Deformities Associated with Neurofibromatosis
8 Deformities Associated with Neurofibromatosis
Von Recklinghausen’s disease of the nervous system (neurofibromatosis [NF1]) is to be differentiated from von Recklinghausen’s disease of bone (hyperparathyroidism). It is a common autosomal dominant disorder (▶ Fig. 8.1). NF1 arises from an abnormality of neural crest differentiation and migration during early embryogenesis.
Fig. 8.1 (a) Mother and daughter with NF1, the former with multiple cutaneous nodules (fibroma molluscum) and the latter with a cutaneous plexiform neurofibroma (elephantiasis neuromatosa). (b) Back view of the daughter showing a high thoracic curve.
There are two discrete forms: NF1 and NF2. NF1 is the most common with an incidence rate of 1 in 4,000 live births. NF1 is caused by a defect in the gene for the protein neurofibromin which is a tumor suppressor gene. It was NF1 that was described by von Recklinghausen but like so many medical conditions that are referred to eponymously it was originally described about a century before von Recklinghausen’s description. 1 Early work by Smith recorded both the clinical and postmortem findings and he thought that the tumors were connected to minute nerve branches. 2 Interestingly he did not mention involvement of the skeleton. Then Virchow 3 studied carefully the pathology of the nodules and determined that these were true neoplasms and not neuromata. Von Recklinghausen himself demonstrated nerve elements in the fibrous tissue tumors and correlated the nerve and skin lesions. 4 Chauffard then demonstrated that the pigmented skin lesions were characteristic of the disease as were the tumors. 5 The diagnostic value of these café-au-lait spots (melanin deposits in the basal skin layers) was reaffirmed by Thannhauser who recommended a careful search elsewhere for other lesions when the characteristic skin lesions were noted. 6 Whitehouse increased the diagnostic value of café-au-lait spots by recommending that at least six should be present 7 and the axilla is a favorite location. NF1 may involve virtually every system and organ in the body and so clinical manifestation may vary considerably. It is rather like scoliosis itself, not so much being a matter of having the condition but rather how much of it you have.
That the condition can affect all components of the musculoskeletal system supports the opinion that the pathogenesis involves embryological or developmental failure of both ectodermal and mesodermal tissues. 8– 10
NF1 is diagnosed when two of the seven signs listed in ▶ Table 8.1 are present.
There should be six café-au-lait spots measuring 1.5 cm in the adolescent and 0.5 cm in the young confirming that the activity of these areas of pigmentation and indeed the tumors are further related to sex hormone status 11 (▶ Fig. 8.2).
Lisch nodules are hamartomata in the iris present in three-quarters of cases of NF1 with an increasing prevalence rate with age so that all adults have them.
There should be two or more neurofibromata or one plexiform neurofibroma and, interestingly, these can involve the nerves and blood vessels of the gut as well as the musculoskeletal system.
All orthopaedic surgeons have known about the osseous changes in von Recklinghausen’s disease with long bone pseudarthroses being common along with dystrophic scoliosis although the latter, perhaps surprisingly, is not sufficiently specific enough to be a diagnostic criterion.
Sphenoid dysplasia is also distinctive.
An affected first degree relative is also diagnostic but there is a high rate of mutation approaching 50%. 12
There is a definite risk of the tumors becoming malignant and changing into the highly lethal neurofibrosarcoma, variably estimated at between 5% and 10%.
Fig. 8.2 Multiple cutaneous café-au-lait spots. Note also the surface appearance of a short angular thoracolumbar scoliosis with a lot of rotation.
8.2 Axial Skeletal Lesions in Neurofibromatosis
With the passage of time more attention was naturally focused on skeletal radiology in NF1. Prevalence rates of skeletal involvement in less than 30% of cases 13 soon rose to more than 50% 14 as the importance of spinal involvement became recognized. Hunt and Pugh, 15 in documenting the skeletal lesions in NF1, put much weight on spinal involvement. In addition to describing these spinal deformities, tibial pseudarthroses, and orbital defects, they stressed the importance of attenuated pencilled ribs, enlarged intervertebral foramina, and in particular the scalloped appearances of the vertebrae themselves (▶ Fig. 8.3). An early single case report of scalloping due to pressure from an adjacent neurofibroma 16 led to the initial view that dysplastic bodies and enlarged foramina were due to pressure from local tumor involvement but this was soon repudiated by the observation that scalloping occurred most commonly without any local tumor tissue. 17 Meanwhile, operatively and radiographically it had been demonstrated that these characteristic skeletal lesions often occurred in association with local meningoceles and dural enlargements (ectasia) such that these were incriminated in the pathological process. 15, 18– 22 There is undoubtedly an association, but the relative paucity of such radiological findings in the great majority of cases lends support to the view of Heard and Payne 17 that these dystrophic changes are primary, and therefore in some way associated with the underlying condition, while the meningeal dilatations are merely filling the greater available space. There is certainly no evidence of neurofibromatous tissue as a causative factor at the apex of spinal deformities. 23
Fig. 8.3 PA tomogram of the upper thoracic spine showing the rib pencilling (vertical arrow) and the vertebral scalloping (horizontal arrow) typical of neurofibromatosis.
