You are on page 1of 102

Orthop Clin N Am 38 (2007) ix

Preface

Anthony A. Stans, MD Guest Editor

Scoliosis treatment has undergone tremendous change and innovation during the past 10 years. From elementary distraction-compression instrumentation methods have evolved sophisticated instrumentation systems that give the surgeon tremendous power to move individual vertebrae three-dimensionally through space. Even as instrumentation techniques become more advanced, there is renewed interest in nonoperative scoliosis treatment. At risk of becoming a lost art, cast treatment is achieving encouraging results in the extremely challenging infantile and juvenile patient groups. Just as a national, randomized, prospective study is getting under way to question whether brace treatment is better than no treatment at all, centers are reporting favorable results in compliant patients treated with a brace fashioned by a skilled orthotist. The aims of this issue of the Orthopedic Clinics of North America were to identify several controversial and exciting aspects of scoliosis treatment and to recruit nationally and internationally renowned authorities to share their wisdom and

recommendations on these topics. Contributing authors include two past Scoliosis Research Society presidents, and several authors who will undoubtedly be future SRS presidents. It has been a privilege to work with the wonderful authors who have contributed to this edition on scoliosis, and I would like to acknowledge them and thank them for their hard work and excellent articles. Deb Dellapena, the editor of this issue, has been patient, encouraging, and a pleasure to work with. Finally, I would like to thank my familydLena, Haakon, AnnaSoa, and Brittadfor their patience and support through early-morning editing sessions. Anthony A. Stans, MD Division of Pediatric Orthopedics Department of Orthopedic Surgery Mayo Clinic 200 First Street SW Rochester, MN 55905, USA E-mail address: Stans.Anthony@mayo.edu

0030-5898/07/$ - see front matter 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.ocl.2007.07.004

orthopedic.theclinics.com

Orthop Clin N Am 38 (2007) 469475

Advances in Scoliosis Brace Treatment for Adolescent Idiopathic Scoliosis


William J. Shaughnessy, MD
Department of Orthopedic Surgery, Mayo Clinic and Mayo School of Medicine, Rochester, MN 55905, USA

Brace treatment of adolescent idiopathic scoliosis remains an excellent option for skeletally immature individuals with scoliosis curve magnitudes between 20 and 40 . Among options available for managing spinal disorders in children are observation, bracing, and surgery. The use of a spinal orthosis was rst documented in the sixteenth century by Pare [1]. The lack of imaging techniques and the absence of an understanding of spinal growth led to unpredictable results with orthotic treatment until the second half of the twentieth century [2,3]. As the natural history of scoliosis and the eects of growth and development became better understood, the indications for bracing narrowed considerably. Bracing is signicantly limited by such factors as neuromuscular disease, skin insensitivity, spasticity, osteoporosis, and congenital spinal disorders. Despite these limitations and with a more rened understanding of treatment indications, brace treatment continues to play an important role in the management of adolescent idiopathic scoliosis. The goal of brace treatment is to prevent curve magnitude progression and to avoid the need for surgical spinal arthrodesis. Surgery is indicated for Cobb angles of greater than 45 to 50 . While other nonoperative treatments have been reported, none have shown ecacy in treating adolescent idiopathic scoliosis. Methods that have not shown success include electrical stimulation, exercise, and manipulation [4]. Brace treatment is used to prevent curve progression during growth. As such, orthotic treatment ends with skeletal maturity and is not useful

in adults. The brace must allow continued growth of the spine while simultaneously preventing the spinal deformity from becoming worse. It should be assumed that scoliosis will not be cured by brace treatment. An acceptable brace result is for the deformity to be no worse at skeletal maturity than when brace treatment started. During the past 20 years there have been many studies of brace treatment for adolescent idiopathic scoliosis. In addition to the Milwaukee brace, multiple variants of the thoracolumbar sacral orthosis (TLSO) brace have been described and reported on, each variant being named for the city of its origin [521]. Most of these studies support the eectiveness of bracing in preventing curve progression and the subsequent need for surgery [5,6,810,13,14,1618,20,21]. Several studies have suggested that bracing may not be eective [11,12,15,19,21]. While these studies do not show brace ecacy, they have been helpful. Like electrical stimulation, ineective bracing reveals the natural history of scoliosis, so these studies have helped to rene the natural history and indications for successful bracing. Natural history Few studies have attempted to document the natural history of progression in adolescent idiopathic scoliosis. To correctly judge the ecacy of any treatment, some knowledge of the natural history is necessary. Several studies have identied the population at risk of progression [2227]. This includes those with curves of greater than 20 at the time of peak adolescent skeletal growth, classically reported as Risser 0 or 1 maturity and premenarchal to 1 year postmenarchal in females
orthopedic.theclinics.com

E-mail address: shaughnessy.william@mayo.edu

0030-5898/07/$ - see front matter 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.ocl.2007.07.002

470

SHAUGHNESSY

[2527]. For this group at risk, Rogala and colleagues [24] found progression among 79% of immature patients with curves between 20 and 30 . Bunnell [26] found progression of 5 or more in 68% of those with a Risser sign of 0. Lonstein and Carlson [27] reported the same 68% progression rate of greater than 5 for curves of 20 to 29 in those with a Risser sign of 0 or 1. Nachemson and colleagues [18] performed a randomized, multicenter study comparing brace treatment to observation in 10- to 15-year-old females with thoracic curves. The investigators documented progression of at least 6 in 66% of the observation group. The best available data show, therefore, that 66% to 79% of immature females with curves of 20 to 35 will progress more than 5 if observed. Progression has been shown to be greater for curves of larger magnitude and less probable for curves smaller than 20 , Risser 3, or greater maturity, and for females more than 1 year postmenarchal [27]. It is not known what percentage of the at-risk population would progress to surgery if left untreated. The ultimate goal of brace treatment is not so much to prevent 5 of progression as it is to prevent the need for surgery. Virtually all studies, including those noted above, use curve progression of more than 5 , rather than surgical rates, as an endpoint. Unfortunately, none of these natural history studies followed untreated patients to surgery or to curve magnitudes of more than 45 , when surgery would ordinarily be recommended. It can be reasonably assumed, however, that this information can be gleaned from well-done brace studies with poorly compliant patients or ineective braces. Such studies suggest that 60% to 79% of immature females with curves larger than 25 may progress to surgery [28]. It can also be assumed that results with lower progression and surgical rates than these indicate a benecial treatment eect and demonstrate the ecacy of brace treatment. While a randomized, blinded study would be helpful, the ethical and logistical diculty in performing such a study may be insurmountable. Indications for brace treatment The results of bracing in early studies were inconsistent because many individuals now known to be at low risk for progression were included [29]. These included patients with skeletal maturity at brace initiation and small curve magnitudes that were not likely to progress. Some studies also

included individuals with large curve magnitudes and nonidiopathic curves, which are not likely to respond to brace treatment. As the natural history of scoliosis becomes better understood, the identication of those patients at risk of progression becomes more accurate [27,3032]. Patients at greatest risk of progression, and therefore those most likely to benet from bracing, include growing children (Risser 02 and, if female, either premenarchal or less than 1 year postmenarchal) with primary curve magnitude of 25 to 40 . Patients with curve magnitudes of between 20 and 25 with documented progression should also be considered for brace treatment [17,29]. Initiation of brace treatment involves tting the orthosis, obtaining a standing posterior-anterior radiograph with the patient wearing the brace, and increasing the number of hours in the brace over the course of 1 to 2 weeks. It is important that the curve magnitude on the rst in-brace radiograph be less than 50% of the original, prebrace curve magnitude [8]. Time-in-brace is increased 1 to 4 hours each day until full-time brace wear is achieved. Full-time wear is prescribed initially for all patients, except those using the Charleston and Providence nighttime bending braces. Patients are encouraged to be active in sports and allowed to be out of the brace if the sport cannot be performed in the brace. Decreased orthotic eectiveness in overweight patients has been demonstrated and may be a contraindication to bracing. ONeill and colleagues [33] showed that individuals with body mass indices greater than the 85th percentile were three times more likely to fail brace treatment than those who were not overweight. Brace treatment is continued until growth is completed. This is determined by a variety of indicators: typically unchanged height measurements on two successive dates 6 months apart, 18 to 24 months postmenarchal, Risser 4 status, or skeletal maturity on bone-age determination. Bracing results The use of a brace for scoliosis is complex, in part because success depends on more than just the skill of the surgeon. Quite the contrary, bracing success depends also on the dedication of the patient and family, the skill of the orthotist, the ecacy of the brace, growth, compliance with wear, regular follow-up evaluations, and an environment that encourages the brace to be worn for

ADVANCES IN SCOLIOSIS BRACE TREATMENT

471

what may be a prolonged period. Given this multitude of factors aecting success, it is not at all surprising that the published results of bracing span a continuum ranging from seemingly ineective to excellent. In general, most brace studies show a similar trend of curve magnitude improvement in the brace, followed by slow loss of correction until the end of brace treatment, when the curve is similar to that at the start of bracing [5,25,34,35]. Several studies have also identied a trend of decreasing brace ecacy with increasing curve size [8,17,28]. In most reported series, full-time Milwaukee brace treatment produces better results than natural history studies. Carr and Noonan reported failure rates requiring surgery in 39% and 42%, respectively [5,19]. While Noonan questioned the ecacy of brace treatment, surgical rates among his patients were still considerably better than natural history studies and several recent brace studies in the literature. In one of the most complete brace studies to date, Lonstein and Winter [17] reported a 40% rate of curve progression of more than 5 and a surgical rate of 22% in a study of Risser 0-to1 Milwaukee brace patients. The majority of patients had prebrace curves between 20 and 39 . The investigators noted a signicant dierence between the reported radiographic failure rate of 40% and their own natural history study predicting 68% progression. Others have reported on the use of the Milwaukee brace, but the studies are limited by incomplete follow-up, small numbers, the inclusion of nonidiopathic scoliosis and patients of ages now known to not benet from bracing [3436]. Patient self-image issues, noncompliance, and dissatisfaction with the cosmetic appearance of the neck ring ultimately led to the development of underarm braces [37,38]. Underarm braces include a variety of devices collectively known as TLSOs. In general, the results of TLSO treatment are similar to the Milwaukee brace studies. Several series have evaluated the results of such devices. Bassett and colleagues [39] studied a mixed population of juvenile and adolescents with curves of 20 to 39 treated with the Wilmington TLSO and found 30% progressed 5 or more. Emans and colleagues [8] reported 295 patients, 4 to 18 years of age, treated with a Boston brace. Only 12% of patients required surgery, although the failure rate was greater for curves greater than 40 when bracing was started. Emans also noted less favorable results in noncompliant patients. More

recently, Spoonamore and colleagues [21] reported the results of a study using the Rosenberger brace, a front-opening TLSO. The investigators found 56% of their patients progressed more than 5 and 31% required surgery or met surgical criteria. While the investigators acknowledge that the Rosenberger brace is eective in preventing radiographic and clinical failure in some, they noted failure in others. Compliance with brace treatment was not reported. The investigators noted this as a possible explanation for the unsatisfactory results. Compared with nearly identical brace studies showing progression and surgical rates of 80%, however, the Rosenberger brace appears to produce good results. The reasons for the disparity among bracing results are unknown, but dierent rates of brace compliance are likely the major culprit. Driven largely by the desire of many patients to avoid daytime and school-time bracing, nighttimeonly braces have been developed that require only 8 to 10 hours of wear. The Charleston and Providence braces use bending or aggressive molding to apply corrective forces to the spine [6,20]. Several studies have reported the ecacy of these two novel braces. DAmato and colleagues [6] found that only 21% of patients progressed using the Providence brace, provided the curve apex was caudal to T8. Katz and colleagues [16] compared the Charleston to the Boston brace. Both braces were more eective with smaller prebrace curves of 25 to 35 than with larger curves. More than 5 of progression was noted in 29% and 47% of the Boston and Charleston groups respectively. With larger prebrace curves of 36 to 45 , both braces did less well. The investigators found greater than 5 of progression in 43% of the Boston group and 83% of the Charleston patients. The investigators recommend the Boston brace for large curve magnitudes and those with thoracic curves [16]. Again recognizing that progression of 5 may be an arbitrary measure of bracing failure and that radiographic failure may not indicate clinical failure, Gepstein and colleagues [10] found no dierence in surgical rates between the Charleston brace (11%) and a TLSO (13%). Surgical rates for both braces, however, are well below that expected from natural history studies. The only published prospective controlled brace study was performed by Nachemson and Peterson [18]. The investigators compared TLSO bracing to observation in a group of 10- to 15year-old (skeletal age) girls with idiopathic scoliosis and single curves between 25 and 35 . Only

472

SHAUGHNESSY

36% of the braced patients showed progression of more than 5 , compared with 66% in those observed. Using survival analysis, the TLSO was found to be statistically eective in altering the natural history (P!.0001). While the study is routinely cited as clearly and dramatically demonstrating the eectiveness of bracing, it has been criticized for not being randomized or blinded and for the inclusion of only single thoracic curves of limited size, up to 35 [21]. Unfortunately, there are many problems with most bracing studies in the literature. Juveniles, adolescents, and mature individuals are included. Dierent curve types, magnitudes, and lengths of follow-up are reported. Nonidiopathic scoliosis is occasionally included. Even the landmark study by Nachemson has been criticized for including thoracic curves and not including curves of greater than 35 . To assess the eectiveness of each orthosis and to compare one study to others requires consistent study parameters. The Scoliosis Research Society (SRS) has recently published a stringent, uniform set of inclusion criteria for bracing studies [29]. The standards for adolescent idiopathic scoliosis brace studies include an age of 10 years and older, Risser 0 to 2, primary curve size of 25 to 40 , no prior treatment, and, if female, either premenarchal or less than 1 year postmenarchal. The SRS also denes progression and sets standards for reporting curve patterns, curve rotation, compliance, and follow-up. It is to be hoped that additional studies using the SRS criteria will allow for an objective comparison of future brace studies. The rst and only published study using the recent, strict SRS criteria is a comparison of the TLSO to the Providence nighttime orthosis at a single institution [28]. The poor results for both braces raise interesting questions. In the TLSO group, 85% showed progression of more than 5 and 79% of braced patients required surgery. For the Providence brace, 69% progressed and 60% required surgery. The results are so similar to natural history studies that it can be argued the brace programs cited are ineective and that the study is a natural history study of adolescent idiopathic scoliosis. Similar assumptions were made for patients treated with electrical stimulation in the past, leading those so treated to be included in observation rather than treatment populations [18,27,40,41]. The investigators speculate that the poor results might be due to demographics, genetics, referral patterns, or compliance, but the reasons remain unknown. The

extreme variation in reported brace success rates suggests that some braces and brace programs are very good and that others have no eect on the natural history of adolescent idiopathic scoliosis. Further study is necessary to determine why one center reports a surgical rate of 79% and another 22%.

Bracing in males Brace treatment in boys may not be as successful as in girls. For boys willing to wear an orthosis, however, brace treatment can be successful. Most studies of scoliosis orthosis treatment include both males and females, but the number of males is small, statistical power is lacking, and the results are dicult to interpret. Karol [15] reported that brace treatment in males was ineffective largely because of poor compliance. A recent study conrmed the decreased compliance in males when compared with females but found bracing useful in both genders, with progression of more than 5 in 31% of the compliant males and 21% of the compliant females [42]. Both results are two to three times better than expected from natural history and other brace studies.

Compliance As with other medical treatments, success depends upon treatment actually being performed. In the case of bracing for adolescent idiopathic scoliosis, it is often incorrectly assumed that brace patients are compliant with their treatment. In fact, compliance with bracing may run the range from absolute nonuse to full-time bracing. The experience of wearing a brace fulltime is substantially dierent, and more obvious to others, than that of taking daily medication or adhering to a diet. Unlike other orthopedic surgical treatments, such as surgery, braces are accessible and easily removed by the patient. Poor compliance is not surprising given the negative eects of bracing on self-image [37]. Those who do not wear their braces should be expected to have results similar to natural history studies. Compliant patients are likely to show the best results and the most signicant dierence when compared with natural history. Given this reality, compliance is a major determinant of bracing success. Compliance may explain the dierence between brace treatment results in males and those in females [15,42].

ADVANCES IN SCOLIOSIS BRACE TREATMENT

473

Compliance, the ability of the patient to wear a brace for the prescribed time, has a dramatic eect on brace treatment eectiveness. Emans and colleagues [8] noted poor results for noncompliant patients in his Boston brace study. Wiley and colleagues [43] also found that brace wear of more than 18 hours per day led to less progression than if the brace was worn 12 hours per day or less. In contrast, Allington and Bowen [44] found no dierence between full-time and part-time bracing using the Wilmington brace. Until recently, compliance was dicult to measure and often relied upon patient or parent self-reporting. With the availability of pressure transducers and temperature sensor loggers, reliable brace wear records may be recorded. Using instrumented braces, several studies have determined that brace wear varies from 8% to 90% of the prescribed time, averaging only 65% to 75% of the time prescribed [45,46]. Patients and parents over-reported their hours of brace wear by 150%. Additionally, there was an inverse correlation between age and compliance. Ten-year-olds wore their braces 84% of the prescribed time compared with only 77% for 12-year-old patients and 60% for 14-year old patients. Rahman and colleagues [47] reported on the association between brace compliance and outcomes using a temperature sensor and data logger in Wilmington TLSO braces. The investigators found striking dierences between the compliant and noncompliant patients. Among those with radiographic progression more than 5 , patients wore their braces only 62% of the time prescribed. For those who did not progress, compliance was better, logging 85% of the prescribed time (P.004). Additionally, in those who wore their brace more than 90% of the time prescribed, progression of more than 5 was seen in only 11%, better than any other group in the literature. The results indicate that those who are compliant with brace treatment have signicantly more favorable outcomes. Factors aecting brace compliance are not well understood and have not been adequately studied. The role of the treating physician, orthotist, and parents require additional attention. Finally, while the SRS has recommended that all noncompliant brace patients be included in further brace studies, it will be important to distinguish between brace failure, and failure of the patient or brace program to use the brace. There is a tendency to label bracing as ineective when noncompliant patients raise the average radiographic progression rates in some studies. Similar standards are not routinely applied

elsewhere in medicine or in orthopedics. As an example, patients who refuse antibiotics for the treatment of osteomyelitis are not usually included in studies of antibiotic ecacy, nor are they cited as failures of drug treatment. While it is useful to know how often brace patients wear their orthosis, the question that physicians, patients, and parents want to know is, If this brace is worn, what will the back look like and can surgery be avoided? Summary Brace treatment is the treatment of choice for some, but not all, patients with adolescent idiopathic scoliosis. Indications for bracing include curve magnitudes of 20 to 40 in growing individuals with Risser 0 to 2 skeletal maturation and, if female, either premenarchal or less-than1-year postmenarchal status. Natural history and brace studies suggest that 68% to 79% of these individuals show progression of their scoliosis and similar numbers require surgery if untreated. In contrast, those that are compliant with brace treatment show radiographic progression of more than 5 in just 11% and rarely require surgery. The importance of wearing the braced so-called compliancedcannot be overstated. Poor compliance is the most likely source of many poor bracing results but it is dicult to measure as patients and their parents routinely report more brace time than they actually achieve. Bracing is less eective in obese patients and in those with curve magnitudes greater than 40 . Bracing is not eective in skeletally mature individuals or in those with nonidiopathic scoliosis. Good results have been reported with Milwaukee, TLSO, and bending braces. Successful brace treatment prevents radiographic progression of scoliosis and avoids the need for spine surgery. In general, if brace treatment is not eective for a patient or group of patients, one or more of the following problems is likely present: The The The The brace does not t or is not eective. patient is not wearing the brace. time-in-brace is insucient. diagnosis is incorrect.

It becomes the role of the treating physician and orthotic specialist to sort these issues out and to make appropriate changes. Brace treatment is a team activity that requires the active, willing participation of the prescribing physician, orthotist, patient, family, and friends.

474

SHAUGHNESSY

If the physician does not prescribe and follow up, if the orthotist does not properly t the brace, if the patient does not wear the brace, and if the family and friends are not supportive, the treatment will not be eective. The literature discussion about bracing ecacy assigns far too much credit or blame to the orthosis and not nearly enough responsibility to those involved with the prescribing, manufacturing, and, most importantly, the wearing of the brace. References
[1] Pare A. Opera ambrosil parei. Paris: Apud Jacobum Dupuys; 1582 [in French]. [2] Arkin AM. Correction of structural changes in scoliosis by corrective plaster jackets and prolonged recumbency. J Bone Joint Surg Am 1964;46:3354. [3] Barr JS, Buschenfeldt K. Turnbuckel brace: 3 point pressure brace for corrective treatment of ambulatory cases of scoliosis. J Bone Joint Surg 1936;18:7605. [4] Lonstein J, Wilson S, Beattie C. Results of stimulator treatment of 332 cases of adolescent idiopathic scoliosis. Orthopaedic Transactions 1989;13:92. [5] Carr WA, Moe JH, Winter RB, et al. Treatment of idiopathic scoliosis in the Milwaukee brace. Long term results. J Bone Joint Surg Am 1980;62:599612. [6] DAmato CR, Griggs S, McCoy B. Nighttime bracing with the Providence brace in adolescent girls with idiopathic scoliosis. Spine 2001;26:200612. [7] Danielsson AJ, Nachemson AL. Radiologic ndings and curve progression 22 years after treatment for AIS: comparison of brace and surgical treatment with matching control group of straight individuals. Spine 2001;26:51625. [8] Emans JB, Kaelin A, Bancel P, et al. The Boston bracing system for idiopathic scoliosis. Follow-up results in 295 patients. Spine 1986;11:792801. [9] Gabos PG, Bojescul JA, Bowen JR, et al. Long-term follow-up of female patients with idiopathic scoliosis treated with the Wilmington orthosis. J Bone Joint Surg Am 2004;86:18919. [10] Gepstein R, Leitner Y, Zohar E, et al. Eectiveness of the Charleston bending brace in the treatment of single-curve idiopathic scoliosis. J Pediatr Orthop 2002;22:847. [11] Goldberg CJ, Dowling FE, Hall JE, et al. A statistical comparison between natural history of AIS and brace treatment in skeletally immature adolescent girls. Spine 1993;18:9028. [12] Goldberg CJ, Moore DP, Fogarty EE, et al. AIS. The eect of brace treatment on the incidence of surgery. Spine 2001;26:427. [13] Green NE. Part-time bracing of AIS. J Bone Joint Surg Am 1986;68:73842. [14] Howard A, Wright JG, Hedden D. A comparative study of TLSO, Charleston, and Milwaukee braces for idiopathic scoliosis. Spine 1998;23:240411.

[15] Karol LA. Eectiveness of bracing in male patients with idiopathic scoliosis. Spine 2001;26:20015. [16] Katz DE, Richards BS, Browne RH, et al. A comparison between the Boston brace and the Charleston bending brace in AIS. Spine 1997;22:130212. [17] Lonstein JE, Winter RB. The Milwaukee brace for the treatment of AIS. A review of one thousand twenty patients. J Bone Joint Surg Am 1994;76:120721. [18] Nachemson AL, Peterson LE, members of Brace Study Group of the SRS. Eectiveness of treatment with a brace in girls who have AIS. J Bone Joint Surg Am 1995;77:81522. [19] Noonan KJ, Weinstein SL, Jacobson WC, et al. Use of the Milwaukee brace for progressive idiopathic scoliosis. J Bone Joint Surg Am 1996;78:55761. [20] Price CT, Scott DS, Reed FR, et al. Nighttime bracing for AIS with the Charleston bending brace: longterm follow-up. JPO J Pract Orthod 1997;17:7037. [21] Spoonamore MJ, Dolan LA, Weinstein SL. Use of the Rosenberger brace in the treatment of progressive AIS. Spine 2004;29:145864. [22] Duval-Beaupere G. Pathogenic relationship between scoliosis and growth. In: Zorab PA, editor. Scoliosis and growth. Edinburgh (UK): Churchill Livingstone; 1971. p. 5864. [23] Brooks HL, Azen SP, Gerberg E, et al. Scoliosis: a prospective epidemiological study. J Bone Joint Surg Am 1975;57:96872. [24] Rogala EJ, Drummond DS, Gurr J. Scoliosis: incidence and natural history. A prospective epidemiological study. J Bone Joint Surg Am 1978;60:1736. [25] Fustier T. Evolution radiologique spontanee des scolioses idiopathiques de moins de 45 degrees en periode de croissance. Etude graphique retrospective de cent dossiers du Centre de readaptation fonctionnelle des Massues. [Thesis]. Universitie Claude-Bernard, Lyon, France. 1980. [26] Bunnell WP. The natural history of idiopathic scoliosis before skeletal maturity. Spine 1986;11:7736. [27] Lonstein JE, Carlson JM. The prediction of curve progression in untreated idiopathic scoliosis during growth. J Bone Joint Surg Am 1984;66:106171. [28] Janicki JA, Poe-Kochert C, Armstrong DG, et al. A comparison of the thoracolumbosacral orthosis and Providence orthosis in the treatment of adolescent idiopathic scoliosis: results using the new SRS inclusion and assessment criteria for bracing studies. J Pediatr Orthop 2007;27:36974. [29] Richards BS, Bernstein RM, DAmato CR, et al. Standardization of criteria for adolescent idiopathic scoliosis brace studies. Spine 2005;30:206875. [30] Bunnell WP. The natural history of idiopathic scoliosis. Clin Orthop Relat Res 1988;229:205. [31] Lonstein JE. Natural history and school screening for scoliosis. Orthop Clin North Am 1988;19:22737. [32] Peterson LE, Nachemson AL. Prediction of progression of the curve in girls who have adolescent idiopathic scoliosis of moderate severity. J Bone Joint Surg Am 1995;77:8237.

ADVANCES IN SCOLIOSIS BRACE TREATMENT

475

[33] ONeill PJ, Karol LA, Shindle MK, et al. Decreased orthotic eectiveness in overweight patients with adolescent idiopathic scoliosis. J Bone Joint Surg Am 2005;87:106974. [34] Keiser RP, Shuebarger HL. The Milwaukee brace in idiopathic scoliosis. Evaluation of 123 completed cases. Clin Orthop Relat Res 1976;118:1924. [35] Mellencamp DD, Blount WP, Anderson AJ. Milwaukee brace treatment of idiopathic scoliosis. Late results. Clin Orthop Relat Res 1977;126:4757. [36] Edmonsson AS, Morris JT. Follow-up study of Milwaukee brace treatment in patients with idiopathic scoliosis. Clin Orthop Relat Res 1977;126:5861. [37] Clayson D, Luz-Alterman S, Cataletto MM, et al. Long-term psychological sequelae of surgically versus nonsurgically treated scoliosis. Spine 1987;12:9836. [38] Fallstrom K, Cochran T, Nachemson A. Long-term eects on personality development in patients with adolescent idiopathic scoliosis. Inuence of type of treatment. Spine 1986;11:7568. [39] Bassett G, Bunnell W, MacEwen G. Treatment of idiopathic scoliosis with the Wilmington brace. Results in patients with a twenty to thirty-nine degree curve. J Bone Joint Surg Am 1986;68:6025. [40] Akbarnia B, Keppler L. Lateral electrical surface stimulation for the treatment of adolescent idiopathic scoliosis. Analysis based on progression risk. J Pediatr Orthop 1986;6:36970.

[41] Sullivan JA, Davidson R, Renshaw TS, et al. Further evaluation of the scolitron treatment of idiopathic adolescent scoliosis. Spine 1986;11:9036. [42] Yrjonen T, Ylikoski M, Schlenzka D, et al. Results of brace treatment of adolescent idiopathic scoliosis in boys compared with girls: a retrospective study of 102 patients treated with the Boston brace. Eur Spine J 2007;16:3937. [43] Wiley JW, Thomson JD, Mitchell TM, et al. Eectiveness of the Boston brace in treatment of large curves in adolescent idiopathic scoliosis. Spine 2000;25:232632. [44] Allington NJ, Bowen JR. Adolescent idiopathic scoliosis: treatment with the Wilmington brace. A comparison of full-time and part-time use. J Bone Joint Surg Am 1996;78:105662. [45] Takkemitsu M, Bowen JR, Rahman T, et al. Compliance monitoring of brace treatment for patients with idiopathic scoliosis. Spine 2004;29: 20704. [46] Nicholson GP, Ferguson-Pell MW, Smith K, et al. The objective measurement of spinal orthosis use for the treatment of adolescent idiopathic scoliosis. Spine 2003;28:224350. [47] Rahman T, Bowen JR, Takemitsu M, et al. The association between brace compliance and outcome for patients with idiopathic scoliosis. J Pediatr Orthop 2005;25:4202.

Orthop Clin N Am 38 (2007) 477484

Casting and Traction Treatment Methods for Scoliosis


Jacques L. DAstous, MD, FRCS(C)a,b,*, James O. Sanders, MDc
a

Department of Orthopaedic Surgery, University of Utah Orthopaedic Center, 590 Wakara Way, Salt Lake City, UT 84108, USA b Shriners Hospitals for Children Intermountain, Fairfax Road & Virginia Street, Salt Lake City, UT 84103, USA c Shriners Hospitals for Children, 1645 West 8th Street, Erie, PA 16505, USA

Casting for infantile or early-onset scoliosis Casting for scoliosis was common until Paul Harrington introduced eective spinal instrumentation. Since then, casting largely has been relegated to history books along with racks, corsets, and other medieval implements. As its role diminished, knowledge of casting techniques disappeared, experienced surgeons passed on, and the remaining historic relicsdcasting tablesdfell into disuse and were discarded. To a large degree, the abandonment of casting is justied. Instrumentation is now solid and secure, provides excellent curve correction, and allows rapid mobilization and return to activity. A rigid cast can create pressure sores, signicant rib or mandibular deformities, and constrict the chest. The historical term for superior mesenteric artery syndrome is cast syndrome. Many of these problems seem to be the result of indiscriminate casting of all types of scoliosis and improper technique combined with a limited understanding of spinal, and, particularly, chest wall deformities. We have used serial casting in a small group of our patients for several years, and our early results indicate that in selected patients and with proper technique, casting plays a useful role in our treatment armamentarium for early-onset scoliosis.

* Corresponding author. Shriners Hospitals for Children Intermountain, Fairfax Road & Virginia Street, Salt Lake City, UT 84103. E-mail address: jdastous@shrinenet.org (J.L. DAstous).

With few exceptions, we have conned scoliosis casting to patients who had infantile scoliosis and do not believe it has a role in adolescent, neuromuscular, or congenital scoliosis with the rare indications for postoperative stabilization, temporary correction, or immobilization to assess pain relief. Scoliosis casting comes in several varieties. The most commonly used method in the United States is that of Risser [1]. Indeed, body casting in this country often is called Risser casting. Risser developed two types of casts. The initial method used turnbuckles, whereas the latter, termed a localizer cast, used a three-point mold and a pusher. Although it is possible to obtain signicant curve correction with this technique, it does not account suciently for rotational abnormality, and, especially in younger children with exible bones, it can cause signicant rib deformities and chest constriction. Scoliosis in young children has two classications. The traditional classication used by the Scoliosis Research Society divides these curves into infantile and juvenile, based on whether the curve was diagnosed before or after the age of 3 years. The classication of Dickson [2] divides curves into early onset (diagnosed before age 5 years) and late onset (diagnosed after age 5 years). Dicksons logic was that curves occurring after age 5 years were less likely to result in pulmonary failure. We prefer the former classication because of Pehrsson and colleagues [3] work showing increased mortality in the infantile and juvenile curves and because younger age of treatment, in our experience, seems to improve the results.
orthopedic.theclinics.com

0030-5898/07/$ - see front matter 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.ocl.2007.03.006

478

DASTOUS & SANDERS

Infantile idiopathic scoliosis occurs in two basic types: resolving and progressive. Mehta [4] was able to distinguish resolving scoliosis from progressive scoliosis by using the rib vertebral angle dierence (RVAD) on an early supine radiograph. The RVAD nearly follows the 80/20 rule: 80% of curves with an RVAD of 20 or more continue to progress, whereas only 20% of those with lesser angles progress. This was conrmed by other investigators [57]. More recently, Mehta [8] described her results of casting in 136 patients who had infantile scoliosis using the technique of Cotrel and Morel [9] with the philosophy that early rapid growth, if guided by the cast, would assist an initially curved spine to straighten. She grouped her patients into four physiologic patterns: a sturdy phenotype with good muscle mass and tone; a slender phenotype with more delicate features, ligamentous laxity, and more rapidly progressive curves; those with known syndromes; and those with unknown syndromes. Her program consisted of cast changes under anesthesia in younger patients every 8 to 16 weeks until the curve was nearly resolved followed by an underarm brace that may be weaned if the patients curve correction continues. Treatment was related to the patients phenotype; the sturdy patients responded the best, followed by the slender phenotype, and patients with unknown syndromes responded the worst. Even more important was the age at treatment onset. Casting resulted in full correction in 94 patients, whereas 42 had only partial correction. Treatment was much more likely to be successful if started before the age of 2 years. Dr. James Sanders learned the basic Cotrel technique from Dr. Albert Sanders and began using it on several patients presenting with progressive infantile scoliosis. The technique was developed further in conjunction with Dr. Jacques DAstous at the Salt Lake City Shriners Hospital through study and personal instruction by Min Mehta. At the Erie Shriners Hospital, working with Dr. Joe Khoury and an excellent casting team, the technique has become eective and reproducible. The rst requirement is a proper casting table. Some hospitals still have vintage Risser or Cotrel tables, although they are a rarity. Although we initially worked on a Risser frame, the Risser frame is large for a small child. We are indebted to Min Mehta for her design of a smaller, practical frame that leaves the head, arms, and legs supported but the body free. We have identied a few principles that we believe help to create a better cast (Box 1).

We have a routine of doing the initial casts at 2month intervals with a schedule based upon how fast the child is growing. Typically, casts are changed every 2 months for those aged 2 years and younger, every 3 months for those aged 3 years, and every 4 months for those aged 4 years and older. It is particularly helpful to be accommodating for the familiess busy schedules when booking a date for the casting. We do not know when to stop casting, but have aimed for curves nearing 10 and then proceed with brace tting. Occasionally, children are given a brace during the summer months with resumption of casting in the fall. Our early results show signicant promise with the typical left thoracic infantile curve being the most likely to respond; however, we lack comparison with other treatment modalities (eg, bracing, growing rods, or the Vertical Expandable Prosthetic Titanium Rib [VEPTR, SYNTHES Spine, Paoli, PA]), and still requires longer follow-up prior to arriving at any denitive conclusions. As Fig. 5 shows, most patients have a signicant response, particularly younger ones. The most signicant failure is a girl who has Prader-Willi syndrome and a sti double thoracic curve. Until we have further follow-up, we cannot distinguish those patients who are most likely to respond from those who are likely to fail serial casting, although Mehtas nding that younger children respond better corresponds with our experience. (Fig. 6) Traction in early-onset scoliosis Despite the fact that traction is rarely used today, it does remain an important option for the treatment of children with spinal disorders [10]. Halo-gravity traction can restore coronal and sagittal balance in severely decompensated curves and may decrease the neurologic risks associated with the surgical correction of these severe deformities, be it by casting, subcutaneous rods, VEPTR, or denitive instrumentation and fusion. Patients who have severe spinal deformities and secondary cor pulmonale may benet from a period of preoperative traction to allow aggressive pulmonary toilet. Improvements in vital capacity, arterial blood gases, and right heart failure may decrease postoperative complications and can help with patient selection in dicult cases [1113]. Historical background Traction is one of the oldest methods for correction of spinal deformity, going back to

CASTING AND TRACTION TREATMENT FOR SCOLIOSIS

479

Box 1. Guidelines for applying an effective cast One must have a casting table that stabilizes the body for the anesthesiologist while positioning the trunk and extremities for effective casting. Light head halter and pelvic traction assist in stabilizing the patient and in narrowing the body (Fig. 1). A mirror slanted under the table is useful for visualizing the gibbus, the posterior cast, and the molds (Fig. 2). Traction should not be large. Although traction can correct the curve while applied, the position cannot be retained in the cast once traction is released and the body recoils. The purpose of traction is to align the trunk and narrow the torso slightly for good cast tting. Proper casting requires an excellent purchase on the pelvis. Without this basic foundation, the cast will slip, create sores, and fail to support the curve. Only a small amount of padding is necessary with a well-tting cast. Mehta uses direct plaster over crepe paper on the skin with occasional felt pads for signicant bony prominences, but we use a thin layer of webril over a silver-impregnated body shirt in addition to thin layers of felt on signicant bony prominences. If there is a lumbar curve, we believe that exion of the hips to decrease lumbar lordosis facilitates curve correction. The cast must not push the ribs toward the spine and narrow the space available for the lung. Rather, the posteriorly rotated ribs must be moved anteriorly to create a more normal chest conguration (Fig. 3). Counterrotation is applied through the pelvic mold and upper torso or shoulders. The original Cotrel/Morel technique uses rotational straps to obtain correction. In most young children, using ones hands works well, but we use the straps in more difcult curves, particularly of the lumbar spine. Although the Cotrel/Morel technique and Mehtas modications use an over-the-shoulder cast, we have had excellent success staying below the shoulders because most infantile curves have low apices, typically at T10 to T11. Finally, well-positioned anterior and posterior windows are helpful in dynamic curve correction and cast comfort (Fig. 4).

Fig. 1. A proper casting table provides good patient stability with support of the head and lower extremities, full access to the torso, and the ability to provide light traction.

Fig. 2. Rotational correction occurs by rotating the rib or lumbar prominence anteriorly while providing counterrotation through the pelvis and upper torso or shoulders. A slanted mirror is helpful in applying correct molds.

480

DASTOUS & SANDERS

Fig. 5. Early results with cast treatment.

Fig. 3. The correct anterior rotation done manually or with wide straps over the prominence. Forces applied laterally will cause signicant rib and chest deformity and must be avoided.

