Aim: To derive predictors of good glycaemic control in patients presenting with diabetic ketoacidosis (DKA) followed prospectively in a specialized clinic.
Methods: One hundred and sixty-one adult patients were admitted during a 31-month period and followed for at least 12 months. After 1 year, the patients were classified into three groups: good control (GC) (HbA1c < or = 7%), intermediate control (IC) (HbA1c 7-9%) and poor control (PC) (HbA1c > 9%). Characteristics of patients in the three groups were compared both at baseline and during follow-up.
Results: At 12 months, 36% of the patients were classified as GC, 27% as IC and 37% as PC. GC patients had higher fasting serum C-peptide levels 0.7 +/- 0.54 compared to 0.38 +/- 0.29 and 0.16 +/- 0.21 nmol/l, respectively, for the IC and PC patients (p < 0.0001). A higher proportion GC patient had a C-peptide level greater than 0.33 nmol/l than that for IC and PC patients (86, 61 and 19%, respectively; p < 0.0001). Exogenous insulin was safely discontinued in 50, 30 and 3% of patients, respectively, in the GC, IC and PC groups (p < 0.0001). Compliance with life-style interventions was higher in the GC than that in IC and PC patients (87, 41 and 5%, respectively; p < 0.0001). In the logistic regression analysis, predictors of good glycaemic control were having baseline fasting serum C-peptide value > or =0.33 mmol/l, OR: 3.01 (95% CI 1.07-8.55, p = 0.03) and compliance with life-style interventions OR 12.66 (95% CI 3.73-51.57, p = 0.0001).
Conclusion: Among adult patients with DKA, significant predictors of good glycaemic control are preserved beta-cell function and compliance with life-style modifications.