Pediatric bony craniovertebral junction abnormalities: Institutional experience of 10 years.

OBJECTIVE: To study the clinical features and treatment outcome of pediatric patients with bony craniovertebral abnormalities.
MATERIALS AND METHODS: The authors studied 189 consecutive cases of pediatric bony craniovertebral junction abnormalities operated between 2001 and March, 2010.
RESULTS: The pathologies were developmental (n = 162), traumatic (n = 18) and tuberculous (n = 9). Surgical procedures included transoral decompression (n = 118), occipitocervical fusion (OCF, n = 139), C 1 -C 2 fusion (n = 45), and posterior fossa decompression (n = 5). Methods for OCF included contoured stainless steel rods (n = 86), titanium lateral mass screws and plates (n = 47) and steel wires (n = 6). Constructs of all patients of posterior fixation with contoured rods and wires or lateral mass screw and rod who could be followed up were either stable/fused or were fused and stable. No implant failure was noticed among these two surgical procedures. However, 6 patients with C 1-C 2 fusion had broken wires on follow-up requiring repeat posterior fixation. Good neurological outcome was observed even in poor-grade patients. No significant effect on the curvature or growth of the spine was observed at follow-up.
CONCLUSIONS: Pediatric craniovertebral junction anomalies can be managed successfully with good outcomes using a low cost contoured rod and wires.

Introduction

The craniovertebral junction (CVJ), being a transition zone between the cranium and the spine, has a complex bony anatomy and intricate relationship with the major neurovascular structures. Therefore, treatment of various types of abnormalities in this region poses many challenges, especially in the pediatric age group. The clinical manifestations are often delayed into the second and third decade because they are usually subtle and easily missed in children unless looked for specifically.[1-3]

Delayed symptom reporting is another hurdle in early diagnosis. The immaturity of bones and the implications of the posterior fusion on bone growth are major concerns while deciding the management strategies in this age group. Moreover, the radiological picture is confusing as the ossification of the bones is completed only by 9 years of age.[4] The incidence of different types of abnormalities varies with the demographic environment[56] and has ill-defined genetic factors.[78] Different types of abnormalities with a complex pathological bony anatomy need individual management decisions tailored for that particular case. The Indian subcontinent with its varied demographic profile of the population along with a high incidence of infectious pathologies like tuberculosis shows a broad spectrum of these abnormalities. Increased accessibility of health care and rising awareness among the physicians resulted in a higher number of pediatric patients being investigated and referred for the management of these abnormalities. In our study, we analyzed the various CVJ abnormalities in patients up to 20 years of age to study the entire epidemiological, clinical and management profile.

Materials and Methods

The clinical data, operation notes and imaging data of patients were analyzed. We reviewed the case records, discharge summaries and the follow-up OPD files of the patients treated at our institute from 2001 to Mar, 2010. Information regarding clinical features, radiological findings, surgical procedures, type of implants used, bone fusion and neurological status at follow-up was retrieved. Patients up to 20 years of age (age range, 2-20 years; mean age, 13. 27 years; 138 males and 51 females) operated by 5 different surgeons were included in this study. Investigation included dynamic X-rays of the CVJ, plain CT scans with 3-dimensional reconstruction on a 128-multislice spiral CT scanner, and MRI (1.5 T) of the CVJ with whole spine screening. Basilar invagination was diagnosed when the tip of the odontoid process was at least 2.5 mm above Chamberlain's line.[9] The diagnosis of atlantoaxial dislocation (AAD) was made when the atlantodental distance was > = 4 mm for children <9 years of age and > = 3 mm for children >9 years of age seen on lateral X-rays and plain reconstructed CT scans. CVJ tuberculosis was diagnosed based on the clinical profile, systemic investigations which include chest X-ray, erythrocyte sedimentation rate and the Mantoux test and specific CT and MRI features of the CVJ. The radiological profile of CVJ tuberculosis included prevertebral collection, odontoid erosion and presence of granulation tissue.[10-12] Patients diagnosed as having CVJ tuberculosis received antitubercular chemotherapy for 18 months with neck immobilization in a Philadelphia cervical collar. Halo fixation was used if the radiology showed severe destruction or instability. Surgery was considered if there was no response to 3 months of chemotherapy, neurological deterioration or a persistent severe bone deformity. Immediate surgery was considered in cases with tuberculous abscess with respiratory difficulty or in cases having rapid neurological deterioration.[1013] In patients with irreducible fixed AAD or basilar invagination, a short-term traction for 1-2 days was applied.[1415] To start with, 7-8% of the total body weight was applied with gradual increments in weight up to a maximum of 7 kg. Serial lateral cervical X-ray films were taken to monitor reduction. In cases of basilar invagination, a partial descent of the odontoid process by traction made the transoral decompression easier while a complete descent rendered the need for a transoral procedure unnecessary. In reducible AAD, traction was applied in the operating room before anesthesia. In partially reducible AAD, a longer period of traction up to a week was applied for achieving reduction. The operative techniques are described below.