Hippocrates and possibly even earlier. Sayre described gravity-assisted traction before application of a corrective plaster cast for scoliosis in 1876. Harrington, in the 1950s, devised a form of internal traction, the Harrington rod to obtain and maintain correction of the scoliosis. In 1959, Nickel and Perry designed the halo device to stabilize the cervical spine in patients undergoing cervical spine fusion. Cotrel introduced the

concept of dynamic traction in 1960 to treat adolescent idiopathic scoliosis and loosen up rigid curves before surgery. Kane devised halo-femoral traction in 1967, whereas Stagnara is credited for the introduction of halo-gravity traction in 1969. In the early 1990s, Dubousset combined Stagnaras elongation turnbuckle cast with halo-gravity traction. He believed that this was the best way of obtaining correction in a rigid early-onset scoliosis; to maintain the correction, he used a custom-fabricated modular, hard plastic adjustable cervicothoracolumbosacral orthosis called corset Garchois (Jean Dubousset, MD, personal communication, 1999). Biomechanics The spine is a complex structure composed of vertebrae, ligaments, and intervertebral discs; biomechanically, it behaves as a viscoelastic material with a reproducible stress/strain curve. The soft tissues undergo creep deformation when continuous traction is applied, and this leads to gradual correction of the spinal deformity with realignment of the coronal and sagittal balance [12]. From the work of White and Panjabi, we know that transverses forces are eective for curves less than 50 , whereas traction forces are more eective for curves greater than 50 . A combination of axial traction and transverse forces was the most eective means of correction for small and large curves [14]. Casting, bracing, or surgery is required to maintain the correction.

Fig. 4. (A) The anterior window allows chest and abdominal space while capturing the anterior ribs to prevent their deformity. (B) The posterior window on the concavity allows the curve to settle into the defect and improve the rotation. The window does not cross the midline.

Techniques Halo-femoral and halo-pelvic traction have been used for the treatment of severe scoliosis. Head halterpelvic traction is used for the initial traction radiograph to assess the exibility of the

CASTING AND TRACTION TREATMENT FOR SCOLIOSIS

481

Fig. 6. Younger patients seem to have a better response than older children. (A) Eight-month-old girl who has infantile idiopathic scoliosis just before the start of casting, T6L1 51 , rib phase 2, RVAD 40 . (B) Radiograph showing the correction in the second cast. (C) The same patient at age 3 years with scoliosis completely corrected; careful follow-up is still required.

scoliosis/kyphosis and again during the application of the Risser cast. The discussion is limited to halo-gravity traction because this is the most common type of traction used in our patients who have early-onset scoliosis. Axial skeletal traction to the spine may be applied by a halo. In children younger than 2 years of age, because of their thin calvarium, 10 to 12 pins should be used, and the torque should not exceed 2 inch-pounds or nger tightness [15]. Multiple pins (68 pins) should be used in older children, and 6 to 8 inch-pounds of torque are applied. The pins are placed under general anesthesia if the childs overall condition allows. It is not necessary to shave the scalp or make a skin incision. The skin and subcutaneous tissues are inltrated with 1% xylocaine with epinephrine. The anterior pin is placed 1 cm above the eyebrows or between the supraciliary ridge and the frontal prominence. It is important to avoid the anterior frontal sinus with the most medial pin. Placement too medially also can cause supraorbital or supratrochlear nerve damage. Placement too far laterally (behind the hairline) impinges on the temporalis and masseter muscles, leading to possible skull penetration and diculty with mastication. The posterior pin is placed 1 cm above and posterior to the pinna, below the equator of the skull. The proper ring size allows 1 to 2 cm of clearance between the skin and halo. This

alleviates problems caused by edema and facilitates proper pin care. The screws should be 180 opposite from each other whenever possible. The pins should be retightened once at 24 hours after halo application. Further tightening is dangerous, possibly leading to skull penetration. The exact method of pin care is less important than meticulous daily attention, taking care to remove all encrusted material and to examine for erythema or drainage. The pins are cleansed once a day with half-strength hydrogen peroxide or plain soap and water. Pins should be replaced if erythema and drainage do not improve with meticulous pin care or oral antibiotics. The hair and scalp should be washed at least once a week. Stagnara [16] is credited with the development of halo-gravity traction; it is the mostly widely used method of halo traction. This method uses the patients own body weight as countertraction. Depending on the childs size, 3 to 5 pounds is applied, with daily additions of 1 to 2 pounds to a maximum of 30% to 50% of body weight. During the day, countertraction may be obtained by the use of a wheelchair or walker. At night, we use a bed in maximum reverse Trendelenburg position, with blocks to elevate the head of the bed, which provides a 45 incline (Fig. 7). Alternatively, a pelvic sling or a Circoelectric bed can be used. This method allows the child to be upright during the day, preventing decubiti and

482

DASTOUS & SANDERS

Fig. 7. Bed position.

in

maximum

reverse

Trendelenburg

osteopenia and promoting renal drainage. It has been used safely in children younger than 2 years of age. If a wheelchair is used, it may be necessary to place counterweights on the footrests to prevent tipping backward (Fig. 8). Halo-walker traction is used to allow ambulation in traction (Fig. 9). We have not found it necessary to use a spring scale for traction as described by Sink and colleagues [12], although it may act as a safety mechanism against excessive traction forces.

Fig. 9. Halo-walker traction is used to allow ambulation in traction.

Indications Our indications for the use of halo-gravity traction in early-onset scoliosis are early-onset scoliosis greater than 80 , scoliosis associated with kyphosis, and to loosen up the scoliosis and kyphosis before VEPTR or subcutaneous rod implantation, because both of these methods are kyphogenic and are mechanically disadvantaged in the presence of kyphosis. Contraindications Relative contraindications to halo-gravity traction include short sharp rigid kyphosis, C-spine abnormalities or instability because of their obvious neurologic implications, cranial defects or thin skull, and age younger than 18 months because of pin penetration problems. Complications Several complications have been associated with the use of halo-gravity traction; the most common one in our series is pain and weakness of neck musculature. Other complications include pin tract infections, pin loosening, and cranial nerve palsy (abducens, oculomotor, glossopharyngeal, hypoglossal). It is important to ask about double vision, diculty swallowing, voice hoarseness, and tongue weakness. The cause of cranial nerve symptoms is believed to be the result of stretching or kinking of the nerve. Alternatively, an interference with the blood supply to the nerves

Fig. 8. Halo wheelchair.

CASTING AND TRACTION TREATMENT FOR SCOLIOSIS

483

Fig. 10. A 3-year-old girl who had severe early-onset kyphoscoliosis. Halo-gravity traction improved the coronal and sagittal balance and decreased the kyphosis, allowing the use of a VEPTR device to maintain the correction.

secondary to traction may lead to cranial nerve palsy. Other reported complications include skull penetration and brain or epidural abscess, paraplegia, paraparesis, brachial plexus palsy, superior mesenteric artery syndrome, and avascular necrosis of the odontoid [1720]. Practical advice A thorough hair wash with shampoo is done the night before surgery. If the patient has long hair, it should be braided. Keep a chart at the bedside to document the patients weight; chart the amount of weight on bed traction and chair traction, and note all increases in weight. Perform a quick neurologic examination and repeat at 2, 4, and 8 hours after increasing the traction weights. Case example A 3-year-old girl presented with severe earlyonset kyphoscoliosis. Halo-gravity traction improved the coronal and sagittal balance and decreased the kyphosis, allowing the use of a VEPTR device to maintain the correction (Fig. 10) until the denitive fusion and instrumentation are undertaken.

cases and if started before 2 years of age, casting may correct the deformity completely, whereas in more severe cases, it allows the child to grow, minimizing spinal and chest wall deformities, until he/she is old enough to undergo denitive treatment of the spinal deformity. Acknowledgments In addition to Dr Sanders, casting in Erie has been done by Drs. Joseph Khoury and Shyam Kishan. We are indebted to them for their assistance and insights. Marcie Fitzgerald PA, has been instrumental in helping us track our early results in these patients. For the most part, the casting at Intermountain was done by Dr. DAstous and Michael Pond, PA. We would like to recognize Michael Pond, without whose enthusiasm and multiple talents this work would have been near impossible. Finally, we thank Robert Eldridge, CPO, who designed and built our casting table and Matt Lowell, PT and the orthotic team who adapted the chairs, walkers, and beds for halo-traction. References
[1] Risser JC. Scoliosis treated by cast correction and spine fusion. Clin Orthop Relat Res 1976;116:8694. [2] Dickson RA. Early-onset idiopathic scoliosis. In: Weinstein SL, editor. 1st edition. The pediatric spine: principles and practice, vol 1. New York: Raven Press, Ltd; 1994. p. 4219. [3] Pehrsson K, Larsson S, Oden A, et al. Long-term follow-up of patients with untreated scoliosis. A study of mortality, causes of death, and symptoms. Spine 1992;17(9):10916.

Summary The presence of a signicant spinal deformity in an infant or toddler with several years of growth remaining has the potential to produce a severe deformity with signicant cardiopulmonary consequences. We present alternative, albeit labor-intensive, modalities to bracing, subcutaneous rods, and VEPTR instrumentation. In milder

484

DASTOUS & SANDERS

[4] Mehta MH. The rib-vertebral angle in the early diagnosis between resolving and progressive infantile idiopathic scoliosis. J Bone Surg Br 1972;54:23042. [5] Ceballos T, Ferrer-Torrelles M, Castillo F, et al. Prognosis in infantile idiopathic scoliosis. J Bone Joint Surg Am 1980;62(6):86375. [6] Ferreira JH, Janeiro R, de James JI. Progressive and resolving infantile idiopathic scoliosis. The dierential diagnosis. J Bone Joint Surg Br 1972;54(4): 64855. [7] Thompson SK, Bentley G. Prognosis in infantile idiopathic scoliosis. J Bone Joint Surg Br 1980;62(2): 1514. [8] Mehta MH. Growth as a corrective force in the early treatment of progressive infantile scoliosis. J Bone Joint Surg Br 2005;87(9):123747. [9] Cotrel Y, Morel G. [The elongation-derotation-exion technic in the correction of scoliosis]. Rev Chir Orthop Reparatrice Appar Mot 1964;50:5975. [10] Crawford H. Traction. In: Weinstein SL, editor. Pediatric spine surgery. Philadelphia: Lippincott, Williams & Wilkins; 2001. p. 6981. [11] Arlet V, Papin P, Marchesi D. Halo femoral traction and sliding rods in the treatment of a neurologically compromised congenital scoliosis: technique. Eur Spine J 1999;8(4):32931.

[12] Sink EL, Karol LA, Sanders J, et al. Ecacy of perioperative halo-gravity traction in the treatment of severe scoliosis in children. J Pediatr Orthop 2001; 21(4):51924. [13] Swank SM, Winter RB, Moe JH. Scoliosis and cor pulmonale. Spine 1982;7(4):34354. [14] White AAPM. Clinical biomechanics of the spine. 2nd edition. Philadelphia: JB Lippincott; 1990. [15] Mubarak SJ, Camp JF, Vuletich W, et al. Halo application in the infant. J Pediatr Orthop 1989; 9(5):6124. [16] Stagnara P. [Cranial traction using the Halo of Rancho Los Amigos]. Rev Chir Orthop Reparatrice Appar Mot 1971;57(4):287300. [17] Garn SR, Botte MJ, Nickel VL. Complications in the use of the halo xation device. J Bone Joint Surg Am 1987;69(6):954. [18] Rozario RA, Stein BM. Complications of halopelvic traction. Case report. J Neurosurg 1976; 45(6):7168. [19] Tredwell SJ, OBrien JP. Avascular necrosis of the proximal end of the dens. A complication of halopelvic distraction. J Bone Joint Surg Am 1975; 57(3):3326. [20] Wilkins C, MacEwen GD. Cranial nerve injury from halo traction. Clin Orthop Relat Res 1977;126:10610.

Orthop Clin N Am 38 (2007) 485496

New Strategies and Decision Making in the Management of Neuromuscular Scoliosis


John Sarwark, MDa,b,*, Vishal Sarwahi, MDc
The Childrens Memorial Hospital, 2300 Childrens Plaza, Box 69, Chicago, IL 60614, USA b Department of Orthopaedic Surgery, Northwestern University, Evanston, IL 60208, USA c Department of Orthopaedic Surgery, Monteore Medical Center/Albert Einstein College of Medicine, 3400 Bainbridge Avenue, MAP6, Bronx, NY 10467, USA
a

Neuromuscular scoliosis occurs in patients with traumatic spinal injuries and in patients with one or several of a large number of neurological or muscular diseases, including cerebral palsy, postmeningitis encephalopathy, posttraumatic encephalopathy, poliomyelitis, myelomeningocele, spinal muscle dystrophy, muscular dystrophies, and myopathies. The incidence of scoliosis due to neuromuscular causes may range between 25% to 90% compared with 2% to 4% for idiopathic scoliosis in the general population. Natural history Most children with cerebral palsy survive to adulthood [1]. Crichton and colleagues [2] reported a mean 30-year survival rate of at least 87%, varying from 95% in patients with hemiplegia or monoplegia to 83.5% in those with quadriplegia. The incidence of scoliosis in adult institutionalized patients varies from 25% to 64% [3,4]. There is an inverse relationship between ambulation and incidence of scoliosis, but a direct correlation between severity of involvement and the incidence of scoliosis [511]. The incidence and severity increases with the degree of involvement, degree of mental retardation, and functional status [3]. Thometz and Simon [12] showed that scoliosis in patients with cerebral
* Corresponding author. The Childrens Memorial Hospital, 2300 Childrens Plaza, Box 69, Chicago, IL 60614. E-mail address: j-sarwark@northwestern.edu (J. Sarwark).

palsy may progress even after skeletal maturity, in contrast to patients with mild and moderate idiopathic scoliosis. Patients with curves less than 50 had curves increase about 0.8 per year, while those with curves greater than 50 had curves increase at 1.4 per year. Lumbar and thoracolumbar curves in bedridden quadriplegic patients were most likely to progress. Progressive deformity interferes with general health and well-being, ambulation, sitting balance, and wheelchair transfers. This disability can lead to decubiti, costopelvic impingement pain, and worsening of pulmonary status. Majd and colleagues [13] demonstrated that a correlation exists between deformity size, functional decline, and decubiti. Patients with spastic cerebral palsy have the highest incidence of scoliosis [14]. Almost 70% of these patients have scoliosis that tends to progress even after skeletal maturity [1214]. Curves are worse in nonambulators, in patients with total body involvement, and in patients with thoracolumbar curves [12,14,15]. Most scoliosis presents before 10 years of age. Saito and colleagues [14] found that patients with curves greater than 40 before 15 years of age ended up with larger curves. They recommended that fusion should be considered for curves greater than 40 before age 15 years. Hsu [16] studied progression of scoliosis in patients with Duchenne muscular dystrophy (DMD). Curve progression was seen in all patients and those with curve greater than 40 experienced decreased sitting tolerance, decreased vital capacity, loss of function, and limited use of hands. Bracing may not be a denitive method of preventing progression in neuromuscular scoliosis,
orthopedic.theclinics.com

0030-5898/07/$ - see front matter 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.ocl.2007.07.001

486

SARWARK & SARWAHI

but bracing may be used for improving trunk support, which perhaps can allow a delay the date of surgery to a better time or improve seating functionality. A custom total contact thoracolumbosacral orthosis is the most common orthosis used and most eective in providing truncal support. The physician and patient families must understand that despite brace treatment, curve progression may continue to occur. Braces may also cause further pulmonary restriction and may interfere with feeding, especially gastronomy tube feeding. As scoliosis worsens, loss of sitting balance makes patients functionally quadriplegic. Independent sitters may lose their sitting ability and dependent sitters may need more support, including constant repositioning. Nursing demands increase in terms of dressing, positioning, and feeding. When scoliosis is greater than 60 , care becomes especially dicult. Natural history after surgery Surgery is indicated in children with progressive scoliosis or large curves that limit overall function [8,9,11]. However, decision making for surgery is challenging in the presence of pulmonary dysfunction, poor nutritional status, feeding disorders, seizures, immune compromise, and coagulopathies. Surgery is associated with a risk of life-threatening complications. Multiple appropriate consultations should be sought in the preoperative and perioperative management phases. A number of long-term follow-up studies have shown improvements in sitting position, quality of life, and lung function; improvements in spinal curves; and improvements in pelvic obliquity [17 22]. Spinal fusion, even in the more severely affected pediatric patients with cerebral palsy, has a documented high satisfaction rate among parents and caregivers, who appreciate the benets of this procedure, especially those resulting in easier care and improvements in the patients sitting ability, physical appearance, and comfort [19]. Kalen and colleagues [15] found that pulmonary function and caregiver time for institutionalized cerebral palsy patients with scoliosis did not dier from pulmonary function and caregiver time for similar patients without scoliosis. Cassidy and colleagues [23] additionally found no dierence in use of pain and pulmonary medicines, in decubiti, and in caring time between institutionalized patients who underwent scoliosis surgery and those who did not.

The complication rate of spinal fusion surgery for neuromuscular scoliosis is higher than that for idiopathic scoliosis [2429]. Comstock and colleagues [19] have reported a 68% complication rate in children with totally involved cerebral palsy. Fourteen percent of their patients had pulmonary complications, 19% had wound complications, 18% had hardware complications, and 19% had curve progression or pseudoarthrosis. Despite high complication rates, studies have shown high rates of caretaker satisfaction after surgery [1820,30]. Jones and colleagues [30] arrived at a similar conclusion after prospectively evaluating caregiver satisfaction in children with total body involvement who underwent scoliosis surgery. Parents reported higher satisfaction with the patients overall health status after surgery. They found that the most common ndings were improvement in appearance, sleeping comfort, pain relief, and self-attitude. Results showed less pain, greater happiness, and less likelihood of feeling sick and tired among patients, and greater satisfaction among parents. Complications did not signicantly aect questionnaire results. Larsson and colleagues [20] reported on a prospective study of 82 patients with neuromuscular scoliosis who underwent spine fusion and instrumentation. The group consisted of patients with cerebral palsy [21], myelomeningocele [22], Rett syndrome [8], myopathies [5], spinal muscular atrophy [9], and other diseases, who were followed up for 9.5 years. Fifty-ve of these patients understood verbal instructions and 27 did not. For the whole group, the investigators noted improvements in sitting balance, weight distribution when sitting, Cobb angle, lung function, activities of daily living, time used for resting, and seating supports in a wheelchair. Long-term follow-up results showed further improvements in sitting balance and activities of daily living and reductions in time required to provide care. The results between the patients who understood verbal instructions and patients who did not were almost the same except for time required to give care, which had decreased in those who understood verbal instructions but did not change for those who did not understand verbal instructions. The most important determinants for reduced life expectancy in patients with neuromuscular scoliosis are the coexistence of pulmonary compromise, epilepsy, feeding disorders requiring gastrostomy or nasogastric tube, signicant neurological involvement, severe cognitive abnormality, and the lack of ambulation [1,3137]. Tsirikos

MANAGEMENT OF NEUROMUSCULAR SCOLIOSIS

487

and colleagues [38] demonstrated a relatively long mean predicted survival for children with severe spastic cerebral palsy who underwent scoliosis surgery. Although their patients had severe impairments, the investigators demonstrated a mean survival of 11 years 2 months. The most accurate determinants for survival rates were the number of postoperation days in the intensive care unit and the presence of excessive thoracic hyperkyphosis. The investigators showed that a child with kyphosis one standard deviation above the mean (85.6 ) was 122% more likely to die than a child with a kyphosis one standard deviation below the mean (26 ). Also, children who spent 30 days in the intensive care unit were at a 10 times higher risk than those in the intensive care unit for 5 days. In this study, such variables as age, sex, ambulation status, cognition, pelvic obliquity, lumbar lordosis, blood loss, surgical time, and length of stay, did not signicantly predict life expectancy. Before surgery, a multidisciplinary evaluation is necessary to optimize medical management. This involves evaluations of seizure control medications, nutritional supplementation, control of gastroesophageal reux, and methods for enhancing gastric motility. If swallowing status is not known before surgery, such a study should be performed to assess risk of aspiration. A simple step in workup is weighing the patient. If the child body weight is less than the fth percentile, nutritional supplementation may be required. This should be performed parenterally for all patients at risk. Jevsevar and Karlin [39] found an increased postoperative infection rate, longer length of intubation, and longer hospital stays in patients with cerebral palsy who had preoperative serum albumin measurement less than 35 g/L and total blood lymphocyte count less than 1.5 g/L. While serum albumin and prealbumin are used, some investigators have found them unreliable indicators of the nutritional status of a severely involved child. In fact, some contend that a child who is overweight as a result of gastrostomy tube feeding is more likely to have complications than an underweight child [26,40].

of pelvic obliquity. Multiple other reasons are attributed to this increased bleeding, including poor nutritional status, venous pooling, impaired connective tissue function, and antiseizure medication, such as valproic acid. Valproic acid is known to decrease platelet count and factor VIII levels [41,42]. Brenn and colleagues [43] studied 17 patients with cerebral palsy and compared them with 17 patients with idiopathic scoliosis. They found that children with cerebral palsy develop signicant alterations in coagulation parameters early, although they had normal coagulation proles. The baseline prothrombin time and partial prothrombin time, although within normal limits, were signicantly higher in patients with neuromuscular scoliosis than in patients with idiopathic scoliosis. After 15% loss of blood volume, investigators found dierences between prothrombin time, partial prothrombin time, maximum amplitude on thromboelastography, ionized calcium, and serum magnesium levels. Their report indicated that increased bleeding caused an apparent coagulopathy instead of transient hypercoagulable state seen in normal states. Children with cerebral palsy have increased bleeding that starts earlier in the procedure despite a normal coagulation prole.

Monitoring Ashkenaze and colleagues [44] reported unreliable somatosensory spinal evoked potential (SSEP) data in neuromuscular scoliosis. Twentyeight percent of their patients did not produce reliable monitoring data and two neurologic decits were missed by monitoring. DMD and poliomyelitis patients had reliable monitoring data and the use of sublaminar wires in their study produced transient changes. Padberg and colleagues [45] found that use of multiple recording sites, including a subcortical site, increased the reliability of SSEP, and the addition of another modality, such as the neurogenic motor evoked potential, improved the quality of their data. Eighty-nine percent of their 74 patients had baseline SSEPs. The investigators recommended monitoring for neuromuscular scoliosis, especially for those with preserved lower extremity function, preserved bowel and bladder function, or both. DiCindio and colleagues [46] recently reported on a pilot study using transcranial motor evoked potential in nine patients with neuromuscular scoliosis. They found consistent reproducible

Bleeding Children with neuromuscular scoliosis incur greater intraoperative blood loss than children with idiopathic scoliosis. Blood loss is correlated with the number of levels operated and the degree

488

SARWARK & SARWAHI

responses that were low-amplitude signals and were present in all of their patients, including those with cerebral palsy. In this study, all of the mild-tomoderate cerebral palsy children and 86% of patients without cerebral palsy had monitorable responses for at least one modality. Patients with cerebral palsy who also had weight-bearing ability or usable motor function, had reliable monitoring data. Multimodality monitoring is useful in children with neuromuscular scoliosis. However, severely involved cerebral palsy patients with severe mental retardation may not benet from it because of lack of reliable data.

Ventilator Improved home pulmonary noninvasive support has added greatly to the health of children with neuromuscular spine deformities. The preoperative pulmonary function test is useful if the patient can perform it. Arm span instead of height is used in these patients to calculate the predictive value. Padman and McNamara [47] reported a correlation between decreased vital capacity and increased complications in neuromuscular scoliosis. Rawlins and colleagues [48] found 19% pulmonary complications in patients undergoing reconstructive spinal surgery when the vital capacity was less than 40% of predicted capacity. Yuan and colleagues [49] found that children more than 13 years old with neuromuscular scoliosis or forced expiratory volume in 1 second (FEV1) less than 40% predicted were likely to require prolonged postoperative mechanical ventilation. Children less than 13 years old with neuromuscular scoliosis and FEV1 less than 40% predicted were also likely to need ventilator postoperation. Patients with DMD have compromised lung and heart function and a higher risk of complications [5053]. Marsh and colleagues [50] reported 30% complication rate in DMD patients irrespective of their preoperative forced vital capacity. Almenrader and Patel [54] reported that 40% of their patients with DMD needed ventilators. Morris [51] has considered forced vital capacity less than 25% and left ventricular ejection fraction of less than 50% a contraindication for elective surgery. Gill and colleagues [55] studied myopathic patients who were on nighttime ventilatory support and had a mean vital capacity of 20%. They did not report any pulmonary complications and concluded that ventilatory support for

respiratory failure presented no increased risk of complications. Their patients returned to their preoperative level of function and had smooth postoperative recovery. Almenrader and Patel [54] have made a case for early extubation and use of noninvasive continuous positive airway pressure if necessary. In their series, 76.2% of patients were safely extubated at the end of surgery while only 9.5% needed ventilation for more than 3 days. Their complication rate of 9.5% was lower than the rate that has been previously reported. They suggested that early extubation was possible in cases with forced vital capacity greater than 30% and (1) where transespohageal hemodynamic monitoring was used; (2) where postoperative pain control was optimized using epidural analgesia and intravenous morphine; and (3) where ultra-shortacting analgesics, such as remifentanil, and rapidly eliminated inhalation anesthetics, such as sevourane, were used.

Anterior instrumentation In selected curves that do not include the pelvis or with pelvic obliquity of less than 15 , stopping fusion in the lumbar spine avoids complications that can result when fusion extends to the sacropelvis and permits improved mobility [56 58]. The authors feel that, by stopping fusion in the lumbar spine, the need for an anterior fusion at L5-S1 is obviated and the patient is allowed the exibility to compensate through the lumbosacral and the sacropelvic joints. This residual mobility may allow more comfortable seating and compensation, especially when the hips are dislocated or subluxated. It also avoids the problem of loosening of the pelvic xationdthe windshield eect. Anterior release leads to improved curve correction and provides a larger surface area for fusion. This is especially useful in myelomeningocele patients and patients with severe osteopenia. It also prevents crankshaft phenomenon in skeletally immature patients. The authors have reviewed the risks and complications of anterior procedure in neuromuscular scoliosis in the past [25]. Basobas and colleagues [56] and Tokala and colleagues [57] have reported on anterior instrumentation in neuromuscular scoliosis as an alternative in ambulatory as well as nonambulatory patients Hopf and colleagues [58] have also reported on 33 neuromuscular patients treated

MANAGEMENT OF NEUROMUSCULAR SCOLIOSIS

489

by anterior instrumentation. Seventeen patients underwent secondary posterior instrumentation. Tokala and colleagues [57] performed anterior instrumentation for patients who had short exible curves with minimal pelvic obliquity (!15 ) and who were ambulatory. The investigators found that anterior instrumentation provided satisfactory clinical and radiological outcome. In the authors series, 30% of patients underwent anterior instrumentation alone for neuromuscular scoliosis while others underwent either combined or staged procedures for correction of scoliosis [25]. Patients treated with both anterior and posterior instrumentation showed greater correction of scoliosis and improved sitting balance in comparison with patients treated with anterior or posterior xation alone [59].

Pelvic xation Broom and colleagues [60] considered it necessary to fuse to the pelvis when (1) the case showed truncal decompensation (plumb line falls lateral to the sacroiliac joint), (2) the xed pelvic obliquity was greater than 15 , and (3) the sacrum was part of the curve. There are multiple methods of sacropelvic xation, namely the Galveston (L-rod) technique, the Dunn-McCarthy (S-rod) technique, transiliac screws, intrasacral rods, iliosacral xation, iliac screws, and L5-S1 pedicle instrumentation [61 70]. Unit rods are eective as is the Luque-Galveston technique of pelvic xation, with or without anterior procedures [1,60,61,64,69]. Sacral screws, iliosacral screws, and L5 and S1 pedicle screws have become alternatives. These are placed independently of the rods, which do not need to be bent as in the Galveston technique. Screws are placed in the posterior ilium, which allows bonegraft harvesting. Camp and colleagues [28], in a biomechanical and clinical study, reported that the Galveston method of sacral xation is better than the Cotrel-Dubousset sacral xation. They found clinical failure rates of 44% for sacral screws, 28% for iliosacral screws, and 0% for the Galveston technique. Gau and colleagues [63] reported on 58 patients undergoing Luque-Galveston instrumentation and found a mean pelvic correction of 53% (range: 17 8 ). Yazici and colleagues [70] reported a mean correction of 81% (range: 27 5 ) using Isola-Galveston instrumentation. Miladi and colleagues [66] showed a mean 71%

correction (range: 21 6 ) using iliosacral screws with Cotrel-Dubousset instrumentation, while Neustadt and colleagues [67] reported a mean 50% correction, which worsened to 37% at follow-up. Peelle and colleagues [68] reported better correction of pelvic obliquity using iliac screws compared with correction using the Galveston technique. Vialle and colleagues [71] and Huang and Lenke [72] have described using intraoperative asymmetric halo-femoral traction to correct pelvic obliquity. Using this technique, the nal corrections of scoliosis and pelvic obliquity were significant. Takeshita and colleagues [73] compared two groups of patients treated with and without intraoperative halo-femoral traction. They found that the curve corrected to an average 37 from average 87 in the halo-femoral traction group and to 35 from 67 in the control group. Average correction of pelvic obliquity was 78% in the halofemoral traction group, compared with 52% in the control group. There are many potential complications with the Galveston technique. Sink and colleagues [74] reported a high incidence of proximal xation pullout (32%) and distal migration of Galveston rods (12%). They felt this was secondary to kyphosing forces created by a long posterior lever arm. This lever arm produces considerable cantilever force at the lumbosacral junction, causing failure of xation or junctional kyphosis. Failure was seen even in the presence of anterior fusion. Several investigators have suggested that fusion extending to the pelvis should be avoided for ambulatory patients without severe pelvic obliquity [7,21,65,75,76]. However, Tsirikos and colleagues [77] recommended fusion to the pelvis even in ambulatory patients and did not nd it to be detrimental. The authors normally do not extend instrumentation across the sacropelvic joint into the pelvis. The authors make every attempt to instrument in the midline and extend fusion down to S2. This means sometimes performing facet joint osteotomies after opening the canal to bring the pelvis level with or at a 15 angle to the horizontal. The S1 screws are placed converging toward the midline with the tip getting a purchase in the S1 endplate. This allows for a strong tricortical purchase and, in the authors experience, requires a C-arm to ensure adequate placement in the lateral plane. The S2 screws are placed divergent, aiming cephalad, to obtain a purchase in the sacral ala as described by Edwards [78]. In

490

SARWARK & SARWAHI

cases with pelvic obliquity greater than 15 , the authors instrument to the pelvis.

Unilateral hip dislocation A unilateral hip dislocation may cause problems with sitting imbalance, pressure ulcers, hip pain, and pelvic obliquity, which may inuence development and progression of scoliosis. There is believed to be a dynamic interplay between hip dislocation, pelvic obliquity, and scoliosis. However, the exact relationship is unclear in the literature. Pritchett [79] reported a direct correlation between pelvic obliquity and the presence and severity of scoliosis. Lonstein and Beck [80], on the other hand, did not nd any correlation between the dislocated hip and the amount of pelvic obliquity. Recently, Senaran and colleagues [81] did not nd correlation between unilateral hip dislocation and scoliosis progression. They reported that the mean rate of curve progression in unilateral hip dislocation was 12.9 /y, whereas in the control group it was 12.2 /y. The progression of pelvic obliquity, however, was signicantly increased in the hip dislocation group in their study.

pseudoarthroses or hardware failures with improvement in sitting ability noted by caregivers, who also expressed high degree of satisfaction with the procedure. Seven children with open triadiate cartilages had a posterior spinal fusion only with only a 3 loss of correction at maturity and no loss of correction of pelvic obliquity. Westerlund and colleagues [69] reviewed their experience with the unit rod in skeletally immature patients. They also concluded that correction was maintained at maturity and that this technique may avoid anterior surgery and its associated risks. Long-term problems with the unit rod are loss of lower lumbar and pelvic xation, which potentially can lead to pseudoarthrosis; failure of construct; and worsening of deformity [74,82,84]. The distal xation usually gets loose, presenting as a windshield wiper eect seen on radiographs, and is believed to be due to xation across an unfused sacroiliac joint. The rod can also pull out at the inferior aspect of the construct. This xation can be augmented by using pedicle screws at L5 and incorporating them in the unit rod construct. Erickson and colleagues [62] showed this to be biomechanically superior and also suggested that addition of cross-links added to the stiness of the construct. Fusion to L5 Unit rods, although popular, need exposure of the crest bilaterally for insertion into the pelvis. This increases operating time, increases blood loss, and limits area available for harvesting bone graft. Also, in spina bida patients, the pelvis is usually hypoplastic and may not be suciently thick to insert the rod [85]. The neuromuscular pelvis is osteopenic and thus does not provide adequate xation. The unit rods corrective force employs a cantilever principle and is based o an osteopenic pelvis. In large deformities, the force can lead to loss of xation or breakage. In the presence of stable lumbosacral articulation, xation into lumbar 5 allows correction without the need to span the sacroiliac joint. McCall and colleagues [86] reported their experience with the U-rod with pedicular xation at lumbar 5 and found that it oered stability equal to the unit rod. They concluded that fusion only to L5 in neuromuscular scoliosis was equivalent to fusion to the pelvis and that the presence of mobile lumbosacral joint may assist in seating and transfer activities.

Unit rod Bell and colleagues [6] rst described use of the unit rod in 1989. This method has become popular for spine xation and has been shown to provide excellent correction. Several studies have documented the ecacy of this method in comparison to older instrumentation systems [62,82,83]. The unit rod has a prebent pelvic section for the Galveston xation technique and allows cantilever force to correct the pelvic obliquity and the scoliosis simultaneously using a sublaminar wiring technique. The advantages of the unit rod are its strong xation and good correction characteristics, its suitability for relatively quick application, its ability to correct pelvic obliquity, and, because it xes into the pelvis, its usefulness as a good foundation. Dias and colleagues [82] reviewed 31 patients who underwent unit rod instrumentation with sublaminar wires and posterior spine fusion. The mean curve in their series was 79 preoperatively, 19 postoperatively, and was maintained at 18 at 2.8 years follow-up. The pelvic obliquity was 25 , which was corrected to 3 and remained unchanged at follow-up. They reported no

MANAGEMENT OF NEUROMUSCULAR SCOLIOSIS

491

Pedicle screw xation Wimmer and colleagues [87] reviewed their experience with the Luque-Galveston and the Isola instrumentations in 52 patients. They found comparable results in terms of correction. The Luque group had a scoliosis correction of 54% and a 64% correction in pelvic tilt, while the Isola group had correction of 57% and 63% improvement in pelvic tilt. They found greater correction in smaller curves (!60 ) than in larger curves (O100 ) in both groups. Gaine and colleagues [88] have also reported similar corrections with the Luque-unit rod and Isola instrumentations in patients with DMD. Vialle and colleagues [71], using a hybrid system, reported a mean scoliosis correction of 62%. Westerlund and colleagues [69] reported a nal correction of 66% using Luque-Galveston instrumentation. Onimus and colleagues [89] reported comparable correction with Cotrel-Dubousset instrumentation and pelvic xation with iliosacral screws. Teli and colleagues [90] reviewed 56 patients operated using hybrid pedicle screw and hook systems. The pseudoarthrosis rate was 1.8%. That compares to rates varying from 1.5% to 10% for the unit rod and Luque-Galveston systems. The loss of correction of scoliosis, kyphosis, and lordosis in their series was 0% to 6% at followup. However, the loss of lordosis was more in patients who had posterior-only surgery. The investigators concluded that hybrid systems provided satisfactory correction of coronal deformity, sagittal deformities, and pelvic obliquity, which improved quality of life in patients with neuromuscular scoliosis. Peelle and colleagues [68] felt that a rod construct with iliac screws allows additional screw xation points in the sacrum and lower lumbar vertebrae, which possibly allows for a stronger construct. They use iliac screws with sacral screws and even lumbar pedicle screws in all their patients with spastic neuromuscular scoliosis. Because this patient population shows progression of spinal deformity following skeletal maturity, rigid xation and fusion is indicated [21,27,62,67,89,90]. Pedicle screws have the disadvantage of limited purchase in an osteoporotic spine. Larger diameters should be used and, preferably, all levels should be instrumented, which allows greater points of xation as well as application of greater corrective forces. Pedicle screws also provide three-column xation, which helps achieve greater correction of rotation. In presence of exaggerated

lordosis or larger curves, the authors prefer to use reduction screws around the apex, which allows easier rod capture and reduction and decreases the risk of screw pullout intraoperation. The keys to greater alignment are extensive releases. This approach may involve osteotomies at the apical and juxta-apical region. The osteotomies must include the facet joints. In severe, rigid curves or curves with a large amount of decompensation, the authors do not hesitate in performing a vertebral column resection at the apex. This posterior-based approach is feasible with pedicle screws and saves the anterior exposure. Resection at one or two levels provides great mobility and improved alignment [85]. Staging of surgery is acceptable practice. In the authors experience, halo traction between stages is poorly tolerated and also involves prolonged hospital stay, which is not feasible in some settings. An alternative is internal distraction between stages. Thus the authors achieve the maximum correction feasible in stage one and then lock the rod in place and close the incision. The patient is brought back to the operating room 2 to 3 weeks after an aggressive nutritional regimen and hematologic regimen, including procrit, are put in place. Sometimes the patient is discharged to a rehabilitation facility or even home. In the second stage, more releases (usually osteotomies) are performed and, if need be, a vertebral column resection is performed. The correction is then attempted again. A similar approach has been reported by Buchowski and colleagues [91].

Staging McDonnell and colleagues [29], in a mixed group of patients, reported a higher incidence of complications with single-stage anteroposterior surgery and with patients who had neuromuscular scoliosis. Shuebarger and colleagues [92] and Powell and colleagues [93] reported less morbidity, better correction, and reduced cost with single-stage anteroposterior surgery. Their study group, however, included multiple diagnoses, including neuromuscular diseases. Ferguson and colleagues [94] found a lower complication rate with single-stage surgeries in neuromuscular scoliosis. OBrien and colleagues [95] found comparable results and complications between the one-stage and two-stage groups. Hopf and colleagues [58] reported increased pulmonary

492

SARWARK & SARWAHI

complications with single-stage procedures. Tsirikos and colleagues [83] more recently compared patients who had surgery done in a single stage with those who had surgery done in two stages. They found that two-stage procedures have significantly less blood loss, shorter surgical time, and fewer complications than one-stage procedures. Meanwhile, no deaths occurred following twostage procedures and the rates of infection associated with two-stage procedures were the same as those associated one-stage procedures. Staging also did not increase hospitalization time or stay in the intensive care unit. The investigators concluded that two-stage procedures are safer and produce consistent results in cerebral palsy patients while single-stage procedures have increased technical and medical complications, increased morbidity and mortality, and can be overwhelming for both the patient and the surgical team.

