0 19 149 different when compared with the results of surgical Posterolateral fusion 4 0 39 129 CT Scan exploration. Interbody fusion 4 0 15 90 One point of concern with this study is that surgical Posterolateral fusion 4 0 31 74 exploration of a lumbar fusion is not commonly indicated TP5True-positive represents pseudarthrosis at exploration and by ra- today. Indications for late re-exploration of a lumbar fusion diographical criteria. FN5False-negative or pseudarthrosis at exploration (more than 1 year after initial fusion) are commonly radio- and BSF-3 by radiographical criteria, FP5false-positive is healed fusion graphic pseudarthrosis, persistent pain or recurrent symp- at exploration and less than BSF-3 by radiographic criteria, TN5true neg- toms, instrumentation failure, or progressive degeneration ative is healed fusion at exploration and BSF-3 by X-ray film criteria. at another spine level [6,8–10,28–31]. Martin et al. [31] found the rate of late reoperation was 21.5% after lumbar fusion. The indications for late reoperation more than year after fusion are not significantly different than the Branti- falsely predicted pseudarthrosis in 6% and the CT scans in gan FDA-IDE series in 2000 and 10-year follow-up in 8%. The false-positive test result was not significantly 2004 [1,2]. A statistical problem of this study is the small different between the two modalities in assessment of either incidence of pseudarthrosis. A larger sample size would interbody fusion (chi-square test5.4624, p5.4965) or clarify our results [32]. Another concern is that this study posterolateral fusion (chi-square test5.012, p5.9609). may have been better if all patients had both CT and plain The plain radiographic diagnosis of a successful inter- radiographic studies. However, statistical comparison of body arthrodesis was enhanced by a radiolucent implant patients with and without CT studies showed no significant filled with cancellous iliac crest graft and standardized differences in fusion grades. In practical clinical use for our radiographic technique including the Ferguson view of study group, admittedly without statistical support of the interbody space. In other studies, accuracy in determin- efficacy, helical CT scans were more often performed in ing fusion was diminished because of the metallic posterior patients in whom the plain films showed indeterminate and anterior interbody instrumentation obscuring the bone evidence of fusion. With these results, the additional cost graft. Metallic implants obscure the bone graft within, mak- and increased radiation exposure with helical CT scanning ing it difficult to see remodeling or trabeculations that have may not be justified. Our results indicate that when the formed across the interspace between the end plates. When plain films show strong evidence either of fusion or pseu- cortical bone chips or synthetic filler is mixed with the darthrosis, the helical CT is unlikely to provide useful cancellous bone graft, the graft is more radio-opaque and new information. remodeling or resorption must be interpreted to assess the progress of bone fusion. Plain X-ray films are dependent on obtaining a true lateral and a true Ferguson parallel view of the interbody end plates. Suboptimal films when seen Conclusion should be repeated. A helical CT scan with sagittal and coronal reconstruc- In a study of surgical exploration after PLIF and postero- tions has become the preferred imaging to assess both pos- lateral fusion using a radiolucent interbody fusion device, terolateral and interbody fusion [12,14,26,27]. CT coronal X-ray and CT methods performed very similarly in evalu- reconstructions are particularly helpful in evaluation of ating lumbar fusion success with no significant differences the continuity of the posterolateral fusion. With the fine de- in accuracy between the two methods. Both x-ray and tail of bone trabeculations within the interspace, a helical helical CT scans predicted pseudarthrosis with 100% sensi- CT scan would seem to be the ideal adjunct imaging for tivity and interbody and posterolateral fusion success with evaluation of suspected pseudarthrosis. However, a helical a 100% negative predictive value. Our results indicate that CT scan may overestimate the significance of lucencies in when plain films show strong evidence of fusion or pseu- the interbody space and may not correctly predict fusion darthrosis the helical CT is unlikely to provide useful at exploration. Comparison of true positives and true new information.

Fig. 2. A patient with true-positive pseudarthrosis of L1–2 fusion. Plain X-ray ﬁlms graded ‘‘1D,’’ CT scan graded ‘‘1D,’’ and L1–2 pseudarthrosis was found at exploration. (A) An anteroposterior view of L1–2; no posterolateral fusion mass is visible (Lenke ‘‘D’’) and lucencies around the interbody cages (BSF-1) are seen. (B) A lateral view showing subsidence of the cages with lucencies between the end plate and the cages (BSF-1). (C) A CT coronal with poor posterolateral fusion mass on only one side (Lenke ‘‘C’’). (D) A CT coronal through interspace showing lucencies around the cages (BSF-1). (E) A CT sagittal view with subsidence and lucencies around an interbody cage (BSF-1). ARTICLE IN PRESS

