There is controversy regarding the relationship of polyneuropathy syndromes to the presence of serum antibody binding to myelin-associated glycoprotein (MAG). Using standard ELISA methodology, we identified 74 sera that appeared to have high titers of IgM binding to MAG and found that only 34% of these sera stained MAG using Western blot methodology. Follow-up studies showed that two factors greatly influence concordance between ELISA and Western blot testing for anti-MAG antibodies. Sera with high titers of binding to both MAG and histone H3 identified by ELISA rarely stain MAG on Western blot. In addition, sera analyzed by ELISA often bind to impurities in the semipure MAG that is frequently used in ELISA assays. Further purifications to separate MAG from other contaminants improved concordance between ELISA and Western blot results to 85% to 90%in a retrospective analysis, as well as in a prospective study of 49 additional sera. Patients with a polyneuropathy and serum IgM binding to MAG preparations by ELISA but not byWestern blot methodology had several different clinical syndromes, including gait disorders and asymmetric motor neuropathies. Patients with IgM binding to MAG by both assay methods usually had a distal, sensory-motor, symmetric polyneuropathy with some features of demyelination on electrodiagnostic testing.