Abstract Antibodies against several neural antigens have been associated with different chronic immune-mediated neuropathies but their practical clinical relevance remains unclear. To determine the possible diagnostic usefulness of these antibodies we reviewed the clinical correlate of IgM antibodies to the myelin-associated glycoprotein (MAG), sulfatide, the gangliosides GM1, GM2, GD1a and GD1b in 539 consecutive patients examined for neuropathy or related diseases in our Neuropathy Clinics and tested for these antibodies in our laboratory since 1985. 302 patients (56%) had an established diagnosis of definite or possible chronic immune-mediated neuropathy while 237 had a neuropathy of non-immune-mediated origin or of unknown aetiology or a closely related disease. Antibodies to one or more antigen were more frequent ( χ 2 = 63.32; p < 0.00001) in patients with chronic immune-mediated neuropathy (37.7%) than with other neuropathy or related diseases (7.2%) and their presence was associated in 87% of the patients with an immune-mediated neuropathy, incrementing by 31% the probability of having this form. Testing for MAG permitted to identify 24.8% of patients with an immune-mediated neuropathy, GM1 an additional 9.9%, while GM2, GD1b, GD1a and sulfatide altogether an additional 3% of the patients. Concerning clinical correlations, all 75 patients with anti-MAG IgM had neuropathy and IgM monoclonal gammopathy (PN + IgM) with a positive predictive value for this neuropathy of 100%. A similarly high predictive value for neuropathy (91.4%) was observed among 269 patients with IgM monoclonal gammopathy including 103 patients without neuropathy. Anti-sulfatide IgM, though rare, were also significantly and constantly associated with PN +IgM and permitted to identify few patients not bearing anti-MAG IgM. Anti-GM1 IgM were significantly associated with multifocal motor neuropathy (MMN) (29.2%) but where also found in a few patients with other immune or non-immune neuropathies or related diseases with a positive predictive value for MMN of 25.5%. Anti-GM2 IgM were also significantly associated with MMN and increased the sensitivity (36.2%) for MMN obtained with anti-GM1 IgM only, without affecting its specificity and positive predictive value. Anti-GD1a, GD1b, though not significantly more frequent in patients with immune-mediated neuropathy, were associated in 80 to 100% of patients with these neuropathies. In conclusion anti-neural IgM antibodies may help in identifying patients with a chronic immune-mediated neuropathy, even if only anti-MAG and anti-sulfatide IgM appear to be strictly associated with a definite clinical syndrome.