8 G.R. Fogel et al. / The Spine Journal - (2007) - References [15] Siambanes D, Mather S. Comparison of plain radiographs and CT scans in instrumented posterior lumbar interbody fusion. Orthopedics [1] Brantigan JW, Neidre A, Toohey JS. The lumbar I/F cage for 1998;21:165–7. posterior lumbar interbody fusion with the variable screw placement [16] Shah RR, Mohammed S, Saifuddin A, Taylor BA. Comparison of system: 10-year results of a Food and Drug Administration clinical plain radiographs with CT scan to evaluate interbody fusion follow- trial. Spine J 2004;4:681–8. ing the use of titanium interbody cages and transpedicular instrumen- [2] Brantigan JW, Steffee AD, Lewis ML, Quinn LM, Persenaire JM. tation. Eur Spine J 2003;12:378–85. Lumbar interbody fusion using the Brantigan I/F cage for posterior [17] Brantigan JW, Steffee AD. A carbon fiber implant to aid interbody lumbar interbody fusion and the variable pedicle screw placement lumbar fusion. Two-year clinical results in the first 26 patients. Spine system: two-year results from a Food and Drug Administration 1993;18:2106–7. investigational device exemption clinical trial. Spine 2000;25: [18] Lenke LG, Bridwell KH, Bullis D, Betz RR, Baldus C, 1437–46. Schoenecker PL. Results of in situ fusion for isthmic spondylolisthe- [3] Kim KS, Yang TK, Lee JC. Radiological changes in the bone fusion sis. J Spinal Disord 1992;5:433–42. site after posterior lumbar interbody fusion using carbon cages [19] Zdeblick TA. A prospective, randomized study of lumbar fusion. impacted with laminar bone chips: follow-up study over more than Preliminary results. Spine 1993;18:983–91. 4 years. Spine 2005;30:655–60. [20] Frymoyer JW, Hanley EN Jr, Howe J, Kuhlmann D, Matteri RE. A [4] Miura Y, Imagama S, Yoda M, Mitsuguchi H, Kachi H. Is local bone comparison of radiographic findings in fusion and nonfusion patients viable as a source of bone graft in posterior lumbar interbody fusion? ten or more years following lumbar disc surgery. Spine 1979;4: Spine 2003;28:2386–9. 435–40. [5] Stromberg L, Toohey JS, Neidre A, Ramsey M, Brantigan JW. Com- [21] Lehmann TR, Spratt KF, Tozzi JE, et al. Long-term follow-up of plications and surgical considerations in posterior lumbar interbody lower lumbar fusion patients. Spine 1987;12:97–104. fusion with carbon fiber interbody cages and Steffee pedicle screws [22] Kuslich SD, Danielson G, Dowdle JD, et al. Four-year follow-up and plates. Orthopedics 2003;26:1039–43. results of lumbar spine arthrodesis using the Bagby and Kuslich [6] Brodsky AE, Kovalsky ES, Khalil MA. Correlation of radiologic as- lumbar fusion cage. Spine 2000;25:2656–62. sessment of lumbar spine fusions with surgical exploration. Spine [23] Pearcy MJ, Whittle MW. Movements of the lumbar spine measured 1991;16(6 suppl):S261–5. by three-dimensional X-ray analysis. J Biomed Eng 1982;4:107–12. [7] Cleveland M, Bosworth D, Thompson F. Pseudarthrosis in the lumbo- [24] Luk KD, Chow DH, Evans JH, Leong JC. Lumbar spinal mobility sacral spine. J Bone Joint Surg Am 1948;30:302–12. after short anterior interbody fusion. Spine 1995;20:813–8. [8] Hilibrand AS, Dina TS. The use of diagnostic imaging to assess [25] Lee CK, Langrana NA. Lumbosacral spinal fusion. A biomechanical spinal arthrodesis. Orthop Clin North Am 1998;29:591–601. study. Spine 1984;9:574–81. [9] Blumenthal SL, Gill K. Can lumbar spine radiographs accurately [26] Cook SD, Patron LP, Christakis PM, Bailey KJ, Banta C, Glazer PA. determine fusion in postoperative patients? Correlation of routine Comparison of methods for determining the presence and extent of radiographs with a second surgical look at lumbar fusions. Spine anterior lumbar interbody fusion. Spine 2004;29:1118–23. 1993;18:1186–9. [27] Williams AL, Gornet MF, Burkus JK. CT evaluation of lumbar inter- [10] Kant AP, Daum WJ, Dean SM, Uchida T. Evaluation of lumbar spine body fusion: current concepts. AJNR Am J Neuroradiol 2005;26: fusion. Plain radiographs versus direct surgical exploration and obser- 2057–66. vation. Spine 1995;20:2313–7. [28] McAfee PC, Cunningham BW, Lee GA, et al. Revision strategies for [11] Lang P, Genant HK, Chafetz N, Steiger P, Morris JM. Three-dimen- salvaging or improving failed cylindrical cages. Spine 1999;24: sional computed tomography and multiplanar reformations in the as- 2147–53. sessment of pseudarthrosis in posterior lumbar fusion patients. Spine [29] Wild A, Pinto MR, Butler L, Bressan C, Wroblewski JM. Removal of 1988;13:69–75. lumbar instrumentation for the treatment of recurrent low back pain [12] Burkus JK, Transfeldt EE, Kitchel SH, Watkins RG, Balderston RA. in the absence of pseudarthrosis. Arch Orthop Trauma Surg 2003; Clinical and radiographic outcomes of anterior lumbar interbody fu- 123:414–8. sion using recombinant human bone morphogenetic protein-2. Spine [30] Nottingham P, Ball H. Hardware removal surgery in patients with 2002;27:2396–408. persistent low back pain. Presented at the North American Spine [13] Chafetz N, Cann CE, Morris JM, Steinbach LS, Goldberg HI, Ax L. Society Meeting. 2005. Philadelphia, PA. Pseudarthrosis following lumbar fusion: detection by direct coronal [31] Martin BI, Mirza SK, Comstock BA, Gray DT, Kreuter W, Deyo RA. CT scanning. Radiology 1987;162:803–5. Reoperation rates following lumbar spine surgery and the influence [14] Santos ER, Goss DG, Morcom RK, Fraser RD. Radiologic assess- of spinal fusion procedures. Spine 2007;32:382–7. ment of interbody fusion using carbon fiber cages. Spine 2003;28: [32] Smith JE, Winkler RL, Fryback DG. The first positive: computing pos- 997–1001. itive predictive value at the extremes. Ann Intern Med 2000;132:804–9. Elsevier Editorial System(tm) for The Spine Journal

