Management of split cord malformation in children: the Lyon experience
Split cord malformation (SCM) is a rare congenital spinal abnormality. Clinical presentation varies. Other congenital defects can be associated. Management is surgical.
Material and methods
We retrospectively reviewed all our SCM cases and reported our experience for its management. From 1990 to 2014, 37 patients were operated. Five situations lead to the diagnosis (orthopedic disorders (n = 8), orthopedic and neurological disorders (n = 16), pure neurological disorders (n = 5), no symptoms except cutaneous signs (n = 7), antenatal diagnosis (n = 1)). Scoliosis was the most common associated condition. The level of the spur was always under T7 except in one case. There were more type I (n = 22) than type II (n = 15) SCM.
Patients with preoperative neurological symptoms (n = 21) were improved in 71.4%. Five out of nine patients that had preoperative bladder dysfunction were improved. Eleven patients needed surgical correction of the scoliosis.
For us, the surgical procedure is mandatory even in case of asymptomatic discovery in order to avoid late clinical deterioration. In any case, the filum terminale need to be cut in order to untether completely the spinal cord. In case a surgical correction of a spinal deformity is needed, we recommend a two-stage surgery, for both SCM type. The SCM surgery can stop the evolution of the scoliosis and it may just need an orthopedic treatment with a brace.
KeywordsDiastematomyelia Diplomyelia Split cord malformation Surgery Tethered cord syndrome Bone spur
Cerebro spinal fluid
- CT scan
computed tomography scanner
Magnetic resonance imaging
Split cord malformation
Diastematomyelia (DSM) is a rare congenital spinal abnormality that was first reported in the literature by Ollivier in 1837 , sometimes also identified in literature as diplomyelia (DM). In 1992, Pang et al. proposed a new classification, the split cord malformation (SCM), with two types (I and II). According to the authors, type I SCM consists of two hemicords, each contained within its own dural sac and separated by a dura-sheathed rigid osteocartilaginous median septum. In type II SCM, the two hemicords are housed in a single dural tube and separated by a non-rigid fibrous median septum .
Not exceptionally, SCM may remain longly asymptomatic until accidentally recognized. In most cases, however, it is suggested by cutaneous markers , namely hypertrichosis with the typical faun tail that, in our experience, is a pathognomonic sign or may present with a combination of neurological, orthopedic, or urological signs and symptoms of variable severity. Other congenital malformations such as myelomeningocele (MM) or lipomyelomeningocele may be associated in a few subjects.
Usually, the diagnosis is made the pediatric age. Females outnumber male .
The management is controversial in asymptomatic accidentally recognized subjects while the surgical correction is preferred in symptomatic cases. The surgical correction consists of removing the bone spur that separates the two hemicords in order to relieve the pressure on the adjacent nervous and vascular structures with the aim of reducing or curing the neurological deficits in symptomatic patients or at least preventing their progression. In children presenting with progressive scoliosis or kyphosis, the main aim of the operation is to stop the evolution of the spinal deformities. It is debated whether the malformation needs to be corrected in asymptomatic subjects or in children with non-progressive spinal deformity, although several authors consider the surgical correction mandatory even in these cases [9, 10, 11].
In this article, we retrospectively review all SCM cases dealt within our department from 1990 to 2014.
Patients and methods
All the cases of patients who underwent SCM surgery in our department between 1990 and 2014 were considered for this retrospective analysis. The study was approved by the local ethical committee. All patients’ data for this retrospective study were anonymized; therefore, the study was exempted from the need for patient consent. Patients’ gender, age at diagnosis, clinical presentation, associated conditions, type of the malformation (type I or II), spur level, age at operation, surgical modality, post-operative complications, and outcome (orthopedic, neurological, and urinary tract) were collected.
Number of patient
Mean age at diagnosis (years)
1 month–16 years
Patient under 2 years old
SCM type I/type II
Clinical presentation at diagnosis
Number of patient (%)
Bilateral leg paresis
Unilateral leg paresis
Urologic disorders (MM excluded)
Description of the cutaneous signs at diagnosis
Number of patient (%)
All patients underwent the surgical correction immediately after diagnosis, even when asymptomatic. The infant with antenatal diagnosis was operated at 1 month of age because he suffered two episodes of urinary infection that were attributed to bladder dysfunction. There were no differences in clinical presentation between patients with type I and type II malformation.