Transoral decompression

Transoral excision of the odontoid process and C 2 body was performed in patients with ventral compression because of basilar invagination or irreducible odontoid dislocation causing compression on the cervicomedullary junction. Transoral decompression was done in 118 patients. Under general anesthesia, traction was either applied or continued if already applied before. The patient was placed in a supine position with slight extension of the head. His/her mouth was opened widely with a retractor (Boyle- Davis's or Crockard's). The palate and uvula were retracted posteriorly and up either by hitching them with silk suture to an infant feeding tube inserted through the nose or by an in-built device in the Crockard's palatal retractor. Incision was performed in the posterior pharyngeal wall in the midline from the lower margin of the clivus to the C 2 body under fluoroscopy guidance. Then using a high-speed pneumatic drill (Legend, Midas Rex), the anterior arch of the atlas, the odontoid process and the lower portion of the clivus were removed. The tip of the odontoid process was lifted off the dura mater, the apical and alar ligamentous attachments were incised and the odontoid process was taken out. The posterior longitudinal ligament and tectorial membrane were also excised and the dura mater was visualized. Hemostasis was achieved using surgicel and/or gelfoam and the posterior pharyngeal wall was closed by interrupted 2-0 vicryl sutures in 3 layers (muscle, submucosa, and mucosa).

Posterior fixation

Occipitocervical fusion (OCF) was done in 139 patients [Figure 1]. Contoured stainless steel rods and sublaminar wires were used in 86 patients. The contouring was performed at the time of surgery using indigenous instruments. Occipito cervical titanium lateral mass screws, rods or plates were used in 47 patients. The type of implant used for the OCF depended on the patient's affordability. Posterior decompression alone was performed in 5 cases with Arnold-Chiari malformation with syrinx with a stable CVJ abnormality where it was felt that instrumentation was not required. Posterior decompression with posterior fixation was done in 26 patients having significant dorsal compression. C 1 -C 2 fusion with steel or titanium wire was performed in 45 patients. Following surgery, patients were electively ventilated overnight for 12 h and extubated the next day. Postoperatively neck immobilization was achieved using a Philadelphia collar for a period of 3 months. Feeding till postoperative day 5 was done through a Ryle's tube. After 5 days, oral feeds were started if the posterior pharyngeal wall wound was healthy. All patients underwent postoperative X-rays and plain CT scans with bone windows. Lateral cervical X-rays were done at 3 months of follow-up to look for bone fusion. In the postoperative period, patients were mobilized as early as possible with a hard cervical collar. Patients were followed up at 3 months, 6 months and then at 1-year intervals. Lateral X-rays of the CVJ in the neutral, flexion and extension position were done during every follow-up visit to assess bone fusion alignment and the curvature of the spine. Stability was defined as: the absence of movement on flexion–extension lateral CVJ X-rays. Fusion was defined as the obvious appearance of bridging bone between the fixed segments on lateral CVJ X-rays [Figure 2]. Stability with fusion was defined as absence of movement on flexion–extension lateral CVJ X-rays along with obvious appearance of bridging bone between the fixed segments on lateral CVJ X-rays. Clinical outcome of the patients was considered as stable, improved or worsened depending on their status on last follow up.