Manuscript Draft

Manuscript Number:

Title: Outcomes of Posterior Lumbar Interbody Fusion with 9mm Width Lumbar I/F Cage and the Variable Screw Placement System

Article Type: Clinical Study

Section/Category: Outcomes/Measurement

Keywords: PLIF; Interbody fusion orthosis; Cage; Pedicle screw; Degenerative disc disease.

Corresponding Author: Dr. Guy R. Fogel, MD

Corresponding Author's Institution: South Texas Orthopedics and Spinal Surgery Associates

First Author: Guy R. Fogel, MD

Order of Authors: Guy R. Fogel, MD; John S. Toohey, MD; Arvo Neidre, MD; John W Brantigan, MD

Manuscript Region of Origin:

Abstract: Background Context Posterior lumbar interbody fusion (PLIF) with the Lumbar I/F Cage and posterior Variable Screw Placement System (VSP) has increased the successful fusion rate to nearly 100% at the four lowest levels, L2-3 through L5-S1. In the design of the cage, only the surface area of the opening for bone graft contact with the endplates varied with the size of the cage. The 11mm cage was the optimal design allowing 5.5mm of bone graft opening or twice the width of each strut. The 9mm cage has only 3.5mm width for bone grafting or proportionately 36% less surface area than the 11mm cages and 54% less than the 13mm cages. In the narrow cage, the bone healing potentially is compromised by the smaller surface area. For this reason, it is clinically important to study the clinical and fusion success of these narrow cages. In some instances, when space is limited, only 9mm width cages may be used at a fusion level. Clinical results of 9mm Lumbar I/F cage fusion and VSP have not been reported. Purpose The purpose is to report the functional outcomes, fusion rate, and complications related to PLIF when more narrow 9mm interbody cages are used. Study Design/Setting The setting is a retrospective single-arm cohort study of consecutive PLIF surgical patients at a single center. Patient Sample A review of 373 of 425 patients who underwent PLIF with lumbar I/F cage and VSP from 1999 to 2002 identified 90 patients who had PLIF with 9mm Lumbar I/F cages and VSP at a fusion level. Minimum follow- up was 24 months (Mean 36 months (median 33 months, range 24-65 months). Methods The 90 patients were evaluated for Clinical success and Fusion success at last follow-up. The results were compared to the results of the 373 patients reviewed, and historical groups of the original IDE study and the 10-year follow-up study. Outcome Measures Clinical success was determined with a modified Prolo score evaluating pain, function, medication usage and economic status. Fusion success, determined by evaluation of plain radiographs and helical cut CT scans, was defined by continuous bone bridging the fusion area with no lucency. Results Prolo scores were Excellent in 33 (37%), Good in 34 (38%), Fair in 20 (22%), and Poor in 3 (3%). 74 patients (82%) had clinical success, and 16 (18%) were clinically unsuccessful. Fusion was successful in 87 (97%) and failed in three patients (3%). Two failed fusions at the L1-2 level, were revised and failed fusion again, and the third was complicated by infection and was never explored. Conclusions PLIF with 9mm width Lumbar I/F Cage and VSP had 97% fusion success. There is no statistical difference to the fusion success with PLIF and two 11 or 13 mm width Lumbar I/F cages and VSP. Mechanical failure did not occur with 9mm cage usage. Authors recommend routine use of bilaterally placed Lumbar I/F cages of maximal width allowable. These results show clinical and fusion success was not diminished by use of 9mm width cages. * Title Page

Outcomes of Posterior Lumbar Interbody Fusion with 9mm Width Lumbar I/F Cage and the Variable Screw Placement System Guy R. Fogel M.D. John S. Toohey M.D. Arvo Neidre M.D. John W. Brantigan M.D. South Texas Orthopaedic and Spinal Surgery Associates 9150 Huebner Road San Antonio Texas, 78240 (210)561-7234 FAX: (210)561-7240 e-mail: [email protected]

Corresponding author: Guy R. Fogel M.D. Device status: The Lumbar I/F Cage and the VSP Spinal Fixation System are approved by the United States Food and Drug Administration. Dr. Brantigan has a financial interest in the Lumbar interbody carbon fiber Cage. No Financial assistance was provided from any source for this study. BLINDED Manuscript in MS Word