8.3 Spinal Deformities in Neurofibromatosis
8.3.1 Pattern of Deformity
Once again the chief confusing factor as regards the description of the deformity has been an analysis of anteroposterior (AP) and lateral radiographs of the patient and not the deformity (see ▶ 3). In this respect kyphoscoliosis comes through as the main spinal deformity, but it is not possible, as with other causes of scoliosis, 24 for the condition of kyphoscoliosis to really exist. Roaf in his classic text 25 likened the curve pattern in NF1 to that in idiopathic scoliosis, namely a lordoscoliosis, although also recognizing a common shorter and more angular curve (▶ Fig. 8.4). He further pointed out 26 “if kyphosis means an increase in length of the posterior elements of the vertebral column relative to the anterior elements the use of the term is certainly erroneous” in relation to kyphoscoliosis. When Weiss, in attributing recognition of the high incidence of scoliosis in von Recklinghausen’s disease to another dermatologist Engman, described spinal curvatures, 27 he wisely did not go much further.
Fig. 8.4 PA view of the short sharp angular curve typical of von Recklinghausen’s disease. There are black dots over the middle of the bodies and triangles over the spinous processes confirming lordosis. Note that there are only four really angulated and rotated vertebrae in this curve.
Subsequent studies of the pattern of deformity showed that it can be divided into two groups of cases: patients with a diagnosis of NF1 who have been examined for the presence of a spinal deformity, and patients presenting as the result of a spinal deformity in the clinic in association with NF1. It is only possible to look at the prevalence rate of spinal deformity in the first group, but as Scott pointed out 28 even this group are selected to a degree by presenting to hospital as a result of one or more components of their condition. Nonetheless, in these studies of patients with NF1, the prevalence rate of spinal deformity ranges from 20 to 40%. 22, 28– 31 Most of these studies could find no particular pattern of deformity, 22, 28, 29 although the thoracic curve was the most common site encountered. Chaglassian 31 looked at 141 cases and found 37 cases of scoliosis (26%), the majority having long idiopathic-type curves with only 16 short, sharp angular ones. Although he could detect no particular curve pattern it is interesting that the average age of presentation of the 21 long curves was 7 years, while that of the 16 short curves was surprisingly 10 years. These 141 cases were all diagnosed according to the criteria of Crowe et al 32 and therefore this series did not contain any formes frustes of neurofibromatosis. Laws and Pallis looked at 18 unselected cases of NF1 and found a spinal deformity in seven (39%), but none were short angular curves. 22 Scott 28 could find no uniformity in length or direction of thoracic curves. However, in seven cases he observed evidence of a congenital bony anomaly in the form of fused ribs or hemivertebrae.
Meanwhile, Cobb had clearly stated that there were two types of scoliosis in neurofibromatosis: a milder idiopathic type and a particularly progressive short angular dystrophic type. 33 Moreover, he stated that the diagnosis could be made radiographically on the basis of this very characteristic short angular curve. He also pointed out that later in childhood some curves did appear congenital but there were always earlier films that showed normal vertebrae. Apart from prevalence rate all that can be said from these studies is that the thoracic curve is the most commonly encountered and it is more often of a long idiopathic type than a short angular dystrophic type.
