Scoliosis and hydrocephalus in myelocele patients

Transcription

1 J Neurosurg 50: , 1979 Scoliosis and hydrocephalus in myelocele patients The effects of ventricular shunting PETER HALL, M.B., B.S., RICHARD LINDSETH, M.D., ROBERT CAMPBELL, M.D., JOHN E. KALSBECK, M.D., AND ALONSO DESOUSA, M.D. Division of Neurosurgery and Department of Orthopaedics, Indiana University Medical Center, Indianapolis, Indiana ~" Developmental scoliosis is a common cause of increasing disability and deformity in long-term myelocele survivors, and is believed to result from a paralytic collapsing spine. The possible etiological role of compensated hydrocephalus and hydromyelia was assessed by determining the effect of ventricular shunting on 11 myelocele patients with developmental scoliosis. After successful shunting, one patient with a 47 ~ curve continued to deteriorate. Three cases with curves greater than 60 ~ were stabilized for short periods, but eventually required spinal fusion. Seven cases with curves less than 55 ~ were improved from a mean scoliosis of 29 ~ to 13 ~ during a 20-month follow-up period. Several patients had pre-existing shunts that were found to be non-functional on shuntogram. These findings suggest that the spinal complications of hydrocephalus may be more common than previously recognized in myelocete patients and that advanced developmental scoliosis may be avoided by early recognition and ventricular shunting. KEY WORDS 9 hydrocephalus 9 hydromyelia 9 syringomyelia 9 scoliosis 9 myelodysplasia S PINAL deformities are a major problem in the long-term management of myelocele patients.14 Scoliosis is particularly common, occurring in 50% to 70% of these patientsy '22'24 It may either be congenital (the result of vertebral anomalies present at birth) or, more frequently, developmental (generally considered to result from a paralytic, collapsing spine)? 1'21'22'24 The mechanism by which developmental scoliosis arises remains unclear. Unapposed action of the iliopsoas has been suggested, 22'~s but it does not explain the thoracic location of most of these curves. A congenital disturbance of paraspinous muscle innervat/on is also unlikely, as the curves are found above the myelocele and at birth the spinal neurons appear histologically normal above the neural plate deformity. 19 Congenital dislocation of the hips has also been implicated, 22 but recently a similar incidence of hip dislocation has been found in myelocele patients with and without scoliosis? Myelocele patients may also develop spasticity and increasing weakness of the extremities, frequently in association with developmental scoliosis. 7"s'~2 We have found these patients characteristically have compensated hydrocephalus and hydromyelia communicating with the ventricles. 7 9 As ventricular shunting decompresses the spinal cysts, and often results in some reduction of spasticity and improvement of strength, 7 we have suggested previously that the hydromyelia might be the basis for both the spinal curves and the extremity weakness? If hydromyelia is an etiological factor in developmental scoliosis, ventricular shunting should improve or stabilize these curves before secondary structural changes occur. This report presents a short-term follow-up study of the effects of ventricular shunting on developmental scoliosis in myelocele patients. Clinical Materials and Methods A group of 1 1 myelocele patients was selected from the cases with scoliosis referred by the Orthopaedic Myelodysplasia Clinic at Riley Children's Hospital between October, 1974, and October, 1976, for evaluation of possible hydromyelia. Patients were included in this study if their scoliosis measured 15 ~ or greater and had increased within the year before the shunt was placed. Cases were excluded if primary vertebral anomalies were also present, or if a spinal fusion of the developmental curve had been performed before shunting or within 6 months after the shunt. ]74 J. Neurosurg. / Volume 50 / February, 1979

2 Scoliosis and hydrocephalus in myelocele patients The 11 cases were selected from a total of 29 patients with scoliosis who were shunted during this time. The other cases were excluded because of co-existing diastematomyelia (one case); a preoperative scoliosis of less than 15 ~ (five cases); a stable curve 1 year before the shunt was placed (two cases); immediate spinal fusion after ventricular shunting for advanced curves (seven cases); shunt infection requiring removal (one case); and failure to obtain a follow-up review (two cases). The selected group consisted of six boys and five girls, aged 6 to 16 years. The scoliosis was assessed in all patients by a standard radiological technique with the patient sitting. Cobb's method was used to measure the curves. Patients with pre-existing shunts underwent positive contrast shuntograms? 