Researchers Develop Predictive Model for Kidney Function Decline in Adults With T2D


Researchers developed a reliable prediction model for kidney function decline in adults with type 2 diabetes and early to moderately progressed chronic kidney disease, which was externally validated and made publicly available via a web-based application.

Researchers developed a robust and well-calibrated prediction model for kidney function decline that accurately predicts future estimated glomerular filtration rate (eGFR) values in adults with type 2 diabetes (T2D) and early to moderately progressed chronic kidney disease (CKD).

According to the prognostic study published in JAMA Network Open, the model was found to be reliable and accurate in predicting the decline in kidney function up to 5 years after baseline examination. A public web-based application with the results and prediction model was developed, which the authors hope will improve the prediction of individual eGFR trajectories and disease progression of CKD.

Between February 2010 and December 2019, researchers conducted a prognostic analysis using baseline and follow-up data from 3 prospective multinational cohort studies: PROVALID, GCKD, and DIACORE. The study included a total of 4637 adult participants between the ages of 18 and 75 years who had T2D and mildly to moderately impaired kidney function, defined as a baseline eGFR of at least 30 mL/min/1.73 m2. The data were analyzed between June 2021 and January 2023.

The researchers chose the following variables as predictors for eGFR rates, which are readily available from routine clinical care visits: age, sex, body mass index, smoking status, hemoglobin A1c level, hemoglobin level, serum cholesterol level, mean arterial pressure, urinary albumin-creatinine ratio, and intake of glucose-lowering, blood pressure–lowering, or lipid-lowering medication.

The study included 4637 White adults—57.8% men—with T2D and CKD, with a mean (SD) age of 63.5 (9.1) years and a follow-up of 5.0 (0.6) years. The model development cohort included 3323 combined participants from the PROVALID and GCKD studies, and the external validation cohort included 1314 participants from the DIACORE study.

The researchers found that updating the prediction model’s random coefficient estimates with baseline eGFR values resulted in better predictive performance, especially seen in the calibration curve (calibration slope at 5 years: 1.09; 95% CI, 1.04-1.15). The prediction model had good discrimination in the validation cohort, with the lowest C statistic at 5 years after baseline (0.79; 95% CI, 0.77-0.80). The model also had a predictive accuracy of R2 ranging from 0.70 (95% CI, 0.63-0.76) at year 1 to 0.58 (95% CI, 0.53-0.63) at year 5.

“We addressed several common methodological limitations of prediction modeling studies, such as the lack of external validation, to ensure generalizability of application to unseen cohorts and inclusion of predictors that are not routinely available in primary clinical care visits (eg, genetic information and serum biomarkers),” the authors said.

They gave an example of one study, which incorporated genetic covariates along with conventional ones to explore the link between known genetic variations and eGFR trajectories. However, the utilization of genome-wide genotyping is not feasible in routine clinical settings and thus, not appropriate for universal application.

“Furthermore, the incremental usefulness of molecular biomarkers in addition to traditional clinical predictors for improved prediction is still under investigation,” they added. “Therefore, we restricted the model to data obtained in primary clinical care visits among individuals with CKD and type 2 diabetes.”

According to the authors, some strengths of this study include the use of baseline eGFR values as part of the outcome vector rather than a covariate, the use of a prespecified set of predictors, the inclusion of interactions with follow-up time as main outcomes, and the rigorous internal-external validation of the model. They also noted the web-based application as a strength, as it provides a user-friendly prediction tool that can be used to identify high-risk patients for recruitment in clinical studies.

Some limitations of this study include the fact that all 3 large-scale cohort studies were conducted in European countries, the creatinine assays were not standardized across cohorts, the data points grew sparser at later time points, and the medications used were not up to date with current treatments for CKD.

While not explicitly listed as a limitation by the authors, it is important to note the study consisted entirely of White participants.

“Despite its complexity, the prediction model was robust, well calibrated, and suitable for implementation in a web-based application, revealing the potential of a publicly available online tool that can be used by patients, caregivers, and primary health care professionals to predict individual eGFR trajectories and disease progression up to 5 years after baseline,” the authors concluded.


Gregorich M, Kammer M, Heinzel A, et al. Development and validation of a prediction model for future estimated glomerular filtration rate in people with type 2 diabetes and chronic kidney disease. JAMA Netw Open. 2023;6(4):e231870. doi:10.1001/jamanetworkopen.2023.1870

Related Videos
Matthew Crowley, MD, MHS, associate professor of medicine, Duke University School of Medicine.
Susan Spratt, MD, senior medical director, Duke Population Health Management Office, associate professor of medicine, division of Endocrinology, Metabolism, and Nutrition,
Stephen Nicholls, MD, Monash University and Victorian Heart Hospital
Amal Agarwal, DO, MBA
Dr Robert Groves
Dr Robert Groves
Jeremy Wigginton, MD
Deepak L. Bhatt, MD, MPH, Icahn School of Medicine at Mount Sinai
Jeremy Wigginton, MD
Related Content
© 2023 MJH Life Sciences
All rights reserved.