Vertical expandable prosthetic titanium rib Hell and colleagues [96] reported on 15 children with progressive scoliosis who had a vertical expandable prosthetic titanium rib (VEPTR) implantation at a mean age of 6 years. Six of these children had severe thoracolumbar scoliosis and pelvic obliquity due to neuromuscular scoliosis. The investigators found that the thoracolumbar curve reduced in size and pelvic obliquity improved both subjectively and on radiographs. The patients had better sitting ability in the wheelchair and families and patients reported better sitting abilities and radical cosmetic improvements.

anterior spine surgery in 111 patients and found that 21.6% of these children had major and 22.5% had minor complications with pulmonary complication being the most frequent [25]. Pulmonary complications stem from lung immaturity, pain, and the inability to participate in pulmonary care, which can range from atelectasis, to prolonged ventilatory support or tracheostomy. Staged anteriorposterior procedures do not decrease the incidence of respiratory complications. Children who require long-term intensive care for respiratory problems postoperatively have the poorest long-term survival [38]. The next commonly seen complications are pancreatitis, ileus, superior mesenteric artery syndrome, gall bladder disease, and poor gastric motility. Wound complications occur more often in patients with neuromuscular scoliosis than in average implant patients. Szoke and colleagues [26] described an 8.7% rate of postoperative wound infection; 7 of 15 were deep infections with 1 late infection that needed implant removal. The implant is usually not removed even in the presence of infection as it provides a stable environment for the spine, which in turn helps with the healing and control of the infection. Implant removal, however, is more commonly a practice in patients with myelomeningocele if the patient develops a wound infection.

Summary Murphy and colleagues [24] found that children with neuromuscular scoliosis have longer lengths of stay (9.2 versus 6.1 days), higher total charges ($66,953 versus $47,463), more diagnoses (6.3 versus 2.5), and more total procedures (4.2 versus 3.0) compared with those with idiopathic scoliosis. In their study, 1570 children with neuromuscular scoliosis underwent spine fusion, which cost $105 million in hospital charges and totaled 1444 days in the hospital. They concluded that children with neuromuscular scoliosis experience signicantly more complicated and costly hospitalizations than do children with idiopathic scoliosis. The outcome of spinal fusion in neuromuscular scoliosis demonstrates a consistently high satisfaction rate among parents and caretakers. It is somewhat more dicult to determine the functional benets in individual patients, especially in those with the most severe involvement. Even in those children, a predicted 70% survival

Complications Sarwahi, Sarwark and colleagues [25], in comparing children with neuromuscular and idiopathic scoliosis to children with only neuromuscular scoliosis, found frequent problems with the development of postoperative pneumonia (3.5% versus 0.7%), respiratory failure (24.1% versus 9.2%), urinary tract infections (5.3% versus 0.7%), and surgical wound infections (1.3% versus 0.3%). Complications of scoliosis surgery in children with neuromuscular and idiopathic scoliosis also include implant failure, pulmonary complications, gastrointestinal complications, infections, wound problems, and death. The complication rates vary from 44% to 80%, with a perioperative death rate of 0% to 7%. The authors reported on the complication rates of

MANAGEMENT OF NEUROMUSCULAR SCOLIOSIS

493

rate at 11 years has been documented following surgery. Scoliosis is common in children with neuromuscular diseases, especially cerebral palsy. Most of these children need stabilization to enhance their quality of life by improving sitting ability. Studies continue to show a high family satisfaction and caregiver satisfaction following spine reconstructive surgery for neuromuscular scoliosis and other spine deformities. References
[1] Evans PM, Evans SJ, Alberman E. Cerebral palsy: why we must plan for survival. Arch Dis Child 1990;65:132933. [2] Crichton JU, Mackinnon M, White CP. The life expectancy of persons with cerebral palsy. Dev Med Child Neurol 1995;37:56776. [3] Madigan RR, Wallace SL. Scoliosis in the institutionalized cerebral palsy population. Spine 1981;6: 58390. [4] Samilson RL, Bechard R. Scoliosis in cerebral palsy: incidence, distribution of curve patterns, natural history, and thoughts on etiology. Curr Pract Orthop Surg 1973;5:183205. [5] Bradford DS. Neuromuscular spinal deformity. In: Bradford DS, editor. Moes textbook of scoliosis and other spinal deformities. Philadelphia: WB Saunders; 1987. p. 271305. [6] Bell DF, Moseley CF, Koreska J. Unit rod segmental spinal instrumentation in the management of patients with progressive neuromuscular spinal deformity. Spine 1989;14:13017. [7] Banta JV, Drummond DS, Ferguson RL. The treatment of neuromuscular scoliosis. In: Zuckerman JD, editor. AAOS Instructional Course Lectures, vol. 48. Rosemont (IL): American Academy of Orthopaedic Surgeons; 1999. p.55161. [8] Herring JA. Disorders of the brain. In: Herring JA, editor. Tachjians pediatric orthopaedics. Philadelphia: WB Saunders; 2002. p. 1121248. [9] Labelle H, Grimard G. Pediatric spine. In: Koval KJ, editor. Orthopaedic knowledge update 7: home study syllabus. Rosemont (IL): American Academy of Orthopaedic Surgeons; 2002. p. 58192. [10] Rinsky LA. Surgery of spinal deformity in cerebral palsy. Twelve years in the evolution of scoliosis management. Clin Orthop Relat Res 1990;253:1009. [11] Renshaw TS. Cerebral palsy. In: Morrissy RT, Weinstein SL, editors. Lovell and Winters pediatric orthopaedics. Philadelphia: Lippincott Williams & Wilkins; 2001. p. 56399. [12] Thometz JG, Simon SR. Progression of scoliosis after skeletal maturity in institutionalized adults who have cerebral palsy. J Bone Joint Surg Am 1988;70:12906. [13] Majd ME, Muldowny DS, Holt RT. Natural history of scoliosis in the institutionalized adult cerebral palsy population. Spine 1997;22:14616.

[14] Saito N, Ebara S, Ohotsuka K, et al. Natural history of scoliosis in spastic cerebral palsy. Lancet 1998; 351(6):168792. [15] Kalen V, Conklin MM, Sherman FC. Untreated scoliosis in severe cerebral palsy. J Pediatr Orthop 1992;12:33740. [16] Hsu JD. The natural history of spine curvature progression in the nonambulatory Duchenne muscular dystrophy patient. Spine 1983;8:7715. [17] Benson ER, Thomson JD, Smith BG, et al. Results and morbidity in a consecutive series of patients undergoing spinal fusion for neuromuscular scoliosis. Spine 1998;23:230818. [18] Bridwell KH, Baldus C, Iring TM, et al. Process measures and patient/parents evaluation of surgical management of spinal deformities in patients with progressive accid neuromuscular scoliosis (Duchennes muscular dystrophy and spinal muscular atrophy). Spine 1999;24:13009. [19] Comstock CP, Leach J, Wenger DR. Scoliosis in total-body involvement cerebral palsy: analysis of surgical treatment and patient and caregiver satisfaction. Spine 1998;23:141225. [20] Larsson E, Aaro S, Normelli H, et al. Long-term follow-up of functioning after spinal surgery in patients with neuromuscular scoliosis. Spine 2005; 30(19):214552. [21] Lonstein JE. Spine deformities due to cerebral palsy. In: Weinstein SL, editor. The pediatric spine: principles and practice. Philadelphia: Lippincott Williams & Wilkins; 2001. p. 797807. [22] Lonstein JE, Akbarnia BA. Operative treatment of spinal deformities in patients with cerebral palsy or mental retardation. J Bone Joint Surg Am 1983;65: 4355. [23] Cassidy C, Craig CL, Perry A, et al. A reassessment of spinal stabilization in severe cerebral palsy. J Pediatr Orthop 1994;14:7319. [24] Murphy NA, Firth S, Jorgensen T, et al. Spinal surgery in children with idiopathic and neuromuscular scoliosis. Whats the dierence? J Pediatr Orthop 2006; 26(2):21620. [25] Sarwahi V, Sarwark JF, Schafer MF, et al. Standards in anterior spine surgery in pediatric patients with neuromuscular scoliosis. J Pediatr Orthop 2001;21:75660. [26] Szoke G, Lipton G, Miller F, et al. Wound infection after spinal fusion in children with cerebral palsy. J Pediatr Orthop 1998;18(6):72733. [27] Boachie-Adjei O, Lonstein JE, Winter RB, et al. Management of neuromuscular spinal deformities with Luque segmental instrumentation. J Bone Joint Surg Am 1989;71:54862. [28] Camp JF, Caudle R, Ashman RD, et al. Immediate complications of Cotrel-Dubousset instrumentation to the sacro-pelvis: a clinical and biomechanical study. Spine 1990;15:93241. [29] McDonnell MF, Glassmann SD, Dimar JR, et al. Perioperative complications of anterior procedures

494

SARWARK & SARWAHI

[30]

[31]

[32]

[33]

[34]

[35] [36]

[37]

[38]

[39]

[40]

[41]

[42]

[43]

[44]

[45]

on the spine. J Bone Joint Surg Am 1996;78(6): 83947. Jones KB, Sponseller PD, Shindle MK, et al. Longitudinal parental perceptions of spinal fusion for neuromuscular spine deformity in patients with totally involved cerebral palsy. J Pediatr Orthop 2003;23: 1439. Eyman RK, Grossman HJ, Chaney RH, et al. The life expectancy of profoundly handicapped people with mental retardation. N Engl J Med 1990;323: 5849. Eyman RK, Grossman HJ, Chaney RH, et al. Survival of profoundly disabled people with severe mental retardation. Am J Dis Child 1993;147:32936. Hutton JL, Cooke T, Pharoah PO. Life expectancy in children with cerebral palsy. BMJ 1994;309: 4315. Hutton JL, Colver AF, Mackie PC. Eect of severity of disability on survival in northeast England cerebral palsy cohort. Arch Dis Child 2000;83: 46874. Anderson TW. Life expectancy in cerebral palsy. Lancet 1996;348(9040):1516. Plioplys AV, Kasnicka I, Lewis S, et al. Survival rates among children with severe neurologic disabilities. South Med J 1998;91:16172. Strauss DJ, Shavelle RM, Anderson TW. Life expectancy of children with cerebral palsy. Pediatr Neurol 1998;18:1439. Tsirikos AI, Chang W, Dabney K, et al. Life expectancy in pediatric patients with cerebral palsy and neuromuscular scoliosis who underwent spinal fusion. Dev Med Child Neurol 2003;45:67782. Jevsevar DS, Karlin LI. The relationship between preoperative nutritional status and complications after an operation for scoliosis in patients who have cerebral palsy. J Bone Joint Surg Am 1993;75:8804. Lipton GE, Miller F, Dabney KW, et al. Factors predicting postoperative complications following spinal fusions in children with cerebral palsy. J Spinal Disord 1999;12:197205. Chambers HG, Weinstein CH, Mubarak SJ, et al. The eect of valproic acid on blood loss in patients with cerebral palsy. J Pediatr Orthop 1999;19:7925. Winter SL, Kriel RL, Novacheck TF, et al. Perioperative blood loss: the eect of valproate. Pediatr Neurol 1996;15:1922. Brenn BR, Theroux MC, Dabney KW, et al. Clotting parameters and thromboelastography in children with neuromuscular and idiopathic scoliosis undergoing posterior spinal fusion. Spine 2004; 29(15):E3104. Ashkenaze D, Mudiyam R, Boachie-Adjei O, et al. Ecacy of spinal cord monitoring in neuromuscular scoliosis. Spine 1993;18:162733. Padberg AM, Russo MH, Lenke LG, et al. Validity and reliability of spinal cord monitoring in neuromuscular spinal deformity surgery. J Spinal Disord 1996;9:1508.

[46] DiCindio S, Theroux M, Shah S, et al. Multimodality monitoring of transcranial electric motor and somatosensory-evoked potentials during surgical correction of spinal deformity in patients with cerebral palsy and other neuromuscular disorders. Spine 2003;28(16):18516. [47] Padman R, McNamara R. Postoperative pulmonary complications in children with neuromuscular scoliosis who underwent posterior spinal fusion. Del Med J 1990;62:9991003. [48] Rawlins BA, Winter RB, Lonstein JE, et al. Reconstructive spine surgery in pediatric patients with major loss in vital capacity. J Pediatr Orthop 1996;16: 28492. [49] Yuan N, Skaggs DL, Dorey F, et al. Preoperative predictors of prolonged postoperative mechanical ventilation in children following scoliosis repair. Pediatr Pulmonol 2005;40:4149. [50] Marsh A, Edge G, Lehovsky J. Spinal fusion in patients with Duchennes muscular dystrophy and a low forced vital capacity. Eur Spine J 2003;12: 50712. [51] Morris P. Duchenne muscular dystrophy: a challenge for the anaesthetist. Paediatr Anaesth 1997;7:14. [52] Rice JJ, Jeers BL, Devitt AT, et al. Management of the collapsing spine for patients with Duchenne muscular dystrophy. J Med Sci 1998;167:2425. [53] Mubarak SJ, Morin WD, Leach J. Spinal fusion in Duchenne muscular dystrophy: xation and fusion to the sacropelvis? J Pediatr Orthop 1993;13:7527. [54] Almenrader A, Patel D. Spinal fusion surgery in children with non-idiopathic scoliosis: Is there a need for routine postoperative ventilation? Br J Anaesth 2006;97(6):8517. [55] Gill I, Eagle M, Mehta JS, et al. Correction of neuromuscular scoliosis in patients with preexisting respiratory failure. Spine 2006;31(21):247883. [56] Basobas L, Mardjetko S, Hammerberg K, et al. Selective anterior fusion and instrumentation for the treatment of neuromuscular scoliosis. Spine 2003; 28:S2458. [57] Tokala D, Lam KS, Freeman B, et al. Is there a role for selective anterior instrumentation in neuromuscular scoliosis? Eur Spine J 2007;16:916. [58] Hopf CG, Eysel P, Dubousset J. Operative treatment of scoliosis with Cotrel-Dubousset-Hopf instrumentation. New anterior spinal device. Spine 1997;22:61827 [discussion: 627618]. [59] Brown JC, Swank S, Specht L. Combined anterior and posterior spine fusion in cerebral palsy. Spine 1982;7:5703. [60] Broom MJ, Banta JV, Renshaw TS. Spinal fusion augmented by luque-rod segmental instrumentation for neuromuscular scoliosis. J Bone Joint Surg Am 1989;71:3244. [61] Bulman WA, Dormans JP, Ecker ML, et al. Posterior spinal fusion for scoliosis in patients with cerebral palsy: a comparison of luque rod and unit rod correction. J Pediatr Orthop 1996;16:31423.

MANAGEMENT OF NEUROMUSCULAR SCOLIOSIS

495

[62] Erickson MA, Oliver T, Baldini T, et al. Biomechanical assessment of conventional unit rod xation versus a unit rod pedicle screw construct. A human cadaver study. Spine 2004;29(12):13149. [63] Gau YL, Lonstein JE, Winter RB, et al. LuqueGalveston procedure for correction and stabilization of neuromuscular scoliosis and pelvic obliquity: a review of 68 patients. J Spinal Disord 1991;4:399410. [64] Maloney WJ, Rinsky LA, Gamble JG. Simultaneous correction of pelvic obliquity, frontal plane, and sagittal plane deformities in neuromuscular scoliosis using a unit rod with segmental sublaminar wires: a preliminary report. J Pediatr Orthop 1990; 10:7429. [65] McCarthy RE. Management of neuromuscular scoliosis. Orthop Clin North Am 1999;30:43549. [66] Miladi LT, Ghanem IB, Draoui MM, et al. Iliosacral screw xation for pelvic obliquity in neuromuscular scoliosis: a long-term follow-up study. Spine 1997; 22:17229. [67] Neustadt JB, Schuebarger HL, Cammisa FP. Spinal fusions to the pelvis augmented by CotrelDubousset instrumentation for neuromuscular scoliosis. J Pediatr Orthop 1992;12:4659. [68] Peelle MW, Lenke LG, Bridwell KH. Comparison of pelvic xation techniques in neuromuscular spinal deformity correction: Galveston rod versus iliac and lumbosacral screws. Spine 2006;31(20):23928 [discussion: 2399]. [69] Westerlund LE, Gill SS, Jarosz TS, et al. Posterioronly unit rod instrumentation and fusion for neuromuscular scoliosis. Spine 2001;26(18):19849. [70] Yazici M, Asher MA, Mardacker JW. The safety and ecacy of Isola-Galveston instrumentation and arthrodesis in the treatment of neuromuscular spinal deformities. J Bone Joint Surg Am 2000; 82A:52443. [71] Vialle R, Delecourt C, Morin C. Surgical treatment of scoliosis with pelvic obliquity in cerebral palsy. The inuence of intraoperative traction. Spine 2006;31(13):14616. [72] Huang MJ, Lenke LG. Scoliosis and severe pelvic obliquity in a patient with cerebral palsy. Surgical treatment utilizing halo-femoral traction. Spine 2001;26(19):216870. [73] Takeshita K, Lenke LG, Bridwell KH, et al. Analysis of patients with nonambulatory neuromuscular scoliosis surgically treated to the pelvis with intraoperative halo-femoral traction. Spine 2006;31(20): 23815. [74] Sink EL, Newton PO, Mubarak SJ, et al. Maintenance of sagittal plane alignment after surgical correction of spinal deformity in patients with cerebral palsy. Spine 2003;28:1396403. [75] Whitaker C, Burton DC, Asher M. Treatment of selected neuromuscular patients with posterior instrumentation and arthrodesis ending with lumbar pedicle screw anchorage. Spine 2000;25:23128.

[76] Wild A, Haak H, Kumar M, et al. Is sacral instrumentation mandatory to address pelvic obliquity in neuromuscular thoracolumbar scoliosis due to myelomeningocele? Spine 2001;26:E3259. [77] Tsirikos AI, Chang WN, Shah SA, et al. Preserving ambulatory potentialin pediatric patients with cerebral palsy who undergo spinal fusion using unit rod instrumentation. Spine 2003;28(5):4803. [78] Edwards CC. Reduction of spondylolisthesis. In: Bridwell KH, Dewald RL, editors. Textbook of spinal surgery. Second edition. Philadelphia: Lippincott-Raven; 1997. p. 131735. [79] Pritchett JW. The untreated unstable hip in severe cerebral palsy. Clin Orthop Relat Res 1983;173: 16972. [80] Lonstein JE, Beck K. Hip dislocation and subluxation in cerebral palsy. J Pediatr Orthop 1986;6: 5216. [81] Senaran H, Shah SA, Glutting JJ, et al. The associated eects of untreated unilateral hip dislocation in cerebral palsy scoliosis. J Pediatr Orthop 2006; 26(6):76972. [82] Dias RC, Miller F, Dabney K, et al. Surgical correction of spinal deformity using a unit rod in children with cerebral palsy. J Pediatr Orthop 1996;16: 73440. [83] Tsirikos AI, Chang W, Dabney KW, et al. Comparison of one-stage versus two-stage anteroposterior spinal fusion in pediatric patients with cerebral palsy and neuromuscular scoliosis. Spine 2003;28(12): 13005. [84] Sanders JO, Evert M, Stanley EA, et al. Mechanisms of curve progression following sublaminar (Luque) spinal instrumentation. Spine 1992;17:7819. [85] Guille JT, Sarwark JF, Sherk HH, et al. Congenital and developmental deformities of the spine in children with myelomeningocele. J Am Acad Orthop Surg 2006;14(5):294302 Review. [86] McCall RE, Hayes B. Long-term outcome in neuromuscular scoliosis fused only to lumbar 5. Spine 2005;30(18):205660. [87] Wimmer C, Wallnofer P, Walochnik N, et al. Com parative evaluation of Luque and Isola instrumentation for treatment of neuromuscular scoliosis. Clin Orthop Relat Res 2005;439:18192. [88] Gaine WL, Lim J, Stephenson W, et al. Progression of scoliosis after spinal fusion in Duchennes muscular dystrophy. J Bone Joint Surg 2004;86B: 5505. [89] Onimus M, Manzone P, Lornet JM, et al. Surgical treatment of scoliosis in bed-ridden patients with cerebral palsy. Rev Chir Orthop Reparatrice Appar Mot 1992;78:3128. [90] Teli M, Elsebaie H, Biant L, et al. Neuromuscular scoliosis treated by segmental third-generation instrumented spinal fusion. J Spinal Disord Tech 2005;18(5):4308. [91] Buchowski JM, Bhatnagar R, Skaggs DL, et al. Temporary internal distraction as an aid to

496

SARWARK & SARWAHI

correction of severe scoliosis. J Bone Joint Surg Am 2006;88(9):203541. [92] Shuebarger HL, Grimm JO, Bui V, et al. Anterior and posterior spinal fusion: staged versus same-day. Spine 1991;16:9303. [93] Powell ET IV, Krengel WF III, King HA, et al. Comparison of same-day sequential anterior and posterior spinal fusion with delayed two-stage anterior and posterior spinal fusion. Spine 1994;19:12569. [94] Ferguson RL, Hansen MM, Nicholas DA, et al. Same-day versus staged anterior-posterior spinal surgery in a neuromuscular scoliosis population:

The evaluation of medical complications. J Pediatr Orthop 1996;16(3):293303. [95] OBrien T, Akmakjian J, Ogin G, et al. Comparison of one-stage versus two-stage anterior/posterior spinal fusion for neuromuscular scoliosis. J Pediatr Orthop 1992;12:6105. [96] Hell AK, Campbell RM, Hefti F. The vertical expandable prosthetic titanium rib implant for the treatment of thoracic insuciency syndrome associated with congenital and neuromuscular scoliosis in young children. J Pediatr Orthop B 2005;14: 28793.

Orthop Clin N Am 38 (2007) 497509

Surgical Treatment of Congenital Scoliosis


Daniel J. Hedequist, MD
Department of Orthopedic Surgery, Childrens Hospital Boston, Harvard Medical School, 300 Longwood Avenue, Hunnewell 2, Boston, MA 02115, USA

The surgical options for congenital spine deformities are numerous and depend on the type of anomaly, the degree of deformity, and the age of the patient. The mainstay of surgical treatment remains early diagnosis before severe curvature and deformity exist. Occasionally, patients present with large deformities that require more signicant procedures; however, early limited arthrodesis remains the safest and most reliable procedure. Patients who have signicant decompensation of the spine at a young age may benet from a fusionless procedure, as do patients who have coexisting rib deformities and chest wall insuciency. This article reviews the general surgical principles that need to be followed to treat these patients safely. The numerous surgical procedures available for treating these patients also are reviewed. Preoperative evaluation The preoperative evaluation of a patient who has a congenital spine abnormality begins with a history and physical examination. The history includes a thorough past medical history given the association of congenital spine deformities with other organ system abnormalities (ie, of the cardiac and genitourinary systems). The physical examination should begin with a complete skeletal examination, with special attention focused on the neurologic examination and spinal examination. The neurologic examination should include basic testing of motor strength, sensory function, and reexes. Subtle signs should be looked for to assess for underlying ataxia or myelopathy. The spinal examination begins with the skin in the spinal midline, assessing for nevi, hemangiomas,
E-mail address: daniel.hedequist@childrens.harvard. edu

dimples, or hairy patches, which may be signs of underlying spinal dysraphism. Coronal and sagittal proles should be evaluated, as should general exibility of the deformity with manual hyperextension testing and manual side-bending tests. The spinal examination also should focus on the cervical spine given the association of Klippel-Feil syndrome with congenital scoliosis. Radiographs should include standard 36 standing radiographs in both planes as well as supine bending lms. Most deformities can be classied readily with plain lms, and any associated rib fusions are evident on the radiographs. The bending lms give a sense of the amount of exibility in adjacent areas of the spine. Some surgeons recommend spot lms of the deformity to further delineate the anatomy; however, the advent of newer imaging modalities make these lm unnecessary. Plain lms of the pelvis, looking at the triradiate cartilages and the Risser sign of the patient, as well as a left hand and wrist radiograph for bone age, are important in the pubescent age child as an indicator of growth when crankshaft is a concern [1,2]. MRI of the entire spine and brainstem is mandatory in any child undergoing surgical treatment of a congenital spine deformity. The presence of underlying spinal dysraphism is believed to be upward of 30% in children who have congenital scoliosis [3,4]. MRI scans also may be studied to look at the anatomy of the deformity. The presence of any cord compression is best studied by MRI, and this largely has replaced myelography as the imaging modality of choice (Fig. 1). The coronal and sagittal images may help to dene the segmentation and formation defects in the anterior spine. The axial views also can be viewed to look at the pedicle anatomy of the patient when contemplating transpedicular instrumentation.
orthopedic.theclinics.com

0030-5898/07/$ - see front matter 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.ocl.2007.05.002

498

HEDEQUIST

correlated with the intraoperative anatomy in patients undergoing hemivertebra excision [5]. These scans are especially helpful in dening the posterior anatomy anddwhen studied before the operationdcan aid the surgeon greatly in the operating room. The scans also have been shown to be eective in more complex deformities; the author and colleagues recommend them in all patients undergoing a deformity operation for congenital scoliosis. The radiation dosage required for a CT scan is a concern and can be reduced greatly by following a pediatric protocol; the software for creating 3-D images is readily available, and no additional radiation is needed. CT scans also have been shown to aid in assessing pulmonary volume before and after vertical expandable prosthetic titanium rib (VEPTR) placement [6,7]. General surgical principles Monitoring
Fig. 1. MRI of a patient with a congenital dislocation of the spine. Note the angular kyphosis and resulting spinal cord compression.

CT scanning has been studied and shown to be eective as a preoperative imaging study for congenital spine deformities (Fig. 2). In all cases, preoperative three-dimensional (3-D) CT scans

The use of neurologic monitoring is mandatory for surgical procedures in patients who have congenital spine deformities. The risk for neurologic injury associated with spine surgery is highest in patients who have congenital spine deformities [8]. Kyphotic deformities place the spine at most risk. Preoperative MRIs can help to rule out any coexisting spinal dysraphism. The author and colleagues use motor-evoked potential and somatosensory-evoked potential to minimize the risk for a neurologic decit [911]. At the authors institution, a wake-up test also is performed on all patients at the end of any procedure; this has been shown to be eective in even the youngest of patients [12,13]. Communication with the anesthesiologist and monitoring technician is crucial to minimize the chance that anesthetic agents impair the ability to monitor the spinal cord. If a patient has a coexisting spinal axis abnormality, referral to a neurosurgeon is paramount. In general, cord detetherings may be done at the same time as hemivertebra resections or in situ fusions. More complex operations requiring osteotomies or corrections with instrumentation should be done after addressing any associated dysraphism by the neurosurgeon. Positioning For most patients undergoing a posterior fusion with instrumentation, standard positioning on a radiolucent spinal bed with headpiece, chest bolster, and iliac crest pads suces. Usually,

Fig. 2. 3-D CT of a patient who has multiple hemivertebra.

SURGICAL TREATMENT OF CONGENITAL SCOLIOSIS

499

younger patients are placed on a radiolucent at operating table with gel rolls to support the torso. The author and colleagues have found that patients who have signicant kyphotic deformities may be managed best with a Mayeld head rest or cervical tongs to have better control of the head; instrumenting or working on the upper thoracic spine in these patients can be extremely dicult. Patients who are undergoing open or thoracoscopic anterior surgery may be placed in the standard lateral decubitus position. Thoracoscopic surgery also may be done with the patient in the prone position with the anesthesiologist using diminished tidal volume ventilation, thereby avoiding the pitfalls of single-lung ventilation [14]. Patients who are undergoing simultaneous anterior and posterior procedures should be placed in the lateral decubitus position (convex side up) on a at radiolucent operating frame with the entire anterior and posterior elds draped out [15]. These patients should be placed near the edge of the operating room table so that when posterior eld self-retaining retractors are placed, the handles do not hit the table and block the ability to retract (Fig. 3). Instrumentation The use of spinal instrumentation for congenital spine deformities was described rst by John Hall and colleagues [16], who used Harrington instrumentation. Instrumentation greatly aided in obtaining fusion in these patients, and it was believed to be safe and eective; however, the

original article described diculties in using this instrumentation in smaller children and cautioned against it. Before the use of instrumentation, the pseudarthrosis rate and loss of curve correction were high in these patients [17]. Newer-generation implants were manufactured to be size specic and were reported on rst in small children undergoing hemivertebra resection [18]. Further studies have shown that instrumentation is feasible and safe in smaller patients who have congenital spine deformities [19,20]. The rate of union and correction maintenance for posterior arthrodesis is greater with instrumentation than without instrumentation. The author and colleagues recommend using titanium instrumentation in all patients who have congenital spine deformities. This allows for further imaging of the spinal axis if needed as well as the ability to clearly image other parts of the thorax and abdomen with MRI in patients who have coexisting organ abnormalities, such as congenital heart disease or urologic abnormalities. Pedicle screw instrumentation of the thoracic spine has become more commonplace as reports have documented the safety and ecacy of segmental pedicle instrumentation [2124]. Pedicle screw placement has been shown to be technically possible in children as young as 1 year of age [25]. Their use in congenital spine deformities has been described in children undergoing hemivertebra resection. Hemivertebra resection remains the best scenario because frequently the instrumentation is placed at the level above and below the hemivertebra, which frequently are anatomically normal, rendering placement straightforward. The author and colleagues routinely use pedicle screws if the preoperative imaging studies suggest that this is feasible; however, the anatomic starting points can be dicult to nd when posterior element anatomy is abnormal. It is helpful to study the pedicle anatomy on the preoperative CT scan. Surgeons should be prepared to use hook or screw xation at available levels to create spinal stability for the maintenance of curve correction and arthrodesis. Allograft Autograft has long been the bone graft of choice in patients who have congenital scoliosis. Limitations of autograft include donor site morbidity when harvesting the iliac crest as well as the small volume available secondary to the size of the patient. Local grafts from rib excisions during anterior exposure or from hemivertebra excisions

Fig. 3. Intraoperative photo of a patient undergoing hemivertebra resection by way of simultaneous anteriorposterior exposures. Note the ability to work in the anteriorposterior elds at the same time allows for maximum visualization and control of the spine.

500

HEDEQUIST

are an important source of autograft and should be used when possible. The author and colleagues have extensive experience with allograft at their institution in patients who have congenital scoliosis. The rate of fusion and infection rate have been shown to be the same as in other series of pediatric deformity surgeries [26]. The author and colleagues use freeze-dried corticocancellous bone chips. These are readily available from numerous bone banks, are inexpensive, and there has never been a reported case of HIV transmission from freeze-dried allograft in spinal surgery. Crankshaft The crankshaft phenomenon is caused by continued anterior growth of the spine in the face of a solid posterior arthrodesis [1]. This leads to continued curve progression and deformity and was described rst in adolescent idiopathic scoliosis. Crankshaft phenomenon exists in congenital scoliosis as well; however, it is less predictable than in idiopathic scoliosis [27]. The potential for crankshaft to occur depends on the quality and number of discs anteriorly. Disc spaces that are clear and denable signify the potential for normal anterior growth; less denable disc space signies less growth potential. In children undergoing surgery for congenital spine deformities who are at risk for crankshaft (Risser 0, open triradiate cartilage), the disc spaces should be studied on the preoperative radiographs, CT scan, and MRI. Signicant growth potential requires an anterior open or thoracoscopic procedure to arrest growth. Traction In the past, preoperative traction was believed to be contraindicated in congenital spine deformities given the potential for traction-induced paralysis. Multiple investigators have revisited traction in pediatric spinal deformity and have shown it to be ecacious as a preoperative tool [2830]. Traction may be halo gravity, halo femoral, or halo pelvic (Fig. 4). The author and colleagues have found halo-gravity traction to be eective because the patients are allowed out of bed and into a wheelchair and are allowed to ambulate behind the wheelchair. Being out of bed and ambulatory has a positive eect on maintaining bone density, which may help to minimize implantbone failures. Traction should start slowly and be increased gradually with careful attention given to the neurologic examination. Any

neurologic decit or neurologic complaint by the patient calls for a diminishment of traction. Traction also may be used after an anterior release to get more correction of a deformity before a posterior instrumented fusion. The amount of time needed for traction to be ecacious is variable, and it should be used from 2 weeks to 2 months if needed. A nutrition consult should be done before the hospitalization in these patients because they are, or frequently become, nutritionally depleted. Although traction has been shown to be safe and ecacious, the author and colleagues do not recommend applying traction in cases of short, rigid, and angular deformities, especially when associated with kyphosis. These deformities do not have a positive response to traction and are at high risk for developing a neurologic decit. In situ fusion Fusion in situ remains the most reliable and safe operation for congenital spine deformities. The hallmark of treatment for congenital deformities is early diagnosis and recognition of a deformity that will be progressive. In situ fusion aords no correction of a deformity; however, it is an eective operation when done correctly on the appropriate patient. Patients with a fully segmented hemivertebra with no associated deformity may be good candidates for this procedure. The classic indication is a patient who has a unilateral bar or a unilateral bar with contralateral hemivertebra who is diagnosed early, before any signicant deformity [31,32]. Patients undergoing an in situ fusion need to have an anterior procedure when there is the potential for anterior growth. This can be done thoracoscopically or as an open procedure. The author and colleagues recommend the use of posterior instrumentation because it negates or minimizes the time needed for a cast or brace and improves fusion rate. Patients who have congenital kyphosis that is due to failure of formation may be fused in situ posteriorly. This allows for anterior growth over time to correct their deformity. The ideal indication is a deformity that is due to a congenital failure of formation that is less than 50 in a patient with signicant growth remaining [33,34].

Convex hemiepiphysiodesis Convex hemiepiphysiodesis is a procedure that requires growth to obtain correction over time [35]. The classic indication is a patient with a fully

SURGICAL TREATMENT OF CONGENITAL SCOLIOSIS

501

Fig. 4. A patient in wheelchair halo-gravity traction.

segmented hemivertebra without any associated deformity. The procedure is done as an anterior and posterior fusion on the convex side of the curvature. The concave side retains its growth potential and, in theory, allows for some correction of the deformity; however, in practice, the amount of correction is negligible. It is not a reliable procedure for many patients if any correction is desired. It is contraindicated if there is no concave growth potential. Hemivertebra excision Hemivertebra excision remains an enticing surgical option for a hemivertebra causing a progressive curve and deformity. The authors and colleagues have found that the ideal patient remains a child younger than 5 years of age with a fully segmented hemivertebra at the thoracolumbar junction, lumbar spine, or lumbosacral spine. In

general, most hemivertebra in the thoracic spine are not problematic. Hemivertebra at the cervicothoracic or cervical region occasionally become problematic, and excision in these areas has been described [36]. Earlier publications on hemivertebra excision reported the occurrence of neurologic decits; however, recent series demonstrated that hemivertebra excision can be performed safely [15,21,3638]. Hemivertebra may be resected by an anteriorposterior procedure or by a posterioronly procedure. The author and colleagues described simultaneous anteriorposterior approaches with excision in two separate studies at their institution; however, they have moved to a posterior-only excision because Ruf and Harms [19] reported success using this technique. The excision of a hemivertebra by way of an anteriorposterior technique can be done reliably by way of simultaneous exposures. The patient should be placed in the lateral position with

502

HEDEQUIST

a bolster under the concave side of the deformity and the table slightly broken. The anterior exposure should be undertaken as a standard thoracotomy or with the transthoracic-retroperitoneal or retroperitoneal approach, depending on the level of the hemivertebra. Once the anterior exposure is performed, the posterior spine is exposed, and a spot radiograph is taken to conrm the correct level. The author and colleagues place pedicle screws bilaterally at the levels above and below the hemivertebra. Laminar hook sites also are prepared on the convex side of the spine for compression across the deformity. Resection begins with the anterior vertebral body after the discs above and below have been removed. The body should be resected posteriorly toward the canal, and the posterior vertebral body should be resected back to the posterior longitudinal ligament. Using a diamond-tipped burr can facilitate safe resection of the posterior body. Then, the pedicle, as it enters the vertebral body, can be resected under direct visualization. Posterior resection begins with resection of the hemilamina, which generally is done with Kerrison rongeur. The transverse process can be resected, and the resection can continue by working down the pedicle. The nerve roots above and below the hemivertebra pedicle need to be protected. The remaining pedicle can be resected through both incisions. The bolster is removed and the table unbroken; the wedge resection can be closed down with downward pressure on the spine through the anterior incision. The rst rod is inserted into the laminar hooks, which are used to compress across the deformity, achieving correction. The author and colleagues have found that the pedicles in young patients who have congenital scoliosis do not tolerate compression reliably, given the small pedicle size and small screw diameter. The remaining two rods are placed in the pedicle screws and seated. This allows for maximum xation with the three rods. The rod diameter is 3.2 or 3.5 mm in younger children. The author and colleagues have not had any implant-related problems with the three-rod technique and have adapted it to the posterior-only resections as well. Resection by way of a posterior-only approach is a challenging operation secondary to blood loss and the need to work around the cord or cauda from the posterior side. These factors make visualization more dicult than anteriorposterior procedures. The spine is exposed posteriorly, and the pedicle screws are placed into position. Resection begins by dissecting over the transverse

process with a Cobb elevator and following this plane with a Cobb elevator down the lateral aspect of the vertebral body to the anterior portion, which will help with retraction. Then, the hemilamina can be resected followed by the transverse process. The pedicle resection can begin, with care taken to place a retractor on the medial pedicle wall protecting the dural contents. The pedicle can be resected by burring down the cancellous bone and resecting the walls with rongeurs, taking care to protect the nerve roots above and below. Moving anteriorly down the pedicle, the author and colleagues have found it useful to use a bipolar cautery to seal the epidural veins, which lie on the inner aspect of the pedicle and vertebral body. Continued anterior protection of visceral structures anteriorly and anterolaterally is necessary as one moves forward with the vertebral body resection. The discs above and below can be resected, taking care to move across the midline to the concave side to get a complete wedge resection. Complete correction of the deformity can be performed using laminar hooks to compress across the convexity. Rods can be placed into bilateral pedicle screws (Fig. 5). The author and colleagues have found it advantageous to place the child in a custom-molded TLSO for 3 months; children younger than 3 years of age may be candidates for a Risser-type cast for a period of 4 weeks. A comparison of posterior only-resection versus simultaneous anteriorposterior resection yields results that are similar with regard to curve correction, blood loss, and fusion. Earlier reports of posterior-only resection with segmental xation had a higher implant failure rate; however, the author and colleagues experience with the threerod technique has not borne this out. They prefer to resect hemivertebra from a posterior-only procedure when feasible, because this avoids an anterior procedure. Relative contraindications include the surgeon being uncomfortable with a posterior-only technique or a patient who is at risk for signicant bleeding (eg, congenital heart disease) that would render visualization from a posterior procedure dicult.

