#2525 THE 2021 CKD EPI GLOMERULAR FILTRATION RATE ESTIMATING EQUATION AND RISK OF HOSPITALIZATION FOR KIDNEY DISEASE

Abstract

Abstract Background and Aims The revised Chronic Kidney Disease Epidemiology Collaboration (CKD EPI) equation for estimated glomerular filtration rate (eGFR) reduced CKD prevalence among Asians, but few studies have evaluated its impact on healthcare utilization. We aimed to evaluate the impact of the 2009-eGFRcr(ASR) [1] and 2021-eGFRcr(AS) [2] equations on the risk of hospitalization for kidney disease among multi-ethnic Asians. Method Retrospective cohort study of adults who consulted at Singapore's largest healthcare cluster of public healthcare institutions and had a baseline serum creatinine performed in 2014. eGFR was calculated using both eGFRcr(ASR) [1] and 2021-eGFRcr(AS) [2] equations and categorized as G1 (≥90 ml/min/1.73 m2), G2 (60-89 mL/min/1.73 m2), G3A (45-59 mL/min/1.73 m2), G3B (30-44 mL/min/1.73 m2), G4 (15-29 mL/min/1.73 m2), G5 (<15 mL/min/1.73 m2). Electronic medical records were used to identify baseline demographic and clinical information, and the outcome of hospitalization for kidney disease including acute kidney injury, chronic kidney disease, dialysis or transplant in the next three years until December 2017. Multi-variable logistic regression models were used to compare the impact of the eGFR equations on risk of hospitalization for kidney disease. Model discrimination was assessed using the area (AUC) under a receiver operating characteristic (ROC) curve. Results Among 33007 Asian adults included in the study, the median age was 69.0 years (interquartile range 63.0, 76.0). The majority were Chinese (79.5%; Indian 7.2%, Malay 10.2% and other race 3.2%), female (53.3%), had diabetes (56.4%) and received statin therapy (71.6%) at baseline. Hospitalization for kidney disease occurred in 1915 (7.1% of 26874 individuals) after excluding those with missing covariates (n = 5133). Hospitalization incidence increased with worsening kidney function (Table 1) regardless of the estimation equation used. After adjusting for age, sex, ethnicity, comorbidities (cardiovascular disease, diabetes, hypertension, cancer, systolic and diastolic blood pressure, body mass index) and medications (statin, ACE inhibitor, angiotensin receptor blocker, loop and thiazide diuretic), eGFR categories were independently associated with hospitalization for both eGFR equations (Table 2). Compared to G1, the risks of hospitalization for G2, G3A, G3B and G4 were incremental at approximately 2, 6, 10 and 20 times, respectively, while the risk of hospitalization for G5 was 100 times that of G1. The AUCs were 0.859 (95% CI: 0.849-0.869) for the model with eGFRcr (ASR) and 0.861 (95% CI: 0.852-0.871) for 2021-eGFRcr (AS) but the difference was not significant (p = 0.11). Conclusion Lower eGFRs ascertained by both the 2009 eGFRcr(ASR) and 2021-eGFRcr(AS) equations were independently associated with greater risks of hospitalization for kidney disease. The revised 2021-eGFRcr(AS) equation did not significantly alter model discrimination.