Posterior lumbar interbody fusion (PLIF) was introduced by Cloward [1-4] more than fifty years ago. Limitations of PLIF included inadequate compression strength of allograft bone[5] and low fusion rates[6]. The Lumbar I/F cage (DePuy Spine, Raynham, MA) was designed to enhance posterior lumbar interbody fusion results by providing the mechanical strength to provide load-sharing during fusion and to allow the best biological healing with autologous bone graft[7, 8]. In an Investigational Device Exemption (IDE) study of the Lumbar I/F Cage, fusion success was reported in 176 of 178 patients (98.9%)[9]. In the 10-year follow-up of the IDE study, Fusion success was reported in 29 of 30 (96.7%) at 10 years[10]. Based on this study, the Lumbar I/F Cage was approved by the FDA for treatment of degenerative disc disease at one or two contiguous levels from L2 to the sacrum, accompanied by posterior-lateral fusion and pedicle screw fixation using the VSP Spinal System. Since the release of the lumbar I/F cages in 1999, favorable results have been reported in Japan [11, 12], United States clinical studies[10, 13], in a active military population[14], and in Europe [15.] When the Lumbar I/F Cage was first designed, the optimal dimensions were unknown. The 11 mm cage prototypes had the opening of 5.5 mm width and the combined strut width of 5.5 mm (or 2.75 mm each)[16]. In the design of the 9 mm cage size the strut width was retained because of the concern for subsidence if the strut was weakened. This made the opening for bone graft 36% less than the standard 11 mm width cage. Because bone healing was potentially compromised with the 9 mm width cages by the proportionally greater reduction in bone graft surface area, it is clinically important to study the clinical and fusion success of these narrow cages. In a single surgical practice, we have reviewed 373 of 425 patients who underwent PLIF with the Lumbar I/F Cage and VSP pedicle screw fixation from 1999 to 2002. Of these, 90 were identified who had PLIF at one or more levels using one or two 9 mm wide cages. The purpose of this retrospective review is to report the results of the 9 mm cages and VSP. Materials and Methods Description of the Lumbar I/F cage The Lumbar I/F cage is manufactured from poly-ether-ether ketone reinforced with “chopped” or short carbon fibers. The cage is filled with autologous bone and placed in a disc space channel. This allows bone fusion from endplate to endplate across the bone graft within the cage. The cage material has approximately the same modulus of elasticity as cortical bone and is one-tenth the stiffness as compared with metals. This flexibility minimizes stress shielding of the bone graft and potentially decreases motion between the graft and the endplates during healing[9]. Because the cages are radiolucent, bridging bone can be accurately assessed by plain radiographs. The width and height of the Lumbar I/F cage are 9mm, 11mm, or 13mm. The lengths are 21 and 25mm. There are 9X9, 9X11, 9X13, 11X11, 11X13, and 13X13mm. The two cage lengths are 21 and 25mm lengths. The 13X13 is only available in 25mm in length. The implant has an open architecture for packing of autologous bone, allowing a large surface area of bone contact (Figure 1). Description of the Variable Screw Placement System (VSP) The VSP (DePuy Spine, Raynham, MA) includes pedicle screws and plates. The screw has a cancellous thread ending in an integral nut, above which a machine thread with tapered and locking nuts allows the screw to be anchored securely to the plate. The lumbar I/F cage and the VSP have been previously described [9]. Clinical study design From February 1999 to October 2003, we performed PLIF using the Lumbar I/F Cage with VSP pedicle screw fixation on 425 patients in our practice. During the summer of 2004, we conducted a retrospective chart review of all patients. We had a minimum of one year of full clinical and radiographic data on 377 of 425 patients (87%). While one year is not long enough to define an exact rate of clinical success, it is sufficient follow-up to identify complications and failed fusions. The current study group is comprised of consecutive 9mm cage patients with a minimum of two years follow-up. All 425 patients had disabling back and/or radicular pain refractory to conservative management with moderate to severe degenerative change, with instability and adjacent segment deterioration based on magnetic resonance imaging (MRI) or discogram. 49 patients (54%) had 75 prior lumbar surgeries (average 1.5 surgeries/patient). In the previously operated patients, there was an average interval of 5 years (range 1-18 years) from the most recent surgery to the current surgery. Demographic data was recorded including age, gender, neurologic function, number and type of prior back surgeries, and diagnosis at each lumbar level. Patients were examined and data recorded before surgery, at surgery, and at post-operative intervals after surgery. Evaluations included ratings of pain, function, economic status, and medication. Patient satisfaction, sensory and motor function, tension signs, reflexes, and radiographic findings were recorded at each follow-up interval. The mean follow-up was 36 months (range 24-65 months). Clinical success was defined according to previous published literature parameters used over many years and modeled after an expanded Prolo scale[9, 10, 13, 17]. ]. The 5-point Likert scales for pain, function, economic status, and medication usage are added to a combined 4- and 20-point scale as used for the IDE study of the Lumbar I/F Cage [8]. This study defined a patient’s result as a clinical success when the final rating was excellent (17 to 20 points) or good (13 to 16 points), or fair (9 to 12 points) if the patient achieved a minimum improvement of three points or more in the combined 20 point scale. Fusion status was recorded for each surgically treated segment at each follow-up interval. All patients had two year follow-up plain static radiographs including Anterior-posterior, lateral, and Ferguson views. Figure 1 shows a typical serial AP Ferguson radiographs demonstrating a solid fusion. Figure 2 shows the features of a pseudarthrosis with discontinuity in the bone graft and lucency adjacent to the cage. 48 patients had thin cut helical CT scanning of the involved lumbar segments at an average 26 mos.