A clinical stratification tool for chronic kidney disease progression rate based on classification tree analysis

Area of Science:

• Nephrology
• Public Health
• Clinical Epidemiology

Background:

• Registry studies identify chronic kidney disease (CKD) risk factors but lack individual prognosis.
• Sequential kidney function measurements are often missing in current risk assessments.
• Predicting differential annual decline in glomerular filtration rate (GFR) is crucial for patient management.

Purpose of the Study:

• To identify demographic and clinical characteristics that discriminate patients with varying annual GFR decline.
• To develop a predictive model for CKD progression using routine clinical data.

Main Methods:

• Observational retrospective study using data from the Prevention of Progressive Renal Insufficiency Project registry.
• Inclusion criteria: patients with at least four serum creatinine measurements between July 2004 and June 2010.
• Classification tree analysis (CTA) applied to demographic and laboratory data for subgroup identification.

Main Results:

• CTA identified seven patient subgroups with distinct annual GFR decline rates.
• Proteinuric patients with baseline eGFR >33 mL/min/1.73 m² showed the fastest progression (-3.655 mL/min/1.73 m²/year).
• Non-proteinuric patients <67 years, especially those with diabetes, exhibited faster progression; older females showed stable eGFR.

Conclusions:

• Routine clinical variables can effectively stratify patients by CKD progression rate.
• Stratification aids in optimizing follow-up schedules and therapeutic interventions.
• Personalized risk stratification supports proactive planning for dialysis and transplantation.