Data are odds ratio (OR) and 95% CIs for miglustat use vs not. Longitudinal analyses of follow-up data were performed using generalized linear models for repeated measures with independent working correlations for ordinal multinomial models. Probabilities for severity of disease (disease progression) were modeled for the American Speech-Language-Hearing Association National Outcome Measurement System swallowing domain (ASHA-NOMS) and Rosenbek aspiration-penetration scale (PAS) outcomes. Models accounted for the association of miglustat use outcomes, seizure history, duration of neurologic symptoms, and duration of follow-up in the overall cohort, the early childhood onset subgroup (ECO; age <6 years), and late childhood onset subgroup (LCO; age ≥6 years and <15 years).