To the Editor Isobe and colleagues¹ studied the association of HLA genetic burden with mainly magnetic resonance imaging (MRI)–based disease phenotypes in a large cohort of 586 patients with multiple sclerosis (MS). Analyses were conducted separately for men and women using linear regression; age and disease duration were only included as covariates if this decreased the Akaike information criterion of the models. Several associations were reported, but P values were rather moderate. Age at onset, known to be associated with HLA-DRB1*1501 alleles,² was nominally associated in women but not in men. In addition, cerebral white matter volume was nominally associated in women and spinal gray matter fraction in men. Only the association of subcortical gray matter in women survived correction for multiple testing.