Figure 1

(a) Lateral cervical X-rays of a CVJ showing bone fusion at 1 year of follow-up in a 10-year-old girl following lateral mass screw and rod fixation. (b) Lateral cervical Xrays of a CVJ showing transoral odontoidectomy and posterior fixation using a contoured rod and wire construct in a 3-yearold girl. (c) Lateral cervical X-rays of a CVJ showing bone fusion in a 5-year-old girl who underwent transoral odontoidectomy and posterior fusion with contoured rod and wires. The construct was removed as the patient developed sinus tract formation discharging pus at 6 months of follow-up. However, the CVJ was stable due to bone fusion. (d) MRI of a CVJ with tuberculosis in a 17-year-old boy showing granulation tissue and abscess anterior to the odontoid process with erosion of the odontoid process causing compression over the odontoid process. In view of the rapidly developing clinical condition, the abscess was drained orally followed by posterior fixation and fusion

Figure 2

(a) X-ray CVJ extension (pre op), (b) X-ray CVJ flexion (pre op) Showing reducible AAD, (c-e) X-rays CVJ of same patient post-op showing stability following C1-C2 wiring

Results

Progressive weakness of all four limbs(88%) was the major presenting complaint followed by sensory symptoms(numbness/paraesthesias) in 70.3%. Five patients were detected to have CVJ abnormalities when they underwent cervical X-rays following a traumatic injury. Head tilt was seen in 26% patients. Cerebellar signs were seen in 25.3% [Table 1]. Only 7 patients had a syndromic association; 5 of them had Down's syndrome and 2 patients had associated Klippel–Feil syndrome. AAD was the most common radiological abnormality (86%) followed by basilar invagination in 48% of patients. In our series, 29% had assimilation of atlas. The characteristic MRI features of CVJ tuberculosis include AAD, fragmentation or destructive lesion of the C 2 body and/or odontoid process and/or lateral mass of the atlas, with retropharyngeal, usually multiloculated, paraspinal abscess with an irregular enhancing rim causing varying degrees of cervicomedullary compression. There is bone fragmentation of the vertebral end plates and obliteration of fat planes around vertebral bodies. Involvement of the lateral mass of the atlas and odontoid process is a common finding and was noted in the present series.

Table 1

Clinical and radiological features and the procedures performed

Outcome

One hundred and thirty eight patients were followed up (mean follow up: 20.3 months, range: 1–84 months). On last available follow up, 86% patients had improved clinically post operatively whereas 10% patients were stable. Worsening was observed in 4% patients. Improvement in power continued even 2 years after surgery.

Bone fusion

All 53 patients with contoured rods and wires and 25 patients with lateral mass screw and rod fixation being followed up were either stable/fused or fused and stable. No implant failure was noticed among these two surgical procedures. 21 patients with C 1 -C 2 fusion had either stable constructs or had fused. 6(22.2%) patients with C 1 -C 2 fusion had broken wires on follow-up requiring repeat posterior fixation.

Complications

There were 5 deaths directly related to the operative procedure; 2 died of meningitis following cerebrospinal fluid leak and the other 3 patients had ventilator dependence postoperatively resulting in chest infection, septicemia and death. Two patients having associated Arnold- Chiari malformations died of sudden-onset respiratory distress after discharge. Morbidity included cerebrospinal fluid leak in 7 patients, wound infection in 19 patients and hemiparesis due to a residual odontoid compression requiring repeat surgery in 1 patient. Six patients had broken wires. Three of these patients underwent occipital lateral mass fixation while repeat wiring was performed in 2 patients. One patient underwent posterior fixation by contoured rod and wire.

