Changes in CSF flow after one-stage posterior vertebral column resection in scoliosis patients with syringomyelia and Chiari malformation type I.

OBJECT: Phase contrast-cine MRI (PC-cine MRI) studies in patients with syringomyelia and Chiari malformation Type I (CM-I) have demonstrated abnormal CSF flow across the foramen magnum, which can revert to normal after craniocervical decompression with syrinx shrinkage. In order to investigate the mechanisms leading to postoperative syringomyelia shrinkage, the authors studied the hydrodynamic changes of CSF flow in the craniocervical junction and spinal canal in patients with scoliosis associated with syringomyelia after one-stage deformity correction by posterior vertebral column resection.
METHODS: Preoperative and postoperative CSF flow dynamics at the levels of the foramen magnum, C-7, T-7 (or apex), and L-1 were assessed by electrocardiogram-synchronized cardiac-gated PC-cine MRI in 8 adolescent patients suffering from severe scoliosis with syringomyelia and CM-I (scoliosis group) and undergoing posterior vertebral column resection. An additional 8 patients with syringomyelia and CM-I without spinal deformity (syrinx group) and 8 healthy volunteers (control group) were also enrolled. Mean values were obtained for the following parameters: the duration of a CSF cycle, the duration of caudad CSF flow (CSF downflow [DF]) and cephalad CSF flow (CSF upflow [UF]), the ratio of DF duration to CSF cycle duration (DF%), and the ratio of UF duration to CSF cycle duration (UF%). The ratio of the stationary phase (SP) duration to CSF cycle duration was calculated (SP%). The maximum downflow velocities (VD max) and maximum upflow velocities (VU max) were measured. SPSS (version 14.0) was used for all statistical analysis.
RESULTS: Patients in the scoliosis group underwent one-stage posterior vertebral column resection for deformity correction without suboccipital decompression. The mean preoperative coronal Cobb angle was 102.4° (range 76°-138°). The mean postoperative Cobb angle was 41.7° (range 12°-75°), with an average correction rate of 59.3%. During the follow-up, 1 patient with hypermyotonia experienced a significant decrease of muscle tension and 1 patient with reduced anal sphincter tone manifested recovery. A total of 5 patients demonstrated a significant decrease (> 30%) in syrinx size. With respect to changes in CSF flow dynamics, the syrinx group was characterized by slower and shorter downflow than the control group, and the difference was more significant at the foramen magnum and C-7 levels. In patients with scoliosis, CSF downflow at the foramen magnum level was significantly restricted, and a prolonged stationary phase indicated increased obstruction of CSF flow. After posterior vertebral column resection, the peak velocity of CSF flow at the foramen magnum increased, and the downflow phase duration was markedly prolonged. The parameters showed a return to almost normal CSF dynamics at the craniocervical region, and this improvement was maintained for 6-12 months of follow-up.
CONCLUSIONS: There were distinct abnormalities of CSF flow at the craniocervical junction in patients with syringomyelia. Abnormal dynamics of downflow could be aggravated by associated severe spinal deformity and improved by correction via posterior vertebral column resection.