The Normal Position of the Conus Medullaris Does Not Exclude a Tethered Spinal Cord

17 The Normal Position of the Conus Medullaris Does Not Exclude a Tethered Spinal Cord

R. Shane Tubbs and W. Jerry Oakes

Since the inception of the term tethered spinal cord (TSC), most have intuitively accepted that the spinal cord must be located distal to its normal termination.1 Many have declared coni located inferior to the L2 vertebral level and in patients with symptoms of lower spinal cord dysfunction as abnormally low and thus tethered. However, we have previously described a small series of patients in whom symptoms of a TSC were evident clinically yet radiologically their conus was found to lie at a widely acceptable normal anatomical level.^2,3 In fact, after a review of the extant literature prior to our publication in 1993, several authors had operated on patients for symptoms of a distally TSC in whom a conus was either retrospectively or prospectively found to terminate at a “normal” vertebral level.^4–10 Justifying surgical treatment for bona fide symptoms of a tethered cord seems appropriate in the fact that some authors advocate operative intervention for patients who are asymptomatic and are merely found to have fat within the filum terminale.¹¹ Of importance is where should the “normal” conus terminate? To some, below the L1–2 disk space is abnormally low, whereas to others, below the inferior border of L2 is abnormally displaced. From sonograms, one group concluded that a conus termination at the L3 level is indeterminate for abnormality.¹² This is confounded by the fact that the conus at times has no definite tip but gradually tapers to a very thick filum.^13,14 Operatively, however, the caudal end of the spinal cord is identified by the lowest coccygeal nerves as described by Yamada et al.^15,16

Embryology of the Conus Medullaris and Filum Terminale

The human embryo has a distinct tail bud, which gradually regresses, and by 9 weeks gestation it is no longer seen externally.¹³ The tip of the vertebral coccygeal segments contains an epidermal cell rest, the caudal cell mass. Secondary neurulation involving this cell rest apparently only gives rise to the filum and ventriculus terminalis, which may be identified from days 43 to 48, at which time it lies adjacent to the coccyx.¹⁷ By 5 years of age, this “ventriculus terminalis” was found in only 2.6% of 418 children in a magnetic resonance imaging (MRI) study.¹⁸ The filum terminale is composed of primarily pia mater, a neural crest derivative.¹⁹ The filum terminale is created when the caudal neural tube regresses between the ventriculus terminalis and the caudal cell mass and is first seen on day 52.^17,20 The distal conus may also arise from the process of secondary neurulation.²¹ In the term infant, the inferior tip of the conus has been described as being located at the L2–3 interspace in 98% of the cases and at the L3 level in 1.2% of the cases. Presumably, by 3 months of age, the tip of the conus achieves its adult position by “ascending” to the L1–2 interspace.¹ Robbin et al, with sonography, have refuted this idea of ascension during childhood and think that by ∼19 weeks gestation the conus has achieved its adult position.12 Therefore, with these data it is unlikely that significant additional ascent of the conus medullaris will occur after birth. Wilson and Prince have concluded that a conus positioned at L2–3 should be considered normal at any age.²² DiPietro has shown that with sonography, children less than 2 months old had a mean conus termination at the lower third of the L1 vertebral body, children from 1 to 4 years of age had a mean conus termination at the upper third of the L1 vertebral body, and children greater than 4 years of age but less than 13 years also had a mean termination of the conus at the upper third of the L1 vertebral body.²³ Their study also concluded that as the conus ascends throughout childhood, the criteria for conus level should alter depending on the age of the patient. One study using sonography has found that in 92.1% of term babies the conus was above the L2 vertebral level and at the L2–3 disk level in only 6.3%.²⁴ Reimann and Anson reviewed 801 adult spinal cords and found that the conus medullaris is found above the L2–3 disk level in ∼98%, whereas the mean conus lies at the lower third of the L1 vertebrae.²⁵ Saifuddin et al also found the mean termination of the conus at the lower third of the L1 vertebrae.²⁶ It is also said that the conus may tend to terminate differently in various races.²⁷

