Prenatal ultrasound examinations can in most open spinal neural tube defects (NTD) in infants be declared before the  moment of delivery when the lower half of the trunk and legs make their appearance. The lesions vary in size and location and that in addition to the asymmetrically altered neurological condition results in no two infants appearing similar.

The designation spina bifida occulta is a radiologic one and refers to a defect of the spinous process and/or associated posterior neural arch, usually in the thoracic or lumbar spine, at one level only. It is allegedly found in 5 to 36 percent of the general population and as a rule has few if any clinical implications. The significant forms of closed spinal dysraphism are characterized by a more disrupted appearance of the posterior arch components at two or more levels. In the clinical context, spinal dysraphism is an umbrella term used to designate all the forms, open and closed, of spina bifida. It implies, however, splaying of the pedicles and laminae associated with nonfusion and various defects and disorganization of bony spinal elements. The dysraphism may thus be occult  with or without cutaneous hallmarks or open, with rudimentary neuroectodermal tissue visible somewhere along the spinal axis.

Meningocele, the simplest form of open neural tube defect, is characterized by a cystic lesion which consists of meninges only and contains cerebrospinal fluid (CSF) which is in continuity with that in the spinal canal but has no neural tissue within its confines. The neural tissue may or may not be visible at the base of the lesion, but in any event, the unperturbed spinal cord runs through the ventral component of the dysraphic spinal canal. This entity is one-tenth as frequent as myelomeningocele and is rarely associated with hydrocephalus, but sometimes will camouflage a "second lesion."

The far more common form of open tube defect is the myelomeningocele (a term used synonymously with spina bifida aperta, spina bifida cystica and open neural tube defect). As the term implies, there is some form of cyst apparent, even if it collapses at birth or shortly after. Rudimentary dura and leptomeninges have developed around and are attached to the malformed neural tube. This visible abnormal tissue is usually covered by a reactive response consisting of gliosis and dysplastic axons and cell bodies. At the rostral end of the sac, the spinal cord can be traced as it exits from the spinal canal; caudally, the placode may end within the meninges, extend out a thickened filum or penetrate the spinal canal to continue on its course. The actual circumstance is determined by the level of the lesion. Multiple roots, which appear quite healthy, are found for the most part on the ventral surface of the placode, as they seek their normal dural penetration sites. There are, in addition, roots which wander aimlessly to the lateral and dorsal dural encasement and end blindly in that membrane, subcutaneous tissues or occasionally superficial layers of the paravertebral musculature. The interior of the sac is frequently honeycombed by arachnoid strands and at least one or two anomalous vessels, usually veins.

Myeloschisis, or rachischisis, is a term reserved for the infrequent circumstance of a large filleted placode without any encasing meninges. The lesion is usually oblong. with the open central canal constituting a median furrow and crossing the thoracolumbar junction. In its worst form the condition allows the examiner to peer up the rostral cleaved spinal canal. Early hydrocephalus and a severe neurological deficit are present.

Given this description. it is understandable that myelomeningocele has been described as the "most complex, treatable, congenital anomaly consistent with life. Its management taxes the emotions of the parents, the spirit of the child, the patience of the surgeon, the reserve of the physiotherapist and the ingenuity of the orthotist.

Incidence

During the past tow decades, if indeed not longer, there has been a substantial global change in the incidence of open neural tube defects. For example, the defect historically has been more prevalent in the United Kingdom, especially Northern Ireland, Wales and Scotland, where the incidence during the 1970s was approximately 2 per 1000 live births. In North America, the incidence has been just less than I per 1000, varying from the east coast (where it is higher) to slightly lower in the west. Myelomeningocele is much less common in Japan (0.2 per 1000) and rare in Manila (0.03 per 1000). By the 1980s anecdotal and reported experience from many North American and European centers indicated that the incidence had declined by at least 40 percent. In some regions of the United Kingdom the risk had dropped to 0.6 per 1000. This decline seems to be more than can reasonably be accounted for by antenatal screening and termination of pregnancy.

One of the curiosities of the condition, which lends credibility to the role of environmental causative factors, is its changing incidence in migratory populations. For example, Irish couples born in Boston have relatively high rates compared with other Bostonians but lower rates than first-generation immigrants from Ireland. The incidence among Japanese in Hawaii is higher than in Japan itself. In Australia, immigrants from Britain have higher rates and immigrants from the rest of Europe lower rates than native Australians. Given these international geographic variations. clinicians may find the rules breaking down at the local level within North America. Regions with alleged high incidence of the condition may represent nothing more than areas where young couples are settling and raising families or where there are pockets of low socioeconomic conditions and poor nutrition.

Epidemiology

It is no small comfort to parents of an infant disabled by spina bifida to learn that the cause of the condition remains unknown. There is no identifiable cause for neural tube defects. A study of the available clues yields an interesting blend of genetic, geographic, circumstantial, environmental and dietary factors. It is not always easy (or wise) to separate one alleged influence from another.

There is no sex preference for affected offspring, and earlier suggestions that children with myelomeningocele were born to women at the extremes of the childbearing years or to primiparous mothers are suspect. On the other hand, it has been observed on occasion that the pregnancy which immediately preceded the delivery of a child with a tube defect ended in spontaneous or therapeutic abortion. As yet there are no supporting data to indicate whether the incidence is any greater for this particular group of mothers than it is for the obstetric population at large.

