Metabolic and immunogenetic prediction of long-term insulin remission in African patients with atypical diabetes

Aims We aimed to characterize a cohort of 'atypical' diabetic patients of sub-Saharan African origin and to analyse possible determinants of long-term remission. Methods Over 6 years, we studied the clinical and therapeutic profile of 42 consecutive patients undiagnosed or untreated prior to inclusion presenting with cardinal features of diabetes mellitus. We measured insulin secretion and sensitivity at inclusion. Immunogenetic (anti-GAD, anti-ICA and HLA class II) markers of Type 1 diabetes were compared with a 90-non-diabetic unrelated adult African population. Results Twenty-one ketonuric patients (age 42 +/- 9 (sd) years; body mass index (BMI) 26 +/- 3 kg/m(2)) were initially insulin-treated (IT), and 21 non-ketonuric patients (age 38 +/- 8 years; BMI 26 +/- 5 kg/m(2)) had oral and/or diet therapy (NIT). Insulin could be discontinued in 47.6% (10/21) IT with adequate glycaemic control (HbA(1c) 6.7 +/- 1.3%), while insulin was secondarily started in 38.1% (8/21) NIT in expectation of better control. The initial basal (odds ratio (OR) 9.1, 95% confidence interval (CI) 1.3-64.4) and stimulated C-peptide (OR 8.17, 95% CI 1.5-44.1) were independently associated with remission. Insulin resistance was present in all the groups, more marked in the insulin-treated NIT. Anti-GAD antibodies and ICA were rare, but 38.1% IT vs. 1.1% controls had Type 1 diabetes HLA susceptibility haplotypes (P < 0.001) without significant difference between the subgroups. Conclusion Prolonged discontinuation of insulin is frequent in African diabetic patients initially presenting with signs of insulinopenia. In our patients, long-term insulin therapy was not associated with immunogenetic markers of Type 1 diabetes. The initial measure of insulin secretion seemed a good predictor of long-term remission.