(range 8-48mos) following the index PLIF operation. The CT scans were examined by a radiologist who was blinded with regard to the clinical history and plain radiograph findings. The most common type of pseudarthrosis was the locked pseudarthrosis. Locked pseudarthrosis is a lucency in the mid portion of the graft within the cage with boney continuity from the endplate to the graft. Figures 3 and 4 shows typical Helical CT findings for fused and non-healed levels. Fusion success was defined by previous published parameters[9, 10, 13]. A level was regarded as fused if there was radiographic evidence of bone bridging the disc space without lucency. If a lucency was seen to extend across the cage, the level was considered not fused. For patients undergoing multiple-level fusion, all surgically treated segments must be fused for the patient to be considered a fusion success. Statistical Methods Demographics, clinical and fusion success were compared to the larger reviewed patient group, the IDE study[9], and the ten year follow-up study[10] using chi square analysis. Results Of the 90 9mm cage patients, there were 31 men and 59 women. Average age was 54 (range 34 to 79). Insurance was workers compensation in 18 and private insurance in 68. 22 patients had a positive smoking history with an average 21-pack/year histories. There were 75 prior surgeries in 49 patients at an average interval of 5 years (range 1-18years) for an average 1.5 prior lumbar surgeries per patient. Clinical Success Clinical success was achieved in 74 of 90 patients (82%). This included 33 (37%) Excellent, 34 (38%) Good, and 20 (22%) Fair results. Clinical success was not achieved in three (3%) with Poor results. Fusion Success At the final radiographic evaluation, fusion success was achieved in 87 patients of 90(97%) (194 of 197 levels) and unsuccessful in two patients at the L1-2 level and one patient complicated by infection that never had radiographic or surgical exploration proof of healed fusion. Indications for 9mm Width Lumbar I/F Cage The primary indication for the use of the 9mm I/F Cage was inadequate width exposure of the disc space to accommodate a larger cage. From the operative note, access was limited by previous surgery to one or both side of the disc space in twenty-six, and nerve root anomaly with low position of the exiting nerve root in five. Of 197 PLIF levels, 116 levels had 9mm wide I/F cages. All disc levels could be accessed. Eleven levels received one 9mm wide cage. Six of these had a single 9mm cage at that level and five were accompanied by a wider cage (11X11) on the countralateral side. 104 levels received two 9mm cages. From Table 1, the 9mm cages were used in approximately 50% of the levels where PLIF was used. Complications Two patients had retropulsion of a cage that required revision. In one patient, a lumbar I/F cage at L2-3 level with two cages, retropulsed into the canal in the third month post-operatively. This was identified and watched. At six months, the patient was complaining of back pain and hip pain on the same side as the mal-positioned cage. The cage was removed and the pedicle screw instrumentation removed, and the posterior-lateral fusion was augmented with BMP. A second patient had a 9mm cage retropulsed at L4-5 with radiculopathy and was quickly repositioned two weeks after the index surgery. Both patients with retropulsion went on to heal the fusion. Two patients with a CSF leak and one with an epidural hematoma had a second procedure to correct the problem. Two had infection requiring debridement. Two patients had screw loosening requiring revision with BMP augmentation of the posterior-lateral fusion. There was one patient with persistent post-operative foot drop without cause on extensive diagnostic evaluation. Further Surgery 49 (54%) patients had 69 further surgeries at an average interval of 2 years (range 1 month to 5 years). Twenty-four had posterior implant removal and exploration of fusion with findings of complete fusion healing. Sixteen had additional PLIF level surgery. Helical Cut CT Scans to evaluate fusion 48 patients had Helical CT scans in the course of their care after fusion. Fusion success on the helical CT scan occurred in 37 of 48 patients (186 of 197 levels). Eleven patients had a single level reported as a non-healed fusion. Three of the eleven levels involved at least one 9mm cage. Both of the L1-2 levels with helical CT non-union had two 9mm cages. The third case was an L5-S1 with one 9mm cage and an 11mm cage. Eight of the eleven non-unions were of the locked pseudarthrosis type. One at L1-2 showed lucency between the graft within the cages and the endplates. The last two were single level fusions complicated by infection that was associated with subsidence of the cages through the endplate. One of the infected cases explored 21 months after the index procedure was found to be a healed fusion. The other was not explored and later died without evidence of complete healing of the involved L5-S1 level. There were four CT scans interpreted as delayed union with 2-3mm subsidence and complete boney union around the cages. Seven of the eleven with non- healed levels on CT were explored within 6 months and found to have a healed fusion. One L1-2 nonunion was explored and found to have very little motion, and was treated with pedicle screw implant removal and BMP augmentation of posteriolateral fusion. This did not heal and recently the patient was revised with pedicle screw instrumentation and BMP augmentation of posteriolateral fusion. The second L1-2 non-union was revised once with pedicle screw instrumentation and BMP augmentation of posteriolateral fusion, then subsequently with pedicle screw implant removal and BMP augmentation of posteriolateral fusion. Figure 5 is a Helical CT scan assessed as non-healing fusion L1-2. The helical CT evaluation before re-exploration was compared with the findings at fusion exploration. Thirteen patients had Helical CT scan within six months of fusion exploration. Four of thirteen patients had a non-healed fusion level on CT. At exploration, one patient had a retropulsed non-healed cage removed and BMP added to the posterolateral fusion. The other twelve of the scanned patients (26 levels) were considered fused by exploration. Of all 35 patients explored (68 levels) with VSP implant removal, 61 levels (90%) were considered fused. Discussion Outcome measures At the time of the original Brantigan carbon fiber cage IDE in 1990, there was little agreement regarding optimum outcome measures in a spinal fusion study. To our knowledge, the Prolo Scale has not had a thorough examination of its statistical validity. However, the Prolo Score has been widely accepted for study of interbody fusion cages and is used in this report to maintain continuity with the original IDE and the 10 year follow-up of the Lumbar I/F Cage. We believe the Prolo Scale produces results that may be