The ethnic differences are well established. There is a high frequency of the condition in the United Kingdom, northern India and Egypt; an intermediate frequency in much of Europe; and a low occurrence in blacks.

Studies on parental consanguinity as well as on twins have produced little useful information. Family studies, however, are beginning to indicate that the mother of one affected child or the offspring of an affected female have a 3 to 5 percent risk of recurrence for any form of spinal dysraphism.

The list of environmental factors is endless, and attempts have been made to relate the incidence to geography, opportunity, diet and social mores. The social structure of affected families has been examined, but the evidence is conflicting.

An interesting observation is the existence of seasonal trends. In Great Britain, the incidence of neural tube defects has, over a survey period of 10 years, been greater in births occurring in December, January and February than it is during June, July and August. This relation to season is reversed in Australia.

Hence, epidemiologic studies have thus far yielded interesting observations but few conclusions as to trends in the occurrence of myelomeningocele.

Etiology

Present clues suggest that neural tube malformations arise as the result of exposure of embryos genetically at risk to additional intrauterine environmental triggers. As it appears unlikely that anything can change one's genetic background, preventive measures taken against neural tube defects should be directed at eliminating culpable factors in the environment or at least avoiding them.

The search for a responsible teratogen has considered a number of alleged dietary causes, including canned cooked meats, white bread, canned peas, ice cream, tea and blighted potatoes. The thesis that folic acid deficiency might in some way be associated with faulty neural tube closure has been carefully scrutinized during the last 8 years. Four structured interventional trials showed that multivitamin or folic acid supplementation given particularly in the periconceptual period to women who had children with neural tube defects reduced the recurrence of the defects in subsequent pregnancies. It is not only becoming clear that it is possible to prevent some of the serious congenital abnormalities associated with the neural tube, but also, as stated more recently, women who received periconceptional vitamin supplementation had nearly 50 percent fewer congenital malformations of all types. All women of childbearing age who are capable of becoming pregnant should consume 0.4 mg of folic acid per day for the purpose of reducing their risk of having a pregnancy affected with spina bifida.

Embryology

Conceptually, any developmental defect involving the central nervous system can be identified as a neural tube defect. Thus, disorders of neural tube fusion should properly be called neurulation defects, the process of neurulation being completed by stage 12 (26 to 30 days of gestational age). The critical period in this process is between stages 8 and 9, when folding and closure of the neural tube occurs. In particular, most defects of neurulation happen between 18 and 21 days of gestation.

The theories that explain neurulation defects are grouped into those that hold that the neural folds never approximate and those that hold that there is a rupture of the fused tube, the latter condition more properly being termed a postneurulation event. Developmental arrest or overgrowth of neuroepithelium are proffered explanations for faulty neurulation. Alternatively, reopening of the closed tube because of increased pressure in its lumen is suggested as the defect responsible for several midaxial developmental lesions of the brain, spinal cord and spinal column. While most evidence relating to this problem has been obtained from experimental studies, pathologic assessments of human fetuses support a concept of nonclosure of the neural tube characterized by eversion, overgrowth of neural tissue, and abnormal cell orientation. Thus, the nature of the cranial or spinal defect is determined by whether it is covered with skin; if it is not, it is a neurulation defect; if it is, it has arisen after neurulation.

Many suggestions attempt to explain the coincident anomalies of the central nervous system in infants with myelomeningocele. In particular, it is suggested that hydrocephalus and the Chiari malformation arise because of developmental mechanical influences either up- or downstream from the neurulation fault or as the consequence of a well-orchestrated, longitudinal sequence of events. A unified theory of the cause of the Chiari malformation has been proposed. It is based on abnormal neurulation, failure of apposition, and transient occlusion of spinal neuroceles, leading to failure to maintain distinction of the primitive ventricular system. In the case of the rhombencephalon, this alters the inductive effect of pressure on the surrounding mesenchyme that in turn results in a small posterior fossa.

Early Detection

In 1972. Brock and Sutcliffe recognized that the concentration of alpha-1-fetoprotein (AFP) in amniotic fluid was substantially increased in a large majority of 37 third-trimester pregnancies surveyed in which the product had been a child with spina bifida, anencephaly or hydrocephalus. Human AFP is normally synthesized by embryonal liver cells, the yolk sac and the gastrointestinal tract. It presumably enters the amniotic fluid from fetal capillaries and in the case of open tube defects, does so via a direct communication between CSF and amniotic fluid between the fifth and ninth weeks. As AFP also crosses the placental barrier and may be found in maternal serum, early serologic recovery from mothers at risk is also possible. AFP elevation reaches its peak at 12 to 14 weeks of gestation: thus the amniotic fluid should be examined by the sixteenth week. Supplemental gel-acetylcholinesterase testing, when applied to amniotic fluid samples with positive AFP, substantially reduces the number of false-positive results. with only a small loss in the detection of open neural tube defects. While AFP is not a specific marker of a neural tube malformation, its measurement via blood and amniotic fluid testing is recommended for women at high risk of having a child with a neural tube defect (that is, those who have a child, a parent or a sibling with such a defect.