Eleven patients needed surgical correction of the scoliosis that was carried out in all of them only after having removed the malformed bone spur or fibrous band separating the two hemicords. The remaining seven subjects with scoliosis could be managed with a brace.
Postoperative clinical controls were made at 3 and 6 months after the surgical correction and then, every year, every 2 years and every 5 years, depending on the clinical evolution. Postoperative spinal MRI or CT studies were carried out at 3 months and then yearly throughout the entire postoperative follow-up. For the vesical function evaluation, the necessity of complementary examination was accessed by our urologist. Patients were referred to the orthopedist in order to follow up the spine growth and to check for scoliosis if not present at diagnosis. In the cases where the scoliosis was present at diagnosis, the orthopedist’s follow-up consisted in a yearly outpatient consultation with a full spine X-ray. Follow-up ranged from 2 to 20 years (mean 10.2 years).
Under general anesthesia, all the patients (symptomatic and asymptomatic) underwent to the correction of the malformation either in prone or genupectoral position. After a midline incision (approximately 10–15 cm) centered on the level of the abnormality, a subperiosteal exposure of the spinal process and lamina was performed. Then, a laminectomy or laminotomy (at least four laminae) was done in order to expose the spur. Laminotomy was preferred in young children.
The filum terminale was cut at the lower lumbar level and sent for anatomopathological examination. Macroscopically, this structure was tight with an abnormally fibrotic consistence. When present, other associated abnormalities such as lipomas or dermal sinuses were removed as well in order to minimize the risk of retethering. If the lesion was too far from the SCM, a second incision was performed.
No postoperative wound drain was used, and the patient is allowed to get up the day after the surgery in function of pain (no postoperative mobilization restriction).
No perioperative neurophysiological monitoring was used because it was not available at the time of the surgeries.
We did not have any perioperative mortality, surgical complications, and blood transfusion.
Postoperative complications were transient. We observed three cases (8.1%) of pseudomeningocele that were resolved conservatively with compression dressing and diuretic medication for 1 week. Four children cases (10.8%) complained of transitory motor worsening. There were two wound infections (5.4%) without meningitis that were treated with standard antibiotics.
In no case, there was regrowth of the bony spur at the late MR or CT control.
The neurological long-term outcomes were overall good. The 16 patients with no preoperative neurological symptoms were not worsened and had stable neurological condition. In the group of patients with preoperative neurological symptoms (n = 21), 15 (71.4%) were improved, 5 (23.8%) stabilized, and 1 (4.8%) worsened. Generally, motor deficit improvement was observed after the first 6 postoperative months and continued during the first 2 years after the operation.
The long-term urinary function outcome was good as well. No patient who did not have preoperative bladder dysfunction worsened. If we excluded the seven MM patients, five out of nine patients with preoperative bladder dysfunction were improved clinically, three were stabilized, and one worsened during the follow-up period.
Embryology and associated conditions
The actual incidence of DSM is difficult to estimate as this rare congenital malformation has been described in literature under different definitions and often underreported or only mentioned as an additional finding of other more common malformative conditions such as MM or lipomyelomeningocele. Indeed, for many years, DSM was distinguished from DM as they were considered to represent two different pathological conditions. DM was regarded to be a spinal cord abnormality characterized by a real duplication of the cord, contained in a unique dural sac, in which each hemicord had its own paramedian and lateral nerve roots. On the other hand, the term DSM was reserved to cases where the spinal cord was divided into two hemicords, each one being contained in its own dural sac. In 1992, Pang published a unified theory in which the distinct definitions of DSM and DM were abandoned in favor of the unique denomination of SCM . The author propounded that all cases of SCM had the same embryological origin and that the presence of one or two dural sacs, the presence or absence of paramedian nerve roots, the presence of a rigid osteocartilaginous median septum, or the presence of a non-rigid fibrous median septum did not imply a different ontogenetic error. The unified theory of Pang has gained a wide favor in recent times after a long debate that was characterized by the proposal of several theories, namely the split notochord syndrome theory, the accessory neurenteric canal theory, the hydromyelic theory, the twinning theory, the overgrowth theory, and the neuroschistic blebs theory [2, 3, 4, 5, 6, 7]. By examining the findings observed in his patients and taking into account the embryologic development, Pang concluded that the SCM originates from a splitting of the notochord in agreement with the experimental conclusion of Rilliet [2, 7]. According to Pang’s theory, the splitting of the notocord is caused by the formation of adhesions between the ectoderm and the endoderm leading to an accessory neurenteric canal that condenses an endomesenchymal tract which divides the notochord finally resulting the formation of two hemineural plates . The two types of SCM described by Pang, according to the presence of a spur that can be osseous cartilaginous or fibrous [2, 9]. Three different positions of the placode in SCM are described: the first one in which the placode is terminal, the second one in which the placode is not terminal, and the third in which a fibrous band travels with the fibrous septum to the base of the meningeal outpouching with the placode at the level of the malformation [12, 13]. In our series, the neural placode was situated caudally to the level of the split cord in all cases, and we never saw spinal roots on the median side of the hemicord. SCM is associated with other neural tube defects such as MM in 5 to 36%. In our series, seven patients (18.9%) had associated MM, five (13.5%) had a lipoma, and three (8.1%) had a dermal closed fistulous tract.