Partial correction with instrumentation and fusion The earliest series of patients who had congenital scoliosis that was treated with posterior fusions were fraught with loss of correction and

SURGICAL TREATMENT OF CONGENITAL SCOLIOSIS

503

Fig. 5. (A) Standing posteroanterior radiograph of a patient who has a fully segmented hemivertebra. (B) Standing postoperative radiograph after posterior-only excision of the hemivertebra with excellent curve correction.

pseudarthrosis given the lack of implants [17,31,39]. The use of instrumentation for congenital spine deformities has evolved into modern segmental implants with improved maintenance of correction and arthrodesis rate [20]. The ability to aord any correction of a congenital anomaly depends on the type of anomaly and the amount of normally segmented spine adjacent to the anomaly. The principles of stable vertebra apply when trying to balance a decompensated spine in congenital scoliosis. Studying presurgical standing radiographs and bending lms can help to dene stable zones of instrumentation and fusion. Frequently, it is possible to obtain a balanced spine by way of posterior (only) arthrodesis and instrumentation. Positioning of the patient, thorough facetectomies, and stable instrumentation frequently achieve an arthrodesis with a balanced spine. Care must be taken not to apply distractive forces with implants or to attempt signicant corrections with instrumentation in the case of a rigid deformity. Transpedicular segmental instrumentation can achieve large curve corrections and possibly prevent crankshaft in patients who have idiopathic scoliosis; however, segmental screw xation frequently is not possible given abnormal anatomic landmarks and pedicle anatomy in congenital deformities, and surgeons should be familiar with all available posterior implant anchors.

Occasionally, osteotomies (eg, Smith-Peterson or pedicle subtraction) are necessary to obtain spinal balance; these also may need to be done in concert with anterior releases. Osteotomies place the patient are at a higher risk for neurologic decit and should be done only in concert with neurologic monitoring and the ability to achieve solid, stable xation. The planning of the osteotomies is facilitated by preoperative CT scans (Fig. 6). These are highly challenging cases and should be done only by experienced hands; occasionally, the author and colleagues have found it useful to have their neurosurgical colleagues available when decompression is needed. Although vertebral column resections are needed rarely, when a patient has a severe deformity that cannot be managed by other means, they have been performed safely with good results [29,40]. Vertebral resection also may be indicated when a kyphotic deformity is creating spinal cord compression that needs to be addressed. Frequently, kyphotic deformities are best operated on through posterior approaches, such as a costotransversectomy. Vertebral column resections may be done as a staged anteriorposterior operation or as a posterioronly resection. Usually, circumferential cases are staged because of the signicant blood loss and the technically demanding nature of the procedure. The anterior procedure begins with a thorough exposure, and it should

504

HEDEQUIST

Fig. 6. (A, B) Preoperative standing radiographs of an adolescent patient who has progressive congenital kyphoscoliosis. (C) 3-D CT scan of the patient showing the apical kyphoscoliotic anomaly. (D, E) Postoperative standing radiograph after posterior wedge osteotomy with excellent correction of the deformity and spinal balance.

include an osteoperiosteal ap to safely dissect around to the far side of the spine and to protect against visceral structures. Resection of the anterior body is performed all the way back to the posterior longitudinal ligament. Once complete resection is done, gelfoam should be place over

the dura followed by abundant autograft and allograft. The posterior resection usually can be performed a week later. Stable xation is placed rst during the posterior procedure, which is followed by resection of the posterior elements and careful correction of the deformity. Neurologic

SURGICAL TREATMENT OF CONGENITAL SCOLIOSIS

505

Fig. 6 (continued )

monitoring is essential during these cases; frequently, multiple wake-up tests are needed. The author and colleagues have found it benecial to keep the arterial blood pressure elevated during correction of the deformity to prevent cord ischemia. Posterior-only vertebral column resections have come into favor to avoid any anterior surgery. Frequently, this is the best option if there is a corresponding kyphotic deformity that would make access to the vertebral column dicult through the standard anterior approach [29,40,41]. Growing rods In the past, patients younger than 5 years of age who had congenital deformities involving long sections of the spine or with large compensatory curves in normally segmented spine were treated with early arthrodesis to create a balanced spine. Recent studies showed that early arthrodesis over a longer section of the thoracic spine before 5 years of age may be associated with a signicant reduction in pulmonary function and chest height. This had led to interest in the use of growing rods for select congenital spine anomalies (Fig. 7).

Recent reports documented successful treatment of early-onset deformities with growing rods [4244]. Growing rods rely on normally segmented areas of spine to maintain growth while instrumentation aids in curve control. Proximal and distal anchors are placed by using claw foundations, and spanning submuscular rods are used to gain correction while allowing for growth. Lengthenings are done every 6 months. The ideal patient is younger than 5 years of age and has a congenital deformity involving a long section of the spine in which an arthrodesis would have a detrimental aect on growth. In general, the proximal and distal anchor points need to be established in reasonably normal areas of spine to allow for sucient xation. Patients who have progressive curves that have congenital anomalies that will not allow for stable anchor placement or associated congenital rib fusions that need to be addressed with thoracostomy may be better served with VEPTR placement. Expansion thoracoplasty and vertical expandable prosthetic titanium rib Occasionally, congenital spine anomalies are associated with congenital rib fusions [45]. These

506

HEDEQUIST

Fig. 7. (A) Preoperative radiograph of a 4-year-old patient who underwent a previous hemivertebra resection. This patient has had progressive decompensation above and below the resection site. (B) Postoperative radiograph after placement of a dual growing-rod construct.

may act together and cause progressive curvature of the spine with progressive diminishment in pulmonary function and hemithorax volume. The resultant syndrome, termed thoracic insuciency, causes limitation of pulmonary function [7]. The progressive loss of hemithorax height is believed to be responsible for this loss of function. Expansion of the hemithorax by opening wedge osteotomies into the area of congenital rib fusions or adhesions was shown to have a positive eect on lung function and lung volume [6,46]. The expansion of the hemithorax is held in place by a VEPTR device. This device anchors proximally around the second or third ribs. Distally, the VEPTR may anchor into one of three places: around distal ribs, into the lumbar spine, or around the ilium (Fig. 8). The device can be expanded through connectors every 6 months. The eect is one of increasing thorax height, improving lung function, and improving lung volume. The device also allows for improvement in Cobb measurements over time and control of corresponding spinal curvature. The ideal indication is a patient who has a constricted hemithorax secondary to congenital rib fusions. In general, the rib fusions create a sti hemithorax; expansion thoracoplasty and VEPTR do not improve chest wall stiness and, in fact, may make chest wall

compliance worse. Patients who do not have rib fusions and have a compliant chest wall may be better served with other treatment modalities.

Summary Successful surgical treatment of patients who have congenital spine anomalies relies on applying general principles to the correct operation. A thorough preoperative assessment is mandatory and includes a standard history and physical examination as well as obtaining appropriate imaging tests. The correct surgical procedure depends on the anomaly itself and the degree of deformity. Early recognition and surgical treatment before severe deformity develops is the hallmark of treatment. Curvatures associated with a moderate degree of deformity can be managed successfully by fusion with instrumentation. Correction can be achieved by positioning and facetectomies. Occasionally, osteotomies are needed to obtain a balanced spine. More severe deformities may require traction with osteotomies or vertebral column resection. Curvature in younger children involving long, normally segmented areas of the spine may be managed best by growing rods. Patients with signicant growth

SURGICAL TREATMENT OF CONGENITAL SCOLIOSIS

507

Fig. 8. (A) Intraoperative photograph of a patient who underwent expansion thoracoplasty with placement of a rib-torib VEPTR and rib-to-spine VEPTR. Pre- (B) and postoperative (C) radiographs after expansion thoracoplasty with VEPTR placement for congenital rib fusions with congenital scoliosis.

remaining and associated chest wall anomalies may be best served by expansion thoracoplasty and VEPTR.

References
[1] Dubousset J, Herring JA, Shuebarger H. The crankshaft phenomenon. J Pediatr Orthop 1989; 9(5):54150. [2] Sanders JO, Browne RH, McConnell SJ, et al. Maturity assessment and curve progression in girls with idiopathic scoliosis. J Bone Joint Surg Am 2007;89(1):6473. [3] Belmont PJ Jr, Kuklo TR, Taylor KR, et al. Intraspinal anomalies associated with isolated congenital

hemivertebra: the role of routine magnetic resonance imaging. J Bone Joint Surg Am 2004;86(8):170410. [4] Prahinski JR, Polly DW Jr, McHale KA, et al. Occult intraspinal anomalies in congenital scoliosis. J Pediatr Orthop 2000;20(1):5963. [5] Hedequist DJ, Emans JB. The correlation of preoperative three-dimensional computed tomography reconstructions with operative ndings in congenital scoliosis. Spine 2003;28(22):25314 [discussion: 1]. [6] Emans JB, Caubet JF, Ordonez CL, et al. The treatment of spine and chest wall deformities with fused ribs by expansion thoracostomy and insertion of vertical expandable prosthetic titanium rib: growth of thoracic spine and improvement of lung volumes. Spine 2005;30(Suppl 17):S5868.

508

HEDEQUIST

[7] Campbell RM Jr, Smith MD, Mayes TC, et al. The eect of opening wedge thoracostomy on thoracic insuciency syndrome associated with fused ribs and congenital scoliosis. J Bone Joint Surg Am 2004;86(8):165974. [8] MacEwen GD, Bunnell WP, Sriram K. Acute neurological complications in the treatment of scoliosis. A report of the Scoliosis Research Society. J Bone Joint Surg Am 1975;57(3):4048. [9] Lubicky JP, Spadaro JA, Yuan HA, et al. Variability of somatosensory cortical evoked potential monitoring during spinal surgery. Spine 1989;14(8):7908. [10] Dawson EG, Sherman JE, Kanim LE, et al. Spinal cord monitoring. Results of the Scoliosis Research Society and the European Spinal Deformity Society survey. Spine 1991;16(Suppl 8):S3614. [11] Mooney JF 3rd, Bernstein R, Hennrikus WL Jr, et al. Neurologic risk management in scoliosis surgery. J Pediatr Orthop 2002;22(5):6839. [12] Brustowicz RM, Hall JE. In defense of the wake-up test. Anesth Analg 1988;67(10):1019. [13] Jones ET, Matthews LS, Hensinger RN. The wakeup technique as a dual protector of spinal cord function during spine fusion. Clin Orthop Relat Res 1982;168:1138. [14] Sucato DJ, Elerson E. A comparison between the prone and lateral position for performing a thoracoscopic anterior release and fusion for pediatric spinal deformity. Spine 2003;28(18):217680. [15] Hedequist DJ, Hall JE, Emans JB. Hemivertebra excision in children via simultaneous anterior and posterior exposures. J Pediatr Orthop 2005;25(1):603. [16] Hall JE, Herndon WA, Levine CR. Surgical treatment of congenital scoliosis with or without Harrington instrumentation. J Bone Joint Surg Am 1981;63(4):60819. [17] Winter RB, Moe JH, Lonstein JE. Posterior spinal arthrodesis for congenital scoliosis. An analysis of the cases of two hundred and ninety patients, ve to nineteen years old. J Bone Joint Surg Am 1984; 66(8):118897. [18] Lazar RD, Hall JE. Simultaneous anterior and posterior hemivertebra excision. Clin Orthop Relat Res 1999;364:7684. [19] Ruf M, Harms J. Hemivertebra resection by a posterior approach: innovative operative technique and rst results. Spine 2002;27(10):111623. [20] Hedequist DJ, Hall JE, Emans JB. The safety and ecacy of spinal instrumentation in children with congenital spine deformities. Spine 2004;29(18): 20816 [discussion: 2087]. [21] Nakamura H, Matsuda H, Konishi S, et al. Singlestage excision of hemivertebrae via the posterior approach alone for congenital spine deformity: followup period longer than ten years. Spine 2002;27(1): 1105. [22] OBrien MF, Lenke LG, Mardjetko S, et al. Pedicle morphology in thoracic adolescent idiopathic

[23]

[24]

[25]

[26] [27]

[28]

[29]

[30]

[31]

[32]

[33]

[34]

[35]

[36] [37]

[38]

[39]

scoliosis: is pedicle xation an anatomically viable technique? Spine 2000;25(18):228593. Kim YJ, Lenke LG, Bridwell KH, et al. Free hand pedicle screw placement in the thoracic spine: is it safe? Spine 2004;29(3):33342 [discussion: 342]. Kuklo TR, Lenke LG, OBrien MF, et al. Accuracy and ecacy of thoracic pedicle screws in curves more than 90 degrees. Spine 2005;30(2):2226. Ruf M, Harms J. Pedicle screws in 1- and 2-year-old children: technique, complications, and eect on further growth. Spine 2002;27(21):E4606. Hedequist D, Emans J. Congenital scoliosis: a review and update. J Pediatr Orthop 2007;27(1):10616. Kesling KL, Lonstein JE, Denis F, et al. The crankshaft phenomenon after posterior spinal arthrodesis for congenital scoliosis: a review of 54 patients. Spine 2003;28(3):26771. Sink EL, Karol LA, Sanders J, et al. Ecacy of perioperative halo-gravity traction in the treatment of severe scoliosis in children. J Pediatr Orthop 2001; 21(4):51924. Suk SI, Chung ER, Lee SM, et al. Posterior vertebral column resection in xed lumbosacral deformity. Spine 2005;30(23):E70310. Rinella A, Lenke L, Whitaker C, et al. Perioperative halo-gravity traction in the treatment of severe scoliosis and kyphosis. Spine 2005;30(4):47582. Winter RB, Moe JH. The results of spinal arthrodesis for congenital spinal deformity in patients younger than ve years old. J Bone Joint Surg Am 1982; 64(3):41932. McMaster MJ, David CV. Hemivertebra as a cause of scoliosis. A study of 104 patients. J Bone Joint Surg Br 1986;68(4):58895. McMaster MJ, Singh H. Natural history of congenital kyphosis and kyphoscoliosis. A study of one hundred and twelve patients. J Bone Joint Surg Am 1999;81(10):136783. McMaster MJ, Singh H. The surgical management of congenital kyphosis and kyphoscoliosis. Spine 2001;26(19):214654 [discussion: 2155]. Uzumcugil A, Cil A, Yazici M, et al. Convex growth arrest in the treatment of congenital spinal deformities, revisited. J Pediatr Orthop 2004;24(6): 65866. Ruf M, Jensen R, Harms J. Hemivertebra resection in the cervical spine. Spine 2005;30(4):3805. Polly DW Jr, Rosner MK, Monacci W, et al. Thoracic hemivertebra excision in adults via a posterior-only approach. Report of two cases. Neurosurg Focus 2003;14(2):e9. Shaw BA. Neurological risk management in scoliosis surgery. J Pediatr Orthop 2003;23(4):564 [author reply: 564]. Winter RB, Moe JH, Lonstein JE. The surgical treatment of congenital kyphosis. A review of 94 patients age 5 years or older, with 2 years or more follow-up in 77 patients. Spine 1985;10(3):22431.

SURGICAL TREATMENT OF CONGENITAL SCOLIOSIS

509

[40] Suk SI, Chung ER, Kim JH, et al. Posterior vertebral column resection for severe rigid scoliosis. Spine 2005;30(14):16827. [41] Smith JT, Gollogly S, Dunn HK. Simultaneous anterior-posterior approach through a costotransversectomy for the treatment of congenital kyphosis and acquired kyphoscoliotic deformities. J Bone Joint Surg Am 2005;87(10):22819. [42] Blakemore LC, Scoles PV, Poe-Kochert C, et al. Submuscular Isola rod with or without limited apical fusion in the management of severe spinal deformities in young children: preliminary report. Spine 2001;26(18):20448. [43] Akbarnia BA, Marks DS, Boachie-Adjei O, et al. Dual growing rod technique for the treatment of

progressive early-onset scoliosis: a multicenter study. Spine 2005;30(Suppl 17):S4657. [44] Thompson GH, Akbarnia BA, Kostial P, et al. Comparison of single and dual growing rod techniques followed through denitive surgery: a preliminary study. Spine 2005;30(18):203944. [45] Tsirikos AI, McMaster MJ. Congenital anomalies of the ribs and chest wall associated with congenital deformities of the spine. J Bone Joint Surg Am 2005; 87(11):252336. [46] Campbell RM Jr, Smith MD, Hell-Vocke AK. Expansion thoracoplasty: the surgical technique of opening-wedge thoracostomy. Surgical technique. J Bone Joint Surg Am 2004;86(Suppl 1): 5164.

Orthop Clin N Am 38 (2007) 511520

Lenke 1C, King Type II Curves: Surgical Recommendations


B. Stephens Richards, MDa,b,*
a

University of Texas Southwestern Medical Center, 5323 Harry Hines Boulevard, Dallas, TX 75390, USA b Texas Scottish Rite Hospital for Children, 2222 Welborn Street, Dallas, TX 75219, USA

The goals of surgical stabilization for spinal deformity are to achieve a well-balanced spine in which the patients head and trunk are centered over the pelvis, to preserve optimal function of the patient, and to achieve an appearance that is acceptable to the patient and family. In recent years, there has been substantial development of spinal instrumentation systems and surgical techniques, both of which allow for increasing amounts of scoliosis correction. All pedicle screw posterior constructs have gained increased use by spinal deformity surgeons over the past decade [19]. To a lesser extent, thoracoscopic anterior instrumentation also has been used [1013]. With these newer devices, an increase in curve correction is obtained and greater reduction of the rib prominence is possible; with the use of anterior implants, preservation of more inferior motion segments is possible because of shorter construct lengths. The objectives of this article are twofold. The rst is to discuss the surgical experience in patients who have Lenke 1C (King type II) curves, including the experience with newer instrumentation. These patients present a signicant challenge to the deformity surgeon because they have structural thoracic deformities and signicant, but nonstructural, lumbar curves. Over the years, selective instrumentation and fusion of the thoracic curve have been the primary approach to preserve motion of the lumbar segments, and thus, improve the long-term outlook. This strategy has been

undertaken with the expectation that the uninstrumented lumbar curve will allow for a well-balanced spinal column postoperatively. Unfortunately, spinal imbalance following surgery was recognized in some patients and has been the topic of numerous reports in the literature. The second objective of this article is to present basic strategies that may be helpful when planning for surgery in these patients.

Lenke 1C (King type II) curves: the terminology In 1983, King and colleagues [14] introduced a radiographic classication system for adolescent idiopathic scoliosis in which ve curve types were described. The two more common curve patterns were single thoracic curves (type III) and primary thoracic curves accompanied by similar, but slightly smaller, lumbar curves (type II). This classication system provided useful recommendations for surgical intervention using Harrington instrumentation. The thoracic curves were fused down to the stable vertebra (the vertebra in the thoracolumbar area that is bisected by the center sacral vertical line [CSVL]), and the overall spinal balance remained intact. With the evolution of multiple hook or screw implant systems in the late 1980s, which allowed for increased three-dimensional curve correction, similar surgical recommendations were followed. Unexpectedly, in some patients who had King type II curves, postoperative spinal imbalance occurred following selective thoracic fusions. At the same time, diculty in distinguishing some curve patterns as King type III curves or King type II curve patterns led to inconsistencies in the reporting of the literature.
orthopedic.theclinics.com

* Texas Scottish Rite Hospital for Children, 2222 Welborn Street, Dallas, TX 75219. E-mail address: steve.richards@tsrh.org

0030-5898/07/$ - see front matter 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.ocl.2007.05.004

512

RICHARDS

The Lenke classication system for adolescent idiopathic scoliosis was introduced in 1997 and provided, among other things, a better method to dierentiate King type II curves from King type III curves [15,16]. King types II and III curves are included in the Lenke type 1 curve pattern, in which the primary deformity is a structural main thoracic curve. If a lumbar curve is also present, it is nonstructural (deemed nonstructural if it measures !25 on a supine bend radiograph). Lumbar spine modier A, B, or C describes the variations seen in these lumbar curves (Fig. 1). If the CSVL is located between the pedicles, the lumbar modier is A. If the CSVL does not touch any part of the apical vertebral body of the lumbar curve, the lumbar modier is C. All others are classied as having a lumbar modier B; these curves generated the most confusion in dierentiating King type II curves from King type III curves. Even today, some studies report single overhang curves (like King type III) as Lenke 1A/1B curve patterns [17], whereas others report thoracic curves with signicant nonstructural lumbar components (like King type II) as Lenke 1B/ 1C curve patterns [18,19]. Without question, the Lenke 1C curve pattern is representative of the originators initial intention of a King type II curve and is the primary deformity that is most at risk for postoperative spinal imbalance

following selective thoracic instrumentation and fusion [1,20]. Furthermore, if the lumbar curve measures at least 25 on the preoperative bend radiograph, it is deemed structural, and the scoliotic deformity is classied as a Lenke 3C curve type. Unlike Lenke 1C curves, patients who have Lenke 3C curves nearly always require fusion of the thoracic and lumbar curves. Spinal deformity surgeons would agree that Lenke 1A and 1B curves require fusion of the thoracic curve only, and most, but not all, agree that selective fusion of Lenke 1C curves is justied [1,19,21]. This article attempts to provide some general guidelines to help the deformity surgeon decide when selective fusions are appropriate for Lenke 1C curves. Lenke 1C (King type II) curves: the literature When selective posterior thoracic fusion to the stable vertebra was undertaken in King type II curve patterns during the Harrington instrumentation era, curve correction approaching 40% was consistently reported, and spinal balance was maintained [2224]. With the advent of the newer derotation hook instrumentation systems in the latter 1980s, selective thoracic fusion continued to be the most common approach to treatment in this population, often with satisfactory results (Fig. 2).

Fig. 1. (A) In Lenke 1A curves, the CSVL is located between the pedicles at the lumbar spine up to the stable vertebra. (B) In Lenke 1B curves, CSVL touches the apex of the lumbar curve, between the medial border of the lumbar concave pedicle and the concave lateral margin of the apical vertebral body. (C) In Lenke 1C curves, the CSVL falls completely medial to the entire concave lateral aspect of the lumbar apical vertebral body. Lenke 1C curves are most clearly representative of King type II curves.

LENKE 1C, KING TYPE II CURVES

513

Fig. 2. This patient underwent selective thoracic fusion to the stable vertebra (T12) with excellent balance. The thoracic curve correction approximated the preoperative supine side-bending radiograph. (A, B) Preoperative radiographs. (C, D) Radiographs taken 2 years postoperatively.

Over time, reports of spinal imbalance surfaced, manifested by a shift in the patients trunk or head (or both) to the left of midline [2536]. This undesirable nding was believed to be the result of too much correction of the thoracic curve, a nding usually not seen with Harrington instrumentation. The uninstrumented lumbar curve below did not spontaneously correct to the extent expected by review of the preoperative bend radiographs. Explanations that have been proposed to explain this nding include the improper choice of distal fusion levels, unrecognized lumbar curve stiness and progression, and the use of incorrect hook/screw patterns. Instead of two points of xation (as used with Harrington distraction instrumentation), multiple hooks that rigidly attached the rod to the spine imparted stronger corrective forces to the scoliotic deformity. Once the rod was seated in the hooks and the scoliosis was corrected by the 90 rod rotational maneuver, the forces imparted to the exible spine below had a deleterious eect on its ability to spontaneously balance the instrumented thoracic curve. The lumbar curves remained larger after surgery than the instrumented thoracic curves, and the patients decompensated to the left. This was accompanied by postoperative persistence of obliquity in the lumbosacral region (between L4 and the pelvis), a nding that often was identied preoperatively

butdbased on the bend radiographsdwas expected to improve following surgery [30,34,37]. Over time, in most cases in which there was some postoperative imbalance, improvement occurred as the uninstrumented lumbar curve adapted (Fig. 3). Failure to properly distinguish King type II curves from true double major curves was responsible for some instances of postoperative imbalance following selective thoracic fusions. Guidelines were developed to help dierentiate King type II and true double major curve patterns [29]. Relative ratios between the thoracic and lumbar curves with regard to their size, rotation, and deviation from the midline were used to assist the surgeon in choosing the appropriate candidate for selective fusion. These relative ratios were assessed preoperatively on a standing radiograph (ie, thoracic curve parameter/lumbar curve parameter). If the ratios were less than 1.0, both curves required fusion. If the ratios were greater than 1.2 for curve size and for apical vertebral displacement and greater than 1.0 for curve rotation, selective thoracic fusion could be performed safely. These guidelines were used regularly with the newer hookrod instrumentation systems before the Lenke classication system was developed. Following the newer classications development, less emphasis was placed on the relative ratios

514

RICHARDS

Fig. 3. (A, B) This 12-year, 10-month-old girl underwent selective thoracic fusion to the stable vertebra. (C, D) Six weeks after surgery, she had signicant imbalance to the left. She was not braced for this. (E, F) Her balance improved spontaneously over time, as seen at the 4-year postoperative evaluation.

between the thoracic and lumbar curves and more emphasis was again placed on the assessment of lumbar curve exibility (structural/nonstructural) and on the lumbar modiers [16,38]. If the lumbar curve bends to less than 25 , then it is considered a compensatory nonstructural curve (Lenke 1C or

King type II curve pattern) rather than a structural curve (Lenke 3C or double curve pattern). With true double major curves (Lenke 3C), both curves needed to be included in the posterior fusion to achieve a balanced spine using segmental xation systems.

LENKE 1C, KING TYPE II CURVES

515

Selective anterior instrumentation has been compared with posterior hookrod instrumentation in patients who had Lenke 1B and 1C curves [5,1820,39,40]. In those undergoing anterior fusion, the distal extent of the fusion usually stopped at the end vertebra of the thoracic curve Cobb measurement, thus preserving distal motion segments (Fig. 4). When posterior fusion was used with hookrod constructs, the selective fusion usually extended to the stable vertebra. Thoracic curve correction (and spontaneous lumbar curve correction) is slightly greater with anterior fusion, but postoperative spinal balance is similar in both groups. The best coronal balance following surgery is noted when the thoracic curve is corrected intraoperatively to a measurement close to, but not more than that of, the preoperative thoracic push-prone Cobb measurement. In Lenke 1C curves with satisfactory balance after surgery, patients who undergo posterior fusion have approximately 40% thoracic curve correction. Selective posterior fusions using all-hook constructs also have been compared with pedicle screw constructs in patients who had Lenke 1C curve patterns (Fig. 5) [1]. The total levels

instrumented did not vary signicantly between these two groups. The amount of correction of the thoracic curve obtained surgically and the amount of spontaneous lumbar curve correction were greater in the pedicle screw group. The incidence of postoperative coronal decompensation, dened as a greater than 20-mm shift to the left of the C7 plumbline, was higher in the hook group (average shift, 21 mm) than in the pedicle screw group (average shift, 17.3 mm). This report suggested that increased thoracic curve correction, when using pedicle screws, does not lead to the same amount of postoperative imbalance as noted in previous studies using hook constructs. The investigators postulated that screw xation oers better control of the end vertebrae, which prevents decompensation. They recommended that correction of the thoracic curve with pedicle screws should still mimic the correction obtained on a preoperative push-prone radiograph to minimize potential decompensation of the unfused curve. Not all scoliosis surgeons agree that selective thoracic fusion in Lenke 1C (King type II) curves is justied [19,21]. A 2003 multicenter study described factors that inuenced surgeons decisions

Fig. 4. (A) Preoperative Lenke 1C curve type with 58 thoracic curve magnitude. (B) The anterior instrumentation ended at the end vertebra of the thoracic curve (T11), resulting in satisfactory postoperative balance.

516

RICHARDS

Fig. 5. (A) Preoperative thoracic curve measures 60 . (B) Selective thoracic fusion using pedicle screws distally to the stable vertebra resulted in satisfactory postoperative balance.

about whether to perform a selective thoracic fusion or to include both curves in the fusion [19]. Two hundred and ve Lenke 1B and 1C curves were reported on, all referred to as King-Moe type II curves. Anterior instrumentation and fusion were performed in 75% of the patients. Posterior instrumentation and fusion were undertaken in the remaining 25%. The investigators categorized a fusion as being selective if the implant stopped at L1 or above and nonselective if the implant extended to L2 or below. The incidence of including the lumbar curve in the fusion ranged from 6% to 33% at the dierent patient care sites. Lenke 1C curves had a much higher rate of nonselective fusion when compared with Lenke 1B curves. Factors associated with nonselective fusion included larger preoperative lumbar curve magnitude, greater displacement of the lumbar apical vertebra from the CSVL, and a smaller thoracic/lumbar curve magnitude ratio. The most agreed upon radiographic determinant for choosing the extent of distal fusion was the absolute magnitude of the lumbar apical vertebral deviation from the CSVL. In those patients who underwent nonselective fusion, the average displacement exceeded 3 cm from midline. The investigators found that side-bending correction of the lumbar curve to

less than 25 was not a sucient criterion to perform a selective fusion in some cases. Surgical recommendations for Lenke 1C (King type II) curve patterns When surgery is deemed necessary, all patients who have Lenke 1C (King type II) curves should be considered for selective thoracic fusion because most would have satisfactory outcomes; however, a small percentage of patients will be identied preoperatively who may not qualify for selective thoracic fusion. The criteria used to identify them are listed later. For most patients who are candidates for selective fusion, recommendations for the distal level of fusion vary little, regardless of the type of posterior instrumentation used. If hookrod constructs are used, the fusion should be extended to, but never past, the stable vertebra. Fusing one or two levels lower than this tends to shift the patients trunk further to the left. If the CSVL bisects a disk space instead of a vertebra, the instrumentation should end on the vertebra immediately above. If the thoracic curve is fused shorter than this when using hooks, postoperative adding-on of the thoracic curve has been reported (Fig. 6). If posterior pedicle screw xation is used,

LENKE 1C, KING TYPE II CURVES

517

Fig. 6. (A, B) Preoperative radiographs of a 13-year-old girl who underwent selective thoracic fusion ending one level above the stable vertebra. (C, D) Two weeks postoperatively, her balance was excellent. (E, F) One year following surgery, adding-on of the thoracic curve was evident radiographically.

518

RICHARDS

similar criteria should be followed, despite the knowledge that pedicle screws may generate greater curve correction. Overcorrection, or excessive straightening, of the thoracic curve must be avoided. This point requires a great deal of emphasis. A straighter spine immediately after surgery is not a better result if it leads to postoperative decompensation. When hookrod constructs are used, the more successful outcomes with regard to spinal balance are found in patients whose mean thoracic correction at 2-year follow-up is modest (40%). If rotation is used as the corrective maneuver, the rotation maneuver of the primary rod should be less than a complete 90 . If done this way, the uninstrumented lumbar curve has an excellent chance to compensate satisfactorily [18,22,27, 32,36]. Translational correction or distraction/ compression correction, instead of rotational correction, may reduce the chance of creating postoperative imbalance [41]. With the use of posterior pedicle screw constructs, surgeons may consider trying to achieve an increased amount of thoracic curve correction because greater amounts of spontaneous lumbar correction have been reported. Despite this, the surgeon should be cautious when considering correcting the thoracic curve beyond the amount noted on a preoperative pushprone radiograph, the test most often referred to

as being predictive of the outcome [18]. If a pushprone radiograph has not been obtained, the thoracic curve correction noted on the preoperative supine bend radiograph may give an estimation of desired correction. In the past several years, more experience has been gained with selective anterior thoracic fusion in Lenke 1C curves using instrumentation on the convexity of the thoracic curve. If anterior instrumentation and fusion are considered, the distal level should stop at the end vertebra of the measured thoracic curve. This will allow for a greater number of residual mobile segments below the fusion than can be expected when posterior instrumentation is used. As noted with posterior instrumentation, the desired correction of the instrumented segment should remain limited (approximating the push-prone thoracic curve correction) to avoid spinal imbalance. Some patients who have Lenke 1C patterns may not be candidates for selective thoracic fusions. These patients have large lumbar curves that approximate the thoracic curve size and have signicant rotation, and they may present clinically with a notable convex-side clinical paralumbar prominence. Although the exibility of these lumbar curves (as noted on side-bending radiographs) is sucient to still be classied as Lenke 1C, the lumbar deformity may be too prominent

Fig. 7. (A, B) Radiographs taken before selective thoracic fusion was performed, despite the notable kyphosis at the thoracolumbar junction. (C, D) Six months postoperatively, the junctional kyphosis was more noticeable radiographically.

LENKE 1C, KING TYPE II CURVES

519

to be left unaddressed. Although patients who have larger lumbar curves (O50 ) have been treated successfully with selective fusions, some warrant strong consideration for fusion of thoracic and lumbar curves. Although no absolute lumbar curve magnitude has been identied that always requires lumbar fusion, the surgeon should be cautious about performing selective fusions when a large amount of paraspinal lumbar prominence is noted clinically, the lumbar curve exceeds 50 to 55 , and the lumbar apical vertebra is displaced more than 3 cm away from midline. The sagittal contour at the thoracolumbar junction also must be assessed preoperatively from the radiographic and clinical perspectives [42]. If prominent kyphosis is noted in this location, the lumbar curve may need to be included in the fusion to avoid exacerbating this localized sagittal deformity following surgery (Fig. 7). With Lenkes classication system, if the T10L2 kyphosis measures more than 20 , the thoracolumbar region is considered structural, and the recommendation is made to extend the fusion distal to this level. Some spinal deformity surgeons may elect to carry the fusion further inferiorlydeven with lesser amounts of localized thoracolumbar kyphosisdto limit the chance of a postoperative junctional problem in the sagittal plane [19]. Summary Lenke 1C (King type II) curves present a signicant challenge to the scoliosis surgeon. Selective thoracic fusions preserve motion of the lumbar spine and often lead to a satisfactory outcome, particularly if overcorrection of the thoracic curve is avoided. Still, spinal imbalance following surgery remains the most signicant preoperative concern and requires detailed planning. References
[1] Dobbs MB, Lenke LG, Kim YJ, et al. Selective posterior thoracic fusions for adolescent idiopathic scoliosis: comparison of hooks versus pedicle screws. Spine 2006;31:24004. [2] Lee S-M, Suk S-I, Chung E-R. Direct vertebral rotation: a new technique of three-dimensional deformity correction with segmental pedicle screw xation in adolescent idiopathic scoliosis. Spine 2004;29:3439. [3] Lenke LG. Debate: Resolved, a 55 degrees right thoracic adolescent idiopathic scoliotic curve should be treated by posterior spinal fusion and segmental instrumentation using thoracic pedicle screws: Pro:

[4]

[5]

[6]

[7]

[8]

[9]

[10]

[11]

[12]

[13]

[14]

[15]

[16]

[17]

[18]

Thoracic pedicle screws should be used to treat a 55 degrees right thoracic adolescent idiopathic scoliosis. J Pediatr Orthop 2004;24:32934 [discussion: 3381]. Liljenqvist UR, Halm HF, Link TM. Pedicle screw instrumentation of the thoracic spine in idiopathic scoliosis. Spine 1997;22:223945. Potter BK, Kuklo TR, Lenke LG. Radiographic outcomes of anterior spinal fusion versus posterior spinal fusion with thoracic pedicle screws for treatment of Lenke Type I adolescent idiopathic scoliosis curves. Spine 2005;30:185966. Storer SK, Vitale MG, Hyman JE, et al. Correction of adolescent idiopathic scoliosis using thoracic pedicle screw xation versus hook constructs. J Pediatr Orthop 2005;25:4159. Suk SI, Kim WJ, Lee SM, et al. Thoracic pedicle screw xation in spinal deformities: are they really safe? Spine 2001;26:204957. Suk SI, Lee CK, Kim WJ, et al. Segmental pedicle screw xation in the treatment of thoracic idiopathic scoliosis. Spine 1995;20:1399405. Suk SI, Lee SM, Chung ER, et al. Selective thoracic fusion with segmental pedicle screw xation in the treatment of thoracic idiopathic scoliosis: more than 5-year follow-up. Spine 2005;30:16029. Newton PO. Thoracoscopic treatment of pediatric spinal deformity. Seminars in Spine Surgery 2003; 15:24451. Newton PO, Parent S, Marks M, et al. Prospective evaluation of 50 consecutive scoliosis patients surgically treated with thoracoscopic anterior instrumentation. Spine 2005;30:S1009. Sucato DJ. Thoracoscopic anterior instrumentation and fusion for idiopathic scoliosis. J Am Acad Orthop Surg 2003;11:2217. Wong HK, Hee HT, Yu Z, et al. Results of thoracoscopic instrumented fusion versus conventional posterior instrumented fusion in adolescent idiopathic scoliosis undergoing selective thoracic fusion. Spine 2004;29:20318 [discussion: 2039]. King HA, Moe JH, Bradford DS, et al. The selection of fusion levels in thoracic idiopathic scoliosis. J Bone Joint Surg Am 1983;65:130213. Lenke LG, Betz RR, Harms J, et al. A new and reliable 3-dimensional classication system of adolescent idiopathic scoliosis. St. Louis (MO): Scoliosis Research Society; 1997. Lenke LG, Betz RR, Harms J, et al. Adolescent idiopathic scoliosis: a new classication to determine extent of spinal arthrodesis. J Bone Joint Surg Am 2001;83:116981. Kuklo TR, OBrien MF, Lenke LG, et al. Comparison of the lowest instrumented, stable, and lower end vertebrae in single overhang thoracic adolescent idiopathic scoliosis: anterior versus posterior spinal fusion. Spine 2006;31:22326. Dobbs MB, Lenke LG, Walton T, et al. Can we predict the ultimate lumbar curve in adolescent idiopathic scoliosis patients undergoing a selective

520

RICHARDS

[19]

[20]

[21]

[22]

[23]

[24]

[25]

[26]

[27]

[28]

[29]

[30]

fusion with undercorrection of the thoracic curve? Spine 2004;29:27785. Newton PO, Faro FD, Lenke LG, et al. Factors involved in the decision to perform a selective versus nonselective fusion of Lenke 1B and 1C (King-Moe II) curves in adolescent idiopathic scoliosis. Spine 2003; 28:S21723. Edwards CC 2nd, Lenke LG, Peelle M, et al. Selective thoracic fusion for adolescent idiopathic scoliosis with C modier lumbar curves: 2- to 16-year radiographic and clinical results. Spine 2004;29:53646. Lenke LG, Betz RR, Haher TR, et al. Multisurgeon assessment of surgical decision-making in adolescent idiopathic scoliosis: curve classication, operative approach, and fusion levels. Spine 2001;26:234753. McCance SE, Denis F, Lonstein JE, et al. Coronal and sagittal balance in surgically treated adolescent idiopathic scoliosis with the King II curve pattern. A review of 67 consecutive cases having selective thoracic arthrodesis. Spine 1998;23:206373. McCance SE, Winter RB, Lonstein JE. A King type II curve pattern treated with selective thoracic fusion: case report with 44-year follow-up. J Spinal Disord 1999;12:2625. Winter RB, Lonstein JE. A meta-analysis of the literature on the issue of selective thoracic fusion for the King-Moe type II curve pattern in adolescent idiopathic scoliosis. Spine 2003;28:94852. Arlet V, Marchesi D, Papin P, et al. Decompensation following scoliosis surgery: treatment by decreasing the correction of the main thoracic curve or letting the spine go. Eur Spine J 2000;9:15660. Bassett GS, Hensinger MC, Keiper MD. Eect of posterior spinal fusion on spinal balance in idiopathic scoliosis. J Pediatr Orthop 1989;9:6724. Benli IT, Tuzuner M, Akalin S, et al. Spinal imbalance and decompensation problems in patients treated with Cotrel-Dubousset instrumentation. Eur Spine J 1996;5:3806. Bridwell KH, McAllister JW, Betz RR, et al. Coronal decompensation produced by Cotrel-Dubousset derotation maneuver for idiopathic right thoracic scoliosis. Spine 1991;16:76977. Lenke LG, Bridwell KH, Baldus C, et al. Preventing decompensation in King type II curves treated with Cotrel-Dubousset instrumentation. Strict guidelines for selective thoracic fusion. Spine 1992;17:S27481. Mason DE, Carango P. Spinal decompensation in Cotrel-Dubousset instrumentation. Spine 1991;16: S394403.