Simultaneously, high resolution ultrasound performed after the 20th week will detect about 80 percent of fetuses with open spina bifida. Unlike AFP testing, the ultrasound may also uncover some cases of skin-covered meningocele or lipomyelomeningocele. Together, amniocentesis and level II ultrasound will detect 96 to 100 percent of cases of spina bifida aperta. But, practically, the confirmation of this diagnosis with the associated hydrocephalus seldom occurs before the 24th week of gestation at which point the pregnancy and fetal management decisions become most pending.

It must also be appreciated that AFP examinations alone will not detect skin-covered abnormalities such as the lipomyelomeningocele, dermal sinus or diastematomyelia. Similarly, a small fat and caudally positioned myelomeningocele may escape notice during ultrasound examination by virtue of the minimally distorted local anatomy. In either of these circumstances, there are less devastating neurological consequences: when the specific operative repair is completed, the outcome is promising.

Such screening programs have allegedly reduced the frequency of spina bifida cystica, but only if the affected parents agree to termination of the pregnancy. Moreover, since as many as 90 percent of affected infants with myelomeningocele are born into families previously unaffected, these pregnancies would not be caught in anything other than a universal routine screening program.

Associated Anomalies

There are a multitude of possible associated anomalies in the child with an open neural tube defect of the spine, but most invoke the developing nervous system or its mesodermal surround. That is, congenital anomalies in other major body systems are the exception rather than the rule.

Hydrocephalus is present in 80 percent of children with myelomeningocele and relates to any of the described variants of the sylvian aqueduct, as well as the Chiari malformation, which itself is almost invariably present. In addition to the structural features of the latter, the brain may also show micropolygyria, cerebellar dysgenesis, agenesis of the corpus callosum, cysts within the septum pellucidum or behind the third ventricle or a midline lipoma.

The skull anomalies include craniolacunia, a small posterior fossa with an enlarged foramen magnum, a low attachment of the tentorium and a low-lying torcular. These facts should be borne in mind by the surgeon performing a decompression for the Chiari malformation; the surgery of that condition usually necessitates not posterior fossa exposure but rather upper cervical laminectomy.

The vertebral anomalies may extend far beyond the obvious region of the myelomeningocele. The possible combination of bony abnormalities, including anterior or posterior fusion defects, hemivertebrae, butterfly vertebrae, widened interpediculate distance and partial or complete vertebral fusion, is dazzling and may exist at other levels than merely that involved with the obvious surface lesion. In its worst form the spinal axis may appear telescoped, so that the head seems impacted on the trunk and the lumbar spine seems almost nonexistent. The adjacent ribs may sometimes be absent or may show bifurcation or duplication.

Similarly, segments of spinal cord beyond that obviously involved have been examined and associated lesions noted. The frequently associated "second lesion" in many instances results in delayed deterioration of neurological function. The upper cervical cord usually shows the characteristic features of the Chiari malformation, but in addition, there may be hydromyelia extending through its length and occasionally into the midthoracic components. The relation of hydromyelia to untreated or inadequately treated hydrocephalus and secondary neurological deterioration is now well established.

Alternatively, doubling of the spinal cord has often been encountered. This may or may not be associated with genuine diastematomyelia. The second lesions, which often lie rostral to the obvious defect and may not be visible at the time of the first operation, include the preceding lesions as well as epidermoid or dermoid cysts, intraspinal lipoma and assorted bands and adhesions.

Obviously, any combination of these is possible. For example, an infant with an obvious lumbar myelomeningocele may have at a higher level a small meningocele manqué, dermal sinus or subcutaneous lipoma. In this circumstance treatment priority is given to the life-threatening lesion or the one which is productive of the most severe neurological defect. Alternatively, infants may have a complex dysraphic condition at the site of the open defect, consisting of any combination of myelomeningocele, lipomyelomeningocele, diastematomyelia and ectopic development of renal or intestinal tissue.

There are case reports of associated systemic anomalies in the gastrointestinal, pulmonary, craniofacial and cardiovascular systems, but these are infrequent. Perhaps the most common organ abnormality in children with myelodysplasia is that in the genito­urinary system, in which hydroureter, hydronephrosis or ureteral reflux is a secondary effect of the long-standing neurogenic sphincter.

Clinical Assessment

In some circumstances the neurosurgeon will be asked to provide prenatal counselling to parents whose fetus at 30 weeks gestation has a myelomeningocele and associated hydrocephalus. It has been agreed that by then termination of pregnancy is no longer an option, and a sense of urgency may arise that the fetus should be delivered as soon as possible so that corrective measures can be taken and further damage to the nervous system minimized. The surgeon should stand fast in favour of a pregnancy taken to term that produces a baby whose pulmonary maturity will not create additional hazards. Only at that stage can the anatomic upsets of the ultrasonography be reliably matched with functional performance and treatment decisions made rationally.

Otherwise, the neurosurgeon is most frequently contacted by the referring pediatrician from the delivery suite concerning the child with a myelomeningocele. The general neonatal assessment having been concluded, the infant is kept warm and placed prone, and a moistened, sterile saline dressing is applied to the open defect.

