Risk factors for clinical and radiological worsening following Chiari malformation type I surgery in the pediatric population

Abstract

OBJECTIVE Foramen magnum (FM) decompression with or without duraplasty is considered a common treatment strategy for Chiari malformation type I (CM-I). The authors’ objective was to determine a predictive model of risk factors for clinical and radiological worsening after CM-I surgery. METHODS A retrospective review of electronic health records was conducted at an academic tertiary care hospital from 2001 to 2019. A multivariable Cox proportional hazards regression model was used to determine the risk factors. The Kaplan-Meier estimate was plotted to delineate outcomes based on FM size. FM was measured as the preoperative distance between the basion and opisthion and dichotomized into < 34 mm and ≥ 34 mm. Syrinx was measured preoperatively and postoperatively in the craniocaudal and anteroposterior directions using a T2-weighted MRI sequence. RESULTS A total of 454 patients (231 females [50.9%]) with a median (range) age of 8.0 (0–18) years were included in the study. The median duration of follow-up was 21.0 months (range 3.0–144.0 years). The model suggested that patients with symptoms consisting of occipital/tussive headache (HR 4.05, 95% CI 1.34–12.17, p = 0.01), cranial nerve symptoms (HR 3.46, 95% CI 1.16–10.2, p = 0.02), and brainstem/spinal cord symptoms (HR 3.25, 95% CI 1.01–11.49, p = 0.05) had higher risk, whereas those who underwent arachnoid dissection/adhesion lysis had 75% lower likelihood (HR 0.25, 95% CI 0.10–0.64, p = 0.004) of clinical worsening postoperatively. Similarly, patients with evidence of brainstem/spinal cord symptoms (HR 7.9, 95% CI 2.84–9.50, p = 0.03), scoliosis (HR 1.04, 95% CI 1.01–2.80, p = 0.04), and preoperative syrinx (HR 16.1, 95% CI 1.95–132.7, p = 0.03) had significantly higher likelihood of postoperative worsening of syrinx. Patients with symptoms consisting of occipital/tussive headache (HR 5.44, 95% CI 1.86–15.9, p = 0.002), cranial nerve symptoms (HR 2.80, 95% CI 1.02–7.68, p = 0.04), and nonspecific symptoms (HR 6.70, 95% CI 1.99–22.6, p = 0.002) had significantly higher likelihood, whereas patients with FM ≥ 34 mm and those who underwent arachnoid dissection/adhesion lysis had 73% (HR 0.27, 95% CI 0.08–0.89, p = 0.03) and 70% (HR 0.30, 95% CI 0.12–0.73, p = 0.008) lower likelihood of reoperation, respectively. The Kaplan-Meier curve showed that patients with FM size ≥ 34 mm had significantly better clinical (p = 0.02) and syrinx (p = 0.03) improvement postoperatively when the tonsils were resected. CONCLUSIONS These results showed that preoperative and intraoperative factors may help to provide better clinical decision-making for CM-I surgery. Patients with FM size ≥ 34 mm may have better outcomes when the tonsils are resected.