Autoimmune diabetes is a disorder in which the (3-cells in the pancreatic islets of Langerhans are specifically destroyed resulting in absolute insulin deficiency; typically this is a childhood-onset disease, Type 1 Diabetes (T1D). Type 2 Diabetes (T2D) is a metabolic disorder usually developing in adults resulting from defects in insulin secretion and action. Latent Autoimmune Diabetes in Adults (LADA) is a form of diabetes that shares autoimmune disease pathology with T1D but a clinical presentation similar to T2D; LADA patients develop diabetes as adults (>25 years) and do not immediately require insulin treatment for survival. They are therefore often misdiagnosed with T2D. The aims of this work were to characterise immunological and genetic aspects of LADA using a large cohort collected from various patient repositories the United Kingdom to determine if it is a separate disease entity or an age-related extension of T1D. Both T1D and LADA are characterised by autoantibodies to the islet cell protein glutamic acid decarboxylase 65 (GADA) at diagnosis. The persistence and titre of GADA post-diagnosis in LADA was examined at 0.5, 3 and 6 years. GADA persisted in 93% of patients for 6 years; GADA litres decreased between 0.5 and 3 years post-diagnosis and either stabilised or increased again between 3 and 6 years. GADA titre was not associated with age at diagnosis, glycaemic control, β-cell function or other clinical features. GADA titre at 0.5 years was associated with a greater likelihood of requiring more intensive antihyperglycaemic therapy but did not predict therapy or insulin requirement at 3 and 6 years. Autoantibodies against IA-2 plus GADA compared to GADA alone at diagnosis predicted increased therapy requirement by 3 and 6 years and insulin requirement by 3 years postdiagnosis. Variants of the Human Leukocyte Antigen (HLA) genes DRB1 and DQB1, are associated with susceptibility for T1 D. An analysis of these variants in LADA (n = 378) revealed that the predisposing and protective variants in LADA are similar to those reported in T1D; DR3 (in linkage disequilibrium, LD with DQ2) and DR4 (in LD with DQ8) were the main predisposing variants whereas DR2 (in LD with DQ6) was most the protective against LADA. 85% of LADA patients possessed the DR3 and DR4 specificities, compared with 95% seen in T1D, suggesting a reduced predisposition in LADA compared with T1D. Synergistic effects of the DR3 and DR4 specificities occurred in LADA and the DRB1*0401 allele within the DR4 specificity was predisposing to disease, as seen in T1D. No other predisposing variants were identified in LADA. As reported for T1D, DR11, DR13, DQ5, DQ7 and DQ9 were protective against LADA; DQ6 was positively correlated with age at diagnosis. Association analysis of the insulin gene region in LADA (n = 400) showed that the variable number of tandem repeats (VNTR) locus primarily confers susceptibility to disease. Overall, the short Class I alleles predisposed to disease whereas longer Class III alleles conferred dominant protection, as reported in T1D. Fine-structure analysis showed that the Class I haplotypes 'IC+/ID+' and 'ID-' both conferred susceptibility for LADA - unlike in T1D, where the ID- haplotype has been reported to have protective effects. The Class III 'Protective' and Very Protective' haplotypes, conferred equal protection in LADA, as reported forTID. In conclusion; GADA persist post-diagnosis but are not markers for disease progression of LADA. Patterns of susceptibility at the HLA and insulin gene regions in LADA are similar to that reported for T1 D. LADA is likely to represent an age-related extension of T1D rather than a separate disease entity.