The neurological assessment presupposes that the child is not cold or affected by maternal sedation. The cranium is assessed for the overt features of hydrocephalus, and the first serial head circumference plotting is obtained. The nature of the infant's cry and the arm and hand function are noted in case these features deteriorate subsequently as a result of the Chiari malformation.

The lesion site and size are next determined. Forty-five percent of lesions cross the thoracolumbar junction, 20 percent lie over lumbar segments, 20 percent cross the lumbosacral region, 10 percent lie entirely over the sacrum and the rest are found at more rostral levels. Occasionally, more than one lesion will be encountered on the spinal axis, with the more caudal one being larger and more clinically significant. The circumferential shape, placode size, sac integrity and extent of marginal epithelialized dura are all noted for surgical reconstruction. The spinal column is examined for evidence of early scoliosis, kyphosis, telescoping and visible and palpable prominent laminae scattered at the lateral margins of the defect.

The neurological assessment of the trunk and lower limbs is based on the segmental innervation of the lower limb muscles and an awareness of the different patterns of neurological abnormality that may be encountered. The examiner must be aware of stereotyped patterns of reflex spinal movement, usually elicited when somatic tissue at or below the level of the lesion is stimulated. If it is necessary to arouse the child, stimulation is best applied to the scalp or the skin of the shoulders and arms. The child who is hungry and crying will show spontaneous and active movement of intact muscle groups in the legs corresponding to the preserved neural segments. The infant with a lesion at T12 or above will have fail legs: L1 to L3 function is required for hip flexion, L2 to L4 for knee extension and L5 to S2 for hip extension and knee flexion. Plantar flexion and the function of the intrinsic muscles of the feet require preservation of the sacral roots.

Neurogenic hip and foot deformities are assessed, along with joint mobility and muscle bulk loss in the thighs and calves. Whatever the final results of the motor and sensory testing, the surgeon should expect that the discrepancy will be asymmetrical. If the discrepancy covers more than a few root segments, one must be suspicious that there is an associated lesion within the defect itself (such as cord duplication, lipoma or diastematomyelia) or another fault at more rostral levels.

Sensory examination is too reliant upon subjective interpretation to be a benchmark for future comparisons. Moreover, a case can be made for not utilizing any form of noxious stimulation during a child's sensory examination given its psychological impact and thus low yield in an era characterized otherwise by exquisite neuroimaging detail.

In almost every instance the child with a myelomeningocele will have some form of urinary sphincter disturbance. One cannot always rely on the character of the anal sphincter at birth to declare the degree and nature of that impairment. The constant leakage of meconium from a patulous anus usually predicts later difficulties, but bowel training is often more readily accomplished in this circumstance. Frequently, dribbling of small volumes of urine which increases with crying or movements is indicative of future incontinence, whereas periodic micturition with a good stream suggests a possibility of partial continence.  If one leg is normal, then normal bladder function can be expected. Infants who have neither voluntary nor reflex function in muscles innervated from S2 to S4 tend to have complete bladder paralysis. However, the reliable determination of bladder function is often not possible for months after operation, even though early urologic assessment and urodynamic testing are completed.

Approximately 30 percent of children with a myelomeningocele have congenital scoliosis: half of these develop evidence of the scoliosis between 5 and 10 years of age. The initial examination of the spine does not necessarily show the scoliosis, but if it is present, one must be suspicious of major bony structural abnormalities at higher levels in the spinal column. Kyphosis attributed to wide separation and eversion of the pedicles and rudimentary laminae may exist in conjunction with the scoliosis or be independent of it. It is most treacherous when it involves the same segments as the myelomeningocele and interferes with its proper surgical repair. Lumbar lordosis is often not present at birth but develops in later years to compensate for the upright posture, particularly in the child who is sitting and is dependent on a wheel­chair.

There is less dogma now about the laboratory investigations to be performed on the neonate. Most radiologic assessments of the nervous system and its envelope, with the exception of ultrasound examination of the cerebral ventricles, are not required during the early days and have little influence on the initial treatment plans: the urinary tract should receive early radiologic attention instead. Similarly, the bacteriologic review of the surface of the sac and its contents will almost always yield some organism, raising a question about directing treatment toward that organism: but prophylactic antibiotic care seems to have little influence on the infant's subsequent risk of meningitis and ventriculitis.

The examination of the infant with a meningocele is conducted similarly, but normal neurological function below the level of the lesion is expected and, except in rare instances. no evidence of hydrocephalus.

This first testing is only the beginning of a series of continuing evaluations, the emphasis of which will change as the child's problems change; they will later be concerned with the child's needs for education, employment, and social integration.

Radiologic Evaluation

The radiologic examination of the spine of a neonate with a myelomeningocele yields so little that it is not an obligatory preoperative test. At this stage the spine is immature, and radiologic analysis shows nothing more than multi segmental canal widening, with a variety of anterior fusion defects, hemivertebrae, block vertebrae. etc.- features not essential for the early treatment decisions. As time passes the child will be subjected to several roentgenographic examinations of the spine for curvature, growth, and perhaps the search for the "second lesion," so it is pointless to obtain films in the neonatal period.

The same attitude applies to routine skull films taken at this time. They will inevitably show craniolacunia and as hydrocephalus increases, splitting of the vault sutures. Hydrocephalus is better monitored with repeated ultrasound examinations.

