Association of HLA class II markers with autoantibody-negative ketosis-prone atypical diabetes compared to type 2 diabetes in a population of sub-Saharan African patients

Aim: We investigated the association of HLA DRB1 and DQB1 alleles, haplotypes and genotypes with unprovoked antibody-negative ketosis-prone atypical diabetes (A- KPD) in comparison to type 2 diabetes (T2D). Methods: A- KPD and T2D sub-Saharan African patients aged 19-63 years were consecutively recruited. Patients positive for cytoplasmic islet cell, insulin, glutamic acid decarboxylase or islet antigen-2 autoantibodies were excluded. Odds ratios were obtained via logistic regression after considering alleles with a minimum frequency of 5% in the study population. Bonferroni correction was used in the case of multiple comparisons. Results: Among the 130 participants, 35 (27%) were women and 57 (44%) were A- KPD. DRB1 and DQB1 allele frequencies were similar for both A- KPD and T2D patients; they did not confer any substantial risk even after considering type 1 diabetes susceptibility and resistance alleles. We found no association between A- KPD and the derived DRB1*07-DQB1*02:02 (OR: 0.55 [95%CI: 0.17-1.85], P=0.336); DRB1*11-DQB1*03:01 (OR: 2.42 [95%CI: 0.79-7.42], P=0.123); DRB1*15-DQB1*06:02 (OR: 0.87 [95%CI: 0.39-1.95], P=0.731) and DRB1*03:01-DQB1*02:01 (OR: 1.48 [95%CI: 0.55-3.96], P=0.437) haplotypes. Overall, we did not find any evidence of susceptibility to ketosis associated with DRB1 and DQB1 genotypes (all P>0.05) in A- KPD compared to T2D. Similar results were obtained after adjusting the analysis for age and sex. Conclusion: Factors other than DRB1 and DQB1 genotype could explain the propensity to ketosis in A- KPD. These results need to be confirmed in a larger population with the perspective of improving the classification and understanding of the pathophysiology of A- KPD.