compared with literature reports of outcomes over many years [9, 10]. Demographic profile This cohort had a variety of socioeconomical and etiological backgrounds similar to the IDE study for the Lumbar I/F Cages. In the IDE study there were 326 prior surgeries in 170 patients (average 1.9 surgeries per patient). Statistical analysis indicated that the following factors had no effect on clinical or fusion success: gender, numbers of levels treated surgically, previous surgery, smoking history, and obesity [9]. The rate of previous surgery of the current 9 mm cage group was significantly less than the original IDE cage study (chi square 24.7616, p= 0.0001). The demographic profile should not adversely bias our conclusions regarding clinical and fusion success. Biomechanics When the Lumbar I/F Cage was first designed, the optimal dimensions were unknown. It appeared that the average width that would allow bilateral cages to be placed was 11 mm. Possible failure modes included subsidence into the vertebral bodies (which argued for maximal width of the struts) and fusion failure (which argued for maximal width of the opening for bone graft surface area). In the design of the 11 mm cage prototypes tested mechanically in cadaver spines the two concerns were given equal weight: the opening was 5.5 mm wide and the strut width was 5.5 mm (or 2.75 mm each) 15. This design performed well in cadaver spines and bore physiologic loads without subsidence. In the design of larger and smaller cage sizes, the width of the struts was retained at 2.75 mm because of the concern for subsidence if the support area had been reduced. Thus the cage 9 mm in width has an opening for bone graft of 3.5 mm wide, proportionately 18% narrower overall than the 11 mm cage but 36% smaller in surface area for bone graft. The 9 mm cage has a bone surface area 54% less than the 13 mm cage. Because bone healing was potentially compromised with the 9 mm side cages by the proportionally greater reduction in bone graft surface area, it is clinically important to study the clinical and fusion success of these narrow cages. Biomechanical studies of the Lumbar I/F cages demonstrated several differences between the 9mm and the 13mm cages [16]. In static compression, the larger cages appeared stronger. With a 45-degree compressive shear force, the 9 mm cages were stronger. In fatigue tests, the 9 mm cages were stronger. Fatigue strength is defined as the load that the cages can bear to 5 million loading cycles without failure. Using two cages, the fatigue strength has a two-fold safety factor compared with maximum loads of daily living. Using a single cage, the fatigue strength safety factor is nonexistent over the maximum loads of daily living. The results of this 9mm cage study are important because the narrow cages do not have a negative clinical effect on bone healing judged by fusion success. A second major finding of this study is mechanical failure did not occur with 9mm cage usage. Two fusion failures at the L1-2 level occurred with two 9mm cages. Our previous study showed an unexpectedly low fusion success of 7 of 12 cases (59%) at L1-2, making this level an independent risk factor for failure [18]. Clinical success The IDE study for the Lumbar I/F Cage reported clinical success in 79 of 91 (86.8%). This included 52% excellent, 33% good, 15% fair and no poor results [9]. In the ten-year follow-up report, clinical success was maintained in 29 of 33 (87.8%)[10]. When the clinical results of our larger clinical series were compared with the IDE results using the chi-square statistic, clinical success rates were not statistically different (chi-square = 0.555, p=0.4563). Clinical success in the current 9 mm cage series was not statistically different at 81% from the IDE study (chi-square = 0.4518, p=0.5015). The 9mm cohort includes fewer Excellent and more Fair results than the IDE study. Fusion success Santos et al. reported that helical CT scans of carbon fiber cage anterior fusions identified twice as many fusion failures as seen on plain radiographs [19]. These were anterior fusions without posterior instrumentation and the results do not directly relate to the PLIF and VSP construct used in this cohort. The authors have found the Helical CT scan to be as valuable as the plain radiographs in detecting non-union. In our practice, helical CT scanning was used in clinically symptomatic patients at an appropriate interval following surgery, to further evaluate questionable lucencies through the fusion cage seen on plain radiographs. Overall, comparisons of Helical Ct scans with the fusion results at exploration indicated a sensitivity of 89.9%, specificity of 100%, a positive predictive value of fusion of 100%, and an overall accuracy of 95%. The Helical CT scan correctly confirmed the radiographic identification of the three patients with non-union. Comparison of radiographs including the Ferguson views of each fused level with exploration yielded a sensitivity of 92.6%, a specificity of 100% and a positive predictive value of fusion of 100%. These results indicate that the radiologic fusion interpretation as defined in this study is sufficiently accurate to be used for the assessment of fusion status. Although this study had independent review of CT scan results, it did not include an independent review of the plain radiographs. These were reviewed by the authors. The accuracy of our radiographic observations was confirmed by surgical exploration. In the IDE study of the Lumbar I/F Cage, fusion success was achieved in 91 of 91 (100%) of the degenerative disc disease cases and 176 of 178 (98.9%) of the entire study group [9]. In the 10- year follow-up of the IDE study, Fusion success was reported in 37 of 37 patients (100%) at 24 months and in 29 of 30 (96.7%) at 10 years [10]. When the fusion results of our larger series are compared with the IDE results using the chi-square statistic, the fusion rates were not statistically different (chi-square=0.1414, p=0.7069). When the fusion results of the 9mm group are compared with our larger clinical series(chi-square=0.847, p=0.5358), the IDE study (chi-square=1.5942, p=0.2067), and the 10 year follow-up (chi-square=0, p=1.000), the fusion rates were not statistically different. Conclusion Fusion success is not diminished by the use of 9 mm width Lumbar I/F Cages. The clinical results may be somewhat diminished. The authors continue to recommend routine use of bilaterally placed Lumbar I/F cages of maximal width allowable. If there is any question about width of exposure or safe retraction of a nerve root, a surgeon should not hesitate to use a 9mm cage at that position.