The characteristic features of hydrocephalus associated with myelomeningocele are now well established on computed tomography (CT) examination. The expansion of the lateral ventricles, often with minimal dilatation of hook-shaped frontal and expanded occipital horns as well as mild-to-moderate dilatation of the third ventricle, is apparent. The fourth ventricle is not visualized. The region of the cerebellar vermis appears unduly prominent and the petrous bones show scalloping.

Many of these same features are now becoming apparent on ultrasound examination as expertise with the technique develops. This scan similarly shows on sagittal sections smaller anterior horns, so that the ventricle adopts a scimitar outline. The coronal views demonstrate a teardrop appearance of the ventricles and apparent ablation of the interhemispheric fissure. Moreover, this particular study permits the surgeon to determine the dynamic change in ventricular size and thus gauge the need and time for the shunting procedure.

It must be stressed that the neurosurgeon must not abdicate responsibility for the child's continuing care several years later, when it appears that the neurological condition is stable. A few children will deteriorate or demonstrate new neurological signs. At that time, thorough neuroradiologic evaluation following the' 'top down" approach will be required.

Operative Care

Decisions

Faced with an immediate decision which must consider the rights of the child and its future quality of life, the involved physician is on the one hand accused by activists of participating in a "God squad" and on the other reminded by colleagues that children with NTD are high on the list of those who are dubiously rewarded for staying alive. So the physician is again challenged by the incompatibility of his dedicated aims of prevention of suffering and preservation of life.

What is the immediate decision? With occasional exceptions, most experienced paediatric neurosurgeons would concentrate first on repair of the exposed spinal lesion rather than initiate treatment of the associated hydrocephalus. The thrust of the decision, then, is whether the infant's spinal lesion should be closed immediately, on a delayed basis or not at all. It must be abundantly clear that surgery on the myelomeningocele sac is not designed to repair a faulty spinal cord which, at least in the exposed portion, has not matured beyond the fourth week of pregnancy. No one can improve the neurological disability related to that embryologic disorder: but operation limits the possibilities of retrograde ascending meningitis and preserves neurological function intact.

The initial enthusiasm of the Sheffield physicians who in 1959 began a program of immediate closure in all neonates with NTD has now waned. Lorber lost enthusiasm when he re-examined their large series and calculated that only 7 percent "had less than grossly crippling disability and may be considered to have a quality of life not inconsistent with self-respect, earning capacity, happiness and even marriage. Consequently, he developed selective criteria for children who should not be treated, who included those with (1) paralysis at L2 to L3 or above. (2) marked hydrocephalus. (3) kyphosis. and (4) other major congenital abnormalities or birth injuries. Adhering to such criteria. many experienced surgeons calculated that less than 30 percent of the untreated would be alive at the end of the first year. But the contentious corollary to this statement is that while it is true that a large majority of untreated infants do not live long, a significant minority do live, and what is the quality of their survival? It is suggested that they become significantly more impaired from nontreatment than if they had been operated upon early. The arguments for and against a selective treatment policy for infants with myelomeningocele continue to be examined.

Aims

The goals of early operative care for the newborn with an open NTD are to preserve all neural tissue. especially that which may be functional and must not be allowed to dry, to reconstitute normal anatomic barriers and thus minimize retrograde contamination and to control early progressive hydrocephalus. If such treatment is to be instituted, then one might presume the earlier the better. As the threat of superficial infection on the placode accelerates during the first 36 to 48 h. many surgeons would prefer to repair the myelomeningocele within that time. But, some recommend a deliberate delay in operative closure for up to 5 days if necessary, to be certain that the baby is well studied and stable and, most important, that the consequences of the birth defect and the treatment aims are well understood by the infant's family. If operation is planned after a more extended interval, then a wait of 6 weeks is usual, by which time a clean, epithelialized scar has developed and repair proceeds after a ventricular puncture has confirmed sterility of the CSF.

Operative Techniques of Myelomeningocele Closure

The anaesthetic management of the neonate who is about to have early repair of a myelomeningocele is concerned at all times with the maintenance of body temperature and proper fluid balance prior to and during surgery. An intravenous infusion of 10 percent dextrose in water solution should be started. The child is nursed prone in an Isolette with a sterile, constantly moistened saline dressing on the exposed tissue.

Intubation is performed with the child lying on the left side; then it is placed prone on small bolsters. Temperature drift is controlled by an overhead heating lamp during induction, and heating blankets are applied after positioning.

The fundamental principles of neurosurgical technique are observed during the closure of a myelomeningocele. These include tissue debridment, meticulous dissection and manipulation of potentially viable neural tissue, gentle bipolar coagulation where necessary, and the reconstitution of the usual tissue barriers with respect for the patency of the intervening spaces.

Although wide skin exposure may be required at the conclusion of the procedure, it is recommended that until that time, as much of the skin as possible be towelled off, in part to maintain a degree of surgical asepsis, but also to keep the infant warm. The technique that follows, while not the only method of repairing a myelomeningocele, will allow easy closure in approximately 95 percent of cases. In the other 5 percent, the neural plaque dissection and dural closure proceeds similarly, but a reconstructive plastic surgeon may have to design cutaneous flaps for skin approximation.

