Co‐occurrence analysis of the correlation of antibodies to glycolipids or ganglioside complexes with clinical features: A retrospective cohort study

Abstract Aims Various combinations of antibodies to glycolipids or ganglioside complexes (GSCs) are found in Guillain–Barré syndrome (GBS) and Miller Fisher syndrome (MFS). Specific combinations of co‐occurring antibodies are thought to be associated with specific clinical phenotypes of GBS/MFS. However, the clinical implications of antibody combinations have not been fully investigated. This study was performed to identify probable antibody combinations and to examine correlations between these combinations and clinical symptoms. Methods The subjects were 315 patients with GBS ( n = 226) or MFS ( n = 89) with anti‐glycolipid‐antibodies. The co‐occurrence of anti‐glycolipid antibodies was analyzed using ‘KH Coder,’ a software program for identifying co‐occurrence networks. Correlations between subgroups of co‐occurring antibodies and clinical features were also analyzed. Results Five unique autoantibody subgroups were identified. In each subgroup, the main autoantibodies were Anti‐ GQ1b , − GT1a antibodies (subgroup 1), Anti‐ GM1 , − GA1 , ‐ GalNAc ‐ GD1a , and ‐ GD1b antibodies (subgroup 2), Anti‐ GM1 / GM2 , − GM1 / GalNAc ‐ GD1a , and other Anti‐ GSC antibodies (subgroup 3), Anti‐ GM1 / GT1b , − GM1 / GalNAc ‐ GD1a , and other Anti‐ GSC antibodies (subgroup 4), and Anti‐ GD1a / GQ1b , ‐ GD1a / GT1a , and other Anti‐ GSC antibodies (subgroup 5). Antecedent infection ( P = .0376), clinical diagnosis ( P < .0001), clinical phenotype ( P < .0001) and cranial nerve involvement ( P < .001) differed significantly across subgroups. Subgroup 1 most frequently had respiratory antecedent infection, contained MFS cases, and had cases that developed cranial nerve deficiencies. Subgroup 3 contained the largest proportion of GBS cases. Conclusion These results suggest that co‐occurrence of anti‐glycolipid antibodies can be classified into five subgroups. Each subgroup exhibits different clinical characteristics.