Models Effective for Predicting Nephropathy in Type 2 Diabetes

Female kidney failure, computer artwork.
Researchers conducted a systematic review and external validation study to assess the quality and accuracy of prognostic models for nephropathy and to validate these models in external cohorts of patients with type 2 diabetes.

Several models for predicting nephropathy in patients with type 2 diabetes showed good performance, according to a study in BMJ.

Researchers conducted a systematic review and external validation study to assess the quality and accuracy of prognostic models for nephropathy and to validate these models in external cohorts of patients with type 2 diabetes.1

The investigators searched PubMed and Embase from inception until June 16, 2020, for relevant studies. External validation of the models was done with use of the Hoorn Diabetes Care System cohort (mean [SD] age, 62.6 [12.1] years; 52.7% male) of 11,450 patients with type 2 diabetes. For each validated model, 3 prediction horizons of 2, 5, and 10 years were assessed.

A total of 41 studies were included and accounted for 64 prediction models that predicted a nephropathy outcome, of which 46 were developed in patients with diabetes and 18 in the general population.

Of the identified models, 21 could be externally validated—15 in patients with type 2 diabetes and 6 in the general population. For predicting the risk of albuminuria, diabetic kidney disease, and chronic kidney disease, discrimination and calibration showed considerable variation across horizons and models. Multiple models performed well, with C statistics >0.80 for the 3 outcomes.

Calibration had the same variation among outcomes and studies compared with discrimination, but to a lesser extent between horizons. The difference in performance regarding horizons within a model was smaller than the performance between models for most cases. Models for patients with diabetes had similar general performance compared with those developed in the general population.

The discriminatory ability of the models developed in patients with diabetes varied considerably, with C statistics ranging from 0.50 to 0.96 in the Hoorn DCS cohort. Performance on the 2-year horizon was generally best within a model, followed by the 5-year and 10-year horizons.

In a secondary external validation using the Genetics of Diabetes Audit and Research in Tayside Scotland cohort (GoDARTS; mean age, 65.2 [11.1] years; 56% male), models developed for diabetic kidney disease especially showed good discrimination compared with those developed for chronic kidney disease.

Of the 5 validated models for albuminuria, the model from Basu et al2 had the best combination of discrimination and calibration. For diabetic kidney disease, models from Afghahi et al3 and Nelson et al4 performed best. For chronic kidney disease, the model from Saranburut et al5 performed well and had better calibration than other models.

The study had several limitations. The authors excluded articles that were not in English or Dutch. Additionally, the researchers could not validate models that predicted end stage renal disease, owing to limited events in the study population.  The Hoorn DCS cohort included patients with generally well-controlled diabetes, which could have potentially affected the study results as well.

“This study identified several suitable models that will contribute to preventing or postponing renal decline and ultimately end stage renal disease in people with diabetes,” stated the researchers.


  1. Slieker RC, van der Heijden AAWA, Siddiqui MK, et al. Performance of prediction models for nephropathy in people with type 2 diabetes: systematic review and external validation study. BMJ. Published online September 28, 2021. doi:10.1136/bmj.n2134
  2. Basu S, Sussman JB, Berkowitz SA, Hayward RA, Yudkin JS. Development and validation of risk equations for complications of type 2 diabetes (RECODe) using individual participant data from randomised trials. Lancet Diabetes Endocrinol. 2017;5(10):788-798. doi:10.1016/S2213-8587(17)30221-8
  3. Afghahi H, Cederholm J, Eliasson B, et al. Risk factors for the development of albuminuria and renal impairment in type 2 diabetes—the Swedish National Diabetes Register (NDR). Nephrol Dial Transplant. 2011;26(4):1236-1243. doi:10.1093/ndt/gfq535
  4. Nelson RG, Grams ME, Ballew SH, et al. Development of risk prediction equations for incident chronic kidney disease. JAMA. 2019;322(21):2104-2114. doi:10.1001/jama.2019.17379
  5. Saranburut K, Vathesatogkit P, Thongmung N, et al. Risk scores to predict decreased glomerular filtration rate at 10 years in an Asian general population. BMC Nephrol. 2017;18(1):240. doi:10.1186/ s12882-017-0653-z