Other studies of spinal deformity in NF1 concern patients who have presented as a result of a spinal deformity. 34– 44 These therefore represent an extremely selected group from which it would be quite incorrect to infer anything about prevalence rates and patterns of deformity. Quite naturally in this selected group the most common curve encountered is the short angular dystrophic curve. This situation is analogous with studies of idiopathic scoliosis in which data derived from scoliosis clinics describe the thoracic curve site as the most common with a significant female preponderance, whereas data from community screening studies reveal a more even gender ratio and the lumbar spine is now found to be preponderant. 45
Thus it is the short sharp angular dystrophic curve that attracts the attention of the scoliosis surgeon. 34– 44 Veliskakis looked at 55 patients with a spinal deformity and found 43 patients with short angular curves with wedging that often looked like a hemivertebra. 37 There was an equal gender distribution but there were more left-sided curves. In contradistinction to Chaglassian et al, 31 he found that the short curves appeared, as expected, at an early age and progressed more rapidly. Kyphosis was noted to be common and the severity of the kyphosis was proportional to the severity of the lateral spinal curvature. This implies that AP and lateral views of the patient were assessed, which gave the spurious appearance of a kyphosis. (If you’re not with us go back and read ▶ 3 again). Dawson et al looked at 41 cases and followed them for 5 years. 38 Twenty-one were sharp angular curves, nine were pseudo-idiopathic, and the appearance of a kyphosis was seen in 14 patients. The 14 cases that progressed only did so by 30 degrees in 5 years. Moe et al reported 100 dystrophic curves out of a total of 112 cases of spinal deformity in neurofibromatosis. 42 They noted that all dystrophic curves progressed but only seven were lordotic. Winter et al then reviewed 102 patients, 80 of whom had dystrophic changes. 43 They described 31 as having kyphoscoliosis, 49 as scoliosis only, with no kyphotic component (he accepted 49 degrees of kyphosis as normal), and only 5 lordoscoliotic curves.
These reports demonstrate clearly that it is not possible to accurately define curve pattern from AP and lateral views of the patient. We have already seen (see ▶ 2 and ▶ 3) the spurious appearance of spinal deformities when inappropriate views of them are taken, but there is still much that can be learned about a spinal deformity from an AP view of the patient. Roaf 25, 26 clearly understood the spine in three dimensions; he stressed the simple geometrical point that if there is rotation of a scoliosis with the vertebral bodies toward the convexity and the spinous processes toward the concavity then the line of the vertebral bodies is longer than the line of the posterior elements which, in conjunction with the direction of rotation, implies that the front of the spine is longer than the back and therefore these curves are all lordoscolioses. Moreover, in these studies of patients with neurofibromatosis and a spinal deformity, 34– 44 all the AP radiographic illustrations confirm the presence of a lordoscoliosis. Vlok, however, recognized these radiographic features, and in his report of 21 cases 44 noted a short sharp angular dystrophic curve in 9 cases, all of which had a lordosis. Furthermore, the average curve magnitude of the lordoscolioses was significantly greater at 74 degrees than those who had a true kyphosis in which average curve magnitude was only 25 degrees (there being no buckling or “spinning” potential with a kyphotic spine).
Although by far the most common deformity in NF1 is the lordoscoliosis, this does not imply that these spines cannot be kyphotic. However, it does indicate that the area of kyphosis exists alone or lies above or below the scoliosis, where the vertebrae are not rotated. We therefore have the elementary principle that the pattern of spinal deformity in NF1 is similar to idiopathic spinal deformities. The thoracic lordoscoliosis seen in the NF1 patient is comparable to the idiopathic scoliotic deformity, while the thoracic kyphosis, with rotated lordoscoliosis below, is comparable to the Scheuermann’s deformity albeit more angular. 46 The prevalence rate of clinically significant deformities in neurofibromatosis is however 150 times greater than in otherwise normal children (30% and 0.2%, respectively), and this is entirely attributable to the dystrophic nature of the vertebrae. The forces acting on the spine, which 98% of normal children can resist, cannot be resisted by the dystrophic vertebrae of any more than 70% of patients with von Recklinghausen’s disease. Moreover, if the dystrophic process is particularly obvious, 34– 44 then the short sharp angular curve is produced, whereas with little or no dystrophic change the longer idiopathic-type curve is produced. 22, 28– 31 In the former situation the spine fails locally, whereas in the latter the spine fails over a greater area.
It is far better to be less prescriptive about scoliosis in NF1 and not simply try to divide the curves into angular dystrophic and long idiopathic type curves because there is clearly a spectrum from one to the other; the more evidence of systemic von Recklinghausen’s disease, the more dystrophic features there are whereas the less the evidence of NF1, the more the patient resembles a normal child with the more prevalent long idiopathic type deformity. In von Recklinghausen’s disease dystrophic curves always require an anterior spinal fusion in addition to a posterior one to try and mitigate a strong tendency for curve progression after surgery. In addition, knowing the unpredictability of the von Recklinghausen’s scoliosis, you should always carry out a front and back fusion regardless of whether you think the curve is dystrophic or not. If you don’t do this, then the von Recklinghausen scoliosis will make you regret that you had not done so. Do not forget the aphorism of the philosopher Santayana(6) “those who cannot remember history are condemned to repeat it.”