In case of associated MM, we proceeded to repair this malformation within the first 48 h of life to decrease the risk of infectious complications, while the surgical correction of SCM, when located far from the placode, was delayed of a few months. Similarly, delayed surgical correction was performed in case of an associated lipomyelomeningocele even when recognized at birth. In this last instance, however, attention was paid to not postpone the operation excessively late. Actually, in our experience, the surgical treatment of a lipomyelomeningocele was less risky during the first year of life than in older age when the increased adherence of the abnormal fat tissue to the nerve roots make the surgical manipulation more tedious and dangerous in case of a dermal closed fistulous tract; it must be resected during the correction of the SCM.
Concerning the spur, in our series, it was located in the lumbar region in the great majority of the cases and always inferior to T7. We have not observed cervical spur as others . The cervical localization is mostly accompanying Klippel-Feil syndrome that results from a more severe mesenchymal trouble with fusion of the vertebral bodies .
Split cord malformation and tethered cord
The tethering of the spinal cord in SCM depends on the spur (bony or fibrous) which impedes the up and down movement of the spine and on the nearly always association with a fatty or tight filum terminale as we proved in our series by systematic anatomopathological examination of this structure. This dual cause of spinal cord tethering should then be managed possibly in the same operation .
Surgical technique and indications
The timing and technique of SCM correction were and still are a subject of debate. It is generally accepted that early surgery leads to better results so that some authors will advocate prophylactic surgery [11, 12, 16, 18, 19]. Other authors, however, recommend a more careful attitude because of the surgical risk, especially in still asymptomatic children . Nevertheless, such a cautious attitude should be weighted taking into account the possibility of non-reversible neurological deficit or progression of an associated scoliosis. Actually, in 78% of our patients the diagnosis of SCM was due to an actual clinical progression that suggests the evolutive nature of the malformation in the great majority of cases. A further demonstration of the risk of a wait and see policy is provided by many reports of the literature related to patients becoming symptomatic in adult life [12, 21]. In this series, we adopted a prophylactic approach which appears to be justified by the results we achieved. None of our asymptomatic children (n = 7) were damaged by the operation. All but one patient with neurological deficit at the time of surgery were stabilized or improved. Only one patient who had already a neurological deficit at the time of the treatment presented a worsening with an incomplete recovery. Among patient with neurological deficit and scoliosis, the surgery did worsen the neurological status. Nearly 90% of patients with sphincterian disorders before the surgery (n = 9) were either stabilized or improved, a result which further supports the hypothesis that the neurosurgical treatment prevents the natural worsening of such a function in SCM.
The surgical treatment of SCM is technically demanding and requires an expert surgeon to avoid spinal cord traumatism during procedure. The malformation should be approached from the nearest intact superior and inferior spinal cord segment moving towards the area of the bone spur or splitting fibrous band slowly. In SCM type I, the bony spur must be removed before opening the dura mater. The dissection of the spur from the dura mater must be done with extreme caution in order not to damage both hemicords. When the bony spur is removed, the dura mater is opened on both sides and suspended. Then, the goal is to recreate only one dural sac. The margins of the internal sheet are sutured on the midline in front of the two hemicords to achieve a water-tight suture of the anterior wall of the unique sac to be created. The posterior wall is finally reconstructed suturing the lateral margins above the two hemicords. We do not use artificial dural patch because, in our experience, it is large enough to avoid the use of this material to recreate a large perimedullary space in order to reduce the risk of late retethering unlike in the surgery of conus lipomas. We consider this maneuver relevant to prevent the re-formation of the bone spur which was propounded to depend on an osteogenic activity of the dura mater in contact with the base of the removed spur. Indeed, we have never observed a regrowth of the bony spur when it is removed completely as others had .