Figure 1, Serial Ferguson views of two 9mm cages demonstrate progression to excellent fusion at 1 year on radiographs. Arrows demonstrate bridging bone across the interspace through the fusion cage.

Figure 2,Serial Ferguson views of L4-5 show two 9mm cages and a locked pseudarthrosis with lucency through mid-substance of each interbody graft. Arrows point to bridging bone interrupted by lucency at one year post-op. This was a healed fusion at exploration.

Figure 3, Helical CT at 9 months post-op shows bridging bone across interbody space on coronal and sagittal reconstructions at three interspaces.

Figure 4, 4A Helical CT Coronal reconstruction with arrow demonstrates L2-3 bridging bone across interbody space and a single 9mm cage. 4B Sagittal reconstruction: large arrow demonstrating subsidence into the L3 body. There is lucency between graft, cage, and endplate of L2 and L3. A thin arrow points to bridging bone posteriorly across the interspace. This was a healed fusion at exploration. Clinically this was a delayed fusion after subsidence of the 9mm cage into the body of L3.

Figure 5, A 73 year old woman with previous L3-S1 healed fusion has non-union of L1-2. This level was revised with pedicle screws, BMP, and 14 mos. later revised with implant removal and BMP again. 5A Coronal CT reconstruction reveals lucency surrounds both interbody cages at L1-2. 5B Sagittal CT reconstruction: superior arrow demonstrates lucency around the cage and inferior arrow shows subsidence into L2 body. Metallic dot in the disc space is the cage position marker.

References

1. Cloward, R.B., The treatment of ruptured lumbar intervertebral discs by vertebral body fusion. I. Indications, operative technique, after care. J Neurosurg, 1953. 10(2): p. 154-68. 2. Cloward, R.B., Lesions of the intervertebral disks and their treatment by interbody fusion methods. The painful disk. Clin Orthop, 1963. 27: p. 51-77. 3. Cloward, R.B., Spondylolisthesis: treatment by laminectomy and posterior interbody fusion. Clin Orthop, 1981(154): p. 74-82. 4. Cloward, R.B., Posterior lumbar interbody fusion updated. Clin Orthop, 1985(193): p. 16-9. 5. Brantigan, J.W., B.W. Cunningham, K. Warden, P.C. McAfee, and A.D. Steffee, Compression strength of donor bone for posterior lumbar interbody fusion. Spine, 1993. 18(9): p. 1213-21. 6. Brantigan, J.W., P.C. McAfee, B.W. Cunningham, H. Wang, and C.M. Orbegoso, Interbody lumbar fusion using a carbon fiber cage implant versus allograft bone. An investigational study in the Spanish goat. Spine, 1994. 19(13): p. 1436-44. 7. Brantigan, J.W., Pseudarthrosis rate after allograft posterior lumbar interbody fusion with pedicle screw and plate fixation. Spine, 1994. 19(11): p. 1271-9; discussion 1280. 8. Brantigan, J.W. and A.D. Steffee, A carbon fiber implant to aid interbody lumbar fusion. Two- year clinical results in the first 26 patients. Spine, 1993. 18(14): p. 2106-7. 9. Brantigan, J.W., A.D. Steffee, M.L. Lewis, L.M. Quinn, and J.M. Persenaire, Lumbar interbody fusion using the Brantigan I/F cage for posterior lumbar interbody fusion and the variable pedicle screw placement system: two-year results from a Food and Drug Administration investigational device exemption clinical trial. Spine, 2000. 25(11): p. 1437-46. 10. Brantigan, J.W., A. Neidre, and J.S. Toohey, The Lumbar I/F Cage for posterior lumbar interbody fusion with the Variable Screw Placement System: 10-year results of a Food and Drug Administration clinical trial. Spine J, 2004. 4(6): p. 681-8. 11. Miura, Y., S. Imagama, M. Yoda, H. Mitsuguchi, and H. Kachi, Is local bone viable as a source of bone graft in posterior lumbar interbody fusion? Spine, 2003. 28(20): p. 2386-9. 12. Hashimoto, T., K. Shigenobu, M. Kanayama, M. Harada, F. Oha, Y. Ohkoshi, et al., Clinical results of single-level posterior lumbar interbody fusion using the Brantigan I/F carbon cage filled with a mixture of local morselized bone and bioactive ceramic granules. Spine, 2002. 27(3): p. 258-62. 13. Stromberg, L., J.S. Toohey, A. Neidre, M. Ramsey, and J.W. Brantigan, Complications and surgical considerations in posterior lumbar interbody fusion with carbon fiber interbody cages and Steffee pedicle screws and plates. Orthopedics, 2003. 26(10): p. 1039-43. 14. Molinari, R.W. and T. Gerlinger, Functional outcomes of instrumented posterior lumbar interbody fusion in active-duty US servicemen: a comparison with nonoperative management. Spine J, 2001. 1(3): p. 215-24. 15. Tullberg, T., B. Brandt, J. Rydberg, and P. Fritzell, Fusion rate after posterior lumbar interbody fusion with carbon fiber implant: 1-year follow-up of 51 patients. Eur Spine J, 1996. 5(3): p. 178-82. 16. Serhan, H. Presentation of mechanical testing of the Brantigan Cage. in the Orthopaedics and Rehabilitation Devices Advisory Panel Meeting of the Department of Health and Human Services, Public Health Service,Food and Drug Administration. December 11, 1997. Bethesda, Maryland. 17. Prolo, D.J., S.A. Oklund, and M. Butcher, Toward uniformity in evaluating results of lumbar spine operations. A paradigm applied to posterior lumbar interbody fusions. Spine, 1986. 11(6): p. 601-6. 18. Fogel, G.R., J.S. Toohey, A. Neidre, and J.W. Brantigan, Outcomes Of L1-2 Posterior Lumbar Interbody Fusion With the Lumbar I/F Cage And the Variable Screw Placement System Reporting Unexpected Poor Fusion Results at L1-2. Spine, 2005. 19. Santos, E.R., D.G. Goss, R.K. Morcom, and R.D. Fraser, Radiologic assessment of interbody fusion using carbon fiber cages. Spine, 2003. 28(10): p. 997-1001.