Moreover, in an excellent review from the Royal National Orthopaedic Hospital in London, 47 they pointed out that modulation from initially idiopathic type curves to dystrophic ones occurs commonly and in the 91 cases they reported 80% of children under the age of 7 modulated and 25% of those over the age of 7. There are two particularly important discriminating factors, one obviously being patient age, and the other the presence of rib pencilling. All the more reason therefore for assuming dystrophism in all NF1 spinal deformity cases.
While the otherwise normal child who develops an idiopathic spinal deformity does not do so in the cervical region, the dystrophic vertebral situation in the patient with von Recklinghausen’s disease facilitates the production of cervical spine deformities. 48, 49 Yong-Hing 49 from Dean MacEwen’s center in Wilmington noted that of 56 patients, 17 had cervical deformities and 15 of these occurred in association with significant deformities lower down in the spine. The majority of these were angular kyphoses, although lordoses were not uncommon. These deformities so high in the spine tend to resist rotation, but in the cervicothoracic junction serious rotational abnormalities can be produced in the presence of an underlying lordosis.
There are therefore four patterns of spinal deformity which can be encountered in the patient with neurofibromatosis:
An inconsequential nonstructural lumbar curve consequent upon leg-length inequality caused by local hemi-hypertrophy.
The idiopathic-type long lordoscoliosis.
The dystrophic short angular lordoscoliosis.
An angular thoracic or cervical kyphosis that can occur alone, or above a compensatory lordosis that has rotated to the side.
8.3.2 Neurological Involvement
Neurological deficits secondary to a spinal deformity are very prevalent in association with NF1, second only to congenital spine deformities. 50 In a superb report by Curtis et al of 8 cases, with a review of the literature, 51 they found 32 cases of NF1 with paraplegia from the end of the 19th century. 52– 71 Analysis of these cases demonstrated two clear patterns of spontaneous neurological involvement affecting the spine in von Recklinghausen’s disease. The most common situation is a low cervical or high thoracic paraplegia in association with a local angular kyphosis for which laminectomy is disastrous. In other words, the last thing angular dystrophic kyphosis in the growing child wants is removal of the posterior spinal column which of course is the only local tension member and if removed rapidly leads to accelerated progression of the deformity. If a laminectomy is required for removal of an intradural lesion, then it is mandatory to carry out a concomitant spinal fusion procedure. This clearly cannot be a posterior fusion because there are no posterior elements left and so anterior or lateral intertransverse fusion is necessary, and preferably both. Cobb stated so succinctly 33 “the spinal deformity resulting from neurofibromatosis may be horrible, but laminectomy without stabilization will make it a nightmare.” The second type of neurological problem is that associated with local tumor formation, is much less common than in association with a spinal deformity itself, and can arise anywhere in the spine. It is only in this group that laminectomy, followed by tumor removal, is of benefit. Again there should be no hesitation in carrying out a concomitant spinal fusion.
8.4 Management of Spinal Deformities in Neurofibromatosis
Retrospective analyses of the treatment of these spinal deformities have produced interesting but in no way surprising information that tells more about the natural history of these deformities than the efficacy of treatments prescribed. Conservative treatment has been in the form of traditional Milwaukee bracing and, of course, the different “responses to treatment” are attributable solely to whether the curve was more of the mild idiopathic type or tended toward the more progressive angular dystrophic variety. Even now, when it has been confirmed that there is no evidence base in support of orthotic treatment for any kind of scoliosis 72 except perhaps a circumferential torso support device for the collapsing neuromuscular curve, the Milwaukee or one of its underarm forms still has its proponents. Of course the most recent brace trial 73 had a number of serious flaws—no randomization, more of the progressive thoracic curves in the control group than the treated group biasing natural history, and stopping before the attainment of spinal maturity to mention but a few. Even forgetting about these discrepancies the results were extraordinary. Using an increase in Cobb angle by 6 degrees as failure the control group was significantly worse than the brace group with the electrical stimulation (LESS 74) group worst. Not only should orthotic treatment for idiopathic scoliosis be binned or kicked into the long grass but we are talking about scoliosis in NF1 and it is positively harmful in those children whose curves are known to have significant progression potential. Fiddling about with bracing for a number of months or years simply delays surgical treatment for a condition that most definitely requires it. Not surprisingly it was the short angular dystrophic curve that responded “less well” to brace treatment.