5 None were found to have functional shunts. Ventricular size was assessed in all patients by computerized tomography, angiography, or ventriculography. The ventricles were normal in one patient, mildly or moderately enlarged in nine, and markedly expanded in one. All the patients were clinically compensated. Seven patients were investigated for hydromyelia by myelography or radioisotope ventriculography before shunting. Hydromyelia was found in all seven cases. One patient (Case 1) underwent Pantopaque and air myelography -- both of which revealed an enlarged cervical cord. On percutaneous needling of the cord, cerebrospinal fluid was aspirated from a hydromyelic cavity. Injection of air directly into the hydromyelia resulted in rapid ventricular filling. This patient had normal-sized ventricles. Another patient (Case 2) also underwent Pantopaque myelography which revealed a normal-sized cord. However, radioisotope ventriculography revealed hydromyelia in the patient. Seven patients (Cases 1-3, 6-8, and 11) had radioisotope ventriculograms before shunting by a method described previously; 9 all were positive for hydromyelia. Two patients (Cases 2 and 9) had radioisotope ventriculograms after shunting, neither of which showed flow through the hydromyelia. Four patients (Cases 4, 5, 9, and 10) were not investigated for hydromyelia as the association of compensated hydrocephalus and developmental scoliosis was considered sufficient indication for shunting. All 11 patients were treated by ventricular shunts. Ventriculoperitoneal shunts were placed in 10 patients and a ventriculoatrial shunt in one using low- or medium-pressure Holter or Heyer-Schulte valves. Postoperatively, the scoliosis was not treated in any other way. Results The effect of ventricular shunting on the scoliosis of these patients is shown in Table 1. One patient with a preoperative curve of 47 ~ continued to progress to 73 o over a period of 14 months after the shunt. Three patients with scoliosis greater than 60 ~ were stable for periods of 6, 10, and 12 months, and then underwent TABLE 1 The effect of ventricalar shunting on developmental scoliosis Case Spinal Curve Follow-Up No. Level Period Preop Postop (mos) progressive postoperative 1 T10-L3 47 ~ 73 ~ 14 stable postoperative 2 T ~ 70 ~ 10 3 T ~ 65 ~ 12 4 T9-L5 85 ~ 85 ~ 6 improved postoperative 5 T5-L2 31 ~ 5" 24 6 Tll-L3 15 ~ 4* 24 7 T ~ 10 ~ 19 8 T ~ 40 ~ 20 9 T12-IA 15 ~ 0 ~ T4-L2 28 ~ 10* 7 11 TI0-L3 37 ~ 22 ~ 32 mean 28.9 ~ 13 ~ 20.4 spinal fusions. The remaining seven patients (all with scoliosis of less than 55 ~ showed an improvement after shunting from a mean of 29* to a mean of 13 ~ during an average follow-up period of 20 months (Fig. 1). Neither age nor the level of the myelocele appeared to influence the outcome of shunting, although the one failure (Case 1) was our oldest patient, aged 16 years, who also had had normal ventricles. Preoperatively, six patients had developed deterioration of gait; of these, one had become nonambulatory. An additional patient developed weakness and wasting of the upper extremities. All seven patients showed some recovery of function after shunting, which was marked in two patients. The nonambulatory patient improved to the point of limited ambulation with the aid of a walker and braces. The neurological improvement did not correlate entirely with the outcome of the scoliosis to shunting; one patient (Case 1), whose scoliosis worsened, showed considerable improvement of strength in the upper extremities. In the unselected group of 29 shunted patients, ventricular shunts were complicated by symptomatic subdural hematomas in two patients; one of these also had a shunt infection. Both were treated successfully following shunt removal. One of the I I selected patients in this series died 8 months after the shunt procedure of an unrelated cause. Discussion Ventricular shunting produced a short-term improvement in seven of 1 1 patients with progressive developmental scoliosis. This suggests that hydrocephalus is an etiological factor, since the natural history of these curves is to increase? '~'21'25 This finding supports the conclusion of our earlier study s in which an association was found between developmental scoliosis and the presence of unshunted, J. Neurosurg. / Volume 50 / February, 1979 ]75