The filum terminale is an elastic formation usually less than 2.0 mm wide,^13,28 which allows the conus medullaris to ascend during flexion of the spine. Distally, the filum travels to fuse with the dorsal dura mater then continues with this dural sheath as the coccygeal ligament.¹⁹ If it is insufficiently elastic [e.g., abnormally thick, fat laden (decreased elasticity due to increased fibrous tissue)], this supposedly reduces the ability of the cord to move cephalad and thus places undue stress on the distal conus.²⁹ Incidental fat within the filum is seen in ∼3.7 to 17% of the normal adult population.^17,30 In one of our series, fat was found in 91% of TSC patients.³⁰ Although the presence of fat may demonstrate possible pathology, it is the amount of fibrous tissue within this fat that dictates the degree of viscoelasticity of the filum. In addition, the lack of denticulate ligaments along the caudal cord allows for more mobility if caudal forces are placed upon them.³¹ We have found with cadaver studies that the denticulate ligaments, even cephalically, do little to abort either cranial or caudal traction on the spinal cord.³² Supposed failure of involution of the terminal spinal cord and/or failure of the lengthening process of the filum terminale results in the TSC.^1,33

Tethered Cord Syndrome with a Normally Positioned Conus

This entity, first described in 1953³⁴ with a normally positioned conus, is diagnosed on a clinician’s ability to interpret the filum terminale on sagittal MRI as being taut. A more objective finding is a dorsally displaced conus and filum on MRI.^15,16,35 The term tethered has unfortunately been used both for cases in which the tip of the conus medullaris is at normal levels as well as for cases in which it is at abnormal vertebral levels.^1,4 In one series of tight fila, the tip of the conus resided at L2 or above.¹³ Breig³⁶ has shown that the distal spinal cord becomes attenuated with flexion of the pelvis and that this biomechanical feature causes insult to the spinal cord during motion and increase in growth. However, these studies were performed in cadavers that had lost a great deal of their viscoelastic characteristics. We and others have found the distal cord to be hyperemic (indicating stress/tension) following sectioning of fila in patients with a normally or an abnormally positioned conus.^1,17,37 Yamada et al³⁸ have eloquently shown in a cat model that the TSC produces changes in the redox activity of cytochrome a,a₃ with a decrease in oxidation metabolism, presumably from mitochondrial anoxia.

The term tethered usually implies abnormally positioned.¹³ Most definitions of normal cord termination levels are based on the anatomical measurements of Barson, who, quite intuitively,concluded that his findings in embryos were minimally inaccurate due to the hyperextension of the specimens during dissection.39 Many series have suggested that incontinence in children who essentially have normal MRI scans of the lumbosacral region and usually have cystometrographic findings indicative of a neurogenic bladder can be successfully treated by simple sectioning of the filum terminale.^5–8,34,37 These series have had patients with both normal and fat-infiltrated fila.^{5–8,30,34,40} Improvement of bladder function in these series showed improvements of 58%, 44%, 59%, and 67% at 1, 26, 13, and 20 months, respectively. Most patients in these series demonstrated little if any elements of occult spinal dysraphism. However, one should be clear that this is a diagnosis of exclusion and should be carefully considered.

In patients who present with symptoms indicative of TCS yet with imaging demonstrating a “normally” positioned conus medullaris, which can be visualized on sonography by the nineteenth week, the approximate same frequency of the following is observed (patients with normally positioned cord versus those with low-lying conus): cutaneous signatures of occult spinal dysraphism (46% vs 52%), abnormalities of the extremities (39% vs 32%), bony abnormalities (100% vs 95%), dysraphic abnormalities (62% vs 78%), and neurological abnormalities (77% vs 87%).³ We stress that in our series none of the patients presented solely with urological dysfunction.^2,3 One criticism of the normally positioned conus in a patient with symptoms of a TSC is that these individuals may have additional vertebrae in the lumbar region thereby distorting this diagnosis if one does not number vertebrae from a cephalad to caudal direction. Often imaging of the lumbosacral region is only obtained in evaluation of these patients. The degree of potential error introduced into the aforementioned study by the possible occurrence of undetected transitional vertebrae or six lumbar vertebrae was found to be negligible.²² In a review of 1614 lumbosacral spine radiographs the incidence of transitional vertebrae was found to be 8%.²²

In support of biomechanical stress on the conus medullaris without alteration of its termination site at a normal level is the simple concept that a static midsagittal MRI of the lumbosacral region is all revealing in the absence of flexion or extension of the spine. Even with Barson’s historic publication, he stated, “Our data represents the vertebral segments as though they were of constantly equal length both in respect of position and time…. This is obviously not so, although no adequate physical measurements exist of the precise size of the various vertebral segments at varying gestational ages and the normal ranges to be expected.” Barson also makes an important query with regard to conus position by asking whether conus level is a result of an unusually short vertebral column or an abnormally long spinal cord, or whether both factors play a part.³⁹

