The following is a summary of “A klotho-based machine learning model for prediction of both kidney and cardiovascular outcomes in chronic kidney disease,” published in the March 2024 issue of Nephrology by Wang et al.

Researchers conducted a prospective study to develop a machine learning (ML) model based on serum Klotho for predicting end-stage kidney disease (ESKD) and cardiovascular disease (CVD) in patients with chronic kidney disease (CKD). 

They used five ML models, trained on data from 400 patients with non-dialysis CKD. These Models predicted ESKD and CVD risks at 3, 5, and 8 years. The dataset was divided into training (70%) and internal validation (30%) sets. Features included 47 clinical variables, including serum Klotho. The best-performing model was used to identify risk factors for each outcome.

The result showed that Lasso regression model showed the highest accuracy (C-index=0.71) for ESKD prediction,  including features like eGFR, urinary microalbumin, serum albumin, phosphate, parathyroid hormone, and serum Klotho (AUC=0.930, 95% CI: 0.897-0.962). For CVD, the Random Survival Forest (RSF) model achieved the highest accuracy (C-index=0.66), incorporating age, hypertension history, calcium, tumor necrosis factor-alpha, and serum Klotho (AUC=0.782, 95% CI: 0.633-0.930).

Investigators concluded that ML models using serum Klotho successfully predicted ESKD and CVD risk in high performance patients with chronic kidney disease, indicating their clinical utility

Source: karger.com/kdd/article/doi/10.1159/000538510/897931/A-Klotho-based-Machine-Learning-Model-for