3 P. Hall, et al. FIG. 1. Upper." Radiographs taken before shunting. Left: At 2 years the spine is straight. Scoliosis first appeared at 4 years. Right: At 6 years the scoliosis measures 31 ~ The patient was shunted shortly after this. Lower." Radiographs taken after shunting. Left: Three months after the shunt the scoliosis was unchanged. Center." One year after the shunt the scoliosis has significantly improved. Right: Two years after the shunt, at 8 years of age, the remaining scoliosis measures 5 ~ clinically compensated hydrocephalus in myelocele patients. Previous studies have demonstrated a relationship between the level of the myelocele and both hydrocephalus at birth and scoliosis developing in the first two decades of life. From the findings of a study of the natural history of scoliosis in myelocele patients, Shurtleff, et al.? 5 predicted that 80% to 90% of patients with thoracic or high lumbar myeloceles would develop scoliosis of 30 ~ or greater by age 20 years, compared to only 9% of those with sacral lesions. A relationship has also been documented between the incidence of hydrocephalus at or shortly after birth and the myelocele level. Lorber 2~ found ventriculographic evidence of hydrocephalus in 96% of patients with thoracolumbar myeloceles, compared with 60% of those with sacral myeloceles. Laurence 18 reported clinical or radiographic evidence of moderate ]76 J. Neurosurg. / Volume 50 / February, 1979

4 Scoliosis and hydrocephalus in myelocele patients FIG. 2, This patient with a sacral myelocele developed deterioration of gait. The patient was initially treated with a suboccipital craniectomy and a cervical myelotomy as the ventricles were normal. The patient failed to improve and was reinvestigated. A ventriculogram showed normal ventricles (left). A myelogram showed persistent hydromyelia (right). A ventricular shunt was then placed with lasting improvement of gait. or severe hydrocephalus in 89% of thoracolumbar and 40% of sacral myelocele patients. An association between scoliosis and diminished walking ability in myelocele patients has been reported by Hoffer, et al. ~2 In their group of 40 patients with lumbar myeloceles, 14 were classified as "community ambulators;" three of the 14 had scoliosis, and three had hydrocephalus. In contrast, 19 of the 40 patients were non-ambulatory; nine of the 19 had scoliosis and 10 had hydrocephalus. Interestingly, while the level of high- or low-lumbar paralysis did not have any relationship to the ambulatory status of these 40 patients, 14 patients developed a significant decline in walking ability between the ages of 9 and 17 years. While these studies suggest a relationship between hydrocephalus, developmental scoliosis, and deterioration of gait in myelocele patients, we do not believe the enlarged ventricles directly cause developmental scoliosis because the hydrocephalus was mild in most of the patients in our study. Furthermore, scoliosis is not common in cases of congenital hydrocephalus from other causes, except those treated by lumbar shunts. 18 An alternative explanation is that hydrocephalus may indirectly cause these curves, possibly by means of hydromyelia, which has frequently been found at autopsy. 2,' In our earlier series we found hydromyelia communicating with the ventricles in 14 of 15 cases with developmental scoliosis? as well as in myelocele patients presenting with progressive weakness and spasticity of the extremities. 7 Other authors have also reported hydromyelia as a cause of progressive decline in motor function, a,xv,23,~8 and respiratory failure 13 in myelocele patients. Recently an association between scoliosis and syringomyelia in three cases of communicating hydrocephalus treated by lumboureteral shunts has been reported by Fischer, et al. 5 These authors have suggested that scoliosis may result from syringomyelia precipitated by lumbar shunts, and not directly from arachnoiditis. Determination of ventricular flow by radioisotope ventriculography in unshunted myelocele patients has revealed that the hydromyelia characteristically acts to compensate the hydrocephalus, 9-1~ supporting the hydrodynamic theory of myelodysplasia developed by Gardner. 8 Hydromyelia may apparently compensate a few cases so well that the size of the ventricles remains normal, as shown in Fig. 2. The critical importance of this compensatory action of hydromyelia in myelocele patients has been demonstrated when the spinal central canal is inadvertently occluded. We have previously reported one fatality from acute hydrocephalus following occlusion of the hydromyelic cavity with Pantopaque introduced during positive contrast ventriculography. 7 Winston, et al.? 7 have reported acute hydrocephalus complicating ligation of the spinal cord in myelocele patients undergoing correction of kyphosis. J. Neurosurg. / Volume 50 / February,

