Characterizing Dysphagia: Eating Assessment Tool-10 Severity Categories After Cervical Spine Surgery

This study proposes and assesses the validity of postoperative dysphagia severity categories based on patients' Eating Assessment Tool-10 (EAT-10) scores and quality-of-life patient-reported outcome measures (PROMs) collected before and after cervical spine surgery.

A prospectively collected multi-institutional quality registry including 2003 patients was retrospectively reviewed. Patients were categorized based on EAT-10 score severities, with thresholds determined by existing literature and significant differences in PROMs: none (0-2), mild (3-4), moderate (5-15), and severe (16-40). Correlations between EAT-10 severity and established PROMs were assessed at baseline, 3-, and 12-month follow-ups. The Dunn test using Bonferroni adjustment was used to identify differences in PROMs between EAT-10 severities. Multivariate linear regression assessed how EAT-10 severity affects PROMs at baseline, 3-, and 12-month follow-ups.

Among 1659 patients with 3-month EAT-10 data, there were 1183 no, 149 mild, 245 moderate, and 82 severe dysphagia patients. Neck Disability Index and EuroQol 5-Dimension scores were significantly different between none vs mild (P < .001; P = .001), mild vs moderate (P < .001; P < .001), and moderate vs severe (P = .035; P = .024) at 3 months. Among 1405 patients with 12-month EAT-10 data, there were 1044 no, 100 mild, 194 moderate, and 67 severe dysphagia patients. Neck Disability Index scores were significantly different between none vs mild (P < .001) and moderate vs severe (P = .004) but not for mild vs moderate (P > .9) at 12 months. EuroQol 5-Dimension scores were significantly different between none vs mild (P = .004) and mild vs moderate (P = .005) but not for moderate vs severe (P = .08). In multivariate regression, EAT-10 severity categories independently predicted worse PROMs at all time points.

We propose none (0-2), mild (3-4), moderate (5-15), and severe (16-40) dysphagia categories based on EAT-10 scores. Evidence from univariate and multivariate analyses strengthens the validity of our proposed EAT-10 severity categories, providing statistical justification for their incorporation into clinical practice.