Cobb 33 stated that it was better to fuse earlier than to wait and accordingly most of these reports favored early surgical intervention in the nature of posterior fusion with or without instrumentation. Rapp and Glock, however, disagreed with those that said the scoliosis always needed early fusion and emphasized that age played a major part in determining treatment. 75 They were essentially emphasizing that growth velocity was the final common denominator and that fusion would be all the more definitive the closer to the end of adolescent growth. Consequently early results of posterior fusion were not encouraging and in Dawson’s series of 27 operations, the average correction was only 23 degrees, 6 of whom subsequently lost correction and a further 6 developed pseudarthroses. 38 In Stagnara’s series of 37 cases there was little difference in postoperative progression between posterior fusion and posterior fusion along with metalwork which averaged about 23 degrees. 39 He did however perform some extensive anterior and posterior fusions but the numbers were not sufficient to attribute benefit. In Winter’s series of 102 patients 43 he tried to divide kyphosis into two groups, above or below 50 degrees. For those with a kyphosis of less than 50 degrees a posterior fusion alone was satisfactory but for those more than 50 degrees there was a two-thirds incidence of pseudarthrosis and in such cases he recommended quite rightly an additional anterior fusion. However, in Professor Tanner’s wonderful book Growth at Adolescence, 76 he nicely describes the growth velocity curve of normal children along with standard centile charts to record serial measurements, 77 and not those with notoriously progressive deformities such as von Recklinghausen’s disease. These unfortunate children do not behave according to Tanner’s normal child and that is why aggressive curves such as neurofibromatosis, early onset idiopathic, and congenital present earlier with a much worse progression potential than their idiopathic counterpart. Although George Rapp from Indianapolis meant well by telling us to pay attention to adolescent growth 75—the simple fact is that you have to step in surgically as and when the patient tells you, which is sooner rather than later regardless of precise chronological age. Moreover, 90% of height gain during adolescence occurs in the legs and not much in the spine, so concern about stunting spinal growth is not really justified.
Have we made much progress since these earlier reports mainly in the 1970s? Reviewing now the last 20 or so years of publications concerning spinal deformities in association with NF1 one sees the same sort of results replicated. There is also the same tendency to divide curves into the dystrophic and nondystrophic varieties when there is no clear discriminatory point between them. We can certainly all recognize the very dystrophic curve (short, sharp, angular, considerable apical wedging, rib pencilling, large foramina, lots of rotation) but it is much more difficult to fix a point down the pathway to dystrophism to confirm two discrete categories. Our basic message to all young and aspiring scoliosis surgeons is to assume they are all of the dystrophic variety (which they may well be) and treat them in the same aggressive surgical manner—anterior and posterior.
Surprisingly, at the height of the French revolution, CD instrumentation was regarded as being the universal panacea for all scolioses and was used as the first segmental instrumentation for neurofibromatosis. In Miami, Harry Shufflebarger reported on 11 NF1 patients, 10 with idiopathic-like curves and 1 dysplastic. Correction in all three planes was very satisfactory but the dysplastic NF1 patient required further anterior surgery. 78 Then Holt and Johnson from the Leatherman centra in Louisville in 1989 reported five dystrophic NF1 patients treated with CD instrumentation with only one posterior and the rest front and back procedures. Despite this, three patients showed significant progression that required further surgery and they quite rightly talked about the “tendency to extraordinary progression with growth.” 79
In 1999 the spine surgeons from the Rizzoli Institute in Bologna reported on 56 cases of dystrophic curves divided, as Winter had originally suggested, into those with a kyphosis less (group 1) or more than 50 degrees 80 (group 2). Despite stating that they had previously recommended anterior and posterior surgery 81 they carried out posterior surgery alone in 19 cases and only combined anterior and posterior in the remaining 6! The average age of the children was 13 years but did go down in range to 4 years. Only 10 of these posterior fusion patients achieved stabilisation, the initial Cobb angle being 71 degrees with 33 degrees of kyphosis going down to 45 degrees and 25 degrees, respectively, at surgery, deteriorating with time to 54 degrees and 31 degrees. Even in group 2 children they still did 11 posteriorly and 20 anteriorly and posteriorly. Failure to stabilize the curve occurred in more than 50% of those who underwent posterior fusion alone and less than 25% in those who had anterior and posterior fusions. They warned again that the severe dystrophic curve always requires combined anterior and posterior stabilization, particularly in younger patients, “even if the sagittal curves do not become pathologic by the time of presentation.”