[31] McCall RE, Bronson W. Criteria for selective fusion in idiopathic scoliosis using Cotrel-Dubousset instrumentation. J Pediatr Orthop 1992;12:4759. [32] Moore MR, Baynham GC, Brown CW, et al. Analysis of factors related to truncal decompensation following Cotrel-Dubousset instrumentation. J Spinal Disord 1991;4:18892. [33] Patwardhan AG, Rimkus A, Gavin TM, et al. Geometric analysis of coronal decompensation in idiopathic scoliosis. Spine 1996;21:1192200. [34] Richards BS. Lumbar curve response in type II idiopathic scoliosis after posterior instrumentation of the thoracic curve. Spine 1992;17:S2826. [35] Richards BS, Birch JG, Herring JA, et al. Frontal plane and sagittal plane balance following CotrelDubousset instrumentation for idiopathic scoliosis. Spine 1989;14:7337. [36] Thompson JP, Transfeldt EE, Bradford DS, et al. Decompensation after Cotrel-Dubousset instrumentation of idiopathic scoliosis. Spine 1990;15: 92731. [37] Schwender JD, Denis F. Coronal plane imbalance in adolescent idiopathic scoliosis with left lumbar curves exceeding 40 degrees: the role of the lumbosacral hemicurve. Spine 2000;25:235863. [38] Lenke LG, Edwards CC 2nd, Bridwell KH. The Lenke classication of adolescent idiopathic scoliosis: how it organizes curve patterns as a template to perform selective fusions of the spine. Spine 2003; 28:S199207. [39] Betz RR, Harms J, Clements DH 3rd, et al. Comparison of anterior and posterior instrumentation for correction of adolescent thoracic idiopathic scoliosis. Spine 1999;24:22539. [40] Lenke LG, Betz RR, Bridwell KH, et al. Spontaneous lumbar curve coronal correction after selective anterior or posterior thoracic fusion in adolescent idiopathic scoliosis. Spine 1999;24:166371 [discussion: 1672]. [41] Goshi K, Boachie-Adjei O, Moore C, et al. Thoracic scoliosis fusion in adolescent and adult idiopathic scoliosis using posterior translational corrective techniques (Isola): is maximum correction of the thoracic curve detrimental to the unfused lumbar curve? Spine J 2004;4:192201. [42] Lowe TG, Lenke L, Betz R, et al. Distal junctional kyphosis of adolescent idiopathic thoracic curves following anterior or posterior instrumented fusion: incidence, risk factors, and prevention. Spine 2006; 31:299302.

Orthop Clin N Am 38 (2007) 521529

Classication of Operative Adolescent Idiopathic Scoliosis: Treatment Guidelines


Peter S. Rose, MD, Lawrence G. Lenke, MD*
Washington University Orthopedics, Washington University School of Medicine, Washington University Medical Center, 660 South Euclid Avenue, Campus Box 8233, Suite 11300 West Pavilion, St. Louis, MO 63110-1093, USA

Idiopathic scoliosis may present in infant, juvenile, or adolescent patients. Infantile idiopathic scoliosis presents in children age 0 to 3 years and accounts for approximately 1% of idiopathic curves. Juvenile idiopathic curves present at between 4 and 9 years of age and account for 12% to 31% of curves [1]. Adolescent idiopathic scoliosis (AIS) constitutes most remaining cases, presenting in patients between age 10 years and adulthood. This article focuses on the classication and treatment of patients who have AIS. Idiopathic scoliosis is a diagnosis of exclusion made after a patient has been evaluated for underlying congenital, syndromic, or neurologic causes. By denition, a coronal curve of 10 or more must be present to diagnose scoliosis, with lesser curves being termed spinal asymmetry to avoid overdiagnosing patients who have minimal deformities. Approximately 80% of curves ultimately are diagnosed as idiopathic [1]. The prevalence of AIS is estimated at approximately 2% to 3%; however, the prevalence decreases with increasing curve severity. For example, 20 curves have a prevalence of approximately 0.5%, whereas 40 curves have a prevalence of only 0.1%. AIS is more common in girls than in boys, with a ratio of 9:1 in curves larger than 40 . Treatment of AIS may be observational, bracing, or surgical. Observation may be indicated in curves less than 25 in skeletally immature

patients and for curves up to 45 to 50 in skeletally mature patients. Bracing is considered in skeletally immature patients with curves between 25 and 45 to 50 . Surgical treatment is considered in patients with progressive curves greater than 40 that fail or cannot tolerate bracing and those with curves greater than 45 at skeletal maturity. Classication Researchers have sought to classify AIS curves to allow comparison of outcome and to provide treatment guidelines. The King-Moe classication, published in 1983 after an evaluation of more than 400 patients who had AIS, was the rst broadly applied classication of patients in the era of instrumented spinal fusions [2]. Although limited in that it only analyzed thoracic curves and only addressed coronal plane deformity, this work led to the development of the rst AIS treatment guidelines. This classication system suffered from only fair intra- and interobserver reliability, but provided the gold standard for curve classication for 2 decades [3,4]. In 2001, Lenke and colleagues [5] provided a more comprehensive AIS classication system. This system provided for analysis of all AIS curves rather than just thoracic curves, set forth objective criteria for determining curve exibility, and addressed coronal and sagittal plane deformities. The Lenke classication system consists of a three-tiered analysis of curves based on curve type, lumbar modier, and sagittal modier. Evaluation requires upright long-cassette coronal and lateral radiographs as well as side-bending
orthopedic.theclinics.com

* Corresponding author. E-mail address: lenkel@wudosis.wustl.edu (L.G. Lenke).

0030-5898/07/$ - see front matter 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.ocl.2007.06.001

522

ROSE & LENKE

radiographs. Once a curve is classied, this system provides a template for the selection of fusion levels by curve type. Curve classication The rst step of classication is to identify the curve type (Table 1). The proximal thoracic (PT), main thoracic (MT), and thoracolumbar/lumbar (TL/L) curves are measured using the Cobb technique. The larger of the latter two is considered the major curve. Lesser curves are termed minor and are considered structural if they remain at least 25 on side-bending lms. Additionally, hyperkyphosis in the PT region (T2T5R20 ) or thoracolumbar junction (T10L2R20 ) renders these curves structural, regardless of their sidebending measurement. PT curves are always minor, but they may be structural or nonstructural depending on their bending characteristics and local kyphosis. Second, the lumbar spine modier is determined by the position of the apex of the lumbar curve with respect to the center sacral vertical line (CSVL) (Fig. 1). The CSVL is the vertical line from the geometric center of S1, which determines the coronal position of the lumbar spine in relation to the pelvis. For greatest accuracy, a shoe lift should be used if there is more than 2 cm of pelvic obliquity. The stable vertebra is the most cephalad vertebra below the major curve whose center is most closely bisected by the CSVL. If the CSVL falls between the pedicles of the vertebra up to the stable vertebra, the curve has a lumbar modier A. If the CSVL touches the pedicle of the apical vertebra, lumbar modier B is used. Finally, if the CSVL falls outside of the lateral aspect of the lumbar apical vertebral body (or bodies if the apex is a disc), lumbar modier C is used. When in doubt as to the correct lumbar modier, modier B is used. Third, the sagittal thoracic modier is selected as , N, or , based on the T5T12
Table 1 Description of 6 Lenke Curve Types Curve type 1 2 3 4 5 6 Proximal thoracic Nonstructural Structural Nonstructural Structural Nonstructural Nonstructural Main thoracic Structural Structural Structural Structural Nonstructural Structural

kyphosis. If sagittal kyphosis is more than 10 , the or hypokyphotic modier is used; for 10 to 40 , the N or normal modier is used; and for more than 40 of kyphosis, the or hyperkyphotic modier is used. In total, the six curve types, three lumbar modiers, and three sagittal modiers dene 42 curve patterns (one might expect 54, but because all operative TL/L curves are completely deviated from the midline and have type C lumbar modiers, only 42 result). A curve is specied as type lumbar modier sagittal modier, such as 1AN, 5CN, and so forth. The type 1 (single thoracic) curve is the most common (see Table 1). A recent study found that ve classications account for 60% of surgically treated curves (1AN 19%, 1BN 11%, 2AN 10%, 5CN 10%, and 1CN 8%). Overall, lumbar modier A was the most common (41%) followed by B (37%) and C (22%). Normal kyphosis was seen in 75% of curves, with 14% hypokyphotic and 11% hyperkyphotic [6]. This classication system has been evaluated by the developers, an independent group of surgeons from the Scoliosis Research Society, and an international group of surgeons. On premeasured radiographs, this system had better intra- and interobserver reliability than the KingMoe system [7,8]. Both systems have decreased reliability on unmeasured radiographs, consistent with the known variability of Cobb angle level selection and measurement [3]. Treatment guidelines The Lenke classication system is designed to allow surgeons to plan operative treatment based on this curve classication. The general principle of the system is that the major curve and structural minor curves should be fused and that nonstructural minor curves may be allowed to correct spontaneously following fusion of the structural curves.

Thoracolumbar/lumbar Nonstructural Nonstructural Structural Structural Structural Structural

Description Main thoracic Double thoracic Double major Triple major Thoracolumbar/lumbar Thoracolumbar/lumbar main thoracic

Prevalence 51% 20% 11% 3% 12% 3%

OPERATIVE ADOLESCENT IDIOPATHIC SCOLIOSIS

523

Fig. 1. (A) Lenke classication system. (B) Lumbar spine modier rules. (From Lenke LG, Betz RR, Harms J, et al. Adolescent idiopathic scoliosis: a new classication to determine extent of spinal arthrodesis. J Bone Joint Surg Am 2001;83-A(8):116981; with permission.)

524

ROSE & LENKE

In preparation for selection of fusion levels, a careful assessment of the patients medical status, activity level, and clinical appearance should be made. It is the authors practice to obtain pulmonary function tests preoperatively on all patients. Known cardiopulmonary compromise may dictate that a posterior fusion is preferred over an anterior approach to avoid pulmonary insult of a thoracotomy. Similarly, the clinical appearance of shoulder height and symmetry is important in the selection of fusion levels. Simple correction of a thoracic curve in a patient with level shoulders before surgery often results in postoperative shoulder imbalance. Careful preoperative consideration will lead to the selection of a more cephalad fusion level to allow better control over postoperative shoulder balance. Type 1: main thoracic curves The general guideline for these commonly encountered curves is to perform a selective anterior or posterior fusion of the MT curve, regardless of the lumbar spine or sagittal thoracic modier. Segmental spinal instrumentation allows excellent correction of the MT curve and the ultimate lumbar curve response after may be predicted by the preoperative push-prone radiograph [9]. In the absence of shoulder height considerations, the nonstructural proximal curve need not be included in the fusion [10]. Instrumented posterior spinal fusion from the neutral vertebra above to the stable vertebra below (or stable minus one or two levels as dened below) is the traditional treatment of these curves. Posterior segmental spinal instrumentation and fusion are indicated in type 1 curves with a normal or hyperkyphotic sagittal prole and in patients with diminished pulmonary function (in whom this approach avoids violation of the chest cavity). Pedicle screw xation allows for true three-dimensional deformity correction through a vertebral derotation maneuver [11]. For the patient with right shoulder elevation, proximal fusion to T4 or T5 usually allows leveling of the shoulders in a right thoracic curve. Strong consideration should be given to extending the upper instrumented vertebra (UIV) to T4 or even T3 if the shoulders are level before surgery to prevent a high left shoulder postoperatively. If the left shoulder is high before surgery, the UIV likely needs to be T2 to maintain adequate control of shoulder alignment.

In selecting the lower instrumented vertebra (LIV), the stable vertebra is dened as the most proximal vertebra whose pedicles are bisected by the CSVL. The stable minus one vertebra is one level proximal to this and generally is the vertebra touched by the CSVL in the lateral body/pedicle region. Alternatively, the lower end vertebra (LEV) of the MT curve is another landmark vertebra, with the LIV typically one level caudal to that. Most type 1 MT curves can be fused to the stable minus one vertebra. Suk and colleagues [12] analyzed the fusion level in 42 patients with single thoracic curves followed for a minimum of 2 years. In this study, uniformly good results were obtained when patients were fused to the neutral vertebra when there was no more than a two-level dierence between the neutral vertebra and the end vertebra. In general, this corresponded to the stable minus one level. In patients with more than a two-level dierence between the end vertebra and the neutral vertebra, success was predictable when fusions were extended to the neutral minus one level (generally consistent with the stable minus two level). When shorter fusions were attempted, there was a high likelihood of lumbar decompensation. Consideration to limit the fusion to the stable minus two vertebra should be taken in these curves based on assessment of curve exibility, intraoperative correction, and residual angulation and parallelism of the LIV disc. Anterior spinal fusion may be selected in type 1 curves, particularly when there is hypokyphosis or a C lumbar modier. Anterior fusion may be preferred in a patient who is skeletally immature to minimize the risk for crankshaft phenomenon from subsequent growth. Because anterior fusion tends to increase kyphosis, whereas posterior fusion tends to decrease it, anterior fusion may not be suitable in patients with preoperative highnormal or hyperkyphosis [13]. In general, the results of appropriately selected anterior fusion are equivalent to those of posterior fusion [14]. Retrospective data showed that an anterior approach to 1A/B curves can result in a mean shorter fusion of 1.5 vertebral levels [15]. The spontaneous correction of the PT curve may be better following anterior fusion [16]; however, the use of segmental thoracic pedicle screw xation increased instrumented curve correction, improved spontaneous lumbar compensatory curve correction, and improved thoracic torsion and rotation [17].

OPERATIVE ADOLESCENT IDIOPATHIC SCOLIOSIS

525

Particular consideration should be paid to the treatment of the type 1C curve (Fig. 2). Selective thoracic fusion may be considered for these curves because the lumbar curve, although large and deviated, generally is exible with good potential for spontaneous correction. When contemplating whether a 1C curve is a candidate for a selective thoracic fusion, clinical and radiographic analysis is required. The thoracolumbar junction should not be hyperkyphotic, the left shoulder should not be high, and the thoracic

prominence and shift should dominate over the lumbar deformity. Comparison can be made to the ratio of the Cobb measurement, scoliometer measurement, and apical translation and rotation of the MT and TL/L regions; ratios greater than 1.25 predict successful thoracic fusion in these patients [1821]. A combined anterior and posterior approach to a type 1 curve is rarely required in the era of posterior segmental spinal instrumentation [22]. Relative indications include severe, sti curves

Fig. 2. 5C curve treated with selective anterior fusion. (AG) Preoperative radiographs and clinical photographs. (A) Anteroposterior (AP). (B) Left side bending. (C) Right side bending. (D) Lateral. (E) Close posteroanterior (PA). (F) Forward bend lumbar. (G) Forward bend thoracic. (HK) Postoperative radiographs and clinical photographs. (H) AP. (I) Lateral. (J) Close PA. (K) Forward bend.

526

ROSE & LENKE

Fig. 2 (continued)

(O90 100 upright, O60 70 residual curve on bending lms), severe lordosis or hyperkyphosis, and highly skeletally immature patients. Type 2: double thoracic curves These curves are distinguished by structural PT and MT regions; they are the second most commonly encountered operative AIS curves. Treatment is posterior fusion of both curves. Determining the UIV requires careful assessment of shoulder height before surgery. Increasing correction of the right MT curve tends to elevate the left shoulder. Patients with a high left shoulder require instrumentation and fusion extending proximally to T2. In patients with level shoulders, the UIV should be T2 or T3, depending on the correction and resulting shoulder balance achieved. In patients with a high right shoulder, T3 usually can be used as the UIV. During correction, compression is placed on the convex side of the PT curve and distraction on the concavity. Because the PT curve usually is hyperkyphotic, convex compression is applied before concave distraction to avoid worsening the local kyphosis. Suk and colleagues [23] reported results in accordance with these recommendations in analyzing 40 patients with PT curves greater than 25 ; they found improved results with fusion of the proximal curve whenever the shoulders were level or higher on the side of the proximal thoracic

curve. In preoperative evaluation, the clavicle angle provides the best radiographic assessment of shoulder balance [24]. The LIV is selected similarly to the selection rules of type 1 curves. Generally, the most proximal lumbar vertebra intersected by the CSVL may serve as the LIV. Combined anterior and posterior fusions are rarely needed in the treatment of double thoracic curves. Type 3: double major curves Type 3 curves include a major MT curve and a structural TL/L curve. Type 3 curves are the third most common (11%); they present most often with a C lumbar modier. Treatment generally requires posterior instrumentation and fusion of both curves. Hyperkyphosis of the thoracolumbar junction is present frequently (T10L2O20 ) and should be normalized intraoperatively. The UIV selected may be T35, depending on the size and stiness of the nonstructural PT curve and any shoulder asymmetry. The LIV is usually L3 or L4 with the level determined by the most proximal lumbar level intersected by the CSVL. The surgical goal should be horizontalization, centralization, and neutralization of the level chosen as the LIV; critical attention should be paid to making the LIV disc horizontal. Some 3C curves can be treated with selective thoracic fusion. Prerequisites to this approach are

OPERATIVE ADOLESCENT IDIOPATHIC SCOLIOSIS

527

lack of junctional kyphosis and a clinical deformity that predominates in the thoracic spine. Additionally, a rare severe and sti thoracic or lumbar curve may require anterior release or circumferential fusion. Type 4: triple major curves Type 4 curves are rare, accounting for approximately 3% of all operative curves. The PT, MT,

and TL/L curves are all structural; the MT or TL/L may be the major curve. Treatment is by posterior segmental instrumentation and fusion from the upper thoracic spine to L3 or L4. The UIV generally is T2 or T3, with the level determined in the same manner as the UIV for type 2 double thoracic curves. The LIV should achieve horizontalization, centralization, and stabilization using the same criteria as the LIV in type 3 double major curves. In the rare case of a large and sti

Fig. 3. (AF) 1CN curve treated with fusion of major curve only. Preoperative radiographs and clinical photographs. (A) Anteroposterior (AP). (B) Left side bending. (C) Right side bending. (D) Lateral. (E) Close posteroanterior (PA). (F) Forward bend. (GJ) Postoperative radiographs and clinical photographs. (G) AP. (H) Lateral. (I) Close PA. (J) Forward bend.

528

ROSE & LENKE

MT or TL/L curve, anterior release and fusion may be needed. Optimal correction is achieved by pedicle screw xation of all three curves to allow secure xation for correction and realignment. Type 5: thoracolumbar/lumbar curves Type 5 curves have only the TL/L region as a structural curve, which may be fused anteriorly or posteriorly (Fig. 3). When fused anteriorly, treatment is from the upper end vertebra (UEV) to the LEV (to include all levels with convex discs). With transpedicular xation, these curves can be treated posteriorly over the same levels or to one additional caudal level. In cases in which some residual MT or PT curve is expected to remain, the UIV may be left deliberately tilted to correspond to the amount of residual MT curve expected. When fusing only the TL/L region of the curve, secondary correction of the compensatory thoracic curve tends to elevate the left shoulder. Additionally, there is little change in the rib deformity with limited TL/L fusions, so the clinical appearance of the thoracic region before surgery must be acceptable to the patient. Sanders and colleagues [25] reported on 49 patients who underwent selective anterior fusion of thoracolumbar curves with a minimum 2-year follow-up. Successful selective fusion was dependent on patient maturity and curve characteristics. Forty-two of 43 patients with closed triradiate cartilages had a successful outcome following selective curve fusion. Similarly, 42 of 44 patients who had a thoracolumbar/thoracic Cobb angle ratio of at least 1.25 or a thoracic curve that bent up to 20 achieved a good outcome with selective fusion of the thoracolumbar curve. Type 6: thoracolumbar/lumbarmain thoracic curves These curves combine a TL/L curve with a structural minor, MT curve. General treatment recommendations are to treat the MT and TL/L curves posteriorly. The UIV varies from T35 with the MT curve location and shoulder height. The LIV is the most proximal lumbar vertebra touched by the CSVL, usually L3 or L4. A selective anterior fusion of only the TL/L curve can be considered (similar to a type 5 curve). Such treatment presupposes that the residual thoracic curve has acceptable radiographic appearance and that thoracic and shoulder clinical asymmetry are acceptable. Patients should have closed triradiate cartilages to consider such an approach [25].

Outcomes The Lenke classication is designed as a tool to guide the surgical treatment of AIS curves. Studies have found good intra- and interobserver reliability of the system, particularly when the errors of Cobb measurement are eliminated [3,4,26]. A retrospective analysis of 606 surgical curves treated before the development of this classication was performed to assess how well this new system would predict treatment. Overall agreement between the classication system and surgical treatment was 90%, with individual curve types varying between 75% and 95%. Similarly, Puno and colleagues [27] retrospectively applied the Lenke classication treatment recommendations to 183 patients with a greater than 2-year followup. Better correctiondin avoiding undercorrection of the structural secondary curves and in avoiding unnecessary fusion of nonstructural curvesdwas seen in patients treated in accordance with the Lenke criteria recommendations. Disagreement remains in the selection of fusion levels. In a roundtable discussion of 28 scoliosis surgeons, exact treatment levels varied widely, even when agreement on curve types was reached [26]. Similarly, in analyzing data of 203 patients from the Harms Study Group, Newton and colleagues [28] reported considerable variations of selective fusion of Lenke 1B and 1C curves across the ve study sites.

Summary The Lenke classication system provides a comprehensive and biplanar system to evaluate patients who have AIS. The system has improved intra- and interobserver reliability compared with the King-Moe system and has the ability to direct surgical treatment for successful management. Curve exibility and the potential for selective fusions may be assessed objectively. Additionally, the system provides for grouping of similar curve types to facilitate research into various treatment methods. Reports in the literature suggest that the classication guidelines provide appropriate guidance for the selection of fusion levels.

References
[1] Newton PO, Wenger DR. Idiopathic scoliosis. In: Morrissy RT, Weinstein SL, editors. Lovell and Winters pediatric orthopaedics. 6th edition.

OPERATIVE ADOLESCENT IDIOPATHIC SCOLIOSIS

529

[2]

[3]

[4]

[5]

[6]

[7]

[8]

[9]

[10]

[11]

[12]

[13]

[14]

[15]

Baltimore (MD): Lippincott Williams & Wilkins; 2006. p. 693762. King HA, Moe JH, Bradford DS, et al. The selection of fusion levels in thoracic idiopathic scoliosis. J Bone Joint Surg Am 1983;65:130213. Niemeyer T, Wolf A, Kluba S, et al. Interobserver and intraobserver agreement of Lenke and King classications for idiopathic scoliosis and the inuence of level of professional training. Spine 2006;31:21037. Richards BS, Sucato DJ, Konigsberg DE, et al. Comparison of reliability between the Lenke and King classication systems for adolescent idiopathic scoliosis using radiographs that were not premeasured. Spine 2003;28:114856. Lenke LG, Betz RR, Harms J, et al. Adolescent idiopathic scoliosis: a new classication system to determine extent of spinal arthrodesis. J Bone Joint Surg Am 2001;83:116981. Lenke LG, Betz RR, Clements D, et al. Curve prevalence of a new classication of operative adolescent idiopathic scoliosis: does classication correlate with treatment? Spine 2002;27:60411. Lenke LG, Betz RR, Bridwell KH, et al. Intraobserver and interobserver reliability of the classication of thoracic adolescent idiopathic scoliosis. J Bone Joint Surg Am 1998;80:1097106. Ogon M, Giesinger K, Behensky H, et al. Interobserver and intraobserver reliability of Lenkes new scoliosis classication system. Spine 2002;27:85862. Dobbs MB, Lenke LG, Walton T, et al. Can we predict the ultimate lumbar curve in adolescent idiopathic scoliosis patients undergoing a selective fusion with undercorrection of the thoracic curve? Spine 2004;29:27785. Cil A, Pekmezci M, Yazici M, et al. The validity of Lenke criteria for dening structural proximal thoracic curves in patients with adolescent idiopathic scoliosis. Spine 2005;30:25505. Lee SM, Suk SI, Chung ER. Direct vertebral rotation: a new technique of three-dimensional deformity correction with segmental pedicle screw xation in adolescent idiopathic scoliosis. Spine 2004;29:3439. Suk SI, Lee SM, Chung ER, et al. Determination of distal fusion level with segmental pedicle screw xation in single thoracic idiopathic scoliosis. Spine 2003;28:48491. Rhee JM, Bridwell KH, Won DS, et al. Sagittal plane analysis of adolescent idiopathic scoliosis: the eect of anterior versus posterior instrumentation. Spine 2002;27:23506. Betz RR, Harms J, Clements DH, et al. Comparison of anterior and posterior instrumentation for correction of adolescent thoracic idiopathic scoliosis. Spine 1999;24:22539. Kuklo TR, OBrien MF, Lenke LG, et al. Comparison of the lowest instrumented, stable, and lower end vertebrae in single overhang thoracic

[16]

[17]

[18]

[19]

[20]

[21]

[22]

[23]

[24]

[25]

[26]

[27]

[28]

adolescent idiopathic scoliosis: anterior versus posterior spinal fusion. Spine 2006;31:22326. Kuklo TR, Lenke LG, Won DS, et al. Spontaneous proximal thoracic curve correction after isolated fusion of the main thoracic curve in adolescent idiopathic scoliosis. Spine 2001;26:196675. Potter BK, Kuklo TR, Lenke LG. Radiographic outcomes of anterior spinal fusion versus posterior spinal fusion with thoracic pedicle screws for treatment of Lenke Type I adolescent idiopathic scoliosis curves. Spine 2005;30:185966. Lenke LG, Betz RR, Bridwell KH, et al. Spontaneous lumbar curve coronal correction after selective anterior or posterior thoracic fusion in adolescent idiopathic scoliosis. Spine 1999;24:166371. Lenke LG, Bridwell KH, Baldus C, et al. Preventing decompensation in King type II curves treated with Cotrel-Dubousset instrumentation. Strict guidelines for selective thoracic fusion. Spine 1992;17:S27481. Edwards CC II, Lenke LG, Peelle M, et al. Selective thoracic fusion for adolescent idiopathic scoliosis with C modier lumbar curves: 2- to 16-year radiographic and clinical results. Spine 2004;29:53646. Lenke LG, Edwards CC II, Bridwell KH. The Lenke classication of adolescent idiopathic scoliosis: how it organizes curve patterns as a template to perform selective fusions of the spine. Spine 2003;28:S199207. Luhmann SJ, Lenke LG, Kim YJ, et al. Thoracic adolescent idiopathic curves between 70 and 100 . Is anterior release necessary? Spine 2005;30:20617. Suk SI, Kim WJ, Lee CS, et al. Indications of proximal thoracic curve fusion in thoracic adolescent idiopathic scoliosis: Recognition and treatment of double thoracic curve pattern in adolescent idiopathic scoliosis treated with segmental instrumentation. Spine 2000;25:23429. Kuklo TR, Lenke LG, Graham EJ, et al. Correlation of radiographic, clinical, and patient assessment of shoulder balance following fusion versus nonfusion of the proximal thoracic curve in adolescent idiopathic scoliosis. Spine 2002;27:201320. Sanders AE, Baumann R, Brown H, et al. Selective anterior fusion of thoracolumbar/lumbar curves in adolescents: when can the associated thoracic curve be left unfused? Spine 2003;28:70613. Lenke LG, Betz RR, Haher TR, et al. Multisurgeon assessment of surgical decision-making in adolescent idiopathic scoliosis: curve classication, operative approach, and fusion levels. Spine 2001;26:234753. Puno RM, An KC, Puno RL, et al. Treatment recommendations for idiopathic scoliosis: an assessment of the Lenke classication. Spine 2003;28:210214. Newton PO, Faro FD, Lenke LG, et al. Factors involved in the decision to perform a selective versus nonselective fusion of Lenke 1B and 1C (King-Moe II) curves in adolescent idiopathic scoliosis. Spine 2003;28:S21723.

Orthop Clin N Am 38 (2007) 531540

Anterior and Thoracoscopic Scoliosis Surgery for Idiopathic Scoliosis


Vidyadhar V. Upasani, MDa, Peter O. Newton, MDb,*
Department of Orthopedic Surgery, University of California San Diego, 3020 Childrens Way, MC5054, San Diego, CA 92123, USA b Department of Orthopedic Surgery, Rady Childrens Hospital San Diego, 3030 Childrens Way, Suite 410, San Diego, CA 92123, USA
a

Orthopedic management of idiopathic scoliosis is based on knowledge of the natural history of this spinal disorder, identifying those patients with a high likelihood of developing worsening deformity. Long-term studies of untreated scoliosis have associated severe deformity with poor prognosis, increased morbidity and mortality associated with worsening cardiopulmonary compromise, increased back pain, and psychosocial issues related to the deformity [1,2]. Surgical treatment, with instrumentation and spinal arthrodesis, was shown to achieve long-lasting deformity reduction at the cost of spinal exibility [3,4]. Although posterior spinal instrumentation and fusion continues to be the most commonly used approach in the surgical treatment of idiopathic scoliosis, anterior procedures, both open and thoracoscopic, have proven to be a viable option in selected patients. Anterior scoliosis surgery Open anterior surgical techniques for spinal deformity correction were described rst in the late 1960s. Dwyer and colleagues [5] developed a vertebral screw and exible cable system to achieve coronal deformity correction by applying a compressive force along the convexity of the spine. The Zielke system, developed in the 1970s, replaced the Dwyer cable and used an anteriorthreaded rod to treat short- and long-segment

deformities [6]. Since that time, open anterior surgical techniques and instrumentation have advanced considerably. Current implant systems use rigid, single- or dual-rod constructs to achieve spinal realignment and stabilization. During the early 1990s, anterior video-assisted thoracoscopy was described by Regan and Mack and colleagues [7,8]. This technique provides an alternate approach, through a limited chest wall dissection, to achieve multilevel thoracic spinal release and instrumentation. Once mastered, this technically challenging approach theoretically aords several advantages over the traditional thoracotomy, including reduced pulmonary morbidity and postoperative pain, better visualization, and improved cosmesis. The ultimate goals of all anterior spinal deformity correction procedures remain the same. The primary concern is to visualize the spine and obtain exposure that will allow spinal release and instrumentation without damaging the surrounding neurovascular structures. Secondarily, a thorough disc and ligamentous excision must be performed to mobilize the spine maximally. Lastly, a solid interbody arthrodesis must be achieved after instrumentation and correction of the spinal deformity. In general, surgical fusions should be as short as possible to minimize the loss of spinal exibility, yet be long enough to ensure optimal correction and lasting spinal balance. Indications

* Corresponding author. E-mail address: pnewton@chsd.org (P.O. Newton).

Surgical indications for patients who have adolescent idiopathic scoliosis are dependent on
orthopedic.theclinics.com

0030-5898/07/$ - see front matter 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.ocl.2007.05.003

532

UPASANI & NEWTON

various factors. The risk for deformity progression based on gender, bone age, and curve magnitude plays a critical role when deciding which patients need operative treatment [911]. Studies of the natural history of this disease indicate that surgical treatment should be considered in immature patients once the Cobb angle of their major curve exceeds 40 to 45 , whereas more mature patientsdwith a lower risk for curve progressiondmay be observed until their major curve exceeds 50 [1,2,12]. Curve pattern, trunk deformity (axial plane rotation), and balance also should be considered when recommending surgical correction, because a single curve may create greater trunk shift than a balanced double or triple curve. Combined anterior and posterior scoliosis procedures are primarily indicated in the treatment of large (O75 ) or rigid (bend correction !50 ) spinal deformities. The anterior approaches (open or thoracoscopic) enable maximal spinal mobilization before instrumentation, by allowing access to the main anterior stabilizing structures of the spine, including the annulus, the intervertebral disc, and the anterior longitudinal ligament. The degree to which spinal exibility can be increased is dependent on the complete release of these stabilizing structures. In the most severe cases, resection of the rib head and costovertebral joint may be required to optimize spinal mobility. Another indication for these combined approaches is in the treatment of immature (Risser 0 with an open triradiate cartilage) patients. Unbalanced ablation of the posterior growth potential, with continued anterior growth, is believed to cause a slowly progressive rotational (crankshaft) deformity [13,14]. To prevent this problem, patients are treated often with an anterior release and fusion and a posterior instrumentation and fusion. Combined procedures also are indicated in patients who are at increased risk for pseudoarthrosis formation, such as those with bone-healing deciency syndromes (eg, neurobromatosis, Marfan syndrome) or a history of irradiation. An anterior discectomy provides a large surface area of cancellous bone and allows for a circumferential fusion to increase the likelihood of forming a solid arthrodesis. Anterior instrumentation has been associated with the potential to save one to three distal fusion levels in the treatment of isolated major thoracic, thoracolumbar, or lumbar curves [15]. This procedure also has been found to be kyphogenic and

ideally suited in the treatment of patients who have a hypokyphotic or lordotic thoracic spine [16]; however, in the treatment of thoracolumbar/ lumbar curves, structural grafts are required to maintain normal thoracolumbar sagittal alignment. Double-rod, double-screw constructs also may allow additional control of the sagittal plane in patients who have thoracolumbar curves [17]. Anterior thoracoscopic techniques are more amenable to the treatment of curves with certain characteristics. For example, smaller curves (usually !70 ) with greater than 50% exibility can be treated appropriately with a single rodscrew construct. Single structural thoracic curves and those double or triple curves in which only the thoracic component is structural also are treated more readily with a single-rod anterior thoracoscopic procedure.

Contraindications Impaired preoperative pulmonary function and the presence of comorbidities associated with intrathoracic or intra-abdominal visceral abnormalities are two of the absolute contraindications to anterior scoliosis procedures. The pulmonary status of the patient must allow single-lung ventilation; transthoracic and thoracoscopic approaches require selective deation of one lung to allow adequate space within the chest cavity to expose and instrument the spine. Vertebral body osteopenia, although rare in the adolescent idiopathic scoliosis population, is seen commonly in patients who have neuromuscular disorders and may limit anterior instrumentation options. Vertebral body size is another consideration; it may limit adequate xation in small or underweight patients. Obesity (O6070 kg) may be another relative contraindication to anterior thoracic instrumentation, because patients with an increased body mass may be more prone to overstressing a singlerod anterior construct, resulting in implant failure or loss of deformity correction. Dual-rod anterior constructs may be more appropriate in obese patients; they were shown to be more sti biomechanically [18] and clinically equivalent [17] to single-rod instrumentation in achieving deformity correction. Specic to anterior thoracoscopic surgery, the existence of intrathoracic pleural adhesions from prior thoracotomy procedures or pulmonary infections should be considered a relative contraindication. Although minor adhesions can be

ANTERIOR SURGERY FOR IDIOPATHIC SCOLIOSIS

533

divided, a near complete pleural symphysis between the chest and lung can make adequate lung collapse nearly impossible. In addition, although children weighing less than 30 kg have been treated safely with the anterior thoracoscopic approach, the relative benet of this minimally invasive technique seems to be reduced in small patients [19]. If visualization is inadequate at any point during the endoscopic procedure, conversion to an open approach must be considered. A rigid spinal deformity or one that is too closely approximated to the rib cage also is dicult to treat with a thoracoscopic anterior procedure. Preoperative radiographs should be reviewed to ensure a minimum working distance of 2 to 3 cm for thoracoscopic procedures. Open anterior release and instrumentation The thoracic spine is accessed most commonly through a single or double anterolateral thoracotomy in patients who have idiopathic scoliosis. Usually, a single thoracotomy is adequate to access seven or fewer levels between T4 through T12, whereas a double thoracotomy may be required to achieve exposure when more than seven levels are going to be fused. In idiopathic scoliosis, the spine usually is approached from the convexity of the curve (a right-sided thoracotomy). Typically, the thoracolumbar/lumbar spine (T10L5) is approached through a thoracoabdominal incision. This incision should cross the costochondral junction before turning obliquely across the abdominal wall toward the lateral border of the rectus abdominis sheath. A low thoracotomy, with a tenth-rib resection, can facilitate access to the thoracic cavity, the thoracolumbar spine, and the retroperitoneal space.

