ABSTRACT
INTRODUCTION@#The risk for diabetes progression varies greatly in individuals with type 2 diabetes mellitus (T2DM). We aimed to study the clinical determinants of diabetes progression in multiethnic Asians with T2DM.@*MATERIALS AND METHODS@#A total of 2057 outpatients with T2DM from a secondary-level Singapore hospital were recruited for the study. Diabetes progression was defined as transition from non-insulin use to requiring sustained insulin treatment or glycated haemoglobin (HbA1c) ≥8.5% when treated with 2 or more oral hypoglycaemic medications. Multivariable logistic regression (LR) was used to study the clinical and biochemical variables that were independently associated with diabetes progression. Forward LR was then used to select variables for a parsimonious model.@*RESULTS@#A total of 940 participants with no insulin use or indication for insulin treatment were analysed. In 3.2 ± 0.4 (mean ± SD) years' follow-up, 163 (17%) participants experienced diabetes progression. Multivariable LR revealed that age at T2DM diagnosis (odds ratio [95% confidence interval], 0.96 [0.94-0.98]), Malay ethnicity (1.94 [1.19-3.19]), baseline HbA1c (2.22 [1.80-2.72]), body mass index (0.96 [0.92-1.00]) and number of oral glucose-lowering medications (1.87 [1.39-2.51]) were independently associated with diabetes progression. Area under receiver operating characteristic curve of the parsimonious model selected by forward LR (age at T2DM diagnosis, Malay ethnicity, HbA1c and number of glucose-lowering medication) was 0.76 (95% CI, 0.72-0.80).@*CONCLUSION@#Young age at T2DM diagnosis, high baseline HbA1c and Malay ethnicity are independent determinants of diabetes progression in Asians with T2DM. Further mechanistic studies are needed to elucidate the pathophysiology underpinning progressive loss of glycaemic control in patients with T2DM.