In 2005 a review of neurofibromatosis in terms of diagnosis and treatment was published from London and again categorized the curves into dystrophic and nondystrophic types although emphasized that that should be based on a meticulous assessment of the spine with both plain films and magnetic resonance imaging (MRI) to highlight radiologically undetectable dysplastic features that would determine prognosis and surgical planning 47 (▶ Table 8.2). When MRI scanning of the whole spine is carried out to identify vertebral dysplasia then in one-third of cases of NFI initially classified on plain films as having nondystrophic curves did have typical vertebral dysplasia and hence were therefore dystrophic curves. Nondystrophic curves could be managed as for idiopathic scoliosis whereas dystrophic curves required anterior and posterior fusion of the entire structural curve with abundant autologous bone graft. Assessing carefully these MRI scan features indicating dystrophic change markedly assists in the division between dystrophic and nondystrophic curves and indicate that when there were three or more dysplastic features present the risk of curve progression was significantly increased in 85% of patients, with rib pencilling being the most important single factor.
From the above it doesn’t look as though there are many scolioses in von Recklinghausen’s disease that are not dystrophic (15% or less) in which case the message—treat them all as dystrophic anyway—sounds like even better advice (no apologies for telling you again).
China seems to have become an important source of scoliosis literature in recent years and in 2009 the Changhai Orthopaedic Department 82 reported on 19 patients with NF1 noting that posterior instrumented fusion alone was not adequate to correct scoliosis. They say because of weak bone structure but while this is true (dystrophic change) time allows spinal growth to significantly alter vertebral shape, particularly at the apex of the curve where the vertebrae can look like hemivertebrae, and so accelerates and potentiates early and progressive deformity. They looked at whether extension of the fusion beyond the usual neutral to neutral vertebrae would enable posterior instrumented fusion to be effective. Their patients were aged more than 10 years and their scolioses were less than 90 degrees. There were 16 dystrophic and 3 nondystrophic curves. In the dystrophic curves the initial Cobb angle was 68 degrees with 30 degrees of kyphosis and this was reduced to 27 and 30 degrees at a minimum of 2 years’ follow-up. A pseudarthrosis only occurred in one patient.
This goes back to the old notion of end-to-end posterior fusion (parallelism), rather than neutral to neutral, for younger idiopathic curves. 83 Although this seemed to be successful in their series the Cobb angle or degree of kyphosis did not seem excessive whereas with aggressive dystrophic curves both measures are usually considerably greater. The power of transpedicular instrumentation seems to result in a well corrected and stable curve, at least in Li’s series, 82 but the real message of having to go end-to-end should be to mandate an anterior and posterior fusion.
Then the Cairo Group reported 32 cases of dystrophic NFI again divided into two groups according to the angle of pseudokyphosis, less than and more than 45 degrees. 84 All underwent what they describe as aggressive anterior and posterior surgery with an average of four apical disks removed. When the deformity is as angular as can occur in NFI then the spine above and below the apex angulates away from the surgeon but it is usually possible to remove at least four intervertebral disks including the growth plates. Taking out first the apical three disks usually allows a degree of curve improvement bringing the disks above and below more accessible. Then posterior transpedicular instrumentation with some sublaminar wires was the basis of the second stage. Their results were perhaps surprisingly good with a preoperative Cobb angle of 100 degrees being reduced to 40 degrees in both groups and with no significant loss of correction over a 3-year follow-up. 84
In 2010 16 patients with dystrophic NFI deformities were reported having undergone corpectomy and circumferential spinal fusion alleging that vertebral body resection had not previously been investigated 85 although we reported on vertebral body resection for severe dystrophic curves from the Leatherman spine center back in 1988. 86 Leatherman’s closing wedge resection (▶ Fig. 6.19) has become the model for all subsequent techniques. 87, 88 There was a correction of Cobb angle in the frontal and lateral planes of 90 degrees and 70 degrees respectively down to 50 degrees each with just a few degrees lost in the frontal plane and 13 degrees in the lateral plane at 7-year follow-up. 85
Rib head protrusion into the central canal in NFI cases has been described several times with or without spinal cord dysfunction since the first description in 1986 89, 90 (▶ Fig. 8.5) Clearly if this occurs then rib resection and spinal canal decompression is required as part of the surgical strategy.
Fig. 8.5 The rib head (A) has dislocated through the enlarged foramen and is compressing the cord (B).