Table(s) in MS Word

Table 1 Current Fusion Levels 9mm by % of Levels Level (PLIF) Total % Total % Level Level Treated L1-2 6 3% 4 67% 1Level 67 55% L2-3 33 17% 16 48% 2 Level 36 30% L3-4 47 24% 33 70% 3 Level 14 12% L4-5 60 30% 40 67% 4 Level 4 3% L5-S1 51 26% 23 45% Total Levels 197 116 fused With adjacent 15 13.2% Lami With IR 7 6.1% * IR= Implant removal; Lami=Laminectomy; PLIF= posterior Lumbar Interbody Fusion, PLF= posteriolateral fusion. * Disclosure - Conflict of Interest Form(s)

THE SPINE JOURNAL

A copy of this form must be uploaded along with your manuscript FOR EACH AUTHOR before an article will be considered for publication.

DISCLOSURE FOR (article title):

Outcomes of L1-2 Posterior Lumbar Interbody Fusion with the Lumbar I/F Cage and the Variable Screw Placement System: Reporting Unexpected Poor Fusion Results at L1-2 FDA APPROVAL STATUS - If a device or drug requiring FDA approval is identified as an important component of your article, you must indicate the FDA status for use as it will be discussed. Please list the name of the devices and drugs requiring FDA approval and check the appropriate status for use as it is discussed in the article.

a. Device/Drug: Lumbar I/F Cage______XX__Approved for this indication ___Approved but not for this indication ___Investigational/Not Approved

b. Device/Drug: VSP Variable Screw Placement System______XX _Approved for this indication___ Approved but not for this indication ___ Investigational/Not Approved

c. Device/Drug ______Approved for this indication ___Approved but not for this indication ___ Investigational/Not Approved

___This article does not discuss or include any applicable devices/drugs.

RESEARCH FUNDING - Please indicate the source of any funding received for the research discussed in this manuscript.

______None______Grant#______Indicate type: ___Federal/state agency ___Nonprofit foundation ___ Commercial entity

______Grant#______Indicate type: ___Federal/state agency___ Nonprofit foundation ___Commercial entity

CONFLICT OF INTEREST - The North American Spine Society strives to ensure balance, independence, objectivity and scientific rigor in all its publications. All authors are expected to disclose any financial interest or other relationship with the manufacturer(s) of any commercial product(s) and/or provider(s) of commercial services discussed in an article. The intent of this disclosure is not to prevent an author with a significant financial interest or other relationship from being published, but rather to provide readers with information on which they can make their own judgments. It remains for the audience to determine whether the author’s interests or relationship may influence the article with regard to exposition or conclusion. The Spine Journal and the North American Spine Society do not view the existence of these interests or commitments as necessarily implying bias or decreasing the value of your article.

XX _ Do not have a financial relationship that creates, or may be perceived as creating, a conflict related to this article.

- OR -

___ Do have a financial relationship that creates, or may be perceived as creating, a conflict related to this article (please indicate type and provide name of commercial organization):

___ Grant Research Support (______Grant #______)

___ Consultant (______)

___ Speaker's Bureau (______)

___ Stockholder (______)

___ Board Member (______)

___ Other Support (______)

NAME OF AUTHOR Guy R. Fogel MD______

* Disclosure - Conflict of Interest Form(s)