The dissection begins in a circumferential fashion, as close to the margin of the placode as possible. This dissection is best performed with magnified vision. The incision is carried through the epithelialized meninges, which leads directly into the intradural space. This is neurosurgical "turf." from which the CSF can be drained, the roots slackened and the entire malformed placode, as well as the proximal and distal cord segments, identified. Anomalous roots are accurately defined, and should they end blindly in the dome of the sac. they are divided. Attempts are made to preserve all small arteries, but inevitably one or two laterally displaced, engorged veins will have to be sacrificed. Finally, the possible presence of a second lesion can be determined and appropriately managed.

The perimeter dissection is completed laterally first, leaving the cephalic and caudal poles until the end. Once the placode and adjoining tissues have been freed, the cuff of epithelialized dura is trimmed back flush with the placode. There need not be undue anxiety about residual cells of epithelium adhering to the placode. These seldom give rise to delayed epidermoid or dermoid cysts; rather. the subsequent appearance of an epidermoid or dermoid cyst is more likely related to the dysraphic syndrome generally, in which the cysts were coexistent.

With the placode isolated and the filum divided if necessary, the surgeon can then determine the fit of the placode into the saucer-deformed spinal canal. At the same time, an estimate is made of the amount of dura which must be freed laterally and rolled centripetally like two book flaps. While it is agreed that there are viable cells in the plaque. particularly rostrally, there is no guarantee that efforts to reconstitute the neural tube by microsuture of its everted lips will preserve these cells or improve their function. Moreover, such a technique seems to add to the axial bulk of the lesion, which in turn can be threatened by the overlying tension of the closed dura and skin.

The dura is easily identified from within and separated from its skin attachment, leaving a narrow margin about the dome which nourishes the frail skin edges. The extradural plane is easily recognized as the flaps are rolled medially and the characteristic epidural fat within the spinal canal is encountered. The adherent dura, especially at the rostral end of the fascial and bony defect, must be freed in order that it can be properly approximated. From time to time there will not be enough such material for a watertight closure: in that circumstance. fascia lata or another substitute should be used. The subdural space needs to be as capacious as possible. The dura is closed in a watertight fashion with a fine continuous suture such as 6-0 black silk. Its integrity can be tested with a Valsalva manoeuvre.

The skin is then closed in the direction in which it will most easily reapproximate - in the axial plane, transversely, diagonally or even in a triradiate fashion. The skin is widely undermined in the plane between the subcutaneous tissue and muscle, well around to the anterior abdominal wall if necessary. Haemostasis is obtained, and the tension on the skin edges is tested in the various directions in which it might close. The subcutaneous tissues are approximated with an interrupted dissolvable suture and the skin itself can usually be united with a running intracuticular suture.

A number of other treatment options have been recommended. Many of these concern the reconstitution of the tissue barriers and the associated efforts to induce a degree of spinal stability. For example. it is suggested that fascial flaps be dissected free, rolled medially and closed over dura. This is a difficult task, as the fascia is limited in amount, is widely displaced from the midline just at the site where it is maximally needed. and tends to shred as it is dissected. Moreover, it may form a significant compressive layer of tissue lying on the dural tube. Consequently, the recommendation here is for a meticulous dural exposure and closure to provide the CSF barrier.

Severe open defects are often associated with prominent protruding laminae, which should be removed in order to minimize the compressive effect on the overlying skin. The suggestion that the laminae be fractured and rolled medially in order to protect the tube is too artificial and, as it is difficult to hold these elastic structures in place, they often migrate and compress the skin. Occasionally, a child with a large defect and pronounced kyphosis will require spinal surgery to reduce vertebral angulation, minimize skin ulceration and prevent secondary ventilatory insufficiency. The "kyphectomy" (excision of one or more vertebral bodies) can be carried out simultaneously.

Finally, the reconstructive surgeon may design attractive methods of using various myocutaneous flaps for skin coverage. These are required in a few instances, and when used they heal perfectly. They also provide additional muscle cover over the repaired neural tissues.

After operation, the child is returned to the Isolette and infant feeding schedules begin within 24 h. Prophylactic antibiotics are not prescribed.

The technique for repair of a meningocele is similar to that just described. The sac is usually smaller and dissection of nervous tissue easy, if it is required at all. Any minor herniations of cord substance and/or roots into the sac are isolated and placed back within the spinal canal. The neck of the sac is often small and it is a simple matter to turn a small leaf of dura and close it over the opening of the neck.

Hydrocephalus

Exactly 80 percent of all infants with myelomeningocele have hydrocephalus apparent at birth or develop it within the next few weeks, the larger and more rostral the lesion, the greater the likelihood and the converse applies, so that infants with small sacral lesions have a 50 percent risk. Whether or not the myelomeningocele sac acts as a reservoir for the CSF, which then is ablated at the time of repair, is a moot point. The clinical features of hydrocephalus, if not initially apparent, will become obvious in most instances within 5 to 14 days of the spinal surgery. The diagnosis is confirmed with ultrasound examination of the head and if there is any doubt about the sterility of the CSF, a ventricular sample of CSF is analyzed before placement of a ventriculoperitoneal shunt. The latter is inserted in the standard fashion, with the peritoneal catheter descending subcutaneously in the lateral flank or anterior thorax, so that it lies somewhat remote from the myelomeningocele skin closure.