After exposure of the spine has been achieved, the discs to be resected, and, if indicated, the vertebrae to be instrumented are veried using intraoperative uoroscopy or portable radiograph. A thorough discectomy is performed at each subsequent level, with excision of the annulus and the anterior longitudinal ligament. Incomplete disc resection has been associated with suboptimal mobilization of the spine and an increased incidence of implant failure and pseudoarthrosis formation. At times, visualization of the posterior disc may require resection of the rib head down to the base of the transverse process. The cartilaginous superior and inferior endplates also must be separated completely from the adjoining vertebral bodies, and the bony endplates should be decorticated with sharp curettes. Fixed-head vertebral body screws, ranging from 5 to 7 mm in diameter, are used commonly in patients who have idiopathic scoliosis. Prior to screw placement, the superior and inferior endplates, the anterior cortex, and the anterior aspect of the spinal canal should be understood clearly for each vertebral body. Biomechanically, pronged staples and bicortical screw xation signicantly increased construct stiness during single-screw anterior vertebral body instrumentation [20]. In addition, juxta-endplate screws provided better xation than did screws placed in the traditional midvertebral location [21]. This screw position seems to increase xation strength by butting the screw threads up against the superior (or inferior) endplate of the vertebral body. For dual-rod constructs, a two-hole vertebral body staple should be placed carefully to allow appropriate positioning and trajectory of both screws (Fig. 1A). All screw tips should be palpated on the concavity of the deformity to ensure that

Fig. 1. (A) A two-hole vertebral body staple is used to allow appropriate positioning and trajectory of dual-screw implants. (B) Dual-rod construct after deformity correction and rod implantation has been achieved.

534

UPASANI & NEWTON

they are not excessively long and are not causing injury to adjacent neurovascular or visceral structures. After all screws have been placed, the single or dual rods (Fig. 1B) are prebent to the desired coronal and sagittal contour of the spine, loosely anchored within the screw heads, and then rotated into position to correct scoliosis and to restore the normal sagittal contour of the spine. Beginning distally, the disc spaces are wedged open, and the interspaces are packed with bone graft. Many bone-grafting materials are available: structural and nonstructural grafts, autologous grafts, fresh-frozen and freeze-dried allografts, demineralized bone matrices, and various synthetic bone substitutes. Regardless of the material used, it is important to ensure that both bony endplates are well decorticated and able to provide good vascularity to the graft material. Because the anterior approach tends to be kyphogenic, a structural graft may be necessary in the thoracolumbar and lumbar spine to preserve the natural lordosis. Derotation, translation, and compression maneuvers may be used during rod insertion to achieve deformity correction. Scoliosis correction can be achieved directly by cantilevering the rod into the vertebral screws or by rolling the precontoured rod from scoliosis into the sagittal plane. After tightening the proximal screws, further coronal correction can be obtained by sequentially compressing between screws along the

convexity of the curve. After the instrumentation is complete, the patients neurologic function should be assessed with a wake-up test or spinal cord monitoring. Intraoperative radiographs also should be obtained to conrm screw placement and evaluate the initial deformity correction. Case example 1 A 14-year-old girl who has a progressive scoliotic deformity presented after failed treatment with a brace. Preoperative posteroanterior (PA) (Fig. 2A) and lateral (Fig. 2B) radiographs revealed that she was Risser 3 and had a 43 thoracolumbar curve with a 32 compensatory right thoracic curve. One-year after a left open anterior spinal fusion from T11 to L3 with dual-rod instrumentation, PA (Fig. 2C) and lateral (Fig. 2D) radiographs revealed a 15 thoracolumbar curve and an 11 thoracic curve.

Thoracoscopic anterior release and instrumentation Similar to open anterior procedures, the left lateral decubitus position allows optimal access to the right thoracic spine and enables greater circumferential access to the vertebral bodies and discs during the thoracoscopic approach. Maintaining spatial orientation is more dicult

Fig. 2. A 14-year-old girl who had a progressive scoliotic deformity presented after failed treatment with a brace. Preoperative PA (A) and lateral (B) radiographs revealed that she was Risser 3 and had a 43 thoracolumbar curve with a 32 compensatory right thoracic curve. One year after a left open anterior spinal fusion from T11 to L3 with dualrod instrumentation, PA (C) and lateral (D) radiographs revealed a 15 thoracolumbar curve and an 11 thoracic curve.

ANTERIOR SURGERY FOR IDIOPATHIC SCOLIOSIS

535

Fig. 3. A harmonic scalpel is used to create a longitudinal opening of the pleura.

Fig. 5. The segmental vessels are coagulated and divided with the harmonic scalpel to minimize bleeding while broadening the exposure.

during endoscopic spinal surgery. Positioning the surgeon and the assistant anterior to the patient with the video monitor properly aligned and oriented behind the patient allows access to the spine from the most natural viewing perspective. Single-lung ventilation with a double-lumen endotracheal tube is performed to deate the right lung selectively before port placement. The location of the ports is determined using anatomic landmarks and uoroscopic guidance to optimize access to all motion segments planned to be instrumented. The number of ports required depends on the type of deformity being treated and the number of levels being instrumented. Generally, three portals along the posterior axillary line are used for instrumentation, and two portals along the anterior axillary line are used for exposure and release of the anterior spine. Anterior axillary ports allow greater exposure of the

concave aspects of the deformity during disc excision and retraction of the great vessels. Angled optics should be used to ensure that the tip of the working instrument is visualized at all times. A fan retractor is placed through one of the posterior axillary ports to retract and protect the deated lung. After conrming anatomic levels using intraoperative uoroscopy, a harmonic scalpel is used to create a longitudinal opening of the pleura (Fig. 3). The initial pleural opening is performed supercial to the segmental vessels (Fig. 4). Limited exposure of the discs is accomplished by retracting the pleura between the vessels. To broaden the exposure anteriorly, the segmental vessels are coagulated and divided with the harmonic scalpel (Fig. 5). To achieve optimal hemostasis, the energy from the harmonic scalpel is applied slowly over a 3- to 5-mm length of the vessel.

Fig. 4. The initial pleural opening is performed supercial to the segmental vessels. A segmental artery and vein is seen crossing the midportion of the vertebral body.

Fig. 6. A packing sponge is inserted between the anterior spine and pleura to create a space and protect adjacent neurovascular structures.

536

UPASANI & NEWTON

After dividing the pleura, the loose areolar tissue is divided, and a space is created between the anterior spine and pleura using packing sponges (Fig. 6). Directly lateral to the anterior spine, the azygos vein and other great vessels can be identied and must be avoided during the discectomy procedure (Fig. 7). Circumferential exposure of the spine and discs should be completed before excising the discs. Distal exposure to the T12L1 disc space may require division of the diaphragm insertion. To accomplish this, the longitudinal pleural incision is extended onto the inferiorly retracted diaphragm, and blunt dissection is used to strip the diaphragm o the anterior aspect of the spine. Disc excision is initiated by performing a circumferential annulotomy using the ultrasonic blade. An up-biting rongeur is used rst to remove the most anterior and concave aspect of the annulus. A complete discectomy requires optimal visualization deep into the disc space, ensuring that the integrity of the posterior longitudinal ligament is maintained and that the neural elements are protected. The superior and inferior cartilage endplates are dissected sharply, and an angled curette or rongeur is used to decorticate the bone. An endplate shaver inserted into each intervertebral space may be used to ensure that complete disc excision has been performed and that the spine is mobile (Fig. 8). Hemostasis is maintained by placing an oxidized cellulose agent in each intervertebral space. To prepare for screw/rod implantation, a 15-mm port is placed between the ribs through the skin incisions along the posterior axillary line (Fig. 9). Each screw should be started in the midaspect of the vertebral body just anterior to the rib

Fig. 7. The contralateral segmental vessels and azygos vein can be identied directly under the anterior spine.

head articulation (Fig. 10). An awl is used rst to determine the screw path. Then, the vertebral body is tapped, and a ball-tipped calibrated probe is used to determine screw length. The screws should achieve bicortical purchase; however, excessive screw penetration should be avoided to protect adjacent neurovascular structures (aorta). After insertion of all of the screws, the xed-angle screw heads are aligned for rod insertion, and the intervertebral spaces are packed with bone graft using a tubular plunger device. Deformity correction is accomplished by cantilevering a precontoured rod into position (Fig. 11), and segmental compression is performed using an endoscopic compressing device (Fig. 12). Following rod insertion, the pleura is reapproximated and closed over the instrumentation using the EndoStitch device. Beginning distally, the suture needle is passed through both sides of the cut pleura or diaphragm, and an

Fig. 8. (A) An endplate shaver is inserted into each intervertebral space to ensure that a complete discectomy has been performed and that the spine is optimally mobilized. (B) The posterior longitudinal ligament can be visualized by opening the disc space to ensure that the spinal cord is protected.

ANTERIOR SURGERY FOR IDIOPATHIC SCOLIOSIS

537

Fig. 9. A 15-mm port is placed between the ribs through skin incisions along the posterior axillary line to prepare for instrumentation.

Fig. 11. The precontoured rod is cantilevered into position to obtain deformity correction.

Complications externally tied knot is slid down securely into place. The suturing device allows a double-ended needle to be passed from one jaw to the other, and a simple running closure of the pleura is performed (Fig. 13). Case example 2 A 12-year-old girl who had adolescent idiopathic scoliosis and a progressive deformity presented with preoperative PA (Fig. 14A) and lateral (Fig. 14B) radiographs that revealed she was Risser 0, had a 43 thoracic curve, and had a 28 thoracolumbar curve. One year after an anterior thoracoscopic release, T6 to T12 single-rod instrumentation, and fusion with iliac crest bone graft, PA (Fig. 14C) and lateral (Fig. 14D) radiographs revealed a 19 thoracic curve and a 10 thoracolumbar curve. The incidence of major complications, including death, paraplegia, or deep wound infections after open or thoracoscopic procedures to instrument the anterior thoracic/thoracolumbar spine, are less than 1% [22]. As would be expected, pulmonary complications account for more than 50% of the morbidity associated with these procedures [23,24]. Most of time, these pulmonary issues are related to postoperative pleural eusions, pneumothorax, atelectasis, or excessive chest tube drainage. Preoperative pulmonary function studies can be obtained to prevent or plan for postoperative respiratory issues. An evaluation of pulmonary function after open versus thoracoscopic anterior procedures revealed that pulmonary function recovered more quickly after the less invasive procedure; this dierence was maintained at the 2-year follow-up [25]. Other

Fig. 10. A tap is used to ensure that the vertebral body screw insertion is juxta-endplate to provide for superior xation.

Fig. 12. Segmental compression is performed to obtain coronal plane deformity correction.

538

UPASANI & NEWTON

Fig. 13. The EndoStitch device is used to suture the pleura over the instrumentation.

infrequently reported complications after anterior scoliosis surgery include injury to the great vessels, ureter, or spinal cord and development of a retroperitoneal hematoma or brosis.

Outcomes When evaluating anterior thoracoscopic release procedures, several animal studies reported the ability to achieve similar amounts of spinal mobilization compared with open techniques

[2628]. A radiographic analysis also demonstrated a similar ability to achieve coronal and sagittal plane correction when comparing the two techniques [29]; however, the thoracoscopic procedure is technically challenging, and the steep learning curve described by Newton and colleagues [30] must be overcome to perform this procedure safely and ecaciously. A series of 112 consecutive cases of thoracoscopic anterior release and fusion with more than 2-years of follow-up revealed that clinical failures were exceedingly rare and that the primary goals of increasing spinal exibility and achieving a solid arthrodesis occurred in most cases [31]. Several studies evaluated surgical outcomes in patients who had idiopathic scoliosis by comparing anterior and posterior instrumentation methods. In 1999, Lenke and colleagues [32] reported a greater main thoracic curve and spontaneous thoracolumbar/lumbar curve correction following selective anterior fusion compared with posterior instrumentation. Multiple investigators conrmed these ndings and demonstrated the ability of anterior instrumentation to achieve superior radiographic results with the fusion of fewer vertebral levels [3335]; however, these early studies compared anterior instrumentation techniques with posterior hook or hybrid constructs.

Fig. 14. A 12-year-old girl who had adolescent idiopathic scoliosis and a progressive deformity presented with preoperative PA (A) and lateral (B) radiographs that revealed she was Risser 0, had a 43 thoracic curve, and a 28 thoracolumbar curve. One year after an anterior thoracoscopic release, T6 to T12 single-rod instrumentation, and fusion with iliac crest bone graft, PA (C) and lateral (D) radiographs revealed a 19 thoracic curve and a 10 thoracolumbar curve.

ANTERIOR SURGERY FOR IDIOPATHIC SCOLIOSIS

539

In 2005, Potter and colleagues [36] evaluated posterior instrumentation with thoracic pedicle screws; in some cases, posterior surgery provided better coronal and axial correction compared with thoracic anterior instrumentation. Anterior scoliosis procedures with single-rod constructs also have been associated with an increased rate of pseudoarthrosis and an increased rate of implant failure compared with posterior instrumentation [37]. Dual-rod constructs were developed to provide more rigid xation and were found to increase mechanical stiness [17,18]; however, vertebral body size restrictions make it dicult to place these implants in the adolescent thoracic spine. Two-year follow-up has been reported for thoracoscopic anterior instrumentation. In a series of 50 consecutive patients [38], curve correction averaged 60%, with an average operating time of 5.8 hours. This initial series of patients suggested that thoracoscopic instrumentation was a viable option in the treatment of adolescent idiopathic scoliosis; however, success remained dependent on patient selection and the surgeons technical ability. Eighteen of the rst 20 consecutive patients of this series now have more than 5-year follow-up; deformity correction and absolute pulmonary function have been maintained, and successful bony fusion has occurred at 92% of the motion segments (unpublished data).

contraindications for anterior versus posterior surgical intervention (for thoracic and thoracolumbar curve patterns) have been dened to some degree, there remains appropriate exibility in the decision-making process, allowing the surgeon to make an optimal recommendation for each patient based on surgeon experience and patient needs.

References
[1] Nachemson A. A long term follow-up study of nontreated scoliosis. Acta Orthop Scand 1968;39(4): 46676. [2] Nilsonne U, Lundgren KD. Long-term prognosis in idiopathic scoliosis. Acta Orthop Scand 1968;39(4): 45665. [3] Andersen MO, Christensen SB, Thomsen K. Outcome at 10 years after treatment for adolescent idiopathic scoliosis. Spine 2006;31(3):3504. [4] Benli IT, Ates B, Akalin S, et al. Minimum 10 years follow-up surgical results of adolescent idiopathic scoliosis patients treated with TSRH instrumentation. Eur Spine J 2007;16(3):38191. [5] Dwyer AF, Newton NC, Sherwood AA. An anterior approach to scoliosis. A preliminary report. Clin Orthop Relat Res 1969;62:192202. [6] Moe JH, Purcell GA, Bradford DS. Zielke instrumentation (VDS) for the correction of spinal curvature. Analysis of results in 66 patients. Clin Orthop Relat Res 1983;(180):13353. [7] Mack MJ, Regan JJ, Bobechko WP, et al. Application of thoracoscopy for diseases of the spine. Ann Thorac Surg 1993;56(3):7368. [8] Regan JJ, Mack MJ, Picetti GD 3rd. A technical report on video-assisted thoracoscopy in thoracic spinal surgery. Preliminary description. Spine 1995; 20(7):8317. [9] Biondi J, Weiner DS, Bethem D, et al. Correlation of Risser sign and bone age determination in adolescent idiopathic scoliosis. J Pediatr Orthop 1985;5(6): 697701. [10] Peterson LE, Nachemson AL. Prediction of progression of the curve in girls who have adolescent idiopathic scoliosis of moderate severity. Logistic regression analysis based on data from The Brace Study of the Scoliosis Research Society. J Bone Joint Surg Am 1995;77(6):8237. [11] Lonstein JE, Carlson JM. The prediction of curve progression in untreated idiopathic scoliosis during growth. J Bone Joint Surg Am 1984;66(7):106171. [12] Weinstein SL, Ponseti IV. Curve progression in idiopathic scoliosis. J Bone Joint Surg Am 1983; 65(4):44755. [13] Dubousset J, Herring JA, Shuebarger H. The crankshaft phenomenon. J Pediatr Orthop 1989; 9(5):54150.

Summary The anterior surgical treatments for idiopathic scoliosis continue to evolve and provide advantages over posterior procedures in specic instances. Open and thoracoscopic anterior approaches allow direct access to the anterior stabilizing structures of the spine, enable mobilization of a rigid deformity, and provide a large surface area for arthrodesis; however, these procedures are associated with increased rates of pulmonary compromise, and long-term studies have not been completed to determine their ability to maintain deformity correction relative to modern posterior segmental pedicle screw constructs. Thoracoscopic procedures are technically demanding and surgeon experience must be considered before recommending this procedure to carefully selected patients; however, they do provide a more cosmetically appealing alternative to a large midline posterior or anterolateral thoracotomy scar. Although the indications and

540

UPASANI & NEWTON

[14] Sanders JO, Little DG, Richards BS. Prediction of the crankshaft phenomenon by peak height velocity. Spine 1997;22(12):13526 [discussion: 13567]. [15] Lowe TG, Betz R, Lenke L, et al. Anterior single-rod instrumentation of the thoracic and lumbar spine: saving levels. Spine 2003;28(20):S20816. [16] Lowe TG, Alongi PR, Smith DA, et al. Anterior single rod instrumentation for thoracolumbar adolescent idiopathic scoliosis with and without the use of structural interbody support. Spine 2003;28(19): 223241 [discussion: 22412]. [17] Hurford RK Jr, Lenke LG, Lee SS, et al. Prospective radiographic and clinical outcomes of dual-rod instrumented anterior spinal fusion in adolescent idiopathic scoliosis: comparison with single-rod constructs. Spine 2006;31(20):23228. [18] Lowe TG, Enguidanos ST, Smith DA, et al. Singlerod versus dual-rod anterior instrumentation for idiopathic scoliosis: a biomechanical study. Spine 2005;30(3):3117. [19] Early SD, Newton PO, White KK, et al. The feasibility of anterior thoracoscopic spine surgery in children under 30 kilograms. Spine 2002;27(21): 236873. [20] Lowe T, OBrien M, Smith D, et al. Central and juxta-endplate vertebral body screw placement: a biomechanical analysis in a human cadaveric model. Spine 2002;27(4):36973. [21] Horton WC, Blackstock SF, Norman JT, et al. Strength of xation of anterior vertebral body screws. Spine 1996;21(4):43944. [22] Faciszewski T, Winter RB, Lonstein JE, et al. The surgical and medical perioperative complications of anterior spinal fusion surgery in the thoracic and lumbar spine in adults. A review of 1223 procedures. Spine 1995;20(14):15929. [23] Anderson PR, Puno MR, Lovell SL, et al. Postoperative respiratory complications in non-idiopathic scoliosis. Acta Anaesthesiol Scand 1985;29(2): 18692. [24] Weis JC, Betz RR, Clements DH 3rd, et al. Prevalence of perioperative complications after anterior spinal fusion for patients with idiopathic scoliosis. J Spinal Disord 1997;10(5):3715. [25] Faro FD, Marks MC, Newton PO, et al. Perioperative changes in pulmonary function after anterior scoliosis instrumentation: thoracoscopic versus open approaches. Spine 2005;30(9):105863. [26] Newton PO, Cardelia JM, Farnsworth CL, et al. A biomechanical comparison of open and thoracoscopic anterior spinal release in a goat model. Spine 1998;23(5):5305 [discussion: 536].

[27] Wall EJ, Bylski-Austrow DI, Shelton FS, et al. Endoscopic discectomy increases thoracic spine exibility as eectively as open discectomy. A mechanical study in a porcine model. Spine 1998;23(1):915 [discussion: 156]. [28] Huntington CF, Murrell WD, Betz RR, et al. Comparison of thoracoscopic and open thoracic discectomy in a live ovine model for anterior spinal fusion. Spine 1998;23(15):1699702. [29] Newton PO, Wenger DR, Mubarak SJ, et al. Anterior release and fusion in pediatric spinal deformity. A comparison of early outcome and cost of thoracoscopic and open thoracotomy approaches. Spine 1997;22(12):1398406. [30] Newton PO, Shea KG, Granlund KF. Dening the pediatric spinal thoracoscopy learning curve: sixtyve consecutive cases. Spine 2000;25(8):102835. [31] Newton PO, White KK, Faro F, et al. The success of thoracoscopic anterior fusion in a consecutive series of 112 pediatric spinal deformity cases. Spine 2005; 30(4):3928. [32] Lenke LG, Betz RR, Bridwell KH, et al. Spontaneous lumbar curve coronal correction after selective anterior or posterior thoracic fusion in adolescent idiopathic scoliosis. Spine 1999;24(16):166371 [discussion: 1672]. [33] Kovac V, Puljiz A, Smerdelj M, et al. Scoliosis curve correction, thoracic volume changes, and thoracic diameters in scoliotic patients after anterior and after posterior instrumentation. Int Orthop 2001;25(2):669. [34] Kuklo TR, Lenke LG, Graham EJ, et al. Correlation of radiographic, clinical, and patient assessment of shoulder balance following fusion versus nonfusion of the proximal thoracic curve in adolescent idiopathic scoliosis. Spine 2002;27(18):201320. [35] Kuklo TR, Lenke LG, Won DS, et al. Spontaneous proximal thoracic curve correction after isolated fusion of the main thoracic curve in adolescent idiopathic scoliosis. Spine 2001;26(18):196675. [36] Potter BK, Kuklo TR, Lenke LG. Radiographic outcomes of anterior spinal fusion versus posterior spinal fusion with thoracic pedicle screws for treatment of Lenke Type I adolescent idiopathic scoliosis curves. Spine 2005;30(16):185966. [37] Betz RR, Harms J, Clements DH 3rd, et al. Comparison of anterior and posterior instrumentation for correction of adolescent thoracic idiopathic scoliosis. Spine 1999;24(3):22539. [38] Newton PO, Parent S, Marks M, et al. Prospective evaluation of 50 consecutive scoliosis patients surgically treated with thoracoscopic anterior instrumentation. Spine 2005;30(Suppl 17):S1009.

Orthop Clin N Am 38 (2007) 541545

Fusionless Treatment of Scoliosis


James T. Guille, MDa,b,*, Linda P. DAndrea, MDa,b, Randal R. Betz, MDb,c
Division of Spinal Disorders, Brandywine Institute of Orthopaedics, 600 Creekside Drive, Suite 611, Pottstown, PA 19464, USA b Temple University School of Medicine, 1500 N. Broad Street, Philadelphia, PA 19140, USA c Shriners Hospital for Children, 3551 North Broad Street, Philadelphia, PA 19140, USA
a

Scoliosis is a complex three-dimensional spinal deformity that results from both known and unknown causes in patients of all ages. Young children who present with large curves have a high tendency to progress as the child grows, but this progression can vary. More is known about the natural history of curve progression in adolescent idiopathic scoliosis, which is dependent on the patients skeletal maturity, the curve pattern, and the curve magnitude. Children with congenital or neuromuscular forms of scoliosis can have an unpredictable course, with most being progressive. Curves that present in the growing child may be amenable to a variety of treatments to address curve progression during growth. The standard of care currently for skeletally immature patients with progressive scoliosis measuring greater than or equal to 25 is a thoracolumbosacral orthosis. These braces are used in an attempt to prevent curve progression, but the results can be variable. Brace wear can be associated with many problems. As most braces exert their eect via pressure on the rib cage, their inuence on the chest wall in the growing child creates concern. Some children have a problem with the stigmata associated with wearing a brace, especially children who have to wear a brace for many years.

* Corresponding author. Division of Spinal Disorders, Brandywine Institute of Orthopaedics, 600 Creekside Drive, Suite 611, Pottstown, PA 19464. E-mail address: guille@brandywineortho.com (J.T. Guille).

Also, while brace treatment is noninvasive and preserves growth, motion, and function of the spine, it does not correct an established deformity. While most orthopaedists, families, and patients agree that it is reasonable to wear a scoliosis brace for 1 or 2 years if it means preventing an operation, a more dicult situation is encountered in the very young child who faces the prospect of wearing a brace for many years with no guarantee of a favorable outcome. It is in these children that fusionless treatment options hold the greatest potential. To date, most fusionless treatment options have centered on addressing a progressive scoliosis in the growing child. The fusionless treatment of an established larger curve in the patient who is skeletally mature or is nearly so has been limited, but may hold some potential [1,2]. Fusionless scoliosis surgery may provide substantial advantages over both bracing and denitive spinal fusion. The goal of this procedure is to harness the patients inherent spinal growth and redirect it to achieve correction, rather than progression, of the curve. Several methods of treatment of scoliosis without fusion have evolved and include (1) anterior vertebral body stapling, (2) anterior spinal tethering, (3) convex scoliosis tethering, (4) mechanical modulation of spinal growth, and (4) internal bracing. The anterior fusionless techniques are theoretically more advantageous than external bracing because they address the deformity directly at the spine and not via the chest wall and ribs, and because they eliminate problems with patient noncompliance.
orthopedic.theclinics.com

0030-5898/07/$ - see front matter 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.ocl.2007.07.003

542

GUILLE

et al

Furthermore, minimally invasive thoracoscopic access to the anterior thoracic spine is less extensive than that for posterior instrumented surgery. For the purposes of this article the term fusionless scoliosis surgery is used to describe anterior spinal procedures that control the progression of scoliosis during growth, but other fusionless treatments are available. These more established fusionless procedures employ posterior implants (growing rods and the vertical expandable prosthetic titanium rib) that can control the progression of spinal deformity but do not correct an established deformity in the younger child. These posterior growing systems are fraught with complications and take a multiyear commitment of semiannual surgery from the family and patient to be successful. Single or dual growing-rod techniques and the vertical expandable prosthetic titanium rib are not only more invasive than anterior fusionless scoliosis surgery but are associated with higher rates of complications. Anterior fusionless scoliosis surgery avoids multiple procedures, as well as the requirement for an eventual fusion, by oering a single intervention that may provide a more permanent solution to the spinal deformity. Furthermore, correction of a spinal deformity in the absence of a rigid fusion mass spanning multiple vertebral motion segments may ameliorate some of the long-term problems related to spinal fusion with instrumentation, such as adjacent level degeneration. To date, many models of fusionless surgery involving both the genesis of iatrogenic scoliosis and its treatment have been studied [310]. A review of the recent literature shows a marked increase in interest in the topic and number of publications on this subject in the past 2 years [11]. However, much of what has been done has involved experimental methods in animal models, with no transition to the clinical realm or a patient series. The authors rst discuss the advent of anterior vertebral body stapling and its clinical experience, followed by experimental works on other fusionless scoliosis treatment options. Anterior vertebral body stapling The rst recent study of a large patient series is that of Betz and colleagues [12], who reported on the Philadelphia Shriners Hospital experience with anterior vertebral body stapling in patients with adolescent idiopathic scoliosis. This is a landmark study in that previous attempts to correct scoliosis with anterior fusionless techniques had

been disappointing [13,14]. Convex apical vertebral body (hemiepiphyseal) stapling theoretically aords immediate and possibly reversible cessation of growth of the anterior vertebral physes [15,16]. Animal studies using a rat tail model conrm its ability to modulate vertebral growth plates with skeletal xation devices [1719]. In 1951, Nachlas and Borden [13] were initially optimistic about their ability to create and correct lumbar scoliosis in a canine model using a staple that spanned several vertebral levels. Many of the dogs exhibited some correction, and some of the animals exhibited arrest of their curve progression. Some of the staples failed because they spanned three vertebrae. The enthusiasm for this new treatment was lost after the application of their stapling technique in three children with progressive scoliosis yielded poor results. Other investigators have also been dissatised with convex stapling as a means of controlling progressive scoliosis. In 1954, Smith and colleagues [14] presented disappointing results for human patients with congenital scoliosis. The scoliosis correction was limited because the children had little remaining growth and the curves were severe, with considerable rotational deformity. Some staples broke or loosened, possibly because of motion through the intervertebral discs. James M. Ogilvie, MD, reported (personal communication) that in 1997 he began performing anterior vertebral body stapling with thoracoscopic assistance in six patients: three with infantile scoliosis, two with juvenile scoliosis, and one with spina bida. Preoperatively, the patients curves had progressed despite bracing for the previous year. At 2-year follow-up, four of the six curves stabilized following the procedure. In two patients the staples partially dislodged, requiring another operation to replace the staple. A likely cause of disappointing results in previous series and experiments was the implant (staple). While the concept of stapling the anterior vertebral endplates/physes for growth modulation and curve stabilization seems sound, the staples designed for epiphyseal stapling about the knee are prone to dislodge in the spine because they are not designed for movement in the spine. To address this concern, Medtronic Sofamor Danek (Memphis, Tennessee) has designed a specic spine staple called the Nitinol (Nickel Titanium Naval Ordinance Laboratory) staple, which has 510K approval from the US Food and Drug Administration specically for use as an anterior

FUSIONLESS TREATMENT OF SCOLIOSIS

543

spinal staple. The uniqueness of this staple is that it is made out of a shape memory alloy in which the prongs are straight when cooled but, when warmed to body temperature, clamp down in a C shape in the bone for secure xation. Before its use in humans, the Nitinol staple had been tested in a goat model by Braun and colleagues [10] and was shown to be safe and have utility for arresting iatrogenic scoliosis in the goat. In 2003, Betz and colleagues [12] reported on the use of the Nitinol staple in 21 skeletally immature patients with adolescent idiopathic scoliosis. Indications for the procedure were either brace noncompliance or the inability of the brace to prevent progression of the curve. They found the procedure to be safe and eective, with the results comparable with that of what would have been expected from bracing. In 2005, this group reported on 39 patients and their increased experience with the procedure [20]. Eighty-seven percent of those patients older than 8 years at the time of stapling who had a curve of 50 or less with at least 1 year of follow-up had stabilization of their curve. No curve less than 30 at the time of stapling progressed more than 10 at follow-up [20]. Experience from more than 80 patients to date has taught that the entire Cobb angle of all of the curves needs to be stapled (Figs. 1 and 2), and staple implants with tines proportional to the length of the staple (Fig. 3) have yielded better results than previous designs. Relative contraindications include curves above T2 or below L4, very small vertebral body size, thoracic kyphosis greater than 40 , and coronal curves above 45 . MRI scans of several of these patients have shown that the intervertebral discs remain hydrated and are normal in appearance (Fig. 4). The technique of anterior vertebral body stapling has been published [12,20,21].

Fig. 1. Posteroanterior radiograph of a 7-year-old girl with idiopathic scoliosis having a 25 thoracic and 40 lumbar curve.

Tethering procedures Braun and colleagues [6,7,13] have done much work with the goat model in the creation of experimental scoliosis with a posterior tether of the ribs, and correction of this deformity with an anterior fusionless technique using bone anchors with ligament tethers. Brauns group and Newton and colleagues [22,23] have also shown that if the bone-anchorligament-tether does not fail, it causes vertebral body wedging and scoliosis correction via growth modulation. Lowe and

colleagues [24] looked at the role of posterior tethers in sagittal plane (kyphotic) deformities in sheep. While the posterior tether resulted in less kyphosis, this was accompanied by signicantly less motion from heterotopic ossication. The authors are unaware, however, of any published series on the use of bone-anchorsligament-tethers in human patients. Braun and colleagues [22] recently compared shape memory staples with bone-anchorsligament-tethers in the fusionless treatment of iatrogenic scoliosis in the goat. Results of this study demonstrate greater ecacy and integrity of a bone-anchorexible-ligament-loop-tether compared with a more rigid shape memory alloy staple in the fusionless treatment of a progressive experimental scoliosis. In contrast to the more rigid staple base, the ligament loop used with the bone anchor provided a more exible tether spanning the disc space. This increased exibility was likely associated with decreased forces during spinal motion. This decrease in force potentially protected the bone anchor from loosening. Whereas the staple demonstrated no signicant

544

GUILLE

et al

Fig. 4. MRI of the patient in Fig. 2 showing normalappearing well-hydrated discs. Fig. 2. At the age of 12 years, both curves completely corrected in girl shown in Fig. 1.

change in axial pullout strength between the two time-points, the bone anchor showed a signicant increase in pullout strength. It is postulated that this dierence was related to two factors: (1) the rigidity of the portion of the implant spanning

the motion segment and (2) the quality of the xation to bone. The staple, though made of shape memory alloy, has a relatively rigid base spanning the disc space compared with the ligament-loop bone-anchor construct. Additionally, the smooth tine is suboptimal for xation to bone and relies primarily on the mechanical crimping eect of the deployed shape memory alloy staple. The authors do not necessarily believe that the lack of rigid xation seen with staples is bad. The associated halo that is seen around the tines on radiographs in long-standing cases shows that there is motion, and perhaps this is why there is such a low rate of staple failure and why patients are able to maintain spinal mobility.

Summary The recent investigations of convex anterior vertebral body stapling, both in animal models and in juvenile and adolescent scoliosis have oered promising early results with use of improved implants and techniques. The use of a shape memory alloy staple tailored to the size of the vertebral body, the application of several staples per level, the instrumentation of the Cobb levels of all curves, and the employment of

Fig. 3. Intraoperative uoroscopy image showing ideal placement of the proportional-sized staples.

FUSIONLESS TREATMENT OF SCOLIOSIS

545

minimally invasive thoracoscopic approaches all oer substantial improvements over previous fusionless techniques. Patient selection may also play a role in the current success of these fusionless treatments, with perhaps the ideal candidates for this intervention possessing smaller and more exible curves. Still, reports on the clinical success of these stapling procedures are based on shortterm results. Long-term results of the eects on the instrumented motion segments and adjacent spine are not yet available. References
[1] Guille JT, Betz RR, Balsara RK, et al. The feasibility, safety, and utility of vertebral wedge osteotomies for the fusionless treatment of paralytic scoliosis. Spine 2003;28:S26674. [2] Maruyama T, Kitagawa T, Takeshita K, et al. Fusionless surgery for scoliosis: 217 year radiographic and clinical follow-up. Spine 2006;31:23105. [3] Braun JT, Hines JL, Akyuz E, et al. Relative versus absolute modulation of growth in the fusionless treatment of experimental scoliosis. Spine 2006;31: 177682. [4] Braun JT, Homan M, Akyuz E, et al. Mechanical modulation of vertebral growth in the fusionless treatment of progressive scoliosis in an experimental model. Spine 2006;31:131420. [5] Braun JT, Akyuz E, Ogilvie JW. The use of animal models in fusionless scoliosis investigations. Spine 2005;30:S3545. [6] Braun JT, Akyuz E, Udall H, et al. Three-dimensional analysis of 2 fusionless scoliosis treatments: a exible ligament tether versus a rigid-shape memory alloy staple. Spine 2006;31:2628. [7] Braun JT, Ogilvie JW, Akyuz E, et al. Creation of an experimental idiopathic-type scoliosis in an immature goat model using a exible posterior asymmetric tether. Spine 2006;31:14104. [8] Puttitz CM, Masaru F, Barkley A, et al. A biomechanical assessment of thoracic spine stapling. Spine 2007;32:76671. [9] Wall EJ, Bylski-Austrow DI, Kolata RJ, et al. Endoscopic mechanical spinal hemiepiphysiodesis modies spine growth. Spine 2005;30:114853. [10] Braun JT, Ogilvie JW, Akyuz E, et al. Fusionless scoliosis correction using a shape memory alloy staple in the anterior thoracic spine of the immature goat. Spine 2004;29:19809.

[11] Cunningham ME, Frelinghuysen PH, Roh JS, et al. Fusionless scoliosis surgery. Curr Opin Pediatr 2005; 17:4853. [12] Betz RR, Kim J, DAndrea LP, et al. An innovative technique of vertebral body stapling for the treatment of patients with adolescent idiopathic scoliosis: a feasibility, safety, and utility study. Spine 2003;28: S25565. [13] Nachlas IW, Borden JN. The cure of experimental scoliosis by directed growth control. J Bone Joint Surg [Am] 1951;33:2434. [14] Smith AD, von Lackum HL, Wylie R. An operation for stapling vertebral bodies in congenital scoliosis. J Bone Joint Surg [Am] 1954;36:3428. [15] Roaf R. The treatment of progressive scoliosis by unilateral growth-arrest. J Bone Joint Surg [Br] 1963;45:63751. [16] Roaf R. Vertebral growth and its mechanical control. J Bone Joint Surg [Br] 1960;42:4059. [17] Akyuz E, Braun JT, Brown NAT, et al. Static versus dynamic loading in the mechanical modulation of vertebral growth. Spine 2006;31:E9528. [18] Mente PL, Aronsson DD, Stokes IA, et al. Mechanical modulation of growth for the correction of vertebral wedge deformities. J Orthop Res 1999;17: 51824. [19] Stokes IA, Spence H, Aronsson DD, et al. Mechanical modulation of vertebral body growth. Implications for scoliosis progression. Spine 1996;21: 11627. [20] Betz RR, DAndrea LP, Mulcahey MJ, et al. Vertebral body stapling procedure for the treatment of scoliosis in the growing child. Clin Orthop Relat Res 2005;434:5560. [21] DAndrea LP, Guille JT, Betz RR. Intervertebral stapling for spinal deformity. In: Vaccaro A, Albert T, editors. Spine surgery: tricks of the trade, in press. [22] Braun JT, Akyuz E, Ogilvie JW, et al. The ecacy and integrity of shape memory alloy staples and bone anchors with ligament tethers in the fusionless treatment of experimental scoliosis. J Bone Joint Surg [Am] 2005;87:203851. [23] Newton PO, Faro FD, Farnsworth CL, et al. Multilevel spinal growth modulation with an anterolateral exible tether in an immature bovine model. Spine 2005;30:260813. [24] Lowe TG, Wilson L, Chien J-T, et al. A posterior tether for fusionless modulation of sagittal plane growth in a sheep model. Spine 2005;30: S6974.

Orthop Clin N Am 38 (2007) 547552

The Use of Growth-Sparing Instrumentation in Pediatric Spinal Deformity


John T. Smith, MD
Department of Orthopaedics and Pediatrics, The University of Utah School of Medicine, Primary Childrens Medical Center, 100 North Medical Drive, Suite 4550, Salt Lake City, UT 84113, USA

The management of severe spinal deformity in the growing child remains a challenging problem. Methods to manage or correct spinal deformity in young children date back to Hippocrates. Nonoperative methods range from orthotics to casting to traction; however, in certain circumstances, these techniques cannot eectively prevent deformity progression or are not tolerated by the child and surgical methods are required. Current options for surgical management of spinal deformity in the growing child include denitive spinal fusion with or without instrumentation, selective fusion (hemiepiphysiodesis), growth modulation (staples or growth tethers), spinal instrumentation without fusion (growing rods, Luque trolleys), or more recently, the use of the vertical expandable prosthetic titanium rib (VEPTR). Historically, all of these methods have a signicant complication rate and despite advances in technology and instrumentation, remain problematic. This article provides an overview of current methods and outcomes for spinal instrumentation in the growing spine.