Discussion

CVJ abnormalities are a complex group of abnormalities to treat. The challenges and the concerns in the pediatric age group are manifold considering the growing age of the patients. Our series is the largest series of bone CVJ abnormalities in the pediatric age group treated in 5 years with a reasonable follow-up. In this study, we evaluated various procedures performed in a wide spectrum of bone CVJ abnormalities in challenging conditions. The association with syndromes in our series is rare, having 5 patients with Down's syndrome and 2 patients with Klippel–Feil syndrome. Among 189 patients, only 7 patients had an association with syndromes while CVJ tuberculosis requiring surgical treatment is another distinct feature in our series. The number of tuberculous abnormalities we treat is much larger than that represented in the series as most of the patients are treated as outpatients with antitubercular chemotherapy and neck immobilization. Another factor while considering the treatment strategy is the demographic profile of the patients. As the majority of patients come from a low socioeconomic background, low-cost instrumentation made of stainless steel contoured rod was used while in other patients titanium instrumentation was used. However, a concern for stability in flexion, extension and axial rotation has been raised in a contoured rod and wire construct.[16] Good bone fusion, and thereby stability, has been achieved in all patients followed up in our study except for 6 patients who underwent posterior fixation by C1-C2 wiring. Winegar et al.,[17] carried out a meta analysis of occipital cervical fusion techniques by carrying out a literature search of articles using PubMed and CINAHL/Ovid. Among the patients identified within the cited articles, the use of posterior screw/rod instrumentation constructs were associated with a lower rate of postoperative adverse events (33.33%) (P < 0.0001), lower rates of instrumentation failure (7.89%) (P < 0.0001), and improved neurological outcomes (81.58%) (P < 0.0001) when compared with posterior wiring/rod, screw/plate, and onlay in situ bone grafting techniques. The surgical technique associated with the highest fusion rate was posterior wiring and rods (95.9%) (P = 0.0484), which also demonstrated the shortest fusion time (P < 0.0064). Posterior wiring and onlay in situ bone grafting was associated with a high rate of fusion, a relatively higher time to fusion, and a high rate of instrumentation failure when compared with the other techniques studied. The lowest failure rate was following screw/rod techniques. They concluded that regardless of the technique used when attempting occipitocervical stabilization procedures, attention to the detail of the particular technique is crucial. Apostolides et al.,[18] in their series observed a 97% bone fusion rate while Kalra et al.,[19] reported bone fusion rates of 83.8% with a contoured rod and wire construct.[18-20] This shows that the concerns raised by the biomechanical studies regarding the stability offered in different directions and axes are not actually transforming to clinically worse outcomes. By using a contoured rod made of titanium, MRI can be performed as a diagnostic modality in the event of postoperative neurological deterioration. From the above observations, we can conclude that contoured rod and wire is still an option to be considered for posterior fusion. Another purpose of the study was to see long-term efficacy of this procedure, particularly in pediatric patients in relation to their growth. The neurological outcome after the compression at the cervicomedullary junction has been relieved surgically appears good in this group of abnormalities. There were 21 patients with a mean follow-up of 76 months and 4 patients with a maximum follow-up of 7 years. In this group of patients, we have not observed any significant change in the lordotic cervical curvature in the follow-up. The standard methods of sagittal alignment measurements of the cervical spine include Cobb's angle measured by lines passing along the inferior end plates of C 2 and C 7 or the angle drawn along the posterior end plates of C 2 or C 7.[2122] Though none of these measurements were used in our study, the authors observed that there is no significant change in the sagittal alignment of the cervical spine so as to have any clinical implications.

No spinal deformity was seen because of the natural growth being restricted due to instrumentation. Our own observation shows that all forms of constructs ultimately fail though larger well-designed studies are required to prove this. It is adequate bone fusion which provides a stable spinal structure in the long term. These concerns of a negative impact on growth have led people to advocate the removal of the metallic construct after 1–2 years. However, authors believe that even though bone fusion may take place within 3–6 months, ongoing strengthening of the bone keeps going on for a much longer period. Thus, there is no need for an additional support by a construct. Moreover, none of the cases in our series have shown progressive deformity.

Conclusions

Pediatric developmental craniovertebral abnormalities are a different subset of abnormalities and have a satisfactory outcome, and they have to be evaluated individually according to the patients. They seem to be more common in the Indian subcontinent. A contoured rod and wire construct besides being cheap offers good stability resulting in good bone fusion in this subset of patients. The concerns regarding the effect of fusion procedures on bone growth and curvature of the growing spine are debatable.