Photo Credit: renatahamuda
The following is a summary of “Use of consensus clustering to identify distinct subtypes of chronic kidney disease and associated mortality risk,” published in the December 2024 issue of Nephrology by Qin et al.
Chronic kidney disease (CKD) has varied causes and outcomes. A data-driven clustering approach can identify distinct subgroups with specific mortality risks.
Researchers conducted a retrospective study to explore CKD subtypes.
They used unsupervised consensus clustering to classify CKD subtypes based on 45 baseline characteristics from 6,526 National Health and Nutrition Examination Survey (NHANES) participants (1999–2018). They assessed the associations between CKD subgroups and mortality outcomes, including all-cause, cardiovascular, cancer, and other cause-related deaths.
The results showed that 6,526 individuals with CKD were classified into 4 clusters. Cluster 1 (n=508) had favorable cardiac and kidney markers, lower cancer prevalence, higher obesity, and lower medication use. Cluster 4 (n=2,029) had the worst cardiac and kidney function markers. Clusters 2 (n=1,439) and 3 (n=2,550) fell in between. Hazard ratios (HR) for all-cause, cardiovascular, and other cause mortality increased from cluster 1 to cluster 4. Sensitivity analysis revealed heterogeneity within subgroups with similar baseline kidney function and mortality risks.
They identified distinct CKD subgroups with varying mortality risks through consensus clustering. This suggested that further investigation of these subgroups could enhance precision medicine in CKD management.
Source: nature.com/articles/s41598-024-81208-1#Abs1
