Cervicomedullary CSF space obliteration as a risk factor of reoperation in Chiari malformation type I

In patients with Chiari malformation type 1 (CM1), it is unclear whether severe cervicomedullary CSF space crowding on preoperative MRI is associated with postoperative outcomes. We tested whether a binary classification scheme of cervicomedullary CSF space crowding (obliteration vs. non-obliteration) was associated with Chiari-directed reoperation in patients with CM1 after index surgery.

We compared outcomes with and without cervicomedullary space obliteration (CMSO) in a single-center, single-surgeon retrospective cohort of patients undergoing surgery for CM1. CMSO and non-CMSO groups were compared via Chi-square testing and multivariate logistic regression. Interrater reliability (IRR) of identifying CMSO was assessed by polling 68 attending pediatric neurosurgeons using 15 MRIs with varying degrees of cervicomedullary CSF crowding.

CMSO existed in 154 of the 272 included patients (56.6%). At baseline, the group with CMSO had statistically higher rates of ataxia (16.9% vs. 7.6%, p = 0.02), motor weakness (10.4% vs. 2.5%, p = 0.01), swallowing difficulties (30.3% vs. 11.1%, p < 0.001), sleep apnea (20.1% vs. 9.3%, p = 0.01), and cranial neuropathies (9.7% vs. 1.7%, p = 0.01) than those without CMSO. They also had a higher risk of requiring reoperation (aOR 2.53; 95% CI 1.20-5.73; p = 0.02) and postoperative hydrocephalus (aOR 4.23; 95% CI 1.18-27.17; p = 0.03). The IRR survey demonstrated acceptable agreement of identifying CMSO (κ = 0.789).

A dichotomous scheme for rating cervicomedullary CSF space obliteration in CM1 as "obliterated" or "non-obliterated" was associated with Chiari-directed reoperation in this cohort and demonstrated substantial IRR. These findings support further prospective evaluation of CMSO as a simple radiographic descriptor of disease severity.