In case where the SCM is near the filum, this structure is sectioned from the lower aspect of the operative field prior suturing the posterior wall of the dural sac. If the SCM is far from the filum terminale, a second additional incision at the L5-S1 level is required.
Concerning the risk of postoperative deficits and their relationships to the type of SCM, some authors have reported a major incidence in case of type I . Even if our surgical results were not different in the two types of SCM, we agree on the major surgical difficulties when dealing with type I SCM. Furthermore, the level of the spur can condition the surgical risk with an increased neurological risk in higher location.
Besides sensorimotor deficits and vesico-sphincter anomalies, SCM is often associated with orthopedic abnormalities, namely scoliosis, which often are progressive especially when there is a malformation of vertebrae like a butterfly vertebra. About one third of our children presented with a progressively worsening scoliosis. In cases, surgical correction of the scoliosis is needed; we strongly recommend untethering the spinal cord both at the level of the SCM and the filum prior to the orthopedic procedure in order to prevent neurological complications. Also, neuroimaging studies of cervicocranial junction should be obtained before the operation on the spine in order to exclude a possible, though rare, Chiari type I malformation which also may require a preventive correction. Even though some authors do not consider the treatment of SCM necessary prior to the surgical management of scoliosis [23, 24], we prefer to avoid further constraints and straining forces on an untreated malformation; the blood supply of which could be already in a precarious equilibrium and unable to tolerate further ischemic insult such as those that may result from a mechanical stretching of the spine.
Furthermore, if the correction of the spine deformation must be made, we recommend a two-stage surgery as described by Winter in 1974 . First of all, the complication rate described for the one-stage procedure is quite high [23, 24]. In particular, the longer duration of the one-stage procedure might enhance the risk of infection that is already favored by the use of osteosynthesis materials. Also, it should be emphasized that adopting the two-stage procedure, the clinical response to the initial operation to correct the SCM may be so favorable that the subsequent orthopedic correction on the scoliosis might not be necessary anymore as we noticed in 40% of our cases. Moreover, the correction of the spinal deformity in a one-stage procedure may be lower than in a two-stage one . One other argument against the two-stage surgery is that most of the time, the age at surgery is low and therefore the decision to perform the surgery for the scoliosis is not taken. Once again, some patient will improve their scoliosis after the tethering; therefore, our orthopedic team follow the patient closely after the surgery to see if the deformation improves or not.
SCM is a rare and complex congenital abnormality which is difficult to understand from an embryological point of view. In our opinion, surgical treatment is mandatory, and we recommend early surgery to prevent clinical evolution, especially vertebral deformation and neurological and vesico-sphincter deficit. Most of the time, surgery helps symptomatic patients improving their clinical evolution. We recommend always completing the surgery of the malformation with the filum terminale section. In case of severe scoliosis, we recommend a two-stage surgery. To finish, it is important to remember that patients with SCM generally have a good neuropsychological development. This point is of tremendous importance when the diagnosis is done before birth in order to provide good prenatal counseling to parents.
Compliance with ethical standards
Conflict of interest
The authors report no conflict of interest concerning the materials or methods used in this study or the findings specified in this paper. This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors
- 1.Ollivier C (1837) Triate des maladies de la moelle épinière. 3rd edn. pp 189Google Scholar
- 8.Fardon DF (2001) From the clinic...diastematomyelia, when examination of the skin predicts spinal deformity. Spine J Off J North Am Spine Soc 1Google Scholar
- 17.Barutcuoglu M, Selcuki M, Selcuki D, Umur S, Mete M, Gurgen SG, Umur (2015) Cutting filum terminale is very important in split cord malformation cases to achieve total release. Childs Nerv Syst ChNS Off J Int Soc Pediatr Neurosurg 31:425–432. https://doi.org/10.1007/s00381-014-2586-1 CrossRefGoogle Scholar