Longitudinal scoliosis behavior in Chiari malformation with and without syringomyelia.

Objective: The objective of this study was to understand the natural history of scoliosis in patients with Chiari malformation type I (CM-I) with and without syringomyelia.

Methods: A retrospective review of data was conducted. Patients with CM-I were identified from a cohort of 14,118 individuals age 18 years or younger who had undergone MRI over an 11-year period at the University of Michigan. Patients eligible for study inclusion had a coronal curve ≥ 10° on radiography, associated CM-I with or without syringomyelia, and at least 1 year of clinical follow-up prior to any surgery. Curve magnitude at initial diagnosis, prior to posterior fossa decompression (PFD; if applicable), and at the last follow-up (prior to any surgical correction of scoliosis) was recorded, and clinical and radiographic characteristics were noted. The change in curve magnitude by 10° was defined as curve progression (increase by 10°) or regression (decrease by 10°).

Results: Forty-three patients met the study inclusion criteria and were analyzed. About one-third (35%) of the patients presented with symptoms attributed to their CM-I. The mean degree of scoliosis at presentation was 32.6° ± 17.7°. Twenty-one patients (49%) had an associated syrinx. The mean tonsil position below the level of the foramen magnum was 9.8 ± 5.8 mm. Patients with a syrinx were more likely to have a curve > 20° (86% vs 41%, p = 0.002). Curve magnitude remained stable (≤ ±10°) in 77% of patients (33/43), progressed in 16% (7/43), and regressed in 7% (3/43). Mean age was higher (14.8 ± 0.59 years) among patients with regressed curves (p = 0.026). All regressed curves initially measured ≤ 20° (mean 14° ± 5.3°), and none of the patients with regressed curves had a syrinx. The change in curve magnitude was statistically similar in patients with (7.32° ± 17.7°) and without (5.32° ± 15.8°) a syrinx (p = 0.67). After a mean follow-up of 3.13 ± 2.04 years prior to surgery, 27 patients (63%) ultimately underwent posterior fossa or scoliosis correction surgery. For those who eventually underwent PFD only, the rate of change in curve magnitude prior to surgery was 0.054° ± 0.79°. The rate of change in curve magnitude was statistically similar before (0.054° ± 0.79°) and after (0.042° ± 0.33°) surgery (p = 0.45) for patients who underwent PFD surgery only.

Conclusions: The natural history of scoliosis in the presence of CM-I is variable, though most curves remained stable. All curves that regressed were ≤ 20° at initial diagnosis, and most patients in such cases were older at scoliosis diagnosis. Patients who underwent no surgery or PFD only had similar profiles for the change in curve magnitude, which remained relatively stable overall, as compared to patients who underwent PFD and subsequent fusion, who demonstrated curve progression. Among the patients with a syrinx, no curves regressed, most remained stable, and some progressed. Understanding this variability is a first step toward building a prediction model for outcomes for these patients.