Abstract Background The Major Histocompatibility Complex is the main genetic contributor to susceptibility to type 1 diabetes (T1D); genome-wide scans have consistently mapped increased predisposition to this region. The highest disease risk has been associated with HLA-DR3 and HLA-DR4. In particular, the DR3-positive ancestral haplotype 18.2 was reported as highly diabetogenic. We aimed to corroborate whether this haplotype increases the susceptibility conferred by the DQ2-DR3 alleles in a Mediterranean population. We also searched for additional susceptibility factors to the classic DQ2-DR3 and DQ8-DR4. Results Genetic MHC markers were analysed in a case-control study with 302 T1D patients and 529 ethnically matched controls. DR3-TNFa1b5 carrier rate was significantly higher in DR3-positive heterozygous T1D patients than in DR3-positive heterozygous controls (p = 0.0019; odds ratio OR [95% confidence interval CI] = 2.26 [1.3–3.93]). This data was confirmed analysing the allelic frequency, which includes the information corresponding to the DR3-homozygous individuals (p = 0.001; OR = 2.09) and by using the Arlequin software to check the DR3-positive haplotypes (p = 0.004;OR = 1.93). The present results provide strong evidence of a second susceptibility region in the ancestral haplotype 18.2 in the Spanish population. Moreover, we searched for T1D susceptibility factors in addition to the MHC classical ones, within the DR2-DQ6/DR3-DQ2/DR4-DQ8 negative population. Several genetic markers in both MHC class II (DQA1*0101-DQB1*0501 [p = 0.007;OR = 2.81], DQA1*0201-DQB1*0202 [p = 0.03; OR = 2.35]) and III (TNFa2b1 [p = 0.01 OR = 2.74], BAT-2*2 [p = 0.004; OR = 3.19]) were found. These different alleles associated with T1D were not independent and we observed linkage disequilibrium among them leading us to describe two new risk haplotypes (DQA1*0101-DQB1*0501-TNFa2b1 and DQA1*0201-DQB1*0202- BAT-2*2). Finally, we studied a T1D susceptibility/protection marker located in extended class I, D6S2223; however, no association was observed in our population. Conclusion Our results suggest that other associated MHC haplotypes might present susceptibility factors in loci different from HLA-class II and that the class II molecules are not necessarily the universal etiologic factor in every MHC haplotype.