Anatomical Subtypes and Swallowing Dysfunction as Predictors of Disease Severity in Pediatric Laryngomalacia: A Cohort Study

Background Laryngomalacia (LM) is the most common congenital laryngeal anomaly. While typically self-limiting, it increasingly presents as a complex aerodigestive disorder involving swallowing dysfunction and gastroesophageal reflux. Our aim is to evaluate the prevalence of swallowing abnormalities in LM, correlate anatomical subtypes and swallowing function with disease severity, and identify prognostic factors for surgical intervention. Methods This observational cohort study included 138 pediatric patients (aged ≤ 24 months) with endoscopically confirmed LM at a tertiary academic center between March 2022 and March 2024. Patients underwent flexible fiberoptic laryngoscopy and/or bronchoscopy and were classified by anatomical type (Hollinger and Konior Types A, B, C). Swallowing function was assessed clinically and via Fiberoptic Endoscopic Evaluation of Swallowing (FEES) using the Penetration-Aspiration Scale (PAS). Disease severity was graded using International Pediatric ORL Group (IPOG) criteria. Results The cohort (mean age 6.6 ± 6.09 months; 55.1% male) exhibited anatomical types: A (30.4%), B (39.1%), and C (30.4%). Swallowing abnormalities were present in 65.2% of patients. Type B morphology was strongly associated with severe disease (p < 0.001) and severe stridor (p < 0.001). Clinical swallowing abnormalities were significantly associated with higher disease severity (p < 0.001). Supraglottoplasty was required in 13.0% (n = 18) of the cohort; notably, 100% of surgical candidates had Type B anatomy. Synchronous airway lesions, primarily laryngeal clefts (15.2%) and tracheomalacia (10.9%), were frequent comorbidities. Conclusion Type B laryngomalacia is a distinct prognostic indicator for severe disease and the need for surgical intervention. Swallowing dysfunction is highly prevalent and correlates with disease severity, necessitating systematic instrumental assessment in management algorithms.