Validation of a risk prediction model for early chronic kidney disease in patients with type 2 diabetes: Data from the German/Austrian Diabetes Prospective Follow-up registry

Aim: To validate a recently proposed risk prediction model for chronic kidney disease (CKD) in type 2 diabetes (T2D). Materials and Methods: Subjects from the German/Austrian Diabetes Prospective Follow-up (DPV) registry with T2D, normoalbuminuria, an estimated glomerular filtration rate of 60 ml/min/1.73m(2) or higher and aged 39-75 years were included. Prognostic factors included age, body mass index (BMI), smoking status and HbA1c. Subjects were categorized into low, moderate, high and very high-risk groups. Outcome was CKD occurrence. Results: Subjects (n = 10 922) had a mean age of 61 years, diabetes duration of 6 years, BMI of 31.7 kg/m(2), HbA1c of 6.9% (52 mmol/mol); 9.1% had diabetic retinopathy and 16.3% were smokers. After the follow-up (similar to 59 months), 37.4% subjects developed CKD. The area under the curve (AUC; unadjusted base model) was 0.58 (95% CI 0.57-0.59). After adjustment for diabetes and follow-up duration, the AUC was 0.69 (95% CI 0.68-0.70), indicating improved discrimination. After follow-up, 15.0%, 20.1%, 27.7% and 40.2% patients in the low, moderate, high and very high-risk groups, respectively, had developed CKD. Increasing risk score correlated with increasing cumulative risk of incident CKD over a median of 4.5 years of follow-up (P < .0001). Conclusions: The predictive model achieved moderate discrimination but good calibration in a German/Austrian T2D population, suggesting that the model may be relevant for determining CKD risk.

Automated summary

The aim of this study was to validate a recently proposed risk prediction model for chronic kidney disease in type 2 diabetes. The results showed that the model achieved moderate discrimination and good calibration in a German/Austrian population, suggesting its relevance in determining CKD risk.