5 P. Hall, et al. Ventricular shunting has proven to be a good method of treatment of hydromyelia, shown by an absence of ventricular fluid flow through the spinal cord on radioisotope ventriculography/a decrease in spinal cord diameter on myelography/ and clinical improvement in extremity strength and spasticity. 7,8 The improvement in developmental scoliosis seen in seven of these cases after ventricular shunting, therefore, supports our theory that hydromyelia may be the etiology of these spinal curves. In conclusion, we believe that the most common clinical manifestation of unshunted hydrocephalus in the older myelocele patient is the onset of developmental scoliosis. Many patients may also develop increasing extremity weakness and spasticity. These deficits probably do not arise directly from the hydrocephalus but may arise from the associated hydromyelia. Macrocephaly and signs of increased intracranial pressure are not commonly present and, in fact, the ventricular size may be normal on radiological investigation. Ventricular shunting may not be indicated as a routine in all myelocele patients, but we believe that when developmental scoliosis is diagnosed, evidence of hydrocephalus or hydromyelia should be sought. Ventricular shunting may then be tried as a primary method of treatment, although advanced scoliosis with secondary structural changes of the vertebral column will continue to require surgical correction and fusion of the spine. Acknowledgments The authors thank Ms. Anita Lightfoot for help in preparation of the manuscript and Ms. Christine A. Bertelson for editorial assistance. References 1. Barden GA, Meyer LC, Stelling FH: Myelodysplastics -- fate of those followed for twenty years or more. J Bone Joint Surg 57A: , Cameron AH: The Arnold-Chiari and other neuroanatomical malformations associated with spina bifida. J Pathol Bacteriol 73: , Day AL, Maniscalco JE, Geissinger JD, et al: Communicating hydromyelia. Surg Neurol 7: , Emery JL: The back lesions, lipomas and dermoids, in: Proceedings of the American Academy of Orthopaedic Surgeons Symposium on Myelomeningocele, Hartford, Connecticut, November, St. Louis: CV Mosby, 1972, pp Fischer EG, Welch K, Shillito J: Syringomyelia following lumboureteral shunting for communicating hydrocephalus. Report of three cases. J Neurosurg 47:96-100, Gardner W J: Myelocele: rupture of the neural tube? Clin Neurosurg 15:57-79, Hall PV, Campbell RL, Kalsbeck JE: Meningomyelocele and progressive hydromyetia. Progressive paresis in myelodysplasia. J Neurosurg 43: , Hall PV, Kalsbeck JE, Wellman HN, et al: Clinical radioisotope investigations in hydrosyringomyelia and myelodysplasia. J Neurosurg 45: , Hall PV, Kalsbeck JE, Wellman HN, et al: Radioisotope evaluation of experimental hydrosyringo- 178 myelia. J Neurosurg 45: , Hall PV, Lindseth RE, Campbell RL, et al: Myelodysplasia and developmental scoliosis. A manifestation of syringomyelia. Spine 1:48-56, Hensinger RN, MacEwen GD: Congenital anomalies of the spine, in Rothman RH, Simeone FA (eds): The Spine. Philadelphia: WB Saunders, 1975, Vol. 1, pp (see pp ) 12. Hoffer MM, Feiwell E, Perry R, et al: Functional ambulation in patients with myeiomeningocele. J Bone Joint Surg 55A: , James HE, Schut L, Pasquariello PP: Communication of hydromyelic cavity with fourth ventricle shown by combined Pantopaque and air myelography. Case report. J Neurosurg 38: , James JIP: Spinal deformities in myelomeningocele. J Bone Joint Surg 60B:3-4, 1978 (Editorial) 15. Keucher TR, Campbell RL: Direct shunt injection (shuntogram). Presented at the Annual Meeting of the American Association of Neurological Surgeons, New Orleans, Louisiana, April 23-27, 1978 (Paper No. 70) 16. Kushner J, Alexander E Jr, Davis CH Jr, et al: Kyphoscoliosis following lumbar subarachnoid shunts. J Neurosurg 34: , Lassman LP, James CCM, Foster JB: Hydromyelia. J Neurol Sci 7: , Laurence KM: The natural history of spina bifida cystica. Detailed analysis of 407 cases. Arch Dis Child 39:41-57, Lendon RG: Neuron population in the spinal cord of children with spina bifida and hydrocephalus. Dev Med Child Neurol (Suppl 15):50-54, Lorber J: Systematic ventriculographic studies in infants born with meningomyelocele and encephalocele. The incidence and development of hydrocephalus. Arch Dis Child 36: , Mackel JL, Lindseth RF: Scoliosis in myelodysplasia. J Bone Joint Surg 57A:1031, 1975 (Proceedings) 22. Raycroft JF, Curtis BH: Spinal curvature in myelomeningocele: natural history and etiology, in: Proceedings of the American Academy of Orthopaedic Surgeons Symposium on Myelomeningocele, Hartford, Connecticut, November, St. Louis: CV Mosby, 1972, pp Schurr PH: Tethered cord and syringomyelia: a problem in differential diagnosis. Dev Med Child Neurol 18 (Suppl 37):165, Sharrard WJW: The kyphotic and lordotic spine in myelomeningocele, in: Proceedings of the American Academy of Orthopaedic Surgeons Symposium on Myelomeningocele, Hartford, Connecticut, November, St. Louis: CV Mosby, 1972, pp Shurtleff DB, Goiney R, Gordon LH, et al: Myelodysplasia: The natural history of kyphosis and scoliosis. A preliminary report. Dev Med Child Neurol 18 (Suppl 37): , Turnbull FA: Syringomyelic complications of spina bifida. Brain 56: , Winston K, Hall J, Johnson D, et al: Acute elevation of intracranial pressure following transection of non-functional spinal cord. Clin-Orthop 128:41-44, 1977 Address reprint requests to: Peter Hall, M.B., B.S., Division of Neurosurgery, Indiana University Medical School, 1100 West Michigan Street, Indianapolis, Indiana J. Neurosurg. / Volume 50 / February, 1979