The literature of the past ten years shows that the introduction of highly purified heterologous and, lastly, homologous insulins has notably lowered the production of IgG and IgE specific insulin antibodies, but has not succeeded in completely eliminating clinical manifestations of the immune or hyper-immune response to insulin therapy. In particular, insulin allergy with or without lipodystrophy is still seen. Among the factors of insulin immunogenicity, there is a possible genetic control of the immune response in type I diabetes: determining HLA halloantigens (A, B, C, D) might identify specific immune response genes (Ir genes). Initial researches, performed until now almost exclusively upon diabetics treated with conventional heterologous insulin, seem to indicate a positive relationship between haplotype HLA - B15 - DR4 and an elevated immune response, whereas haplotypes HLA - B8 - DR3 and HLA - B18 - DR3 might protect against the formation of anti-insulin antibodies. Antigens D/DR3 and D/DR4 are known to be primitively associated to susceptibility for type I diabetes, whereas antigens B8, B15, B18 are secondarily associated to the rise in frequency of DR3 and DR4 for the "linkage disequilibrium" existing between alleles of B and D loci. The results of HLA typing are presented in 2 groups of insulin-dependent diabetics (ID) followed from an immunological viewpoint during therapy with monocomponent heterologous insulin for over 5 years. The first group is composed of 50 patients with low IgG anti-insulin antibody titers (less than 1 mU/ml, Christiansen: low responders); the second group is made up of 23 patients with high IgG anti-insulin antibody titers (greater than 2.5 mU/ml, Christiansen: high responders) and includes 5 subjects with insulin allergy (associated or not with insulin lipoatrophy) and high levels of insulin specific IgE antibodies. A study of the frequencies of various HLA-B antigens in both groups of patients, in regard to a control group of piemontese population, in relation to the intensity of association (relative risk) and to the statistical importance of frequencies, shows only a possible protective effect of the HLA-B18 phenotype (linkage disequilibrium with HLA - DR3) towards the production of anti-insulin antibodies and hyperimmune clinical manifestations, such as allergy. Reliable conclusions are not possible between low and high responders for the other phenotypes (HLA - B7, B8, B15) commonly implicated. HLA-B12 was noted in 3 of 5 patients with allergy, in 2 cases associated with B8.(ABSTRACT TRUNCATED AT 400 WORDS)