It does not necessarily follow that the 20 percent of infants who do not show evidence of hydrocephalus do not have ventriculomegaly. A substantial number have enlarged ventricles, but in a non­progressive fashion and unassociated with features of intracranial hypertension. In the absence of overt clinical signs and with a nonthreatened, healing spinal wound, the surgeon can afford to continue assessing the child before deciding upon the need for a shunting device. If the child reaches its first birthday and has not required shunting, it is quite unlikely that it will be necessary to place a shunt in the next few years.

If hydrocephalus is moderately severe at birth or if a delayed closure of the spinal defect is planned, a shunt system may be inserted at the same time as or prior to the myelomeningocele closure. In the latter circumstance, preoperative bacteriologic analysis of the CSF is advised.

Shunt systems inserted in patients with hydrocephalus associated with myelodysplasia are more likely to become infected than those inserted for all other causes of obstructive hydrocephalus. Infection within these shunt systems is usually due to Staphylococcus epidermidis, S. aureus or Escherichia coli and most often appears within the first 2 years. The infected shunt is managed in a standard fashion, with appropriate intravenous antibiotic therapy and externalization and/or removal of the shunt, with eventual shunt replacement.

It had been believed that as many as 50 percent of myelodysplastic children might outgrow the need for their shunt system by their seventh birthday. While it is true that shunt breakage or obstruction in the older child may not declare itself with traditional symptoms. follow-up magnetic resonance imaging (MRI) studies are now revealing a galaxy of changes about the brain base or within the spinal canal in this particular group of children. Many of these alterations relate to faulty CSF circulation that in turn is causing other neurological signs. These can be relieved by a shunt revision.

Chiari Malformation

The Chiari malformation is deemed to be present in all children with myelodysplasia. Recent studies indicate that it becomes clinically significant in 10 to 20 percent of these children, especially those aged 3 months or less. During the early postoperative days, the nursing staff may document clinical evidence of the symptomatic malformation with events usually beginning after the CSF shunt has been placed. Prolonged feeding, choking, regurgitation of fluid through the nose, stridor and a change in the lustiness of the cry and then cyanotic episodes and apnea are characteristic. It must not be assumed that these are the symptoms of shunt malfunction: they are the signs of progressive brain stem failure.

Clearly, there is a part of the Chiari malformation, intrinsic cellular disorganization within the brain stem and upper cervical cord, about which nothing can be done. But notwithstanding the enlarged, scooped foramen magnum, there is compression on the herniated tissues in the upper cervical spine. Not infrequently. a thick, circumscribing and compressive extradural band lies beneath the arch of the atlas. The compressive factor of the Chiari malformation should be recognized for what it is and surgical decompression performed before yielding to neurogenic vocal cord paresis and submitting the child to long-term tracheotomy. The surgical decompression is a cervical laminectomy taken as far caudad as necessary to positively identify cervical cord tissue.

Postoperative Care

Assuming that the operative technique has been meticulous, the surgeon should not be dismayed if the motor function in the legs is somewhat worse following surgery. This is often the case for the first 10 to 14 days, by which time the motor and sphincter performance should have returned to what was seen preoperatively. The creed with regard to operative care is that the child's postoperative leg function is rarely improved by surgery but is seldom worsened.

The main interest in the early days following operation, therefore, is directed at preserving wound integrity as the effective barrier against retrograde infection. If there was not luxuriant skin for closure and the surgical wound is thus frail, the infant is best mobilized in the face-down position, with the legs restrained (if they have active movement), so that tension on the incision is lessened. By the tenth day, the fate of the wound is usually known.

It is not a major calamity if the wound separates over a portion of its distance, provided that there is no CSF leakage and that hydrocephalus, if present, has been treated. Once the shunt has been inserted, the spinal repair wound will close in, albeit slowly, with the application of topical dressings.

For those children who will develop hydrocephalus, the problem, if not apparent at birth, will almost always declare itself within 5 to 14 days of spinal surgery. It usually becomes manifest with the typical signs of infantile hydrocephalus, but occasionally anterior fontanel tension, spread skull sutures, and increasing head circumference will not appear; rather, the spinal operative wound disgorges a volume of clear, colorless fluid. Even under these circumstances, and when the head circumference remains well within accepted growth parameters, CT head examination will show a mild to moderate degree of ventriculomegaly. Shunt insertion is required.

Once the spinal operative wound has healed, the child can be turned from the prone position and placed in a regular cot and the other baseline investigations can be obtained. As at least 90 percent of patients will be incontinent of urine, it is of some importance that the urologist assess the child as soon as possible, to determine the degree of sphincter impairment, the need for assisted bladder evacuation (by regular intermittent catheterization), and ureteral reflux, infection, and bladder trabeculation. Frequently, although not always, an infant will from the outset require antibiotic prophylaxis. Exhaustive surgical procedures aimed at urinary diversion have now given way to the much more acceptable practice of intermittent bladder catheterization and home screening programs for urinary tract sepsis.

Bowel management programs are somewhat more easily achieved and take advantage of normal gastrointestinal physiologic responses supplemented by proper nutrition and laxatives and enemas as required.