Normal growth of the spine and thorax Understanding the normal growth of the spine and thorax is critical to understanding the decision process for surgical management of severe spinal deformity in the young child. Dimeglio [1] has contributed signicantly to current understanding of the growing spine. The prenatal development of the spinal column is the result of

E-mail address: john.smith@hsc.utah.edu

a complex process of segmental development and dierentiation resulting in formation of each vertebra. With a minimum of three growth zones per vertebra, there are at least 100 growth plates within the vertebral column that contribute to the ultimate morphology of the adult spine. This complex process of segmental development oers the potential for developmental errors, producing an innite variety of growth alterations and deformity. A fourth dimension of spinal growth is the thorax. Thoracic distortion is a common sequela of spinal deformity, resulting in alteration of pulmonary growth and respiratory function. The primary eect of scoliosis on pulmonary function is believed to be mechanical, producing impaired movement of the chest wall, decreased volume of the thorax, and reduced compliance [2]. In addition, it has been postulated that most alveolar growth and multiplication occurs until about age 8 years and that reduction of thoracic volume may impair alveolar development and growth [3]. In early-onset scoliosis (before age 5 years), Pehrsson and colleagues [4] reported on a cohort of patients who have unoperated scoliosis; there was a signicantly increased risk for mortality in patients who have infantile (P!.001) and juvenile (P!.01) scoliosis but not in patients who have adolescent scoliosis. It is believed that the increased risk for death is due to respiratory failure. Therefore, any treatment strategy for severe deformity must consider the eect a given treatment would have on the growth and shape of the thorax and its consequence on pulmonary function. The rate of normal spinal growth varies with age. There is a rapid phase of spinal growth from birth to age 5 years and a slower phase of
orthopedic.theclinics.com

0030-5898/07/$ - see front matter 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.ocl.2007.03.009

548

SMITH

growth between 5 and 10 years, which is followed by a period of rapid growth until completion of puberty (14.5 years in girls, 16.5 years in boys). The thoracic spine averages 11 cm in length at birth and reaches about 28 cm in men and 26 cm in women at skeletal maturity. The lumbar spine averages 7 cm at birth and grows to an average of 16 cm in men and 15.5 cm in women [1]. Early fusion of the growing spine eectively stops the longitudinal growth of the spine. Dimeglio and Bonnel [5] published tables that predict the overall eect of early fusion on ultimate sitting height based on the age of fusion. For example, each thoracic vertebra contributes 2.5% to the overall sitting height at skeletal maturity. From these data, the surgeon can calculate the resultant eect of fusion based on the number of vertebral segments fused and the age of the child. This awareness is critical to the decision process of growth-sparing procedures and the associated risks compared with the ultimate loss of height with a denitive fusion procedure.

Patient evaluation Severe spinal deformity in children results from many etiologies and is associated with a variety of comorbidities. The preoperative evaluation of this population requires a multidisciplinary team of physicians. Although the orthopedic surgeon often is asked to be the team leader, additional expertise is virtually always needed, including radiology, pulmonary medicine, cardiology, nutrition, and others. When considering a management strategy that involves signicant risks, repeated surgeries, and prolonged growth, the childs health and the familys ability to cope with and endure this form of treatment must be evaluated thoroughly. Indications The use of growth-sparing instrumentation is indicated in patients who have progressive spinal deformity that cannot be controlled by nonoperative means, such as bracing, and where there is signicant spinal growth remaining. The surgeon must consider the risk versus benet of this course of treatment. Factors, including age, curve severity, curve type, growth remaining, and the childs general health and ability to tolerate multiple surgeries, must be considered in relationship to the benet of continued growth of the spine. Instrumentation options The rst use of spinal instrumentation without fusion dates back to the introduction of the Harrington rod. Harrington [8] concluded that fusion should not be considered in children younger than 10 years of age. Moe and colleagues [9] reported that a single Harrington rod placed subcutaneously through limited incisions without fusion resulted in preservation of growth in the instrumented spine. Further studies of this technique with newer-generation implants have shown improvement in Cobb angles and sagittal contour [10]. Luque trolleys have been used to control progressive curves using segmental instrumentation. Rinsky and colleagues [11] reported that there was an average 32% loss of correction at a mean follow-up of 28 months, with minimal preservation of spinal growth and a high incidence of rod breakage; they concluded that these early results were discouraging. Pratt and colleagues [12] reported that the use of a Luque trolley alone resulted in a signicant loss of correction over

The outcomes of early fusion Historically, early fusion of progressive curves has been advocated in young children with progressive curves. The concept was that early fusion would allow for a smaller number of spinal segments to be fused preserving adjacent spinal growth and a balanced spine; however, recent long-term follow-up studies suggest that early fusion had deleterious eects on pulmonary function. Goldberg and colleagues [6] reported that children whose spinal deformity required early surgery (mean age, 4.1 years; 11 patients) had recurrence of their deformity and impaired respiratory function at skeletal maturity. Emans and colleagues [7] reported on a series of 13 patients who underwent fusion of at least four thoracic vertebras before 5 years of age; they demonstrated signicant impairment of pulmonary function. The extent of pulmonary impairment correlated closely with the number of thoracic vertebrae fused. Further operative intervention was required in 3 patients. Quality-of-life questionnaires indicated that 7 of 13 patients had persistent pain, and 3 of 13 patients complained of major functional and cosmetic limitations. These papers have resulted in recent increased interest on growth preservation of the spine and thorax in young children with severe spinal deformity.

GROWTH SPARING INSTRUMENTATION

549

Fig. 1. (A) Five-year-old boy who had infantile (early-onset) scoliosis that failed nonoperative brace management. (B) Postoperative posteroanterior radiograph after placement of dual growing rods using a hybrid hook pedicle screw construct and tandem connectors for rod expansion.

time and noted only 32% of predicted spinal growth over the time of treatment. The signicant incidence of rod breakage and hook failure with single growing-rod constructs led Akbarnia and colleagues [13] to develop the dual growing-rod technique (Fig. 1). This technique establishes anchor points in the upper and lower spine using hooks or pedicle screws and

passes rods subcutaneously, attached by parallel connectors that allow for rod distraction with growth. Early results with this technique in a heterogenous group of patients documented the ability to preserve spinal growth and improve the Cobb angle over time. Thompson and colleagues [14] compared single with dual growing rods; dual growing rods provided superior outcomes.

Fig. 2. (A) Five-year-old girl with a history of an omphalocele and progressive early-onset scoliosis resulting in pelvic obliquity. She had failed brace treatment. (B) Standing posteroanterior (PA) view of the spine showing scoliosis and pelvic obliquity measuring 87 . (C) Standing PA view of the spine after placement of bilateral percutaneous rib to pelvis VEPTR constructs.

550

SMITH

Fig. 3. (A) Twenty-month-old child who had a severe gibbus deformity associated with myelodysplasia and early skin breakdown over the gibbus deformity. (B) Sagittal MRI showing the gibbus deformity. Lateral (C) and anteroposterior radiographs (D) demonstrating short-term follow-up after placement of bilateral rib to pelvis VEPTR constructs. This technique allows for control of the deformity while avoiding poor midline skin.

Additionally, they reported that the apical fusion was detrimental to the overall growth of the spine and should be avoided. A new alternative for the treatment of progressive scoliosis is the use of VEPTR. The VEPTR was developed by Dr. Robert Campbell, Jr. for the treatment of thoracic insuciency syndrome (TIS), dened as the inability of the thorax to support normal respiration and lung growth [15]. Typically, TIS is the result of the tethering eect of congenital rib fusions and

scoliosis that produces a constriction of the hemithorax and reduced lung volumes; it also can result from other syndromes associated with a hypoplastic thorax. This reduced volume restricts growth of the lung, especially during the critical ages of lung growth during the rst 8 years of life. The interdependent relationship between scoliosis and chest wall deformity led to the development of expansion thoracoplasty and stabilization using the VEPTR device [16]. This

GROWTH SPARING INSTRUMENTATION

551

procedure is able to promote lung growth and spinal growth without fusion and avoids the deleterious eects of early fusion [7]. Gollogly and colleagues [17] documented an increase in lung volumes following expansion thoracoplasty measured by CT scan. The VEPTR device also has been used to treat progressive scoliosis without fused ribs without the addition of an expansion thoracoplasty. As a minimally invasive alternative to growing rods, Smith developed the bilateral percutaneous VEPTR technique. In this technique, the VEPTR devices are attached to the ribs through a small proximal midline incision. Distally, the implant is seated on the pelvis using a modied Dunn McCarthy hook or is anchored to the spine using laminar hooks (Fig. 2). The devices are expanded at 6-month intervals during growth. Smith and Smart [18] reported the early pilot data with this technique, noting satisfactory control of curve progression and a complication rate that is similar to other growing instrumentation techniques. Long-term follow-up data are lacking with this technique. A recent review of mid-term follow-up data suggested that this procedure should be used with caution in ambulatory children because of a forward shift in sagittal balance and resultant crouched gait (JT Smith, unpublished data, 2007). It has been useful in children who have neuromuscular disease, such as spina bida (Fig. 3) [19]. Complications remain problematic, as with all fusionless systems.

Future directions The ability to control progressive spinal deformity and children using fusionless spinal instrumentation systems remains a signicant challenge for the future. Current methods are fraught with an unacceptable rate of complications, most notably loss of xation, implant breakage, and device migration. There is a signicant need for self-expanding implants and better methods for xation to the spine and chest wall, obviating the need for multiple surgeries for expansion of growing instrumentation.

References
[1] Dimeglio A. Growth in pediatric orthopaedics. In: Morrissy RT, Weinstein SL, editors. Lovell and Winters pediatric orthopaedics. 5th edition. Philadelphia: Lippincott, Williams and Williams; 2001. p. 503.

[2] Jones RS, Kennedy JD, Hasham F, et al. Mechanical ineciency of the thoracic cate in scoliosis. Thorax 1981;36:456561. [3] Davies G, Reid L. Eect of scoliosis on growth of alveoli and pulmonary arteries on the right ventricle. Arch Dis Child 1971;46:62332. [4] Pehrsson K, Larsson S, Oden A, et al. Long term follow up of patients with untreated scoliosis. A study of mortality, causes of death and symptoms. Spine 1992;17(9):10916. [5] Dimeglio A, Bonnel F. Le rachis en coissance [The spinal column in growth]. Paris: Springer Verlag; 1990. [6] Goldberg CJ, Gillic I, Connaughton O, et al. Respiratory function and cosmesis at maturity in infantileonset scoliosis. Spine 2003;28(20):2397406. [7] Emans J, Kassab F, Caubet JF, et al. Earlier and more extensive fusion is associated with diminished pulmonary function. Outcome after spinal fusion of 4 or more thoracic spinal segments before Age 5. Scoliosis Research Society Annual Meeting. Paper 101, Buenos Aires (Argentina), September 7, 2004. [8] Harrington PR. Scoliosis in the growing spine. Pedriatr Clin North Am 1963;10:22545. [9] Moe JH, Kharrat K, Winter RB, et al. Harrington instrumentation without fusion plus external orthotic support for the treatment of dicult curvature problems in young children. Clin Orthop Relat Res 1984;185:3545. [10] Blakemore LC, Scoles PV, Poe-Kochert C, et al. Submuscular isola rod with or without limited apical fusion in the management of severe spinal deformities in young children: preliminary report. Spine 2001;26:20448. [11] Rinsky L, Gamble JG, Bleck EE. Segmental instrumentation without fusion in children with progressive scoliosis. J Pediatr Orthop 1985;5:68790. [12] Pratt RK, Webb JK, Burwell RG, et al. Luque trolley and convex hemiepiphysiodesis in the management of infantile and juvenile idiopathic scoliosis. Spine 1999;25:153847. [13] Akbarnia BA, Marks DS, Boachie-Adjei O, et al. Dual growing rod technique for the treatment of progressive early onset scoliosis: a multicenter study. Spine 2005;30:S4657. [14] Thompson GH, Akbarnia BA, Kostial P, et al. Comparison of single and dual growing rod techniques followed through denitive surgery. A preliminary study. Spine 2005;30:203944. [15] Campbell RM, Smith MD, Mayes TC, et al. The characteristics of thoracic insuciency syndrome associated with fused ribs and congenital scoliosis. J Bone Joint Surg Am 2004;85:399408. [16] Campbell RM, Smith MD, Mayes TC, et al. The eect of opening wedge thoracostomy on thoracic insuciency syndrome associated with fused ribs and congenital scoliosis. J Bone Joint Surg Am 2004;86:165974. [17] Gollogly S, Smith JT, Campbell RM. Determining lung volume with three dimensional reconstructions

552 of CT scan data: a pilot study to evaluate the eects of expansion thoracoplasty on children with severe spinal deformities. J Pediatr Orthop 2004;24:3238. [18] Smith JT, Smart MP. Treatment of progressive spinal deformity using a bilateral dual VEPTR construct

SMITH

from ribs to pelvis without thoracoplasty: a pilot study of 10 patients. IMAST. Podium presentation, 2004. [19] Smith JT, Smart MP. VEPTR in spina bida. IMAST. E-Poster. Scoliosis Research Society. Athens (Greece), July 9, 2006.

Orthop Clin N Am 38 (2007) 553562

Scoliosis Associated with Neurobromatosis


Alvin H. Crawford, MD, FACSa,*, Jose Herrera-Soto, MDb
a

Cincinnati Childrens Hospital, 3333 Burnet Avenue, ML 2017, Cincinnati, OH 45229-3039, USA b Orlando Regional Medical Center, 86 West Underwood Street, Orlando, FL 32806, USA

Neurobromatosis type 1 (NF-1), also called von Recklinghausen disease or peripheral neurobromatosis, is a multisystemic disease that primarily aects cellular growth of neural tissue [1]. The entity is one of the most common dominantly inherited gene disorders of the nervous system in humans, aecting 1 in 4000 individuals [2]. There is no gender or ethnic predilection [1,3]. Approximately 50% of all cases of NF-1 are due to new mutations, which is 100-fold higher than the usual rate for a single locus and may reect the huge size of the NF-1 1ocus (estimated at 350,000 base pairs). The gene responsible for NF-1 has been identied and is localized to chromosome 17q [3]. The product for this gene is neurobromin. Neurobromin is involved in controlling cellular growth and its dierentiation. Therefore, it works as a tumor suppressor gene [4]. Prenatal testing is now possible in some families, but its use is minimal becausedin most patientsdmutations have not been easy to identify. Patients who have NF-1 develop Schwann cell tumors called neurobromas and pigmentation abnormalities [2]. Cloning the gene has allowed the creation of animal models that ultimately may be used to develop more eective therapy against the disease. NF-1 is a multisystemic disease. It may manifest as abnormalities of the nervous tissue, bones, soft tissue, and skin. The manifestations of NF-1 vary from person to person and range from subclinical to severe. Individuals who carry the gene eventually exhibit some clinical feature of the disease. The penetrance for NF-1 nears 100% during adulthood. Skeletal abnormalities are

common in NF-1; most patients present with some type of bony dysplasia. The orthopedic complications usually appear early. They include spinal deformities, such as scoliosis or kyphosis, congenital tibial dysplasia with bowing and pseudarthrosis of the tibia, forearm, other bones, as well as overgrowth phenomenon of an extremity, and soft tissue tumors [2,3,5,6]. The clinical criteria for the diagnosis were determined in 1987 by the National Institute of Health Consensus Development Conference (Box 1) [7]. To establish the diagnosis for NF-1, two of the seven criteria must be present. These criteria are useful, even in young children. Type 2 neurobromatosis (NF-2) also is an autosomal dominant disorder. NF-2 is not as common as NF-1, aecting about 1 in 100,000 individuals. The NF-2 gene is located on chromosome 22 [8]. Characteristically, individuals who have NF-2 present bilateral schwannomas of the vestibular portion of the eighth cranial nerve. Schwannomas of other peripheral nerves, meningiomas, and ependymomas also are common [9]. No patients who have NF-1 have the eighth cranial nerve tumors [3]. NF-2 does not seem to have orthopedic manifestations. Segmental NF is another form of NF in which features of NF-1 (ie, cafe au lait spots, freckling, and neurobromas) are seen in only one segment of the body. This is believed to be due to somatic mosaicism for the NF-1 gene mutation. A nal form, schwannomatosis, was described recently; it involves multiple deep and painful schwannomas. It may be genetically distinct from NF-1 and NF-2. Imaging studies

* Corresponding author. E-mail address: alvin.crawford@cchmc.org (A.H. Crawford).

Complete spinal radiographs should be scrutinized closely for thoracic lordosis and
orthopedic.theclinics.com

0030-5898/07/$ - see front matter 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.ocl.2007.03.008

554

CRAWFORD & HERRERA-SOTO

Box 1. Diagnostic criteria for neurobromatosis type 1 More than six cafe au lait spots measuring at least 15 mm in adults and 5 mm in children Two or more neurobromas of any type or one plexiform neurobroma Freckling in the axillary or inguinal regions Optic glioma Two or more Lisch nodules (ie, iris hamartomas) A distinctive bony lesion, such as sphenoid wing dysplasia, or thinning of the cortex of a long bone with or without pseudarthrosis A rst-degree relative with NF-1, as suggested by these criteria Since the consensus panel meeting, specic kinds of learning disabilities and MRI abnormalities (especially in children) have been associated specically with NF-1. The pediatric orthopedist encounters other conditions associated with cafe au lait spots, namely, Watsons syndrome, brous dysplasia (ie, McCune-Albright syndrome), LEOPARD (lentigines, electrocardiogram [EKG] abnormalities, ocular hypertelorism, pulmonary stenosis, abnormalities of genitalia, retardation of growth, and deafness) syndrome, and Noonans syndrome. Only the gene for Watsons syndrome has been linked to the NF-1 locus.

Table 1 Nine radiographic characteristics of dystrophic deformity as seen in 457 children Characteristics Rib penciling Vertebral rotation Posterior vertebral scalloping Vertebral wedging Spindling of transverse processes Anterior vertebral scalloping Widened interpediculate distance Enlarged intervertebral foramina Lateral vertebral scalloping % incidence 62 51 31 36 31 31 29 25 13

From Durrani AA, Crawford AH, Choudry SN, et al. Modulation of spinal deformities in patients with neurobromatosis type 1. Spine 2000:25:6975; with permission.

receive radiographic imaging of the cervical spine. Other reasons for obtaining cervical spinal radiographs in a patient who has NF-1 include the evaluation of torticollis or dysphagia possibly secondary to tumors. If instability or subluxation is suspected, tomograms, CT scans, or MRIs are appropriate. MRI should be performed to evaluate any progressive spinal deformity in NF-1 [3].

Spinal pathology A high percentage of patients who have NF-1 develop some type of spinal disorder during their lifetime [3,13,14]. We discuss our approach to individuals who have NF-1 under the following topics: cervical, thoracic/thoracolumbar, lumbosacral, and spinal canal problems. Cervical spine changes and associated complications Features of the cervical spine in patients who have NF-1 have not received enough attention in the literature [13,15]. The cervical spine should be evaluated at the initial scoliosis assessment. Cervical abnormalities occur more frequently when scoliosis or kyphoscoliosis is present in the thoracolumbar region where the examiners attention is focused on the more obvious deformity. Evidence of dystrophic changes may be present on a true cervical lateral view. The manifestations of NF-1 can be observed as dystrophic changes in the vertebral body or they can be due to pathologic alignment [16]. If any suspicious area is noted on plain radiographs, right and left oblique views should be obtained to look for widening of

cervicothoracic kyphosis. Spinal deformities include dystrophic and nondystrophic changes. The radiographic appearance of dystrophic changes include scalloping of the vertebral borders, severe rotation of the apical vertebra, widening of the spinal canal, enlargement of the neural foramina, hypoplastic or absent pedicles, paraspinal masses, spindling of the transverse process, and rotation of the ribs, which resemble twisted ribbons (Table 1) [6,1012]. All patients who have NF-1 who require endotracheal anesthesia, who undergo halo traction, or who present with neck tumors should

SCOLIOSIS ASSOCIATED WITH NEUROFIBROMATOSIS

555

the neuroforamina. These may represent dumbbell lesions caused by enlarged neurobromas exiting from the spinal canal. MRI is the denitive study to evaluate these lesions. The most common cervical abnormality observed is a severe cervical kyphosis, which, in itself, is highly suggestive of the disorder. In the study of Yong-Hing and colleagues [17] 17 patients who had NF-1 had cervical abnormalities. Of these, 7 patients were asymptomatic, whereas the rest had limited motion or pain in the neck. Four patients had neurologic decits that were attributed to cervical instability. Four of the 17 patients required fusion of the cervical spine. Curtis and colleagues [18] described 8 patients who had paraplegia and NF-1. Four of these patients had cervical spine instability or intraspinal pathology in the cervical spine. Attention also should be given to the atlantoaxial region. Isu and colleagues [19] described three patients who had NF-1 who had C1C2 dislocation with neurologic decit. All patients improved after decompression and fusion. Problems often occur after excision of cervical tumors when excision includes resection of the laminae and posterior elements. Postoperatively, the spine becomes unstable and progressive kyphosis develops. The best prevention is to avoid complete laminectomies or perform spinal instrumentation with abundant bone grafting at the time of decompression to prevent instability [20]. Combined anterior and posterior spinal fusion is recommended for cervical kyphosis if there is severe deformity and instability [21]. Some patients may need fusion in the cervical spine as well as in the thoracic region. This group of patients may benet from a trap door sternal split approach if anterior fusion is needed (Fig. 1) [22,23]. This approach allows anterior exposure of the lower cervical and upper thoracic spine. Scoliosis Scoliosis is the most common osseous defect associated with NF-1 [3]. Weiss [24] and Gould [25] were the rst to emphasize the high incidence of spinal deformities. It may vary in severity from mild and nonprogressive to severe curvatures. The exact etiology is unknown, but it has been suggested to be secondary to osteomalacia, localized neurobromatous tumor eroding and inltrating bone, endocrine disturbances, or mesodermal dysplasia [1]. In a general orthopedic clinic, 2% of patients who have scoliosis have neurobromatosis,

whereas in a neurobromatosis clinic, 10% to 20% of patients have some disorder of the spine. All preadolescent children who have neurobromatosis should be evaluated with scoliosis screening or the Adams forward bend test to exclude a spinal deformity. Usually, scoliosis occurs earlier in children who have neurobromatosis compared with idiopathic scoliosis. Two primary types of scoliosis are observed in persons who have neurobromatosis: dystrophic and nondystrophic [1]. Dystrophic scoliosis in NF-1 is characterized by early-onset, rapid progression that is dicult to treat. Nondystrophic scoliosis behaves similar to idiopathic scoliosis, and the treatment is similar. Natural history It is important to understand the natural history of a condition to make good clinical and surgical treatment decisions. Calvert and colleagues [26] presented a series of treated (n 34) and untreated (n 32) patients who had NF-1 and scoliosis. Seventy-ve percent of patients in the non-treated group had kyphoscoliosis. The investigators reported that patients who had severe anterior vertebral scalloping, noted on the lateral view, progressed an average of 23 per year for scoliosis and kyphosis. All other patients had an average rate of scoliosis progression of 7 and kyphosis progression of 8 per year. Despite solid spinal fusions, some dystrophic patients demonstrate curve progression. A study by Wilde and colleagues [27] showed that vertebral subluxation, disc wedging, and peripheral skeletal dystrophy are additional prognostic features that predict the progression of deformity after arthrodesis of the spine. Patients who had hyperkyphosis (kyphosisO50 and sharply angulated over three vertebrae), despite successful spinal fusion, were noted to progress up to 38 (range, 20 60 ) in one study [27]. Nondystrophic scoliosis The most common curvature is nondystrophic, similar to curves observed in adolescent idiopathic scoliosis in pattern and behavior [3]. This form usually involves 8 to 10 spinal segments. Most often, the deformity is convex to the right. Some of these curves acquire dystrophic changes (ie, modulate) over time. Early MRI of all progressive curves is recommended to evaluate for spinal cord pathology and to examine apparent nondystrophic curves for dystrophic changes. With early advanced imaging, we believe that the modulation

556

CRAWFORD & HERRERA-SOTO

Fig. 1. (A, B) Cervicothoracic deformity in a patient who has neurobromatosis. (C, D) CT scan of the upper thoracic spine demonstrates rib head subluxation. (E, F) Postoperative lms following anterior cervicothoracic fusion performed through a sternal splitting trap door procedure and posterior spinal instrumentation and fusion.

of some nondystrophic to dystrophic deformities can be detected before typical dystrophic changes are seen on plain radiographs. Durrani and colleagues [28] reported that modulation occurred in about 65% of patients overall. Modulation occurred in 81% of patients who presented with scoliosis before 7 years of age and in 25% of those diagnosed after 7 years of age. In this study, rib penciling acquired through the modulation period was the only factor that inuenced progression of the deformity. The rate of progression for modulated scoliosis and kyphosis was 12 and 8 , respectively, versus 5 and 3 for nonmodulated spines. Nondystrophic curvatures of less than 25 should be observed. Curves between 25 and 40 can be treated with brace successfully [1,29]. Once beyond 40 , surgery by posterior spinal fusion is indicated (Fig. 2) [3]. Dystrophic scoliosis Dystrophic scoliosis is characterized by earlyonset, rapidly progressive curves that are dicult to treat [30,31]. The dystrophic curve is a shortsegmented, sharply angulated type that includes fewer than six spinal segments. It has a tendency to progress to a severe deformity. Dystrophic curves may be associated with kyphosis and have a higher incidence of neurologic injury [3,30]. Dystrophic vertebral changes develop over time; on plain lms, they are manifest by vertebral scalloping, rib penciling (rotation of the rib about 90 ), spindling of the transverse processes, severe apical rotation, severe vertebral wedging, and

wide nerve root foramina (see Table 1). Most of these patients present with associated neurobromas that envelope the bone or come from the canal (dumbbell lesion). Dystrophic curves are found most commonly in the thoracic region (Fig. 3) [1]. Enlargement of the spinal canal caused by intraspinal tumors or dural ectasia is common. Dural ectasia is an expansion of the thecal sac, probably due to increased hydrostatic pressure. It erodes the bony and ligamentous structures causing vertebral scalloping and meningocele formation [32]. Dystrophic curvatures of less than 20 should be observed for progression at 6-month intervals [13]. Bracing of progressive dystrophic curvatures is contraindicated simply because it has not been found to be eective [29,33,34]. For adolescent patients with dystrophic curvature greater than 20 to 40 of angulation, a posterior spinal fusion with segmental spinal instrumentation is recommended [13,34]. In more severe dystrophic scoliosis, anterior fusion often is performed in addition to posterior fusion, to increase the fusion rate and to reduce the risk for progression despite solid posterior fusion. Infantile, juvenile, and young adolescent patients who have dystrophic scoliosis represent challenging treatment groups. Posterior spinal fusion alone is now believed to be contraindicated in young patients who have progressive dystrophic deformities. A prevalence of pseudoarthrosis of up to 60% has been reported in the literature after posterior spinal fusion only [13,30,35,36]. Although some surgeons may perform posterior

SCOLIOSIS ASSOCIATED WITH NEUROFIBROMATOSIS

557

Fig. 2. This patient who has nondystrophic scoliosis shows several characteristics of possible early modulation, such as a high apex left curve that is progressing rapidly. Other osseous changes diagnostic of dystrophic curves, such as rib penciling and vertebral scalloping, are not present (A,B). Nondystrophic scoliosis can be treated similarly to idiopathic adolescent scoliosis, with posterior spinal instrumentation and fusion (C,D).

fusion with abundant autologous bone graft and pedicle screw instrumentation with good longterm results, the more predictable and successful procedure is anterior and posterior spinal fusion with segmental instrumentation and bone grafting [13,22,29,31,33,34,3740]. The anterior fusion can be performed by thoracotomy or thoracoscopy

[4143]. It is even appropriate to fuse the young child who has dystrophic scoliosis greater than 40 , based on the known natural history of certain progression for this curve type [1,13,34]. An early fusion causes minimal stunting of growth that is due to poor growth potential of the involved segments [1].

Fig. 3. This dystrophic deformity has a similar curve location to that illustrated in Fig. 2; however, the curves are much more severe, short, sharp, and associated with dystrophic osseous changes.

558

CRAWFORD & HERRERA-SOTO

In theory, the use of subcutaneous, so-called growing rods allows for further growth, although some have questioned their value on the basis of the small amount of growth achieved and the number of procedures required. The authors have used submuscular growing rods on earlyonset dystrophic curves with or without anterior annulotomies with a great deal of optimism. If the curve is less than 60 and exible, the rods are inserted directly with fusion of the cranial and caudal anchors only followed by lengthening every 6 months (Fig. 4). In large sti curves, anterior annulotomies without fusion (to preserve growth) have been performed. Our early results are extremely encouraging. This is a promising technique made especially useful because most dystrophic curves are early onset. Often, bracing following surgery is appropriate [44,45]. Bracing may need to be extended to the cervical region in cases of severe dysplastic curves that are instrumented into the upper thoracic and cervicothoracic region. Cervical bracing, halo

vest, or Minerva casting may help to prevent the possibility of screw/hook pullout. This is especially true for dysplastic curves that have low bone mineral density [4]. Preoperative halo traction may be benecial for the treatment of severe curves, including those with kyphoscoliosis [34,44,46,47]. It allows gradual and controlled soft tissue relaxation and curve correction before surgery or between staged surgeries; however, it is contraindicated in patients who have cervical kyphosis. Daily neurologic evaluations are mandatory to avoid spinal or cranial nerve injuries. When staged procedures are planned, interim halo-femoral traction has been used successfully [48]. Nutrition is also paramount during this time. We use supplemental nasojejunal feeding in between stages to decrease the protein depletion that is seen in staged patients [49,50]. During surgery, careful posterior subperiosteal dissection is performed by using Bovie electrocautery dissectiondrelying less on subperiosteal elevatorsdto reduce the risk for inadvertently

Fig. 4. (A) An early onset patient was treated unsuccessfully with brace-cast-brace treatment. (BD) Subcutaneous growing rods were placed to prevent further progression and to allow growth of the spine until denitive surgery could be performed closer to skeletal maturity (E).

SCOLIOSIS ASSOCIATED WITH NEUROFIBROMATOSIS

559

plunging through areas of laminar defects or hypoplasia and directly damaging the spinal cord. Anterior spinal dissection may be complicated by venous lakes and engorgement of saccular, almost sinusoidal, vessels, which are dicult to control in and around the vertebral bodies. Extensive blood loss from the blood vessels in the cancellous bone of the vertebral bodies is distinctly possible. We recommend a putty of Avitine, Gelfoam, and Thrombin to assist in achieving hemostasis. Surgicel and Flosseal also help with hemostasis. Discectomies should be performed with Bovie dissection and the use of a rongeur through the annulus brosis instead of sharp dissections of the endplate apophysis. Sharp dissection may cause signicant bleeding from the often friable cancellous matrix of the vertebral bodies. Care should be taken to preserve the segmental vessels as much as possible. If sectioning of segmental vessels is necessary, the authors recommend staging procedures to allow reperfusion of the anterior column. Lordosis Lordoscoliosis also has been identied and may represent a subset of dystrophic deformity. It is a poorly reported entity, but it may have signicant long-term implications. Often, there is associated posterior element erosion and dural ectasia [6,51]. Thoracic lordosis has been associated with decreased pulmonary function [52]. Kyphosis Kyphoscoliosis is dened as scoliosis accompanied by a kyphosis of greater than 50 . It may occur by gradual scoliotic rotation and progression or it can be found early in the disease with an abrupt angular kyphotic curve [51]. Kyphosis observed in individuals who have NF-1 is distinguished by acute anteroposterior angulation. Vertebral bodies may be deformed so severely that they are confused with congenital deformities. Occasionally, weakening of the spinal stabilizers (eg, facets, pedicles, and ligaments) by dural ectasia with meningocele formation gives rise to kyphosis with subluxation, and, in extreme cases, dislocation of the spine [53]. Even with severe hair pin turn angulation, neurologic status usually remains normal, and the spinal cord often is protected because the spinal canal is widened by dural ectasia. Traction may be dangerous when performed on rigid deformities because it increases tension

on the midapical spinal cord. Traction following anterior release is safe when monitored appropriately. For curves greater than 50 , anterior surgery (intervertebral discectomy, rib strut grafting, and bone chip grafting) is recommended, followed by posterior segmental instrumentation one or two levels above and below the end vertebrae [1,13,27,44]. For severe deformities (ie, O80 90 ), indenite bracing may be required until solid union is achieved, even after anterior and posterior spinal surgery. Severe kyphosis is the most common cause of neurologic decits in NF-1 [3]. Because of the association of paraplegia with kyphosis, physicians have tended to perform laminectomies. Laminectomy alone for kyphotic cord compression is absolutely contraindicated [3]. Posterior vertebral element resection does not relieve the anterior compression on the spinal cord, and, in fact, has not been shown to improve the condition [44]. The oending neurobromas usually are anterior, and decompression should be performed anteriorly. The removal of posterior element predisposes the spine to instability and there is loss of valuable bone stock required for fusion. Anterior decompression with posterior spinal fusion and instrumentation should be performed [44,54]. Spondylolisthesis Spondylolisthesis/spondyloptosis is a rare disorder that is associated most often with a pathologic luxation of the vertebra caused by erosions or elongation of the pedicles or pars from foraminal neurobroma or dural ectasia [1,12,13,5558]. This leads to forward progression of the anterior spinal column elements (Fig. 5) [59]. Surgery is indicated with the progression of the spondylolisthesis from grade II to grade III or with the presence of pain that cannot be controlled suciently by nonoperative means. Anterior and posterior stabilization is recommended for progressive deformity [55]. An alternative to instrumentation is to perform posterior spinal fusion and autologous bone grafting augmented with pantaloon casting. It is imperative to reassess the fusion mass approximately 6 months postoperatively and repeat bone grafting performed if necessary. Soft tissue abnormalities Subcutaneous plexiform neurobromas often overlie the incision area. When posterior spinal fusion surgery is performed, one should be aware

560

CRAWFORD & HERRERA-SOTO

Fig. 5. (A, B) NF-1 also can cause deformity in the low lumbar spine, such as scoliosis associated with severe spondylolisthesis. (C) CT sagittal reconstruction conrms grade IV spondylolisthesis, posterior vertebral body scalloping, and additional osseous changes. (D) 3-D reconstruction rendering illustrates dural ectasia in this region.

of the thin laminae that often are eroded by dural ectasia surrounding the spinal cord in the thoracic region. The laminae may be inadequate to accept hooks; pedicle screws may be necessary. Considerable bleeding may occur with dissection around subcutaneous vascular tumors. The authors recommend using monopolar and bipolar electrocautery for subperiosteal exposure of the posterior elements. Meningoceles, pseudomeningoceles, dural ectasia, and dumbbell lesions are related to the presence of neurobroma or abnormal pressure phenomena in and around the spinal canal neuraxis. High-volume myelography or MRI should be used in the investigation of all dystrophic curves before surgical treatment [3,32]. Occasionally, these intraspinal elements may compromise the cord directly when instrumentation and stabilization are attempted, or they may cause erosive changes in the bone, preventing primary fusion.

Pitfalls in spinal deformities in neurobromatosis type 1 There is a known increased prevalence of osteoporosis among patients who have NF-1 [4]. This can create problems in xation of dystrophic curves [29]. The bone is soft and nonresistant to manipulation.

Paraplegia is an uncommon nding in patients who have dystrophic curves [18,33]. It is more prevalent in patients who have severe vertebral angulation (kyphosis), vertebral subluxation, and soft tissue tumors in the spinal canal [18]. A rare, but important, cause of paraparesis in scoliotic patients is spinal cord compression due to rib penetration [6062]. It is imperative to evaluate for such a condition in the preoperative period. A CT scan is the most sensitive tool to diagnose intraspinal rib dislocation. A resection of the rib will prevent or improve paraparesis in most patients who have dislocation. Failure to recognize intraspinal lesions in patients who have neurobromatosis and undergo manipulation and instrumentation of the spine may result in neurologic compromise. Preoperative radiography, CT, or MRI is essential for patients who have dystrophic vertebral elements and curvatures requiring fusion. Be especially alert to the possibility of subluxated ribs in the spinal canal. It the surgeons responsibility to stabilize the spine with the most expedient, safe, and permanent method without causing neurologic injury [3].

References
[1] Crawford AH. Neurobromatosis. In: Weinstein SL, editor. The pediatric spine, principles and practice.

SCOLIOSIS ASSOCIATED WITH NEUROFIBROMATOSIS

561

[2] [3]

[4]

[5] [6]

[7] [8]

[9]

[10]

[11] [12]

[13]

[14]

[15]

[16]

[17]

[18]

[19]

[20]

2nd edition. Philadelphia: Lippincott Williams & Wilkins; 2001. p. 47190. Crawford AH, Schorry EK. Neurobromatosis update. J Pediatr Orthop 2006;26(3):41323. Crawford AH, Parikh S, Schorry EK, et al. The immature spine in type-1 neurobromatosis. J Bone Joint Surg Am 2007;89:12342. Lammert M, Kappler M, Mautner VF, et al. Decreased bone mineral density in patients with neurobromatosis 1. Osteoporos Int 2005;16(9):11616. Gabriel KR. Neurobromatosis. Curr Opin Pediatr 1997;9(1):8993. Crawford AH, Schorry EK. Neurobromatosis in children: the role of the orthopaedist. J Am Acad Orthop Surg 1999;7(4):21730. National Institute of Health Consensus Development Conference. NF-1. 1988. p. 1728. Rouleau GA, Wertelecki W, Haines JL, et al. Genetic linkage of bilateral acoustic neurobromatosis to a DNA marker on chromosome 22. Nature 1987; 329(6136):2468. Mautner VF, Tatagiba M, Lindenau M, et al. Spinal tumors in patients with neurobromatosis type 2: MR imaging study of frequency, multiplicity, and variety. AJR Am J Roentgenol 1995;165(4):9515. Vitale MG, Guha A, Skaggs DL. Orthopaedic manifestations of neurobromatosis in children: an update. Clin Orthop Relat Res 2002;401:10718. Crawford AH. Neurobromatosis in children. Acta Orthop Scand Suppl 1986;218:160. Crawford AH Jr, Bagamery N. Osseous manifestations of neurobromatosis in childhood. J Pediatr Orthop 1986;6(1):7288. Crawford AH. Pitfalls of spinal deformities associated with neurobromatosis in children. Clin Orthop Relat Res 1989;245:2942. Akbarnia BA, Gabriel KR, Beckman E, et al. Prevalence of scoliosis in neurobromatosis. Spine 1992; 17(8 Suppl):S2448. Haddad FS, Williams RL, Bentley G. The cervical spine in neurobromatosis. Br J Hosp Med 1995; 53(7):3189. Stone JW, Bridwell KH, Shackelford GD, et al. Dural ectasia associated with spontaneous dislocation of the upper part of the thoracic spine in neurobromatosis. A case report and review of the literature. J Bone Joint Surg Am 1987;69(7):107983. Yong-Hing K, Kalamchi A, MacEwen GD. Cervical spine abnormalities in neurobromatosis. J Bone Joint Surg Am 1979;61(5):6959. Curtis BH, Fisher RL, Buttereld WL, et al. Neurobromatosis with paraplegia. Report of eight cases. J Bone Joint Surg Am 1969;51(5):84361. Isu T, Miyasaka K, Abe H, et al. Atlantoaxial dislocation associated with neurobromatosis. Report of three cases. J Neurosurg 1983;58(3):4513. Garg S, Hosalkar H, Dormans JP. Quadriplegia in a 10 year-old boy due to multiple cervical neurobromas. Spine 2003;28(17):E33943.