In our original report of patients with TCS and the conus in a normal position,² we reviewed our experience over 12 years with 73 patients with TCS. Of these, 13 (18%) had a cord termination at or above the L1–2 vertebral disk space. The mean age for this group was 11.3 years. These patients otherwise displayed characteristics usually associated with the patient with an abnormally low conus. The most common neurological complaints were lower extremity weakness and bladder dysfunction, each occurring in 40% of patients. Bowel dysfunction occurred in 30%. Six of these patients had cutaneous signatures indicative of occult spinal dysraphism; four had lumbosacral hemangiomas; one had a lumbosacral subcutaneous lipoma, and one had a midline lumbar skin tag. Five of these patients presented with extremity abnormalities, including three with leg-length discrepancy and two with foot deformities. Adipose tissue was found in 92% of patients either radiologically or histologically.

Radiologically and radiographically, several bony anomalies were identified in this population of patients. Ten of the 13 had bony abnormalities. These abnormalities included scoliosis (three patients), bifid vertebrae (10 patients), hemivertebrae, and segmentation errors. Other nonbony dysraphic anomalies included meningocele manqué in two patients, intradural lipoma in four patients, split cord anomaly (without median septa) in two patients, and terminal syringomyelia in two patients. Twelve of the 13 patients were operated. The average follow-up at that time was 2.2 years. Three patients presented neurologically normal. Seventy-five percent of patients with lower extremity weakness improved with surgery. For urinary complaints, at follow-up, two patients had normal urinary control, one patient had improved control, and one patient in whom there was a neurogenic bladder had no postoperative change. Of patients presenting with bowel complaints, all three had cessation of complaints at follow-up. Two patients presented with back pain that had resolved after their operation. The one patient that did not undergo surgery presented with low back pain and radicular leg pain that had improved at 6 months follow-up. Our follow-up publication on TCS and the normally positioned conus compared this subgroup to patients with TCS and a low-lying conus medullaris, which consisted of 60 patients.

Moufarrij et al have reported a series of patients with findings of TSC in which three patients had a cord termination at the L2 vertebral level.9 Only one of these patients had a fatty filum terminale. Presenting symptoms in these three children were gait disturbance with lower-extremity weakness, leg cramps with foot inversion, and progressive kyphoscoliosis. These patients were noted at operation to have the cut ends of their fila retract 1 to 3 cm. Raghavan has reported 25 patients with TSC in which four had coni superior to the middle segment of the L2 vertebrae.¹⁰ Two of the four presented with urinary incontinence, and fatty fila were found in three of the four, one of which did not present with urinary difficulties. Interestingly, tonsillar ectopia was found in five of nine patients with imaging of the craniocervical junction. Curiously, we have found previously that there is no significant difference in the level of cord termination between controls and children with a Chiari I malformation.⁴¹

Urinary Dysfunction

Although our description of TSC with a conus in the normal position did include patients who had urinary incontinence, this was only in four patients. In our report, three of the four patients had improved bladder function at postoperative follow-up.² Other sporadic reports in the literature have shown that urinary complaints, especially incontinence, are dealt with by sectioning the filum terminale in the face of a conus that is in a “normal” position.^7,8,34,40,42 Khoury et al have discussed sectioning of the filum for this scenario.^5,6 In their first publication, 23 children were operated on, all with normally positioned coni per myelography. Approximately 72% had resolution of their preoperative incontinence.⁵ However, urinary incontinence unrelated to TSC must be considered. For example, some children may have persistent nocturnal enuresis secondary to developmental delay and not TSC.

Conclusion

There is a subset population of the patient with TSC in whom the tip of the conus lies at even liberally accepted normal levels. We would encourage clinicians treating patients with symptoms of TCS not to treat the patient based simply on imaging but imaging coupled with clinical symptoms and physical exam. It is important to point out that in lieu of the many publications of conus termination, one accept that there is no one single “normal” position of the terminal cord but rather a normal range. In addition, one may consider a patient with a supposedly normally positioned conus to actually have caudal descent of the cord in that, if not placed under stress it would have terminated superior to its current location. This would infer that earlier intervention might address an abnormally taut cord prior to its displacement. Perhaps a superior way of interpreting “tethered” cord is to view this as abnormal tension on the cord and not necessarily elongation of the distal cord. Finally, we stress that, with our experience, this group is the exception and not the rule with regard to TSC.