Similarly, the orthopaedic surgeon must be asked to assess the child's spinal anomalies and the possibilities of neurogenic hip dislocation, fractured femur, and varieties of neurogenic clubfoot. Stretching and/or casting manoeuvres for the latter frequently begin during the infant's first few weeks of life. Projections are made with regard to the infant's ultimate orthotic needs. The child's ability to ambulate is determined by more than the neurosegmental lesion; factors such as motivation, intelligence, spinal curvature and intractable orthopaedic deformities all substantially influence the child's ultimate mobility. Not infrequently a child who has required extensive brace assistance to stand and ambulate will, at adolescence, choose a wheelchair for greater mobility.

From the time of the infant's first release from the hospital, the neurosurgeon must be dedicated to his or her professional marriage with this patient. The intense level of neurosurgical care persists through the child's first two or three years of life and is then supplanted by more involved urologic and/or orthopaedic treatments. During those first few years, the CSF shunt system is frequently blamed for every problem that befalls the child. Children with spina bifida are no more or less susceptible than other children to the common maladies of childhood, and the clinician should be forewarned that, while shunts do malfunction, the infant may be ill from a urinary tract or upper respiratory tract infection.

A "second age" of spina bifida care frequently begins after the child's fifth birthday. It is personified by a child who has been walking, with whatever assistive devices, and who shows deterioration in gait. Worsening spasticity may be present in these children, or even in those who are severely paraparetic, or the child's scoliosis may worsen. These changes are commonly seen after a period of rapid growth. The belief that these are all "inevitable" should be discounted until the neurosurgeon has had an opportunity to study what clearly is a dynamic problem. The' 'top down" investigative approach, which begins with a CT brain scan, will document the compromised level of the nervous system. The worsening may be related simply to a malfunctioning shunt, with an expanded cerebral ventricular system and the ultimate channelling of CSF into a progressively enlarging hydromyelia. A straightforward shunt revision may reverse the problem. A second lesion, as noted above, may have been hidden from view at the time of the first operation and may be uncovered now by MRI. Additional surgery on this second lesion or even on the severely tethered repair site, usually has a favourable outcome.

The long-term follow-up is best managed through a multidisciplinary outpatient clinic. There the neurosurgeon, orthopaedist, urologist, orthotist and therapist can assess the child together and discuss with one another changed directions for the child's care.

Long- Term Results

7 percent of infants would eventually become self-sufficient and that the next 20 percent would also be of normal intelligence but able to earn only in a sheltered workshop environment. After early operative care, 50 percent of children would survive to the teen years and I5 percent would be minimally handicapped. 75 percent of surviving infants having normal intelligence, bladder and bowel control being achieved by school age in 90 percent of survivors, and more than 80 percent of children becoming community ambulators by school age.

Clearly, there are many determinants of outcome, such as the size and location of the spinal defect, the tempo of hydrocephalus and its treatment, complicating infections, a symptomatic Chiari deformity, surgical philosophy, rehabilitation efforts and family motivation. The larger rostral lesions, early and uncontrolled hydrocephalus, ventriculitis, brain stem compromise with pulmonary aspiration and shunt failure are all related to early demise. The mortality during the child's first decade is 14 percent. Sometimes this occurs precipitously. Over longer periods of time, pulmonary complications, pneumonia and renal failure claim life.

The relation of intellectual development to hydrocephalus has recently been analyzed in detail, as hydrocephalus decreases the intellectual potential. Thus, the child without hydrocephalus or whose hydrocephalus is arrested has a better prognosis for a normal IQ than selectively treated but dependent patients with hydrocephalus, who have a 34 percent incidence of retardation. A common outcome of early hydrocephalus is an uneven growth of intelligence during childhood, with nonverbal intelligence developing less well than verbal. In this regard, myelodysplastic children perform poorly on tasks of persistent motor control and eye­hand coordination. While intellectual potential is negatively influenced by ventriculitis or shunt infection, it is not affected by either the type or the number of shunts required for its control.

One of the primary aims of early operative care of NTD is to preserve neurosegmental levels and thus influence ambulation. Success with ambulation is subject to many of the same factors as intelligence. Theoretically, a preserved L3 level should be compatible with erect posture and orthotic-assisted gait. but, practically, it is children with L5 levels who will be long-term ambulators and enjoy a degree of self-confidence while negotiating public facilities. It must also be recalled that as the neurological level is frequently asymmetrical, so will be the gait performance and required orthoses.

Other associated nervous system symptoms may require assessment in the first few years. These include strabismus and other disjunctive eye movement disorders and seizure problems.

There is increasing concern for the downstream effects of long-term myelodysplasia on the child and its family. It may be a long while before the devastating news of the child's birth defect is finally accepted by the parents, who not unexpectedly suffer guilt, frustration, and unhappiness over their misfortune. Marital instability is all too frequent, as one parent gives full attention to the disabled child at the expense of the partner and the other siblings. Never a year passes without major decisions to be made for the child-more surgery, bladder care, changing mobility aids, coordination of appointments and hospitalizations with personal needs, and educational requirements. Parents show increasing maturity and sophistication and can be greatly assisted by social service agencies and lay parent organizations.