[21] Yonezawa I, Arai Y, Tsuji T, et al. Anterior fusion and posterior correction of severe cervical kyphosis using pedicle screw xation in a patient with neurobromatosis: a case report. J Spinal Disord Tech 2003;16(5):4936. [22] Mulpuri K, LeBlanc JG, Reilly CW, et al. Sternal split approach to the cervicothoracic junction in children. Spine 2005;30(11):E30510. [23] Nazzaro JM, Arbit E, Burt M. Trap door exposure of the cervicothoracic junction. Technical note. J Neurosurg 1994;80(2):33841. [24] Weiss RA. A curvature of the spine in von Recklinhausens disease, von Recklinghausens disease in a Negro. Arch Dermatol Syph 1921;iii:14451. [25] Gould EP. The bone changes occurring in von Recklinhausens disease. Q J Med 1918;11:2218. [26] Calvert PT, Edgar MA, Webb PJ. Scoliosis in neurobromatosis. The natural history with and without operation. J Bone Joint Surg Br 1989;71(2):24651. [27] Wilde PH, Upadhyay SS, Leong JC. Deterioration of operative correction in dystrophic spinal neurobromatosis. Spine 1994;19(11):126470. [28] Durrani AA, Crawford AH, Chouhdry SN, et al. Modulation of spinal deformities in patients with neurobromatosis type 1. Spine 2000;25(1):6975. [29] Kim HW, Weinstein SL. Spine update. The management of scoliosis in neurobromatosis. Spine 1997; 22(23):27706. [30] Sirois JL 3rd, Drennan JC. Dystrophic spinal deformity in neurobromatosis. J Pediatr Orthop 1990; 10(4):5226. [31] Holt RT, Johnson JR. Cotrel-Dubousset instrumentation in neurobromatosis spine curves. A preliminary report. Clin Orthop Relat Res 1989;245:1923. [32] Tsirikos AI, Saifuddin A, Noordeen MH. Spinal deformity in neurobromatosis type-1: diagnosis and treatment. Eur Spine J 2005;14(5):42739. [33] Winter RB, Moe JH, Bradford DS, et al. Spine deformity in neurobromatosis. A review of one hundred and two patients. J Bone Joint Surg Am 1979; 61(5):67794. [34] Halmai V, Doman I, de Jonge T, et al. Surgical treatment of spinal deformities associated with neurobromatosis type 1. Report of 12 cases. J Neurosurg 2002;97(3 Suppl):3106. [35] Betz RR, Iorio R, Lombardi AV, et al. Scoliosis surgery in neurobromatosis. Clin Orthop Relat Res 1989;245:536. [36] Shen JX, Qiu GX, Wang YP, et al. Surgical treatment of scoliosis caused by neurobromatosis type 1. Chin Med Sci J 2005;20(2):8892. [37] Singh K, Samartzis D, An HS. Neurobromatosis type I with severe dystrophic kyphoscoliosis and its operative management via a simultaneous anteriorposterior approach: a case report and review of the literature. Spine J 2005;5(4):4616. [38] Hopf CG, Eysel P. One-stage versus two-stage spinal fusion in neuromuscular scolioses. J Pediatr Orthop B 2000;9(4):23443.

562

CRAWFORD & HERRERA-SOTO

[39] Parisini P, Di Silvestre M, Greggi T, et al. Surgical correction of dystrophic spinal curves in neurobromatosis. A review of 56 patients. Spine 1999;24(21): 224753. [40] Al-Sayyad MJ, Crawford AH, Wolf RK. Early experiences with video-assisted thoracoscopic surgery: our rst 70 cases. Spine 2004;29(17):194551 [discussion: 52]. [41] Al-Sayyad MJ, Crawford AH, Wolf RK. Videoassisted thoracoscopic surgery: the Cincinnati experience. Clin Orthop Relat Res 2005;434:6170. [42] Sucato DJ, Elerson E. A comparison between the prone and lateral position for performing a thoracoscopic anterior release and fusion for pediatric spinal deformity. Spine 2003;28(18):217680. [43] Stagnara P, Biot B, Fauchet R. [Critical evaluation of the surgical treatment of vertebral lesions due to neurobromatosis. 31 cases (19541973)]. Rev Chir Orthop Reparatrice Appar Mot 1975;61(1):1738 [in French]. [44] Winter RB, Lonstein JE, Anderson M. Neurobromatosis hyperkyphosis: a review of 33 patients with kyphosis of 80 degrees or greater. J Spinal Disord 1988;1(1):3949. [45] Sink EL, Karol LA, Sanders J, et al. Ecacy of perioperative halo-gravity traction in the treatment of severe scoliosis in children. J Pediatr Orthop 2001; 21(4):51924. [46] Rinella A, Lenke L, Whitaker C, et al. Perioperative halo-gravity traction in the treatment of severe scoliosis and kyphosis. Spine 2005;30(4):47582. [47] Mehlman CT, Al-Sayyad MJ, Crawford AH. Eectiveness of spinal release and halo-femoral traction in the management of severe spinal deformity. J Pediatr Orthop 2004;24(6):66773. [48] Mooney JF 3rd. Perioperative enteric nutritional supplementation in pediatric patients with neuromuscular scoliosis. J South Orthop Assoc 2000; 9(3):2026. [49] Lalueza MP, Colomina MJ, Bago J, et al. Analysis of nutritional parameters in idiopathic scoliosis patients after major spinal surgery. Eur J Clin Nutr 2005;59(5):7202. [50] Funasaki H, Winter RB, Lonstein JB, et al. Pathophysiology of spinal deformities in

[51]

[52]

[53]

[54]

[55]

[56]

[57]

[58] [59]

[60]

[61]

[62]

neurobromatosis. An analysis of seventy-one patients who had curves associated with dystrophic changes. J Bone Joint Surg Am 1994;76(5):692700. Winter RB, Lovell WW, Moe JH. Excessive thoracic lordosis and loss of pulmonary function in patients with idiopathic scoliosis. J Bone Joint Surg Am 1975;57(7):9727. Winter RB. Spontaneous dislocation of a vertebra in a patient who had neurobromatosis. Report of a case with dural ectasia. J Bone Joint Surg Am 1991;73(9):14024. Shuebarger HL. Cotrel-Dubousset instrumentation in neurobromatosis spinal problems. Clin Orthop Relat Res 1989;245:248. Toyoda K, Taguchi T, Kaneko K, et al. High-grade L5 spondylolisthesis associated with dural ectasia in neurobromatosis. J Orthop Sci 2005;10(2):2336. Bensaid AH, Dietemann JL, Kastler B, et al. Neurobromatosis with dural ectasia and bilateral symmetrical pedicular clefts: report of two cases. Neuroradiology 1992;34(2):1079. Wong-Chung J, Gillespie R. Lumbosacral spondyloptosis with neurobromatosis. Case report. Spine 1991;16(8):9868. Winter RB, Edwards WC. Case report. Neurobromatosis with lumbosacral spondylolisthesis. J Pediatr Orthop 1981;1(1):916. Mandell GA. The pedicle in neurobromatosis. AJR Am J Roentgenol 1978;130(4):6758. Khoshhal KI, Ellis RD. Paraparesis after posterior spinal fusion in neurobromatosis secondary to rib displacement: case report and literature review. J Pediatr Orthop 2000;20(6):799801. Deguchi M, Kawakami N, Saito H, et al. Paraparesis after rib penetration of the spinal canal in neurobromatous scoliosis. J Spinal Disord 1995;8(5): 3637. Gkiokas A, Hadzimichalis S, Vasiliadis E, et al. Painful rib hump: a new clinical sign for detecting intraspinal rib displacement in scoliosis due to neurobromatosis. Scoliosis 2006;1:14. Major MR, Huizenga BA. Spinal cord compression by displaced ribs in neurobromatosis. A report of three cases. J Bone Joint Surg Am 1988;70(7): 11002.

Orthop Clin N Am 38 (2007) 563572

Spinal Deformities in Marfan Syndrome


Constantine A. Demetracopoulos, MD, Paul D. Sponseller, MD*
Department of Orthopaedic Surgery, Johns Hopkins Medical Institutions, The Johns Hopkins Hospital, 601 North Caroline Street, Room 5253, Baltimore, MD 21287, USA

Marfan syndrome (MFS) is an autosomal dominant disorder of connective tissue that displays complete penetrance but variable expression [1]. Genetic studies localized the FBN1 gene responsible for MFS disease to chromosome 15q21 [1,2]. More than 135 mutations in the brillin gene have been identied, and 70% of aected individuals have an identiable FBN1 mutation [3]. Patients who have Marfan disorder are believed to have a defect in the synthesis, secretion, or incorporation of brillin, a glycoprotein that is a major component of various types of connective tissue, including bone [3]. Thus, MFS is a multisystem disease, and manifestations are seen most commonly in the ocular, cardiovascular, and skeletal systems. The prevalence of MFS is estimated at two to three cases per 10,000 individuals [4]. Approximately 15% to 25% of patients do not have a family history of MFS and may represent de novo mutations [3,4]. MFS is a pleiotropic condition, and ndings tend to become more apparent with age. One end of the spectrum represents patients who are aected only mildly, whereas the other end represents patients who have a severe neonatal form who may encounter cardiovascular complications during the rst year of life. The most apparent clinical ndings in MFS involve the skeleton (arachnodactyly, scoliosis and thoracic lordosis, dolichostenomelia, sternal deformities, and joint laxity). Although these ndings are not specic when found individually, the presence of multiple ndings should prompt a referral to genetics or the echocardiography laboratory.

Diagnosis Although genetic testing is available, it is complex and not highly sensitive. As a result, MFS remains a clinical diagnosis. According to the diagnostic (Ghent) criteria, the diagnosis can be made with the presence of at least one major criterion in two organ systems and the involvement of a third system or one major criterion in an organ system and the involvement of a second system when there is a positive family history of MFS or documentation of a FBN1 mutation (Box 1) [5]. Major criteria are named as such because they are infrequent in other conditions and carry high diagnostic specicity when present. The dierential diagnosis includes homocystinuria, congenital contractual arachnodactyly, Stickler syndrome, Ehlers-Danlos syndrome, MASS (mitral, aortic, skin, and skeletal manifestations) phenotype, Loeys-Dietz syndrome, and SchprintzenGoldberg syndrome [6]. Early diagnosis is of the utmost importance to initiate prophylactic bblockade therapy, which has been shown to be effective in slowing the rate of aortic dilatation and reducing the development of aortic complications in patients who have MFS [7]. Vertebral morphology The classic Marfan spine features may include increased vertebral scalloping, a higher prevalence of lumbosacral transitional vertebrae, lengthened transverse process distance, and a reduction in pedicle width and laminar thickness [8,9]. Mean pedicle widths from L1L3 were smaller than the smallest available pedicle screw (5 mm) in a study of 32 patients who had MFS. In addition, laminar thickness from L1L5 was signicantly less in patients who had MFS than
orthopedic.theclinics.com

* Corresponding author. E-mail address: psponse@jhmi.edu (P.D. Sponseller).

0030-5898/07/$ - see front matter 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.ocl.2007.04.003

564

DEMETRACOPOULOS & SPONSELLER

Box 1. The Ghent diagnostic criteria for Marfan syndrome Skeletal system Major criteria: presence of at least four of the following: Pectus carinatum Pectus excavatum requiring surgery Reduced upper to lower segment ratio or arm span to height ratio>1.05 Wrist sign (thumb and fth digit overlap circling the wrist) and thumb sign (distal phalanx protrudes beyond border of clenched st) Reduced extension at the elbows (<170 ) Scoliosis>20 Pes planus Protrusio acetabuli of any degree (ascertained on radiographs) Minor criteria Pectus excavatum Joint hypermobility High arched palate Facial Dolichocephaly Malar hypoplasia Enophthalmos Retrognathia Down-slanting palpebral ssures Ocular system Major criterion Ectopia lentis Minor criteria Flat cornea Increased axial length of the globe (>23.5 mm) Cardiovascular system Major criteria Dilatation of the ascending aorta Dissection of the ascending aorta Minor criteria Mitral valve prolapse with or without mitral valve regurgitation Dilatation of the main pulmonary artery in the absence of valvular or peripheral pulmonic stenosis (age <40) Calcication of the mitral annulus before the age of 40 years

Dilatation or dissection of the descending thoracic or abdominal aorta before the age of 50 years Pulmonary system Minor criteria (only) Spontaneous pneumothorax Apical blebs Skin and integumentary system Minor criteria (only) Striae atrophicae Recurrent or incisional hernia Neurologic system Major criterion Lumbosacral dural ectasia by CT or MRI Family/genetic history Major criteria First-degree relative who independently meets the diagnostic criteria Presence of mutation in FBN1 gene Presence of haplotype around FBN1 inherited by descent and unequivocally associated with diagnosed MFS in the family The index case requires major criteria in two different organ systems and involvement of a third. If the FBN-1 mutation is present, then one major criterion and involvement of a second organ system sufces. A relative of the index case requires one major criterion in family history and one major criterion in an organ system and involvement of a second organ system.

in normal controls [9]. Many patients also have signicant scalloping of the sacrum inside the spinal canal (Fig. 1). Therefore, careful preoperative planning of xation anchors, possibly including CT, may be necessary.

Scoliosis and kyphosis Because the spine depends upon a balance of forces during growth, conditions that disrupt the musculoskeletal matrix will cause spinal deformity. In a large cross-sectional study, the

SPINAL DEFORMITIES IN MARFAN SYNDROME

565

Fig. 1. Osseous scalloping in MFS.

prevalence of scoliosis was 63% in MFS [10]; however, many of the curves were minor, and only 10% to 20% required treatment of any kind. The curve patterns were similar to those seen in idiopathic scoliosis, with the most common types being thoracolumbar and thoracic, although there was a higher rate of triple curves [10]. Dierences were seen in the sagittal plane, however (see later discussion). Scoliosis progresses at a faster rate in MFS than in the general population in all age groups, with the greatest progression occurring in patients who are 3 years old or younger (19 17 per year) [10]. Virtually all curves greater than 30 in immature patients reach at least 40 at maturity. Curves greater than 40 increase during adulthood, at a slightly higher rate than idiopathic scoliosis. Curves greater than 50 progress at a mean rate of 3 4 per year in adulthood [10]. Patients who had MFS and scoliosis had more back pain than those who did not have scoliosis. Brace treatment does not seem to have as much success in MFS as in idiopathic scoliosis. One

study reported a 17% success rate for bracing scoliosis curves in MFS. Thus, bracing is recommended mainly for growing children with curves in the range of 15 to 25 . Patients with curves of 25 to 45 may be oered the option of using the brace; however, patients and physicians should be aware that there is only a one in ve chance of controlling the curve successfully. Bracing is not recommended for curves greater than 40 , except as a temporizing option; the curve will likely increase [11]. Infantile scoliosis is a special challenge to treat in MFS. Bracing has a limited role, especially in smaller curves. It is mainly useful in promoting upright posture in patients with a coexistent kyphosis. Also, surgery is especially challenging for patients younger than 4 years of age. In addition to typical diculties encountered when operating on the young child for scoliosis, because of thin body habitus and a paucity of local, durable soft tissues, instrumentation is hard to cover in young patients who have MFS. The use of a growing rod using iliac xation distally and strong proximal anchorage has been successful in these cases (Fig. 2). Eventual fusion will be necessary. Sagittal plane deformities also are common in MFS. The mean kyphosis in the population that has MFS is greater than in the general population. Forty percent of patients who have MFS have a kyphosis greater than 50 [10]. Dierent types of sagittal alignment were found: hypokyphosis, hyperkyphosis, and thoracolumbar kyphosis with compensatory thoracic lordosis [10]. Sponseller and colleagues [10] proposed a classication

Fig. 2. Infantile scoliosis in MFS. (A) Preoperative anteroposterior (AP) radiograph. (B) Four-year postoperative AP radiograph. (C) Preoperative lateral radiograph (D) Four-year postoperative lateral radiograph.

566

DEMETRACOPOULOS & SPONSELLER

system for the sagittal deformity associated with MFS (Fig. 3). Dural ectasia Dural ectasia (DE), a widening of the dural sac and nerve root sleeves in the caudal portion of the spine, is rare in the normal population but it has a high prevalence (56% to 92%) in the adult population that has MFS (Fig. 4) [1214]. The dural sac is enlarged mainly below the level of L5 [1318]. Among 73 patients who were diagnosed with MFS, Rose and colleagues [19] found DE

to be the second most common major manifestation. DE worsens with age, and a recent study on a murine model of MFS suggests that DE is the result of hydrostatic pressure exerted upon an inherently weakened dura [20]. In children, symptoms may be mild or not present at all. In a prospective patient-control study of patients between 4 and 18 years of age, Knirsch and colleagues [21] found MRI evidence suggestive of DE in up to 40% of children. The consequences of DE include bony erosion that may lead to anterior meningocele or posterior meningocele. Clinically, patients who have DE may present with headache, proximal leg pain, leg weakness and numbness, abdominal pain, and genital and rectal pain [22]. DE also has been associated with moderate to severe back pain; however, as many as 41% of patients who have DE have no back pain at all [23]. Sometimes DE is completely asymptomatic [2330]. There are several denitions of DE. Fattori and colleagues [31] used morphologic criteria consisting of bulging of the dural sac, lack of epidural fat at the level of the posterior wall of at least one posterior vertebral body, and the presence of radicular cysts on MRI to grade the degree of dural abnormality. Villeirs and colleagues [32] used CT of the spine to establish a quantitative method of assessing for DE by dening the normal ratio of the transverse diameter of the dural sac to the

Fig. 3. Types of sagittal proles in patients who have MFS. (A) In type I sagittal alignment, the transition zone between the kyphotic and lordotic segments is at or above L2. Subtype IA corresponds to normal kyphosis and lordosis, subtype IB describes thoracic kyphosis less than 20 , and subtype IC refers to thoracic kyphosis greater than 50 . (B) In type II sagittal alignment, the transition zone is caudad to L2. Subtype IIA refers to a thoracolumbar kyphosis that extends caudad to L2; subtype IIB indicates a reversal of normal contours with thoracic lordosis, thoracolumbar kyphosis, and low lumbar lordosis. (From Sponseller PD, Hobbs W, Riley LH 3rd, et al. The thoracolumbar spine in Marfan syndrome. J Bone Joint Surg Am 1995;77(6):868; with permission.)

Fig. 4. DE in MFS.

SPINAL DEFORMITIES IN MARFAN SYNDROME

567

transverse diameter of each lumbar vertebral body. Oosterhof and colleagues [33] used MRI to calculate dural sac ratios in the anteroposterior plane; they found that abnormal dural sac ratios at L3 and S1 could be used to identify MFS with 95% sensitivity and 98% specicity. Ahn and colleagues [15] suggested that the gold standard for the diagnosis of DE is a dural volume greater than 7 cm3 when measured caudal to the inferior endplate of L5; however, because techniques to calculate dural volume from reconstructed sacral spine MRIs are not widely available, Ahn and colleagues [15] used MRI and CT to identify two major criteria and two minor criteria indicative of DE in patients who have MFS. Major criteria included a sagittal diameter of the spinal canal at S1 or below greater than the sagittal midaxis of L4 or above and the presence of an anterior meningocele. Minor criteria included scalloping at S1 and an enlarged L5 nerve root sleeve diameter. When at least one major or two minor criteria were present, DE was diagnosed with a sensitivity of 87% and a specicity of 95% using MRI, and a sensitivity of 83% and

a specicity of 90% using CT. In a separate study, Ahn and colleagues [34] examined the eectiveness of conventional radiography to detect DE in patients who had MFS. They determined that the presence of an interpediculate distance at L4 of at least 38.0 mm, a sagittal canal diameter at S1 of at least 18.0 mm, or a scalloping value at L5 of at least 5.5 mm can detect DE with a specicity of 91.7% and a sensitivity of 57.1%. Weigang and colleagues [35] and others have evaluated the ecacy of the MRI and CT criteria proposed for the diagnosis of DE (Table 1). Weigang and colleagues [35] found only two patients who had MFS who satised all criteria for DE using the three methods proposed by Villeirs and colleagues, Oosterhof and colleagues, and Ahn and colleagues in a group of 18 individuals who had MFS and 23 individuals who did not have MFS. Moreover, DE was found in up to 44% of normal controls. Because a gold standard was not used, the sensitivity and specicity of each method cannot be calculated from this study; however, it suggests that the various methods available to assess for DE can produce varying

Table 1 Comparison of four methods for detecting dural ectasia on MRI in individuals who do and do not have Marfan syndrome Method Oosterhof et al [33] Villeirs et al [32] Fattori et al [31] Criterion Increased dural sac ratioa at L3 and S1 Increased spinal canal indexb or dural sac indexc Presence of a rectilinear dural sac, bulging of the dural sac, lack of epidural fat at the posterior wall of vertebral bodies, radicular cysts, or anterior sacral meningocele Sagittal diameter of the spinal canal at or below S1 is greater than the sagittal diameter at the midaxis of L4 or above MFS 94% 18% 92% Non-MFS 44% 0% 0%

Ahn et al [15]

72%

44%

a Dural sac ratio is the ratio of the anteroposterior (AP) diameter of the dural sac to the AP diameter of the vertebral body. b Spinal canal index is the sum of the transverse diameters of the spinal canal measured from L1S1 divided by the sum of the transverse diameters of the vertebral bodies from L1L4. c Dural sac index is the sum of the transverse diameters of the dural sac measured from L1S1 divided by the sum of the transverse diameters of the vertebral bodies from L1L4. Data from Weigang E, Ghanem N, Chang XC, et al. Evaluation of three dierent measurement methods for dural ectasia in Marfan syndrome. Clin Radiol 2006;61(11):976.

568

DEMETRACOPOULOS & SPONSELLER

results, which makes the diagnosis of DE challenging. In a similar study, Habermann and colleagues [36] evaluated the available MRI criteria for the diagnosis of DE in children and adolescents. They determined that abnormal ratios of the midsagittal diameter of the vertebral bodies to the sagittal dural sac diameters at L5 and S1 and a sagittal dural sac width at S1 greater than at L4 were the only statistically signicant criteria for DE when children and adolescents who had MFS were compared with normal controls.

Surgical correction of kyphoscoliosis Spine surgery is required in about 10% to 15% of patients (Fig. 5) [10,37]. A medical specialist

should be involved to comanage cardiac medications and other issues and to assist with anticoagulation management if there is an articial heart valve. An MRI or CT should be done of the area of interest before any spine surgery to visualize the bony and dural structures. Intraoperative complications include increased bleeding and cerebrospinal uid leak. A recent retrospective review of patients who had MFS who underwent surgical correction for spinal deformity found that patients who underwent posterior fusion for scoliosis had an average blood loss of 2150 mL, whereas patients who required revision had an average blood loss of 5300 mL [38]. We have found aprotinin to be helpful in decreasing blood loss. Cerebrospinal uid leak occurs because of the DE, and dural tears occurred in 8% of patients

Fig. 5. Severe kyphoscoliosis in a 16-year-old who has MFS. (A) Preoperative clinical appearance. (B) Lateral radiograph shows a 95 scoliosis. (C) CT shows bridging of vertebrae. (D) Clinical coronal appearance. (E) Coronal lm showing a 95 kyphosis. The patient underwent apical wedge resections and fusions from a posterior approach. (F) Postoperative sagittal appearance. Postoperative lateral (G) and coronal (H) radiographs.

SPINAL DEFORMITIES IN MARFAN SYNDROME

569

Fig. 6. Adding-on deformity after correction of scoliosis in a 12-year old.

undergoing posterior fusion [38,39]. It can be minimized by keeping the patient in a slight Trendelenburg position intraoperatively and minimizing dissection inside the canal. If a leak occurs,

closure is possible sometimes; if not, a dural patch and postoperative bed rest for 2 to 3 days have been successful. Postoperative complications include failure of xation, adding on of curvature, and pseudarthrosis [37,40]. In a review of 39 patients who had MFS and underwent surgical correction, Jones and colleagues [38] reported a xation failure rate of 21% and a pseudarthrosis rate of 10%. When correcting spinal deformity, it is important not to fuse too short a segment of spine (Fig. 6). If patients have dysplastic vertebrae, fusions that need to extend to L4 may be better o carried to the sacrum to prevent late pain and decompensation (Fig. 7). The surgeon probably should not use selective thoracic fusion for double major curves unless the lumbar curve is minor. Lipton and colleagues [41] analyzed data of patients who had MFS and underwent posterior spinal arthrodesis. They compared postoperative curve progression between patients who had primary and secondary curves fused completely and patients who had complete fusion of the primary curve and partial fusion of the secondary curve. Preoperative curve characteristics were not compared between groups; therefore, it is unclear whether any dierences existed between the two groups preoperatively. Postoperatively, none of the 7 patients who had both curves fused progressed more than 10 at last follow-up; however, 11 of the 16 patients who had partial fusion of the secondary curve progressed more than 10 at last follow-up. It also is important to take into account any unusual kyphosis that may be

Fig. 7. Fusion in a patient who has MFS with dysplastic vertebrae that was carried to the sacrum to prevent late pain and decompensation. (A) Preoperative anteroposterior (AP) radiograph. (B) Postoperative AP coned view.

570

DEMETRACOPOULOS & SPONSELLER

present. Failure of xation is common and is due to ligamentous laxity that allows hooks to dislodge from laminae. In a review of 23 patients who had MFS and underwent surgery for spinal deformity, Di Silvestre and colleagues [39] found a distal hook dislodgement rate of 22%. Problems also may occur with pedicle screw xation, with screws at risk for losing xation because of thin pedicles. The surgeon should critically assess the xation and work to achieve a stable proximal and distal foundation in every patient. Spondylolisthesis The prevalence of spondylolisthesis in the general population is approximately 3%, and, of this population, the mean slip is approximately 15% [42]. In one study, 6% of the population that had MFS had spondylolisthesis of the fth lumbar vertebra or the rst sacral vertebra with a mean slip of 30% [10]. It is at this level where DE, sacral scalloping, and potential cerebrospinal uid leaks have the greatest clinical importance. Although the frequency of spondylolisthesis in patients who have MFS may not be markedly higher than in the general population, if a slip is present, the altered tissue properties allow greater forward slip to occur. In patients who have MFS and scoliosis of at least 10 , mean slip was 60% [10]. Cervical spine Hobbs and colleagues [43] conducted a radiographic analysis of the cervical spine in patients who had MFS, which revealed an increased prevalence of focal kyphosis and a slightly increased atlantoaxial movement with exion and extension. Herzka and colleagues [44] reported three cases that suggested that children who have MFS might be at risk for atlantoaxial rotatory subluxation when muscle tone is attenuated by general anesthesia or muscle relaxants. Special attention to intubation and positioning, intraoperatively and postoperatively, may be warranted, and rotatory subluxation should be included in the differential diagnosis for patients who have MFS and neck pain. Bone density Abnormal brillin in the population that has MFS may play a role in bone mineralization. Carter and colleagues [45] demonstrated that

there is a reduced axial bone mineral density (BMD) in men and women who have MFS; they postulated that this may be due to mutations of the brillin gene or environmental issues, such as reduced exercise that leads to suboptimal peak bone mass. Giampietro and colleagues [46] reviewed 51 patients who had MFS and found that the average BMD in the femoral necks of adult male patients was signicantly less than in normal controls; however, average BMD measurements for female and adolescent patients who had MFS were within normal limits. Therefore, more research may be necessary before implementing widespread BMD evaluations by Dual Energy X-Ray Absorptiometry scans.

Screening and athletics It is recommended that patients who have MFS avoid physical stress, because many aortic dissections occur during these events [4749]. Pyeritz and Dietz [47] recommended that the physician counsel patients on an individual basis. For example, a child with only a slightly dilated aortic root does not need outright restriction but should be counseled away from competitive athletics. Conversely, an older patient with a dilated aortic root should be advised against any sort of strenuous exertion, especially activities with sudden stops (eg, basketball). Furthermore, reports of increased atlantoaxial translation in MFS suggest they should avoid contact sports, which place high-impact loads on the cervical spine. Other activities, such as low-intensity isokinetic sports, should be encouraged.

Summary MFS is a disease of connective tissue that aects multiple organ systems. Skeletal manifestations are common at an early age and oer clinicians the opportunity to make a timely diagnosis. Thus, the surgeon examining the spine must be aware of the manifestations of MFS and be able to recognize aected individuals. Spinal deformities are frequent, progress with age, and are less responsive to conservative treatment. Surgical management presents unique intraoperative and postoperative challenges because complication and revision rates are greater in patients who have MFS; however, with attention to detail, excellent results are possible.

SPINAL DEFORMITIES IN MARFAN SYNDROME

571

References
[1] Dietz HC, Pyeritz RE. Mutations in the human gene for brillin-1 (FBN1) in the Marfan syndrome and related disorders. Hum Mol Genet 1995;4 Spec No:1799809. [2] Tsipouras P, Del Mastro R, Sarfarazi M, et al. Genetic linkage of the Marfan syndrome, ectopia lentis, and congenital contractural arachnodactyly to the brillin genes on chromosomes 15 and 5. The International Marfan Syndrome Collaborative Study. N Engl J Med 1992;326(14):9059. [3] Giampietro PF, Raggio C, Davis JG. Marfan syndrome: orthopedic and genetic review. Curr Opin Pediatr 2002;14(1):3541. [4] Judge DP, Dietz HC. Marfans syndrome. Lancet 2005;366(9501):196576. [5] De Paepe A, Devereux RB, Dietz HC, et al. Revised diagnostic criteria for the Marfan syndrome. Am J Med Genet 1996;62(4):41726. [6] Beighton P, de Paepe A, Danks D, et al. International nosology of heritable disorders of connective tissue, Berlin, 1986. Am J Med Genet 1988;29(3): 58194. [7] Shores J, Berger KR, Murphy EA, et al. Progression of aortic dilatation and the benet of long-term betaadrenergic blockade in Marfans syndrome. N Engl J Med 1994;330(19):133541. [8] Tallroth K, Malmivaara A, Laitinen ML, et al. Lumbar spine in Marfan syndrome. Skeletal Radiol 1995;24(5):33740. [9] Sponseller PD, Ahn NU, Ahn UM, et al. Osseous anatomy of the lumbosacral spine in Marfan syndrome. Spine 2000;25(21):2797802. [10] Sponseller PD, Hobbs W, Riley LH 3rd, et al. The thoracolumbar spine in Marfan syndrome. J Bone Joint Surg Am 1995;77(6):86776. [11] Sponseller PD, Bhimani M, Solaco D, et al. Results of brace treatment of scoliosis in Marfan syndrome. Spine 2000;25(18):23504. [12] Pyeritz RE, Fishman EK, Bernhardt BA, et al. Dural ectasia: a common pleiotropic feature of the Marfans syndrome. Am J Hum Genet 1986;39S: A75. [13] Fishman EK, Zinreich SJ, Kumar AJ, et al. Sacral abnormalities in Marfan syndrome. J Comput Assist Tomogr 1983;7(5):8516. [14] Pyeritz RE, Fishman EK, Bernhardt BA, et al. Dural ectasia is a common feature of the Marfan syndrome. Am J Hum Genet 1988;43(5):72632. [15] Ahn NU, Sponseller PD, Ahn UM, et al. Dural ectasia in the Marfan syndrome: MR and CT ndings and criteria. Genet Med 2000;2(3):1739. [16] Robinson L, Dominguez R, Cabrera J, et al. Multiple meningeal cysts in Marfan syndrome. AJNR Am J Neuroradiol 1989;10(6):12756. [17] Savini R, Cervellati S, Beroaldo E. Spinal deformities in Marfans syndrome. Ital J Orthop Traumatol 1980;6(1):1940.

[18] Strand RD, Eisenberg HM. Anterior sacral meningocele in association with Marfans syndrome. Radiology 1971;99(3):6534. [19] Rose PS, Levy HP, Ahn NU, et al. A comparison of the Berlin and Ghent nosologies and the inuence of dural ectasia in the diagnosis of Marfan syndrome. Genet Med 2000;2(5):27882. [20] Jones KB, Myers L, Judge DP, et al. Toward an understanding of dural ectasia: a light microscopy study in a murine model of Marfan syndrome. Spine 2005;30(3):2913. [21] Knirsch W, Kurtz C, Haner N, et al. Dural ectasia in children with Marfan syndrome: a prospective, multicenter, patient-control study. Am J Med Genet A 2006;140(7):77581. [22] Foran JR, Pyeritz RE, Dietz HC, et al. Characterization of the symptoms associated with dural ectasia in the Marfan patient. Am J Med Genet A 2005;134(1): 5865. [23] Ahn NU, Sponseller PD, Ahn UM, et al. Dural ectasia is associated with back pain in Marfan syndrome. Spine 2000;25(12):15628. [24] Duncan RW, Esses S. Marfan syndrome with back pain secondary to pedicular attenuation. A case report. Spine 1995;20(10):11978. [25] Ho NC, Hadley DW, Jain PK, et al. Case 47: dural ectasia associated with Marfan syndrome. Radiology 2002;223(3):76771. [26] Nallamshetty L, Ahn NU, Ahn UM, et al. Dural ectasia and back pain: review of the literature and case report. J Spinal Disord Tech 2002;15(4):3269. [27] Raftopoulos C, Delecluse F, Braude P, et al. Anterior sacral meningocele and Marfan syndrome: a review. Acta Chir Belg 1993;93(1):17. [28] Smith MD. Large sacral dural defect in Marfan syndrome. A case report. J Bone Joint Surg Am 1993; 75(7):106770. [29] Sonier CB, Buhe T, Despins P, et al. Sacral pseudomeningocele and Marfans disease. One case. J Neuroradiol 1993;20(4):2926. [30] Stern WE. Dural ectasia and the Marfan syndrome. J Neurosurg 1988;69(2):2217. [31] Fattori R, Nienaber CA, Descovich B, et al. Importance of dural ectasia in phenotypic assessment of Marfans syndrome. Lancet 1999;354(9182):9103. [32] Villeirs GM, Van Tongerloo AJ, Verstraete KL, et al. Widening of the spinal canal and dural ectasia in Marfans syndrome: assessment by CT. Neuroradiology 1999;41(11):8504. [33] Oosterhof T, Groenink M, Hulsmans FJ, et al. Quantitative assessment of dural ectasia as a marker for Marfan syndrome. Radiology 2001;220(2): 5148. [34] Ahn NU, Nallamshetty L, Ahn UM, et al. Dural ectasia and conventional radiography in the Marfan lumbosacral spine. Skeletal Radiol 2001;30(6): 33845. [35] Weigang E, Ghanem N, Chang XC, et al. Evaluation of three dierent measurement methods for dural

572

DEMETRACOPOULOS & SPONSELLER

[36]

[37]

[38]

[39]

[40] [41]

[42]

ectasia in Marfan syndrome. Clin Radiol 2006; 61(11):9718. Habermann CR, Weiss F, Schoder V, et al. MR evaluation of dural ectasia in Marfan syndrome: reassessment of the established criteria in children, adolescents, and young adults. Radiology 2005; 234(2):53541. Robins PR, Moe JH, Winter RB. Scoliosis in Marfans syndrome. Its characteristics and results of treatment in thirty-ve patients. J Bone Joint Surg Am 1975;57(3):35868. Jones KB, Erkula G, Sponseller PD, et al. Spine deformity correction in Marfan syndrome. Spine 2002; 27(18):200312. Di Silvestre M, Greggi T, Giacomini S, et al. Surgical treatment for scoliosis in Marfan syndrome. Spine 2005;30(20):E597604. Birch JG, Herring JA. Spinal deformity in Marfan syndrome. J Pediatr Orthop 1987;7(5):54652. Lipton GE, Guille JT, Kumar SJ. Surgical treatment of scoliosis in Marfan syndrome: guidelines for a successful outcome. J Pediatr Orthop 2002; 22(3):3027. Fredrickson BE, Baker D, McHolick WJ, et al. The natural history of spondylolysis and spondylolisthesis. J Bone Joint Surg Am 1984;66(5):699707.

[43] Hobbs WR, Sponseller PD, Weiss AP, et al. The cervical spine in Marfan syndrome. Spine 1997;22(9): 9839. [44] Herzka A, Sponseller PD, Pyeritz RE. Atlantoaxial rotatory subluxation in patients with Marfan syndrome. A report of three cases. Spine 2000;25(4): 5246. [45] Carter N, Duncan E, Wordsworth P. Bone mineral density in adults with Marfan syndrome. Rheumatology (Oxford) 2000;39(3):3079. [46] Giampietro PF, Peterson M, Schneider R, et al. Assessment of bone mineral density in adults and children with Marfan syndrome. Osteoporos Int 2003; 14(7):55963. [47] Pyeritz RE, Dietz HC. Marfan syndrome and other microbrillar disorders. In: Royce PM, Steinmann B, editors. Connective tissue and its heritable disorders: molecular, genetic, and medical aspects. 2nd edition. New York: Wiley-Liss; 2002. [48] Bain MA, Zumwalt RE, van der Bel-Kahn J. Marfan syndrome presenting as aortic rupture in a young athlete: sudden unexpected death? Am J Forensic Med Pathol 1987;8(4):3347. [49] Maron BJ, Epstein SE, Roberts WC. Causes of sudden death in competitive athletes. J Am Coll Cardiol 1986;7(1):20414.

You might also like